id stringlengths 36 40 | text stringlengths 50 3.89M | source stringclasses 1 value | added stringdate 2025-04-07 03:56:57 2025-04-07 04:04:00 | created stringlengths 0 16 | metadata dict |
|---|---|---|---|---|---|
ffa2a234-da32-4c59-b516-55230ec92248.181 | *2.3. Data Analysis*
Transcripts were analysed by two experienced qualitative researchers using the framework approach described by Ritchie and Spencer [30]. Framework analysis uses a systematic approach of
data management to provide coherence and structure within the analysis process [30,31]. Passages of text representing repeated themes were identified and assigned headings according to the context and coded to several relevant categories to reduce the likelihood of missing key points. The data were then synthesised in sub-headings identified from the thematic analysis [30]. This approach is useful in enhancing the rigour, transparency and validity of the analytic process [32]. The analysis was both inductive, with categories emerging purely from the data and deductive, with categories derived from prior knowledge [33].
| doab | 2025-04-07T04:13:02.974495 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 181
} |
ffa2a234-da32-4c59-b516-55230ec92248.182 | *2.4. Ethical Considerations*
All study participants were provided with an information sheet outlining details of the research and provided written consent to participate in the research. Some focus group members knew each other but not all. The issue of confidentiality was discussed with the focus group participants. Due to the sensitivity of the topic and the possibility of participants becoming distressed as a result of the focus group discussions and interviews, referral procedures were put in place prior to data collection to address the needs of distressed participants. After the data collection procedures were completed, all participants were provided with a list of professional counselling agencies where they could access support. Participants were offered a follow-up, private counselling session to discuss any issues that were raised by the interviews, with costs covered by the African Communities Council of South Australia (ACCSA). Each participant was reimbursed thirty dollars. Ethics approval was provided by the Flinders University Social and Behavioural Research Ethics Committee (SBREC project number 5480).
| doab | 2025-04-07T04:13:02.974577 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 182
} |
ffa2a234-da32-4c59-b516-55230ec92248.183 | **3. Results**
As African migrants, many of the youths participating in this study experienced insecurity and displacement because of civil unrest in their countries of origin (Table 1). Social identity for these youths is beyond the physical construct that often shapes identity discourses. Identity for these participants was not just about their race (a social construct used here as a point of analysis), gender, ethnicity or country of origin. It is innate in nature and involved matters that are important for everyday life. There were multiple identities, and it relates to how they felt both in- and outside their physical body. The concept of 'opportunity' appeared to be important, regarded as a gift holding special meaning and privilege. As such, it came with the expectation that opportunity is used for reciprocating benefits to family in Africa and the Australian society more generally. However, the extent to which individuals acted on these opportunities in meaningful ways varied, affecting a person's sense of purpose and belonging in Australia. Although there appeared a general feeling of being torn between two places among respondents, the loss of social networks following migration and cultural differences between African and Australian societies shaped the experience of belonging. The demographic details of the participants and the identified themes are presented with reference to salient quotations from respondents (Table 2).
**Table 1.** Characteristics of focus group youth participants.
**Table 2.** Characteristics of face-to-face interview participants.
| doab | 2025-04-07T04:13:02.974662 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 183
} |
ffa2a234-da32-4c59-b516-55230ec92248.184 | *3.1. Freedom and Opportunities*
Respondents tended to recognise that the rights afforded to them upon settlement within Australia provided them with the opportunity to satisfy their personal needs and ambitions in a way that was not possible given their situation in Africa. This was expressed in terms of gratefulness, particularly considering the limited opportunities afforded to those still living in Africa. For example, there were views that showed a feeling of freedom from previous experiences of education.
In Australia, even though everywhere people have their own tough times and all that but I'm free here. No more running away from war. Free education. I have a lot of opportunities to do things that I want to do than when I was back home. (Respondent 23)
Then, I also have this opportunity where I can go to any university without being denied for my right to study . . . it's a very important gift that God gave to us, so to me I think we are lucky to be here. (Respondent 17)
Through these views, not only did the participants perceive freedom as understood by the mainstream Australian community, but they also understood it in their own unique experiences. They linked freedom to their past experiences and integrated it to shape their self-construct in Australia. Freedom to study in Australia was particularly perceived as important for girls given experiences of girls' education in Africa. The following quote shows how the freedom to education was perceived through a gendered lens.
It's when you have freedom; you have the opportunity to study whatever you want, because that's the hardest thing in Africa for a young girl, just to finish even year 11. To reach that stage it's really hard where here you can study as many courses as you want, so it's a privilege to be here, that's what I always tell my friends. (Participant 12)
However, participants acknowledged that the availability of opportunities tended to be an insufficient condition for using those opportunities in an instrumentally meaningful and purposeful way. Rather, support external to the individual is required to use and benefit from these opportunities.
I think in Australia there are a lot of opportunities in education and all this kind of thing, but the downside of it is even if there's education, if there's no help to cope, you know, to cope with it then it becomes hard to use that opportunity. So, yes, there are opportunities, but we need help, something or someone to keep pushing us to get in there. (Respondent 14)
Some participants revealed pessimism because they felt their lives have not improved since coming to Australia despite the available opportunities in Australia. This view has profoundly shaped their self-outlook as expressed in the below quote.
When we were in Africa, life was really, really terrible and then when we come to Australia—me personally, I thought I was going to have a better life. I've been here for nine years now. I wake up every morning; the only thing I see shining is the sun, but my life's not shining. (participant 14)
Additionally, many participants acknowledged that even if there are opportunities to study, they also experienced limited opportunities for employment, which negatively impacted their experience of living in Australia.
Then you try to get a job and then no-one offering you job, so you find yourself a bit depressed, I guess. Then if you find yourself a bit depressed, what do you do? Drink up, hang out with your friends, go out, make yourself feel good. (Respondent 21)
What I know so far, there are a lot of young people actually get frustrated from school and finding no jobs or getting an appropriate job, so things like that, and people who are actually going under the trauma of homelessness and into drugs, into other things which are negative actually to the young people (respondent 15)
| doab | 2025-04-07T04:13:02.974784 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 184
} |
ffa2a234-da32-4c59-b516-55230ec92248.185 | *3.2. Freedom and Family Relationships*
Another common consensus among participants was around differences in culture from the older generation in relation to the rights and freedom of expression and how those disagreements came to shape the identity of some African youths in Australia. In some African cultures, youths are expected to follow the views of their parents and elders. Many youths in the study, however, perceived themselves as embodying Australian culture and were free to make an individual choice to do whatever they wish. Sometimes the 'freedom' in Australia put youths in the direct course of cultural conflict with parents and caused tensions within families. For example, the below quote captures this sentiment:
We are in Australia everyone is equal, and we know our rights and wrongs so that we argue with parents, like 'this is wrong. This is what you're supposed to say to me because you are wrong' and then they don't accept that. That's how they take it to the community, because Africans consider that as being rude, you are not respecting your parents by talking back at them, but this is not what Australian culture says. Everyone is equal (Participant 12)
I've got pressure from my culture, like family at home, and then pressure from outside because I want to be—socially interact with others and they, my family, don't want to accept that . . . I'm with my aunt here, and we have a lot of disagreement and stuff. Like you're a girl, you don't go out and meet friends except for close families, like a family that they know. Like I have friends from different places in Australia and stuff, but they don't accept that they find it hard because they have different culture, like 'no'. (Participant 10)
When participants were asked how the cultural conflicts within the family setting affected them, there was consensus around disengagements from their communities. Participants revealed that disengaged youths did not 'fit' in their communities or in the Australian culture. These missing pieces of identity from the country of origin and Australia made some participants define their own self-construct in ways that were comfortable and appropriate for themselves, for example, by going out to find a social group to identify with. Participants described the process of finding a sense of belonging in such a position as a "confused" situation. The following quote demonstrates this sentiment.
For me, I have to go to meet friend just to clear thing out . . . A lot of people, I think they isolate themselves from going out in the community or in a tribe with different people or getting involved in something . . . It always goes back to family, you know? Follow your culture with your family, listen to them, what they want you to do or just ignore them and do what you want to do to yourself. So, you're between confused what-exactly how you're going to help yourself to get out of that issue (Participant 10)
However, there remained tension between expectations for personal responsibility for action and the role of others in supporting this action. This reflected the recognition that personal freedom, representing the ability to define oneself and one's own course of action, might result in very different consequences for an individual according to their personal resources and ability to seek out opportunities and support.
The good thing is that it's you make your choices. At the end of the day, it's up to you. Nobody makes the choices for you, but it's up to you and if you're willing to listen to anybody, take in their advice and sort of work out your life and stuff it's up to you . . . You're free, but all of that comes with consequences too. (Respondent 16)
| doab | 2025-04-07T04:13:02.975066 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 185
} |
ffa2a234-da32-4c59-b516-55230ec92248.186 | *3.3. Not Belonging and Being Torn between Two Places*
Some respondents also noted conditions where opportunities could not be used in meaningful ways, making it difficult for individuals to find the means to enact their desired role within Australian society. The inability to fulfil personal ambitions resulted not only in a feeling of being trapped within Australian society, but also the experience of becoming marginalised from others because of barriers to engaging in meaningful education and employment.
Most youths are happy from outside, but inside they are not happy because they don't have jobs. (Respondent 1)
Because you wake up every day and then, you know, you do the same thing, and you see the same people and the same things happening over and over again . . . You know you try to get—a job, trying to get to know people. Some of them tend to, you know, disengage from you and all that stuff. (Respondent 21)
The above quotations reveal the role employment opportunities and social relationships play in forming a meaningful self-concept of belonging. The inability to engage in employment or develop bonds with others led to the marginalisation and the experience of 'not belonging' by respondents. The experience of 'not belonging' can have a profound impact on an individual, as one of the participants demonstrated when narrating the experience of another youth in the community.
So, I know one of the persons who has committed suicide. Two weeks before he died there was kind of—he said 'I want to go back to Africa', you know? He told family 'I want to go back. I don't want to live here. This place is no good'. (Respondent 7)
Although it is not clear what the phrase 'this place is no good' means, other respondents made it clear that settling in Australia came with a feeling of being torn between two places, their country of origin and Australia. This is not to say that belonging is simply associated with connection to place, but also recognises that family and social ties may still exist within Africa and that there is a constant need to accommodate cultural beliefs from both African and Australian societies.
Yeah if you see—I have some people here you see—if you look at that you seem happy but inside you are not happy because I know—I have some cousins and my sister is here, but our mum is in Africa. (Respondent 1)
Let's say if you want to marry now, you want a woman, you've got to have at least forty to sixty grands to marry the woman. You have to give them to the girl's family and if you don't have those you're not going to get any girl. So that's our—you know, it's just our culture sometimes, it's just different. Yeah, difficult. (Respondent 9)
Previously, it was mentioned that the inability to make use of educational and employment opportunities resulted in the experience of being trapped. Likewise, the experience of being trapped tended to emphasise the subjective feeling of not belonging, related to the perceived difference between African and Australian societies and a loss of family and social support networks.
Am I happy? I can say I am but like in the sense that I've got everything around me but in terms of that I'm not really—like I'm not really happy. Like back at home I was happy. Like we didn't have food and all this stuff, but I still was happy. It's just like maybe you can put it this way, here like you've been just surrounded by a fence around you, and you just want to escape, you know what I mean? (Respondent 21)
| doab | 2025-04-07T04:13:02.975350 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 186
} |
ffa2a234-da32-4c59-b516-55230ec92248.187 | *3.4. Using Available Resources*
The study findings demonstrate that African youths in Australia draw on available resources when negotiating identities in Australia. For example, finding other youths from another community for social networks and a sense of belonging. They also draw on available services for youths in the community to maximise their access to resources that support their needs. Lastly, engaging in sport is another key activity through which African refugee youths negotiate their identities in Australia. Some participants who did not have their 'ethnic' communities in Australia revealed that they negotiated their identity through their friends in order to find a community for social activities and belonging. The following quote demonstrates this sentiment.
I don't have my community here, like tribe; you know how Sudan, they have different tribes. So for me, I don't have that community here so I get involved in other people's communities, like other tribes. If they have program and stuff, I go there, just with my friends. (Participant 10)
The respondents in the current study also had a belief that there is a moral impetus not only to make use of available opportunities but also to actively seek support from others in obtaining and making effective use of those opportunities.
If you are youth here in Australia, you are a very young person, you get so many supports and so many help, and you should be seeking to get that help. There are a lot of sports activities. If you're good at it Australian people will take you, you know. For example, if you want to make music and you're really good at it, you can get a scholarship. If you're good at basketball as African youth, you can get a scholarship. (Respondent 20)
Additionally, drawing on individual strengths of resilience and determination were important for the participants in this study. They acknowledged there were challenges of settling in Australia. In order to achieve their goals and contribute to the Australian community, however, participants emphasised the need to remain positive and focussed. The following two quotes demonstrate these views.
I think to be a young African youth in Australia—well, for me it's a good gift, and it's a great gift because no matter what happened to you if you're still following your dream then I think it will be a good gift. (Respondent 17)
If you're still alive, you are rich because one day you might become rich but if you're dead, even though you are rich nobody will call you as rich because you are a dead person. That keeps other kids to push their lives, because they know one day, they will be rich. (Participant 11)
| doab | 2025-04-07T04:13:02.975851 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 187
} |
ffa2a234-da32-4c59-b516-55230ec92248.188 | **4. Discussion**
This qualitative study attempted to further an understanding about how African youth negotiated their identity within urban settings of South Australia. Even though it is well acknowledged that the psychological construction of identity occurs throughout adolescence [1], youths, especially those with cultural backgrounds such as in the current study, require additional negotiations to incorporate their cultural, racial, and ethnic identity [3,5] within their new society. This is in recognition that, to migrants, migration to a new setting brings with it a range of novel challenges and stressors [34,35] compared to the general community in the host nation population. For youths settling into a new host culture, this also affects the social construct of their identity.
The findings of the current study revealed that African youths valued opportunities for self-determination, in particular the availability of formal education, freedom of speech and individual action, and freedom from oppression and war. The ability to realise personal ambitions was of significant importance for these youth. This is of particular relevance given that personal accomplishments and achievements have been found to predict personal and ethnic self-esteem among Australian immigrant youth, with personal self-esteem being the single major predictor of immigrant psychological health [19]. The presence of, and effective use of, opportunities enabled youth to realise ambitions and in doing so, promoted a sense of fulfilment and belonging. However, the ability to obtain and make use of opportunities was made difficult in the context of social exclusion and economic hardships. This study revealed that increased individual freedom would lead to different consequences depending on the motivations and resources of individuals. Because of the effect of economic disadvantage and the loss of social networks on arrival to Australia, social exclusion and economic hardship may limit individuals' capacity to make use of opportunities, which in turn could further limit the attempts towards social inclusion and economic security. Within Australia, similar obstacles to social integration, including separation of family members, lack of access to housing and education support, and employment challenges have been observed amongst refugees and immigrants from the Horn of Africa [4]. Although discrimination and racism have been demonstrated to negatively affect an immigrant's self-concept, self-esteem, and access to economic resources [3,16], this was not discussed by respondents within this study. However, this may have been a consequence of the action of more covert forms of discrimination due to the perceived cultural and racial difference between African migrant youths and their host society [14].
It is well acknowledged that upon resettlement, migrants, especially those with a refugee background, tend to pass through many phases of readjustment leading to outcomes ranging from integration through to marginalisation, with the direction of this outcome dependent on the presence or absence of resources in a person's surrounding environment. Evidence also suggests that social inclusion, freedom from discrimination, and access to economic resources are important in promoting mental health during this resettlement process [16]. With its focus on African migrant and refugee youths, this study recognises the importance of these resources in resettlement, but conceptualises them as resources that can be drawn upon to realise their goals and ambitions, meet basic needs, and address the demands imposed by resettlement. This resonates with Dermot Ryan and colleagues' 'resource-based model of migrant adaptation', which extends previous conceptualisations of migrant adaptation to recognise the role played by social environments in creating, or otherwise alleviating, stressful conditions impacting on resettlement. This is particularly relevant for those operating
from a public health understanding as it takes the emphasis away from the role of individuals in managing demands and coping with stress and places it on the social conditions that create stressful demands and examines how social policies shape such conditions. Here, resources are the focus of our attention, including personal resources such as problem-solving and social skills, material resources such as housing and employment, social resources such as emotional and tangible support from others, and cultural resources that enable navigation of daily activities within a particular cultural environment. Importantly, it should be recognised that an individual's resources are likely to be reduced, for various reasons, during each stage of the migration process, including resource losses endured prior to migration and during the migration process itself [24]. It is recognised that upon resettlement, migrants, especially those who are refugees, tend to arrive with histories of trauma and abuse, loss of status and social networks, language barriers, unemployment, financial problems, psychiatric disorders, and addiction to alcohol and drugs [4,17,18]. Thus, it becomes important to build the personal, material, social, and cultural resources of African youths upon arrival, attempting to replenish those lost during the process of migration. By building resources related to education and socioeconomic status, individuals will be more likely to successfully negotiate their own identity within a foreign country [36]. Our study has highlighted education, employment, and supportive social networks as resources important in the minds of African youth, enabling them to realise their needs and goals for life in Australia. It also suggests that a resource-based approach for enhancing adaptation of migrant youths provides a useful framework for public health action enabling youth to negotiate their own identity, alongside psychological, anthropological, and sociological understandings of identity formation [1]. A few limitations of this study are worth noting. One limitation is that the study was unable to elicit details regarding the nature of goals, needs, and demands experienced by these youth and the resources required to address them. It was also unable to identify the processes by which discrimination and racism impacted on the use of and access to resources, which is a key limitation given its potential to erode personal resources [3,16] and increase aversive demands [24]. Research shows that societal responses to refugees can act to preserve or undermine their human social resources and influence the experience of resettlement [37]. Furthermore, although we identified education as an important resource for African youths, we were not able to examine how this might relate to gaining meaningful employment and the impact of this on identity formation as a dynamic process [36]. This is an important consideration, as African migrants elsewhere have been observed to encounter significant barriers to finding employment related to English language comprehension and their ability to navigate job networks [38,39]. It was also less clear as to how histories of interrupted schooling and cultural and social language differences may have impacted on educational outcomes [40]. Further examination of these issues and the extent to which social policies can influence individual resources may prove useful in framing public health action to facilitate positive settlement outcomes for migrant and refugee youths.
| doab | 2025-04-07T04:13:02.976046 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 188
} |
ffa2a234-da32-4c59-b516-55230ec92248.189 | **5. Conclusions and Research Contribution to the Body of Knowldge**
Attempting to understand the way in which the socio-cultural and economic conditions of resettlement influence the negotiation of identity among African refugee youth within urban South Australia, this study has highlighted the narratives between the meaningful use of life opportunities and the experience of fulfilment and belonging within Australia, with social and economic resources as necessary for individuals' ability to make effective use of these opportunities. These findings call for consideration of the socio-cultural and political context surrounding resettlement and how this influences the negotiation of identity among migrant and refugee youths. They also resonated with a resource-based model of migrant adaptation, which shows promise in directing public health action, given its focus on social conditions that may promote or minimise psychological stress during the process of resettlement. Additionally, the recognition of African youths as a vulnerable group when it comes to negotiating identity is critical in understanding suicide in order to develop appropriate health
policies and design effective interventions to prevent suicide among these and similar populations in Australia and elsewhere in settings where African migrants have been resettled.
Although this study is limited in its ability to examine the precise nature of resources that may act as enablers or barriers to the realisation of personal ambitions and needs, it has provided direction for future research aimed at examining these issues and how they might be addressed through public health policy. The contributions of this paper to the body of knowledge are as follows:
**Author Contributions:** W.M. contributed to the methodology, analysis, writing—original draft preparation, writing the final version, reviewing and editing of the manuscript. L.M. contributed to the conceptualization, analysis, writing—original draft preparation, writing the final version, reviewing and editing of the manuscript. All authors have read and agreed to the published version of the manuscript.
**Funding:** This research received no external funding.
**Acknowledgments:** The authors acknowledge the participants of this study.
**Conflicts of Interest:** The authors declare no conflict of interest.
| doab | 2025-04-07T04:13:02.976584 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 189
} |
ffa2a234-da32-4c59-b516-55230ec92248.191 | *Article* **Migrant Women's Access to Sexual and Reproductive Health Services in Malaysia: A Qualitative Study**
**Tharani Loganathan 1,\* , Zhie X. Chan <sup>2</sup> , Allard W. de Smalen 2,3,4 and Nicola S. Pocock 2,5**
Received: 25 June 2020; Accepted: 24 July 2020; Published: 26 July 2020
**Abstract:** Providing sexual and reproductive health (SRH) services to migrant workers is key to fulfilling sustainable developmental goals. This study aims to explore key informants' views on the provision of SRH services for migrant women in Malaysia, exploring the provision of SRH education, contraception, abortion, antenatal and delivery, as well as the management of gender-based violence. In-depth interviews of 44 stakeholders were conducted from July 2018 to July 2019. Data were thematically analysed. Migrant workers that fall pregnant are unable to work legally and are subject to deportation. Despite this, we found that insufficient SRH information and contraceptive access are provided, as these are seen to encourage promiscuity. Pregnancy, rather than sexually transmitted infection prevention, is a core concern among migrant women, the latter of which is not adequately addressed by private providers. Abortions are often seen as the only option for pregnant migrants. Unsafe abortions occur which are linked to financial constraints and cultural disapproval, despite surgical abortions being legal in Malaysia. Pregnant migrants often delay care-seeking, and this may explain poor obstetric outcomes. Although health facilities for gender-based violence are available, non-citizen women face additional barriers in terms of discrimination and scrutiny by authorities. Migrant women face extremely limited options for SRH services in Malaysia and these should be expanded.
**Keywords:** migrant health; access to health; sexual and reproductive health; contraception
| doab | 2025-04-07T04:13:02.976763 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 191
} |
ffa2a234-da32-4c59-b516-55230ec92248.192 | **1. Introduction**
Migration of women is an important component of international migration, with women comprising nearly half (48%) of the 258 million international migrants worldwide in 2017 [1]. In Malaysia, 19% of the 2.0 million documented migrant workers in 2019 were women [2]. The country also houses an estimated 2 to 3 million undocumented migrants [3], which increases the number of female migrants significantly. Although women represent a substantial proportion of less skilled migrant workers in Malaysia, appropriate migration and gender-sensitive policies are still lacking. As a result, female migrants are more vulnerable and prone to human rights violations [4,5].
Sexual and reproductive health and rights (SRHRs) are fundamental human rights, which lies with the right of individuals and couples to freely decide the number, timing, and spacing of children and have adequate information to make those decisions, and the right to attain the highest standard of sexual and reproductive health [SRH] [6]. SRHR was conceptualised during the 1994 International Conference on Population and Development (ICPD) in Cairo, where reproductive health was defined as a "state of complete physical, mental and social well-being, not merely the absence of disease and infirmity, in all matters relating to the reproductive system, and its functions and processes" [6] and was subsequently built on evolving international agreements [7].
SRH addresses a wide range of health issues, including contraception, unintended pregnancies, unsafe abortions, gender-based violence (GBV), pregnancy and childbirth complications, human immunodeficiency virus (HIV) and other sexually transmitted infections (STIs), and infertility and reproductive cancers, and are seen as essential elements to achieve social, economic and environmental development goals [7].
Maternal and child health services are the cornerstone of Malaysia's public health system and are available nationwide as part of the integrated primary care services provided at public health clinics, maternal and child health clinics, and community clinics under the Ministry of Health (MOH), Malaysia. The scope of maternal and child health services includes maternal and perinatal health services (pre-pregnancy, antenatal, intrapartum and postnatal care), child health services (childhood immunisation and health, development and growth assessment), and woman's health services (family planning services, and cervical and breast cancer screening) [8]. SRH services are also provided by the National Population and Family Development Board, under the purview of the Ministry of Women, Family and Community Development, non-governmental organisations like the Federal Reproductive Health Associations Malaysia (FRHAM), and private practitioners [9].
Improving access to SRH is central to development, as reflected under target 3.7 of the Sustainable Development Goals (SDGs) which calls for "universal access to sexual and reproductive health-care services, including for family planning, information, and education" by 2030 [10]. Although providing SRH services to marginalised communities including migrant workers are key to fulfilling the SDGs [11], the 2017 Voluntary National Review of SDGs by the Malaysian government did not identify migrant workers as a vulnerable group to improve delivery of healthcare services [12].
Women emigrating for employment face intersecting vulnerabilities of gender, social class, and ethnicity [13] and often encounter physical, psychological, and sexual violence [14,15]. Despite the ratification of the Convention on the Elimination of Discrimination against Women (CEDAW) and commitments to the ICPD Programme of Action [16], Malaysia has not fully recognised the migrant workers' SRHRs [17,18]. Female migrant workers in Malaysia still face SRHR-related difficulties, mainly through the prohibition of pregnancy during employment [19,20].
In its concluding observations on the combined third to fifth periodic reports of Malaysia, the Committee on the Elimination of Discrimination against Women were concerned about the barriers faced by non-citizen women, including female migrant workers, when accessing healthcare [21]. Financial constraints are a major healthcare access barrier, as healthcare charges for non-citizens are considerably higher when compared to citizens for services at public facilities [17,18,22]. In addition, healthcare personnel are required to report undocumented migrants seeking medical care to the Immigration department, deterring women from seeking needed care due to fear of arrest and detention [23,24].
Therefore, this study aims to explore key informants' views on the provision of SRH services for low-skilled, documented, and undocumented migrant women in Malaysia, including SRH education, contraception, abortion, antenatal and delivery, and the management of GBV.
### **2. Materials and Methods**
We used qualitative methods in an exploratory, iterative design to explore policy and the provision of SRH services for migrant workers in Malaysia.
| doab | 2025-04-07T04:13:02.976915 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 192
} |
ffa2a234-da32-4c59-b516-55230ec92248.193 | *2.1. Definition of Terms*
This study focuses on low-skill, low-wage migrant workers who cross international borders for employment. Documented or "regular" migrants are authorised to enter, stay, and partake in employment in a country, and also have legal documents, such as valid passports and work permits. Undocumented or "irregular" migrants are those who enter the country, reside, or partake in employment without authorisation, including those who may have entered the country legally, but have violated either the terms of their visa or over-stayed beyond the authorised period [9–11].
Refugees, asylum-seekers, foreign wives, and expatriates are not included in this study.
#### *2.2. Sampling and Recruitment*
We conducted 37 in-depth interviews with 44 individuals from July 2018 to July 2019 in Malaysia (Table 1). Most interviews were conducted on a one-on-one basis, while several were conducted in small groups of 2 to 3 participants from the same organisation.
**Table 1.** Characteristics of the study participants (*n* = 44).
<sup>1</sup> Only 1 of the 4 migrant workers interviewed identified himself as a worker only. Others were also members of civil society organisations (2) or trade unions (1). <sup>2</sup> Government or government-linked organisation.
The health and welfare of migrant workers in Malaysia are contentious, with issues concerning migrant workers' SRHR and their immigration status being particularly sensitive. As such, we did not specifically target female migrant workers for interviews. We interviewed multiple stakeholders, including members of civil society organisations (CSOs), international organisations, academia, industry, medical doctors, and migrant representatives to obtain a broader understanding of SRHR for this vulnerable population.
Migrant representatives interviewed represented the interests of migrant workers and were able to speak broadly on migrant workers' experiences in Malaysia. We interviewed representatives of workers from major migrant-sending countries to Malaysia (Indonesia, the Philippines, Nepal, and Bangladesh). We also interviewed medical doctors from the public sector, private sector, and CSOs, who provided SRH services to migrant populations. In addition, we interviewed representatives of CSOs that primarily worked on migrant women's rights and welfare.
Participants were recruited purposively from a previous migrant health stakeholder workshop in Kuala Lumpur [25], and subsequently from snowball sampling of interviewees and further stakeholder recruitment through LinkedIn, until researchers agreed that new interviews would not yield additional information, as thematic saturation was reached.
#### *2.3. Data Collection and Analysis*
In-depth interviews averaged from 1 to 1.5 h and were conducted at physical locations chosen by participants or via telephone. Interviews were conducted either in English or Bahasa Malaysia (Malay language) depending on the participants' preference, by the multi-lingual research team. The majority of interviews were conducted in English, with only 5 out of 37 interviews conducted in Malay.
Semi-structured interview guides were developed to seek participants' perspectives on SRH health services for migrant women in Malaysia, and these questions were tailored towards the participants' professional and organisational backgrounds. The interview guides were constructed based on literature review and discussion among the research team. Concurrent data analysis informed data collection and further refinement of question guides. Interviews with stakeholders from different backgrounds allowed triangulation of findings. Audio recordings were transcribed verbatim.
We conducted thematic analysis as described by Braun and Clarke, where themes or patterns of meaning within data were identified and reported using six phases: becoming familiar with the data, generating initial codes, searching for themes, reviewing themes, defining themes, and producing the report [26].
Data analysis was conducted in an immersive, exploratory, and inductive manner. The first and second authors reviewed and analysed transcripts independently, with regular discussions between researchers to refine codes and identify new themes. Transcripts were coded into emerging themes using NViVo 12 Plus, (QSR International, Melbourne, Australia) and quotations were extracted into Microsoft® Excel® for Office 365, (Microsoft, Redmond, WA, USA). Interviews in Bahasa Malaysia were analysed in the same language, while extracted quotations were translated to English for publication.
| doab | 2025-04-07T04:13:02.977260 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 193
} |
ffa2a234-da32-4c59-b516-55230ec92248.194 | *2.4. Reflexivity*
Interviews were conducted by a medical doctor and academic researchers, who could be perceived as trusted authority figures. To counter possible power imbalances, especially among migrant workers and their representatives, participants chose interview times and their locations.
| doab | 2025-04-07T04:13:02.977579 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 194
} |
ffa2a234-da32-4c59-b516-55230ec92248.195 | *2.5. Ethics*
Participant information sheets were distributed, which detailed the benefits and potential risks of the study, as well as patient rights and study procedures, including audio recording, confidentiality, and data storage. Verbal and written informed consent were sought from all participants before interviews. All participants agreed to be audio recorded and quoted anonymously in publications. Audio recordings and electronic transcripts were stored in secure data servers, while printed transcripts and notes were stored in a locked cupboard. Study participation was voluntary, and we explained that participants could refuse to answer questions or terminate interviews at any point.
This study was approved by the Medical Ethics Committee, University Malaya Medical Center, Malaysia and the Medical Research and Ethics Committee, Ministry of Health, Malaysia (Approval numbers: UM.TNC2/UMREC-238 and NMRR-18-1309-42043).
| doab | 2025-04-07T04:13:02.977621 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 195
} |
ffa2a234-da32-4c59-b516-55230ec92248.196 | **3. Results**
Study results are presented on the health policy context, followed by findings on SRH services, such as SRH education and contraception, abortion, antenatal care and delivery, and GBV. Table 2 summarises the major study results.
**Table 2.** Summary of major study findings.
| doab | 2025-04-07T04:13:02.977704 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 196
} |
ffa2a234-da32-4c59-b516-55230ec92248.197 | **Health Policy and Employment Contract Clauses**
| doab | 2025-04-07T04:13:02.977742 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 197
} |
ffa2a234-da32-4c59-b516-55230ec92248.198 | **Sexual and Reproductive Health Education and Contraception**
| doab | 2025-04-07T04:13:02.977766 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 198
} |
ffa2a234-da32-4c59-b516-55230ec92248.202 | *3.1. Health Policy and Employment Contract Clauses*
| doab | 2025-04-07T04:13:02.977789 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 202
} |
ffa2a234-da32-4c59-b516-55230ec92248.203 | 3.1.1. Mandatory Health Screening and the Prohibition of Pregnancy
To obtain and renew work permits in Malaysia, documented migrant workers must undergo mandatory pre-employment medical examinations within the first month of arrival, and subsequently, annual medical examinations. These medical examinations are conducted at private clinics approved by the Foreign Workers Medical Examination and Monitoring Agency (Fomema) and include screening for a list of communicable and non-communicable diseases like HIV/AIDS, syphilis, tuberculosis, leprosy, hepatitis B, malaria, diabetes mellitus, hypertension, and also pregnancy for women. Female migrant workers testing positive for pregnancy will fail their medical examinations, and consequently will be denied work permits and are subject to deportation.
Most participants agreed that prohibiting pregnancy during employment is an infringement of a woman's reproductive rights and is discriminatory against women. This migrant representative described how pregnancy is equated with illness in mandatory health screening.
*"The women who are pregnant, they are considered [as having] an illness. Pregnancy is an illness. They failed [the FOMEMA medical examination] and they have to be sent back. It is like they discriminate [against] us as a woman. This is our reproductive right." MW-1*
This interviewee expressed discomfort with the government-mandated screening for pregnancy, as it does not fulfil the purpose of a pre-employment medical examination to ensure "fitness to work" and to protect the public from communicable disease.
*"To get your work permit, you have to pass the medical screening, but the medical screening is not only screening for contagious disease, but also for pregnancy. For me, personally, it becomes a problem when it infringes the reproductive right [of migrants]. Other screenings make sense, that is something that is needed to ensure public health for everyone. For the workers themselves to be 'fit to work' and for the health of society, they have to be free from contagious disease–that makes sense! But, reproductive health issue–that concerns reproductive rights. It infringes human rights." CSO-8*
| doab | 2025-04-07T04:13:02.977811 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 203
} |
ffa2a234-da32-4c59-b516-55230ec92248.204 | 3.1.2. Employment Contracts Prohibit Relationships, Marriages and Pregnancy
Several participants shared that employment contracts expressly forbid sexual relationships, marriages, and pregnancy. This medical practitioner explained that while both men and women are expected to be celibate, women are especially vulnerable because of the possibility of pregnancy.
*"Most of the migrant workers, especially the women, when they sign up agreements [employment contracts] with their companies, they are not allowed to get pregnant or be sexually active [throughout employment]. A lot of women have come to me and say, 'My boss shouldn't know this!' Because you are not allowed to have sex. It doesn't make sense! You are staying in this country for two years or more, and you are not allowed to have sex? Men and women are the same. But for the men, you don't see much consequences because they don't get pregnant! They don't have to worry about getting pregnant! Women have a more vulnerable position because they fear they will get pregnant." MD-12 CSO*
The immediate termination from employment is a direct consequence of pregnancy. This stakeholder informed that this practice, while legal, is inherently discriminatory against women.
*"Migrant workers who are pregnant, they lose their job almost immediately. So, these are some of the concerns that people are afraid of* . . . *In terms of why is there a discriminative practice? If the woman is pregnant, you automatically lose the job. That is questionable." IO-2*
3.1.3. Prohibiting Pregnancy Forces Women to Become Undocumented
Unlike expatriates from a professional, managerial, or highly skilled technical backgrounds, less skilled migrant workers are not allowed to bring family members or to get married in Malaysia in policy. This participant explains that this denial of the right to family results in unregistered marriages among non-citizens.
*"Reproductive rights, it is actually a basic of human right. You cannot say that [when] you come here, only the expatriate can have the family, non-expatriate cannot. This is human nature, you know? They got married, but they are not allowed to get married here. That is why there is a lot of 'nikah bawah tanah' [underground*/*unregistered marriages], so they get their own 'imam' [priest]* . . . *" IO-1*
Participants explained that migrant women who are pregnant and opt to keep their babies are driven to become undocumented migrants. This medical practitioner expresses surprise that many migrant women opt to deliver their babies in Malaysia despite the severe consequences.
*"They will automatically be illegal migrants, because the moment they are pregnant, they will lose their visa and if they lose their visa, they become illegal migrants. But somehow, many of them do deliver locally." MD-9 PRIVATE GP*
| doab | 2025-04-07T04:13:02.977961 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 204
} |
ffa2a234-da32-4c59-b516-55230ec92248.205 | *3.2. Sexual and Reproductive Health Education and Contraception*
| doab | 2025-04-07T04:13:02.978147 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 205
} |
ffa2a234-da32-4c59-b516-55230ec92248.206 | 3.2.1. Employers do not Provide Access to Family Planning
Although pregnancy is prohibited among low-skilled migrant women employed in Malaysia in policy, those interviewed informed that there is little support from employers in terms of providing SRH education or services, either in terms of preventing STIs or providing family planning services. This interviewee explained that the prevalent moral attitude in Malaysia—that providing family planning encourages sexual promiscuity—may explain employers' attitudes.
*"The thing that upsets me is that there is very little recognition that women migrant workers who come here are young and usually sexually active. It's a fact of life. We have actually tried, through our NGOs, to promote the information on contraception, and access to contraception for these people. But people [employers] are very cagey about this! It all has got to do with the idea that: 'Oh, they are only here to work, you know. They are not supposed to have boyfriends or relationships.' And therefore, 'Why should we give them any information on contraception? It will only make them bad workers.' But the reality is, many of them are sexually active. And then, of course, if they don't have access to contraception, they get unwanted pregnancies. And of course, for them to terminate their contract halfway, it's a real waste because they made arrangements to do a two to four year contract with the factories, intending to earn and send money home. But the moment they are found to be pregnant, you know, they have two choices; They either have an abortion, or they are sent back." MD-9 PRIVATE GP*
This participant implied that providing family planning services is an important investment for both employers and workers, as unwanted pregnancies may result in job loss.
Nevertheless, civil society organisations have approached employers and embassies to provide SRH awareness for migrant workers with mixed success. This participant illustrates the best practices of multinational companies that invest in their employee's health by training migrant community leaders to ensure the continuous education of new recruits.
*"There were programmes done by our NGO with a few companies, where we train their community leaders. So, we will start talking about, 'What is the menstrual cycle? How to prevent STDs? About contraception and everything'. So, these community leaders will keep training new people [newly recruited migrant workers]. So, they know where to get contraception and will come to the clinic to get this [SRH services]" MD-12 CSO*
| doab | 2025-04-07T04:13:02.978171 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 206
} |
ffa2a234-da32-4c59-b516-55230ec92248.207 | 3.2.2. Migrants Pay Out-of-Pocket for Sexual Reproductive Health Services
Family planning is freely available to local patients at public clinics, as part of a comprehensive package of maternal and child health services for citizens. This medical doctor explained that financial constraints may deter some female migrants from seeking contraception at public clinics, as non-citizens must pay for services.
*"To be honest, migrants have to pay for the contraception-for injectable hormonal therapy or any sort of contraception-they have to pay! As opposed to locals, where contraception is free. So, the problem still comes back to financial issue. So, if they are willing to pay and can a*ff*ord, and if they understand the importance to not conceive within the next two years, then they will pay for it. But most of them-no [they won't pay]." MD-13 PUBLIC CLINIC*
Migrant workers pay out-of-pocket for contraception at private clinics, as SRH services are not covered by the government-mandated migrant health insurance (SPIKPA) or employer-provided healthcare.
3.2.3. Private Practitioners Promote Expensive Contraceptives and Fail to Provide Information on SRH
Medical practitioners interviewed informed that although a wide range of contraceptives are available at private clinics, most migrants prefer injectable hormonal contraceptives, especially the commonly available Depo-Provera injections. This interviewee explained that private doctors do not sufficiently advise women on contraceptive options, such as on the use of long-acting contraceptives like intrauterine contraceptive devices (IUCDs) or implants, because these options are less lucrative than injectable hormonal contraceptives.
*"Not many GPs [general practitioners] even want to talk about it! But they keep telling them to use Depo-Provera because it is profitable! In a year, if you are coming [to the clinic for] 4 times. So, RM 60* × *4* = *RM 240. [Whereas, the] IUCD is RM 200 for 4 years. So, you are not going to see her for the next few years. It [the IUCD] is more economical for the woman, but it is less profitable for the doctors!" MD-12 CSO*
While the private practitioners interviewed acknowledged that there is a substantial market for contraceptives among migrant women due to perceived need, participants explained that the awareness and willingness-to-pay are low for the prevention of STIs, specifically the use of condoms.
*"I got quite a number of them coming for depo injections [Depo-Provera injections]. Contraception, in the form to prevent pregnancy-yes. But to prevent STDs [sexually transmitted diseases], they have to buy la* . . . *Condoms and all that, they have to just find ways to buy it* . . . *But I have had quite a number of them who come in for depo injection. So, they do know about it, and they do come." MD-2 PRIVATE GP*
This participant implied that migrants were not willing-to-pay for condoms, as this was not seen as essential. Likewise, there is very little information provided by medical practitioners regarding the use of condoms in the prevention of STI.
| doab | 2025-04-07T04:13:02.978445 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 207
} |
ffa2a234-da32-4c59-b516-55230ec92248.208 | *3.3. Abortion*
3.3.1. Migrant Women's Abortion Decisions Linked with Financial Security and Employer Support
Migrant women may lose formal employment and face deportation, as a consequence of pregnancy. Migrant women who chose to continue with their pregnancy in Malaysia are likely to become undocumented. Since the economic and social costs of pregnancy are substantial, this participant explained that migrant women that opt to continue with their pregnancy are usually in stable, committed relationships with relative financial security.
*"Migrants pay for antenatal care at private clinics themselves. So usually, the ones who are willing to keep a child, they know it's going to cost them. So, they should have some 'back up' money or husbands who are ok, and then they can a*ff*ord. Maybe he is taking home RM 1800 to RM 2000 a month. So, from all his work, he can a*ff*ord it. Then they go ahead. There are some who will come and say, 'No I can't, I can't a*ff*ord it'. Then some are like girlfriend*/*boyfriend, but he might be married, she might be married, you know* . . . *'accidents', you know. This group will come and ask if they can get a medical abortion." MD-2 PRIVATE GP*
Migrant workers are generally in Malaysia for the short term, with employment contracts lasting 2 to 4 years. This participant explained that many of the relationships formed by migrant workers in Malaysia are impermanent. Without support from a partner, pregnancies are unwanted and result in abortion.
*"Basically, when they arrive [in Malaysia], they may have a husband back home. But, after few months, no more. We heard from other Filipinos, that mostly after they separate from [their husband], they find someone else here. And then when they get pregnant, they just abort it." MW-3*
Several participants shared that some domestic workers are highly valued by their employers, and that these employers are supportive of their employees' pregnancy. Examples were given of employers sending workers back to their home countries for delivery, with the option to return to Malaysia for employment. Others gave examples of employers supporting their workers by bringing them to private clinics for antenatal follow-up. This participant shared that some employers support their domestic workers in having an abortion, as this would mean the domestic worker could keep her job.
*"In terms of unwanted pregnancies, they cannot be pregnant and stay in the work. But fortunately, many of the private employers want to keep their maids. Very often their maids are quite well-paid and they [the employers] are happy with them. And if they are pregnant, the employer [would] actually bring her along [to the clinic], and then you [as the doctor], would do the termination because she wants to continue working." MD-9 Private clinic*
It was unclear whether domestic workers in this position received pre-abortion counselling or advice from providers on their options.
| doab | 2025-04-07T04:13:02.978631 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 208
} |
ffa2a234-da32-4c59-b516-55230ec92248.209 | 3.3.2. Health Providers have Negative Attitudes towards Abortion
Although abortion is legal in Malaysia, the prevalent negative perception of termination of pregnancies has led to the widespread belief that it is illegal, even among healthcare providers. Many healthcare providers view contraception and abortion as sensitive topics and opt not to be part of the network of private healthcare providers offering safe abortion services. This participant explained that the opposition towards abortion is related to cultural norms, as the abortion laws in Malaysia are fairly liberal.
*"All of them who are at the top level [government] say: 'Oh, yeah, we have to recognise the law.' The law in Malaysia is almost identical to the English law on abortion. So, what happens on the ground, seems to be not so much an o*ffi*cial policy, but all 'cultural opposition' to make reproductive health and particularly contraception accessible to single women, and to make safe abortion accessible to women in general." MD-9 PRIVATE GP*
Abortions are rarely conducted at public healthcare facilities. While a selected number of private clinics provide safe abortions, these options are expensive and maybe unaffordable for low-wage migrant workers. Thus, migrant women may opt to perform illegal, self-induced abortions, which are likely to be unsafe. Medical practitioners interviewed informed that migrant women do present at the emergency departments of public hospitals with complications of unsafe abortions.
*"I have never seen any migrants coming to us for abortions [at public clinics]. They do it by themselves–self-induced. They have their own traditional ways of doing it, you know, by drinking vinegar and certain traditional medication, or they will try to induce trauma to the stomach! So, when they do present to us, it is already–not there [pregnancy terminated]. So, we had to refer them to the [public] hospital for a D&C [Dilation and Curettage]. Curettage is to clear o*ff *whatever is left behind." MD-13 PUBLIC CLINIC*
| doab | 2025-04-07T04:13:02.978833 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 209
} |
ffa2a234-da32-4c59-b516-55230ec92248.210 | 3.3.3. Medical Abortion Unavailable Legally
Malaysia has not legalised the use of medical abortions (i.e., non-surgical abortions, where oral medications are given to terminate pregnancy). While "abortion pills" are available for purchase online without a doctor's prescription, these pills are illegal and unregulated. Medical practitioners interviewed explained that medical abortions are a safe alternative and cheaper than surgical abortions. This medical doctor cautions patients against purchasing "abortion pills" from unknown sources and would make discreet referrals to a professional network of CSOs that facilitates safe abortions.
*"We have to advise them on medical or a surgical abortion. A surgical abortion will cost them almost RM 700 to RM 1000, which most of them don't have. So, instead of them harming themselves [unsafe abortion], we will actually tell them that 'The procedure is not available here [at this clinic]. Don't trust anybody, don't Google, don't find [abortion services] anywhere! Here are the contact details, where you can get pills online. But there's a possibility of not fully recovering. You will then need to see these certain doctors [who are] providing surgical [abortion]!' So, usually that is how we refer them to XXX." MD-12 CSO*
| doab | 2025-04-07T04:13:02.978984 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 210
} |
ffa2a234-da32-4c59-b516-55230ec92248.212 | 3.4.1. Migrants OPT for Private Clinics and Traditional Midwives for Antenatal Care
Due to immigration regulations, pregnant migrant workers in Malaysia inevitably become undocumented. While healthcare providers at public healthcare facilities will not deny patients necessary medical care, they are obliged to report undocumented workers to the police and immigration authorities. This medical practitioner explains that because of these restrictions, migrant women tend to opt for private healthcare.
*"They [migrants] tend not go to the 'Klinik Kesihatan' [public clinics for antenatal care], because they have to pay quite a bit for it. Some of them are scared that if they go there, and they [health authorities will] inform immigration department and they will be deported. So, they don't want to go to the government side. So, they don't get any [antenatal] follow up, they don't get anything. Sometimes you [would] ask them, 'Do you have antenatal records [home-based antenatal book given to patients at public clinics]?' No records, you know, that makes it very di*ffi*cult. But there are apparently some [private] clinics or some smaller maternity centres, who have their own follow-up for foreigners. So, they have their own [antenatal] book and they can go in for deliveries." MD-2 PRIVATE GP*
As the lack of antenatal follow-up and records prove problematic for the management of pregnancy and delivery, some more established private maternity centres provide more detailed follow-up for non-citizens.
Mainly due to the cost of private healthcare, some migrants prefer to deliver babies at home with the help of untrained traditional birth attendants. As shared by this migrant representative, this practice is likely done out of desperation, not cultural preference, and is linked with poor obstetric outcomes.
*"Some, they prefer to go to the traditional midwives. In some cases, that's why they pass away during delivery, because they don't want to go to the hospital. Because of the lack of documents and also because the payments are very high. So, they prefer to use the 'dukun beranak' [traditional midwife]. I found one [lady like that] last year, passed away in XXX. We had sent her to the hospital, but it was too late already. The baby also passed away." MW-1*
| doab | 2025-04-07T04:13:02.979076 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 212
} |
ffa2a234-da32-4c59-b516-55230ec92248.213 | 3.4.2. Delayed Booking, Incomplete Antenatal Follow-Up and Poor Obstetric Outcomes
Most stakeholders explained that due to healthcare costs, non-citizens tend to present late for booking and default follow-up at antenatal clinics. Doctors interviewed observed that late presentations could result in poor obstetric outcomes and avoidable complications. These complications would inevitably incur additional financial expense, as more advanced treatment may be necessary.
*"For migrants, when they present, it is already 30 weeks? 32 weeks? I even had one patient last week [who] presented at 36 weeks! So, that was the first time ever that I saw her. So, whatever that has happened, has happened! It is irreversible. For example, that is something we called: IUGR, which is 'Intrauterine growth restriction'. So, when that already occurs, nothing can be done! So, the baby may be born–with low birth weight from premature delivery. Then they will have a lot of complications! Like sepsis and all! So, all of these actually contribute to more financial burden to the patients! Because they will require a NIC [neonatal intensive care] admission for a long time!" MD-13 PUBLIC CLINIC*
| doab | 2025-04-07T04:13:02.979236 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 213
} |
ffa2a234-da32-4c59-b516-55230ec92248.214 | 3.4.3. Hospital Delivery Linked to Deportation
The government policy to report undocumented migrants to the police has resulted in incidents of women being detained immediately after delivery. Participants explained that conditions at detention centres are unsuitable for post-partum women and newborn babies. This interviewee explains that linking healthcare with deportation is a human rights violation.
*"We had a case of a migrant worker [who] was admitted to the hospital due to deliver. Within less than 24 h, both mother and the baby were already at the XXX Detention Camp. We [the CSO] needed to get intervention from the Embassy. They shouldn't detain the baby inside there because there are not such facilities, and besides, the mother was still very fragile, and shouldn't be detained. The Immigration Department persisted with their decision but [with] expedited repatriation. The Indonesian embassy refused to bear the expenses [of repatriation], so we [the CSO] had to find money for them. Because, the Indonesian government also has certain [financial] constraints. This was actually a very challenging situation for us." IO-1*
| doab | 2025-04-07T04:13:02.979330 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 214
} |
ffa2a234-da32-4c59-b516-55230ec92248.216 | 3.5.1. One Stop Crisis Centre Linked with Police
One Stop Crisis Centres (OSCC) were established in Malaysia since 1996 to assist survivors of gender-based violence (GBV), to obtain comprehensive care from multiple agencies in a common venue. The OSCC are located at the emergency departments of Ministry of Health hospitals in Malaysia and provide immediate treatment, while facilitating protection, counselling, medico-legal, and social support services for survivors of GBV, rape, sodomy and sexual assault, domestic violence, and child abuse.
Legally, citizens and non-citizens can use GBV-associated healthcare services, which are available free of charge regardless of citizenship status at OSCCs. However, reporting violence to the police is a pre-requisite to seeking care at the OSCC. This participant explains that the procedure for reporting to the police is not always consistent. Survivors of violence are supposed to go directly to the OSCC, and the police report should be done at the hospital. However, some survivors are asked by hospitals to go to the police station first before coming to hospital for treatment.
*"I have heard di*ff*erent information, at di*ff*erent times. Previously, I have heard [that] people should just go to the emergency [department] and then be referred to the OSCC. Then the police report will be lodged there. So, the police will go [there], to take the report. But I have also heard another story when they go to the emergency [department] and want to be directed to the OSCC, and they were asked to lodge [a police] report first, before they come [into OSCC]." CSO-8*
This participant explained that fear of the police is a hindrance faced by many non-citizen women seeking care or justice. Law enforcement personnel were described as lacking sensitization in dealing with GBV, and migrant women face additional discrimination.
*"I would say on the whole, there is definitely a lack of sensitisation amongst the police. I think in general, when it comes to gender-based violence, there is a lot of 'victim blaming' and those kinds of attitudes that are pretty pervasive. For non-Malaysian women, there is another layer of discrimination and some xenophobia. So, I think the quality of services is even lower for them! And then sometimes, if it is a situation where the employer has not done what they need to do to renew the work permit or the visa, then they might be afraid to go to the police because they can get reported to immigration! So, that is often a reason for women not to access help." CSO-7*
Study participants informed that undocumented migrants were particularly afraid to come forward to report incidences of violence, due to their immigration status.
### 3.5.2. Limited Shelters for Non-Citizens
While government and CSOs provide shelters, those interviewed informed that there is a shortage of shelters specifically designed for survivors of GBV. Study participants informed that government shelters provided by the Welfare Department are general shelters, which may also house the homeless or elderly populations, and may lack comprehensive case management of GBV.
Furthermore, not all government shelters accept non-citizen women. Participants informed that government shelters only accept non-citizens that have been issued protection orders by the courts.
*"There are two types of shelters, shelters run by the NGO and then shelters by [the] government, especially [the] Women Ministry [Ministry of Women, Family and Community Development] and the Jabatan Kebajikan [Welfare Department]. But government shelters that takes migrants are limited to migrant workers who have already been given a protection order; after the case has been determined by [the] police and court. Let's say the charges [pending] can be categorised as human tra*ffi*cking, then* . . . *the person will be given a protection order or an interim protection order during the investigation. Only then, will they be put in the government shelters." CSO-8*
Participants informed that migrants may also be reluctant to obtain refuge and protection at shelters, as they would have to make a report with the police. As government shelters are limited for migrant women, CSOs are an important source of assistance, also providing legal aid and counselling.
#### **4. Discussion**
In Malaysia, female migrant workers are subject to regulation of their reproductive rights with pre-employment and annual screenings for pregnancy, and face termination from employment if found pregnant.
Premature dismissals from employment are financially detrimental to both employers and migrant workers. Nevertheless, we found that information and access to family planning are seldom supplied to migrant workers by employers and not provided for by the government [20]. The state and employers essentially deny that migrant workers are sexually active adults, with the intent of avoiding being seen as promoting promiscuity by raising the topic of SRH. This outdated approach must change towards a pragmatic one, whereby migrant workers, including men, are provided with education and access to low-cost contraceptives. The low contraception prevalence in Malaysia (33.1% for all methods and 23.3% for modern methods) compared to the global estimate for 2019 (48.5% for all methods and 44.3% for modern methods) [27,28] may be explained by social, cultural, and structural barriers and lack of knowledge on contraception [29–32].
Our findings suggest that the choice of contraceptive methods among migrants may be influenced by the perceived risk of pregnancy and its consequences borne by women; hence, female-controlled methods like injectable steroids may be preferable, with less uptake of male-controlled barrier methods like male condoms. Poor uptake of condoms may also be explained by a worrying lack of awareness of STD and HIV prevention [33,34].
Although Malaysia has relatively liberal abortion laws, its interpretation is subject to cultural and religious resistance in the predominantly Muslim nation [35,36]. The Penal Code Act 574 (revised 1997) Section 312 permits safe abortion if performed by a registered medical practitioner and the medical practitioner determines that continuance of the pregnancy endangers the life of the pregnant woman or harms her physical or mental health [37–39].
In 2014, a 24-year-old Nepali migrant worker who opted for an abortion for fear of losing her job was arrested when police raided the clinic where she had her abortion. This Nepali worker was the first woman charged and convicted for having an abortion in Malaysia, although her conviction was subsequently acquitted [40–42]. This case illustrates the plight of migrant women under restrictive immigration laws and labour practices, as even after her innocence was proved and despite being no longer pregnant, the Nepali worker was dismissed by employers for being a "bad role model" [43,44].
Many medical practitioners, especially public sector providers, have conservative views and exercise personal judgement that restricts a woman's access to safe abortion [35]. While abortion services are available at certain private clinics, we found that financial constraints were a likely barrier for less skilled migrant women. Furthermore, the lack of information on where to obtain safe abortions and the underlying social stigma [35,45–47] are plausible drivers for migrant women to seek unsafe abortions, including unregulated medical abortions.
Medical abortion is a non-invasive, effective method for early pregnancy termination (within 49 days of the last menstrual period), that gives control to the woman rather than the healthcare provider [48]. Despite the recent classification of misoprostol and mifepristone as essential drugs–"where permitted under national law and culturally acceptable" by the World Health Organization [49,50], the Ministry of Health, Malaysia has yet to approve their use for medical abortions [36,37].
Notably, no participants undertook pre- or post-abortion counselling, either for their decision-making and feelings around abortion or on contraception post-abortion. We have no evidence for the latter in developed countries on increasing contraceptive uptake and acceptability [51].
Prohibition of pregnancy may result in avoidance of needed care due to apprehensions of job loss and deportation, and this may lead to treatment delays or unsafe abortions. It is accepted as a given by employers and healthcare providers that migrant women will want to terminate pregnancies so they can retain employment. Yet, the legal basis to prohibit pregnancy is unclear, as pregnancy as a clause for dismissal from employment is not specifically included in Malaysia's Employment Act or the Immigration Act [52,53]. Women are effectively coerced by policy and employment contracts into abortions, and this may curtail their reproductive rights.
Migrant workers face complex barriers in accessing healthcare in Malaysia, including financial constraints, the need to present legal documents like passports and work permits at public facilities, language barriers, discrimination, and physical inaccessibility [23,54]. Immigration policies in Malaysia essentially deny maternal and child health services for migrant workers at public facilities. Previous studies support our findings that migrant women are late in initiating antenatal care, while many never attend antenatal clinics and have home births with untrained birth attendants [19,55]. These factors may lead to delivery complications, as migrant women only seek care when critically ill, necessitating more advanced and expensive care [56]. As seen in other settings, migrant women are at higher risk of poor obstetric outcomes, including increased maternal and neonatal mortality, as compared to local women [57,58].
Malaysia successfully lowered maternal mortality through health system strengthening and meticulous auditing of all maternal deaths, including non-citizen deaths, through confidential enquiry into maternal deaths (CEMD) [59–61]. Unfortunately, while non-citizen maternal deaths are investigated in Malaysia, maternal mortality among migrants are not captured in national statistics reported internationally [56,60], raising questions if definitive measures to reduce risk in this group have been attempted.
States have the sovereign right to govern migration within national boundaries. However, the detention of new mothers and their babies for immigration offence may conflict with international laws and conventions. According to the Bangkok Rules or "the United Nations Rules for the Treatment
of Women Prisoners", non-custodial measures are preferred to the detention of vulnerable pregnant women and minor children [62]. Malaysia has ratified both the CEDAW and the Convention on the Rights of the Child (CRC) [16] and is under obligation to provide reasonable care and cater to the special needs of pregnant women, breastfeeding mothers, and mothers with children in custody [63].
In Malaysia, OSCC provides integrated services for victims of GBV at public hospitals [64,65]. While services are available to all women, in theory, barriers remain in practice for non-citizen women. Migrant women, especially those with precarious legal status, are reluctant to report violence or seek medical treatment, for fear of arrest and detention for an immigration offence. Lack of uniform implementation, seen here with confusion regarding the need for victims to report violence at police stations before seeking treatment, is an example of a shortfall in service. We would like to stress the importance of gender-sensitisation training among law enforcement agents, in terms of improving gender sensitivity and reducing discrimination against vulnerable non-citizen women [44,66].
This study has several limitations. Due to the sensitive nature of this study, we had difficulties obtaining interviews with migrant workers, employers, and policy stakeholders. Nevertheless, we were able to triangulate study findings by interviewing diverse key informants, including medical doctors, representatives of civil society organisations, trade unions, and academia. While the qualitative nature of this study precludes the generalisation of findings, we were able to illustrate the landscape of SRH services for migrant women in Malaysia by examining different stakeholder viewpoints and perspectives. We were also unable to fully explore the management of STIs and HIV/AIDS among migrant populations, an important component of SRH which would need dedicated future study.
This study has several strengths. Ours is one of few studies in Malaysia that explore the access to SRH services among vulnerable female migrant workers. We hope that this work will provide a vital understanding of some of the barriers faced by this vulnerable population and opportunities for intervention. We suggest that future quantitative research be conducted to fill the gap in SRRH data in Malaysia disaggregated by citizenship, especially on contraceptive usage, abortion, utilisation of SRH services, and maternal mortality.
| doab | 2025-04-07T04:13:02.979427 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 216
} |
ffa2a234-da32-4c59-b516-55230ec92248.217 | **5. Conclusions**
This study shows that the SRHR of migrant workers remains severely curtailed in Malaysia. Political will is necessary to revise restrictive immigration laws and labour policies to enable low-skill migrant workers to fulfil their SRHR. We suggest that instead of the discriminatory prohibition of pregnancy during employment, that all migrant workers are provided with access to SRH education and low-cost contraception by employers. All pregnant women, including non-citizens, should also have equal access to antenatal and delivery care at public healthcare facilities, and healthcare access should be decriminalised. A more inclusive, rights-based approach to healthcare access would have population-wide benefits, and this would put Malaysia towards the path of meeting the SDG target of 3.7 for universal access to SRH services.
**Author Contributions:** Conceptualization, T.L.; methodology, T.L., Z.X.C. and N.S.P.; software, T.L. and Z.X.C.; validation, T.L., Z.X.C., A.W.d.S., and N.S.P.; formal analysis, T.L. and Z.X.C.; investigation, T.L., Z.X.C. and N.S.P.; resources, T.L. and N.S.P.; data curation, Z.X.C.; writing—original draft preparation, T.L.; writing—review and editing, T.L., N.S.P., Z.X.C. and A.W.d.S.; project administration, Z.X.C.; funding acquisition, T.L., and N.S.P. All authors have read and agreed to the published version of the manuscript.
**Funding:** This research was funded by The Asia Pacific Observatory (APO) on Health Systems and Policies [grant number IF034-2020] and the China Medical Board's Equity Initiative [grant number IF055-2018].
**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.
| doab | 2025-04-07T04:13:02.980249 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 217
} |
ffa2a234-da32-4c59-b516-55230ec92248.219 | **A System Model of Post-Migration Risk Factors A**ff**ecting the Mental Health of Unaccompanied Minor Refugees in Austria—A Multi-Step Modeling Process Involving Expert Knowledge from Science and Practice**
#### **Nicole Hynek 1,\* , Arleta Franczukowska <sup>2</sup> , Lydia Rössl <sup>3</sup> , Günther Schreder <sup>1</sup> , Anna Faustmann <sup>3</sup> , Eva Krczal <sup>2</sup> , Isabella Skrivanek <sup>3</sup> , Isolde Sommer <sup>4</sup> and Lukas Zenk <sup>1</sup>**
Received: 23 June 2020; Accepted: 12 July 2020; Published: 14 July 2020
**Abstract:** Various studies have indicated that unaccompanied minor refugees (UMRs) have a higher risk of suffering from mental health problems than do accompanied minor refugees and general population norm. However, only a few studies provide data on UMRs regarding post-migration risk factors, their interrelations, and their influence on mental health. In this study, system models of post-migration risk factors for mental health and their interactions were developed in the case of Austria. In three consecutive interactive workshops with scientists and practitioners, fuzzy-logic cognitive mapping techniques were used to integrate the experts' knowledge. The resulting final system model consists of 11 risk factors (e.g., social contacts in the host country, housing situation, or professional health care services). The model provides a deeper insight into the complexity of interrelated direct, indirect, and reciprocal relations, as well as self-reinforcing triads. This systemic approach provides a sound basis for further investigations, taking into account the inherent complex multifactorial dependencies in this topic.
**Keywords:** system models; expert knowledge; fuzzy-logic cognitive mapping; unaccompanied minor refugees; mental health; post-migration risk factors
| doab | 2025-04-07T04:13:02.980492 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 219
} |
ffa2a234-da32-4c59-b516-55230ec92248.220 | **1. Introduction**
During the refugee movements in the years 2014 to 2016, European countries faced a rapidly growing number of refugees and asylum seekers, resulting in a record level of 1.3 million asylum applications in 2015 [1]. Of these, 8277 asylum applications were filed by under-age refugees [2]. Although the number of asylum-seeking applications has been decreasing since 2016, asylum seekers and refugees still represent a persistent issue and an important field of action, especially in terms of the vital matter of ensuring their mental health and well-being.
Refugees may suffer from mental disorders, as they often cope with dire situations and life events that affect their mental health. In particular, numerous studies have highlighted the special vulnerability of refugee children and adolescents to mental health problems and psychiatric disorders [2–8]. In a systematic review [3] that included 47 studies from 14 European countries and published from 1990 to 2017, covering a sample size of 24,786 refugee or asylum-seeking minors, the estimated point prevalence of diverse psychiatric disorders and mental health problems for children and adolescents was reported as follows: For posttraumatic stress disorder (PTSD), between 19.0% and 52.7%; for depression, between 10.3% and 32.8%; for anxiety disorders, between 8.7% and 31.6%; and for emotional and behavioral problems, between 19.8% and 35.0%. Despite the highly heterogeneous evidence base, it can be assumed that up to one-third of refugee and asylum-seeking children and adolescents suffer from depression, anxiety disorders, or emotional or behavioral problems, and up to one-half could be affected by PTSD [3]. Moreover, various studies showed that unaccompanied minor refugees (UMR) had a higher risk of suffering from mental health problems and psychiatric disorders than did the accompanied minor refugees and general population norm (cf., [3]). Out of this group, female UMRs appear to be exposed to a higher risk for developing mental health problems, PTSD and depression [4]. Several risk factors can influence the point of prevalence of mental health problems in minors. The literature differentiates between factors of pre-migration, e.g., traumatic exposure to poverty, violence, and war; factors of peri-migration, e.g., separation, sexual abuse, and trafficking; and factors of post-migration, e.g., access to education, social support, the asylum application process, discrimination, acculturation, insecure living conditions, and uncertainty about the future [9–13]. The latter factors are those that can be triggered by country-specific policies of the host countries through the creation of conditions that either hinder or facilitate the integration of refugee children and adolescents and, consequently, either reinforce or weaken their mental health. A comparison of approaches to the integration of UMRs established by EU Member States (MS) shows that MS generally give priority to the care of UMRs. Thereby, MS apply similar accommodation arrangements, appoint a representative (e.g., guardian), provide access to education, the labor market, social welfare assistance and health care (including emergency treatment, basic medical care, and in many cases additional specialized medical care and counseling) [5]. Despite these similarities, integration policies and processes generally depend on the country-specific environment and, hence, post-migration risk factors may be more or less subject to social, political, and economic conditions.
One of the main challenges for the integration of UMRs found in the above-mentioned comparison refers to a lack of specialized and trained staff [5]. This highlights the importance of political and practical recommendations for coping with post-migration risk factors, taking into account the specific conditions affecting the situation of UMRs in a foreign country.
Given the complex interplay between post-migration risk factors, mental health status, and country-specific integration conditions, this study focuses on Austria as a country recording a high share of (unaccompanied) minor asylum applicants [1]. However, in the case of Austria, only a few studies provide data on post-migration risk factors and their influence on the mental health of UMRs. Among others, the Institute for Empirical Social Research (IFES) carried out an exploratory study on the living conditions of UMRs in eastern Austria [14]. This study included quantitative and qualitative interviews with 66 UMRs who spoke about their living situations and future expectations. The study provides insights into essential framework conditions such as accommodation, access to education and work, the financial situation and daily routines of UMRs, and their future perspectives, even if the influence on mental health was not specifically surveyed. Other available publications on specific post-migration risk factors are mostly from non-profit-organizations (e.g., Asylkoordination Österreich, Caritas), international institutions (e.g., European Migration Network), and government institutions and ministries (e.g., Ombudsman Board, the Federal Ministry of Education, Science and Research), though they focus mainly on practical and/or legal matters.
Moreover, significant interrelationships between risk factors have not yet been sufficiently investigated. Consequently, the interactions and dynamics of risk factors of this complex domain
of research may be inadequately addressed and implemented in the development of strategies or measurements. In the case of UMRs, this could not only raise costs for secondary and tertiary professional care but also hinder their sustainable integration into society. Thus, evaluating the effectiveness of appropriate measures to ensure the mental health of UMRs requires more than the mere aggregation of parts of the interrelated and multidimensional factors contributing to vulnerabilities of UMRs.
In view of the topicality and relevance of the subject and the scarce availability of data at national and international levels, this study aims to identify post-migration factors, their interrelationships, and potential influence on mental health. An interdisciplinary and systemic approach incorporating experts' knowledge and experiences was applied to obtain different perspectives from both a scientific and a practical point of view. The final goal of this study was to develop a knowledge-based system model that integrates different knowledge and describes post-migration risk factors as a network of interacting factors. In research and practice, knowledge-based system modeling techniques are used to retrieve the knowledge and represent how individuals organize knowledge, link concepts within a knowledge domain, and understand complex problem situations [15–17]. Different types of problems can benefit from this approach of incorporating experts' knowledge, if, for instance, scientific data on a specific topic is limited or if the problem involves many parties and has no clear solution or clearly correct answers and is, therefore, complex [18–20].
In this study, we incorporated the approach of system modeling, as described in the section *Materials and Methods*. Based on our final system model developed within three workshops as described in the *Procedure* section, the research results are presented and discussed within the context of the specific legal and policy framework in Austria. The developed system model calls for a more detailed analysis and should currently only provide insights into future needs for action and further investigation as described in the *Limitations* and *Conclusions*.
#### **2. Materials and Methods**
#### *2.1. The Technique of Fuzzy-Logic Cognitive Mapping*
In this study, the technique of fuzzy-logic cognitive mapping (FCM), a commonly used form of semi-quantitative system modeling, was applied. It is an established technique for eliciting, capturing, and diagramming structured knowledge on interrelated issues of a knowledge domain held by individuals or groups [21,22]. The result of the process of system modeling is a cognitive map that takes the form of a system and provides a visual representation of a person's existing understanding on a particular subject. A cognitive map has three characteristics [23]. The first characteristic includes both the direction and nature of causality; the second characteristic reveals the strength of the connections, and the third characteristic reflects a feedback mechanism that captures the effect of a change in a node on another node that, in turn, affects other nodes along the path. In other words, these cognitive maps contain nodes, i.e., concepts that can be linked qualitatively (e.g., low, medium, high) or quantitatively (e.g., between −1 and 1) to other concepts. This connectivity allows researchers to uncover trends by measuring the degree of conceptual agreement in the cognitive maps produced by individuals [15] or by groups jointly defining, expanding, discussing, and collaborating on the concepts and structures of a system [18].
Furthermore, by applying mapping techniques in a group environment, the knowledge of the group can be extended by developing and discussing the emerging system models. That is, collaborative system modeling can act as scaffolding tools to create an environment in which diverse group members can share their knowledge [24]. Thus, by generating a representation of the problem to solve a particular situation, individual contributors can build on each other, and actions or ideas can be taken up or complemented by other group members [25].
| doab | 2025-04-07T04:13:02.980630 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 220
} |
ffa2a234-da32-4c59-b516-55230ec92248.221 | *2.2. Participants*
The participants who were invited to the first workshop consisted of five female scientists working in the fields of migration studies and health economics (see A1–A5 in Table 1). All of them have several years of experience in research projects on health-related migration issues at the same university but are partly assigned to different departments. The expert group of practitioners invited to the second workshop consisted of three male and two female practitioners who worked directly with UMRs in social or medical fields (see B1–B5 in Table 1). All of them had several years of experience in health-related migration issues in their various fields of expertise. They were not acquainted with each other and had no official organizational connections. The participants in the third workshop were the same scientists who had taken part in the first workshop, including one additional scientist from the field of evidence-based medicine (see C6 in Table 1).
**Table 1.** Participants from workshop 1 to workshop 3.
A = scientists of workshops 1 and 3, B = practitioners of workshop 2, C = additional scientists of workshop 3.
| doab | 2025-04-07T04:13:02.981300 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 221
} |
ffa2a234-da32-4c59-b516-55230ec92248.222 | *2.3. Procedure*
Study phases I–III were conducted to create the final system model of post-migration risk factors in Austria that contains integrated information and assumptions about the relationship between elements in the system of risk factors. The aim was to obtain different perspectives from both a scientific and practical point of view. The following phases were implemented.
| doab | 2025-04-07T04:13:02.981396 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 222
} |
ffa2a234-da32-4c59-b516-55230ec92248.223 | 2.3.1. Phase I: Co-Developing System Models
In workshop 1 with five scientists and workshop 2 with five practitioners, each group co-developed a system model of risk factors of UMRs' mental health upon arrival in Austria. The research question was as follows: "Which influencing factors negatively affect the mental health of UMRs, whereby UMR is used to define persons under the age of 18 who came to Austria without care between the years 2003 and 2018?" In the case of unclear points, the facilitators supported the groups; otherwise, they kept away from the task execution and simply observed the interaction of the participants. The co-development of a system model and its relationships were elaborated on in four tasks:
• Task 1: Individual elicitation of influential factors.
Using moderation cards, each participant freely wrote up to 10 factors associated with UMRs' mental health that he or she considered essential based on his or her knowledge and experience.
• Task 2: Collaborative identification of the main factors for the system model.
After the individual elicitation of up to 10 risk factors, the participants presented those risk factors to the group and jointly clustered the cards regarding the topics that were similar and most relevant. In a collaborative effort, they labeled these clusters on new moderation cards and defined the names of the main factors to indicate more general concepts.
• Task 3: Collaborative drawing of relations between factors.
After a detailed explanation of how to create FCM, each expert group built a system model and collaboratively drew directed relations between the identified factors. A positive relationship meant a positive effect, i.e., if the value of factor A increases, then the value of factor B also increases, while if the value of factor A decreases, then the value of factor B decreases. A negative relationship meant a negative effect, i.e., if the value of factor A increases, then the value of factor B decreases, while if the value of factor A decreases, then the value of factor B increases. The participants placed the factors written on cards on a flip chart in an arrangement of their choice and drew the relationships. *Int. J. Environ. Res. Public Health* **2020**, *17*, x FOR PEER REVIEW 5 of 17 a positive effect, i.e., if the value of factor A increases, then the value of factor B also increases, while if the value of factor A decreases, then the value of factor B decreases. A negative relationship meant a negative effect, i.e., if the value of factor A increases, then the value of factor B decreases, while if
• Task 4: Collaborative decision-making of the strength of effects. the value of factor A decreases, then the value of factor B increases. The participants placed the factors written on cards on a flip chart in an arrangement of their choice and drew the relationships.
After sketching the essential directed relations, the groups decided upon the strength of each relation. For the group of practitioners, a distinction was made between strong, medium, and light relations. A strong positive relation was marked, for example, with three plus signs ("+++"), while a slightly negative relationship was marked with a minus sign ("−"). Considering that scientists are accustomed to numerical expressions of correlations, they evaluated the interactions between the factors as decimals ranging from −1 to +1. Figure 1 depicts the two system models developed in workshop 1 and workshop 2. Task 4: Collaborative decision‐making of the strength of effects. After sketching the essential directed relations, the groups decided upon the strength of each relation. For the group of practitioners, a distinction was made between strong, medium, and light relations. A strong positive relation was marked, for example, with three plus signs ("+++"), while a slightly negative relationship was marked with a minus sign ("−"). Considering that scientists are accustomed to numerical expressions of correlations, they evaluated the interactions between the factors as decimals ranging from −1 to +1. Figure 1 depicts the two system models developed in workshop 1 and workshop 2.
**Figure 1.** Part of the group system model of post‐migration risk factors in Austria co‐developed in workshop 1 with scientists (**a**) and workshop 2 with practitioners (**b**). **Figure 1.** Part of the group system model of post-migration risk factors in Austria co-developed in workshop 1 with scientists (**a**) and workshop 2 with practitioners (**b**).
#### 2.3.2. Phase II: Consensus‐Based Evaluation of the Main Risk Factors 2.3.2. Phase II: Consensus-Based Evaluation of the Main Risk Factors
Workshop 3 was designed to enable the group of scientists of this study, as a larger group of six people, to integrate the factors and assumptions of the two groups' system models from workshops 1 and 2. The task was to jointly identify those factors the participants regarded as the main risk factors of this study. Thus, all the factors identified in the scientists' and practitioners' system models were evaluated until an agreement was reached. The co‐evaluation of the risk factors followed three criteria: (1) The degree of impact a factor has on UMRs' mental health, i.e., the weighted degree; (2) the impact's effects on other factors, i.e., the outdegree; and (3) the degree of being influenced by other factors, i.e., the indegree. 2.3.3. Phase III: From Individual Models to a Shared System Model Workshop 3 was designed to enable the group of scientists of this study, as a larger group of six people, to integrate the factors and assumptions of the two groups' system models from workshops 1 and 2. The task was to jointly identify those factors the participants regarded as the main risk factors of this study. Thus, all the factors identified in the scientists' and practitioners' system models were evaluated until an agreement was reached. The co-evaluation of the risk factors followed three criteria: (1) The degree of impact a factor has on UMRs' mental health, i.e., the weighted degree; (2) the impact's effects on other factors, i.e., the outdegree; and (3) the degree of being influenced by other factors, i.e., the indegree.
#### After defining the main risk factors of the final system model of this study, the researchers of workshop 3 rated the relationships between these factors and the impact of each factor on mental 2.3.3. Phase III: From Individual Models to a Shared System Model
health. An Excel workbook was created, with one sheet providing the introduction and an example of the task and a second sheet providing a matrix of the factors identified in phase II of the study. In the matrix, the six researchers individually rated all relations between all factors from 0 (no influence) to 10 (most substantial influence). Additionally, the main risk factors were rated regarding their impact on mental health between 0 (no impact) and 10 (most substantial impact). In this sense, the scientists' mental models, i.e., their knowledge, beliefs, and assumptions about the risk factors of UMRs, were individually gathered to calculate a shared system model of post‐migration risk factors of UMRs in Austria. This shared system model is the final model of this study and was calculated by After defining the main risk factors of the final system model of this study, the researchers of workshop 3 rated the relationships between these factors and the impact of each factor on mental health. An Excel workbook was created, with one sheet providing the introduction and an example of the task and a second sheet providing a matrix of the factors identified in phase II of the study. In the matrix, the six researchers individually rated all relations between all factors from 0 (no influence) to 10 (most substantial influence). Additionally, the main risk factors were rated regarding their impact on mental health between 0 (no impact) and 10 (most substantial impact). In this sense, the scientists' mental models, i.e., their knowledge, beliefs, and assumptions about the risk factors of UMRs, were individually gathered to calculate a shared system model of post-migration risk factors of UMRs in Austria. This shared system model is the final model of this study and was calculated by aggregating and filtering the data from all scientists. This procedure allows for the identification of the main relational patterns between the factors. The following thresholds were defined to count a relation in the final system model: Two-thirds of the group (i.e., at least four persons) indicated a value higher than 0 (from 0, no influence, to 10, most substantial influence), and the aggregated average score was higher than 5.0. Afterward, the calculated matrix was visualized as a system model, using the freely available software package Visone. Phase III ended with a sense-making workshop; the researchers evaluated the final system model to identify missing factors or correlations and discussed relational patterns by referring to the current literature. The results of this sense-making process are described in the Discussion and Conclusions sections. *Int. J. Environ. Res. Public Health* **2020**, *17*, x FOR PEER REVIEW 6 of 17 aggregating and filtering the data from all scientists. This procedure allows for the identification of the main relational patterns between the factors. The following thresholds were defined to count a relation in the final system model: Two‐thirds of the group (i.e., at least four persons) indicated a value higher than 0 (from 0, no influence, to 10, most substantial influence), and the aggregated average score was higher than 5.0. Afterward, the calculated matrix was visualized as a system model, using the freely available software package Visone. Phase III ended with a sense‐making workshop; the researchers evaluated the final system model to identify missing factors or correlations
#### **3. Results** and discussed relational patterns by referring to the current literature. The results of this sense‐ making process are described in the Discussion and Conclusions sections.
#### *3.1. Experts' System Models on Post-Migration Risk Factors* **3. Results**
Figure 2 depicts the system models of the expert groups in workshops 1 and 2 of this study as well as the final system model evaluated in the course of this study. In all three system models, the positive relations between the nodes are represented as green arrows in three categories: Medium influence (5.0–5.9, light green), strong influence (6.0–6.9, green), and powerful influence (>7.0, dark green). Likewise, the red arrows, i.e., the light red, orange, and red ones in the system models of workshops 1 and 2, represent the negative relations between the factors, whereby they were applied only in these two networks. Additionally, the scientists of workshop 3 rated the main risk factors of the final system model in terms of their impact on mental health. In Figure 2, these ratings are visualized as blue nodes in three categories: Medium impact (5.0–5.9, light blue), strong impact (6.0–6.9, blue), and powerful impact (>7.0, dark blue). *3.1. Experts' System Models on Post‐Migration Risk Factors* Figure 2 depicts the system models of the expert groups in workshops 1 and 2 of this study as well as the final system model evaluated in the course of this study. In all three system models, the positive relations between the nodes are represented as green arrows in three categories: Medium influence (5.0–5.9, light green), strong influence (6.0–6.9, green), and powerful influence (>7.0, dark green). Likewise, the red arrows, i.e., the light red, orange, and red ones in the system models of workshops 1 and 2, represent the negative relations between the factors, whereby they were applied only in these two networks. Additionally, the scientists of workshop 3 rated the main risk factors of the final system model in terms of their impact on mental health. In Figure 2, these ratings are visualized as blue nodes in three categories: Medium impact (5.0–5.9, light blue), strong impact (6.0– 6.9, blue), and powerful impact (>7.0, dark blue).
**Figure 2.** Illustrations depict the system models of post‐migration factors developed in workshops 1 and 2. The main factors identified in workshop 3 are illustrated in the center, and the final system model developed throughout the three workshops can be seen below. **Figure 2.** Illustrations depict the system models of post-migration factors developed in workshops 1 and 2. The main factors identified in workshop 3 are illustrated in the center, and the final system model developed throughout the three workshops can be seen below.
The derivation of the main risk factors of the final system model during the third workshop can be found in Table 2. Congruent factors were identified and maintained, similar factors were summarized, and rarely named factors were removed. As can be seen in Table 2, some factors were not included in the final system model, even though they were named in both workshops.
**Table 2.** Clustered post-migration risk factors identified in the three workshops.
The factors of the scientists were identified in workshop 1 (N = 5), the factors of the practitioners in workshop 2 (N = 5), and the main factors in workshop 3 (N = 6). The main factors are presented as short names. Professional care services <sup>1</sup> (social and health), and language and education <sup>2</sup> are single risk factors that have been split into two main risk factors.
As in the case of the factor *future perspectives*, the scientists argued that this factor was already covered by the remaining risk factors, such as *income security* and *residence security*. They agreed to exclude the factors *criminal conduct* and *prior information about the dangers of flight and circumstances in the EU* because these had been classified as a mix of post- and pre-migration factors.
Furthermore, the factor *discrimination* was not included in the final system model because it has been considered the possible output of the determinant *political and social climate of the host country*. The factor *contact with family and family remittances*, which includes, among others, financial family support, was initially subsumed under the factor *income security* and excluded afterward. The scientists of workshop 3 considered that the host country cannot influence this factor, and it was ambiguous as the factor of *family reunification*, which was also subsequently excluded. The two factors were excluded for the following reasons: First, the two factors may exert both a positive impact and a negative impact
on the mental health of UMRs. For example, the prospect of family reunification can be both a positive future perspective and a source of pressure, whichever is more likely is related to the status of the UMR and the expectations of the family. The same applies to *financial support* for or from the family in the country of origin. Second, the factors have a strong linkage to the circumstances and conditions of the country of origin. Hence, it could not be clearly distinguished as a post-migration risk factor.
Table 3 lists the factors of the final system model depicted in Figure 2 and ranks them by the average score, as determined by the researchers, of the indicated impact on mental health. The network metric indegree calculates the incoming arrows (the factors that influence this specific factor), while outdegree calculates the outgoing arrows (the factors that this specific factor influences). The absolute number measures the amount of incoming or outgoing arrows, while weight measures the percentage of the weighted relations, i.e., the strength of the influence. In this context, a factor with a high absolute indegree is influenced by many other factors, while a high weighted indegree indicates that the factor is strongly influenced by other factors with a high impact, regardless of the amount of the other factors. For instance, the factor social contacts is influenced by many other factors (absolute indegree of 6) that are also highly weighted (weighted impact of 21%). The outdegree signifies the extent to which a specific factor influences other factors, e.g., health care influences a high number of other factors (absolute outdegree of 5) that are also highly weighted (weighted impact of 17%).
**Table 3.** Network metrics of the final system model.
The main risk factors are presented in short names.
| doab | 2025-04-07T04:13:02.981438 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 223
} |
ffa2a234-da32-4c59-b516-55230ec92248.224 | *3.2. Main Risk Factors of the Final System Model*
The resulting final system model, as depicted at the bottom of Figure 2, consists of *11 main risk factors*, listed and described below in descending order of their impact on the mental health of UMRs. For better orientation, the impact on UMRs' mental health on the respective factor is listed in parentheses:
the facilitation of family reunification, organization of leisure and recreational activities, and asset management [26].
| doab | 2025-04-07T04:13:02.982360 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 224
} |
ffa2a234-da32-4c59-b516-55230ec92248.225 | *3.3. Description of the Final System Model*
The final system model comprehends post-migration risk factors on the individual and societal levels as well as factors combining characteristics of both. However, a strict differentiation of individual and societal factors is not possible. As depicted in Figure 2, at the bottom left of the final system model, most of the factors are on an individual level or lie within the individual's sphere of influence. Furthermore, they are linked to the general and individual conditions in the country of origin, e.g., *social contacts and relationships in the host country, German language skills*, *daily structure and leisure activities*, and *sociocultural adaptation*.
The factor *social contacts and relationships in the host country*, left in the model, was rated as a factor having a powerful impact on the mental health of UMRs. This factor also showed the highest number of six incoming ties: *Political and social climate in the host country*, *German language skills*, *sociocultural adaptation*, *daily structure and leisure activities*, *professional social care services*, and *access to education and training in the host country*. On the lower left in the model, a triad with reciprocal relations consists of the factors *sociocultural adaptation*, *German language skills*, and *social contacts and relationships in the host country*. A closer look at the factors *sociocultural adaptation* and *social contacts and relationships in the host country* reveals the reciprocal influence and direct as well as indirect levels of impact: While *sociocultural adaptation* describes the ability and individual competence to adjust to new social and cultural settings, the factor *social contacts and relationships in the host country* describes a possible consequence of this
competence, but also has an influence on the development and extension and application of this ability (sociocultural adaptation). Social contacts, exchanges and relationships are one possible way of acquiring and applying knowledge and skills about language, culture, and the societal context. Furthermore, *sociocultural adaptation* has a powerful effect on *German language skills* and, therefore, both indirectly and directly influence *social contacts and relationships in the host country*.
The factor German language skills has the highest number of strong reciprocal relationships with other factors. Among these factors are access to education and training in the host country, social contacts and relationships in the host country, and sociocultural adaptation. Only the influence on income security is one-directional. In addition, the factor German language skills is indirectly influenced by political and social climate in the host country via the powerful factor social contacts and relationships in the host country, and itself indirectly influences other factors that have a powerful impact on the mental health of UMRs: Daily structure and leisure activities via access to education and training in the host country, as well as professional health care services and housing situation via income security. The factor daily structure and leisure activities is influenced mainly one-sidedly by other factors, with one exception: It has a medium influence on social contacts and relationships in the host country as part of a reciprocal relation.
The factors *professional social care services* and *residence security* describe societal factors and are characterized by a high outdegree. The factor *political and social climate in the host country* shows several noteworthy characteristics: Although it has only a medium impact on the mental health of UMRs, this factor directly influences factors with a strong and powerful effect and indirectly affects numerous other (very) strong factors. No other factor influences this factor in the model but, through direct—and primarily indirect—relations, it indicates a high degree of impact on factors that can strongly influence the mental health of UMRs.
*Professional social care services* is a factor with a high outdegree (five relations with other factors) and a low indegree; it is influenced only by *political and social climate*. All these relations are one-directional. The factors *residence security* and *income security* have a substantial effect on the mental health of UMRs and are of high relevance due to their influence on factors with a powerful impact. Most notably, *income security* influences three factors that have a powerful impact on the mental health of UMRs: *housing situation*, *daily structures and leisure activities*, and *professional health care services*. Furthermore, it is influenced by *residence security* and *access to education and training in the host country* (by shaping job opportunities) as well as by *German language skills*. *Residence security* exerts a medium influence on the factors *housing situation* and *access to education and training in the host country* and is influenced in a common way by *political and social climate*. Furthermore, this factor has a strong impact on *income security*.
Particularly striking is the frequency of reciprocal relationships as well as the strength and number of indirect and direct relations among the five factors of *social contacts and relationships in the host country*, *access to education and training in the host country*, *daily structure and leisure activities*, *German language skills*, and *sociocultural adaptation*. These factors and their interactions can also be influenced—at least in part—at the individual level. This cluster is characterized by an accumulation of five factors with a high indegree as well as high outdegree and high intensity of the (reciprocal) effect to other nodes. This also means that if one factor were to change, all the other factors would be influenced as well.
#### **4. Discussion**
The further comparison of the main factors with international and country-specific studies on post-migration risk factors of UMRs points out that while many of the identified factors correspond to the results of international research, the multifactorial impact and relations of factors are seldom addressed. A look at the final system model, including its factors and relationships, reveals that the mental health of UMRs results from a complex interplay between individual and societal-level factors as well as factors combining different characteristics.
The societal factor *professional social care services* has a powerful impact on UMRs' mental health and various other highly relevant post-migration risk factors. Since UMRs must get by without family backing, support programs connecting them to members of their own and new cultures are recommended to reduce psychological distress [28,29]. In Austria, care for UMRs takes place under a guardianship arrangement, irrespective of the UMRs' residence status. Until they obtain their residence title, UMRs receive care and social benefits from basic welfare support for foreigners in need of aid and protection and, afterward, from the child and youth welfare system, though provisions can vary from province to province [30,31]. A special challenge is the long waiting periods for the appointment of guardianship, which can lead to delayed diagnosis and treatment of any mental health illness [14]. Furthermore, the basic welfare support legislation in Austria does not define minimum requirements related to qualifications for UMRs' counselors [26,28]. Considering the societal factor *professional health care services*, which also strongly influences UMRs' mental health, the present situation is as follows: In most EU countries, health examinations of newly arrived refugee children are provided on either a mandatory basis or a voluntary basis [29]. In Austria, UMRs receive health insurance coverage as part of basic welfare support. Once their residency status has been determined, costs are covered by the statutory health insurance and by the child and youth welfare system. Thus, UMRs get the same access to health care as do Austrian child citizens [30]. However, nothing is known about medical assessments specifically designed to identify UMRs' special health care needs [28], and there is a general short supply of psychotherapeutic and psychiatric services [28,30].
Due to its highest number of relations to other factors and assumed medium impact on UMRs' mental health, *access to education and training in the host country* represents a critical factor in the final system model. According to the 2017 data, which, however, do not differentiate between UMRs and accompanied children and adolescents, 18,468 asylum seekers were registered in compulsory general schools and secondary general schools in Austria [31]. The UMRs subject to compulsory schooling generally attend schools as "exceptional pupils or students," regardless of their residence status. This status allows them to take beginners and tutorial language courses to acquire *German language skills*. However, access to vocational training is granted only to a UMR with a residence title [28]. In 2016/2017, transitional courses were established for young refugees without compulsory schooling at vocational secondary and general higher schools [31]. Education and training play an important role in daily structuring, as shown in our final system model. *Daily structure and leisure activities* have been considered to have a powerful impact on UMRs' mental health and are influenced by various other risk factors. In Austria, the Basic Welfare Support Agreement stipulates the structuring of daily activities. It covers items such as sports, recreation, individual and group activities, and household tasks [28].
The final system model supposes a powerful relation between the *housing situation* and UMRs' mental health. For center or campus housing, evidence of the impact of living arrangements on the mental health of UMRs is consistent. Living in a reception center was associated with slightly higher risks of internalizing and/or externalizing problems [32], while minors placed in a reception center for adults had higher levels of psychological distress than did those in a youth reception center [33]. However, differences are no longer significant when one adjusts for the outcome of the asylum application [33]. In one study, living in foster care was associated with higher PTSD symptoms than was living more independently [34]. By contrast, another study found that refugee minors living with a family had significantly less depressive symptoms than did those with other living arrangements [8]. PTSD symptoms were also significantly higher among those in low-support living arrangements [35], while living in unsupported accommodation was associated with psychological ill-health [36]. In Austria, UMRs' accommodation is provided within the framework of either basic welfare support or child and youth welfare [28]. The low accommodation standards in the area of basic welfare support are criticized compared to other socio-educational institutions running under the child and youth welfare system. In particular, serious differences in accommodation standards among provinces were noted [31].
In the final system model, uncertain application processes and missing *residence security* were associated with poor mental health. Considering this result, international studies have conflicting findings. Yet, ultimately, quicker resolution of asylum claims to reduce insecurity and related distress is recommended [26]. In Austria, UMRs' entitlement to various kinds of benefits depends on the respective residence status [28]. The final system model presented also illustrates the connection between *residence security* and *income security*, which is related to various other risk factors. On an international scale, one study found a significant association between general hassles such as economic hardship and depression [37,38]. In addition, based on a quantitative survey, Reference [39] found an association between household income and mental health. In Austria, access to employment is granted only to UMRs with a residence title. Generally, to take up employment, UMRs must have reached the age of 15 and completed nine years of compulsory schooling [28]. However, several factors mentioned in the literature are not shown in the system model, e.g., the direct relation of income to social contacts. The model depicts the impact of *access to education and training in the host country* on *income security* but does not include the relationship in the other direction.
The factor *social contacts and relationships in the host country* has been considered a very relevant post-migration risk factor due to its powerful impact on UMRs' mental health and its numerous direct and indirect connections to other factors, as well as due to the fact that it is part of the previously described triad and cluster. The state of research confirms the influence of social contacts and relationships on mental health. Reference [40] found that the number of friends is related to reduced internalizing behavior problems, though no influence on mental health was found in another study, which also found that being exposed to bullying or marital discord were factors that did not influence mental health [41]. As a result of a longitudinal study of unaccompanied refugees in Norway, Reference [8] stated that the level of daily, general difficulties related to family, friends, and school/work is an independent predictor of depressive symptoms. Further studies found that different daily stressors are associated with higher anxiety, depression, or PTSD symptoms [9,37].
In this regard, the *political and social climate in the host country* has a strong influence on social contacts and relationships. Our assumptions are in line with studies from Sweden and Denmark [13,29] that found an association between the experience of discrimination and lower rates of social acceptance, poorer peer relations, and mental health problems among young refugees. Several international studies confirm the influence of *sociocultural adaptation* on the mental health of UMRs. The study of [42] reported that depression decreases when the level of cultural competence increases. Perceived lower levels of discrimination mediated this effect. The same study, and others, also found an association between acculturation-specific difficulties and depression [37,42]. These findings confirm the stated reciprocal relations between *sociocultural adaptation* and *social contacts and relations in the host country*. The acculturation process should include the development of intercultural competence, describing the ability to effectively communicate and interact with people from various cultures based on one's intercultural knowledge, skills, and attitudes [43].
*German language skills* are a central prerequisite for broad social and societal participation, which are reflected in the final system model as part of the triad with *social contacts and relationships in the host country* and *sociocultural adaptation*. *Sociocultural adaptation* and *German language skills* are factors with reciprocal relations that are relevant in terms of their direct and indirect influences on various factors in the final model. The connections between the language skills and mental health of UMRs have hardly been researched so far. On an international level, only one study, carried out in Denmark, showed a positive effect of self-assessed Danish spelling competency on internalizing behavior [40,44]. In the final system model, language skills gain great importance because of the many indirect relationships and chains of action associated with them. The factor *German language skills* is not only part of the triad of factors at the individual level but is also linked to numerous factors at the societal level. UMRs attending school have, for the duration of the extraordinary status (a maximum of two years), the opportunity to participate in a language course of 11 h per week [14]. German courses for children
that are not compulsory for schooling are often held only once or twice a week, which affects UMRs' acquisition of language skills and hinders rapid integration processes [31].
As illustrated in the system model by the factor *German language skills*, these multifactorial connections can cause the emergence of vicious circles that follow their own dynamics and can thus further endanger, for example, the mental health, social integration, and well-being of UMRs. For example, a good knowledge of German is essential for building *social contacts and relationships in the host country* and facilitating *sociocultural adaptation*. At the same time, *social contacts and relationships in the host country* are relevant for language training (*German language skills*) and for developing the necessary understanding for *sociocultural adaptation*.
#### **5. Limitations**
While there is evidence of the higher prevalence of mental health problems and psychiatric disorders among UMRs (cf., [3]), the results regarding the influence of pre- and post-migration factors on UMRs' mental health are mixed. In this study, we adapted system modeling techniques as an exploratory tool for post-migration risk factors impacting the mental health of unaccompanied minor refugees in Austria. Our system model is intended to contribute to a better understanding of the multi-layered interdependencies between individual and societal factors in refugee-receiving countries, as we showed in the case of Austria.
However, our research is subject to some limitations that must be pointed out. Regarding the connections between post-migration factors, the final system model displays only connections with a value of 5.0 or higher to provide a clearer picture of strong interrelations between the main risk factors of the current study. However, lower weighted connections might be relevant for understanding the complex system of post-migration factors and should be considered when one is planning adequate support structures for UMRs. This involves an effect of *German language skills* on *professional health care services*, of *residence security* on *professional social care services*, and of *housing situation* on *daily structure and leisure activities*, as well as a reciprocal relationship between *access to education and training in the host country* and *sociocultural adaptation*.
Furthermore, some of the factors identified in workshop 1 and workshop 2 were excluded in the further course of the third workshop. For instance, the factor *contact with family and family remittances* or the factor of *family reunification* were not integrated into the final system model, despite their assumed relevance as an important determinant of children's mental health in workshop 1 and workshop 2 as well as in the literature on family support and family situation [36,42,45]. However, both factors seemed ambiguous to the scientists of workshop 3, as they may exert both a positive and negative impact on the mental health of UMRs, as described in more detail in the results. Moreover, the scientists of workshop 3 considered that these factors could not be clearly classified as a post-migration risk factor. In addition, the factors *sociocultural adaptation* and *social contacts and relationships in the host country* were discussed in the two workshops for the finalization of the model. While these two factors may have a strong influence on resilience and the integration process, their characteristics and strengths are already developing in the country of origin. These examples of factors that have been excluded indicate that a complete separation between pre- and post-migration risk factors cannot ultimately be maintained.
The participants in the scientists' workshop were selected because of their research focus on migration or health, but additional expertise from the fields of psychotherapy, biopsychosocial health, or the legal field could be involved in future studies in order to further expand the system model. We also could have invited practitioners to the third workshop to integrate the perspectives of scientists and practitioners as the final step of model development. These changes in the composition of the workshop participants also could have led to slightly different results. Moreover, we also could have invited UMRs or their families to one of the three workshops to broaden the perspectives of scientists and practitioners through the voices of UMRs. The process of involving refugees in system modeling, however, requires additional sensitivity regarding the way their knowledge based on experience is
drawn out, for example, to avoid emotional stress. In this study, we focused on the group of scientists and decided to expand their knowledge elements with those of practitioners. The inclusion of UMRs' perspective in a further step of system development could be an outlook for a further study.
Overall, the results of the system model depend on the expertise and knowledge of the workshop participants as well as on group-dynamic factors inherent in the discussion process, and additional expertise and workshops could further enrich the results. An objective, generally valid system model for the risk factors of UMR after migration can hardly be created in three workshops. Knowledge-based system modeling represents an explorative approach that aims to integrate the knowledge of the experts involved in system modeling. Thus, it should be emphasized that the figures and tables presented to illustrate the results could have created a sense of objectivity and generalizability, but this was not our intention. We would like to emphasize that system modeling, as we have done, is an exploratory approach, but it has nevertheless produced interesting results. Therefore, the model must still be checked empirically, and the results must be compared more deeply to the existing literature and empirical findings from similar research so the model can be adopted.
Moreover, gender differences have not been considered in the workshop discussions. In fact, female UMRs appear in small sample sizes or are even not considered in research studies on mental health and associated risk factors. However, research suggests that gender might affect the nature of traumatic events experienced during flight, as well as the challenges faced on arrival in terms of adequate accommodation or sociocultural adaptation [4,8]. The integration of gender aspects would be a further step in refining the model.
Finally, it must be noted that UMRs were defined as "children and adolescents under 18 years of age migrating into another country without a parent or caregiver". Consequently, the final system model focuses only on UMRs under the age of 18. Since living conditions and received support change when UMRs reach the age of majority, there is a need for a separate model for refugees over 18. In general, UMRs' transition to adulthood usually requires relocation to accommodation for adults and the loss of guardians and key relationships. Nationwide measures or procedures specifically designed to address refugees' needs before, during, or after their transition to adulthood are not in place in Austria [28].
#### **6. Conclusions**
Overall, our final system model of post-migration factors provides a sound basis for further investigations. The consideration of complex multifactorial dependencies illustrates, on the one hand, the limitations of this model. On the other hand, it indicates the benefit of this approach in terms of generating further questions and identifying research needs. The method this paper describes can be used in similar settings in which specialized knowledge from different fields of expertise is needed to explain and understand complex interrelations. By showing the relationships between factors through the final system model, it is possible to put diverging results into a concrete context, to show further influencing factors, and, thus, to uncover starting points for further questions and research needs. While in a standard focus group setting, researchers may develop a system model based on the experts' views, a collaborative approach toward fuzzy-logic cognitive mapping allows participants to establish their model in the meeting, reflect their current beliefs and assumptions, and run scenarios to evaluate the completeness of their prior beliefs and knowledge [21]. Applying the FCM can also be fruitful at a later implementation stage of the model, as inter-professional collaboration and cooperation among different institutions is required to enhance UMRs' mental health. Further in-depth analysis may shed some light on the relevance of the conjunction of the systemic framework and individual characteristics and resources in relation to pre-, peri-, and post-migration factors, e.g., personal and family background, socio-economic conditions in the country of origin, and migration experiences.
Nevertheless, the system model shows interdependencies that, so far, empirical research has omitted. As next steps, we encourage researchers to empirically test the relationships described in the system model of post-migration factors. For example, an analysis of the triad between *social contacts and relationships in the host country*, *German language skills*, and *sociocultural adaptation* should reveal valuable insights into the reinforcing effect of these factors on UMRs' mental health. As discussed previously, *sociocultural adaptation* contains characteristics of both pre- and post-migration factors. In the model, this factor is embedded in reciprocal relationships and exerts strong direct and indirect effects on other factors. Therefore, we propose exploring the post-migration determinants of sociocultural adaptation skills, i.e., to explore those dimensions of social adaptation skills that may develop in the host country. Further, we recommend investigating the relationships between the factors influencing *daily structure and leisure activities*, which has a strong effect on UMRs' mental health in our model. To the best of our knowledge, empirical evidence of the importance of this factor, as well as its interplay with other preand post-migration factors, is scarce.
At the practice level, the system model can be used as a basic framework for determining preventive strategies as well as strategies to restore or enhance the mental health of UMRs. The model underpins the necessity of a holistic approach to address UMRs' mental health problems. Providing easy access to psychotherapeutic support, for example, represents an important protective factor. However, implemented as a single measure, it might not be as effective as a comprehensive approach dedicated to integrating systemic factors such as health and social services, and housing, with the strengthening of individual resources such as social contacts and relationships and the structuring of daily and leisure activities. Refining the model would allow for the definition of institutional interfaces between health and social care providers and for the planning of collaborative action to provide seamless and comprehensive care and support. Given this paper's methodological emphasis, the required in-depth substantive dialogue on this issue will take place in a subsequent publication.
**Author Contributions:** Conceptualization, N.H., A.F. (Arleta Franczukowska), G.S., and L.Z.; data curation, formal analysis, investigation, methodology, validation, visualization, N.H., G.S., and L.Z.; funding acquisition, A.F. (Anna Faustmann), L.R., E.K., I.S. (Isolde Sommer), and L.Z.; project administration, A.F. (Anna Faustmann), E.K., I.S. (Isolde Sommer), and L.Z.; resources, L.R., A.F. (Arleta Franczukowska), N.H., G.S., and L.Z.; software, N.H. and L.Z.; supervision, N.H., A.F. (Arleta Franczukowska), and L.Z.; writing—original draft, N.H. and A.F. (Arleta Franczukowska); writing—review and editing, L.R., G.S., E.K., I.S. (Isabella Skrivanek), A.F. (Anna Faustmann), I.S. (Isolde Sommer), and L.Z. All authors have read and agreed to the published version of the manuscript.
**Funding:** The project was funded by the Danube University Krems.
**Acknowledgments:** Open Access Funding by the University for Continuing Education Krems.
**Conflicts of Interest:** The authors declare no conflict of interest.
| doab | 2025-04-07T04:13:02.982426 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 225
} |
ffa2a234-da32-4c59-b516-55230ec92248.227 | **Asylum Seekers and Swiss Nationals with Low-Acuity Complaints: Disparities in the Perceived level of Urgency, Health Literacy and Ability to Communicate—A Cross-Sectional Survey at a Tertiary Emergency Department**
Received: 13 March 2020; Accepted: 16 April 2020; Published: 17 April 2020
**Abstract:** *Background:* Emergency departments (EDs) are being increasingly used for low-acuity conditions and as primary care providers. Research indicates that patients with the status of asylum seeker (AS) may be seeking care in EDs at higher levels than nationals. The aim of this study was to identify disparities in the use of emergency care between AS and Swiss nationals (SN) with non-urgent complaints. *Methods:* Data were obtained from a survey in the period 01/12/2016–31/07/2017 of walk-in low-acuity patients attending the ED of the University Hospital Bern (Switzerland). AS and a gender, age-matched control group of SN of ≥16 years of age were included. Sociodemographic and survey data comprised information about health-seeking behavior in the home and reception country, knowledge of health care systems (HCSs), barriers to care and perceived acuity of the visit. Furthermore, attending physicians assessed the level of urgency of each case. *Results:* Among AS patients, 30.2% reported that they had no knowledge of the Swiss HCS. In total, 14.2% considered that their medical needs were non-urgent. On the other hand, 43.4% of the attending physicians in the ER considered that the medical needs were non-urgent. This contrast was less pronounced in SN patients. The majority of AS (63.2%) and SN (67.6%) patients sought care from the ED without first contacting a GP. In 53.8% of cases, an interpreter was needed during the ED consultation. *Conclusions:* Several factors associated with health-seeking behavior in the ED differed between AS and SN patients. Measures to increase health literacy, provision of easily accessible primary care services and intercultural-trained staff could improve quality of care and reduce the usage of EDs as primary care providers.
**Keywords:** health-seeking behavior; access to health care; emergency department; refugee; asylum seeker; non-urgent complaints; migrants
| doab | 2025-04-07T04:13:02.984769 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 227
} |
ffa2a234-da32-4c59-b516-55230ec92248.228 | **1. Introduction**
In recent years, there has been an unprecedented increase in the numbers of individuals experiencing forced migration, with many seeking refuge in countries throughout Europe [1].
**Karsten Klingberg <sup>1</sup> , Adrian Stoller <sup>1</sup> , Martin Müller <sup>1</sup> , Sabrina Jegerlehner 1,2 , Adam D. Brown <sup>3</sup> , Aristomenis Exadaktylos <sup>1</sup> , Anne Jachmann 1,**† **and David Srivastava 1,**†**, \***
In particular, Switzerland has witnessed a sharp rise in persons seeking asylum [2]. The net population growth from AS and refugees in countries such as Switzerland has important public health implications since there will be a greater health care demand [3]. Research on hospital-based emergency department (ED) utilization in Norway has shown that immigrants use emergency health care services significantly more often than nationals [4]. These findings are consistent with other studies in Europe that show that recent immigrants are more likely than local nationals to seek care from EDs and out-of-hours GP services [5–8]. During the asylum process in Switzerland, every individual is granted with universal health care coverage and a GP-based model of care, which gives the asylum seekers access to a GP including regular consultations in the asylum center. Additional ED visits are possible without prior GP consultation and there are no co-payments necessary.
General trends showing that hospital-based EDs face increasing levels of visits throughout the world [9–13]. This means that EDs are confronting growing pressure to meet the needs of patients with insufficient resources, resulting in a variety of "supply" problems such as overcrowding, boarding, higher morbidity, and staff burn out [9,14].
Patients seeking care for non-acute medical issues appear to make up a large percentage of ED visits, ranging up to 62%, with a mean of 37% [15]. Triage data from a study of North African patients who had recently migrated to Switzerland showed that they were less likely to need highly urgent care [16]. In a recent interview-based study of low-acuity ED patients in Germany, two factors were identified [17]: firstly, patients felt it would be more convenient to present in the ED, as this did not require an appointment and was not restricted to office hours. Secondly, patients believed they would receive more specialized advice. In addition, poor health and socioeconomic status have been shown to be important factors that influence the threshold of ED use for non-urgent complaints [18]. These findings are supported by Kraaijvanger et al., who showed that health concerns, access to medical tests (e.g., X-rays, blood tests, etc.) and convenience are strongly associated with ED visits for non-acute issues [19].
Identifying the underlying factors contributing to consultations for non-urgent complaints could guide stakeholders and policy makers in implementing measures for equal and effective health care—especially for this vulnerable population. To understand and manage the influencing factors that force AS to seek help in the ED for non-urgent complaints will help to improve services and quality of care for those who may be unable to navigate in a new health care system and have to use the ED as an entry point to the health care system. Therefore, the purpose of this study was to conduct interviews with AS and SN patients, in order to understand the different factors that influence consultations in the ED for non-urgent conditions.
#### **2. Materials and Methods**
#### *2.1. Study Design, Setting and Participants*
We conducted a prospective cross-sectional, controlled, single-center study. Data were collected from 01/12/2016 to 31/07/2017 among patients attending the ED of the University Hospital Bern (Inselspital) in Switzerland. The Inselspital is one of the largest hospitals in Switzerland, with a catchment area of 1.8 million people. More than 45,000 patients are treated in the ED each year [20]. Eligible AS were matched for a predefined period with a group of SN as controls. The two groups were matched by sex, age and triage category. The STROBE (STrengthening the Reporting of OBservational studies in Epidemiology) guideline for cross-sectional studies was employed [21].
Walk-in AS patients attending the ED during the study period and of 16 years of age or older were asked to participate in the survey. Their asylum status was defined by the official Swiss identification card ("F": provisionally admitted foreigners, "N": permit for asylum seekers, or "S" people in need of protection). The study was restricted to patients who had no life-threatening or highly urgent problem as defined by the category in the Swiss Emergency Triage Scale (STS) (Range from 1: acute life-threating to 5: non-urgent problem [22]).
The criteria for eligibility in the SN control group included registration with Swiss citizenship, together with the same triage category, age +/−10 years as well as same sex as the matching AS. Efforts were made to obtain a close temporal match, e.g., recruiting a successive control patient shortly after successfully recruiting an AS study patient. The control group was recruited in a predefined, reduced period. Exclusion criteria for both groups were critically ill patients by the STS (STS < 3), the need for expedited diagnostic testing as judged by the attending consultant, transport by ambulance or patient's refusal to participate in the study.
We designed a survey with questions regarding prior and current health-seeking behavior based on current literature and with the help of a psychologist (YB). The survey was available both as a printed and as a protected web-based version on a tablet.
Trained medical students acted as interviewers and conducted the survey in the waiting area or the treatment area without interfering with medical care. The students received an allowance per questionnaire but were not involved in the treatment.
In cases, where communication between the participant and the interviewer was not possible due to language barriers, accompanying persons or professional interpreters were consulted by phone or in person, to ensure that the participant understood the consent form and the survey questions. These interpreters then also facilitated the medical care (not part of the study).
| doab | 2025-04-07T04:13:02.985042 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 228
} |
ffa2a234-da32-4c59-b516-55230ec92248.229 | *2.2. Measures*
The survey consisted of a nineteen-item patient questionnaire and a nine-item physician questionnaire. The patient questionnaire covered questions about the patient's demographics and socioeconomic status (SES), including education, language skills and current employment situation. Furthermore, all participants were asked about their health-seeking behavior, including previous visits to a general practitioner (GP), the perceived level of urgency of the current visit, their motives in seeking help in the ED and their knowledge of the Swiss health care system (HCS). The attending physician assessed the urgency of the consultation and the discharge status.
| doab | 2025-04-07T04:13:02.985510 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 229
} |
ffa2a234-da32-4c59-b516-55230ec92248.230 | *2.3. Statistical Analysis*
The distribution of categorical variables is given with the absolute number and the relative number as a percentage. The distribution of continuous variables, such as age or length of stay, is described as medians with interquartile (IQR) ranges. Unless otherwise stated, the Chi-squared test was used to test for significant differences between the study and control groups for categorical variables and the Mann–Whitney U test was used for continuous variables, as they were not normally distributed. Data were entered using Microsoft Office Excel 2016 for Windows 10 (Version 1805, Microsoft Corporation, Redmond, WA, United States). All statistical analyses were performed with IBM SPSS Statistics Version 25 (Armonk, New York, NY, United States). Statistical significance was considered at a *p* value smaller than 0.05. Graphs were created using Microsoft Office Excel 2016 for Windows 10 (Version 1805, Microsoft Corporation, Redmond, WA, United States).
| doab | 2025-04-07T04:13:02.985571 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 230
} |
ffa2a234-da32-4c59-b516-55230ec92248.231 | *2.4. Compliance with Ethical Standards*
Study participation was voluntary, free of any compensation and individual verbal and written patient consent was obtained before answering the survey. Patient-related information was anonymized prior to analysis. The study was presented to and approved by the regional ethics committee of the Canton of Bern, Switzerland (06.10.2016, KEK-BE: 2016-01662). The study—including data collection and extraction, anonymization, analysis, and storage—was performed in accordance with Swiss law, the standards of the local ethics committee and the Declaration of Helsinki [23].
| doab | 2025-04-07T04:13:02.985658 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 231
} |
ffa2a234-da32-4c59-b516-55230ec92248.232 | **3. Results**
### *3.1. Demographics*
In total, 557 AS patients were admitted to the ED during the study period. In total, 168 patients were excluded because of admission by ambulance services. Another 38 were excluded because of a higher triage category (STS 1 or 2). In total, 351 AS met the inclusion criteria. In total, 237 patients of the eligible collective could not be included due to circumstances in the ED, e.g., study team unavailable. The analysis excluded four individuals who refused consent and four patients who, after giving consent, did not answer the questions during the interview for unknown reasons. Questionnaires with partially missing answers were included in the study. Missing answers are marked accordingly in the results section. A total of 106 AS were included in the analysis. The control study group of 68 SN was subsequently recruited during a restricted period—after matching for age (+/−10 years), gender and STS.
The demographic characteristics of the groups are summarized in Table 1. The AS patients were predominantly from two geographical regions (cumulatively 70.8%), with Eastern Africa accounting for 40.6% and Western Asia for 30.2% of cases, followed by Southern Asia for 17%, Northern Africa for 5.7% and Western Africa for 4.7%. Only one individual originated from Southern Europe. More than half of the AS were unemployed and nearly two-thirds of the patients were males. There was a significant difference (*p* < 0.001) between the groups in work status (unemployed vs. employed/self-employed vs. other (student/retired/housewife/man)).
**Table 1.** Study population demographic characteristics.
The level of education (see Figure 1) was less than 9 years in 41.5% of the AS and 16% reported that they had no formal education at all. The level of the reported formal education was significantly lower in the AS group than in the control group (*p* < 0.001).
*Int. J. Environ. Res. Public Health* **2020**, *17*, x 5 of 11
**Figure 1.** Formal education (years in school), numbers in % (missing answers: asylum seeker (AS) 2, control group (CG) 2). **Figure 1.** Formal education (years in school), numbers in % (missing answers: asylum seeker (AS) 2, control group (CG) 2). **Figure 1.** Formal education (years in school), numbers in % (missing answers: asylum seeker (AS) 2,
#### *3.2. Health Care Knowledge 3.2. Health Care Knowledge* control group (CG) 2).
Almost one-third of the AS (*n* = 32/106; 30.2%) reported that they had had no knowledge of the Swiss HCS, with the greatest unawareness being within 3 months of arrival. Within this group, only four persons (*n* = 4/13; 30.8%) reported that they had received information about the HCS at the reception center. This number rose to 28 persons (*n* = 28/55; 50.9%) among those with a length of stay of between three months and two years. With increasing length of stay in Switzerland, AS patients reported that they had acquired health Almost one-third of the AS (*n* = 32/106; 30.2%) reported that they had had no knowledge of the Swiss HCS, with the greatest unawareness being within 3 months of arrival. Within this group, only four persons (*n* = 4/13; 30.8%) reported that they had received information about the HCS at the reception center. This number rose to 28 persons (*n* = 28/55; 50.9%) among those with a length of stay of between three months and two years. *3.2. Health Care Knowledge* Almost one-third of the AS (*n* = 32/106; 30.2%) reported that they had had no knowledge of the Swiss HCS, with the greatest unawareness being within 3 months of arrival. Within this group, only four persons (*n* = 4/13; 30.8%) reported that they had received information about the HCS at the reception center. This number rose to 28 persons (*n* = 28/55; 50.9%) among those with a length of stay
care knowledge from family and friends. In contrast, knowledge acquired using the provided media (e.g., printed brochures, general advertisement, internet-based information sources) was very low at 4.7% overall. After more than 2 years in Switzerland, 11 persons (*n* = 11/38; 28.9%) still reported no knowledge of the health care system. This was compared to 7.4% of the SN control group. Figure 2 shows an overview of the knowledge and source of information regarding the Swiss With increasing length of stay in Switzerland, AS patients reported that they had acquired health care knowledge from family and friends. In contrast, knowledge acquired using the provided media (e.g., printed brochures, general advertisement, internet-based information sources) was very low—at 4.7% overall. After more than 2 years in Switzerland, 11 persons (*n* = 11/38; 28.9%) still reported no knowledge of the health care system. This was compared to 7.4% of the SN control group. of between three months and two years. With increasing length of stay in Switzerland, AS patients reported that they had acquired health care knowledge from family and friends. In contrast, knowledge acquired using the provided media (e.g., printed brochures, general advertisement, internet-based information sources) was very low at 4.7% overall. After more than 2 years in Switzerland, 11 persons (*n* = 11/38; 28.9%) still reported no
HCS within the AS group. Figure 2 shows an overview of the knowledge and source of information regarding the Swiss HCS within the AS group. knowledge of the health care system. This was compared to 7.4% of the SN control group. Figure 2 shows an overview of the knowledge and source of information regarding the Swiss HCS within the AS group.
**Figure 2.** Health literacy of asylum seekers and source of knowledge of the Swiss health care system (HCS), dependent on the length of stay in Switzerland, total numbers (missing answers: 1). **Figure 2.** Health literacy of asylum seekers and source of knowledge of the Swiss health care system (HCS), dependent on the length of stay in Switzerland, total numbers (missing answers: 1).
#### *3.3. Health Care Utilization 3.3. Health Care Utilization*
In total, 26.4% of the group of AS patients had a GP in their country of origin, which increased to 67.9% in the reception country. Occasional use of the ED in their pre-migration countries (1–3 times a year) was reported by 27.4% of AS patients versus 35.3% of SN patients. In general, there is a significant increase in the usage of primary care services in Switzerland (McNemar Test for GP usage in home country vs. Switzerland (*p* = 0.001) and ED usage (*p* = 0.029)), even though the use of EDs was already more pronounced than GP use in their home countries. However, a high rate of missing responses to these questions warrants special care when interpreting these results. The comparison between the utilization of health care within the pre-migration country and Switzerland is displayed in Figure 3. In total, 26.4% of the group of AS patients had a GP in their country of origin, which increased to 67.9% in the reception country. Occasional use of the ED in their pre-migration countries (1–3 times a year) was reported by 27.4% of AS patients versus 35.3% of SN patients. In general, there is a significant increase in the usage of primary care services in Switzerland (McNemar Test for GP usage in home country vs. Switzerland (*p* = 0.001) and ED usage (*p* = 0.029)), even though the use of EDs was already more pronounced than GP use in their home countries. However, a high rate of missing responses to these questions warrants special care when interpreting these results. The comparison between the utilization of health care within the pre-migration country and Switzerland is displayed in Figure 3.
*Int. J. Environ. Res. Public Health* **2020**, *17*, x 6 of 11
**Figure 3.** Chart of general practitioner (GP) and emergency department (ED) consultations of AS in their home country and in Switzerland, numbers in %. **Figure 3.** Chart of general practitioner (GP) and emergency department (ED) consultations of AS in their home country and in Switzerland, numbers in %.
#### *3.4. Barriers to Care 3.4. Barriers to Care*
Communication between the AS patient and the physician without an interpreter was possible in 45.3% of cases. More than half (53.8%) of the AS patient group were not able to communicate directly with the physician, as they did not speak a national language (e.g., German, French, Italian) or English. In 70.2% of these cases, accompanying persons, such as family and friends, acted as interpreters. Communication between the AS patient and the physician without an interpreter was possible in 45.3% of cases. More than half (53.8%) of the AS patient group were not able to communicate directly with the physician, as they did not speak a national language (e.g., German, French, Italian) or English. In 70.2% of these cases, accompanying persons, such as family and friends, acted as interpreters.
Approximately one-third (35.8%) of the AS patients had tried to consult the physician in their asylum center (15.1%) or their attending GP (20.8%) before consulting the ED. Nevertheless, the majority of AS patients (63.2%) and of the control group (67.6%) sought care from the ED without first contacting a GP or a physician from their asylum center. The reasons given for these decisions are shown in Table 2. The most common response (29.9%) by the AS patients was not having a GP, followed by consultations outside GP opening hours (25.4%). In total, 19.4% of the AS stated previous poor experiences with GP services or expected better care at the ED. In contrast, SN only half as often (13%) reported that they had no GP and the most common reason given (32.6%) was that they had directly consulted the ED because their visit was outside GP opening hours. Approximately one-third (35.8%) of the AS patients had tried to consult the physician in their asylum center (15.1%) or their attending GP (20.8%) before consulting the ED. Nevertheless, the majority of AS patients (63.2%) and of the control group (67.6%) sought care from the ED without first contacting a GP or a physician from their asylum center. The reasons given for these decisions are shown in Table 2. The most common response (29.9%) by the AS patients was not having a GP, followed by consultations outside GP opening hours (25.4%). In total, 19.4% of the AS stated previous poor experiences with GP services or expected better care at the ED. In contrast, SN only half as often (13%) reported that they had no GP and the most common reason given (32.6%) was that they had directly consulted the ED because their visit was outside GP opening hours.
**Table 2.** Reasons why patients did not try to or use GP before presenting to the emergency department. (Patients with direct ED consultation N-AS: 67; N-CG: 46; multiple answers possible.) **Reasons Study Group** *n* **(%) Control Group** *n* **(%)** No GP 20 (29.9) 6 (13) There were no significant differences (*p* = 0.223) between AS patients and SN with respect to seeking care during working hours (08:00–18:00, weekdays Monday–Friday). It is striking that more than half of the consultations in both AS and SN patients took place within general working hours. The length of stay in the ED did not differ significantly between the two groups (*p* = 0.141)—the median LOS of the SN control group was 3:22 h (IQR: 2:40–5:25 h) and the median LOS of the AS group was 3:09 h (IQR: 2:01–5:00 h)—neither did the respective proportions of patients receiving inpatient versus outpatient treatment (*p* = 0.892).
**Table 2.** Reasons why patients did not try to or use GP before presenting to the emergency department. (Patients with direct ED consultation N-AS: 67; N-CG: 46; multiple answers possible.) Expected better treatment in ED 13 (19.4) 7 (15.2) Highly urgent problem 15 (22.4) 10 (21.7)
Previous bad experience 13 (19.4) 5 (10.9)
*Int. J. Environ. Res. Public Health* **2020**, *17*, x 7 of 11
#### *3.5. Patient Perceptions of Medical Urgency* Over one-third (34.9%) of the AS patients perceived a need for treatment within one hour,
Over one-third (34.9%) of the AS patients perceived a need for treatment within one hour, whereas the attending physician assigned this level of urgency to only 11.3% of cases. In addition, 14.2% of the enrolled AS patients regarded their problem as non-urgent and this assessment was shared by the attending ED physician in 43.4% of these cases. whereas the attending physician assigned this level of urgency to only 11.3% of cases. In addition, 14.2% of the enrolled AS patients regarded their problem as non-urgent and this assessment was shared by the attending ED physician in 43.4% of these cases. In total, 22.1% of the SN patients perceived a need for treatment within one hour, whereas the attending physician assigned this level of urgency to 29.4% of cases. In addition, 19.1% of the enrolled
In total, 22.1% of the SN patients perceived a need for treatment within one hour, whereas the attending physician assigned this level of urgency to 29.4% of cases. In addition, 19.1% of the enrolled SN patients regarded their problem as non-urgent and this assessment was shared by the attending ED physician in 27.9% of these cases. The estimated level of urgency from the patient's perspective compared to the physician's perspective is displayed in Figure 4. SN patients regarded their problem as non-urgent and this assessment was shared by the attending ED physician in 27.9% of these cases. The estimated level of urgency from the patient's perspective compared to the physician's perspective is displayed in Figure 4. The subjective level of urgency by patients and physicians did not influence the initially allocated triage level. All patients included in this study were categorized in STS 3–5 prior to consultation and the item "subjective level" was used only to display the perceived level of urgency.
**4. Discussion** AS and SN differed in their reasons for seeking care in the ED, their knowledge of the Swiss HCS, and their perceptions of medical urgency. However, other factors, such as length of stay, The subjective level of urgency by patients and physicians did not influence the initially allocated triage level. All patients included in this study were categorized in STS 3–5 prior to consultation and the item "subjective level" was used only to display the perceived level of urgency.
discharge type, and time of visit did not differ between the two groups.
#### **4. Discussion**
AS and SN differed in their reasons for seeking care in the ED, their knowledge of the Swiss HCS, and their perceptions of medical urgency. However, other factors, such as length of stay, discharge type, and time of visit did not differ between the two groups.
Although we are unable to determine from this study whether these findings are representative of AS patients generally, the differences observed between the AS and SN patients may shed light on the observed high usage of the ED by AS patients in recent years (e.g., Müller et al. 2016 [3]). These findings suggest that the factors associated with seeking care in the ED among AS patients are multifactorial and reflect potential gaps in health care knowledge, linguistic barriers, and perceptions of acuity and care. The differences in health care utilization, represented by the use of PC and ED care in the home country versus Switzerland, might not be due to differences in the individual health-seeking behavior but represent effects of the different national health care systems and the availability of primary and emergency care.
Attempts to help guide and inform health-seeking behavior by AS patients will benefit from improvements in a range of social and cultural factors that influence the dissemination of health care information, coupled with training for medical staff working with the AS community.
Firstly, AS patients' lower levels of knowledge of the Swiss HCS could be due to a lack of education and general health care experience, as the usage of primary care services in their home countries was significantly lower than in Switzerland. Furthermore, information about health care appears to be gradually acquired through informal networks such as friends and family members. Such findings suggest that policy makers and those working in health care promotion may be able to reduce non-urgent visits through public health campaigns such as peer group interventions [24].
A second key factor contributing to ED use among AS patients may be actual or anticipated poor care from GPs. AS patients were more likely to say that they had had a poor experience with a GP or expected to receive better care in the ED. These findings warrant further investigation, as the specific nature of the poor experiences and the perceptions around quality of care cannot be determined from this study. A combination of public health campaigns and culturally informed training for GPs may help to improve perceptions and actual experiences of care during GP visits [25].
A third difference between AS and SN patients was around perceptions of urgency. Even though the majority of both groups sought care in the ED without prior consultation by a GP, the perceived level of urgency among SN was closer to the assessment of the attending physicians. One reason that AS patients may be more likely to seek care in the ED is that they believe their symptoms are of higher acuity. Education of AS patients on how to identify acute symptoms and better access to primary care for non-urgent complaints may help to improve quality of care [26].
More generally, these findings provide further information about the major gaps in interpretation in health care contexts for AS patients. As these data show, a majority of AS patients are not able to communicate easily with their medical providers, and often rely on family members and even children to broker these interactions. The reliance on untrained interpreters may be associated with miscommunication, misdiagnosis, and poor outcomes and includes safety concerns in relation with human trafficking [27–29].
The level of education reported by the AS group was significantly lower than in the SN group. Furthermore, the work status differed significantly, as most AS patients were unemployed, but most SN were employed or self-employed. Low socioeconomic status is linked to the overuse of ED care [18]. This result was reproduced here but we did not investigate the nature of the association.
The differences identified between the AS and SN may help to address the increase in non-urgent visits to hospital-based EDs [3,5]. The findings suggest several different approaches to improving the access of this vulnerable population to adequate and equal care.
#### **5. Limitations**
This survey was restricted to AS and SN patients with non-urgent problems and in the context of an urban university hospital in a single-center setting. There are therefore several different selection biases. Furthermore, this study does not cover an entire year, so there might be the risk of seasonality in data collection. Additionally, there were a number of randomly missing answers, which limits the interpretation of the results. The strong skew towards young males in the study group proved to be hard to match in the control group. A large peak at the outmost higher end of the age group admitted to the adult ED department was compounded by the permitted age variation of +/−10 years. The average
age of ED patients in Switzerland is above 50 years [30]. One limitation is the self-reported answers, which may be positively or negatively biased, according to the participant's impression that answers could influence their treatment in the ED or even the asylum process even though all participants were informed in advance that the study would have no influence on their treatment or asylum process.
Another important limitation is the bias due to the German language of the questionnaires. The interviewer decided in case of little or no understanding to involve a translator either in person or mostly by phone. The same was used to obtain consent.
The difference between the use of PC and ED in the home country and Switzerland does not necessarily represent a difference in health-seeking behavior but might be due to different structures with different availabilities in the national HCS.
We did not control for big differences such as education and work status and this may act as a confounder for the measured outcomes in this study.
Despite these limitations, we believe the data help to shed important light on the experiences of AS patients in the HCS. The findings are relevant for all stakeholders involved in clinical care and health care policy and can encourage them to develop and implement new strategies to fill the demonstrated gaps in health care knowledge and improve quality of care.
| doab | 2025-04-07T04:13:02.985729 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 232
} |
ffa2a234-da32-4c59-b516-55230ec92248.233 | **6. Conclusions**
Disparities in knowledge of the HCS in the reception country, language barriers, and the perceived level of urgency of medical care seem to be the main reasons for AS to seek care in ED for low-acuity medical issues. In both groups, the decision to present to the ED was influenced by the unlimited access over 24 h, expectation of better treatment in the ED and the perceived level of urgency.
Measures to increase health literacy and provision of easily accessible primary care could improve quality of care and reduce the usage of EDs as primary care providers to AS. Implementation and usage of a professional interpreting service will relieve family and friends from this role and might provide better and equal care.
**Author Contributions:** Conceptualization, K.K., A.S., S.J., A.J. and D.S.; methodology, K.K., M.M., A.J. and D.S.; validation, M.M. and A.J.; formal analysis A.S., M.M. and A.J.; investigation, K.K., A.S. and M.M.; resources, A.E..; data curation, K.K., A.S., A.J.; writing—original draft preparation, K.K., A.S., A.B., A.J. and D.S.; writing—review and editing, all authors; visualization, K.K., A.S. and A.J.; supervision, A.J. and D.S.; project administration, K.K. and D.S.; funding acquisition, A.D.B., A.E. and D.S. All authors have read and agreed to the published version of the manuscript
**Funding:** The study was partly funded by the Swiss Federal Office of Public Health (FOPH), Nr. 16.925522 and a Fulbright Specialist Grant awarded to AB.
**Acknowledgments:** We thank Yvonne Bogenstätter (YB) for the design of the questionnaire, and Nicole Frischknecht, Jolanta Klukowska-Rötzler, Romana Saredi, and Sabina Utiger for their administrative and technical support.
**Conflicts of Interest:** The authors declare no conflict of interest.
| doab | 2025-04-07T04:13:02.986821 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 233
} |
ffa2a234-da32-4c59-b516-55230ec92248.235 | *Article* **The Relationship Exploration between Public Migration Attention and Population Migration from a Perspective of Search Query**
#### **Chun Li <sup>1</sup> , Jianhua He <sup>2</sup> and Xingwu Duan 1,\***
Received: 8 March 2020; Accepted: 30 March 2020; Published: 1 April 2020
**Abstract:** Rapid population migration has been viewed as a critical factor impacting urban network construction and regional sustainable development. The supervision and analysis of population migration are necessary for guiding the optimal allocation of urban resources and for attaining the high efficiency development of region. Currently, the explorations of population migration are often restricted by the limitation of data. In the information era, search engines widely collect public attention, implying potential individual actions, and freely provide open, timelier, and large-scope search query data for helping explore regional phenomena and problems. In this paper, we endeavor to explore the possibility of adopting such data to depict population migration. Based on the search query from Baidu search engine, three migration attention indexes (MAIs) are constructed to capture public migration attention in cyber space. Taking three major urban agglomerations in China as case study, we conduct the correlation analysis among the cyber MAIs and population migration in geographical space. Results have shown that external-MAI and local-MAI can positively reflect the population migration inner regions and across regions from a holistic lens and that intercity-MAI can be a helpful supplement for the delineation of specific population flow. Along with the accumulation of cyber search query data, its potential in exploring population migration can be further reinforced.
**Keywords:** population migration; search query; Baidu Index; urban agglomeration
| doab | 2025-04-07T04:13:02.986978 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 235
} |
ffa2a234-da32-4c59-b516-55230ec92248.236 | **1. Introduction**
Along with the rise of a city network, which is constructed under the push of different kinds of urban elements flows, the interactions among different cities have been emphasized in the planning of urban areas, including the interaction of population, material, information, technique, etc. Hereinto, population interaction or population migration is one of the most important aspects. The floating of population is not only the flowing of individual human but also the transfer of demand, information, and technique carried by individuals [1,2]. They discriminately impact economic, social, and political development of both resettled areas and out-migrating areas [3,4]. Timely measuring and analyzing of population migration are particularly crucial for suitably planning urban space and distributing urban resources.
Related explorations on population migration have been concerned as hotspots since the 1990s. A larger body of researches have been conducted, such as the labor market performance, social and physical status of migration [5–7], the causes of migration flow [8–10], the consequent impacts of migration [11–13], the changing migration policies [14–16], the classification research of population migration [17,18], the spatial pattern of population migration [19], etc. These researches have been
conducted mainly based on three kinds of data: national censuses data, regional field survey data, and cyber big data. In the traditional migration researches, population censuses and field survey are the principal sources to provide population data [20,21]. For instance, Zhu [22] explored the determined factors in urban area which influence migrants' settlement intention based on the data from a survey on the floating population in the coastal area of Fujian Province. He et al. [23] adopted national census data to examine the distinctive spatial patterns of floating and Hukou population and evaluated their consequent impact on Chinese urbanization and industrialization. With the development of cyber space and the popularization of personal mobile termination, numerous researches have implemented under the assistance of data from cyber space exploration of the change, characteristic, and pattern of population migration [24–27]. For instance, Blumenstock [28] analyzed migration pattern based on mobile phone records and revealed more subtle patterns that were not detected in the government population survey. Zagheni et al. [29] used geolocated data for about 500,000 users of the social network website "Twitter" to predict turning points in migration trends and to improve the understanding of migrant populations.
Those researches have contributed largely to promoting the understanding of the progress of population migration and their impact. However, the deficiencies in migration data still exist. Studies based on national censuses data can explore the migrants in a large range but with a relatively large time interval of ten years, which hinders the short time-series analysis of population migration, and little can be inferred for specific years between censuses and for recent trends [29]. The researches based on field survey can provide detailed migration information, but it asks for a lot of time, manpower, and material resource to deploy, which are expensive for many researches. Simultaneously, the field survey often has a certain spatial location and cannot cover a large spatial scope. The increasing cyber data has opened up a new opportunity to deepen our understanding of population migration. However, studies based on the network big data always need to deal with extensive data and complicated procedures in acquiring and processing the data. At the same time, some data sources are not available openly, such as cellphone signal data and GPS data of resident activities, because those types of data include much individual private information that is protected by national law. A type of data with open, timelier, and easy-taking characteristics is necessary for effectively investigating the migration population.
With the growing application of search engine in cyber space, search query data has been brought out to reflect the preference of public attention, which is generated from the personal behavior of Internet search. This kind of data with opening and timelier characteristic has provided effective support for analyzing regional phenomena and problems [30–33]. In such context, the concern is triggered about its applicability in population migration research. In current information era, most people tend to take migration after an inquiry of destinations. Web search engine as the most widely used Internet tools provides massive information to the migrants and obtains relevant public attention on the specific subject of migration. The relationship between Internet search query data and population migration deserves more attention. However, the relationship between them is still unclear and there are a number of questions to be raised: can the search query data generated from migration-related information search offer some clues about population migration? If they can, how are they related? Do cites with higher cyber search quantity have a larger migration population than the cities with lower search quantity?
Based on these questions, this paper endeavors to answer them and to propose a new angle to analyze population migration. A hypothesis can be made that the search queries generated from individual migration-related search can positively reflect population migration. Based on the search query data from Baidu search engine, we construct a series of migration attention indexes (MAIs) to explore public attention on migration. Taking three main urban agglomeration areas of China as study area, the correlation analysis has been utilized to explore the relationship between MAIs and population migration to test the hypothesis. This paper is organized as follows. Section 2 introduces the study area and data. Section 3 elucidates the methodology of this paper, including the method and indicators that we applied in this paper. Section 4 reports the result of correlation analysis between
MAI and population migration. Section 5 conducts further discussion based on the results in our study area. Last, we conduct the conclusion of this paper.
| doab | 2025-04-07T04:13:02.987136 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 236
} |
ffa2a234-da32-4c59-b516-55230ec92248.238 | *2.1. Study Area*
*2.1. Study Area*
To verify the relationship between search query data in cyber space and population migration in geographical space, we select three urban agglomerations in China as case study: Beijing-Tianjin-Hebei metropolitan region (BTH), the Yangtze River Delta (YRD), and the Pearl River Delta (PRD). There are 38 cities located in these regions, 10 cities from BTH, 16 from YRD, and 9 from PRD, as shown in Figure 1. These regions are chosen based on the following reasons: (1) Extensive population migration can be detected in these areas. In 2015, the migration population in these areas has reached more than 8 million in total, accounting for 30.77% in China. Exploration of migration in these regions can avoid the influence of random migration under the support of large quantities. (2) These regions with relatively higher internet penetration offer adequate search query data. By the end of 2015, internet penetrations of core cities in those three urban metropolitan areas are separately 76.5% for Beijing, 73.1% for Shanghai, 78.4% for Guangzhou, and 83.2% for Shenzhen. More widespread application of the internet can be identified in almost all the provinces cover BTH, YRD, and PRD [34]. (3) They are the most significant areas for China's urban system construction. These areas occupy approximately 5.09% land of China but account for 23.65% and 39.87% of national permanent residents and gross domestic product in 2015. Research on them can offer more information to guide the coordinated development of urban areas in China. *Int. J. Environ. Res. Public Health* **2020**, *17*, x 3 of 18 **2. Study Area and Data**
**Figure 1.** Location of the study area. **Figure 1.** Location of the study area.
To verify the relationship between search query data in cyber space and population migration
shown in Figure 1. These regions are chosen based on the following reasons: (1) Extensive population migration can be detected in these areas. In 2015, the migration population in these areas has reached more than 8 million in total, accounting for 30.77% in China. Exploration of migration in these regions can avoid the influence of random migration under the support of large quantities. (2) These regions with relatively higher internet penetration offer adequate search query data. By the end of 2015, internet penetrations of core cities in those three urban metropolitan areas are separately 76.5% for Beijing, 73.1% for Shanghai, 78.4% for Guangzhou, and 83.2% for Shenzhen. More widespread application of the internet can be identified in almost all the provinces cover BTH, YRD, and PRD [34]. (3) They are the most significant areas for China's urban system construction. These areas occupy approximately 5.09% land of China but account for 23.65% and 39.87% of national permanent residents and gross domestic product in 2015. Research on them can offer more information to guide
The data used in this paper include the population migration data, search query data, and socioeconomic data. (1) There are three kinds of population migration data used in this study: the net inflow population, intercity population flow, and the floating population. The net inflow population
the coordinated development of urban areas in China.
*2.2. Data*
| doab | 2025-04-07T04:13:02.987633 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 238
} |
ffa2a234-da32-4c59-b516-55230ec92248.239 | *2.2. Data*
The data used in this paper include the population migration data, search query data, and socioeconomic data. (1) There are three kinds of population migration data used in this study: the net inflow population, intercity population flow, and the floating population. The net inflow population delineates the total population migrated into the city during a specific period. Intercity population flow is the population migrate from the original city to the terminal city. Based on the prevalent use of series Tencent's applications (e.g., Wechat is the most used software for 79.6% of Chinese netizens), more precise expression on the migration of population in China can be provided by Tencent migration map under the support of enormous user base. Considering the merit of Tencent migration map and avoiding the self-certification of Baidu, we obtained the net inflow population and intercity population flow from Tencent migration map (https://heat.qq.com/qianxi.php) through web crawler technology. Due to the specific Hukou policy in China (which has been regarded as the central mechanism underlying the unsettled nature of the floating population), the floating population has been defined as the population living in the objective city more than six months without local registered Hukou [35]. It was obtained through the deviance calculation of permanent residential population and household population in the local city, which were collected from regional statistical bureaus. (2) We obtain the search query data based on the support of Baidu search engine, which is the most widely used search engine in China and freely provides the search trend of objective terms through Baidu Index (http://index.baidu.com/). The average daily queries of each migration keyword versus the name of the city (e.g., job + Beijing) from 1 January 2015 to 31 December 2015 were collected based on Baidu Index. (3) Relevant socioeconomic data were acquired from regional statistical bureau, including the Tertiary Industrial Output-Value, Participant Rate of Urban Basic Medical Care System, the number of schools, etc. The migration reasons were collected from the dynamic monitoring survey of China's migration population in 2015, which was conducted by the National Health and Family Planning Commission of China.
| doab | 2025-04-07T04:13:02.987834 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 239
} |
ffa2a234-da32-4c59-b516-55230ec92248.240 | **3. Methodology**
We endeavor to verify the relationship between public attention on migration which was provided by search query in cyber space and the population migration in geographical space. Migration attention indexes (MAIs) are proposed to express public attention on migration comprehensively. Based on the different original location of migration search, we construct three MAIs as local-MAI, external-MAI, and intercity-MAI to delineate the public attention generated from local city, attention from external areas, and attention flow among urban areas; then, the correlation analysis is conducted between MAIs in cyber space and urban migrants in geographical space to further verify the aforementioned hypothesis. The framework of this paper can be illustrated in Figure 2.
Specially, the net inflow population, intercity population flow, and the floating population have been collectively adopted to depict the movement of population in this paper. The definition of migration for them can be separately clarified as follows: The net inflow population of a city is defined as population that the city has received from the external areas, which is the result of movement of people with different origins and the same destination; the intercity population flow is also defined as the movement of people, which happens among different cities; and the floating population of a city is defined under the Hukou policy of China (which has been regarded as the central mechanism underlying the unsettled nature of the floating population), of which the migration can be explained as the change in the place of personal residence.
*Int. J. Environ. Res. Public Health* **2020**, *17*, x 5 of 18
**Figure 2.** Research framework. **Figure 2.** Research framework.
| doab | 2025-04-07T04:13:02.988005 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 240
} |
ffa2a234-da32-4c59-b516-55230ec92248.241 | *3.1. Migration Attention Index Based on Baidu Search Query 3.1. Migration Attention Index Based on Baidu Search Query*
To verify the hypothesis that the migration-related search queries from individual users can positively reflect the population migration, three issues should be concerned: (1) what are the main driving factors cause population migration; (2) how to express those factors in cyber space through search query data; and (3) how to synthesize those search query data to comprehensively express public attention on migration in cyber space. For the first issue, based on the dynamic monitoring survey of China's migration population in 2015, we have conducted the statistic of population percentage on different migration reasons to confirm the main factors which cause population migration. For the second issue, a series of search keywords expressing different migration reasons has been selected. The Baidu Index of keywords versus the name of city has been collected to reflect the public attention on migration in cyber space. For the third issue, migration attention indexes (MAIs) have been constructed to integrate public attentions generated based on different migration To verify the hypothesis that the migration-related search queries from individual users can positively reflect the population migration, three issues should be concerned: (1) what are the main driving factors cause population migration; (2) how to express those factors in cyber space through search query data; and (3) how to synthesize those search query data to comprehensively express public attention on migration in cyber space. For the first issue, based on the dynamic monitoring survey of China's migration population in 2015, we have conducted the statistic of population percentage on different migration reasons to confirm the main factors which cause population migration. For the second issue, a series of search keywords expressing different migration reasons has been selected. The Baidu Index of keywords versus the name of city has been collected to reflect the public attention on migration in cyber space. For the third issue, migration attention indexes (MAIs) have been constructed to integrate public attentions generated based on different migration reasons.
#### reasons. 3.1.1. Confirmation of Main Migration Driving Forces
3.1.1. Confirmation of Main Migration Driving Forces To pointedly select search keywords that load public attention on migration. First, we confirm the main reason for population migration based on the dynamic monitoring survey of China's migration population in 2015. The percentage statistics of migrant population based on diverse migration reasons in the three different urban agglomerations are deployed. The results have been shown in Table 1; we can see that work and trade, that study and training, that accompanying transferring of family members, and that relocation are the main migration factors in the study area. To pointedly select search keywords that load public attention on migration. First, we confirm the main reason for population migration based on the dynamic monitoring survey of China's migration population in 2015. The percentage statistics of migrant population based on diverse migration reasons in the three different urban agglomerations are deployed. The results have been shown in Table 1; we can see that work and trade, that study and training, that accompanying transferring of family members, and that relocation are the main migration factors in the study area. The percentages of population who migrate for the four reasons separately occupy 75.70%, 85.39%, and 89.77% in Beijing-Tianjin-Hebei metropolitan region, the Yangtze River Delta, and the Pearl River Delta.
The percentages of population who migrate for the four reasons separately occupy 75.70%, 85.39%, and 89.77% in Beijing-Tianjin-Hebei metropolitan region, the Yangtze River Delta, and the Pearl River Delta. Due to the transferring of family members always accompanying family relocation [36], we have viewed them as one perspective and marked as relocation. Therefore, three main reasons for population migration have been confirmed as *work and trade*, *study and training*, and *relocation*.
**Table 1.** Migration population percentage of different migration reason in the different urban agglomerations.
#### 3.1.2. Selection of Search Keywords from Baidu Index
To better exhibit and exploit search query data, relevant search exploit services based on search query data are produced, typically as Google Trend (www.google.com/trends/) and Baidu Index (http://index.baidu.com/). A series of researches have been conducted to analyze data from Google Trend and Baidu Index; the robustness and effectiveness of them have been assessed [37–39]. In China, compared to Google, which is the largest search engine in the world, Baidu shares more internet search engine market. In 2016, there are 731 million netizens in China and the number of search engine users has reached 602 million [34]. Hereinto, Baidu shares 77.07% of the Internet search engine market, which is more than Google China. Especially, Vaughan and Chen [40] collected and compared the data from Google and Baidu and found that Baidu Index can offer more search volume data than Google Trend did in China. Under such context, the Baidu Index is employed in this paper to obtain public search attention in the cyber space.
Focusing on the three main migration reasons, we endeavor to confirm the search keywords which reflect public attention on migration. The confirmation of search keywords is comprehensively confirmed under five steps. First, according to the least effort principle in network information retrieval behaviors, users incline to choice the search keywords in their common language with brief and straightforward features [21,36,41–44]. We set the candidate keywords with brief structure and expressed them in Chinese. Second, the specific content of candidate keywords was derived from the three main migration reasons. Relevant search terms for them were selected by brainstorming common words used in searching for migration and review of related literature [21,45–47]. Third, we have compared the daily average search query data of designated search keywords with similar words during the same period to confirm that the selected keywords are the most popular search keywords in the related aspects. For example, "租房 (house renting)" has been compared to "出租 (rent)" and "租赁 (lease)"; collecting and organizing their average daily Baidu Index can find that "house renting"(11,795) gets much more attention than "rent"(477) and "lease"(636). Fourth, we sift the candidate words to follow the principle of search query data for each keyword in each city to be delineated as a sequential time series with a yearly resolution. Fifth, the correlation analysis between the last candidate keywords has been conducted and the one with a high correlation with others has been removed to reduce data redundancy. Through the comprehensive consideration of keyword selection, the last keywords can be viewed as not only representing the meaning itself but also including some clues for other potential keywords. Finally, six Chinese keywords from Baidu index have been confirmed to express public attention on migration in cyber space as list in Table 2.
**Table 2.** Selection of search keywords.
### 3.1.3. Construction of MAIs
The migration attention indexes (MAIs) are designed to comprehensively express public attention on migration in cyber space comprehensively. First, we combine the candidate search keywords with the name of objective cities to obtain the cityward migration keywords, such as "school + Beijing", "house price + Shanghai", "recruitment + Shenzhen", etc.; second, the average daily search volume of these cityward keywords are acquired based on Baidu Index from 1 January 2015 to 31 December 2015; third, the population percentages of different migration reasons are viewed as index weight to synthesize the corresponding Baidu Index into MAIs; fourth, according to the origin location of Baidu Index, the *local-MAI*, *external-MAI*, and *intercity-MAI* are separately constructed to express public migration attention on objective cities from internal area of the objective cities, external areas, and other specific cities. The relationship among those indexes can be depicted as follows:
$$\text{MAI}\_{\text{i}} = \text{External\\_MAI}\_{\text{i}} + \text{Local\\_MAI}\_{\text{i}} \tag{1}$$
$$\text{External\\_MAI}\_i = \sum\_{j=1}^{} \text{interity\\_MAI}\_{ij} \tag{2}$$
where *i* is the objective city, *j* is the original city, *MAI<sup>i</sup>* is the total migration attention city *i* has achieved from all regions, and *local-MAI<sup>i</sup>* and *External-MAI<sup>i</sup>* are separately the total migration attention city *i* has received from the urban internal area and external areas. *Intercity-MAIij* is the public migration attention derived from city *j* to city *i*. The formula of those indexes can be shown as follows:
$$\text{Local\\_MAI}\_i = \sum\_{n=1}^{3} \mathcal{W}\_{\text{in}} \times \text{BI}\_n / \text{MAI}\_{\text{max}} \tag{3}$$
$$\text{External\\_MAI}\_i = \sum\_{j=1} \sum\_{n=1}^3 \mathcal{W}\_{ijn} \times \text{BI}\_n / \text{MAI}\_{\text{max}}, i \neq 1 \tag{4}$$
$$\text{Intercrity\\_MAI}\_{\text{ij}} = \sum\_{n=1}^{3} \mathcal{W}\_{\text{ijn}} \times \text{BI}\_n / \text{MAI}\_{\text{max}} \,\text{i} \neq \text{j} \tag{5}$$
where *BI<sup>n</sup>* is the average daily volume of Baidu Index about different search keywords under migration reason *n*; *Win* and *Wijn* are the weights of *BIn*, which are defined by the proportion of people who migrate into city *i* for this reason; and *MAImax* is the max absolute value of MAI indicators.
#### *3.2. Correlation Analysis between MAIs and Population Migration*
#### 3.2.1. Correlation with Urban Migrants
To investigate the relationship between public migration attentions in cyber space and population migration in geographical space, we conduct the correlation analysis between MAIs and urban migrants. In the cyber space, local-MAI, external-MAI, and intercity-MAI were selected to represent public migration attentions with different originations to objective cities; in geographical space, floating population, inflow population, and intercity population flow were collected. Regarding the diverse kinds of migration and different definition of MAIs, the correlation analysis have been conducted from three aspects: (1) the correlation between local-MAI and floating population, which reflects the
relationship between migration attention generated from the local city and actual floating population inside the city; (2) the correlation analysis between external-MAI and inflow population, which explores the relationship between migration attentions received from the external areas and actual inflow population of the objective city; and (3) the correlation analysis between intercity-MAI and intercity population flow, which investigates the relationship between cyber migration attention flows and the actual population flows in the geographic space. Pearson correlation coefficient is employed to test such correlations, the formula can be shown as follows:
$$r = \frac{1}{n-1} \sum\_{i=1}^{n} \left( \frac{\mathcal{X}\_i - \overline{\mathcal{X}}}{\delta\_X} \right) \left( \frac{\mathcal{Y}\_i - \overline{\mathcal{Y}}}{\delta\_Y} \right) \tag{6}$$
where *r* is the correlation coefficient of the two indexes and *n* is the number of cities.
| doab | 2025-04-07T04:13:02.988128 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 241
} |
ffa2a234-da32-4c59-b516-55230ec92248.242 | 3.2.2. Correlation with Urban Migration Attractiveness
Furthermore, we inquired about the relationship between urban external-MAI in cyber space and urban comprehensive attractiveness for migrants (UAM) in geographical space to further test the validity of the proposed indicators. Based on the push-pull theory which has been widely used in analyzing migration action and willing [48–51], we confirmed the UAM from urban pull perspective. The major migration reasons confirmed by the dynamic monitoring survey of China's migration population have been employed as reference in confirming the objective content of UAM, including work and business, study and training, and relocation. The three aspects separately correspond with the three major migration reasons as job opportunity and income level, living condition, and educational opportunity of children. Based on the data availability principle and integrated analysis of previous studies, eight indicators with respect to three aspects of urban pulling power have been selected as shown in Table 3. From job and income perspectives, Tertiary Industrial Output-Value (TIV) [52] and Urban Residents' Per Capita Disposable Income (IPC) [43] were employed to reflect urban job opportunities and income level; Unemployment Rate (UR), Participant rate of Urban Basic Medical Care System (RBM) [53], and Per Capita Living Area (LPC) [43] were utilized to expose the living condition of local residents; Number of Regular Primary Schools (PSN), Number of Regular Secondary Schools (SSN), and Number of University (UN) were applied to reveal educational opportunity for migrants' children [44].
**Table 3.** Indicator system of urban pulling power.
Note: RMB is the abbreviation of Ren Min Bi (China Yuan), which is the basic monetary unit of China.
Last, we adopted the principal component analysis (PCA) to integrate the index system and to obtain the indicator which reflects urban comprehensive attractiveness for migrants. The components with eigenvalues greater than 1 and the cumulative ratio of total variance greater than 85% are extracted and rotated with the varimax method in SPSS 19.0 (International Business Machines Corporation,
New York, USA), so that each factor has the minimum number of high load variables, which can be expressed as follows: Corporation, New York, USA), so that each factor has the minimum number of high load variables, which can be expressed as follows:
$$
\delta LAM\_k = \sum\_{i=1}^{m} \left[ A\_i \cdot \sum\_{j=1}^{n} C\_{ij} \times X\_{kj}^\* \right] \tag{7}
$$
where *UAM<sup>k</sup>* is urban comprehensive attractiveness for migrants of city *k*; *m* is the number of major components which make the cumulative ratio of the total variance greater than 85%; *A<sup>i</sup>* contributes the major components *i* to UAM of the city; *n* is the number of indexes; *Cij* is the contribution of index *j* to the major components *i*; and *X \* kj* is the standardized value of index *j* in city *k*. where *UAMk* is urban comprehensive attractiveness for migrants of city *k*; *m* is the number of major components which make the cumulative ratio of the total variance greater than 85%; *Ai* contributes the major components *i* to UAM of the city; *n* is the number of indexes; *Cij* is the contribution of index *j* to the major components *i*; and *X\* kj* is the standardized value of index *j* in city *k*.
#### **4. Results 4. Results**
#### *4.1. Correlation between External-MAI and Urban Inflow Migrants 4.1. Correlation between External-MAI and Urban Inflow Migrants*
According to the definition of MAI, the migration tendency of the person from the outside areas can be conveyed through external-MAI. Under the assistance of relevant migration data from the Tencent map, we engaged in exploring the relationship between external-MAI and urban migration population. Pearson correlation coefficient was adopted to reveal the relationship between them; the results have been shown in Table 4 and Figure 3. As we could observe, there are significant positive correlations between external-MAI and population migration in the three urban agglomerations. The Pearson coefficients are 0.948, 0.876, and 0.879 separately in BTH, YRD, and PRD, which has a holistic coefficient of 0.844. All of them have passed the significance test at 99% confidence level. Focused on their spatial heterogeneity, the cities of BTH has the highest correlation. According to the definition of MAI, the migration tendency of the person from the outside areas can be conveyed through external-MAI. Under the assistance of relevant migration data from the Tencent map, we engaged in exploring the relationship between external-MAI and urban migration population. Pearson correlation coefficient was adopted to reveal the relationship between them; the results have been shown in Table 4 and Figure 3. As we could observe, there are significant positive correlations between external-MAI and population migration in the three urban agglomerations. The Pearson coefficients are 0.948, 0.876, and 0.879 separately in BTH, YRD, and PRD, which has a holistic coefficient of 0.844. All of them have passed the significance test at 99% confidence level. Focused on their spatial heterogeneity, the cities of BTH has the highest correlation.
**Table 4.** The Pearson coefficient between population migration and external-migration attention indexe (MAI). **Table 4.** The Pearson coefficient between population migration and external-migration attention indexe (MAI).
Note: UA: urban agglomeration; BTH: Beijing-Tianjin-105 Hebei metropolitan region; YRD: the Yangtze River Delta; PRD: the Pearl River Delta. Note: UA: urban agglomeration; BTH: Beijing-Tianjin-105 Hebei metropolitan region; YRD: the Yangtze River Delta; PRD: the Pearl River Delta.
**Figure 3.** Scatter plot of external-MAI and migration population. **Figure 3.** Scatter plot of external-MAI and migration population.
Applying the principal component analysis, we obtained UAM of target cities based on the statistical data; the correlation study was deployed between the comprehensive UAM and external-MAI. As shown in Table 5 and Figure 4, we could observe a significant correlation between the UAM and external-MAI in the study areas. The coefficients of the whole area, BTH, YRD, and PRD are separately 0.829, 0.924, 0.984, and 0.789. The high correlation between them illustrated that urban received external-MAI is highly correlated to the attractiveness of urban itself. The relationship between such a cyber-based index and a traditional statistic-based index can be implied. Applying the principal component analysis, we obtained UAM of target cities based on the statistical data; the correlation study was deployed between the comprehensive UAM and external-MAI. As shown in Table 5 and Figure 4, we could observe a significant correlation between the UAM and external-MAI in the study areas. The coefficients of the whole area, BTH, YRD, and PRD are separately 0.829, 0.924, 0.984, and 0.789. The high correlation between them illustrated that urban received external-MAI is highly correlated to the attractiveness of urban itself. The relationship between such a cyber-based index and a traditional statistic-based index can be implied.
**Table 5.** The Pearson coefficient between urban comprehensive attractiveness for migrants (UAM) and external-MAI. **Table 5.** The Pearson coefficient between urban comprehensive attractiveness for migrants (UAM) and external-MAI.
Note: UA: urban agglomeration; BTH: Beijing-Tianjin-105 Hebei metropolitan region; YRD: the Yangtze River Delta; PRD: the Pearl River Delta. Yangtze River Delta; PRD: the Pearl River Delta.
**Figure 4.** The scatter plot of external-MAI and UAM.
**Figure 4.** The scatter plot of external-MAI and UAM. Furthermore, the Pearson correlation coefficients between the selected indexes and external-MAI have been calculated, as shown in Table 6. We can see that all the two indexes for job opportunities and income levels have the highest correlation with external-MAI in the study area. For the living condition perspective, a positive correlation can be observed between the Participant Rate of Urban Basic Medical Care System and external-MAI in BTH and YRP. However, significant correlations cannot be observed between the unemployment rate per capita living area with external-MAI. Paying attention to the education opportunities, significant correlations can be found in BTH and YRD between the three educational indexes and population attention index. In PRD, only the number of primary schools significantly correlates with external-MAI. In the three urban agglomerations, the strongest correlations are depicted between the Tertiary Industrial Output-Value and external-MAI, which reflect job opportunities in the areas being conventionally attractive for the potential migrants. Insignificant low correlation between the unemployment rate per capita living Furthermore, the Pearson correlation coefficients between the selected indexes and external-MAI have been calculated, as shown in Table 6. We can see that all the two indexes for job opportunities and income levels have the highest correlation with external-MAI in the study area. For the living condition perspective, a positive correlation can be observed between the Participant Rate of Urban Basic Medical Care System and external-MAI in BTH and YRP. However, significant correlations cannot be observed between the unemployment rate per capita living area with external-MAI. Paying attention to the education opportunities, significant correlations can be found in BTH and YRD between the three educational indexes and population attention index. In PRD, only the number of primary schools significantly correlates with external-MAI. In the three urban agglomerations, the strongest correlations are depicted between the Tertiary Industrial Output-Value and external-MAI, which reflect job opportunities in the areas being conventionally attractive for the potential migrants. Insignificant low correlation between the unemployment rate per capita living area with external-MAI can be detected.
area with external-MAI can be detected.
**Table 6.** Correlation coefficient between external-MAI and urban pulling indicators.
Note: \*: Pearson correlation is significant at the 0.01 level. TIV: Tertiary Industrial Output-Value; IPC: Urban Residents' Per Capita Disposable Income; UR: Unemployment Rate; RBM: Participant rate of Urban Basic Medical Care System; LPC: Per Capita Living Area; SSN: Number of Regular Secondary Schools; PSN: Number of Regular Primary Schools; UN: Number of University.
### *4.2. Correlation between Local-MAI and Floating Population Inner City*
The results of correlation analysis between local-MAI and local floating population have been shown in Table 7 and Figure 5. We can see that, no matter in the whole study area or the individual urban agglomeration, high correlation coefficients were gained. Especially in the YRD, the relevant coefficient has arrived at 0.950. PRD has a relatively lower value but is still higher than 0.75. Divided by the median value of local-MAI and local floating population, the cities in the study area can be divided into four types. Thereinto, 78.95% of them has high-high or low-low features. For the cities with higher-than-average floating population and higher-than-average local-MAI, there are three located in the BTH (Beijing, Tianjin, and Baoding), two in YRD (Shanghai and Suzhou), and two in PRD (Shenzhen and Guangzhou).
**Table 7.** The Pearson coefficient between local-MAI and floating population.
Note: UA: urban agglomeration; BTH: Beijing-Tianjin-105 Hebei metropolitan region; YRD: the Yangtze River Delta; PRD: the Pearl River Delta. *Int. J. Environ. Res. Public Health* **2020**, *17*, x 12 of 18
**Figure 5.** The scatter plot of local-MAI and floating population. **Figure 5.** The scatter plot of local-MAI and floating population.
**Figure 6.** Scatter plot of external-MAI and Local-MAI.
To explore the relationship between intercity-MAI and intercity population flow, the results have been shown in Table 8 and Figure 7. As we could notice, the average value of intercity-MAI is 1.00, Guangzhou-Shenzhen has the highest intercity-MAI at 5.19; and Shenzhen-Chengde has the lowest index of 0.04. For the individual urban agglomeration, the intercity-MAI among Beijing, Tianjin, and Shijiazhuang has the highest top three values in BTH. The same level of intercity-MAI
*4.3. Correlation between Intercity-MAI and Intercity Population Flow*
local-MAI also has a higher external-MAI
To further excavate information from MAI, the relationship between local-MAI and external-
To further excavate information from MAI, the relationship between local-MAI and external-MAI has been explored; the results are shown in Figure 6. There is a highly positive correlation between the two indexes, of which the r is 0.7538 and *p* is 0.01. It is shown that the city with higher local-MAI also has a higher external-MAI To further excavate information from MAI, the relationship between local-MAI and external-MAI has been explored; the results are shown in Figure 6. There is a highly positive correlation between the two indexes, of which the r is 0.7538 and *p* is 0.01. It is shown that the city with higher local-MAI also has a higher external-MAI
**Figure 5.** The scatter plot of local-MAI and floating population.
*Int. J. Environ. Res. Public Health* **2020**, *17*, x 12 of 18
**Figure 6.** Scatter plot of external-MAI and Local-MAI. **Figure 6.** Scatter plot of external-MAI and Local-MAI.
| doab | 2025-04-07T04:13:02.988847 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 242
} |
ffa2a234-da32-4c59-b516-55230ec92248.243 | *4.3. Correlation between Intercity-MAI and Intercity Population Flow*
*4.3. Correlation between Intercity-MAI and Intercity Population Flow* To explore the relationship between intercity-MAI and intercity population flow, the results have been shown in Table 8 and Figure 7. As we could notice, the average value of intercity-MAI is 1.00, Guangzhou-Shenzhen has the highest intercity-MAI at 5.19; and Shenzhen-Chengde has the lowest index of 0.04. For the individual urban agglomeration, the intercity-MAI among Beijing, Tianjin, and Shijiazhuang has the highest top three values in BTH. The same level of intercity-MAI To explore the relationship between intercity-MAI and intercity population flow, the results have been shown in Table 8 and Figure 7. As we could notice, the average value of intercity-MAI is 1.00, Guangzhou-Shenzhen has the highest intercity-MAI at 5.19; and Shenzhen-Chengde has the lowest index of 0.04. For the individual urban agglomeration, the intercity-MAI among Beijing, Tianjin, andShijiazhuang has the highest top three values in BTH. The same level of intercity-MAI can be found in YRD for Shanghai, Hangzhou, and Suzhou. In PRD, such level interactions are observed betweenGuangzhou, Shenzhen, and Foshan.
**Table 8.** The Pearson coefficient between intercity-MAI and intercity population flow.
Under the correlation analysis of these two indexes, a moderately positive correlation can be observed in the study area (See Table 8). For the three urban agglomerations, PRD has the highest correlation among them and the correlation in BTH and YRD represents a relatively lower level. There are 59 pairs of cities that have a high-high correlation pattern (high intercity-MAI and high intercity population flow), there into 15 pairs in BTH, 24 pairs in YRD, and 20 pairs in PRD; 147 pairs of cities exhibited the low-low correlation pattern, of which, in BTH, YRD, and PRD, are separately 46, 69, and 32. These two kinds of correlation patterns occupy 75% of the total. Although relevant correlation coefficients of intercity-MAI are relatively limited, it can capture the interaction trend of population flow at an acceptable level.
population flow at an acceptable level.
observed between Guangzhou, Shenzhen, and Foshan.
can be found in YRD for Shanghai, Hangzhou, and Suzhou. In PRD, such level interactions are
**Table 8.** The Pearson coefficient between intercity-MAI and intercity population flow.
**Region Three UAs BTH YRD PRD**
Sig(2-side) 0.0000 0.0000 0.0000 0.0000
Under the correlation analysis of these two indexes, a moderately positive correlation can be observed in the study area (See Table 8). For the three urban agglomerations, PRD has the highest correlation among them and the correlation in BTH and YRD represents a relatively lower level. There are 59 pairs of cities that have a high-high correlation pattern (high intercity-MAI and high intercity population flow), there into 15 pairs in BTH, 24 pairs in YRD, and 20 pairs in PRD; 147 pairs of cities exhibited the low-low correlation pattern, of which, in BTH, YRD, and PRD, are separately 46, 69, and 32. These two kinds of correlation patterns occupy 75% of the total. Although relevant correlation coefficients of intercity-MAI are relatively limited, it can capture the interaction trend of
**Figure 7.** The scatter plot of intercity-MAI and intercity population flow. **Figure 7.** The scatter plot of intercity-MAI and intercity population flow.
#### **5. Discussion 5. Discussion**
#### *5.1. Evaluation of MAIs 5.1. Evaluation of MAIs*
The massive population migration is the specific phenomenon and the inevitable driving force promoting the urbanization of population in China and many developing countries. The collection of urban MAIs can obtain the public intention of migration based on individual search actions and can offer exploration of population migration. Depending on the MAIs, we analyzed the correlation relationship between local-MAI and external-MAI; a high correlation has been discovered. It implied that the city with relatively higher local-MAI has a higher external-MAI. Migration may be active in the high-high cities, such as Shanghai, Beijing, and Shenzhen. According to the dynamic monitoring survey of China's migration population in 2015, the proportions of floating population inside these three cities separately reached 40.26%, 38.02%, and 67.51%, which are much higher than the average value of China at 18.00%. Besides, they have separately occupied 12.22%, 11.99%, and 8.64% (ranked top 3) of the whole inflow population of the three urban agglomerations, which has the most dynamic The massive population migration is the specific phenomenon and the inevitable driving force promoting the urbanization of population in China and many developing countries. The collection of urban MAIs can obtain the public intention of migration based on individual search actions and can offer exploration of population migration. Depending on the MAIs, we analyzed the correlation relationship between local-MAI and external-MAI; a high correlation has been discovered. It implied that the city with relatively higher local-MAI has a higher external-MAI. Migration may be active in the high-high cities, such as Shanghai, Beijing, and Shenzhen. According to the dynamic monitoring survey of China's migration population in 2015, the proportions of floating population inside these three cities separately reached 40.26%, 38.02%, and 67.51%, which are much higher than the average value of China at 18.00%. Besides, they have separately occupied 12.22%, 11.99%, and 8.64% (ranked top 3) of the whole inflow population of the three urban agglomerations, which has the most dynamic migration in China. The predominant roles of them in attracting population outside the city are declared. Active migration movement can be detected to support the hypothesis derived from the correlation relationship between local-MAI and external-MAI.
Analyzing the correlation of external-MAI with UAM, the reasonability of external-MAI can be verified through the high correlation with conventional statistics analysis. Based on the push-pull theory of migration, in the cities with higher urban pulling force, more influx of population can be observed. Through the calculation of UAM, which depicted urban pulling force, the city with higher external-MAI can observe higher UAM. The feature of external-MAI coincides with the setting of push-pull theory. Further, exploring the relationship between external-MAI and the indexes which reflect urban migration attractiveness, there are significant correlations between tertiary industrial output-value and urban residents' per capita disposable income with the external-MAI in the whole study area. Most of the cities with higher job opportunities and income levels receive more migration attentions from the outside area. This finding coincides with the dynamics monitoring survey of migration population suggesting a migration reason in Table 1 (*work and trade* as the predominant migration reason), which can represent the ability of MAI indexes in capturing the impact of migration reasons. With respect to the indexes described urban living conditions, there are no significant correlations between the population migration attention and unemployment rate or per capita living area, this results may correspond to the great exception of potential migrants for their future urban condition, which can be explained by the Todaro migration model from the perspective of development economics. Todaro migration model argues that the migration of population is based on the "expected profit" of migrants. The hardships in urban life have not obtained enough attention from potential migrants, particularly for the rural-urban migrants who lack the necessary information as they enter a new different world [43]. Further, the more schools a city has, the more public migration attention it receives. The positive correlation existing between the education indexes (the number of primary schools, secondary schools, and universities) and external public migration attention exposes that the educational opportunities also intensify the level of urban migration attention. In PRD, the focus of educational concern only derives from the consideration of secondary schools; significant correlation has not been observed between the number of the other two levels of schools in the area, which may be attributed to the relatively lower education level of Guangzhou Province (the administrative province that PRD belongs to) than the other two urban agglomerations.
Besides, we further adopted the neoclassical theory in population migration to explore the reasonability of MAIs. The per capita disposable income of urban residents, which has been viewed as the direct index depicting the possibility for migrants to improve economic benefit, has been adopted to conduct the correlation analysis with external-MAI; the results have shown that the external-MAI has significant positive correlation with the per capita disposable income of urban residents (with the correlation coefficients 0.650, 0.945, 0.752, and 0.780 separately in the whole study area, in BTH, in YRD, and in PRD). The reasonability of MAIs can be further identified.
#### *5.2. Relationship with Migrants*
With the correlation analysis of external-MAI and urban migration population, we could observe a significantly positive correlation. The resource endowment gap between different urban areas (e.g., economic development level, environmental quality, promotion of opportunities for individuals, etc.) triggers personal develop exception and forges migrant needs in flowing among diverse regions [22,43]. Collecting information about the targeted city by employing the search query engine is an efficient and low-cost approach to supplement requisite information before deploying actual migration to external areas. As noted as the correlation results of external-MAI and urban inflow migrants in the study area, we can accept the hypothesis that the migration-related search queries from individual users were able to positively reflect urban inflow migrants. External-MAI can be applied to reflect urban inflow migrants on the annual scale.
The high correlation between local-MAI and the floating population inside cities was a signal to prove their close relationship. Nowadays, the floating population inner city has become an influential part in enacting urban planning and policy. Generally speaking, the floating population lived with relatively weaker urban amenities than the local population [54,55]. The desire of improving current conditions was more intensive for them, which was delineated by the high demand for new job opportunities, study chance, and the possibility of improving living quality, etc. Driven by such basic needs, the corresponding search query can be brought into the cyber space and raises the high correlation between local-MAI and floating population inside the city.
Intercity migration has already become one of the significant migration models in current China. We analyzed the correlation between intercity-MAI and intercity population flow in 2015; a similar positive correlation can be observed as 0.57 (*p*-value 0.00) in the whole area. The results show that the representativeness of intercity-MAI for population flow between different cities was effective, but the correlation relationship was relatively limited. It might be caused by two main reasons: (1) The selection of search keywords cannot cover every reason for migration flows. A unique keyword system may exhibit some deviation in reflecting the driving force of every population flow interaction; (2) migration movement has a lagging feature. It may happen a few months, years, or a much longer
time after the search action. It also may be canceled or indefinitely delayed after information acquisition through searching, which makes the relatively lower correlation between intercity-MAI and intercity population inflow. Generally speaking, the correlation between intercity-MAI and population flow is still on an acceptable level. It can be a supplement for the population flow research of insufficient data. In future work, the construction of a targeted search keywords system for objective population flow can be adopted to remedy such drawbacks.
Besides, for the three MAI indexes, the different correlation coefficients in the three urban agglomerations revealed that regional disparity exists. We calculated the variance (VAR) and coefficient of variation (CV) of relevant correlation coefficients of three MAI indexes, as shown in Table 9. It can be seen that all the VARs are lower than 0.01 and that all the CVs are lower than 10%. The robustness of external-MAI, local-MAI, and intercity-MAI in reflecting population can be partly ensured in the study area.
**Table 9.** The variance (VAR) and coefficient of variation (CV) of different types of MAI.
Furthermore, we have tested the significance of slope of the three trend lines, which were separately fitted based on external-MAI and inflow population, local-MAI and floating population, and intercity-MAI and population flow to identify whether MAI indexes could steadily reflect the migration situation in different urban agglomerations. All the significances of slopes have been rejected by significant testing at a significant level of 0.05 (sig = 0.43, 0.19, and 0.86 for external-MAI, local-MAI, and intercity-MAI). The null hypothesis could be accepted as there is no significant deviance between the slopes. Although the representations of MAI are diverse in different regions, the deviances are nonsignificant.
| doab | 2025-04-07T04:13:02.989457 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 243
} |
ffa2a234-da32-4c59-b516-55230ec92248.244 | **6. Conclusions**
Migration population has immense potential to push urbanization process in current China and other developing countries. Exploration of population migration based on multisource data can bring more information about the noticeable driving force of urban development. In the information and network era, the MAI indexes can reveal how the public put their attention on migration-related items. Based on the cyber-based indexes, we explore the relationship between public migration attention in cyber space and urban migration population in geographical space inner region, across region, and between regions. The results can answer the questions mentioned in the introduction that search query data based MAI indexes can positively reflect the situation of migration population inner region and across region and, for the population flow, that it is an alternative supplement and support when relevant data is deficient.
Population migration is a complex process driven by diverse forces; this paper conducted a series of analyses from the perspective of search query data in cyber space. However, some limitations exist: First, the selection of continuous search keywords is limited by the short period of data acquisition from the search query engine. Following the incremental collection of search query data, more suitable search keywords should be selected to cover different aspects of public migration attention to thus better delineate the difference and characteristic of urban migrant population; second, this paper focus on the panel data analysis; future work will emphasize on the time-series analysis and excavate more information from a dynamic perspective.
**Author Contributions:** C.L. participated in all phases; J.H. helped in conceiving and designing the research; X.D. helped in paper organization and language correction. All authors have read and approved the final manuscript. **Funding:** This research was funded by China Postdoctoral Science Foundation, grant number 2019M663592.
**Conflicts of Interest:** The authors declare no conflict of interest. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; and in the decision to publish the results.
| doab | 2025-04-07T04:13:02.990267 | 17-11-2022 17:28 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa2a234-da32-4c59-b516-55230ec92248",
"url": "https://mdpi.com/books/pdfview/book/6288",
"author": "",
"title": "Migration, Resilience, Vulnerability and Migrants’ Health",
"publisher": "MDPI - Multidisciplinary Digital Publishing Institute",
"isbn": "9783036555577",
"section_idx": 244
} |
ffa52fe3-258a-4af9-9030-900a85a489ae | #### A Byzantine Book on Dream Interpretation The Oneirocriticon of Achmet and Its Arabic Sources
Maria Mavroudi
BRILL
A BYZANTINE BOOK ON DREAM INTERPRETATION
## **THE**
# **MEDIEVAL MEDITERRANEAN**
**PEOPLES, ECONOMIES AND CULTURES,** 400-1453
EDITORS
HuGH KENNEDY (St. Andrews) PAUL MAGDALINO (St. Andrews) DAVID ABULAFIA (Cambridge) BENJAMIN ARBEL (Tel Aviv) MARK MEYERSON (Toronto) LARRY **J.** SIMON (Western Michigan University)
VOLUME 36
# **A BYZANTINE BOOK ON DREAM INTERPRETATION**
*The* Oneirocriticon of Achmet *and Its Arabic Sources*
BY
### MARIA MAVROUDI
BRILL LEIDEN· BOSTON· KOLN 2002
This is an open access title distributed under the terms of the CC BY 4.0 license, which permits any non-commercial use, distribution, and reproduction in any medium, provided no alterations are made and the original author(s) and source are credited. Further information and the complete license text can be found at https://creativecommons.org/licenses/by/4.0/
The terms of the cc license apply only to the original material. The use of material from other sources (indicated by a reference) such as diagrams, illustrations, photos and text samples may require further permission from the respective copyright holder.
Illustration on the cover. The group of dream interpreters mentioned in the Oneinorition as depicted in the only illustrated copy of the work, MS 3632 of the Bologna University Library, fol. 443v. (© Bologna University Library.)
This book is printed on acid-free paper.
#### Die Deutsche Bibliothek - CIP-Einheitsaufnahme
Mavroudi, Maria:
A Byzantine book on dream interpretation : The Oncirocriticon of Achmet and its Arabic sources / by Maria Mavroudi. - Lciden ; Boston ; Köln : Brill, 2002 (The medieval Mediterrancan ; Vol. 36) ISBN 90-04-12079-3
#### Library of Congress Cataloging-in-Publication Data
Library of Congress Cataloging-in-Publication Data is also available
Copyright 2002 by Maria Mavroudi. Published by Koninklijke Brill NY, Leiden, The Netherlands.
All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission from the publisher.
> Authorization to photocopy items for internal or personal use is granted by Brill provided that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive, Suite 910 Danvers MA 01923, USA. Fees are subject to change.
#### PRINTED IN THE NETHERLANDS
### TABLE OF CONTENTS
#### VI TABLE OF CONTENTS
#### ACKNOWLEDGMENTS
This book is a revised version of the doctoral dissertation that I submitted to the Committee on Byzantine Studies at Harvard University in March of 1998. Some of its conclusions can also be found, though in a briefer form, in the dissertation of John Lamoreaux, entitled "Dream Interpretation in the Early Medieval Near East" submitted to the Program of Religion at Duke University in 1999. As Dr. Lamoreaux makes no reference to my dissertation, I am assuming that we both reached similar conclusions independently, after examining much of the same material, and that our agreement corroborates the arguments presented here. Dr. Lamoreaux's dissertation is on a different topic and consequently examines our common subjects only in passing; conversely, I treat only briefly problems that he deals with in greater detail.
The dissertation that became this book owes a great deal to the guidance of my academic adviser, Thor Sevcenko, and my other readers, Wolfhart Heinrichs, Angeliki Laiou, and Ioli Kalavrezou. I am indebted to Toufic Fahd for his generosity in sharing with me some of his unpublished work and for his warm hospitality, and to Eleni Bassoukou-Kondyli for helping me choose the dissertation topic. My thanks are also due to M. J. Kister, who provided me with a photo reproduction of Ibn Qutayba's Ankara manuscript; to Erich Trapp for supplying me with the unpublished entries of his dictionary in progress regarding words that occur in the *Oneirocriticon;* to Paul Magdalino for sending me his then unpublished article "The Road to Baghdad in the Thought World of Ninth Century Byzantium"; and to Margaret Sevcenko for being an intelligent copy-editor.
I also benefited from the support of various institutions. The Harvard Center for Middle Eastern Studies provided me with a grant to study at the Arabic Language Institute of the American University in Cairo during the summer of 1992. Most of the microfilms of Arabic manuscripts I consulted were purchased with a U. S. Department of Education Pre-doctoral Research Award. My research would not have been the same without the excellent resources of Dumbarton Oaks, where I spent a year as Junior Fellow (1994-95). A Whiting Fellowship in the Humanities from Harvard (1995-96) freed me from teaching obligations to allow me to write my dissertation in the course of the following year. The final corrections to the manuscript were made during a second
fellowship at Dumbarton Oaks (2000-2001).
I dedicate this work to my parents as a small token of appreciation for what I owe them.
M.M.
#### LIST OF ABB RE VIA TIO NS
#### LIST OF ABBREVIATIONS
#### AUTHOR'S NOTE
Old Testament passages in Greek are quoted according to the critical text in A. Rahlfs, *Septuaginta* (Stuttgart: Deutsche Bibelgesellschaft, 1979). The New Testament passages in Greek are quoted according to the critical text in E. Nestle, K. Aland et al., *Novum Testamentum Graece* (Stuttgart: Deutsche Bibelgesellschaft, 1979). For the translation of passages from both, *The Holy Bible: King James Version* (New York: Ivy Books, 1991) has been used. The Qur'an is quoted according to the text and numbering of verses in M. Pickthall, *The Glorious Koran* (Albany: State University of New York Press, 1976) unless otherwise noted, and the translations are either from Pickthall, or from A. J. Arberry, *The Koran Interpreted* (London: Oxford University Press, 1964), depending on which was judged more understandable in the context of the Arabic dreambook where the verse was quoted.
Arabic terms that are italicized are defined at first usage. Some of them can even be found in English dictionaries.
#### CHAPTER ONE
#### THE AUTHOR OF THE ONEIROCRITICON AND HIS SOURCES
In addition to the numerous references to dreams and dream interpretation that are dispersed throughout Byzantine literature, two groups of texts exclusively dedicated to dreams and oneiromancy survive from the Byzantine centuries. The first group comprises theoretical treatises on the nature of sleep and dreams, the lengthiest of which are the Christian approach to dreams by the bishop of Ptolemais Synesios with its fourteenth-century commentary by Nikephoros Gregoras and the two commentaries on Aristotle's views on dreams, one by Themistios and the other by Michael of Ephesos. Shorter works are John Italos's explanation of the ivory and horn gates of dreams mentioned in the Odyssey, Michael Psellos's brief treatise on dreams, and the emperor Manuel Paleologos's epistle on dreams addressed to Andreas Asan. All these texts are philosophical in nature and do not concern us here.
The second group consists of texts that view dreams as a means for predicting the future. These can be subdivided into two categories. The first category comprises manuals that combine astrology with dream interpretation by determining the significance of a dream according to the phases of the moon and the position of the stars; they are brief (one or two printed pages each) and anonymous.2 The second category comprises texts that provide a "key to dreams," consisting of a catalogue of objects or activities that occur in dreams and an interpretation of what they mean for the dreamer's future. They are longer than the astrological dreambooks. The name of an author usually appears
For references to editions of Synesios, Gregoras, Themistios, and Manuel Paleologos, see G. Guidorizzi, "I prontuari oniromantici bizantini," Istituto Lombardo di Lettere. Rendiconti 111 (1977), p. 136. For references to editions of the works by Italos, Psellos, and Michael of Ephesos, as well as a discussion of their contents, see T. Ricklin, Der Traum der Philosophie im 12. Jahrhundert. Traumtheorien zwischen Constantinus Africanus und Aristoteles (Leiden, Boston, Cologne, 1998), pp. 270-307.
<sup>2</sup> Guidorizzi, "I prontuari oniromantici bizantini," p. 139, mentions six such texts, including two unpublished ones. For additional ones, see below, n. 31.
<sup>1</sup> There is no technical term in English for this kind of dreambook. The French term is "clef des songes." For the most recent survey of Byzantine dreambooks in this category (a total of nine items), and the identification of hitherto unknown manuscripts that afford the opportunity for further textual emendations to already published texts, see Th. Dettorakes, "Ta byzantina oneirokritika: dyo nea cheirographa," Palimpsēston 16 (1996), pp. 65-74.
#### 2 CHAPTER ONE
in the title, though it is almost always a false attribution. The subject of this study, the so-called *Oneirocriticon of Achmet,* is a work of this kind.
The oldest and most extensive work on the interpretation of dreams that survives in the Greek language is the *Oneirocritica* in five books written by Artemidoros in the second century A.D., during the age of the Antonines. 4 His work belongs to a much earlier tradition of Greek dream interpretation. According to its preface, the material was collected through research and experience, and from older works on dream interpretation, whose authors are frequently quoted by name. The representatives of this older tradition lived as much as seven hundred years before Artemidoros's time. 5 The five books of the *Oneirocritica* cover three hundred and twenty-four Teubner pages. <sup>6</sup>
Next in chronological order come the eight surviving Byzantine dreambooks. <sup>7</sup> They are brief and schematic. Seven of them do not take up more than twenty pages of the volumes in which they appear. Only the so-called *Oneirocriticon of Achmet,* two hundred and forty-one Teubner pages long, is comparable to the work of Artemidoros in length, breadth, and detailed treatment of dream symbols. <sup>8</sup>
The evidence found in or about the *Oneirocriticon* does not allow us to date it with great accuracy. Only a *terminus post* and a *terminus ante quern* that are removed from each other by approximately two centuries can be established. The work includes a chapter on icons,9 which indicates that it was composed
8 Critical edition *Achmetis Oneirocriticon,* ed. F. Drexl (Leipzig, 1925), hereafter cited as Drexl when identifying quotations in notes and text.
9 Chapter 150, Drexl I 05, 12 ff.
<sup>4</sup> On Artemidoros and his age, see e.g., Artemidoros, *The lnte1pretation of Dreams,* introduction, trans. and commentary by Robert J. White (Park Ridge, N. J., 1975), pp. I ff. Also Artemidoros, *Traumhuch,* introduction, trans. and commentary by F.S. Krauss and M. Kaiser (Basel and Stuttgart, 1965), pp. I ff.
*<sup>5</sup>*On Greek dream interpretation before Artemidoros, see D. de! Como, *Graecorum de re onirocritica scriptorum reliquiae* (Milan, 1969).
<sup>6</sup> Most recent critical edition, *Artemidori Daldiani Onirocriticon Lihri V,* ed. R. Pack (Leipzig, 1963 ), hereafter cited as Pack.
<sup>7</sup> Six of them have been studied, dated, and translated in S. M. Oberhelman, "The Oneirocritic Literature of the Late Roman and Byzantine Eras of Greece: Manuscript Studies, Translations and Commentaries to the Dream Books of Greece during the First Millennium A.D., with Greek and English Catalogues of the Dream Symbols and with a Discussion of Greek Oneiromancy from Homer to Manuel the Palaeologian ,"Ph.D. diss., University of Minnesota, 1981. A seventh, which survives in *Marc. gr.* 608 but is still unpublished, is mentioned in Guidorizzi, "I prontuari oniromantici bizantini," p. 138. See also the more recent German translation of the shorter Byzantine dream books by K. Brackertz, *Die Volks-Traumhiicher des hyzantinischen Mittelalters* (Munich, 1993).
after the end of iconoclasm and the restitution of icons in 843. <sup>10</sup>
Three manuscripts from the eleventh century indicate that the *Oneirocriticon* enjoyed a fairly wide circulation before the century's end, providing the *terminus ante quern.* The first of the three is *Paris. Suppl. gr.* 690, which contains an abridged version of the *Oneirocriticon* and constitutes the oldest surviving manuscript of the work. G. Rochefort dated the manuscript between 1075 and 1085 on the basis ofpaleography. 11 The second is *Laurent. Plut.* 87, 8 (11th century), the oldest surviving manuscript of Artemidoros. An eleventhcentury hand copied two passages from the *Oneirocriticon* on the margins of this manuscript as commentary to passages from Artemidoros with similar contents. 12 On the basis of these two manuscripts S. Oberhelman established ca. 1075 as the *terminus ante quern* for the composition of the work. 13 Supporting it also is a third manuscript from the eleventh century, *Patmiacus* 6, <sup>14</sup> which is the oldest surviving manuscript of the anthology known as the *Florilegium Baroccianum. <sup>15</sup>*A lengthy passage of the *Oneirocriticon <sup>16</sup>*is excerpted in "A6yo~ MS DEpt £vunviffiv" (chapter 24, On Dreams) among quotations from the Old and the New Testament, Basil of Caesarea, Gregory of Nazianzus, John Chrysostom, St. John of the Ladder, St. Ephraim, Achilles Tatius and a letter of Euryviades to Kirnon, both Athenians of the fifth century B.c. <sup>17</sup>The composition of the *Florilegium Baroccianum* has also been placed in the
<sup>15</sup>M. Richard, *Dictionnaire de Spiritualite,* vol. 5 (Paris, 1962), s.v. "Florileges spirituels," col. 494 3°. Richard does not explain the reasons for such a date.
<sup>16</sup>Drexl 1, 16-2, 22.
<sup>10</sup>Earlier studies on the *Onefrocriticon* consider the beginning of the reign of Caliph al-Ma'miin (813) as a *terminus post quem.* The caliph is mentioned in the *Oneirocriticon* seven times (Drexl 15, 19; 16, 2; 23, 25; 29, 18; 57, 13; 99, 7; 111, 25).
<sup>11</sup>G. Rochefort, "Une anthologie grecque du Xle siecle: Le *Parisinus Suppl. gr.* 690," *Scriptorium* 4 (1950), pp. 3-17.
<sup>12</sup> Pack designated this hand as L1• The passages comment on Artemidoros's theory of dream interpretation and can be found on fols. 7r (Drexl 240, 21-241, 17) and 8r (Drexl 240, 9-12), corresponding to Pack 16, I 0 and 18, 17; cf. Pack, pp. vi-vii; also S. Oberhelman, "Two Marginal Notes from Achmet in the Cod. *Laurent. Plut.* 87, 8," *BZ* 74 (1981 ), pp. 326-27, who presents this information as his own discovery without reference to Pack.
<sup>13</sup> S. Oberhelman, *The Oneirocriticon of Achmet: A Medieval Greek and Arabic Treatise on the lnte1pretatio11 of Dreams* (Lubbock, Tex., 1991), p. 13.
<sup>14</sup>A. Komines, *Patmiake Bibliotheke, etoi katalogos t1!11 cheirographon kodikon ti's Hieras Mones Hagiou Ioannou tou Theologou Patmou,* vol. I (Athens, 1988), pp. 8-12.
<sup>17</sup>E. Sargologos, *Une traite de vie spirituel/e et morale du Xie siecle: le florilege sacro-profane du manuscrit* 6 *de Patmos* (Asprovalta and Thessaloniki, 1990), p. 838. The excerpt is identified by D. Christi des, 'To apospasma tou Sirim sto *Florilegium Baroccianum," Hel/enika* 43:2 ( 1993), pp. 219-25.
eleventh century. The most recent author, other than Achmet, cited in the *Florilegium* is Patriarch Photius ( d. 898). <sup>18</sup>The text of the *Oneirocriticon* mentions a few financial and administrative terms that confirm its dating between 843 and 1075, but it is impossible to narrow down this time span based on their occurrence.
The influence that the *Oneirocriticon of Achmet* exerted on subsequent works on dream interpretation was considerable. Dreambooks produced in medieval and Renaissance Europe are heavily indebted not only to Artemidoros, but also to the *Oneirocriticon.* Its full text was translated into Latin in the twelfth century, that is, almost four hundred years earlier than the full text of Artemidoros, <sup>19</sup>and through this translation found its way into several European vernacular languages from the late thirteenth century onward. 20 The *Oneiro-*
<sup>19</sup>The complete text of the *Oneirocriticon* was translated into Latin by Leo Tuscus in 1176; that of Artemidoros, by Janus Comarius in 1539.
<sup>2</sup>° For the influence of the *Oneirocriticon of Achmet* on Latin and vernacular dreambooks, see F. Berrio!, *Exposicions et significacions des songes* (Geneva, 1989), pp. 36-42. He dates the Anglo-Norman version of the *Oneirocriticon* in manuscript *Berlin* 968.Q to the late 13th century. For a critical edition of the Middle French version, see M. G. Glover, "Critical Edition of the Middle French Version of Achmet ibn Sirin's Oneiromancy Found in MS Fran~·ais 1317 fols. 51 r-106v, Paris, B ibliotheque N ationale, Entitled ( cy commence la table des) Exposicions et significacions des songes par Daniel et autres exposez," Ph.D. diss., Birkbeck College, London, 1992 (inaccessible to me). On the influence of the *Oneirocriticon* in medieval and early modem Europe, see L. Thorndike, *A History of Magic and Experimental Science,* vol. 2 (New York, 1943), pp. 290-304; T. Fahd, "L' oniromancie orientale et ses repercussions sur I' oniromancie de l' occident medieval," *Oriente e Occidente nel medioevo: filosofia e scienze: convegno internazionale, 9-15*
<sup>18</sup>The reason is the difficulty of establishing exactly when a technical term appeared for the first time, became obsolete, or changed in meaning. These changes came about gradually, and often our archival information is insufficient for drawing a definite conclusion. The terms are *nomismata, miliaresia* and *folleis* (Drexl 208, 19-209, 22), which, in the context of the *Oneirocriticon,* mean "gold coin," "silver coin" and "bronze coin" respectively. All three terms had been current since late antiquity, but gradually became obsolete after Alexios Komnenos' s coinage reform of 1092. In the passage on coins from the abridgment of *Paris. Suppl. gr.* 690 (fol. 128v, col. 1) *nomisma* has the general meaning of "coin" (which was current even before 1092) and therefore cannot be used either to confirm or to reject the dating of 1075-85 ascribed to the manuscript. The word *charistike* in the context of the *Oneirocriticon* (Drexl 76, 1-3) indicates a royal gift other than cash; at least that is the meaning this word has in the tenth-century military treatise attributed to Constantine Porphyrogenitus; see Constantine VII (attributed to), *Three Treatises on Imperial Military Expeditions,* ed. and trans. J. Haldon (Vienna, 1990), p. 239. *Charistike* is also an administrative term indicating the giving of a monastery to a private person or institution on a conditional basis for a limited period. Its first known use is in a 908 act of Leo VI (r. 886-912), though the main evidence comes from a somewhat later period; its last known use is in the *diataxis* of Michael Attaleiates (1077). *Paroikos* (Drexl 52, 23-24) in classical Greek means "neighbor"; its meaning in the Old and New Testament is "foreign, alien," and in early Christian literature "sojourner, temporary resident." From the tenth century through to the end of the Byzantine Empire, *paroikos* signifies a dependent peasant, and this is the meaning of the word as it is used in the *Oneirocriticon.*
*criticon* is therefore the most important Byzantine dreambook in terms of both size and literary impact on the genre of dream interpretation.
The *Oneirocriticon* can be dated to the late ninth or the tenth century, a period of increased intellectual activity in Byzantium known as the Macedonian Renaissance. The role of the classical, Hellenistic and late-antique heritage on the literary output of this period has been discussed at length in scholarly publications. 21 Determining the *Oneirocriticon's* sources and especially its relationship to the second-century Greek text of Artemidoros can shed further light on its character.
Investigating the sources of the *Oneirocriticon* is important not only on account of the literary impact of the work and its significance for cultural and intellectual history, but also because of the potential misunderstandings that might arise when contemporary scholars use the text as a source for the study of everyday life in Byzantium. Before tapping into the wealth of information it provides on the material culture of the Middle Byzantine period and the mind and soul of the Byzantine, it is imperative to know where the interpretations offered in this dreambook came from and what changes, if any, they underwent in the process of transmission.
#### *Earlier Scholarship*
The discussion of the sources and authorship of the *Oneirocriticon* began in 1577, when Johann Loewenklau (Johannes Leunclavius, 1533-93) published his Latin translation of the Greek text. Loewenklau based his translation on a single Greek manuscript, then in the possession of the Hungarian humanist Janos Zsamboky (Johannes Sambucus)22 and today in the Osterreichische
*aprile 1969* (Rome, 1971), pp. 347-74; S. Collin-Rosel, ed. "Le *Liber Thesauri Occulti* de Pascalis Romanus," *AHDLMA* 30 ( 1964 ), pp. 111-98. The influence of the *Oneirocriticon ofAchmet* on a Slavonic dreambook has yet to be determined. The work survives in a Dubrovnik manuscript from 1520 and is attributed to the prophet Daniel. It is different, however, from other known Old Slavic versions of Daniel's dreambook. The immediate ancestry of the Dubrovnik dreambook can be traced to a combination of Byzantine and Latin (Italian) traditions (information provided by Adelina Angusheva and brought to my attention by Professor I. S evcenko ).
<sup>21</sup>For bibliography on the subject, see *ODB,* s.vv. "Encyclopedism" and "Renaissance."
<sup>22</sup> Johannes Sambucus or Janos Zsamboky of Nagyszombat(frnava (1531-84) was a Hungarian humanist, physician, and historiographer to the Viennese court. He traveled extensively in Italy in search of Greek manuscripts. In 1578 his collection of Greek manuscripts was purchased by the Imperial Library in Vienna and formed the basis for the Greek manuscript collection of today's Nationalbibliothek. See H. Gerstinger, "Johannes Sambucus als Handschriftsammler," *Festschriji*
#### 6 CHAPTER ONE
Nationalbibliothek *(Vindob. philos. et philol. gr.* 297), which was copied at the beginning of the sixteenth century.23 Loewenklau's translation was published under the title, *Apomasaris Apotelesmata sive de significatis et eventis insomniorum, ex Indorum, Persarum Aegyptiorumque disciplina* (The "Apotelesmata"24 of Apomasar, or on the Meaning and Consequences of Dreams from the Teaching of the Indians, Persians and Egyptians).25 The name of the author is given as "Apomasar" on the Greek manuscript itself, but its first folio with the title of the work is missing. The specification "from the Indians, Persians and Egyptians" was taken from the chapter headings in the Greek text. In his twelve-page preface to the translation, Loewenklau briefly explained how he found the manuscript and discussed dream interpretation in ancient Greek and Latin literature. He observed that in the text the Arabs were nowhere mentioned. The Persians were the "Magians" known from Herodotus; the Indians, also called "Gymnosophistae," were known from Plutarch's life of Alexander the Great. As for the Egyptians, Loewenklau admitted that he knew nothing about their science and sacred literature and would rather not repeat the abundant but unscholarly information that was circulating at the time. <sup>26</sup> Loewenklau did not know who the author Apomasar was, but according to information provided by J. Camerarius, he must have been the Arab scholar known in the West as Albumasar, "qui vero nomine laphar adpellabatur," or more accurately Abu Ma'shar Ja'far b. Mul)ammad b. 'Umar al-Balkhi, an Arab astrologer of the ninth century known for his copious quotations in his writings from Indian, Egyptian and Persian sources.27 Loewenklau notes that Albumasar was an Arab and a Muslim, and further remarks that the auth.or of
24 The Greek word *apotelesmata* literally means "outcomes," "results"; as a technical term in astrology it means the results of certain positions of the stars on human destiny.
<sup>25</sup>Further specifications: "Depromptus e Io. Sambuci v.c. bibliotheca liber, Io. Leunklavio interprete. Francofurti, excudebat Andreas Wechelus, 1577."
26 Praefatio, pp. 7-8.
<sup>27</sup>On Abii Ma'shar Ja'far b. Mul).ammad b. 'Umar al-Balkhi (787-886), see Sezgin, *GAS,* vol. 7, pp. 139-51 and 328-29; also *DSB,* s.v. "Abii Ma'shar al-Balkhi, Ja'far ibn Mul).ammad."
*der Nationalbihliothek in Wien* (Vienna, 1926), pp. 251-400; idem, "Aus dem Tagebuch des kaiserlichen Hofhistoriographen Johannes Sambucus (1531-84 )," *Sitzungsberichte der osterreichischen Akademie der Wissenschaften, phil.-hist. Klasse* 248:2 ( 1965); idem and A. Vantuch, "Die Briefe des Johannes Sambucus (Zsamboky) 1554-1584," *Sitzungsberichte der osterreichischen Akademie der Wissenschaften, phil.-hist. Klasse* 255 ( 1968); A. Vantuch, *Jan Samhucus* (Bratislava, 1975).
<sup>23</sup> F. Drexl, *Achmets Traumbuch, Einleitung und Probe eines kritischen Textes,* Inauguraldissertation (Freising, 1909), p. 15. For the most recent description of the manuscript, see H. Hunger, *Katalog der griechischen Handschriften der osterreichischen Nationalhibliothek,* 4 vols. in 7 (Vienna, 1961-1994), vol. I, p. 392.
the Greek text could not have been either, even if he was not a Greek, because of the obvious Christian character of numerous passages in the book. Loewenklau was certain that the true identity of the dreambook's author could been found, if only the first folio of Zsamboky's manuscript had not been missing.
The text in *Vindoh. philos. et philol. gr.* 297 begins with the last phrases from the fourth introductory chapter of the *Oneirocriticon,28* which is immediately followed by the interpretation of the first dream symbol, the Resurrection of the Dead, according to the Indians, the Egyptians and the Persians. The first page of the manuscript, as it now stands, contains the following chapter headings: "EK tffiv ivoffiv £pµriv£ia 7tEpt avacrtcl<JEW<;" (From the Indians, Interpretation on the Resurrection); "£K tffiv al. yunticov 7tEpt avaataaEcoi:; oµoicoi:;" (From the Egyptians on Resurrection, as Well); "E:K tffiv 7tEpaffiv 7tEpl ava<Jtcl<JECO<;" (From the Persians on Resurrection); "E:K tffiv ivoffiv nEpt napaoEiaou" (From the Indians on Paradise). In the space left after the third chapter heading, "f:K tffiV 7tEpaffiv 7tEpl avaata<JECO<;" (From the Persians on Resurrection), a later hand has added the word anoµciaapoi:; *(Apomasaros),* a Hellenized form of the name of Abu Ma'shar, which is also repeated in the upper margin of the same page. Many of his works had been translated into Greek before the end of the tenth century. 29 In
<sup>28</sup> The text begins in the middle of a word: "pwv O'lfEt<;· Ka\ yap rn\ au10<; rcoUT]v [sic] ElXE rc60ov rcpo<; mu<; 0£0u<; ... " (Drexl 3, 17 ff.).
<sup>29</sup>Only one of these translations has received a critical edition. This is Abu Ma'shar's handbook on casting horoscopes entitled *Kitiih tahiiwil sin/ a/-mawiilid* (Book of the Revolutions of the Years of Nativities). translated into Greek as *Peri tes ton et on enallages.* Only the first five books survive; critical edition Abu Ma'shar, *A/humasaris: De revolutionihus nativitatum.* ed. D. Pingree (Leipzig, 1967). On the date of the translation, see ibid., p. viii. The Byzantine astrological compendium, *Ta mysteria tou Apomasar* (The Mysteries of Apomasar), contains excerpts from further works by Abu Ma'shar: Book 1 of the *Mysteries* discusses elections and seems to be based on his *Kitah al-ikhtiyiiriit* (Book of Elections) and possibly also his *Kitah al-siham* (Book of Lots). Book 2 (partially published in *CCAG,* vol. 4, pp. 124-27 and *CCAG,* vol. 5: 1, pp. 142-55), is a work by his student Abu Sa'id Shadhan b. Bal)r, who quotes him profusely. The title of the Arabic work is *Kitah mudhakariit ahi Ma'sharfi asrar 'ilm al-nujiim wa-su'al ahi Sa'id Shadhiin ihn* Ba~r *'an ah/ Ma'shar wa-jawiihat ahi Ma'shar /ahu hi-ma ajahahu* (Book of the Deliberations of Abu Ma'shar on the Secrets of Astrology and the Questions of Abu Sa'id Shadhan b. Bal)r to Abu Ma'shar and the Answers of Abu Ma'shar by Which He Responded to Him). See M. Ullmann, *Die Natur- und Geheimwissenschaften im Islam* (Leiden, 1972), p. 323. On the date of the Byzantine translation, see D. Pingree, "Classical and Byzantine Astrology in Sassanian Persia," *DOP* 43 ( 1989), p. 227. Book 3 of the *Mysteries* consists for the most part of lengthy selections from Abu Ma'shar's *Kitah al-madkhal al-kahir 'ala 'ilm ahkam al-nujiim* (Great Introduction to the Science of Astrology). The Greek translation of yet another work by Abu Ma'shar, his *Kitah qiriiniit a/-kawakih fl a/-huruj al-ithna 'ashara* (Book of the Conjunctions of the Planets in the Twelve Signs) is included in the first book of another Byzantine astrological compendium, the *Introduction*
#### 8 CHAPTER ONE
them, he refers to the use of horoscopes by the Babylonians, Persians, Indians, and Egyptians. Conceivably, someone familiar with Abu Ma'shar's predilection for Indian, Persian, and Egyptian sources read the first chapters of the *Oneirocriticon* in the acephalous *Vindob. philos. et philol. gr.* 297, associated its contents with the Arab astrologer and wrote the name "Apomasaros" in the manuscript margin, where Loewenklau subsequently found it. Abu Ma'shar was very well known to Latin scholars, because his *Introduction to Astrology,* translated into Latin in the eleventh century, had become a standard textbook in the Latin West. That is why his name was familiar to Camerarius, Loewenklau' s informant. Loewenklau transferred the marginal note of the Greek manuscript to the title of his Latin translation, thus making Abu Ma'shar the author of the dreambook.
We know that Abu Ma'shar did write a book on dream interpretation, but it is now lost, 30 and there is reason to doubt that the *Oneirocriticon* is the translation of that work. First, the title of the lost dreambook, *Kitiib tafsir al-maniimiit min al-nujum* (Book on Dream Interpretation by the Stars), indicates that its contents combined dream interpretation with astrology and explained how to interpret dreams according to the phases of the moon and the position of the stars, a method known from texts that survive in Greek manuscripts, <sup>31</sup>but never mentioned in the *Oneirocriticon.* Second, the Indians to whom the genuine works of Abu Ma'shar refer are clearly Hindus, while the Indians quoted in the *Oneirocriticon* are Christians. Third, the *Oneirocriticon* includes a chapter on the interpretation of the planets in the following sequence: Sun, Moon,
*to Astrology* by AJ:imad the Persian (published in *CCAG,* vol. 2, pp. 123-30). For information on the works of Abii Ma'shar and their Greek and Latin translations, see Pingree's article in *DSB,* s.v. "Abii Ma'shar al-Balkhi, Ja'far ibn MuJ:iammad."
<sup>3</sup>° *Kitiih tafslr al-maniimiit min al-nujum;* see Ibn al-Nadim, *Kitiih a/Jihrist,* ed. G. FIUgel, J. Ri:idiger and A. MUiier (Leipzig, 1871 ), p. 277, I. 20. Ullmann suggested anew that the *Oneirocriticon* is a translation of this work by Abii Ma'shar, but the text he cited is the Latin translation by Loewenklau. See Ullmann, *Die Natur- und Geheimwissenschaften im Islam,* p. 324.
<sup>31</sup>For example, *Vat. gr.* 1056, fol. I 16r (published in *CCAG,* vol. 5:3, pp. 88-90); also, *Paris. gr.* 2417, fol. 165r (published in *CCAG,* vol. 8:1, pp. 152-53). Both are chapters from longer astrological works. Dreambooks that connect the veracity of dreams to the phases of the moon survive in several manuscripts, among them the following: *Athen. Bihl. Nat.* 1275 (ed. A. Delatte, *AnecdotaAtheniensia,* vol. 1 [Liege and Paris, 1927], pp. 182-83); *Athen. Bihl. Nat.* 1350 (ed. ibid., pp. 204-5); *Paris. gr.* 2315 (ed. ibid., p. 546); *Paris. gr.* 2511 (ed. ibid., pp. 525-26; and ed. F. Drexl, *Bayerische Blatter fur das Gymnasialschulwesen* 1923, pp. 214-15); *Vat. gr.* 342 and *Bero/. gr.* 168 (ed. S. G. Mercati, "Onirocriticon lunare secondo i codici Vaticano greco 342 e Berlinense greco 168," *BZ* 32 [1932], pp. 263-66); *Vat. gr.* 573. See Guidorizzi, "I prontuari oniromantici bizantini," p. 139; also T. Gregory, "I sogni e gli astri," *I sogni nel medioevo,* ed. idem (Rome, 1985), pp. 111-48.
Venus, Mercury, Mars, Jupiter, Saturn. 32 In antiquity and the Middle Ages the planets were given several different sequences, but the one that appears in the *Oneirocriticon* does not correspond to any of them. 33 These planets were evidently listed in the Arabic source used by the *Oneirocriticon* by someone without astrological expertise, according to the order in which he could remember them. It is improbable that an accomplished astrologer such as Abu Ma'shar would overlook the basic principles of his science. Last but not least, the third introductory chapter of the *Oneirocriticon,* which is missing from *Vindob. philos. et philol. gr.* 297 and is ascribed to the dream interpreter of the Persian king, denounces astrology and declares that dream interpretation is an easier and more accurate way to predict the future. 34 It is unlikely that an astrologer would quote such a condemnation of his art in his own work.
Loewenklau predicted that the dreambook he translated would not be pleasing to "good and erudite men," for it is inferior to the dreambook of Artemidoros, which is on a similar subject, although Apomasar wrote long after Artemidoros. He observed that, like Artemidoros, Apomasar suggested different interpretations of a dream depending on the social situation of the dreamer, as well on the season of the year and time of day at which a dream was dreamt. Loewenklau agreed that although much of the text's contents was pure superstition, it should be tolerated, for, as he put it, "Don't we read pagan and foreign texts that are filled with superstitions? And don't we forgive their superstitions, like a mark on the skin that does not mar the body?"
Twenty-six years later, in 1603, the Greek text was printed for the first time. The editor was N. Rigault (Rigaltius) who published a number of dreambooks in a single volume. 35 In his two-page introduction, Rigault admitted that he knew little about the author, Achmet. He was an Arab physician, none other than the one whose seven books on medicine were mentioned by Gesner,
<sup>12</sup> Drexl 129, 12-18.
<sup>&</sup>quot; See 0. Neugebauer. *The Exact Sciences in Antiquity,* 2nd ed. (Providence, R. I., 1957), p. 169.
<sup>14</sup> Drexl 2, 25-3, 11.
<sup>15</sup>*Artemidori Daldiani et Achmetis Sereimif Oneirocritica. Astrampsychi et Nicephori versus etiam Oneirocritici. Nicolai Rigaltii ad Artemidorum Notae* (Paris, 1603). This was published "ex officina Claudii Morelli," but the same volume, with the same title and contents was printed that same year by another Parisian printer, Marcus Orry ("Apud M. Orry"). See V. F. Goldsmith, *A Short-Title Catalogue of French Books, 1601-1700, in the Library of the British Museum* (London, 1969), nos. 913 and 914.
a piece of information which is correct according to Ianus Antonius Saracenus in his notes to Dioscorides. 36 Rigault added that twenty-six or twenty-seven years earlier Loewenklau had published a Latin translation, but the text was in several passages corrupt and mistakenly attributed to Apomasar, as Loewenklau himself later admitted in his *Annales Sultanorum Othmanidarum.* Rigault concluded:
And, so that I do not conceal or disguise anything, I have adduced two codices of the royal library, but there was nowhere I could obtain the name of Achmet with certainty, for they are both aKE<j>aA.ot. Except that in one of the two an inscription with the name of Achmet *["Achmetis titulus"]* was placed before the text by a more recent hand. Moreover, in the Greek copy, from which Leo Tuscus translated the text into Latin and dedicated [his translation] to Hugo Echerianus17 around the year 1160, the name of the author was missing, as it is possible to deduce from the Italian translation by Tricassus Mantuanus. From that translation I have transcribed the prologue which you see here, not because I believe that this prologue was
I was unable to identify !anus Antonius Saracenus, aside from a single reference to one of his publications, *De peste commentarius* (Lyons, 1572); see *Short Catalogue of Books Printed in France and of French Books Printed in Other Countries from 1470 to 1600 Now in the British Museum* (London, 1924). H. M. Adams, *Catalogue of Books Printed on the Continent of Europe. 1501-1600 in Cambridge Libraries.* vol. 2 (Cambridge, 1967), p. 387, mentions a work published by Johannes Antonius Saracenus, *Summa Pont. Leoni X. J.A.S. equitis et oratoris pro repuh. Senen.* (Rome, [1513?]); however, since Janus is not identical to Johannes, and the two works are rather removed from each other in terms of date, place of publication, and subject matter, it is possible that *De peste* and the *Summa* were not by the same author.
17 *Sic.* instead of "Etherianus."
<sup>16</sup> There is no evidence in the Greek text itself that would connect it to medicine, except for the fact that in some of its later manuscripts it is copied together with medical texts. This is done in *Cantahrig. (Trinity College) gr.* 1386 (0 8.1 l, 6102); *Paris. gr.* 2419; *Bononiensis (Bihl. Univ.)* 3632. Rigault's observation only shows how much in his own mind dream interpretation was associated with medicine. Konrad Gessner's first volume of the *Bihliotheca Universalis.* which is an index of authors, includes the following entries: "Achmeti opus de proportionibus singularibus. Achmeth medici filij Habramij peregrinantium viatica libris septem Graeco sermone descripta. Habet eos hoc tempore Venetijs illustrissimus vir D. Diegus Hurtadus a Mendozza Caesareae majestatis legatus" (C. Gesner, *Bihliotheca Universalis* !Zurich, 1545], pt. l, *Quae omnis generis authorum nomina cum lucuhrationihus singulorum iuxta literarum seriem proponit,* p. 3). The second entry clearly refers to Abil Ja'far AI:imad b. Ibrahim b. Abl Khalid al-Jazzar (on him, see below, n. 43). I was unable to locate any reference to any other author named A ch met in Gesner. The second volume of the *Bihliotheca Universa/is.* the *Pandectarum sive partitionum universa/ium lihri XX!. sive Bihliothecae Tomus II* (Zurich, 1548), contains a classified arrangement of the contents of the *Bihliotheca* and supplements. Only *lihri* I-XIX were issued; *fiber* XX, which dealt with the subject of medicine, was never published; Ii *her* XXI, which dealt with theology, was published separately in 1549; see J. Christian Bay, "Conrad Gesner (1516-1565), The Father of Bibliography: An Appreciation," *The Papers of the Bibliographical Society of America* 10:2 (April 1916), p. 66. This means that Gesner never published any further information on "Achmeth," nor did he connect Abil Ja'far AI:imad b. Ibrahim b. Abl Khalid al-Jazzar with dream interpretation.
written by Achmet, for I rather think it is by Leo, but in order that nothing should seem to be missing from our edition.18
Soon after the first edition of the Greek text appeared, Barth observed that "Achmet, son of Sereim, is not the author of the Greek compilation, but rather a barbarian *(barbaricus)* dream interpreter." 39 Somewhat later, in 1688, Du Cange rejected Achmet's authorship of the Greek dreambook. 40 The problem was further discussed by Peter Lambeck in his commentary on the holdings of the Imperial Library in Vienna, published some seventy years after the appearance of the first edition of the *Oneirocriticon.41* In discussing *cod.* 143,42 Lambeck pointed out that Rigault's identification of Achmet based on Gesner was wrong. Rigault, according to Lambeck, confused Achmet, a dream interpreter by profession, "with Achmet the Syrian, son of Abraham, grandson of Khalid, a physician by profession."43 He then ran a chronological check
19 Kaspar von Barth, *Casparis Barthi Adversariorum commentariorum lihri LX* (Frankfurt, 1624), vol. I, bk. 31, chap. 14.
40 Du Cange, s.v. "Maµouv."
<sup>41</sup>P. Lambeck, *Commentariorum de augustissima Bihliotheca Caesarea Vindohonensi,* 8 vols. (Vienna, 1665-1679), vol. 7, codd. CXLII-CXLVI, cols. 562-88.
*<sup>42</sup>*Today *Vindoh. philos. et phi/of. gr.* 111 (13th century), which includes the opening chapter of the work, entitled "i':v ov6µan toil itmpoi; Kat WU 1.JlOU Kat toil eryiou 7tV£Uµatoi;, 13tl3Aiov ovnpoKpUlKOV 07t£p crnvfi~i:: Kat CJ1JVEta~£ 'Axµi::'t 1.Jtoi; L11pdµ, 0 OV£tp0Kph11i; toil 7tpolto1J avµ(3ouA.ou Maµouv" (In the name of the Father and of the Son and of the Holy Spirit, Book on dream interpretation which was put together and composed by Achmet, son of Sereim, dream interpreter to Caliph Mamoun). For a detailed description, see Hunger, *Katalog der griechischen Handschr(ften der osterreichischen Nationalhihliothek,* vol. I, p. 222.
41 This is Abu Ja'far AI:imad b. Ibrahim b. Abi Khalid al-Jazzar (d. 979). He was from Kairouan, Tunisia. His *Kitah i'timad a/-adwiya a/-muji-ada* (Book of Reliance on Simple Medicines) and *Zad a/-musafir* (Supplies for a Traveler) were translated into Greek presumably by Constantine the African. Lambeck must have thought he was Syrian because the title of the Greek translation of *Zad a/-musafir* in *Vindoh. med. gr.* 30 reads as follows: Afrtll itoAv1puA.11wi; *[sic* I ii Lupwv 13il3A.oi;. Bil3A.oi; A.i::yoµi':v11 n':i 'E\$68ta wu 'Aito811µouvwi;, auni::0i::"iaa itapa "Eitpou Baysa<IJap *wii* "'El311v "EA.ys11s<ip, µnal3Af10i::"iaa di; 1T,v i':Ucilia yA.ronav itapa Kwva1av1ivou itpw-1aa11Kpi *wv wii* 'P11yi vou. Au111 13il3A.oi; ~v ii itciA.m 0puUouµi':v11, ii rnt LUptKoi; i::ua16xwi; K£KA11µi':v11 (This is the famous book of the Syrians. The book called "Supplies for the Traveler," composed by Abu Ja'far b. al-Jazzar, translated into the Greek language by Constantine of Rhegion, the *protasecretis.* This is the long famous book, which is appropriately called Syrian). Moreover, the index of the work in the same manuscript is titled LUV'tayµa lii':A.toi; [lii':Atov? I EK LUpwv ao<iJou yi':voui; (Composition of the book from the wise nation of the Syrians). See Lambeck, *Commentariorum,* vol. 6:2, cols. 285 ff. The titles in *Vindoh. med. gr.* 30 can be found in Lambeck's description of manuscript 29, which is today *Vindoh. med. gr.* 20. For a recent
<sup>18</sup> On the second unnumbered page of the preface addressed to the reader following the title page: "Achmetis f. Seirim Oneirocritica, nunc primum graece in lucem edita. Ex bibliotheca Regis Christianissimi."
#### 12 CHAPTER ONE
using chapter 19 of the *Oneirocriticon* (Drexl 15, 18ff.), where Achmet is presented as a dream interpreter to Caliph al-Ma'mun, who died in 833. Lambeck cited a number of Greek sources, as well as *Georgii Elmacini Historia Saracenica,* published in 1625, that attest to al-Ma'mun's interest in Greek learning.44 However, Lambeck observed that the Arab revolt that overthrew the reigning caliph described in chapter 192 (Drexl 148, l 7ff.) fits better into the reign of al-Amin, the brother of al-Ma'mun, who was deposed by al-Ma'mun's revolt in 813. Since in chapter 19 Achmet speaks in the first person, Lambeck believed that the text furnishes two chronological indications that allow us to date it to the beginning of the ninth century and, more specifically, to the reign of al-Ma'mun. Lambeck concluded that, as is clear from the *Oneirocriticon* itself, Achmet, though the dream interpreter to the Muslim al-Ma'mun, was himself a Christian. He must also have been of Greek origin, since he wrote the *Oneirocriticon* in Greek, a language that al-Ma'mun, to whom the *Oneirocriticon* is dedicated in the first chapter, knew well.45
Lambeck went on to comment on three other codices that contain Achmet's text.46 The last one is the codex that once belonged to Zsamboky, the one from which Loewenklau had made his Latin translation attributed to Apomasar. Lambeck pointed out that a note anoµacraprn; next to the chapter heading *"£K* 'tcOV IlEpcr&v 7t£pt avaO"'tcl0"£(l)-;" (From the Persians on Resurrection) must have been responsible for the mistaken attribution.
There was considerable confusion and disagreement about the identity of Achmet among the early scholars. One reason for their difficulties was the dearth of available publications on Islamic history and literature. The last volume of Lambeck's *Commentarii* was published in 1679. Eighteen years later, in 1697, a milestone of Oriental studies and a forerunner of today's *Encyclopaedia of Islam,* the *Bibliotheque orientate* by Barthelemy d'Herbelot
description of both manuscripts, see Hunger, *Katalog der griechischen Handschriften der osterreichischen Nationalhihliothek,* vol. 2, pp. 64-66 and 80-82; for lbn al-Jazzar, see Sezgin, *GAS,* vol. 3, pp. 304-7; for the Greek translations, see H. Hunger, *Byzantine /ogotechnia. He logia kosmike grammateia ton hyzantinon,* vol. 3 (Athens, 1994), pp. 134-35 [Greek translation with updated bibliography of *Die hochsprachliche profane Literatur der Byzantiner* (Munich, 1978)].
<sup>44</sup> Lambeck, *Commentariorum,* vol. 6:2, col. 574: *"Georg* ii *Elmacini Historia Saracenica,* a Thoma Erpenio in Latinam linguam translata, & a Jacobo Golio A.C. 1625 Lugduni Batavorum typis Erpenianis Arabice ac Latine in folio edita."
<sup>45</sup> The first chapter of the *Oneirocriticon* does contain a dedication to a *despotes,* but no name is given. Though al-Ma'mun was interested in Greek learning and sponsored translations of Greek works into Arabic, we have no reason to assume that he knew Greek himself.
<sup>46</sup> CXLIV, CXL V and CXL VI, today *Vindoh. philos. et philol. gr.* 162, 287 and 297.
(1625-95) appeared.47 It was published posthumously, its author having died two years earlier, and was a work of unprecedented breadth. It is, however, marred by errors that are only too easy to criticize today, after three hundred years of additional scholarship. Among them, it lists at least three different figures by the name of Sirin: first, "Ebn Sirin ... , qui a ete fort estime pour l 'austerite de sa vie" (no connection with dream interpretation mentioned); second, "Abou Abdallah Mohammed Ben Sirin," also known as "al-Salemi," who wrote a dreambook in fifty chapters (Paris, Bibliotheque du Roi, ms. 1034) based on that of "Abou Ishak al-Kermani,.; third, "Mohammed Ben Sirin," who made an Arabic translation and commentary of the Greek dreambook of Artemidoros.48 Mul:iammad b. Sirin was a Muslim scholar of the seventh century known for the piety and austerity of his life. From the tenth century onward he was credited with the authorship of a number of dreambooks. The translator of Artemidoros into Arabic was I:Iunayn b. Isl:iiiq (9th century). D'Herbelot's misleading information passed from one scholarly work to the next, and injected confusion into subsequent discussions on the relationship of the Greek *Oneirocriticon* to Arabic dreambooks for more than two hundred years.
Between 1766 and 1782, Lambeck's *Commentarii* was published again, this time augmented and improved with notes by Adam FrantiSek Kollar (alias Colarius, 1718-83). The new editor consulted d 'Herbelot and repeated what he said, namely, that Ibn Sirin, also known as "Alsalemi," wrote a dreambook (the one that survived in Bibliotheque du Roi in Paris as ms. 103449) based on the work of al-Kirmani. Kollar observed that the Arabic dreambook in fifty chapters differed from the Greek *Oneirocriticon,* which is divided into a larger number of chapters, and while Ibn Sirin was an Arab and a Muslim, Kollar
<sup>47</sup> B. d'Herbelot, *Bihliotheque orientate, ou Dictionaire universe/, contenant genera/ement tout ce qui regarde la conoissance des peuples de /'Orient* (Paris, 1697). The work was reprinted several times in the course of the 18th century.
<sup>48</sup> A full account of d'Herbelot's entries on dream interpretation and correction of the mistakes can be found in Steinschneider,"Ibn Shahin und Ibn Sirin zur Literatur der Oneirokritik,,. *ZDMG* 17 (1863), pp. 234-35.
<sup>49</sup> This is the call number given by d'Herbelot. However, the manuscript had already been given a new call number, 1212, and was described under the new call number in the catalogue that was current when Kollar was writing. According to Steinschneider, "Ibn Shahin und Ibn Sirin,,. p. 235, d'Herbelot's 1034 had become 1210 (today it is 2742). However, 1210 is titled *Ta'hlr al-Ru'ya* and its older number was 4573; 1212 (today 2744) bears the older number 1034 and a Latin note stating that its title is "Ketab alescharat fi elm al-ebarat," it is divided into fifty chapters, and is the work of "Abou Abdallah Mohammed Ben Sirin,,. based on the method and principles of "Abou Ishak al-Kermani.,.
#### 14 CHAPTER ONE
considered the Greek work to be a compilation that used Achmet, Apomasar and other Muslim writers, but was written by a Christian Greek who lived later than his Arab models. Kollar believed this compiler to have been Symeon Seth (or his son) who flourished in the eleventh century and who knew Arabic. <sup>50</sup> A number of works by Symeon Seth had already been published before Kollar's time.<sup>51</sup>
It is easy to exclude Symeon Seth as a candidate, however. He was a scientist and writer at the court of the emperor Alexios I Komnenos ( 1081-1118), on whose orders he translated *Kalila wa-Dimna,* a collection of originally Indian fables, from Arabic into Greek. 52 He was also familiar with Arabic medicine, as is evident in his medical writings. Not much is known about his life, though it is certain that he flourished in the second half of the eleventh century.53 Kollar suggested that he might have been the compiler of the *Oneirocriticon* because he was known to have translated works from Arabic into Greek and to have lived after the reign of Caliph al-Ma'mun, who is mentioned in the Greek dreambook. However, Symeon's style is more sophisticated than that of the *Oneirocriticon,* as is apparent not only from his grammar and syntax but also from the allusions in his texts that indicate familiarity with ancient Greek literature.54 In addition, the *Oneirocriticon* enjoyed a wide circulation
52 The work is known in Greek as *Stephanites kai Ichnelates;* its title informs us about the circumstances of its translation. See Symeon Seth, *Stephanites und Jchnelates: Uherliefesrungsgeschichte und Text,* ed. L. 0. Sjoberg (Stockholm, 1962), p. 151: ~uyypa<)n'l ... £~ EAAT1Vtcr0Eicra oi:: EV Krovcrtavtt VO'U1tOAEt 7tpocrta~Et toii aotoiµou ~acrtA.Ero~ 1rnpoii 'AA.E~iou toii Koµv'Tlvoii (A work ... translated into Greek in Constantinople by order of the praiseworthy emperor lord Alexios Komnenos).
<sup>53</sup>See *ODB,* s.v. "Seth, Symeon."
54 Besides the New Testament and the Psalter, *Stephanites kai Ichnelates* includes quotations from Choerilos Samios, Demosthenes, Galen, Hesiod, Homer *(Iliad),* Pindar and Theognis; see Sjoberg, *Stephanites und lchnelates,* p. 246. The Greek translation of *Kalila wa-Dimna* was rather sophisticated, as shown in the recent study by H. Bassoukos-Kondylis, *Stephanites kai lchnelates, traduction grecque (Xie siecle) du livre Kalila wa-Dimna d'lhn al-Muqaffa' (VIiie siecle ). Etude lexico/ogique et litteraire* (Louvain, 1997); for a brief presentation of its arguments, see eadem, *"Stephanites kai /chnelates.* Traduction grecque de *Kalila wa-Dimna," Museon* 103 (1990), pp. 139-49. In his *Conspectus rerum naturalium (EvvalJIL<; rwv ¢vcmcwv),* Symeon Seth refers to Plutarch, Aristotle and the Peripatetics, the Stoics, John Philoponos, Proclus, Ptolemy,
<sup>50</sup> *Petri Lahecii Hamhurgensis Commentariorum de augustissima Bihliotheca cresarea vindohonensi,* ed. A. F. Kollar, 8 vols. (Vienna, 1766-82), vol. 7, cols. 561-64, n. (A).
<sup>51</sup>For early publications of Symeon Seth's works, see K. Krumbacher, *Historia tes hyzantines logotechnias,* vol. 2, trans. S. Soteriades (Athens, 1900), pp. 420-21. The more recent work by Hunger, *Die hochsprachliche profane Literatur der Byzantiner,* gives bibliography that postdates Kollar's. Kollar mentions that Fabricius had published Seth's writings in *Bihliotheca Graeca,* 14 vols., 1st ed. (Hamburg, 1708-28), vol. 13.
in the 1080s, if not earlier. An abridged version of it appears in *Paris. Suppl. gr.* 690; it is quoted in the eleventh-century *Florilegium Baroccianum,* and its excerpts appear in the margins of the eleventh-century *Laurentianus Plut.* 87, 8. In short, both chronological and stylistic evidence tells us that Symeon Seth could not have been the compiler of the *Oneirocriticon. <sup>55</sup>*
Meanwhile, a new catalogue of the Arabic manuscripts in the Bibliotheque du Roi had been published. In the *Catalogus Codicum Manuscriptorum Bihliothecae Regiae* that appeared in 1739, there are no remarks regarding the Greek connection of the Arabic manuscript mentioned by Kollar. However, in the entry on manuscript 1210 (today BN *arahe* 2742) we read the following:
ms. 1210: A bombycine codex that arrived in the library of Colbert through the services of Wansleben56 in the year 1676. It contains the dreambook *(oneirocriticon)* of Muhammed, son of Sirin, who was born in Amida in Mesopotamia and died on the last day in the year of the Hijra 762. This work is exactly the same *(idem omnino est)* as the one produced in Greek by [the author] by the name Achmet son of Sirin. <sup>57</sup>
The information incorporated into this entry is probably related to two Latin notes on the flyleaves of the manuscript. The first note, inscribed "Joseph Ascari 1735," informs us that Mut.iammad b. Sirin was from the Mesopotamian city of Amida (today Diyarbakir in Turkey), on the banks of the river Tigris. He died in 762 H. (1358 A.o.). On the next page a second Latin note by a different hand reads: "Cod. 4573, Oneirocriticon autore Muhamed f. Sirin. Idem qui graece editus est nomine Achmet f. Sirim,"58 which must have been
Plato, and Stephanus of Alexandria, and shows his acquaintance with their theories; in his *De utilitate corporum caelestium (nt:pi xpEim; rwv ovpaviwv awµarwv)* he mentions Gregory of Nyssa, Plotinus, Ptolemy, Poseidonius, Hippocrates, Galen, and Aristophanes; see A. Delatte, *Anecdota Atheniensia,* vol. 2 (Liege and Paris, 1939), p. 127.
<sup>55</sup> Kollar added that there might soon be a new edition of the *Oneirocriticon.* Jacob Tollius (1630-96) had compared Rigault's text with the four manuscripts in the Viennese Imperial Library, and had made numerous emendations. but had died before publishing his improved edition. His manuscript was in the hands of his heirs and Kollar thought they would probably have it published, though Kollar himself saw no need for a new edition, on account of the eminently superstitious nature of the *Oneirocriticon.* In fact Tollius's work was never printed.
<sup>56</sup> Johann Michael Wansleben (1635-79) was a German Orientalist who served as Colbert's envoy to Egypt between 1672 and 1676, when he was recalled to Paris. On his life, see A. Pougeois, *Vans/eh, savant orientaliste et voyageur: sa vie, sa disgrace, ses oeuvres* (Paris, 1869).
<sup>57</sup>*Catalogus Codicum Manuscriptorum Bihliothecae Regiae* (Paris, 1737), p. 230.
<sup>58 &</sup>quot;Cod. 4573, dreambook by the author Mu~ammad, son of Sirin. The same who is published in Greek by the name Achmet, son of Sirim."
what misled the cataloguer into stating that the work, and not simply the author, was the same as in the Greek edition.59
The catalogue entry and the notes on the manuscript further complicate the problems of authorship and date. If the author lived in the fourteenth century, as is claimed in the note on the Paris manuscript and catalogue, how could he have served Caliph al-Ma'mun, who lived in the ninth? Clearly, there is something wrong either with the dates given for the author, or with the identification of the two texts, or both.
Wlistenfeld, who included an entry on Ibn Sirin in his *History of the Arab Physicians,60* was the first to have noticed that dreambooks by later authors were attributed to lbn Sirin (d. 728). <sup>61</sup>The sources and authorship of the *Oneirocriticon* were discussed again by N. Bland, in an article that appeared in 1856.62 Bland drew the facts for his discussion primarily from Persian works on dream interpretation that had been translated from the Arabic. He gave a detailed exposition of the theoretical principles of Islamic dream interpretation according to the dreambooks with which he was familiar. He pointed out which of these principles were inspired by Greek philosophy and medicine63 and made cursory mention of some similarities between Islamic dreambooks and the work of Artemidoros. 64 However, he did not investigate the subject, limiting himself only to a note: "[Artemidoros is] the authority most frequently named by Arabian writers on Tabfr, and the resemblance of the two systems is the most strongly traced in his writings."65 In a long appendix to the article, he drew up a preliminary list of Islamic works on dream interpretation that he knew either from manuscripts (mostly Persian or Turkish) or, more often, from the list of sources given by the authors of dreambooks that he had read. He also appended the table of contents of a number of Islamic dreambooks that he had examined. Finally, he wrote two pages "On
<sup>59</sup> About the Latin notes on this manuscript, see also Fahd, "L'oniromancie orientale," pp. 364-65.
<sup>6</sup>° F. Wtistenfeld, *Geschichte der arabischen Aerzte und Natwforscher* (Giittingen, 1840).
<sup>61</sup>Note again the connection between dream interpretation and medicine; Wtistenfeld, ibid., pp. 10-11, no. 20, mentions that Ibn Sirin was a well-known authority on Islamic law, *}Jadith,* and dream interpretation, but cites no medical writings; nor does he refer to any medical activities by Ibn Sirin.
<sup>62</sup> N. Bland, "On the Muhammedan Science of Til.bfr, or Interpretation of Dreams," *IRAS* 16 (1856), pp. 118-71.
<sup>63</sup> Ibid., pp. 124, 128, 142.
<sup>64</sup>Ibid., p. 138.
<sup>65</sup> Ibid., p. 124, n. 3.
the pretended Greek and Latin version oflbn Sirin's Oneirocritics," in which he mentioned the twelfth-century Latin translation by Leo Tuscus and that of Loewenklau, as well as the edition of the Greek text by Rigault and the problematic attributions made until then. He observed that Ibn Sirin was not a contemporary of Caliph al-Ma'mun and deduced that the "Sereim" of the Greek *Oneirocriticon* cannot be identified with the Arab Ibn Sirin. Moreover, the references in the Greek text to Christian notions show that the work was not composed by a Muslim. Bland judged that "the arrangement [of the *Oneirocriticon],* as well as the whole character of the composition, is far from Oriental"66 and concluded:
On the whole it is reasonable to suppose the Greek to have been the original of the work, or perhaps that it was compiled in Arabic by some Christian, probably of Syria, from various native sources, and of these, especially, the Khabar al Mamuni [a work that is known only by its title], which would account for the frequent mention of Mamun. lbn Sfrfn's name may have been assigned to it as its author, from the numerous interpretations it contains of his.67
In 1863 M. Steinschneider, a scholar known for the enormous breadth of his studies on the medieval translations from other languages into Arabic and vice versa, published a response to Bland.68 In it he attempted to ascertain the dates and identities of the Arab writers who had been proposed as the authors or sources of the Greek text. He was very careful not to repeat the suppositions and mistakes made by other scholars, but he briefly reviewed the previous literature on the identity of Ibn Sirin and pointed out the mistakes made in d'Herbelot's *Bibliotheque orientale* that had subsequently been perpetuated in the works of scholars such as Kollar and Bland. He also established the source of Artemidoros's influence on Arabic dream interpretation.69 According to d'Herbelot, Ibn Sirin had prepared an Arabic translation of and commentary on the work of Artemidoros. Steinschneider, on the other hand, quoted the tenth-century bibliographical work, *al-Fihrist* of Ibn al-Nadim, which stated that I:Iunayn b. Isl).aq had translated the work of Artemidoros in five books into Arabic.70 As for the Greek *Oneirocriticon,* Steinschneider mentioned the entry of the 1739 Paris catalogue of Arabic manuscripts, according to which
<sup>66</sup> Ibid., p. 170.
<sup>67</sup> Ibid., p. 171.
<sup>68</sup> Steinschneider, "Ibn Shahin und Ibn Sirin," pp. 227-44.
<sup>69</sup> Ibid., pp. 235 and 241.
<sup>70</sup> Ibn al-Nadim, *Kitiih al-fihrist,* ed. G. Fliigel, et al., p. 255, II. 9-10.
#### 18 CHAPTER ONE
the text in ms. 1210 was identical with the Greek work. Steinschneider could not consult the manuscript himself, but suggested that the Arabic text might be a translation from the Greek and called for its re-examination by an Arabist to end "the doubts and confusion that have reigned until most recently."71 He prudently concluded that several questions concerning Arabic dream interpretation could only be answered by examining the beginnings of this kind of literature in Islam and carefully investigating not only the contents of dreambooks, but also their connection to the whole of Arabic literature.
Between 1883 and 1895 a new catalogue of Arabic manuscripts in Paris was prepared by De Slane. The old Bibliotheque du Roi had become the Bibliotheque Nationale and new call numbers had been assigned to the older acquisitions (it was then that the old Bibliotheque du Roi ms. 1210 became BN *arabe* 2742). It is not clear whether De Slane was aware of the debate surrounding this manuscript, or whether he examined it with Steinschneider's suggestion in mind; in any case, the new catalogue included no comment on the alleged connection between BN *arabe* 2742 and the Greek *Oneirocriticon.72*
In his 1898 collation of Rigault' s 1603 edition of the *Oneirocriticon* with further Greek manuscripts, C.-E. Ruelle states that "Achmet or Ahmed Abou-Mazar, son of Seirim, an Arab physician, lived at the beginning of the ninth century. He was a Christian and fulfilled the function of a dream interpreter in Babylon, at the side of Caliph Al Mamoun, the son of the famous Haroun al Raschid."71 In 1900, F. Cumont briefly examined the question whether 'Axµchri-; 6 11£pcrri-; (A!Jmad the Persian), author of a collection of astrological texts, was the same as the author of the *Oneirocriticon.* His conclusion was negative. 74 In 1909, Drexl, who would later prepare a critical edition of the Greek *Oneirocriticon,* published the preliminary results of his research on the text. 75
<sup>71</sup> Steinschneider, "lbn Shahin und lbn Sirin," p. 237.
<sup>72</sup> W. Ahlwardt, *Verzeichnis der arahischen Handschriften der kg/. Bihliothek zu Berlin* (Berlin, 1887-99), nos. 4263-89, is likewise silent about Ibn Sirin's Greek connection.
<sup>73</sup> C.-E. Ruelle, "La clef des songes d' Achmet Abou-Mazar. Fragment inedit et bonnes variantes," *Revue des hudes grecques* 7 (1894), p. 305. It is evidently from Ruelle's article that 0. Gotthardt, *Uher die Traumhiicher des Mittelalters,* Konigliches Luthergymnasium zu Eisleben, no. 334 (Eisleben, 1912), p. 3, copied the following information: "Achmet ist im 9. Jahrhundert Leibartzt und zugleich Hoftraumdeuter des Kalifen Mamun (oder al-Mamun) gewesen."
<sup>74</sup> See *CCAG,* vol. 2, p. 122, n. I. This theory is given more credence than the opinions expressed by other scholars on the identity of the *Oneirocriticon' s* author by Thorndike, *History of Magic and faperimental Science,* vol. 2, p. 292. On Achmet the Persian, see D. Pingree, "Historical Horoscopes," *JAOS* 82 (1962), pp. 487-502.
<sup>75</sup> Drexl, *Achmets Traumhuch.*
In the chapter, "Is Achmet the author of the Greek dreambook?,"76 Drexl reviewed the previous literature on the subject77 and concluded, not without frustration: "I am not risking a final conclusion regarding this question now; possibly [such a conclusion] is completely impossible. "78 In his 1923 introduction to the critical edition of the Greek text, Drexl mentioned the alleged identity of BN *arabe* 2742 with the *Oneirocriticon.* However, he interpreted De Slane's silence on the question as an indication that the two works were different. 79 His conclusion was that the attribution of the *Oneirocriticon* to Achmet was spurious.80
Fahd's research facilitated enormously all future examination of the connection between the Greek *Oneirocriticon* and Arabic dream interpretation. In his *Divination arabe* and in a number of articles, 81 Fahd examined the
79 Drexl, *Achmetis Oneirocriticon,* Prolegomena, p. ix.
80 A. Fischer, "Die Quitte als Vorzeichen bei Persern und Arabern und das Traumbuch des 'Abdal-B.ani an-Nabulusi," *ZDMG* 68 (1914), 275-325, offered some clarifications regarding the dates and identities of authors and works on Islamic dream interpretation, as well as references to available publications on the subject. However, the problems presented by the Greek *Oneirocriticon* were not discussed at any length. The Greek text is briefly mentioned on p. 304, n. 2. Abdel Daim, *L'oniromancie arabe d'apres lbn Sirin* (Damascus, 1958), p. 26, revisited the question by examining all previous scholarship on the topic and suggesting that the *Oneirocriticon* had been written by I:Iunayn b. lsJ:iaq, the famous Christian physician and translator from Arabic into Greek who was active in the court of 9th-century Baghdad.
81 T. Fahd, *La divination arabe. Etudes religieuses, sociologiques et folkloriques sur /e milieu natif de /'Islam* (Leiden, 1966; rpt. without the bibliography, Paris, 1987); idem, ed. *Artemidore d'Ephese, Le livre des songes. Traduit du grec en arabe par Jjunayn b. lslJiiq (mort en 2601873)* (Damascus, 1964); idem, "Les songes et leur interpretation selon l'Islam," *Sources Orienta/es* 2. *Les songes et /eur interpretation* (Paris, 1959), pp. 127-58; idem, "La traduction arabe des *Oneirocritica* d'Artemidore d'Ephese," *Arabica* 7 (1960), pp. 87-89; idem, "Le reve dans la societe musulmane du Moyen Age," in *Les reves et !es socieres humaines,* ed. G.E. von Grunebaum and R. Caillois (Paris, 1967), pp. 335-65; Spanish trans., *Los suefios y las sociedades humanes* (Buenos Aires, 1964), pp. 193-230; English trans., *The Dream and Human Societies(Berkeley,* 1966), pp. 351-79; idem, "Les corps de metiers au IVe/Xe siecle a Baghdad d'apres le chapitre *Xlld'al-Qiidirift t-ta'bir* de Dinawari," *JESHO* 8:1 (1965), pp. 186-212; idem, "L'abeille en Islam," *Traite de bio/ogie de l'abeille,* ed. R. Chauvin, vol. 5: *Histoire, Ethnographie et Folklore* (Paris, 1968), pp. 61-83; idem, "Ja'far a~-~adiq et la tradition scientifique arabe," *Le Shi'isme imamite* (Paris, 1970), pp. 131-142; idem, "La connaissance de l'inconnaissable et l'obtension de !'impossible dans la pensee mantique et magique de !'Islam," *Bulletin des Etudes Orienta/es* 44 (1992/93), pp. 33-44; idem, *EI2,* s.v. "ru'ya,"; idem, "L'oniromancie orientale"; idem, "Anges, Demons et Djinns en Islam," *Sources Orienta/es* 8 (Paris, 1971), pp. 155-213; idem, "I:Iunayn ibn IsJ:iaq est-ii le traducteur des *Oneirocritica* d'Artemidore d'Ephese?," *Arabica* 21:3 (1975), pp. 270-84.
<sup>76</sup> Ibid., pp. 2-5.
<sup>77</sup>Drexl mentioned the opinions of a number of scholars (Reiske, Fabricius-Harles and Casiri); all repeat earlier opinions and information without stating anything new.
<sup>78</sup> Ibid., p. 5.
#### 20 CHAPTER ONE
beginnings of Arabian dream interpretation and the development of the literary genre of dreambooks, and contributed an inventory of lost and surviving Arabic dreambooks82 based on existing bibliographical compilations83 and especially on years of research in Turkish manuscript collections. He recorded 158 Arabic titles and 23 Turkish and Persian works that seem to be translations from the Arabic, for a total of 181 titles of works, some lost, some still extant. The inventory, arranged alphabetically by author, also gives dates whenever possible, lists the existing manuscripts of a work, and summarizes the conclusions that Fahd reached after a necessarily short examination of each text. In this way, he made the investigation of the overwhelming bulk of Arabic dreambooks manageable. In his entry on Ibo Sirin, for example, Fahd listed translations of Ibo Sirin's work in Persian, Turkish, Greek and Latin, and added: "One thing is certain: all these treatises do not resemble each other. We are convinced that, after a detailed comparative study, one would manage to identify their various compilers. "84
A new interest in the *Oneirocriticon* produced a number of publications in the 1980s. In 1986, K. Brackertz published an annotated German translation of the Greek text of the *Oneirocriticon.* In his introduction, he briefly examined the question of sources, and concluded that the name of Ibo Sirin attached to the Greek work is a pseudonym, in the way several Arabic dreambooks were attributed to him. Brackertz added that the *Oneirocriticon* is not a translation from the Arabic, but a work written by a Christian Greek who drew upon Arabic sources, as is clear from the several Christian references in the text. <sup>85</sup> He discounted the claim of the author of the *Oneirocriticon* that he used Indian, Persian and Egyptian sources, remarking that the only Indian and Egyptian elements consisted in a little local color. The single Indian reference, a statement that the elephant is only hunted in India, appears in chapter 269. A comparison with the Indian dreambook of Jagaddeva yielded no borrowing from true Indian dream interpretation. The Egyptian local color is limited to the word pharaoh and references to the Nile (Drexl 152, 16) and Cleopatra (Drexl 153, 1).
<sup>82</sup> Fahd, *La divination arahe,* pp. 330-67.
<sup>81</sup> lbn al-Nadim, *Kitiih al-fihrist;* English trans., *The Fihrist of al-Nadim: A Tenth-Century Survey of Muslim Culture,* trans. B. Dodge, 2 vols. (New York, 1970); J:Iajji Khalifa, Mu~tafah b. 'Abd Allah, *Kash/* al-~unun *'an al-Asiimi wa-al-Funun!Lexicon Bihliographicum et Encyclopedicum,* ed. and trans. G. Fltigel, 7 vols. (London, 1835-58); Ahlwardt, *Verzeichnis der arabischen Handschriften,* nos. 4263-89; Brockelmann, *GAL.*
<sup>84</sup> Fahd, *La divination arahe,* p. 356.
<sup>85</sup> K. Brackertz, trans., *Das Traumhuch des Achmet hen Sirin* (Munich, 1986), p. 10.
Brackertz found the Persian-Arabian contributions to the work, on the other hand, in references to polygamy, legitimate wives and concubines, musk and perfumes (Drexl 19, 5 ff.), Arabian horses (Drexl 110, 23 ff.), camels as riding and pack animals (Drexl 185, 25 ff.), a mill driven by a camel (Drexl 149, 21 ), and feathers in a dream indicating rank (Drexl 231, 18). 86 Brackertz added that a number of loan words from Arabic and the mention of sugarcane and cotton, both introduced into the Mediterranean world by the Arabs, also suggested that the Greek author had used Arabic sources. 87 That the author was Christian is apparent, according to Brackertz, in quotations from the Bible and biblical words in the text, and the contents of chapters 5 to 10, which treat elements of the Christian faith, such as the Resurrection of the Dead, Paradise, Hell and the Angels. 88 In addition, chapter 11 discusses the prophets, apostles, teachers, martyrs and various church officials. Dream interpreters include Joseph, the foster father of Christ (Drexl 2, 5) and the prophet Daniel (Drexl 2, 6). The interpretation of fat and lean cows when the dreamer is a king (Drexl 189, 25-27) remind us of the pharaoh 'swell-known dreams and their interpretation by Joseph in the Old Testament. 89
Brackertz was very careful to discuss the Arabic sources of the *Oneirocriticon* solely on the basis of the Greek text itself, and avoided confusing his readers by citing Arabic works that have an undetermined relationship to the Greek text. He found that, although the author of the *Oneirocriticon* did draw on Arabic sources for his book, he was mainly indebted to the Greek tradition of dream interpretation. He not only knew, but also used, the second century A.D. work of Artemidoros, for the general principles of dream interpretation, and also for several interpretations of dream symbols. The same rationale for the interpretation of a given dream is adduced both in Artemidoros and in the *Oneirocriticon. <sup>9</sup>°* Finally, Brackertz observed that the language and style of
<sup>86</sup> Brackertz does not explain why he considers this an exclusively Arabian interpretation. He probably repeats it from K. Latte, review of *Achmetis Oneirocriticon* rec. Drexl, *Gnomon* 2 ( 1926 ), p. 419. The Greek word m:i::pov (which Brackertz renders by *Feder)* means both "feather" and "wing." The interpretation of wings as denoting a high rank is mentioned in Artemidoros ii.68 (Pack 192, 8-11 ).
<sup>87</sup> Brackertz, *Traumhuch des Achmet.* p. 12. In that, he follows the opinions expressed by K. Dietrich in his review of *Achmetis Oneirocriticon* rec. Drexl, in *Orientalistische Literaturzeitung* 30: 10 ( 1927), pp. 881-84.
<sup>88</sup> Brackertz does not mention that these concepts are not exclusively Christian.
<sup>89</sup> Ibid., pp. 12-13.
<sup>90</sup> Ibid., p. 14.
the *Oneirocriticon,* more elevated than that found in the rest of the surviving Byzantine dreambooks, indicate the influence of Artemidoros ("sein Vorbild undModell"), of the New Testament, and of vernacular forms and expressions.<sup>91</sup>
In 1991, S. Oberhelman published an English translation of the *Oneirocriticon* accompanied by an extensive introduction and commentary , 92 both based on his 1981 dissertation.93 He identified the author of the *Oneirocriticon* as "a Christian Greek who used the *nom de plume* of Achmet to project an air of erudite, cosmopolitan leaming."94 He believed that the three sources used for compiling the *Oneirocriticon* were Arabic works, Artemidoros, and Byzantine sources:
[Achmet's] Arabic sources provided the schema of chapter listings, exemplary dreams, and some symbols and their interpretative meanings. Certain sections were borrowed from Byzantine Christian sources, especially the earlier dreambooks ascribed to Daniel and Astrampsychus. Finally, the *Oneirocriticon* of Artemidorus was the provenance of both many of Achmet's symbols and ... his methodology and theories on dreams.95
Oberhelman was persuaded that pseudo-Achmet had direct knowledge of Artemidoros and used his work extensively.96 To demonstrate the relationship between the two works, he drew up a comparative list of dream symbols from them with parallel interpretations. The list is not exhaustive because of constraints of space and in places it is faulty and misleading, but it remains a handy tool for research on the exact relationship between the Byzantine *Oneirocriticon* and Artemidoros.
The translations of Brackertz and Oberhelman made the Greek text more accessible to scholars, and their introductions reexamined its problems in some detail. Several questions, however, still remained unresolved. In his 1987 review of *Das Traumbuch des Achmet hen Sirin,* G. Strohmaier called for a new
96 Ibid., pp. 18-19.
<sup>91</sup>Ibid., p. 18.
<sup>92</sup> Oberhelman, *Oneirocriticon of Achmet.*
<sup>93</sup> Oberhelman, "Oneirocritic Literature of the Late Roman and Byzantine Eras of Greece."
<sup>94</sup> Oberhelman, *Oneirocriticon of Achmet,* p. 12
<sup>95</sup> Ibid., p. 20. Concerning the Arabic material, Oberhelman notes: "It would appear that Achmet used Ibn Shahln's *[sic]* dreambook as a schematic model for the presentation of his own material" (p. 17) because "the parallels and arrangement of material ... are too exact not to suppose some relationship" (p. 18). Appendix II of Oberhelman's publication is the table of contents of Ibn Shahin's dreambook, which has been lifted (without acknowledgment) from Bland, "Muhammedan Science ofTabfr." Ibn Shahin was a 15th-century author, which makes it impossible for the author of the *Oneirocriticon* to have used his work.
examination of the sources of the Greek text in order to determine its relationship with Arabic dream interpretation,97 and in a 1985 article G. Dagron summarized the current state of research regarding the authorship and sources of the Byzantine *Oneirocriticon* in the following words:
If all specialists ... agree today that they recognize in the *Oneirocriticon* the work of a Byzantine influenced by Islam, nobody has yet elucidated the complicated problem of its borrowings from Arabic literature of the ninth-tenth centuries. Regarding this point, the research of N. Bland, M. Steinschneider and A. Fischer remained without conclusion and, unfortunately, without an echo.98
Clearly, any investigation of the sources of the *Oneirocriticon* should not remain limited to the Greek text alone. One ought to examine the Greek text side by side with an Arabic dreambook and see what conclusions a detailed comparison yields. But which Arabic dreambook should one use? The only Muslim dream interpreter expressly mentioned in the Greek text, lbn Sirin, is the alleged author of several dreambooks with different titles that belong to different centuries. Moreover, none of the surviving Arabic dreambooks listed in Fahd's inventory is earlier than the beginning of the eleventh century, which suggests that no Arabic dreambook that clearly antedates the Greek text was known to him.99 The only viable alternative is to examine the old Paris, Bibliotheque du Roi ms. 1210, now BN *arabe* 2742, that had been reputed to be *omnino idem* with the Greek text.
Scholars who have examined the Greek text of the *Oneirocriticon* have all insisted on its Christian character, especially in chapters 5-12 which discuss a number of Christian concepts. One would think that the most fertile testing grounds, even if the introductions of the two works were identical, would be the passages discussing religion in each dreambook. Both Islam and Christianity
<sup>97</sup>G. Strohmaier, review of *Das Traumbuch des Achmet ben Sirin,* iibersetzt und erlautert von Karl Brackertz, Miinchen (C.H. Beck) 1986, *Klio* 69 (1987), pp. 654-55.
<sup>98</sup> G. Dagron, "Rev er de Dieu et parter de soi. Le reve et son interpretation d' a pres !es sources byzantines," *I sogni nel medioevo,* ed. Gregory, p. 49.
<sup>99</sup> The oldest Arabic dreambook surviving in its entirety mentioned in Fahd's inventory is that of al-Dinawari, which, according to its introduction, was completed in 1006. A still older one, that of Ibn Qutayba (d. 889), was known to Fahd from a unique Turkish manuscript, Ankara University, Oil ve Tarih-Cografya Fakiiltesi Ktp., *Ism. Saib Sine-er* I, 4501, fols. 180b-217, which, however, contained only a theoretical introduction and no interpretations of specific dream symbols. A manuscript with the full text of lbn Qutayba's dreambook was made known to the scholarly world in 1974 through the study of M. J. Kister, "The Interpretation of Dreams: An Unknown Manuscript oflbn Qutayba's *'/biiriit al-Ru'yii," /OS* 4 (1974), pp. 67-103. The dreambook of Ibn Qutayba, together with several other manuscripts from the collection of Professor Yahuda, was given to the Hebrew University of Jerusalem by his heirs after his death in 1951. The text remains unpublished.
#### 24 CHAPTER ONE
are monotheistic religions and share a number of concepts, as well as of holy figures. But could the Christian interpretation of the Last Judgment, of Paradise and Hell, angels and prophets be identical with the Muslim interpretation of the same topics? None of the researchers of the Greek text had thought it was possible, and yet that is exactly the case.
The Resurrection of the Dead is the first religious dream discussed in the *Oneirocriticon.* It is also the first religious dream interpreted in BN *arabe* 2742 after godhead itself. The interpretation in the two dreambooks is identical and parts of the text in the two languages read almost word for word the same. *wo* The introduction to the Arabic text, on the other hand, turns out to be very different from its Greek counterpart, and the examination of further entries shows that, while the dream interpretations sometimes converge, at other times they are completely different. It soon becomes clear that BN *arabe* 2742 contains numerous nuggets, but is definitely not identical with the Greek text.
Another Parisian manuscript, currently BN *arabe* 2744, was claimed by d'Herbelot to be the source of the Greek text, a claim repeated by Kollar but not in subsequent manuscript catalogues. Consequently, it has not attracted much scholarly attention. An examination of BN *arabe* 2744, of other Parisian manuscripts on dream interpretation, and of other Arabic dreambooks, both published and unpublished, leads to the conclusion that all texts of Arabic dream interpretation resemble each other to a greater or lesser extent, but that no two among them-or at least of those that I examined-are identical. Additional complications arise from the fact that treatises with the same title and purportedly by the same author can have slightly different contents, <sup>101</sup>
<sup>100</sup>BN *arahe* 2742, fol. 4; and Drexl 3, 26-4,3.
<sup>101</sup>Such is the case, for example, with BN *arahe* 2742 and 2743, both of which are attributed to Ibn Sirin and bear the title *Ta'hlr al-ru'yd.* Although they share some word-for-word passages, they are not identical throughout. A treatise with the same title and purportedly by the same author was published in Pakistan (Mul:Jammad b. Sirin, *Ta'hir al-ru'yd.* Arabic text and Urdu translation, [s. I., 1970]). The same work has also been translated into French by D. Penot, *L'inte1pretation des reves. Manuel d'oniromancie musulmane* (Lyons, 1992); and into Italian by I. Zilio Grandi, *II lihro de/ sogno veritiero* (Turin, 1992). Except for Zilio Grandi (on p. xxiv), no published version of the work indicates which manuscript or manuscripts it was based on. The five versions that I have examined are similar, though they are not all equally long; they vary in chapter sequence and do not always phrase an interpretation in the exact same words. A *terminus post quern* for their composition can be established, since all three narrate the dream dreamt by Alphonso IV (I 065-1109), king of Castille, Leon, and Galicia, before the battle of Saragossa, which took place in 1086 (Ibn Sirin, *Ta'hir al-ru'yd,* p. 124; Ibn Sirin, *lnte1pretation des reves,* trans. Penot, pp. 139-42; Ibn Sirin, *Lihro de/ sogno,* trans. Zilio Grandi, pp. 99-100). The same dream is also quoted and translated into French by Fahd according to Ibn al-Athir *(La divination arahe.* pp. 299-300).
and, conversely, that the same treatise is often referred to by different titles or authors. 102 Dating each dreambook and ensuring that its attribution to the author mentioned in its title is genuine can create other vexing problems. As was mentioned earlier, many are attributed to lbn Sirin (d. 728), who certainly never wrote anything on the subject and whose renown as a dream interpreter emerged only in the course of the ninth century, gradually to become synonymous with the art of dream interpretation itself. <sup>103</sup>
The Greek text of the *Oneirocriticon* is connected with the Arabic dreambooks in the way that the Arabic dreambooks are connected with each other: it resembles a number of them, both in structure and in content, but is not identical with any. Moreover, the slightly disturbed order of the Greek chapters and the repetition of certain interpretations in various parts of the *Oneirocriticon* indicate that more than one Arabic dreambook was used for its compilation. If the Greek author based the *Oneirocriticon* on a single Arabic text, then it was one that had been compiled on the basis of more. Either case implies that in order to obtain a better picture of how the Byzantine compiler used his Arabic sources it is important to compare the Greek text with as many Arabic dreambooks as possible.
Because it is impossible to compare the *Oneirocriticon* with each and every one of the several dozen surviving Arabic works, choosing which Arabic works to compare with the Greek can only be arbitrary. I selected five, using two criteria: the early date of their composition, which would place them chronologically as close as possible to the *Oneirocriticon,* and their accessibility to the larger scholarly community, which would facilitate further discussion, especially concerning the interpretation of the hundreds of dream symbols in the Greek and the Arabic tradition that are impossible to cover in the present study.
Using these criteria, the first two of the five chosen Arabic dreambooks are among the earliest surviving Arabic works on dream interpretation, though even they cannot be safely considered to antedate the Greek text, which possibly constitutes the earliest surviving document belonging to the rich tradition of
Other editions of a work under the same title that I have not examined are Ibn Sirin, *Kitiih ta'hlr a/-ru'yii* (Cairo, [n. d.]); Ibn Sirin, *Kitiih ta'hlra/-ru'yii* (Baghdad, [1900?]).
<sup>102</sup> Such is the case with Ibn Qutayba's dreambook; see Kister, "Interpretation of Dreams," p. 67.
<sup>101</sup> See Fahd, *La divination arahe,* pp. 312-15; also £/ 2, s.v. "Ibn Sirin." The most recent and detailed discussion concerning the connection of Ibn Sirin with dream interpretation can be found in J. Lamoreaux, "Dream Interpretation in the Early Medieval Near East." Ph.D. diss., Duke University,1999, pp. 32-41.
#### 26 CHAPTER ONE
Arabic dream interpretation. The other three Arabic works definitely postdate the *Oneirocriticon;* they were written between the eleventh and the early eighteenth century. As they incorporate much of the earlier tradition on dream interpretation, however, comparing them with the Greek text can still be fruitful. Moreover, these works constitute the three major sources on Arabic dream interpretation currently available in print and are housed in a number of university libraries in North America. The five Arabic sources are:' <sup>04</sup>
1. Abu Muqammad 'Abd Allah b. Muslim b. Qutayba (828-89), *'lbiirdt al-ru'yd* (Interpretations of Dreams). The work survives in two manuscripts: Hebrew University, *Yahuda ar.* 196, and Ankara University, Dil ve Tarih-
Ie» Lamoreaux, "Dream Interpretation in the Early Medieval Near East," established the identity of Abii al-'Abbas Al)mad b. Khalafb. Al)mad al-Sijistani (no. I 12 in Fahd, *la divination arabe,* p. 354) as the last Saffarid amir of Sijistan (963-1009) and thus firmly placed his work among those written in the second half of the !0th century. Lamoreaux also identified a manuscript of al-Qayrawani's dreambook (I Ith century), which was known to Fahd only by title (no. 93), and was able to summon evidence indicating that a further work, the dreambook by al-Mu'afiri (or Ma'afiri), listed by Fahd (no. 81) as surviving in two manuscripts without an indication as to its chronology, could be dated to the !0th century. After Lamoreaux's work became known to me, I contemplated whether it was worth procuring the manuscripts of these three authors to compare them with the *Oneirocriticon,* on account of their early date. I decided against it since all three most likely postdated the *Oneirocriticon;* their dating, combined with Lamoreaux's report on their contents, indicated that none of them could possibly be the direct source of the Greek text. Al-Sijistani's work is extremely concise (Lamoreaux, p. 60) and for this reason promised few new interpretations. Al-Qayrawani's work (ibid., pp. 86-96) picked from earlier sources only what in his opinion was firmly grounded in the Muslim tradition and "wholly eschewed" dream Jore derived (or believed to be derived) from the other religious communities of the ancient and contemporary Near East, as well as everything inherited from the Greco-Roman tradition (ibid., p. 94). The author of the *Oneirocriticon* relied on Islamic sources that not only included, but even gave a prominent position to, the kind of material that al-Qayrawani deliberately avoided in his compilation, so the additional material I could collect from al-Qayrawani's work also promised to be minimal. The dreambook of Ma'afiri interprets exclusively parts of the human body and is very concise (ibid., p. 341), which suggests that it too contained very little, if any, material comparable to the interpretations found in the *Oneirocriticon.* My examination of Arabic dreambooks, both those that I chose to compare with the *Oneirocriticon* and others (manuscripts in the Bibliotheque Nationale and printed works), indicated that Arabic dreambooks, based as they always were on earlier sources, repeated much of the same material. The re.petitiousness and ultimate cohesion of the Arabic tradition on dream interpretation are also commented on by Lamoreaux (pp. 128-74), based on his *in situ* examination of manuscripts in Turkey and the Middle East, so few new interpretations could be expected from either al-Sijistani or al-Qayrawani. My experience had also taught me that the oldest Arabic compilations did not always preserve the passages most closely connected with the Greek *Oneirocriticon,* since I had sometimes been able to find in the two later dreambooks that I had chosen (written in the 15th and 17th-18th century) interpretations comparable to the Greek ones that were missing from the two earlier (9th- and early-I I th-century) Arabic works I had at my disposal. Finally, all but one of these manuscripts had the disadvantage of being in Turkish libraries, from which obtaining microfilms is particularly complicated and time-consuming.
Cografya Fakiiltesi Ktp., *Ism. Saib Sincer* I, 4501, fols. 180a-217b. 105 Parts of the text are missing in both manuscripts, but it seems that the one belonging to the Hebrew University preserves a version closer to the archetype than the one in Ankara. *w6*
2. Abii Sa'id (or Sa'd) Na~r b. Ya'qiib al-Dinawari (d. ca. 1020), <sup>107</sup>*Kitab al-Qadiri ft al-ta<bir* (The Book Dedicated to Caliph al-Qadir on Dream Interpretation). Since there is no critical edition available, I compare the Greek text with the oldest surviving manuscript, Istanbul, Siileymaniye Ktp. *Esad Efendi* 1833 (12th century). 108 Supplementary readings and references are taken from BN *arabe* 2745. <sup>109</sup>
3. Abii 'Ali al-I:Iusayn b. I:Iasan b. Ibrahim al-Khalili al-Dari, *Muntakhab al-kalamfi ta/sir al-af:ilam(Selection* of Statements on the Exegesis of Dreams). The author seems to be otherwise unknown. <sup>110</sup>The text used is the edition
<sup>107</sup>Al-Dinawari was a tax collector at Nishapur. The sultan Yamin al-Daula relied on him for his correspondance with the caliph al-Qadir bi-I-Lah. For his literary output and further references, see Khayr al-Din al-Zirikli, *al-A'lam,* 10 vols., 2nd ed. (Cairo, 1954-59), vol. 8, p. 353 (s.v. \_,........,..:. ...,\_,~ 0-:> ); see also Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp.
104-13.
108 I would like to thank Professor Fahd for generously informing me about the results of his unpublished research on the text of al-Dinawari and for lending me his microfilm of *Esad Efendi* 1833, which othe1wise would have been impossible to consult.
109 Unfortunately, the manuscript stops at *hab* 2 of the 12th *fa:f/.* The references to BN *arabe* 2745 are supplied because not only is it easily accessible, as it is housed in a European library, but it is also the manuscript of reference used in one of the most important articles on the history of Arabic dream interpretation, Fahd's "Les songes et leur interpretation selon l'Islam."
<sup>110</sup>Al-Dari's name seems to have remained unrecorded in medieval biographical dictionaries and bibliographies, and the only mention of him in contemporary standard reference works is in *GAL,* vol. S I, p. 361, where he is identified only as the author of *al-Muntakhab,* which survives in BN *arabe* 2749 (the manuscript used by Fahd for the excerpts of Ibn Qutayba in *la divination arabe).* The date of the composition of *al-Muntakhab* is discussed in J. Lamoreaux, "Some Notes on the Dream Manual of al-Dari," *Rivista degli Studi Orientali* 70 (1996), p. 52. Lamoreaux gives the early 11th century as a *terminus post quern* and 1214 as a *terminus ante quern,* with the likelihood that it was composed in the earlier part of this period. The elusive identity of the author of *a/-Muntakhab* has also been discussed by I. Zilio Grandi, "II problema della visione di dio secondo ii manuale di onirocritica *Muntakhab al-kalamfi tafslr a/-afilam," Annali di Ca' F oscari*
<sup>105</sup> Ibn Qutayba is one of the great Sunni polygraphs of the 9th century, with an interest in both theology and belles lettres. For a detailed account of his life and work on dream interpretation, see Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 46-58; see also *EI2,* s.v. "Ibn ~utayba, Abu Mu~ammad 'Abd Allah b. Muslim al-Dinawari." In the past, doubts have been expressed concerning the authenticity of Ibn Qutayba's dreambook. For a re-examination of the problem and proof of the work's authenticity, see Lamoreaux's Appendix 2.
<sup>106</sup>For the problems in the manuscript tradition of Ibn Qutayba's dreambook, see ibid., p. 47; also Kister,"Interpretation of Dreams," p. 69. I would like to express my gratitude to Professor Kister for kindly making available to me a reproduction of the Ankara manuscript.
#### 28 CHAPTER ONE
published under the name of Mui)ammad b. Sirin, *Tafsir al-alfldm al-kabir al-musamma muntakhab al-kalam ft tafsir al-alflam* (The Great Exegesis of Dreams, also called Selection of Statements on the Exegesis of Dreams ). <sup>111</sup>
A major source for *al-Muntakhab* was the work of Abii Sa'id 'Abd al-Malik b. 'Uthman al-Wa'i~ al-Khargiishi (d. 1015), <sup>112</sup>*al-Bishara wa-al-nidhdra ft ta'bir al-ru'ya* (The Glad Tidings and Warnings in the Interpretation of Dreams). This treatise is the third oldest surviving Arabic dreambook after al-Dinawari's. Several copies of it exist, but no *stemma* has ever been prepared, and it has never been published. <sup>113</sup>I used the British Library copy, BL *Or.* 6262. Al-Khargiishi, who lived at the same time and in the same city as al-Dinawari, based many of his interpretations on those of his compatriot. <sup>114</sup>In its turn, *al-Muntakhab* quotes al-Khargiishi's interpretations almost in their entirety, sometimes adding further material. <sup>115</sup>Given the occasional additions
111 Mul]ammad b. Sirin, Taf~lr *a/-afJ/iim a/-kahfr a/-musammii muntakhah al-kaliim fl ta/sir al-afJ/iim* (Cairo, 1963). Other relatively recent editions are: Mul]ammad ibn Sirin, *Mukhta,mr ta'tir al-aniimft ta'hir al-maniim. Wa-hi-hi nuhdhah min Ki tab a/-kalamft tafsir al-a!Jlam lil-'al/ama a/-ma'ruf hi-ihn a/-Niihulusi* (Beirut, 1969); *Muntakhah al-kaliimft ta/sir al-afJ/iim* (Cairo, 1972), printed together with the work of al-Nabulusi, *Ta' fir a/-aniimft ta'hfr al-maniim; Ta/sir a/-alJ/iim* (Amman, 1995). The work was first printed in the 19th century. For the earlier printings and their more recent descendants, see Lamoreaux, "Some Notes on the Dream Manual of al-Dari," p. 47.
112 Al-Khargushi was a celebrated preacher and ascetic from Nishapur. Besides *al-Bishiira wa-al-nidhiira,* he also wrote a biography of Mul]ammad and an account of sufism. See *E/2 ,* s.v. "al-Khargu~hi." See also Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 96-104.
111 See Fahd, *La divination arahe,* p. 358, no. 128.
114 This is made evident by the number of textual parallels between them. Cf. al-Dinawari, *maqii/a* 7, *Esad Efendi* 1833, fol. 28b, lines 2-6, with the introduction of al-Khargushi, BL *Or.* 6262, fol. 6a [ 12]; cf. the anecdotes on dreaming of crucifixion in al-Dinawari, *fa\$1* 11, *hiih* 60, *Esad Efendi* 1833, fol. 168a and al-Khargushi, BL *Or.* 6262, fol. 150b [300]; cf. the interpretation of the cross, including the quotation from Artemidoros, in al-Dinawari, *fa\$!* 11, *hiih* 59, *Esad Efendi* 1833, fol. l 67b and al-Khargushi, BL *Or.* 6262, fol. 151 a [30 I]; cf. the interpretation of the elephant, including the quotation from Artemidoros, in al-Dinawari,/a\$1 21, *hiih* 104, *Esad Efendi* 1833, fols. 224a-b and al-Khargushi, BL *Or.* 6262, fols. 142b-143b [344-346]; cf. the interpretation of the planets in al-Dinawari,fa,~I 15, *hiih* 18, *Esad Efendi* 1833, fols. 152b-153a and al-Khargushi, BL *Or.* 6262, fol. 197b [394], among others.
115 On the relationship between *a/-Muntakhah* and the dreambook of al-Khargushi, and a discussion of the remaining sources used for *al-Muntakhah,* see Lamoreaux, "Some Notes on the Dream Manual of al-Dari," pp. 47-52.
<sup>27:3 ( 1988),</sup> pp. 69-70. Zilio Grandi examines the chapter on dreaming of God and suggests that the work must have been written during the 12th or 13th century, a time when Islamic mysticism reached its peak (ibid., p. 79). However, this chapter was copied verbatim from the equivalent chapter of al-Khargushi's *al-Bishiira wa-al-nidhiira* (BL *Or.* 6262 fol. 9b [17]-12a [22]) and can therefore be dated with certainty to the years between 1006 and 1020. The sufi references in *al-Bishiira* are understandable, since al-Khargushi was an ascetic and had also written a history of sufism.
and the easy accessibility of the work in print, I chose to compare the *Oneirocriticon* with *al-Muntakhab* instead of al-Khargushi's dreambook, despite the latter's earlier date.
5. Ghars al-Din Khalil b. Shahin al-Z:ahiri (1410-68), 116 *al-lshiiriit ft 'ilm al-'ibiiriit* (Intimations on the Science of Interpretations). This work exists in several manuscript versions and a number of printed editions. <sup>117</sup>Ibn Shahin's passages will be quoted from the Cairo edition of 1991. <sup>118</sup>In the introduction to his dreambook, Ibn Shahin says that he based his compilation both on older works and on his own experience, and lists the written sources that he used: <sup>119</sup> "I have relied on the books of the ancients and the sayings of the venerable dream interpreters, such as the *Kitiib al-u:jul* (Book of Sources) by Daniel the Wise; 120 the *Kitiib al-taqsim* (Book of Classification) by Ja'far al-Sadiq; 121 the *Kitiib al-jawiimi'* (Book of Epitomes) by Mul)ammad b. Sirin; 122 the *Kitiib al-dustur* (Book of Constitutions) by Ibrahim al-Kirmani; 123 the *Kitiib al-irshiid* (Book of Guidance) by Jabir al-Maghribi; <sup>124</sup>the *Kitiib al-ta'bir* (Book of Dream Interpretation) by Isma'il b. al-Ash'ath; 125 the *Kitiib kanz al-ru,yii* (Book of the Treasure of Dreams) by al-Ma'muni; 126 the *Kitiib bayiin al-ta'bir* (Book on the Elucidation of Dream Interpretation) by 'Abdus; 127 the *Kitiib jumal*
I. 122 Fahd, *La divination arabe,* pp. 355-36, no. 117, and especially p. 356, n. 3; Lamoreaux, no.
<sup>116</sup>Ibn Shahin's father was a *mamliik* of the sultan Sayf al-Din Tatar. Ibn Shahin studied in Cairo and had a brilliant administrative career. He is the author of several works, including a description of Egypt; see £/2, s.v. "Ibn Shahin al-Zahiri."
<sup>117</sup>See Fahd, *La divination arabe,* p. 351, no. 102.
<sup>118</sup>Khalil b. Shahin, Taf~ir *al-al;/iim a/-musammii a/-ishiiriit ft 'ilm al-'ibiiriit.* 2 vols. in one (Cairo, 1991 ).
<sup>119</sup>Ibid., p. 8.
<sup>120</sup>Fahd, *La divination arabe,* p. 335, no. 24. Daniel the Wise is the prophet Daniel; dreambooks attributed to him are also known in Greek, Latin, and various European languages.
<sup>121</sup>Fahd, *La divination arabe,* p. 338, no. 39. The attribution is clearly spurious. Ja 'far al-Sadiq (d. 765) was the sixth imam of the Shi'a; his name is linked with the occult sciences. See Fahd, *"1* a'far al-Sadiq et la tradition scientifique arabe," pp. 131-42; see also Ullmann, *Die Natur- und Geheimwissenschaften im Islam,* pp. 195-96. See also no. 2 in the more recent catalogue of early Islamic dreambooks in Lamoreaux, "Dream Interpretation in the Early Medieval Near East," appendix 1 (subsequent references to numbers preceded by Lamoreaux's name refer to the number assigned to each Arabic author in this catalogue).
<sup>123</sup>Fahd, *La divination arabe,* p. 345, no. 67; Lamoreaux, no. 5.
<sup>124</sup>Fahd, *La divination arabe,* pp. 337-38, no. 38; Lamoreaux, no. 6.
<sup>125</sup>Fahd, *La divination arabe,* p. 332, no. 9; Lamoreaux, no. 29, now lost.
<sup>126</sup>Lamoreaux, no. 32.
<sup>127</sup> Fahd, *La divination arabe,* p. 331, no. 4; Lamoreaux, no. 44, now lost.
*al-dala'il* (Book of the Groups of Signs); 128 the *Kitab mabddi' al-ta'bir* (Book of the Principles of Dream Interpretation); 129 the *Kitab kafi al-ru'ya* (Book of Adequate Dreams), 130 the *Kitab al-ta'bir* (Book of Dream Interpretation) by al-Tamiisa (u-'"",j-Cl..b.J I); the *Kitab muqarma.t al-ru'ya* (The Fine and Narrow Written Book of Dreams); 131 the *Kitab tul}fat al-muluk* (Book of the Gem of Kings); 132 the *Kitab minhdj al-ta'bir* (The Way to Dream Interpretation) by Khalid al-l~fahani; 133the *Kitab muqaddimat al-ta'bir* (Book of the Introduction to Dream Interpretation); 134 the *Kitab l}aqa'iq al-ru'ya* (Book of Facts on Dreams); 135 the *Kitab al-wajiz* (Concise Book) by MuI:iammad b. Shamawayh; 136 the *Kitab al-ta'bir* (Book of Dream Interpretation) by Abii Sa'id al-Wa'i?-; <sup>137</sup> the *Kitab kamil al-ta'bir* (Book of Perfection on Dream Interpretation) by Shaykh Abii'l-Fa91 I:Iubaysh b. Ibrahim b. AI:imad al-Nuqayshi; the *Kitab al-ishara ila 'ilm al-'ibdra* (Book oflntimation on the Science oflntepretation) by Abii 'Abd-Allah b. AI:imad b. 'Umar al-Salimi; 138 the *Kitab al-durr almuna'Pamfi al-sirr al-mu'a'Pam* (Book of Arranged Pearls Regarding the Sublime Mystery) by MuI:iammad al-Qurashi al-Na~ibi; and other authors like these, such as Shaykh AwI:iad al-Din 'Abd al-Latif al-Dimyati, Shaykh
<sup>131</sup>Fahd, *La divination arabe,* p. 361, no. 153; Lamoreaux, no. 60, now lost. The Qaramifah (Carmathians) were a Muslim sect influenced by gnosticism; they received their name from the founder I:Iamdan Qarmaf. The name Qarmaf in Aramaic means "he of the two red eyes," probably signifying a teacher of secret doctrines. The title of the dreambook has been translated according to one of the meanings of the root *q-r-m-f* in Arabic, namely "to write in fine and narrow characters," which was also a characteristic of old Manichean books. An interesting Greek example is the so-called Mani Codex, now in Cologne, an uncial manuscript of unusually small format that contains the biography of Mani (for references, see B. L. Fonkic and F. B. Poljakov, "Paliiographische Grundlagen der Datierung des ktilner Mani-Codex," *BZ* 83 [1990], pp. 22-30). Another meaning of the root *q-r-m-{* in Arabic is "to copy out a charm"; see C. E. Bosworth, *The Medieval Islamic Underworld: The Banu Siisiin in Arabic Society and Literature,* 2 vols. (Leiden, 1976), vol. l, p. 90. The dreambook's title could also mean "The Carmathian Book of Dreams," which would imply that it claims to be transmitting gnostic or apocryphal wisdom. On the sect of the Carmathians, see the *Concise Encyclopedia of Islam,* s.v. "Qarami!ah."
<sup>128</sup> Fahd, *La divination arabe,* p. 361, no. 155; Lamoreaux, no. 58, now lost.
<sup>129</sup> Fahd, *La divination arabe,* p. 361, no. 156, now lost.
<sup>130</sup> Fahd, *La divination arabe,* p. 361, no. 157; Lamoreaux, no. 59, now lost.
<sup>132</sup> Fahd, *La divination arabe,* p. 354, no. 112, still extant; the author is Abu al-'Abbas AJ:imad b. Khalaf b. AJ:imad al-Sijistani; Lamoreaux, no. 61 (among the anonymous works) and no. 35.
<sup>133</sup> Fahd, *La divination arabe,* p. 344, no. 64; Lamoreaux, no. 53, now lost.
<sup>134</sup> Fahd, *La divination arabe,* p. 361, no. 152; Lamoreaux, no. 62, now lost.
<sup>135</sup>Lamoreaux, no. 63.
<sup>136</sup> Fahd, *La divination arabe,* p. 352, no. 104; Lamoreaux, no. 54.
<sup>137</sup>Lamoreaux, no. 37.
<sup>138</sup> Fahd, *La divination arabe,* p. 352, no. 103.
'Abd al-Qadir al-Ashmuni, Shaykh Yusuf al-Karuni al-Sikandari, Shaykh Mul).ammad al-Fir'awni, Shaykh I:Iasan al-Ramli, Shaykh Nur al-Din al-Karkhi al-Ghazawi, Shaykh Taqi al-Din al-Maqdisi, Shaykh Sharaf al-Din al-Karaki, Shaykh Shams al-Din I:Iamdun al-Safadi, etc. To these I have added the truthful dreams that I and my companions have had that were fulfilled as clearly as the dawn of the morning .... " lbn Shahin's array of sources is impressive. Though most of the works he lists are no longer extant, they are also mentioned in other sources, such as the work of the bibliographer I:Iajji Khalifa and in introductions to dreambooks other than lbn Shahin's, which indicates that their titles are genuine.
6. 'Abd al-Ghani b. Isma'il al-Nabulusi (1641-1731), <sup>139</sup>*Ta'tzr al-anamfi ta/sir al-manam* (The Perfume of the Creation on the Exegesis of Dreams), which exists in several manuscripts and printed editions. <sup>140</sup>The Cairo edition of 1940 is used here. <sup>141</sup>In the concluding chapter of his dreambook, al-Nabulusi gives his sources: <sup>142</sup>al-Dinawari's *al-Qadir! Ji al-ta'bir* (Book Dedicated to Caliph al-Qadir on Dream Interpretation) and the dreambook of Mul).ammad b. Abi Bakr Mal).mud b. Ibrahim, known as lbn al-Daqqaq al-Muqri, entitled *al-f:iukm wa-al-ghayat ft ta'bir al-manamat* (Decision and Objectives in the Interpretation of Dreams); 143 the dreambook of Abu 'Ali al-I:Iusayn b. I:Iasan b. Ibrahim al-Khalili al-Dari called *al-Muntakhab* (The Selection); <sup>144</sup>the dreambook of Jalal al-Din 'Abd-Allah b. I:Iazim b. Sulayman al-Muzani
<sup>139</sup>Al-Nabulusi was a poet, theologian, and an author of belles lettres. Born and raised in Damascus, he traveled to Baghdad, Palestine, Lebanon, Egypt, and the Hijaz before returning to his native city, where he died. For a list of his works, see al-Zirikli, *al-A'liim,* vol. 4, pp. 158-59 (s.v. ~~I .J-7 ~I ~).For further biographical information, see also *EI2,* s.v. "Abd al-Ghani b. Isma'il al-Nabulusi."
<sup>140</sup>Fahd, *La divination arabe,* p. 348, no. 85. To the editions cited by Fahd add two that were published since: 'Abd al-Ghani al-Nabulusi, *Ta' fir al-aniimft ta'bir al-maniim, Qiimiis a/-aflliim.,* ed. I:Ianan MuJ:iammad Nur Tabbarah (Beirut, 1993); 'Abd al-Ghani al-Nabulusi, *al-'iibir ft al-ta'bir ft U\$iil kayftyat ta'bir al-ru'yiift a/-maniim,* ed. MuJ:iammad 'Abd al-RaJ:iim (Beirut, 1996).
<sup>141</sup>'Abd al-Ghani al-Nabulusi, *Ta'(ir a/-aniimft ta'bir a/-maniim, wa-bi-hiimishih awwaluhuma Muntakhab al-kaliim ft tafsir al-aflliim li-Muflammad ibn Sirin wa-thanihuma a/-ishiiriit ft 'ilm al-'ibiiriit, Ii-Khalil ibn Shahin* a/-~iihiri. 2 vols. (Cairo, 1940); in this edition the dreambook of al-Nabulusi occupies the top half of each page in both volumes; *al-Muntakhab* can be found the bottom half of each page in the first volume, and the dreambook of Ibn Shahin in the bottom half of each page in the second volume.
<sup>142</sup>Vol. 2, p. 350.
<sup>143</sup>Fahd, *La divination arabe,* p. 335, no. 25; see also ibid., p. 348, no. 83, and Lamoreaux, no. 33, where the title is different, but the name of the author almost the same.
<sup>144</sup> Lamoreaux, no. 55.
(~j...11) al-Shaf"i, entitled *al-Isharahfi 'ilm al-'ibarah* (Intimation on the Science of Interpretation); 145 the dreambook of Abii 'Abd-Allah Mul).ammad b. 'Umar al-Salimi, entitled *Ki tab al-ishiirah ila al-' ibiira* (Book of Intimation on Interpretation); 146 the dreambook of Shihab al-Din Abii al-'Abbas Al).mad b. al-Shaykh Jamal al-Din Abi al-Faraj 'Abd al-Ral).man al-Maqdisi al-I:Ianbali, entitled *al-Badr al-munir fi 'ilm al-ta'bir* (The Shining Full Moon on the Science of Dream Interpretation); 147 the dreambook of Abii Tahir Burhan al-Din Ibrahim b. Yal).ya b. Ghanim al-Maqdisi al-I:Ianbali, entitled *al-Mu' lam* or *al-Mu'allam 'ala f:iuruf al-mu'jam* (The Marked or The Inspired <Book> in Alphabetical Order), and its abridgment by Mul).ibb al-Din Abii I:Iamid Mul}ammad al-Maqdisi al-Shafi'i, entitled *al-Muf:ikamfi ikhti\$ii\$ al-mu'lam* (Referee in Distinguishing the Signs). 148 All of al-Nabulusi's sources are extant, but only *al-Muntakhab* is published.
#### *The Author of the* Oneirocriticon
The attribution of the *Oneirocriticon* to "Achmet, son of Sereim" appears in the title of the work in a number of Greek manuscripts. However, there is no mention of Achmet, of Islam, or of the Arabs in the first four chapters of the *Oneirocriticon,* where the information on the work's sources is presented in the form of four separate prologues written in the first person and strung together. The first prologue mentions no author and only states that the *Oneirocriticon* is a compilation of Indian, Persian and Egyptian sources. The second prologue is attributed to one Syrbacham, dream interpreter to the king of the Indians, the third is said to be by Baram, dream interpreter to king Saanisan of Persia, and the fourth is attributed to Tarphan, dream interpreter to the Egyptian pharaoh.
Although Islam and the Arabs are not mentioned in the first four chapters, they do appear in subsequent passages. They are mentioned for the first time in chapter 19, in a narrative on a dreamer's dream and his consultation with a dream interpreter; the chapter's purpose is to exemplify how dream interpreta-
<sup>145</sup> This is possibly Fahd, *La divination arabe,* p. 347, no. 80, where the author is named 'Abd Allah Sulayman b. I:Iazim. The title of the book is the same, and it is said to be an abridgment of a longer work by the same author.
<sup>146</sup> Fahd, *La divination arabe,* p. 352, no. 103.
<sup>147</sup> Ibid., p. 346, no. 72.
<sup>148</sup> Ibid., p. 338, no. 42.
tion is supposed to be conducted and to confirm that the interpretations offered in theory earlier in the Oneirocriticon have proved to be reliable in practice. The story in chapter 19, which is also narrated in the first person, contains two Arabic names, Achmet, son of Sereim, and Caliph al-Ma'mun.
' Ελθών τις ανθρωπος ήρώτησέ μοι τω 'Αχμετ τώ νίφ Σηρείμ, τφ όνειροκρίτη του πρωτοσυμβούλου Μαμούν είδον έν όράματι, ότι οί τρίχες των σκελών μου έδασύνθησαν και ηύξήνθησαν, και ταύτας τη ψαλίδι έκούρευον. και άπεκρίθην αυτώ ότι το λογάριόν σου και ό πλουτός σου ἐπληθύνθησαν καὶ ὅσον ἔκοψας των τριχών σου, τοσούτον κακοδιοικείς αύτά. και εύρέθη ούτως το πραγμα. "Υ
A certain man came and asked me, Achmet, the son of Sereim, the dream interpreter of the caliph Mamoun: "I saw in a dream that the hair on my legs grew longer and thicker and that I was cutting it with a pair of scissors." I replied: "Your wealth and riches have increased, but you mismanage them in proportion to the quantity of hair that you cut." And it was discovered that things were indeed so.
Taking into account the four prologues, this short anecdote constitutes the fifth instance of first-person discourse in the Oneirocriticon. Twelve more examples of actual consultation are inserted in the remaining pages, usually under the heading "ερώτησις" or "έρώτημα" (Question). Of the thirteen anecdotes about consultations, only the one in chapter 19 is narrated in the first person. It is also the only place in the whole book where the full name, "Achmet, son of Sereim," is mentioned. 50 Ten of the thirteen anecdotes give the name as "the dream interpreter Sereim," or simply "Sereim,"151 and the remaining two say only "the dream interpreter" (o overookoittne).152 In chapter 19, Ibn Sirin speaks in the first person, not because the compiler of the Oneirocriticon wants to assume his identity, but because his words are quoted directly in the same way the Indian, Persian and Egyptian dream interpreters were quoted earlier.
Why is Ibn Sirin allowed to speak in the first person only this one time, and after that is referred to in the third person? Arabic dreambooks frequently contain narrations or direct quotations like the thirteen anecdotes incorporated
<sup>149</sup> Drex1 15, 18 ff.
ISo "Achmet, son of Sereim" is mentioned only in Drexl I5, 18. "Sereim" is mentioned in Drexl 23, 25; 24, 2; 29, 25; 30, 1; 57, 13; 92, 2; 99, 8; 102, 2; 111, 26; 137, 24; 148, 19; 156, 7.
<sup>151</sup> Chap. 20, Drexl 16, 1-10; chap. 36, Drexl 29, 28 ff.; chap. 96, Drexl 57, 11 ff.; chap. 139, Drexl 92, 1 ff .; chap. 144, Drexl 99, 6 ff .; chap. 147, Drexl 102, 1; chap. 153, Drexl 111, 24; chap. 176, Drex1 137, 22 ff .; chap. 194, Drex1 148, 18 ff .; chap. 199, Drexl 156, 6 ff.
<sup>152</sup> Chap. 26, Drexl 23, 23 ff.; chap. 164, Drexl 217, 1 ff.
#### 34 CHAPTER ONE
into the Greek *Oneirocriticon.* Whenever the Arabic narrations describe the way a specific dream had been interpreted either by the Prophet himself or by another important figure, such as Ibn Sirin, the authenticity of the interpretation is guaranteed by a "chain of authorities" *(isndd)* which precedes the narration, just as a similar chain of authorities precedes the *f:zadith,* or traditions of the Prophet Mul).ammad. The thirteen anecdotes in the Greek text have been stripped of their preceding chains of authorities, but the story in chapter 19 apparently had a chain of authorities that reached all the way back to the protagonist himself, that is, to lbn Sirin, and this is why it is narrated in the first person. The chains for the other twelve must have reached back only as far as someone who had heard the story, which explains why they are narrated in the third person. <sup>153</sup>
Unlike the other dream interpreters, Syrbacham, Baram and Tarphan, Achmet, the son of Sereim, is the name of a recognizable historical figure, the seventh-century scholar Mul).ammad ibn Sirin. 154 • Axµ£1 is the Arabic AQ.mad (~I), transliterated into Greek letters. The transformation of "Sirin" to "Sereim" (pronounced in Greek as "Sirim") is easy to explain: the confusion between v (n) and µ (m) is frequent in Greek manuscripts written in the minuscule. Its several variants found in the Greek manuscript tradition indicate the difficulty the Greek scribes had in accurately reading and copying such a foreign name. Ibn Sirin was not, however, a contemporary of Caliph al-Ma'miin (813-33), but must have become associated with him in later literature, at a time sufficiently removed from the actual lifetime of both.
Achmet, son of Sereim, is the only name of a dream interpreter mentioned in the *Oneirocriticon* that does not appear to be fictional. This is apparently why it was attached to the title of the work, probably by one of the scribes who copied the Greek text. The work thus became known as *The Oneirocriticon*
<sup>153</sup> Similar anecdotes pertaining to astrological questions addressed to Abu Ma'shar are contained in the 10th-century Byzantine translation of Abu Sa'id Shadhan's *Mudhakarat,* which constitutes the second book of the Greek *Ta µvcnrjpw wii 'Anoµaaap* (The Mysteries of Apomasar). A number of these anecdotes were published in *CCAG,* vol. 5:1, pp. 144 ff. Two of them are preceded by short chains of authorities: £tn£V 6 'Arroaai 't on drrov i:tj} 'Arroµacmp on dm: µot 6 Mouxouµ£t 6 UtO~ WU Mffi<JE(l)~ WU Xopacrµtcit on ... (Abu Sa'id said that I said to Abu Ma'shar that Mut,iammad b. Musa al-Khuwarizmi told me that. .. ) (ibid., p. 146). Also Eh£v 6 . Arroµacmp on ~Koucra WU Mouxouµn WU UtOU WU ernoouA.ou Myovw~ on ... (Abu Ma'shar said, "I heard Mut,iammad b. 'Abd Allah [b. 'Umar b. al-Bazyar] say ... ") (ibid., p. 147).
<sup>154</sup> The spelling of "At,imad" (....\_.\_\_,.I ) in Arabic is very close to the spelling of "Mut,iammad" (~ ). Both names stem from the same root, *1.z-m-d;* the name of the Prophet Mut,iammad is also said to be At,imad.
of Achmet, although the author/compiler makes no attempt to link himself to that name. We may therefore surmise that the title appearing in some of the Greek manuscripts attributing the Oneirocriticon to "Achmet, son of Sereim, the dream interpreter to Caliph Mamoun," is a later invention, based on the information provided in chapter 19. Indeed, this title is not fully supported by the Greek manuscript tradition, since it appears only in some Greek manuscripts. The various titles of the work recorded in the Greek manuscripts are listed below in chronological order:
Paris. Suppl. gr. 690 (11th century), fol. 125r:< O>verpoxpirov Ίνδών, Περσών και Αίγυπτίων (From the Oneirokrites of the Indians, Persians and Egyptians).
Vindob. philos. et philol. gr. 111 (13th century), fol. 1r: ἐν ὀνόματι τοῦ πατρός και του νίου και του άγίου πνεύματος. Βιβλίον όνειροκριτικόν δπερ συνήξε και συνέταξεν άχμέτ νίός σηρείμ ό όνειροκρίτης του πρωτοσυμβούλου μαμμούν. Πρόλογος των όνειράτων (In the name of the Father and of the Son and of the Holy Spirit. Book on dream interpretation which Achmet, son of Sereim, dream interpreter to Caliph Mammoun [sic], put together and composed).
Marc. gr. 299 (14th-century hand on the flyleaves of a 10th-century manuscript), fol. 5r: Περι όνίρων [sic] (On dreams).
Vindob. philos. et philol. gr. 162 (end of 14th century), fol. 8r: βιβλίον όνειροκριτικόν ὅπερ συνήξε και συνέταξεν αχμετ ό νίος σηρείμ ό όνειροκρίτης του πρώτου συμβούλου μαμούν. Πρόλογος των όνειράτων σύν θεώ άγίω ούτως (Book on dream interpretation which Achmet, son of Sereim, dream interpreter to Caliph Mamoun, put together and composed. Prologue to the dreams, by holy God, thus).
Paris. gr. 2511 (end of 14th century), fol. 7r: βιβλίον όνειροκριτικόν δπερ συνήξε και συνετάξατο άχμέτ, νίος σερήμ του όνειροκρίτου του πρωτοσυμβούλου Μαμούν (Book on dream interpretation which Achmet, son of Serem, dream interpreter to Caliph Mamoun, put together and composed).
Paris. gr. 2419 (15th century), fol. 295: όνειροκρίτης συρίμ (Oneirokrites Syrim [sic]).
Vat. gr. 573 (15th century), fol. 120: βιβλίον όνειροκριτικόν ὅπερ συνῆξε και συνέταξεν αχμετ υίδς σηρείμ ό όνειροκρίτης του πρωτοσυμβούλου uaunv (Book on dream interpretation which Achmet, son of Sereim, dream interpreter to Caliph Mamen [sic] put together and composed).
<sup>155</sup> The Greek word oneirokrites can mean both "dream interpreter" and "dreambook."
Athos, Iviron 4285.165 (15th century, excerpts only), fol. 146a: Ονειροκριτικόν κατά Πέρσας Ίνδούς και Αίγυπτίους (Dreambook according to the Persians, Indians and Egyptians).
Ambros. gr. 592 (O 94 sup) (15th century), fol. 42v. No title. The text of the Oneirocriticon in this manuscript begins with an excerpt corresponding to chapter 124 of the critical edition (Drexl 73, 19 ff.) and gives the title of this chapter as " Έκ των ινδών περι νευμάτων δι' όφθαλμών και περι νυμφεύσεως" (From the Indians on nodding with the eyes and on getting married).
Borbon. gr. 356 (III.E.34) (15th century), fol. 1. No title. The first several folia (probably a whole quire) are missing.
Bononiensis (Bibl. Univ.) 3632 (15th century), fols. 442-446: The top margin of the page with the table of contents has the following inscriptions: συρήμ (Syrem); slightly further down: πίναξ του όνείρατος ινδών, περσών, αίγυπτίων (Table [of contents] of the dream [sic] by Indians, Persians, Egyptians). The first chapter in the table of contents is: πρώλογος του σηρίμ του όνηρωκρίτου (Prologue by Serim the dream interpreter). The end of the table of contents is signaled as follows: τελος του όνυροκρίτου συρήμ του όνηρωκρίτου έκ του λόγων ηνδών περσών και εγηπτίων. σηλβαχάμ ονυροκρίτου/ βαραάμ ονυροκρίτου/ ταρφάν ονυροκρίτου (End of the Oneirokrités Syrim the Oneirokritēs from the teachings of Indians, Persians and Egyptians. Of the dream interpreter Selbacham [sic]/ the dream interpreter Baraam [sic]/ the dream interpreter Tarphan). The beginning of the text is inscribed συρίμ ό όνηρωκριτικ < > (Syrim the dream interpreting [sic]). On the left side of this inscription: ἐκ τὸν ήνδὸν. περσον (From the Indians, Persians ... ). On the right side: και εγυπτίον ( ... and Egyptians).
Leidens. Voss. 49 (end of 15th century), fol. 1: άρχη σύν θεώ του όνειροκρίτου (Beginning, with [the help of] God, of the Oneirokrites).
Cantabrigiensis (Trinity College) gr. 1386 (0 8.11, 6102) (15th-16th century), fol. 5: πίναξ του παρόντος κριτικού των όνείρων βιβλίου (Table of contents of the present book on the interpretation of dreams). Fol. 8: βιβλίον όνειροκρίτης.156 όπερ συνήξε και συνέταξεν άχμέτ υίδς σερείμ όνειροκρίτου (Book Oneirokrites; which Achmet, son of Sereim the dream interpreter, put together and composed). The part of the title after the semi-
<sup>1.60</sup> M. R. James, The Western Manuscripts of the Library of Trinity College, Cambridge, vol. 3 (Cambridge, 1902), p. 398, read this as βιβλίον όνειροκριτικής, but what is written is clearly όνειροκρίτης. The accent is clearly placed on ι; η and ς of the ending are placed on top of each other above the letter T.
colon, found in other manuscripts, is here incorporated into the text of the introduction: Tou rrpc&wu cruµpouA.ou µaµouv rroA.M Komacrav't°'; Eupfo8m 't0 OEcr7tO'tlJ µou 'ti]v aKptPfi £pµ11v£ia 'tWV ovEipmv ... (Of Caliph Mamoun after I greatly toiled to find for my master the exact interpretation of dreams ... ) *[sic,* without punctuation].
*Vindob. philos. et philol. gr.* 287 (15th-16th century), fol. 1: No title.
*Bero!. gr.* 171 *(Phil. gr.* 1575) (I 6th century), fol. l: PtPA.iov· 6vEtpoKp11- 't£tKov: axµe't utou crnpi]µ 'tou 'tfii; rrapoucrrii; cro<j>iai; ot~i; Ti 'tou µ£A.A.ovwi; EKPacrti; rrpoytvrocrKE'tat (Book; on dream interpretation: by Achmet son of Seirem of the present wisdom [sic] through which the outcome of the future becomes known in advance).
*Vindob. philos. et philol. gr.* 297 (I 6th century). The first folio with the title of the work is missing. A later hand attributed the work to arroµacrapoi; (Abu Ma'shar).
*Paris. gr.* 2427 (16th century). This is a direct copy of *Vindob.* 297. It therefore also omits the title of the work and the first chapter and attributes the work to arroµacrapoi; (Abu Ma'shar).
*Paris. gr.* 2538 (16th century). The first folio with the title is missing (this is Rigault's codex).
*Zagora,* (Bibliotheke Zagoras, Thessaly, Greece) 89 (1594), fol. 4: PtPA.iov ovEtpoKptnKov orrEp cruvfisE Kat cruve'taSE na'. E'tEt [sic] 6' EcrEtpi]µ 6 ovEtpoKpi 'tlli; wu rrpmwcruµpouA.ou Maµouv (Book on dream interpretation which Eseirem, dream interpreter to Caliph Mamoun, put together and composed in the year 1601 [sic]).
BL *Additicius* 8240 (17th century), fol. 124v. No title. The manuscript contains an excerpt corresponding to chapter 247 of the critical edition (Drexl 203, 27 ff.) and gives the title of this chapter: EK 'tWV 7tEpcrO:iv Kat at yu7t'timv 7t£pt ota<j>6pmv dorov (From the Persians and Egyptians on various goods).
*Hierosol.* (of St. Sabbas) *gr.* 555 (17th century), fol. I:' Ev 6v6µan 'tOU 7tmpoi; Kat 'tOU UlOU Kat 'tOU ayiou 7tVEUµmoi;· PtPA.iov OVEtpOKpt 'tlKOV, orrEp cruvfisEv Kat cruv£'tas£v 'Axiµ 6 rrprowi; fiµO:iv ovnpoKpi 'tlli; 'tou rrpffiwu cruµpouA.ou Maµµouv K'tA. (In the name of the Father and of the Son and of the Holy Spirit; book on dream interpretation, which Achim [sic], our foremost dream interpreter to Caliph Mammoun *[sic],* put together and composed, etc.).
*Hierosol.* (of the Patriarchate) *gr.* 220 (17th century), fol. 2: EV 6v6µan 'tOU rrmpoi; Kat 'tOU UlOU Kat 'tOU ayiou 7tVEUµmoi; vuv Kat ad. BtPA.iov
#### 38 CHAPTER ONE
ov£tpoKptnKov, 'tO 67totov 'tO £µcit;co~£v Kat 'tO £Kaµ£v 'Axa£'t [sic] o u1o-; L£tpi]µ 6 6v£tpoKpt Til-; wu 7tpcowcruµ~out.ou Maµouv· Kat 7tpol.oyo-; wu ov£tpoKpt wu aATJ8Ecr'tmo-; (In the name of the Father and of the Son and of the Holy Spirit, now and forever; a book on dream interpretation, which Achaet [sic], the son of Seirem, dream interpreter to Caliph Mamoun, compiled and made. And a most truthful prologue to the *Oneirokrites).*
*Petropolitanus Bib!. Acad. scient. graec.* 161 (olim Instituti Archaeolog. Constantinopol.) (18th century), fol. 35:'EK 'tOU ov£tpoKpt'tou wu Tapa<j>av wu cro<j>ou (From the *Oneirokrites* by Taraphan [sic] the Wise).
In addition to the Greek manuscripts that preserve the *Oneirocriticon,* we should also consider some indirect evidence for its title. The excerpt of the *Oneirocriticon* contained in the eleventh-century anthology known as the *Florilegium Baroccianum* is inscribed "'tou Ltptµ" (by Sirim). 157 In the twelfth century, Pascalis Romanus, who used the Greek texts of both the *Oneirocriticon* and Arternidoros to compile his *Liber Thesauri Occulti,* enumerates his sources in rather vague terms: 158
Collectus autem est liber iste ex divina et humana scriptura, tam ex usu experimenti quam ex ratione rei, de Latinis, Grecis et Caldeicis et Persis et Pharaonis et Nabugodonosor annalibus in quibus multifarie sompnia eorum sunt exposita. Fuerunt enim Pharao et Nabugodonosor amatores futurorum et quia prophetas non habebant, velud gentiles, dedit eis Deus per tegumentum sompnii futura conspicere.
This book has been compiled from the divine and human scripture, both from experience and from logical deduction, from Latin, Greek and Chaldean and Persian [writings] and from the annals of the pharaoh and Nabuchodonosor, where their dreams are explained in many passages. For both the pharaoh and Nabuchodonosor loved to know the future and, since they were gentiles and did not have prophets, God gave them [the grace] to know the future through the veil of a dream.
Though an informed reader will realize that both Artemidoros and the *Oneirocriticon* are implied in this enumeration, the names Artemidoros, Achmet, and Sirim are not mentioned. The information provided by Pascalis Romanus suggests that his copy of the *Oneirocriticon* did not mention the
<sup>157</sup> Critical edition by Sargologos, *Un traite de vie spirituelle et morale du Xie siecle,* p. 838 (chap. 24, 20). The date of the *Florilegium Baroccianum* is deduced from the age of its oldest manuscript, *Patmiacus* 6 (11th century).
<sup>158</sup> Collin-Roset, "Le *LiberThesauri Occulti,"* p. 147, §II. The same text is also quoted from Bodleian *Digby* 103 in C. H. Haskins, "Leo Tuscus," *BZ* 24 (1923-24), p. 47.
name of its compiler. This hypothesis is corroborated by the similarity of Pascalis Romanus's text to the versions of the Greek text preserved in *Paris. Suppl. gr.* 690 (eleventh century) and *Marc. gr.* 299 (on a fly leaf written by a fourteenth-century hand), neither of which gives the Greek author's name.
The second Latin version of the *Oneirocriticon* from the twelfth century by Leo Tuscus is not an adaptation but a translation. It is therefore reasonable to assume that Leo faithfully reproduced not only the text but also the title of his model. The dedication reads as follows: 159 "Ad Hugonem Eterianum doctorem suum et utraque origine fratrem Leo Tuscus imperatoriarum epistolarum interpres de sompniis et oraculis" (Leo Tuscus, the interpreter of imperial letters, <dedicates this work> on dreams and oracles to Hugo Eterianus, his teacher and brother on both [the maternal and the paternal] side). The designation of the *Oneirocriticon* as "de somniis et oraculis" implies that the title found in Leo's Greek model did not give the name of the author. A statement from his introduction suggests that the Greek title of the work he translated was *Oneirokrites* (which in Greek can mean both "dream interpreter" and "dreambook"), though the rhetorical manner in which it is phrased does not allow us to draw such a conclusion with certainty:
Ex eo igitur tempore pectus sollicitudine percussi sub corde ignitos versavi carbones cogitando uti Jene esset annon si <sup>160</sup>*onirocriti* Grecorum philosophis ariolanti loqui latine persuaderem enucleatim atque inoffensam perspicuitatem figmenti sompnialis tuo favore nostrorum Tuscorum desiderio breviter reserarem.
Therefore, since that time, I have anxiously smitten my chest and have been stirring live coals under my heart thinking whether or not it would be useful if the *Oneirokrites* who divines for the philosophers of the Greeks spoke in Latin, and if I quickly disclosed for your sake the smooth lucidity of the images of dreams at the request of our Tuscans.
The titles seem to fall into five categories:
1. Titles that do not give the name of an author *(Paris. Suppl. gr.* 690, Leo Tuscus, *Marc. gr.* 299, *Leidem. Voss.* 49, *Athos,Jviron* 4285.165). This category is represented by the two oldest pieces of evidence that survive (the eleventh-
<sup>159</sup> I am copying from the oldest surviving manuscript of this translation, Bodleian *Digby* 103, fol. 59r.
<sup>160</sup> In the manuscript Wolfenbtittel, *Guelpherh. lat.* 2917, quoted in C.H. Haskins, *Studies in the History of Mediaeval Science,* 2nd ed. (Cambridge, Mass., 1927), p. 217: "cogitando utile esse si" = "thinking it will be useful if .... "
century *Paris. Suppl. gr.* 690 and the twelfth-century evidence of Leo Tuscus), as well as three later Greek manuscripts. <sup>161</sup>
2. Titles that attribute the work to "Sirim" without giving further information about the author *(Florilegium Baroccianum,Paris. gr.* 2419, *Bononiensis (Bibi. Univ.)* 3632).
3. Titles that attribute the work to a7toµacmpoc; *(Vindob. philos. et philol. gr.* 297 and its direct copy, *Paris.gr.* 2427).
4. One title that attributes the work to Tarphan *(Petropolitanus Bibi. Acad. scient. graec.* 161 ).
5. Titles that attribute the work to "Achmet, son of Sereim, dream interpreter of Caliph Mamoun" *(Vindob. philos. et philol. gr.* 111, *Vindob. philos. et philol. gr.* 162, *Bero!. gr.* 171 [Phil. gr. 1575), *Vat. gr.* 573, *Paris. gr.* 2511, *Cantabrig. gr.* 1386 (0 8.11, 6102), *Zagara* 89, *Hierosol.* 555 and *Hierosol.* 220), the most numerous category, but also the most recent, since its earliest member belongs to the thirteenth century.
The analysis suggests that the *Oneirocriticon* was originally circulated anonymously. This is supported by our two oldest versions (the abridgment contained in *Paris. Suppl. gr.* 690 and Leo Tuscus's translation), and is corrosborated by Pascalis Roman us' s compilation, as well as the attribution of the *Oneirocriticon* to Tarphan, Sirim, or Achmet; all these attributions must have been made by the scribes who copied the text and who wanted to attach an author's name to an originally anonymous work. The original title of the work, according to the earliest surviving testimony, *Paris. Suppl. gr.* 690, was "6v£tpoKpi n1c;," which is also implied by Leo Tuscus. This title is recorded in the *Leidensis,* and is also repeated in conjunction with the names of putative authors in the *Petropolitanus* (Tarphan), *Paris. gr.* 2419, and the *Bononiensis* (Sirim). In addition, the *Cantabrigiensis* clearly calls the work "PtPA-iov 6v£tpoKpi -cric;," though this might be the mistake of the copyist, who possibly misread the abbreviation for the ending "-t Kov" found in his model. <sup>162</sup>
<sup>161</sup>*Leidens Voss.* 49 (L) was considered by Drexl to be one of the best Greek manuscripts of the *Oneirocriticon,* but a new study of the manuscript tradition might change this evaluation.
<sup>162</sup>The same conclusions about the spurious attribution of the *Oneirocriticon* to Achmet are set forth much more briefly in Lamoreaux, "Dream Interpretation in the Early Medieval Near East," without reference to my dissertation. I consider the concurrence of our opinions as a corroboration of the arguments presented here.
#### The Author's Sources
The first chapter of the Oneirocriticon, where the compiler explains the reason for writing his book and the method of its composition, reads as follows (Drex1 1, 3-14):163
α΄ Πρόλογος τών όνειράτων
Πολλά κοπιάσας πρός το άνερευναν τφ δεσπότη μου τὴν ἀκριβῆ έρμηνείαν των όνειράτων, καθώς αυτός δι' ἐπιθυμίας εἶχε πολλής, ευρον ἐκ τῶν ποιησάντων τὴν τοιαύτην ἀκρίβειαν κατ ἀλήθειαν, ἤτοϊ Ινδών, Περσών καὶ Αἰγυπτίων, οῖ τὴν άλήθειαν ἀκριβολογησάμενοι καὶ λεπτολογήσαντες ἐξέθεντο καὶ έλογογράφησαν τὴν παρουσαν έρμηνείαν. καὶ ἐξ ἐκάστου τούτων ἐκλεξάμενος κεφαλαιωδώς ἐξεθέμην τῶν τριών τὰς κρίσεις καὶ λύσεις ἐν ἐκάστῳ κεφαλαίως, ώσαν και ό ἐμὸς δεσπότης γνοὺς τὴν περὶ τὸ αὐτὸ κεφάλαιον ἐκάστου κρίσιν καὶ λύσιν συλλογίσηται καὶ μάθῃ τὸ ἀληθὲς καὶ πειραθῆ τοῦ γλυκέος καὶ βαθέος και πεποθημένου και δυνατού της παρούσης σοφίας, δι΄ ής ή του μέλλοντος έκβασις προγινώσκεται.
#### 1. Introduction to Dreams
I have greatly labored in order to investigate the accurate interpretation of dreams for my Lord (despotes)-for he was very zealous about such matters-and have discovered that great and true precision has been achieved by some, namely the Indians, Persians and Egyptians. For, having weighed accurately and put in fine detail the truth, they have set forth and written down the present explanation. I have extracted summarily from each of those sources and have arranged the interpretations and solutions of all three of them in every chapter, so that my Lord, too, having learnt the interpretation and solution of each <dream> from that very chapter, may contemplate and perceive the truth, and experience the sweetness, profundity, satisfaction, and power of the wisdom contained therein; for through this wisdom the outcome of future events can be known in advance.
Subsequently, the compiler quotes three more prologues, also written in the first person; they are by "Syrbacham, dream interpreter to the king of India," "Baram, dream interpreter of Saanisan, king of Persia," and "Tarphan, dream interpreter of the pharaoh, king of the Egyptians." The royal dream interpreters speak thus:
<sup>166</sup> The English translation of passages from the Oneirocriticon are taken from Oberhelman, Oneirocriticon of Achmet, with emendations by me. Oberhelman's translation should be used with caution, since it is suffers from carelessness and an insufficient knowledge of Greek. The German translation by Brackertz, Traumbuch des Achmet, is much more reliable.
β΄ Εκ του λόγου των Ίνδών
Συρβαχαμ ό όνειροκρίτης του τών Ίνδών βασιλέως: Σοφία μεγίστη [διαν] ή περί των όνειράτων κρίσις και λύσις και προφητεία άπο θεού πασιν εύαγγελιζομένη, καθό που γέγραπται έν τοις άγίοις εὐαγγελίοις, ότι προς τον ἀγαπῶντά με ἐγώ καὶ ὁ πατήρ μου ἐλευσόμεθα καὶ μονὴν παρ᾽ αὐτῷ ποιήσομεν. τούτο δὲ τελειούται δι΄ όράματος. καὶ μαρτυρεῖ Ίωσὴφ ὁ καταπιστευθείς τὴν Μαρίαν, τὴν μητέρα τοῦ φωτὸς, δι' όράματος μηνυθείς, όμοίως δὲ καὶ Δανιήλ και οί πλείους των προφητών δι' όράσεως το θείον φως έδιδάχθησαν. οις βεβαιούσθωσαν πάντες, ότι θείόν τι μήνυμα περί πάντων, άγαθών τε καὶ φαύλων, παντι το λαφή των όνειράτων όψις έστίν. μὴ λογιζέσθω δέ τις, ότι μία τίς ἐστι κρίσις και λύσις όνείρατος ένὸς τῶν διαφόρων προσώπων. διότι τὰ τῶν βασιλέων όνείρατα οίκείαν έχουσι κρίσιν και λύσιν, και των άρχόντων και πλουσίων όμοίως οικείαν, και των απόρων τε και πτωχών αναλόγως οικείαν; όμοίως δε αλλη κρίσις όνείρων γυναικών και άλλη άνδρών και άλλη κρίσις έν θέρει και αλλη ἐν χειμῶνος ώρα. δι' ών καὶ μᾶλλον ὀφείλει ὁ ὀνειροκρίτης πυκνός τις είναι και πολυμαθής και τον θείον φόβον έχων άεί. οίς και μαλλον ή κρίσις ἐστίν ἀσφαλής, διότι ἀπὸ θεοῦ κεχαρίτωται. οὐ μόνον δὲ τοῖς ἀχαθοῖς ὁ θεῖος ονειρος προδείκνυται άλλά και τοις πονηροίς και άμαρτωλοις, και ταυτα διά το πλούσιον έλεος του θεού, καθώς τρέφει και τούς άρνουμένους και τους βλασφημουντας αύτόν. νύν ουν έν τη δυνάμει της άγίας ανάρχου και άχωρίστου τριάδος τῆς έρμηνείας ἀπάρχομαι. 164
#### 2. From the Account of the Indians
Syrbacham, the dream interpreter of the king of the Indians <said>:
"The interpretation and decipherment of dreams is very great wisdom and prophecy brought by God as glad tidings to all, as is written in the Holy Gospels, that 'to the one who loves Me, My Father and I will come and tarry with him' [John 14:23]. This is fulfilled through a vision. Joseph, the one entrusted with Mary, the mother of Light, who was informed by a vision, bears witness to it.165 Likewise, Daniel and most of the prophets were taught the divine light through <prophetic> visions. By these <examples> all should rest assured that the appearance of dreams is for everybody a divine message concerning everything, both good and evil. And let no one reckon that the interpretation and solution to a dream are the same for different individuals. For the dreams of kings have their own interpretation and solution, and those of the nobles and the wealthy likewise their own, and those of the destitute and the poor their own accordingly. Similarly, the interpretation of women's dreams differs from that of men's dreams.The interpretation is also different during the summer and during the winter season. For these reasons the dream interpreter ought to be someone very wise and extremely learned and God fearing always. Thereby the interpretation is very sound, because it is bestowed by
<sup>164</sup> Drexl 1, 15-3, 24.
<sup>165</sup> See Matthew 1:18-25, 2:13-14 and 2:22,
the grace of God. For God-sent dreams appear not only to the good, but also to the wicked and sinful, because of God's bountiful mercy, just as he takes care even of those who deny and blaspheme him. Now, by the power of the eternal and inseparable Holy Trinity, I begin my interpretation."
#### γ ' Εκ του λόγου των Περσών
Βαράμ ό όνειροκρίτης Σαανισάν βασιλεί των Περσών Έκ τῆς κρίσεως των όνειράτων μεγίστην γνώσιν και πρόγνωσιν έφευρών, ζωής τε και θανάτου, πενίας τε και πλούτου, νόσου και ύγείας, χαράς και λύπης, νίκης έχθρων και ήττης, έλάττονι κόπω το μέλλον άληθώς ἐκμανθάνω ὑπέρ του μεγίστου κόπου της άστρονομίας, δηλονότι έαν άρα ό όνειροκρίτης έστιν άκριβέστατος, έγκοπος γαρ και πολλάκις σφαλλομένη της άστρονομίας ή κατάληψις. πολλοί γάρ, ώς ἐγώ πεῖραν ἔσχον, τῶν ἀστρονομησάντων εἰς ἀλλήλους διέφερον τότε καὶ διηνέχθησαν έτερος τον έτερον άνατρέποντες. ή δε κρίσις των όνειράτων, ην ἐγώ ἐξεθέμην, παντὶ τρόπω τὸ ἀναμφίβολον ἔχει. καὶ ὁ ταύτην μετέρχεσθαι βουλόμενος λεπτότητα διανοίας και πόθον πρός τους κυρίους των άστέρων ἐχέτω. νῦν ουν αُρχομαι περὶ πάντων ἀληθώς καὶ βεβαίως. 16
#### 3. From the Account of the Persians
Baram, dream interpreter to Saanisan, king of the Persians <said>:
"Having discovered through the interpretation of dreams very great wisdom and foreknowledge about life and death, poverty and wealth, sickness and health, joy and sorrow, victory over enemies and defeat, I truthfully study the future at much less toil than if I were to use the very laborious process of astrology, that is, if the dream interpreter is very precise. For the comprehension of astrology is toilsome, frequently yields wrong results, and is also wearisome and very lengthy. Indeed, as I know from experience, many astrologers at times disagreed with each other and each excelled in refuting the other. On the other hand, the interpretation of dreams which I have set forth is in every way unambiguous. And let him who wishes to pursue it have a sharp mind and love for the lords of the stars. So now I begin <to talk> truthfully and with certainty about everything."
#### δ' Έκ του λόγου των Αίγυπτίων
Ταρφαν ό όνειροκρίτης Φαραώβασιλέως των Αίγυπτίων: ' Ηρεύνησα '' περὶ ών εύρον οί βασιλείς των Αίγυπτίων έν ταις κρίσεσι των όνειράτων, περὶ ών οὐκ ϊσχυσεν έτερός τις ούτως καταλαβείν ταιντα ώσπερ έγώ δια τον έμον δεσπότην πολλή γαρ ήν και συνεχὴς αὐτῷ τῶν ἀνειράτων ή ὅψις. καὶ γὰρ καὶ αὐτὸς πολύν είχε πόθον πρός τους θεούς, οϊ πάντα, όσα καθ έκαστον αύτω συμβήσεσθαι έμελλε, προεδείκνυον ἐν ἀνείρασιν. απερ ἐγὸ ἀκριβώς διέλυον ἀεί: καὶ αὐτὸς
<sup>166</sup> Drexl 2, 25-3, 11.
<sup>167</sup> Drexl 3, 14 has "inocuvnou"; I prefer "ipe unoa" on the basis of Vat. gr. 573, as well as the use of the first person singular in this chapter.
#### 44 CHAPTER ONE
()t' £µou TIJV aA.i]0£tav itpocyiyvcocrKE 't£ Kat £UptcrK£. vuv ouv EK'tt0T]µt 1au1a, 7t£pt WY £yoo E7t£tpa0riv Kato\. apxa1m <l>apacovt'tat Kato\. Km' a-uwui; cro<1>oi. Kat itav'ta, ocra EVOEXE'tat 0ccop£l v wui; av0pcimoui;, £pµT]V£UCOV EK'ti0T]µt. <sup>168</sup>
4. From the Account of the Egyptians
Tarphan, the dream interpreter of the pharaoh, king of the Egyptians <said>: "I have researched what the kings of the Egyptians have found in the interpretation of dreams, which nobody else was able to comprehend as well as I have, for the sake of my Lord. For he used to have many and frequent dreams. For he used to have great love for the gods, who foreshadowed every single thing that was about to happen to him in dreams, which I have always interpreted accurately, and through me he learned and knew the truth beforehand. So now I set forth what I and the ancient pharaonites and their wise men know from experience. And I set forth the interpretation of everything that is possible for humans to dream."
As each of the dream interpreters declares that his art draws its power and veracity from its association with the divine, it becomes evident that the Indian interpreter, who quotes a passage from the Gospels and mentions the dreams of Joseph, the husband of Mary, is a Christian, and all subsequent chapters of the *Oneirocriticon* that have an openly Christian flavor, are also labeled "Indian." The Persian dream interpreter worships the stars; the Egyptian dream interpreter and his master believe in several unnamed gods. In other words, the first four chapters of the *Oneirocriticon* claim that the work was compiled on the basis of Christian Indian and pagan Persian and Egyptian sources. The question that arises is whether the designations "Indian," "Persian" and "Egyptian," as well as the interpretations presented under these labels, existed in the Arabic sources of the *Oneirocriticon* or were invented by the Greek compiler. The answer to this question would reveal something about the Greek compiler's approach to his source material.
Neither the compiler of the *Oneirocriticon* nor his Byzantine readers understood "Indians" to refer to the inhabitants of the Indian subcontinent. In tenthcentury Greek, the word "Indian" could be taken to mean, more or less, an eastern Christian. All eastern nations, including Christian ones, are called "Indian" in Greek sources as early as the fourth century .169 W arks such as
<sup>168</sup> Drexl 3, 12-24.
<sup>169</sup>See I. Shahid, *Byzantium and the Arabs in the Fourth Century* (Washington, D.C., 1984), pp. 86-106, discussing the diplomatic mission ofTheophilos the "Indian" among Oriental Christians, ca. 356. Christians existed on the Indian subcontinent long before European colonization. According to legend, their church was founded by the apostle Thomas. The earliest physical remains of Christianity were discovered in southern India and consist of five stone crosses inscribed in
Palladios's *On the Races of India and the Brahmans,* Pseudo-Kallisthenes' *Alexander Romance* and its numerous versions from the fourth century onwards, 170 and the immensely popular romance *cum* hagiography *Barlaam kai loasaph,* portray India as the home of a pious and wise people. At times, and especially from the seventh century onwards, India was confused with Christian Ethiopia, which was called "inner India" in earlier sources and through which Indian goods were imported to Byzantium. The same confusion of India with the lands south of Egypt is evident in Artemidoros' s ninth-century translation into Arabic by ijunayn b. lsl)aq, where the Greek "India" *C* Ivoia) is rendered as "the land of Nubia" (~\_,\_\_:,\_JI ..i~). <sup>171</sup>It is from this literary world that the Christian Indians of the *Oneirocriticon* hail.
The only possibly genuine Indian trace in the Byzantine dreambook is the name of the Indian dream interpreter, Syrbacham, which could be a corruption of the Sanscrit sri Brahmanah, meaning "the reverend Brahman." 172 The most likely manner for such an Indian element to have ended up in a Middle Byzantine text is by way of an Arabic source, since the Arab acquaintance with the Indian subcontinent and its learning began with the first Arab conquest of Sindh under Mul)ammad b. Qasim in 712. Indeed, a comparison of the "Indian" chapters of the *Oneirocriticon* with the interpretations given in the Indian dreambook of Jagaddeva, 173 which was written later than the *Oneirocriticon* but is the earliest independent Indian dreambook that survives, 174 shows that
Pahlavi; they have been assigned to the 6th or 7th century on paleographic grounds. See M. K. Kuriakose, *History of Christianity in India: Source Materials* (Madras, 1982), pp. 1-9. The Christian Indian tradition is clearly not connected with anything found in the *Oneirocritirnn.*
<sup>170</sup> Both Palladios and the Alexander Romance were copiously excerpted in the 10th-century *Suda Lexicon.*
<sup>171</sup> Cf. Artem. ii, 12; Pack 123, 2; Arabic text in Fahd, *Artemidore d'Ephese, Le livre des songes,* p. 224, 3. The fact should hardly be surprising, since the medieval translators of classical Greek texts into Arabic did not understand Greek in the way the ancient authors did, but in the way the contemporary Byzantines did, because they learned their Greek from Byzantine teachers. To give one example, I:Iunayn b. IsJ:iaq is said to have learned Greek after spending a couple of years in the lands of the Byzantines; see *E/2,* s.v. "I:Iunayn b. IsJ:ia~."
<sup>172</sup> Suggested by G. Dagron, "Formes et fonctions du pluralisme linguistique a Byzance (IX-XII siecle)," *TM* 12 ( 1994), p. 237. Another Indian name could possibly be masked under "Syrbacham"; this could be Varahamihira, a famous Indian astrologer and diviner of the 6th century c.E. with whom Arabic science became acquainted mainly through the works of Abu Ma'shar (787-886) and al-Blriini (973-1048). On Varahamihira, see Ullmann, *Die Natur- und Geheimwissenschaften im Islam,* p. 301; see also D. Pingree, *Jyotil,l.fostra: Astral and Mathematical Literature,* A History of Indian Literature, ed. J. Gonda, vol. 6:4 (Wiesbaden, 1981 ), pp. 74-76.
*m* Text and translation in J. von Negelein, *Der Traumschlussel des Jagaddeva. Ein Beitrag zur indischen Mantik* (GieBen, 1912).
<sup>174</sup> Indian dream interpretation was usually treated in a chapter of longer works on omens. The
#### CHAPTER ONE
similarities are rare and should not be attributed to any direct influence. 19 Aside from the titles of the chapters labeled "from the Indians," India is mentioned only once in the Oneirocriticon, in the chapter on the interpretation of elephants according to the Persians and Egyptians (Drexl 221, 5-6): 'O έλέφας είς ύψηλον ανδρα ξένον έξουσιαστὴν πολύπλουτον κρίνεται δια το μὴ πανταχού θηρασθαι πλὴν ἐν Ίνδία (An elephant is interpreted as an exalted man who is a very wealthy foreign ruler, because it is hunted nowhere else but in India). The same interpretation of elephants also occurs in Arabic dreambooks, without the justification quoted in the Oneirocriticon. However, one of the Arabic dreambooks does state that the significance of an elephant dreamt in India is different from that of the same dream dreamt in other parts of the world.176 The "Indian" chapters of the Oneirocriticon that discuss religious notions are in fact copied or adapted from the Muslim
175 For some comments on the connection of the Oneirocriticon with Indian dream interpretation, see Negelein, Der Traumschlüssel des Jagaddeva, pp. XX-XXI. Negelein points out the similarity between Jagaddeva I, 17 and the Oneirocriticon, chap. 301 (Drexl 241, 1-14). K. Latte, review of Achmetis Oneirocriticon, Gnomon 2 (1926), p. 419, remarks: "Die Berührungen mit dem von Negelein in seiner Ausgabe des Traumschlüssels des Jagaddeva zusammengestellten Material sind spärlich (vgl. etw 61, I=J. I 48 p. 60 N., aber auch Artemidor I 33 p. 34, 23 Hercher; 77, 24=J. II 63.76.25, 12 und Negelein p. XIV 2)."
176 Al-Muntakhab, p. رؤية الفخيل في غيـر بلاد الهند شـدة وفــز ع (The dream of an elephant outside of India <signifies> difficulty and distress).
earliest independent treatise on dreams (svapna) that survives is the Svapnacintamani by the Gujarati scholar Jagaddeva (ca. 1175), but the interpretation of dreams goes back in Sanskrit literature to the Rgveda. The most common Sanskrit text on the subject is the Svapnadhyaya of unknown date, attributed to Brhaspati (Jupiter). The manuscripts of this work have been inventoried in D. Pingree, Census of the Exact Sciences in Sanskrit. Series A, vol. 4, Memoirs of the American Philosophical Society 146 (1981), pp. 250-51; none of them is earlier than the 17th century. However, this should not be considered as conclusive evidence for the date of the work, since Indian manuscripts earlier than the 17th century are extremely rare (see Pingree, Jyotihsastra, p. I 18). For a comparison of ancient Greek with ancient Indian dream interpetation, see R. Stuhrmann, "Der Traum in der altindischen Literatur im Vergleich mit altiranischen und griechischen Vorstellungen," Ph.D. diss., Eberhard-Karls-Universität Tübingen, 1982. This study does not discuss any Greek dreambooks later than Artemidoros and concentrates mainly on the theoretical aspects of dream interpretation, though a few specific dream symbols are also considered. The relationship between Indian and Islamic dream interpretation has never been studied. Islam reached India as early as 712 and was firmly established in the Punjab and Kashmir by the first half of the 11th century. It is therefore possible that Islamic influences made their way into the twelfth-century dreambook of Jagaddeva. Regarding Indian influences on Islamic dream interpretation, the source and authenticity of the interpretations attributed to the Indians in the eleventh-century dreambook of al-Dinawari still need to be studied. On Indian dream interpretation, see Pingree. Jyotihsastra, chap. 4, "Divination," and esp. p. 77. I would like to thank Professor Michael Witzel of Harvard University for numerous bibliographical references and for a very useful discussion on Indian dream interpretation.
interpretations of analogous symbols in Arabic dreambooks.
As for the purportedly Persian and Egyptian interpretations in the *Oneirocriticon,* many of them can be found in Arabic dreambooks, though no items peculiar to the material culture of ancient Egypt or Persia are mentioned in the Greek text. <sup>177</sup>But a handful of interpretations, mainly those that discuss religion and social customs, appear as specific to these two cultures. It is important to decide whether these details reflect a Byzantine or an Arabic image of Persia and Egypt and how accessible the pertinent information was to a Byzantine or Arabic author of the ninth and tenth centuries.
The Persian dream interpreter, Baram, and his king, Saanisan, have genuinely and almost generically Persian names. Saanisan appears to be a corruption of Sasan, an ancestor of the last dynasty to rule Persia before the Arab conquests. His identification in the *Oneirocriticon* as "king of the Persians" is inaccurate, however, because, though of noble lineage, Sasan never became king. But Arabic sources refer to the last imperial house of Persia as Banii Sasan (the clan of Sasan), and it was a genuine Persian name with royal associations that was widely known, even to those who were not knowledgeable in history. <sup>178</sup> Baram is Bahram, also a common Persian name, borne by the great warrior god of Zoroastrianism, 179 and by six Sasanian kings and several notables of
178 Al-Mas'iidi, the 10th-century geographer, records a popular belief that the last ancient Persian to perform the pilgrimage to Mecca was Sasan, whose ritual murmuring *(zamzama)* gave the name to the sacred well of Zamzam. Thus the name "Sasan" that occurs in the *Oneirocriticon* might have been introduced to the Arabic source of the Greek text by a Muslim author because of both its royal Persian connotations and its Islamic acceptability. See al-Mas'iidi, *Les prairies d'or,* trans. B. de Meynard, P. de Courteille and C. Pellat (Paris, 1962), vol. 1, p. 215, §574 (II, 148). Alternatively, assuming that the name Sasan already occurred in the Arabic source of the *Oneirocriticon,* it might have been mentioned in order to remind the Arab reader of an association with apocryphal wisdom and magic, since "banii Sasan" was the name applied to tricksters and beggars, who were often practitioners of magic and quack doctors. Cf. *E/2,* s.v. "Sasan" and Bosworth, *Medieval Islamic Underworld,* vol. I.
179 Interpreted in ancient Greek and Hellenistic sources as Ares (Mars) or Herakles (Hercules). See M. Boyce, *A History of Zoroastrianism,* vol. 2 (Leiden, 1982), pp. 40-41.
<sup>177</sup> According to Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 256-57, the interpretations that the *Oneirocriticon* ascribes to the Persians and Egyptians cannot be found in the extant Arabic dreambooks. Even in the chapters ascribed to the Indians, what seems to have been based on Islamic sources are only the interpretations presented at the very beginning; the rest might have been invented by the Greek author under the influence of the general idea in his Arabic source. Both these statements stand in need of correction. For examples of interpretations attributed to the Persians and Egyptians that have their equivalents in Arabic dreambooks, see the discussion on the interpretation of heads in chapter 5; also the interpretations of worshiping a star or a tree, receiving something from the pharaoh and sleeping in the same bed with him quoted below in this chapter. An example of a Christian chapter from the *Oneirocriticon* whose contents all have parallels in Islamic interpretations is chap. 149 (Drexl 103, 25-105, 11); see the discussion on the interpretation of priests and priestly duties in chapter 7.
the Sasanian and later periods. 180
The religion of the Persians, as it can be gleaned in chapter 13 (Drex) 10, 5s11, 9), is Sasanian Zoroastrianism :
ιγ ' Έκ του λόγου τών Περσών περι πίστεως
Έάν τις ίδη κατ' όναρ είδώλοις έαυτον προσκυνουντα τοις όπό τοις όνόμασι των μεγίστων ἀστέρων, ἢ τοῦ ήλίου ή τὸ εἴδωλον, ἐπιδεηθήσεται βασιλέως καὶ είσακουσθήσεται και έγγιει τώ βασιλεί όσον τω ήλιακο είδώλω κατά τους υπνους ... Εί δε το είδωλον ή είς πρόσωπον τῆς σελήνης και ίδη τουτό τις, ἐπιδεηθήσεται του πρώτου ἀνθρώπου του βασιλέως και εἰσακουσθήσεται προσεγγίζων αὐτῶ, ὅσον τῷ σεληνιακώ εἰδώλω κατά τοὺς ὕπνους, ἐάν τις Ἱδη κατ΄ όναρ προσκυνουντα έαυτον τω είδώλω της Αφροδίτης, είς πρώτον ανθρωπον τῆς αὐγούστης εὐδοκιμήσει. δμοίως καὶ ἐπὶ τῶν λοιπῶν ἀστέρων καὶ τῶν μεγιστάνων του βασιλέως ή αύτη κρίσις.
Εάν τις ίδη προσκυνούντα έαυτόν τώ πυρί τω έν τοις ναοις άσβέστω, εί μέν βασιλεύς ή και ίδη, ότι ώσει καπνός δυσώδης έξελήλυθε του πυρός, πολεμηθήσεται παρ' έχθρών και τραπήσεται: εί δε καπνός ήν ευώδης, έξάκουστον νίκην κατ΄ ἐχθρών λάβη. εἰ δὲ τις ϊδη τοῦτο τῆς κοινότητος τοῦ λαοῦ, δουλεύσει βασιλεί και άποβήσεται άναλόγως της εύωδίας ή δυσωδίας και γυνὴ έάν ή, πρός τον ανδρα αύτης όμοίως της όσμης αποβήσεται αύτη. Έάν τις ίδη, ότι ἔκτισε ναὸν καὶ ἔθετο πῦρ ἐν αὐτῷ εἰς προσκύνησιν τοῦ λαοῦ, εἰ μέν ἐστι βασιλεύς, ἐν βουλή αὐτοῦ βασιλέα ποιεῖ καὶ ὁ λαὸς ἀποδέχεται αὐτόν. 180
13. From the Account of the Persians on Faith
If someone dreams that he was worshiping the idols that are named after the greatest of the stars, should the idol be that of the sun, he will beseech the king and [his wish] will be granted and he will approach the king as closely as he did the idol in the dream .... If the idol depicted the moon and someone sees it, he will beseech the first minister to the king and [his wish] will be granted, and he will approach him as far as he approached the idol of the moon in his sleep. And if someone dreams that he was worshiping the idol of Venus, he will become the most trusted person of the queen. And likewise, the same interpretation applies to the remaining planets and the king's magistrates.
If someone dreams that he was worshiping the eternal flame in the temples, if he is king and saw something like a malodorous smoke arise from the fire, he will be attacked by enemies and will be routed; but if the smoke is sweet-smelling, he mav
<sup>180</sup> This name appears in the Greek sources as Βαράνης, Βαραάνης and Βαραράνης, but also as Bapau, e.g., in Theophylaktos Simocatta; for a list of occurrences, see Theophylaktos Simocatta, Historiae, ed. C. de Boor and P. Wirth (Stuttgart, 1972), p. 320. On the kings by that name in history and mythology, see Encyclopaedia Iranica, s.v. "Bahrām."
<sup>181</sup> Drexl 10, 5-11, 9. In addition, the Persians appear as star worshipers in chap. 168, "From the Persians on Stars" (Drexl 131, 23-132, 3), and in the introduction to the Oneirocriticon by the Persian dream interpreter (Drexl 2, 25-3, 11) quoted above.
gain a famous victory over them. If a commoner dreams this, he will serve a king and the result will be analogous to the foulness or sweetness of the smell. If the dreamer is a woman, it will turn out for her likewise in regard to her husband. If someone sees that he built a temple and set up a fire for public worship, if he is king he will appoint someone as king according to his will and the people will accept him.
In chapter 169, "From the Persians on the Stars" (Drexl 131, 23~132, 3) it is said that incense burning and sacrificing formed part of the ritual worship of the stars.
The image of Zoroastrianism conveyed in these chapters is accurate. <sup>182</sup>In Zoroastrian belief the sun is the symbol of Ohrmazd, the creator and principle of good. The moon is associated with light and water and conveys health and growth to plants. 183 Venus (Anahid) was the tutelary divinity of the Sasanian house. <sup>184</sup>Her cult originated in that of the Assyro-Babylonian Ishtar, the Lady of the planet Venus and of love and war. Anahid was absorbed by Zoroastrianism in the time of the Achaemenians and was worshiped down to Islamic times as Banu Pars ("the Lady of Persia"). <sup>185</sup>Fire is considered sacred, and the maintenance of perpetual flames is central to ritual worship, <sup>186</sup>which also included sacrifice of animals and other offerings. 187 Further chapters indirectly supply a very few details about the beliefs and customs of the Persians. In chapter 6, "From the Persians on Resurrection," it is implied that resurrection
<sup>182</sup>Chap. 159, "From the Persians and Egyptians on Fire" (Drexl 120, 17-122, 25), also includes a "Zoroastrian" interpretation, though its importance is negligible, considering the total length of the chapter: To 1tUp Eii; µqiami; KpivEml KpicrEti;• Eii; 8EWV yap avciyEml 7tpOCTOOJta (Drex] 120, 18-19) .... Ei i\E <Hi!] n<;> on 7tp0CTEKUV1]CTE 'tc\i 7tUpt, µqtmcivoov 0E118Tianm Kat aKoua8ilaEi:m (Drexl 121, 2-3 ). (Fire bears the greatest interpretations, for it refers to gods .... If <someone dreams> that he worshiped fire, he will beseech nobles and be heard).
<sup>183</sup>See *Encyclopaedia Iranica,* s.v. "Astrology and Astronomy in Iran," esp. § § "Sun and Moon," "Stars" (p. 865).
<sup>184</sup>See M. Boyce, *Zoroastrians: Their Reliiious Beliefs and Practices* (London, 1979), pp. 115 and 142.
<sup>185</sup>Ibid., pp. 61-62 and 163.
<sup>186</sup>See A. Christensen, *L'lran sous /es Sassanides* (Copenhagen, 1936), pp. 140-41: "L'Avesta distingue cinq sortes de feu .... *B:ir:izisavah* est le feu du temple (appele le *feu Vahram)* et aussi le feu destine a !'usage ordinaire .... "See also ibid., p. 157: "Correspondant aux grades du regime patriarcal des anciens Iraniens, ii y avail un feu de maison, un feu de clan ou de village *(iidhuriin)* et un feu pour chaque canton ou province. Ce dernier est appele feu de *Varhriin* (Vahram). Tandis que le feu de maison etait entretenu par le miinbadh, le maitre de maison, deux pretres au mains etaient necessaires pour le service de *l'adhuriin* et le feu de Varhriin demandait un corps de prerres plus nombreux sous la direction d'un mobadh."
<sup>187</sup> See Boyce, *Zoroastrians,* pp. 53, 66, 75-76, 86, 164, 173-74, 211.
was part of Zoroastrian eschatology; a number of chapters attributed to the Persians and Egyptians together mention polygamy .188 The only inaccuracy that turns up is the implication that the Persians bury their dead, which is found in a chapter attributed to both the Persians and the Egyptians. 189 In fact they disposed of dead bodies by exposing them. <sup>190</sup>
The Greek-speaking world had direct contacts with Persia from antiquity until the Arab conquest. Greek historians from Herodotus to Zonaras wrote about ancient Persia in detail, 191 and information on Persian religion, culture, and wisdom can be found in a vast number of Greek sources. 192 But, in the ninth or the tenth century, only a devoted antiquarian would have been able to piece together an accurate picture of Zoroastrianism, and such a task would have been even more daunting, if not impossible, for someone like the compiler of the *Oneirocriticon,* whose only textual references come from the Bible. On the other hand, Zoroastrianism was part of contemporary reality in the Muslim world. After the Islamic conquest its adherents were accorded the status of *dhimml,* 193 and functioning Zoroastrian temples could be found in the heartland of the caliphate until the eleventh century, as is evident from the information provided by the Arab geographers. 194 Moreover, aspects of the Zoroastrian
191 On the Greek sources pertinent to Iranian history, see the introduction to each chapter in R. Frye, *The History of Ancient Iran,* Handbuch der Altertumswissenschaft 3:7 (Munich, 1983).
192 Most of these references have been collected and studied in J. Bidez and F. Cumont, *Les mages hellenises. Zoroastre, Ostanes et Hystaspe d'apres la tradition grecque,* 2 vols. (Paris, 1938).
194 See B. M. Tirmidhi, "Zoroastrians and Their Fire Temples in Iran and Adjoining Countries from the 9th to the 14th Centuries," *Islamic Culture* 24 (1950), pp. 271-84. For the survival of Zoroastrianism under Islamic rule, see M. Boyce, *Zoroastrianism: Its Antiquity and Constant*
<sup>188</sup> Drexl 20, 18-19; 56, 6-12; 64, 21-23; llO, 15-17; 120, 22-24; 173, 20; 175, 27-176, 3; 178, 7-8; 182, 25. For polygamy in Zoroastrianism, see Christensen, *L' Iran sous les Sassanides,* pp. 317 ff. For ancient Egyptian polygamy as described in medieval Arabic sources, see Murta<.la ibn al-Khafif, *L' Egypte de Murtadi.fils du Gaphiphe,* ed. G. Wiet, trans. P. Vattier (Paris, 1953), p. 34.
<sup>189</sup> Chap. 122, Drexl 86, 21-89, 11.
<sup>190</sup> It is known that in the 9th-10th century, Zoroastrians erected high walled enclosures ( *dakhma,* funerary towers) for that purpose, in order not to offend Muslims and to shield the dead from the risk of violation; see Boyce, *Zoroastrians,* pp. 90-92 and 157-58.
<sup>193</sup> The *ahl al-dhimma* ("the protected people") or *ahl al-kitiib* ("people of the Book") were non- Muslims who could not be forcibly converted to Islam and, though subject to social restrictions, were guaranteed protection and autonomy of institutions by the ruler in exchange for special taxes. Their religions were respected because, according to the Qur'an, they possessed a scripture containing divine revelation. The Qur'an did not clearly include Zoroastrians among the "people of the Book" but, although the matter was open to interpretation, the status of *dhimmi* was conferred upon them; see R. Frye, *The Golden Age of Persia* (New York, 1975), p. 135.
religion were treated, though never in much detail, by authors such as Ibo al-Nadim in the *Kitab al-fihrist,* 195 Ibo Tahir al-Baghdadi in *al-Farq bayn al-firaq,* 196 and Shahrastani in *Kitab al-milal wa-al-nif:zal. <sup>197</sup>*Further proof that whoever conceived the interpretations in the *Oneirocriticon* was familiar with Zoroastrianism as a living religion can be found in chapter 12, which discusses the significance of dreams that involve abandoning Christianity in favor of Judaism, Islam or Zoroastrianism (Drexl 8, 10-12 and 9, 6-11). <sup>198</sup> Including such a dream indicates that Zoroastrianism was a religion practiced in the author's milieu, as was the case in Muslim lands, but not in Byzantium.
The most important indication that the Greek compiler of the *Oneirocriticon* did not invent the "Persian" interpretations but copied them from an Arabic model is the fact that these interpretations can also be found in Arabic dreambooks. The interpretation of the sun, the moon and the planets as representing the king, his prime minister and members of the royal court found in the *Oneirocriticon* is also found in every single Arabic dreambook, 199 along with the interpretation of fire as king. 200 The portrayal of Persians as star worshipers, however, may not have been based on direct and accurate information obtained by the Arabic author of the Persian interpretations that ended up in the *Oneirocriticon,* but simply inspired by a tendency detectable in the Arabic sources to consider all pagan religions-that is religions other than Judaism, Christianity, and Islam-as cults of the stars with rituals that included the worship of and sacrifices to idols.201
The *Oneirocriticon* is less knowledgeable about ancient Egypt than about Zoroastrian Persia. Aside from the name of the pharaoh and a reference to
*Vigor* (Costa Mesa, Calif. and New York, 1992), pp. 149-62.
<sup>195</sup> Ibn al-Nadim, *Kitah al-fihrist;* Ibn al-Nadim, *Fihrist,* trans. Dodge.
<sup>196</sup>Ibn Tahir al-Baghdadi, *Moslem Schisms and Sects (al-Far* ti *bain al-Fira!i), Being the History of the Various Philosophic Systems Developed in Islam,* trans. K. Chambers Seelye and A. Halkin, 2 vols., Columbia Oriental Studies 25 (New York, 1920).
<sup>197</sup> For further references, see *EI2,* s.v. "Madjus," esp. pp. 1116-18.
<sup>198 •</sup> Eciv 'tl~ rni:i, on µciyo~ EYEVE10, oi'irn~ <i>tA.cipyupo~ Kat <i>tA.OltAOU'to~ yEvl\crE"tat (If someone dreams that he became a magian, he will become avaricious and niggardly). The same interpretation can be found in Arabic dreambooks.
<sup>199</sup>See chap. 166, and especially Drexl 127, 3-7 and 129, 12-18.
<sup>200</sup> Cf. al-Nabulusi, vol. 2, p. 267, s.v . ....>"~ :L..l-0.. ~ <..:.Ll ..JL.:JI ~<..:ii LSl..J j..oJ . 0~ *).:.JI* 0'¥ 0~ (If someone dreams that he was worshiping fire, he will seek to be employed in the service of a king, for the fire is a king).
<sup>201</sup>See J. Hjarpe, "Analyse critique des traditions arabes sur les sabeens l)arraniens," Ph.D. diss., University of Uppsala, 1972, pp. 43-61, and especially §3: "La notion du 'paganisme' comme astrolatrie et idol:ltrie."
#### CHAPTER ONE
ancient Egyptian polytheism, Egyptian references are limited to "the water of the Nile" and "the carafe of Cleopatra." By the time the Oneirocriticon was compiled, the Nile irrigated Muslim territories, and its interpretation as power and money (Drexl 152, 16-19) is consistent with that found in Arabic dreambooks.202 It is unclear whether or not the "carafe of Cleopatra" (to βαυκάλιον της Κλεοπάτρας, Drexl 153, 1) reflects something found in the Arabic sources of the Oneirocriticon (although I was unable to find a similar passage in them) and whether it was a real object203 or invented in order to evoke Egyptian local color.204 Alchemical writings ascribed to Cleopatra exist both in Greek and in Arabic, which implies that the reference to Cleopatra might be explained by her connection, not with Egypt, but with the occult sciences. 205
Both the Byzantines and the Arabs considered ancient Egypt to be the homeland of men versed in apocryphal wisdom, such as Hermes Trismegistus, whose writings had been translated from Greek into Arabic. But the Greek books on dreams called βίβλοι αί "Ωρου και" Ισιδος (Books of Horos and Isis) that are recorded by two second-century authors, Lucian and Dio Chrysostomos, are unlikely to have furnished any of the "Egyptian" material in the Oneirocriticon, since they did not contain interpretations of dream symbols but stories about incubations and supernatural visions.206 Knowledge of the religion and customs of ancient Egypt was very limited among both the
ومن شرب من نهر النيل فانه ينال ذهبا بقدر ما شرب : 289 Al-Muntakhab, chap. 41, p. 289 (Whoever drinks from the river Nile will receive gold commensurate with what he drank). من راى فى اللنام انه يشرب منه فانه ينال : نيل مصر (نهر ) . Al-Nābulusī, vol. 2, s.v. . Whoever sees) ذهبا بقدر ما شرب ۾ ومن راي نهر النيل نال سلطانا وقوة . in his dream that he drank from its water will obtain gold commensurate with the amount he drank. If someone dreams of the river Nile, he will obtain sovereignty and power).
203 Baukalion is "ein ursprünglich besonders in Alexandreia übliches gläsernes oder tönernes bauchiges Gefäß mit engem Hals, das beim Füllen und Ausgielten einen glucksenden Ton gab. In byzantinischer Zeit bezeichnet es auch einen Wasserkrug" (Brackertz, Traumbuch des Achmet, n. 371).
200 Latte, review of Achmetis Oneirocriticon, ed. F. Drexl, p. 419; Brackertz, Traumbuch des Achmet, p. 12.
205 See Sezgin, GAS, vol. 4, p. 70; also M. Ullmann, "Kleopatra in einer arabischen alchemistischen Disputation," WZKM 64 (1971), pp. 158-75, where the dialogue of Cleopatra with her students translated by Ullmann (p. 169) reveals that the "water of the Nile" was an allegorical name used in alchemy for the philosopher's stone: "Drauf frug sie: Und warum wird er [=der Stein des Goldes] Wasser des Niles (mã) an-Nil ) genannt? Sie antworteten: Weil der Nil, wenn er schwillt, die Saaten befruchtet. Dann gibt es viele Früchte, und der Segen zieht ein. So auch unser Wasser: Wenn es in den (alchemistischen) Prozeß eintritt, wird es ganz und gar ein Gutes."
200 See Del Corno, Graecorum de re onirocritica, pp. 70 and 151.
Byzantines and the Arabs. Unlike Zoroastrianism, the old pagan religion of ancient Egypt had long been extinct by the tenth century and information on it could only be gleaned from books written in the Hellenistic era and later. <sup>207</sup> The vagueness and rarity of the Egyptian references in the *Oneirocriticon* reflect this ignorance.
The title of pharaoh that is applied to Egypt's ruler in the *Oneirocriticon* was known to Christians through the Old Testament and to Muslims from the Qur'an. The preoccupation of the pharaoh with dream interpretation mentioned in the introductory chapter attributed to Tarphan is consistent not only with the story of the pharaoh' s dreams interpreted by Joseph, which is related both in the Old Testament and in the Qur'an, but also with Arabic lore about pharaonic Egypt. According to the medieval Arabic sources, the kings of Egypt used to dream constantly about imminent events. Their dreams were interpreted by their high priests and were always truthful, because of their lofty and mighty position. <sup>208</sup>
Tarphan, the name of the pharaoh's dream interpreter, is curious. It could stem from the Arabic root *t-r-f(c.J\_;:.) <sup>209</sup>*or *t-r-f(c.J..),:.),210* though no adjective
<sup>207</sup> The most extensive single Greek source on the religion and society of Egypt before Alexander's conquest was probably the second of the nine books of Herodotus's history, but this work was not widely read in Byzantium, and the compiler of the *Oneirocriticon* was definitely unaware of the information it provides. Byzantine chroniclers who included Egyptian history in their treatment of the pre-Christian era drew from the Old Testament, and from Hellenistic and late-antique sources. Their main goal, however, was not to convey ethnographical information, but to record the succession of Egypt's kings. In the Arabic-speaking world, ancient Egypt was treated more often in geographical than in historical works. For a collection of the information on ancient Egyptian religion contained in the writings of Greek and Latin authors, see Th. Hopfner, *Fontes historiae religionis Aegyptiacae.* 5 vols. (Bonn, 1922-25). The information contained in the Arabic sources on ancient Egypt can be found in the introduction to Murta<;la ibn al-Khafif, *Egypte de Murtadi,* ed. Wiet. More recent bibliography in U. Haarmann, *Kitah anwar 'ulwiyy al-ajram Ji a/-kashf 'an asrar a/-ahram ta'lif Ahl Ja'far Muf:iammad h.' Ahd a/-'Azfz al-/jusayni al-Idrisi* (Beirut, 1991 ). See also lbn WaJ:ishiyya, *Ancient Alphabets and Hieroglyphic Characters Explained with an Account on the Egyptian Priests and Their Classes, Initiation and Sacrifices,* trans. Joseph Hammer [von Purgstall] (London, 1806). This work is based mainly on Hellenistic sources translated from Greek; see T. Fahd, "Sur une collection d'alphabets antiques reunis par Ibn WaJ:ishiyya," in *Le dechiffrement des ecritures et des langues,* ed. J. Leclant (Paris, 1973), pp. 105-19. Cf. also U. Sezgin, "Al-Mas'udi, Ibrahim b. Wa~ifshah und das *Kitah* a/-'aja'ih-Aigyptiaka in arabischen Tex ten des 10. Jahrhunderts n. Chr.," *Zeitschrift fur Geschichte der arahisch-islamischen Wissenschaften* 8 (1993), pp. 1-70.
<sup>208</sup> For references to primary sources, see Murta<;la ibn al-Khafif, *Egypte de Murtadi,* ed. Wiet, pp. 35-36 and 67.
<sup>200</sup> The root *t-r-f* in verbal pattern I means "to live in luxury"; in pattern IV "to make someone effeminate" and "to surround with luxury"; the noun *taraf* means "luxury, opulence".
<sup>210</sup> The root *f-r-f* has a variety of meanings in the verbal patterns and nouns that are formed
tarfan or tarfan is recorded in Arabic dictionaries. At any rate, it is not clear why a Greek- or an Arabic-speaking person would consider this name appropriate for an ancient Egyptian dream interpreter.211
The chapter from the Egyptians on faith gives very vague information about their religion (Drexl 11, 10-26). The gods of the Egyptians remain unnamed, as they did in the introductory chapter:
ιδ' Έκ του λόγου των Αίγυπτίων περὶ πίστεως δμοίως
Έάν τις ίδη κατ' όναρ προσκυνούντα έαυτον θεούς ή είδωλα ή ζώα ή δένδρα, ουτος εύρήσει χάριν πρός τον Φαραώ και τους μεγιστάνους αύτου άναλόγως της εύγενείας των είδώλων και ζώων και δένδρων. ἐαν δὲ προσκυνήσει τὸν θρόνον του Φαραώ καθώς και οί αρχοντες αύτου, ούτος αρχων γενήσεται όμοιος τῶν ἀρχόντων αὐτοῦ. ἐὰν δὲ προσκυνήσῃ τῇ ῥάβδῳ τοῦ Φαραώ ἢ βαστάσῃ αύτην ή καθίση ἐπὶ τοῦ ἄρματος αὐτοῦ ἤτοι τοῦ δίφρου αὐτου, ούτος πρώτος σύμβουλος έσται αύτφ. ἐαν ίδη, ὅτι ἐκάθισεν ἐπὶ ὕπα τοῦ Φαραώ ήτοι σελλαρίω, εί μεν έν βουλή του Φαραώ τουτο ἐποίησε, γυναϊκα οἰκείαν δώσει αὐτω, εί δὲ έξω της βουλής αύτου, ούτος έπιβήσεται κατά του κορασίου αύτού και φωραθήσεται.
Εαν ίδη τις, ότι συγκαθεύδει τφ Φαραώ η ό Φαραώ συγκοιμαται αύτφ τρόπω γυναικός, μυστικός έσται το Φαραώ και πρώτος των άποκρύφων μυστηρίων αύτου.
14. From the Account of the Egyptians on Faith
211 Unless the name Tarfan is somehow connected with the name of Tarifa al-Kāhina, one of the most celebrated prophetesses of pre-Islamic Arabia. Fahd, La divination arabe, pp. 142-44, tentatively connects her name with the biblical Hebrew word teraphim which means, among other things, "donneur d'oracle." It is even conceivable that an Arab author might have borrowed the name of the Egyptian dream interpreter from the list of the ancestors of the pharaoh whose dreams were interpreted by Joseph. Tapoav (ترفـــان ) would then represent a misreading of "Tharwān" (شروان). See al-Tabarī, Tarīkh al-rusul wa-al-mulūk, ed. Muhammad Abū Fadl Ibrahīm (Cairo, 1960): فاما غيره فإنها الملك بمصر وفرعونها الريان بن (Another) الوليد بن ٿروان ٻن آراشة بن قاران بن عمرو ٻن عملاق بن لاوذ بن سام بن نوح. account gives the full name of the king and pharaoh of Egypt at that time as al-Rayyan b. al-Walid b. Tharwan b. Arashah b. Qaran b. 'Imlāq b. Lud b. Shem b. Noah), W. M. Brinner, trans., The History of al-Tabari, vol. 2: The Prophets and Patriarchs (Albany, N. Y., 1987), p. 153. In the Arabic spelling the ductus of "Thawran" is almost identical with that of "Tarfan," and it only takes misplacing one dot and mistaking و for ف to read the second name instead of the first.
from it. Verbal pattern I means "to blink, to squint"; IV means "to feature or tell something new or novel or original"; and V "to be on the extreme side". The noun tarf means both "glance" and "eye". In astrology, it is the name of one of the houses of the moon. It can also mean "a generous man" or "a nobleman (with respect to ancestry) up to the most remote forefather." If we interpret the ending -an as that of a dual, tarfan could mean "the two eyes."
If someone dreams that he was worshiping gods or idols or animals or trees, he will find favor before the pharaoh and his nobles in proportion to the nobility of the idols and animals and trees. If he performs the *proskynesis* in front of the throne of the pharaoh just as his magistrates do, he too will become a magistrate like them. If he performs the *proskynesis* or lifts the pharaoh's staff, or if he sits on his carriage, that is his chariot, he will become his first councilor. If he dreams that he sat on a horse that belonged to the pharaoh-that is, a saddle horse-if he did so with the pharaoh's consent, the pharaoh will give him a woman from his family; if without his consent, the dreamer will sleep with a girl that belongs to the pharaoh and will be caught.
If he dreams that he slept with the pharaoh or that he had intercourse with him as if he were a woman, the dreamer will become the pharaoh's private secretary and be the first man privy to his hidden secrets.
The chapter on faith according to Egyptian sources is much shorter than the corresponding chapters according to Indian and Persian sources, and does not include many details about the Egyptian religion, possibly because the author knew little about it. The interpretation of worshiping a tree quoted at the beginning of the passage (Drexl 11, 11-25) can be found in the dreambook of al-Dinawari concerning the Sabians: <sup>212</sup>
J..::-.J ...:.Ll i~I ~~GI lSl.J .:,,\_..\*;;~JI~~.:,,\_.. ~J.J ~ i,?I>-:' l,il..;l.... 0~ JI ~\_;-..'.u J..::-.J \_)I '7'~ JI 0::-::>-:'~l lSl.J <t..:;I~ ':f""L...:. . ..i..:..1 JS
On dreaming of adoring a star or a tree: Whoever dreams that he worships some of the stars is a Sabian, for his persuasion is that of the Sabians. Or he will curry favor with a distinguished man or he will vacillate between the persuasion of each one.
A similar interpretation is repeated in *al-Muntakhab: <sup>213</sup>*
.:,,\_.. rA'J ~L.:.JI 0-=.J ~..1 ~.! ...:.Ll ;;~JI~~ ...:.LS: lSl.J 0Ll ~j\_)I ol..41 01 µJ dJj <sup>~</sup>0:--:>l..;l.... Jl...U \_\_.JW <l.JI ~J 0-=l.ll i\_,\_i.JI ·~.I..:' 0...9~ ~ J..::-.J L.i,.,;. \_)I '7'~ 4-:i-=>-l...:. 01 ~ J..i.:;
And if he dreams that he worships a star or a tree, a man of his religion belongs to the religion of the Sabians, and they are from the people that God Almighty
<sup>212</sup>Al-Dinawari,fa~/ 8, *biib* 71, *Esad Efendi* 1833, fol. 83a.
<sup>213</sup> *Al-Muntakhab,* p. 330.
#### 56 CHAPTER ONE
described as "swaying between this and that" [Qur'an 4: 143]. And it is said this dream signifies that the dreamer will curry favor by serving a man in a lofty position who will disdain his religion.
The similarity between the Greek and Arabic interpretations of worshiping a tree indicates that the image of the ancient Egyptians in the *Oneirocriticon* owes something to the image of the Sabians in medieval Arabic literature. Arabic authors occasionally applied the name *Sabi,* to various groups of polytheists past and present. 214 Sabians are usually said to be star worshipers, while the Egyptians of the *Oneirocriticon* are not. 215 But it seems that several Arabic authors thought of the ancient Egyptians as Sabians,216 and al-Mas'iidi in *Kitab al-tanbih* clearly says so in his enumeration of Sabian groups, in which he includes the Chaldean or Babylonian Sabians (Mandaeans), the Buddhists of China, the ancient Greeks, and the Egyptian Sabians, "the last of whom can be found nowadays in f.Iarran." <sup>217</sup>
The remaining interpretations in the Egyptian chapter on faith, which make up more than half of its total, discuss dreams about the pharaoh. The pharaoh (Ar. *fir' awn* ) of the tale of Moses in the Qur'an is generally presented in Islam as the epitome of evil. The interpretations of dreaming of the pharaoh in Arabic dreambooks are accordingly inspired by the relevant Quranic passages and are totally different from the interpretations quoted in the *Oneirocriticon.<sup>218</sup>*
218 Al-Dinawari.faol"i 8, *bah* 86 *(Esad Efendi* 1833, fol. 66a): lSI.; .:\_,.......J .:...-:..JI J.l-L w>-"~ JS rL....y1 JL..:.. .:\_,.........,. -'""~ JL..:.. •J-'-" <sup>0</sup> <sup>1</sup>*L..S* <Lo~J rL....~ JL..:.. •J-'-" ~ JL..JI .:\_,.........,. L..:."""~ ~J ·~ J~ .....:.Ll L..,,W..JI ~I~~ J~ ....:.I lSI.; .\_:iLl \*~.)JU. *JS* d.J:i.s:J <Lo~J \* ... \*Pl~ uy.:J J~J ~YIJ ~I~ w-'"'.;-iJI d.J.:i ·~ ~J ~J ~J ol-'"'J .\_:iLl 4-:JIJ ~J ~J-" JI o.Jl.,. ~ ~ o.r7~1J oJ-'-"i.SYIJ 4-.:.Ll\_,.ijl dJ-01 ~ lSI.; jLl ..:.ll:...A. ~ .c..>-::'-"" (Every pharaoh is an enemy of religion. If someone dreams of a pharaoh in good condition, this bodes ill for the leader and people of that place. Likewise, ill in the condition of the pharaoh bodes well for the condition of the leader and his people. And likewise for the enemies of man. And if he dreams that he was transformed into one of the pharaohs of the world,
<sup>214</sup> The question of who the Sabians of the Arabic sources were is a complicated one, and need not be addressed here; see D. Chwolsohn, *Die Ssabier und der Ssabismus.* 2 vols. (St. Petersburg, 1856); for criticism of Chwolsohn's views, see J. Hjiirpe, "Analyse critique des traditions arabes sur Jes sabeens J:iarraniens," and *E/2 ,* s.v. "Sabi"' and "Sabi'a."
<sup>215</sup> It is possible that the Egyptians are not described as star worshipers in order to avoid giving them a characteristic already attributed to the Persians.
<sup>216</sup> See Murta<.la ibn al-Khafif, *Egypte de Murtadi,* ed. Wiet, p. 60.
<sup>217</sup> Chwolsohn, *Die Ssabier,* vol. 1, p. 214. For a French translation of the passage in question, see B. Carra de Vaux, trans., *Mm;:oudi. Le livre de l'avertissement et de la revision* (Paris, 1897), p. 221.
But if we realize that the pharaoh is a sovereign and look into the interpretations of the dreams about caliphs and kings given in the Arabic dreambooks, it is possible to find analogous passages. Here are some interpretations of dreams featuring caliphs quoted in the dreambook of Ibn Shahin (nos. 369 and 375):
$$\begin{array}{c} \mathsf{c} \leftarrow \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \\\mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \\\mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \mid \mathsf{c} \sqcap \mathsf{c} \\\end{array}$$
Whoever sees that the caliph dressed him or transported him or gave him a mount or gave him one of the worldly commodities, indeed he will attain sovereignty and might and glory analogous to the gift.
~-"=\_,I o\_r-al ~ ~~ ......:.µ .i..:.-1\_, .\_,.ti~~ U#I\_, *\_,\_A* .....:.I (.SI.; 0-4J -~.;~~\_,I ~I~ 0-4 ~IJ-"I \::-'~..JI L.1 *µ\_,* ~ ~ L..:.~
Whoever sees that he and the caliph are in the same bed, indeed he will become a partner in his command or <the caliph> will entrust <the dreamer> with authority over a place. And it is said either that he will marry a woman from the house of the caliph or that the caliph will give the dreamer a slave girl.
The inclusion of interpretations about the pharaoh in a chapter on faith can partly be explained, perhaps, by the statement of lbn Tahir al-Baghdad! that some among the pagans worship particular individuals, such as Jamshid, Nimriid b. Kana'an and Fir'awn.219 But, most important, it conforms to the arrangement of Islamic dreambooks. There, dreams about caliphs are interpreted in the first, religious, chapters, right after the holy figures, because their title in Arabic literally means that they are the successors of the Prophet as heads of the Muslim community. Moreover, the first four, known as the "orthodox" or "rightly guided" caliphs, are indeed regarded as holy figures and had all been close companions of MuQ.ammad during his prophetic mission. The *proskynesis* of the pharaoh is reminiscent both of Byzantine ceremonial220 and of that
he will obtain power and his pretensions will become known, his religion will decay, and his conduct will concur <with that of> the pharaoh regarding evil and sin and he will forsake <religion> and will die in godlessness .... And if he dreams that any of the dead pharaohs and Persian kings and tyrants live in his country or at a place while he governs it, indeed a tyrant's conduct will appear there). The same interpretations are repeated almost verbatim in al-Nabulusi, vol. 2, p. 121, s.v . ..)~~.
<sup>219</sup>See Ibn Tahir al-Baghdad!, *Moslem Schisms and Sects,* trans. Chambers Seelye and Halkin, vol. 2, p. 345.
<sup>220</sup> See *ODB,* s.v. "Proskynesis."
followed at the caliphal court.221 As for the pharaoh's staff, sceptres were understood to be symbols of imperial power both in Byzantium and in the Muslim lands.222
The following passage of the Oneirocriticon is tantalizing because it is unclear whether it refers to the Egyptians or the Persians (Drex1 51, 19-27):
πδ' Έκ των Περσων και Αίγυπτίων περὶ ἐγκάτων καὶ σπλάγχνων
Εάν τις ΐδη το ήπαρ αυτού ότι διά του άφεδρώνος έξήλθεν, εί μέν έστι μέγιστος και πλούσιος, ό οίκονόμος αύτου άπολείται και πασα ή έπιθυμία αύτου, διότι το ήπαρ διανέμει πάση σαρκιે τὴν τροφήν. ούτως γὰρ ἔκριναν καὶ οἱ νόμοι ήμῶν πάλαι, ήνίκα ό νεανίας έκεινος άρπαξ ώφθη της θυγατρός του άρχιερέως ήμων, ίνα φάγωσι το ήπαρ αύτου οίωνοί, διότι πασων έπιθυμιών αίτιον το ήπάρ έστιν.
84. From the Persians and the Egyptians on Intestines and Internal Organs
If someone dreams that his liver came out through his anus, if he is very powerful and wealthy, his steward and every object of his desire will be destroyed, for the liver distributes nourishment to the entire body. Our laws have also decreed thus in the past, when that young man was discovered to have abducted the daughter of our high priest, that birds of prey should devour his liver, because the liver is responsible for all desires.
The story of the young man who was condemned to have his liver devoured by birds of prey does not appear in either Arabic or Greek sources, though the story of Prometheus, who received the same punishment for having stolen fire from the gods, is frequently referred to in texts that constituted standard Byzantine school reading. 223 The liver ( كيد) was considered to be the seat of the soul and responsible for desires by the Arabs, as well. 24 Whether the Greek compiler invented this story or found it in his Arabic sources is unknown.
All this evidence indicates that, despite the claim made in the first prologue
<sup>221</sup> See D. and J. Sourdel, La civilisation de l'Islam classique (Paris, 1968), p. 379.
<sup>222</sup> See ODB, s.v. "Insignia" and "Scepter"; Sourdel and Sourdel, La civilisation de l'Islam classique, p. 368.
<sup>223</sup> See also Brackertz, Traumbuch des Achmet, p. 250, n. 170.
<sup>224</sup> This belief appears to antedate knowledge of Greek science and philosophy among the Arabs, as is evident from the following historical occurrence connected with the rise of Islam. In the third year of the hirra, the Meccans marched against the Muslims of Medina. The Prophet went out to meet them on Mount Uhud and the Muslim army was beaten. The wives of the Quraysh, who had been brought to the battlefield in order to give courage to the fighters by their presence and their chanting, mutilated the Muslim dead. Hind, the wife of Abu Sufyan, publicly plucked out the liver of the Prophet's uncle Hamzah and tried to eat it. The Prophet subsequently received a revelation according to which mutilation was forbidden to Muslims; see M. Pickthall, The Glorious Koran (Albany, N. Y., 1976), p. 61.
of the Oneirocriticon that it was compiled on the basis of Indian, Persian and Egyptian sources, genuine elements from these sources are few. In addition, the image of Persia and ancient Egypt conveyed in the Oneirocriticon seems to depend on knowledge about these two civilizations current among the Arabs, but not among the Byzantines. So apparently it was the Arabic sources of the Oneirocriticon that contained dream interpretations purportedly offered by Persian and Egyptian interpreters.225 The Greek compiler must have retained these Persian and Egyptian chapters more or less as he found them in the Arabic sources. His own intervention consisted in Christianizing the Muslim chapters and labeling them "Indian," to make them both exotic and acceptable to his Christian readers. It is possible, however, that the Arabic sources included chapters on, or at least mention of, Indian dream interpretation, from which the Greek compiler might have borrowed the name of the Indian dream interpreter, Syrbacham.
#### The Author's Patron
The first chapter of the Oneirocriticon quoted earlier informs us that it was compiled for a despotes (translated as "lord"). The last paragraph of the last chapter in the Oneirocriticon defends the reliability of its interpretations in the presence of the dream interpreter's patrons who, in the final phrase of the work, are said to be emperors (basileis).226
Εν δε ταις μακραις έκβάσεσι τών χρονίων όνειράτων ἐγγράφως ἐσημειούμεθα έκαστον όναρ. και ὅτε ή ἔκβασις ἔλθη,22 ἀνεμιμνήσκομεν τὴν δεσποτείαν ήμῶν ήμεῖς οἱ κριταὶ τῶν ὀνειράτων λέγοντες: "ὅτι τόδε εἶδες καὶ τόδε κρίνεται καὶ τότε απεξέβη.""" ταυτα δὲ εἰς τοὺς ἀντιφιλονεικοῦντας εἰς ἔλεγχον αὐτῶν ἐποιοῦμεν. διὰ τοῦτο πιστευτέον, ὅτι πάντα τὰ ἐνεῖρατα ἀγγελία καὶ προγνώσεις είσίν από θεου ύπέρ τε κακου ύπέρ τε άγαθου έν παντι τώ λαώ. πλέον δε έν τούτω πιστώθητι, ώ αντιλέγων, ἐν τῆ ταχεία καὶ καθημερινῆ κρίσει καὶ ἐκβάσει άπαραλλάκτω. και έκ τούτων διδάσκου τας μακράς και χρονίας έκβάσεις, άς
<sup>225</sup> We have seen that al-Dinawari also claimed to have based his dreambook on a variety of foreign sources, and he was, at least in part, sincere. Ibn Shahin's enumeration of sources (Introduction, p. 8) includes the work of Shaykh Muhammad al-Fir'awnī (Muhammad the Pharaonite), who is not otherwise known. Could it be another dreambook claiming ancient Egyptian sources?
<sup>226</sup> Drexl 241, 15-26.
<sup>227</sup> Paris. gr. 2419, fol. 295: ήλθεν.
<sup>228</sup> Drex1 241, 19: πορεύει; Paris. gr. 2419, fol. 295: αποξεύει [sic]; Bononiensis (Bibl. Univ.) 3632, fol. 444r: απεξέβη; Dighy 103, fol. 127v: pervenit eventus.
#### CHAPTER ONE
έπι των δεσποτών ήμων και βασιλέων σημειούμεθα.
For old dreams that take a long time to come true, we recorded in writing each dream. And whenever the dream came true, we, the interpreters of dreams, reminded our Lord (despotes) saying, "You have dreamt such and such, and it is interpreted as such and such, and it came true (anegesm)." We did this in order to refute our opponents. Therefore we must believe that all dreams are messages and portents from God about both good and evil for all the people. Moreover, you should be persuaded by this, O critic, by the unambiguous quick and every day interpretation and outcome [of dreams]. From those, you should learn about the chronic and long term outcomes, which we record for our Lords (despotai) and emperors.
The word despotes, used twice in the first chapter of the Oneirocriticon to designate the compiler's patron (Drex11, 3 and 1, 10), and repeated in the last chapter of the work, generally means "lord" or "master" and could be applied to anyone in a position of authority, such as princes, bishops, and emperors.209 Beginning in the sixth century it was also the usual form of address for an emperor. 300 How, then, should the word despotes be understood in the context of the Oneirocriticon? In the fourth prologue, Tarphan, the dream interpreter to the pharaoh, king of the Egyptians, also refers to his master as despotes (Drexl 3, 16). This time the term is clearly used in the sense of a reigning monarch.231 The literary sources, 22 as well as the numismatic and sigillographic
222 Cf. especially the preface to the Geoponika (Proem. 11), a 10th-century treatise on agriculture dedicated to the emperor Constantine VII (r. 945-59), which addresses him as despotes, instead of
<sup>229</sup> See Lampe, s.v. "δεσπότης. "
<sup>20</sup> F. Dölger, Byzantinische Diolomatik, 20 Aufsätze zum Urkundenwesen der Byzantiner (Ettal, 1956), p. 131; L. Bréhier, "L'origine des titres impériaux à Byzance," BZ 15 (1906), pp. 161-78, esp. 176.
<sup>231</sup> Internal evidence from the Oneirocriticon, especially when its text is compared with Artemidoros and Arabic dream interpretation, also suggests that it might have been compiled at the request of a royal patron. Both the introduction (Drexl 2, 10-15) and the final chapter (Drexl 240, 7-20) of the Oneirocriticon explain that the significance of a dream depends largely on the sex and social position of the dreamer. Accordingly, a variety of interpretations for the same dream are given, and the possibility that the dreamer is a king is often mentioned (Drexl 10, 26; 33, 27 ff .; 59, 20-24; 76, 6-9; 87, 18-21; 174, 2-5; 175, 3-8; 179, 8 ff.; 207, 20-22; 214, 25-215, 5; 225, 23-27 are some examples). A variety of interpretations for the same dream according to the identity of the dreamer is also offered in Artemidoros and in Arabic dreambooks, but the Oneirocriticon interprets royal dreams more frequently than its ancient Greek and Arabic counterparts do, including the 11th-century dreambook of al-Dinawari, which is dedicated to the reigning caliph al-Oadir bi-'llah. The Oneirocriticon also includes the following instructions on how to address an emperor (Drexl 59, 20-24): τώ γάρ βασιλεί ούκ ένι είπειν "οί συγγενείς σου" ἀλλά "οί δούλοι σου", διότι από θεού το βασιλεύειν αύτφ (For it is not possible to say to a emperor "your relations" but "your servants," for kingship has been bestowed upon him from God).
evidence of the ninth and tenth centuries, also indicate that the title *despotes,* routinely and by itself, designates the reigning emperor in the period during which the *Oneirocriticon* was compiled. <sup>233</sup>
If the *Oneirocriticon* was indeed compiled for an emperor, as both its first and its last chapter suggest, it could be any one of those who reigned between 843 and the eleventh century, since internal evidence and its manuscript tradition suggest that it was written between these two dates. Among them, Leo VI (886-912) seems the likeliest candidate. He was an educated man who was interested in literature as well as the occult sciences and divination, as confirmed by three pieces of evidence. The first is an incident narrated in the tenth-century chronicle known as *Theophanes Continuatus.234* After an eclipse of the moon the emperor summoned Pantaleon, metropolitan of Synnada, to interpret this portent for him. The metropolitan's interpretation was that the eclipse pertained di:; to O£Ut£pov np6cromov (to the second person, i.e., the second most powerful person after the emperor). The eclipse proved to have prefigured the downfall of the minister Samonas, an Arab captive who managed to exercise great influence at the Byzantine court during Leo's reign. Other evidence of Leo's partiality to the occult is the horoscope cast for his newborn son Constantine (the future Constantine VII) which must have been commissioned by him, since it is written in terms that flatter the infant's parents. 235 Finally, dream interpretation was removed from the list of evil practices with legislation that was passed during Leo's reign. 236 Two military manuals of the tenth century
<sup>234</sup>*Theophanes Continuatus,* ed. I. Bekker (Bonn, 1838), p. 376, 8-19.
235 D. Pingree, "The Horoscope of Constantine VII Porphyrogenitus," *DOP* 27 (1973), pp. 219-31.
using another imperial title:. *A/..),,'* Etnuxiotc;, ffi oucmomtE OEmi:ota Kffiv<navii VE (May you prosper, 0 most just Emperor Constantine) in *Geoponica,* ed. H. Beckh (Leipzig, 1895), pp. 2, 28-29.
<sup>233</sup> See Ph. Grierson, *Catalogue of the Byzantine Coins in the Dumbarton Oaks and the Whittemore Collection,* vol. 3 (Washington, D.C., 1973), pp. 176-83. For 10th-century seals where the word 1iEcm6TI)s alone means "emperor," see N. Oikonomides, *A Collection of Dated Byzantine Lead Seals* (Washington, D.C., 1986), p. 65 *(no.* 59), p. 68 (no. 62), and p. 73 (no. 69). See also the ornamental crown dedicated to the church of Saint Sophia by the emperor Leo VI (r. 886-912) inscribed with the legend AEQN i'.ELI10THL *(Leo despotes).* Today this crown belongs to the treasury of the church of San Marco in Venice. At a later date, it was topped with a piece of ornamental rock crystal and a statuette of the Virgin. The ensemble is generally known as the "Grotto"; see M. Carrieri et al., *Le tresor de Saint-Marc de Venise* (Milan, 1984), pp. 117-21.
<sup>236</sup> See G. Calofonos, "Manuel II Palaiologos: Interpreter of Dreams?" *ByzF* 9 (1991 ), p. 450; Dagron, "Rever de Dieu et parler de soi," p. 39, n. 10, with further references to Byzantine legislation concerning dream interpretation; *Syntagma ton theion kai hieron kanonon,* ed. G.
further suggest the importance attached to dream interpretation in imperial circles as a means for boosting the morale of soldiers going into battle, which makes the possibility that the *Oneirocriticon* was compiled for an imperial patron seem even more likely.237
In conclusion, the *Oneirocriticon* was compiled in the tenth century, most probably for an emperor, by an author who opted not to disclose either his own identity or that of his patron. A possible reason for this reticence might have been the official condemnation of various forms of divination, including dream interpretation, by Byzantine law. Though the author based his work on Arabic sources, he had no intention of expressly connecting it with the Arabs, or Islam, but only with Indian (which, in his and his readers' understanding, meant eastern Christian), Persian, and Egyptian sources; its attribution to Achmet resulted from a scribe's misunderstanding of the information given in its chapter 19.
Railes and M. Poties, vol. 1 (Athens, 1852), p. 192.
<sup>237</sup>The earlier manual is Leo Vi's *Constitutiones Tacticae (PG* 107, col. 1061A). The later is the treatise, *On Imperial Expeditions,* published as "Appendix ad librum I" in Constantine VII (attributed to), *De cerimoniis aulae byzantinae,* ed. J. Reiske, vol. 1 (Bonn, 1829), pp. 444-508; new edition in Constantine VII (attributed to), *Three Treatises on Expeditions,* ed. and trans. Haldon, text (C), pp. 94-151. For the importance of the occult sciences, and especially magic, in the tradition of ancient and medieval Greek military manuals, see E. L. Wheeler, "Magic in Late Antique Warfare and Byzantine Military Theory," *Twenty-Second Annual Byzantine Studies Conference. Abstracts of Papers, October 24-27, 1996* (Chapel Hill, N. C., 1996), p. 78.
#### CHAPTER TWO
#### THE LANGUAGE OF THE ONEIROCRITICON
F. Drexl, the editor of the Oneirocriticon, attributed its breaches of the rules of high-style Greek to the scribes who had copied the text. On those grounds he introduced corrections that sometimes violated the manuscript tradition, arguing that such changes were warranted by "the sense of the text" and "the style of the author." At least two of his reviewers must have agreed with him, because they also proposed improvements to the text that disregarded the manuscript tradition.2 At the same time, however, Drexl was not oblivious to the connection of the text's language to Modern Greek, and attempted to localize its composition based on its use of four words: avannov (woolen blanket), ήπατοπνεύμων (= ήπαρ + πνεύμων = liver and lung), λυποπούλι (a kind of bird') and ψιχίον (fingertip), which occur as ανάπλα, σκωτοφλέμονα, κλαψοπουλι and ψίχα in the contemporary dialects of the eastern Aegean (Crete, Mytilene, and Asia Minor). 3
Phaidon Koukoules considered the language of the Oneirocriticon in two publications: one, originally an address to the Academy of Athens delivered in 1922, for which he must have used Rigault's text of 1604; and the other a review of Drexl's critical edition published in 1926. In both publications, Koukoules' overriding concern was to prove the direct dependence of the Oneirocriticon on Artemidoros and to exclude the possibility that it was translated or paraphrased from Arabic or Syriac texts. In his opinion, the
Achmetis Oneirocriticon, ed. Drexl, p. xv.
<sup>2</sup> Ch. Charitonides, review of Achmetis Oneirocriticon, ed. F. Drexl, EEBS 8 (1931), pp. 231-34; A. D. Nock, review of Achmetis Oneirocriticon, ed. by F. Drexl, JHS 47 (1927), pp. 149 ff. According to Nock (p. 149), the Oneirocriticon was edited "in a most satisfactory manner."
<sup>3</sup> For the exact meaning of this word, see Appendix 4, s.v. "λυποπούλι."
<sup>4</sup> In Modern Greek σ(υ)κώτι = ήπαρ and φλεμόνι = πνεμόνι = πνεύμων. Sœ Drexl, Achmetis Oneirocriticon, p. 251, index rerum et verborum, s.v. "ήπατοπνεύμων."
<sup>5</sup> Drexl, Achmetis Oneirocriticon, p. vii; also ibid., pp. 243-65, index rerum et verborum, s.v. "ανάπλιον," "ήπατοπνεύμων," "λυποπούλι," "γιχίον." Koukoules, in his review of Achmetis Oneirocriticon, ed. F. Drexl, Laographia 9 (1926), pp. 288-89, rightly insists that the occurrence in texts of words that survive in dialects today does not necessarily mean that the text was composed where the dialect is spoken, since many of the words that only survive in contemporary dialects were possibly common in the past.
#### 64 CHAPTER TWO
language of the *Oneirocriticon* furnished proof of its exclusively Greek antecedents, since it contains "no forced expressions" and "indicates no dependence on a foreign model."6 His scholarly objective was to prove that the tradition of ancient Greek dream interpretation was preserved in Byzantium and passed down from there to modern Greek dreamlore,7 an objective consistent with the broader intellectual trends of the time, which were concerned with proving the continuity of Greek civilization from antiquity through Byzantium into modern times.
K. Dietrich expressed the opposite opinion in his review of Drexl 's critical edition,8 referring to an earlier suggestion by Bland, the first scholar who systematically studied Islamic dream interpretation, that the *Oneirocriticon* had been compiled "by some Christian, probably of Syria, from various native sources. "9 Dietrich adduced evidence from both the content and the language of the Greek text to support this view, which Bland had not discussed at any length. He singled out the interpretation that wearing sandals made from cowhide means marrying a Byzantine woman ('Pwµaia) and from camelhide means marrying an Arab (£mm Ti yuvi) *EK* 'tOU y£voui:; 'tcOV • Apa~wv ). <sup>10</sup>He found that, at the time of the composition of the work between the ninth and the eleventh century, such cross-cultural marriage prospects were possible only in Syria. The reference in the *Oneirocriticon* to the *semanterion,* a long piece of iron, bronze, or wood struck with a hammer to summon the Christian flock to church, was viewed by Dietrich as an indication of the Syro-Palestinian provenence of the text, since this instrument had been in use in Palestinian monasteries since the seventh century, and is called ~uA.ov 'tOU Kpoucrµawi:;
<sup>6</sup> According to the account of Koukoules' address published in *Athena* 35 (1924), p. 237, "At the beginning, Mr. Phaidon Koukoules examined the opinions expressed until now about the author of the so-called *Oneirocriticon of Achmet* and deduced that he is not an Arab, as is believed, but a Byzantine Greek and a Christian. The speaker arrived at this conclusion ... because of the text's language and expression, which is similar to that of later dreambooks that definitely did not draw their material from it .... " Also Koukoules, review of *Achmetis Oneirocriticon,* p. 286: "It is out of the question that this is a translation of an Arabic or Syriac model, because the overall phrasing of the text, which is fortunately long enough, indicates no forced expressions and no dependance on a foreign model."
<sup>7</sup> Koukoules observed that the traditions on dream interpretation that survive in Mane in the southwest Peloponnese can be found in the works of Hippocrates, Artemidoros, Astrampsychos, Ps.-Achmet, Ps.-Nikephoros, and Ps.-Daniel in *Athena* 35 ( 1924), p. 238; see also Ph. Koukoules, *He neoellenike hermeneia ton oneiron kai he oneirokritike paradosis* (Athens, 1954) [non vidi].
<sup>8</sup> Dietrich, review of *Achmetis Oneirocriticon,* ed. by F. Drexl, *Orientalistische Literaturzeitung* 30: 10 (1927), cols. 881-84.
<sup>9</sup> Bland, "On the Muhammedan Science of Tabfr," p. 171.
<sup>10</sup> Drexl 178, I 6ff.
(sounding board). Dietrich added that cotton and sugar, both mentioned in the text, were brought to the Mediterranean by the Arabs, 11 thus implying that the Greek text must have been written at a place subject to Arab political and cultural influence. Dietrich also pointed out that the word KOXAa *(kochla* = kohl) that occurs in the Greek text is closer to the pronounciation of its Syriac equivalent *(kuf:zla),* than to its Arabic one ( *kuf:zl). <sup>12</sup>*It should therefore be regarded as a loan word from Syriac rather than from Arabic and as proof of the Graeco-Syrian provenence of the *Oneirocriticon.* 13 Finally, Dietrich identified four loan words from Arabic <sup>14</sup>and concluded that the author of the
<sup>12</sup>Drexl 33, 17 ff.: £av 1u; 101), on doo:; 16 A.qoµ£VOV *K(JxA.a* £µ~ciM£l mi:i; c'xi>0aA. µoti; aurnu rrpoi; 16 <\JaV£tcr0m <!Jroi; EV 6\$0aA.µoti;, ourni; K£v6ool;oi; E<J'Cat (If someone dreams that he put on his eyes the substance called *kochla* so that more light would appear in his eyes, he will be vainglorious). Recipes containing antimony for the cure of eye diseases were well known in Graeco-Roman medicine. The use of antimony for eye diseases continued among the Arabs and in fact was so widespread that it was used even by quack doctors. See Bosworth, *Medieval Islamic Underworld,* vol. 1, p. 146. For *kuhl* in the sense of "eye medication" in general, see al-Kindi, *The Medical Formulary or Aqriihiidhin of al-Kindl,* ed. M. Levey (Madison, Milwaukee, and London, 1966), p. 181, fol. 129a; p. 175, fol. 127b.
13 Dietrich, review of *Achmetis Oneirocriticon,* col. 884: "Dieses Arabische *ku!Jl* lautete nun im Syrischen *ku!Jlii,* kommt also der Form *KoxA.a* des griechischen Textes ganz nahe, so daB wir damit ein wichtiges Kriterium ftir den syrisch-griechischen Ursprung des griechischen Textes gewinnen."
<sup>14</sup><;ouA.cirrwv = olvoi; EK craqcip£wi; (Drexl 150, 22) from *ju/ah, julliih* Uulep); 1;ourra = sagum (Drexl 177, I) *fromjuhhah* (a long outer garment, open in front, with wide sleeves); <!Japai; = arabisches RoB (Drexl 110, 24; 111, 26; 181, 6 ff.; 182, 5; 182, 9; 183, 7) from *faras* (horse, mare); xcicrowv (noun) = Seidengewebe (Drexl 175, 16; 180, 11) from *khazz* (silk, silk fabric). The word xacrotav (noun: Drexl 175, 16; 180, 11) or xacrotai; (adjective: Drexl 115, 3; 170, 13; 177, 8; 204, 15), which is attested in several Byzantine sources from the 10th century onward, means both "silk" and "felt," reflecting the variety of meanings possible for the Arabic word *khazz,* from which xcicrowv originated. Lane, *Arabic-English Lexicon,* gives the following meaning:"j...:;. A certain kind of cloth ... , well known ... , woven of wool and silk: and also a kind of cloth entirely of silk; ... or it is the name of a certain beast [thought by Golius to be the beaver]: and afterwards applied to the cloth made of its fur .... " This double meaning seems to have been the source of some confusion to modern scholars of Byzantine Greek. Drexl' s interpretation of the word (in *Achmetis Oneirocriticon,* p. 264, index rerum et verborum) is: "xcicrowv (\*xcicrornv, \*xcicroov?) = textum sericum." However, the text of the *Oneirocriticon* twice indicates that the meaning of the word cannot be "silk": 1a OE arro E:piou Kcicrrnpoi; 11 1cOV aA.A.wv ljrnt xacrota A.wpw1a ... (170, 12-13); £i OE lO\] ni; 10 WlOUWV [ Ka~ciot]. on xacrOtOV ~v. £Uprjcr£t JtAOUWV £1; avoprov JtOVT)pcOV Ota 16 £1; E:piou KUVOJto1ciµou dvm au16 (115, 2-4). Brackertz (p. 119) translates this passage, "1st dieses ein Chasdion, wird er von Schurken viel Geld bekommen, weil das Kleid aus der Wolle des Bibers gemacht ist," and explains Chasdion as "ein arabisches oder persisches Wort. Darunter ist ein kostbarer Stoff (sowie das daraus gefertigte Kleid) zu verstehen, der aus der Wolle des Bibers, d. h. wohl der Unterwolle, oder der des Wassermarders, gewebt ist. Reiske ilbersetzt in senem Kommentar zu *'de cerimoniis'* Bd. II, S. 712 das Wort mit 'sammet' oder 'Pilsch'. Das *Chasdion* wurde in der Regel nur von adligen oder reichen Leuten
<sup>11</sup>Bciµ~al; occurs in several instances, among them Drexl 154, 23; 155, 21 ff.; 171, 19; 172, 4; 175, 23. Icixap occurs in Drexl 150, 21; 152, 19; 197, 7; 206, 8.
*Oneirocriticon* was a Syrian Greek and that his text was an adaptation based on an Arabic original. <sup>15</sup>
Though the second of these conclusions is valid, the evidence collected by Dietrich does not help us identify the place where the *Oneirocriticon* was written, or the language spoken there. Regarding sandals from various hides as representing women belonging to different ethnicities, the Greek text does not reflect the reality of any particular place; it faithfully reflects its Arabic sources. In the Arabic dreambooks I have examined, the dream of possessing sandals made of cow hide is consistently interpreted as meaning a man will marry a non-Arab woman(~ I .:,,....a :; I j-0 I), which, for a Muslim man, was perfectly possible not only on the frontier, but throughout the Muslim world. <sup>16</sup> Sandals of camelhide as representing an Arab woman is equally common. <sup>17</sup> As for the reference to the *semanterion* in the *Oneirocriticon,* it does not necessarily indicate that the Greek text was written in Syria-Palestine, since sounding boards were used throughout the Byzantine Empire, as well as in lands where the Byzantine cultural influence was felt. They are known to have been in use in Kiev an monasteries in 1091, and eleventh- and twelfth-century monastic *typika* record the use of sounding boards in monasteries in or near Constantinople, such as the Kecharitomene convent and the monastery of Euergetes. Church bells, introduced into Byzantium in the ninth century, did not eliminate the use of *semanteria. <sup>18</sup>*
<sup>16</sup>For the image of Byzantine women in Muslim sources, see N. M. el Cheikh Saliba, "Byzantium Viewed by the Arabs," Ph.D. diss., Harvard University, 1993, pp. 71-78.
17 Al-Dinawari's interpretation of the sandals (J\_....:,) *(Esad Efendi* 1833, fol. 250b; repeated verbatim in *al-Muntakhah,* p. 145, and al-Nabulusi, vol. 2, p. 307) is the following: ..J-..o U..:.Ll: .:, I ~I ..J-..o ~ .ri-:J I J~ (If [the sandal] is made of cowhide, she will be a non-Arab). In Ibn Shahin (no. 4641) this interpretation is given for a slipper: .i.4- ..J-..o oj\_,\_.->j..11 ..:...:.Ls: lj) *µ\_,* .:,I L....\_,\_......:i.. ~~ ;;l.>-"l l+iµ j..o-11 \_,I ~I .i.4- ..J-..o ..:...:.LS .:,I\_, ~1 ol\_r..ol !+i\--0 \_µ1 . J~ I .i.4- *j-o* 4J.....:i .:,L.S (It is said that if the slipper is made of cowhide, indeed the woman will be a non-Arab; and if it is of sheep- or goathide she will be an Arab woman, especially if her sandal is from camelhide).
18 See *ODB,* s.v. "Bell," "Bell Tower," "Semantron."
getragen" (n. 299). Oberhelman translated it otherwise (p. 159): "If someone dreams that his caftan was made of cotton cloth, he will receive wealth from wicked men: for it is made from the wool of the *kynopotamos."* He reasons in nn. 429-30: "Sophocles translates *chasdion* as 'silk cloth.' But I take it in the modem Greek *(chases)* sense of 'cotton cloth'; this is supported by *erion* ('wool') later in the sentence"; *"Kynopotamos* is not known; Drexl equates the word with the Latin *fiber* ('beaver'), but *erion* casts doubt on this, unless *erion* is to be rendered as 'fleece' or 'skin'. Also, Achmet uses *kastor* for beaver."
<sup>15</sup>Dietrich, review of *Achmetis Oneirocriticon,* col. 883.
Cotton and sugar are also mentioned in Greek texts of the first Christian centuries and were therefore known to the Mediterranean world long before the Islamic conquests, which took place some two and a half centuries before the *Oneirocriticon* was written. 19 The word *kochla* is not a transliteration of an Arabic or Syriac equivalent, but belonged to the medieval Greek vocabulary: it occurs in yet another tenth-century text, *Basilica* 2.5.25 of Leo VI. <sup>20</sup>It is referred to in the *Oneirocriticon* as "the so-called *kochla."* Accompanying a word with the expression "so-called ... " seems to have been the author's standard way of introducing vernacular words regarded as inappropriate for a text with literary aspirations. 21 It is a way of apologizing for their use in the text and of informing his readers that he is perfectly aware of a word's lack of literary status but uses it for the sake of accuracy and clarity.22 The Arabic loan words that appear in the *Oneirocriticon* are much more numerous than the four identified by Dietrich. The question is whether they were coined by the Greek author of the *Oneirocriticon* or were already part of the vernacular language.
Brackertz found that, in choice of topic as well as in language, the
<sup>19</sup>For citations of sugar and cotton in Greek authors of the 1st-3rd centuries, see Liddell-Scott, s.v. "miqap" and s.v. "13aµl}aKOEtlii]~." The Arabic word for cotton is *qu{n,* which is very different from the Greek *bambax.* For a dismissal of the legend that the Arabs brought sugarcane to the lands they overran in the 7th and 8th centuries and for the appearance of Indian cotton in Asia and the Middle East in the early centuries of the Christian era, see A. Watson, *Agricultural Innovation in the Early Islamic World* (Cambridge,1983), pp. 26 and 34; see also idem, "The Imperfect Transmission of Arab Agriculture into Christian Europe," *Kommunikation zwischen Orient und Okzident. Al/tag und Sachkultur* (Vienna, 1994), pp. 199-212, which considers agricultural developments exclusively in Western Europe. For sugar, see also J. Stannard, "Byzantine Botanical Lexicography," *Episteme* 5 (1971), p. 175, and E. 0. von Lippmann, *Geschichte des Zuckers* (Leipzig, 1890).
<sup>20</sup> The word Ti KoxA.a is not found in any Greek dictionary; the entry in Sophocles is K6XA.o~. -0\J, 6, ii. The word occurs in *Basilica* 2. 5. 25 in the genitive: tfi~ e~ eyKauatfi~ ECJKE\Jaaµevri~ KoxA.ou, which, in the nominative, could be both ii KoxA.a (vernacular) and ii KOXAo~ (learned); see the remarks by Drexl, "K6XA.a bei Achmet," *Philologische Wochenschrift* 46:8/9 (1926), 240. The word also occurs in the 11th-century Greek translation of the work by al-Razi, *De pestilentia* (1tEpt A.otµtKii~). published in *Alexandri Tralliani Medici Lib. XII, Rhazae de pestilentia libel/us ex Syrorum lingua in Graecam translatus, Jacobi Goupyli in eosdem castigationes* (Paris, 1548), p. 255, I. 2; p. 258, I. 13.
<sup>21</sup> He writes Eloo~ 'to A.ey6µevov KoxA.a. Cf. ~uytov 'to A.ey6µevov Kaµitavov (12, 16), 6µoiro~ Kat eitt *wii* A.eyoµevou Kaµitavoii (13,1), 'ta A.ey6µeva µavtKEA.ta (114, 6), 'tO A.Ey6µevov CJKapaµ<iyytov (114, 26), 'ta A.Ey6µeva oimia (170, 14), 'to A.Ey6µevov µa~tA.A.<iptov (173, 18), Kat ii p~ T\wt ii A.Eyoµevri piirya (231, 3).
<sup>22</sup> The author of the *Oneirocriticon* also informs the readers that he is not ignorant of the learned language by using both the vernacular and the learned word for several items: eitt 'tOU apµaw~ auwii Tjwt 'tOU 1ii\$pou auwii ( 11, 17-18), xrniiva Tjwt lirnA.otlia (88, 5), l}ouvou~ Tj'tot wuµl3a~ (98, 14) CJ\JVEXO>CJEV Tjwt ECJJtEtpEV (108, 3), etc.
#### 68 CHAPfERTWO
*Oneirocriticon* belongs well within the tradition of classicizing Byzantine literature, even though it includes forms and expressions from the New Testament and the vernacular language. 23 Drexl regularized much that seemed to him "irregular" in its vocabulary, grammar, and syntax, and this gives us a somewhat inaccurate impression of the text. 24 But the main reason for the conclusions of Koukoules and Brackertz was their conviction that the *Oneirocriticon* derived directly from the Greek text of Artemidoros. Drexl's classicizing choices in establishing a critical text preclude the possibility of any thorough linguistic or stylistic analysis unless a new edition is prepared, but one can make some provisional remarks. 25 In determining a text's level of style one looks for the following characteristics, here summarized by I. Sevcenko:
For [a working Byzantinist] a work in high style is one that uses periodic structure; its vocabulary is recondite, puristic and contains *hapax legomena* made up on a classicistic template; its verbal forms, especially its pluperfects, are for the most part Attic; its Scriptural quotations are rare or indirect and its classical ones, plentiful. In a work of middle style, periods are rarely attempted and fill-words and cliches, more abundant; it requires the use of a patristic lexicon; and its Scriptural quotations are more frequent than its classical ones. A work in low style uses largely paratactic structures; its vocabulary contains a fair number of words unattested in standard dictionaries or corning from languages other than Greek; its verbal forms are not Attic; its Scriptural quotations, more frequently than not, come from the New Testament and Psalter. <sup>26</sup>
The most complicated structures are attempted, predictably, in the introductory chapter (Drexl 1, 1-14), but throughout the *Oneirocriticon* structures are paratactic, reflecting not only the linguistic ability of its author in Greek, but possibly also the structure of speech in Middle Arabic, the language in which his sources were written. 27
As for vocabulary, in discussing several hundred dream symbols, the *Oneirocriticon* mentions numerous objects that surrounded the dreamer in his
<sup>23</sup> Brackertz, *Traumbuch des Achmet,* p. 18.
<sup>24</sup> Nock, review of *Achmetis Oneirocriticon,* p. 150, for example, found that the author of the *Oneirocriticon* "writes a literary Greek with some interesting loan words."
<sup>25</sup> The following examination is based on the existing critical text and its apparatus. I did not verify the readings under consideration in any additional manuscripts.
<sup>26</sup>I. Sevcenko, "Levels of Style in Byzantine Prose," *JOB* 31 (1981 ), p. 291.
<sup>27</sup> Classical Arabic has a very finely tuned system of hypotaxis. However, the closer to middle Arabic one comes, the more paratactic the language. The linguistic level of the Arabic dreambooks I have examined is generally closer to middle than to classical Arabic.
everyday life that are rarely, if at all, mentioned in other sources of the period. This makes the text a mine of linguistic information especially pertinent to material culture. The otherwise unattested words are not necessarily the invention of the author, however; their rarity, or uniqueness, is due rather to the nature of the surviving sources.
Drexl appended to his edition a list of 136 *hapax legomena* that occur in the text and the critical apparatus. 28 The dictionaries of Demetrakos and Kriaras that have appeared since indicate that 46 of those words are also found in other texts. <sup>29</sup>The recent *Lexikon zur byzantinischen Griizitiit, besonders des 9.-12. Jahrhunderts,* adds 19 more words to the list. 30 This means that only 71 words from Drexl's list are otherwise unattested. To these one should add at least 2 more words not listed by Drexl, and not in any dictionary, 31 bringing the total to 73 *hapax* words. 32 This count is not definitive: the dictionaries of Kriaras and Trapp are still in progress, and a number of Byzantine texts remain unpublished and consequently unavailable to lexicographers. But whatever the number of *hapax* words, their character is immediately recognizable. Most of them belong to the vernacular, referring to everyday objects that must have been part of standard contemporary parlance, not words invented by the author.
In addition to KoxA-a, sOU7ta, souA.amov' <j>apai; and xacrblOV singled out by Dietrich, three more loan words from the Arabic occur in the *Oneirocriticon,33*
Jo E. Trapp et al., *Lexikon zur hyzantinischen Griizitiit, hesonders des 9.-12. Jahrhunderts.* I. -2. Fasc. (a-oucrauxEvoc;) (Vienna, 1994-). The lemmata of words beginning with T\ were published in *JOB* 35 (1985), pp. 149-170. I would like to thank Professor Trapp for kindly making available to me the unpublished lemma ta containing references to the *Oneirocriticon* ( OOlpEacrnKffic; and ff. in Appendix 4).
Jt oucriov (170, 14) = *vestis talaris* (according to Drexl's index rerum et verborum. See, however, below, n. 34); rtotvaA.tcrTI']c; (129, 18) = *punitor.*
<sup>J</sup>2 Appendix 4 includes Drexl's list of *hapax /egomena,* signaling the words that occur in additional texts according to the dictionaries of Demetrakos, Kriaras and Trapp.
JJ Four, if we include the word *charzanion* ( 171, 9; 205, 15), the etymology of which is problematic. According to Sophocles, s.v. "xapt;civwv," it means either "strap," (as in the *Oneirocriticon,* Drexl 171, 8-11 and 205, 15-17), or "a kind of ornament." *Charzanion* also occurs in Constantine VII (attributed to), *De cerimoniis aulae hyzantinae,* ed. J. Reiske, 2 vols. (Bonn, 1829-30), vol. 1, pp. 623, 12 and 624, 5-7. In his commentary on the text (vol. 2, p. 733), Reiske concludes that *charzanion* must denote some kind of female head ornament, and connects the Greek word with the Arabic *hirz* (amulet) or *kharaz* (something hung around the neck). In my
<sup>28</sup> See Appendix 4.
<sup>29</sup> D. Demetrakos, *Mega lexikon tes hellenikes glosses* (Athens, 1936-50). This dictionary includes words from ancient, medieval and modern texts. More specialized is the still incomplete E. Kriaras, *Lexiko tes mesaionikes hellenikes demodous grammateias, I 100-1669,* 14 vols. (Thessaloniki, 1968-). Though the *Oneirocriticon* lies outside the time span covered by this dictionary, it is occasionally referred to in its entries.
two of them in a part of the text attested only in *Vat. gr.* 573: ouofov (Drexl 170, 14 ff.); 34 avairnpa.; *(Vat. gr.* 573, fol. 201 v) 35 and saµcipa *(Vat. gr.* 573, fol. 201 v).36 Eight others (six, if we count only once the verbs and nouns that stem from the same foreign root) might have entered the medieval Greek vocabulary through Byzantium's contact with the Arabs, but are phonetically closer to the equivalent Persian word, suggesting that they should be considered loan words from Persian, not from Arabic. They are: smptKism *(zatrikizo;l92,* 28); smptKlOV *(zatrikion;* 192, 22); <sup>37</sup>Ka~aOtV *(kabadin;* 88, 5; 114, 26; 218, 12);38 rmA.ouMKtv *(paloudakin;* 198, 4); 39 'tsuKavism *(tzykanizo;* 112, 21
34 Drexl, *Achmetis Oneirocriticon,* p. 258, index rerum et verborum, gives the meaning of this word as *vestis talaris;* Dietrich, review of *Achmetis Oneirocriticon,* col. 884, wonders about its etymology. I believe that either it is a loan word from the Arabic *washy* or, vice versa, the Arabic is a loan word from the Greek, but in any case the meaning of *housion* in Greek and *washy* in Arabic is the same, viz. "many-colored ornamentation, embroidery; embroidered or painted fabric." This meaning is also supported by the context of the word (Drexl 170, 15-17): £&.v 'ili'\1 *w;* on qiopEl oucriov, EUpi\crEt 1tAOU'tOV 1tOAUCJUAAEK'tOV Ota TO 'tOU xurovrn; EK PEAOVT\<; 1tOAUCJUVaK'tOV £pyov (If someone dreams that he was wearing an *housion* he will find wealth gathered in abundance, because of the abundant needlework required for this <kind of> tunic).
35 The full text reads as follows: d 8E £cm v civaKapcii;, civaµ<1>tP6A.oii;, di; 8civmov Kpi vEmt · oµoiroi; Kat Et<; -ri)v 1t0AlV Eii; +\v nai.~oucrtv civaKapciOE<;, ElavmtKOV µEAAEl YEVECJElat, Kat clKOUCJEJi\crEmt· *i\* {)€ 1;aµcipa Kat "ta aAf..a, El<; uppEt<; Kat cittµia<; KplVOV'rat (If it is an *anakaras,* it is undoubtedly interpreted as death. Likewise, in the city where *anakarades* are played, there will be a plague, and it is going to become known. As for *zamara* and the rest, they are interpreted as insults and dishonesty). (From the chapter EK trov CTEpcrrov 1tEpt cicrµci-roiv, opxriµcii:oiv, auA.i\crEro<; Kat Kt8cipai;, fols. 20lr ff.= Drexl 207, 14-26). *Anakaras* is the Arabic word *naqqara* (a small drum having a hemispheric body of copper or wood). Kriaras, *Lexiko tes mesaionikes hellenikes demodous grammateias,* cannot decide whether the word entered the Greek language from Arabic, medieval Latin *(nacara)* or Italian *(nacchera). Anakaras* occurs in many 14th-century texts. Its attestation in the 10th-century *Oneirocriticon* is the earliest that I know of and implies that a medieval Latin and Italian provenance of the word should be ruled out. Supporting an Arabic provenance is the prosthetic ci- at its beginning, corresponding to the Arabic elision of the article together with the noun: *an-naqqiira.*
36 From the Arabic *zammiira* or *zummiira* (a woodwind instrument consisting of two pipes). In Kriaras, *Lexiko tes mesaionikes hellenikes demodous grammateias,* vol. 10, appendix *prosthekes kai beltioseis,* p. \*81, the word is explained as "dooi; µaKptai; <1>A.oy£pai;" and its etymology is given as coming from the Albanian *zamare* or Vlach *dzamara.* Again, its attestation as early as the 10th century and its phonetical identity with the equivalent Arabic word prove that the Albanian and Vlach etymology should be excluded.
37 The Arabic equivalent is *sha.tranj* (chess), an originally Indian, then Persian, and finally Arabic word.
38 The Arabic equivalent is *qaba'* (an outer garment with full-length sleeves); see *De cerimoniis,* ed. Reiske, vol. 2, p. 880. Various etymologies have been suggested for the word, including
opinion, *charzanion* means "strap" in every text, including *De cerimoniis.* Its etymology is possibly Armenian, because John the Grammarian, patriarch of Constantinople (837-43), belonged to the family of Morocharzanioi, who were probably of Armenian origin; cf. Ioannes Skylitzes, *Synopsis historiarum,* ed. H. Thurn (Berlin-New York, 1973), p. 84, I. 93; also Du Cange, s.v. "xap1;civ10v."
ff.); 't/;UKclVlOV *(tzykanion;* 112, 20 apparatus); 40 OOUµclKlV *(doumakin;* 196, 6; 227, 10).41 The question is whether these loan words were borrowed to render Arabic words that did not exist in Greek, e.g., words for objects that were not known in the Byzantine world, or whether they were already familiar to Byzantine readers and referred to objects that were part of their surroundings.
Of the seven loan words from Arabic and eight from Persian that occur in the *Oneirocriticon,* only two, *housion* (a kind of garment) and *paloudakin* (a kind of sweet) are otherwise unattested. This indicates that at least fourteen out of the fifteen were well integrated into the Byzantine vocabulary. The context of the remaining two in the *Oneirocriticon* indicates that they were also known to its readers-both words are introduced with 'tO A.eyoµevov or 'ta A.eyoµeva ("the so-called"). Since the passages on the interpretation of both *housion* and *paloudakin* do not include descriptions of the objects designated by these two words, one can conclude that the reader was assumed to be familiar with them, since their appearance and properties justify the interpretation: ti 0£ 'tpffiytt yA.uKtcrµa crapaKT]VtKov 'tO A.eyoµevov 1taA.ouoaK1v, eupiicrtt vocrov Ota 'ta xpffiµma Kat 'tO *nw* avaA6yco~ ni~ ~pfficreco~ (If he eats a Saracen sweet, the *so-calledpaloudakin,* he will find sickness
40 In Persian *chawgiin* <.:i~~) and Arabic ~awlajiin ( .:i~\_,\_..,) =polo mallet. The Greek *tzykanion* is phonetically closer to.the Persian *chawgiin.*
Persian and Slavic. For a full discussion, see P. B. Golden, "The Byzantine Greek Elements in the Rasiilid Hexaglot," *Archivum Eurasiae Medii Aevi* 5 (1985-87), pp. 95-96.
<sup>39</sup> From Middle Persian *piiludag* (~J\_,Jt.,,.), from which both the Arabic *fii/Udhaj* and medieval Greek *paloudakin* derive. The Middle Persian *pii/Udag* is actually a passive participle of *piiludan* (to strain, filter, purify). D. N. Mackenzie, *A Concise Pah/evi Dictionary* (London, 1971), p. 64, gives its meaning as "starch jelly, flummery," while Latte, review of *Achmetis Oneirocriticon,* p. 420, n. I, gave the meaning of this word as "siiBe Speise aus feinem Mehl, Honig und verschiedenen Gewiirzen." The New Persian form is *piiludah* (•Jyt.,,.); both New Persian and Arabic dictionaries interpret *pii/Udah* and *fiiludaj* as a "sweet made with.honey"; it is a gelatinous sweet, similar to the Turkish delight more widely known in the Western world. Though the Greek form *paloudakin* only occurs in the *Oneirocriticon,* it is possible that the sweet meant is the same as the *palodaton* mentioned in the 12th-century vernacular poems by Ptochoprodromos: rpavii.a, craxapiini TE Kat TO TE nal,,w0iiwv (Koraes, *Atakta,* 5 vols. [Paris, 1828-35], vol. I, p. 283; see also Koukoules, *Byzantinon bias kai politismos,* 6 vols. [Athens, 1948-57], vol. 5, p. 120). The more recent edition of Ptochoprodromos by D. C. Hesseling and H. Pernot, "Poemes prodromiques en grec vulgaire," *Verhandelingen der Koninklijke Nederlanse Akademie van Wettenschappen, Affd. Letterkunde,* n.s. I !:I (1910), p. 60, verse III, 283b, omits this verse from the main text and includes it only in the critical apparatus.
<sup>41</sup>The origin of the word (which means "tail") is Iranian, and it also occurs in Georgian and Armenian (though not in Arabic). It is unclear which language provided it as a loan word to Byzantine Greek. For a full discussion, see Golden, "Byzantine Greek Elements in the Rasiilid Hexaglot," p. 83.
#### 72 CHAPTER TWO
commensurate with what he ate, because of the colors and the fire). 42 The color usually interpreted as sickness in the *Oneirocriticon* is yellow.43 This interpretation indicates that *paloudakin* is yellow and its preparation requires the use of fire; but the elliptical reference to these characteristics implies that the readers of the *Oneirocriticon* are already familiar with them. The reference to *housion* in the text leads to a similar conclusion: Kat 'tU AtyOµ£Va OUOfo £ts clVol't£pov 1tclV'tCOV 'tcOV £tpT]µ£vrov 1tAOU'tOV Kpi V£'tat. £av l01J 'tls, on <j>opd oucriov' eupi]crn 1tAOU'tOV 1tOA '\.l<JUAA£K'tOV 8ta 'tO 'tOU Xl 'tcOVOs EK ~£AOVT]s 1tOA'\.lcrUvaK'tOV epyov (The so-called *housia* are interpreted as wealth greater than in any of the aforementioned items. If someone dreams that he is wearing an *housion,* he will find wealth accumulated in abundance because of the abundant needlework required for this <kind of> tunic).44 The *Oneirocriticon* further interprets *housia* decorated with red dots, or colored in gold, blue and yellow. It is improbable that such a long passage (Drexl 170, 14-25) would be dedicated to an object unknown to both author and reader, especially when one considers that in other instances the author did omit interpretations found in his sources he thought too specifically Muslim. For example, the Arabic dreambooks begin with a chapter on the interpretation of godhead. The *Oneirocriticon* does not include such a chapter, probably because it would have been too complicated to disguise as Christian the Muslim interpretation of godhead and its properties. It is possible that he omitted the interpretation of objects that were known only in Muslim lands, but neither *paloudakin* nor *housion* was among them.
The Arabic sources of the *Oneirocriticon* obviously did not cause the
<sup>42</sup> Drexl 198, 3-5. The interpretation of *faludhaj* in Arabic dreambooks is similar to that of *paloudakin;* cf. al-Nabulusi, vol. 2, p. 165, s.v. c;<iyLl : ~ c;<i\_,JW I JJ~ ~ ~ *'J* <sup>J</sup> .~Li.JI..\_;,->--"~ JJ ~..! *<....:.'J* ~ j..,;.J JI il..:..11 (There is nothing good for whoever eats a *faludhaj* in his dream or acquires it. For this probably indicates the sickness of hemiplegia *[falij]).* Other Arabic dreambooks state that the *faludhaj* has the same interpretation as another sweet, the khabl~. The fresh dates used in the preparation of the khahi~ are interpreted as follows *(al-Muntakhab,* p. 132; repeated almost verbatim in Ibn Shahin, no. 4383): ufa "'--::'-' ~ ....:..... ..,..,\_1~ IJ . .\_;,\_,..11 ~ J~ <..:ii ~~J o\_r-L.:JI.:.,..... ~ U ~(The fresh dates [of the khabl~] are interpreted differently. Some of the [dream interpreters] dislike them because they are yellow, and it is said that this is interpreted as sickness). The interpretation of *faliidaj* as sickness is understandable even without the reasoning adduced in Arabic dreambooks, since in Arabic medicine it was often used in perscriptions for ulcers and coughs; see al-Kindi, *Medical Formulary,* ed. Levey, p. 311, no. 219.
<sup>43</sup> Drexl 115, 8-9; 116, 23-24; 117, 26, etc.
<sup>44</sup> Drexl 170, 14-17.
introduction of any neologisms into the Greek vocabulary. It is true that some of the Oriental loan words *(zatrikion, anakaras, zamara)* are possibly attested in Greek for the first time in the *Oneirocriticon,* while others occur in tenthcentury texts such as *De cerimoniis.* Since our loan words from the Arabic belong to the vernacular, the problem of when they entered the Greek vocabulary is complicated by the nature of the sources. The vernacular was not written until the 12th century. Traces of it found in Byzantine texts before that date are either efforts to clarify a passage (especially in texts aiming at offering practical advice, such as military and technical treatises) or slips of the tongue. But the Arabic loan words referring to everyday objects were not introduced in the Greek language through literature but through contacts on a subliterary level and were not written unless they were well established in oral communication.45 Therefore, the author of the *Oneirocriticon* did not invent new words, which implies that he did not introduce objects unknown to his Byzantine readers.
The presence of loan words from foreign languages in the *Oneirocriticon* is easily accounted for without taking into consideration the Arabic provenance of the Greek text. The Byzantine vocabulary included many loan words from a variety of languages, many more than philologists and lexicographers will ever be able to count, because, though they were current in oral communication, they were avoided by authors striving for high style. They come from the languages of almost every culture that the Byzantines came in contact with, including Latin. They occur in texts not only when the author did not know any better, but also when he did, but wanted to be clearly and immediately understood. They are found not only in practical manuals on court ceremonial, warfare, medicine and, in our case, dream interpretation, but also in legal texts, such as the tenth-century *Book of the Eparch.46*
A search for learned words in the *Oneirocriticon* yields poor results. In its
<sup>45</sup> For the analogous example of a loan word from Greek into Slavonic, see I. Sevcenko, "To Call a Spade a Spade, or the Etymology of *Rogalije," Harvard Ukrainian Studies* 20 (1996), pp. 607-26.
<sup>46</sup> The *Book of the Eparch* includes regulations on the activities of Constantinopolitan guilds. Its enumeration of Syrian textiles includes loan words from the Arabic *(chareria, sophoria, audia, phouphoulia, thalassai, chamia, bagdadikia* ). A couple of them are otherwise unattested, but there is no reason to doubt that these were common words in the 10th century. See the commentary of J. Nicole, *Le livre du Prefet* (Geneva,1893), p. 29; rpt. in Dujcev, *To eparchikon bib/ion* (London, 1970). The most recent edition of the *Book of the Eparch* by J. Koder, *Das Eparchenbuch Leans des Weisen* (Vienna, 1991), p. 94, §5. 2, does not make any comment on these particular words; see, however, Koder's remarks on the language and style of the *Book of the Eparch,* pp. 58-64.
#### CHAPTER TWO
241 pages there are only two words - κάρα (head)4 and φάσηανον (sword)40that clearly belong to the realm of learned literature and especially poetry. Their presence in the Oneirocriticon, however, should not be surprising, since they both occur in ancient texts that were among the staples of a Byzantine elementary education.40 A third word, ήγεμονικόν (intellect),30 a term from ancient Stoic philosophy, was further developed by Christian theologians, can be found in Byzantine lives of saints, 51 and is also explained in the tenth-century Suda Lexicon.
Frequently, two words are used for the same object. One word is (or attempts to be) learned, the other not. A list of these instances would include: ἐπὶ τοῦ άρματος αύτου ήτοι του δίφρου αύτου (11, 17-8) ἐπλ ὕππω τοῦ Φαραώ ήτοι σελλαρίω (11, 20)· οί βραχίονες και αί κνημαι αύτων ήτοι οί μύες (44, 8) περί λωβών ήτοι κελεφών (65, 7-8) χιτώνα ήτοι διπλοϊδα (88, 5): καβάδι ήτοι σκαραμάγγιν (88, 5-6): πίθηκον ήτοι μιμώ (90, 26): βουνούς ήτοι τούμβας (98, 14): συνέχωσεν ήτοι έσπειρεν (108, 13): πρός τὴν εὐχὴν τῆς αὐγῆς ὅτοι τὸ περιόρθριον (111, 28)· ἐλαύνει τὴν σφαῖραν ήτοι τζυκανίζει (112, 21)· θώρακα ήτοι λωρίκιον (113, 28)· άναβόλιον ήτοι σάβανον (115,5): περί τα κρυπτά κάλυμμα ήτοι περισκέλισμα (115, 17) σποδιάν ήτοι στάκτην (119, 12) το όχημα αύτου ήτοι ό δίφρος
<sup>47</sup> Drexl 234, 2-4: πολλάκις γαρ [ό ταών] και είς βασιλέα μικρὸν κρίνεται διὰ τὸ κόσμιον και τὴν εὐπρέπειαν τοῦ πτερου και τῆς κάρας λόφον (The peacock is frequently also interpreted as a minor king because of the beauty and dignity of its plumage and the tuft on its head).
<sup>48</sup> Drexl 223, 1: εί δε ίδη, ότι φασχάνω έπληξεν αύτην [ = τὴν αρκτον] ... (If he dreams that he wounded the bear with a sword ... ). For the occurrence of phasganon in a Greek-Arabic-Coptic scala, see H. Munier, La scala copte 44 de la Bibliothègue Nationale de Paris. Transcription et vocabulaire, vol. 1 (Cairo, 1930), p. 117, fol. 57v. Φάσγανον appears in the same list as vernacular terms for arms and armor, such as κληβάνιν [sic], λουρίκην [sic], μανικέλια, χαλκοτουβία, κούκουρον.
<sup>49</sup> A rank-and-file Byzantine author would have read two books from the Iliad, Hesiod, some Pindar, three tragedies of Sophocles and three of Euripides, three comedies of Aristophanes, some Demosthenes, some Aelius Aristeides, and the eulogy of St. Basil by Gregory of Nazianzus. See A. Dain, "A propos de l'étude des poètes anciens à Byzance," Studi in onore di Ugo Enrico Paoli (Florence, 1956), pp. 195-201. The word kápa occurs in Sophocles (Antigone, Oedipus Rex, etc), as well as in Homer (in its Ionian form, κάρη). Φάσγανον is used in the fliad (see Liddell-Scott, s.v. " κάρα," "φάσχανον"). Both words are explained in Byzantine dictionaries, such as the Suda (for κάρα it only has κάρη, κάρητος without giving a synonym) and the Etymologicum Magnum.
<sup>50</sup> Drexl 36, 16: διότι ή διάκρισις τῆς εὐοσμίας καὶ τῆς δυσωδίας διὰ τῆς ρινός ἐστι τῶ nyeuovuk@ (Because the intellect can distinguish between good and bad odors through the nose).
<sup>51</sup> For example, in the 10th century life of Basil the Younger; cf. A. N. Veselovskij, "Razyskanija v oblasti russkogo duchovnogo sticha," Shornik Otdelenija russkogo jazyka i slovesnosti Imperatorskoj Akademii nauk 46 (1889-90), p. 70, 1. 7.
αυτού (122, 16): ό ήλιος ήτοι ό δίσκος (127, 26): είδε τους άστέρας έν αληθέσι ήτοι έν τύπω γραμμάτων (132, 9-10) έβράχη ό τόπος έξωθεν του άέρος ήτοι ανευ βροχής δι' έτέρου είδους (133, 29-134, 1)· ώκιμον ήτοι βασιλικόν (158, 14) τα δε άπο έρίου κάστορος ήτοι χάσδια λωρωτά (170, 11-12): βέλη ήτοι σαγίτας (204, 7): θώρακα ήτοι λωρίκιον (204, 21) περι βήλων και καλυμμάτων ήτοι ταπήτων (214, 5-6) ἐπικαλύμμασί τισιν ήγουν διαπετάσμασιν (214, 20) περὶ κλώσεως ήτοι νήσεως ἀτράκτου (215, 20-21) δεσμών σιδηρών ήτοι κουρκούμων (221, 9-10) όφιν μέγιστον ήτοι δράκοντα (228, 11-12): ή ρώξ ήτοι ή λεγομένη ρώγα2 (231, 3).
In only three cases is ήγουν-ήτοι used to introduce explanations: είς τον μέγιστον ναόν είς το Μέκκε, ήγουν έν τη σκηνή του ' Αβραάμ (29, 20) (in the great temple in Mecca, that is, the tent of Abraham); oi ώμοπλάται είς την γνησίαν των γυναικών ανάγονται του άνδρος, ήτοι τὴν μητέρα των τέκνων αυτού (50, 5) (the shoulder blades refer to the legitimate one from among the wives of a man, that is, the mother of his children); icoso ήτοι προσευχίτην (92, 4) (a priest, that is, a person performing prayers). In these cases nyovv- nov explain notions peculiar to the Muslim faith-the Great Mosque of Mecca, polygamy, and the leading of prayer by an imam.
A number of these glosses are words common in biblical or ecclesiastical literature: διπλοΐδα (a kind of cloak; Job 20:14; Psalms 108:29; etc); περιόρθριον(dawn; in the context of the Oneirocriticon, morning prayer); σαβανον (a kind of garment); δίφρος (chariot; Job 29:7; Proverbs 9:14; etc); δίσκος (disk, i.e. the sun disk); διαπετάσμασιν (coverings); δράκοντα (large serpent; Psalms 73:13, 14: 90:13: 103:26; 148:7).54 Others belong to technical language: τζυκανίζει (to play polo); λωρίκιον (cuirass); χάσδια λωρωτά (felt garment with stripes); σαγίτας (arrows); κουρκούμων (muzzles). At least two of them seem to explain too literal, and therefore infelicitous or
<sup>32</sup> Drexl (Achmetis Oneirocriticon, p. 260, index rerum et verborum) gives the meaning of rhox as animal reptile; see, however, Koukoules, review of Achmetis Oneirocriticon, p. 291, stating that rhox is not an animal repitle, but a kind of poisonous spider called rhax by the ancient Greeks and roga in the modern Greek dialects of Bithynia and Saranta Ekklisies in Thrace and roba on Cyprus and Kos.
<sup>53</sup> The Great Mosque in Mecca is called "the tent of Abraham" because, according to the Qur'an and Muslim tradition, Abraham and Ishmael rebuilt the Ka'ba (originally established by Adam) and called mankind to make the pilgrimage to it. A few meters away from the Black Stone that is housed in the Kaba is the Magam Ibrahim ("the standing place of Abraham"), a stone with the indentation of a footprint which, according to tradition, is the footprint of Abraham, impressed in the stone during the rebuilding of the Ka ba.
<sup>54</sup> For references to texts, see Lampe, s.v. "περιόρθριον," "σάβανον," "δίσκος," and "διαπέτασμα. "
incomprehensible, translations from the Arabic. These two instances are Ev at.ri8fo1 and £~w8£v 'tOU a£po<;. The meaning of EV at.ri8fo1 in the phrase cl8£ 'tOU<; am£pa<; EV at.ri8fo1, ll'tOl EV MC\) ypaµµa't(J)V is obscure, and I have been unable to locate a parallel Arabic passage that could clarify it.55 The problematic £~w8£v 'tou a£po<; (outside of the air) in the phrase E~paxri *b* 't07t0<; £~w8£v 'tOU a£po<; ll'tOl avcu ~poxfi<; 81' E't£pou ctboU<; (the place got wet *outside of the air,* that is without rain <but> in some other way) might be the Greek author's too literal rendition of the Arabic • lw I <sup>~</sup>4.>L..:.. *(khiirijan <an al-hawii,)* or .1,W I ~ *.>L;.. (khiirij al-hawii,)* which would mean "beside the climate"; *khiirij* or *khiirijan <an* means "beside", "apart from," as well as "outside of'; *al-hawii,* means both "air" and "climate, or atmosphere." <sup>56</sup>
The synonym offered for "muscles" in the phrase, oi ~paxiov£<; Kai ai Kvf]µm mhffiv ilw1 oi µu£<; (the arms and their *knemai* [= legs, calves of the legs], i.e., muscles)57 is also problematic. The meaning "muscle" for the word *kneme* does not appear in any dictionary of ancient, Byzantine or modern Greek. However, the lemma on *kneme* from the *Etymologicum Magnum* (12th century, but based on earlier compilations) reads as follows: KNHMH: I1apa 'tO Kl vffi Kl vi]crw, Kl vi]µ ri. Kat cruyK07ti}' Kvi]µri, OlOVEl 'ta 'tfj<; Kl VYJO"EW<; ahta, 'ta omcr8£v 'tOU O"KEAOU<;. A£y£'tm 8E Kat ii iyvuri· µuffiv bf:, napa 'tO crapKffi8£<; Kat vrnpffi8£<; ... *(Kneme* ... is called ... "a muscle" because of its muscular and sinewy quality). 58 It seems that the Greek author supplied a
<sup>55</sup> Oberhelman, *Oneirocriticon of Achmet,* p. 171, circumvents the problem by translating "he looked up and saw the stars arranged in the form of letters," which omits "Ev aA.ri0£cn." Brackertz, who generally is a much more careful translator, renders the phrase as "[er] schaue die Sterne in ihrer wahren Gestalt oder in Form von Schriftzeichen." This translation presupposes two emendations to Drexl's text that are not warranted by the manuscript tradition: EV aA.ri0£cn to EV aA.ri0Eic;t and i\rnt (that is) to il (or). The text possibly makes a reference to the Arabic *'ilm a/-l;iuruf* (the science of letters), which is closely connected with astrology, as it is based on arithmomancy, the know ledge of the natural properties of the letters according to alchemy ( *'ilm* a/-khawa~~) and their astrological conjunctions *(qiranat);* see *EI2,* s.v. "J:lurUf." Alternatively, the Greek phrase could have originated in a faulty reading given by the Arabic manuscript that the author of the *Oneirocriticon* had in front of him. The intended Arabic phrase could have been *l..S* I .J J-=\_;-h.J~ ~~ *.i..S* r ~I (he saw the stars as signs [guiding] his way). If the last word, J-=.rb (way, road) is miswritten or misread as J-=~ (truthful), then the Arabic phrase given above could result in a Greek translation such as the one we see in the *Oneirocriticon:* dliE rnuc; acr-r£pac; Ev aA.riefot.
<sup>56</sup> There seems to be a lacuna in the text. In order to make good sense, the text should read: "If the place got wet for a reason besides the climate," which in Arabic would be something like .:, I • I *w* I ( *) . ..:..* ~ ~ I~ *.:,l..S.l* I .:,Ll:. Could it be that the word ~ ~, or an equivalent, like <sup>~</sup> ~I or~· "for a reason," was missing from the Arabic text used by the Greek author?
<sup>57</sup> Drexl 44, 8.
gloss on knēmē that he had found in a dictionary containing an entry similar to that recorded in the Etymologicum Magnum. The author's possible acquaintance with the lemmata of the lexicographical tradition that accumulated in the Etymologicum Magnum might also have caused his use of the word πρόβατον (sheep) in the following phrase: εί δε εύρη τις έριον το άπο κουράς προβάτων, εύρήσει πλούτον πιστον και ίσχυρον άπο μεγίστων άνδρών διά το του προβάτου κέρας (If someone finds wool from the shearing of sheep, he will find reliable and mighty wealth from very great men, because of the sheep's horns).39 Πρόβατα do not have horns. But the Etymologicum Magnum clarifies that point with πολλάκις κατά κοινού έπι πάντων των βοσκημάτων είρηται ή λέξις (The word is often said in general of all kinds of cattle).60
Some examples of the use of o, n, to λεγόμενον (the so-called) before a word to show the author's awareness that it belongs to everyday parlance are: ζύγιον ή το λεγόμενον καμπανον (a balance or the so-called steelyard; Drexl 12, 16); όμοίως και ἐπὶ τοῦ λεγομένου καμπανοῦ (likewise regarding the so-called steelyard; Drexl 13, 1); είδος το λεγόμενον κόχλα (a product called kohl; Drexl 33, 18); ἐνεδύσατο τὸ λεγόμενον κλιβάνιον μονομερές (he wore the so-called klibanion monomeres;61 Drex1 114, 1-2); τα λεγόμενα μανικέλια (the so-called manikelia; 6 Drexl 114, 7); καβάδι το λεγόμενον σκαραμάγγιον (kabadi, the so-called skaramaggion;63 Drex1 114, 26); οίνον άπο σαχάριτος το λεγόμενον ζουλάπιν (wine made of sugar, the so-called julep;64 Drex1 150, 21-22); και τα λεγόμενα ούσία (and the so-called housia;
61 The meaning of klibanion is given by Drexl as vestimenti genus but is in fact a kind of breast plate. See Koukoules, review of Achmetis Oneirocriticon, p. 291: "Το κλιβάνιον το γενικώς ώς vestimenti genus χαρακτηριζόμενον είναι είδος θώρακος. Περὶ τούτου παραβλητέα τὰ ὑπὸ του Du Cange έν λ. κλίβανον και κλιβανοφόρος λεγόμενα ώς και οί των μεταγενεστέρων ποιημάτων χρυσοκλιβανιασμένοι." An exact definition is given in E. McGeer, Sowing the Dragon's Teeth: Byzantine Warfare in the Tenth Century (Washington, D.C., 1995), p. 369: "klivanion: sleeveless, waist-length cuirass, usually of scale armor."
62 The manikelia were arm-guards made of thick cotton or coarse silk that covered the lower arm from the elbow down, as well as the back of the hand. See McGeer, Sowing the Dragon's Teeth, p. 69.
63 A kind of garment; cf. above, n. 38.
64 Julep in English (and the Persian and Arabic terms from which the word originated) can signify either a soft drink such as rosewater, or an alcoholic drink. In the context of the Oneirocriticon the meaning of the word is evidently the latter one. For recipes on how to prepare alcoholic
<sup>58</sup> Etymologicum Magnum, ed. T. Gaisford (Oxford, 1848: rpt. Amsterdam, 1962), s.y. "κνήμη, "
<sup>59</sup> Drexl 172, 3-4.
<sup>60</sup> Such is the use of the word in Homer (also explicated in the old scholia to the Iliad; see Scholia Graeca in Homeri Iliadem (Scholia Vetera), ed. H. Erbse, vol. 4 (Berlin, 1974), p. 538: scholia to Ξ 124), Hesiod, and the Septuagint; see Liddell-Scott, s.v. "πρόβατον."
Drexl 170, 14); 'to A.cyoµEvov µa~tA.A.cipwv (the so-called pillow; Drexl, 173, 18); yA.uKtcrµa crapaKTJVtKov 'to A.cyoµEvov 1taA.ouOciKtv (a Saracen sweet, the so-called *paloudakin;* Drexl 198, 4); Ti A.EyoµEVTJ p{jyya (the so-called *r6ga; <sup>65</sup>*Drexl 231, 5).
The tendency to give both a learned and a vernacular word for the same thing can also be observed in other Byzantine technical manuals of the tenth century, such as the agricultural treatise *Geoponika66* and the manual on war tactics attributed to the emperor NikephorosPhocas.67 A third Byzantine manual of the tenth century, the *Poliorketika,* which discusses siege engines using more elevated language than that of the *Oneirocriticon,* is introduced with a warning to the reader that the style of the text that follows is neither purist nor ornate, because the author deliberately chose to emphasize the meaning rather than the form of his text, following the example of several antique writers.68 Such an introduction serves as both a declaration that the author is capable of writing in a higher style than that of the text at hand and an acknowledgment that the majority of his readers, whom he must accommodate because of the practical nature of his work, would not have been able to understand a more complicated language.
Besides choice of vocabulary and grammatical usage,69 an author's level of sophistication is also revealed in the kind of literature with which he is familiar. The *Oneirocriticon* contains no references to works other than the Bible; even though it is directly quoted only twice and not quite verbatim, biblical vocabulary and forms of expression permeate the entire text. The two direct scriptural quotations are introduced with a phrase signaling that their source is biblical: Ka86 1tOU yeypa1t'tat EV 'tOt~ ayiot~ dmyyEA.tot~. O'tt 1tp0~ 'tOV aya1tCOV'tcl
68 For the text, see R. Schneider, "Griechische Poliorketiker," *Abhandlungen der koniglichen Gesellschaft der Wissenschaften zu Gottingen,* philol.-hist. Klasse, n. F. 11 (1909), pp. 8-JO (200, 14-204, 4). For a new edition of the text based on an older manuscript, see D. Sullivan, *Siegecraft* (Washington, D.C., 2001).
69 For instances of grammatical solecisms in the *Oneirocriticon,* see *Achmetis Oneirocriticon,* ed. Drexl, pp. 265-69, index grammaticus.
*zoulapia,* see *Paris. gr.* 2419.,fol. 154v *(CCAG,* vol. 8:1, p. 47).
<sup>65</sup> A kind of spider; cf. above, n. 52.
<sup>66</sup>*Geoponica,* ed. Beckh, II. 27 (title): nepi crt'to~oA.iov T\'tot ciipdov· IV. 14 (title) cO<J't£ 'tOV aU'tOV ~O'tp"UV exnv Ota\$6pov~ paya~ ('tO"\l'tEcr'tl KOKKO"U~)· IV. 15.8 'tlVE~ OE d~ cripmov, 'tO"U'tEcr'ttv d~ Eljfl]µa· IV. 15.13 E~at0puicra~ 'tOV'tfon Otmjlu~a~, etc.
<sup>67</sup>L'tPO't11YtK1fEK0£crt~ Kai I:uv'ta~t~ NtKT1\$6pov D.ecrit6'to"U (for the text, see McGeer, *Sowing the Dragon's Teeth):* Kai cravOciA.ta ijyovv µov~<iKta, ehe 'ta A£y6µeva Ev Tft <J"UVT10£i~ 't~Ep~ouA.ta (I, 22-23); oipdA.et 'tO <J'toµa ri]~ napm<i~e~, Tjyovv 'tO µ£'troitov (III, 18-19); Kai 6 µEv itpcii'to~ opotvo~, ijyovv 'to cr'toµa ri]~ itapm<i~e~ (III, 60-1 ); etc.
με έγώ και ό πατήρ μου έλευσόμεθα και μονήν παρ' αὐτῷ ποιήσομεν ο ( ... as is written in the Holy Gospels: "To the one who loves Me, My Father and I will come and tarry with him"[John 14:23]); 11 έαν ίδη τις, οτι ό πάλαι νεκρός άνέζησε ..., τουτο είς σώσμα του νεκρού κρίνεται, πληροφορούμενον έκ τῆς θείας γραφής λεγούσης, ὅτι ὁ θεὸς οὐκ ἔστι νεκρών, άλλα ζώντων72 (If someone dreams that a person who is already dead returns to life ... this is to be interpreted as the dead man's salvation, being confirmed by the Holy Writ: "He is not the God of the dead, but of the living" [Matthew 22:32; Mark 12:27; Luke 20:38]).73 Neither of the two passages copies the corresponding quotations verbatim from the New Testament, suggesting that the author most likely quoted both of them from memory.
The echoes of the Old and especially the New Testament in the author's choice of vocabulary and expression are many. Drexl identifies three instances where the phrasing of the New Testament has clearly influenced the author of the Oneirocriticon.14 A further example of biblical influence is the frequent use of the word μυστήριον (secret), 35 as well as the wording of phrases like: είς μέγιστον κριτὴν κολληθήσεται (he will devote himself to a great judge); 6 στρατείαν νέαν έξ έτέρων κταται γλωσσών (he will obtain a new army <made up of soldiers> from other nations); ? ουρησεν αίμα έπ' όψει της
73 The word order is slightly different from that in the New Testament; of. Matthew 22:32: Eyo είμι ό θεὸς Αβραάμ, καὶ ὁ θεὸς Ἰσαὰκ καὶ ὁ θεὸς Ίακώβ' οὐκ ἔστιν ὁ θεὸς νεκρῶν, ἀλλὰ ζώντων; Mark 12:27: ούκ έστιν θεός νεκρών, άλλά ζώντων; Luke 20:38: θεός δὲ οὐκ ἔστιν νεκρών, άλλα ζώντων πάντες γαρ αυτώ ζωσι.
14 Drexl 83, 26:ούτος είς μετάνοιαν ήξει και τας έκειθεν κτίζει έαυτοῦ μονάς. Cf. John 14:2: ἐν τῆ οἰκία τοῦ πατρὸς μου μοναὶ πολλαί είσι; Drexl 144, 21: ἀπέχει τὸν μισθὸν αύτου. Cf. Matthew 6:2: ἀμὴν λέγω ήμῖν, ἀπέχουσι τῶν μισθών αὐτῶν (repeated in Matthew 6:5 and 6:16); Drexl 173, 4: ότι ούπω έλήλυθεν ή ώρα της εύπραγίας αύτου. Cf. John 7:30: ότι ούπω ἐληλύθει ή ώρα αὐτου (repeated in John 8:20) and John 13:1: εἰδώς ὁ Ιησους ὅτι ήλθεν αύτου ή ώρα ...
75 Already observed by Nock in his review of Achmetis Oneirocriticon, p. 150.
76 Drexl 132, 16. Cf. Matthew 19:5: κολληθήσεται τη γυναικι αύτου; Luke 15:15: πορευθείς ἐκολλήθη ἐνὶ τῶν πολιτῶν τῆς χώρας ἐκείνης; Acts 10:28: ὡς ἀθέμιτόν ἐστι ἀνδρὶ Ίουδαίω κολλασθαι άλλοφύλω; Acts 17:34: τινὲς δὲ ἄνδρες κολληθέντες αὐτῷ ἐπίστευσαν; etc.
77 Drexl 23, 20. For γλώσσα in the sense of "people, nation," cf. Revelation 6:9: ἐκ πάσης φυλής και γλώσσης και λαού και έθνους; Revelation 7:9: οχλος πολύς... έκ παντός έθνους
<sup>70</sup> Drexl 2, 2.
<sup>71</sup> The quotation is slightly changed from John 14:23: ἐάν τις ἀγαπᾶ με τὸν λόγον μου τηρήσει, και ό πατήρ μου άγαπήσει αύτον και πρός αύτον έλευσόμεθα και μονήν παρ αύτω ποιησόμεθα ("If a man loves me, he will keep my words: and my Father will love him, and we will come unto him, and make our abode with him"),
<sup>72</sup> Drex1 83, 18.
yft~ (he urinated blood on the face of the earth); 78 Ti yuvii eft~ xapa~ i::crtm (the woman will be pleasant); 79 7t00E ~ O~El~ au'tfl~ 7tp0~ "CO aypEUCJat avopa (her feet will be swift in chasing a man). 80 These examples are not references to the Scriptures, but were triggered by his close familiarity with the Bible and other religious literature.
The influence of biblical texts is evident in the grammar, as well. The author uses classical grammatical constructions current in New Testament Greek, such as passive verbs accompanied by nouns in the accusative81 and subjects in the neuter plural construed with verbs in the third-person singular (Attic syntax). Examples of the first are: Et µev E:crn pamA.Eu~. 7tapa-CJKEuacr0iJcrEtat rd *apµara l(ai* rov *arparov* Kata trov ex0prov aU'tOU (18, 18-19) (If he is king, he will make preparations in his weaponry and army against his enemies); eav 'Ct~ l01J, on *'l'OV fva o<fJOaA.µov* E'tU<j>A.roeri (33, 6-7) (If someone dreams that he was blinded in one eye ... ); eav 'Ct~ l01J, on Etpauµmicr0ri *n]v yA.roaaav* auwu (40, 5) (If someone dreams that he was wounded in his tongue ... ); EOV "Ct~ lOlJ 7tEptKEKaAuµµEVTJV KOpTJV *'l'O np6aomov* (76, 24) (If someone dreams of a maiden covered in the face ... ); eav lOlJ n~. on EK07tTJ *ra <Tl(EATJ* Kat *WV<; n6'5a<;* (69, 7) (If someone dreams, that he was amputated at the thighs or legs ... ) etc.82 This use of the accusative occasionally extends to sentences without a passive verb: eav 0£ lOlJ, on iopot *ra<; µaaxaA.a<;* (27, 19) (If he dreams that he was sweating in the armpits ... ); auto~ 0£ a7ta0i]~ µ£vn *rl]v opamv* (33, 12) (But he himself remains unharmed regarding his vision).
Attic syntax only occasionally occurs in the New Testament,83 and not at all in modern Greek, but the author of the *Oneirocriticon* was probably inspired
<sup>1</sup>ml cj>uA.ciiv 1ml A.aciiv 1rnl yA.oocrcrciiv; also Revelation 10: 11; 11 :9; 13:7; 14:6; 17:5.
<sup>78</sup> Drexl 30, 24. Cf. Genesis 10:5: 1rnl 1rnA.u1j1et ti]v Oljltv tii~ yii~; Genesis 10: 15: 1rnl EK<iA.uljlEV 'ti]v O'lfl v ti;~ yi;~; etc.
<sup>79</sup>Drexl 115, 7. Cf. Hebrews 12: 11: xapa~ dvm (qualitative genitive)= it is pleasant.
<sup>80</sup> Drexl 60, 23-24. Cf. Romans 3:15 (from Isaiah 59:75; Proverbs 1:16): o~e'i~ oi itooe~ autciiv EKXEat alµa.
<sup>81</sup>F. Blass, *Grammar of New Testament Greek,* trans. H. St. John Thackeray (London and New York, 1898), §34, 6. Biblical examples of the phenomenon include: oeoeµevo~ toil~ itooa~ (John 11:44); 1itecj>8apµEvot tov voiiv (I Timothy 6: 5); pepavncrµevot tel~ Kapoia~. A.EA.ouµEvot to crciiµa (Hebrews 10:22).
<sup>82</sup> This rule is not followed consistently; cf. EelV tol] 'tt~ on 'tel O"KEATJ autoii E't~aKicr8T]crav *i\* 07tEK07tl] 6 7t0U~ OU'tOU .•• (68, 11-12).
<sup>83</sup> See F. Blass and A. Debrunner, *A Greek Grammar of the New Testament and Other Early Christian Literature,* trans. R. W. Funk (Chicago, 1961), pp. 73-74 ( § 133).
to use it by an almost equivalent phenomenon in Arabic84 (but not in Syriac85). Several examples of a subject in the plural construed with a verb in the singular can be found in the *Oneirocriticon:* 'tcl aiooia auwu cl7tEK07tTJ (his genitals were cut off; Drexl 58, 11); 86 016n 'ta aiooia ayroyor; £cnt wu cm£pµmor; (because the genitals are the conduit of sperm; Drexl 58, 27); µaKpovom1cr£t 'ta 'tEKva auwu (his children will suffer a long sickness; Drexl 59, 2); £npficr0r] f\ E\JfUYTJ 'ta yovata auwu (his knees were burnt or frozen; Drexl 60, 16); OUK ap£cr£t 'tcl ooyµata auwu 'tc\) A.ac\) (his opinions will not be to the people's liking; Drexl 65, 12); qiav£pro8ficrnm 'ta µumfipta auwu (his secrets will be revealed; Drexl 69, 19); 87 ota <1>1A.ov£tKiai; ~TJµtro8ficr£'tat 'ta xpfiµma auwu (he will suffer monetary damage because of a dispute; Drexl 92, 20); 'ta OtKfiµma auwu E7tAatUV8TJ Kat £µqaA.uv8ri Kat 7tAEloV E<i>ffi'ttcr8TJ (his buildings became wider and bigger and better lit; Drexl 101, 19-20); on 't£8£µEA.irotm 'ta PacriA.Eta (the foundations for the imperial palace were laid; Drexl 103, 4 ); 88 ta youv qiuA.A.a 'tcOV 0£.voprov Eir; tfiv OlCl8£crtV KplVEtat tcOV av8pomrov (the leaves of the trees are interpreted as the disposition of men; Drexl 107, 6-7); ti µ£v £qiu'tprocr£ ta o£vopa (if the trees sprouted; Drexl 108, 11); ta &pµata aqiopiav *ano* £x8prov crriµaiv£t (weapons signify not fearing the enemy; Drexl 113, 13); £Ppaxri ta 1µcina amou (his clothes became wet; Drexl 146, 8); Eir; EVOO~O't£pa Kat A.aµnpotEpa anopficrEtm au'tc\) 'ta npciyµma (things will turn out for him even more gloriously and illustriously; Drexl 180, 24 ); £v totr; auto tr; anopficrEtm autc\) ta npciyµma (things will turn out for him in the same way; Drexl 180, 27-181, l); Kat ta Kotva trov fanrov Eir; £Mnova £uy£v£tav Kat oo~av £Kpi8TJ (common horses are interpreted as lesser nobility and glory; Drexl 181, 8-9).89
The language of the *Oneirocriticon* is occasionally tinted with peculiarities
<sup>84</sup> In Arabic when the verb precedes a subject in the third person, the verb always remains in the singular, even if the subject is in the plural; see R. Blachere and M. Gaudefroy-Demombynes, *Grammaire de I' arabe classique* (Paris, 1975), pp. 300 ff. (§ 247).
<sup>85</sup> In Syriac a plural subject requires a plural verb; see T. Noldeke, *Compendious Syriac Grammar,* trans. J. A. Crichton (London, 1904; rpt.Winona Lake, 2000), p. 255 ( §321).
<sup>86</sup> However, cf. E:0A.acr0ricmv "ta aiSo1a (58, 14) and on anEKonricrav "ta aiSo1a mhou (58, 26).
<sup>87</sup> However, cf. E:<1>avricrav "ta Kpuma (Drexl 68, 19).
<sup>88</sup> However, cf. 1>t£1a~mo Kncr0ijvm vf:a ~acril..Eta ... , £i µE:v E"tEAEtw0ricrav ... (103, 2-4).
<sup>89</sup> Attic syntax can also be found in other Byzantine dreambooks which, like the *Oneirocriticon,* were written in a language close to the vernacular, but were not translations of Arabic originals.
#### 82 CHAPIBRTWO
that do not occur either in the Old or the New Testament, nor do they conform to the rules of syntax and grammar in what is considered regular Greek. These peculiarities should therefore be attributed to the influence exercized on the *Oneirocriticon* by the language of its source. Deciding in which language this source was written can help us understand the process of disguising Islamic dream interpretation as Christian. Although Islamic works on dream interpretation in Arabic form the ultimate source of the *Oneirocriticon,* it is not immediately clear whether the Byzantine author translated and at the same time adapted Arabic Islamic material to which he had direct access or simply translated a Christian source that was, in its tum, based on Islamic interpretations, unaware of its Islamic origin. The languages of the Christian Orient in which a Christianized version of an Islamic dreambook serving as the immediate source for the Greek *Oneirocriticon* could have been written are Armenian, Georgian, Coptic or Ethiopic, Syriac, and Arabic.90 All, with the exception of the last one, were exclusively used for the literary expression of Christians, and if the source of the *Oneirocriticon* were written in any one among them besides Arabic, the Byzantine author would have been translating from an already Christianized text. However, the occasional peculiarities of grammar, syntax, and vocabulary in the Greek text indicate that its source was indeed in Arabic, since they closely render equivalent Arabic expressions that do not exist in any other Middle Eastern language, including Syriac, which is the language of the Christian Middle East most akin to Arabic. The resulting Greek, where it is influenced by the wording of the Arabic, is awkward and sometimes even unintelligible. Some examples follow.
<sup>90</sup> The Armenian tradition of dream interpretation has not been examined in any scholarly publication that I know of. No such works are mentioned in R. W. Thomson, "Let Now the Astrologers Stand Up: The Armenian Christian Reaction to Astrology and Divination," *DOP* 46 (1992), pp. 305-12; rpt. in idem, *Studies in Armenian Literature and Christianity* (Aldershot, Hampshire and Brookfield, Vt., 1994), no. 11; dream interpretation as a method of divination practiced in Armenia is mentioned on p. 308 and n. 36. Contemporary scholarship does not do justice to the rich Armenian tradition of dream interpretation; dreambooks translated from the Arabic, as well as dreambooks ascribed to Ibn Sirin, were apparently circulating in Armenian (see Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 280-81), but no work on dream interpretation in Armenian, Coptic or Ethiopic is mentioned in K. Brockelmann, et al., *Geschichte der christlichen Literaturen des Orients* (Leipzig, 1907). No Georgian work on dream interpretation is mentioned in P. M. TarchniSvili, J. Assfalg, *Geschichte der kirchlichen georgischen Literatur* (Vatican City, 1955). Nothing on dreambooks is mentioned in E. Khintibidze, "Byzantine-Georgian Literary Contacts," *Bedi Karthlisa: Revue de Kartvelologie* 36 (1978), pp. 275-86. For an overview of the early Christian position on dream interpretation, including Greek, Latin, and a brief mention of Coptic and Syriac Christian works, see Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 230-42, who concludes that the consistently negative or mistrustful position of the church prevented the composition of dream manuals.
Chapter 117 of the Oneirocriticon opens with the phrase (Drexl 69, 2-3): Τα σκέλη και οι πόδες κίονές είσι του σώματος και χωρις αυτών πίπτει το σώμα τιμωρούμενον έν τω έλκεσθαι (The legs and the feet are the pillars of the body; without them the body falls and is condemned to creep).
The verbs of punishment in classical Greek are never construed with a preposition.31 In modern Greek, which is probably closer to what the author of the Oneirocriticon spoke in everyday life, the verbs of punishment are construed with the preposition eis>se. The combination timoroumenos en is definitely not correct Greek. In Arabic, however, the verb hakama, ("to condemn," "to punish") is construed with a preposition followed by the penalty inflicted: hakama bi. One of the possible translations of the preposition bi in Greek is en.
The Oneirocriticon contains the following narration of a dream (Drex) 148, 18-21): Είδον και όναρ, ότι πάντες οι κέραμοι της πόλεως ανευ ύετου ἔρρεον ἔνθολον και πάντες ἐδέχοντο τὸ ὕδωρ ἄνευ ἐμοῦ καὶ τῶν ἐμῶν ανθρώπων (I dreamt that all the roof tiles of the city were streaming with turbid <water>, though there was no rain, and that everybody was collecting this water, except for me and my people).
The word ένθολον does not fit very well into the Greek sentence. Instead of an adjective in the accusative case one would normally expect an adverb. The anomaly of the Greek text is explicable, however, if we assume that it was translated from the Arabic. Arabic does not have proper adverbs; instead, it employs adverbial accusatives. 2 Syriac does not have a proper equivalent of the Greek accusative 3 and Syriac adverbs are normally formed with an adjective and the suffix -ith, or, more rarely, a word compounded with a preposition."
<sup>91</sup> The verbs of punishment are usually construed with the genitive of the punishment inflicted: τιμώ τινι φυγής (I condemn someone to exile); ἔκριναν αὐτὸν θανάτου (they condemned him to death).
<sup>92</sup> The adverbial accusative "amply makes up for the want of adverbs in Arabic" ; see C. P. Caspari, A Grammar of the Arabic Language, 3rd ed., trans. W. Wright (Cambridge and New York, 1979), vol. 2, pp. 109 ff., § 43-44; vol. 1, pp. 288 ff., § 364. This phenomenon was called compliment circonstanciel de manière (e.g., مــتـــمــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ Demombynes, Grammaire de l'arabe classique, p. 294, §238a; cf. also pp. 207-8, §134, "noms au cas direct de valeur adverbiale."
<sup>95</sup> Syriac has no cases marked with terminations; the nominative and the oblique cases of Greek and Latin are recognized by the context or are expressed by the constructive state, by the influence of a transitive verb, or by some particle; cf. G. Phillips, A Syriac Grammar (Cambridge and London, 1866), p. 53.
<sup>94</sup> On the formation of adverbs in Syriac, see T. Nöldeke, Compendious Syriac Grammar, pp. 98-101; also Phillips, Syriac Grammar, pp. 132-33, 177.
The irregularity of the Greek text in this passage proves that it was translated from Arabic and not from Syriac.
In chapter 169 of the *Oneirocriticon* we read the passage (Drexl 132, 15-20): eav l01J 1u;, on U1t0 acn£prov oomouµEvor; 01t001']µ£l Ota "CE ~ripar; El TE 8aA.acr<JT]r; Ka'tEUOOOU'tat, de; µ£y1crwv Kpl 'ti]V Kat oi KatOV KoA.A.118i]cr£'tat Kat Eupi]crn xaptv 1tap. auwu. £i OE Ti Mor; auwu £y£v£'t0 *Kata* 1tAclVT\V OKa'tEUOOoYtEpOV, de; av8pro1tOV \j/EUOOKpt'ti]V f}~Et Kat 8A.1pi]cr£'tat 1tap' auwu (If someone dreams that, guided by the stars, he had a successful journey by land or sea, he will devote himself to95 a great and fair judge and will find favor with him. If his journey became more difficult because of an error, however, he will go to a false judge and will be grieved by him). The second half of the excerpt includes an obvious mistake: d 0£ Ti 6oor; amou £y£v£w Kma 7tA.avriv aKmEuoomtEpov (If his journey became more difficult because of an error, however. .. ). According to the rules of Greek syntax, the case, gender, and number of the predicate should agree with those of the subject. In our example, the predicate is in the accusative case while its subject is in the nominative. In Arabic the verb .;L..:. (~ara = to become; *egeneto* in the Greek text) is construed with the predicate in the accusative case, even though its subject is in the nominative. Again the irregular syntax of the Greek text suggests a slavish adherence to the syntax of an Arabic model. Syriac does not have an equivalent to the Greek accusative case. If we postulate that the *Oneirocriticon* was translated from Syriac, we would not be able to account for the accusative case of the predicate in Greek.
Chapter 218 of the *Oneirocriticon* contains the phrase (Drexl 171, 8-9): 6µoiroi; £av m'.noi; EKEtvor; o pamA.Eur; *£w\j/£* nva µayKA.cipw il xapl;civw ... (Likewise, if the king himself beat someone with whips or straps). In this phrase not only the object of the verb *typto* (to beat) but also the instrument with which the action of the verb is carried out is in the accusative
<sup>95</sup> The Greek KoA.A116i]crE1m (he will devote himself to) could be attributed to the influence of biblical language on the vocabulary of the Greek author, since KoA.Acioµm (to devote myself to) occurs in the New Testament (e.g., Matthew 19:5 KoA.A116i]crE1m 11] yuvmKt mli:oii; Luke 15:15 7t0pEu6d~ EKOAAT]611 EVt 'tclJV 1toAt'tclJV TI]~ xwpa~ EKEiv11~; Acts 10:28 ffi~ ci6£µti:6v fon *a* vii pt' louliai{\l 1COAAcicr6m ,. . ciA.A.o<IJUA{\l; Acts 17:34 n VE~ OE avlipE~ 1COAA116£v1E~ aui:<jl E:nicri:Eucrav). In the New Testament, however, this verb is consistently accompanied by its object in the dative, while in the *Oneirocriticon* it is followed by the preposition Ei~ + accusative. The phrasing in the *Oneirocriticon* could possibly reflect a close rendering of the Arabic verb Jl....:; *(ta'allaqa)* which is regularly followed by the preposition *hi* (possibly rendered in Greek as Ei~) and means "to be devoted to." The verb *ta'allaqa* is form V of the root Jl..c *('aliqa),* the primary meanings of which are to hang, to be suspended, to cling, to stick, to adhere to, the same as the literal meanings of the verb KoA.A.cioµm in Greek.
case. But the instrument in Greek is never expressed in the accusative; it is always expressed either with a dative or with a preposition and an oblique case, as is done throughout chapter 218, with the exception of the phrase quoted above. <sup>96</sup>Though the Arabic verb *qaraba* (to strike) is usually accompanied by the object of the verb in the accusative and the instrument of striking prefixed by the preposition *bi* (with) (e.g., ..bJ-'-"-:' l..1.:>j ~fa= I hit ~ Zayd with a whip), it is also possible to construe *qaraba* with two accusatives, one expressing the object and one the instrument (Lb\_,\_\_....,.. l..1.:>j ~~=I hit Zayd with a whip97). It is therefore possible to account for the irregular instrumental accusative of the Greek text if we assume that the Greek author was translating from the Arabic and following closely the syntax of the Arabic source. However, no explanation can be provided if we postulate that he was translating from a Syriac source. Syriac does not have a proper accusative and expresses the instrument by prefixing a *lomadh* to the noun.
In the *Oneirocriticon,* dreaming of a pleasant smell is consistently interpreted as gaining an excellent reputation, while dreaming of an offensive smell signifies losing one. 98 The connection between this symbol and its meaning does not come from the ancient oneirocritic tradition-Artemidoros says nothing about dreaming of perfumes and smells-and seems arbitrary. It becomes obvious, however, once we realize that the Arabic expression:;~ ~ • a . ,, *(sum< a <atira),* which actually means "excellent reputation" but can be literally translated as a "fragrant reputation," must have given rise to the interpretation in Arabic dreambooks of a fragrant smell as indicating a good reputation. Reliance on puns and the alternative meanings of words for interpreting dreams is a method mentioned in ancient Greek, Jewish, Islamic and Byzantine texts on dream interpretation.
<sup>96</sup> Drexl 171, 3: E1:\Jq>0Ji ~ouvn'ipot1; (He was beaten with straps of raw ox-hide); Drexl 171, 67: wptcrE "tu<j>6fivai nva ~ouvEupot1; (He commanded that someone be beaten with straps of raw ox-hide); Drexl 171, 10: E'tulj/E nva pci~oo;i (He beat someone with a stick); Drexl 171, 13: µEta 01ta0Ji~ EOCOKE n vt (He gave someone [a blow J with a sword).
<sup>97</sup> See Lane, *Arabic-English Lexicon,* s.v. '-:"~(col. 1777).
<sup>98</sup>Ei OE lOl:J, on KaltVOU *£n1'.i]ae11* 6 OlKO~ ... El OE 6 KaltVO~ EUOOO!l~ ~v. A.Oyot~ yAuKEat tl]v E~ouaiav autou napE~Et EtEpot~. EUpljaEt OE Kat <1>iJµ11v avcil.oyov tii~ toU KaltVOU EUCOOta~ Kat tii~ i]out11to~ (121, 23-26). Ei. OE Ei.~ ouaroo11 n11Mv 1tEpt1tCHEl, ~aput£pav 6A.tljltV Kat <j>i]µ11v KaKi]v Acl~Et Ota tl]v ouacooiav (134, 8-10). Ei. OE lOl:J, on EUpE o£vopov Kttpov, Eupi]crEt avopa EUyEvii, ltAOUO'lOV, EU<\>11µov Ota tl]v EUCOOiav Kat to EUKapnov (155, 25-26). El OE lOl:J, on E<j>ay£ O'Kopoov *i\* Kp6µµuov *i\* npaaov ... <1>iJµ11v KaKi]v E~Et Ota tl]v OUO'OO'µtaV aUtcOV (160, 8-10). EaV lOl:J n~, On fo6tE pE<j>clVOU~ ... , KaKO<\>llµO~ Eatat ava µfoov 1'.aou Kat µtcr11to~ Ota to tii~ EpEUYll~ o\Jaoaµov (162, 8-10). El OE (lOl:J nooopna) ouacooiav €xovta, <1>iJµ11v KaKi]V Eupi]crEt Ota to OUO'cOOE~ EV ttji Epyqi toU µtcr6ou autou (176, 6-8).
#### 86 CHAF'TER TWO
Besides the examples indicating that Arabic and not Syriac was the language from which the author was translating, others conform to patterns common to most Semitic languages. For instance, in the phrase 1mA.aioac; £av E1t't(OOEV £-repov, nµi]oEt -rov neo6v-ra (If, while wrestling, he made his opponent fall, he will honor his fallen opponent),99 a verbal form otherwise unattested in the Greek language was concocted from the stem of the perfect tense of the verb nimm (to fall) 100 and used in the sense of "to make someone fall." Such a grammatical twist was probably instigated by the mechanism of producing verbs from a root according to a pattern that is typical of all Semitic languages. Form I of the root b li ... *(s-q-.t)* means "to fall"; form IV means "to cause someone to fall." The spelling of the third person singular in the imperfect is identical in both forms ( b li . ,, *.:),* though in each case the word is vocalized differently. Instead of emmoev the correct Greek verb would have been av£-rpE\JIEV or KmEpt'lfEV (which is close to the Modern Greek £ptl;E), but it seems that the author of the *Oneirocriticon* was unable to think of an alternative to *nin-rw,* the primary meaning of the Arabic root *s-q-.t,* and since it was inappropriate for his translation, he resorted to emmoev. <sup>101</sup>
The *Oneirocriticon* includes at least one passage where the words at the end of each colon conform to a rhythmical pattern. This phenomenon of literary prose has been broadly defined by S. Skiminaw2 and has been studied mainly
99 Drexl 72, 18.
100 In the perfect JtEJt'tOJKa. The stem m:w- is used in other derivatives of nimw, such as mwcru;, Jt'tooµa.
101 A similar way of thinking may have resulted in the use of PacrtA.Euw in the following passages: pacrtA.EucrEt µiav £1; au-rwv (110, 16) (He will make one of them queen); ti µ£v EXEt uiov, PacrtA.dcrn au-rov, d 0£ µr,, faEpov pacrtA.Eucrn (175, 6-7) (If he has a son, he will make him king; if he does not have a son, he will make someone else king). The Arabic equivalent of this verb is dL which in form I means "to reign" and in form IV "to make someone reign." The third person singular in the imperfect of both forms is spelled ~· The primary meaning of PacrtA.EUW (the only possible in classical literature) is "to reign"; however, the meaning "to make someone king" occurs at least three times in patristic Greek texts (Basil of Seleucia, Ephraim the Syrian, and John of Nikiou. They all lived in areas where Semitic languages were spoken, and especially the works of Ephraim were originally written in Syriac and then translated into Greek. For references to these texts, see Lampe, s.v. "PacrtA.Euw").
102 See S. Skimina, *L' erat actuel des hudes sur le rhythme de la prose grecque II* (Lwow, 1930), pp. 5-6: "Un texte, divise en unites logiques, presente a la fin de chaque phrase ou de chaque proposition une clausule qui se distingue au point de vue rythmique du reste du texte. C'est le propre du rythme de la prose artistique, ou plut6t d'une tendence souvent inconsciente de chaque ecrivain et que Jes theoriciens ont remarque et condifie depuis longtemps." The most recent and comprehensive publication on the subject is Horandner, *Der Prosarhythmus in der rhetorischen*
in Byzantine authors of the high style, such as Synesios, 100 Procopios of Caesarea, 104 and Agathias. 105 Each author prefers a different rhythmical pattern. As to their variety, Skimina enumerated eighteen different types belonging to three different forms that are applied to the last two to six syllables of a colon or semi-colon.100 The four examples of rhythmic phrases in the Oneirocriticon could be extended to comprise the last twelve to fifteen syllables at the end of the colons where they appear (the phrases that present rhythmical patterns are italicized):107
πολύτροπός έστιν ή των δένδρων κρίσις πολυσχεδής 108 γάρ έστιν ή των δενδροκάρπων φύσις, τα γουν φύλλα των δένδρων είς τὴν διάθεσιν κρίνεται των ανθρώπων τα γαρ εὐθαλὴ καὶ εὐτραφὴ εὐθύτητα γνώμης διασημαίνουσιν, τα δε άσθενη και ροώδη και μεμαραμμένα γνώμης διακρίνουσιν άσθενείς τρόπους.
The interpretation of trees is complex, since fruit-bearing trees by nature have many component parts. The leaves of the trees are inrpreted as the disposition of men; those that are thriving and luxuriant indicate straightness of intent, but those that are declining and falling off and withering away signify the weakness of one's intent.
The rhythmical patterns discernible in the above excerpt can be analyzed as follows:
xxx--- // xxxx-- ( twelve syllables, caesura after the 6th : ------- // -------- // ---------- // -------- /109 xxxx--- || xxxx ---- (fifteen syllables, caesura after the 7th: -------- || ( ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------
xxxxx-~~ // xx-~ (twelve syllables, caesura after the 8th: ~~~~~~~ // ~~~ ) xxxx-~~ // xxx-~ (twelve syllables, caesura after the 7th:-------- // ----- )
Literatur der Byzantiner (Vienna, 1981), including study of several authors from the Middle Byzantine period.
103 N. Terzaghi, "Le clausole ritmiche negli opuscoli di Sinesio," Didaskaleion 1 (1912), pp. 205-319.
104 H. B. Dewing, "The Accentual Cursus in Byzantine Greek Prose with Especial Reference to Procopius of Caesarea," Transactions of the Connecticut Academy of Arts and Sciences 14 (1910), pp. 415-66; A. W. De Groot, Untersuchungen zum byzantinischen Prosarhythmus (Groningen, 1918).
105 G. Franke, Quaestiones Agathianae (Bratislava, 1914). For remarks on later Byzantine authors, see W. Horandner, Der Prosarhythmus, pp. 47-152.
106 Skimina, L'état actuel, p. 9; see also Hörandner, Der Prosarhythmus, p. 33.
107 Drexl 107, 5-12,
108 Corrected by Charitonides, review of Achmetis Oneirocriticon, p. 233, to πολυσχιδής.
109 The regular caesura in a Byzantine dodecasyllabic verse is after the fifth or the seventh syllable; cf. P. Maas, "Der byzantinische Zwölfsilber," BZ 11 (1903), pp. 278-323.
#### 88 CHAPTER TWO
Every colon or semi-colon is preceded by a rhythmic phrase in two parts. The first part ends with a dactyl ( ---) and the second part with a trochee (--). The trochee of the first rhythmic phrase in the quotation (Kpicru;) rhymes with the trochee of the rhythmic phrase that immediately follows it (<!Jucru;) *(homoioteleuton).* The last rhythmic phrase in the quotation constitutes a regular Byzantine dodecasyllabic verse with the caesura after the seventh syllable. <sup>110</sup> Given that this is the only instance of colons ending with rythmical patterns that I was able to identify in the *Oneirocriticon,* it might well be accidental. Or, if it is intentional, its rarity indicates that the author could not produce it with ease, so even in this case it does not necessarily reflect the author's familiarity with texts of the high style. The occurrence of rhythmical prose at the end of colons has been observed in texts read by the educated but not erudite Byzantine public; they include the chronicles of Theophanes and George the Monk and the writings of St. John of the Ladder. 111 Rhythmical patterns extending longer than the last six syllables of a colon also occur in the New Testament. <sup>112</sup>
Our author's work, though far from erudite, betrays a man with a Byzantine elementary education, familiar with such basic texts as Homer, 113 and possibly
<sup>111</sup>See Skimina, *L' hat actue/,* p. 24 (summarizing information originally found elsewhere). For elements of poetry, including rhythm, in the work of St. John of the Ladder, see J. Duffy, "Embellishing the Steps: Elements of Presentation and Style in the *Heavenly Ladder* of John Climacus," *DOP* 53 (1999), pp. 1-17.
<sup>113</sup>It is even possible that the author of the *Oneirocriticon* had read some patristic texts that contain words such as aip£atapx11s ("heresiarch" ; Drexl 8, 21) and A.oyoµaxos ("contender
<sup>&#</sup>x27;'0 The Byzantine dreambook attributed to Patriarch Nikephoros is written in Byzantine dodecasyllabic verses, but the dodecasyllable of the *Oneirocriticon* cannot be found there, which indicates that it was neither lifted nor copied from somewhere else; see Nikephoros, ed. Guidorizzi, V. 56: KOltEVm Mvopa lt'tWCHV avopciiv µ11vu£t (repeated verbatim in v. 111 of the dreambook attributed to Patriarch Germanos, ed. Drexl). The interpretation of trees in the remaining Byzantine dreambooks do not coincide with the rhythmical passage of the *Oneirocriticon,* either: Nikephoros, ed. Guidorizzi, v. 78: 211pciiv <)>avE:v-rcov CiEvopcov E:v K£vots n6vot or K£vot oi K6itot. Daniel, ed. Drexl, vv.121-123 and 127: l'>E:vopov Kapito<)>6pov l0£lV KEPOos a11µaivn./l'>E:vopa s11pa t0£lV clltO'tuxiav a11µaivn, aAA.a oi] Kat l;11µiav Kat oouA.qi KEpoo<;/Mvopa EKptl;coµE:va l0£lV ij Ka'ta"tEµVHV ltOAEµOUs Kat lt'tWCHV K'tllVWV Kat av6pcimcov O!lAOl/ ... i'lf:vopov av6ouvm l0£lV ltCTO'l KaA.Ov EO''tlV. 312: KA.aoov clltO CiEvopou KA.aam <)>tA.iav amilaaaem <011A.ot>. Anonymous dreambook from *Paris. gr.* 2511, ed. Drexl, vv. 55-57: l'>E:vopa io£tv Kapno<)>6pa daooov a11µaivn, s11pa OE clltO'tuxiav./ l'>E:vopa K01t'tOµ£va l0£lV EA£U6Epots µEv t;11µiav, OoUAOls OE KEPOos 011A.ot/Mvopa i.&tt v EKpt1,;ouµ£va 1t0A.£µov Kat lt'tWCH v avepcilltcov a11µai V£t. Manuel Paleologos, ed. Delatte, chap.14: Ta µf:ylO'-ra OE:vopa £ts £uy£vous *[sic]* avopas KptV£'tat Kat Y£VVatOUs Kat {j lt'tWO'ls au-rciiv 6avmov Eµ<)>atVH.
<sup>112</sup> Matthew 16:17-19; Matthew 28: 18-20; Luke 11:2-5; John 1:1-8; John 6:26-59; Romans 3:24-26; etc. For an analysis, see K. Metsakes, *Byzantine hymnographia apo ten epoche tes Kaines Diathekes heos ten Eikonomachia* (Athens, 1986), pp. 41-42.
with access to philological tools such as dictionaries, as is evident from his gloss on the word Kv~ri; he may have dabbled in learned literature, though he was far from having mastered it. His readings from the Bible and especially, if not exclusively, the New Testament must have been much more extensive, since this is the most salient textual influence on his work, evident in the quotations, language and expressions employed. He also seems to have been familiar with ecclesiastical literature.
His Greek is influenced by Arabic, the language of his model, but this is not an unusual fault among translators. The Arabic translation of Artemidoros by J:Iunayn b. IsQ.aq is a pertinent example. <sup>114</sup>Before deciding which language -Greek or Arabic-the author of the *Oneirocriticon* appears to have known better, it is important to keep in mind that too slavish a translation (reflecting word by word the language of the prototype and therefore resulting in solecisms in the language into which one translates) can be a choice made not because the translator has not mastered the language into which he is translating, but because he is unsure about the meaning of words in the language from which he is translating. This problem can be compounded when translating from a language such as Arabic, which, like Greek, has a literary idiom rather different from its spoken vernacular. Mastering the intricacies of classical Arabic required special schooling even for a native speaker, which individuals received in varying degrees. Among the four examples of too literal renderings from Arabic into Greek resulting in irregular Greek expressions, three are slavish renderings of grammatical phenomena that occur in literary Arabic, but which modern dialects tend to do away with, suggesting that at least some of the solecisms in the Greek translation might have been caused by the author's unease with classical Arabic.
One instance could be interpreted as an indication that the author of the *Oneirocriticon* was a native speaker of Greek who had learnt Arabic as a
about words, quibbler" and also "contender against the Logos", i.e. Arian; Drexl, 8, 21). The latter word was also used by Gregory of Nazianzus (see Lampe, s.v. "A.oyoµaxo~" ), who was characterized as "the Christian Demosthenes" and was by far the most widely read patristic author in Byzantium.
<sup>114</sup>I:Iunayn must have had native fluency in Arabic, because Arabic was the vernacular of his home town, al-l:lira in Iraq. At the same time, he belonged to the Syrian Nestorian church, where Syriac was the language of the liturgy and of high Christian education. He translated from Greek into both Arabic and Syriac, and his predilection for Syriac, as opposed to Arabic, should not be considered as a sign of greater fluency in Syriac. As I:Iunayn himself explained in his writings, compared with Syriac or Greek or Persian, Arabic in his time lacked an adequate scientific vocabulary, which he and his students and colleagues helped foster with their translations. For further comments and references, see *EI2,* s.v. "I:Iunayn b. lsJ:ia~ al-'Ibadi."
second language. The following passage from the Oneirocriticon includes an expression that is incomprehensible in Greek: 15 ' Eάν τις Ίδη, ότι αι ρίνες αυτού έφράγησαν και ούκ όσφραίνεται, εί μέν έστι βασιλεύς, νοείτω, ότι κινδυνεύσει έν αύτφ ό άναφέρων ἐνώπιον αὐτοῦ πρώτος τὰς κοσμικὰς διοικήσεις, διότι ή διάκρισις της εύοσμίας και της δυσωδίας διά της ρινός ἐστι τφ ήγεμονικώ (If someone dreams that his nostrils were obstructed and he could not smell, if he is king, he should know that the chief among those who report to him about secular administration" will risk his life [?] because the intellect can distinguish between good and bad odors through the nose). The meaning of the phrase κινδυνεύσει έν αύτω (rendered in our translation as "he will risk his life") is unclear. The combination κινδυνεύω ev (to risk in) does not appear in Greek dictionaries, and contemporary translators disagree as to how it should be understood. 17 Once the Arabic provenance of the Greek text is taken into consideration, however, κινδυνεύσει خـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ ه (khāṭara bi-nafsihi) or جازف بنفسه (jāzafa bi-nafsihi) which is an expression meaning "to risk one's life," an expression the Greek author seems not to have known, since he translated each word literally and came up with "to risk in oneself."18
It is impossible to ascertain where the Oneirocriticon was composed. Though the profile of its author (a tenth-century Greek who had learnt Arabic and was conversant with biblical and ecclesiastical literature) seems to point to the monastic milieu of Sinai and Palestine, any suggestion that this was the world where the author belonged would be a conclusion hastily drawn. Not only monks living in Muslim lands but also former prisoners of war, traveling merchants, and residents of the eastern frontier of the empire all had the opportunity to learn Arabic. Biblical and ecclesiastical readings interested both men of religion and secular individuals. Even Constantinople cannot be excluded as a possibility, especially if the Oneirocriticon was compiled for an emperor, as suggested earlier. Unless further evidence appears it is unlikely we will ever know anything more definitive about the identity of the Oneirocriticon's author or the date and locale of its composition.
<sup>115</sup> Drex1 36, 12-16.
<sup>116</sup> It appears that the phrase ὁ ἀναφέρων ἐνώπιον αὐτοῦ τὰς κοσμικὰς διοικήσεις renders an office that did not exist in the Byzantine court (vizier, wali?).
<sup>117</sup> Oberhelman, Oneirocriticon of Achmet, p. 107, translates it as "will fall into danger along with him." Brackertz, Traumbuch des Achmet, p. 51, renders it as "in Ungnade bei ihm fallen wird."
<sup>118</sup> khāṭara = to risk = κινδυνεύσει; hi = in = ἐν; nafsihi = himself = αὐτῷ (ἑαυτῷ).
#### CHAPTER THREE
#### THE MANUSCRIPT TRADITION, TRANSLATIONS, AND EDITIONS
When Drext published the critical edition of the Oneirocriticon, he knew of sixteen Greek manuscripts and based his text on eight of them. To these one can add another seven that contain parts and versions of this work, some long some short. I made no systematic effort to identify all of its extant manuscripts and therefore a few more manuscripts containing at least parts of the Oneirocriticon may have escaped my attention. When I set out to reexamine the manuscript tradition, my objective was not to establish a foundation for a new critical edition but to circumvent the limitations imposed by the existing one that make it difficult to use for investigating problems such as the identity and linguistic ability of the Greek author and the exact relationship of the Greek text to equivalent Arabic works. As grateful as one ought to be for the existing critical edition, especially considering the complicated and contaminated manuscript tradition of the work in question, it has also to be acknowledged that Drexl's text is, like all scholarly publications, very much a product of its time. As a result, its usefulness for my purposes was often compromised, not by the limitations of the editor's philological expertise, but by his adherence to criteria that today, after eighty more years of scholarship, seem anachronistic.
Drexl's text is regarded as inadequate, mainly because at the beginning of the twentieth century, the editorial technique adopted for Byzantine texts was the same as that applied to classical Greek texts: the existing manuscripts were classified into families of older and better (closer to the archetype) or recentiores et deteriores. They were subsequently arranged in a stemma and the choice of the different readings was often made mechanically, dictated by the stemma. Since then scholars have become aware that Byzantine scribes did not copy all kinds of texts with the same reverence. For texts such as the Old and the New Testament, the writings of the Fathers of the Church and the canon of ancient authors every effort was made to ensure that the text copied was as close to the archetype as possible, often by collation and correction on the basis of more than one manuscript. Volksliteratur, on the other hand, along with scientific
<sup>1</sup> Cf. H.G. Beck, "Überlieferungsgeschichte der byzantinschen Literatur," Geschichte der Textüberlieferung der antiken und mittelalterlichen Literatur, ed. H. Hunger et al., vol. 1 (Zurich, 1961), pp. 425-510.
#### 92 CHAPfER THREE
and pseudo-scientific texts destined for practical consultation, was treated in a less formal way. The scribe, who in many cases was copying for his own use, felt free to deviate from his model at will by adding, subtracting, or paraphrasing, depending on what he judged to be appropriate for the future user's purposes. This "irreverent" attitude accounts for the confused and contaminated tradition of Byzantine texts on medicine, pharmacology, alchemy, astronomy, astrology, dream interpretation, and other forms of divination. The different attitude of the scribes toward different kinds of texts makes it not only inappropriate but also impossible to adopt the criteria developed for classical philology to the editing of Byzantine texts of the "unrevered" category. The weaknesses in Drexl' s critical edition of the *Oneirocriticon* well illustrate this problem.
In the following pages I will try to demonstrate the importance of the newly surfaced data, including extant Arabic texts on dream interpretation, for textual criticism of the *Oneirocriticon,* and the need for a new critical edition. I will begin by presenting the seven Greek manuscripts that were unknown to Drexl. Discussion of the remaining sixteen will be limited to observations not made in Drexl's publications and resulting from my own examination of the relevant manuscripts. The Latin versions of the *Oneirocriticon* will also be discussed, as their importance in shaping W estem medieval tradition on dream interpretation can hardly be overestimated. The problems of Drexl' s edition will be pointed out, and the guidelines along which a new critical edition can be planned will be indicated.
#### *Greek Manuscripts*
*Paris. Suppl. gr.* 690. I Ith century, parchment, 258 fols., 24 cm x 19 cm, anthology of texts. 2
*Paris. Suppl. gr.* 690 is a collection of 94 texts representing a wide variety of authors and literary genres. It has been in the Bibliotheque Nationale since 1859, but the only adequate description of it was not published until 1950. 3 It is important for the manuscript tradition of numerous Greek works. It is a *de luxe* volume, though it has no illuminations and its ornamentation is simple,
<sup>2</sup>For a detailed description of the manuscript and a report on its contents, see G. Rochefort, "Une anthologie grecque du Xie siecle," pp. 3-17.
*<sup>3</sup> Paris. Suppl. gr.* 690 was brought from Mount Athos to Paris by Minas Minoides in 1842 and was obtained by the Bibliotheque Nationale upon his death in 1859; see G. Rochefort, "Une anthologie grecque du Xie siecle."
consisting of titles and initials written in gold ink. According to G. Rochefort, it must have been written between 1075 and 1085.
Fols. 123v-132v comprise a collection of short texts on dream interpretation that was discussed by D. Gigli in 1981; 4 it includes an abridged version of the *Oneirocriticon* on fols. 125-129, which constitutes its oldest known surviving version.5
The abridgment in the Paris anthology bears the title "'EK -r&v 'Ivo&v, CTEpcr&v Kai Atyu1nimv" (From the Indians, Persians and Egyptians); no author is indicated.6 The distinction between the dream interpretations of the Indians, Persians and Egyptians is made only in the title and does not appear in the chapter headings, as it does in Drexl's text. Moreover, the titles and the order of the various paragraphs are in several instances very different from the critical edition, and some of the paragraphs are missing altogether. The interpretations given are schematic and comparable to those in the dreambook of Daniel. Gigli observes that "a compendium of this type would correspond to a popular dissemination, in which schematic treatises are easier to use and to consult" and points to passages from the abridgment where the epitomist admits that his work is not meant for the professional dream interpreter (fol. 125r, col. 1): CTEpt yap 'tfi<; Ota<jlopa<; 'tcOV ElOcOV Kat 'tfi<; rrpo<; au-ra KpicrEm<; 7totKtATJ<; ofuric; £mmrrficraµEv, •0 ovEtpoKpi'tl) TI]v WU'tTJ<; Kpicrtv rrapaxmpficrav-rE<; (We have been silent on the subject of the difference between the kinds [of dreams] and their various interpretations, ceding the judgment about this to the dream interpreter).<sup>7</sup>
Gigli collected all the instances where the text of *Paris. Suppl. gr.* 690 implies the possibility of new variants, or contains notes and interpretations that are absent from the other manuscripts of the *Oneirocriticon. <sup>8</sup>*In particular, he identified a number of variants in *Paris. Suppl. gr.* 690 that confirmed readings that were rejected by Drexl and were only included in the apparatus to the critical edition, or that offer a better text than the other Greek manuscripts
<sup>4</sup> D. Gigli, "Gli onirocritici del cod. *Paris. Suppl. gr.* 690," *Prometheus* 4 (1981), pp. 65-86, esp. 79-86, and 173-88.
<sup>5</sup> Drexl incorrectly considered *Vindob. philos. et phi/of. gr.* 111 of the 13th century to be the oldest surviving Greek manuscript of the text.
<sup>6</sup> Gigli, "Gli onirocritici del cod. *Paris. Suppl. gr.* 690," p. 85, suggested that the lack of author's name and the "generic" title of the work indicate that the epitomist distanced himself from his model. However, further evidence from the Greek and Latin manuscript tradition indicates that the epitomist simply copied the title of the work as he found it in his model.
<sup>7</sup> Ibid., p. 79.
<sup>8</sup> Ibid., pp. 83-85.
#### 94 CHAPTER THREE
known to him.9 A further instance shows that Drexl was correct to insert in his main text an interpretation not attested by the Greek manuscripts but found in the Latin translation of Leo Tuscus. '° With reference to the two main branches in the manuscript tradition of the *Oneirocriticon* distinguished by Drexl, *x* (= *codd. meliores)* and *y* (= *codd. deteriores),* Gigli noted that the abridgment includes readings that belong to both branches. He therefore suggested that the epitomist based his abridgment either on one single model with contaminated readings from both branches of the manuscript tradition as described by Drexl, or on a number of different models belonging to both the *x* and *they* branches of the tradition.
Gigli identified a few dream interpretations that are included in *Paris. Suppl. gr.* 690 but are absent from the standard version of the text published by Drexl. Some of them are similar to interpretations offered by Artemidoros and pseudo-Nikephoros. Others are particular to *Paris. Suppl. gr.* 690 and do not appear anywhere else in the extant Greek dreambooks. Gigli considers the variants that are close to Artemidoros and pseudo-Nikephoros as additions made by the epitomist, who presumably knew these two texts. <sup>11</sup>
If we examine *Paris. Suppl. gr.* 690 in the light of the evidence presented by Arabic dream interpretation, however, this conclusion must be modified. At least five out of nine interpretations that occur only in *Paris. Suppl. gr.* 690 and no other extant Greek source can be found in Arabic dreambooks. <sup>12</sup>These interpretations must already have existed in the model for the abridgment. Moreover, Arabic dreambooks contain all four interpretations that occur in the Paris abridgment and in pseudo-Nikephoros. Since the phrasing in *Paris. Suppl. gr.* 690 is far removed from the phrasing in Nikephoros, it is safe to conclude that the source of these interpretations in the abridgment is not Nikephoros but a version of the *Oneirocriticon* longer than the one printed by Drexl. The five interpretations that occur in *Paris. Suppl. gr.* 690 and Artemi-
<sup>9</sup> E.g. *Paris. Suppl. gr.* 690, fol. 126v, col. 2: ii µi:v Kapfo Ei~ avopa y£povi:a Kat cp£t8wA6v; in Drexl 154, 16: avopo~ µqicri:ou ... cpnowA.ou, but in apparatus rBRSTV: y£pov1:0~ cp£lOOJAou. Further examples in Gigli, "Gli onirocritici del cod. *Paris. Suppl. gr.* 690," pp. 81-85.
<sup>10</sup>*Paris. Suppl. gr. 690,* fol. 128, col. 2: wcra-ui:w~ Kat o oaKtUAto~ ~acriAEtov £~oucriav Kat yuvmKo~ Kat i:£Kvoiv oi]A.wmv E)'.El (Likewise, a ring suggests royal power and signifies a wife and children); cf. Drexl 211, 22-23.
<sup>11</sup>Gigli, "Gli onirocritici de! cod. *Paris. Suppl. gr.* 690," p. 85.
<sup>12</sup>I have searched in the five Arabic dreambooks I chose as the basis for comparing the *Oneirocriticon* to Arabic dream interpretation, two of which are available only in manuscripts; of the remaining three published versions, only one has an index of dream symbols. Given these conditions, it is possible that I may have missed some interpretations in the Arabic.
doros present a somewhat more complicated case. I have been able to locate four out of five in Arabic. The phrasing of two of those interpretations in Paris. Suppl. gr. 690 and the Greek text of Artemidoros is very close, however; therefore, it is not out of the question that the epitomist was familiar with this second-century text. 13
The fact that the additional interpretations of Paris. Suppl. gr. 690 can be found in Arabic dreambooks should keep us from assigning them to the epitomist. Since the purpose of an abridgment is to shorten a work, it is unlikely that the epitomist would add even a few further interpretations from different sources at the same time as he was eliminating hundreds of others from his model.44 It is more likely that all the additional interpretations, including those that Gigli attributed to the epitomist's knowledge of Artemidoros and Nikephoros, existed in the epitomist's model, again suggesting a version of the Oneirocriticon longer than the one found in Drexl's critical edition. Such a conclusion is supported by two additional pieces of evidence: first, one of the examples collected by Gigli that confirm readings rejected by Drexl for the critical text but reproduced in the apparatus of the edition indicates that the epitomist's model was a text longer than the one printed in the critical edition.15
<sup>13</sup> For a detailed catalogue of interpretations missing from Drexl's text and the corresponding Arabic interpretations, as well as the possible connection between the abridgment of Paris. Suppl. gr. 690 and the Greek text of Artemidoros, see Appendix 1.
<sup>14</sup> The only phrase in the abridgment that at first sight points to the use of additional sources is in Paris. Suppl. gr. 690, fol. 128r, col. 2: ΜΑΡΓΑΡΟΙ· Τινές δε δάκρυα και κλαυθμον είπον, ή φόβον από έξουσίας (Pearls: Some said that [they signify] tears and wailing, or fear of an authority). However, the phrase "τινές δε είπον" ( "some said," which could be a direct translation of the Arabic قـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ dreambooks) must already have existed in the postulated extended version from which the abridgment was made. Gigli ( "Gli onirocritici del cod. Paris. Suppl. gr. 690," p. 84) mentions that the interpretation of pearls as tears occurs in two other Byzantine dreambooks, those attributed to Nikephoros and to Germanos: μαργαρίται δηλούσι δακρύων ροήν (Pearls signify the flowing of tears). For the relevant texts, see Nikephoros, "Das Traumbuch des Patriarchen Nikephoros," ed. F. Drexl, Festgabe Albert Erhard (Bonn and Leipzig, 1922), pp. 94-118, verse 184; and Nikephoros, Pseudo-Nicephoro: Libro dei sogni, ed. G. Guidorizzi (Naples, 1980), verse 68. Also Germanos, "Das Traumbuch des Patriarchen Germanos," ed. F. Drexl, Laographia 7 (1923), pp. 428-48, verse 138. The interpretation of pearls as tears is also found in Islamic dream interpretation: see al-Nabulusi, vol. 2, p. 198; also Appendix 1. For the interpretation of pearls as φόβον άπο έξουσίας (fear of authority), see Drexl 211, 1-9.
<sup>15</sup> See Gigli, "Gli onirocritici del cod. Paris. Suppl. gr. 690," pp. 83-85. Drexl 239, 8 ff: Περί στρουθίων, codd: περι τρωγλητών και διαφόρων στρουθίων. This was rejected by Drexl because only orpovelia are discussed in the subsequent paragraph. However, Paris. Suppl. gr. 690, fol. 129r, col. 1: ΣΤΡΟΥ ΓΛΙΓ [sic] · Οί τρωγλοδύται και τα λοιπά στρουθία ἀτιμίαν ἐκ φίλων δηλούσι. Cf. also Vat. er. 573. fol. 212v: περί στρουγλιτών και διαφόρων στρουθίων (the chapter talks about στρουθία and όρνιθας έναλίους ή λιμνίτας ή άπλώς φιλύδρους).
Second, additional evidence from *Vat. gr.* 573 and the twelfth-century Latin adaptation of Pascalis Romanus also leads to the conclusion that the Greek archetype was longer than Drexl' s critical edition. <sup>16</sup>
The Greek archetype was an adaptation of Arabic material, the foreign character of which left clear traces in some of its passages. The abridgment in *Paris. Suppl. gr.* 690, on the other hand, provides a text that is not only shorter, but also purged of its exotic details, so it has lost much of its foreign flavor. References to concepts that would have been unfamiliar and possibly irrecognizable to a Byzantine reader were eliminated. For example, the epitomist eliminated the numerous allusions to polygamy that can still be found throughout the critical edition; 17 he chose to retain the interpretation of the five fingers as representing acts of piety, though their interpretation as the five daily prayers, an Islamic religious duty, was eliminated. 18 Further on, only three of the planets are given a specific interpretation; the remaining four, which, in the critical
<sup>18</sup>*Paris. Suppl. gr.* 690, fol. 125v, col. 2: oi oaK1UAOt El<; 1a £pya Kpivov.m •ii<; 1tlCT1EW<; wu avOpcimou. (The fingers are interpreted as a person's acts of faith). Cf. Drexl 45, 11 ff: Ai XEtpE<; Kal. oi oaK1UAOt d<; 1a £pya •ii<; 1ttCT1EW<; 1ciiv av0poonwv OtaKpi vovmt ... Kat npciiw<; 6 µ£ya<; 1ciiv OaKTUAWV Kpi VE1at El<; 1i]v Euxiiv WU op0pou, OEU1Epo<; i\wt 6 Atxav6<; Et<; 1i;<; 1phri<; wpa<; 1i]v Euxi]v Kpi VE mt, 1pi *W<;* 6 µfoo<; •ii<; EK1TI<; wpa<;, 6 *an'* auwu i\wt 6 1£mpw<; •ii<; EVcl1TI<; wpa<;, n£µmo<; 6 EAclXtcrW<; •ii<; Ecr1tEpt vii<; EUXii<; (The hands and fingers signify a person's acts of faith .... The thumb signifies the morning prayer; the forefinger, the prayer of the third hour; the middle finger, the prayer of the sixth hour; the ring finger, the prayer of the ninth hour; the little finger, the evening prayer). For further discussion on the interpretation of fingers, see chapter 7.
<sup>16</sup>See also the introductory note to Appendix 2 and the interpretation of eating honey discussed in chapter 5.
<sup>17</sup> There is, however, one exception. In the interpretation of ribs (fol. 126r, col. I) a reference to polygamy was retained, probably because the epitomist failed to understand the meaning of the term yvi]<nm yuvatKE<; (legitimate wives), which makes sense only in the context of Islamic law, with which he was unfamiliar. rvi]<nm yuvaiKE<; must correspond to the women whom it is lawful for one to marry according to Islam, while ai EK y£vou<; cruyyEVEt<; (blood relatives) must be the women that a Muslim is prohibited from marrying because of their degree of kinship to him (see Qur'an 4:23-24). The interpretation of ribs in the abridgment reads as follows: ai itAEupat yuvaiKa<; crrtµaivoucrw, ai µEv iivw µ£y1cr1m yvricria<;, ai OE Ka1w yuvaiKa<; cruyyEvEi<; (The ribs signify women; the upper, larger ribs legitimate wives, and the bottom ribs women who are relatives). Cf. Drexl 55, 5-8: ai nAEupai datv ai yuvatKE<;. ai µEv *iivw,* ai µ£y1cr1m, yvi]<nm yuvmKE<; datv, ai OE Ka1w ai EK y£vou<; Eicrl.v cruyyEvEi<; (The ribs are the women; the upper, larger ribs are the legitimate wives, and the bottom ones are the relatives from the same family). The Arabic dreambooks I have checked all agree that "a rib is a woman, because she was created from a rib" (Ibn Qutayba from Jerusalem, *Yahuda ar.* 196, fol. 30a). The creation of Eve from the rib of Adam as narrated in Genesis 2:21-23 was known to Muslims through religious traditions and theological commentaries; see, e.g., Ibn Qutayba, *Kitab al-Ma'arif* (Cairo, 1935), p. 7. I have been unable to find an Arabic dreambook that differentiates between upper and lower ribs, though I believe that this differentiation was made in the Arabic source of the Greek passage. Artemidoros does not write anything about dreaming of ribs.
edition, correspond to officials not of the Byzantine but of the caliphal court, are given a more general interpretation.19 Paris. Suppl. gr. 690 also omits the name of a "Saracen" sweet that is interpreted as sickness because of its color, leaving the shortened interpretation somewhat obscure.20 Further abbreviated interpretations that occur in Paris. Suppl. gr. 690 are similarly obscured by the omission of details indispensible for understanding the passage."
It is impossible to know when the abridgment was made. The text in Paris. Suppl. gr. 690 was not created ad hoc. A model for it already existed, as is obvious from scribal errors.22 The only secure terminus ante quem for the
20 Paris. Supp. gr. 690, fol. 128r, col. 2: ή βρώσις των φοινίκων άγαθόν έστιν, άλλά και του μέλιτος και του σάχαρ. το δε γλύκισμα νόσον διά το χρώμα (Eating dates, as well as honey and sugar, is a good thing, but a sweetmeat signifies sickness because of its color). Cf. Drexl 198, 3-5: εί δὲ τρώγει γλύκισμα σαρακηνικὸν τὸ λεγόμενον παλουδάκιν, εύρήσει νόσον διά τα χρώματα και το πύρ αναλόγως της βρώσεως (If he is eating the Saracen sweetmeat called paloudakin, he will find sickness commensurate with what he ate because of the colors and the fire). In the abridgment, note that vooov in the accusative is an inconsistency in syntax, probably carelessly carried over from its model.
21 E.g., Paris. Suppl. gr. 690, fol. 128r, col. 1: τα ποδόρτια είς μισθοποίησιν πτωχών κρίνονται, άναλόγως της εύωδίας ή δυσωδίας αύτών (Slippers are interpreted as giving charity to the poor, depending on their pleasant or foul smell). Cf. Drexl 176, 5-8: ἐάν τις ϊδη, ότι περιεβάλετο ποδόρτια νέα εύωδίαν έχοντα, μισθοποιήσει πτωχοις έκ τῆς ὑπάρξεως αύτου, εί δε δυσωδίαν έχοντα, φήμην κακὴν εύρήσει δια το δυσώδες έν τώ έργω του utotov autov (If someone dreams that he was wearing new slippers that had a pleasant smell, he will give charity to the poor out of his own fortune. If they had a foul smell, he will earn a bad reputation on account of his charity because of the foul smell). Also Paris. Suppl. gr. 690, fol. 128v, col. 1: ό ίστος το μεν τούτου συνήθως κέρδος σημαίνει, a phrase that is virtually incomprehensible. Cf. Drexl 215, 9-10: ἐἀν ἴδη τις, ὅτι ὑφαίνει ἐν οϊκῳ αὐτοῦ συνήθως, εύρήσει έν τω έργω αὐτοῦ κέρδος ἀνάλοΥον του ἔργου τῆς ἀνατάσεως (If someone dreams that he was weaving in his house, if he usually does this, he will find profit in his work in proportion to the amount of work stretched out on the loom).
22 Changing a preposition in the first part of a composite verb in Paris. Suppl. gr. 690, fol.
<sup>19</sup> Paris. Suppl. gr. 690, fol. 127r, col. 2: ό ήλιος είς πρόσωπον κρίνεται του βασιλέως, ή σελήνη είς το δεύτερον από του βασιλέως πρόσωπον. Ή 'Αφροδίτη είς το της αύγούστης λογίζεται. Οί λουποί πλανήται είς τούς μεγιστάνας (The sun is interpreted as the person of the king, the moon as the second most important person after the king. Venus is reckoned as the queen. The rest of the planets as noblemen). Cf. Drexl 129, 12-18: ὁ ήλιος είς μέγαν βασιλέα κρίνεται .... Και ή σελήνη, ώς είρηται, είς πρόσωπον του δευτέρου έστιν από του βασιλέως και ή 'Αφροδίτη είς πρόσωπον της Αύγούστης: και ό Ερμής είς πρόσωπον του πρώτου των γραφέων του βασιλέως κρίνεται και ό Άρης είς τον πρώτον πολεμιστήν του βασιλέως και ό Ζεύς είς τον πρώτον του πλούτου και της διοικήσεως του βασιλέως και του χρυσίου και ό Κρόνος είς <τὸν> πρώτον ποιναλιστὴν διακρίνεται (The sun is interpreted as a great king .... The moon, as was mentioned, is the person who is second in authority after the king, and Venus is the queen. Mercury is interpreted as the chief among the scribes of the king, Mars as the chief warrior of the king, Jupiter as the one in charge of the wealth, administra- tion and gold of the king, and Saturn is interpreted as the chief punisher and chastiser). For further discussion on the interpretation of the moon and the stars, see chapter 5.
composition of this model remains 1075-85, the date of *Paris. Suppl. gr.* 690.
*Vat. gr.* 573.23 14th and 15th century, paper, 380 fols., 22 cm x 14 cm., anthology of 31 miscellaneous texts, put together from 7 parts of various manuscripts in 7 different hands.
The *Oneirocriticon* is found on fols. 120r-213v, which belongs to part 5 of the manuscript (fols. 46-214). 24 The watermarks of this part suggest that it was written in the fifteenth century. <sup>25</sup>It is followed by a prayer that was pronounced before asking for a dream to be interpreted, a list of the days and hours most auspicious for dreams, and an anonymous lunar dreambook (fols. 213v-214v).26 The scribe copied the text of the *Oneirocriticon* from a model with numerous lacunae. He therefore left a number of pages blank in his copy, apparently hoping to find a model that preserved the missing passages so he could complete the text later.
The text of the *Oneirocriticon* as it stands in *Vat. gr.* 573 cannot be put into either of the two families of texts identified by Drexl in his critical edition *(x* = *codd. meliores; y* = *codd. deteriores).* It contains readings that indicate its affinity with both families, though examples putting it close to they family are more numerous. 27Vat. *gr.* 573 contains a few interpretations that are absent
<sup>23</sup>This report is based on my examination of fols. l 20r-2 l 4v in microfilm, and on the description of the manuscript by R. Devreesse, *Codices Vaticani Graeci,* vol. 2 [codd. 330-603], (Rome, 1937), pp. 469-77.
24 Ibid., p. 477.
25 The watermarks are discussed ibid.
26 For the prayer and the lunar dreambook in *Vat. gr.* 573, see also the description of the Oxford *Baroccianus* 206 (13th-14th century), fol. 232 in *CCAG,* vol. 9, p. 24. The *incipit* and *desinit* of this section of *Baroccianus* 206 seem to indicate that its text might coincide with that of *Vat. gr.* 573, but I have not examined the *Baroccianus* and cannot be sure.
27 Examples from the first ten pages of the critical edition that show *Vat. gr.* 573 to belong toy: *Vat. gr.* 573 has nunoc; for nuKvoc; (2, 16); EvT)AAayµ£vT) for iouayyioA.tl;;oµ£vT) (2, I); Bapaaµ for Bapaµ (2, 26); PA.Emov for npopA.Emov (5,6); 1:0u A.aou npooiouxioi:m il uniopiouxioi:m for wu A.aou imiopiouxioi:at (9, 13); rrA.ouoiac; for rrpomaoi.ac; (10, 2); <1>aviopw8i]oio1m for <1Jwpa0i]oioi:at
<sup>126</sup>v, col. 2: £av nc; iiop£a lO\J E<J'tOAtoµ£vov £v i:tvt 16nC(l £v c\i OUK £~ioonv iiop£a npo£pxioo0m instead of cio£pxrnem. Cf. Drexl 103, 26-27: £av nc; lO\J Km' ovap, on ci<JiiA.0iov iiopiouc; Eo't0Atoµ£voc; EV 101tC(l, OU OUK ~v 1unoc; cio£pxrn0m aui:ov ... Transposition of words in *Paris. Suppl. gr.* 690, fol. I 28r, col. I: 10 KaµioA.auKtov Tj1:0t ioi.c; 1T]v Kc<IJaA.T]v 'tOV 8c<J1tO/;;ovi:a 'tOU 6pciiv1:0c; avayio1m instead of 10 KaµioAaUKlOV cic; i:T]v Kc<IJaA.T\v Tj1:0t 10V 8rnn61;;ov'ta 'tOU 6pciiv1:0c; avayio1at. Cf. Drexl 168, 15-17: £av lO\J nc;, on <1>opioi KaµcA.aUKlOV ... 'tOU10 cic; 1T]v U7tcp£xouoav au1:0u Kpi. VC'tat Kc<IJaA.T\v. The abridgment must have been in circulation long enough for the word ni.mt v to change into mKpi.av *(Paris. Suppl. gr.* 690, fol. 127r, col. I; see Gigli, "Gli onirocritici de! cod. *Paris. Suppl. gr.* 690," p. 83).
from Drexl's text and can be as long as half a manuscript page. The longest additions are the following:
1. Vat. gr. 573, fols. 162v-163r, to be added after Drexl 105, 11, pu0 Tlepì iepewv (chapter 159, "On Priests"). In Vat. gr. 573 the excerpt is titled Περι ίερέων και μοναχών. ἐκ τῶν Ινδών (On Priests and Monks, from the Indians):
Γυνὴ ἐαν ἴδη ὅτι ἐγένετο ἱερεὺς καὶ συνελειτούργει ἱερεῖ, ταχύ ἀποθανεῖται, και σωθήσεται. ἐαν τις ἴδη ὅτι συνάντησε τινὰ μοναχὸν φορωντα σταυρὸν ύπογεγραμμένον ώς μεγαλόσχημον, εί μεν ώμίλησεν αὐτώ γλυκέα καὶ χρηστά, ίδου κρατείτω ούτως εί δε μὴ, τούναντίον. εἰ δὲ ἔτυψεν αὐτὸν ὁ τοιοῦτος μοναχός, εύρήσει χάριν ό ίδών, είς τὴν ψυχὴν αύτού, παρὰ άγίου. ό γὰρ μοναχὸς ό σεσημειωμένος είς άνθρωπον του θεου κρίνεται. και άπερ τυφθη έξ έκείνου, δωρεά έστι άγίου τινός. Εί δὲ ἔστι μοναχὸς ἐκτὸς σημείου σταυροῦ ὑπογεγραμμένου ή μεγαλοσχήματος μόνου ρασοφόρου, εί μὲν ἔστι γέρων, εἰς καιρὸν κακόν κρίνεται. και εί μεν όμιλει αὐτώ καλώς καὶ εὐτάκτως, ἢ ἐδίδου αὐτω τὶ δόμα, ώς πρὸς τὸ δόμα, μέλλει εὑρεῖν, ἀναλόγως του δόματος καὶ τοῦ πράγματος, μεμετρημένον καλόν. εί δὲ ἔστιν ὁ μοναχὸς νέος, καὶ ἐπολέμει καὶ αὐτὸς μετ' αυτού, ώς τον γέροντα είς νόσημα άνίατον μέλλει έμπεσειν. εί δε έκρουεν ούτος τον μεγαλόσχημον, μέλλει ούτος δώσειν θυσίαν και παράκλησιν προς άγιον τινὰ ό ιδων αναλόγως του δαρμου. εί δὲ ε̃δαιρε τον νεον, ζημιωθήσεται παρ' έχθρού.
If a woman dreams that she became a priest and performed the liturgy together with a priest, she will die and be saved. If someone dreams that he met a monk wearing the sign of the cross like monks of the higher order {do, and] if [the monk | talked to [the dreamer] in sweet and kind words, let the dreamer hold on to them. If not, the opposite. If this monk beat him, the dreamer's soul will receive the grace of a saint, for a monk wearing the cross is interpreted as a man of God. And whatever beating the dreamer suffered from the monk is a spiritual benefit from a saint. If the monk did not wear the sign of the cross, and did not have the long habit of monks of the higher order, but was only wearing his cassock, if he was old, this is interpreted as a bad circumstance. If he talked to the dreamer in a nice and orderly manner, or gave him something, the dreamer will obtain something moderately good, commensurate with the [monk's] gift. If the monk was young and fought with the dreamer, the dreamer will fall ill with an uncurable sickness, as if he were an old man. If the dreamer beat the monk with the long habit of the higher order, he will make an offering and address a request to a saint commensurate with the blows. If he beat the young [monk], he will be harmed by an enemy.
2. Vat. gr. 573, fol. 183v, to be added after Drexl 157, 28, at the end of o' ' EK των Περσών και μόνον περὶ ἀπώρας παντοίας (chapter 200, "From the Persians Alone, on All Kinds of Fruit"):
<sup>(11, 23).</sup> In the same pages there are only three examples that indicate the affinity of Vat. gr. 573 with x: Vat. gr. 573 has βασιλέως for φαραώ (4, 12); αγαπηθήσεται for άγαθοποιηθήσεται (9, 13); εί δὲ οὐκ ἐτελείωσε (11, 7).
Περὶ καστάνων: τὰ κάστανα, εἰς ἡμέρας καὶ νύκτας κρίνονται: ἐαν ἴδῃ τις ὅτι ἔτρωγε κάστανα χλωρά, γινωσκέτω ὅτι ήμέρας καλάς μέλλει διαδραμείνι ἐαν γουν είσι πολλά, έστω ή ζωή αυτού πλείων εί δε μή, όλίγη ει δε ήν ξηρα, ήμέραι κακαι είσι και στεναί. εί δε είδεν τινά ή έδίδου αύτώ, έτεραι ήμέραι ή καλαί ή κακαι είσιν ώς πρός τα κάστανα. ώσαύτως και τα σκωλήκια του μεταξίου, είς ήμέρας κρίνονται. ἐαν ϊδη τις ὅτι ἔλαβεν ἀπὸ τουούτων καρπων είτε αμύγδαλα από βασιλέως, εί μέν είσι χλωρά, είτε μηνας είτε ήμέρας κατα το μέτρον τούτων, μέλλει διαβιβάσειν καιρόν καλόν. εί δέ είσι ξηρά, γινωσκέτω ὅτι μέλλει ἀποθανεῖν ὡς εἴρηται εἰς τοσούτον καιρὸν κατὰ τὸ μέτρον.
On chestnuts: chestnuts are interpreted as days and nights. If someone dreams that he was eating fresh chestnuts, let him know that he is about to go through good days. If they are many, his life will be longer. If they were dry, the days will be bad and dire. If he dreamt of someone or gave him [chestnuts (?)],28 these are other days, either good or bad, depending on the [kind of] chestnuts. Likewise, silk worms are interpreted as days. If someone dreams that he received such fruits or almonds from a king, if they are fresh he will have good times for months or days commensurate to the number [of the nuts]. If they are dry, let him know that he will die, as was said, in a period of time commensurate fwith the number of nuts].
3. Vat. gr. 573, fol. 209v, to be added after Drexl 230, 21, in σπβ' Έκ των Ίνδών, Περσών και Αίγυπτίων περὶ σφηκών καὶ μελισσών ἀγρίων και ήμέρων (chapter 282, "From the Indians, Persians and Egyptians on Wasps and Wild and Domesticated Bees" ):
ἐὰν Ίδη τις ὅτι εὐρεν μέλι μετὰ κηρίου αὐτοῦ, ἐνόλοΥον τοῦ μέλιτος πλούτον. πλήν δέ, μετα βίας όλίγης. εί δε εύρε μέλι και έφαγεν αὐτό, γινωσκέτω ὅτι διασυντόμως θέλει ίδεῖν τὸ μέλλον αὐτοῦ, κὰν τε καλὸν, εἴτε ἐναντίον. εἰ δε ΐδη ότι ευρεν ή ήρεν ή έφερον αύτφ μέλι έψητον ήτοι γλύκισμα μετα άρρωμάτων, εύρήσει θλίψιν και στενοχωρίαν, άναλόγως του γλυκίσματος. εί δε έκτὸς ἀρρωματων ἐστίν, ἐλαττοτέρα ή κρίσις ἐστὶ τῆς θλίψεως καὶ στενοχωρίας, όμοίως ἐαν ϊδη τὶς ὅτι εὐρε κηρία σβεστὰ μικρὰ καὶ ἀπῆρεν αυτά, γινωσκέτω ό τοιούτος, ότι διασυντόμως μέλλει αποθανειν. εί δὲ εἰσὶν άπτόμενα, εύρήσει χαράν και άγαλλίασιν, αναλόγως των κηρίων, πλήν έκ του καπνού. εί δε μὴ, τουναντίον. εί δὲ ϊδη ὅτι ἔσχε λαμπάδας, αί λαμπάδες, εἰς καιρούς κρίνονται, και άρχοντας. εί δε σβεσται είσιν, είς μηνας τόσους, αναλόγως των λαμπάδων, μέλλει αποθανείν.
If someone dreams that he found honey with its honeycomb, he will find wealth commensurate with the honey, though through the use of some violence. If he
<sup>28</sup> There seems to be a lacuna in this part of the Greek text.
found honey and ate it, let him know that he will soon meet with his future, whether good or bad. If he dreamt that he found or took or was brought baked honey, that is, a dessert with spices, he will find sorrow and trouble commensurate with the dessert. If it was without spices, the interpretation of sorrow and trouble is more moderate. Likewise, if someone dreams that he found small, extinguished candles and took them, let such a person know that he will soon die. If they were lit, he will find joy and delight commensurate with the candles. Save for the smoke [sic]. If not, the opposite [?].20 If he sees that he had torches, the torches are interpreted as time and noblemen. If they were extinguished, he will die in as many months as there are torches.
4. Vat. gr. 573, fol. 201r-v, to be added after Drexl 207, 26, in σνβ' 'Eκ των Περσών περὶ ἀσμάτων καὶ ὀρχημάτων, αὐλήσεως καὶ κιθάρας (chapter 252, "From the Persians on Songs and Dances, Playing the Flute and the Cithara"):
ἐαν Ίδη τις ὅτι ἔκρουε βούκινα εἰς αὐλὴν τινὸς, ἤτοι εἰς κάστρον, γινωσκέτω είς όντινα παίζουσι τα τοιαύτα, μέλλει άποθανείν ή κινδυνεύσειν ό βίος αύτου και ή δόξα αυτού, ή και είς φυλακὴν ἐμβληθήσεται. εἰ δὲ εἰς πόλεμον ἐστὶν ή πόλις ἐκείνη, στενοχωρηθήσεται παρ՝ ἐχθρῶν. εἰ δὲ ἔστιν ἀνακαράς, ἀναμφιβόλως, είς θάνατον κρίνεται. όμοίως και είς τὴν πόλιν είς ήν παίζουσιν άνακαράδες, θανατικόν μέλλει γενέσθαι, και άκουσθήσεται. ή δε ζαμάρα και τα αλλα, είς ύβρεις και ατιμίας κρίνονται.
If someone dreams that he was sounding horns in someone's court, that is, in a castle, let him know that the person for whom such instruments were played will die, or his life and glory will be endangered, or he will be thrown into prison. If that city is at war, the enemy will bring it to dire straits. If the instrument is an anakaras, it is undoubtedly interpreted as death. Likewise, in a city where anakarades are played, a plague will take place and will be heard. The zamara" and the other [instruments] are interpreted as offense and dishonesty.
Three additional interpretations are not unique to Vat. gr. 573, but are also found in Paris. Suppl. gr. 690. Right after the paragraph on the interpretation of crabs and other crustaceans (Drexl, chap. 300) and before the last paragraph on the theory of dream interpretation (Drexl, chap. 301), two brief interpretations are inserted: Περὶ ψύλλων καὶ φθειρών (On Lice and Fleas) and περὶ αλατος (On Salt). These interpretations also appear at the very end of the relevant passage in Paris. Suppl. gr. 690 (fol. 129r, col. 1): Αί ψύλλαι είς ἐχθροὺς πτωχούς κρίνονται. Ωσαύτως καὶ οἱ όλίγοι φθεῖρες. Οἱ δὲ
<sup>29</sup> Possible lacuna in the Greek text.
<sup>30</sup> A small drum having a hemispheric body of copper or wood.
<sup>31</sup> A wind instrument consisting of two pipes.
#### 102 CHAPTER THREE
7t0AA0l 7tAOtYtoV 8riA-oucn. To 8£ aA-ai:; 1rn't a'U'tO £l~ 1tAOU'tOV ii8uv EKpteri (Lice are interpreted as poor enemies. A few fleas [are interpreted] likewise, while many fleas indicate wealth. Salt is also interpreted as pleasant wealth). *Vat. gr.* 573 contains none of the other interpretations found in *Paris. Suppl. gr.* 690, but absent from Drexl. It is impossible to know whether the brief interpretations on lice, fleas and salt at the end of the *Oneirocriticon* were added by the scribe of *Vat. gr.* 573 who, as is indicated by the blank spaces left at various places in the manuscript, was looking out for a copy that would allow him to fill in the lacunae in his text, or whether it already existed in his model. But whoever made this addition seems to have had both a long and an abridged version of the *Oneirocriticon* at hand, but not to have compared the two very carefully, probably because he did not expect to find anything new in the abridgment. Possibly, after a brief examination of the two texts, he only spotted the three additional interpretations that appear at the very end of the abridgment under a separate heading. Since finding the other additions to the abridgment requires a thorough and careful examination of the two texts in their entirety, he added to the text that he was copying only those extra interpretations that were easy to spot.
*Bononiensis (Bibliothecae Universitatis)* 3632. 14 December 1440, paper, 475 fols., 29.6 cm x 21.9 cm. A number of epistolary formulae appear on the first six folia; most of the rest of the first half of the manuscript (up to fol. 266) contains an assortment of medical texts. The second half is an anthology of texts pertaining to the occult sciences: astronomy (including a glossary on the Arabic names of the signs of the zodiac on fols. 327v-329), magic, and geomancy.32 The *Oneirocriticon* is on fols. 442r-467r.
This manuscript contains only part of the *Oneirocriticon.* The sequence of chapters differs greatly from the one in Drexl 's critical edition,33 but is the
<sup>32</sup> For descriptions of *Bononiensis (Bibi. Univ.)* 3632, see Chester Charlton McCown, *The Testament of Solomon* (Leipzig, 1922), pp. 21-25; also A. Olivieri and N. Festa, "Indice dei codici greci delle Biblioteche Universitaria e Comunale di Bologna," *Studi Italiani di Fi/o/of?ia Classica* 3 (1895), pp. 442-56; rpt. C. Samberger and D. Raffin, eds., *Catalol?i codicum f?raecorum qui in minoribus bibliothecis ita/icis asservantur in duo volumina collati et novissimis additamentis aucti* (Leipzig, 1965). A description of the contents from fol. 266 onward can be found in *CCAG,* vol. 4, pp. 39-46.
<sup>33</sup> On the index of chapters on fol. 442 is written: a rrpciJA.oyo~ wu crripiµ wu ovriproKpi wu· <sup>~</sup> EK WV flVOWV O'KE'lfl~ KptO'EW~ OVEtpfrrrov· *y* EK WU A6you WV flVOWV 1tEpt Kpt0'£0~ ov£iprov· 8 EK wv riv8rov 7tEpt vE<)>EA.wv· E itEpt avEµrov· ~ itEpt i]A.iou Kat O'EAT]vri~ Kat acr-rEpwv· itEpt ~poxT]~· Tl 1tEpt 1tUAOU Kat OlO'W8ia~· e 8tiiYT\O'T\~ ov£ipmo~· l 8tiiYflO'l~ E'tEpa; ta 1tEpt 0Epµov u8mwv· t~ 7tEpt rro-rripiou· ty itEpt rrA.i]0ou· t8 itEpt avacrmcrEw~· tE itEpt rrapa8icrou·
same as that in *Paris. gr.* 2419, which allows us to conclude that the two manuscripts belong to the same branch of the manuscript tradition.
The *Bononiensis* is the only illustrated manuscript among the Greek manuscripts of the *Oneirocriticon* that I have examined. The portraits of the four dream interpreters mentioned in the text (Syrbacham, Baram, Tarphan and Sereim), each labeled with his name, appear on fol. 443v, between the table of contents and the beginning of the text. Other parts of the manuscript containing different texts are also illustrated with analogous portraits of authors. The organization of the pictorial composition on fol. 443v and a comparison with the other portraits in the manuscript indicate that this illustration was not conceived to accompany the text of the *Oneirocriticon;* rather, it was borrowed from the illustrations of some other text, and the figures were relabeled to make it relevant to the context of the *Oneirocriticon.*
British Library *Additicius* 8240. 34 17th century, paper, 149 fols., 15 x 9 cm., written in various hands. The manuscript contains an assortment of dreambooks. It was put together from a number of different manuscripts, and contains a collection of various texts. It once belonged (as no. 421) to Frederic North, fifth Earl of Guilford ( 1766-1827), 35 and before him to the eighteenth-century Phanariot intellectual Nikolaos Karatzas. 36
The volume, as it is bound today, contains a wide variety of texts ranging from a description of the Holy Land combined with a number of *gnomologia,* to a treatise on meter in ancient Greek poetry, and an exposition of the doctrinal differences between the Orthodox and the Latins. It also contains three texts
36 See G. Papazoglou, "Cheirographa tou Nikolaou Karatza eis ten bibliotheken tou Brettanikou Mouseiou (allote cheirographa Guilford)," *EEBS* 49 (1994-98), pp. 247-48.
u; rrEpi yEi:v11<;· ti; rrEpt avyyi:/...wv· tT\ rrEpt rrpo<)>111c0v arrocrw/...wv Kat otoacrKa/...wv· 18 rrEpt rricr1mw<;· K rrEpt' KPT\WV, etc.
<sup>34</sup> The most comprehensive description of the manuscript and its contents can be found in *CCAG,* vol. 9:2, pp. 28-29. Its description in *List of Additions Made to the Collections in the British Museum in the Year 1831* (London, 1833), p. 6, is inadequate. My information is either found in the relevant manuscript catalogues or is based on my examination of the manuscript.
<sup>35</sup> Frederick North, fifth Earl of Guilford, was a well known philhellene and chancellor of the Ionian University of Corfu. Upon his death in 1827 his various collections, including his collection of manuscripts, were bequeathed to the Ionian University, but were recovered by his executors, because the university failed to comply with certain conditions set down in the bequest. Some of the Guilford manuscripts ended up in the British Museum as *Add.* 8240, 20016-17, 20036-7, 27430-1; see *Dictionary of National Biography,* vol. 14, p. 610. On the manuscripts owned by Frederic North, see also B. Mpompou-Stamate, "He bibliotheke Guilford sten Kerkyra (1824-1830)," *Ho Eranistes* 20 (1995), pp. 97-162; and M. Paize-Anagnostopoulou, "He bibliotheke tou N. Karatza kai to cheirographo *Paris. Suppl. gr.* 1375," *Hellenikn* 49:1 (1999), pp. 69-70.
#### CHAPTER THREE
on dream interpretation,37 one of which, on fols. 124v-126v, is an excerpt from the Oneirocriticon (fol. 124v inc.: Έκ τών Περσών και Αἰγυπτίων περὶ διαφόρων εἰδών. Έαν ἴδη τις ὅτι ἀπεπλύθη καὶ ἔβαλεν ίμάτια ... ; fol. 126v des .: τέξει θυγατέρα). This excerpt corresponds to Drexl 203, 27-205, 26 (chapter 247).38 Compared with the text in Drexl's critical edition, the excerpt in Additicius 8240 presents frequent variant readings. These variants link BL Additicius 8240 to Rigault's edition of 1603 and its source, Paris. gr. 2538.
The excerpt from the Oneirocriticon is immediately followed by two sentences that do not appear in Drexl's edition and seem to belong to another text:
της αγάπης ϊδιον φησί των τα κατ΄ αυτήν ζηλούντων, μίαν αποδείξαι γνώμην.των νοητών διαβάς, τὴν ούσίαν ὁ νούς ἀνόητος γίνεται, θεώ προσβάλλων ὑπέρ πασαν δτι κατ΄ ούσίαν, και γνώσιν, και νόησιν, δια την ήδονὴν ἀγαπώμεν τα πάθη, ώς δια τὴν όδύνην, τὴν ἀρετὴν, φεύγωμεν [sic].
He says that a characteristic of love is that those who strive for it produce a uniform opinion.
The mind, after going beyond the essence of mental things, becomes mindless with regard to God, applying above all <essence> that as far as essence, knowledge and understanding are concerned, we are fond of passions because of pleasure, as we avoid virtue because of pain.
It has been impossible to locate these sentences in any of the texts available through the electronic database of the Thesaurus Linguae Graecae.39 They seem to belong to a philosophical or patristic work. A possible explanation for copying these lines as if they also belonged to the Oneirocriticon is that in one of the ancestors of BL Additicius 8240, the excerpts from the Oneirocriticon must have appeared just before the text to which these two sentences belong.40 The end of the excerpt from the Oneirocriticon as it appears in BL Additicius 8240 must have come at the end of a verso page, and the two unidentified
<sup>37</sup> Besides the excerpt from the Oneirocriticon, the texts on dream interpretation from BL Additicius 8240 are: fols. 31v (olim 347)-32v: Διάλεξις, ήτοι έρμηνεία περί των της σελήνης; fol. 120 (οlίm 49)-124r: Νικηφόρου πατριάρχου Κωνσταντινουπόλεως όνειροκριτικόν κατά άλφάβητον δια στίχων.
<sup>38</sup> The text in BL Additicius 8240 presents slight variations from Drex1's text. For example, its incipit in Drexl reads: ' Eav iδη τις ότι απεπλύθη και έβαλε νέα ίμάτια ...
<sup>39</sup> Thesaurus Linguae Graecae [computer file]. CD ROM no. D. (Irvine, Calif., 1992).
<sup>40</sup> The Oneirocriticon (or excerpts from it) were copied together with patristic and philosophical works in Paris. Suppl. gr. 690, Vat. gr. 573, Vindob. philos. et philol. gr. 162, Athos, Iviron 4285.165.
sentences copied at the very beginning of a recto page. At a later time a certain number of folia containing the end of the *Oneirocriticon* (or its excerpts) and the beginning of the unidentified text, including its title, must have become separated from the manuscript. A scribe copying from this ancestor of BL *Additicius* 8240 must have joined the two excerpts without noticing that some folia were missing and under the misapprehension that both belonged to the same text, since they ran from one page to the next with no new title in between.
Among the interpretations of a lance found in Drexl 205, 8-9 and Drexl 205, 9-10, a phrase that looks like a marginal note which crept into the main text has been inserted. This phrase is placed in the manuscript within parentheses and reads as follows: aviip E1tt<)>8ovo~ µ£ya lWKOV i::i8£ *[sic].* Kat yci'l:COV *[sic]* £lvm KEKAi]pCO'l:at, anapat'l:T]WV 1:0 OUCi'l:UXT]µa, Kat 8av<h(\) µ6A.t~ naucr6µi::vov (An enviable man although [ti 8£] /dreamt of [doi::] a great evil and ... allotted the misfortune to be inevitable and would hardly stop by death).41
*Petropolitanus Bihl. Acad. scient. graec.* 161 (olim Instituti Archaeolog. Constantinopol.). 1731-1740, paper, 41fols.,14.5 cm x 10.2 cm. Various texts, most of them on divination.
This manuscript possibly preserves excerpts from a vernacular paraphrase of the *Oneirocriticon* in fols. 35-39v. According to the description of the manuscript in the *Catalog us Codicum Astrologorum Graecorum,* 42 these folia contain a dreambook titled' EK wu ovi::tpoKpi wu wu Tapa<)>av wu 43 ao<)>ou (From the dreambook of Taraphan the Wise). Inc.:. Eav l01J~ nva~ onco~ avaa'!:T]8ij VEKpo~ - des.[fol. 39]: µna a<jlEnia~ KlVTJO'El Katµ£ Ota<jlopav. T£A.o~. I have not been able to examine this manuscript.
*Zagara* (Bibliotheke Zagoras, Thessaly) 89. 44 1594, paper, 245 fols., 19 cm x
<sup>41</sup>The passage is so corrupt that an accurate translation is impossible.
<sup>42</sup>*CCAG,* vol. 12, p. 48. The manuscript is also described in I. N. Lebedeva, *Opisanie rukopisnogo otdela hihliotheki Akademii nauk SSSR,* vol. 5 (Leningrad, 1973), p. 153, but there the text is only designated as "interpretation of dreams."
<sup>43</sup> Tapa<1>av1:0u' in *CCAG,* vol. 12, p. 48. Tap<1>av is the name of the Egyptian dream interpreter mentioned in the *Oneirocriticon* (Drexl 3, 13).
<sup>44</sup>Described in K. I. Dyobouniotes, "Katalogos ton kOdikon tes bibliothekes tes Zagoras," *Neos Hel/enomnemon* 13 (1916), pp. 451-52; rpt. as a separate volume (Athens, 1920); it is projected that it will shortly be reprinted again in J. Declerek, J. Naret, C. de Vocht, eds., *Catalogi manuscriptorum graecorum qui in periodico "Neos Hellenomnemon" olim puh/ici iuris facti,*
13 cm. Oneirocriticon (fols. 4r-154v), Physiologos (fols. 155r-237v), collection of oracles (fols. 238r ff.). Incipit (fol. 4r): βιβλίον όνειροκριτικόν δπερ συνήξε και συνέταξε αχα΄. έτει ό Έσειρήμ ό όνειροκρίτης του πρωτοσυμβούλου Μαμούν.
The date axa (1601) obviously resulted from a misreading of the name Axuèr.40 A note on fol. 154r informs the reader that the text was copied in the year 6875 (of the world era = 1367 of the Christian era): Έγράφη το παρόν βιβλιαρίδιον το καλούμενον Ονειροκριτικόν κατά το έν Θηβύλαις και έδόθη τω όσιωτάτω ἐν μοναχοῖς κυρίω Γρηγορίω τω Χρυσοχόου, τω ἐκ Τρίκκης τῆς Θεσσαλίας ἐν τῷ ζωοε΄. Οἱ ἀναγινώσκοντες εὐχεσθαί μοι έν Κυρίω τις γράψαντι (The present booklet called Oneirocriticon was copied according to [the copy] in Thebylai46 and was given to the most devout among monks master Gregory Chrysochoou from Trikke in Thessaly in the year 6875. You who read it pray to the Lord for me who wrote it).
The manuscript also contains a second text. It is a vernacular version of the Physiologos, with the title (fol. 155r): Φυσιολογία νέα τὴν όποίαν τὴν ἔκαμεν ὁ Δαμασκηνὸς ἀπὸ τῶν παλαιῶν φιλοσόφων τὰ βιβλία καὶ ἔχει των χερσαίων ζώων και τών θαλασσίων είς φράσιν κοινήν (New Physiologos which was composed by the Damascene from the books of the old philosophers; it contains [information] on the land and sea animals expressed in the vernacular). The Physiologos is accompanied by a note on fol. 241r that gives the year 7102 of the world era (1594 of the Christian era): Και ἐγράφη το παρόν έν έτει ζρβ΄ (The present was written in the year 7102).
The manuscript is all written in the same hand. The note giving the year 1367 must have been copied from the model used by the scribe of the Zagora manuscript, as was the poem that precedes it (fols. 153v-154r); this poem, which gives biographical details on the scribe of the model, is designated as στίχοι ούς ευρον έν τφ ανθιβολαίω του όνειροκρίτου (Verses that I found in the model of the dreambook).47 Zagora 89 belonged to the eighteenthcentury patriarch of Constantinople Kallinikos III, who donated it to the school
adhuc usui sunt (Brussels, 1981--). For the most recent developments in the collection of the library of Zagora, see J.-M. Olivier, Repertoire des catalogues et de bibliothèques de manuscrits grecs de Marcel Richard (Turnhout, 1995), p. 851.
<sup>45</sup> I assume that the scribe of Zagora 89, who copied the manuscript in the year 1594, understood the date 1601 to be of the world era, that is 3906 в.с.( !).
<sup>46</sup> Thebylai (Little Thebes?) is not known to me from anywhere else.
<sup>47 &</sup>quot;Verses that I found in the model of the dreambook."
of the town, according to a note on fol. 4r. 48 The dream symbols in the manuscript are presented in loosely alphabetical order, which still reflects the original thematic arrangement. For example, the sequence of the dream symbols that begin with "a" is avcicrmcrn;, ayyEAot, aATJ<i>Tt *[sic],* ava8£v8pa~, aifoia, atµa Kat EAKO~, avEµot, acm.~pE~, aKav8a, aµ<1>iam~, aµa~a etc. I know of no other manuscript of the *Oneirocriticon* where the dream symbols have been rearranged in alphabetical order.49
*Marc. gr.* 299. 10th century, parchment, 196 fols., 30.2 cm x 24 cm, collection of texts on the occult sciences, mainly alchemy. 50 This manuscript is the oldest known surviving Greek alchemical codex and was part of Cardinal Bessarion's bequest to the Bibliotheca Marciana *(olim* 440).
The current binding includes at the beginning of the volume a number of flyleaves covered with writing by later hands. On one of these flyleaves (currently numbered Sr) one can read, in a fourteenth- or fifteenth-century handwriting, the introductory chapter from the abridgment surviving in *Paris. Suppl. gr.* 690, dedicated to the theoretical principles of dream interpretation. This text corresponds approximately to part of chapter 301 in the critical edition of the *Oneirocriticon* (Drexl 240, 8-241, 14 ).
The manuscripts known to Drexl are presented here in the order in which he discussed them in the introduction to his critical edition; the first eight manuscripts were those he actually used and have a place in his stemma. The purpose of presenting them is to give a schematic picture of the problems inherent in the fragmentary character of the Greek manuscript tradition and to provide a
<sup>48</sup> On the interest of Kallinikos in the school of Zagara, see A. Chrysoberges, *Ho patriarches Kanstantinoupoleos Kallinikos ho tritos ho ek Zagoras ( 1757)* (Zagara, 1995), esp. pp. 131 ff.
<sup>49</sup> I would like to thank Mr. Theodoros Roussis, the librarian of the Zagara Public Library, for allowing me to examine the manuscript. Microfilms of manuscripts from peripheral Greek libraries are kept in the National Library in Athens; however, no reproduction of *Zagara* 89 exists there.
<sup>50</sup> The most recent description of the manuscript can be found in E. Mioni, *Bihliothecae Divi Marci Venetiarum codices graeci manuscripti,* vol. I: *Thesaurus Antiquus, Codices 1-299* (Rome, 1981), pp. 427-33. For an older but more detailed description of the manuscript by 0. Lagercrantz, see *CMAG,* vol. 2, pp. 1-22. The manuscript has been dated to the end of the 11th century by Mioni, to the 11th century by Lagercrantz-Zanetti, the 12th century by Morelli, and the I Oth-11 th century by Berthelot (for references to the corresponding publications, see Mioni, *Bihliothecae Divi Marci,* p. 427). I am following the date given by H. D. Saffrey in R. Halleux, *Les a/chimistes grecs 1: Papyrus de Leyde, Papyrus de Stockholm. Recettes* (Paris, 1981 ), p. xiv. After an examination of a microfilm of *Marc. gr.* 299, Professor Boris Fonkic was also in favor of a 10th-century date. I would like to thank Professor Fonkic for his advice.
quick reference to the Greek manuscripts. I added my own comments on the manuscripts that I have examined. More information on each manuscript can be found in Drexl's dissertation, as well as in the relevant library catalogues.51
*Vindob. philos. et philol. gr.* 111 ( = R). 52 13th century, parchment. Contains only the *Oneirocriticon.* Compared to Drexl's edition, the text is considerably abbreviated. From chapter 30 and on, there are several lacunae, which can be as long as one, two, or three pages.
*Vindob. philos. et philol. gr.* 162 ( = S). End of the 14th century, paper. Contains prophecies and excerpts from Anastasios Sinaites, Maximos Homologetes, Ioannes Pediasimos, etc. Some chapters, especially the long ones, have lacunae.
*Vindob. philos. et philol. gr.* 287 ( = T). 53 End of the 14th century and 15th-16th centuries, paper. Collection of astrological texts. Contains only parts of the *Oneirocriticon.*
*Vindob. philos. et philol. gr.* 297 ( = V). 16th century, paper. Contains only the *Oneirocriticon.* Possibly copied by' Icoavvrii; tEpEui; 6 EuA.orriµ£voi; 6 [o]\xovoµoi; ap-rrii; . 54 In 1562 this manuscript was acquired by J. Sambucus.55 He made it available to J. Loewenklau, who based the 1577 translation of the *Oneirocriticon,* the *Apomasaris apotelesmata,* upon it. The manuscript lacks the first folia, where the title and possibly the name of the author would have appeared, but the name Apomasaros was added by a later hand right after the interpretations of the first dream symbol, the resurrection of the dead, according to the Indians, Persians and Egyptians. 56 This note on the manuscript must have been the source for the title of the translation.
*Leidens. Voss.* 49 ( = L). 57 End of 15th century, paper. Contains only the *Oneirocriticon;* has several lacunae, the most extensive being one page long.
<sup>51</sup> See Drexl, *Achmets Traumbuch.* References to library catalogues can be found in this publication, as well as in the prolegomena to the critical edition of the *Oneirocriticon.* My references are limited to library catalogues that superseded the ones mentioned by Drexl.
<sup>52</sup> The four Viennese manuscripts are described in Hunger, *Katalog der griechischen Handschriften,* vol. 1, pp. 222, 265, 386, 392.
<sup>53</sup> For a description of this manuscript, see also *CCAG,* vol. 6, pp. 51-53.
<sup>54</sup> See Drexl, *Achmets Traumbuch,* p. 14.
<sup>55</sup> Bought in Salerno for eleven ducats. See Hunger, *Katalog der griechischen Handschriften,* vol. 1, p. 392.
<sup>56</sup> Drexl, *Achmets Traumbuch,* p. 5
<sup>57</sup> Described in K. A. de Meyier, *Codices Manuscripti 6: Codices Vossiani graeci et misce/lanei* (Leiden, 1955), p. 157.
The last folio substitutes chapters with interpretations lifted from the dreambook attributed to Manuel Paleologos. Compared with the edition of this text by Delatte according to its version in *Paris. gr.* 2419 (where it is also copied immediately after a version of the *Oneirocriticon),* the excerpts in *Leidensis* are very few and appear in no particular order.
*Paris. gr.* 2511 ( = P).58 End of 14th century, paper. Anthology of texts. Contains only parts of the *Oneirocriticon.* The text is partly a paraphrase.
*Ambros. gr.* 592 ( = A). 15th century, paper. Anthology of various texts. Contains part of the *Oneirocriticon* (about 27 pages).
*Bero/. gr.* 171 ( = B). 16th century, paper. Contains only the *Oneirocriticon.* The text in this manuscript is very close to *Paris. gr.* 2538, used by Rigault for the *editio princeps.* The two manuscripts have the same lacunae, but the Berlin manuscript is not a direct copy of *Paris. gr.* 2538.59
These manuscripts were arranged by Drexl into the following stemma:
<sup>58</sup> Described in detail in *CCAG,* vol. 8:4, pp. 70-72.
<sup>59</sup> F. Drexl, "Die Berliner Achmethandschrift," *BZ* 24 (1924), pp. 307-12.
The manuscripts that Drexl knew of, but did not examine are:
*Paris. gr.* 2538. 16th (15th?) century, paper. Contains only the *Oneirocriticon,* copied by Georgios Grammatikos. 60 Rigault based the 1603 edition on this manuscript. The text has several lacunae.<sup>61</sup>
*Paris. gr.* 2427. 16th century, paper. Contains only the *Oneirocriticon,* chaps. 1-70. Attributes the text to Apomasaros, because it is definitely a direct copy of *Vindob. philos. et philol. gr.* 297 (the source for Loewenklau's translation); *Paris. gr.* 2427 has the exact same lacunae as *Vindob. philos. et philol. gr.* 297 and even cuts the missing words in the middle in exactly the same way as they appear in the Vienna manuscript. It has several blank pages at the end, probably intended for copying the complete text from the Vienna manuscript, an intention that was never carried out.
*Paris. gr.* 2419. 62 15th century, paper. Collection of astrological texts copied by Georgios Midiates,63 which contains part of the *Oneirocriticon.*
*Borbon. gr.* 356 (III.E.34). 64 15th century, paper. Contains only the *Oneirocriticon.* The first folia are missing. The text begins with a chapter numbered 15 and entitled EK 'teov CTepcrffiv Kett Ai yuJt'ttrov 7t£pt 'tp1xrov Kett uJt~vTJi; Kett µetcrxetA.ffiv Kett +\~TJi;. Dept yeveiwv (From the Persians and Egyptians on Hair and Mustaches and Armpit <Hair> and Pubic <Hair>. On Beards) which corresponds to Drexl's chapter 42 (27, 11 ff.). The manuscript contains a complete text of the *Oneirocriticon,65* though several phrases or paragraphs
63 See Hunger, *Repertorium der griechischen Kopisten 800-1600,* vol. 2, no. 87 (with bibliography).
64 Described in S. Cirillo, *Codices graeci manuscripti regiae hihliothecae Borhonicae,* vol. 2 (Naples, 1832), pp. 469-70. The newer catalogues by G, Pierleoni, *Catalogus codicum graecorum Bihliothecae Nationalis Neapolitanae* (Rome, 1962), and E. Mioni, *Catalogus codicum graecorum Bihliothecae National is Neapolitanae,* vol. I: I (Rome, 1992), cover only manuscripts II.A. I to II.C.38 and do not include a description of this manuscript.
65 According to Drexl, *Achmetis Oneirocriticon.* p. xiii, the manuscript contains only chapters 15-192, but this is a mistake; the error was generated by the different method of numbering the chapters followed in *Borbon, gr.* 356, where the same subject when treated according to the Indians, Persians, and Egyptians is only numbered once, whereas in the critical edition it can be
<sup>60</sup> See H. Hunger, *Repertorium der griechischen Kopisten 800-1600,* vol. 2 (Vienna, 1989), no. 104.
<sup>61</sup>See Ruelle, "La clef des songes d'Achmet Abou-Mazar," pp. 305-12, who identified the manuscript as Rigault's source and listed some of the lacunae, Beyond the description of this manuscript given by Drexl, further descriptions have been published by H. Lebegue, *CMAG,* vol. I, p. 62; by F, Cumont, *CCAG,* vol. 8: I, p. 20; see also A. Delatte, *Anecdota Atheniensia,* vol. I, p. 446; and idem, "Le lexique de botanique de *Paris. gr.* 2419," *Serta /eodensia* (Liege-Paris, 1930), p. 59.
<sup>62</sup> Detailed description in *CCAG,* vol. 8: I, pp. 20-63.
are missing from various chapters. Moreover, the sequence of chapters does not coincide with either Drexl or Rigault or *Paris. gr.* 2419 and *Bononiensis (Bibi. Univ.)* 3632. Finally, the contents of the chapters themselves are also different from Drexl's text. <sup>66</sup>
*Cantabrig. gr.* 1386 (0 8.11, 6102). 15th-16th century, paper. The greatest portion of the manuscript is taken up by Actuarius, *De Urinis.* The only parts of the *Oneirocriticon* are an index of chapters (incomplete) and part of chapter 1.
*Athas, Iv iron* 4285 .165. 15th century, paper. Anthology of nineteen texts belonging to various authors and literary genres. Contains only five chapters of the *Oneirocriticon* (= Drexl, chapters 221-26).
*Hierosol.* (of the patriarchate) *gr.* 220. 17th century, paper. Collection of texts on divination. Vernacular paraphrase of the *Oneirocriticon.*
*Hierosol.* (of St. Sabbas) *gr.* 555. Dated 26 April 1685. Paper. Most of the manuscript is taken up by a vernacular paraphrase of the *Oneirocriticon,* accompanied by three very brief ~povwA.Oyta.<sup>67</sup>
#### *Translations*
Among the several translations of the *Oneirocriticon* that were made at different times and into different languages, two medieval renderings from Greek into Latin are the most important for textual criticism. They were both made during
numbered as many as three times. The concluding chapter in *Borbon. gr.* 356 is numbered 193, but its contents correspond to Dre xi's chapter 20 l.
<sup>66</sup> To give an illustration of the problem, the contents of the first folia in *Borbon. gr.* 356 and their corresponding passages from Drexl' s edition are: [ tE] EK i:ciiv itEpaciiv Kat ai yuni:iwv 7tEpt i:ptxciiv Kat uni\vric; Kat µaaxaA.ciiv Kat i\13ric;. CTcpt yEvEiwv (= Drexl 27, 11-27, 25 = µ[3. Despite the title in *Borbon. gr.* 356, nothing on beards is included); t<; itEpt i:ptxciiv 1:0u AOl7tOU aroµmoc; (= Drexl µy); EK 't:OOV 7t£paciiv Kat aiyuni:iwv 6µ6<1>wva 7tEpt i:ptxciiv 't:OU AOl7tOU aroµmoc; (= Drexl µ8); 7tEpt i:ptxciiv yuvmKO<; (= Drexl 217, 8 ff.= a~E); lT\ EK 't:OOV ivociiv 7tEpt aA.n<l>ii<; (= Drexl 18, 5-15 =Ky); EK 't:OOV 7tEpaciiv 7t£pt clAEllj/ECll<; (= Drexl 18, 16-22 = K); EK 't:OOV aiyuni:iwv oµoiwc; (= Drexl KE); EpcO'tf\at<; (= Drexl 15, 17-24 = t0); EpcO'tf\at<; &Uri ( = Drexl 16, 1-10 = K, though narrated in the first person); t0 EK 't:OOV i vociiv itEpt µoaxou auv0Ei:ou £1'.ioaµiac; (includes Drexl K<; and Ks under the same title); K itEpt 0uµtmripiwv Kat KaitviaECllV ( = Drexl 20, 1-8 = KT\); oµoiwc; ( = Drexl 20, 9-19 = A.); Ka 7tEpt Koupa<; (= Drexl 20, 20-26 = A.); EK i:ciiv itEpaciiv itEpt Koupac; (= Drexl 21, 1-8 =A.a, to which is added Drexl 16, 21-24, to which is added Drexl 17, 16-19); K[3 EK i:ciiv ivociiv itEpt yEvnaooc; (= Drexl 22, 25-23 = A.Ii, to which is added Drexl 16, 19-21 and 16, 24-17, 2); itEpt i:ptxciiv (= Drexl 23, 10-16 =A.£).
<sup>67</sup> Neither of the two Jerusalem manuscripts is included in K. W. Clark, *Checklist of Manuscripts in the Libraries of the Greek and Armenian Patriarchates of Jerusalem Microfilmed for the Library of Congress, 1949-50* (Washington, D.C., 1953).
the reign of Manuel Komnenos (1143-80), an emperor with a very keen interest in the occult sciences. The two translations were made almost simultaneously and can be dated with accuracy.
The earliest Latin translation is not complete. Excerpts of the *Oneirocriticon* were translated by Pascalis Romanus and incorporated into his book *Liber Thesauri Occulti,* a work on dream interpretation that, according to its title, was published in Constantinople in 1165.68 The only information we have on Pascalis Romanus is furnished by his own prefaces to the four works of his that have survived. All four are translations or adaptations of Greek texts. <sup>69</sup> Through these prefaces we can verify his presence in Constantinople from 1158 to 1169. It is not known whether he was a Greek born in Rome or a Roman who became acquainted with the Greek world late in life or a Roman born and raised among Greeks.70 Most recently T. Ricklin has argued that he had connections with the Venetian society of Constantinople.71 He was a lowranking clergyman and had some knowledge of medicine, though not enough to qualify as a proper physician. <sup>72</sup>
The *Liber Thesauri Occulti* is divided into three books. The first book discusses the nature of sleep and dreams based on ancient and medieval sources of a scientific, philosophical and literary character. 73 It also contains some infrequent and unacknowledged borrowings from the *Oneirocriticon.74* The remaining two books are a *clef des songes* entirely composed of passages from Arternidoros and the *Oneirocriticon* translated from Greek into Latin. 75 The interpretations are arranged thematically, inspired by, but not exactly copying, the organization of the material found in the two Greek works. The arrangement of the interpretations in the *Liber Thesauri Occultiwas* devised by Pascalis Romanus himself. <sup>76</sup>
<sup>68</sup> Edited with an introduction by Collin-Roset, "Le *Liber Thesauri Occulti."* For a detailed discussion of the work and its sources (mainly concentrating on its philosophical aspects concerning the nature of dreams) in connection with 12th-century Byzantine philosophical pursuits, see Ricklin, *Der Traum der Philosophie,* pp. 247-322.
<sup>69</sup> Their titles are *Disputatio Judeorum contra sanctum Anastasium, Liber thesauri occulti, Cyranides, Ystoria Beate Virginis Mariae.* For more on these works, see Collin-Roset, "Le *Liber Thesauri Occulti,"* pp. 113-17; and Ricklin, *Der Traum der Philosophie,* pp. 248-50.
<sup>7</sup>° Collin-Roset, "Le *Liber Thesauri Occulti,"* p. 112.
<sup>71</sup> Ricklin, *Der Traum der Philosophie,* p. 322.
<sup>72</sup> Collin-Roset, "Le *Liber Thesauri Occulti,"* pp. 112-13.
<sup>73</sup> Ibid., pp. 125-31.
<sup>74</sup> Ibid., p. 130.
<sup>75</sup> Ibid., p. 131.
<sup>76</sup> Ibid., pp. 138-139.
The *Liber Thesauri Occulti* is from Pascalis Romanus's own pen up to book III, chapter 2. The remainder (book III, chapters 3-15) reproduces word for word the Latin translation of the *Oneirocriticon* by Leo Tuscus, which was finished in 1176.77 The editor of the *Liber Thesauri Occulti,* Simone Collin-Roset, concludes: "We therefore definitively accept that the dreambook by Pascalis Romanus remained unfinished; that he abandoned it in 1165, or a little later, in order to translate the Cyranides ... ; and that at least ten years later, a scribe attempted to supplement it by attaching to it [the text of] Achmet translated by Leo Tuscus. "78
Collin-Roset carefully noted the source of each interpretation given in books II and III of the *Liber Thesauri Occulti* and quoted the corresponding passage in either Artemidoros or Achmet. For that purpose, however, she did not use the Greek text of either work, but only their Latin translations.79 Out of hundreds of interpretations she lists eighteen which she was unable to identify in the works of Artemidoros and Achmet. Collin-Roset then proceeded to compare the eighteen additional interpretations of Pascalis Romanus with the only French translation of an Arabic dreambook available to her, Pierre Vattier's seventeenth-century translation of a twelfth-century author, 'Abd al-RaJ:iman b. Na~r al-Shiraz! (or al-Shayzari) al-Nabarawi. 80 She managed to match two of the eighteen interpretations with passages from Vattier's translation. Remarking that nothing in the works of Pascalis Romanus shows that he knew Arabic, she concluded that he possibly used Greek translations of Arabic treatises that have not survived, or he found some of these interpretations in the popular
<sup>77</sup> Ibid., p. 135.
<sup>78</sup> Ibid., p. 139.
<sup>79 &</sup>quot; ••• nous avons cite Artemidore d 'a pres la traduction de Cornarius reproduite par N. Rigault, et Achmet d'apres celle de Leo Tuscus contenue dans le manuscrit d'Oxford [Bodi. Digby 103] ou !'on trouve egalement le *Liher thesauri occu!ti* et, en cas de necessite celui du manuscrit 337 de Carpentras. En cas de lacune de la traduction de Leo Tuscus, nous avons eu recours a celle de J. Leunlcavius, reproduite et completee par N. Rigault" (Collin-Roset, ibid., p. 125). A direct reference to Drexl's critical edition is only given once (p. 140). The combined texts of Leo Tuscus, Leunclavius, and Rigault are still different from Drexl's, which explains why Collin-Roset was unable to identify a few passages.
<sup>80</sup>*L' onirocrite mussulman, ou la doctrine et !'interpretation des songes selon /es arabes, par Gahdorrhachaman fils de Nasar. De la traduction de P. Vattier* ... *sur le manuscrit arahe du "recueil de ce qui se peut dire de meilleur sur !'interpretation des songes* (Paris, 1664). This book is extremely rare today. I have been unable to locate a copy in any American library. There is one in the British Library and another in the Bibliotheque Nationale; see also A. Fischer, "Die Quitte als Vorzeichen bei Persern und Arabern und das Traumbuch des 'Abd-al-~ani an-Nabulusi," pp. 305-6. The Arab author was a physician to \$alaJ:i al-Din al-Ayyiibi; see Fahd, *La divination arabe,* pp. 354-55, no. 116.
tradition of the Mediterranean, which was likely to have been influenced, even to a limited degree, by its equivalent Arabic tradition. <sup>81</sup>
A direct comparison of the additional interpretations in Pascalis Romanus listed by Collin-Roset with the Greek texts in their most recent critical editions reveals that eight of the eighteen, including the two matched by passages in the dreambook of 'Abd al-Raqman b. Na~r, can be found in either the *Oneirocriticon* or Artemidoros. 82 A ninth interpretation corresponds to a passage from *Vat. gr.* 573, which is missing from the critical edition of the *Oneirocriticon.* Where did the remaining nine come from? At least the ones that do not represent Pascalis Romanus's own understanding of passages from Artemidoros83 were probably in the version of the *Oneirocriticon* that Pascalis Romanus used for the compilation of the *Liber Thesauri Occulti,* which must have been longer than Drexl' s text.
Our examination of the abridgment from *Paris. Suppl. gr.* 690, and the text found in *Vat. gr.* 573, already furnished proof for the existence of a version of the *Oneirocriticon* which was longer than the text published by Drexl. Examination of the *Liber Thesauri Occulti* corroborates this conclusion by showing that the model of Pascalis Romanus was, at least in part, closer to the Paris abridgment than Drexl's text is. The introductory chapter to book II of the *Liber Thesauri Occulti,which* immediately precedes the interpretations, is very close to the introductory chapter of the Paris abridgment that includes remarks on the theory of dream interpretation found in the last chapter of Drexl's edition. Moreover, two passages of the Paris abridgment identified by Gigli as missing from Drexl's text can be found in the introduction to book II by Pascalis Romanus.84 Therefore, both the order of presentation and passages from the *Liber Thesauri Occulti* confirm that the particularities of the Paris abridgment were extant in its model and should not be attributed to the epitomist.
The *Liber Thesauri Occulti* is not only the earliest translation of Artemidoros and the *Oneirocriticon* into Latin; 85 it also furnishes important textual evidence about the form and content of a more extensive version of the *Oneirocriticon.* It should therefore be used in any future critical edition of the Greek text.
<sup>81</sup> Collin-Roset, "Le *Liber Thesauri Occulti,"* p. 137.
<sup>82</sup> See Appendix 2.
<sup>83</sup> This might be the case with nos. 3, 13 and 18 (see Appendix 2).
<sup>84</sup> For the outcome of dreams in twenty units of time (e.g., twenty years, twenty months, twenty weeks, twenty days, or twenty hours) and the significance of pearls and black clothes for those who wear them regularly, see Appendix 1.
<sup>85</sup> Collin-Roset, "Le *Liber Thesauri Occulti,"* p. 131.
The earliest complete translation of the *Oneirocriticon* was made in 1176, seven years after the one by Pascalis Romanos. The translator was Leo Tuscus, who is known to us from other sources as well. 86 He and his brother Hugo Etherianus, to whom the preface to the translation of the *Oneirocriticon* is addressed, were natives of Pisa. They both came to Constantinople to seek a career during the reign of Manuel Komnenos, since this emperor was known for employing Latins in his services. Hugo, Leo's older brother, was a layman with a thorough theological education and author of several theological treatises that won him a cardinal's hat shortly before the end of his life in 1182. He arrived in Constantinople before 1166 and soon caught the attention of the emperor, who consulted him on doctrinal questions and ordered him to write his long, three-book treatise on the procession of the Holy Ghost. This treatise, written in collaboration with his brother Leo,87 was Hugo's most important work and was frequently mentioned by later Greek theologians.
Leo was an interpreter in the imperial chancery. Besides the *Oneirocriticon* he also translated the liturgy of St. John Chrysostom and wrote a theological treatise, *De haeresihus et praevaricationibus Graecorum.* In his preface to the translation of the *Oneirocriticon,* he gives the reason for undertaking this work: 88 Hugo had dreamt that the emperor Manuel was riding on the bronze horse of the statue that surmounts the column called Augustiana in Constantinople. He was surrounded by several erudite Latins and was reading a booklet in Latin. In the dream Manuel singled Hugo out for special attention, after Hugo had interrupted his reading. The meaning of this dream remained obscure for the two brothers until 1166, when Manuel ended by decree the theological controversy over the inferiority of the Son to the Father started by Demetrios of Lampe. Manuel's decision was influenced, according to Leo, by reading Hugo's booklet, *De Filii Hominis Minoritate ad Patrem Deum.* It was the outcome of Hugo's dream that gave Leo the idea of translating the *Oneirocriticon* into Latin. 89 Based on a reference to Manuel's campaign against the
<sup>86</sup> The facts presented here on the lives and works of Leo Tuscus and his brother Hugo Etherianus are taken from A. Dondaine, "Hugues Etherien et Leon Toscan," *AHDLMA* 27 (1952), pp. 67-134.
<sup>87</sup> Ibid., p. 68.
<sup>88</sup> The full text of the introduction was published by Haskins, *Studies in the History of Mediaeval Science,* pp. 217-18 and idem, "Leo Tuscus," pp. 45-47.
<sup>89</sup> In a recent article, Schreiner hypothesized, "Der AnlaB zur Ubersetzung des beriihmten Traumbuchs des Achmet, den Leo schildert, wirkt eher wie eine Parodie, aber die Form scheint dem Kaiser doch geschmeichelt zu haben: Hugo, der Bruder, hatte im Traum Kaiser Manuel gesehen, wie er in der Ubersetzung des Buches las"; see P. Schreiner, "Der Austausch von
Turks, Haskins dated the translation of the *Oneirocriticon* by Leo Tuscus to 1176.90
Several manuscripts of Leo Tuscus's Latin translation are known, <sup>91</sup>but no critical edition has ever been published. The oldest manuscript that survives, *Digby* 103 in the Bodleian Library in Oxford, is from the twelfth century; it is also our oldest manuscript for Pascalis Romanus's *Liber Thesauri Occulti.92*
In addition to the the twelfth-century translations of Pascalis Romanus and Leo Tuscus, a third translation of the *Oneirocriticon* into Latin was made by J. Loewenklau, and published in 1577. Loewenklau's translation attributed the original text to Apomasar. In the 1603 edition of the *Oneirocriticon,* Rigault reprinted Loewenklau's translation opposite the Greek text, supplementing it with his own Latin rendering of the passages that were missing from Loewenklau' s Greek manuscript but that existed in Rigault's Parisian source.
The evidence collected so far indicates that the translations of the *Oneirocriticon* into the vernacular European languages were all based on these three Latin renderings of the work. They can therefore be divided into three groups, depending on which Latin translation provided their source. The first and most recent group is represented only by printed texts; it comprises translations made from Loewenklau's 1577 edition. The members of this group can be easily identified, since they all attribute the work to Apomasar.93
Of the members of the second group, those based on the translation of Leo Tuscus, none mentions an author's name. Following Leo Tuscus's translation, their title refers to the alleged Indian, Persian and Egyptian sources used for
92 On the importance of this manuscript, see Ricklin, *Der Traum der Philosophie,* pp. 307-22.
Motiven und Ideen," in *Europa Medievale e mondo Bizantino.Contatti effettivi e possibilita di studi comparati,* ed. G. Arnaldi and G. Cavallo (Rome, 1995), p. 3. Schreiner does not explain why he considers the motive for the translation to function as a parody. The book that Manuel was reading in Hugo's dream was in any case most likely to have been the latter's *De Fi/ii Hominis Minoritate,* rather than a translation of the *Oneirocriticon.*
<sup>90</sup> Haskins, "Leo Tuscus," pp. 45-47.
<sup>91</sup>*Wolfenbuttel* 2917; Bodleian *Digby* !03; modern copy in *Ashmolean* 179; BL *Harleian* 4025; Bibliotheca Casanatense, C. vi. 5 (new no. 1178); the printed catalogue of this library, M. Ceresi et al., *Catalogo dei Manoscritti della Biblioteca Casanatense,* 5 vols. (Rome, 1949-58), only covers manuscripts 1-500 and does not include the old C.vi.5; *Vat. lat.* 4094; BN *lat.* 7337; *Vindobon.* 5221; *Marburgensis* B 21; listed in Thorndike, *History of Magic and Experimental Science,* vol. 2, pp. 292-93; Haskins, *Studies in the History of Mediaeval Science,* p. 217; *Achmetis Oneirocriticon,* ed. Drexl, p. ix. A further list of extant manuscripts in N. F. Palmer and K. Speckenbach, *Triiume und Kriiuter. Studien zur Petroneller "Circa Instans" Handschr(ft und zu den deutschen Traumbuchern des Mittelalters* (Cologne and Vienna, 1990), p. 125.
<sup>93</sup> For the German members of this group, see Palmer and Speckenbach, *Triiume und Kriiuter,* p. 127. The first French translation from Loewenklau's Latin appeared in 1581.
the composition of the Greek original, or it echoes Leo's designation of the work from its content as "De somniis et oraculis."94 The earliest of them is possibly the thirteenth-century BN *franr,;ais* 1553.95 The earliest printed one is also the earliest printed version of the *Oneirocriticon* in any language: an Italian translation of Leo Tuscus' s text that appeared in Venice in 1525 under the title, *Expositione degli Jnsomnii secondo la lnterpretatione de lndi, Persi, ed Egyptii, tradutta da Greco in Latino per Leone Toschano, ed al presente data in luce per ii Tricasso Mantuano ad Alessandro Bicharia Patricio Pavese.* Several reprints were made, as well as a French translation from the Italian, which first appeared in 1552.96
The third group seems to have been based on both Pascalis Romanus and Leo Tuscus. The oldest representative of this group is possibly the Anglo-Norman dreambook preserved in *Berlin* Q. 968 of the Staatsbibliothek preussischer Kulturbesitz in Berlin.97 While this manuscript is from the middle of the fourteenth century, the translation itself is older, possibly from the end of the thirteenth century. 98 The translation is anonymous; 99 it was made for an aristocratic lady, Alice de Courtenay or, as the name was later deciphered, Alice de
96 Thorndike, *History of Magic and Experimental Science,* vol. 2, p. 297, mentioned an anonymous Latin dreambook titled *Expositio somniorum* from BN */atin* 16610, which he dated to the 12th-13th century. According to Thorndike's report, this work had been influenced by the Latin translation of the *Oneirocriticon* by Leo Tuscus and until then had remained unpublished. Thorndike's information on the text is repeated by Lamoreaux, "Dream Interpretation in the Early Medieval Near East," p. 283. The dreambook in question is in fact a copy of Pascalis Romanus's *Liher thesauri occulti,* known to and utilized by Collin-Rosel, the editor of the text; see Collin-Rose!, "Le *Liher Thesauri Occu/ti,"* pp. 118-19.
99 The manuscripts of this dreambook give no information on the identity of the translator. However, Suchier identified him as the Carmelite monk Jean Golein; see Suchier, "Altfranzosische Traumbticher," p. 133.
<sup>94</sup> For example, the French translation is called "Exposicion et significacion des songes" [sic].
<sup>95</sup> Three French and two Czech manuscripts are known. See W. Suchier, "Altfranzosische Traumbticher," *Zeitschrift furfranziisische Sprache und Literatur* 67 (1956-57), 129-67. The text of BNfran~ais 1317 was published by Berrio!, *Exposicions et significacions des song es.* Berriot's introduction repeats information from secondary literature (sometimes with mistakes). His edition of the text is supplemented with readings from *Berlin* Q. 968 (whose text is based both on Leo Tuscus and Pascalis Romanus, which Berrio! seems to be unaware of), BN fran~ais 24.432, *Cmpentras latin* 337, the 1581 printed French edition of Denis Du Val, and the Latin translation that accompanied Rigault's Greek text in the 1603 edition. However, Berrio! does not discuss the connection between the texts used, nor does he give a *stemma codicum.* I have been unable to consult the more recent cri\ical edition by M. G. Glover, "Critical Edition of the Middle French Version of Achmet ibn Sirin's Oneiromancy."
<sup>97</sup> Ibid., p. 135.
<sup>98</sup> Berrio!, *E.1posicions et significacions des songes,* p. 42.
Couty. 100 Its title reads *so/om* [ = *selon] ceo qe Daniel le Prophete le fist,* which gives the false impression that the work is yet another recension of the dreambook of Daniel, versions of which existed in Arabic, Greek, Latin and a number of European vernacular languages. It seems that this translation generated a whole group of dreambooks attributed to Daniel. The connection of the *Oneirocriticon* to a branch of the vernacular tradition of Daniel's dreambook has so far been ignored by contemporary research. <sup>101</sup>
Besides its versions in Latin and other European languages that could be characterized as translations, the *Oneirocriticon* served as a source for the compilation of original treatises on dreams. Establishing which Latin or vernacular version furnished the immediate source for each author requires further research, but the contents of two treatises indicate the author's familiarity with the interpretations contained in the *Oneirocriticon.* The first is William of Aragon's *De prognosticatione sompniorum. <sup>102</sup>*According to the editor of the text, William seems to have known the *Oneirocriticon* by way of Leo Tuscus' s Latin translation. 103 The fifteenth-century author Venancius of Moerbeke based part of his treatise on prognostications *(De presagiis futurorum libel/us)* on William's work. w4 The second is the treatise on dreams by the Milanese physician, mathematician, and philosopher Girolamo Cardano ( 1501-
<sup>100</sup>For the reading "Alice de Courtenay." see Steinschneider, "Ibn Shahin und Ibn Sirin," p. 239; Berriot, *Exposicions et significacions des songes,* p. 42, reads "Alice de Couty."
<sup>101</sup>The recent publication of S. Fischer, *The Complete Medieval Dreambook: A Multilingual, Alphabetical "Somnia Danie/is" Collation* (Bern and Frankfurt-am-Main, 1982), fails to mention anything on the subject. Speckenbach in his study of medieval German dream books, *Triiume und Kriiuter,* does not connect the "Dritte Uberlieferungsgruppe" (pp. 133-35) of the dream book of Daniel with similar passages from the Greek *Oneirocriticon,* though he is aware of the *Oneirocriticon* in ten Latin, three Old French and two Czech manuscripts (pp. 125-27). For an additional Slavonic manuscript, see chapter I, n. 20.
<sup>&#</sup>x27;°2 R. Pack, "De prognosticatione sompniorum Libellus Guillelmo de Aragonia adscriptus," *AHDLMA* 33 (1966), pp. 237-93; Spanish translation, *Arnaldus de Villanova, De la inte1pretacion de los suenos* (Madrid, 1975); see also R. Pack, "Addenda to an Article on William of Aragon," *AHDLMA* 35 (1968), pp. 297-99. About the attribution of William's work to Arnald of Villanova, a 13th-century physician who interpreted dreams for the kings of Aragon and Sicily, see Thorndike, *History of Magic and Experimental Science,* vol. 2, pp. 301-2.
im Pack, "De prognosticatione sompniorum," p. 247. Given that Pack detects a possible influence of Artemidoros on the work of William, and that Artemidoros, except for the utilization of his work in the 12th-century *Liber thesauri occulti* by Pascalis Romanus, was translated into Latin for the first time in 1539, the sources of William's possible acquaintance with Artemidoros should be investigated.
<sup>104</sup>See R. Pack, "A Treatise on Prognostications by Venancius of Moerbeke," *AHDLMA* 43 (1976), pp. 311-22.
76), whose *Somniorum Synesiorum omnis generis insomnia explicantes, Libri IV,* was published in 1562. <sup>105</sup>
Each translation was also, to a greater or lesser extent, an adaptation which further removed the new product from the original text. The problems in the manuscript tradition of all Latin or vernacular renderings are the same as for the Greek manuscript tradition: no two manuscripts preserve an identical text. Modern scholarly works mention, but do not sufficiently explore, the relationship between the Greek *Oneirocriticon* and its renderings into other European languages, as well as the influence of the *Oneirocriticon* on original works in these languages. 106 Much remains to be done in this field of research, as well as on the subject of the relationship of the body of Byzantine dream interpretation to its counterparts in the rest of medieval Europe. *w7*
#### *Editions*
The *editio princeps* of the Greek text appeared in Paris in 1603. 108 The volume comprised the Greek text and Latin translation of Artemidoros, Achmet, Astrampsychos and Nikephoros. 109 In his preface to Achmet, the editor N. Rigault noted that he consulted two Greek manuscripts from the collection of the Bibliotheque du Roi, without giving their call numbers. Rigault's text reproduces faithfully the text of *Paris. gr.* 2538 of the fifteenth or the sixteenth
<sup>107</sup>The main handicap of recent researchers seems to be their inability to read Greek. Both Collin-Roset and Berriot, the editors of a Latin and a French version of the *Oneirocriticon* respectively, did not use Drexl' s critical edition but relied instead on the Latin translation that accompanies Rigault's text for their comparisons. Let us hope that the recent English and German translations of Drexl's text will help remedy the situation, though no real textual study can be undertaken without the Greek text. A partial French translation has been prepared by Anne-Marie Vincent-Bernardi, "L' *Oneirocriticon* d' Achmet: Traduction et commentaire," 2 vols., Ph.D. diss., Universite Aix-Marseille I, 2000.
<sup>108</sup>It appeared in the same year from two print shops: "ex officina Claudii Morelli" and "apud Marcum Orry." The *National Union Catalog* (pre- 1956), vol. 22 (1968), only records the Orry printing (NA 0439075). More copies of the second than of the first printing seem to be available in university libraries in the United States. In the printing "ex officina Claudii Morelli," pp. 27-30 of the *Oneirocriticon* are repeated between pages 274 and 275. This defect is absent from the Orry printing. Otherwise, the two are identical.
<sup>109</sup>*Artemidori Daldiani et Achmetis Sereimif Oneirocritica. Astrampsychi et Nicephori versus etiam Oneirocritici. Nicolai Rigaltii ad Artemidorum Notae* (Paris, 1603).
<sup>105</sup> See Fahd, "L'oniromancie orientale," pp. 350-51; for a number of passages establishing the relationship between Cardano and the *Oneirocriticon,* see 0. Gotthardt, *Uher die Traumhiicher des Mittelalters,* pp. 11-13.
<sup>106</sup> Some remarks in Fahd, "L'oniromancie orientale," pp. 347-74; also E. Ploss, "Byzantinische Traumsymbolik und Krimhilds Falkentraum," *Germanisch-Romanische Monatschriji* 39 ( 1958), pp. 218-26 and Gotthardt, *Uher die Traumhiicher des Mittelalters.*
#### 120 CHAPTER THREE
century that is still today at the Bibliotheque Nationale. The second manuscript that Rigault claims to have consulted was identified by Ruelle as *Paris. gr.* 2427. <sup>110</sup>However, this manuscript was acquired by the Bibliotheque du Roi much later, in 1719. Ruelle's identification should therefore be rejected. <sup>111</sup> Drexl assumed that the second manuscript mentioned by Ruelle, possibly copied from *Paris. gr.* 2538, has been lost, but it is possible that this second manuscript never existed.
Deciding whether Rigault had used one or two manuscripts for his edition has implications for the importance that should be assigned to the readings of the *editio princeps.* If we accept that only one manuscript was used, then all discrepancies from *Paris. gr.* 2538 should be attributed to Rigault's own intervention. If we grant that two manuscripts were used, we should treat the *editio princeps* as an indirect source for the readings of a manuscript that is no longer extant. Rigault does not seem to have been very thorough in his search for Greek manuscripts to prepare his edition of the *Oneirocriticon: Paris. gr.* 2511 and 2419, which contain excerpts from Achmet, including the beginning of the text in which Rigault was interested, were already in the Bibliotheque du Roi in 1594, but Rigault did not use them, 112 nor does he claim to have used the second manuscript for the reconstruction of the text; he only claims to have consulted it to establish the identity of the author-to no avail, since the second manuscript was, like the first, missing its beginning pages. We know that in the early days of printing, editors sometimes tried to disguise the inadequacy of their work by making false claims about the sources of their texts. <sup>113</sup> Rigault would not have been the first editor to base his text on a single manuscript, while announcing that he had consulted two. <sup>114</sup>It is certainly improbable that a manuscript disappeared from the royal collection of France, especially one whose presence there is otherwise unrecorded and whose disappearance cannot be accounted for.
The Greek text in Rigault's edition was accompanied by Loewenklau's Latin translation, published in 1577. Loewenklau had based his translation on *Vindob.*
<sup>110</sup>Ruelle, "La clef des songes d' Achmet Abou-Mazar," p. 306.
<sup>111</sup>Drexl, "Die Berliner Achmethandschrift," p. 312.
<sup>112</sup>Ibid., n. 3
<sup>113</sup>For the practices of editors and printers in the first centuries of printing, see A. Dain, *Les manuscrits,* 3rd ed. (Paris, 1975), p. 162.
<sup>114</sup>Robotel, who edited the text of Aelian the Tactician in 1552, pretended that he had found in the Marciana two "satis vetusta exemplaria" ; in reality, all he had used was a copy ( an6ypm1>ov) of *Marc. gr.* 516 (14th century). See Dain, ibid.
*philos. et philol. gr.* 297, the text of which has several lacunae, when compared with *Paris. gr.* 2538. The parts missing from Loewenklau's text were supplemented by Rigault's own Latin translation. Since both the Viennese and the Paris codex lacked the first folio, Rigault substituted the first two missing chapters by translating back into Latin the Italian translation of Leo Tuscus's text, which had been available in print since 1525.
Rigault's edition has a number of drawbacks, including its reliance on only one or two manuscripts. <sup>115</sup>However, the plans of later scholars to improve on Rigault did not soon materialize. <sup>116</sup>In 1894, C.-E. Ruelle in a brief article compared Rigault's text with *Paris. gr.* 251 land *Paris. gr.* 2419. As a result, he made a number of emendations to Rigault's text, listed a few of its lacunae and called for a new critical edition. <sup>117</sup>The call was taken up by Drexl, who laid the foundation for a new critical edition in his dissertation of 1909, 118 and published the final product in 1925. <sup>119</sup>
Drexl inventoried sixteen Greek manuscripts of the *Oneirocriticon,* no two of which has exactly the same text. Seven out of the sixteen manuscripts contained only selected chapters. 120 Two contained a vernacular paraphrase; 121 another had only a table of contents and a paraphrase of most of what is chapter l in Drexl 's edition. 122 The remaining five copy the text from beginning to end, but all have lacunae of various lengths and a varying sequence of chapters. Unavoidably, Drexl's final product (or any editor's for that matter) is a pastiche of chapters forming a version that does not exist in any surviving manuscript. The challenge for an editor in such a situation is to establish criteria that would allow him to choose the sequence of chapters and the readings that are the closest to the archetype.
120 T, P (which contains a text so altered in places, that it is almost a paraphrase; Drexl, *Achmetis Oneirocriticon,* p. xii), A, *Paris. gr.* 2427, *Paris. gr.* 2419, *Borbon. gr.* 356 (III. E. 34) and *Athos, lviron* 4285.165
<sup>115</sup>Cf. the criticism in F. Drexl, "Studien zum Text des Achmet," *BZ* 33 ( 1933), p. 13.
<sup>116</sup> We have seen that a posthumous edition of the text by J. Tollius (1630-96) was never printed. For other aborted plans for a new edition and corrections of Rigault's text, see Drexl, *Achmets Traumhuch,* pp. 6-7.
<sup>117</sup> Ruelle, "La clef des songes d' Achmet Abou-Mazar," pp. 305-12.
<sup>118</sup>Drexl, *Achmets Traumhuch.*
<sup>119</sup>Drexl, *Achmetis Oneirocriticon.* Negative comments on the editorial method in Latte, review of *Achmetis Oneirocriticon,* ed. F. Drexl, pp. 413 ff. and V. de Falco, review of *Achmetis Oneirocriticon,* ed. F. Drexl, *BZ* 27 (1927), pp. 113 ff.
<sup>121</sup>*Hieroso/.* (of the Patriarchate) *gr.* 220 (17th century); *Hieroso/.* (of St. Sabbas) 555 (17th century).
<sup>122</sup> *Cantahrig. (Trinity College) gr.* 1386 (0 8.11,6102) of the 15th or 16th century.
#### 122 CHAPTER THREE
The most obvious criterion is the age of each item in the manuscript tradition. The oldest Greek manuscript known to Drexl was R *(Vindob. philos. et philol. gr.* 111 of the thirteenth century), a text which, compared with Drexl' s edition, has several lacunae. The Latin translation of Leo Tuscus was a century older than the oldest Greek manuscript available. Drexl knew a number of manuscripts preserving Leo's texe 23 and his initial intention, which he never carried through, was to print the Latin translation side by side with the Greek text, 124 but he did use the Latin translation as the basis for his reconstruction of the Greek text. The oldest surviving manuscript of the translation known to Drexl was the twelfth-century *Digby* 103, but Drexl used the fourteenth-century Wolfenbi.ittel *Guelpherb. lat.* 2917 instead. 125 The oldest Greek manuscript, R, was thought to belong to the *codices recentiores et deteriores* and was assigned a secondary role in the text's reconstruction.
Drexl consulted only eight of the sixteen Greek manuscripts he had been able to locate in library catalogues. Excluding the remaining eight without examination must have been a decision that resulted from their inaccessibility coupled with the seemingly poor promise of their contents. <sup>126</sup>The eight Greek manuscripts that were consulted plus the *editio princeps* by Rigault were arranged in a *stemma codicum.* Surprisingly, the Latin translation of Leo Tuscus
<sup>123</sup> In his 1909 dissertation, Drexl notes that during his search for Greek manuscripts he also found two Latin manuscripts of Leo Tuscus' s translation, Wolfenbtittel, *Guelpherb. lat.* 2917 and Bodleian, *Digby* 103. In his 1925 critical edition he also cites the catalogue of Leo Tuscus's manuscripts by Haskins, *Studies in the History of Mediaeval Science,* and an additional one, *Marburg.* B 21.
<sup>124</sup> Drexl, "Studien zum Text des Achmet," p. 13
<sup>125 &</sup>quot;Cod. Guelpherb. optime conservato sum usus" (Drexl, *Achmetis Oneirocriticon,* p. xv). I suspect that the accessibility of the manuscript was not unrelated to this choice, at a time when manuscripts could not be photoreproduced but had to be consulted in the library where they belonged or sent to the interested scholar by the librarian in charge. At the end of the prolegomena to the critical edition, Drexl thanked the librarians in Berlin, Vienna, Leiden and Paris who had generously sent manuscripts to Munich for his use (pp. xv-xvi). The parchment of *Digby* 103 is in excellent condition and very legible; however, the back of the binding is worn and the manuscript must be handled with care. Perhaps this is why it was not sent to Munich.
<sup>126</sup> The two Jerusalem manuscripts contained a vernacular paraphrase of the text. The Athos manuscript contained only five chapters (221-26). All three manuscripts must have been extremely difficult to consult at the time. I do not know if it was difficult to gain access to Naples, *Borbon. gr.* 356, which was not consulted, though it contains almost all of the Greek text. The Cambridge manuscript only preserves a partial table of contents and part of chapter I. The three Paris manuscripts that were not consulted by Drexl had been examined by Ruelle, who published his emendations in 1894 (Ruelle, "La clef des songes d' Achmet abou-Mazar" ; cf., however, the criticism in Latte, review of *Achmetis Oneirocriticon,* ed. F. Drexl, p. 413). Moreover, Drexl consulted Rigault's edition, which probably, in Drexl's eyes, rendered superfluous the consultation of its source, *Paris. gr.* 2538.
that constituted the basis for the reconstruction of the Greek text is not assigned a position in this *stemma.* The only comment that Drexl makes about the relationship of the Latin translation to the Greek manuscripts is that Leo must have used a manuscript close to A *(Ambros. gr.* 592 of the fifteenth century), a conclusion he illustrated with fifteen examples. <sup>127</sup>Interestingly, the portion of the text that A preserves (only 27 Teubner pages) is shorter than that contained in any of the remaining seven manuscripts. Moreover, the Latin translation was not assigned a *siglum* in the critical apparatus; the choices made on the authority of the Latin text are almost never discussed in the apparatus, and it is impossible for the reader of the critical edition to obtain a clear picture of the manuscript tradition. <sup>128</sup>
Drexl distinguished two branches of the tradition, *x* and *y.* Branch *x,* represented by manuscripts ALPT, was considered to preserve older and better readings. The reason for such a conclusion, though not clearly stated, must have been the relationship of A, which belonged to branch *x,* to Leo Tuscus's translation. It is surprising that Drexl places P in this branch since, in his words, "the context in this codex is in several places changed to such a degree, that it is possible to call it a paraphrase." <sup>129</sup>
Drexl then explains the method by which he reconstructed the Greek text. <sup>130</sup> Codices AL TP, which represent an older and better tradition, were preferred, especially when they agreed with each other and were confirmed by Leo Tuscus. But, whenever APT preserved no text (since they only contain parts of the work) then L was added, but caution was exercised because L had been copied by an "ignorant and negligent scribe," so the readings of L were retained only if confirmed by Leo Tuscus and a manuscript from branch *y.* Otherwise, the matter was decided according to the meaning of the passage and style of the author. The few readings that were attested by branch *y* alone were decided on after consulting Leo Tuscus and considering the meaning of the passage and style of the author.
Drexl 's editorial method meant that one single manuscript of the Latin translation provided the authority on which chapters, such as those containing the anecdotes, were transposed, and readings emended; no thought was given to the possibility that the Latin tradition itself might present the exact same
<sup>127</sup> Drexl, *Achmetis Oneirocriticon,* pp. viii-ix.
<sup>128</sup>Cf. criticism in Latte, review of *Achmetis Oneirocriticon,* ed. F. Drexl, p. 414.
<sup>129</sup>Drexl, *Achmetis Oneirocriticon,* p. xii.
<sup>130</sup> Ibid., pp. xiv-xv.
#### 124 CHAPTER THREE
problems that the Greek tradition of the *Oneirocriticon* did. Drexl' s exposition makes clear that in several instances a reading was decided on, not by strict adherence to the manuscript tradition, but by what the editor considered appropriate to the meaning of the passage and style of the author. In other words, the editor made several arbitrary decisions in defiance of the manuscript tradition. As de Falco pointed out in his review of the critical edition, Drexl often corrected, not a corrupted reading of the manuscripts, but the author of the *Oneirocriticon* himself, both in the arrangement of the chapters and in the grammar and syntax of the text, anachronistically introducing forms that occur only in older phases of the Greek language. <sup>131</sup>
The evidence furnished by Arabic dreambooks sometimes corroborates readings preserved in the *y* branch of the Greek tradition that was rejected by Drexl as belonging to an inferior family of manuscripts. For example, in Drexl's edition the paragraph on interpreting dreams about playing polo *(tzykanizein)* is incorporated into chapter 154, "From the Persians and Egyptians on Excellent Horses." <sup>132</sup>However, all of the manuscripts in classy *(deteriores)* and L, from class *x,* discuss polo in a separate chapter, titled "From the Indians on Balls" or simply "On Balls" or "On Balls, That Is *Tzykanion." <sup>133</sup>*The chapter "On Balls" in class *y* also includes a paragraph on saddles. <sup>134</sup>In the Arabic dreambooks an interpretation of polo is never found in the chapters on horses. In the work entitled *al-Muntakhab, <sup>135</sup>*polo (or, rather, the polo mallet, ~aulajiin), is interpreted in a chapter on riding implements, which is found in a part of the dreambook far removed from the chapter on horses. In another Arabic dreambook, that of Ibn Shahin, the polo mallet is interpreted in the chapter on
<sup>131</sup>Cf. de Falco, review of *Achmetis 011eirocritico11,* ed. F. Drexl, p. 114: " ... mi pare che tavolta ii Drexl correga non una corruzione <lei codici ma proprio Achmet, oppure ristabilisca non la vera lezione, ma piuttosto quella lezione che si richiederebbe in un testo di parecchi secoli anteriore al nostro." Ibid., p. 115: "A confortare la mia opinione che in questo testo non si debba pretendersi una rigida precisione di forma, e opportuno citare qualche luogo, che ben dimostra come ii trattato risulti variamente elaborato. L' afferma innanzi tutto l 'au tore stesso nel brevissimo proemio, dove prima dichiara di attingere EK <wv notricrciv1cov <iiv 10tau1riv ciKpi~nav Kma aAi\0nav, T\10t' lvlioov Kat CTepcroov Kat Aiyun1icov, o'i <iiv aA.i\0nav aKpt~OAOYT\O'cXµevot Kat A.rn10A.oyi\crav1e~ E1;£0evw Kat £1..oyoypci<Pricrav <Tiv napoucrav Epµriveiav, quindi prosegue: Kat £1; EKacrwu 10u1cov EKA.el;ciµevo~ Ke<IJaA.mcoliw~ El;e0£µriv <wv <ptoov <a~ Kpicret~ Kat A.ucrn~ EV EKacr1cµ Ke<IJaA.aicµ K"CA. Ma nell'uso delle fonti egli non ha segulto un criteria rigoroso e preciso." De Falco gives a number of incontrovertible instances where the editor "hyper-corrected" the text.
<sup>132</sup>pvli' 'EK *"CWV* nepcrwv Kat Ai yun1icov nepi euyevwv 'inncov; Drexl 112, 20-113, 10.
<sup>133</sup> EK <wv' lvliwv nepi cr<IJaipa~; nept cr<Paipa~; nepi cr<Paipa~ T\wt </;uKaviou.
<sup>134</sup> Drexl 113, 5-10.
<sup>135</sup> *Al-Muntakhah,* p. 338.
arms and armor. 136 A chapter on arms and armor immediately follows the chapter on polo in the Greek tradition. The tradition of class *y* and L is therefore supported by the Arabic evidence, indicating that the Greek interpretation of polo should be separated from the chapter on horses, where Drexl placed it.
A second example is more complicated. Chapter 18 in Drexl' s edition is called 'EK 'tOU Ahyou -rffiv' Ivo&v 7t£pi 'tij~ crriµaofo~ -rffiv -rp1xc0v (From the Account of the Indians on the Interpretation of Hair). This title was invented by Drexl on the basis of P, 7t£pi -rp1xc0v 'tOU crffiµa'tO~ oA.ou (On Hair of the Whole Body), although the text in Pis almost a paraphrase. 137 A and T do not preserve this part of the text. Leo Tuscus has "De membrorum hominis significationibus" (On the Significance of Members of the Human Body) 138 and thus agrees with class *y* and L, which give the title £K -rffiv 'Ivo&v 7t£pt 'tij~ crr1µacria~ -rffiv mu av9po'mou µ£Affiv (From the Indians on the Significance of Members of the Human Body). Neither title is ideally suited to the contents of chapter 18, but the title supported by class y, L, and Leo Tuscus should be preferred, not only because of the overwhelming manuscript evidence but also as *lectio difficilior.*
The Greek text before and after the title of chapter 18 is somewhat incoherent. Based on Leo Tuscus (according to the critical apparatus), Drexl changed the position of three chapters and the title of chapter 18 in order to smooth over that difficulty, but all he managed to do was obscure the evidence about the process of the text's rendering from Arabic into Greek. In contrast, the text in Rigault's edition, where the editor did not bother much with textual criticism and remained closer to his unique manuscript source, gives the following sequence of chapters: after a chapter on judges we read two examples of actual consultation: the dream of a man whose hair between the legs became longer and thicker, and the dream of a second man whose body hair fell out after annointment. These are followed by Drexl's chapter 18, "On the Significance of Members of the Human Body," which discusses dreaming of a known person or someone who looks like a known person, dreaming of a young or an old man, dreaming that one's hair became gray or white, that it became denser and longer or that it was cut. Drexl changed the sequence by inserting chapter 18 after the Egyptian chapter on judges and changing its title to "On the Interpretation of Hair," followed by the two examples of actual consultation on dreams about hair.
<sup>136</sup> Ibn Shahin, nos. 5155-58; chapter on arms and armor in vol. 2, p. 67 ff.
<sup>137</sup> In Drexl's own words, "Contextus in hoc codice plurimis locis adeo commutatus est, ut paraphrasis dici possit" *(Achmetis Oneirocriticon,* p. xii).
<sup>138</sup> I am quoting from *Digby* 103, table of contents on fol. 59v; Drexl gives no information for the reading of *Guelpherh. lat.* 2719.
#### 126 CHAPTER THREE
The examination of Arabic dreambooks illuminates the seeming incoherence of the Greek text and supports the Greek manuscript tradition that was rejected by Drexl. Arabic dreambooks first interpret dreams with religious contents. This group of dreams usually concludes with dreaming of the educated classes, judges, and religious scholars. After religion, Arabic dreambooks discuss humans in a sequence similar to that found in Artemidoros, who begins with birth and youth and goes on to discuss the members of the human body from head to toe. This outline is faithfully followed in the Greek *Oneirocriticon.* After the discussion of judges and the conclusion of the religious dreams the *Oneirocriticon* goes on to discuss hair and parts of the human body. The beginning of this discussion can be found in the latter part of chapter 17, "From the Account of the Egyptians on Judges and Judgments." Chapter 18, "On the Interpretation of Members of the Human Body," which begins, in fact, with dreaming of a known or an unknown person, a young or an old man," has its exact parallel in a chapter from Ibn Qutayba (d. 889): <sup>139</sup>
\_,I~ \_,I~ ~\_)I dJj ~ -..i\_,\_;-o-11 ...4-.;-ll .<GL.....:.....c.I\_, .:\_,L. ••. :i'l/I ~J.; .:\_,LS.:.,)\_, J.1-L ~ l;>Ll .:\_,LS\_, "i~ .:\_,LS.:\_,µ '-""'L...JI ..J-a .. ~\_,I~ ... .;illl ~I\_,.,~~~
Dreaming of Humans and Parts of Their Body. A known man is indeed this very same man in person, or someone with the same name, or his brother, or somebody like him from among the people. And if the man was unknown and was a youth, it is an enemy. If he was an old man, he is the dreamer's luck, and luck is destiny ....
Ibn Qutayba goes on to discuss dreaming of an old lady, a mature woman, and a maiden, which the Greek text does not discuss, at least not in this chapter. <sup>140</sup> But Ibn Qutayba subsequently discusses the parts of the human body beginning with the hair of the head, exactly as it is done in the *Oneirocriticon.*
Drexl also omitted other passages found in the manuscripts, apparently be-
<sup>139</sup> Ibn Qutayba, *hilh* 14, Jerusalem *Yahuda ar.* 196, fol 28a ff. Cf. with Rigault, chap. 20 = Drexl 14, 16 ff. (Rigault's text reflects classy of the Greek manuscripts and is closer to the Arabic text than Drexl's is): EK trov· Ivorov m::p1 tili; cr11µacriai; 1rov civ0pffirrou µEA.rov. · Eav tol,l ni; av0pC01tOV Ka't <sup>0</sup>ovap yvffiptµov ij oµotov mu yvcopiµou, di; 'tOV yvffiptµov ciito~T]crE'tat toU civ0pc01tOU toil ~A.rnoµ£vou. Ei OE VEW'tEpov ciyvffiptcrtov tol,l, £x0p6i; EcrtlV 0 opffiµEvoi;.' Eriv oE y£povta ciyvffiptcrtov, i\ yvffiptµov y£povta, ij tUXTJ £m1 toil oprovtoi; (From the Indians on the meaning of parts of the human body: If someone dreams of a man who was familiar or looked like someone familiar, the dream's outcome regards this familiar man. Ifhe dreams of an unfamiliar young man, the person dreamt of is an enemy. If he dreams of an unfamiliar or familiar old man, he is the destiny of the dreamer).
<sup>140</sup> This discussion occurs later, in the chapter on women (Drexl 76, 10 ff.).
cause they did not meet his criteria for what he considered sound. <sup>141</sup>The extent of these omissions, together with problems that remained unresolved in Drexl's edition, as well as those that have emerged since with the available new data such as the Paris abridgment and Pascalis's *Liber Thesauri Occulti* cannot be addressed until a new critical edition of the text is made that avoids Drexl 's methodological biases. Preliminary investigation indicates that Rigault's text, which reflects classy of the Greek manuscripts, though judged by Drexl to be inferior, is closer to the Arabic dreambooks than is the text of Drexl' s critical edition.
The new critical edition should examine all available Greek manuscripts, the translation by Leo Tuscus, the *Liber Thesauri Occulti,* and the evidence of Arabic dream interpretation. The additional interpretations that are found in the abridgment of *Paris. Suppl. gr.* 690 and the *Liher Thesauri Occulti* should be sought in the Greek manuscripts that Drexl did not examine. All branches of the tradition, Greek, Latin and Arabic, should be taken into consideration. Though the tradition is very contaminated and it is likely that no stemma can be constructed, an effort should be made at least to evaluate the age of the tradition reflected in each existing manuscript and examine the relationship between them. The resulting critical text will necessarily be, like Drexl's text, a pastiche that does not exist in any of the surviving manuscripts. But this time the pastiche will at least be all inclusive, will show greater respect for the manuscript tradition and as a result, it is hoped, will be that much closer to the irretrievably lost archetype.
<sup>141</sup>For example, the critical edition omits a phrase from the introduction found in P, though Pis considered by Drexl to belong to the *x* family of better manuscripts. The phrase is inserted after Drexl 1, 14 *(Paris. gr.* 2511, fol. 7r): EXEl 8£ *i)* apxi) i:ciivoi:: i:ciiv K£\$aA.aiwv i:ciiv ovnpcimv Kat <Tis i:Kcicnou A.uai::ws aui:ciiv, oui:ws. No comment on the omission is made in the critical apparatus.
#### CHAPTER FOUR
#### COMPARING ARTEMIDOROS AND ARABIC DREAM INTERPRETATION WITH THE ONEIROCRITICON
The earliest surviving dreambook written in the Greek language is the secondcentury A.D. work of Artemidoros, which was based on an existing tradition, both written and oral, as the author himself says in the introduction:
έγω δε τουτο μεν ούκ έστιν ό τι βιβλίον ούκ έκτησάμην όνειροκριτικόν πολλὴν είς τουτο φιλοτιμίαν έχων, τουτο δὲ καὶ σφόδρα διαβεβλημένων τῶν ἐν ἀγορᾶ μάντεων ... ἔτεσι πολλοίς ώμίλησα, και ἐν Ἑλλάδι κατὰ πόλεις καὶ πανηγύρεις, καὶ ἐν Ἀσία καὶ ἐν Ἱταλία καὶ τῶν νήσων ἐν ταῖς μεγίσταις καὶ πολυανθρωποτάταις ύπομένων άκούειν παλαιούς όνείρους και τούτων τας αποβάσεις ου γαρ ήν αλλως χρήσασθαι τη κατά ταυτα γυμνασία. ὅθεν μοι περιγέγονεν ἐκ περιουσίας έχειν περὶ ἑκάστου λέγειν [πλείονα μὲν ή προσδοκῆσαι αν τις] ούτως ώς αυτά τάληθη λέγοντα μὴ φλυαρείν, ών δ΄ αν έπιμνησθώ [και] τας αποδείξεις φανεράς και πασιν εύκαταλήπτους άποδουναι [μ ] έξ άπλών, πλὴν εί μή τι είη τουτο σαφές, ώς περιττὴν ήγήσασθαι τὴν περὶ αὐτοῦ ἐξήγησιν.
I have not only taken special pains to procure every book on the interpretation of dreams, but have consorted for many years with the much-despised diviners of the marketplace ... In the different cities of Greece and at great religious gatherings in that country, in Asia, in Italy and in the largest and most populous of the islands, I have patiently listened to old dreams and their consequences. For there was no other possible way in which to gain practice in these matters. As a result, from the superabundance of examples, I am able to discuss each individually [more than anyone might have expected] so as to speak the truth without nonsense, and to prove the truth of my assertions clearly and comprehensively by simple statements. except in cases so obvious that I think an explanation is superfluous.
The earliest author of a dreambook that Artemidoros refers to is Antiphon the Athenian, possibly the same man as Antiphon the Sophist, the rival of Socrates, an authority approximately six centuries older than Artemidoros. Though most of the dream interpreters whom Artemidoros mentions by name are Greek, the era in which he lived and the extent of his travels indicate that he
<sup>1</sup> Greek text in Pack, 2, 11-27; English translation in Artemidoros, Interpretation of Dreams, trans. White, pp. 13-14; cf. also Artemidoros's introduction to Book v: "I carefully collected as many dreams as I could at festal assemblies throughout Greece and Asia as well as Italy ... " (Pack 301, 10-13; Artemidoros, Interpretation of Dreams, trans. White, p. 229).
<sup>4</sup> For bibliographical references on Antiphon's identity, see Artemidoros, Interpretation of Dreams, trans. White, p. 11, n. 19 and p. 67, n. 4; see also del Corno, Graecorum de re onirocritica, pp. 45-50 and 129-32.
This is an open access chapter distributed under the terms of the CC BY 4.0 license.
must have been familiar with material originating in the Egyptian and possibly also the Near Eastern tradition of dream interpretation.' Such contact would account for the common elements that can be found between Artemidoros and the ancient Egyptian and Near Eastern dreambooks that survive.4
Greek influence penetrated the tradition of Jewish dream interpretation, as is evident in the dreams and their explanations recorded in the tractate *Beriikot* of the Babylonian Talmud, which was written down ca. A.D. 600 on the basis of earlier oral tradition. 5 Jewish lore (presumably including elements received from the Greek tradition) obviously influenced Arabic dream interpretation, since methods and principles of this art, such as the interpretation of dreams on the basis of passages from the Torah in the Jewish tradition and from the Qur'an in the Islamic tradition, as well as specific examples of dreams that had been dreamt and correctly interpreted, occur both in the Talmud and in later Arabic dream books. 6 Though Greek paganism did not possess holy books equivalent to the Torah and the Qur'an, Artemidoros also based a number of interpretations on older Greek poetry such as Homer, Euripides and Menander which, by his time, was several centuries old and constituted literary classics.
<sup>3</sup> Artemidoros refers to a dream interpreter whom he calls "the Egyptian" (iv.48, Pack 273, 5-12). The geographical position of the homelands of the dream interpreters Geminus of Tyre (ii.44, Pack 179, 13-14) and Phoebus of Antioch (i.2, Pack 6, 15; ii.9, Pack 11 l, 17; iv.48 Pack 275, 6-10; iv.66, Pack 289, 1-6) mentioned by Artemidoros indicates that they were likely to have been familiar with Near Eastern traditions.
<sup>4</sup> For two Egyptian dreambooks contemporary with Artemidoros, see A. Volten, *Demotische Traumdeutung* (Copenhagen, 1942). Volten in his footnotes gives copious examples of similar interpretations that occur in Artemidoros and the Byzantine dreambooks. For ancient Near Eastern dream interpretation, see A. L. Oppenheim, "The Interpretation of Dreams in the Ancient Near East, with a Translation of an Assyrian Dream-Book," *Transactions of the American Philosophical Society* N. S. 46:3 (l 956), pp. 179-354. The Assyrian dreambook translated by Oppenheim is preserved on cuneiform tablets from the library of Asurbanipal (668-33 s.c.) that yield a rather fragmentary text. Despite the limited volume of the Assyrian material and the chronological distance of several centuries separating the two dreambooks, some loose correspondences between the Assyrian and the Artemidorean interpretations can still be identified, e.g., the penis indicates progeny (Oppenheim, p. 271, Artemidoros v.86); flying is auspicious but also dangerous (Oppenheim, p. 287, Artemidoros ii.68); rivers signify money and social importance (Oppenheim, p. 287, Artemidoros ii.27); etc.
<sup>5</sup>*Beriikot* 9, fols. 55a-60b, A. Cohen, ed. and trans., *The Babylonian Talmud, Tractate Beriikot* (Cambridge, 1921 ), pp. 358-99. For the influence of Greek sources on the Talmudic material, see S. Lieberman, *Hellenism in Jewish Palestine* (New York, 1950), pp. 68-82; and H. Lewy, "Zu dem Traumbuche des Artemidorus," *Rheinisches Museumfiir Philologie,* n. F. 48 (1893), pp. 398-419. Some resemblances between Artemidoros and the Talmudic material have also been pointed out in P. Cox-Miller, *Dreams in Late Antiquity* (Princeton, N .J., 1994), p. 84, n. 44; p. 87, n. 55; p. 88, n. 58.
<sup>6</sup> See Kister, "Interpretation of Dreams," pp. 99-10 I.
Basing interpretations on poetry is a well-known method also employed in Arabic dream interpretation; it is documented in the earliest surviving dreambook in Arabic, the dreambook of lbn Qutayba.
The connection between Muslim and Jewish dream interpretation in the early Islamic period must have become close since Muslims evidently did not hesitate to summon the services of Jewish interpreters of dreams, as is demonstrated by the examples of Caliph Yazid II (r. 720-24) 7 and Caliph Hisham b. 'Abd al-Malik (r. 724-43). 8 Direct Greek influences on Arabic dream interpretation were introduced through the wave of translations from Greek into Arabic in the ninth and tenth centuries that included medical and philosophical texts on dreams, such as those by Galen9 and Aristotle, 10 and especially through the
8 Cf. al-Dinawari, fa~/ 19, *hiih* 24. Before becoming caliph, Hisham dreamt that he was holding nineteen-and-a-half fruits in his hands. He narrated his dream to a learned Jewish dream interpreter, who foretold that he would become caliph and reign for nineteen and a half years.
9 The theoretical introduction to Arabic dreambooks separates pathogenetic dreams as a category that cannot be interpreted and explains their mechanism based on the existence of the four humors in the human body (blood, phlegm, black bile and yellow bile), as advocated in Greek medicine. Ibn Qutayba refers to pathogenetic dreams in Ankara, *Ism. Saih Sincer* I, 4501, fol. 181 b (the equivalent passage is missing from the Jerusalem manuscript). Pathogenetic dreams are the sixth class of dreams in al-Dinawari's introduction *(maqii/a* 7, *Esad Efendi* 1833, fol. 28b; BN *arahe* 2745, fol. 38a). They are also discussed in *a/-Muntakhah* (p. 24 = *a/-hishiira,* BL *Or.* 6262 fol. 6a [= 12]). See also *al-Muntakhah,* p. 7: *cSJ-0 ,)* ~.J i"~~I ..:..,~I d.Jj .j-o.J .J-o ~.JI \_u.W ~\_;.\_,<>\_;~I .\_:,1.J ~~.\_:,I ...J~.J l..ii........ ..:...\_;l..,..,, .L.........JI .\_:,tS: .\_:,L.........:.~1 d.Jj ~I L.J li..:. J.:;.i.Jl.J ~.\_:,~I J~ .JI\"~ .\_,<>\_;~I .J-o ..::...tl.b.J .;l..7--•:.81 .~I <sup>o</sup>~I *cS \_;.: cS..9-"* ..J Ls: -.\_~\_'.L., .. ..b *..}....-:* ~ .i.:....c. .\_:,L..\_, .. \_:,Y I LA I *J-0* l: .'.u d.J j .j-o .J Y-J J...o .JL.:; ~ .J .;J.rJ-1 .J Jl\_,\_.\_...J I .J ..:...Ll.l;J I *c>.>-: c>.J* IJ\_,\_....J I .J o \_;.L.:.J I *c>.>-:* c>.JI ~I .J ~.)I *c>.>-:* .\_,.,\_,...b.). I .J ~~I .L...,,\_ .. :;,YI *c>.>-:* .~l.J ..:...IJ.J\_;-,JI *c>.>-:* J.Jr.Jl.J i"l....,...,Jl.J \_;l:.Jl.J ~I *c>.>-:* .L...::.,:1 .U ~.J'G ~ l:j.) I .J-o t\_,\_:J I 1.J..+--0 ~(Confused dreams are when one dreams that heaven became a ceiling and fears that it is going to fall on one, or that the earth was transformed into a mill and was rotating, or that the trees grew from the sky, or that the stars rose from the earth, or that the devil was transformed into an angel and an elephant into an ant. Dreams like that have no interpretation. Such dreams are dreamt by people with ailing dispositions. So the sanguine will see redness, the one with moistness in him will see moisture, the choleric will see yellowness, the melancholic will see darkness and blackness. Someone who is hot will see the sun and fire, while someone who is cold will see coldness and someone with a full stomach will see the heavy things inside him. This kind of dream has no interpretation, either). Part of this description repeats verbatim al-Dinawari's fifth class of dreams *(maqii/a* 7, *Esad Efendi* 1833, fol. 28b). A similar explanation for the mechanism of pathogenetic dreams is expounded in al-Nabulusi (vol. I, p. 3), who mentions *al-Muntakhah* among his sources. These passages repeat the principles expounded in Galen's opusculum, *De dignotione ex insomniis* (G. Guidorizzi, ed., "L'opusculo di
<sup>7</sup> See G. Strohmaier, "Der Kalif Yazid II. und sein Traumdeuter: Eine byzantinische Legende Uber den Ursprung des Ikonoklasmus," *Jahrhuch fur Geschichte des Feuda/ismus* 3 (1979), pp. 11-17.
ninth-century translation of Artemidoros by I:Iunayn b. Isqaq.
Foreign influences notwithstanding, dream interpretation was an ancient practice among the Semitic peoples and a widespread method of divination in Arabia long before the advent of Islam. It was the only one among the divinatory practices of the pagan period that was officially sanctioned by the new monotheistic religion. Sura 12 of the Qur'an depicts Joseph as a divinely inspired dream interpreter, and the Prophet and several of his companions were also purportedly gifted interpreters of dreams. Distinguished scholars from among the next generation of Muslims, such as Ibn al-Musayyab, also became renowned dream interpreters. Consequently, dream interpretation acquired religious overtones, especially since several interpretations were based on verses from the Qur'an and the *fladith. <sup>11</sup>*
<sup>10</sup>On the influence of Aristotle on Arabic dream interpretation and the theory of dreaming, see Fahd, *La divination arabe,* pp. 331-32 (on Aristotle) and p. 345 (on al-Kindi). Also H. Gatje, "Philosophische Traumlehren im Islam," *ZDMG* 109 (1959), pp. 258-85; idem, "Die "inneren Sinne" bei Averroes," *ZDMG* 115 (1965), pp. 255-93; M. Wali-ur-Rahman, "Al-Farabi and His Theory of Dreams," *Islamic Culture* 10 ( 1936), pp. 137-52; idem, 'The Psychology of al-Farabi," *Islamic Culture* 11 (1937), pp. 228-52; R. Walzer, "Al-Farabi's Theory of Prophecy and Divination," *Journal of Hellenic Studies* 77: I (1957), pp. 142-48; Ibn Sina, "A Unique Treatise on the Interpretation of Dreams by Ibn Sina," ed. M. 'Abdul Mu'id Khan, *Avicenna Commemoration Volume* (Calcutta, 1956 [?]), pp. 255-307; English trans. in idem, *"Kitiibu ta'bir-ir-ru'yii* of Abu 'Ali b. Sina," *lndo-lranica* 9:4 ( 1956), pp. 43-57; G. E. Pruett, "Through a Glass Darkly: Knowledge of the Self in Dreams in lbn Khaldiin's *Muqaddima," Muslim World* 75 (1985), pp. 29-44 (useful only because it summarizes Ibn Khaldiin's views about dreams; unfortunately, the article does not mention the sources that shaped them); M. Jevolella, "Songe et prophetie chez Maimonide et dans la tradition philosophique qui l'inspira," *Maimonides and Philosophy. Sixth Jerusalem Philosophical Encounter 1986,* ed. S. Pines and Y. Yovel (Dodrecht-Boston-Norwell, Mass., 1986), pp. 174-84. Still, however, the arduous task of collecting instances that coincide with Greek theories on the nature of sleep and dreams from the introductions of Arabic dreambooks and tracing how they ended up there remains to be undertaken. The Arabic and the ancient Greek traditions on dream interpretation seem to coincide even on subjects that are not covered by Artemidoros but are discussed in other Greek texts on dreams, such as the truthfulness of a dream dreamt while sleeping on one's right or left side, and the intensity and veracity of the dream depending on the season of the year.
<sup>11</sup>For a more detailed history of Arabic dream interpretation than what is covered here, see Fahd, *La divination arabe,* pp. 247-329; idem, "Les songes et leur interpretation selon !'Islam," pp. 127-58. Also Ch. Magdi, *Die Kapitel iiber Traumtheorie und Traumdeutung aus dem Kitiih* at-ta~rirfi *'ilm at-ta/sir des l)iyii' ad-Din al Djaziri (7.113. Jahrhundert)* (Freiburg im Breisgau, 1971 ), pp. 7-25. For a detailed exposition on the relationship between dream interpretation and the Muslim religious sciences, see Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 175-227. The connection of dream interpretation with the religious sciences is evident from its classification in the curriculum of knowledge by medieval authors. In the *Rasii'il lkhwiin*
Galeno *De dignotione ex insomniis," Bolletino de! Comitato per I' edizione dei classici graeci e latini,* N. S. 21 [ 1973], pp. 81-105). Though the Arabic translation of this work does not survive, it is certain that one was made, since it is mentioned by the 17th-century Ottoman bibliographer I:Iajji Khalifa; see Fahd, *La divination arabe,* p. 338.
#### 132 CHAPTER FOUR
Before the end of the eighth century, know ledge of dream interpretation among the Arabs seems to have been preserved and transmitted through oral tradition. The first Arabic dreambook ever written was apparently the *Dustur fi al-ta'blr* (Constitutions on Dream Interpretation) or *Kitiib al-ta'bir* (Book of Dream Interpretation) by Abu IslJaq Ibrahim b. 'Abd-Allah al-Kirmani. <sup>12</sup>It was composed at the order of Caliph al-Mahdi (r. 775-85), as is related in a number of Arabic dreambooks. 13 One of the earliest sources, the eleventhcentury dreambook of al-Dinawari, says:
LL..).J ~!Ji J>-""I ~J ..:,LS:~ <LLJI .\_:,I~\_; ~).11 ->-:'-"I ~J+ll csl\_;J b~,'.u ~ ~J .\_:,~~I~ .,?L,.?JI <LLJI *..l...:'L* 0-7 ~.J-:'I rl~l:i .;--olJ ~J JJ; \_;:...:.yl:i rAJ-:>.I ~ I.ii\_, ~J ~ .U\_,.i.J ~I ~ ..U.J-::'-'-" JW b•~ ~I\_,"---:'~ ..::......:...,1 ~I d.l.:; .U ..::....U\_,\_j \* ~ .J-AIJ IJ.J-'-"-" .~I~ '7'L:..S: ~l:....; b.;--olJ
Al-Mahdi, the prince of the faithful, may God be pleased with him, dreamt that his face became black. He woke up in terror and ordered that Ibrahim b. 'Abd-Allah al-Kirmani be summoned to him from SirJ:ian. <sup>14</sup>Al-Mahdi narrated his dream to al-Kirmani, who said that a girl would be born to the caliph, according to the saying of God Almighty: "Whenever any of them is given the good tidings of the birth of a female his face is darkened and he is wroth inwardly" [Qur'an 16:58]. Indeed, on that night a daughter was born and al-Mahdi was delighted about it;
12 See Fahd, *La divination arahe,* p. 345, and "The Dream in Medieval Islamic Society," p. 357; Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 41-46.
13 For references to this, as well as a further anecdote demonstrating al-Kirmanl's expertise on dream interpretation, see Lamoreaux, "Dream Interpretation in the Early Medieval Near East," p. 42.
*al-Jafa',* an encyclopedia compiled by a group of scholars in the 10th century, we read the following: "The sciences of the religious law that were created for the healing of souls and for the quest for future life are of six kinds: (I) The science of the Revelation, (2) the allegorical interpretation of the Qur'an, (3) reports and traditions (of the Prophet and other recognized authorities), (4) jurisprudence, ordinances and laws, (5) prayers, sermons, asceticism and mysticism, (6) the interpretation of dreams" ; see F. Rosenthal, *The Classical Heritage in Islam* (Berkeley and Los Angeles, 1975), p. 56. Ibn Khaldiin also classified dream interpretation among the *'uliim al-shari'a* (for a quotation of the passage and the relevant reference, see Fahd, *La divination arahe,* p. 248, n. 3). For a commentary on both passages, see Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 225-27.
<sup>14</sup>Lamoreaux, "Dream Interpretation in the Early Medieval Near East," p. 42, n. 41, relying on a different manuscript *(Chester Beatty* 3569) reads here "al-Sirjan," a toponym which he considers to be the name of a city. I have been unable to locate a city by this name. In favor of the reading "al-SirJ:ian" is the fact that Wadi SirJ:ian (:Eupµa"iov it£1itov in Greek sources) is a valley in northern Arabia running from the south end of the I:Iawran toward the southeast; since antiquity and throughout the Middle Ages it was used as a major communication and trade route *(EI2,* s.v. "SirJ:ian").
he gave al-Kirmani a generous gift and ordered him to compile a book on dream interpretation. is
No manuscript containing al-Kirmani's dreambook survives. However, his interpretations have been quoted by later compilers of dreambooks, and especially Abu 'Abd-Allah MuJ::iammad b. AJ::imad b. 'Umar al-Salimi (fl. end of the fourteenth century) <sup>16</sup>in his *al-Ishiira ila 'ilm al-'ibara* (The Intimation on the Science of Interpretation). Al-Salimi informs his readers that he based his work on that of al-Kirmani, but added new data. <sup>17</sup>More than a generation later, the dreambook of al-Salimi, including the interpretations attributed to al-Kirmani, was used as a source for the *Kitab al-ishiirat fi 'ilm al-' ibarat* (Book of Intimations on the Science of Interpretations) by Ibn Shahin (d. 1468), who added 30 chapters to al-Salimi's 50, for a total of 80 chapters. <sup>18</sup>
The oldest surviving Arabic dreambook was written by Abu MuJ::iammad 'Abd-Allah b. Muslim b. Qutayba (d. 889), and is known under two slightly different titles: *'lbarat al-ru'ya* (Interpretation of Dreaming) and *Ta'bfr alru>ya* (Dream Interpretation). 19 Ibn Qutayba seems to have relied on an already well-developed written tradition. He was a contemporary of I:Iunayn b. IsJ::iaq (d. 873) and might have been familiar with the latter's translation of Artemidoros. Indeed, some of Artemidoros' s interpretations can also be found in lbn Qutayba's dreambook. However, given the earlier Arabic familiarity with the Jewish oneirocritic tradition, which was heavily influenced by the Greek, and the possibility that elements of Near Eastern lore found their way into Artemidoros' s work, it is conceivable that the interpretations shared by Artemidoros and lbn Qutayba had been absorbed into the Arabic tradition even before I:Iunayn's translation.20 The matter cannot be decided without a study of Ibn Qutayba's dreambook in connection with the Arabic translation of Artemidoros and a thorough search for textual parallels.
The second oldest Arabic book on dream interpretation that survives is possibly *al-Qadir! fi al-Ta'blr* (The Book Dedicated to Caliph al-Qadir on Dream Interpretation), written by Abu Sa'id (or Sa'd) Na~r b. Ya'qub al-
is Al-Dinawari, *fa!f/* 6, *biib* 19 *(Esad Efendi* 1833, fol. 47b; BN *arabe* 2745, fol. 79r).
<sup>16</sup>Fahd, *La divination arabe,* p. 352, no. 103.
<sup>17</sup> In some manuscripts Ibn Sirin appears as the author of the treatise; all information from Fahd, *La divination arabe,* p. 352, no. 103.
<sup>18</sup> Ibid., p. 351, no. 102.
<sup>19</sup>See Kister, "Interpretation of Dreams," p. 67.
<sup>20</sup> Ibid., p. 99.
Dinawari (d. ca. 1009). According to its introduction, it was finished in 1006 and was dedicated to Caliph al-Qadirbi-1-Lah (r. 991-1031), whence its title. Al-Dinawari's dreambook was extremely popular and became the most important source for several subsequent authors.21 Fahd has inventoried more than 30 manuscripts of it, the oldest of which dates to 1202. The dreambook of Ibn Qutayba is rather brief, but al-Dinawari's work is remarkably extensive and detailed; in fact it constitutes one of the longest such works ever written. It is divided in 30 sections *(fa!f[),* and subdivided into 1,396 chapters *(biib).*
Approximately contemporary with al-Dinawari is the dreambook written by Abu AJ:imad Khalaf b. AJ:imad al-Sijistani (937-1009), the last Saffarid amir of Sijistan and a cultured patron of literary endeavors. This work is known under various titles, but the original one must have been Tu~fat *al-mu!Uk* (The Jewel of Kings, or the Prized Possession of Kings). It is deliberately concise, because it was intended as a quick reference.22
A dizzying number of Islamic dreambooks survive from the eleventh century and later. Fahd, who inventoried the authors and titles of both lost and extant dreambooks, ended with a total of 181 titles. 23 We will compare the *Oneirocriticon* with Arabic dream interpretation using five of them: Ibn Qutayba, al-Dinawari, *al-Muntakhab,* lbn Shahin and al-Nabulusi. The first two were chosen because of their early date; the remaining three because of their impor tance for preserving earlier material and because their printed editions make them relatively accessible.
Most modern scholars who have studied the *Oneirocriticon* have concluded that its author was familiar with both the Greek text of Artemidoros and Arabic dream interpretation. Only Dagron briefly mentioned the possibility, without, however, discussing it in any detail, that the influence of Artemidoros might have reached the *Oneirocriticon* through Arabic sources. 24 We will ex-
<sup>21</sup> Fahd, *La divination arabe,* pp. 336-37, no. 30. *Al-Qadir/* was also translated into Persian and Turkish; see *EI2,* s.v. "al-Dinawari, Abu Sa'id (Sa'd) Na~r b. Ya'qiib."
<sup>22</sup> Al-Sijistani's dreambook was known to Fahd from two manuscripts (see Fahd, *La divination arabe,* p. 354, no. 112), but the identification of its author with the well-known person of the amir and consequently the possibility for dating it escaped him. For the identification of the author and some additional details on his work, see Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 58-64; on p. 60 he lists three additional manuscripts.
<sup>23</sup> Fahd, *La divination arabe,* pp. 330-67.
<sup>24</sup> Dagron, "Formes et fonctions du pluralisme linguistique a Byzance," p 237: "<L'> auteur <de *l'Oneirokritikon>* connait l'arabe et retrouve peut-etre par ce canal une partie de la tradition antiqued' Artemidore."
amine here, therefore, whether the passages that have been attributed to the influence of Artemidoros in the Oneirocriticon are in fact closer to similar passages in Arabic dreambooks, or whether they were indeed inspired by the author's direct knowledge of the original text; and what exactly the relationship of the Oneirocriticon is to Arabic dream interpretation, particularly those aspects of the Oneirocriticon that appear as specifically Christian and Byzantine. Passages with Christian interpretations will be examined using their corresponding Muslim interpretations in Arabic dreambooks. Instances that indicate the author's familiarity with elements of the Byzantine imperial ideology will also be juxtaposed to equivalents from Arabic dreambooks.
#### The Arabic Translation of Artemidoros
The first three of Artemidoros's five books on dream interpretation are dedicated to a certain Cassius Maximus. Only these three books were meant for publication; the introduction to the fourth book states that the last two books were composed after the first three had been completed and were intended for the exclusive use Artemidoros's son and namesake, who was an apprentice dream interpreter. This introduction also mentions that the previous three had met with approval as well as criticism among the reading public.25 The last two books discuss subjects that were either omitted or insufficiently explained in the first three books. The father warns the son that, in order to have an advantage over his rival dream interpreters and diviners, he should keep the contents of these last two books to himself, because "once they become the common property of everyone, it will be obvious that you know nothing more than anyone else."26 Artemidoros included all he intended the public to know about his art in the first three books, which can therefore be viewed as a complete work, even without books 4 and 5.
The Arabic translation of Artemidoros survives in a unique manuscript, Ar.
<sup>35</sup> Pack 237, 17-22: και γάρ νυν αίσθάνομαί τινων τοις βιβλίοις έγκαλούντων ώς τών μεν έν αύτοις γεγραμμένων άληθείας μέν ούδ΄ ότιουν αποδεόντων, ού πάντων δε έξειργασμένων ούδε ήτιολογημένων, έστι δε ών και παραλελειμμένων αναγκαίων όντων Th incolege (As a matter of fact, I am now aware of certain men who charge that these books, while they leave nothing at all to be desired with regard to their accuracy, are nonetheless incomplete and do not delve far enough into the causes. And, indeed, they maintain that certain things still demanded by the subject at hand have been omitted).
<sup>26</sup> Introduction to book iv. Pack 238, 1-6; see also Artemidoros. Interpretation of Dreams, trans. White, p. 8.
*yazma* 4726 now in the Istanbul University Library. It was discovered by Fahd in 1959 and edited by him in 1964.27 The manuscript has been dated both to ca. 1200 and to the fourteenth century. 28 Its title page, written by a later hand,29 does not give the name of the translator; it tells its reader that the work is a "book on the interpretation of dreams by the wise Artemidoros in three sections" (..::...~L.:i...a ~ V"J->..l.AL.b) f *:<..,,* 11 ..::...LWI ~ '-:-'L...::..S: 1),30 and it contains the translation only of books 1-3. 31
The tenth-century bibliographical compilation, *Kitab al-fihrist,* by Ibn al-Nadim, mentions that the famous translator I:Iunayn b. Isi)aq (d. 873) had translated Artemidoros's work on dream interpretation in five books. 32 The text discovered by Fahd therefore generated a number of questions: who was the translator? how many books of Artemidoros's work had been translated into Arabic? was there an intermediary Syriac translation? how useful was the Arabic text for the retrieval of the Greek archetype?
The characteristics of the translation were analyzed by Fahd and by F. Rosenthal in somewhat different terms. According to Fahd, the overall impression of the reader is that it is a faithful and exact translation, and for this reason it is an indispensable aid to the editor of the Greek text. 33 The translation
<sup>27</sup> On the discovery, see Fahd, "La traduction arabe des *Oneirocritica,"* pp. 87-89. The text was published by Fahd, ed., *Artemidore d' Ephi'se;* an exhaustive index to the Arabic text with cross references to the Greek text was compiled by E. Schmitt, *Lexikalische Untersuchun{?en zur arabischen Ubersetzung von Artemidors Traumbuch* (Wiesbaden, 1970).
<sup>28</sup> H. Ritter and A. Ate§ dated the manuscript to ca. 1200 according to Fahd, ed., *Artemidore d' Ephese,* p. xxiii. A 14th-century date is given by F. Rosenthal, "From Arabic Books and Manuscripts, XII: The Arabic Translation of Artemidorus," *JAOS* 85 (1965), p. 139.
<sup>29</sup> See Fahd, "La traduction arabe des *Oneirocritica,"* p. 88.
<sup>30</sup> The first folio of the manuscript with the title of the work is reproduced in Fahd, *Artemidore d' Ephi'se,* pl. I.
<sup>31</sup> For a description of the manuscript, see Fahd, "La traduction arabe des *Oneirocritica* ," pp. 87-89, and *Artemidore d' Ephese,* pp. xxii-xxiv. Fahd's edition includes three plates that reproduce a total of five folia from the manuscript. The folia that currently comprise the manuscript seem to have belonged to a larger collection, from which they were extracted and bound separately for the library of Sultan Abdiil-Hamid II (1879-1909). The first twelve folia of the manuscript are missing (Fahd, "La traduction arabe des *Oneirocritica,"* p. 88, and *Artemidore d' Ephese,* p. xxi) as is indicated by an early "foliotage en minuscules grecs" that runs throughout the manuscript. The condition of the last folio, which is filled with certificates of purchase and other formulas from the 16th century and later, indicates that there were no further folia (Fahd, "La traduction arabe des *Oneirocritica,"* p. 88). The first part of the introduction (the address to Cassius Maximus, Pack 1,1-3,8) is missing.
<sup>32</sup> Ibn al-Nadim, *Kitiib a/-fihrist,* Fliigel et al., eds., p. 255: .... ,.L~ .\_,..=-L..... *->".JJ.l....:'--"l..b)* .~1.j-: ~ <l.i.:. -=.,~LL~ ~j.)1 ~ *'-:'L:;S* ~I .>-a .UJ ~j.)1
<sup>33</sup> Unfortunately, Fahd's edition of the Arabic translation appeared too late for Pack to consult
is so literal that several terms, for which the translator was unable to procure an Arabic equivalent, are simply transcribed from Greek into Arabic. The literal rendering of the Greek often obscures the sense of the Arabic text, especially in the passages where Artemidoros referred to Graeco-Roman institutions that were unknown in the Middle Ages. In these cases the translator, who was incapable of grasping the meaning of the Greek text, rendered the original word by word, often losing sight of the sense in the process. Parts of sentences that exist in the Greek original are missing from the Arabic translation, either because the translator could not understand them, or because they were absent from his Greek copy or because they were lost in the course of the transmission of the Arabic text. The Greek text is sometimes not rendered literally. Composite Greek words are replaced by two or more Arabic words, but composite Greek expressions are also sometimes rendered with a single Arabic word. The language of the translator is burdened with foreign terms and expressions modeled after Greek or Syriac usage -in fact, Syriacisms are so frequent that Fahd wondered whether the Arabic translator was working from a Syriac intermediary. The vocabulary and expression of the Arabic text place it between the classical and the colloquial language, as is usually the case with ninth-century translations. The Syriacisms of the text, understandable for a native speaker of Syriac, as well as the evidence of the *Fihrist,* led Fahd to deduce that the author of the translation at hand was indeed I:Iunayn b. Isi)aq. The Arabic Artemidoros is inferior to I:Iunayn's translations of Galen, so perhaps it should be considered a product of his youth, before he had gained proficiency in Greek and mastery of translation techniques. <sup>34</sup>
Rosenthal agreed with Fahd that the translator was I:Iunayn b. Isi)aq, but dismissed the notion of a Syriac intermediary, since no certain traces of it could be discovered:
The overall translation technique is clearly that of I:Iunayn and his school. The often complicated Greek is rendered throughout into lucid and concise Arabic according to the sense. In fact, Ar[= the Arabic translation] is clearer and simpler than G [=the Greek text], at the expense of any literary ambitions Artemidorus may have had. For instance, Ar. consistently uses the verb *dalla* ["it signifies"] where G never tires of varying the expressions introducing the explanation of dreams. However, the translation is by no means philologically exact in our sense of the term. Ar. is full of mistranslations which can only to a small extent be explained as due to a poor text of Gin the manuscript (or manuscripts, we cannot tell) available to the translator. A frequent cause of loose translation is the wrong
during the preparation of his edition of the Greek text; see R. Pack, "Artemidoriana Graeco-Arabica," *TAPA* 106 (1976), p. 307.
<sup>14</sup> Fahd, *Artemidore d' Ephese,* pp. xiv-xxii.
#### 138 CHAPTER FOUR
analysis of sentence structure or the failure to pay attention to such matters as the cases of nouns. It is obvious that the translator often merely guessed at the meaning. We cannot escape the strange feeling that variant readings in G seemingly suggested by Ar. never really existed in any Greek manuscript but were formed in the mind of the translator trying to decipher the text in front of him or listening to someone reading the text to him aloud (Ar. often suggests a Greek text as it was then pronounced, rather than the classical spelling). All such flaws do not entirely eliminate the possibility of I:Iunayn' s translatorship, as he himself felt truly at home only in the technical language of Galen and, for his know ledge of the technical language of dream interpreters, was presumably restricted to the work of Artemidoros. 35
Rosenthal concludes that, since the translator often resorted to guess work, the value of the Arabic translation of Artemidoros for textual criticism is very limited, and that the Arabic text merits further study only because of its importance for cultural history. 36 But this opinion was evidently not shared by other scholars. After the publication of the Arabic translation, a number of articles suggested textual emendations to be used by a future editor of the Greek text. 37
In his review of Fahd's edition, G. Strohmaier sought to reconcile the problem posed by the absence of a translator's name on the unique manuscript of the translation with the evidence in the *Fihrist.* In view of the Syriacisms of the Arabic text, he suggested that I:Iunayn had initially translated it from Greek into Syriac and then one of his students went from Syriac into Arabic. 38 The problem was further investigated by Ullmann, who listed fourteen names of plants, five names of animals, and eight medical terms that were rendered in the Arabic with different equivalents than in other translations known to have been produced by I:Iunayn and his school. Ullmann concluded that two Arabic translations of Artemidoros had been produced: one by I:Iunayn b.
IsJ:iaq comprising all five books, which was mentioned in the *Fihrist* and is
<sup>35</sup> Rosenthal, "Arabic Translation of Artemidorus," pp. 139-40.
<sup>36</sup>For further comments on the translation reflecting the literature published up to 1968, see Magdi, *Die Kapitel iiber Traumtheorie und Traumdeutung aus dem Kitab* at-ta~r!r *ft 'ilm at-tafs!r des Pi ya' ad-Din al-Jaz!rl (7.1131 Jahrhundert),* pp. 19-22. See also Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 78-86.
<sup>37</sup> See R. Pack, "On Artemidorus and His Arabic Translator," *TAPA* 98 (1967), pp. 313-26; idem, "Artemidoriana Graeco-Arabica," pp. 307-12; G. M. Browne, "Ad Artemidorum Arabum," *Le Museon* 97 (1984), pp. 207-20; idem, "Ad Artemidorum Arabum II," *Le Museon* 103 (1990), pp. 267-82; A. Breen, "Observations on the Arabic Translation of Artemidorus: Book I," *Le Museon* IOI (1988), pp. 179-81; G. W. Bowersock, *Fiction as History: Nero to Julian* (Berkeley, Los Angeles and London, 1994), pp. 145-147.
<sup>38</sup> G. Strohmaier, review of T. Fahd, ed., *Artemidore d' Ephi!se* in *Orientalistische Literaturzeitung* 62 (1967), pp. 270-75.
now lost; and a second one by an anonymous translator that included only three books, which is in fact the surviving Arabic text. 39 In his answer to Ullmann's article, Fahd pointed that al-Dinawari quoted passages from all five books and that his quotations from books 1-3 reproduce the surviving translation, which indicates that it initially comprised all five books. Moreover, Fahd accounted for the discrepancies in the translation of terms listed by Ullmann and demonstrated that only one Arabic translation of Artemidoros existed, and that was the one prepared by I:Iunayn.40
The Arabic translation of Artemidoros is not simply a rendition of the text from one language into another. Artemidoros frequently referred to notions and institutions that were peculiar to the pagan Graeco-Roman civilization of late antiquity. In these passages the Arabic text does not give a translation, but an adaptation of the Greek original. The changes relevant to religion have been collected and discussed by Strohmaier.41 The ancient deities mentioned in the Greek Artemidoros are most often rendered as angels, a practice that seems to be typical of I:Iunayn's school of translation. 42 There are limited examples where the name of the deity is simply transliterated without further explanation, or is omitted and the deity is referred to by his or her attributes. In three instances, the names of Aphrodite and Hermes are rendered with the Arabic names for the corresponding planets.43 In a very few passages, such as the one on dreaming of the crucifixion, it is possible to discern that the translator was a Christian. <sup>44</sup>
I:Iunayn's renderings are not always consistent. The same deity is not always rendered in the same manner.45 In addition, at the beginning the Greek *iereus* (priest) is rendered by the Arabic *imam,* a word with clear Muslim connotations; later on *iereus* is translated as *kahin,* the word for the soothsayers of pagan Arabia. Strohmaier conceded that it is sometimes difficult to decide whether
<sup>39</sup> M. Ullmann, "War l:Iunayn der Obersetzer von Artemidors Traumbuch?," *Die Welt des Islams* 13 (1971 ), pp. 204-11.
<sup>4</sup>° Fahd, "l:Iunayn ibn Isl:iaq est-ii le traducteur des *Oneirocritica* d' Artemidore d'Ephese?," pp. 270-84.
<sup>41</sup>G. Strohmaier, "Die griechischen Gotter in einer christlich-arabischen Obersetzung. Zur Traumbuch des Artemidor in der Version des l:Iunain ibn lsl:ia~,'' in *Die Araher in der a/ten Welt,* ed. F. Altheim and R. Stiehl, vol. 5: I (Berlin, 1968), pp. 127-62.
<sup>42</sup> Ibid., p. 131, n. 2.
<sup>43</sup> Ibid., p. 135.
<sup>44</sup> Ibid., p. 146.
<sup>45</sup> Hermes and Aphrodite are a case in point. For a rendering of their names other than by the equivalent Arabic planets, see ibid., pp. 133-34.
#### 140 CHAPTER FOUR
the Arabic departures from the Greek text merely reflect the translator's limited understanding of the Greek language and his ignorance of the Graeco-Roman context within which Artemidoros's interpretations functioned. 46 But he considers a number of them as the product of a deliberate choice, motivated by the translator's desire to protect the reputation of the ancient Greek authors, to be considerate toward Orthodox circles and to increase the esteem of Christianity in the eyes of Muslims.47 l;Iunayn's adaptations were made at the expense of a more faithful rendering of the Greek text and seem to have been criticized by al-Jfil).i'.?, a Muslim intellectual who was l;Iunayn's contemporary and also frequented the court of Baghdad.48
In spite of the changes meant to make Artemidoros's text more appealing and understandable to medieval readers who followed monotheistic religions, several passages of the Arabic translation are incomprehensible without recourse to the Greek text. The difficulties that later Arabic authors faced in understanding the Arabic translation become apparent if we examine Artemidoros's afterlife in subsequent Arabic dreambooks. Al-Dinawari quoted extensively from the Arabic translation of Artemidoros. Though he generally followed the translation faithfully, he often quoted the Arabic text with slight changes if its meaning was difficult to grasp. Other authors, who were not as literal as al-Dinawari in quoting their sources, did include interpretations that ultimately originated in Artemidoros's Arabic translation, but omitted the problematic phrases altogether.49
<sup>46</sup> Ibid., pp. 130-31.
<sup>47</sup> Ibid., p. 156.
<sup>48</sup> Ibid., pp. 155-56.
<sup>49</sup> A good example is the interpretation of the crucifixion. Stroh maier used the Arabic rendering of the corresponding Greek passage as an example of J:Iunayn' s altering the meaning of the text which should be attributed to his Christian faith (ibid., p. 146): "Wo Artemidor von der Kreuzigungsstrafe redet, haben die Assoziationen, die ein Christ damit begreiflicherweise verbindet, die Obersetzung beeintrachtigt. Filr reiche Leute ist es schlimm, sich im Traum gekreuzigt zu sehen, yuµ VOl yap <Haupouv-tm Kat "tCt~ crap Ka~ U1toAA UOUcrt Ol crmupco0EV"tE~ (because the crucified are stripped naked and Jose their flesh) [Artem. ii.53]. J:Iunain formuliert die Begrilndung pietatvoller: *der Gekreuzigte wird niimlich nackt gekreuzigt, und sein Karper veriindert sich* ("-'.I-:' ~J L.:.L:..>-" ~ ...,..,~1 ,:,I dJ.:iJ [Fahd, ed., *Artemidore d'Ephese,* p. 332, I. 9]). Nun ist es aber nicht mehr unmittelbar einsichtig, warum der Traum for den Reichen ungilnstigt ist, und das muB J:Iunain auch empfunden haben. Er sucht namlich die Bedeutung des Traumes durch einen Zusatz ausdrilcklich sicherzustellen: *und deswegen weist es auf den Verlust ihrer Hahe hin, wenn sie sich gekreuzigt sehen* [.;\_] ~Ll: IJI \_, ljl rA-'J-A I Jl..u......A .\_...k JJ dJ jJ J l\_,\_..,...L:. (Fahd, ibid., p. 332, I. 11)]." Al-Dinawari copied this passage into his own dreambook. The meaning of the phrase.....:....\_,. ~J (his body changes) was apparently hard for him to
So far, the only effort to demonstrate the influence of Artemidoros on Arabic dream interpretation through textual examples has been undertaken by Abdel Daim, in a book published a year before the discovery of the Arabic translation of Artemidoros. Abdel Daim was therefore limited to comparing *al-Muntakhabwith* excerpts from Artemidoros's Greek text in French translation.so However, given the adaptations evident in l;Iunayn's translation, which altered some of the nuances as well as the contents of the Greek original, any investigation of Artemidoros' s influence on Arabic dream interpretation based on the Greek text without recourse to its Arabic version could be misleading.st In addition to the first three books that survive in Istanbul University *Ar. yazma* 4726, the Arabic translation of the last two books can be retrieved, at least in part, from the extensive quotations of Artemidoros given in al-Dinawari and in al-Damiri's fourteenth-century zoological encyclopedia, the *ljaydt al-f:wyawdn* (Life of Animals).s2 This means that almost all l;Iunayn's translation
comprehend, because in both manuscripts of al-Dinawari that I examined it appears as ~.J ....:.~(his color changes); cf. *Esad Efendi* 1833, fol. 167b and BN *arahe* 2745, fol. 205a. *Al-Muntakhah* (p. 161), which copied al-Dinawari, omitted the bothersome phrase and limited the explanation of the interpretation to the phrase ~ *.!l..c.* ~ .\_,.,\_,.Ld I .:, *'ii* [sic] (because the crucified is crucified naked). Al-Nabulusi (vol. 2, p. 40), who also copied al-Dinawari, included the interpretation but omitted its justification, though he did repeat the phrase added by I;Iunayn: ~J .~'ii I ~.J ~.!>-"I J~ ..\_..h J ~ <....:.LJ •(.SJ.! (For the rich it is a bad sign, and indeed it indicates the corruption of their affairs). Ibn Shahin (no. 2276) also omitted the problematic explanation of the interpretation: *\_:,''ii* .,r..i-.i.JI ..\_..h ~" ~ ..,.,...,..~ ~ r-1 L.., .L::•-: •. dJ ~I \_:,I µ.J .L..:.~~ ~ .\_,.,~I (And it is said that crucifixion for the rich who are not high ranking dignitaries is a sign of poverty, because the crucified is crucified naked). For a more detailed analysis of the interpretations on crucifixion in the Greek and Arabic tradition, see also chapter 5.
50 A. Abdel Daim, *L' oniromancie arahe d' apres lhn Sirin,* pp. 150-61. The French translation used by Daim is H. Vidal, *Artemidore: La clef des songes ou /es cinq livres d' interpretation des songes, reves et visions* (Paris, 1921).
*<sup>52</sup>*Al-Damiri also mentions other authors of dreambooks, including al-Dinawari and the Persian Jamasb, who is cited by al-Dlnawari together with Artemidoros. It is therefore possible that al-Damiri lifted his quotations, not directly from I;Iunayn's Arabic translation, but indirectly from
<sup>51</sup>For example, anyone familiar with the Greek text of Artemidoros' s introduction to book i who would read A. Miquel' s summary of the same introduction made from the Arabic translation (A. Miquel, "Reflexions sur le *Livre des songes* d' Artemidore d'Ephese-I;Iunayn b. Isi)aq," *Studia /slamica* 52 [1980], pp. 89-103) would immediately realize that the meaning of several Greek sentences has shifted in its Arabic translation. In his translation of book i, I;Iunayn seems to have rendered certain Greek terms on the theory of dream interpretation, not by choosing an Arabic word that would be as close to the Greek as possible, but by employing the Arabic terminology on dream interpretation that was already current in the 9th century. A case in point is the Greek *enypnion,* rendered with the Quranic *adhghiith a}iliim,which* literally means "confused dreams" (see Strohmaier, "Die griechischen Gotter," p. 149).
survives. The examination of the relationship between the Arabic translation of Artemidoros and later Arabic dreambooks remains to be undertaken, but it lies outside the scope of the present study.
The stylistic particularities of I:Iunayn's translation and the alterations that its passages underwent at the hands of subsequent Arabic authors explain why Artemidoros's interpretations that occur in the *Oneirocriticon* are phrased so differently from the ancient Greek text.53 This is the reason why the Greek *Oneirocriticon* (the Arabic sources of which did not include the Arabic translation of Artemidoros but only Arabic dreambooks that were directly or indirectly informed by Artemidoros's wisdom) does not contain interpretations that perfectly correspond to I:Iunayn's text, either. However, I:Iunayn's choice of words can sometimes elucidate the expressions found in similar interpretations in the *Oneirocriticon.*
#### *The Arrangement of Chapters*
The overwhelming majority of the surviving Greek dreambooks are arranged in alphabetical order. Only three, Artemidoros, the *Oneirocriticon,* and the dreambook of Manuel Paleologos,54 are not-each of these three has its own arrangement. In his introduction, Artemidoros described the sequence in which he would discuss the various dream symbols, because he intended to break
al-Dinawari. Given the use of al-Damiri's quotations in Ullmann's debate with Fahd, it is of some importance to examine whether this is indeed the case. For al-Damiri's sources, see J. de Somogyi, "The Interpretation of Dreams in Ad-Damiri's *ljayiit a/-}Jayawiin," IRAS* (1940-41 ), pp. 2-3.
<sup>53</sup> For a further example, see Lamoreaux, "Dream Interpretation in the Early Medieval Near East," p. 123: Ibn Sina, who relied heavily on the Arabic translation of Artemidoros for the compilation of his own work on dream interpretation, omitted or rephrased several interpretations found in the Arabic version of the Greek text because in these instances I:Iunayn's Arabic made little sense.
<sup>54</sup> Only the systems of Artemidoros and the *Oneirocriticon* will be discussed; the dreambook attributed to Manuel Paleologos does not, at least in its present form, seem to follow any clearly discernible arrangement. The published text (ed. Delatte, *Anecdota Atheniensia,* vol. I, pp. *S* 11-24) is based solely on *Paris. gr.* 2419, which also contains the *Oneirocriticon* just before Manuel's dreambook. Drexl 's critical edition of the *Oneirocriticon* has a different sequence of chapters. It is conceivable that the chapters of Manuel's dreambook, which might have had a discernible arrangement at some earlier stage, were similarly rearranged. The fragments from Manuel Paleologos's dreambook added at the end of the text of the *Oneirocriticon* in *Leidens. Voss.* 49 are so few and so incongruously arranged (in a different way than in *Paris. gr.* 2419), that, even with the additional evidence from the *Leidensis,* one cannot arrive at a conclusion regarding the sequence of chapters originally intended by the author.
away from the method of classification used by his predecessors (asterisks indicate lacunae in the original text):55
Έξης ύποθησόμεθα πώς δεῖ κρίνειν τοὺς όνείρους. ἕξει δὲ τάξιν ή πραγματεία τοιαύτην. ούχ ώσπερ οι παλαιοί από θεών άρξώμεθα, καν άσεβείν τινι δοκώμεν, άλλα πρός το άναγκαιον τῆς ἀκολουθίας ἀποβλέποντες ἀρξόμεθα πρώτον ἀπὸ τοῦ γεννᾶσθαι, ἔπειτα \*\* ἀνατρέφεσθαι, ἐξῆς περὶ σώματος καὶ τῶν ἐν τῶ σώματι μερών προσγινομένων και απογινομένων και αὐξανομένων καὶ μειουμένων και άλλοιουμένων είς μορφὴν έτέραν ή είς ύλην, είτα περὶ διδασκαλίας τεχνών παντοδαπών και έργων και ἐπιτηδευμάτων, εἶτα περὶ ἐφηβίας, περὶ γυμνασίων [κατ΄ είδος], περὶ ἀγώνων, περὶ βαλανείου καὶ λουτροῦ παντοδαπού, περὶ τροφῆς πάσης ὑγρας τε καὶ ξηρας, περὶ μύρων καὶ στεφάνων, περὶ [ἀφροδισίων] συνουσίας, περὶ ὕπνου. ταῦτα μὲν περιέξει ή πρώτη βίβλος: ή δὲ δευτέρα \*\* περὶ ἐγρηγόρσεως ἀσπασμάτων κόσμου παντὸς ἀνδρείου καὶ γυναικείου άέρος και των περὶ ἀέρα, περὶ κυνηγεσίας περὶ πλοῦ περὶ γεωργίας, περὶ δίκης, ἀρχῆς δημοσίας καὶ λειτουργίας, στρατείας, θεών τιμῆς καὶ περὶ θεών, περὶ θανάτου, καὶ εῖ τι ἄλλο προϊών ὁ λόγος ὑπομνήσει.
Next, we shall explain how to classify dreams. The treatment will be arranged in the following order. We shall not begin, as early writers did, with the gods, even if we strike some people as being irreverent but, observing the natural sequence of events, we shall begin first with birth, then \*\* education. Then we shall treat the body and the parts of the body-those which are added, subtracted, grown, diminished, and changed into a different shape or substance. After that, we shall talk about the teaching of trades of all types, pursuits, and occupations; then, about the life of the ephebe, bodily exercises [according to types], contests, the bath, washing of every kind, all food, liquid as well as solid, unguents and garlands, sexual intercourse, and sleep. This material will be treated in the first book. The second book, however. \*\* about waking up, greeting people, all masculine and feminine clothing, climate and climatic phenomena, about hunting, fishing, sailing, farming, about legal proceedings, public office, and liturgies (i.e., public service performed at private expense), military service, the worship of the gods, and about the gods themselves, death, and anything else that will be suggested in the course of the exposition.
The pattern which Artemidoros rejected was the traditional method of organizing knowledge in an accessible form before the introduction of alphabetization. Thematic arrangement of knowledge in antiquity was followed by Aristotle and Pliny, but the Greek-speaking world had already started breaking away from it in the second century A.D.56 By the tenth century, when the Oneir-
56 The earliest known alphabetically arranged Greek dictionary is that by Diogenianos of Heraklea, who lived during Hadrian's reign. For the early history of Greek lexicography, see
<sup>55</sup> i.10; Pack 19, 5-24. The passage is quoted according to Artemidoros, Interpretation of Dreams, trans. White. Whatever Artemidoros's original intentions, the sequence of his five books lacks orderly classification because of the incremental manner in which the books were written: at the beginning Artemidoros intended to write two books, the organization of which is set forth in the introduction to book i. He then wrote a third one to cover the omissions of the first three. He later added books iv-v to include new material, and explain the interpretations already set forth in the first three books.
#### 144 CHAPTER FOUR
*ocriticon* was written, thematic arrangement of the type avoided by Artemidoros had generally been abandoned in favor of an alphabetical sequence of topics. 57 The *Oneirocriticon,* however, is organized thematically.
Alphabetical organization was not as widely employed in the Arabic-speaking world. The nature of the Arabic language, in which the majority of the words can be broken down to a three-consonant root, made it difficult and impractical to employ a system of strict alphabetical order in compiling a dictionary, a difficulty reflected in the history of Arabic lexicography.58 Not surprisingly, dream-interpretation literature in Arabic was always organized by subject from its origins until the thirteenth century, when alphabetical classification of dream symbols was introduced by lbn Ghannam (d. 1294). Even after lbn Ghannam, the standard format of an Arabic dreambook included a long introduction on the theory and principles of dream interpretation, followed by the interpretations of specific dream symbols arranged according to subject in a hierarchical manner, beginning with God, Heaven and its inhabitants, such as angels, prophets, and the companions and followers of the Prophet Mul).ammad, and continuing with earth and its inhabitants, i.e., man and the parts of the human body, human activities, and every-day objects, fauna and flora. <sup>59</sup>
59 Variations within this general frame were possible. The most usual one is to treat the sky and heavenly bodies immediately after the godhead and before the angels and holy figures, as is done by lbn Shahin. Al-Dinawari discusses the godhead and the heavenly inhabitants at the beginning, but, as he organizes his treatise around the full circle of a human life, the hereafter is discussed at the very end, right after death. It is impossible to know whether the Arabic-speaking world invented the hierarchical system of classification independently, as a matter of common sense, or became acquainted with it through the Arabic translations from the Greek, or even earlier, through contact with the Syriac- and Coptic-speaking heirs to the Graeco-Roman intellectual tradition. For the survival of this system of arrangement in medieval Coptic lexicography, see A. Sidarus, "Onomastica Aegyptiaca: La tradition des lexiques thematiques en Egypte a !ravers les ages et les langues," *Histoire, epistemologie, /angage* 12:1 (Paris, 1990), pp. 7-19 and idem, "Les lexiques
Hunger, *Byzantine logotechnia,* vol. 2, pp. 407-30. In Western Europe, thematic lexicography was widely employed by medieval Scholasticism; see T. McArthur, 'Thematic Lexicography," in *The History tif Lexicography,* ed. R. P. K. Hartmann (Amsterdam-Philadelphia, 1986). pp. 157-66.
<sup>57</sup> The tradition of thematic arrangement persisted in medical writings. For example, the medical compilation by Theophanes Nonnos, which dates from the reign of Constantine VII, around the same time as the *Oneirocriticon,* the material is arranged according to subject, following the parts of the human body in order from head to toe. However, this arrangement applies strictly to the human body, and not the universe, as is the case in the *Oneirocriticon.* Nonnos's text was published twice: *Nanni, medici c/arissimi, De omnium particu/arium morborum curatione fiber, nunc primum in lucem editus, et summa diligentia conversus per H. Martium* (Strasbourg, 1568); reprinted with additional commentary in *Theophanis Nonni Epitome de curatione morhorum Graece ac Larine, Ope codicum manuscriptorum recensuit notasque adjecit .!. St. Bernard,* 2 vols. (Gotha and Amsterdam, 1794-95).
<sup>58</sup> See J. Haywood, *Arabic Lexicography* (Leiden, 1965).
The arrangement of the *Oneirocriticon,* which is different than that of Artemidoros and defies its contemporary Byzantine practice, betrays its kinship with Arabic dreambooks.60 A detailed comparison of the organization of the material in the *Oneirocriticon* and the Arabic dreambooks can shed light on the changes made by the Greek author in order to adapt Muslim concepts for Christian readers. It can also explain the position of certain chapters in the *Oneirocriticon* which may at first glance seem arbitrary.
Chapters 5-14 of the *Oneirocriticon* discuss religious notions. With the exception of chapters 6-7 and 13-14, they purport to be based on "Indian," that is, Christian, sources, and indeed appear to present a Christian point of view. They are followed by three chapters "On Judges and Measures" according to the Indians, Persians, and Egyptians respectively (chapters 15-17), and by a number of chapters on hair and various parts and functions of the human body (chapter 18 and further). Chapters 11, 12, 17 and 18 in Drexl's text are somewhat mislabeled, since they discuss more than their titles indicate. The contents of chapters 5-18 are as follows:
5. On the resurrection of the dead according to the Indians
6. On the resurrection of the dead according to the Persians
7. On the resurrection of the dead according to the Egyptians
8. On Paradise according to the Indians
9. On Hell according to the Indians
I 0. On the angels according to the Indians
11. On the prophets, apostles and teachers according to the Indians: Jesus Christ; being ordained patriarch, priest or deacon; reading from the Holy Writ to the people; monastic tonsure.
12. On various faiths according to the Indians: performing the duties of a (Christian) priest; reading books to the people; ringing the sounding-board; building a church
13. On faith according to the Persians
14. On faith according to the Egyptians: interacting with the pharaoh
onomasiologiques greco-copto-arabes du Moyen Age et leurs origines anciennes," *Lingua Restituta Orienta/is: Festgabe fiir Julius Assfalg* (Wiesbaden, 1990), pp. 348-59.
<sup>60</sup> The discomfort of later Greek scribes of the *Oneirocriticon* with this unusual arrangement is evident in their efforts to reorganize the sequence of chapters. Thus, in two Greek manuscripts, *Paris. gr.* 2419 and *Bononiensis (Bibi. Univ.)* 3632, the chapters on heavenly bodies and heavenly phenomena such as the clouds, winds, and rain and their consequences, are made to precede the chapters on godhead and religious notions, evidently following a mindset similar to that of later organizers of Aristotle's works who made the *Physics* precede the *Metaphysics.* In addition, the scribe of *Zagora* 89 attempted to organize the dream symbols alphabetically, without being able to break away totally from their previous thematic arrangement.
17. On judges and judgments according to the Egyptians: having a fever; being sick in the belly; the dreamer's hands becoming longer; the dreamer's hair becoming partly white; the dreamer's legs becoming longer and stronger
18. On hair according to the Indians: dreaming of an unknown man; dreaming of a youth; dreaming of an old man; becoming gray-haired; becoming whitehaired; obtaining longer hair; shaving or cutting the hair
The order in which these Christian dream symbols are presented in the *Oneirocriticon* closely follows the order of Muslim dream symbols in Arabic works. Arabic dreambooks start with a chapter on dreaming of God. The discussion of such a dream unavoidably touches on theological matters, such as God's lack of visual attributes, which is part and parcel of the general iconoclastic attitude prevalent in Islam.61 The *Oneirocriticon* lacks a chapter on dreaming of God, either because the contents of such a chapter proved too difficult for the Greek author to adapt, or because he found it unacceptable to discuss the God of the Christians in the same terms as the God of the Muslims. He therefore started his interpretations with the next chapter of his Arabic model, the resurrection of the dead. The *Oneirocriticon* then treats Paradise, Hell, and the angels, as it is done in the Arabic dreambooks.
Chapter 11, "On the Prophets, Apostles and Teachers" (whose nomenclature and sequence is lifted from St. Paul62) corresponds to a chapter on the prophets and other holy figures, such as the companions of the Prophet Mul).ammad and the next generation of Muslims that is found in Arabic dreambooks. The *Oneirocriticon* discusses dreaming of Jesus Christ in the chapter on the prophets because this is where dreaming of him is interpreted in Arabic dreambooks, since Islam considers him a prophet. 63 The interpretations of being ordained
<sup>61</sup> For the theological approach in the chapter on dreaming of God from *al-Muntakhah* (which was in fact copied from al-Khargushi's *al-bishiira wa-al-nidhiira* ), see I. Zilio Grandi, "II problema della visione di dio secondo ii manuale di onirocritica *Muntakhah a/-kaliimfi tafslr al-alj/iim,"* pp. 69-81 (with an appendix quoting the entire chapter from *a/-Muntakhah* in Italian translation).
<sup>62</sup>0£pt npo<1rr11rov, *anom6/cwv* Kai otfocrKa/cwv. Cor. I, 12:28-31; Eph. 4: 11-12; see also Brackertz, *Traumhuch des Achmet,* p. 236, n. 51.
<sup>63</sup> The interpretation of Jesus in the *Oneirocriticon* seems to be based, not on the Islamic interpretation of 'Isa, but on the Arabic interpretation of godhead and/or of the Prophet MuJ:iammad. Given the impossibility of knowing what exactly was written in the Greek author's Arabic models, the reason for his reluctance to adapt the Muslim interpretations of God but his inclusion of interpretations of Christ, another member of the Christian trinity, can only be speculated upon. It is possible that the Arabic interpretations of godhead were supported by Quranic passages for which the Greek author could not find biblical equivalents. It is also possible that his reticence was
or performing the duties of a priest and reading aloud to the people, which are set forth in chapters 11 and 12, correspond to the Muslim interpretations of leading the people in prayer, delivering the Friday sermon *(khutba)* and reading aloud from the Qur'an *(qirii,a).* Ringing the sounding board and building a church are equivalent to performing the call to prayer *(adhiin)* and building a mosque.
Chapters 13 and 14, which discuss faith and the performance of religious duties according to the Persians and the Egyptians, are appropriately placed in this section of the work; dreaming of the pharaoh, however, which is also discussed in chapter 14, as well as dreaming of judges, measures, and judgments, which are covered in chapters 15-17, appear to be somewhat irrelevant. Their discussion does not seem to provide a smooth transition to chapter 18 and those following, which contain interpretations of the human body.
As we have seen earlier, the interpretations of the pharaoh given in chapter 14 are identical with the Arabic interpretations of the caliph, which are usually discussed in an equivalent position in Arabic dreambooks. 64 Caliphs in Arabic dreambooks are often interpreted in the first, religious chapters, because their title literally means that they are the successors of the Prophet as leaders of the Muslim community. 65 As for chapters 15-17, which interpret judges, dreaming of a judge *(qiicj,l)* can also be inserted among the religious chapters of an Arabic dreambook, because a judge ought to be a superior religious scholar, as his appointment is based on his profound knowledge of the Islamic law.66 Arabic dreambooks then turn to humans and the parts of the human body, as is done in the *Oneirocriticon.*
Reconstructing the sequence of contents in the Arabic text that the author of the *Oneirocriticon* was using yields the following: <sup>67</sup>
triggered by specific arguments in the Christian-Muslim religious polemics of the day. For a summary of the Byzantine arguments against the Islamic view of God and references to the works of the two polemicists (Niketas of Byzantium and Theodore Abu Qurra) who wrote earlier than the author of the *Oneirocriticon* on the subject, see A. T. Khoury, *Polemique hyzantine contre /'Islam (VIII-XIII s.)* (Leiden, 1972), pp. 315-52.
<sup>64</sup> Cf. Ibn Shahin, no. 369.
<sup>65</sup> Other placements of the chapter on caliphs are also possible, such as among the chapters covering human activities. The *Oneirocriticon* places the chapter on the priest in an analogous position.
<sup>66</sup> Cf. Ibn Qutayba's table of contents in Kister, "Interpretation of Dreams," pp. 102-3.
<sup>67</sup> Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 317-18, compares the table of contents from the *Oneirocriticon* with those from Ibn Qutayba and al-Qayrawani (cf. his tables 8 and 9). Useful as they are, these tables contain a few minor inaccuracies. For example, the *Oneirocriticon* is said to be missing the interpretations of reading the Qur'an, calling
On this list at least two chapters, 8 on reading the Qur'an and 10 on idol worship, are repeated as chapters 12 and 15. The recurrence under a different guise in Greek of the same subject indicates that, in all likelihood, the author of the *Oneirocriticon* used at least two different Arabic dreambooks that had a similar, though not identical, sequence of chapters. The table of contents of the *Oneirocriticon* as it is recorded in Drexl' s critical edition yields further repetitions: horses are discussed in chapters 152-54 and also 230-31; trees and plants in 151and198-213; clothes in 156-57 and 215-29; houses and buildings in 142-48 and 254. The overall organization of an Arabic dreambook is still discernible in the *Oneirocriticon,* but the contents of Arabic dreambooks are better organized; for example, they do not repeat the same cluster of dream symbols in two totally different parts of the work. <sup>68</sup>It is rather unlikely that
to prayer, and building a mosque. These in fact appear in a Christianized version in chapters 11 and 12 of the *Oneirocriticon.* Interpretations on hills and mountains are also listed by Lamoreaux as missing, but they are in fact contained in chapters 142 and 143, "On Owning Land and Houses" (Drexl 93, 20-99, 5). In addition, a chapter on the interpretative principles is marked as missing in comparison with al-Qayrawani, but interpretative principles are contained in the *Oneirocriticon* in chapters 2 and 301(Drexl240, 6-241, 26), the final chapter of the *Oneirocriticon,* in exactly the same position as the equivalent chapter in al-QayrawanL
<sup>68</sup> Of the Arabic dreambooks under examination, only the earliest four are thematic. Al-Nabulusi's compilation is alphabetical, but his difficulty in abiding by a strict alphabetical order and of breaking away from the long tradition of thematic arrangement is evident: though he brings together in one chapter all the dream symbols that begin with the same letter, the internal structure of each chapter is not alphabetical but thematic.
the author of the *Oneirocriticon* consulted a single Arabic source that was full of repetitions. It is more probable that the somewhat confused structure of the *Oneirocriticon* was the result of the Greek author's consulting at least two Arabic dreambooks at the same time. Possibly, the hierarchical organization of the Arabic dreambooks was clearer to him at the beginning, where he managed to follow it closely. Later, however, he either lost sight of the system's intricacies, or he did not care to observe them. He therefore combined the structure of the two sources he was using without establishing a consistent organization for his final product.69
Though the sequence of chapters in the *Oneirocriticon* is generally unrelated to that in Artemidoros, a distant echo of Artemidoros 's way of thinking might have survived in one instance: the *Oneirocriticon* discusses courtship, women, and marriage (chapter 124) right after the interpretation of killing and death (chapters 120-23). Artemidoros also discusses marriage (ii.65) after a series of chapters pertaining to death (ii.60-64) and comments on the transition in these words: £nn8ii *Kai* 6 yaµ°'; fotKE eavfrtO? *Kai* Uito eavfrrou crriµai-VE'tat, EV'tffuea KaAfil~ EXEtV i]y11craµ11v £mµvricr9ftvm amou (Since marriage is similar to death and is represented by death, I thought it would be appropriate to treat the subject now).70 One wonders whether this coincidence is the result of several authors' s faithfulness to the tradition of a genre or a trick played by the human subconscious. <sup>71</sup>
#### *Principles of Dream Interpretation*
Both Artemidoros and the Arabic dreambooks begin with an introduction expounding on the principles of dream interpretation. Of the *Oneirocriticon* 's four introductions, only the one ascribed to Syrbacham, the dream interpreter to the "king of the Indians," contains information pertinent to the theory of dream interpretation that is comparable to the contents of the introductions in
<sup>69</sup> Further evidence that the *Oneirocriticon* is a compilation of more than one Arabic dreambook is the obvious inconsistency of chapters 96 (Drexl 57, 11 ff.) and 139 (Drexl 92, 1 ff.) with the principles expounded in the theoretical conclusion of the work in chapter 301 (Drexl 241, 1-14). According to chapter 96, a dream dreamt at the sixth hour of the night was fulfilled six months later, while according to chapter 301 it should have been fulfilled five years later. The dream narratad in chapter 139, which was dreamt in the ninth hour, was fulfilled three months later, while according to chapter 301 it should have been fulfilled a year later.
<sup>70</sup> Pack 188, 20-22.
<sup>71</sup> The relationship between love and death *(eros* and *thanatos)* has been amply explored by the literature of romanticism and modern literary criticism.
#### 150 CHAPTER FOUR
Artemidoros and the Arabic dreambooks. It briefly discusses the provenence of dreams and stresses that the interpretation of a particular dream depends on the wealth, social position and sex of the dreamer,72 as well as on the season of the year during which it was dreamt. Syrbacham's introduction differs from that of Artemidoros in that it presents a Christian justification for dream interpretation. Compared to the extensive theoretical introductions in Artemidoros and the Arabic dreambooks, Syrbacham's introduction is also very meager.13
Supplementary principles on the theory of dream interpretation are presented in the very last chapter of the *Oneirocriticon,* EK'. 'tON I1epcr&v Kai Ai yun'tiffiv 7tEpt 8tacrKE\lfEffii; Kai Kpicre(J)i; OVEtp(J)V oµo<j>rov(J)i; (From the Persians and Egyptians Unanimously, on the Examination and Interpretation of Dreams). Placing an additional chapter on the theory of dream interpretation at the very end of the dreambook defies the arrangement of Artemidoros. It is also unusual, though not unknown, in Arabic works. 74 The similarities between Artemidoros
73 Syrbacham' s introduction occupies I printed page; Artemidoros 's introduction to book i occupies 19 pages; lbn Qutayba's introduction is about a third of the total volume of his work; al-Dinawari's includes 15 lengthy sections ( *maqiiliit);* the introduction to *al-Muntakhah* is 26 printed pages long; the introduction to the brief *Ta'hir al-ru'yii* by pseudo-Ibn Sirin takes up several pages; Ibn Shahin and al-Nabulusi have introductions of 5 and 6 printed pages respectively. The introduction to *al-Muntakhah* is presented in detail by Y. Gouda, *Dreams and Their Meanings in the Old Arab Tradition* (New York, 1991 ), pp. 1-26. It is followed by the interpretation of numerous dream symbols arranged alphabetically and based primarily, but not exclusively, on *al-Muntakhah.* Unfortunately, Gouda's publication adds little to the serious study of Arabic dream interpretation; see the review of Gouda's book by W. M. Brinner in *International Journal of Middle East Studies* 28:2 (1996), pp. 302-4. A combination of the introduction to *al-Muntakhah* and to the dreambook of al-Nabulusi is given in Muhammad M. al-Akili, *Ihn Seerfn' s [sic] Dictionary of Dreams according to Islamic Inner Traditions* (Philadelphia,1992), pp. xv-xxxi. In spite of the title, al-Akili's interpretations are based mainly on the dreambook of al-Nabulusi. For a French translation of the introduction to the *Ta'hir al-ru'yii,* see Ibn Sirin, *Interpretation des reves,* trans. Penot, pp. 1-9; Italian translation in lbn Sirin, *Li bro de/ sogno,* trans. Zilio Grandi, pp. 4-9.
74 I know of only one Arabic dreambook that expounds the theory of dream interpretation in both its initial and its concluding chapter, and that is the dreambook of al-Nabulusi, whose epilogue (vol. 2, pp. 349-62) not only enumerates the sources of the work, but also identifies the most propitious hours for interpreting a dream, exhorts dreamers to narrate their dreams with accuracy, names the sources of dream interpretation, gives some further advice to dream interpreters, and cites a catalogue of dream interpreters *((ahaqiit al-mu'ahhirfn),* evidently copied from the 15th introductory *maqii/a* ofal-Dinawari. To this one could perhaps add al-Qayrawani's *a/-Mumatti'* that discusses the theoretical aspects of dream interpretation both in its first part (which constitutes about half of the total work) and in chapter 29 of the 30 that make up its second part, primarily dealing with specific dream symbols and their interpretation. See Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 86-96 and p. 318 (Table 9). The final chapter of Drexl's edition of the *Oneirocriticon* appears as an introductory chapter in three Greek manuscripts:
<sup>72</sup>Cf. Artemidoros i.9 (Pack 18, 16-20); iv, 2 (esp. Pack 245, 6-23). For a summary of the principles of ancient dream interpretation, see A. BoucM-Leclercq, *Histoire de la divination dans* /' *antiquite,* vol. I (Paris, 1879; rpt. Brussels, 1963), pp. 291-329.
and the *Oneirocriticon* have been obvious to the readers of the two works since the Middle Ages. In the eleventh century, the scribe L 1 of *Laurent. P lut.* 87, 8, the oldest surviving manuscript of Artemidoros, copied on its margins two passages from the *Oneirocriticon* (Drexl 240, 21-241, 17 and 240, 9-12) in order to comment on relevant paragraphs from Artemidoros's introduction (Pack 16, 10 and 18, 17).75 But should this be taken as an indication that the author of the *Oneirocriticon* used the Greek text of Artemidoros as one of his sources?
The *Oneirocriticon* and Artemidoros agree that the interpretation of a dream can vary depending on the identity of the dreamer and the time that it was dreamt. According to Artemidoros (i.9, Pack 18, 16-20):
Aucrt1:EAEi; *b* c'iv E'iT1, OU µovov 8£ Aucn1£A£i; aUa Kat avayKatov i:0 i86vn 1:0V ovEtpov Kat i:0 UitoKptvoµ£vcp, £ii:icr1acr8m i:ov ovnpoKptTI)V i:ii; i:£ £crnv 6 i8wv 1:0V OVEtpov Kat 0 n itpacrcrEt Kat omoi; y£yov£ Kat 0 n EXEl nf]µa Kat omoi; EXEt crffiµmoi; Kat ~crnvoi; iiAtKiai; y£yov£.
It is profitable-indeed, not only profitable but necessary-for the dreamer as well as for the person who is interpreting that the dream interpreter know the dreamer's identity, occupation, birth, financial status, state of health, and age.
In addition, Artemidoros (i.7, Pack 16, 10-17, 2) admonishes dream interpreters as follows:
"En ltclcrt 1:0ti; itp6811Aov aii:iav EK<j>Euyoucrt itpOCTEXElV xpii, £av 1:£ VUK10i; £av 1:£ i]µ£pai; opa8ij, µ118Ev 8ta<j>EpEtv voµi.l;;onai; Eti; ii:poyvwcrtv i;iJv VUK1:a i:fii; i]µ£pai; µ118£ 1:i]v 8EiA11v £crii:£pav i:f\i; 8Ei.A.11i; ii:pwiai;, £av cruµµ£1pwi; ni; £xwv [ Tiii;J 1po<1>fii; Ka8£001J EltEl cii YE aµE1p0t i:po<j>at 000£ itpOi; aVtji tji ECfl ii:ap£xoucrt v t8Et V 1:0 aA118£i;.
Next, one must pay attention to all those dreams in which one can perceive no obvious motivating factor, holding that there is no difference with regard to prediction between night and day, or between late evening and early evening, if a man falls asleep after having consumed a moderate amount of food. For immoderate eating prevents one from seeing the truth even at dawn. <sup>76</sup>
76 For a number of references to ancient literature confirming that in antiquity the dreams dreamt closest to dawn were considered to be the most truthful, see Artemidoros, *Inte1pretation of Dreams,* trans. White, p. 69, nn. 20 and 21. Cf. also the views on the subject that al-Dinawari
*Leidens. Voss.* 49, *Paris. gr.* 2419, and *Bononiensis (Bibi. Univ.)* 3632. Most of the general principles of dream interpretation from the final chapter of Drexl' s critical edition can be found in the introductory chapter of the abridgment of the *Oneirocriticon* from *Paris. Suppl. gr.* 690.
<sup>75</sup> Pack, *Artemidori Daldiani Onirocriticon Libri V,* pp. vi-vii. The similarities between the two works have also been noted on one copy of the *Oneirocriticon, Zagara* 89 (fol. 3b), where we read: TO napov ovnpoKptnKov ffi<; npo<; TO 1:0u 'ApTEµt6c.ilpou aTEAov [sic for aTEAE<;] £miv. 6 yap' ApTEµi6wpo<; Kat muTa ltEpu:xn itpom:n Kat Tel TOU 'A<npaµllfUXO'll (The present dreambook is, compared to Artemidoros, incomplete. For Artemidoros contains not only this, but also what Astrampsychos has).
Both principles are expounded in Syrbacham's introduction (Drexl 2, 10-16"), as well as in the epilogue of the Oneirocriticon (Drexl 240, 8-241, 14). In the Oneirocriticon, however, and especially in its concluding chapter, they are expanded with details regarding the correlation between the time that a dream was dreamt and its fulfillment that are absent from Artemidoros:
Τα όνείρατα, καθά ήδη προείρηται, είς διάφορα κρίνονται πρόσωπα διαφόρως. έφ' ένὶ γὰρ ὀνείρατι αλλη κρίσις ἐπὶ βασιλέως καὶ ἀλλη ἐπὶ λαοῦ, ἀλλη γεωργού και αλλη στρατιώτου, έτέρα του μεγιστάνου καὶ ἐτέρα τοῦ πένητος, αλλη κρίσις ἐν ἀνδράσι καὶ ἄλλη ἐν γυναιξίν, ἄλλη ἐν θέρει καὶ ἑτέρα ἐν χειμώνι, έτέρα ἐν τῇ αὐξήσει τῆς ήμέρας καὶ ἐτέρα ἐν τῇ μειώσει, ἀλλη ἐν αύξήσει της σελήνης και άλλη έν μειώσει. και άλλα των όνειράτων ταχέως καὶ ἄλλα βραδέως κρίνονται. καὶ ταῦτα, καθώς ἤδη προείπομεν, προεκρίθη ἐν έκάστω προσώπω, ήτοι βασιλέως και πένητος, μεγιστάνου και κοινού λαού, γυναικός και άνδρός, γεωργού και στρατιώτου. και ούκέτι χρεία περὶ τοῦ αύτου διαλαμβάνειν έφ έκάστη γαρ κρίσει προηρμηνεύθη έκαστον.
Και έντευθεν έρμηνευτέον περὶ θέρους καὶ χειμώνος χειμών γὰρ και θέρος έν τοῖς δένδροις καὶ φυτοῖς διακρίνονται. ἐαν γὰρ ῖδη τις περὶ δένδρων τι, εϊ έστιν άρχὴ τῆς βλάστης ἐν ἔαρι, ή ἔκβασις τῆς κρίσεως εἰς καλόν ἐστιν, εί δε έν φθινοπώρω και έν φυλλορροία, είς έναντίον ή έκβασις. καὶ ἐν αὐξήσει των ήμερών πλείον άληθεύει των όνείρων ή κρίσις, έν δε τη μειώσει των ήμερών έλάττων ή ἕκβασις. όμοίως και ἐπὶ σελήνης καὶ ἐν τῷ τάχει καὶ ἐν τῆ βραδύτητι ούτως διακρίνεται.
Και <εϊ τις> έρωτα σε περὶ ὀνείρου, ἀντερώτα τὴν ώραν, ἐν ἢ τοῦτο τεθέαται. από πρώτης γαρ ώρας της νυκτός έως τρίτης κρίνεται έως είκοστου έτους ή έκβασις από δε τρίτης ώρας έως έκτης ή έκβασις μέχρι δεκαπέντε η δέκα έτων ή όκτώ από δὲ ἐκτης ἔως ἐνάτης ή ἔκβασις μέχρι πέντε ἢ τεσσάρων ή τριών ἐτῶν· ἀπὸ δὲ ἑνάτης μέχρι πρωὶ ἔως ἐνιαυτοῦ ἢ ἔξ ἢ τριῶν μηνῶν ἢ μηνὸς ἢ δέκα ήμερών ή ἔκβασις τῆς δὲ αὐγῆς αὐθημερὸν ἢ ἐμετὰ> δύο ἢ τρεῖς ήμέρας και ούτως ή έκβασις. περὶ δὲ τῶν ώρων τῆς ήμέρας οὐδὲν όρίζομεν μέχρι γαρ έβδόμης της ήμέρας το πλέον ή έκβασις του όνείρατος, εν ή αν ώρα τις είδεν αύτό. τῆς ήμέρας γὰρ τὰ ὀνείρατα, ὅσα μνημονεύει τις ἀσφαλῶς, εἰς έκβασιν τελείαν είσίν ή είς αναμφίβολον μάλιστα δε τα από ένάτης ώρας μέχρι πρωϊ άληθέστερα και ταχύτερα.
Dreams, as was already mentioned, are interpreted differently for different individuals. Because the same dream has one interpretation for a king and another
attributes to the Christians (Appendix 3).
<sup>11</sup> Drexl 2, 10-16: μὴ λογιζέσθω δέ τις, ὅτι μία τίς ἐστι κρίσις και λύσις ἀνος των διαφόρων προσώπων. διότι τα τών βασιλέων όνείρατα οίκείαν έχουσι κρίσιν και λύσιν, και των άρχόντων και πλουσίων όμοίως οίκείαν, και των άπόρων τε και πτωχών αναλόγως οικείαν όμοίως δὲ ἄλλη κρίσις όνείρων γυναικών καὶ ἀλλη κρίσις ἐν θέρει καὶ ἄλλη ἐν χειμῶνος ώρα (And let no one reckon that the interpretation and solution to a dream is the same for different individuals. For the dreams of kings have their own interpretation and solution, and those of the nobles and the wealthy likewise their own, and those of the destitute and the poor their own accordingly. Similarly, the interpretation of women's dreams is different from that of men's dreams. The interpretation is also different during the summer and during the winter season.)
for the people, one for a farmer and another for a soldier, one for a magnate and another for a poor man, and the interpretation is different for men and women, different during the summer and during the winter, different at the beginning and at the end of the day, different when the moon is waxing and when it is waning. Some dreams are fulfilled sooner and others later. These things, as was already mentioned, have been previously interpreted for each individual, that is for a king and a pauper, a magnate and a commoner, a woman and a man, a farmer and a soldier. There is no more need to discuss the same thing, because each case has been treated <in the chapter of> each interpretation .
Therefore, we should now explain about summer and winter. For these seasons are separated according to the <condition of> the trees and the plants. If someone dreams something about trees, if it is the beginning of growth during spring, the outcome of the interpretation will be good; but if it is autumn and during the falling of the leaves, the outcome will be the opposite. And when the day is at its beginning the interpretation of dreams is more truthful, but when the day is ending their outcome is less so. The interpretations are likewise for the waxing and waning of the moon and the swiftness or slowness <of their fulfillment>.
And if someone asks you about a dream, question him about the time that he had it. For, from the first hour of the night until the third, the fulfillment of the dream will occur in up to twenty years. From the third hour until the sixth the fulfillment <will take place> in up to fifteen or ten or eight years. From the sixth until the ninth hour, the dream will be fulfilled in up to five or four or three years. From the ninth hour until the morning in up to one year or six or three months or one month or ten days. <The dreams> of dawn <will be fulfilled> within the same day, or after two or three days. For the hours of the day we do not make such fine distinctions. The outcome of a dream is more truthful and swift, regardless of the time of day that it was dreamt, until the seventh hour. For the dreams of the day, those that one can remember correctly, are fulfilled completely or unambiguously. Especially the ones dreamt from the ninth hour until the morning are even more truthful and swift <in their outcome>.
The idea that a dream should be interpreted according to the dreamer's station in life is expounded in all five of the Arabic dreambooks. It was stressed by Ibn Qutayba *(Yahuda ar.* 196, 2b inf.):
~L..:.....::.J i""+3~ ~ J""L..:JI Jl\_,\_,,..I ...J~l,i 4J\_,\_...,.I ~ ~ ~\_j.)1 ,).~ ... ..:\_,l....)YIJ ...::..uJYI ...J~t..,.J i""+31Jl)J !'"+"'\_,.......AIJ ~~JIJ ~)~IJ
Dreams change from their basic forms due to the difference in the station of the dreamers regarding their social class, profession, rank, religion, concerns and desires, as well as according to the time and season <that a dream is dreamt>.78
<sup>78</sup> Repeated in Ibn Qutayba's introduction ( *Yahuda ar.* 196, fol. 7b int): *.J.L '-=3.rl* I ~ .;...JJ .r=.YI ..,.UJ ~\_, ..i.:..1\_,J .:\_,\_µ ~LdJ r.t\_,l~IJ ~Ld ... : •. ""J .\_rl.:JI ..::..t...,..A .\_j)l....::..;..l,> l..tL>I L; I~ (Dreams change from their basic meaning owing to differences in the social class of the dreamers, as well as their profession, their rank and their religion. <The same dream> signifies mercy for one and punishment for another). The point is further elaborated by citing a number of actual dreams that had been interpreted in the past according to this premise (ibid., fol. Sa). The passage is repeated verbatim in al-Nabulusi, vol. 2, p. 352.
#### CHAPTER FOUR
The same point is elaborated in detail in al-Dinawari's introduction (12th magāla): 79
واذا ستئلت عن رؤيا ملك عدل او ملك متعزز فللا تؤولها برأيك فان تاويل رؤياه ليس كتاويل رؤيا غيره من الرعيته ...
If you are asked about the dream of a just or a great king, do not interpret it according to your point of view, because the interpretation of his dreams is unlike that of the dreams <dreamt> by the rest of his subjects ....
Al-Dinawari explains that, since kings rule the people, the interpretation of their dreams influences the way they govern. Therefore, the dreams of sovereigns are the most truthful ones, as was proven by the two dreams dreamt by the pharaoh and interpreted by Joseph. 00 The dreams of other people should ورؤيـا كل رجل يـعــبّـر ) also be interpreted according to their station in life ا ( على مـر تـبـتــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ fulfillment applies to somebody else close to the dreamer. 92
If a slave has a dream that does not befit his station, this dream will be fulfilled for his master, because the master owns him. And if a woman dreams of something that does not befit her station, then the dream is for her husband, for she was created from his rib. And the interpretation of a child's dream is for his parents. And whatever dream is dreamt by someone who knows a trade or a craft should be interpreted according to the station of the dreamer and according to his forefathers and their craft.
The fulfillment of a dream for a person other than the dreamer is a concept known to Artemidoros (i.2, Pack 7, 17-8, 5) and also applied in the interpretations of separate dream symbols contained in the Oneirocriticon, though it does not appear in its two chapters that expound the theory of dream interpre-
<sup>79</sup> BN arabe 2745, fol. 41b; Esad Efendi 1833, fol. 29b.
<sup>80</sup> For the special significance of royal dreams as assessed by al-Dinawari, see Fahd, "Les songes et leur interprétation selon l'Islam," p. 144 and J. Lecerf, "The Dream in Popular Culture: Arab and Islamic," in The Dream and Human Societies, ed. G. E. von Grunebaum and R. Caillois (Berkeley and Los Angeles, 1966), p. 366.
<sup>81</sup> BN arabe 2745, fol. 41b; Esad Efendi 1833, fol. 30a.
<sup>82</sup> BN arabe 2745, fol. 42a, Esad Efendi 1833, fol. 30a.
tation. <sup>83</sup>*Al-Muntakhab* sets forth a more condensed version of the same principle,84 closely following the wording of al-Dinawari:
~J.J....,.......... .:.,L.b.L...JI l:>J.J ~ ~ l:ij..)1 .\_,.,~I *0-:* ~.:\_,I ~J <Lol..:.... ~ '-51.J lj) ~I\_, 4-:=-L::. Jl\_,\_,..I ...J)l.:;..:;..~ ~ ~j..)1 .:\_,µ~\_)I <L.J ~ rJ L. ..;:..,l.J ljl :;l\_;-11 ..:.U.:i..s\_, ,<L.JL. ~y ~U ~).(.All <L.J ~ rJ L. ·"-:!.J-/Y J.i.bJ I ~J.J *J.=\_,L.:;\_,* ~ *.J-o* ..::...ih 4-:>''1 ~\_,j.J ~ ).(.All
One of <the duties of a dream interpreter> is to make a distinction between the different dreamers. For the dream of the sovereign should not be interpreted in accordance with the dreams of the subjects, since dreams are different from each other according to the difference in the station of the dreamers. If a slave sees in his sleep something that does not befit his station, then the dream is for his master, for he is the one who owns <the slave>. Likewise, if a woman dreams of something that does not befit her station, then the dream is for her husband, for she was created from his rib. And the interpretation of a child's dream is for his parents.
*Al-Muntakhab* adds that, before giving an interpretation, Ibn Sirin used to spend the better part of the day interrogating the dreamer "about his station, himself, his profession, his kinsfolk and his way of life."85 lbn Shahin likewise claims that dreams should be interpreted according to the identity of the dreamer:s6
~J.J ~ .:.,i\_, .~I r-1c- ~ J-9-'""Y~ *Ll).L* <~I>.:\_,~ .:.,i <~>J .<G\_,~ ~ ->"L.:JI '-5~ *'1\_,* .....,........~ L.\_, *"-:!* ~ L.\_, <L.JL:.. ~ .i.:>.i JS
The dream interpreter must be knowledgeable in the fundamental principles of the science of dream interpretation and differentiate the vision of each person according
<sup>83</sup> Drexl 5, 15-17: £av SE -rofrro 0EwpiJcrlJ rrmliiov, Et<; rrpocrwrrov Tciiv iliiwv yov£wv -rou-ro £ci>paKEv (If a child dreams this, the child dreamt it on account of his parents); Drexl 78, 15 ff: Kat o £av llil] rrmliiov civiJ:.1.tKov, Ei<; rrpocrwrrov Ti litciKptcrt<; Ti\<; µ1]TEpo<; (Whatever an underage child dreams, the interpretation pertains to the child's mother); Drexl 80, 21-23: ocra llil] 7tatliia civiJ:.1.tKa, -rou µ£v apcrEVO<; Et<; rrpocrwrrov Ti\<; µ1]Tp0<; litaKpivnm, -rou Ii£ SiJ:.1.rn<; Et<; rrpocrwrrov -rou rrmpo<; (Whatever underage children see, a boy's dream is interpreted as pertaining to his mother and a girl's dream to her father). Cf. also al-Nabulusi, vol. 1, p. 6 sup.: JI "--;>.) . ..:.IJ ilil *.J-o* ·~ .J-ibJ ~ ol.,r-: .i...:<J ~~I ..s~ .LI .JL .. , ... :81 .JI rl-<-1\_, ..S~ JI ~Jj JI ....:;..i..L..,. JI~'-:-'"""~ JI~ JI~ JI o.i..llJ JI~ (Know that a person may dream on behalf of himself, but also on behalf of other people from among his family, relatives, siblings, parent, or on behalf of somebody else who either resembles him, or is his namesake, or has the same trade, or comes from the same region, or on behalf of his wife or his slave).
<sup>84</sup>*Al-Muntakhah,* p. 25 med.
<sup>85</sup>"~J .\_,~ .J.LJ • ... :;,d .. : .. ~J <.........LJ <JL:.. *.J.L"; al-Muntakhah,* p. 15 inf.
<sup>86</sup> Ibn Shahin, Introduction, *fa.yl* 4, p. 11.
#### 156 CHAPTER FOUR
to the dreamer's station and what is appropriate for him and what befits him, and should not put people on the same footing regarding what they dream.
Arabic dreambooks discuss the correlation between the time that a dream was dreamt and its fulfillment in much more detail than Artemidoros. They agree with the ancient Greek dreambook that the closer to dawn a dream is dreamt the more truthful it is. In addition, they address the importance of the season of the year, as is done in the *Oneirocriticon.* Ibn Qutayba writes the following: 87
Abu MuJ:iammad said: "The <interpretation> of a dream differs according to the time <that it was dreamt> .... " Abu MuJ:iammad said: "The most truthful time for a dream during the night is just before daybreak, and midday for a dream during the day. The most truthful season of the year is that of blossoming and when the fruit spring forth and reach maturity, while the least truthful season is winter. And the dream of the day is stronger than the dream of the night."88
The length of time that it takes for a dream to come true is discussed in a different manner in Artemidoros than in the *Oneirocriticon.* Artemidoros says: <sup>89</sup>
npoi; 8£ wui; nuveavoµ£voui; n:OCi(\) XPOV(\) o\. OVElpot anopai VO'\JCil v E'in:oti; av O'tl ocm nov ovnov EV ci>ptcrµ£vou; xpovoti; yi VE'tat, i:aiii:a Kat EV Wt<; un:voti; 6pa0£vi:a i:oti; miw'ii; an:opaivtt xpovoti; .... "Ocra 8£ i:ciiv OV't(OV EV aopicri:oti; Kat µlj Jl:E7tTjYOO'l XPOVOl<; yi VE'tat, aoptO''tco<; Kat an:opai VEl ....
When men ask how much time it takes for dreams to come true, you should reply that everything that occurs within a definite period of time in real life will also, if it is seen in a dream, occur within the same period of time .... On the other hand, things that do not have a definite or fixed duration in real life, if they are seen in a dream, will not have a definite duration when they come true ....
The *Oneirocriticon,* on the other hand, advocates a strict correlation between the time of night that a dream was dreamt and the length of time it takes for its
<sup>87</sup> Ankara, *Is. Saih Sincer I,* 450 I, fol. l 9b; *Yahuda ar.* 196, fol. Sb.
<sup>88</sup> Repeated in *al-Muntakhah,* pp. 11-12 and ibid., p. 24: cf. also ibid., p. 14; repeated in al-Nabulusi, vol. I, p. 5. As for the Greek tradition, though Artemidoros does not attach any particular importance to the season of the year in which a dream is dreamt (cf. iv.I I), other ancient authors (including Aristotle, whose writings on the subject had been translated into Arabic) believed that springtime was a more propitious time for dreams than winter. See A. Bouche-Leclercq, *Histoire de la divination dans /'antiquite,* vol. I, p. 287.
<sup>89</sup> iv.83 (in Artemidoros, *Interpretation (!f'Dreams,* trans. White, included in iv.84); Pack 299, 15 ff.
fulfillment. Al-Dinawari agrees that the closer to dawn a dream is dreamt the more truthful it is and the sooner it will be fulfilled. In addition, he vaguely defines the length of time that elapses between a dream and its fulfillment, though the units of time he suggests do not coincide with those set forth in the Oneirocriticon: 90
قال رسول الله صلى الله عليه وسلم اصدق رؤيا ماكان باسحار وقال عليه السلام اصدق رؤيا ما كان رؤيا النهار لان الله تعالى خصتى بالوحى نهاراً وقال جعفر الصادق اصدق الرؤيا رؤيا القيلولة ... \* ومن راى في أول الليل رؤيا فان صاحبها يصير اليها الى عشرة ايام او شهر او اكثر ومن راى اخر الليل فهو اسر ع ما يكون وابطأوها الى سنة لان الأماد قد قصرت.
The messenger of God, may the peace and blessings of God be upon him, said: "The most truthful of dreams is the dream at daybreak." He also said, may peace be upon him: "The most truthful dream possible is the dream of daytime, because God, who is exalted, favored me with the revelation during the day." Ja'far al-Sadiq said: "The most truthful dream is the one dreamt during the midday nap ...." Whoever has a dream at the beginnings of the night, it will come to pass in ten days or a month or more. And if he has a dream toward the end of the night, it <will come to pass> as soon as possible, and no later than a year, because these stretches of time are shorter.
Though al-Dinawari disagrees with the Oneirocriticon regarding the units of time within which the fulfillment of a dream should be expected, Ibn Shāhīn quotes Ibn Sirin's view on the topic, and it coincides with the opinion expressed in the Greek work. The Arabic dreambook justifies the utmost limit of twenty years for the fulfillment of a dream by adducing that Joseph's dream (apparently the second of the two mentioned in Genesis, " which is the only one described in the Our'an") was fulfilled twenty years later. Such a conclusion cannot be drawn from the narrative of the Bible or of the Qur'an. The source of this piece of information is not quoted, but it could be exegesis or the popular Oisas al-anbivā' (Stories of the Prophets) that were popular in the Muslim world by the eighth century. 33 The relevant passage in Ibn Shāhīn's introduction reads as follows. 94
<sup>90</sup> Introduction, magala 8, BN arahe 2745, fol. 38b; Esad Efendi 1833, fol. 69a.
<sup>91</sup> Genesis 37:9-11.
<sup>92</sup> Qur'an 12:4.
<sup>93</sup> See T. Nagel, Die Qisas al-anbiyat : Ein Beitrag zur arabischen Literaturgeschichten, Inaugural-Dissertation (Bonn, 1967), p. 26.
<sup>94</sup> Ibn Shāhīn, introduction, fasl 3, p. 10.
واقرب ما تخرج الرؤيا اي تظهر الرؤيا اذا رؤيت آخر الليل فإنه ينتظر بها"، وروي أن ابن سيرين قال: من راى رؤيا اول اللليل فإنه ينتظر بها إلى عشرين سنة مماناً دون ذلك، ويقاس على الليل وعلى السنين ويعرف مـا مضى من الليل وينقص من السنين بقدره مثاله، إذا مضى من الليل نصف ينتظر الرؤيا إلى عشر سنين فـمـاّاً دون ذلك، ويقاس على ذلك ومن راي رؤيا بعد الصبح فإنه ينتظر لها مدة شهر وما دون ذلك، وكذلك رؤيا النهار وقد ظهرت رؤية يوسف عليه السلام بعد عشرين سنة، فلأجل ذلك حد أخر انتظار الرؤيا عشرين سنة.
The soonest that <the predictions of> a dream will come to pass, i.e., become manifest, is when it is dreamt in the latter part of the night, and <the dreamer> should expect that. For it is reported that Ibn Sirin said: "If someone has a dream at the beginning of the night, he should expect its outcome in twenty years or something less than that, and he should correlate the <time of> the night and the <number of> years. He should know what portion of the night has passed and subtract from the number of years according to the example. If half of the night has passed, he should expect <the outcome> of the dream in ten years or something less than that, and he should calculate by analogy to that, and <do> likewise <for> the dream of the day. The vision of Joseph, may peace be upon him, became clear after twenty years, and for this reason the utmost limit for the fulfillment of a dream is twenty years."
Despite the discrepancy between al-Dinawari and the Oneirocriticon, the two works agree that the interpretation of dreams about flora depends on the season of the year that they are dreamt: if dreamt while plants are flourishing they are auspicious, but if dreamt while the foliage, flowers, or fruits are fading they are inauspicious. 38 Moreover, they agree that the interpretation of a dream can vary, depending on whether it was dreamt at the beginning or the end of the day, and the Arabic terms ighal al-nahār wa-idbāruh, which occur in the following excerpt from al-Dinawari's dreambook, closely correspond to the Greek auxēsis kai meiõsis tēs hēmeras:99
<sup>95</sup> فانه سنتظر نها [sic] in the printed text; possibly a dittography.
<sup>96</sup> Sic in the printed text; should possibly be corrected to .
<sup>97</sup> Sic in the printed text; should possibly be corrected to و ما .
<sup>90</sup> This principle is further illustrated in the Oneirocriticon with an anecdote (chapter 199, Drexl 156, 6-19): A man dreamt that he was in an apricot tree eating its fruit. Sirim said that the man would be the beneficiary of a rich man's generosity. A second man had the same dream. This time Sirim said that he would be afflicted with sorrow. The same dream received two different interpretations because the first dream was dreamt in spring, the second in the fall.
<sup>99</sup> Al-Dīnawarī, magāla 9, BN arahe 2745, fol. 39a-b and Esad Efendi 1833, fol. 69a-b.
الرؤيا تتقوى فى السنة سبع اشهر \* ... وذلك اذا دب الما فى عروق الاشجار الى ان يسقط اوراقها وخصوصا في وقت طلوع الشمار وقوة الاغـصـان واستـواء الاوراق <و > لا سـيـمـا فـي رؤيـة النبـات لان الاشـجـار اذا كـانـت فـي الاقبال وكسر منها غصن عاد لجنبه غصنان واذا التقطت منها ورقه خرج بجنبها خمس ورقات ۾ فان راي الرجل في اقبال الاشجار والنبات انه التقط ورقة او اقتصب غصنا اصاب بكل ورقة وبكل غصن درهما ↑ واذا التقطها في ادبار السنة خسر بكل ورق وغصن در هما واصابه هم او ضرب ب قال ابن سيرين والكرماني ان الرؤيا اذا سيل عنها في اقبال السنة فهي خير من ان يسال عنها فى ادبارها لان فى اقبال السنة اقبالا وفى ادبار السنة ادبار الرؤيا وكذلك اذا سيل عنها فى اقبال النهار وادباره …
A dream becomes even stronger during seven months in the year. ... This is, from when the water spreads to the roots of the trees until the leaves fall, and especially during the time when the fruits appear, the twigs are strong, and the foliage of the trees becomes complete. < This principle applies> mainly in the dreams about plants, because the trees, when they are flourishing and a twig is broken, grow two twigs from its place, and if a leaf is taken from them, five leaves come out in its stead. If during the flourishing of trees and plants a man dreams that he was collecting leaves or that he was cutting twigs, for every leaf or twig he will receive a dirham. And if he collected them in the period of the year when <nature> retreats, for every leaf and twig he will loose a dirham and will be afflicted with sorrow and beating. Ibn Sirin and al-Kirmani said: "Indeed, if the interpretation of a dream is requested at the advent of the year, it is better than if it is requested at the end of the year, because the advent is prosperity and the end is adversity regarding a dream. 100 Likewise if the interpretation of a dream is requested at the beginning or the end of the day."101
<sup>100</sup> This remark is based on the several meanings of the words idbar (latter part, conclusion, and adversity) and ighal (advent, prosperity).
<sup>101</sup> The rest of the maqāla discusses individual lunar months of the Muslim calendar (cf. al-Muntakhab, p. 5). Al-Muntakhab and Ibn Shāhīn also claim that springtime is a more propitious season for dream interpretation than the fall, but do not specifically apply this to dreams about plants, as is done in the Oneirocriticon and by al-Dinawari. Cf. al-Muntakhab, p. واعسرف الازمنة فى الدهر فاذا كانت الشجر عند حملها ثمـار ها فإن الرؤيا فى ذلك الوقت مرجوة قوية فيها بطء قليل وإذا كانت الرؤيا عند ادراك ثمار الشجر ومنافعها واجتماع امرها فيان الرؤيا عند ذلك ابلغ وانفـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ الرؤيا عند ذلك دون ما وصفت فى القوة والبقاء دون الغاية واذا سقط ورقها وذهب ثمار ها فان الرؤيا عند ذلك اضعف (You should know about the seasons. If the trees are at fruit bearing, indeed the dream in this period is desirable [and] powerful and will soon be fulfilled. If the dream is dreamt when the fruits on the trees are mature and useful and are harvested, indeed the dream at this time is more intense, effective, sound, and appropriate. And if the trees have produced leaves but their fruits have not yet appeared, the dream at this time is inferior in strength and duration than what I have described, and without purpose. And if the leaves are falling and the fruits are gone, the dream at this time is even weaker). Cf. Ibn Shāhīn, introduction, fași 2, p. 10:
#### CHAPTER FOUR
These examples demonstrate that the theoretical principles of dream interpretation set forth in the last chapter of the Oneirocriticon are equivalent to those found in Arabic dreambooks.
What about the second introductory chapter of the Oneirocriticon attributed to the Christian dream interpreter Syrbacham? According to the practice of both Artemidoros and the Arabic dreambooks, Syrbacham's introduction discusses the provenance of dreams, though much more briefly, since the Oneirocriticon does not make a distinction between false and true dreams and omits the discussion of their respective source with all its philosophical ramifications and medical aetiology.102 As for truthful allegorical dreams, the kind that the Oneirocriticon addresses, Artemidoros believes that their source is the human soul. 105 Arabic dreambooks accord them a divine origin and stress that dream interpretation is a form of prophecy, citing a number of hadith and Quranic passages to fortify their arguments. It is the Islamic point of view that is not only espoused in the Oneirocriticon, but also argued in the same manner, by citing a scriptural quotation and by referring to examples of truthful dreams from the Holy Writ. An examination of the first part of Syrbacham's introduc-
واصدق ما تكون الرؤيا فى الربيع والصيف لا تقدم من الصديث الشريف وقد ذهب بعضهم بأن تفسير ذلك على هذا اللوجه واضعف ما تكون فى الخريف والشقاء، وقد قال ابن سيرين ﻭﻏﻴﺮﻩ ﺍﻗﻮﻱ ﻣﺎ ﺗﻜﻮﻥ ﺍﻟﺮﺅﻳﺎ ﻋﻨﺪ ﺍﺩﺭﺍﻙ ﺍﻟﺸﻤﺎﺭ ﻭﺍﺟﺘﻤﺎﻉ ﺍﻣﺮﻫﺎ، ﻭﺍﺿﻌﻒ ﻣﺎ ﺗﻜﻮﻥ ﻋﻨﺪ سقوط وررقها وذهاب تُمار ها (The most truthful dreams possible <are dream(> in spring and summertime, according to what is set forth in the honorable hadith . Some <of the dream interpreters> are of the opinion that such is the interpretation in this case. The weakest dreams possible <are dreamt> in the fall and in the winter. Ibn Sirin and others said that the strongest possible dream is the one <dreamt> during the season of the maturity of fruits and their harvest. The weakest possible are the dreams dreamt when their leaves have fallen and their fruits are gone).
102 For the Muslim division of dreams into false and truthful and the subdivisions within each group, see G. von Grunebaum, "The Cultural Function of Dream as Illustrated by Classical Islam," The Dream and Human Societies (Berkeley and Los Angeles, 1966), pp. 7-9, based primarily on the introduction to the dreambook of al-Nabulusi. Unfortunately, von Grunebaum does not explore the relationship between the Artemidorean and the Islamic views on the origin of dreams. Al-Nabulusi's division of dreams is lifted from al-Dinawari (introduction, magala 7. Esad Efendi 1833, fol. 28a ff; BN arabe 2745, fol. 36a ff). Fahd in "Les songes et leur interprétation selon I'lslam," pp. 134-37, discusses the origin of dreams in Arabic dream interpretation mainly on the basis of a passage from al-Masfudi's Muruj al-dhahab, and the various categories of dreams according to the division of al-Dinawari. For the division between false and truthful dreams in other Muslim dream interpreters, see Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 61-62 (al-Sijistānī); pp. 99-100 (al-Kharkūshī); p. 109 (al-Dīnawarī; on al-Dinawari's distinction between false and truthful dreams, see also the brief passage from his dreambook translated in Lamoreaux's Appendix IV, no. 1); p. 116 (Ibn Sina). Where relevant. Lamoreaux's discussion includes reference to the Greek sources of the Arabic authors.
103 i.2 (Pack 5, 17-20).
tion (Drexl 1,15-2,10) alongside a passage from an Arabic dreambook is revealing:
Συρβαχάμ ό όνειροκρίτης του των Ινδών βασιλέως. Σοφία μεγίστη [διαν] ή περι των όνειράτων κρίσις και λύσις και προφητεία ἀπὸ θεοῦ πασιν εὐαγγελιζομένη, καθό που γέγραπται ἐν τοῖς ἀγίοις εὐαγγελίοις, ὅτι πρὸς τὸν ἀγαπῶντά με ἐγώ καὶ ὁ πατήρ μου ἐλευσόμεθα καὶ μονήν παρ᾽ αὐτὸ ποιήσομεν. τοῦτο δὲ τελειούται δι΄ όράματος. και μαρτυρεί Ίωσηφ ό καταπιστευθείς τὴν Μαρίαν, τὴν μητέρα του φωτός, δι΄ όράματος μηνυθείς. όμοίως δὲ καὶ Δανιὴλ καὶ οἱ πλείους των προφητών δι' όράσεως το θειον φώς έδιδάχθησαν. οίς βεβαιούσθωσαν πάντες, ότι θελόν τι μήνυμα περὶ πάντων, ἀγαθῶν τε καὶ φαύλων, παντὶ τω λαώ ή των όνειράτων ὅψις ἐστίν.
Syrbacham, dream interpreter of the king of the Indians. The interpretation and deciphering of dreams is very great wisdom and prophecy brought by God as glad tidings to all, as is written in the Holy Gospels, that "to the one who loves Me, My Father and I will come unto him and make our abode with him'' [John 14:23]. This is fulfilled through a vision. Joseph, the one entrusted with Mary, the mother of Light, who was informed by a vision, bears witness to it. " Likewise, Daniel and most of the prophets were taught the divine light through <prophetic> visions. By these <examples> all should rest assured that the appearance of dreams is for everyone a divine message concerning everything, both good and evil.
The same message, this time tailored for a Muslim readership, is conveyed in the introduction of al-Dīnawarī (maqāla 10): 105
قال المعبرون من المسلمون علم الرؤيا هو العلم الاول منذ ابتداء العالم لم يزل عليه الانبياء والرسل عليهم السلام ياخذون به ويعملون عليه حتى كان اكثر نبواتهم بالرؤيا وحيا من الله عز وجل اليهم فى المنام لقول الله عز وجل الذين امنوا وكانوا يتقون لهم البشرى فى الحيـوة الدنيـا فـالرؤيـاا :: الصادقة ولقول النبى صلى الله عليه وسلم ذهب النبوة وبقيت المبشرات من الرؤيا وانما كان ضعف شرف الرؤيا في عهد النبى صلى الله عليه وسلم للوحى الذى كان ينزل عليه عيـانا والا فـمـا كان قـبل النبى صلى الله عليه
<sup>114</sup> Joseph was instructed by dreams three times. The first time was during his betrothal to Mary. Realizing she was pregnant, though he had never touched her, he suspected her of having relations with another man and wanted to annul the engagement. An angel in a dream informed him of the circumstances of Mary's pregnancy and the identity of the child (Matthew 1:8-25). In a second dream an angel warned him of the imminent massacre of the innocents, and advised him to flee to Egypt with Mary and the child (Matthew 2:13-14). The third time he was instructed by a dream (χρηματισθείς κατ' όναρ) to go to Galilee (Matthew 2:22).
<sup>115</sup> BN arahe 2745, fol. 40a-b; Esad Efendi 1833, fol. 70a. Al- Dinawari also elaborates on the connection of dream interpretation with prophecy in his maqala 5, BN arabe 2745, fol. 34b; Esad Efendi 1833, fol. 25b.
<sup>1166</sup> Sic in BN arahe 2745, fol. 40a (should be emended to ?); the variant in Esad Efendi 1833, fol. 70a: ... فى الحيوة الدنيا وفى الاخرة قـالوا الرؤيا الصـادقـة...
#### 162 CHAPTER FOUR
~ u-1 LU ill I *'.>--* ill jJ J l:i J.) I rh 0-" .\_j ~I ~I J y I i \_\_,..h .J.-4 F <sup>J</sup> ~.iL>.YI cHJL. 0-" ~J u-ll...u <L.J~ c.,? F-' ~ <l.JI ~ ......L....-9-;; ..:.WI .J.-4 ~I u *'-:"'.J* Jl..Ll i~I ~ *.....L.....>-:* \_\_,JL......:; *.,\_µ* d.Jj ~J .~.i~YI cHJL. .J.-4 ~J
The Muslim dream interpreters said: "The science of dreams is the first science. Since the beginning of the world, the prophets and the apostles, may peace be upon them, did not cease to practice it, and employed it to the point that most of their prophecies were accomplished through dreams that appeared to them during sleep as a revelation from God Almighty, according to the saying of God 'those who believe and keep their duty <to God>, for them is glad tidings in the present life' [Qur'iin 10: 64-65], and this is the truthful dream; and according to the saying of the Prophet, may the peace and blessings of God be upon him, 'Prophecy has passed, and there remain only the bearers of glad tidings from dreams.' 107 Yet, the weakness of the reputation of dreams at the time of the Prophet, may the peace and blessings of God be upon him, was only due to the revelation that came down to him visibly [and not in his sleep]. But before the Prophet, may the peace and blessings of God be upon him, none among the sciences of the Ancients was more honorable than dream interpretation. For this reason, God Almighty favored Joseph, may the peace and blessings of God be upon him, in the saying of God 'so that We may teach him the interpretation of events' [Qur'iin 12:21], and for this reason Joseph thanked God by saying: 'Oh my Lord! Thou hast given me sovereignty and hast taught me the interpretation of events' "[Qur'iin 12: 101]. <sup>108</sup>
Not only are the ideas expressed in the two passages identical, but there is at least one affinity in their verbal expression as well: in both texts dreams and their interpretation are characterized as "prophecy" *(nubuwa)* and "glad tidings" *(propheteia euaggelizomene,* the participle reflecting the Arabic *bushrii* and *mubashshariit).* The passage from the Qur'an (10:63-64) cited by al-Dinawari is included in the introduction of every Arabic dreambook I have examined, 109
109 lbn Qutayba, *Yahuda ar.* 196, fol. 2, II. 1-2; *al-Muntakhab* p. 8; ibid. p. 19 [ = al-Khargiishl,
<sup>107</sup> This */:tadfth* is repeated in the other four Arabic dreambooks: lbn Qutayba, *Y ahuda ar.* 196, fols. I b-2a; *al-Muntakhab,* p. 7; lbn Shahin, *fa\$!* I, p. 9; al-Nabulusi, vol. I, p. 2. For the quotation of this */:tadfth* in other Arabic authors, see Kister, "Interpretation of Dreams," p. 70, n. 18.
<sup>108</sup>A summary of this passage is given in al-Nabulusi, vol. 1, p. 2: ~j.)J ~I rL:. ..:,LS: U u-""\_,JI *j...o* 4-'.>..i~ ~ FJ oJJI ~.~YI ..::....:.LS:\_i iWI ~.)I i\_,h.JI *j...o* '--:;--oWI \_ii ~.)I LAii~ LlL...,..JI ~j.)1 d~I ~J o.,...,wJI ~,:, ..U\_i il..S..:..YI ~I~~ ~I ~l\_i :;)L.....JI ~I~ il:.YI ~ ~ ~.J....:JI ~ .i\_u L. ~ ~ il:...ll ~ <L.J *'->\_;:,* . i)L....J I (The science of dream interpretation is one of the lofty sciences. The prophets, may the peace and blessings of God be upon them, were wont to consider dreams as a revelation of the sacred laws to them. "Prophecy has passed, and there remain only bearers of glad tidings of the truthful dreams that man sees or that are shown to him in sleep," according to the view found in the */:tadfth* on the authority of <Mu~ammad>, the master of dreams, may the best blessings and the most perfect peace be upon him).
and is therefore very likely to have been cited in the Arabic source used by the author of the *Oneirocriticon.* The scriptural quotation in the Christian text seems to have been chosen as a fitting substitute for this Quranic passage. i *w*
The connection between dream interpretation and prophecy is stressed in all five Arabic dreambooks and is demonstrated through *f:zadlth.* In addition to the one already quoted in the above passage by al-Dinawari, a second is also unfailingly repeated in all five introductions: ;; . . ,, .:.,...... I~~ ~J..rl I ..:,, I b~ I ~I?. I ~.;I J ("Indeed, a dream is one forty-sixth of prophethood" ). <sup>111</sup> In addition, the five Arabic dreambooks abundantly cite the example of Joseph, the son of Jacob who, according to the Qur'an (sura 12), not only had a truthful dream himself, but also interpreted the dreams of other people aided by divine inspiration. <sup>112</sup>Conversely, the *Oneirocriticon* refers to the divine messages sent in a vision to the New Testament Joseph, the husband of Mary, as well as the vision that informed Daniel about the content and interpretation ofNebuchadezzar's dream (Dan. 2:16-49). Given that the Old Testament Joseph is the only example of a dream interpreter in the Qur'an, and as such is frequently mentioned in Arabic dreambooks, it is very likely that the Arabic source of the *Oneirocriticon* did refer to him. Since Joseph is also an Old Testament figure, the author of the *Oneirocriticon* could have included his name in the Greek text without deviating from its Christian character. However, the *Oneirocriticon* omits any mention of Joseph, not only in Syrbacham's introduction, but also in the interpretation of cows as years (Drexl 190, 4-17) and of winemaking as serving somebody powerful and receiving high offices (Drexl 151,
<sup>111</sup>Ibn Qutayba, *Yahuda ar.* 196, fol. 2a; al-Dinawari, *maqala* 5, BN *arahe* 2745, fol. 34b; *Esad Efendi* 1833, fol. 25b; *al-Muntakhab* p. 3 sup. and p. 17 inf.; Ibn Shahin, *fa!f/* I, p. 8 inf.; al-Nabulusl, p. 7. For other passages of this tradition in Arabic authors and a similar tradition in the Talmud, see Kister, "Interpretation of Dreams," p. 71, n. 20. Both traditions are recorded in the *\$a/:Lf/:L* of al-Bukhari (91:2 and 91: 5); see al-Bukhari, *Le receuil des traditions mahometanes par Abou Ahdallah Mohammed ihn lsmafl el-Bokhari,* ed. M. L. Kreh! and T. W. Juynboll, vol. 4 (Leiden, 1908), pp. 348-49.
<sup>112</sup>lbn Qutayba, Ankara, *ls. Saih Sincer* I, 4501, fol. 18 lb (missing from the Jerusalem manuscript); al-Dinawari, *maqala* 10, BN *arabe* 2745, fol. 40a-b; *Esad Efendi* 1833, fol. 70a; *al-Muntakhab,* p. 16; ibid., p. 21 (the story of Joseph's dream is enriched with details, mostly from the Old Testament, that are absent from the Qur'an); Ibn Shahin.fa!// I, p. 8; al-Nabulusi, vol. I, p. 3 (Joseph's gift of dream interpretation was bestowed by God) and pp. 6-7 (the interpretation of the dreams of the butler, the baker, and the pharaoh).
BL *Or.* 6262, fol. 2b (=3)]; Ibn Shahin. introduction, fa.~/ I, p. 8; al-Nabulusl, vol. I, p. 3.
<sup>&</sup>quot; 0 The first part of the two quotations is well matched: "to the one who loves Me" (John 14:23) = "those who believe and keep their duty <to God>"' [Qur'an 10:64]. The second half of each quotation is made to correspond with the other through its meaning: "'My Father and I will come unto him and make Our abode with him' [John 14:23], and this is accomplished through dreams" ='"for them is good tidings in the present life' [Qur'an 10:64], and this is the truthful dream."
13-14 ). These two interpretations could be justified on the basis of both the Old Testament and the Quranic narrative, 113 which is indeed done in the Arabic dreambooks. However, the only time that Joseph is mentioned in the *Oneirocriticon* is in a chapter attributed to the Egyptians, where a chariot is interpreted as proximity to kingship (Drexl 191, 22-26), "because Joseph, when he was freed and governed Egypt," was given a chariot (Gen. 41:43). Joseph's ascent in a chariot in the days of his glory is missing from the Qur'an, though it could be known to Muslims from the *Qi!fWf al-anbiyd,* and other works that fulfilled the desire of the faithful to find out more about the holy figures that are only briefly mentioned in the Qur'an. Indeed, an Arabic interpretation of the chariot identical to the one recorded in the *Oneirocriticon* can be found in the dreambook of al-Dinawari. <sup>114</sup>The reason for omitting Joseph in all other passages of the *Oneirocriticon* can only remain speculative, especially since it is impossible to know what Arabic text (or texts) was used as a source for the Greek. But it is conceivable that the author of the *Oneirocriticon* avoided naming Joseph because he viewed him as having been appropriated by Islam. <sup>115</sup>
Syrbacham' s introduction stresses the difficulty inherent in interpreting a dream, since the correct solution depends on the identity of the dreamer and the time that the dream was dreamt (Drexl 2, 10-16). He concludes as follows (Drexl 2, 16-24):
8t' wv Kat µa:A.A.ov ocj>EtAEl 6 6vE1poKpt•T\i; nuKvoi; ni; dvm Kat no:A.uµa0Tii; Kat 'tOV 8Et0V <1>6~ov EXON aEi. oli; Kat µaUov Ti Kpt<Jli; E<HtV aacj>a:A.Tii;, 8t6n U7t0 8EOU KEXapi'tco'tat. OU µ6vov 8£ wii; aya8mi; 6 8Etoi; OVEtpoi; 7tpo8EiKVU'tal ana Kat wii; novr1poii; Kat aµap'tCOAoti;, Kat 'taU'ta Ota 'tO 7tAOU<JlOV EAEOi; 'tOU 8EOu, Ka8wi; 'tpEcj>El Kat wui; apvouµf:voui; Kat ~Aa<Jcj>T]µOUV'tai; aU'tOV. vuv ouv EV 'tji ouvaµEt tiii; ayiai; avapxou Kat axcopicn:ou 'tptaooi; tiii; £pµrivciai; *anapxo*µat.
For these reasons the dream interpreter ought to be someone very wise and extremely learned and God-fearing always. Thereby the interpretation is very sound, because it is bestowed by the grace of God. For God-sent dreams appear not only to the good, but also to the wicked and sinful, because of God's bountiful mercy, just as
<sup>113</sup>Cows: Genesis 41: 1-32 and Qur'an 12: 43-49; wine-making: Genesis 40:9-13 and Qur'an 12:36 and 12:41.
<sup>114</sup>For a full quotation of the relevant Greek and Arabic passages, see the paragraph on chariots in chapter 8.
<sup>115</sup>Since Joseph is the only dream interpreter mentioned in the Qur'an (the 12th siira narrates his story in 111 verses), his example is pivotal for the Quranic and Islamic justification of dream interpretation, a practice that existed in pre-Islamic Arabia and therefore ran the risk of being condemned as pagan and impious by the new religion. For the rejection of the Islamic image of Joseph in Byzantine religious polemics earlier than the composition of the *Oneirocriticon,* see the anti-Muslim treatise by Niketas of Byzantium, *PG* 105, cols. 7560-757 A.
he takes care even of those who deny and blaspheme him. Now, by the power of the eternal and inseparable Holy Trinity, I begin my interpretation.
In the above excerpt, Syrbacham identifies the two virtues that a dream interpreter should possess in order to avoid errors: he has to be knowledgeable and he has to be God-fearing. The same virtues are required of a dream interpreter in Arabic dreambooks, as is evident from the following passage from Ibn Qutayba:116
فانه «=عـابر الرؤيا› يحــتــاج الى // ان يكون عــالما بكتــاب الله عــز وجل وبحديث رسول الله صلى الله عليه وسلم ليتعبرهما فى التاويل وبامثال العرب والابيات النادرة واشتقاق اللغة والالفاظ المبتذلة عند الاعوام وان يكون مع ذلك اديبًا لطيفًا ذكيًا عارفا بهيات الناس وشمايلهم واقدار هم واحوالهم عالما بالقياس حافظا ولن تغنى عنه معرفة الاصول الا ان يمده الله بتوفيق يسدد حكمه للحق ولسانه للصواب وان يحضره الله تعالى تسديده حتى يكون طيب الطعمة نقيا من الفواحش طاهرا من الذنوب فاذا كان كذلك اهر غ الله عليه من التوفيق ذنوبا فجعل له من مواريث الانيياء نصيباً\*
Indeed, < the interpreter of dreams> must be knowledgeable in the Qur'an and the hadith, in order that he may take them into account in <his> interpretations, as well as in the proverbs of the Arabs and the rare verses of poetry, the etymology of the language and current colloquial speech. Besides, he has to be a man of culture, gentle, sagacious, endowed with a capacity to judge the countenance of people, their character, their rank and station, to have a knowledge of analogy and be a person mindful of the ordinances prescribed by God. For knowledge of the principles of oneiromancy will be of no use to him unless God helps him succeed, so that his decision is directed to the truth and his tongue to that which is proper, and unless God provides him with the right direction so that he is seemly in his conduct and unstained by vile deeds and pure from sins. If he is such, then God generously sends down upon him success and grants him a share from the heritage of the prophets.
Ibn Shahīn expresses his thought in a similar manner:117
و ينبغي أن يكون اللعبر ذا حذاقة وفطنة، صدوقا في كلامه، حسنا في افعاله، مشتهرا بالديانه والصيانه بحيث لا ينكر عليه فيما يعبره لشهرة صدقه، ولذلك ســمـى الله يوسف بـالصـديـق، وأن يكون عــارفــا بـالاصــول فى علم التعسر ...
The dream interpreter must be endowed with sagacity and intelligence, and be sincere in his words and seemly in his deeds, and be well known for his religious obedience and his honor, so that people do not distrust him in what he interprets
<sup>116</sup> Yahuda ar. 196, fol. 2a-b.
<sup>117</sup> Introduction, fast 4, p. 11.
#### CHAPTER FOUR
because of his reputation for veracity. For this reason God named Joseph by the surname "veracious." And indeed the dream interpreter must know the sources of the science of dream interpretation .... 118
Finally, the Oneirocriticon says that truthful dreams are sent by God even to unbelievers and sinners.119 This is in agreement with what one can find in Arabic dreambooks. Ibn Qutayba considers the truthful dreams as either glad tidings for the righteous or warnings for sinners: 120
و اما الرؤيا الصادقة التي ياتيك بها ملك الرؤيا من نسخة ام الكتاب عن اللكوت ويضرب لك فيها الامثال والاشكال ليبشرك بخير قدّمته او ينذرك معصلية هممت بها او يُؤذيك بمكروه اشرفت عليه لتتعوذ بالله منه او ينبهك على تقصير كان منك في شي من فرائض لتتوب اليه فما ينكرها الا معاند جاحد بلسانه ما تستيينه نفسه...
As for the truthful dream with which the angel of dreams comes to you from the copy of the Mother Book [= the Archetype] in the kingdom <of Heaven>10 and impresses in it for you likenesses and forms in order to bring you glad tidings about a good deed you performed, or to warn you about disobedience [= sin] that you are contemplating, or to torment you with an adversity to which you are near, so that you can seek God's protection from it, or to awaken you to your neglect in performing your religious duties, so that you can do penance for it, only a stubborn mu anid [= one who wilfully rejects cogent arguments for no reason], who denies with his tongue what his soul [mind] perceives, will reject it ... 122
وان الكافرون وفساق المؤمنين The introduction to al-Muntakhab also states قد سرون الرؤسا الصادقة (Indeed, the unbelievers and the sinners among the believers do have truthful dreams),123 and further in the work, that these truthful
<sup>118</sup> These sources are enumerated further in the same paragraph (ibid., p. 11), as verses from the Quranic exegesis (tafsir), traditions of the Prophet ( hadith ), the interpretations handed down by the ancient dream interpreters in their books, sayings and verses of poetry.
<sup>119</sup> Brackertz, Traumbuch des Achmet, p. 234, n. 36, correlated this passage with Matthew 5:45: όπως γένησθαι υίοι του πατρός ύμων του έν ούρανοῖς, ότι τὸν ήλιον αὐτοῦ ἀνατέλλει ἐπὶ πονηρούς καὶ ἀγαθοὺς καὶ βρέχει ἐπὶ δικαίους καὶ ἀδίκους (That ye may be children of your Father which is in heaven: for he maketh his sun to rise on the evil and on the good, and sendeth rain on the just and on the unjust); and Luke 6:35: και έσεσθαι νίοι ύψιστου, ότι αυτος χρηστός έστιν έπι τους άχαρίστους και πονηρούς (And ye shall be the children of the Highest: for he is kind unto the unthankful and to the evil).
<sup>130</sup> Ankara, Is. Saih Sincer I, 4501, fol. 182a (missing from the Jerusalem manuscript).
<sup>121</sup> This refers to al-lauh al-mahfuz (the preserved tablet), which is the depository of the decrees or willed events ordained by God. It is kept in heaven. The angel of sleep copies from it what the future has in store and presents it to the dreamer.
<sup>12</sup> A similar passage attributed to Daniel is found in al-Dinawarī, magāla 4, BN arabe 2745, fol. 34a. Cf. al-Dinawari, maqala 6, BN arabe 2745, fol. 36a. Cf. also Ibn Shāhīn, introduction, p. 9 (end of fast 1); al-Nabulusi, vol. 1, p. 5.
<sup>123</sup> Al-Muntakhah, p. 3.
dreams of unbelievers are evidence of their unfaithfulness adduced against them by God. The point is illustrated with these examples:124
وقد يري الكافر الرؤيا الصادقة حجة لله عليه الا ترى فرعون يوسف راي سبع بقرات كما اخبر الله تعالى في كتابه فصدقت رؤياه وراى بتختنصر زوال ملكه وعظيم ما يبتلى به فصدقت رؤياه على ما عبرها له دنيال الحكيم ورای کسیری زوال ملکه فصدقت رؤیاه فاعرف هذا الجری فی التــاویل واعتبر عليه ترشد ان شاء الله تعالى
Sometimes an infidel will have a truthful dream as evidence in favor of God against him. Do you not see that the Pharaoh of Joseph dreamt of seven heifers, as is reported in the book of God the Almighty, and his dream was truthful,128 and that Nebuchadezzar dreamt of the end of his kingdom and the great affliction that he was going to suffer, and his dream came true exactly as Daniel the Wise had interpreted it for him, 120 and that Kisra dreamt of the end of his kingdom and his dream came true?127 Know this principle for dream interpretation and interpret according to it and you will be well guided, God willing.128
These passages show that every concept found in the two chapters of the Oneirocriticon that address the theoretical principles of dream interpretation is also included in the Arabic dreambooks. Occasionally, even the choice of words in the Greek text closely corresponds to bits and pieces of paragraphs found in the Arabic introductions. The relationship with Artemidoros, both in content and in verbal expression, is much more remote. It is therefore safe to conclude that the two theoretical principles on dream interpretation which are common in all three traditions that we are comparing (Artemidoros, the Arabic dreambooks and the Oneirocriticon) found their way into the Byzantine text by way of its Arabic sources.
136 Daniel's interpretation of Nebuchadezzar's dream is told not in the Qur'an, but in the qisas al-anbiya' (Stories of the Prophets).
177 Kisrā is a generic name applied in Arabic sources to the kings of Persia. Among the omina that announced the birth of the Prophet of Islam was a dream either by the king or the high priest of Persia that wild horses led by camels crossed the river and invaded Persian Iraq. The dream interpreters of the court could not interpret the dream. The Persian king then sent a letter to Nu man, the Arab king of a part of Persian Iraq, asking for someone capable of procuring an interpretation. The task was finally delegated to the famous Satih, one of the two legendary diviners of the Arabs, who interpreted the dream just before his death. See Lecerf, "The Dream in Popular Culture: Arab and Islamic," pp. 367-69.
128 A similar passage, enumerating the dreams of infidels that came true, can be found in Ibn Shāhīn, introduction, fasl 4, pp. 10-11.
<sup>124</sup> Al-Muntakhab, p. 17, last 4 lines.
<sup>125</sup> Qur'an 12:43-49.
#### CHAPTER FIVE
#### DREAM INTERPRETATIONS COMMON TO THE ONEIROCRITICON, ARTEMIDOROS AND ARABIC DREAMBOOKS
In addition to general principles governing the art of dream interpretation, the meaning of many dream symbols in the Oneirocriticon and Artemidoros also coincides. S. Oberhelman, the scholar who dealt with the subject most recently, assumes from this that the author of the Oneirocriticon relied heavily on the Greek text of Artemidoros. He compared 285 dream symbols from the Oneirocriticon with the corresponding interpretations in Artemidoros and tabulated the results as follows: he found 124 cases of "specific agreement" (43.5%), 56 cases of "general agreement" (19.5%), 32 instances of "general disagreement" (11%) and 75 instances of "strong disagreement" (26%). According to these figures, between 40 and 60 percent of the Oneirocriticon relies on Artemidoros.2
So much for the interpretations themselves. When it comes to their phrasing in the two texts-the approximation of which would be the strongest evidence that the medieval author had indeed copied the ancient dreambook-the two Greek texts are very different from each other. In over two hundred pages, the editor of the Oneirocriticon identified only one instance where the wording was the same. According to Artemidoros, ήλιος από ανατολής ανίσχων λαμπρός και καθαρός και καταδυόμενος είς δύσιν άγαθὸς πασιν (When Helius, the sun, rises in the east, bright and clear, and sets in the west, it means good luck for all men). The Oneirocriticon version is εί δε τδη <τον ήλιον > ανατείλαντα καθαρώς έν τη κλίνη αυτού, εί ... ούκ έχει γυναίκα, λαμβάνει πλουσίαν (If he sees that the sun rose clearly in his bed, if ... he
<sup>1</sup> Oberhelman, Oneirocriticon of Achmet, Appendix III, pp. 289-95.
<sup>2</sup> Oberhelman (ibid., p. 18) states that the comparative list of his Appendix III "is not exhaustive, simply because of constraints of space," thus implying that there are even more convergences between the two texts than the ones he lists. However, the number of agreements listed must be reduced. In at least two instances (nos. 46 and 47) under the category "specific agreement" the interpretations said to be given by Artemidoros cannot be found, and one cannot help but wonder if there are not more such mistakes.
<sup>4</sup> Drexl, "Achmet und Artemidoros," BZ 36 (1936), p. 299; not an article but a very brief note comprising a total of six lines.
<sup>4</sup> Artem. ii.36; Pack 160, 25-26.
does not have a wife, he will get a rich one).5 Not only does this example involve two very different interpretations, but each text uses a different synonym to convey the verb "to rise." The occurrence of 1m0apoc;/1m0ap&<; (clear/clearly) in both works is probably a coincidence-after all both authors were writing in the same language.
If the author of the *Oneirocriticon* had indeed borrowed heavily from the Greek text of Artemidoros, then it would be striking to find absolutely no textual parallels and no shared verbal expressions in light of the prevalent Byzantine (and general medieval) practice of quoting a source verbatim or with only minor changes. 6 The absence of such textual parallels and shared verbal expressions therefore tells us that the author of the *Oneirocriticon* relied exclusively on Arabic sources and did not borrow directly from Artemidoros. What similarities there are between Artemidoros and the *Oneirocriticon* can be attributed instead to the influence that the Arabic translation of Artemidoros exercised on Islamic dream interpretation. It is conceivable that some of the interpretations of Artemidoros were similar to the Islamic ones even before the Greek text was translated into Arabic, either because of the common Hellenic influence, introduced into the Arabic tradition through the Jewish one, or because of the ancient Near Eastern lore with which Arternidoros might have become acquainted during his travels. 7
At any rate, it is possible to demonstrate that the author of the *Oneirocriticon* did not directly borrow anything from the Greek text of Artemidoros by examining the interpretations from the two texts that are purportedly in "specific agreement." Closer scrutiny reveals that they converge only in their general idea, not in the details. In these cases, the imagery and expression of the *Oneirocriticon* are more akin to the equivalent interpretations found in Arabic dreambooks, as is evident in the examples that follow. Fifteen of the 124 cases
<sup>5</sup> Drexl 128, 15-17.
<sup>6</sup> The Arabic dreambooks that use earlier sources either quote the source verbatim or rephrase it slightly, in a manner that most of the time allows one to detect from where an interpretation was lifted, if the source text is available.
<sup>7</sup> The matter, however, cannot be decided unless the influence of I:Iunayn's translation is studied in detail. The problem is that the earliest surviving Arabic dreambook is contemporary with the Arabic translation of Artemidoros, which makes it possible that lbn Qutayba was familiar with I:Iunayn's text. If it is true that, as Kister argued ("Interpretation of Dreams," pp. 98-99), Ibn Qutayba meant his dreambook to be the Islamic answer to foreign dreamlore, then we should consider that the interpretations he records must deliberately avoid anything found in Artemidoros 's work. The fact that some of Artemidoros's interpretations are similar to those of Ibn Qutayba would then indicate that these interpretations converged even before the Arabic translation of the Greek text.
of"specific agreement" listed by Oberhelman (nos. 4, 9, 10, 21, 22, 23, 33, 37, 38, 45, 46, 47, 78, 83, 101) will be discussed. The choice is arbitrary, but an effort was made to find examples that would yield some further insights into the way the Greek author worked, in addition to demonstrating the course of the interpretation's transmission. Eyebrows, crucifixion and crosses, kings and temples, heavenly bodies and elephants were chosen because either the dream symbol itself or its interpretation seems to have been defined by the cultural surroundings of the dreamer. Eating a man's head and horses were included because their interpretation as dream symbols depends on the etymology of the word that designates them in Greek and/or Arabic, and provides an opportunity to see how this problem was addressed in their transfer from one language to the other. Finally, honey was singled out because it provides evidence pertaining to a philological problem.
#### *Eyebrows*
Oberhelman pointed out that hairless eyebrows in both Artemidoros and the *Oneirocriticon* indicate failure. 8 Artemidoros' s interpretation follows:
, O<iJpuci; 8acr£tat Kat £1'.mv8£ti; nacrt v ciya8ai, µciAtcrTa yuvm~i v· auTm yap U7tEp £uµop<1>iai; Kat µ£A.av1 xpiovTm Tai; 6<1>pfoi;. Tot ycipTot ~8ovcii; rnt £unpa~iai; 8riA.oum. ljltAat 8£: ou µovov cinpa~iai; <rnl> ciri8iai; ciAA.a ica\. n£v8oi; fooµ£vov 7tpObTJAOUO"tV' £8oi; yap naAmov Ent n£v8£t Tai; o<iJpfoi; ljllAOU0"8at.9
Eyebrows that are thick and luxuriant are auspicious for all, especially for women. For women, in the interests of beauty, color their eyebrows with black tincture. Therefore, they signify pleasures and success. Eyebrows that are hairless indicate not only failure and discomfort, but future mourning as well. For it is an ancient custom to shave one's brows while one is in mourning.
The interpretation of hairless eyebrows as symbolizing mourning is tied to a custom specific to the ancient Graeco-Roman world and is therefore omitted from the Arabic dreambooks that picked up Artemidoros's material, as well as from the *Oneirocriticon.* In spite of this omission, the *Oneirocriticon* devotes a full page to dreaming of eyebrows and eyelids, providing a much more detailed interpretation. In chapter 55, which discusses the Indian interpretation, we read the following: <sup>10</sup>
<sup>8</sup> Oberhelman, *Oneirocriticon of Achmet.* Appendix III, no. 4: *Oneir.* 54 = Artem. i.25: hairless eyebrows = failure.
<sup>9</sup> Artem. i.25; Pack 32, 3-8.
Τα βλέφαρα και αί οφρύες κόσμος είσι των όφθαλμών. ἐάν τις ίδη τα βλέφαρα και τας όφρύας αυτού εύειδέστερα, ούτος στερεώτερος έσται έν τη πίστει αύτου και χαρήσεται ἐπὶ τοῖς τέκνοις αὐτοῦ. ἐαν δὲ ἴδη αὐτὰ δυσειδέστερα, εἰ μέν ἐστι βασιλεύς, εἰς νόσον ἐλεύσεται καὶ φόβον μὲν ἐχθρῶν, οὐ νικηθήσεται δέ· ἐαν δὲ τοῦ κοινοῦ λαοῦ, εἰς νόσον μὲν ὕξει, τὴν δὲ πίστιν αὐτοῦ οὐκ απολέσει. έάν τις ίδη κατ όναρ, ότι τας όφρύας αυτου ξυρώ και βαφή κοσμεί, σπουδήν ούτος ἐπὶ τὸ δοξασαι τὰ τέκνα αὐτοῦ ἔξει καὶ δοξάσει αὐτά. ἐάν τις ίδη, ότι ἔπεσον οί όφρύες αὐτου ώς ἐπιλελωβημένου, εἰ μέν ἐστι βασιλεύς, τὰ αρματα αύτου άπολουνται και ό στρατος αυτού κακώς διοικηθήσεται εί δε του κοινού λαου, ού καταπιστευθήσεται τι παρά του λαου.
The eyelids and eyebrows are the embellishment of the eyes. If someone dreams that his eyelids and eyebrows became more beautiful, he will be more resolute in his faith and take pleasure in his children. But if he dreams that they became uglier, if he is emperor, he will become sick and will fear his enemies, but he will not be defeated. If he is a commoner, he will become sick but will not lose his faith. If someone dreams that he was embellishing his eyebrows with a razor and tint, he will zealously try to secure glory for his children, and will <indeed> secure it. If someone dreams that his eyebrows fell out as if he were a leper, if he is emperor his armed forces will be destroyed and his army will be poorly administered. If he is a commoner, people will not trust him with anything.
The next chapter, which covers similar dreams according to the Persians and the Egyptians, reads:11
Αί όφρύες και τα βλέφαρα δόξα του προσώπου είσίν. ἐαν ουν ϊδη τις, ὅτι ὑπερ δ πρώην ήσαν έδασύνθησαν και έκοσμήθησαν, εύρήσει έν μέσω πολλών δόξαν και τιμήν και νίκας κατ' έχθρών: και έαν ή αγαμος, συζευχθήσεται γυναικι και πλουτήσει. ἐάν τις ἴδη, ὅτι ταῦτα ήρημώθησαν καὶ ἐξέπεσον, εἰ μὲν ή βασιλεύς, τα τέκνα αύτοῦ οὐ κληρονομήσουσι τὴν βασιλείαν αὐτοῦ καὶ ή ἐπιθυμία αὐτοῦ στερηθήσεται ἀπ᾽αὐτοῦ καὶ ἐν μέσῳ <αὐτοῦ καὶ > τῆς συζύγου αὐτου ἔχθρα καὶ ῖσος γενήσεται: εἰ δὲ τοῦ κοινοῦ λαοῦ, ὅπερ αν ποθεινότερον κατα κόσμον 12 έχη, απολέσει αὐτὸ.
The eyelids and eyebrows constitute the glory of the face. If someone dreams that his eyelids and eyebrows become more bushy and beautiful than what they previously were, he will find glory, honor, and victories over his enemies before many; and should he be single, he will marry a woman and become wealthy. If someone dreams that [his eyebrows] were plucked out or fell out, if he is king, his children will not inherit his kingdom, the object of his desire will be taken from him, and there will be hatred and enmity between him and his wife. If he is a commoner, he will lose whatever he holds dearest in the world.
<sup>10</sup> Drexl 34, 21-35, 7.
<sup>11</sup> Drexl 35, 10-19.
<sup>12</sup> Drexl 35, 19: κατάκοσμον.
In the Oneirocriticon the condition of the eyebrows in the dream is correlated not only to the general well being of the dreamer, but also to his religious disposition and his relationship with his children and spouse. This dimension of the interpretation does not occur in Artemidoros and is borrowed from Arabic dreambooks. Ibn Qutayba interprets eyebrows as "the beauty of the dreamer's faith") والحاجبان زينته فى الدين).13).13 Further interpretations of eyebrows are given in the dreambook of al-Dīnawarī:14
الحاجبان زينة العين وهما للرجل حُسْن شيمته وجماله وامره وجاهه في دينه وامانته ومكانته ويقع تاويلهما على ما راى فيهما من صلاح او فساد وقال ارطاميدورس اذا كان الطجبان متكاثفي الشعر فهما محمودان من اجل ان النساء يسودون حواجبهن طالبا للزينة والذلك صار ذلك على امر لذيذ واستواء الاعمال(sic). {sic}.
The eyebrows are the embellishment of the eyes and, for a man, signify excellence in his disposition, his beauty, his authority and his glory in religion, honesty and rank. Their interpretation depends on whether [the dreamer] saw them in good condition or in decay. Artemidoros said that if the eyebrows had thicker hair it is a good sign, because women blacken their eyebrows in order to beautify themselves. This is why this became a token of something delightful and of steadiness of affairs\_15
Al-Dinawari's interpretation of eyebrows, including the excerpt from Artemidoros, is repeated verbatim'6 without mentioning the name of either author in the dreambook of al-Nābulusī, who then continues as follows:17
<sup>13</sup> Bāh 14, Yahuda ar. 196, fol. 30a, l. 10.
<sup>14</sup> Esad Efendi 1833, fol. 51a (fasl 6, bāh 50).
<sup>15</sup> Al-Dinawari gives the interpretation of Artemidoros by quoting verbatim the translation of Hunayn b. Ishaq. Al-Muhtakhab, p. 81, essentially paraphrases the interpretations found in al-Dinawari: الحاجبان حسن سمت الرجل وحسن دينه وجاهه والنقصان فيها نقصان فى هذه وقيل اذا كان الحاجبان متكاثفى الشعر فهما محمودان من اجل ان النساء يسودن حواجبهن طلب (The eyebrows are the beauty of a man's manner, the beauty of his faith, and his honor. Any damage to the eyebrows is damage to these things. It is said that if the hair of the eyebrows is dense, this is a good sign, because women blacken their eyebrows in their pursuit of beauty). Ibn Shahin does not give a separate interpretation for eyebrows.
<sup>16</sup> The only change is due to a scribal error, or a voluntary correction of the source: امر الذنـن (a pleasing or delightful affair) became المر زائد (an improving affair).
<sup>17</sup> Al-Nābulusī, vol. 1, p. 153, s.v. حاجب عبن الانسان.
و الحاجبان آبوان او ولدان او شريكان او زوجتان او نائبان او حاجبان وشبه الماجب بالنون المعرفة فان را ى الانسان حاجبيه قد اقترنادل ذلك على الألفة والمجة وبالعكس وسوادهما وغزارة شعرهما اذا لم يفحشا دليل على حسن حال من دلا عليه وبياضها ونزولهما على العين دليل على تغيير حال من دلا عليه من ولد او شريك او زوجة او نائب او صاحب وربمادل ذلك على طول عمر حتى يرى نفسه كذلك والماجبان يدلان على مرتبة فى الدين فما حدث فيهما من صلاح او فساد فانسبه إلى شيمته ووقايته فى دينه وربما دل الحاجب على حفظ من دلت عليه اللعين كالحاجب والوالى والوصي والزوج وهو قوس سهامه اللحاظ من العيون الحسن.
The two eyebrows are two parents, or two sons, or two partners, or two spouses, or two proxies, or two chamberlains.18 The eyebrow looks like a definite num.19 And if a man sees that his two eyebrows are joined, this signifies affection and love and vice versa. Their blackness and abundance, if it is not excessive, indicates the excellence of the situation of whoever is indicated by them. Their whiteness and their descent toward the eyes indicates a change in the situation of whoever is indicated by them, be he a son, or a partner, or a spouse, or a proxy [na'ib] or a friend [sahib].20 Sometimes, it indicates longevity, until one sees himself like that <i.e. with white eyebrows>. The two eyebrows signify elevation in religion, and whatever happens to them. either goodness or decay, should be correlated with <the dreamer's> disposition and protection regarding religion. Perhaps an eyebrow signifies protection of whoever is signified by the eye, such as the chamberlain, the ruler [wali], the authorized agent, or the spouse. The <eyebrow> is a bow and its arrows are the glances of lovely eyes.
As is evident from this passage, Artemidoros, al-Dinawari, al-Muntakhab, al-Nabulusi and the Oneirocriticon agree that thick and black eyebrows are a good sign, while thinning or hairless eyebrows are a bad sign. It can safely be said that this interpretation goes back to Artemidoros, from where the Arabic dreambooks borrowed and further embellished it, most probably by drawing upon older traditions of Near Eastern and Islamic dream interpretation. But the Oneirocriticon also connects the eyebrows with religious righteousness, as was done in the indigenous Arabic tradition represented by Ibn Qutayba, and with family members, as we can read in the dreambook of al-Nabulusi. If the author of the Oneirocriticon is aware of Artemidoros's interpretation of eye-
<sup>18</sup> The word hajib means both "eyebrow" and "chamberlain".
<sup>19</sup> The shape of the Arabic letter nun ( ; ) is like a crescent.
<sup>20</sup> Other possible translations of the word صاحب are "lord" and "disciple".
#### CHAPTER FIVE
brows, it is not via the Greek text of Artemidoros but via its Arabic translation, which was echoed in the Arabic models of the Oneirocriticon. Proof of this recycling is the fact that the Artemidorian interpretation ended up in the Byzantine text accompanied by all the Arabic additions and developments.
#### Crucifixions and Crosses
The interpretation of crucifixion in Artemidoros is the following (ii.53; Pack 183, 6-21):
Σταυρούσθαι πασι μέν τοῖς ναυτιλλομένοις άγαθὸν' καὶ γὰρ ἐκ ξύλων καὶ ήλων γέγονεν ό σταυρός ώς και το πλοίον, και ή κατάρτιος αύτου όμοία έστι σταυρώ. ἀγαθὸν δὲ καὶ πένητι καὶ γὰρ ὑψηλὸς ὁ σταυρωθείς καὶ πολλοὺς τρέφει <οίωνούς>. τα δε κρυπτά έλέγχει εκφανής γαρ ό σταυρωθείς. τους δε πλουσίους βλάπτει: γυμνοί γαρ σταυρούνται και τας σάρκας άπολλύουσιν οί σταυρωθέντες. ἀνδρὶ δὲ ἀγάμῳ γάμον προαγορεύει διὰ τὴν δέσιν πλὴν οὐ πάνυ τι συμφέροντα. ό δε αύτος λόγος και περι φιλίας και κοινωνίας. δούλους δε ἐλευθεροῖ ἀνυπότακτοι γὰρ οἱ σταυρωθέντες. τοὺς δὲ ἐν τῆ οἰκεία διατρίβειν βουλομένους και τους ιδίαν γην γεωργούντας και τούς ἐκβληθῆναί ποθεν φοβουμένους ἐκβάλλει καὶ οὐκ ἐᾶ μένειν ἐν οῖς εἰσιν εῖργει γὰρ ὁ σταυρὸς της γῆς ἐπιβαίνειν. ἐν πόλει δὲ δοκεῖν ἐσταυρῶσθαι ἀρχὴν σημαίνει τοιαύτην, οίος αν ή ό τόπος, ἐν ώ ὁ σταυρὸς ἔστηκε.
Being crucified is auspicious for all seafarers. For the cross, like a ship, is made of wood and nails, and the ship's mast resembles a cross. It is also auspicious for a poor man. For a crucified man is raised high and his substance is sufficient to feed many birds. But it means the betrayal of secrets. For a crucified man can be seen by all. On the other hand, it signifies harm for rich men, since the crucified are stripped naked and lose their flesh. For a bachelor, the dream means marriage, because the connection between the victim and the cross is a bond, but it will not be a very advantageous one. The same holds true for friendships and partnerships. But it means freedom for slaves, since the crucified are no longer subject to any man. For those who wish to spend time in their native land, for those who farm their own land, and for those who are afraid that they will be evicted from some place, it signifies that they will be evicted and that they will not be allowed to remain where they are. For a cross prevents a man from setting his feet upon the ground. To dream that one has been crucified in a city signifies a magisterial position that corresponds to the place where the cross has been set up.
Artemidoros further illustrated the interpretation of a crucifixion with an anecdote (iv.49; Pack 276, 6-12):
οίον <ἔδοξέ τις ἐσταυρώσθαι>, σημαίνοντος τοῦ δοκείν ἐσταυρῶσθαι δόξαν και εύπορίαν δόξαν μεν διά το ύψηλότατον είναι τον έσταυρωμένον, εὐπορίαν δε διά το πολλούς τρέφειν οίωνούς. < ό> Μένανδρος ἐν Έλλάδι ἔδοξεν ἐσταυρωσθαι εμπροσθεν ίερου Διός Πολιέως, και ίερευς ἀποδειχθείς αὐτοῦ ἐκείνου του θεού λαμπρότερός τε και εὐπορώτερος ἐγένετο.
For example, someone dreamt that he was crucified. Crucifixion dreams signify honor and wealth-honor, because the crucified person is in a very high position, and wealth because he provides food for many birds of prey. In Greece, Menander dreamt that he was crucified in front of a temple of Zeus, Guardian of the City. He was appointed priest of this same god and became better known and wealthier as a result.
According to the fifth-century church historian Sozomenos, crucifixion was abolished in the Roman Empire by Constantine the Great.21 Sozomenos's statement should not be taken too literally, since a similar method of punishment, anaskolopismos or phourkisis (hanging from the gallows or the gibbet), remained current well into the Middle Byzantine period.22 Anaskolopismos and phourkisis (or derivative words) are sometimes used interchangeably with staurosis (crucifixion) in Greek sources.23 Chapters 89 and 90 of the Oneirocriticon are entitled έκ τών Ίνδών περι άνασκολοπίσεως ήτοι φουρκίσεως (From the Indians on Impalement, that is Hanging on the Gallows) and ex των Περσών και Αίγυπτίων περὶ σταυρώσεως (From the Persians and Egyptians on Crucifixion). As Oberhelman rightly observed, Artemidoros's interpretations of crucifixion are repeated in both of these chapters, but the Christian Indians call the punishment anaskolopisis, whereas the pagan Persians and Egyptians use the term staurosis24 (Drex1 54, 8-55, 3):
<sup>21</sup> See J. McClintock and J. Strong, Cyclopaedia of Biblical, Theological and Ecclesitastical Literature, 10 vols. (New York, 1869-80), s.v. "Crucifixion."
<sup>22</sup> Cf. the account of the Nika revolt (532) by the 8th/9th-century historian Theophanes Confessor, Chronographia, ed. C. de Boor, vol. 1 (Leipzig and Stuttgart, 1883), p. 184, 4-5: κρατήσας ό έπαρχος τρεῖς δημότας των ατακτούντων, έφούρκισεν αὐτούς. καὶ ὁ μὲν εῖς πάραυτα απέθανεν οι δέ δύο έπεσαν έκ της φούρκας. The same method of punishment was applied during the reign of Basil I (867-86) according to the 10th-century author of the Vita Basilii in Theophanes Continuatus, ed. Bekker, p. 303, 16-18: δς τη βασιλική κελεύσει καθυπουργών ίσαρίθμους των αποσταλέντων δεσμίων έν Μεθώνη φούρκας στηναι προσέταξε και τους ώς ἐξάρχους τῆς φυγής πεμφθέντας ἀνεσκολόπισεν. For further references to the practice in Byzantine texts, see Du Cange, s.v. "φουρκα," "φουρκίζειν," "φουλκίζειν." An illustration of this method of punishment, evidently gleaned from contemporary practice, can be found in the illustrated Job of Patmos (Patmiacus 171, ca. 9th century).
<sup>23</sup> Cf. Suda Lexicon, s.v. "άνασκινδαλευθήναι: άνασκολοπισθήναι, ἀνασταυρωθήναι." Cf. also Liddell-Scott, s.v. "σταυρός"; Lampe, s.v. "άνασκινδαλεύω," "άνασκολοπίζω," "άνασκολοπισμός," "φούρκα," "φουρκίζω"; Du Cange, s.v. "φουρκα," "φουλκίζειν."
#### CHAPTER FIVE
πθ΄ Έκ των Ινδών περί άνασκολοπίσεως ήτοι φουρκίσεως
Έάν τις ϊδη κατ΄ όναρ, ότι κρίσει του ἐξουσιαστοῦ ἐκ τοῦ τραχήλου βία και ανάγκη, ἀναλόγως του ὕψους τῆς κρεμάσεως ἀξιώματι τιμηθήσεται εἰ δε ό τουτο ίδων νοσει ή θλίβεται, άποβαλεί ταυτα βία και τέλος χαρήσεται. όμοίως έαν Ίδη τις, ότι παρέδωκεν έτερον κρεμασθήναι, εί μέν έστι βασιλεύς ή έξουσιάζων λαού, θυμωθήσεται και όργισθήσεται έπι τα παραδοθέντι τα πρώτα και ύστερον τιμήσει αύτόν: ό δε τιμηθείς είς θεόν έξαμαρτήσει. ἐάν τις ίδη, ὅτι ἔφαγε κρέας κρεμαμένου, πλουτήσει ἐφαμάρτως ἐξ ὑψηλοῦ ἀνθρώπου αναλόγως, όσον έφαγεν.
89. From the Indians on Impalement, That Is, Hanging on a Gibbet
If someone sees in his dream that, according to the ruling of the authorities he was violently and forcibly hung by the neck, he will be honored with an office analogous to the height of his hanging. If the person who dreams this is sick or in distress, he will shed <his sickness or distress> through the use of violence and in the end he will be happy. Likewise, if someone dreams that he delivered somebody else to be hung, if he is king or <otherwise> rules people, he will first be angry and irate at the person whom he delivered, and then will honor him, and the one honored will sin against God. If someone dreams that he ate the meat of a hung person, he will sinfully gain wealth from a man in a high position in proportion to how much he ate.
ς Έκ των Περσών και Αίγυπτίων περὶ σταυρώσεως
' Εάν τις ϊδη, ότι έσταυρώθη, εί μέν έστι πτωχός, πλουτήσει, εί δὲ πλούσιος, βία και τιμωρία κληρονομήσει πλούτον πλείονα. ἐάν τις ϊδη, ὅτι ἐδάρη βουνεύροις και ούτως έσταυρώθη, κατά το μέτρον του δαρμού λαου έξουσιάσει ύψούμενος έν αύτωλ. ἐαν ϊδη, ὅτι ἐσταυρώθη καὶ κατῆλθεν, ἀπὸ τοῦ ὕψους τοῦ άξιώματος αὐτοῦ πεσεῖται καὶ ὁ πλοῦτος αὐτοῦ ἀφανισθήσεται. ἐὰν ἴδη, ὅτι ἐσταυρώθη καὶ ἐνεπρήσθη, ὑψωθήσεται καὶ λαοῦ ἄρξει καὶ μετὰ ταῦτα ἀνάγκη πολέμου αποθανείται δια τὴν του πυρὸς βίαν.
90. From the Persians and the Egyptians on Crucifixion
If someone dreams that he was crucified, if he is poor, he will become rich, and if he is rich he will inherit more wealth through violence and punishment. If someone dreams that he was beaten with a whip and was then crucified, he will have power over people commensurate with the beating and will be exalted because of this. If he sees that he had been crucified and came down from the cross, he will
<sup>24</sup> Oberhelman I.9-10: On. 89 = Artem. ii.53, iv.49: crucifixion = honor for the dreamer because he is elevated; On. 90 = Artem. ii.53: crucifixion = a good sign for the poor; see also Oberhelman, Oneirocriticon of Achmet, n. 181.
<sup>25 &#</sup>x27; Eξovolaothe was a title in the Byzantine court mentioned in De cerimoniis; see Sophocles, s.v. "έξουσιαστὴς."
fall from the loftiness of his office, and his wealth will vanish. If he sees that he was crucified and burnt at the stake, he will become exalted and will rule over people and after that he will die by the force of war because of the violence of the fire.
In the Islamic world, crucifixion (salb) refers to the criminal punishment whereby the body of the criminal, either living or dead, is affixed to or impaled on a beam or tree trunk and exposed for a number of days. 46 Arabic dreambooks discuss dreams about this method of execution, and their interpretations are very similar to those of Artemidoros, with the addition of new material. The contents of Arabic dreambooks on the topic correspond to chapters 89 and 90 of the Oneirocriticon much more closely than the equivalent passage of Artemidoros. Ibn Qutayba's interpretation of the crucifixion is the following:27
ومن اكل لحم مصلوب اكل مالا حراما من مال رجل رفيع اذا كان لا اكل اثره وان لم يكن له اثرٌ اغتـاب رجلا رفيـعا بـ ومن راى انـه مصلوب اصـاب رفـعـة من جهة السلطان مع فساد فى الدين.
If someone <dreams of> eating the flesh of a crucified <person>, he will spend unlawful money that belongs to a high-ranking individual if there was a mark on what he ate, and if there was no mark on it, he will slander a high-ranking man. And if he saw that he was crucified, he will receive high rank from the sovereign accompanied by corruption regarding his religious faith.
Later in his dreambook, Ibn Qutayba adds: 39 وان اكل لحم مســـصـلـوب او لحم ابرص او مجدوم اصاب مالا (If someone dreamt that he ate the flesh of a crucified or of a leprous person or a leper, he will obtain unlawful money). And further: 99 و من راي انه صلف اصان من الملك رففة (If someone dreamt that he was crucified he will receive a lofty position from the king).
Al-Dinawari's chapter on dreaming of the crucifixion (fasl 11, bab 59) quotes the interpretations of the Muslims, the Christians and Artemidoros: 100
<sup>26</sup> See El2, s.v. "salb."
<sup>27</sup> Ibn Qutayba, fas! 14 (on humans and their members), Yahuda ar. 196, fol. 30a.
<sup>28</sup> Fasl 46, Yahuda ar. 196, fol. 63a.
<sup>29</sup> Ibid... fol. 63b.
<sup>30</sup> BN arabe 2745, fol. 204b; Esad Efendi 1833, fol. 167b. Al-Dinawari quotes the interpretations of Artemidoros according to the translation of Hunayn b. Ishaq (Fahd, Artémidore d'Ephèse, 332, 1-333, 5). The Arabic text is quoted according to the readings of Esad Efendi with the (invariably better) variants from Fahd's edition of Hunayn's text in parenthesis, marked + in the Arabic and H in the English translation. I translated this passage as literally as possible, in order to demonstrate the changes in the meaning that are evident in the Arabic translation.
قال المسلمون من راى انه صلب ميتا فانه يصيب رفعة فى دنياه مع فساد دينه فان صلب حيا لم يغسد دينه بل نال شرفا ورفعة وسلطانا لقوله تعالى وما قتلوه يقينا بل رفعه الله اليه فان راى كانه مصلوب ولا يدري متى صلب فانه ان كان خرج منه مال اعترف وعاد اليه ذلك المال ويصيب حوايجه فان صلب مقتولا فانه يكذب عليه فى تلك الرفعة فان راى انه ياكل من لحم مصلوب وخرج منه الدم وكان لما ياكله اثر ظاهر فانه يغتاب رجلا مسلطا او ياثم والمفعول به ما جور لخروج الدم وقيل من اكل لحم الصلوب فانه يركب اللبريد ومن راي انه ڀرکب البريد فانه يموت وقالت النصاري من راي انه مصلوب على سور الدينة والناس ينظرون اليه ينال رفعة وسلطانا ويصير الاقوياء والضعفاء تحت يديه فان سال منه دم فان رعيته ينتفعون به ومن راى كانه ياكل لحم المصلوب فانه ينال مـالا ومنفعة من جهة احد الرؤساء وقال ارطاميدورس الصلب فى الرؤيادليل خير لكل من يسير فى البحر وذلك ان الة الصلب هي مركبة من خشب واوتاد كما ان السفينة مركبة من ذلك ودقل السفينة نسبه (ح: يشبه) الة الصلب وهو ايضا دليل خير فى الفقراء وذلك ان المصلوب يعلق ويعتدى (ح: يغتذى) منه حيوان كثير ويدل ايضا على ظهور الاشياء الخفية وذلك ان الذي يصلب ينتشر (ح: يشهر) امره فاما فى الاغنياء فانه دليل ردى وذلك ان المصلوب يصلب عريانا ويتـفيـر لونه ولذلك دل على فساد امورهم اذا راوا كانهم قد صلبوا فاما من كان غير متروج فان ذلك تدل على تزويجه وذلك سبب رباط المصلوب غير ان ليس لكلهم دليلا محمودا… ويدل ايضا فى العبيد على عتقهم وذلك ان من صلب ليس عليه خدمة ولاستة فاما فيمن يريد ان يقيم فى منزله ... وفيمن يخاف ان يتوجه فى ناحية على خروجهم من مكانهم وارضهم وذلك ان الصلب يمنع من الدفن فى الارض ۾ فان راى (ح: الانسان) كانه يصلب فى الدينة فان الرؤيا تدل على رياسة يكون على حسب الموضع الذى نصب فيه الصلب (sic) وقال الاخرون من صلب مقتولا فانه يكذب عليه في تلك الرقعة.
The Muslims said: "If someone dreams that he was crucified while dead, he will obtain a lofty position in this world with decay in his faith. And if he was crucified alive, his faith will not decay, but he will obtain a lofty position and honor and power according to the saying of God Almighty, 'They slew him not for certain, but God raised him up to Him' [Qur'an 4:157-58].31 And if he dreams that he is on the cross and does not know when he was crucified, if he has given away money, it will be acknowledged and his money will be returned, and he will obtain his possessions. If he was crucified after he had been killed, lies will be told to him while he is holding this lofty position. And if he sees that he partook of
<sup>41</sup> A reference to Jesus, who, according to the Qur'an, was not truly crucified but only appeared to be so.
the flesh of a crucified person, and that blood was coming out of it, and there was a clear mark on what he was eating, he will slander a powerful person, or he will sin and the object of his wrongdoing will be commensurate to the outpouring of blood. It is said that whoever ate the flesh of a crucified person will become Master of the Post, and if someone dreams that he became Master of the Post he will die."12
The Christians said: "If someone dreams that he was crucified on the walls of a city and people were looking at him, he will obtain a lofty position and power and the strong and the weak will come under his hands. And if blood was flowing from his body, his subjects will benefit from him. And if someone dreams that he ate the flesh of a crucified person, he will obtain money and gifts from one of the leaders."
Artemidoros said: 11 "Crucifixion in a dream is a good sign for all who travel at sea, and this is because the implement of crucifixion is composed of wood and nails, just as a ship is composed of them, and the mast of the ship refers to [l:I: resembles] the implement for crucifixion. This is also a good sign for the poor, because the crucified is suspended and many animals violate him [I:J: feed on him]. It also signifies the manifestation of secret things, because the affairs of the one who is crucified spreads out [I:J: becomes notorious]. As for the rich, it is a sign of ruin, because the crucified is crucified naked, and his color changes, and for this reason, if they dream they have been crucified, it signifies the decline of their affairs. As for those who are not married, it signifies their marriage, and this is because of the bond of the crucified, but it is not for all of them a laudable sign .... In addition, to the slaves it signifies their freedom, because whoever is crucified does not serve and does not conform to the norm. As for the one who wants to stay in his household ... and the one who is afraid to turn to a direction, <it signifies> their departure from their place and their land, because crucifixion precludes burial in the earth. And if he [I:J: the man] dreams that he was crucified in a city, the dream signifies leadership that will take place depending on the location where the crucifix was raised."
The others said: "If someone was crucified after he had been killed, lies will be told to him while he is holding that lofty position."
Al-Dinawari's Muslim interpretation of the crucifixion as achieving high rank is similar to that of Artemidoros. However, its justification with a Quranic passage lends it a specifically Islamic outlook. In the next chapter, which consists of three narratives about people who actually dreamt of crucifixion and the fulfillment of their dreams, al-Dinawari cited a dream dreamt by Muqammad b. Idris al-Shafi'! (767-820), founder of one of the four schools
<sup>12</sup> Literally "rode the postal horse" or "mastered the post." The position of the Master of the Post C~al;ih *a/-harid)* was one of the most important in the Abbasid government, since the appointee not only supervised the postal system, but also controlled the intelligence service and acted as a confidential agent to the central government.
<sup>11</sup> ii.53; Pack 183, 6-21.
#### 180 CHAPTER FIVE
of Islamic law, as well as a dream about the learned divine Qatada (d. 735) and the dream of Menander, the priest of Zeus, which is copied from Artemidoros (iv.49). <sup>34</sup>
Al-Shafi'i dreamt that he was crucified together with 'Ali b. Abi Talib (598-661), a cousin and son-in-law of the Prophet, who eventually became the fourth "rightly guided" caliph and one of the most revered figures in Islam. The dream signified al-Shafi'i's subsequent importance for Islamic legal thought and the orthodoxy of his views. According to the second anecdote narrated in this chapter, a man dreamt that Qatada was crucified and Ibn Sirin interpreted it as indicating Qatada's future renown. Given the prominence of al-Shafi'i and Qatada in Islamic intellectual history, the truthfulness of the dreams pertaining to them confers additional Islamic validity to the interpretation of crucifixion as representing high rank. 35 It is very difficult to ascertain whether this idea was current in Arabic dream interpretation before the ninthcentury translation of Artemidoros into Arabic. This would be the case if the narratives of the dreams foreshadowing the glory of the two eighth-century figures, Qatada and al-Shafi'i, are contemporary with their purported protagonists, which is not an easy matter to determine. At any rate, Ibn Qutayba's and al-Dinawari's treatment of crucifixion demonstrates that the Muslim interpretations which coincide with those of Artemidoros (whether preexisting independently or borrowed from the Greek dreambook) were thoroughly assimilated into the body of Arabic dream interpretation, from where the author of the *Oneirocriticon* received them.
*Al-Muntakhab* and al-Nabulusi essentially repeat al-Dinawari's interpretations,36 while Ibn Shahin includes some additional material. Most of the interpretations from chapters 89 and 90 of the *Oneirocriticon* can be paralleled with interpretations from the Arabic dream books, as is shown below:
Eav *w;* 'ion Km:, ovap, on KptO"El TOD i':~oumam:OD EKpEµacr0Tt EK TOD 'tpaxiiAOU ~iq Kat avayKlJ, avaMycoi; TOD Uljl01Ji; TI]i; KpEµcicrEcoi; a~ubµan nµTt0iicrE'tal ... oµoicoi; E:av rnn ni;, on JtapEOCOKEV ETEpov KpEµacr0i;vm, Et µEv i':crn ~amAEui; ii i':~oumciscov A,aoD, 0uµco8iicrE1m mt 6py1cr8iicrE1m i':nt n? napaoo0£vn 1a itpciYm Kat Ucr'tEpov nµiicrEl au16v· 6 OE nµTt0Eti; Eti; 0EOV £~aµap1iicrEl.
<sup>34</sup>BN *arahe* 2745, fol. 205b; *Esad Efendi* 1833. fol. 168a.
<sup>35</sup> The anecdote is repeated in lbn Shahin, no. 2271.
<sup>36</sup>*Al-Muntakhah.* chap. 31, pp. 161-62, repeats the interpretations quoted in al-Dlnawarl, including those attributed to the Christians and to Artemidoros, without stating their source. The same is done in al-Nabulusi, vol. 2, pp. 39-40, s.v. ~ . lbn Shahin copies some of al-Dinawari's interpretations from *a/-Muntakhah* (cf. nos. 2272-77), and adds more data.
If someone sees in his dream that according to the ruling of the authorities he was violently and forcibly hung by the neck, he will be honored with an office analogous to the height of his hanging .... Likewise, if someone dreams that he delivered somebody else to be hung, if he is king or rules people, he will first be angry and irate at the person whom he delivered, and then will honor him. And the honored one will sin against God.
These interpretations correspond to the ones found in lbn Shahin, no. 2278: <sup>~</sup>*a.!* L; c ~ .:\_,L...b.L.J I 0--o 1..5~ .c.µ ~ <t....:.11..51 *.J* ..J-o .,?Lo~ I JU\_, .J,L:... ~..i ~ .:.,~ ~.J-' ,.:\_,L.:;. *\_,u\_,* ~.J-' (Al-Kirmani said: "Whoever dreams that he was crucified, he will receive great beneficence and an exalted position and high rank from the king, and possibly there will be a flaw in his faith"). They also correspond to the interpretations from lbn Shahin, no. 2273: <sup>37</sup>Lo\_,\_)~ <LJ\_,\_i.J U~\_, ~.J '-:" . .L . .::.I ~ ~ .c.11..51 *.J j-o.-Q* . ~ ! <lJ I 4....1...J *.J* J..:. ~ *i.* ~ (If someone dreams that he was crucified alive, he will receive high rank and honor, according to the saying of God almighty "They slew him not for certain, but God raised him up to Him" [Qur'an 4:157-58]). And no. 2281: .c.µ ~>-"I dill .:\_,I 1..51 *.J* ..J-o *µ\_,* -~J ~ J~ ~ ~-' '~.J-' l.'114 <t.....:....o JL...:..: (It is said, whoever dreams that the king ordered that he be crucified, he will receive from the king a dignity and high rank. However, he is not good in his faith).
Yet another interpretation found in the *Oneirocriticon* is demonstrably closely related to an interpretation found in Arabic: £av n<; rnlJ, on €<1>ay£ Kp£ai; Kpcµaµ£vou, 1tAOU'tTJ<Jfl £<1>aµap'tffi<; £~ U\jfllAOU av8pcD7tOU avaA-oywc;, ocrov €<1>ay£v (If someone dreams that he ate the meat of a hung person, he will sinfully gain wealth from a man in a high position in proportion to how much he ate). This corresponds to Ibn Shahin, no. 2279: ~I <L..:.l *1..51,J* 0--o-' •.'-' *.o.!\_J* .:\_,l..b.L....1..5,j ..J-o ~ ~,)\_J ,Loi.>-"" ~Lo JS~ <L.:.l.J "-;-'\_,.L:...11 rJ I~ <L..:....o (If someone dreamt that he ate the flesh of a crucified person, he will eat [spend?] unlawful money, and probably he will have influence with a person in a position of authority and will receive benefit from him).
Two more Greek interpretations have Arabic parallels. The first one is: 'Eav n<; '(()lJ, on £maupro8ri, ti µ£v fon 7t'tWXO<;, nA,ou'!T]crfl (If someone dreams that he was crucified, if he is poor, he will become rich). An analogous
<sup>&#</sup>x27; 7 This interpretation is ultimately copied from al-Dinawari; lbn Shahin lifted it from *al-Muntakhah.*
interpretation is found in Ibn Shāhīn, no. 2276: الصلف... للفــقــراء غنـي a وسعفة (The cross ... for the poor signifies great wealth). The second Greek interpretation is: ἐαν ίδη, ὅτι ἐσταυρώθη καὶ κατῆλθεν, ἀπὸ του ὕψους του άξιώματος αυτού πεσείται και ό πλούτος αυτού άφανισθήσεται (If he sees that he had been crucified and came down from the cross, he will fall from the loftiness of his office and his wealth will vanish). The parallel text is ومن راي انه مـصلوب وانقطع حـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ مرتسه (If someone dreamt that he was crucified and that his rope broke, he will fall from his rank).
The special significance of crucifixion for Christianity is apparent in chapter 126 of the Oneirocriticon, "From the Indians concerning Kings and Crosses," and appears as one of the most Byzantine sections of the whole work, as it incorporates elements of Christian and Byzantine imperial ideology. The first half of this chapter interprets dreams about the Christian cross (Drexl 74, 21-75. 18):
ρκς Έκ των Ινδών περί βασιλέως και σταυρού
Εαν ίδη τις, ότι σταυρον ήρεν είς ναόν η είς οικον αύτου, από ύψηλου και μεγίστου, ϊσως και άπό βασιλέως, χαράν μεγάλην εύρήσει και νίκας κατ ἐχθρῶν. ἐάν τις ἴδη, ὅτι ἤρθη ἀπ᾽ αὐτοῦ ὁ σταυρός, εἰς θλῖνιν ἐλεύσεται καὶ καταδουλωθήσεται τω έχθρώ αύτου. έαν δε ίδη, ότι πάλιν έστράφη ό σταυρός, αποβαλεί τὴν λύπην αὐτοῦ.
Εαν ίδη τις, ότι τα τίμια ξύλα του σταυρου του Χριστού εύρε και εύπόρησεν αύτά, ήγουν τον σταυρόν όλον, ούτος βασιλεύσει και πιστότατος έσται. εί δε ίδη, ότι προσεκύνησεν αύτα έν τόπω γνωρίμω, εύρήσει χαράν και πλούτον άπο βασιλέως και δίκαιος κληθήσεται. εί δὲ ϊδη, ὅτι ἀπεκίνησε πρὸς τόπον, ἐν ῶ ώσι ταῦτα, εἰς αἴτησιν ἤξει πρὸς τὸν βασιλέα καί, εἰ μὲν ἴδη, ὅτι ἀπεσώθη ἐν τω τόπω και έτελείωσε τὴν ἐπιθυμίαν αὐτοῦ, τελειωθήσεται καὶ ή αἴτησις αύτοῦ.
Εαν ίδη ό βασιλεύς, ότι νέον σταυρόν τίμιον ἐκτήσατο, ουτος τεκνώσει αρσεν βασιλευον έν τω θρόνω της βασιλείας αύτού. έαν δε ΐδη ό βασιλεύς, ότι έδωκε γυναικι σταυρόν έγκόλπιον, άληθη λόγον δώσει αύτη, τάχα και τεκνώσει μετ΄ αὐτῆς οὐκ ἐνδέχεται δὲ τοῦτο πρὸς ἄλλην πλὴν πρὸς τὴν αὐγοῦσταν τοῦ βασιλέως ίδειν.
Εάν τις ίδη είς τόπον, έν ώ ουκ ήν σταυρός, νεωστι σταθέντα σταυρόν, νεωστι κεφαλὴ ἔρχεται εἰς τὸν τόπον ἐκεῖνον. καὶ ἐὰν ῶσι χριστιανοὶ ἐν τῶ τόπω ἐκείνω, χριστιανὸς ὁμόφυλος ἄρξει αὐτῶν, εἰ δὲ οῦκ εἰσιν, ἀλλόφυλος αρξει αύτων.
126. From the Indians on Kings and Crosses
If someone dreams that he raised a cross in a church or in his house, he will receive from an exalted and very powerful man-possibly even the king-great joy and be victorious over his enemies. If he dreams that the cross was taken from him, he will come to grief and be enslaved by his enemy. If he dreams that the cross is returned, he will be relieved from his grief.
If someone dreams that he found and procured the holy wood of the Cross of Christ-that is, the complete cross-he will become king and be very pious. If he dreams that he adored it at a familiar place, he will find joy and wealth from a king and will be called just. If he sees that he started off to a place where the cross was located, he will go to the king in order to make a request and, if he sees that he reached the place and fulfilled his desire, his request will also be granted.
If a king dreams that he acquired a new precious cross, he will beget a male who will rule upon his royal throne. If the king dreams that he gave a pectoral cross to a woman, he will make a true promise to her and will soon have a child with her. And it is not possible to dream such a thing concerning any other woman but the empress.
If someone dreams that in a place where there was no cross, a cross has recently been erected, a leader will soon come to that place. And if the people of that place are Christian, a Christian of the same race will rule over them. If they are not, a foreigner will.
The interpretation of the cross as a sign of victory seems inspired by the Byzantine veneration of the cross and the ideology surrounding it, the development of which began as early as the fourth century. Constantine's vision of the cross on the eve of his victory over Maxentius in 312 provided the earliest association of the cross with military victory, which was enhanced during the iconoclastic period (8th-9th century). The iconophile Orthodoxy that ultimately prevailed considered the cross as only one among several Christian symbols and no more important than the icons, though the cross did remain a military symbol throughout the tenth century. 38 The author of the *Oneirocriticon* further elaborates the significance of the cross as a sign of victory in chapter 150, which primarily discusses dreaming of icons (Drexl 106, 5-12):
'Eav ni; <sup>0</sup> l01J i::ixova 'tOU KUpiou i]µwv 'I11crou Xptcr'tOU £crmupwµ£vou, on npo- <J£KUVT]<J£V ll i]<J1tacrm:o ll EO£i]9T] m'.rciii;, El µ£v Ecr'tl ~amAf:ui;, Kat OU'tOs £Upi]CT£l virni; EX0pwv Kat xapav EV Tii O~Ti amou, 8t6n 6 m:aupoi; µeta tOU crmupw0£vtoi; EV auttji KUpiou i]µwv 'I l]<JOU XptcrtoU vi Kai; µqicrtai; <Jllµai V£l Kat nacr11i; 0At\jf£WS µi::ta~oA.i]v· Kat fov rnn i::ixova µ6v11v toU KUpiou i]µwv' I11crou XptcrtOU
<sup>38</sup> See *ODB,* s.v. "Cross." For the military importance of the cross as a sign of victory and the evolution of its symbolism, see N. Thierry, "La culte de la croix dans !'empire byzantin du Vile siecle au Xe dans ses rapports avec la guerre contre l'infidele. Nouveaux temoignages archeologiques," *Rivista di Studi Bizantini e Slavi* 1 (1981 ), pp. 205-28.
#### CHAPTER FIVE
ανευ του σταυρού, έστιν ή χαρά έλαττοτέρα, όμοίως και ή νίκη.
If someone dreams of an icon of our Lord Jesus Christ crucified, that he prostrated himself in front of it, or that he kissed it or prayed to it, if he is king, he will find victories against the enemies and joy in his glory, for the cross with our Lord Jesus Christ crucified on it indicates very great victories and a reversal of all sorrow. And if he dreamt of a simple icon of our Lord Jesus Christ, without the cross, the joy will be less, and victory likewise.
Given the demonstrable derivation of the Oneirocriticon from Arabic dream interpretation, on the one hand, and the evident Christian and Byzantine flavor of the interpretation of the cross, on the other, one wonders whether the contents of chapter 126 (and of chapter 150) were invented ex nihilo by the author of the Oneirocriticon, or were the result of the author's creative use of material that already existed in his Arabic sources. The association of the cross with victory can be found in Arabic dreambooks as well, as is evident from the following interpretation found in the dreambook of al-Nābulusī: و من ر ای انه If someone dreamt that he was crucified, if he is a commoner he will conquer and vanquish <his enemies>).40
The cross (salīb), as opposed to the crucifixion (salb), is discussed in only two of the five Arabic dreambooks investigated, and the interpretations of only one of them, al-Nabulusi, can be correlated with the contents of the equivalent chapter in the Oneirocriticon:41
صليب: ... وربما (الصليب) دل على الرجل العظيم الشأن [sic] المفتـرض الطاعة القائم بالدين فمن راى ان معه صليبا تزوج او رزق ولدا وربما كان من الزنا وربما دل الصليب على النكاح الفاسد ...
The cross: ... Perhaps <the cross> signifies a man of great importance who has the authority to impose obedience and uphold religion. And if someone dreamt that he had a cross with him, he will marry or have a son, possibly from fornication,
<sup>39</sup> Vol. 2, p. 39, s.v. صلب .
<sup>40</sup> Such is the translation if we understand the verbs as being in the active voice; if we understand them as being in the passive voice, the translation would be just the opposite ( "he will be conquered and vanquished"). The absence of vocalization in al-Nabulusi's printed text makes it impossible to decide whether the author intended to use the active or the passive voice (both are written identically if left unvoweled).
<sup>44</sup> Al-Nābulusī, vol. 2, p. 49, s.v. صليب. Ibn Shāhīn also includes an entry on the cross (no. 7251), but interprets it as a flaw in religion and a predilection for infidelity (kufr).
Sometimes the cross signifies a marriage contracted through a legally defective contract.
The *Oneirocriticon* and al-Nabulusi agree that the cross indicates a powerful man, future marriage, and male progeny. Al-Nabulusi's interpretations of the cross and especially the negative overtones of the marriage signified by it, moreover, seem to be an amplification of Artemidoros's interpretations of a crucifixion, formulated on the basis of I:Iunayn's Arabic translation: "As for those who are not married, it signifies their marriage, and this is because of the bond of the crucified, but it is not for all of them a laudable sign."42 It is conceivable that the author of the *Oneirocriticon* modeled the interpretations contained in the third paragraph of chapter 126 after an Arabic passage similar to the one quoted in al-Nabulusi. But it is improbable that he was directly inspired by Artemidoros, because the interpretation of the cross as begetting a son is absent from the Greek text and only appears in the Arabic tradition.
The fourth paragraph of chapter 126 (a cross newly erected at a place signifies that a leader will arrive there) also seems to echo the last interpretation of crucifixion originally found in Artemidoros. The choice of words employed in the *Oneirocriticon* is very different from the passage in the ancient Greek text, but the connection between the ancient and the medieval Greek interpretation becomes obvious if we have a look at the Arabic translation of Artemidoros.43
I:Iunayn rendered the Greek *arche* (leadership) with the Arabic *riyiisa,* the root of which is the same as for the word *ra's. Ra's* means both "head" and "leader," the exact equivalent of the Greek word *kephale,* which is used in the *Oneirocriticon.* The exact phrasing of the Arabic source of the *Oneirocriticon* can only be imagined, but it is likely that it was close to that of I:Iunayn. At any rate, given the disparities between the two Greek texts, the interpretation
<sup>42</sup> Fahd, *Artemidore d'Ephi'se,* 332, 11-13.
<sup>43</sup> Artem. ii.53; Pack 183, 19-21: EV rr6/i.n M OoKElV E<Haupwcr8m apxiiv m1µaivn 10tm'.m1v, oloi; Clv ~ 6 16rroi;, Ev c\> 6 cr1aup6i; £cr11jKE (To dream that one has been crucified in a city signifies a magisterial position that corresponds to the place where the cross was set up); see l:lunayn (Fahd 333, 3-5): d..........~..> ~ J.;:; ~j.)1 .:iu ~.ill~~ ...:.LS .:it.........:.~I LSI.> .:iu \*~I ~ ~ ~j.J I ~>-11 ~ ~ 0~ (And if a man dreams that he was crucified in a city, the dream signifies leadership that will be exercised, according to <what befits> the place where the cross was raised); cf. *Oneir.* 126 (Drexl 75. 15-16):' Eav ni; 'ilil,l Eii; 16rrov, Ev c\> ouK ~v crmup6i;, vEwcr11. ma8£vm cri:aup6v, vEwcr11. KE<)>ali.i] £pxe1m Eii; 16v 16rrov EKEtvov (If someone dreams that in a place where there was no cross, a cross has recently been erected, a leader will soon come to that place).
#### 186 CHAPTER FIVE
of the erection of a cross as the arrival of a leader is much more likely to have been introduced in the *Oneirocriticon* via the Arabic tradition than through the Byzantine author's direct knowledge of the Greek text of Artemidoros.
At least part of the material from chapter 126 seems not to have been invented by the author of the *Oneirocriticon,* but to have been developed (with what degree of liberty it is impossible to say) on the basis of existing material. Its second paragraph discusses dreaming about traveling to where the relics of the True Cross are, or in other words going on pilgrimage. The institution of the pilgrimage exists in Islam as well-the pilgrimage to Mecca (~ajj) is in fact one of the most important religious duties, one of the "Five Pillars of Islam," the observance of which is obligatory for all Muslims, and is therefore interpreted in every Arabic dreambook. Central to the rites of the pilgrim, when he or she reaches Mecca, is the circumambulation of the Ka'ba, a large cubic structure covered with a black cloth and containing the Black Stone (al-~ajar *al-aswad).* Though the Ka'ba and the Black Stone are not objects of worship, they represent a sanctuary consecrated to God since time immemorial, and for this reason Muslims in every part of the world orient themselves towards the Ka'ba when they pray. The direction of the Ka'ba, or *qibla,* is usually marked in mosques with a prayer niche (mi~rab). Arab lexicographers often call the cross the *qibla* of the Christians.44
In the *Oneirocriticon* the first two paragraphs on the cross contain interpretations that in Arabic dreambooks are given to the Ka'ba and the *qihla.* The Byzantine interpretations of finding the True Cross as meaning that the dreamer will become a ruler and be very faithful correspond to an interpretation of the Ka'ba quoted by Ibn Qutayba: <sup>45</sup>
L.. \_;.i..i.,> ..J.:i..WI ~ (~ ~ .. !.L.. .. L..11 ~ ~J <~I=> 4-.i\_,\_b ...:.I -sl.; .:.,u <1....::..~ ~I~~~ ~I -s~ .:\_,Lb..L..... '-:"-'""""Lo ~\_)I.:.,~ .;,jJ ~ . ....:.Lb..L.....~ JJl.J.:; JI o~ JI il..')'I ~
If he saw that he was circumambulating the Ka 'ba and performed the rites of the pilgrimage, this is probity in his religion commensurate with the rites that he performed, and the dreamer will come to exercise power.
This interpretation is repeated in all but one of the Arabic dreambooks I
<sup>44</sup> This is probably the result of the Muslim-Christian polemics of the 9th century, when Christian intellectuals under Islamic rule repeatedly clarified that the cross was not itself an object of worship. For further details and bibliography, see *EI2,* s. v. "~alib."
<sup>45</sup> *Fa:j/* 6, *Yahuda ar.* 196, fols. 26b-27a.
examined.46 They all also agree that the Ka'ba signifies a ruler, or generally a person of authority (caliph, imam, etc.), and visiting the Ka'ba is interpreted as entering the presence of the powerful individual signified by it. In al-Dinawari, for example, we read the following: <sup>47</sup>
~I ~J .:\_,U ... ~...ij..:j...il L.....~J \_,I *.;-::;...i* \_,I~ iLl..ll ~~I U-::J..=JI .:\_,....., J~ <t..:.Ll ~ 4.:..... :i....:;.1 .:\_,Ll \*U-::J..=JI ~ J.:..~ 4-.:.Ll ~I .L.........;:.
The Ka'ba in a dream is a caliph or a vizier or leadership or marriage. 48 ••• If <someone dreamt that> he entered the Ancient House,49 he will come to see the caliph, and if he received something from it, he will receive something from the caliph. <sup>50</sup>
This omnipresent Arabic interpretation is similar to one of the interpretations of the cross from chapter 126: "If he sees that he started off to a place where the cross was located, he will go to the king in order to make a request, and if he sees that he reached the place and fulfilled his desire, his request will also be fulfilled."
It seems, therefore, that the interpretations of crucifixion and the cross found in the *Oneirocriticon* have their roots not only in the Arabic version of Artemidoros's work and its echo in subsequent Arabic dreambooks, but also in the purely Muslim interpretations of the pilgrimage and the Ka'ba quoted in the Greek author's Arabic source. Arriving at new interpretations through the use of analogy based on existing material is an acceptable practice that is encouraged both by Artemidoros and by Arabic dreambooks, <sup>51</sup>since it is im-
<sup>46</sup> See al-Dinawari,fa,s/ 8, *hiib* 37 (on the Ka'ba), *Esad Efendi* 1833, fols. 76b-77a; not in *a/-Muntakhab;* lbn Shahin, nos. 948-49; al-Nabulusi, vol. 2, p. 173.
<sup>47</sup> Al-Dinawari,fa.yl 8, *bah* 37 (on the Ka'ba), *Esad Efendi* 1833, fol. 76b, BN *arabe* 2745, fol. 134a.
<sup>48</sup> The interpretation of the Ka'ba as marriage coincides with one of the interpretations of the cross examined above.
<sup>49</sup> The Ka'ba is called "the ancient house" *(a/-bayt a/-'atlq)* because, according to tradition, it was originally founded by Adam, and rebuilt several times by among others Seth, Abraham together with his son Isma'il, aand the descendants of Noah.
<sup>50</sup> The same interpretation is repeated in *al-Muntakhab. bah* 14, p. 56 (the wording is close to that in al-Dinawari, but the order of the sentences is different); Ibn Shahin, no. 941: "51 J .J-oJ ~l..b..L.....J I JI u...,..i.;J I "5~ ~ J ~I (If someone dreams of the Ka'ba he will probably see the caliph or the sultan); al-Dinawari's interpretation is repeated verbatim in al-Nabulusi, vol. 2, p. 172,s.v.~.
<sup>51</sup> On Artemidoros's advice to infer the interpretation of dreams not found in his dreambook by
possible for a single dreambook to contain everything that a human being can dream of
#### Decapitation and Eating a Human Head
Oberhelman notes that chapter 241 of the Oneirocriticon and chapter i.70 of Artemidoros agree that "to eat a man's head means money and wealth.""3 Though neither chapter i.70 nor any other passage in Artemidoros mentions eating a human or any other kind of head,53 the two dreambooks do agree that a head signifies both leadership, because the head controls the rest of the body, 4 and money, because the word "capital" (kephalaion) is derived from
52 "Specific agreement," no. 83.
54 In later Greek, кефαλή means both "head" and "the headman" or "the principal." The
analogy to the dreams he has described, see C. Blum, Studies in the Dream-Book of Artemidorus, Inaugural dissertation (Uppsala, 1936), p. 90. Blum rightly observed that this is a method of reasoning that Artemidoros borrowed from ancient empiricism. Reasoning by analogy (in Arabic qiyas) was also an established method of reasoning in the Muslim world, since it was recognized as one of the sources of Islamic law. See EI 2, s.v. " Kiyās."
<sup>55</sup> Artem. i.70 simply treats dreaming of various kinds of flesh, including human flesh, which is interpreted as follows (Pack 76, 21 and 77, 4-13): το δέ μέγιστον και ύπερβολή άχαθόν, ώς έγω ἐτήρησα, ἀνθρωπίνας ἐσθίειν σάρκας .... πάντων δ' αν εῖη σκαιότατον τὸ παιδὸς ίδίου σαρκας έσθίειν σύντομον γάρ όλεθρον μαντεύεται, εί μη άρα τις από τοιούτων ἐσθίειν νομίσειε μερών του παιδός, ἀφ΄ ών ο παῖς πορίζεται, οίον εί δρομαίου τους πόδας ή χειροτέχνου τας χειρας ή παλαιστου τους φάγοι τοτε γάρ τω μεν παιδι εύπορίαν το όναρ τουτο, τω δε πατρι όνασθαι του παιδός προαγορεύει. των δε άλλων ανθρώπων οί σάρκες έσθιόμεναι άγαθαί. τρόπον γάρ τινα οί ανθρωποι, όταν ώφελουνται από άλλήλων, έσθίουσιν άλλήλους (I have learnt from experience that the best and most auspicious dream by far is the one in which a person eats human flesh .... But eating the flesh of one's own child is the most unlucky of all. For it prophesies sudden death, unless a man dreams that he is eating those parts of his child from which he earns his living-for example, if a man's child is a runner and he dreams that he is eating his feet, or if his son is an artisan and he dreams that he is eating his hands, or if his son is a wrestler and he dreams that he is eating his shoulders. For, in these cases, the dream foretells that the child will become prosperous and that the father will derive benefit from his son. Eating the flesh of other men means good luck. For whenever men are assisted by one another, they are, in a sense, partaking of one another). Chap. 241 of the Oneirocriticon, which Oberhelman claims is in "specific agreement" with this passage from Artemidoros, is called "From the Indians, Persians and Egyptians on Various Foods" and discusses eating bread, honey, sugar, dates, olives and other foods regularly consumed in one's waking life. The only interpretation contained in chapter 241 that could conceivably be correlated with Artemidoros's chapter i.70 is Drexl 198, 5-7: εί δε έφαγεν έξ ανθρώπου κεφαλήν, εύρήσει πλήθος χρυσίου και πλούτον έξ άδικίας (If he eats the head of a human, he will find plenty of gold and wealth through wrongdoing). This interpretation is clearly a far cry from Artemidoros's chapter, which mentions eating other parts of the body, but not heads. Though there are similarities between Artemidoros and the Oneirocriticon regarding the interpretation of heads, they are not in the passages identified by Oberhelman, nor are they in "specific agreement" with each other.
the word for "head" (kephale). Accordingly, Artemidoros notes that dreaming of having a bigger head άρχήν τινα προαγορεύει …και χρημάτων ἐπίκτησιν …. τα γαρ χρήματα κεφάλαια καλείται 5 (foretells some office …and the acquisition of money ... . For, indeed, money is called "capital"). By the same token, in chapter i.35 Artemidoros interprets decapitation as release from debt for debtors and freedom for slaves:
τραπεζίταις δε και δανεισταίς <και> ἐρανάρχαις καὶ ναυκλήροις καὶ ἐμπόροις και πασι τοις χρήματα συνάγουσιν ἀπώλειαν τῶν κεφαλαίων δια τὸ ὁμώνυμον σημαίνει. ἀγαθὸν δὲ καταχρέοις διὰ τὰ αὐτά.56
To bankers, usurers, men who have to collect subscriptions, shipmasters, merchants, and all who collect money, it signifies loss of capital, because the word for "capital" is derived from the word for "head." It is auspicious for debtors for the same reason.
τοις δὲ λοιποίς <δούλοις> πασιν ἐλευθερία σημαίνει τὸ ὅναρ· κυρία γὰρ τοῦ σώματος ουσα ή κεφαλή ἐπειδὰν ἀφαιρεθῆ, χωρισθέντα του δεσπότου τὸν οἰκέτην έλεύθερον σημαίνει έσεσθαι.57
But to other slaves, the dream signifies freedom. For the head is the master of the body, and when it is cut off, it signifies that the slave will be separated from his master and will be free.
The same interpretations of decapitation also appear in the Oneirocriticon, 30 though the wording and the structure of the relevant chapter is very different from the text of Artemidoros.
The interpretation of the head as standing for a leader is repeated several times in the Oneirocriticon: ή γαρ κεφαλή είς τον ἐπέχοντα διακρίνεται (For the head signifies the one who holds power over the dreamer);30 ἐἀν ϊδη
earliest attestation of such a meaning occurs in an author who is roughly contemporary with Artemidoros, the late 1st- or early 2nd-century Hermas, in whose writings it means "head of the household" (see Lampe, s.v. "κεφαλή"). Artemidoros does not directly refer to the double meaning of κεφαλή.
<sup>55</sup> Artem. i.17; Pack 26, 1-17. Dreaming of having a bigger head is not discussed in the Oneirocriticon.
<sup>56</sup> Pack 44, 4-8.
<sup>57</sup> Pack 44, 15-18,
<sup>58</sup> Oberhelman "specific agreement," no. 21: Oneirocriticon 120 = Artem. i.35: decapitation for a slave = his freedom; no. 22: Oneirocriticon 120 = Artem. i.35: decapitation for a debtor = release from debt; already observed by Pack (p. 44, testimonia 17).
<sup>59</sup> Chap. 107. "From the Persians and the Egyptians concerning Blows and Wounds" ; Drexl 64, 20.
τις, ότι το αίμα αύτου έκ της ρινός ή της κεφαλής έρρευσεν, είς την κεφαλὴν ήτοι τὸν ὑπερέχοντα αὐτου νοείτω τὴν ἐπιζήμιον ἐξοδον (If someone dreams that blood flowed out of his nose or head, let him reckon this as an injurious spending 60 incurred by his leader, that is the person who has authority over him);61 ή γαρ κεφαλὴ εί ς τον ύπερέχοντα αύτου διακρίνεται (because the head is interpreted as the one who is above him <in rank>).60
In two instances where animal heads are interpreted as both leadership and money the author of the Oneirocriticon directly invokes the etymological connection of head (kephale) and capital (kephalaion) that was already used by Artemidoros:
όμοίως και έάν ϊδη, ότι κεφαλὴν καμήλου ήρεν, εί μέν έστι έξουσιαστὴς ό ίδών, ἐχθροῦ ὑπαρξιν εὐρήσει, εἰ δὲ τοῦ κοινοῦ λαοῦ, ἀπὸ μεγίστου κερδήσει κέρδος έως δέκα ή έκατον ή χιλίων νομισμάτων, δια το κεφάλαιον της κεφαλής και δια το τέλειον του άριθμου.63
Likewise, if he sees that he took the head of a camel, if the dreamer exercises power, he will find the possessions of an enemy; and if he is one of the commoners, he will gain profit from a powerful man up to ten or a hundred or a thousand coins, because of the capital signified by the head and the perfect number.64
έαν ίδη βασιλεύς, στι ήνεγκαν αύτώ κεφαλὴν χοίρου ή ὅα ἐτρωγεν αὐτήν, τὸ κεφάλαιον του πλούτου του έχθρου αύτου κερδήσει και αύτον νικήσει και πατάξει. 65
If a king sees that the head of a swine was brought to him or that he was eating it, he will win his enemy's capital of money and will defeat and beat him.
In addition, the Oneirocriticon dedicates two chapters to the interpretation of eating a human or an animal head, either raw or cooked, a dream that is never discussed in Artemidoros. The whole of chapter 39, "From the Indians on
<sup>60</sup> The translation "spending" for enooo which in medieval Greek can also mean "outcome" or "death" (cf. Lampe, s.v. "éξoδος" ; Sophocles , s.v. "ξέςοδος" ) is supported by the interpretation of blood as money in the Oneirocriticon (chap. 103, Drexl 61, 1-13) as well as by the phrase that immediately follows the one quoted above (Drex) 61. 19-20): ἐἀν δὲ ἐκ τῆς σαρκὸς ἔροευσε <τὸ αίμα αύτου> πλειστον πτωχεύσει καὶ ὀλιγόβιος ἔσται (If <his blood> flowed out of his flesh he will become poorer and will not live long).
<sup>61</sup> Chap. 104, "From the Persians and the Egyptians on Blood and Wounds"; Drexl 61, 17-18.
<sup>62</sup> Chap. 115, "From the Indians, Persians and Egyptians on Leanness <of the Body>, that Is Slimness"; Drexl 67, 22.
<sup>63</sup> Drexl 188, 5-9.
<sup>64</sup> For 10, 100, and 1,000 as perfect numbers, see Brackertz, Traumbuch des Achmet, n. 455.
<sup>65</sup> Drex1 227, 12-14.
Eating Heads," focuses on the interpretation of both human and animal heads as money (Drexl 25, 12-23):
Έάν τις ϊδη κατ΄ όναρ, στι ἔφαγε μυελὸν κεφαλῆς ἀνθρώπου γνωρίμου, εὑρήσει τον πλούτον και το χρυσίον αύτού παν και κληρονομήσει αύτόν εί δὲ ἀγνωρίστου, εύρήσει πλούτον άπό μεγιστάνου τινός και χρυσίον πλειστον ακόπως. ἐὰν δὲ ἴδη, ὅτι ἐσθίει κεφαλὴν προβάτου ἐψημένην, εὑρήσει πλοῦτον ἀνθρώπου του έν άξιώματι μεγάλω και κέρδος συν κόπω όπτωμένην, πολεμήσει ούτος έχθρον αύτου πρός το σκυλεύσαι τον πλούτον αυτού και νικήσει μεν αὐτόν, μετὰ δὲ τιμωρίας ἀπολαύσει τοῦ πλούτου αὐτοῦ διά την έκ του πυρός όπτησιν έάν τις ίδη, ότι έφαγε κεφαλήν ωμήν, ούτος πλουτήσει πλείστα σύν άμαρτίαις μεγίσταις.
If someone dreams that he ate the brains from the head of a person he knew, he will acquire all of his wealth and money and will be his heir. If <the head belonged to> a person he did not know, he will effortlessly acquire wealth from a powerful person and a great deal of money. If he dreams that he is eating the roasted head of a sheep, he will fight an enemy of his in order to capture his wealth as booty and will defeat him, and will enjoy his enemy's wealth with punishment, because roasting is done on fire.66 If he dreams that he ate a raw head he will become exceedingly wealthy amid very great sins.
Some of these interpretations are repeated in chapter 40, which treats the same topic according to the Persians and Egyptians. Here, too, the head is interpreted as leadership or a person in a position of authority:
Εάν τις Ίδη κατ' όναρ, ότι τρώγει κεφαλὴν προβάτου, ούτος εύρήσει χαραν μεγάλην και πλουτον. ἐαν δὲ ἴδη, ὅτι ἔφαγε ἀνθρώπου κεφαλήν, εἰς τὴν ὑπερέχουσαν αύτφ κεφαλὴν δολιεύσεται και κακουργήσει: Ίσως δὲ καὶ διαδέξεται τον θρόνον αύτου. ἐάν τις ἴδη κατ' ὅναρ, ὅτι κεφαλὰς προβείας ἢ έλαφείας ή των άλλων κερασφόρων ζώων ήνεγκέ τις αύτφ, ούτος μεγάλην άρχην άναδέξεται και δεσμίας κεφαλάς άρχόντων συνάξει κατά το μέτρον της ἐν τῶ ἐνείρατι προσενέξεως, τοῦτο δὲ ἀδύνατόν ἐστι θεωρησαί τινα πλὴν του Φαραώ και του διαδόχου αυτού. έάν τις ίδη πτωχός ή του κοινου λαου ότι ἐξωνήσατο ἢ εὐρε κεφαλὴν κατ΄ ὅναρ, οὐτος εὐρήσει χρυσίου νομίσματα η δέκα ή έκατον ή χίλια.67
If someone dreams that he is eating the head of a sheep, he will find great joy and wealth. If he dreams that he ate a human head, he will use wiles and wickedness against the leader who is his superior and possibly will succeed him on his throne. If someone dreams that a person brought him the heads of sheep or goats or deer or any of the other horned animals, he will be appointed to a position of great
<sup>66</sup> For the interpretation of fire as violence, see Oneirocriticon, chap. 158, Drexl 119, 11-15.
<sup>67</sup> Drexl 25, 24-26, 10.
authority and will gather the captured leaders of noblemen analogous to the <heads> brought to him in the dream. However, it is impossible for anyone but the pharaoh and his successor to have this dream. If a pauper or a commoner dreams that he bought or found a head, he will find ten or a hundred or a thousand pieces of gold.
Arabic dreambooks also interpret the head as representing either money or leadership or a person in a powerful position. And even in Arabic dreambooks these interpretations are justified linguistically, as they are in Greek, because of the multiple meanings of the word ra's ("head", "capital," and "leader")68 and its etymological connection with the words ra'is (leader) and riyasa (lead-و الراس هو الرئيس ومطار اه ership). Consequently, Ibn Qutayba states that فى الوجه فـهـو الجـاه (The head is a leader and whatever <the dreamer> sees on the face is his glory). 9 In another chapter, he connects the eating of heads with money :
فان راى ان رؤس الناس مقطوعة فى بلد او محلة فاين ذلك روساء الناس ياتون ذلك الموضع فبإن اكل منها او نال شَـفَرا او عظامـا او مـخــا او عــيـنا اصاب مالا من روساء الناس.
If someone dreamt that the heads of the people in a land or a quarter of a city were severed. " it is the leaders of the people who will come to that place. If he eats from them, or if he gets hold of any hair or bone or brain or eye, he will obtain money from the leaders of the people.71
<sup>68</sup> The meaning of ra's as both "head" and "leader" coincides with the two meanings of the later Greek kedan, and is reflected not only in chap. 40 of the Oneirocriticon, but also in the interpretation of a helmet (Drexl 114, 5-6): εί δε ΐδη, ότι ἐνεδύσατο περικεφαλαίαν, εὐρήσει δύναμιν και στερέωμα είς την ύπερέχουσαν αυτώ κεφαλήν (If he dreams that he wore a helmet, he will find strength and support in the leader <ranking> above him).
<sup>69</sup> Fast 14, fol. 28a
<sup>70</sup> Here there might be a lacuna in the Arabic text.
<sup>71</sup> Fash 46 (nawadir), fols. 63b-64a. A more extensive version of this interpretation is given in فان رای :al-Dinawari, fașl 11, bāb 53, Esad Efendi 1833, fol. 126b; BN arabe 2745, fol. 202b روس اللناس مقطوعة بيده فى مجلسه فان الناس ينقادون له وياتون ذلك الموضع وربما اجتمع روساهم هناك فاين اكلها نيا او نال شعرها او عظامها فانه يصيب مالا من روساء الناس فوق ما يرجوا وربما اغتاب رئيسا (If someone dreamt that people's heads were severed at his hands before his council, the people will obey him and will come to this place. Possibly their leaders will congregate there. If he ate from them raw, or if he took a hair or a bone, he will obtain money from the leaders of the people beyond what was anticipated, and possibly he will malign a leader). Ibn Shāhīn (nos. 1351-52) also has a version of this interpretation: و من ر ای ر ءوس الناس مقطوعة في بلد او مجلد=محل› او بيت او على باب فإن رؤساء الناس يأتون ذلك الموضع ويجتمعون فيه وان راى انه يأكل منهم او يأخذ شيئك فهو حصول منذفعة ومـال
Similar interpretations can be found in each one of the Arabic dreambooks examined. Al-Dinawari also considers the head as indicative of both wealth وقـال المسلمـون راس الانســان رياســة ورئيـســه هـو الذى 22:and leadership تحت يده ور أس ماله وجده على ايّه حالة راها حسنة كانت اوقبييحة... (The Muslims said: "A man's head is leadership and his leader is the one who has him in his power; <it is also the dreamer's> capital of money and his fortune, and the condition <of this money and fortune> is analogous to the condition of the head that <the dreamer> dreamt of, either good or bad.")73
Further on in his dreambook, al-Dīnawarī repeats: فالراس راس المال (The head is a capital of money). 4 Al-Muntakhab states: و امــا الـر اس فــ الــتــاو يـل As for the head in interpretation, it is the leader of the dreamer, that is, the person who has him in his power, 6 as well as the capital of his money and his fortune). Likewise, in Ibn Shāhīn (no. 1344) we read that وقــد سـدل ر أس الانـســان ر أس مــالـه ر أس: فى : The head of a man signifies his capital). Al-Nābulusī concurs) :Head) المنام هو رياســة الانســــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ this, in a dream, is the leadership over a person, and the leader who has him in his power, and his capital).77
In addition, Arabic dreambooks contain interpretations that closely correspond to excerpts from the Oneirocriticon. The beginning of chapter 39, which interprets dreams about eating the head of a known or an unknown person, closely corresponds to the following passage from Ibn Shāhīn, (no. 1354):
ومن راي انه ياكل رأسا نيئا ففيه وجهان قيل حصول مال او عتب من رئيس وإن كان رأسا معروفا فربما انه ياكل من راس مال صاحب الراس.
74 Fash 11, baib 53, Esad Efendi 1833, fol. 126b; BN arahe 2745, fol. 208b, omits this phrase.
و خير (If someone dreamt that the heads of people in a land or a quarter of a city or a house or in front of a gate were severed, it means that the leaders of the people will come to that place and congregate there; if he dreamt that he ate from them or took something <from their flesh>, this represents the attainment of profit, wealth, and goodness).
<sup>72</sup> Fasl 6, bab 27, Esad Efendi 1833, fol. 48b; BN arahe 2745, fol. 74b.
<sup>73</sup> The same interpretation is repeated in al-Muntakhab, p. 77.
<sup>75</sup> Al-Muntakhab, p. 77.
<sup>16</sup> الذى هو تحت يديه; cf. o èπέχων and ὁ ὑπερέχων from the equivalent Greek interpretations (Drexl 64, 20; 61, 18; 67, 22).
<sup>77</sup> Al-Nābulusī, vol. 1, p. 239, s.v. ر أس .
If someone dreamt that he was eating a raw head, there are two possibilities: It is said that <it signifies> either the attainment of money or rebuke from a leader. And if the head belonged to someone known <to the dreamer>, he will possibly spend the capital that belonged to the owner of the head.
When it comes to animal heads mentioned in the Oneirocriticon, the Arabic dreambooks examined contain passages that, though they do not coincide, still bear a vague resemblance to those references. Ibn Shahin (no. 1374) gives the following interpretation:
وقال الكرماني رؤيا رءوس الحيوان من حيث الجملة مال ورياسة فإن كان مما يؤكل لحمه يكون كسب المال من وجه حل وإن كان مما لا يجوز أكله يكون من وجه حرام.
Al-Kirmani said: "The dreams of animal heads generally <signify> wealth and leadership. If <the animal> has edible meat, the acquisition of wealth will be from a lawful source. If not, it will be from an unlawful source."
Al-Nābulusī interprets animal heads as follows:78
فان راي انه ياكل راس شالة او راس بقرة او ٿور او راس جمل نيئا فانه دفتات رئيسا بنسب الى ذلك الجوهر فان كان مطبوخا او مشوبا فانه يستقييد مالا من الرؤساء او ياكل راس مال احد ينسب الى جوهره والراس من الشاة راس مال اكثره عشرة آلف درهم واقله ألف درهم ... ومن راى انه ياكل راس غنم وكراعه فانه ينال عزا ومالا بالصرى أن يكون من ميراث.
If <someone> dreamt that he ate the head of a sheep or the head of a cow or a bull or the raw head of a camel, he will malign a leader referring to this substance. If <the head> was cooked or roasted, he will acquire wealth from the leaders or will spend the capital of someone who is referred to by this substance. The head of a sheep is capital, at most ten thousand dirhams and at least a thousand dirhams .... If someone dreamt that he ate the head of a sheep, as well as its foot. he will obtain power and money, but it will hardly be through inheritance.
The interpretation of a cooked head of a sheep as booty taken from an enemy given in chapter 39 is not found in any of the five Arabic dreambooks, but it also seems to convey an Arabic interpretation based on etymology: the Arabic word for "sheep" (ghanam) has the same root as the word for booty (ghanima). The interpretation of eating a human head as using wiles and wicked behavior
<sup>78</sup> Ibid., p. 243.
against a leader in chapter 40 has its counterpart in al-Dinawari's dreambook:" (1)فان راي روس الناس مقطوعة بيده ... فان اكلها ... ربما اغتاب رئيسا he saw the heads of people cut in his hand ... < and> if he ate from them ... , he would probably malign a leader). Finally, the last phrase from chapter 40 (If a pauper or a commoner dreams that he bought or found a head, he will find ten or a hundred or a thousand pieces of gold®) is also very similar to a passage from al-Dinawari: فان رای انه اخذ ر اسا فانه مال یصیر الیه اقله الف در هم واكــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ head, this means that money will end up in his possession, and the smallest sum possible is a thousand dirhams, and the greatest is ten thousand dirhams). An analogous interpretation is recorded by Ibn Shahin (no. 1349): و قـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ بعض المعبرين من اصاب رأسا فانه يصيب من عشرة دراهم إلى عشرين الف (Some of the dream interpreters say: "If someone acquired a head, he will acquire from ten up to twenty thousand dirhams").
Chapter 120 of the Oneirocriticon, like Artemidoros, states that decapitation indicates freedom for slaves or release from debt. The relevant passage reads as follows: 82
Εάν τις ίδη κατ' όναρ, ότι άπεκεφαλίσθη γωρισθείσης αύτου της κεφαλής, εί μέν ἐστι δούλος, ἐλευθεροῦται, εἰ δ΄ ἀσθενής, ἰαθήσεται, εἰ δὲ θλῖψιν ἔχει ἢ χρέος, ταυτα ἀποβαλεῖ ἐαν ίδῃ τοῦτο ὁ βασιλεύς, ή μέριμνα καὶ ὁ φόβος αυτού πας και ή φροντις αύτου μεταβληθήσεται. ἐάν τις ΐδη, ὅτι παρὰ γνωρίμου απεκεφαλίσθη, κοινωνήσει και αύτος της χαράς αύτου. έαν ϊδη, ότι δια χειρος παιδός ανηλίκου ἀπεκεφαλίσθη, εἰ μὲν νοσεῖ, ἀποθανείται ταχέως ….
If someone dreams that he was decapitated and that his head was separated from his body, if he is a slave he will be freed, if he is sick he will be healed, if he is grieved or in debt he will be relieved from <grief or debt>. If the emperor dreams this, his concerns and all his fear and trouble will be changed <for the better>. If someone dreams that he was decapitated by someone he knew, that person will also partake of his joy. If he dreams that he was decapitated by the hand of an under-age youth, if he is sick, he will soon die ....
The structure and contents of this chapter correspond closely, not to the text
82 Drexl 70, 23-71, 6.
<sup>79</sup> Fasl 11, hāb 53, Esad Efendi 1833, fol. 126b; BN arabe 2745, fol. 208b.
<sup>80</sup> Drexl 26, 8-10,
<sup>81</sup> Al-Dinawari, fasl 11, bab 53, Esad Efendi 1833, fol. 126b; BN arabe 2745, fol. 208b; al-Dinawari's interpretations are repeated in al-Muntakhab, pp. 77-79, and al-Nabulusi, vol. 1, p. 239-43, s.v. رأس.
of Artemidoros, but to a paragraph found in a number of Arabic dreambooks with slight variations in each. According to Ibn Qutayba, 33
ومن راي ان عنقه ضربت وبان الراس فانه إن كان عبدا عتق وإن كان مريضا شفى وإن كان مدينا قضى دينه وإن كان ضرورة ُ ُ حج وإن كان خائفا امن وإن كان مغموما نفس عنه فإن عرف الضارب عنقه جرى له الذير على يديه او يد سميه او نظيره او شقيقه ...
If someone dreams that he was decapitated and his head was separated <from his body>, if he is a slave he will be freed, if he is sick he will be healed, if he is in debt he will be released from it, if he is in want he will go on pilgrimage, if he is afraid <of something> he will be safe, and if he is distressed he will be relieved. If he knew the person who decapitated him, he will receive favor through him or through someone with the same name or through someone like him or through his sibling....
Identical interpretations couched in almost the same terms can also be found in al-Dīnawarī:84
قال المسلمون من ضربت رقبته وبان عنه راسه فان کان مريضا شفى او کان مديونا قضى دينه او كان ضرورة حج البيت وقضى دين الله تعالى الذى عليه او كان فى خوف وكرب فرج عنه فان عرف الذى ضرب رقبته فان ذلك يجرى على يدى من ضربها فان كان الذى ضربها صبيـا لم يبلغ فـان ذلك راحته وفرجه مما هو فيه من كرب اللرض الى ما يصير اليه من فراق الدنيا وهو موته على تلك الحال ... وان راى ان ملكا او واليا يضرب عنقه فان تاويل الوالى هو الله تعالى ينجيه من همومه ويعينه على اموره فان راى ان ملكا ضرب رقاب رعيته فانه يعفوا عن الذنبين ويعتق رقابهم.
The Muslims said: "If someone was decapitated and his head was severed, if he was sick he will be healed and if he was in debt he will be released from it, and if he was in want he will go on pilgrimage to Mecca and pay his debt to God, and if he was in fear or in grief he will be relieved from it; and if he knew the person who decapitated him, <the fulfillment of the dream> will come through the hand of the person who did it. If the person who did it was a youth under age, this <signifies> comfort and relief from whatever trouble or sickness the dreamer had until he departs this world, i.e., <it signifies> his death in this condition ... And if he saw that a king or a local governor decapitated him, the governor is to be interpreted as God Almighty who will save him from his troubles and will make
<sup>83</sup> Bãh 14, Yahuda ar, 196, fol, 31b,
<sup>84</sup> Fasl 11, hāh 51, Esad Efendi 1833, fol. 126a: BN arabe 2745, fol. 201b
him distinguished in his affairs. And if he saw that a king decapitated his subjects, he will pardon the criminals and manumit their slaves."85
In the last sentence of this passage, the interpretation of decapitation as freedom for slaves is based on etymology. In Arabic the verb "to decapitate" is expressed periphrastically as "to hit the napes" (ضرب الرقــاب). The word raqaba (pl. riqāb) means both "napes" and "slaves."86 Given this coincidence, the interpretation of decapitation as freedom for a slave, found in Artemidoros, could very easily have been either absorbed by Arabic dream interpretation after the translation of Artemidoros, or invented independently by Arab interpreters even before Hunayn's translation was in circulation. A paragraph like the one quoted above from al-Dinawari can also be found in Ibn Shāhīn (no. 1343):
ومن راي أن عنقه ضرب وابان رأسه منه ان كان غنيا نقص ماله، وإن كان فقيرا استغنى، وإن كان عبدا عتق، وإن كان مديونا قضى الله دينه، وإن كان مغموما او مكروبا فرج الله غمه وكربه، وان كان مريضا شفاه الله، وإن كان مريضا ومرضه لا يوجد له طب يدل على موته.
If someone dreamt that he was beheaded and his head was separated from his body, if he is rich his wealth will diminish, and if he is poor he will become rich, and if he is a slave he will be freed, if he is in debt God will settle his debt, if he is grieved or troubled God will relieve him of it, if he is sick God will heal him, and if he is sick and there is no medical treatment to be found for his illness, this signifies his death.
A second paragraph from Ibn Shāhīn (no. 2260) contains similar interpretations:
ومن راي انه قتل مضروب العنق فانه ان کان عبدا عتق لقوله تعالى فك رقبة، وقيل فرج من هم وغم، وإن كان مديونا قضى الله دينه من حيث لا يؤمل، وربما اعطى مالا عظيما، وإن عرف الذى فعل به ذلك نال منه خيرا، وإن كان القاتل امراة او خصيا او صبيا لم يبلغ الحلم او رجلا بلا لحية فإنه يدل على من پانخذ روحه سواء کان بموت او قتل او غيره.
If someone dreamt that he was killed by being decapitated, if he is a slave he will be freed, according to the saying of God Almighty "the freeing of a slave" [Qur'an 90:13], and it is said <it signifies> relief from sorrow and distress, and if <the dreamer> is in debt, God will unexpectedly release him from it, and probably he will receive a tremendous amount of money. If he knows the perpetrator, he will receive wealth and benefits (khayr) from him. And if the killer is a woman or a
<sup>85</sup> The above passage is followed by the interpretations from Artem. i.35.
<sup>86</sup> The same interpretation is repeated in Ibn Shahin, no. 2268.
#### CHAPTER FIVE
eunuch or an underage youth or a man without a beard, this signifies the one who will take his soul, whether by <natural> death, killing or something like that.
Given that Arabic dreambooks preserve passages corresponding, sometimes remotely and other times closely, to the interpretations cited in the Oneirocriticon, and that at least one of the interpretations found in the Oneirocriticon is based on a pun related to the etymology of the Arabic and not the Greek language (ghanam = sheep/ghanīma = booty), it is apparent that the immediate source of the Byzantine interpretations on heads and decapitation is not the Greek text of Artemidoros but Arabic dream interpretation.
#### Kings And Temples
Artemidoros and the Oneirocriticon agree that the utterances of kings who appear in dreams should be given credence without reservation.87 The relevant passage in Artemidoros is the following (ii.69; Pack 195, 3-8):
Των δὲ ἀξιοπίστων λεγομένων, οῖς λέγουσί τι [κατ' ὅναρ] πιστεύειν χρὴ καὶ πείθεσθαι, φημι πρώτους είναι θεούς: άλλότριον γάρ θεού το ψεύδεσθαι. Έπειτα ίερεῖς τῆς γὰρ αὐτῆς τοῖς θεοίς παρὰ τοῖς ἀνθρώποις τετυχήκασι τιμης. 8 είτα βασιλεῖς καὶ ἄρχοντας: "τὸ κρατοῦν γὰρ δύναμιν ἔχει θεουῖ"199
Among the people who are worthy of credence and whose words [in a dream] one must believe and obey, I maintain that the gods are first. For it is contrary to the nature of a god to tell lies. Then priests. For they enjoy the same respect among men as the gods. Then kings and rulers. "For to rule is to have the power of a god."
The Oneirocriticon argues that kings (or rather emperors, as the term basileus is understood in Byzantine Greek) are symbols of the person of Christ, who only speaks the truth, and therefore an emperor in a dream would never tell a lie. The connection of Christ with the emperor is a well-known component of Byzantine imperial and Christian ideology. 90 The interpretation of dreaming
<sup>87</sup> Oberhelman no. 23: On. 126 = Artem. ii.69: Kings who appear in dreams do not tell lies.
<sup>88</sup> According to Homer, Hypsenor, a priest of Scamander (Iliad 5.78), and Laogonus, a priest of Zeus (Iliad 16.605), were honored as gods; see Pack, testimonia, 195, 5-6.
<sup>89</sup> This is a verse by Menander (343-292 в.с.); see Pack, testimonia , 195, 7.
<sup>90</sup> Cf. F. Dvornik, Early Christian and Byzantine Political Philosophy: Origins and Background, 2 vols. (Washington, D. C., 1966), vol. 1, pp. 591-93; 567-68; 736. For the manifestation of this idea in the visual arts, see A. Grabar, L'empereur dans l'art byzantin (Paris, 1936), esp. pp. 98-122, which includes examples of works of art from the Middle Byzantine period, and pp. 244-61.
of emperors is contained in the second half of chapter 126, where the interpretation of the Christian cross is also discussed (Drexl 75, 20-76, 9):
ἐὰν ἴδη τις, ὅτι ἐλάλησε βασιλεῖ γνωρίμα, εἰ μὲν μνημονεύει τῶν λόγων τοῦ βασιλέως, κρατείτω αύτους ἀπαραλλάκτως, ὅτι ὁ βασιλεύς εἰς πρόσωπον τοῦ Χριστου έστιν και ούδέποτε λαλήσει ψευδος ...
Εαν ίδη ό βασιλεύς έτερον βασιλέα των πρό αύτου άρχαίων ή έτέρας γῆς, είς πρόσωπόν έστι του Χριστου και παν, ο έαν ήκουσε παρ' αυτού, κρατείτω βεβαίως. είς καλόν δε έσται τα λαληθέντα: ού γαρ κακόν από του Χριστού.
If someone sees that he spoke with an emperor he recognized, if he can remember the words of the emperor, let him hold on to them steadily, because the emperor is symbolic of the person of Christ and will never utter a lie ...
If an emperor sees another emperor from among the ancient ones before him, or from another land, that emperor signifies Christ, and let <the dreamer> confidently hold on to whatever he heard from him. For whatever was said will result in something good, since no evil comes from Christ. 91
Divinely sanctioned leadership was also exercised in the Muslim lands, a fact reflected in the Arabic dreambooks. Al-Dinawari's chapter on kings and sultans begins with the following interpretation: قال المسلمون السلطان في النوم (The) هو الله تعالى و رؤيتــه على حــال رضــاه دالـة على رضـاه تـعــالى.
92 See Aziz al-Azmeh, Muslim Kingship: Power and the Sacred in Muslim, Christian, and Pagan Polities (London and New York, 1997).
93 Al-Dinawari, fasl 9, bab 1, Esad Efendi 1833, fol. 85b; BN arahe 2745, fol. 146a. The interpretation is repeated in al-Muntakhah, chap. 30, p. 145; al-Nabulusi, vol. 1, p. 289, s.v. سلطان Both of these dreambooks repeat much of the analogous chapter in al-Dinawari. This interpretation seems to have been inspired by a hadith recorded (without an isnad) in al-Dinawari's introduction (maqāla 6, BN arabe 2745, fol. 36a): قال رسول الله ملي الله عليه وسلم خير ما ويرى احدكم في النوم ان يرى ربه او يرى نبيه او يرى ايويه مسلمين قالوا يا نبي الله وهل
The messenger of God, may the peace and blessings of God be upon him, said: "The most auspicious <thing> that one of you may see in his sleep is if he sees his Lord, or His Prophet, or his Muslim parents." <His companions> said: "Oh Prophet of God, does someone see his Lord?" He said: "The sultan, and the sultan is God" ). The problem is how the word sultan should be understood, since the meaning "ruler" or "sovereign" is a later one. Its literal meaning is "strength" or "might," and it is only with this meaning that it can be found in the Qur'an (17:35; 49:29; 51:38).
<sup>91</sup> A different interpretation of ancient kings can be found in chapter 132, which contains interpretations attributed to the pagan Persians and Egyptians on the dead, death and burial (Drex) 87, 18-21): ἐὰν ίδη ὁ βασιλεύς, ὅτι ἐκ τῶν προθανόντων βασιλέων ἀμίλησε ή ήκολούθησέ τινι, θλιψιν πολέμου δέξεται από έχθρον μωρών αγνωρίστων και τά τέλη αύτου είς νίκας αποβήσεται (If the king dreams that he conversed with or followed one of the kings who died before him, he will know sorrow due to war waged by stupid unknown enemies, and the end of this will result in victories).
Muslims say: "The sultan in sleep is God Almighty, and seeing him in a pleasant condition is a token of God's pleasure").
The idea that rulers who appear in dreams speak the truth is also expressed in one of the Arabic dreambooks examined:94 dLo ~j.J ~j,;-J I -:;~I i.!l\_,.L..c.J I (The most truthful of dreams is dreaming of a king or a subject). 95 In Artemidoros the truthfulness of kings ranks only third, after gods and priests. In the *Oneirocriticon* the truthfulness of emperors is absolute, and their identification with Christ, with all its felicitously Byzantine overtones, seems to have been inspired from the Arabic interpretation of the sultan as God, not from Artemidoros 's text.
The Arabic provenance of the medieval Greek chapter on emperors is betrayed by yet another interpretation: Ei µ£v EO"'tt v 0 ~amA.Eu~ ayvffiptcr'tO ~. A.oyt~fo9co ayyEAO~ l<Uptou (If the emperor is unknown, let him be reckoned as an angel of God). 96 The interpretation of a king as an angel in Arabic could be based on the identity in the spelling (d.Lo) and the similarity in the pronunciation of *malik* (king) and *malak* (angel). Or it could be a reflection in the *Oneirocriticon* of an Arabic interpretation ultimately inspired by I:Iunayn's translation of Artemidoros. As Strohmaier has already discussed at length, I:Iunayn's text very consistently renders the pagan deities as angels. Artemidoros states that the gods signify persons in a position of power and authority in three instances. In the first two (ii.36; Pack 163, 10 and ii.69; Pack 195, 3-5),
96 Drexl 75, 23.
<sup>94</sup>*Al-Muntakhab.* p. 14, I. 11.
<sup>95</sup> The interpretation of this passage is admittedly problematic. The phrase is included in an exposition of the principles of dream interpretation that is attributed to lbn Qutayba *(al-Muntakhab,* pp. 12-17), which was translated into French in Fahd, *La divination arabe* (pp. 317-26) from *al-Muntakhab,* BN *arabe* 2749. This exposition is similar in content but does not coincide word for word with any part of either the Ankara or the Jerusalem manuscript of lbn Qutayba; it simply summarizes the introduction to Ibn Qutayba's dreambook. I have been unable to locate a passage in lbn Qutayba's text that would correspond to this phrase and clarify it. (For the lacunae in both the Ankara and the Jerusalem manuscripts of lbn Qutayba, cf. Lamoreaux, "Dream Interpretation in the Early Medieval Near East," p. 47). Fahd seems to have been puzzled by its meaning, as well, because he felt compelled to cite the Arabic text in a footnote to his translation (p. 321, n. I). The only conjecture I can make is that the problem in the Arabic text was created by its author's effort to clarify the meaning of his sentence. Conceivably the phrase should be corrected to J~I .!l\_,.L...11 dL ~:u ~j\_i-J I (the most truthful dream is dreaming of a king of a subject), as opposed to 4..£.~I *:.,....* dL (an angel from among the angels), where the genitive *al-mam/uk* is placed in order to avoid the confusion between *malik* (king) and *malak* (angel), which are spelled identically as dL.. Such dubious words in unvocalized Arabic texts are sometimes clarified with a genitive, e.g . .:,L .......... :dl ~ (backbone of humans) for *\$Ulb* ="backbone" and not *\$alb=* "crucifixion", or ~ 1 "':-""!"L..:.. (eyebrow of the eye) for *l;iijib* ="eyebrow" and not "chamberlain."
he used a verse of Menander to make the point: το κρατουν γαρ δύναμιν έχει θεού (For to rule is to have the power of a god). As Strohmaier has observed, Hunayn rendered this verse in a different way each time." In chapter وذلك انهم مـسلطون على من يحـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ (Fahd, Artémidore 353, 6-7), literally "because they rule over the people who ان قـوة الملائكـة تـشـبـــه قـوة love them." But in chapter ii.36, Hunayn wrote الموالي (Fahd, Artémidore 239, 9), "because the power of angels is like the power of a maulā." The word maulā in Arabic has a number of different meanings, among them "master", "lord", and "God" (المولى)); it is also a form of address to sovereigns (مو لات). In other words, the passages cited above from Artemidoros's chapters ii.36 and iv.69, if read not in the Greek text, but in Hunayn's translation, could be understood as stating that angels are equivalent to sovereigns.
The third instance of Artemidoros connecting gods with persons of authority is in iv.69 (Pack 291, 4 ff): "Οτι οί θεοί τοῖς δεσπόταις τον αυτὸν ἔχουσι λόγον, και έν τφ περι θεών λόγω προείρηκα (That gods are the same as masters has already been demonstrated in the chapter on the gods). In order to prove his previous statements, Artemidoros narrates a relevant dream and concludes his chapter thus: και το όλον δεσπόται γονείς διδάσκαλοι θεοί τον αυτον έχουσι λόγον (Generally speaking, masters, parents, teachers and gods have the same meaning). 98 This part of Artemidoros's text did not survive in the unique manuscript containing the Arabic translation. But most of chapter iv.69, unfortunately without the introductory statement, is quoted by al-و الملائكة يشــبــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ
. والموالمو (Angels are like parents and the maulā).
It is therefore conceivable that the interpretation of the emperor as an angel in the Oneirocriticon through a series of translations and alterations of meaning ultimately originated in Artemidoros's interpretation of gods as humans with authority and power. If this is the case, it would not mean that the author of the Oneirocriticon was necessarily using the Arabic translation of Artemidoros, but only an Arabic source that had absorbed, and also possibly rephrased, the interpretations contained in Hunayn's text.
<sup>97</sup> Strohmaier, "Die griechische Götter," p. 140.
<sup>98</sup> Pack 291, 13-14.
<sup>99</sup> Fasl 3, bah 11, Esad Efendi 1833, fol. 41b, 11. 4-8; BN arahe 2745, fol. 51a.
Oberhelman remarked that Artemidoros and the Oneirocriticon agree that temples should be interpreted as kings. 100 Artemidoros puts this in the following words: Βασιλεύς και ναός και στρατιώτης και ἐπιστολὴ βασιλέως καὶ άργύριον και όσα δμοια τούτοις ύπ' άλλήλων σημαίνεται (A king, a temple, a soldier, a letter from a king, money and other similar things signify one another interchangeably). 101 Conversely, the Oneirocriticon reads thus: 102
Έαν Ίδη τις, ότι ό πρωτεύων ναός του τόπου έπεσεν ή ένεπρήσθη, ἔστι το πάθος διά θανάτου είς πρόσωπον του βασιλέως. εί δε τών κατωτέρων ναών τουτο έπαθε και ίδη τουτο ό βασιλεύς ή έτερός τις, είς θλίψιν ἀποβήσεται του βασιλέως: οί γαρ ναοι πάντες οί θεῖοι εἰς πρόσωπον του βασιλέως ή του κατά τόπον κυριεύοντος κρίνονται και ό καλλωπισμός και ή πτώσις είς έκεῖνον διαβαίνουσιν.
If someone dreams that the most important temple in the land fell or burned down, the accident indicates the death of the emperor. If one of the less important temples suffered this <fate>, and the emperor of someone else dreams it, it will be fulfilled as sorrow for the emperor. For all the sacred temples are interpreted as the person of the emperor or the local ruler, and their beautification or fall is transferred to him.
The most important Muslim shrine is the Great Mosque of Mecca (al-masjid al-haram) and its holy of holies is the Ka ba. This passage of the Oneirocriticon corresponds much more closely to the Arabic interpretations of the Ka'ba الكعبـة في , than to the passage of Artemidoros. Al-Dinawarīʾs interpretation (19) المنام خليـــفــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ الكعبية في الرؤيا خليـفة او امـيـر او وزير 44:amplified in al-Muntakhab is a dream is a caliph, or an amir, '05 or a vizier, and the fall of any of its walls signifies the death of the caliph). The last phrase of the Byzantine interpretation quoted
<sup>100</sup> Oberhelman, no. 33: On. 148 = Artem. iv.31: king = temple.
<sup>101</sup> Artem. iv.31; Pack 265, 11-13.
<sup>102</sup> Oneirocriticon, chap. 148, Drex1 103, 18-24.
<sup>103</sup> Al-Dinawari, fast 8, bab 38, Esad Efendi 1833, fol. 76b, BN arahe 2745, fol. 134a.
<sup>104</sup> Al-Muntakhah, p. 56.
<sup>105</sup> The title of amir was (and is) applied to various high-ranking persons depending on the era and the locality, but it also has a specific meaning in connection with the dar al-Islam, the Muslim oikoumene. Ideally, the whole of Islam is under the dominion of one leader, the caliph (khalifa), who is the viceregent of the Prophet on earth. The leaders of the provinces under him are called amīr (pl. umarā'); see Hughes, Dictionary of Islam, s.v. "Rulers." In this case the "caliph and amir" of al-Muntakhah correspond perfectly to the "emperor or the local ruler" of the Oneirocriticon.
earlier (" all the sacred temples are interpreted as the person of the emperor or the local ruler, and their beautification or fall are transferred to him") is equivalent to one found in Ibn Shāhīn (no. ومن ر اي الكعدة نقصا فهو ) it عــائد عـلى الخليفة او الامـام (If someone dreamt of damage to the Ka'ba, it pertains to the caliph or the imam). 106
The interpretations of kings and temples in the Oneirocriticon are much more relevant to the ones found in Arabic dreambooks than to those in Artemidoros, which suggests that the author of the Oneirocriticon based his relevant chapters on exclusively Arabic material.
#### Horses
Artemidoros interprets dreaming of horses as follows:
ίππον κέλητα ἐλαύνειν καλώς πειθόμενον τῷ ρυτῆρι καὶ αὐτῷ τῷ ἐλαύνοντι άγαθον ἐπίσης πᾶσιν ὕππος γὰρ γυναικὴ μὲν καὶ ἐρωμένῃ τὸν αὐτὸν ἔχει λόγον, ότι και επι κάλλει μέγα φρονεί και τον έλατηρα βαστάζει. δμοιος δέ έστι και πλοίω άλος μεν γαρ ίππους και ό ποιητής τας ναύς λέγει, ήμεῖς δε τον Ποσειδώνα "Ιππιον καλούμεν, και ον έχει λόγον έν γή ίππος, τον αυτόν και έν θαλάσση ναυς. δμοιος δ΄ αν είη και δεσπότη [καὶ] ἐργοδότη καὶ φίλω τρέφοντι και παντι τφ βαστάζοντι. όπως αν ούν ό ϊππος τον ίδόντα φέρη, ούτω καὶ γυνὴ καὶ ἐρωμένὴ καὶ δεσπότης καὶ φίλος καὶ ναῦς τὸν ἰδόντα διαθήσουσι. Ὁ׳'
Riding a racing horse well-one that obeys both the reins and the rider himself-is a good sign for everyone. For a horse means the same as a wife and a mistress, since it takes pride in its beauty and bears its rider. A horse also resembles a ship. For just as the poet calls ships "the horses of the sea,"108 we call Poseidon "Hippios," and a horse is to the land as a ship is to the sea. The horse is also like a master who farms out work, a friend who takes care of someone, and every man who carries a burden. The manner in which the horse carries the dreamer indicates how the dreamer will be treated by his wife, mistress, master, friend, or ship.
The entry on horses in the Oneirocriticon is much more extensive and gives many more details.109 It is structured differently and, though it does include the same interpretation ("riding a tamed horse well means goodness," as Ober-
<sup>106</sup> All three Arabic interpretations are repeated in al-Nābulusī, vol. 2, p. 173, s.v.
<sup>107</sup> Artem, i.56: Pack 64, 11 ff.
<sup>108</sup> Homer, Odyssey 4.708; see Pack, testimonia, 64, 15.
<sup>109</sup> It goes on for two and a half pages (Drex1 110, 23-113, 10). For the sake of brevity I will only examine part of it.
helman observed 10), it does not repeat any word or expression found in the Greek text of Artemidoros:
έάν τις ϊδη, ότι ἐποχούμενος φάρας τάξει διήρχετο ὑπείκοντος τοῦ ὕππου, εύρήσει άξίωμα καὶ φήμην μεγίστην ἀναλόγως τοῦ κάλλους τοῦ ὕππου καὶ τῆς ύποταγής. ἐαν ϊδη τις, ὅτι ἐπωχεῖτο τοιούτω Ἑππφ μέγα καὶ δασὺ καὶ μακρὸν τὸ ουραιον έχοντι, εύρήσει άκολούθους της έξουσίας αύτου άναλόγως του πλήθους καὶ τοῦ μήκους τοῦ οὐραίου. εἰ δὲ εἶχε δύο οὕραια ἢ πλέον, πλείων γενήσεται των άκολούθων αύτφ ή πληθύς. εί δε ίδη, ότι ό αυτος ίππος ήν κούντουρος ή σπανόουρος, εύρήσει μέριμναν και ύστέρημα τῆς ἐξουσίας αὐτοῦ ἀναλόγως της λείψεως των τριχών: εί δε ίδη, ότι ἐποχούμενος τῳ δασυτρίχω καὶ μακροούρω συνέβη κοπηναι τὴν οὐράν, εἰ μέν ἐστιν ὁ ἰδών αὐτεξούσιος καὶ ἄρχει λαοῦ, ταθτα άποβαλεί συντόμως. έαν ίδη τις, ὅτι ἐπωχήσατο τοιούτω μὲν Ἑππω εὐγενεί, χωλαίνοντι δέ, εύρήσει θλίψιν και πρόσκομμα, ἐν οῖς Ἑκαστος ἀγωνίζεται. ἐἰ δε ϊδη, ότι ό τοιούτος εὐγενὴς ϊππος ἀνυπότακτος καὶ σκληροχάλινος ήν, σκληρας άμαρτίας ἐπιβάτης ἔσται ὁ ἰδὸν καὶ πλεῖστα θλιβήσεται ἀναλόγως τῆς σκληρότητος του ίππου. 11
If someone dreams that while seated on a steed, he was proceeding in an orderly fashion as the horse yielded to his commands, he will have honor and great fame equal to the horse's beauty and obedience. If he dreams that this horse had a long and large shaggy tail, he will find followers of his authority in proportion to the size and length of the tail. And if the horse had two or more tails, the number of his followers will be even greater. If he dreams that this horse was dock-tailed or had a hairless tail, he will have anxiety and lack resources by analogy to the loss of hair. If the dreamer does not obey anyone's commands but is himself a leader of people, if he dreamt that, while mounted on a horse, it happened that its long and shaggy-haired tail was cut off, he will shortly see <his position> cut short. If someone dreams that he is riding a well-bred but lame horse, he will experience sorrow and obstacles in whatever he is striving for. If he dreams of a noble horse that is unruly and hardmounted, he will embark on obstinate sinning and be exceedingly distressed by analogy to the harsh nature of the horse.
Let us now compare the Byzantine passage with the interpretation of horses in the dreambook of Ibn Qutayba:
وقال المفسرون الفرس عز وسلطان فمن راى انه على فرس نلول يسير عليه رويدا واداة الفرس تامه اصاب عزا وسلطانا وشرفا ومرؤة في الناس بقدر ذل الفرس له و الشرف وذنب الفرس اتباع الرجل فان كان ذنوبا كثر تبعه وان كان مهلوبا محذوفا قل تبعه وكل عضو من الفرس شعبة من السلطان كقدر العضو في
<sup>110</sup> Oberhelman, Oneirocriticon of Achmet, Appendix III, no. 35: Oneirocriticon 152 = Artem. i.56: "riding a well-tamed horse means goodness."
<sup>11</sup> Drexl 110, 14-111, I.
~JI~~~ <\.\_ju 4..-,> ~ *J".r.A* .\_..le. ....:.1 i..SIJ .J-AJ .~'ii . J"~ I ~\_,...\_..., J..l.i.:i J\_,\_A1
The dream interpreters said: The horse is might and power. Whoever dreams that he was on a tractable horse that was gently behaving according to his wishes and the implements of the horse were perfect for him, he will obtain might, power, honor and valor among the people commensurate with the obedience of the horse to him .... And whichever implement of the horse diminished, it is a lessening of this power and honor. The tail of the horse is the following of a man. If it is bushy tailed, his following will become more numerous. If the tail is plucked out or is clipped, his following will diminish. And every member of the horse's body is a kind of power that corresponds to the role of this member in relation to the rest of the body. And if he dreams that he is on a horse that is disobeying him, he will perpetrate a sin, or horror will afflict him commensurate with the difficulty presented by the horse. <sup>112</sup>
The interpretation of horses in the *Oneirocriticon* closely corresponds to the structure and content of the same entry in lbn Qutayba. The final interpretation of the excerpt from the *Oneirocriticon* quoted above (Drexl 111, 10-13), which is identical to the final interpretation from the cited passage of lbn Qutayba, involves a double pun (£nroxi}crmo 'iJtJtCQ O'KATlpoxaA.i.vcq - O'KATlpa<; aµap·tiac; £m ~ci-tric; EO"tat 113). In Greek, the meaning of the expression crKA.ripac; aµap 'ti. ac; £m~ci-tric; £mm is clear (He will perpetrate a grave sin); but the wording is unusual. The expression £m~ai. vro aµap'ti.a <; (lit. "to ride on sin") occurs in no other Greek text of any period. The choice of words in this instance can be explained by a closer look at the analogous interpretation from lbn Qutayba. <sup>114</sup> An equivalent pun (L... *\_r-£* ~ *.J* = to ride on a horse *I* ~ *.!* . 0 • • ~.JI = to perpetrate a sin) comes naturally in Arabic, because the root ~.J in form I *(rakaba)* means "to ride, to mount", whereas in form VIII *(irtakaba)* it means "to perpetrate (a sin or a crime)." <sup>115</sup>The author of the *Oneirocriticon,* in his effort to convey the Arabic pun in Greek, coined an unusual expression, crKA.ripac; aµap'ti.ac; £m~a'tri<; fo'tat.
<sup>112</sup>*Yahuda ar.* 196, fol. 50b; the same interpretation is repeated verbatim in *al-Muntakhah,* chap. 33, p. 177.
<sup>113</sup> Identified by Oberhelman, *Oneirocriticon of Achmet,* p. 261, n. 406: "Two puns here: one on *epihates* ("rider" and "the one who embarks on"); the other, between *sk/erochalinos* ("hardmouthed") and *skleros* ("obstinate").
<sup>114</sup> Repeated in Ibn Shahin, no. 5571.
<sup>115</sup> Also "to pursue, to practice."
Chapter 152 of the Oneirocriticon ("From the Indians on Horses") concludes with dreaming of a horse that entered a house, a field, or a city quarter. The only analogous dream in Artemidoros (iv.46)116 is mentioned in the story of a man who dreamt that a horse was brought into his bedchamber. The dream signified that the man would lose access to his mistress. The interpretation given in the Oneirocriticon is very different:
Εάν τις ϊδη, ότι ίππος εὐειδὴς εἰσηλθεν είς οἰκον ή ἀγρὸν ἢ ρεγεωνα, ἀλλ ἀγνώριστος καὶ γυμνὸς καὶ ἀχαλίνωτος, ἐλεύσεται ἐκεῖ ἐξουσιαστὴς μέγιστος αναλόγως της ἐπιφανείας του ὕππου· καὶ εἰ μὲν ήν ὁ αὐτὸς ὕππος ῥινοφάλιος τετραπέδιλος, πλέων ἔσται ή ἐξουσία τοῦ μεγίστου.'']
If someone dreams that a beautiful horse entered a house, a field or the quarter of a city, but it was unknown, without a saddle and without a bridle, a great nobleman will come there, analogous to the appearance of the horse. And if this horse had a patch of white on its nose and was shod on all four feet, the noble's power will be even greater.
This passage corresponds to one found in the Arabic dreambook of Ibn Shāhīn (no. 5586):
ومن راي فرسا مجهولا يدخل دارا او ارضا لايعرف صاحبها ولا يعرف صلحبها فانه يؤول بقدوم رجل شريف، وإن عرف المكان كان قدوم ذلك الرجل ال
If someone dreamt that an unknown horse entered a house or a field, and the master <of the house or field> was not known, nor did he recognize the horse, this is interpreted as the arrival of a nobleman, and if the place was known, then this man will arrive at that very place.
Oberhelman observes that both Artemidoros and the Oneirocriticon state that a horse is also interpreted as a woman.118 In Artemidoros, this is briefly mentioned; in the Oneirocriticon, on the other hand, horses are interpreted as women in three different chapters (154, 230, and 231119) with many more
119 "From the Persians and Egyptians on Well-Bred Horses," "From the Indians, Persians and Egyptians on Male and Female Horses," and "From the Persians and Egyptians on the Colors of
<sup>16</sup> Pack 272, 6-15.
<sup>17</sup> Drexl 111, 19-23.
<sup>118 &</sup>quot;Specific agreement," no. 38: "Oneirocriticon 154 = Artem. 4. proem. : horse = woman." I was unable to identify which part of Artemidoros's treatise is meant by Oberhelman's reference to "4.proem." (the introduction to book iv does not contain such an interpretation). But the final phrase of Artemidoros's excerpt does state that a horse in a dream can indicate a woman. As White observed (Artemidoros, Interpretation of Dreams, trans. White, p. 76, n. 59), "Achmes (p. 112, 7-10) also interprets a horse to mean a wife."
details. Chapter 231 expounds on the interpretation of royal saddle horses according to their color:120
ό λευκότατος ϊππος έν θεωρία βασιλέως είς πρόσωπον της αύγούστης κρίνεται. ό λευκός σιδηραίος, εί έστιν βασιλέως σελλάριον, είς έλάτονα της αύγούστης γυναικα κρίνεται και έάν τις ΐδη, ὅτι ἔσχεν αὐτόν, εἰς ἐλάττονα ἐξουσίαν τούτο κρινέτω. σελλάριον δε βασιλικόν έάν ἐστι ροδόχρουν, εἰς γυναϊκα ἐπίχαριν κρίνεται και ό μέλας ίππος ό βασιλικός είς γυναίκα πλουσίαν και θλιβεράν κρίνεται δια το μέλαν.
A completely white horse, if dreamt by the emperor, is interpreted as the person of the empress. A gray horse, if it is a royal saddle horse, is interpreted as a woman of a rank lower than empress. If someone dreams that he possessed this horse, let this be interpreted as a lesser office. If a royal saddle horse is auburn, it is interpreted as a charming woman. A royal black horse is interpreted as a rich and oppressive woman because of the black color.
The interpretation of a horse as a woman can also be found in Ibn Qutayba, و الفرس الانق امـراة ) who states that "a beautiful horse is a noble woman ة ( ). In addition, al-Dīnawarī, in a series of very brief chapters, discusses horses according to their color, and two of his interpretations, that of a black and that of an auburn horse, are almost identical with those found in the Oneirocriticon:
من راي انه رکب فرسا اشهب تزوج امراة هندية وان کان مطيعا تطيعه الزوجة ... قـال ابن سيرين بل هى <=الفرس الاهدم> امراة متدينة موسرة فى ذكر وصيت ... ومن راى انه ركب فرسا اشقر فانه يتزوج امراة ذات فرج وجمال ...الابلق امراة مشهورة الجمال والمال ومن راي انه ركت فرسا ابلق تروج امراة بهذه الصفة
If someone dreamt that he rode a gray horse he will marry an Indian woman, and if the horse was obedient the wife will obey him.122 ... Ibn Sirin said: "A black horse is a devout woman who is wealthy because of her reputation and her renown."123 ... If someone dreamt that he was riding an auburn horse he will marry a woman of joy and beauty.124 ... A piebald horse signifies a woman who
124 Fasl 21, hāh 6, Esad Efendi 1833, fol. 209b. The Greek γυναϊκα ἐπίχαριν seems chosen in order to convey in one word a meaning expressed with two words in Arabic: اذات فرج ج وجمال (a)
Royal Horses and Saddle-Horses."
<sup>120</sup> Drexl 182, 17-23.
#### CHAPTER FIVE
is important for her beauty and wealth. If someone dreamt that he was riding a piebald horse he will marry someone according to this description.155
The examples of interpretations of horses from the Oneirocriticon that can also be found in Arabic dreambooks could easily be multiplied. Any further investigation into the extensive Greek and Arabic passages on horses would only strengthen the already evident conclusion: Artemidoros and the Oneirocriticon share no common interpretation that does not also occur in Arabic dream interpretation. Moreover, the correspondence of the Byzantine interpretations with the Arabic ones is much more direct and sometimes extends even to the choice of words in Greek that closely reflect the Arabic original.
#### Sun
As Oberhelman observed, Artemidoros and the Oneirocriticon agree that "to gleam like the sun means future rule for the dreamer."126 The relevant passage in Artemidoros narrates a dream dreamt by a man and its fulfillment:127
οίον έδοξέ τις Ήλιος γεγονέναι και διά της άγορας διέρχεσθαι έχων ἀκτίνας ένδεκα στρατηγός απεδείχθη της έαυτοῦ πόλεως καὶ μῆνας ένδεκα ἐπιζήσας τη αρχη άπέθανε διά το μὴ τέλειον έχειν των ακτίνων τὸν ἀριθμόν.
For example, someone dreamt that he became Helius, the sun god, and that he passed through the marketplace with eleven sun-rays. He was appointed chief magistrate of his native city and died after eleven months in office, since the number of sun rays was not complete.
The corresponding interpretation from the Oneirocriticon is repeated in two sentences: εί δε ίδη τις, ότι ώς ό ήλιος ακτινοβολών έγένετο, αναλόγως του φωτος βασιλεύσει (If someone sees that he is emitting rays like the sun, he will rule by analogy to the light);138 έαν ίδη τις, ότι ἐξήρχοντο ἀπ' αυτού άκτινες κωλύουσαι τον λαόν του μὴ βλέπειν αύτόν, και αύτος βασιλεύσει (If someone sees that rays are coming out of him that make it impossible for the people to look at him, he will also rule).129 An identical interpretation can be found in Arabic dreambooks. According to Ibn Qutayba,
woman of joy and beauty).
<sup>125</sup> Fasl 21, bab 7, Esad Efendi 1833, fol. 209b.
<sup>126 &</sup>quot;Specific agreement," no. 45: "Oneirocriticon 166 = Artem. iv.49."
<sup>127</sup> Pack 276, 1-4.
<sup>128</sup> Drex1 127, 13-14.
<sup>129</sup> Drexl 127, 17-19.
"If someone dreams that he was transformed into the sun, he will obtain sovereignty by analogy to the rays, if his social origins are appropriate" ( من (زاي انه تحول شمسا اصاب مُلُكا يقدر الشعاع ان كان لذلك اهلا (( {) {) { } and to al-Dinawari, "If someone dreams that he is transformed into the sun, he will obtain mighty sovereignty analogous to his rays" (فــمـن راي فــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ 13/تحول شمسا فانه بصنب مُلَكا عظيما على قدر شعاعها.
The Oneirocriticon and Artemidoros dedicate about two pages each to the interpretation of the sun, 132 but there are no further correspondences between the ancient and the Byzantine material. Artemidoros interprets the sun in a wide variety of ways, while the Oneirocriticon repeatedly insists that the sun should always and only be interpreted as a sovereign: 133 ό ήλιος είς πρόσωπον βασιλέως κρίνεται απαραλλάκτως (The sun is unfailingly interpreted as the person of the king);134 ὁ ήλιος εἰς μέγαν βασιλέα κρίνεται (The sun is interpreted as a great king).135 The condition of the sun in the dream generally refers to the condition of the king, while proximity to the sun means proximity to royalty for the dreamer. The equation of the emperor with the sun is a well-known metaphor of imperial rhetoric inherited from Rome and used throughout the Byzantine millennium. 136 Its presence in the Oneirocriticon is as much a reference to this established Byzantine image as it is a reflection of the interpretations found in Arabic dreambooks. The sun is already interpreted as a king in the earliest surviving Arabic dreambook, that of Ibn Qutayba, where the interpretation is supported by a verse from a pre-Islamic poet: 137
الشمس ملك عظيم و كل ما راه قد حدث بالشمس من تغيراوكسوف او علّة فهو حدث باللك من هم ومرض واشباه ذلك
135 Drexl 129, 8.
136 See H. Hunger, Reich der neuen Mitte. Die christliche Geist der byzantinischen Kultur (Graz, Vienna, Cologne, 1965), pp. 96-103.
137 Fasl 13, Yahuda ar. 196, fol. 28b.
<sup>130</sup> Fasl 13, Yahuda ar. 196, fol. 28b.
<sup>131</sup> Fasl 15, bab 24; Esad Efendi 1833, fol. 154a.
<sup>132</sup> Artem. ii.36, Pack 160, 25-162, 27; Oneirocriticon 166, Drexl 127, 1-129, 11.
<sup>133</sup> In Byzantine Greek the term basileus generally refers to the Byzantine emperor. Only under certain circumstances was it applied to foreign rulers; see ODB, s.v. "Basileus." The chapter of the Oneirocriticon interpreting the sun as a basileus is attributed to the Indians, Persians and Egyptians together. Therefore, in the translations that follow, the term basileus will be consistently rendered as "king."
<sup>134</sup> Drexl 127, 3.
#### CHAPTER FIVE
اصاب مُلكا بقدر الشعاع ان كـان لذلك اهـلا الشمس او ملكه نال من الملك بقدر ما ملك منها فان ملكها وهى سـوداء مظلمة اضطر اليه المُلك فى امر يكون حاله منها [sic] كحالها قـال النابغة الذبياني بانك شمس والملوك كواكب يااذا طلعت لم يبد منهو [sic] كوكب.
The sun is a powerful king and whatever change or eclipse or defect the dreamer saw as having happened to the sun, <an analogous> sorrow or illness or the like will happen to the king. If someone dreams that he is transformed into the sun, he will gain sovereignty analogous to his rays, if his social origins are appropriate. And if someone dreams that he took command or possession of the sun, he will obtain sovereignty analogous to his possession of the sun. And if the dreamer possesses the sun, and it is black or darkened, the realm will be in need of him in a <certain> matter which will be in relation to it <?> like the sun's situation. Al-Nabigha al-Dhubyani said: "All other kings are stars and thou a sun. When the sun rises, lo, the heavens are bare!"138
Almost every interpretation of the sun contained in the Oneirocriticon can be correlated with analogous interpretations found in the Arabic dreambooks, as is evident in the following passages:
Drexl 127, 7-9: ἐαν ίδη τις, ὅτι ἐπλησίασε τφ ήλιακώ δίσκω ἢ ὅτι κατέσχεν αυτόν, εύρήσει χαραν έκ του βασιλέως ανάλογον της έγγύτητος.
If someone dreams that he approached or took hold of the solar disk, he will find from the king joy commensurate with <his> proximity <to the sun in the dream>.
Cf. al-Dīnawarī, fasl 15, bāb 24, Esad Efendi 1833, fol. 154a:
فان راى انه تعلق بها <=الشمس> فانه ينال قوة وخيرا من وزير او كاتب ... فان راي انه ملك الشمس او بمكن منها فانه بكون مقبول القول عند الملك ا لا عظم .
If he dreams that he hung onto the sun he will receive power and beneficence through a royal minister or secretary .... If he dreams that he took hold of the sun or that he had power over it, his views will be well received by the greatest king.
<sup>1.8</sup> Lit.: "no star appears because of him," i.e., the light from the sun is so strong that no other luminary can compete with it, and therefore other luminaries are as if they did not exist; see also Kister, "Interpretation of Dreams," p. 95. Al-Nābigha al-Dhubyānī (end of the 6th century) is one of the most famous poets of pre-Islamic Arabia. The line cited by Ibn Qutayba is from a poem addressed to his patron, King al-Nu'man b. al-Mundhir Abū Qābūs of Hīra, after he fell out of favor with him. It is cited according to the translation of R. A. Nicholson, A Literary History of the Arabs (Cambridge, 1930; rpt. Surrey, 1993), pp. 122-23.
Drexl 127, 11-13: εί δέ τις ίδη, ότι ἐμάχετο ἢ ἐπολέμει τῶ ήλίω ἀκτινοβολούντι, βασιλέα πολεμήσει. εί δε τού κοινού λαού, άπὸ βασιλέως τιμωρηθήσεται.
If someone dreams that he was fighting or was at war with the gleaming sun, he will fight a king. If he is a commoner, he will be punished by the king.
Cf. al-Dīnawarī, fasl 15, bāb 24, Esad Efendi 1833, fol. 154b:
فان راي انه ناز ع الشمس فانه يخرج على الملك خارجيا ان كان كذلك اهلا.
If he is fighting the sun, he will revolt against the king, if his social background is appropriate.
Drexl 127, 16-17: γυνὴ ἐαν Ίδη, ὅτι ἐγέννησε ἤλιον, βασιλέα τέξεται, ὅτις αν ή,
If a woman dreams that she gave birth to the sun, she will give birth to a king, whoever she might be.
Cf. al-Dīnawarī, fasl 15, bāb 24, Esad Efendi 1833, fol. 154b:
فان طلعت <الشمس> في فرجها فانها تلد ملكا.
If <a woman dreams that> the sun rose from her genitals, she will give birth to a king.
Drexl 127, 20-22: Εαν ίδη τις, ότι τὴν σφαιραν του ήλίου κρατεί άνευ ἀκτίνων ουσαν, εί μέν έστιν άσθενής ή έν θλίψει ό ίδών, ύγιαίνει και χαρήσεται ….
If someone dreams that he was holding the sphere of the sun without rays, if the dreamer is sick or troubled, he will get well and be happy ....
Cf. Ibn Shāhīn, no.127:
ومن راي انه اخذ الشمس بيده لکن لا من السماء و لا نور لها و لا شعاع و أنها لم تكن مظلمة يحصل له الفرج من الغموم.
If someone dreamt that he took the sun in his hand but not from the sky, and <if the sun> had no light or rays but neither was it dark, he will be relieved from his sorrows.
Drexl 127, 24-26: εί δὲ ΐδῃ τις, ὅτι ἀκτινας ἐξέλαμψεν ὁ ὕλιος ἐν τῷ οϊκω αίτου ένδον άσυνήθως, χαράν και πλούτον εύρήσει παρά του βασιλέως αναλόγως του φωτός ....
If someone dreams that the sun unexpectedly shone with its rays inside his house, he will receive from the king joy and wealth commensurate with the light.
Cf. al-Dinawarī, fasl 15, bāb 24, Esad Efendi 1833, fol. 154a:
فان طلعت <الشمس> فى بيته نال سعدا من السلطان.
If the sun rose in his house, he will receive good fortune from the sovereign.
A similar interpretation is attributed by al-Dīnawarī to the Christians (fast 15, bāb 24, Esad Efendi 1833, fol. 155a):
فان راي ان الشمس طلعت فى الدار فاضاءت الدار كلها فانه ينال عزا وكرامة وذكرا وشرفا وجاها ومرتبة
If someone dreamt that the sun rose in <his> house and cast light on all of it, he will obtain power, honor, reputation, distinction, glory and high rank.
Drexl 127, 26-29: εί δὲ ΐδη, ὅτι ὁ ήλιος ήτοι ὁ δίσκος εἰσηλθεν ἐν τῶ οϊκφ αυτου, εί μέν ἐστι βασιλεύς, γεννήσει βασιλέα, εἰ δὲ μέγιστος ἢ του κοινοῦ λαου, ύψωθήσεται και χαρήσεται, τάχα και βασιλείαν ύποδέξεται.
If someone dreams that the sun, that is, the solar disk, entered his house, if he is a king, he will beget a king; if he is a nobleman or a commoner he will rise <to a position of power> and rejoice, and perhaps will even receive royal status.
Cf. al-Muntakhab, p. 217 sup:
وان اخذها فى كفه او ملكها فى حجره او نزلت عليه فى بيتـه تورها ووضيآائها تمكن من سلطانه وعز مع ملكه ان كان ممن يليق
If <the dreamer> took the sun in his palm or took hold of it in his lap, or <the sun> descended upon <the dreamer's> house with its light and shine, <the dreamer> will come to possess sovereign power and will be strong in his dominion, if he is fit for such a thing.
Drexl 128, 2-5: ἐὰν ϊδη, ὅτι ἐν όδῳ ἐθέρμαινεν αὐτὸν ή τοῦ ήλίου φλόξ, εὐρήσει πλούτον παρά του βασιλέως άναλόγως της θερμάνσεως εί δε ΐδη, ότι πλειστα απέκαυσεν αύτόν, ευρήσει τιμωρίαν άναλόγως τῆς καύσεως.
If he sees that the flame of the sun warmed him while on a trip, he will receive wealth from the king by analogy to the warmth. If he sees that it burnt him severely, he will receive punishment analogous to the burning.
Drexl 128, 22-24: ἐκ πάντων γνώθί τε, ὅτι ὁ ὕλιος εἰς βασιλέα κρίνεται καὶ εῖ τις έξ αύτου ώφεληθή ή βλαβή, άναλόγως προσδοκάτω.
Overall you should know that the sun is interpreted as a sovereign. If someone incurs benefit or damage from the sun, he should expect things to turn out accordingly. Cf. al-Dīnawarī, fasl 15, bāb 24, Esad Efendi 1833, fol. 154b:
فان راى خرجت من الشمس نار احرقت نجو ما من حواليها فان الملك يطرد حاشيته .
If he saw that fire came forth from the sun that burnt stars among those around it, the king will dismiss his entourage.
#### Cf. also al-Muntakhab, p. 218 inf:
فان راي نارا خرجت من اشمس فأحرقت ما حواليها فان اللك يهلك اقواما مر: حاشيته.
If he sees fire come forth from the sun and burn what is around it, the king will destroy members of his entourage.
Drexl 128, 6-10: Έαν ίδη τις τον ήλιον έν τφ ούρανω άφωτιστον και ανευ ακτίνων, το πάθος και ή άδοξία είς πρόσωπόν έστι τοῦ βασιλέως εί δέ ἐστιν ὁ ίδων ἐκ τῶν μεγιστάνων, καταφρονήσει του βασιλέως εί δὲ ϊδη, ὅτι ὁ λαὸς όμοίως αυτω έθεώρει τουτο, όμοίως και ό λαος ἐξουθενήσει τὸν βασιλέα.
If someone sees the sun in the sky without light and rays, the damage and ignominy should be reckoned at the expense of the person of the king. If the dreamer is one of the noblemen, he will disdain the king. If he dreams that the people were also looking at this sight with him, the people will likewise destroy the king.
Cf. al-Dīnawarī, fasl 15, bāb 24, Esad Efendi 1833, fol. 154b:
فان رای الشمس خمدت فانه فساد فی مملکته باق.
If he dreamt that the sun died out, this is permanent disorder in his realm.
Cf. also al-Dīnawarī, fast 15, bāb 24, Esad Efendi 1833, fol. 154b:
فان راى الشمس وليس لها شعاع فانه ينقص من هيبة الملك بقدر ما نقص من شعاعها فان رای ان شعاعها يقع عليه فان کان صاحب الرئيا سلطانيا فان هيبة الملك يوضع عنه وان كان قائدا فان جاهه يذهب فان كان واليا فانه يعزل.
If he saw the sun and it had no rays, this is damage to the dignity of the king commensurate with the damage to the rays of the sun, and if he saw that its rays fell upon him, if he is a royal <person> he will be stripped of his royal dignity, and if he is a military officer (ga'id) his glory will vanish, and if he is a local governor (wali) he will be deposed.
Cf. also Ibn Shahin, no.131:
ومن راي الشمس على الارض ولا ضوء لها يدان على عزل ملك ذلك المكان.
If someone dreamt that the sun was upon the earth without shining, this signifies the removal of the king from that place.
Drexl 128, 10-15: έαν ίδη τις, ότι του ήλίου έγένετο έκλειψις, θλίψιν και πόλεμον ἐχθρῶν εὑρήσει ὁ βασιλεύς: εἰ δὲ ἴδῃ, ὅτι ἀπεκαθαρίσθη, εἰς νίκας έσται του βασιλέως, εί δε μή, τούναντίον. εί δέ τις ϊδη, ότι νεφέλαις έκαλύφθη, εύρήσει θλίψιν ό βασιλεύς και νόσους άναλόγως της καλύψεως.
If someone dreams that the sun was eclipsed, the king will face sorrow and war against enemies. If he dreams that the sun became visible again, this indicates that the king will be victorious. If not, the opposite. If someone dreams that <the sun> was covered with clouds, the king will face sorrow and sickness commensurate with the covering.
The same interpretation is repeated in Drexl 128, 19-20:
έαν ίδη τις, ότι νεφέλαις λεπταις έκαλύφθη ό ήλιος, θλιψιν και νόσημα ελάττονα είς τον βασιλέα νοείτω εί δε ίδη, ότι άπεκαθαρίσθη, είς ύγείαν και χαραν μεταβάλλεται ό βασιλεύς.
If someone dreams that the sun was covered with thin clouds, he should reckon <this> as sorrow and a minor sickness of the king. If he sees <the sun> become clear again, the king will return to health and joy.
Cf. al-Dinawarī, fasl 15, bāb 28, Esad Efendi 1833, fol. 156b:
و من راي سحاب غطي الشمس حتى ذهب نور ها فان اللك يمرض
If someone dreamt that clouds covered the sun until its light vanished, the king will become sick.
Cf. al-Muntakhab. p. 216 inf:
وان راي بها کسوفا او غشاها سحاب او تراکم علیها غبار او دخان حتی نقص نورها... كان ذلك دليلا على حديث يجرى على المضاف اليها اما من مرض او هم او غم او كرب او خبر مقلق.
If someone saw that the sun was eclipsed, or that a cloud concealed it, or that dust clouds or smoke gathered upon it until they obscured its light ... this signifies that
something will happen to the person who is symbolized by the sun, either sickness or trouble or grief or worries or disquieting news.
Drexl 128, 15-19: εί δε τδη αυτον άνατείλαντα καθαρώς έν τη κλίνη αύτου, εί μεν ούκ έχει γυναικα, λαμβάνει πλουσίαν, εί δε έχει, συμφιλιάσει πλουσίαν καὶ εὑρήσει ἐξ αὐτῆς χαρὰν εἰς βασίλεια καὶ ἀξίωμα καὶ ὕψος ἀναλόγως τοῦ φωτός.
If someone sees <the sun> rising brightly in his bed, if he does not have a wife, he will get a rich one. If he does have a wife, he will make love to a rich woman and, through her <influence>, will gain favor in the palace as well as an office and high rank commensurate with the light.
Cf. al-Dīnawarī, fasl 15, bāb 24, Esad Efendi 1833, fol. 154b:
فان<الشمس> طلعت فى بيته تزو ج.
If the sun rises in his house he will get married.
Cf. also Ibn Shahin, no. 129:
ومن راى الشمس مضيئة قد طلعت في بيته خاصة يخطب امرأة من اقاربه وإن راها طلعت فى بيت غيره يخطب امرأة من الأجانب وفى كلاهما يحصل له خير ومنفعة من اهل تلك المراة.
If someone dreams of the sun shining and just rising in his house only, he will propose to a woman from among his relatives. And if he sees the sun rising in the house of one of his friends, he will propose to a woman from outside <his own> clan and both of them will receive wealth and benefits from the family of the woman.
Drexl 128, 25 ff: Έαν ίδη τις, ότι ό ήλιος και ή σελήνη και οί άστέρες πάντες μετα του οικείου φωτὸς συνήχθησαν εἰς ἕνα τόπον, καὶ εἶδεν, ὅτι ἐξουσίασεν αύτων, βασιλεύσει ή γενήσεται τοιούτος, ώστε ύπ' αυτού διοικείσθαι τον βασιλέα και πάντα .... εί δε τδη τις, ότι ό ήλιος και ή σελήνη και οί άστέρες συνήχθησαν αφώτιστοι και σκοτεινοί και έξουσίασεν αύτων, εί μέν έστι μεγιστανος, είς απώλειαν και τιμωρίαν έκ παντος ήξει δια το σκότος, εί δέ έστι βασιλεύς, από πάντων κυκλούμενος πολεμηθήσεται και είς θλιψιν μεγάλην ຖັ້ຍເ.
If someone dreams that the sun, the moon and all the stars with their proper light gathered in one place, if he saw that he had power over them, he will either become king or become so <powerful> that the king and everyone else will be controlled by him ... If someone dreams that the sun, the moon and the stars were gathered <but were> dark and without light, he will by all means end up destroyed and punished, because of the darkness <of the heavenly bodies>. And if he is king, he will be surrounded and attacked by all and will come into great sorrow.
#### CHAPTER FIVE
Cf. al-Dīnawarī, fasl 15, bāb 26, Esad Efendi 1833, fol. 156a:
قال المسلمون من راي ان الشمس والقمر والنجوم اجتمعت في موضم فملكها وكان لها نور فانه مقبول القول عند الملك والوزير واشراف الناس فان لم يكن نور فانها مصيبة لصاحب الرؤيا...فان راى الشمس والقمر والنجوم مجتمعات فى موضعها من السماء بنورها وشعاعها وكان من راهن يصلح اللك والسلطان ملك امر اللك وامــر وزرايه وان لم يكن لهن نور وشعاع فهو دمار صاحب الرؤيا.
The Muslims said: "If someone dreamt that the sun, the moon and the stars were gathered in one place and that he had power over them, and that they were lit, his views will be well received by the king, the minister, and the noblemen among the people. But if they had no light, this is a disaster for the dreamer ... If he dreamt that the sun, the moon, and the stars were gathered in their place in heaven with their light and rays, if the dreamer was suitable for sovereignty and power, he will come to possess the rank of a king and his minister. But if they had no light or rays, this is ruin for the dreamer.
Cf. also al-Muntakhab, p. 219 inf:
ومن راي الشمس والقمر والنجوم اجتمعت فى موضع واحد وملكها وكان لها نور وشعاع فانه يكون مقبول القول عند اللك والوزير والرؤساء فان لم نكن نور فلا خير فنه لصاحب الروّيا.
If someone dreamt that the sun, the moon and the stars were gathered in one place and that he had power over them and that they had light and rays, his views will be well received by the king, his minister and the magistrates. However, if <the heavenly bodies> had no light, this is not a good omen for the dreamer.
Drexl 129, 8-11: Ο ήλιος είς μέγαν βασιλέα κρίνεται και έάν έστιν ό ίδων τας είρημένας όψεις και κρίσεις ή έν τοις τόποις ή έν τοις θέμασι του βασιλέως, αποβήσεται έν αύτφ, εί δὲ άλλης ἀρχῆς καὶ ἐξουσίας, δμοίως ἀποβήσεται εἰς τὴν κεφαλὴν αὐτού.
The sun is interpreted as a very powerful king. If the person who dreamt the aforementioned dreams and their attendant interpretations is either in the lands on the provinces of the king, the outcome <of the dream> pertains to the king. If <the dreamer was in the territory> of another authority and power, the outcome <of the dream> likewise pertains to the local leader.
Cf. al-Muntakhab, p. 216:
وربما دلت«الشمس» على ملك المكان الذى يرى الرؤيا فيه.
Possibly the sun signifies the ruler of the place where the dream was dreamt.
#### Moon
According to Oberhelman's list of "specific agreements," Artemidoros and the Oneirocriticon agree that the moon symbolizes a queen, 139 though why he says so is unclear: neither Artemidoros nor the Oneirocriticon-not even in Oberhelman's own translation-gives such an interpretation. The only analogy between the two dreambooks on the subject is that Artemidoros considers the moon to be the wife of the dreamer (Σελήνη γυναίκα σημαίνει του ίδόντος 40) and the Oneirocriticon that the dreamer will get married if he dreams of the moon in his home or close to him.
In Artemidoros the moon generally stands for a woman. In a dream it can indicate either the wife, mother, daughter, or sister of the dreamer, or that the events foretold in the dream will not take place without the agency of a woman. 14 The word "moon" in Greek is feminine, and the Moon (Selene) was accordingly a female deity. The interpretation of the moon in the Oneirocriticon, on the other hand, which covers approximately one printed page, refers primarily to a man, in fact the second most powerful man after the king (ή σελήνη εἰς πρόσωπον του από βασιλέως δευτέρου έξουσιαστου <κρίνεται>); 42 and again, και ή σελήνη, ώς είρηται, είς πρόσωπον του δευτέρου ἐστὶν ἀπὸ του βασιλέως (The moon, as was mentioned, is the second in command after the king); 13 ή σελήνη είς τον δεύτερον ήτοι τον έλάσσονα του βασιλέως κρίνεται (The moon is interpreted as the man who ranks second after, that is inferior to, the king). 44 This is somewhat surprising, until one realizes that the moon (gamar) in Arabic is masculine. Arabic dreambooks, from where the interpretations of the Oneirocriticon were lifted, interpret the moon primarily
<sup>139 &</sup>quot;Specific agreement," no. 47: "Oneirocriticon 166 = Artem. ii.36: moon = queen."
<sup>140</sup> ii.36, Pack 163, 1.
<sup>141</sup> Artemidoros ii.36, Pack 163, 1-3 and 164, 4-5.
<sup>142</sup> Drexl 127, 4-5.
<sup>143</sup> Drexl 129, 12-13.
<sup>14</sup> Drexl 130, 18-19. The "second in authority after the king" cannot be construed as the queen, because the Oneirocriticon repeatedly and consistently states that the planet Venus symbolizes the wife of the king. Moreover, the word used to refer to "the second most powerful person after the king" in Greek is of masculine gender.
as the royal vizier. Ibn Qutayba states that "the moon seen in the sky in its proper condition is the vizier of the king" (Lo dil <sup>I</sup>*..r;:.,* .J ~I 0~ WI .J .UL:.. ~ .~I ~ ~). <sup>145</sup>According to al-Dinawari, "The full moon is the vizier, since the sun is the caliph or the most powerful king"(~ .;~I ~~I d.lil.J ~~I ~I ~LS ljl ..r;j\_,...ll). <sup>146</sup>*Al-Muntakhab* concurs: "The moon is primarily the vizier of the most powerful king or a ruler other than the most powerful king" (JI r-k-dl dill J-:!j.J ~~I~.; ;; II ~~I dill 0.JJ 0l...b..L...).147
In the course of the ninth and the tenth centuries the vizier *(wazir)* was the caliph's prime minister, usually responsible for financial administration. The name of the office alludes to the Quranic designation of Aaron as the "support" ( *wazir)* of Moses and indicates that its holder was considered to be the caliph's primary assistant and the agent of his authority. The interpretations in the *Oneirocriticon,* lifted from Arabic sources that were written before the end of the tenth century, reflect the prestige of the vizier under the early Abbasids. <sup>148</sup> In the Byzantine court, on the other hand, there existed no single dignitary regarded as second in rank to the emperor. The most influential individuals after the emperor in the ninth and the tenth century held a variety of titles: Bardas was *caesar;* Stylianos Zaoutzes was *basileopator;* Basil the Bastard was *proedros.* The absence of a Byzantine equivalent is reflected in the periphrastic, descriptive, and rather inconsistent manner by which the rank of the vizier is rendered in Greek in the *Oneirocriticon: "b* odrtEpo<; am) wu ~acrt-*Ai\_ox;"* (the second in command after the emperor) or "6 npo:rt:Eucov am) ·i:ou ~acrtAi\_co<;" (the first in command after the emperor) or "6 OEU't:Epo<; fl1:01 6 £A-acrcrcov wu ~acrtA-£co<;" (the man who ranks second after, that is inferior to, the emperor).
What is the source of the interpretation of the moon as marriage? Could it be the influence of Artemidoros's interpretation of the moon as a woman? It is a topos in Arabic literature to liken both male and female beauty to the beauty of the moon. In the words of *al-Muntakhab* (p. 220), <~I> .;~
<sup>145</sup>*Bah* 13, *Yahuda ar.* 196, fol. 29a.
<sup>146</sup>*Fa\$! IS, bah* 11, *Esad Efendi* 1833, fol. !Sib.
<sup>147</sup>*Al-Muntakhah,* p. 220.
<sup>148</sup>For the evolution of the vizierate, see Sourdel and Sourdel, *La civilization de/' Islam classique,* pp. 247-51.
يشيه به ذو الحمال من النساء و الرجال فيقال كانه البدر وكانه فلقة قمر (The light of the moon is like someone who possesses beauty, either woman or man, because it is said that he is like a full moon and like a half moon). Therefore, the brief passages of the Oneirocriticon that interpret the moon as marriage by a man to a woman (and also by a woman to a man) are not influenced by Artemidoros but reflect the logic of Arabic dreambooks. Proof is furnished by the Arabic interpretations that match those of the Oneirocriticon, as is evident from the following examples (Drexl 129, 19-22 and 130, 21-29):
Έαν ίδη τις τὴν σελήνην ἐπιλάμψασαν αὐτφ ἐν τῷ οῖκω αὐτου, εὑρήσει χαρὰν και άντίληψιν άπό του βασιλέως και έάν έστιν αγαμος, λαμβάνει γυναίκα της χαράς και πλουτήσει.
If someone dreams that the moon shone on him in his house, he will receive joy and succor from the king; 49 if he is a bachelor, he will acquire a woman of joy 150 and become rich.
εί δε έγγύθεν ϊδη αύτην ή έν τοις κόλποις ή έν τη χειρι κατάσχη, εί μέν έστι ἀνήρ, λήψεται γυναῖκα χαρίεσσαν ἀναλόγως τῆς πληρώσεως τοῦ πεφωτισμένου δίσκου, εί δε γυνή, όμοίως και αυτή λήψεται ανδρα και ούτως αποβήσεται αυτή ... εί δε ΐδη, ότι ώράθη αυτή σκοτεινή και ἀσέληνος, ἐπὶ κατάρα καὶ απωλεία γυναίκα λήψεται, τάχα και τιμωρηθήσεται παρά του βασιλέως.
If one dreams that one was close <to the moon> or that one was holding it in one's lap or hand, if he is a man he will acquire a charming woman analogous to the fullness of the lit disk. If the dreamer is a woman, she will also acquire a husband, and the fulfillment of her dream will be similar ... If he dreams that <the moon> appeared dark and without light, he will acquire a woman who will bring about a curse and perdition, and <the dreamer> will soon be punished by the king.
These interpretations correspond to the following Arabic passages:151
وانما يكون القمر وزير الملك ما رُئى فى السماء على حاله فان راه عنده او فى حجره او فى يديه تزو ج زوجا بقدر ضوءه ونوره رجل كان او امراة .
The moon seen in the sky in its proper condition is the vizier of the king. But if <the dreamers> see it next to them or in their lap or in their hands, they will marry
<sup>149</sup> Possibly the Greek should read από του πρωτεύοντος από του βασιλέως (from the most powerful man after the king).
<sup>1.50</sup> The meaning of youalka tig xapac is unclear. It could be "a woman with a cheerful character" rendering something like مراة ذات فرج ج or or "a woman meant for pleasure," rendering something like امة or حاربة (a slave girl), as can be found in Ibn Shāhīn, no.149 (quoted below).
<sup>151</sup> Bah 13, Yahuda ar. 196, fols. 28b-29a.
#### CHAPTER FIVE
a spouse analogous to the shine and light of the moon, whether <the dreamer be> man or woman. 152
The same interpretation, accompanied by an additional detail that also appears in the Oneirocriticon, is repeated by al-Dinawari: 53
فان راى في حجره او في يده غير ساقط ولا منقطع فى الارض فانه يتزوج زوجا بقدر ضوء ذلك القمر رجلا راه او امراة ...فان راه كدرا تزوج غير كفو ﺍﻟﻤﻌﺎﺭﺿﺔ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘﺤﺪﺓ ﺍﻟﻤﺘ
If someone dreamt <that the moon> was in his lap or in his hand and it did not fall or break on the earth, he will marry a spouse analogous to the shine of this moon. whether <the dreamer be> man or woman ... and if he saw that the moon was murky, he will marry someone who is not an appropriate match for him.
An even better developed version of this interpretation can be found in Ibn Shāhīn, no. 148:
وقال جابر المغربى: من راى القـمـر فى يده او عنده يدل على انه يخطب امراة، فإن كان القمر هلالا فإنه تكون المراة دون في اصل والنسب، وإن كان نصف القمر مظلما تكون المراة من او لاد الموالى، وإن كان بدرا تكون اعلى منه فى الاصل والنسب، وإن رات هذه الرؤيا امراة يطلبها بعل ويكون حكم ﺫﻟﻚ ﺍﻟﺘﻌﻴﻴﺮ ﻋﻠﻰ ﻣﺎ ﺗﻘﺪﻡ.
Jabir al-Maghribī said: "If someone dreams that the moon is in his hands or near him, it signifies that he will become engaged to a woman. If the moon is a crescent, the woman will be inferior to him in descent and lineage. If one half of the moon is dark, the woman will belong to the offspring of the mawali.154 If the moon is full she will be superior to the dreamer in descent and lineage. If the dreamer is a woman, a husband will propose to her and the interpretation of her dream will be analogous to the form of the moon as was described above.
Ibn Shahin (no. 149) also contains an interpretation that directly corresponds to the structure and contents of Drexl 129, 19-21:
<sup>12</sup> Ibn Qutayba further emphasizes the interpretation of the moon as a spouse by telling the dream of 'A'isha, the daughter of Abu Bakr, one of the most prominent female figures in Islam. She saw that the moon fell in her room, and the interpretation was fulfilled when she married the Prophet (bab 13, Yahuda ar. 196, fol. 29a).
<sup>153</sup> Fasl 15, bab 8, Esad Efendi 1833, fol. 150a.
<sup>1.54</sup> The mawali were non-Arab converts to Islam who were adopted into Arab tribes by becoming "clients" of an Arab patron. The mawali in the Islamic state were second-class citizens, which generated considerable resentment. The mawali system was abolished during the reign of the Umayyad caliph 'Umar ibn 'Abd al-Azīz (d. 720).
$\alpha\_{i}^{\downarrow}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\iota\_{\bullet}$ $\omega\_{\bullet}$ $\omega\_{\bullet}$ $\omega\_{\bullet}$ $\omega\_{\bullet}$
If he dreamt that an immaculate moon rose in his house, it signifies that he will obtain wealth and beneficence from the king or that he will be betrothed to a woman and buy a slave girl.
Passages that strictly interpret the moon as the vizier are:
Drexl 130, 18-21: ή σελήνη είς τον δεύτερον ήτοι τον έλάσσονα του βασιλέως κρίνεται. ἐάν τις ἴδη αὐτὴν ἐν τῷ ούρανῶ, καθώς ἐστι μετὰ τοῦ φωτὸς αὐτῆς πλήρης, εύρήσει χαράν ἐκ τοῦ βασιλέως καὶ πλοῦτον ἐξ αὐτοῦ.
The moon is interpreted as the second, that is, the one immediately next in rank after the king. If someone sees it in the sky as it <usually> is, full and with its attendant light, he will receive joy and wealth from the king.
Cf. al-Dīnawarī, fasl 15, bāb 8, Esad Efendi 1833, fol. 150a:
If he saw the complete moon lighting in his part of the sky, the vizier of the king will benefit the people of that house.
Cf. also Ibn Shāhīn, no. 156:
If someone dreams that he is close to the moon, it signifies that he will obtain from the king or the vizier wealth and benefits.
Drexl 129, 22-25: ἐὰν ἴδη τις, ὅτι αὐτὸς τὸν δίσκον τῆς σελήνης ἦγεν εἰς τὸν οικον αύτου ή ὅτι ἐδέσποσεν αὐτοῦ, κυριεύσει τοῦ πλούτου καὶ τῆς ἐξουσίας του πρωτεύοντος διατασσόμενος αύτούς.
If someone dreams that he led the lunar disk in his house by himself, or that he exercised power over it, he will control and dispose of the wealth and power of the most important <courtier>.
Cf. Ibn Shahin, no. 140:
قال دانياال: يؤول إما بوزير الخليفة او بوزير الملك او بمن يقوم مقامهما فمن راى انه امسك القمر او جعله فى ملكه يدل على انه يكون وزير للملك او مقر نا عنده او خاصا من خواصه.
Daniel said: "< The moon> is interpreted either as the vizier of the caliph or as the vizier of the king or as the person who occupies that position. If someone dreamt that he seized the moon or had it in his possession, it signifies that he will become the vizier of the king or one of the people in his entourage or one of his favorites."
Drexl 129, 25-31: ἐὰν ἴδῃ ὁ βασιλεὺς τὴν σελήνην σκοτισθεῖσαν ἢ ἐκλείψασαν, είς έπιβουλὴν και έναντίωσιν αὐτοῦ ήξει ὁ πρωτεύων ἀπ΄ αὐτοῦ καὶ φωραθησεται έν τφ λαφ αύτου εί δε άλλος τις ίδη τουτο, είς πρόσωπον όμοίως του πρωτεύοντος αποβήσεται. εί δε ίδη, ότι πάλιν ἐκαθαρίσθη καὶ ήλθεν εἰς τὰ φωτα αυτῆς, πάλιν ἀνακαλεῖται καὶ εἰς τὴν αύτοῦ δόξαν ἐπανελεύσεται.
If the king dreams that the moon is darkened or eclipsed, the most important man after himself will oppose and plot against him, and this will become known among his people. If somebody else dreams this, the outcome of the dream will likewise regard the person of the most important man <after the king>. But if he dreams that the moon becomes clear again and that it regains its light, <this man> will be recalled and restored to his glory.
Cf. Ibn Shāhīn, no. 151:
وان راي القمر منخسفا يدل على رداءة حال ملك ذلك الزمان او حال وزير مثل عزل الللك عن مملكته او الوزير عن وزارته خصوصا إذا انخسف بتمامه.
If someone dreamt that the moon was eclipsed, it signifies a bad situation for the king of that time, or for the vizier, such as the removal of the king from his kingdom or the removal of the vizier from his office, especially if the moon was completely eclipsed.
#### The Stars
Artemidoros and the Oneirocriticon concur that the stars signify men in proportion to their size: the bigger the star the more important the man symbolized by it.155 Artemidoros gives this interpretation in a succinct form:156
<sup>(55</sup> Oberhelman, "Specific agreement," no. 46: "Oneirocriticon 166 = Artem. ii.36: the greatest of the stars = powerful men; the smaller stars = insignificant men."
<sup>156</sup> ii.36; Pack 165, 11-14.
ουτε δε καταπίπτοντες είς γῆν οί ἀστέρες είσιν ἀγαθοί (πολλών γὰρ ὅλεθρον μαντεύονται, καὶ ἀξιολόγων μὲν ἀνδρῶν οἱ μεγάλοι, λιτῶν δὲ καὶ ἀσήμων οἱ λεπτοί και άμαυροί) ....
It is inauspicious to see stars falling down upon the earth. For they prophesy the death of many men: large stars, the death of important men; small, dim stars, the death of simple, insignificant men ....
The corresponding interpretation in the Oneirocriticon is far more detailed (Drexl 127, 3-7):
Ο ήλιος είς πρόσωπον βασιλέως κρίνεται απαραλλάκτως και ή σελήνη είς πρόσωπον του άπο βασιλέως δευτέρου έξουσιαστού, ή Άφροδίτη είς πρόσωπον της αύγούστης, και οί λοιποί των μεγίστων αστέρων είς τους μεγιστάνους ανδρας του βασιλέως και οί αλλοι άστέρες είς πάντα τὸν κόσμον.
The sun is unfailingly interpreted as the person of a king, and the moon as the person of the man second in command after the king, Venus as the person of the queen, and the rest of the great stars as the noblemen of the king, and the remaining stars as the common people.
The Oneirocriticon names the seven planets one by one and matches them with specific members of the royal court (Drexl 129, 12-18):
Και ή σελήνη, ώς είρηται, είς πρόσωπον του δευτέρου έστιν από του βασιλέως και ή Αφροδίτη είς πρόσωπον της αύγούστης και δί Ερμῆς εἰς πρόσωπον του πρώτου των γραφέων του βασιλέως κρίνεται και ό "Αρης είς τον πρώτον πολεμιστὴν του βασιλέως και ὁ Ζεὺς είς τον πρώτον του πλούτου και τῆς διοικήσεως του βασιλέως και του χρυσίου και ό Κρόνος είς <τον> πρώτον ποιναλιστὴν και παιδευτήν κρίνεται.
The moon, as was mentioned, is the person who is second in authority after the king, and Venus is the queen. Mercury is interpreted as the chief among the scribes of the king, Mars as the chief warrior of the king, Jupiter as the one in charge of the wealth, administration, and gold of the king, and Saturn is interpreted as the chief punisher and chastiser.
It is immediately apparent that these courtly dignities are not Byzantine, except for the basileus and the augusta, that is, the emperor and the empress. Indeed, this passage translates in Greek as literally as possible the Arabic interpretation of the planets that is found in four out of the five Arabic dreambooks examined. Ibn Qutayba states: 157
والقمر في التكويل وزير الملك والزهرة امراته و عطارد كاتبه و بهرام
<sup>157</sup> Bãh 13, Yahuda ar. 196, fol. 28b.
#### CHAPTER FIVE
صاحب حربه والمشترى صاحب ماله وزحل صاحب عذابه وسائر النجوم العظام اشراف الناس.
The moon in dream interpretation is the vizier of the king; Venus is his wife; Mercury is his secretary; 58 Mars is his minister of war; Jupiter is his treasurer; Saturn is his chastiser. The other great stars are the nobles among the people.
A similar interpretation can be found in al-Dīnawarī (fașl 15, băb 18, Esad Efendi 1833, fols. 152b-153a):
النجوم الخمسة هى زمل والمشترى و المريخ والزهرة وعطارد قوم مشاهير وهم عدول اشراف وقالوا بل زحل صاحب عذاب اللك بيت مـاله والمريخ صاحب حربه ووالى جيــشـه وقـالوا بل هو الشرطى \* والزهرة وهي من المسرح امراة فمن راها خطب امراة جميلة مفتنة الناس لا تكون بينه وبينها قرابة وعطارد كاتبه...
The five planets are Saturn, Jupiter, Mars, Venus, and Mercury, a group of important people and just noblemen. <Dream interpreters> said rather that Saturn is the executioner in the service of the king, Jupiter is the one in control of his financial department and Mars his minister of war and leader of the army. < Other dream interpreters> said, however, that he is the policeman. Venus, because of the comb, is a woman, and whoever dreams of her will be betrothed to a beautiful woman who charms people and is unrelated <to the dreamer>. And Mercury is the secretary <of the king>.
The interpretation found in al-Muntakhab (p. 224) is even closer to the relevant وامبا: paragraph of the Oneirocriticon than the interpretation of al-Dīnawarī الخمسة سنارة فزحل صاحب عذاى اللك والمشترى صاحب مـال اللك والمرسم صاحب حرب الملك والزهرة امراة الملك وعطار دكاتب الملك (As for the five planets, Saturn is the king's chastiser, Jupiter is the king's minister of finance, Mars is the king's minister of war, Venus is the king's wife and Mercury is the king's secretary). Ibn Shāhīn's interpretation is identical, except صاحب خزانة that it calls the one in charge of the financial administration the 159 مو ال الملك (master of the king's treasury of wealth). 159
The dignitaries enumerated in the interpretation of the planets correspond, at least in part, to officials at the caliphal court. In addition to the vizier, two
<sup>158</sup> Lit.: "scribe" (kātih).
<sup>159</sup> Ibn Shāhīn, introductory paragraph to fas! 3, "On Dreaming of the Planets" (before no. 164). For a comparison of the interpretation of the planets in the Oneirocriticon and other Arabic dreambooks that come to the same conclusions, see Lamoreaux, "Dream Interpretation in the Early Medieval Near East," p. 320, table 11.
other officials-the *katib,* or royal secretary (too literally and therefore mistakenly rendered in Greek as *grapheus,* scribe), <sup>160</sup>and the *!fal:zib al-'adhab,* or chastiser <sup>161</sup> -are known to have existed in the Abbasid administration: But in Abbasid times the commander of the army was called *!fal:zib al-jaysh* (master of the army) and not *!fal:zib al-J:zarb* (master of war), and the head of the financial bureaucracy was called .~al:zib *al-kharaj* (master of taxes) and not *!fal:zib al-ma/* (master of wealth), though *bayt al-ma/* (house of wealth) was the term for the department of finance. Moreover, while the list includes the relatively low-ranking *!fal:zib al-' adhab,* it omits important offices of the Abbasid administration, such as the *J:zajib* (chamberlain), the *!fal:zib al-barid* (minister of the post) and the *!fal:zib al-shurta* (chief of police 162). It is impossible to know whether the source of the *Oneirocriticon* included a list of Abbasid offices and, if it did, whether the Greek author would be able to recognize it as such. But the overall interpretation of the planets seems to have been inspired mainly by Chaldaean and Graeco-Roman astrology, which had also been adopted by the Arabs. 163 Whether out of ignorance, lack of imagination, or deliberate choice, the author of the *Oneirocriticon* did not attempt to make the list of dignitaries conform to the titles then in use at the Byzantine court After all, it is contained in a chapter attributed to the Indians, Persians and Egyptians together, which seems to justify its nonconformity to the Byzantine court system.
Further interpretations about the stars given in the *Oneirocriticon* can also be found in Arabic dreambooks, as is evident in the examples given below:
<sup>160</sup>On the duties and education of the Abbasid secretaries, see Sourdel and Sourdel, *La civilization de /'Islam c/assique,* pp. 370-72.
<sup>161</sup>Lit.: "master of punishments." The only mention of the ,~ii~ih *a/-'adhiih* that I found was in the section of al-Dinawarl's dreambook that interprets dreams about the members of the royal entourage and various administrative officials. The most common term for the executioner is *jalliid,* which is also used in al-Dlnawarl. It is unclear what the difference between the two was. Dreaming of them is interpreted as follows (al-Dinawari, *Jae!* 9, *hiih* 8, *Esad Efendi* 1833, fol. 99a, I. 5): rt..:;\_\_.;. ~ \_, J~I J (The *jalliid* is <interpreted as> an abusive man) and *hiih* 14, ibid., ult. line: ,:i,J....o ~..> '-:-'IW I ~L.oJ (The eii~ih *a/-'adhiih* is interpreted as a harmful man); see also Fahd, "Les corps des metiers aux IV/Xe siecle a Baghdad," p. 206, n. 2.
<sup>162</sup>Possibly in al-Dinawarl's interpretation of Saturn one should read 4....b ~I ~L.o (chief of police) instead of~~ I (the policeman). For a description of the Abbasid administration with reference to the relative Arabic terms, see Ph. Hitti, *History of the Arahs,* 10th ed. (New York, 1970), pp. 317-31.
<sup>163</sup> See Bouche-Leclercq, *Histoire de la divination dans* /' *antiquite,* vol. I, p. 227.
#### CHAPTER FIVE
Drexl 131, 3-18: άλλα και το πλήθος των αλλων άστέρων των μεγίστων είς τοὺς εύγενεις και πλουσίους διακρίνονται αναλόγως και τους έγγιστα του βασιλέως κατοικούντας έπι παντος τόπου και πλείονα κρίσιν έχουσιν οι έγγιστα, έλάσσονα δε οί μακράν. ἐαν ἴδη τις τὰ μέγιστα τῶν ἄλλων ἄστρων πολύφωτα συνηχμένα, σύναξιν και χαράν παρά των είρημένων κοινά νοείτων όμοίως και έαν ίδη ταυτα όλιγόφωτα και διεσκορπισμένα ή σκοτεινά, το πάθος είς τον είρημένον λαόν και τους άρχοντας του βασιλέως διακρίνεται εί δε ΐδη, ότι μερικώς ή ἔπαθον ή ἐφωταγώγουν, μερικὴν κρινέτω καὶ τὴν χαραν καὶ τὴν λύπην.
Εαν Ίδη τις, ότι έδέσποσε τών άστέρων πάντων, δεσπόσει πάντων λαών και, ἐάν ἐστιν ἐπιτήδειος, βασιλεύσει. εἰ δὲ ἴδη, ὅτι ἐδέσποσεν αὐτῶν μερικώς, δεσπόσει λαου. εί δε ίδη, ότι ποιμαίνει και διευθετεί και είς κατάστασιν φέρει τα αστρα, έξουσίαν λήψεται λαου παρά του βασιλέως ἀνάλογον του καιρού και τῆς διευθετήσεως των αστέρων.
The multitude of the other big stars is similarly interpreted as the noblemen, the wealthy, and <generally> those who are close to the king in every land. The nearest <stars> have a stronger interpretation than the farthest ones. If someone dreams that the greatest among the rest of the stars were bright and were gathered together, he should reckon a gathering and joy shared by the aforementioned people. Likewise, if someone sees them dimly lit and scattered or darkened, <analogous> suffering will befall the aforementioned people and the magnates of the king. If he sees that the stars suffered partial damage or that they were partially lit, he should reckon the joy or sorrow as partial as well.
If someone dreams that he had power over all the stars, he will have power over all nations, and, if he is suitable, he will become king. If he sees that he only had partial control over them, he will rule a people. If he sees that he is shepherding or arranging the stars in an orderly position, he will receive from the greatest of kings sovereignty over a people analogous to the limpidity of the heavens (6) and the condition of the stars.
#### Cf. al-Dīnawarī, fasl 15, bāb 11, Esad Efendi 1833, fol. 151b:
من راي ساير النجوم في موضعها من السماء مجتمعة فهو يقوم حال الاشراف واجتماع امرهم وحسن شوؤنهم فان راها متفرقة فى السماء فهو يفرق اشراف الناس فى ذلك الموضع فان راى بها حدثا فهو حدث بهولاء \* فان راي انها صافية فهو صلاح احوالهم واذا راي نور ها مطموسا فهو تغير امورهم ومن راى انه ملك النجوم فانه يملك اشراف الناس كلهم بوفان راى النجوم مجتمعة فى داره ولها شعاع ونورها فانه يصيب فرحا وسرورا واجتمع عنده اشراف الناس على سرور فان لم يكن لها نور فهى مصيبة تحمع عند اشراف الناس
<sup>164</sup> o karpo = the weather at night in the dream.
If someone dreamt that the wandering stars gathered in their proper place in heaven, this is the condition of the noblemen, the coalition of their power, and the good order of their affairs. If he sees them scattered all over the sky, <the fulfillment of> this <dream> will divide the noblemen in this place. If he sees that something happened to the stars, something <analogous> will happen to the noblemen. If he dreamt that the stars were limpid, this is goodness in the condition of the noblemen. If he saw that their light was obliterated, this signifies change <for the worse> in their affairs. If someone dreamt that he had the stars in his power, he will have all the noblemen in his power. If he dreamt that the stars with their light and rays gathered in his house, he will receive happiness and joy, and the noblemen will happily congregate in his home. And if the stars did not have light, a calamity will befall the noblemen.
Drexl 131, 18-22: έαν ίδη τις, στι τρώγει αστρα, άναλώσει τους άνθρώπους αναλόγως της βρώσεως. εί δε ίδη, ότι κατέπιεν αυτά άμασήτως, ασυνήθεις ανδρες είσελεύσονται έν τώ οϊκφ αύτου και συναναμιχήσονται αὐτώ φιλικώς.
If someone dreams that he is eating stars, he will destroy people by analogy to what he ate. If he dreams that he swallowed them without chewing, unusual men will arrive at his house and associate with him in friendly terms. 165
Cf. al-Dīnawarī, fasl 15, bāb 14, Esad Efendi 1833, fol. 152b:
قال المسلمون من راي انه اكل الكواكي فإنه بستاكل الناس وناخذ اموالهم ويهلكهم ومن ‹۞۞› من غير اكل يداخله اشراف الناس فى امره و سره.
The Muslims said: "If someone dreamt that he ate the stars, he will destroy people, take their money and ruin them. And if someone <\* \* 166 > without chewing, noblemen will mix with him in his affairs and secrets.167
<sup>165</sup> Eating the stars is also discussed by Artemidoros (ii.36, Pack 165, 19-22), but the interpretations given by him are different from the ones recorded in the Oneirocriticon. According to Artemidoros, dreaming of eating the stars is a good portent only for clairvoyants and people whose profession requires observing the heavens. For the rest it predicts death.
<sup>166</sup> One or two words are illegible in the manuscript.
<sup>167</sup> The corresponding phrase in the Greek interpretation, ανδρες είσελεύσονται έν τφ οϊκφ αυτού και συναναμιτήσονται αύτω φιλικώς (Men will arrive at his house and associate with him on friendly terms), renders both form I and III of the verb dakhala, the basic meaning of which is "to enter." Form I also means "to pay someone a visit" ; in form III, which is found in al-Dinawari, the verb داخلو ه means "they mixed with him in his affairs," or "they mixed with him socially," or "they became intimate with him" (see Lane, An Arabic-English Lexicon, s.v. . 3 د خل.). Form I of the verb is common; form III is rarer, but apparently the author of the Oneirocriticon had داخل (form III) in front of him, a form with which he was probably unfamiliar, so he gave it the meaning of form I, which he did know. The meaning of form III written next to the meaning of form I might have been intended as a correction, but both the old, incorrect, and the new, correct rendering seem to have been retained in the final version.
A correction should be introduced to Drex1's text on the basis of the Arabic parallels. The Oneirocriticon contains the following interpretation (Drex) 129, 4-6): και το σκότος των άστέρων τώ ίδόντι ανθρώπω είσκόσμισιν χρυσίου και πλούτου αύτου σημαίνει (The darkness of the stars signifies for the dreamer the eiskosmisis of his money and wealth). The word eiskosmisis is a hapax in Greek, and its meaning is unclear. 168 One of the manuscripts consulted by the editor (V) has, instead of είσκόσμισιν, the reading είς κορπισμον, which should be read είς σκορπισμόν. The Greek solecism that results from the acceptance of the reading from manuscript V, onucívet eic ("it signifies" expressed with the verb and an unnecessary preposition), corresponds to the perfectly regular Arabic يدل على ("it signifies" expressed with the verb and the preposition with which it is usually construed). The substitution of the strange είσκόσμισις with σκορπισμός (squandering) results in the following interpretation: "The darkness of the stars signifies for the dreamer the squandering of his money and wealth." This would correspond to the interpretation found in the dreambook of al-Nābulusī: و مـن ر اى ان الكواكب قـــد ذهبـت ه من السماء فانه رن (If someone dreamt that the stars disappeared from the sky, his money will disappear).
The sun, the moon, and the stars in the Oneirocriticon are primarily interpreted in connection with the royal court and rulership. This is an important component of the corresponding chapters in the Arabic dreambooks. Every Arabic dreambook, however, beginning with Ibn Qutayba, also mentions that the sun, the moon, and the stars could signify the parents and siblings of the dreamer, according to the interpretation of Joseph's dream that is narrated in Qur'an 12:4-5 and the Old Testament (Genesis 37:5-11).10 It is therefore hard to imagine that the Arabic source of the Oneirocriticon did not mention such an interpretation. But the Oneirocriticon itself completely ignores it, as it does other episodes from the story of Joseph. One can only speculate about the
ίδα See Drexl, p. 249, index rerum et verborum, s.v. "είσκόσμισις." Conceivably, είσκόσμισιν could be corrected to the attested είσκόμισιν or είς κόμισιν (a bringing in, an import, an introduction). However, according to the critical apparatus in Drexl 129, 5, none of the Greek manuscripts that he consulted gave the reading είσκόμισιν or είς κόμισιν.
<sup>169</sup> Al-Nābulusī, vol. 2, p. 194, s.v. كوكب .
<sup>170</sup> Ibn Qutayba, hãh 13, Yahuda ar. 196, fol. 29a; al-Dinawari, fasl 15, bãb 24, Esad Efendi 1833, fol. 154a, for the sun as father; al-Muntakhab, introduction, p. 21; Ibn Shāhīn, no. 164; al-Nābulusī, vol. 2, p. 7, s.v. شمس.
reasons for this omission. It is possible that the author of the Oneirocriticon found the story of Joseph thoroughly appropriated by Islam and therefore avoided referring to it. But he might have chosen to single out the royal interpretation of the heavenly bodies because his dreambook was meant for use by royalty and members of the imperial court.
#### Eating Honey
As Oberhelman has pointed out, 171 both Artemidoros and the Oneirocriticon maintain that honey is interpreted as wisdom. Artemidoros asserts this in an anecdote-172
έδοξέ τις αρτον αποβάπτων είς μέλι έσθίειν. ἐπὶ λόγους φιλοσοφικούς όρμήσας και τὴν ἐν αὐτοῖς σοφίαν εὐπορίσατο καὶ περιεβάλετο χρήματα πολλά ἐσήμαινε γαρ το μέλι την εύέπειαν της σοφίας, ώς είκός, τον πορισμόν δε ό αρτος.
Someone dreamt that he dipped bread into honey and ate it. The man devoted himself to philosophical endeavors, and he acquired not only the wisdom that philosophy contains but great wealth as well. For the honey naturally signified the sweet words of wisdom and the bread signified financial gain.
The Oneirocriticon mentions that honey signifies wisdom in the course of a paragraph that comprises a number of additional interpretations (Drex) 196, 24):
εί δε έφαγε κηρίον μέλιτος από μελισσώνος, εύρήσει, α ουδείς ήλπιζε, και σοφός έσται διά το σοφόν του έργου της μελίσσης εί δε ίδη, ότι έκλασε τον μελισσώνα, έλευθερίαν δούλοις ποιήσει εί δέ έστι δούλος ό ίδών, ἐλευθερουται, εί δε πτωχός, πλουτήσει. ό βασιλεύς έαν ϊδη, ότι προσήνεγκαν αύτω κηρίον μέλιτος και έφαγεν έξ αυτου, είσκομιδάς γλυκείας και χαράν δέξεται ἐκ τοῦ λαοῦ αὐτου εἰς ἔργον γὰρ λαοῦ τὸ κηρίον ἐκρίθη. εἰ δὲ ἴδη, ὅτι κηρίον μέλιτος ήνεγκαν αυτώ και ήν κενόν μέλιτος, αί είσκομιδαί αὐτου είσιν ἐλλειπεῖς.
If he ate a honeycomb from a beehive, he will find what no one had hoped for, and will be wise, because of the wisdom of the bee's work. If he sees that he broke the beehive, he will free slaves. If the dreamer is a slave, he will be freed, and if he is poor, he will become rich. If the king dreams that he was brought a honeycomb and ate from it, he will receive sweet revenues and joy from his people. For the honeycomb is interpreted as the work of the people. If he dreams that he was brought a honeycomb empty of honey, his revenues are wanting.
<sup>171 &</sup>quot;Specific agreement," no. 78: "Oneirocriticon 241 = Artem. v.83: to eat honey means wisdom."
<sup>172</sup> v.83, Pack 322, 10-14.
Arabic dreambooks interpret honey much as the Oneirocriticon does. According to al-Dīnawarī, it signifies lawful money and wisdom: وقــيل الـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ مجموع حلال لقوله تعالى فيه شفاء للناس وقيل هو علم پرزق الانسان (It is said that honey signifies collected lawful money, according to the saying of God Almighty "wherein is healing for mankind" [Qur'an 16:69]. 173 It is also said that it <signifies> knowledge bestowed on people).174
Similar interpretations are repeated throughout the Arabic dreambooks. 155 The significance of honey as freedom for slaves is mentioned in a paragraph from Ibn Shahin (no. 4379) that contains interpretations attributed to al-Kirmānī:
قال الكرمانى من راى انه اكل عسلا او جمعه او يؤتى به الييه فيانه يصيب مالا وغنيمية وفرحا، وان كان عبدا عتق، وإن كان مريضا شفى، وربما دل العسل على كلام البر وطلب القرأن والعلم على وجه حسن ...
Al-Kirmani said: "If someone dreamt that he ate or collected honey, or that honey was brought to him, he will obtain wealth and booty and if he is a slave he will be freed, and if he is sick he will be healed. Possibly honey signifies words of righteousness, and zeal for the Qur'an and for knowledge in a seemly manner ... "
After the paragraph of the Oneirocriticon on dreaming of a honeycomb, Vat. gr. 573 (fol. 209b) adds further interpretations that are missing from the text of the critical edition. A usual philological assumption about works on the popular sciences that have a complicated textual tradition like the Oneirocriticon is that later scribes tampered with the original composition by adding passages which, most of the time, they found in other works on similar topics rather than inventing them themselves. The additional interpretations of Vat. gr. 573, however, are not later additions, but part of the original composition,
<sup>173</sup> Al-Dinawari's quotation that refers to honey comes from the verse 69 of the Quranic "Chapter of the Bee," which points to the bee and its activities as an example of duty, usefulness, and God's providing for the creation. The complete sentence is, "There cometh forth from their [= the bees'] bellies a drink of divers hues, wherein is healing for mankind."
<sup>174</sup> Fasl 23, hāb 39 on شهد (honeycomb), Esad Efendi 1833, fol. 241b.
<sup>175</sup> Honey also signifies knowledge of the Qur'an. See also al-Dinawari's chapter on the bee (fas! 21, hah 146) and on honey (fas! 23, hab 47). Al-Dinawari's interpretations are repeated in al-Muntakhab, pp. 129-30. Parts of al-Muntakhab are repeated in Ibn Shahin. Al-Nabulusi's chapters on honey (vol. 2, pp. 94-95, s.v. عسل) and honeycomb (ibid., pp. 19-20, s.v. شهد) are a combination of al-Dinawari and al-Muntakhah.
since analogous interpretations can be found in Arabic dreambooks. *Vat. gr.* 573 states: £av 'i3i:i n<; on EupEv µ£A.1 µ£'ta KTJpiou mhou, £1.>pi\crn avaA.oyov 'tOU µ£A.t'tO<; 1tAOU'tOV, nA.iiv ()£ µE'ta !)ia<; oA.iyT]<; (If someone dreams that he found honey in its honeycomb, he will find wealth commensurate with the honey, though through the use of some violence). Honey was already interpreted as wealth from booty by al-Kirmani, as is evident in the passage from Ibn Shahin quoted above. A comparable interpretation is also recorded by al-Dinawari: 176 ~ .J-.a J)l:.. JL.,, ct...:.µ *bJ....:>.J* ~I .:iµ (As for the honeycomb by itself, it is lawful wealth from booty).
In addition, *Vat. gr.* 573 (fol. 209b) contains a negative interpretation of desserts prepared with honey: d ()£ 'i3i:i on dJpEv fl ~pEv fl E<jlEpov a\J't0 µEAl E'Jlll'tOV ll'tOl YAUKl<Jµa µ£'ta appwµa't(J)V [sic], Ellpi\crEt 0At\jllV Kat <J'tEVOXWpiav, avaA.oyw<; 'tOU yA.uKicrµmo<; (If he dreamt that he found or took or was brought boiled honey, that is, a dessert with spices, he will find sorrow and trouble commensurate with the dessert). Other desserts are interpreted positively (Drexl 198, 2-3): El 8£ fo0t£t 1tAaKOUV'ta<;, Ellpi}crn aya0ov £meuµ11nKov (If he is eating flatcakes, he will receive something good and desirable). Likewise, lbn Shahin (nos. 4381-4382) quotes a negative interpretation of desserts prepared with honey, as opposed to desserts prepared with sugar: JSL: ct....:.1 (.SI.> .J-.aJ ... 4-..U\_;J ~ct...:.µ~ .l\_,h JSL: ct....:.1 (.SI.>~ .ill j .:i Jj ct...:.µ ~ .J-.a o I \_,h (If someone dreamt that he ate a dessert made with sugar, this is power and a lofty position ... but whoever dreamt that he ate a dessert made with honey, this is less than that). Once more, this additional interpretation from *Vat. gr.* 573 should be considered an integral part of the original composition, since it can also be found in the Arabic tradition.
The closeness of the *Oneirocriticon,* not so much to Artemidoros as to Arabic dreambooks, in this example further indicates that Arabic dreambooks furnished the immediate source of the Byzantine work.
#### *Elephant*
Artemidoros's interpretation of an elephant is the following: <sup>177</sup>
'EA.£<j>a~ i:'l;co µ£v 'haA.ia~ Kat' lv8ia~ 6pwµ£vo~ Kiv8uvov Kat <j>o~ov miµaiv£t 81a 'to XP<i>µa Kat 'to µ£y£8o~· <1>0~£pov yap 'to i;tj)ov, Kat µaA.tcrm 1:01~ µii i]Samv
<sup>176</sup>*Fa#* 23, *bah* 39, *Esad Efendi* 1833, fol. 241a.
<sup>177</sup> ii.12; Pack 123, 2-18.
#### CHAPTER FIVE
αυτού έν' Ιταλία δε δεσπότην σημαίνει και βασιλέα και ανδρα μέγιστον. Όθεν ἐπειδαν βαστάζη ἀφόβως πειθόμενος τω ἐποχουμένω, τὰς ἀπὸ <τῶν> τοιούτων [έργασίας τε καὶ] εὐεργεσίας μαντεύεται: ὅταν δὲ βλάπτη, τὰς ἀπὸ τοιούτων βλάβας, πολλάκις δε [έτήρησα] ἐλέφας διώκων καὶ ἀπειλῶν νόσον προαγορεύει. και καταλαβών μεν και διαχρησάμενος τώ ιδόντι θάνατον σημαίνει, μὴ καταλαβών δε είς έσχατον κίνδυνον έλάσαντα σωθήσεσθαι και γάρ φασι το ζώον ανακείσθαι τω Πλούτωνι. γυναικί δε ούδαμώς άγαθός έστιν όρώμενος ούτε προσιών ούτε βαστάζων. οίδα δέ τινα έν Ιταλία γυναικα πάνυ πλουσίαν και μὴ νοσοθσαν ή έδόκει έλέφαντι όχεισθαι, και ούκ είς μακραν απέθανεν.
Seeing an elephant in any place other than Italy or India signifies danger and fear because of its color and its great size. For the animal is fearsome, especially to those who are unacquainted with it. But seeing an elephant in Italy signifies a master, a king, or a very important person. Therefore, whenever an elephant carries his rider obediently and fearlessly, it signifies benefits from these people. But whenever it causes any trouble, it means trouble from these people. I have observed that if an elephant pursues or threatens, it is often a sign of sickness. And if it catches and kills its victim, it portends death for the dreamer. If the animal does not catch its prey, it signifies that the dreamer will be saved at the very brink of danger. For the elephant is reputed to be an animal sacred to Pluto. It is not at all propitious for a woman to see an elephant approaching or carrying her. I know of a woman in Italy who was quite wealthy and in the best of health who dreamt that she was riding an elephant. Shortly afterwards she died.
Artemidoros's interpretation of an elephant dreamt in Italy as signifying a powerful and important person is repeated in the Oneirocriticon (Drexl 220, 15-221, 22):178
σξη΄ Έκ των΄Ινδών περὶ ἐλέφαντος
Ο ἐλέφας εἰς μέγιστον καὶ πλούσιον ἄνδρα κρίνεται, οὐ μέντοι εἰς ἐχθρὸν μέγιστον, ἀλλὰ μέτριον. ἐὰν ἴδῃ ὁ βασιλεύς, ὅτι ἐλέφαντας ἐθήρευσεν ἢ εὐπόρησεν, πολύπλουτον και πολύχρυσον ανδρα θηρεύσει. και τούτο βασιλεύσι μόνοις ώσπερ άρμόδιον εί δέ τις του κοινού λαου ίδη τουτο, είς τα μέτρα του βασιλείου ύψους προσεγγιεί. έαν ϊδη τις, ότι ἐπωχήσατο ἐλέφαντι, ἐξουσιαστὴς χρηματίσει και χρημάτων και κτημάτων ἐξουσιάσει τινὸς μεγίστου. εἰ δὲ ἴδη, ότι απέθανεν ό ἐλέφας αὐτοῦ, θλῖψιν μὲν οῦτος δέξεται μετρίαν, ἐλαττώσει δέ ούδαμώς πλούτου μέτρον: τφ γαρ ἐλέφαντι πλούτος ή των όστέων χρεία ἐστίν. εί δε ίδη τις, ότι εύρεν έλεφάντειον κόπρον, εύρήσει πλούτον από μεγίστου άνδρὸς καὶ πολύπλουτος ἔσται· εἰ δὲ ὀστέα ἐλέφαντος, πλουτήσει πλουτον ἀπό μεγίστου ανδρός δια χρήματα άναλόγως του πλήθους των όστέων.
<sup>1/8</sup> Cf. Oberhelman, Oneirocriticon of Achmet, Appendix III, "specific agreement," no.101: " Oneirocriticon 268 = Artem. ii.12: elephant = a powerful ruler."
#### 268. From the Indians on Elephants
The elephant is interpreted as a very powerful and wealthy man, though not as a superior but as a moderate enemy. If the emperor dreams that he captured or acquired elephants, he will capture a man who owns enormous wealth and gold. This dream befits only emperors. If a commoner dreams this, he will approach the highness of imperial majesty. If someone dreams that he was riding on an elephant, he will be appointed ruler and will control the money and possessions of somebody powerful. If he dreams that his elephant died, he will experience moderate sorrow. but the extent of his wealth will not in any way be diminished because, for an elephant, the use of its bones is wealth. If someone dreamt that he found dung from an elephant, he will find wealth from a powerful man and will become extremely rich. If he found the bones of an elephant, he will acquire wealth from a rich man through money proportionate to the quantity of the bones.
#### σξθ΄ Έκ των Περσών και Αίγυπτίων περὶ ἐλέφαντος
΄Ο ἐλέφας εἰς ὑψηλὸν ἄνδρα ξένον ἐξουσιαστὴν πολύπλουτον κρίνεται διὰ τὸ μὴ πανταχού θηρασθαι πλὴν ἐν Ίνδία. ἐαν ϊδη ὁ βασιλεύς, ὅτι ἐλέφαντας ἤγαγον αυτώ δεδεμένους ή ένα, ανδρας μεγίστους πολυπλούτους καὶ όλιγοπολέμους έξουσιαστάς δεσμίους δέξεται εί δε ανευ δεσμών σιδηρών ήτοι κουρκούμων, οί είρημένοι μέγιστοι ύποταγέντες έλεύσονται ἐν αὐτῶ. εἰ δὲ ἴδη ὁ βασιλεύς, δτι έκέλευσε τυφθηναι έλέφαντα των όστέων αύτου ένεκα, τοιουτον ανδρα άποκτενεί και αρη τον πλουτον αύτού και θλιβήσεται έν αύτω. εί δε ίδη ό βασιλεύς, ότι επωχήσατο έλέφαντι, και τουτο είς ύποταγήν μεγίστων ἐκρίθη εί δέ τις άλλος ίδη, στι έπωχήσατο έλέφαντι, είς ύψος ήξει και εύρήσει χαραν και πλουτήσει σφόδρα.
Εί δε ΐδη τις, ότι διακονεί έλέφαντι έν τε βρώσει και πόσει, ανδρι μεγίστω δουλεύσει, έξ ου εύρήσει κέρδος. έάν τις ΐδη, ὅτι ἐλέφας ἐξαπίνης εἰσῆλθεν ἐν τῶ οῖκφ αὐτοῦ ξένος, εὑρήσει χαρὰν καὶ θλίψεως πάσης λυτρουται. εἰ δὲ ϊδη τις, ότι εύρε των όστέων αύτου ή του κόπρου, τουτο είς πλουτον διακρίνεται.
#### 239. From the Persians and Egyptians on the Elephant
The elephant is interpreted as a lofty and extremely wealthy foreign ruler, because it is not hunted anywhere else but in India. If the king dreams that either one or several elephants were brought to him in bonds, he will receive in bonds extremely powerful and wealthy men and rulers inexperienced in war. If <the elephants were> without iron bonds, that is, muzzles, the aforementioned powerful men will come to him in submission. If the king dreams that he ordered the slaying of an elephant because of its bones, he will slay such a man and seize his wealth and grieve over it. If the king dreams that he rode an elephant, this is also interpreted as the subjugation of very powerful men; and if anyone else dreams that he rode an elephant, he will become exalted, find joy, and grow exceedingly rich.
If someone dreams that he is attending an elephant, serving him fodder and drink, he will serve a very powerful man from whom he will profit. If someone dreams that an unknown elephant suddenly came into his house, he will find joy
and be relieved from all sorrow. If someone dreams that he found elephant bones or dung, this is interpreted as wealth.
The interpretation of an elephant as a powerful foreigner appeared in every single Arabic dreambook examined. It is inspired by a Quranic chapter (sura 105, also known as "The Chapter of the Elephant") that refers to a well-known historical event of ca. 570: Abrahah, the Christian viceroy of the Abyssinian Negus and ruler of the Yemen, campaigned against Mecca in order to destroy the Ka ba and was turned back through divine intervention. Abrahah's army included at least one elephant, which seems to have greatly impressed the north Arabians. The Qur'an refers to the foreign invaders as "the people of the elephant"; the year of the invasion was subsequently called "the year of the elephant" and is generally considered also to be the year of the Prophet's birth. The interpretation of the elephant in the Oneirocriticon as a powerful foreigner who can be defeated in battle seems to reflect the Islamic perception of the animal. For example, Ibn Qutayba writes the following:179
من راي انه ركب فيلا يملكه وعليه آلة الفيل اصاب سلطانا عظيما اعجميا او قهر سلطانا او اعجميا ... فان راي انه اکل لحم فيل اصاب مالا من سلطان وكذلك كل ما نال من اعضايه او جلده او عظامه او عصبه او شعره # فان رائ انه رکبه فی ارض حرب علی ائی هیئیة کان کانت الدبرة علی اصحاب تلك الارض لقصة اصحاب الفيل... ومن قتل فيلا قهر رجلا ضخما اعجميا.
If someone dreams that he rode an elephant and was able to dominate it and that the elephant was in full riding gear, <the dreamer> will gain a mighty foreign dominion, or will vanquish a sovereign or a foreigner .... And if he dreams that he ate the meat of an elephant, he will obtain money from a sovereign; likewise for whatever part <of the elephant> the dreamer received, whether for its skin or bone or sinew or hair. And if he dreams that he rode <an elephant> in a land at war, the people of the land on whose side the elephant was will be routed, according to what happened in the story of the people of the elephant .... Whoever slays an elephant will triumph over a great foreign man. 180
<sup>179</sup> Bãh 39, Yahuda ar. 196, fols, 54b ff.
<sup>180</sup> A part of Ibn Qutayba's chapter on the elephant is repeated, slightly rephrased, by Ibn Shahin (nos. 5868-69), where it is attributed to the author of the very first Arabic dreambook, al-Kirmanî. It seems that this ancient treatise is indeed the source of the following interpretations: قال الكرمانى من راي انه راكب على فيل ببلس وهو مطيع له فإنه يدل على متابعته ملكا اعجميا، ورؤيا جلد الفيل ولحمه وعظمه وشعره يؤول بحصول مال ومنفعة ونعمة من السلطان. ومن رائ اي انه راكب فيلا فى الصرب فانه يدل على قهر عدو ضخم، وقيل يقهر العدو و هذا القول دليل على على قصة امتحاب القول دليل على قصة اصحاب الفيل
Ibn Qutayba not only gives the interpretation of an elephant that Artemidoros and the *Oneirocriticon* agree on, but also adds two interpretations that appear in the Byzantine text: riding an elephant foretells rulership (while in Artemidoros it only foretells deriving benefits from the powerful person signified by the elephant) and acquiring the bones of an elephant indicates the acquisition of wealth. Al-Dinawari interprets the elephant similarly: "It is said that an elephant is a great king" Cr ; . ;.., dL <L..,j I ~). He also adds a couple of interpretations that correspond to those found in the *Oneirocriticon:* "If someone dreams that he rode an elephant, he will be appointed minister or local governor, and if he takes the elephant's dung he will obtain money <that belongs to the aforementioned officials>" ( cC;J.J ..:,....a .:i...:..1 ,:,l . .Q ~ *'ll* JJ ~) .JJ JL.:. ~ \_; ,:,,L.J .UL..... .J-.c JL..:.). <sup>181</sup>In addition, Ibn Shahin (no. 5865) states that slaying an elephant means slaying a king, which reminds one of an analogous interpretation in chapter 239 of the *Oneirocriticon:* ~ JJ.J..:! o.....:.µ ~ ~ o.....:. I (.51 \_; .J.-.aJ G.b.........IJ-:' JI~~~ dL (If someone dreams that he slew an elephant, this is interpreted as slaying a king with his own hands or through his own intervention).
The following conclusions can be drawn from comparing the texts of Artemidoros, the *Oneirocriticon,* and the Arabic dreambooks: (1) The author of the
riding an elephant with large sacks of straw and that the elephant obeyed him, this signifies that he will follow a foreign king. Dreaming of the elephant's skin, meat, bones and hair is interpreted as acquisition of money, profits, and benefits from the sovereign. If someone dreams that he was riding an elephant to war, it signifies that he will vanquish a great enemy, and it is said that the enemy will be vanquished, and this saying alludes to the story of the people of the elephant" ).
<sup>181</sup>*Fa,I* 21, *hiih* 104, *Esad Efendi* 1833, fol. 224a-b. The second excerpt from al-Dinawarl's chapter is attributed to the Jews. However, the interpretations of the elephant found in the Babylonian Talmud *(Beriikot* IX, 1-2, fol. 54b; trans. Cohen, pp. 371-72) are different from the Jewish interpretations of al-Dinawari. Al-Dinawari also quotes a few interpretations of the elephant attributed to the Christians that do not essentially differ from those of the Muslims and Jews (dreaming of an elephant and not riding him indicates the decline of one's status and financial ruin; but riding one means obtaining sovereignty ( ~ ). A slain elephant means that the king or some other important man will die). *Al-Muntakhah.* pp. 189-90, for the most part repeats the interpretations of al-Dinawari, with a very few additions at the end that do not correspond to anything found in the *Oneirocriticon.* Al-Nabulusi's entry (vol. 2, pp. 130-32) is a combination of al-Dinawari and *a/-Muntakhah* with some additions that closely repeat the interpretations of Ibn Qutayba (e.g., JI d...'..~I.:.,..... :...;..1 .:\_,I ..:JJ~J .:\_,LhL..u.:.,..... ')'L.. '-:'Loi J.:.--0 ~ JSI .....:.I <.SI.; .:\_,I <L...al.!o....c. JI •.i.4- (If someone dreams that he ate meat from an elephant he will obtain money from a sovereign *[su/fiin]* and likewise ifhe took something from its limbs or its skin or its bones).
#### 236 CHAPIBR FIVE
*Oneirocriticon* did not use the Greek text of Artemidoros as a source for his own composition. (2) The Arabic sources used by the author of the *Oneirocriticon* did contain interpretations similar to those found in Artemidoros. However, contrary to the practice followed by al-Dinawari, they did not, in all likelihood, quote I:Iunayn's text verbatim. (3) The author of the *Oneirocriticon* did not always slavishly translate the Arabic text in front of him, but adapted concepts when necessary. At the same time, he did not completely depart from his sources in order to create interpretations *ex nihilo,* but tended to work on the basis of existing material, as is evident in his interpretation of crucifixion, the cross, kings, and temples. In other words, he seems to have adapted existing interpretations, and not to have invented new ones. It is also evident that he made an effort to adapt in Greek some of the Arabic interpretations based on etymology. Sometimes a coincidence helped him (e.g., *ra's* - *riyasa* - *ra's ma/ I kephale* - *kephalaion,* resulting in interpreting heads as leadership and capital). At other times he had to invent a Greek expression *(rakaba farasan -irtakaba ma'#yatan I* E1tE~atVE tn1t(Q crKA:r1poxaA.ivq>-(Jl(A;r1pa~ aµapi"ia~ E:m~a'tT}~ yEvi}crEtm =he was riding on a hardmounted horse-will embark on a course of obstinate sinning). But there are also instances where the etymological game played in the interpretation is lost in Greek *(ghannam* [sheep] and *ghanima* [booty] leading to the interpretation of a sheep's head as booty).
All this can be said with some confidence, since the extant Arabic sources furnish a relatively secure ground for comparison. Other questions, however, such as the Greek author's treatment of the Old Testament and the Quranic story of Joseph cannot be answered, at least not on the basis of the material that has come to light so far, since we do not know exactly what the contents of his Arabic models were. His omission of the initial chapter on God can be hypothesized with some certainty, given the structure and contents of Arabic dreambooks; this omission indicates that he did leave out Arabic passages that he did not want-or did not know how-to Christianize.
#### CHAPTER SIX
#### SYRIAC AND CHRISTIAN ARABIC DREAM INTERPRETATIONS
The Oneirocriticon is not a direct translation but a Christian adaptation of Islamic material. The analysis of certain pecularities in the grammar and syntax of the Greek text demonstrated that its source was written in Arabic. What remains to be addressed is whether the Oneirocriticon was translated into Greek from a Christian (as opposed to Muslim) Arabic source and whether the Byzantine author copied his interpretations from an Arabic version already adapted for Christians or made these adaptations himself at the same time that he was translating an Islamic text. Both Syriac and Christian Arabic dream interpretation must be taken into account, since both Syriac and Arabic were spoken by the Christian communities living in Arab Muslim lands at the time that the Oneirocriticon was composed, and translations from Syriac into Arabic were all the more frequent as Arabic gained ground as the language of communication among Christians. This leads to the following questions: Are there any examples of Christianized dreambooks in Syriac or Arabic that were based on Islamic sources? Was there a tradition of dream interpretation in either Arabic or Syriac that could be labeled as Christian and that was distinctively different from the Muslim tradition of dream interpretation? In other words, if the Greek author of the Oneirocriticon was interested in transmitting Oriental Christian dream interpretations, would he have found any on which to base his work? Moreover, if a Christian tradition did exist, are any of its elements reflected in anything found in the Oneirocriticon?
Little is known about dream interpretation in Syriac. In the same vein as the Oneirocriticon is a brief work on dream interpretation contained in a Syriac manuscript in the British Library, Or. 4434, fols. 78a-93b. Its editor had no doubt that it was translated from the Arabic, since the Syriac word for "chapter" used in the text is the equivalent of the Arabic bab and the style of the work is also "fortement arabisant."2 The date of the text is unclear; the
<sup>1</sup> Published by G. Furlani, "Une clef des songes en syriaque," Revue de l'Orient chrétien 22 (1920-21), 118-44 and 225-48, Syriac text and French translation; Furlani notes that "jusqu' à présent on ne connaissait pas d'overporpriticov en langue syriaque" (p. 118).
<sup>2</sup> Ibid., p. 118.
only *terminus ante quern* that can be established is the date of the manuscript itself, and it was copied in the nineteenth century. <sup>3</sup>
The Syriac dreambook in BL *Or.* 4434 starts with a chapter on the theory of dream interpretation. Seven categories of dreams are identified, according to their causes: dreams in the first four categories are caused by the overflow of one of the four bodily humors;4 dreams in the fifth category are caused by the dreamer's imagination being incited by demons; those of the sixth category are caused by the dreamer's preoccupations during the daytime; dreams in the seventh category, including revelations and visions, come from God according to Mar Ephraim.5 The dream of Joseph, husband of the Virgin Mary , 6 can persuade Christians of the divine provenance of dreams in this category, while the dreams interpreted by Joseph,7 son of Jacob, can persuade the Jews. When waking up from a God-sent dream, the dreamer is calm and remembers the whole dream. A truthful dream is dreamt while sleeping on the left side. The time required for a dream to be fulfilled is commensurate with the time of night at which it was dreamt. If it was dreamt during the night and before the cock crows, it will be fulfilled within twenty years. If it was dreamt after the cock crows but before the sun rises, it will be fulfilled within thirty years. Finally, if dreamt at sunrise, it will be fulfilled in the next one or two days. A dream should be narrated to a wise person or a friend, but never to a woman. Advice is offered on how to diffuse the potential harm of a bad dream. Finally, the dreamer is warned that in the month of February and during the time when leaves grow, dreams are false. 8
These principles of dream interpretation are different from those put forth in chapters 2 and 30 I of the *Oneirocriticon,* which are dedicated to a similar
<sup>3</sup> Description and date of the manuscript in G. Margoliouth, *Descriptive List of Syriac and Karshuni Manuscripts in the British Museum Acquired since 1873* (London, 1899), p. 42; G. Furlani, "Ancora un trattato palmomantico in lingua siriaca," *Rendiconti de/la Reale Accademia dei Lincei. Classe di scienze morali, storiche e filologiche* 27 (1918), p. 316.
<sup>4</sup> The four bodily humors in the Galenic system are blood, phlegm, black bile and yellow bile. Something is missing from the text because only three humors are mentioned: red bile, phlegm and black bile (according to Furlani' s translation, p. 134, *fie/ rouge, matiere pituiteuse, fie/ noir* ). All references to the Syriac dream book will be given according to Furlani 's French translation which accompanies the Syriac text.
<sup>5</sup> Ephraim (306-73) wrote in Syriac, but his works were also translated into Armenian, Greek, Latin, and Old Church Slavonic. For a discussion of Ephraim's ideas about dreams, see E. Beck, *Ephriims des Syrers Psycho/ogie und Erkenntnislehre* (Louvain, 1980), esp. pp. 46 ff., 49 ff., 80 ff.
<sup>6</sup> Matthew 1:18-25; 2:13-14; 2:22.
<sup>7</sup> Genesis 40: 1-42:36.
<sup>8</sup> Compare these with the theoretical principles of Christian dream interpretation quoted in the dreambook of al-Dlnawarl, given in Appendix 3.
discussion. The *Oneirocriticon* supposes that all dreams come from God, and the dream of Joseph, the husband of Mary, and the dreams interpreted by Daniel are given as examples (Drexl 2, 1-24). The time required for the fulfillment of a dream depends on the time of night that it was dreamt, but the time spans given are different from those in the Syriac dreambook (Drexl 241, l -14 ). Finally, the condition of the leaves on the trees is related to the good or bad outcome of a dream, but in a way almost opposite to that in the Syriac dreambook (Drexl 240, 21-29). In short, the theoretical exposition of dream interpretation in the *Oneirocriticon* is very different from that in the Syriac dreambook, even when both works discuss the same subject.
The theoretical principles explained in the Syriac dreambook are, however, closely related to those expounded in a number of Islamic Arabic dreambooks that are, to some extent, adapted from Artemidoros,9 Aristotle,10 and Hippocrates,11 with the addition of Arabian and Islamic precepts. The introduction of the Syriac dreambook is followed by interpretations organized in thirty-one chapters according to subject. 12 The arrangement of the book is again somewhat different from that in the *Oneirocriticon,* in that the planets, natural phenomena that come from the sky (rain, snow, hail, etc.), and the elevations of the earth (mountains, hills, etc.), are placed immediately after the dreams of God, Heaven, the throne of God, 13 and angels, and before the chapter on the Holy Writ, churches, and the sacraments. In spite of the differences with the *Oneirocriticon,* this organization is consistent with the arrangement of chapters in some Arabic dreambooks, such as that of Ibn Shahin. A number of interpretations offered in the Syriac dreambook agree with those that appear in the *Oneirocriticon.* In a short article, Drexl enumerated similar passages in the two texts, noting about ninety interpretations that they had in common, but did not draw any conclusions and avoided saying whether or not he thought that the Greek
<sup>9</sup> i.Proem., Pack 3, 9-19, 4.
<sup>10</sup> Treatises ITi::pl. uitvou Kat £yp11yopai::coc;, ITi::pl. £vuitvicov, ITi::pt Tiic; Ka0' uitvov µavttKT\c;; published in Aristotle, *Parva Naturalia.* Loeb Classical Library (Cambridge, Mass., 1957), pp. 318-85.
<sup>11</sup> Treatise ITi::pt 8tai 111c; ~ *i\* 7t£pt EV1J7tVlCOV; published in Hippocrate, *Du regime'* ed. R. Joly (Paris, 1967), pp. 97-109.
<sup>12</sup>Furlani, "Une clef des songes en syriaque," pp. 135-36.
<sup>13</sup> This is the ..;.~. the Throne of God and the highest sphere of Heaven, according to the Muslim tradition; it appears in almost all the Arabic dreambooks. However, the problem of the relationship between the Syriac dreambook and Islamic dream interpretation will not concern us here.
*Oneirocriticon* relied on Arabic sources by way of a Syriac intermediary .14
BL *Or.* 4434 includes a second dreambook, written in modem Syriac (more commonly known as Neo-Syriac). The publication of this Neo-Syriac dreambook has never materialized, though announced by Furlani in 1921, but according to him, it is similar to the published Syriac text in several respects. <sup>15</sup> It is possible that the contents of the manuscript were translations or abridgments directly or indirectly based on medieval Arabic material; <sup>16</sup>the derivation of both Syriac dreambooks from the same kind of source would explain the similarities between them, as well as the similarities between the *Oneirocriticon* and the published Syriac text.
In addition to the two Syriac dreambooks from BL *Or.* 4434, we know of one further dreambook written by a Christian, though this time not in Syriac but in Arabic. It is called *Ta'bir al-ru'yii 'ala ikhti\$iir* (Dream Interpretation in Abridgment), and constitutes chapter 49 of the *Kitiib al-Dalii'il* (Book of Signs), 17 a miscellany of practical information on various subjects, 18 which was written in Arabic by Ibn al-Bahh11, a tenth-century Nestorian Christian. <sup>19</sup> It has recently been demonstrated that the chapter on dream interpretation in the *Book of Signs* is an abridgment of the Muslim Arabic dreambook by Ibn Qutayba.20 The manner in which this abridgment was made suggests that the
17 Fahd, *La divination arabe,* p. 333, no. 14 and pp. 225-26, and nn. 5 and 6. Facsimile edition in I:Iasan ibn al-Bahliil, *The Book of Indications (Kittib a/-dalti'il) by al-lfasan ibn a/-Bah/Ul (Tenth Century A.o.),* ed. F. Sezgin (Frankfurt am Main, 1985); critical edition in I:Iasan ibn al-Bahliil, *Kittib a/-dalti'il lil-lfasan b. a/-Bahlul,* ed. J. I:Iabbi (Kuwait, 1987). An additional manuscript of the text has recently been discovered.; cf. J. Lamoreaux, "New Light on the Textual Tradition of Bar Bahliil's *Book of Signs," Le Museon* 112 (1999), pp. 227-30.
<sup>18</sup>It contains a calendar of feasts celebrated by Eastern Christians and other religious communities, as well as information on the prediction of the weather, medical advice, instructions on various forms of divination, etc. See J.M. Fiey, "Sur le calendrier syriaque oriental arabe de Bar Bahliil (942/968 A.D.)," *Analecta Bollandiana* 106 (1988), pp. 259-71.
<sup>19</sup>For biographical information in lbn al-Bahliil, see J. Habbi, "Le Livre des Signes de al-I:Iasan b. Bahliil," *Oriens Christianus* 68 (1984), pp. 210-12.
<sup>20</sup>See Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 264-80 and esp. 272 ff. Cf. also idem, "The Sources oflbn Bahliil's Chapter on Dream Divination," *Studia Patristica* 33 (1997), pp. 553-57.
<sup>14</sup> F. Drexl, "Achmet und das syrische Traumbuch des cod. Syr. Or. 4434 des Brit. Mus.," *BZ* 30 (1929-30), pp. 110-13.
<sup>15</sup> Furlani, "Une clef des songes en syriaque," p. 118.
<sup>16</sup>Dream interpretation is a practice still very much alive in the Middle East, and it is easy to buy a dreambook at the bazaars. These printed dreambooks are essentialy versions of medieval Arabic works, and one of the most popular is the *Muntakhab* used in the present study for comparison with the *Oneirocriticon.* On this topic, see also the remarks by J. Lamoreaux, "Some Notes on the Dream Manual of al-Dari," p. 47.
goal was not to Christianize the Muslim manual, but rather to make it appealing to any adherent of a monotheistic religion, whether Judaism, Christianity or Islam. 21 True, both Christian adaptations oflslamic dreambooks that are extant are rather concise, and do not allow space for much detail or comment on the dream symbols interpreted. But neither of the two Christianized versions particularly emphasizes its Christian character, not even the Syriac dreambook which, by virtue of its language, is exclusively addressed to Christians and cannot have aspired to as religiously diverse a readership as lbn al-Bahlill's *Book of Signs,* which was written in Arabic. Compared with the published Syriac dreambook of BL *Or.* 4434 and the abridgment of Ibn Qutayba by lbn al-Bahlill, Muslim dreambooks have a much more pronounced religious character, equal in the frequency and importance of its references to religion to those found in the *Oneirocriticon.*
The evidence, scarce as it is,22 suggests that there were works on dreams written by Christians, both in Syriac and in Arabic, independent of the Muslim tradition of dream interpretation. First and foremost among them are the expositions on the nature of sleep and dreams found as chapters in longer treatises on religious and philosophical subjects, such as the chapters on dreams found in Bar Hebraeus's23 Syriac works, *Candelabrum of the Sanctuary, <sup>24</sup> Book of the Speech of Wisdom, <sup>25</sup>*and *Ethikon.26* Brief discussions on sleep and dreams were part of the literary genre that Bar Hebraeus cultivated with the *Candelabrum* and the *Ethicon;* they are incorporated in works on asceticism, such as the *Scala Paradisi,* as well as in philosophical-theological works, like
<sup>21</sup>See Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 276-77.
<sup>22</sup> The only reference to dream interpretation in Graf' s exhaustive enumeration of Christian Arabic writings occurs in a spurious patristic work, written in the form of questions and answers, purportedly by Gregory of Nazianzus and Basil the Great, where dream interpretation is condemned; see G. Graf, *Geschichte der christlichen arabischen Literatur,* vol. 1 (Vatican City, 1944), p. 326.
<sup>23</sup> Gregory Abii'l-Faraj, also known as Bar Hebraeus (1225-86), was a Syriac scholar and Maphrian of the East (primate of the Monophysite community in the former Persian territories).
<sup>24</sup> On the *Candelabrum of the Sancutuary,* see W. Wright, *A Short History of Syriac Literature* (Amsterdam, 1966), which reprints Wright's article on Syriac literature from the 10th edition of the *Encyclopaedia Britannica,* vol. 22 (1887), pp. 274-75. For the passage on dreams from the *Candelabrum of the Sanctuary,* see Bar Hebraeus, *Psychologie de Gregoire Aboulfaradj, dit Barhebraeus,* ed. and trans. J. Bakos (Leiden, 1948), pp. 68-71.
<sup>25</sup> See Bar Hebraeus, *L' entretien de la sagesse. Introduction aux oeuvres philosophiques de Bar Hebraeus,* trans. H. Janssens (Liege and Paris, 1937), pp. 321-22.
<sup>26</sup> On the *Ethicon,* see Wright, *A Short History of Syriac Literature,* p. 277. For the part on sleep and dreams, see Bar Hebraeus, *Ethicon, Memrii I,* trans. H. G. B. Teule (Louvain, 1993), pp. 38-45.
Nemesius's *Peri physeos anthropou* (On the Nature ofMan).27 Such discussions probably existed in works by other Syrian Christian writers as well. But, insofar as they dealt with dream interpretation, they only addressed theoretical problems and did not, except for well-known biblical examples, offer any concrete interpretations. The theoretical discussion of dreams in the *Oneirocriticon,* which assumes that all dreams are sent from God to all humans, even the greatest sinners, differs from the discussions of these Christian writers.28
Another type of writing on dreams very different from the *Oneirocriticon* is exemplified by an unpublished Syriac work attributed to Jacob of Edessa (ca. 633-708),29 who discussed the veracity of dreams according to the day of the week on which they were dreamt. 30 Similar texts are known in both Greek and Arabic. <sup>31</sup>
One source of information on Christians who wrote about dream interpretation was a list of dream interpreters *(Tabaqat al-mu<abbirin)* compiled by al-f.Iasan b. al-f.Iusayn al-Khallal. This work is now lost, but an abridged version of it is included as the fifteenth (and last) *maqala* to the introduction
29 A critical edition and study of this work (contained in *Mardin* 418) by Abdul Massih Saadi is forthcoming. I would like to thank Mr. Saadi for drawing my attention to Jacob's treatise and providing me with the reference.
<sup>27</sup> Chap. 12, *PG* 40. For a brief discussion on the ideas about dreams expressed in ancient medicine, particularly their tripartite division by Herophilus of Chalcedon (4th/3rd century s.c.) and its influence on later Christian and patristic writers, see H. von Staden, *Herophilus: The Art of Medicine in Early Alexandria* (Cambridge, New York, New Rochelle, Melbourne, Sydney, 1989), pp. 306-10.
<sup>28</sup> Synesios of Cyrene, the pagan gentleman who became a Christian bishop, is the only writer identified as Christian that I know of who argued that all dreams are truthful and God sent in 11Ept i:vuitvi.wv; see Synesius of Cyrene, *Hymni et Opuscu/a,* ed. N. Terzaghi, vol. 2 (Rome, 1944), pp. 143-89.
<sup>30</sup> Given that Jacob of Edessa is a pivotal figure in the intellectual history of Syriac Christianity, it is conceivable that the attribution of this dreambook to him was fabricated in order to legitimize a divinatory method that the church tended to view with suspicion, much like the attribution of Greek dreambooks to one of great fathers of the church, such as Athanasios, Patriarch of Alexandria and Gregory of Nazianzus. Such ascriptions can be found in *Marc. gr.* 608 and *Ambros. gr.* 592 (0 94 sup); see Dettorakes, "Ta byzantina oneirokritika," p. 69.
<sup>31</sup>*Vat. gr.* 573, fol. 213b, followed by a lunar dreambook *(selenodromion).* For an Arabic expose correlating the day of the week that a dream is dreamt to its veracity, cf. the introduction to the dreambook of al-Dinawari, *maqala* 14, *Esad Efendi* 1833, fol. 32a. If the attribution of this work to Jacob is genuine, his good knowledge of Greek, attested by his education and scholarly interests (he studied Greek in Alexandria, wrote a commentary to the Greek Scriptures and revitalized the teaching of Greek at the Syriac monastery of Eusebona) suggests that in this work he might be repeating information originally found in Greek sources.
of al-Dinawari's dreambook completed in Baghdad in 1006.32 In al-Dinawari's abridgment, the one hundred dream interpreters are organized under fifteen categories: prophets, companions of the Prophet Mul).ammad, the generation following the Prophet Mul;lammad, jurists, ascetics, authors of treatises on dream interpretation, philosophers, physicians, Jews, Christians, Zoroastrians, pagan Arabs, diviners, magicians and physiognomists.33 Only seven authors of treatises on dream interpretation are mentioned, and none of them appears to have been a Christian.34
Further on, however, al-Dinawari includes the names of three Christian (Na~arii) dream interpreters: f.Iunayn b. Isl;liiq, the translator; Abu Makhlid; and Zayn al-Tabari. 35 f.Iunayn b. lsl;liiq seems to have been listed by virtue of having translated Arternidoros, though he is not known to have written anything about dreams himself. Abu Makhlid is probably Abu Makhlid b. Bukhtishu' (d. 1025), who belonged to a family of famous Nestorian physicians active in Baghdad during the ninth and tenth centuries.36 The biographical dictionary
34 The seven authors on dream interpretation are (quoting Fahd, *La divination arabe,* p. 342): MllQammad b. Sirin, Ibrahim b. 'Abdallah al- Kirmani, 'Abdallah b. Muslim al-Qutaybi, A1:xi Al}mad Khalafb. ~ad, Mul)ammad b. I:Iammad al-Rlizi al-Khabbaz (=the baker), al-I:Iasan b. al-I:Iusayn al- Khallal, Aqamidiirus al- Yiinani. Five of these authors are known from other sources as well, and are definitely not Christian. Besides lbn Sirin and al-Kirmani, this list also includes Ibn Qutayba (as "al-Qutaybi"). Both BN *arabe* 2745, fol. 45a, and *Esad Efendi* 1833, fol. 33a, agree on the spelling of this name: ~I ~ .J-7 <l.J I ~, probably a corruption of Abii Mul)ammad 'Abdallah b. Muslim b. Qutayba ( ~ .J-7 ~ 0-! <l.J I ~ ~ ~I). Ibn Qutayba is also mentioned in Ibn al-Nadim's *Kitab aljihrist* as an author of a dreambook. His treatise survives in two manuscripts, Jerusalem *Yahuda ar.* 196 and Ankara, *Ism. Saib Sincer* I 4501, fols. 180b-217a. A1:xi Al}mad Khalafb. Al}mad has been identified by Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 58-60, as the last Saffarid amir of Sijistan (937-1006). As for the remaining two, about whom nothing else is known, the name Mul)ammad b. I:Iammad al-Razi al-Khabbaz tells us that he was Muslim; the religion of al-I:Iasan b. al-I:Iusayn al-Khallal, who was also the original author of this list of dream interpreters quoted by al-Dinawari, can only remain a matter of speculation. I have been unable to locate any information on Mul)ammad b. I:Iammad al-Razi al-Khabbaz ( = the baker) or al-I:Iasan b. al-I:Iusayn al-Khallal; on the problem of their identity, see also Lamoreaux, "Dream Interpretation in the Early Medieval Near East," p. 29, n. 6 (on al-Khallal) and p. 31, n. 9 (on al-Khabbaz).
<sup>35</sup>*Esad Efendi* 1833, fol. 33a and BN *arabe* 2745, fol. 45a: '5.!L..:...:..J I ..:,..... ~I ~I\_, ·1.:?~l .J.7j..! J..Ll..... ~I\_, !""'!".;-::...ii ~I .J-7 .:i:;..:..o..
36 Cf. Ibn al-Qif~i, *Ta'rikh al-Jjukama',* ed. A. Miiller and J. Lippert (Leipzig, 1903), p. 435. If the identification of Abii Makhlid with Abii Makhlid b. Bukhtishii' is correct, it follows that al-Khallal,
<sup>32</sup>For a discussion of al-Khallal 's list, see also Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 29-31.
<sup>33</sup> BN *arabe* 2745. fols. 44b-45b; *Esad Efendi* 1833, fols. 32b-33a; translated in Fahd, *La divination arabe,* pp. 341-43.
#### 244 CHAPTER SIX
of lbn al-Qifti, where he is mentioned, does not list any writings from his pen. Zayn al-Tabar! is otherwise unknown, unless his name is a corruption of the name of the ninth-century physician Abu '1-I:Iasan 'AH b. Sahl Rabban al-Tabari (b. 808-d. after 855). The form of the letters for the names Zayn and Rabban (originally Syriac for "our master") are identical in Arabic; the only difference is the placement of the dots: *.J-:! :\_;/* 0-7 ./7
Al-Dlnaw ari 's abridgment of al-Khallal' s list also mentions, in a separate category, the physicians who dealt with dream interpretation. Among them are two Christians: Bukhtishu' and Ahrun. Bukhtishu' could designate any member of this Nestorian family of physicians, including the aforementioned Abu Makhlid. Ahrun is Aaron, a Byzantine priest and doctor who lived around the sixth century. 38 Though neither of the two is known to have written anything on dreams, it is conceivable that they covered the subject in a more general medical treatise, possibly in discussing how to diagnose a disease with the help of a patient's dreams in the tradition of Galen's opusculum, *Peri tes ex enypnion diagnoseos* (On Diagnosis Based on Dreams). 39 The unpublished work, *Kitiib ft al-naum wa-al-ru)ii* (Book on Sleep and Dreams) by a ninthcentury Christian, the famous doctor and translator Qusta b. Luqa, also seems to belong to this kind of medical treatise.40
In addition to these works written by Christians that have no immediate connection with the particular kind of dreambooks to which the *Oneirocriticon*
al-Dinawari, and Abu Makhlid were approximately contemporaries.
<sup>37</sup> I have searched for Zayn al-Tabari in *al-Fihrist,* Ibn al-Qif!i and Ibn Abi U~aybi'a. Fahd does not give any information on him either. Abu '1-I:Iasan 'Ali b. Sahl Rabban al-Tabari was born a Christian and later converted to Islam (see Brockelmann, *GAL,* vol.SI, pp. 414-15); some sources say that he was Jewish. See Ibn Abi U~aybi'a, *Kitah 'uyun a/-anha'fi rahaqat a/-afiha',* ed. A. Muller, 2 vols (Cairo and Konigsberg, 1882-84), vol. 1, p. 308; also *Fihrist of al-Nadlm,* trans. Dodge, vol. 2, p. 956.
<sup>38</sup> See D. Jacquart, "Apropos des sources byzantines d'al-Majusi (Xe siecle): Le livre d'Ahrun," in *Tradizione e ecdotica dei testi medici tardo antichi e hizantini,* ed. A. Garzya, (Naples, 1992), pp. 157-69.
<sup>39</sup> For Ahrun, see Sezgin, *GAS,* vol. 3, pp. 166-68; for the Bukhtishu' family, see ibid., p. 243. Galen concedes that the content of a dream may be used for predicting the future, but it can also indicate the condition of the body and suggest the treatment that should be applied: ·End -mfrm Kat µavuKa nva cruyxcopouµ£v dvm, JtOO<; av -mfrm OtaKpt6£ifl 'tOOV clltO 'tOU crwµmo<; opµcoµ£vcov, OU paowv ditciv.' E6£acrmo youv tt<; 'tO E't£pov 'tOOV O'K£AOOV A.ietvov yqov£vm, Kat 'tOU'tO EKptvav JtOAAOt 'tOOV 1t£pt 'ta 'tOtaU'tO onvrov, 00<; itpo<; 'tOU<; OoUAOU<; 't£tV£tV 'tO ovap, c:\A.A.a itap£AU6fl 'tO O'KEAO<; EKElVO 0 av6pcoito<;, OUO£VO<; T]µrov 1tpOO'OOKi\crav1:0<; 'tOU'tO. Tov µi:v yap JtaAmmfiv EV a'iµmo<; 0£~aµ£vi;\ 06~av'ta i:mavm Kat µ6yt<; auTI;<; U1t£p£xona, nA.fi6o<; a'iµmo<; EX£tv E't£Kµflpaµ£6a, Kat o£tcr6m K£vwcr£co<;. (Galen, "L'opusculo di Galeno *De dignotione ex insomniis,"* ed. Guidorizzi, pp. 81-!05, II. 21-28 of the Greek text).
<sup>40</sup> See Sezgin, *GAS,* vol. 3, p. 273, no. 37.
belongs, we know of at least two works written by Christian authors that, as far as their title would allow one to judge, were dedicated to dream interpretation and must have been, in terms of topics covered, similar to the *Oneirocriticon.* Neither of them is extant, and therefore they cannot be compared with the Greek text. The earlier of the two, the *Kifiiya ft ta'blr al-ru)ii* by Abu Sahl 'Isa b. Yaqya al-Masil:ti al-FaylasUf (d. 1010),41 was written around the same time as the *Oneirocriticon.* The second was written much later, in the course of the thirteenth century, by Bar Hebraeus. Its title, *Kethiihhii dhe-Pushshiik lfelme* (Book on the Interpretation of Dreams) is mentioned in the bibliography of Bar Hebraeus drawn up by his brother Bar Sauma.42 At the beginning of the twentieth century, Bar Hebraeus's dreambook still survived in one unique manuscript, *Edessa* 50; 43 but it was never published, and today that manuscript is considered lost. 44
The only fragments of this tradition, possibly represented by the lost works of al-Masiqi and Bar Hebraeus, are preserved in the dreambook of al-Dinawari, who made a concerted effort to collect dream interpretations from a wide variety of sources, as is evident from the introduction to his extensive work. Except for the Muslim dream interpreters, al-Dinawari's most frequent and extensive quotations come from Artemidoros and reproduce verbatim the extant Arabic translation. This is an indication that al-Dinawari's sources were authentic, at least to the degree that he was able to verify their authenticity. Al-Dinawari also quotes the Greeks (Yunan), the Indians, the Jews, Jamasb,45 the Zoroastrians (Majus), individual Persian kings, such as Kisra Anushirwan, Christians (Na~ara), and Byzantines (Rum).46 The source of these quotations
4' Baumstark, *Geschichte der syrischen Literatur,* p. 318, n. 3.
44 According to information kindly provided by Sebastian Brock, *Edessa* 50 is certainly no longer in the collection where it was recorded by Baumstark. It may have been destroyed in 1915, "the year of the sword," though a few manuscripts of that collection are reputed to be in the church of St. George in Aleppo. Access to the St. George manuscripts is exceedingly difficult. I wish to thank Professor Brock for this information, as well as for providing me with references to Bar Hebraeus's *Candelabrum of the Sanctuary* and *Ethicon.*
45 Jamasb or Jamasf was a contemporary of Zarathustra and an adviser to the mythical king Jushtasf (Vishtaspa, Hystaspes). He is purported to have written on alchemy, astrology, and medicine, as well as a book on dream interpretation. His dreambook is also mentioned by al-Damlrl; see Ullmann, *Die Natur- und Geheimwissenschaften im Islam,* pp. 183-84.
46 Arabic sources use the term "Riim" for both the Byzantines and the Melkite Christians living
<sup>41</sup> Fahd, *La divination arahe,* p. 347, no. 75.
<sup>42</sup> Bar Hebraeus, *The Chronography of Bar Hehraeus,* trans. E. Wallis Budge, vol. 1 (Oxford, 1932), p. xxxiv, no. 20; see also A. Baumstark, *Geschichte der syrischen Literatur, mit Ausschluj3 der christlich-paliistinensischen Texte* (Bonn, 1922), 230, 316-17, 352-53, 377; J.M. Fiey, "Esquisse d'une bibliographie de Bar Hebraeus (t 1286)," *Parole de/' Orient* 13 ( 1986), p. 311.
#### 246 CHAPTER SIX
cannot be examined in the absence of a critical edition of al-Dinawari, nor does it lie within the scope of this study. But it is important to examine if the interpretations that al-Dinawari ascribes to the Christians and the Byzantines coincide with any of the interpretations found in the *Oneirocriticon.* Under the circumstances this examination cannot be exhaustive, especially since the source of each interpretation is not always clearly labeled in al-Dinawari's work. Only a sample of the interpretations will be tested from among those attributed to the Christians and the Byzantines and conclusions will be drawn based on the statistical distribution of the results. The Christian and Byzantine interpretations that are clearly labeled as such in BN *arabe* 274547comprise the sample.
Both the *Oneirocriticon* and al-Dinawari's dreambook include passages explaining the period of time that is required for the event forecast by a dream to be fulfilled. Al-Dinawari quotes the opinions of both the Muslim and the Christian dream interpreters on the topic. The Christian calculations are very different from those in the *Oneirocriticon,* which agrees with the Muslim views quoted in al-Dinawari.48 The same disparity exists between the dream interpretations offered in the *Oneirocriticon* and al-Dinawari's Christian quotations. BN *arabe* 2745 discusses approximately five hundred dream symbols49 and gives many more interpretations; Christian dream interpretations are given for only sixty-three of the dream symbols. They are attributed to two Christian communities, the Na~ara and the Rum, but the vast majority-fifty-three out of sixty-four interpretations-are attributed to the Na~ara and only seven to the Rum (three are attributed to both).
Al-Dinawari also quotes ten narrations of actual dreams dreamt by Christians and their interpretations. Eight of those are attributed to the Na~ara and one to the Rum. A tenth narration is quoted according to both the Na~ara and the Rum, but involves a dream dreamt by one of the Na~ara. Of the sixty-three
in Muslim lands; al-Dinawari, however, seems to restrict its use to the Byzantines, since he records a dream dreamt by the Byzantine emperor Heraclius among the dreams by the Rum.
<sup>47</sup> BN *arahe* 2745 contains the lengthy introductory chapters of the work, and the interpretations of dream symbols up to *fw;l* 12, but it stops at *fa,1/* 12, *bah* 2; the complete work comprises 30 *fu0ul.* BN *arahe* 2745 contains enough of the work for the results of my examination to be statistically valid, and has the advantage of being easily accessible.
<sup>48</sup> For a translation of al-Dinawari's Christian passage on the time required for the fulfillment of a dream, see Appendix 3.
<sup>49</sup> BN *arahe* 2745 includes 469 *ahwah;* each *bah* is dedicated to a dream symbol, but gives several different interpretations, depending on the particular paraphernalia of each dream symbol in a dream.
dream symbols only twenty-seven are also discussed in the *Oneirocriticon,* while none of the narrations is common to both texts.
The following are the Christian interpretations and narrations of dreams included in BN *arahe* 2745; the numbers in parenthesis refer to the folia on which they appear. The dream symbols that also occur in the *Oneirocriticon* are italicized (for translations, see Appendix III).
The interpretations attributed to the Na~ara are: Adam and Eve (49a-b); Abel, Noah ( 49b); Abraham (50a); Isma'H, Isaac, Jacob (50b); Joseph (5 la); Moses and Aaron (52a); David and Solomon (52b ); narration of the dream of a Christian about John the Baptist and its explanation by a bishop (53a); *Jesus Christ,* and narration of the dream of a Christian about Jesus Christ (53b); narration of the dream of another Christian about Jesus (54a); living at the time of the prophets (55a); narration of the dream of a Christian who dreamt of the Holy Spirit (58a); *the angels* (58a-b); *the spiritual angels* (60a); <sup>50</sup> becoming young (70a); *eating brains* (75b); *broken teeth* (83a); *broken upper arm* (89b ); *milking a she-camel and drinking her milk,* drinking the milk of a goat, *drinking the milk of a lioness* (104b); milk of a wolf (105a); *milk of a bitch* (105a); milk of a snake (105a); *eating clarified butter* (106a); eating bread with cheese ( 106b ); <sup>51</sup>*nose bleeding* ( 107b ); narration of the dream of a Christian on prostration (124a); mastering and compiling a book (127a); praying towards the east (142a); *becoming a priest or a monk* (144a); narration of the dream of a Christian who dreamt that he became a priest (144a); being transformed into a king (146b); frequenting the doors of kings and visiting the king, passing by a sultan,52 bringing food to a king or a prelate or disputing with a king (150a-b); becoming a soldier in an army (152a); the cheerful and happy countenance of a judge (154b); *the position of a judge* (154b); *eating the flesh of a young man* (157a); water in one's ears (158b); narration of the dream of a Christian about four men congregated at a house and its interpretation according to the Gospel ( l 65b ); narration of the dream of a Christian about forgiving his neighbor and its interpretation according to the Gospel (175b); dreaming that someone is hurling stones at <the dreamer's> head (190a); <sup>53</sup>
<sup>50</sup>.:,~~J.rl I ~~I (angels of spirit), such as Jibrll, Mika'il and lsra'fil, as opposed to angels created from light; see Lane, *Arabic-English Lexicon,* s.v. ~L...:..J\_;. This is a Muslim classification of angels; I am not aware of it in Christianity.
<sup>51</sup> No dream of eating bread with cheese occurs in the *Oneirocriticon;* on bread alone, see Drexl 196, 18-19; for cheese alone, see Drexl 207, 5-13.
<sup>52</sup>*jlb.L....* .,,h )-,·~I
<sup>5</sup>Jo).7 .. .::JL,. <L...ul\_; .\_j~ l.J..:>.I \_,I\_; *j..o*
*holding a stretched bow* (19lb); *holding an arrow* (192a); *holding a dagger* ( l 94a); *a cuirass* ( l 95a); a shield, 54 *coats of mail for the arms* ( l 95b ); inflicting a wound with a knife55 or an instrument made of iron (200a); <sup>56</sup>*hemorrhage from a wound* (203b ); *being crucified* (205a).
The interpretations attributed to the Rum are: *dreaming of locks of your wife's hair being cut* (78a); *dreaming that you were suckling the milk of your wife* (104a); *milking a cow and drinking her milk* (104b); dreaming that you are eating your vomit (11 la); *collecting the dung of a cow* or a goat (l l 7b); receiving an egg (118a); narration of the dream of Heraclius, king of the Rum, about Mul.iammad (l 19b); having safe and sound feet (120a).
The interpretations attributed to both the Na~iirii and the Rum are: *adultery/fornication (zinii)* (168a); *deflowering a slave girl* (181 a); narration of the dream of a Na~riinl who dreamt that he was disputing with the king ( l 85b ); dreaming that one is being hurled in a catapult by women ( l 90b ).
Of the twenty-seven dreams that are discussed in both texts, only eight have similar interpretations,57 and in all eight cases, the interpretations are not unique. Similar interpretations are also given by Muslim dream interpreters, either in a preceding paragraph in the work of al-Dinawarl itself, or in other Arabic dreambooks; half of these interpretations also occur in Artemidoros.58 In other words, they are not specifically Christian, but rather commonplace.59 Our examination of the Christian interpretations quoted in al-Dinawari leads to
56 The dream of inflicting a wound does not occur in the *Oneirocriticon,* but the dream of receiving a wound does. The two interpretations are different:(.>-'!- GI <.SI.; j-a (.5.;l....o...:JI .,:.,Ju ~ ~ *"'1* J d...'..JL......... ~ ....:.Ll J.o..1-:>. ~ JI ~ (The Na~arii said: "If someone dreams that he inflicted a wound with a knife or a sharp instrument, his bad side will appear; and there is nothing good in this dream" ); cf. Drexl 63, 23-25: £<iv *w;* 'iliu, on µaxaip<;x £7tA.fiy11 7tapci UVO~ yvcopiµou, aya0o7tOtTJ0ficrETat 7tap' m'.noii, El OE: 7tapa ayvWO"'WU, EtpTJV07tOtficrEt µEta £x0poii (If someone dreams that he is wounded with a knife by somebody he knows, he will receive benefit from him; if <the dreamer was wounded> by a stranger, he will make peace with an enemy).
<sup>54</sup> A shield is discussed in the *Oneirocriticon* only if a woman dreams of it (Drexl 205, 23); even so, its interpretation is different from the one given in al-Dinawari.
<sup>55~</sup>
<sup>57</sup> Five of those are attributed to the Na~ara (N) and three to the Riim (R). The dreams with similar interpretations in the two texts are: broken teeth (83a-N), milk of a lioness ( 104b ), holding a dagger (I 94a-N), hemorrhage from a wound (203b-N), being crucified (205a-N), the milk of a woman (104a-R), milking a cow and drinking her milk (104b-R) and collecting the dung of a cow ( 117b-R). For details, see Appendix 3.
<sup>58</sup> The interpretation of broken teeth, hemorrhage, crucifixion, and women's milk.
<sup>59</sup> For details, see Appendix 3.
the conclusion that, whatever their source was, it was unknown to the author of the Oneirocriticon.
It is impossible to verify whether al-Dinawari knew about Christian dream interpretation from oral tradition or from books, and if the latter in what language these books were written. But there is evidence that the Christian interpretations offered by al-Dinawari, which do not cover all the dream symbols discussed in his book, are collected, at least in part, from works that were not specifically dedicated to dream interpretation. Since it is impossible under the circumstances to investigate the source of each one of the Christian interpretations and narratives that occur in al-Dinawari, one example will suffice.
Al-Dinawarī quotes a dream dreamt by the Byzantine emperor Heraclius: 200
راي هرقل ملك الروم ذات ليلة كـأن ملك الخـتـان ظاهر فـقص روياه على اصحابه فلم يلبث ان جاءه رسول صاحب بصري برجل من العرب يقوده فقال يايها الللك ان هذا رجل من اهل الشاء والابل ويزعم انه كان ببلاده حرب وقد خرج رجل ويزعم انه نبي وكان هذا الرجل ابا سفيان بن حرب فقال جردوه فاذا هو مختتن فقال لاصحابه هذا تاويل روياى.
One night Heraclius, king of the Rum, saw in his dream that the kingdom of circumcision was victorious. He related his dream to his companions, and it did not take long before a messenger of the lord of Busra came to him with a Bedouin whom he was guiding. He said <to Heraclius>: "O King, this man is from the people of the sheep and of the camels, and says that in his country there is war and that a man came forth who claims to be a prophet." The <Arab> was Abū Sufyan b. Harb.61 Heraclius said: "Strip him!" and lo and behold, the man was circumcised! And Heraclius said to his companions: "This is the interpretation of my dream!"
This dream of Heraclius is a well-known story that can be found in a variety of Islamic sources. The same anecdote is narrated in the history by al-Tabari (839-923) and in the Kitāb al-aghānī by Abū-1-Farāj al-Isfahānī (897-967). 62
<sup>60</sup> BN arabe 2745, fol. 119b.
<sup>61</sup> The reading of the name is given in Esad Efendi 1833, fol. 73a, l. 14. Abu Sufyan was a rich Meccan merchant who opposed the Prophet during the first years of his mission. He emerged as leader of the Meccans in the years that followed the battle of Badr and conducted the negotiations which resulted in the acknowledgment of the Prophet's authority by the citizens of Mecca. He became Muslim at the time of the conquest of Mecca. His son, Mu'awiya b. Abi Sufyan, became caliph in 661.
<sup>62</sup> Abū Jaffar Muhammad b. Jarīr al-Tabarī, Ta'rīkh al-rusul wa-al-mulūk, vol. 2 (Cairo, 1961), pp. 646-47 [1562/1]. The narration of this anecdote is much more extensive in al-Tabari. The editor of the Ta'rikh states (p. 646, n. 3), that this anecdote is also found in Kitab al-aghani,
A slightly different version than the one narrated by al-Dinawari is included in one of the best-known hadīth collections, the Sahīh of al-Bukhārī.63
In addition, in the very first pages of his dreambook al-Dinawari explains the manner in which he collected the interpretations he cites in his work. His enumeration of sources implies the use of material other than dreambooks and possibly also oral tradition:64
ونقلت الليبه مـقـالات المعــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ والفسرين وفقهاء الدين والزهاد والصالحين واولى العلوم من الفلاسفة والاطباء والمنجمين والشعراء والكهنة والقافة والسحترة وذوى الفراسة والبصيراء والمتاوّلين من حبر ماهر واسقف وراهب وقس عالم وحكماء يونان والروم ونستاك الهند والبـراهمـة والاكــاســرة والموابدة والهـرابدة وحكيت ما الحتجوا به عند التعبيير والتاؤيل من اوضع الدليل من آيات التنزيل والتورية والانجيل واخبار الرسول صلى الله عليه وسلم وما ذكروه من العلل وموجبات عقول اهل النحل بعد ان قابلت حجة كل امّة ذمية من كتابها بكتابها مع اممها ونقلتُها على اكتفاءي بعلم خيـر امـة من علم اهل الذمبة ليعلم ان هذا العلم قديم وانّ منّ الله تعالى به على من اتاه عظيم ولم اعتـمـد فى نظمى على علمى ولا فى نقلى على قـولى و لا احلت فى شى منـه على درايتى و لا رمت بالترتيب سـوى التقريب فانما لى روايتى و حكايتى فقد غاص هولاء الفضلاء النصحاء العقلاء قبلى على الدقائق والحقائق فبقوا وما وبقوا بل توخيت تسهيل السالك الى السائل واعتناء بشدتها وطلبتها عن تفتيش سائر كتبها وافردت لكل منها بابًا اشبعتُه استقصاء واستعابً موسوماً بعدد مذكور فى فصل مشهور معلوم متلوأ بباب بعلاوته مرسوم في رويا معبرة او مجربة هو منشد للضالة ومظنة التاؤيلات الشاذة وفرغت منه في شهر الرمضان سنة سبع وتسعين وثلثمائة حامداً لله ومصليًا على محمد رسول الله ومفوضًا امرى الى الله وهو بصير بالعباد.
I have transferred into [my book] the sayings of the dream interpreters from among the prophets and the rightly guided imams and the generation after the Prophet Muhammad and the commentators of the Qur'an and the legal scholars of Islamic jurisprudence, the ascetics, the holy figures, and the masters of science from among the philosophers, the physicians, the astrologers, the poets, soothsavers, and predictors of the future, magicians, physiognomists, the knowledgeable and
vol. 6, pp. 345-49 (edition of the Dar al-Kutub). I did not have access to this edition and could not verify the reference.
<sup>63</sup> See L. Pouzet, "Le hadith d'Héraclius: une caution byzantine à la prophétie de Muḥammad," in La Syrie de Byzance à l'Islam, VIIe-VIIIe siècles, ed. P. Canivet and J .- P. Rey-Coquais (Damascus, 1992), pp. 59-65.
<sup>64</sup> Quoted from Esad Efendi 1833, fol. 3b-4a, with occasional readings from BN arabe 2745, fols. 1b-3a to improve the text.
the interpreters from among the proficient rabbis, bishops, monks, and scholarly priests, the Greek and Byzantine wise men, the ascetics of India and the Brahmans, the kings of Persia, the pontiffs of the Magians, and the Zoroastrian priests.
I have reported what they advance as an argument in the interpretation and explanation of dreams from the clearest tokens among the verses of the Revelation <i.e. the Qur'an>, the Torah and the Gospel, the reports of the Prophet, may the peace and blessings of God be upon him, and what they mention from among the causes and motives of the minds of the sectarians, after I had collated the argument of every *dhimmf* nation taken from their scripture with their <original> scripture in the opinion of their nations; and I have transferred them to my satisfaction from the science of the *dhimmis* into the best science of the <Muslim> *umma,65* so that it be known that this science is ancient and that the grace of God <manifested> with this science towards the one to whom he gives it is great.
In my compilation I did not rely on my own proficiency, nor in my account <did I rely> on my own word. And I did not refer anything in the book to my own knowledge. In my arrangement I did not seek anything but accessibility. Mine is only transmission and literal quotation, because these excellent, sincere, and clever men before me have dived down to the fine points and the essence <of things>,66 and they lasted, they did not perish. I aimed at facilitating the methods <of answering> the questions and at maintaining their forcefulness and their desired objectives based on scrutiny of the rest of their books.67 And for each one <of these questions> I have singled out an entry (bah) which I filled with minute investigation and thorough study, marked with the number mentioned in the established, known chapter *(fa0l). <sup>68</sup>*This is followed by an entry (Mb) on the corroboration <of the preceding theoretical exposition> titled "On the Interpreted or Attested Dreams." This can serve as a guide for those who were led astray and is the most likely place to find unusual explanations.
I finished this book in the month of Rama<;lan in the year 397 <1006 A.D.>,
66 Following the reading of *Esad Efendi* 1833 ~ .\_,.,L...L (to plunge, to dive for); BN *arahe* <sup>2745</sup>has: ~ JJL<. (to leave behind).
67 The passage in *Esad Efendi* 1833 reads as follows: ... ~ :\_,..c. ~J Y..:.~ 4-'IJ. The same passage in BN *arahe* 2745 reads: ... ~ .J..c. ~J l+'..i...'.;...:. c~IJ.
<sup>65</sup> The translation follows *Esad Efendi* 1833; BN *arahe* 2745 in this instance gives a different text: ;i\_,, I ~ F *c l.i.:iS* <sup>I</sup>~ l+'i.:i'.; J 4-o L. I *t-"* 4-;l.::.S...: 4-;L:.S .\_:,.... .....,.\_. ~ ;i\_,, I *JS* ~ ..:..J..:u ... .:, I *H* L jJ I j.b I F (After I had collated the argument of every *dhimmi* nation taken from their scripture with their <original> scripture through the help of their leader and their trustworthy people to a satisfactory degree by means of the best science of the <Muslim> *umma* <and> the science of the *dhimmis,* so that it be known ... ). If we accept the reading from BN *arahe* 2745, it would imply that al-Dinawari also consulted with "leaders and trustworthy persons" in non-Muslim communities and therefore based his dreambook on both oral and written traditions. It has to be pointed out, however, that the readings in both manuscripts are problematic and the passage in question seems to be corrupt.
<sup>68</sup> I suppose here al-Dinawari means that he is going to comply with the organization of chapters that is obligatory for the genre of *ta'hir.*
#### 252 CHAPTER SIX
praising God and praying for MuJ:iammad the messenger of God and entrusting my affair to God, for he is watchful over his servants.
Even if there is an element of rhetorical exaggeration in this prologue, which is written in rhyming prose and preceded by the flowery praise of the ruler to whom the work is dedicated, its honesty is proved by the abundant quotations that occur throughout al-Dinawari's dreambook.
This evidence suggests that there existed a tradition on dream interpretation among Christians living in Muslim lands, both in dreambooks of the same type as the *Oneirocriticon* and in other types of works. Their tradition was not identical to the Muslim one, at least not in its entirety. It also does not seem to have been as extensive, which might have been the reason for the appropriation of Islamic dreamlore by Christians, as evidenced in the Syriac dreambooks of BL *Or.* 4434 and the abridgment oflbn Qutayba's dreambook by lbn al-Bahlill.
Given these examples of Christianized Islamic dreambooks, we cannot exclude the possibility that the immediate source of the *Oneirocriticon* was an already Christianized dreambook written in Arabic. However, a number of indications furnished by the Greek text suggest that it is more likely that its Christianization occurred when it was translated into Greek. The *Oneirocriticon* does not seem to have been based on an Arabic text intended for Christians living in Muslim lands, because several of the chapters "From the Indians" purportedly offering Christian interpretations contain references to polygamy ,69 and the five fingers are interpreted as the five daily prayers,70 a Muslim religious obligation. Christians, of course, had only one wife and no prescribed number of daily prayers (the formalization of private prayer at set times led
<sup>69</sup> Drexl 43, 15 ff. (rra' EK nov' Ivowv . ai KmaKA£l0£<; 'tWV wµwv di; yuva'iKa<; OtaKpivovmt rraA.A.aKci<;, rro0n vo'ti':pai; 'twv voµiµwv ... Kat xapficr£'tm i':rr't mti; rraA.A.aKa'ii; m'.nou ... £av ni; lOl] Ka't <sup>0</sup>ovap, on EKA.cicr811 Ti Kmad£ii;, µqciA.wi; 0A.tl3ficrnm Kat xwptcr0ficrnm 1WV JtaA.A.aKWV a'inou. £av OE lOl], on £0£pcl1t£UCJ£ 10 KA.acr0£v, Ti 0£parr£ia 1iiv 0£pait£ia cr11µai vn 1wv rraA.A.aKwv ); Drexl 55, 5-7 (rra' EK 1wv' Ivowv · a\. rrkupai dm v ai yuva'iK£<;' al µ£v avw, al µ£ytcrmt, yvi'tcrtm yuvatKE<; dcrtv, ai OE Kcl10l al EK y£voui; dcrtv cruyy£v£'ii;); Drexl 115, 17-18 (10 1t£pt 1a Kpurr1a KciA.uµµa Tj10t 1t£ptcrK£A.tcrµa £ii; rraA.A.aKiiv Tj di; yuva'iKa <Kpivnm>). The examples could be greatly multiplied if we add the chapters purported to be from the Indians, Persians, and Egyptians together, or from the Persians and Egyptians alone; but these are not presented as exclusively Christian interpretations. I did not count as a reference to polygamy each time that the women of the dreamer are mentioned. Most of the time this plural refers to all the kindred women who belong to the same household, which the *Oneirocriticon* often divides into two groups: yuva'iKm yvi'tmm are the women with whom marriage is allowed (cf. Drexl 50, 5: ai wµorrA.ci1m di; 1iiv yv11criav 'tWV yuvmKWV civciyovmt Tjwt 1iiv µri1£pav 1wv 1£Kvwv auwu), and yuva'iKm cruyy£v£'i<; are the women who belong to a man's household but whom he cannot marry. The two categories of women are clearly distinguished in Muslim law.
<sup>70</sup> Drexl 45, 11-25 and 68, 29-30.
to the services of the Hours, but there were six of those daily), but Christians living in Muslim lands must have known that their Muslim fellow citizens were allowed more than one wife, and they cannot have been deaf to the mu'adhdhins calling the faithful to prayer five times daily from the top of the town's minarets.71 The occasionally sloppy adaptation of Muslim notions to Christian ones that is evident in the Oneirocriticon would have been obvious to Christians living among Muslims; they would consequently have found it difficult to accept the resulting product as genuinely Christian. But it would not have been so obvious to Christians living in the Byzantine Empire, who would have been unfamiliar with Islamic practices.
The Christianization of Muslim notions in a work translated from Arabic into Greek is not a phenomenon limited to the pages of the Oneirocriticon. It can be found in other texts as well, notably the Greek translations of the astrological writings of Abū Ma'shar (787-886), also dating from the tenth century.72 Although the institutions of monasticism and priesthood do not exist in Islam, in the Greek anthology of Abū Ma'shar's works, Ta mysteria tou Apomasar (The Mysteries of Apomasar) we read passages such as:13 δηλοί άγαθοσύνην των μοναχών των γερόντων τε και έγχωρίων (It indicates goodness for the monks, both the desert fathers and those residing among men); 8 δηλοί άναγωγήν των βασιλέων της τε πίστεως και του νόμου, έμφαίνων πολέμους πολλούς, ύποκρίσεις τε και κακοσύνην άρχιερέως (It indicates exaltation of the emperors, regarding both religion and the law, manifesting many wars, as well as deceit and corruption of a high priest); 35 έξουσιν οί άρχιερεῖς καὶ οἱ γέροντες ἐκ τῶν βασιλέων δυσκολίαν (The high priests and monks will face difficulty from the emperors); 8 οηλοί ότι οί ανθρωποι καταγίνονται είς τα καλά έργα της πίστεως και είς κτίσματα ἐκκλησιών καὶ γεφυρών καὶ τῶν ὁμοίων (It indicates that men are occupied with good, pious deeds, and with building churches, bridges and
<sup>71</sup> Other obvious Muslim notions that survive in the Oneirocriticon, such as the religious significance of the color green, might have been less well known among the Christian dhimmis .
<sup>72</sup> For the dating of the Byzantine translations of Abu Ma'shar's works to the 10th century, see Abu Ma'shar, Albumasaris: De revolutionibus nativitatum, ed. Pingree, pp. vi and viii; also idem, DSB, s.v. "Abū Ma'shar al-Balkhī, Ja far ibn Muhammad" ; idem, "Classical and Byzantine Astrology," p. 227, n. 2; idem, From Astral Omens to Astrology from Babylon to Bikaner (Rome, 1997), pp. 63-77.
<sup>73</sup> The work is still largely unpublished, though excerpts from it have appeared in CCAG. I am quoting from CCAG, vol. 2, which has excerpts from Marcianus gr. 324; for a description of the manuscript, see ibid., pp. 4-16 (codex 5).
<sup>74</sup> Ibid., p. 123, 21-22.
<sup>75</sup> Ibid., p. 123, 24-26.
<sup>76</sup> Ibid., p. 126, 29-30.
such); " δηλοί ζήτησιν σοφίας και σύνεσιν, μεγαλείον τε και τιμήν των ήσυχαστών, και τών κρατούντων τὴν πίστιν, καὶ οἰκοδομὴν τῶν ἐκκλησιῶν και των εύκτηρίων (It indicates the pursuit of wisdom and intelligence, glory and honor for the hermits and those upholding religion, as well as the building of churches and houses of prayer);78 δηλοί καλλοσύνην [sic] των δουλειών των γερόντων, μεγαλείόν τε της τάξεως των άρχιερέων και μεγιστάνων (It indicates goodness in the business of the gerontes" and glory in the authority of high priests and noblemen). 80
In the Greek translation of another work by Abu Ma'shar, the Peri tes ton eton enallages (On the Revolutions of Nativities), we find the following examples of the translator's efforts to make Muslim notions intelligible to his Christian readers: εύλαβές σχημα έχοντας ή μονάζοντας ή έκκλησιαστικούς (Those wearing the habit, either monks or ecclesiastics); < < O Zeùc σημαίνει... > δικαιοσύνας, άρχάς, πολιτείας, δόξας, προστασίας, ίερωσύνας, πίστεις, νίκας (<Jupiter signifies ... > justice, authority, civil affairs, glory, protection, priesthood, faith, victory);83 εὐφημουμένους δι' ἀρετὴν καὶ ευσέβειαν, τιμαίς αναθημάτων κοσμουμένους ή ίερωσύναις (Praised for their virtue and piety and adorned with the high value of their offerings or with priesthood).84
The translation and adaptation of Arabic texts into Greek is a large subject, and deserves a more extensive discussion than it can be given here. However, these examples indicate that the Byzantine translators of the tenth century did not hesitate to depart from the literal meaning of an Arabic term when rendering Arabic texts into Greek, and were more interested in conveying the meaning of a passage, than in adhering to its letter. They were therefore ready to Christianize the Muslim notions found in their Arabic sources. This means that the adaptations made to the Muslim sources of the Oneirocriticon were not a unique feature of this particular text, but reflect a practice that also appears in other Byzantine texts of this kind.
To summarize, then, the Oneirocriticon is based on sources written in the Arabic language. A Christian tradition of dream interpretation that had several
79 I hesitate to chose between the possible meanings of gerontes: it can mean either "elders" (an honorary title equivalent to the Arabic shaykh) or "desert fathers."
<sup>77</sup> Ibid., p. 127, 35-36.
<sup>78</sup> Ibid., p. 127, 37-p. 128, 2.
<sup>80</sup> CCAG 2, p. 128, 11, 3-4.
<sup>81</sup> Abū Ma'shar, Alhumasaris: De revolutionibus nativitatum, ed. Pingree.
<sup>82</sup> Ibid., p. 247, 17-18.
<sup>83</sup> Ibid., p. 250, 1.
<sup>84</sup> Ibid., p. 251, 5.
elements in common, and was not completely distinct from the Muslim tradition, seems to have existed among Christians living in Muslim lands at around the time that the *Oneirocriticon* was written, as is evidenced primarily by the interpretations labeled "Christian" and "Byzantine" in the dreambook of al-Dinawari. This means that, if the author of the *Oneirocriticon* had been interested in using an eastern Christian work as a source, he could have procured one. Instead, he used none of the sources on Christian dream interpretation that are quoted in al-Dinawari's dreambook, but relied on sources linked to the Muslim tradition. The Christianization of Islamic notions in the *Oneirocriticon* can appear as genuinely Christian only to someone ignorant of the basic tenets of the Muslim faith. Christianizing Muslim notions was a practice followed by more than one Byzantine translator, when rendering Arabic scientific texts into Greek. Given all this, it is unnecessary to postulate a Christian-Arabic dreambook as the immediate source of the Greek *Oneirocriticon.* Rather, it seems likely that the *Oneirocriticon* was based directly on Muslim Arabic dreambooks and that the Byzantine translator was the one who adapted the content of his sources to cater to the expectations of his Christian readers.
#### CHAPTER SEVEN
#### THE CHRISTIAN AND ISLAMIC ASPECTS OF THE ONEIROCRITICON
A number of passages from the Oneirocriticon refer to the Christian and the Muslim faith. They include passages that interpret purportedly Christian dream symbols; passages where one would expect to find Muslim influence (e.g., abhorrence for eating pork or drinking alcohol); passages that appear to have been inspired by the Bible; and finally, passages in the thirteen anecdotes of the Oneirocriticon involving actual consultations with a dream interpreter that constitute the only instances in the Greek dreambook where Muslim Arabs, such as Ibn Sirin and Caliph al-Ma'mun, are mentioned. These passages will be compared to equivalent Arabic interpretations and narratives to elucidate the process by which the Greek Christian author adapted his Arabic Muslim sources.
#### Resurrection of the Dead
The first dream symbol interpreted in the Oneirocriticon comes immediately after the four introductory chapters and is on the Resurrection of the Dead according to the Indians. This interpretation addresses a vital component of Christian eschatology: 1
Έαν τις ἐν ὀνείρω θεάσηται ἀνάστασιν νεκρων, ἐν ὁ τόπω θεωρήση τις κατα τούς ύπνους ανάστασιν νεκρών, έκει δικαιοσύνη πραχθήσεται. ἐαν γὰρ ώσιν έκει άδικοι, τιμωρίαν δώσουσιν, εί δε άδικούμενοι, δικαιωθήσονται ταχέως, διότι έν τη άναστάσει κριτὴς δίκαιός έστιν ό θεὸς μόνος. καὶ διὰ τοῦτο ἔχει τὴν δύναμιν, καθ' ὅσον ἐνδέχεται, αῦτη ή κρίσις.
If someone sees in a dream the Resurrection of the Dead, justice will be effected where he saw it. If unjust men are there, they will be punished; if people are being wronged, they will soon be vindicated, for in the Resurrection of the Dead God alone is a righteous judge. For this reason, this Judgment is as valid as possible.
<sup>1</sup> Drexl 3, 25-4, 3 (ε' ' Έκ των Ινδών έρμηνεία περὶ ἀναστάσεως).
This is an open access chapter distributed under the terms of the CC BY 4.0 license.
The Resurrection of the Dead on the day of God's Last Judgment is also part of Muslim eschatological belief. The Qur'an graphically describes the portents and events of Judgment Day in a number of passages (sūra 22:1-7, 75:1-40, 81:1-19, 82:1-19, 83:4-20, 84:1-19), which inspired the Arabic interpretations of dreams about it. The Arabic interpretations of the Resurrection correspond almost word for word with the entry of the Oneirocriticon. According to Ibn Qutayba:2
من رائ ان القيامة قد قامت بمكان فان العدل ينيسط فى ذلك المكان لاهله ان كانوا مظلومين وعليهم ان كانوا ظالين لان يوم القيبامة يوم الفصل ويوم الجزاء والدين قـال الله عز وجل ونضع الموازين القسط ليـوم القـيـامـة فـلا تظلم نفس شيئا.
Whoever dreams that the Resurrection of the Dead has already happened at a <particular> place, justice will spread to that place to the advantage of its people, if they were oppressed, or against them, if they were oppressors. For the Day of the Resurrection is the Day of Judgment, of punishment and doom. God Almighty said: "And We set a just balance for the Day of Resurrection so that no soul is wronged in aught" [Qur'an 21:47].
The interpretation of al-Dinawarī is similar:3
... فان کان اهل ذلك المكان الذين قامت فيه ظالين انتقم منهم او مظلومين نصروا لان يوم القيامة يوم العدل.
... and if the people of the place where the Resurrection of the Dead occurred were oppressors they will be punished; and if they were oppressed, they will be saved, because the Day of Resurrection is the day of justice.
I ikewise in al-Muntakhab:4
قال الاستاذ ابو سعد رحمه الله قال الله تبارك وتعالى ونضع الميزان القسط ليوم القيـامـة فـلا تظلم نفس شـيـئـا فـمـن راى ان القيامة قد قامت فى مكان فانه ينبسط العدل فى ذلك المكان لاهله فينتقم من الظالمين وينصر المظلومون لان ذلك يوم الفصل والعدل.
Master Abū Sa'd, may God have mercy on him, said: "God Almighty said: 'And We set a just balance for the Day of Resurrection so that no soul is wronged in aught' [Qur'an 21:47]. If someone dreams that the Resurrection of the Dead is
<sup>2</sup> Ibn Qutayba, bah 2, Yahuda ar. 196, fol. 25b.
<sup>3</sup> Al-Dinawari, fast 30, bab 2, Esad Efendi 1833, fol. 273b.
<sup>4</sup> Al-Muntakhab, p. 67.
#### CHAPTER SEVEN
already taking place somewhere, indeed justice will spread there to the advantage of its people. And the oppressors will be punished and the oppressed will be saved, because it is the Day of Judgment and justice."
The interpretation of Ibn Shāhīn (no. 193) is similar:
ومن راي ان القيامة قامت وبسط الله العدل بين الناس يدل على انه إن كان فى اهل ذلك المكان مظلومون سلط الله تعالى على ظالمهم الشدة والمضرة.
If someone dreams that the Resurrection took place and that God spread justice among mankind, if there were oppressed people in the land <where the dream was dreamt>, <the dream> signifies that God Almighty will inflict the oppressors with calamities and miseries.
Al-Nābulusī also states والقـيــامــة عـدل وانصـــاف الــظــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ Resurrection is justice and fairness towards the oppressed by the oppressor).3
Though it is impossible to ascertain whether the Arabic sources of the Oneirocriticon justified the interpretation of the Resurrection with a Quranic passage, it is very likely that they did, especially since such a justification can be found in the earliest of the extant Arabic dreambooks, that of Ibn Qutayba. The wordy justification of the Greek interpretation ( "In the Resurrection of the Dead. God alone is a righteous judge. For this reason, this Judgment is as valid as possible" ) may have been meant by the author of the Oneirocriticon as a substitute for Qur'an 21:47 ("And We set a just balance for the Day of Resurrection so that no soul is wronged in aught" ) that consistently appears in the Arabic dreambooks. 6
#### Paradise and Hell
After the interpretation of the resurrection, the Oneirocriticon proceeds to Paradise (chapter 8) and Hell (chapter 9) according to the Indians. The Christian interpretation of Paradise in the Oneirocriticon is as follows:7
Al-Nābulusī, vol. 2, p. 137, s.v. قدامة .
<sup>6</sup> For additional comparisons of the interpretation on the Resurrection of the Dead from the Oneirocriticon with Arabic interpretations from the dreambooks of al-Khargushi and al-Qayrawani, cf. Lamoreaux, "Dream Interpretation in the Early Medieval Near East," pp. 256; p. 319, table 10.
<sup>7</sup> Drex1 4, 15-5, 3.
#### η ' Έκ τών ' Ινδών περί παραδείσου
Έάν τις ίδη άγαθός ών είσερχόμενον έαυτον έν τω παραδείσω. ουτός έστι σεσωσμένος, διότι θεόθεν αύτφ τουτο το εὐαγγέλιον, έξ ών ἐξειργάσατο ἀγαθῶν εί δε κακούργός έστι, μετανοήσει και σωθήσεται, άλλά και πλούτου κοσμικού τουτο ἐπιτυχίαν σημαίνει. ἐὰν δὲ ἴδη ὅτι ἔφαγεν ἐκ τῶν καρπῶν τῶν δέγδρων. αυτός εύρήσει έν τη πίστει αύτού σοφίαν και γνωσιν, διότι οί καρποί του παραδείσου λόγοι θείοι και άγαθοί τυγχάνουσιν. Έαν δε ίδη, ότι έδωκεν έτέρω ἀπὸ τῶν καρπῶν, οὑτος διδάσκαλος γενήσεται, καθ΄ ὅσον ἐμέρισε καὶ μετέδωκεν. ἐὰν δὲ ἴδη, ὅτι τοὺς καρποὺς μόνον ἔλαβεν, οὐκ ἔφαγε δὲ οὐδὲ μετέδωκεν έτέρω, τα μεν θεία ούτος έδέξατο και ούκ έγνω, ου τελεσφορήσει δέ. εί δ ουκ έφαγε μεν , έδωκε δέ έτέρω, ουτος ώς κενόδοξος κατακριθήσεται, ἐκεῖνος δὲ ὁ λαβών σωθήσεται. ἐὰν δὲ ἴδη ἑαυτὸν εἰσελθόντα ἐν τῷ παραδείσφ και αποθανόντα, ό θάνατος της πίστεως αύτου ζημίαν σημαίνει.
#### 8. From the Indians on Paradise
If a good person dreams that he entered Paradise, he is saved, because this is good news from God on account of his good deeds. If the dreamer is an evil-doer, he will repent and be saved. This dream also signifies the acquisition of worldly wealth. If someone dreams that he ate from the fruits of the trees, he will find in his faith wisdom and knowledge, because the fruits of Paradise are divine and holy words. If he sees that he gave of the fruits to others, he will be a teacher to the extent that he divided and shared <them>. If he sees that he only received the fruit but did not eat or share it with another, he has received God's gifts and did not know it, and therefore he will not bring the fruit to perfection. If he did not eat the fruit, but gave it to another, he will be judged vainglorious, while the recipient of the fruit will be saved. If he dreams that he entered into Paradise and died there. his death signifies the loss of his faith.
Drexl supposed that the interpretation of dreaming of the fruit of Paradise was inspired by two scriptural passages. However, these passages only loosely correspond to the meaning of the interpretation that occurs in the Oneirocriticon, which does not include a verbatim quotation of any phrase or word from the New Testament.
Paradise and Hell, like Judgment Day, are part of Islamic eschatology. Paradise and its delights are described in detail in Our'an 76:12-22, 56: 12-39,
<sup>8</sup> John 14:17: Το πγευμα τῆς άληθείας, δ ό κόσμος οὐ δύναται λαβεῖν, ὅτι οὐ θεωρεί αὐτό, ούδὲ γινώσκει αὐτό· ύμεῖς δὲ γινώσκετε αὐτὸ, ὅτι παρ` ὑμῖν μένει, καὶ ἐν ὑμῖν ëσται (Even the spirit of truth; whom the world cannot receive, because it seeth him not, neither knoweth him: but ye know him; for he dwelleth in you and shall be in you). Also I John 4:6: juzic ἐκ τοῦ Θεοῦ ἐσμέν ὁ γινώσκων τὸν Θεὸν ἀκούει ήμων δς οὐκ ἐστιν ἐκ τοῦ Θεοῦ, οὐκ ακούει ήμων: ἐκ τούτου γινώσκομεν τὸ πνεῦμα τῆς ἀληθείας καὶ τὸ πλάνης (We are of God: he that knoweth God heareth us: he that is not of God heareth not us. Hereby know we the spirit of truth, and the spirit of error).
55: 54-56. Arabic dreambooks use these descriptions to interpret dreams about it. The Arabic interpretations of Paradise in chapter 8 of the Oneirocriticon are all copied from Arabic sources. Ibn Qutayba wrote: 9
فان راي انه دخل الجنة فذلك بشرى من الله بالخير فان اصاب شيئُا من ثمارها واكل ذلك خير يناله فى دينه ودنياه وعلم وير وكذلك اذ ولجها قـال الله عز وجل ادخلوها بسلام آمنين
If <the dreamer> dreams that he entered Paradise, this is a glad tiding from God regarding the good things <that the dreamer will receive>. And if he took any of its fruit and ate it, goodness will extend to him in his religious and worldly pursuits, as well as knowledge and righteousness. And if <the dreamer> enters <Paradise> God Almighty said: "Enter them in peace, secure" [Qur'an 15:46].
A similar interpretation is given by al-Dīnawarī:10
فان راي انه يدخلها واکل من تصارها رزق علما بقدر ما اکل منها لان ٿمار الجنة هو العلم به فان التقطها او اطعمها غيره فانه يفيد غيره من علمه وينتقع به المتعلم ولا يستعمل هو ولا ينتفع به فان راي انه منع ثمار ها فانه فاسد الدين مشرك لقول الله تعالى إنّه من يُشرك بالله فقد حرم الله عليه الحنة.
And if he dreams that he entered <Paradise> and ate from its fruit, he will be endowed with knowledge commensurate with what he ate, because the fruit of Paradise is knowledge. And if he collects them or feeds them to other people, he will teach others from his knowledge and will benefit his apprentices, but he himself will not use it, nor will he benefit from it. And if he is refused the fruit of Paradise, he is a corrupter of religion and a polytheist, according to the saying of God Almighty, "Lo! whoso ascribeth partners unto God, for him God hath forbidden Paradise" [Qur'an 5:72].
Al-Muntakhab interprets dreaming of Paradise likewise:''
قال الاستاذ ابو سعد رحمه الله من راى الجنة ولم ير دخولها فاين رؤياه بشارة له بخير عمل او يهم بعمله… وقيل من راى الجنة عيانا نال ما اشتهى وكشف عنه همه... فان راى كانه ادخل الجنة فقد قرب اجله وموته وقيل إن صاحب الرؤيا يتعظ ويتوب من الذنوب... فان راي انه قيل له انك تدخل الجنة فإنه ينال ميراثا لقوله تعالى وتلك الجنة التى اورثتموها الآية ... فان
<sup>9</sup> Bah 2, Yahuda ar. 196, fol. 25a.
<sup>10</sup> Esad Efendi 1833, fasl 30, hāh 10, fol. 263b.
<sup>11</sup> Al-Muntakhab, pp. 70-71.
رای کانه فی ریاضیها رزق الاخلاص وکمال الدین فان رای کانه اکل من ثمار ها رزق علما بقدر ما اكل وكذلك إن راى انه شرب من مائها وخمرها ولبنها نال حكمة وعلما وغني ... فــان راى انه التـقط ثمار الجنة واطعمها غيره فانه يفيد غيره علما يعمل به وينتفع ولا يستعمله هو ولا ينتقم به …
Master Abu Sa'd, may God have mercy on him, said: "If someone dreams of Paradise, but did not see its entrance, this dream is a glad tiding for him pertaining to a good deed he has performed or <a good deed> he is preoccupied about performing." ... It is said that if someone sees Paradise with his own eyes, he will obtain what he desires and will be relieved from his worries ... If someone dreams that he was made to enter Paradise, his end and death is imminent, and it is said that the dreamer will heed the warning and repent for his sins ... If someone dreams that he was told, "You will enter Paradise," he will receive an inheritance according to the saying of God Almighty, "This is the Paradise that you have been given for an inheritance" [Qur'an 43:72].12 ... If someone dreams that he was in the meadows of Paradise, he will be granted faithfulness and perfection in his religion. If he dreams that he ate from the fruit of Paradise, he will be endowed with knowledge commensurate with what he ate. Likewise, if he dreams that he drank from the water, wine and milk of Paradise, he will obtain wisdom and knowledge and wealth ... If he dreams that he collected the fruits of Paradise and fed them to other people, he will avail others of the knowledge in effect and will benefit <them>, but he himself will not use it, nor will he benefit from it. 13
Ibn Shahin's dreambook repeats similar interpretations of Paradise and its fruit (nos. 213. 215. 223. 228):
(٢١٢) ومن راى انه دخل الجنة فإنه يحصل له فرح وسرور وبشارة من الله تعالى بالخيرات، وقيل امن لقوله تعالى "ادخلوها بسلام امنين". (۲۱۵) ومن راى انه تذاول فاكهة بيده واكل فإنه يتعلم علم الدين ويحصل سيرة المتقين ولا يستفاد منه. (۲۲۳) ومن راي انه قيل له تدخل الجنة يحصل له ميراث. (۲۲۸) و من راي انه قد ناول احدا من فواكه الجنة فإنه يستفيد من علمه.
(213) If someone dreams that he entered Paradise, he will be granted joy, happiness and glad tidings from God Almighty because of <his> good deeds, and it is said that he will be safe, according to the saying of God Almighty "Enter them in peace, secure" [Qur'an 15:46]. (215) If someone dreams that he had the fruits <of Paradise> in his hands and ate from them, he will study the science of religion and will lead a pious life, but will not benefit from it. (223) If someone dreams
<sup>12</sup> Cf. Drexl 4, 19-20: άλλά και πλούτου κοσμικού τουτο έπιτυχίαν σημαίνει (This dream also signifies the acquisition of worldly wealth).
<sup>17</sup> The same interpretations are repeated verbatim in al-Nābulusī, vol. 1, pp. 99-102, s.v. جنة
#### CHAPTER SEVEN
that he was told, "You will enter Paradise," he will receive an inheritance. (228) If someone dreams that he took one of the fruits of Paradise, he will derive profit from his knowledge.
The next chapter of the Oneirocriticon discusses dreaming of the fire of Hell. The Christian interpretation of Hell set forth in this chapter is the following: 41
θ 'Eκ τών Ινδών περί πυρός γεένης
Εάν τις ΐδη, ότι κατεδικάσθη είς πύρ γεένης, ούτος είς απώλειαν τελείαν παραδοθήσεται ώς κακουργος μέγιστος προβλέπων τον ίδιον όλεθρον. πλην και τουτο έκ του θεου το μήνυμα, ίνα μεταγνώ και σώση τὴν ίδίαν ψυχήν. ἐαν δε το μεν πυρ της γεένης έθεάσατο μόνον, ου κατεδικάσθη δε έν αύτώ, είς κοσμικάς περιπεσείται θλίψεις άμαρτάνων έν αύταις. δσον δε έγγίζον ή μακρύνον είδε το πυρ τῆς γεένης, τοσούτον ἐγγίζουσαν ἢ μακρύνουσαν ἐλπιζέτω τὴν θλῖψιν. ἐὰν δὲ τοῦτο θεωρήση βασιλεύς, παρακρινεῖ τὰς κρίσεις αὐτοῦ και άπο θεού τιμωρηθήσεται. του άνδρός, μοιχευομένη και άσελγαίνουσα. ἐαν δὲ τοῦτο θεωρήση παιδίον, είς πρόσωπον των ιδίων γονέων τουτο έώρακεν.
9. From the Indians on the Fire of Hell
If someone dreams that he has been condemned to the fire of Hell, he will deliver himself to total perdition, for he is a great evil doer foreseeing his own destruction, though this is also a message from God, so that he can repent and save his soul. If he only sees the fire of Hell and is not condemned to it, he will be afflicted by worldly sorrows and will sin because of them. And however close or far away he saw the fire of Hell is how soon or late he can expect his sorrow. If a king sees that, he will err in his judgments and will be punished by God. If a woman sees it. she will plot against her husband by committing adultery and acting lasciviously. If a child sees this, the dream refers to his parents.
The interpretation of the fire of Hell as a warning for the dreamer to repent and as an indication that he will perpetrate grave sins, be afflicted by worldly sorrows, and commit adultery can also be found in the dreambook of al-Dīnawarī:15
من راي انه دخل جهنم فانه يرتكب الكبائر فان خرج منها من غير مكروه اصابه وقع في غموم الدنيا ومحتة سلطان لا ينجو منها لقول الله تعالى وراى المجرمون النار فظنوا
<sup>14</sup> Drex1 5. 4-12.
<sup>15</sup> Faș! 30, băb 8; Esad Efendi 1833, fol. 274a. Ibn Qutayba, băb 2, Yahuda ar. 196, fol. 25a, رؤ ية جهنم فى ) laconically states that "the interpretation of Hell is opposite to that of Paradise ﺍﻟﺘﺎﻭﻳﻞ ﺿﺪ ﺭﺅﻳﺔ ﺍﻟﺠﻨﺔ).
انهم مواقعوها ولم يجدوا عنها مصرفا واصابه غرامة وخسران فاحش لقوله تعالى ان عذابها كان غراما راى انه دخلها فانه ياتى الذنوب الكبائر والفواحش التى اوجب الله تعالى عليه بها الحد وينسى ربه وياتى الاثم والبــغى بـفـيـر الدق فليــتق الله سبحانه... فان راى انه ادخل النار فانه يخويه الذي ادخله النار ويحرضه على ار تكاب ذنب عظيم مثل قتل او زناء فليتق الله جل ذكره.
Whoever dreamt that he entered Hell will perpetrate atrocious sins. And if he came out of it unharmed, he will fall into worldly sorrows. And if someone dreamt the fire of Hell nearby, he will fall into distress and punishment from a king, from which there is no escape, according to the saying of God Almighty: "And the guilty behold the Fire and know that they are about to fall therein, and they find no way of escape thence" [Qur'an 18:53]. And a penalty and atrocious damage will befall him, according to the saying of God Almighty: "Lo! The doom thereof is anguish" [Qur'an 25:65]. And this is a warning to him so that he may repent and turn away from <the sin> that he is in. And if he saw that he entered Hell, indeed he is perpetrating the gravest sins and vile deeds, for which God Almighty imposes legal punishment, and <indeed, the dreamer> is forgetting his Lord (6 and is committing crimes and unjust wrongdoings, and let him be afraid of God, praise be to Him ... And if he saw that he was made to enter Hell, indeed, the one who made him enter Hell will misguide him and goad him to commit a grave sin, such as murder or illegal sex, so let him be afraid of God, whose invocation is exalted.17
Similar interpretations, the phrasing of which is very close to that of al-Dīnawarī, are also recorded in chapter 18 of al-Muntakhab:18
فان راى النار من قريب فانه يقع في شدة ومحنـة لا ينجـو منها لقول الله تعالى وراى الجرمون النار فظنوا انهم مواقعوها ولم يجدوا عنها مصرفا واصابه خسران فاحش لقوله عز وجل ان عذابها كان غراما وكانت رؤياه نذيرا له ليتوب من ذنب هو فيه فان راي كانه دخل جهنم فانه يرتكب الفواحش والكبائر الموجبة للحد وقيل انه يقبض بين الناس فان راى كانه ادخل النار فان الذي ادخله النار يضله ويحمله على ارتكاب فاحشة فان راى انه خرج منها من غير إصابة مكروه وقع فى غموم الدنيا.
<sup>16</sup> The text of al-Dinawari in Esad Efendi 1833 gives the reading بنسي رايه (he is forgetting his father), where "father" could be understood either as a parent, or as God the Father. The same interpretation is repeated almost verbatim in the dreambook of al-Nabulusi, where the text reads رينسى (he is forgetting his Lord). I prefer the reading according to al-Nābulusī because it fits the context better.
<sup>17</sup> Repeated almost verbatim in al-Nabulusī, vol. 1, p. 102, s.v. جنبع.
<sup>18</sup> Al-Muntakhab, p. 69.
#### 264 CHAPTER SEVEN
If he sees Hell nearby he will fall into distress and tribulation from which there is no escape, according to the saying of God Almighty: "And the guilty behold the Fire and know that they are about to fall therein, and they find no way of escape thence" [Qur'an 18:53]. And atrocious damage will befall him, according to the saying of God Almighty: "Lo! The doom thereof is anguish" [Qur'an 25:65]. His dream is a warning for him so that he may foreswear the sin that he is into. If he sees that he entered Hell, he will commit adultery and grave sins that call for legal punishment, and it is said that he will be apprehended in front of the people. If he sees that he was made to enter Hell, the one who made him enter will cause him to go astray and prompt him to commit vile deeds. If he sees that he has exited from Hell unharmed, he will fall into worldly sorrows.
lbn Shahin discusses the interpretation of the fire of Hell in connection with the ~irat, the bridge across the infernal fire. Falling from the bridge into the fire of Hell is interpreted as worldly misfortunes. For the mighty, it signifies that they will oppress the people under their power, which is similar to the interpretation of Hell dreamt by a king that is set forth in the *Oneirocriticon.* Falling into the fire of Hell also signifies sinning with the possibility of pardon (nos. 201 and 206):
~J "~J 4.....:...::... ~ *ti=* \_;L.:..J I ~ .b I~ I ..:,,.... .b..L... <GI lSI J .J-oJ ('L \) dill .J--o ').l.........c. .:i....,;.l:: JL:JI ~ .bl~I..:,,.... ~J <WI lSIJ .J-oJ ('Li)-~ <sup>~</sup>*Jj* <GI lSIJ .J--oJ ('1'. i) -~~ '-:-'..,\_:,jJ ~ <sup>F</sup><l:!..1-::' ~ 0~J ... ~ \_,.iil I d.J <sup>~</sup>*>-! fl* J 0-:> .J I ~ liii.J l,. J J.J..:i q . ..Q ~ J-'b J .b I~ <sup>I</sup>
(201) If someone dreamt that he fell from the bridge into the fire <of Hell>, he will fall into trial and tribulation and great misfortune. (203) If someone dreamt that he fell from the bridge into the fire <of Hell>, he will receive an appointment from the king and many oppressions and crimes will take place because of his intervention. (203) If someone dreamt that he was tearful and slipped from the bridge, it indicates negligence in his <observance of> religion; however, pardon is anticipated for him.
#### *Angels and Eunuchs*
Chapter 10 of the *Oneirocriticon* sets forth the Christian interpretation of dreams of angels: <sup>19</sup>
<sup>19</sup> Drexl 5, 18-6, 14 (Chap. 10, "From the Indians on Angels").
Εάν τις ϊδη κατά τους ύπνους τινά των αγγέλων των ώνομασμένων και άπογεγραμμένων έν ταις θείαις βίβλοις, εὐαγγελικὴν εύρήσει χαράν. εἰ γὰρ παράκειται πόλεμος έν τώ τόπω, νίκη ταχεία και ίσχυρά κατά των έχθρών έσται. εί δὲ θλῖψίς ἐστι, μεταβάλλεται εἰς χαράν εἰ δὲ νόσος, ἐπανελεύσεται είς ύγείαν: εί δε πενία, είς πλούτον μετατραπήσεται. ἐαν γυνὴ ἔγκυος ϊδη τουτο, τέξει αρσεν. εί δε βασιλεύς ίδη τουτο και ότι συνωμίλησε τω άγγέλω, παν, όπερ ήκουσε παρά του άγγέλου, αμεταβλήτως αποβήσεται, είτε άγαθον είτε φαύλον. ἐαν δὲ μόνον τοῦτον ῖδη βασιλεύς, οὐ συλλαλήση δὲ τῶ ἀγγέλω, νίκην κατά των έχθρών αύτου σημαίνει και ό λαος αύτου αύξηθήσεται και οί πτωχοί αύτου χορτασθήσονται.
Εί δέ τις όμοίως Ίδη αγγελον μέλλων ύπεραθλείν του θείου όνόματος, ούτος γιγνωσκέτω, ότι σεσωσμένος έστίν. ἐαν δὲ ἴδη τις ἄγγελον τῶν ἀνωνύμων καὶ μὴ ἀπογεγραμμένων, ἔσται αὐτφ χαρά τε καὶ ἀγαλλίασις, μετριωτέρα δέ.
If someone dreams of one of the angels who are listed and named in the Holy Books, he will receive good news of joy. For if a war is being waged in the land, victory over the enemy will be quick and strong. If there is sorrow, it will be changed to happiness. If illness, it will turn into health; if poverty, wealth. If a pregnant woman sees this <dream>, she will give birth to a male. If an emperon sees this, and <dreams> that he talked with the angel, whatever he heard from the angel, whether good or bad, will be fulfilled exactly. If the emperor only saw the angel and did not speak to him, it signifies victory against his enemies, and his people will increase and the needy in his realm will be fed.
If someone is about to suffer on behalf of the Holy Name <i.e., a future martyr> dreams of an angel, let him know that he is saved. If someone dreams of an angel that is not named or listed <in the Holy Books>, he will also have joy and exultation, but less <than in the aforementioned interpretations>.
Angels are central to the religious beliefs of both Christians and Muslims. Arabic dreambooks contain elaborate interpretations of angels, often naming the most important among them and giving a separate interpretation for each. The interpretation given by Ibn Qutayba is similar to that of the Oneirocriticon: 30
ومن راي الملائكة تنزل بمكان فان ذلك نصر لاهل ذلك المكان وفرح من كرب لانها نزلت بنصر الانبياء والتفرج عنهم فجعل نزولهم منالا لذلك ومن راء انها تکلمه بکلام من البر او تعظه او تبشره او تصله او يطير معها او تذهب به فانها شهادة يُرزقها وشرف فى الدنيا وصيت.
Whoever sees the angels descend upon a place, this is indeed victory for the people of that place, as well as joy after sorrow. For angels descend in order to support and watch over the prophets, and their descent <is> a gift to that end. Whoever sees that <the angels> spoke to him in words of righteousness or that
<sup>20</sup> Ibn Qutayba, bah 3, Yahuda ar. 196, fol. 26a.
they admonished him or gave him glad tidings or joined him or that he flew with them or that they carried him off, he will be endowed with martyrdom and honor in this world, as well as repute.
Al-Dinawari gives a more detailed interpretation of angels, as he dedicates a separate chapter to each of the most important ones. In the chapter on Gabriel we read:21
من راى جببريل عليه السلام فى منامه او واحد من الملائكة المعروفين الاشراف الذين سماهم الله تعالى مستبشرا به يكلمه بكلام بر وموعظة او يوصيبه او يبشره فانه ينال شرفا وعزا وقوة وظفرا وبشارة وان كان مظلوما نصر او مريضا شفى او خايفا امن او فى نمم فرج عنه او ضرورة حج وهو دليل على شهادة يرزقها وان عاش طويلا.
If someone dreamt that Gabriel, may peace be upon him, or one of the other known and distinguished angels to whom God gave a name, was cheerful towards <the dreamer> and spoke to him in words of righteousness and exhortation, or admonished him, or gave him glad tidings, <the dreamer> will obtain distinction, might, power, victory, and glad tidings. If <the dreamer> was wronged, he will be assisted; if he was sick, he will be healed; if he was in fear, he will be safe; if he was in distress, he will be relieved from it; and if he is in need, he will perform the pilgrimage to Mecca. It also signifies martyrdom that will be granted <to the dreamer>. even if he lives long.
In al-Dinawari's interpretation of the angels that carry the throne of God, there are further resemblances to the interpretation of angels found in the من راى حملة العرش علييهم السلام فى حرب او اشراف Oneirocriticon: 2 If someone الملائكة كانت رؤيتهم دليل الظفر على العدو ولاغنى بعد الفقر dreamt of the angels that carry the throne of God, may peace be upon them, and he was at war, or <dreamt> of the distinguished among the angels, this signifies victory over the enemy and wealth after poverty). The dream of angels in connection with the birth of a male is mentioned in al-Dinawari's chapter on the common angels:23
فان راى ان الملائكة يپشرونه بغلام مولود رزق ابنا عالما طاهرا تقيا صالحا تقتدى به لقوله تعالى انما انا رسول ربك لاهب لك غلاما زكيا ولقوله ان الله يبشرك بكلمةمنه اسمه المسيح عيسى ابن مريموجيها فى الدنيا والاخرة ومن المقربين.
<sup>21</sup> Fașl 3, hāh 1; Esad Efendi 1833, fol. 40a; BN arahe 2745, fol. 57b.
<sup>22</sup> Fasl 3, hāh 9; Esad Efendi 1833, fol. 40b; BN arahe 2745, fol. 59a.
<sup>23</sup> Fasl 3, hāb 10; Esad Efendi 1833, fol. 40b; BN arahe 2745, fol. 59a-b.
If someone dreamt that the angels gave him the glad tidings of a son that will be born, he will be given a wise, pure, pious, righteous and exemplary son, according to the saying of God Almighty: "I am only a messenger of thy Lord, that I may bestow on thee a faultless son" [Qur'an 19:194] and according to His saying "God giveth thee glad tidings of a word from him, whose name is the Messiah, Jesus, son of Mary, illustrious in the world and in the Hereafter, and one of those brought near <unto God>" [Qur'an 3:45].
The interpretations of the angels recorded in al-Muntakhab are similar to those of al-Dinawari, but the wording of al-Muntakhab is occasionally closer to that of the Oneirocriticon:25
قـال ابو ســعـد رضـى الله عنـه رؤيـة الملائكـة فى النوم اذا كـــانوا مــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ مستجشرين تدل على ظهور شيئ لصاحب الرؤيا وعز وقوة وبشارة ونصر بعد الظلم او شفاء بعد المرض او امن بعد خوف او يسر بعد عسر او غنى بعد فقر او فرج بعد شدة وتقتضى ان يحج صاحبها او يغزو فيستشهد .
Master Abu Sa'd, may God be pleased with him, said: "The dream of angels in sleep, when they are known and cheerful, indicates the appearance of something to the dreamer, power and strength and good news and victory after tyranny or health after sickness or safety after fear or ease after difficulty, prosperity after poverty or joy after distress. And it is necessary that the dreamer perform the pilgrimage to Mecca or that he go to war and die a martyr."
The Oneirocriticon states that whatever the angels announce in a dream will be fulfilled exactly as it is foretold. A similar statement is made by Ibn Shāhīn (no. 72): وإن راى ملكا واخبره بأمر فيكون كذلك (If <someone> dreams that an angel informed him of something, it will happen exactly so). The truthfulness accorded to angelic utterances in dreams both in the Oneirocriticon and in Arabic dream interpretation might have originated in the Arabic translation of Artemidoros. The Greek text of Artemidoros repeatedly states that the most truthful utterances in dreams are those of the gods: τών δε άξιοπίστων λεγομένων, οις λέγουσί τι [κατ' όναρ] πιστεύειν χρὴ και πείθεσθαι, φημί πρώτους είναι θεούς άλλότριον γαρ θεού το ψεύδεσθαι (Among those who are worthy of credence and whose words [in a dream] one must believe and obey, I maintain that the gods are first. For it is contrary to the
<sup>24</sup> These verses of the Qur'an refer to the Annunciation. Sūra 19 of the Qur'an, also called "the Chapter of Mary," gives the longest version of the Annunciation and birth of Jesus, though briefer references to these events can also be found elsewhere in the Qur'an. See G. Parrinder, Jesus in the Qur'an (Oxford, 1995), pp. 67-74.
<sup>25</sup> Al-Muntakhab, p. 35.
nature of a god to tell lies);26 ούτε οί θεοί ούτε αλλοι άξιόπιστοι ψεύδονται, άλλ' οια αν λέγωσιν άληθη λέγουσιν (Neither the gods nor anyone else worthy of credence tells lies and everything they say is the truth).27 These الـنـين يـنـبــنــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ قــولـهم فــى الـرؤيـا ونصـــدقــهم اقــول انـهم او لا الــلائـكـة، وذلك ان الملائكـة لا ريكذيو (As for those whose words we ought to obey and deem as credible in a dream, I say that <these are> first of all the angels, because the angels do not lie); و الملائكة يستبههون بـالأبـاء والمـوالى وكل مـن هو اهـل لان يصـدق The angels are) قوله فانهم لا يكذبون فى الرؤيا وكلما يقولونه فهو حق like parents, sovereigns and anyone else who deserves his word to be believed, because their words are truthful, since they do not lie in a dream and whatever they say is accurate).23 The truthfulness accorded to angelic utterances might have been an idea that originated in Arabic dream interpretation without outside influence, but such a statement is absent from Ibn Qutayba, the Arabic dreambook most likely to have ignored Artemidoros. The insistence of the Oneirocriticon on the truthfulness of angelic utterances could therefore be counted as an additional indication that the author of the Oneirocriticon was using an Arabic text that had been influenced by the Arabic translation of Artemidoros.
The Christian chapter on angels in the Oneirocriticon concludes with a paragraph that interprets dreaming of eunuchs, who are said to be equivalent to angels because they are sexless:30
έαν δέ τις ϊδη εύνουχον άγνώριστον εύειδη εὐήλικον, καὶ οῦτος ὡς ἀγγελος λογιζέσθω δια το καθαρόν και άγγελοπρεπες και άνεπίδεκτον της σαρκικής ἐπιθυμίας καὶ ὁμοίως τὸ ἀγγέλφ ἀποβήσεται αὐτῷ ἀλλὰ καὶ ὅγγελίσηται αυτώ, άποβήσεται ταχέως. όμοίως δὲ εῖ τις ϊδη τινὰ τῶν γνωρίμων εὐνούχων και μεγιστάνων και περιβλέπτων έν άξιώματι, όμοίως, καθά είρηται περι των αγγέλων, άποβήσεται αύτω.
30 Drexl 6, 7-14. The likening of eunuchs to angels is widespread in Byzantine texts. See K. Ringrose, "Living in the Shadows: Eunuchs and Gender in Byzantium," in Third Sex, Third Gender: Beyond Sexual Dimorphism in Culture and History, ed. G. Herdt, (New York, 1994), pp. 95, 97, 100. For eunuchs in Islam, see D. Ayalon, "On the Eunuchs in Islam," Jerusalem Studies in Arabic and Islam 1 (1979), 69-124.
<sup>26</sup> ii.69: Pack 195. 3-5.
<sup>27</sup> jv.71; Pack 292, 4-6.
<sup>28</sup> Fahd, Artémidore d'Éphèse, p. 353, 3-4.
<sup>29</sup> Book iv is missing from the one surviving manuscript with the Arabic translation of Artemidoros, but this portion of Hunayn's text is quoted in al-Dinawari, fas! 3, hab 11; Esad Efendi 1833, fol. 41b, 11. 8-16; BN arabe 2745, fol. 51a, 11. 12-14.
If someone dreams of a handsome eunuch of good stature whom he does not know, let the eunuch be reckoned as an angel, because of his purity and angel-like quality and his insusceptibility to carnal lusts; thus, the result of the dream will be the same as in the case of an angel. And whatever the eunuch should say to the dreamer will soon take place. Likewise, if someone dreams of one of the well known and very powerful eunuchs who are looked upon with honor, the results of the dream will be the same for him as for the angels.
The same interpretation, supported by a similar argument, is given in Arabic dreambooks. According to al-Dinawari,31 ..U ~I *,:\_;'i* ~)l..11 ~,:,;~I ~~I~ ~\_µ(Eunuchs are angels because castration deprived them of carnal passion). Likewise in *al-Muntakhab: <sup>32</sup>*. i.!.ll...a ~I rJL..:J I .J (The castrated servant is an angel). Al-Nabulusi goes to even greater detail: <sup>33</sup>
4.S:.~l .J..o dL ~~I i~.J ~~I~ 4.J '(~ ~ .slJ .:,La dL ~'11 ~l.J ... ~.J-'11 ~ LI.JJ-LA ~I .:,LS .Jl.J ~.JI J~ . r-i-.J ~ .,;.w1.J ;;J~ J.J'lu ...,...1k..l1 dL ~1.J ..1\_,\_.....'11.J ~.)I
If someone dreams of an unknown eunuch who has the manner of the righteous and speaks in words of wisdom, he is an angel who warns or gives glad tidings. And if the eunuch is known, it is none other than he .... The white eunuch is the angel of mercy, while the black or Ethiopian eunuch is the angel of punishment. The first one represents glad tidings, while the second one sorrow and distress.
#### *Prophets, Apostles and Teachers; Jesus and Icons*
Chapter 11 of the *Oneirocriticon* is called "On Prophets, Apostles and Teachers" (Drexl 6, 17-21), a sequence of holy figures taken from the New Testament,34 but only its first paragraph deals with that topic. The second paragraph (Drexl 6, 22-7, 5) interprets dreams about Jesus Christ, and the rest discusses dreaming of being ordained a patriarch, a priest, or a deacon, of reading a book or the Gospels to the people, and of being tonsured or dressed as a monk (Drexl 7, 13-26). The first two paragraphs are also related to chapter 150, "On Icons" (Drexl 105, 12-107, 2).
<sup>31</sup>Al-Dinawari,fa~/9, *hiih* IO; *Esad Efendi* 1833, fol. 90a.
<sup>32</sup>*Al-Muntakhah,* chap. 30, p. 149.
<sup>33</sup> Al-Nabulusi, vol. I, p. 192, s.v. ~
<sup>34</sup> See chapter 4, n. 62.
#### 270 CHAPIERSEVEN
According to chapter 11, dreaming of prophets, apostles, teachers, or martyrs has the same meaning as dreaming of angels, but the fulfillment of these dreams is somewhat less auspicious. Dreaming of Jesus signifies salvation, worldly wealth, and victory. Whatever the dreamer hears Christ say in a dream will be fulfilled exactly, and he who beholds Christ in a dream is blessed. Seeing Christ in an unknown house and entering it to join him signifies the dreamer's imminent death and salvation, along with the attainment of high rank and wealth by the dreamer's offspring. Dreaming of the apostles or patriarchs signifies something similar but less auspicious.
The first half of chapter 150 ("On Holy Icons") also interprets dreaming of Jesus or one of the apostles, prophets, and saints (Drexl 105, 12-106, 4 ). It, too, says that dreaming of Christ signifies victory, that whatever is said by Jesus or one of the other holy figures in a dream will be fulfilled to the last word, that dreaming of them also foretells worldly and spiritual well-being, that dreaming of one of the holy figures is less auspicious than dreaming of Christ himself, and finally that Christ is dreamt about only by the very pious or the very sinful so that they may repent and be saved.
Only the second half of chapter 150 is true to its title and discusses icons (Drexl 106, 5-107, 2). Its interpretations also repeat material found in other parts of the *Oneirocriticon.* Dreaming of an icon depicting Christ's crucifixion predictably signifies victory, since both Christ and the cross have already been associated with victory in chapters 11 and 126 ("On Kings and Crosses"). An icon of Christ without the cross also signifies victory, but is less auspicious. An icon of the apostles, prophets, or saints signifies a lesser victory for an emperor and well-being for a commoner. If the icon talks to the dreamer, whatever the icon says will be fulfilled. An icon with gold revetments signifies sorrow because of the gold, the negative of which is mentioned in several chapters of the *Oneirocriticon. <sup>35</sup>*Dreaming of having an icon made signifies success commensurate with the faithfulness of the depiction on the icon.
In a position corresponding to chapter 11 in the *Oneirocriticon,* Arabic dreambooks interpret dreams about various holy figures, beginning with the prophets, among them Jesus, son of Mary ('Isa b. Maryam). According to the Qur'an, Jesus was one of the prophets who appeared before Mui)ammad and
<sup>35</sup> Chap. 12, "From the Indians on Various Faiths," Drexl 9, 1-3; chap. 240, "From the Indians, Persians and Egyptians on All Kinds of Sheep, Deer and Goats," Drexl 193, 20; chap. 255, "From the Indians, Persians and Egyptians on Golden, Silver and Bronze Coins," Drexl 208, 19-209, 22; chap. 256, "From the Same Sources, Regarding Ornaments of Gold, Silver, Precious Stones and Pearls," Drexl 210, 4-5.
performed various miracles, such as raising the dead and calling down from heaven a table laden with sustenance, that provided proof of his prophethood. The Arabic interpretations of Jesus are conditioned by his image in the Qur'an and do not coincide with the interpretations given in the Oneirocriticon; but neither do the Christian interpretations of Jesus quoted by al-Dinawari coincide with those of the Oneirocriticon, as is evident from al-Dinawari's relevant chapter:36
قال المسلمون من راى عيسى بن مريم فانه يكون رجلا مباركا نفاعا كثير الخير كثير الشرف فى رضا الله تعالى كثير البر والذير صاحب نسك وپرضي بالقليل ويرزق نصرا بالطب لقوله تعالى وابرىء الاكمه والابرص وأحيي الموتى باذن الله. وقالت النصارى من راه فى منامه فانه لا يصيبه مكروه فى تلك السنة وان طلب طلبا اصابه ومهر فيه فان راى امه مريم بنت عمران فانه ينال جاها ومرتبة من الناس ويظفر بجميع حوائجه وان راي امراة هذه الرؤيا وهى حامل ولدت ابنا حكيما ذلك واظهر الله براءتها \* ومن رايكانه يسجد لمريماو عيسى فانه يكلم الملك ويجلس معه.
The Muslims said: "If someone dreams of Jesus, son of Mary, he will be a blessed and useful man, <who will enjoy> many blessings and honors by the favor of God Almighty, and will be very righteous and outstanding, will lead an ascetic life and will be pleased with little, and will be endowed with triumph in medicine, in accord with the saying of God Almighty: 'I heal him who was born blind, and the leper, and I raise the dead, by God's leave' [Qur'an 3: 49]. The Christians [Nasārā] said: "Whoever sees him in his dream, nothing terrible will befall him throughout that year, and if he seeks something, he will attain it and will be skillful in it. If someone dreams of his mother, Maryam, the daughter of 'Imran, he will gain rank and dignity from the people and will be successful in everything he wishes. If a pregnant woman has this dream, she will give birth to a wise son. And if calumnies are spoken against her she will be proven guiltless and God will reveal her innocence. Whoever sees himself prostrate in front of Maryam or Jesus will talk to the king and will sit beside him."37
<sup>36</sup> Al-Dīnawarī, fasl 2, bāb 25, Esad Efendi 1833, fol. 36b; BN arabe 2745, fol. 53a-b.
<sup>37</sup> This last interpretation, attributed to the Christians, seems to reflect an awareness of Byzantine imperial ideology, which drew an analogy between the emperor and Christ. One could also see in it a reflection of the ceremonial genuflexion (proskynesis) before the emperor that was customary at the Byzantine court, though the same practrice was followed in the caliphal court as well (see chapter 1, nn. 220 and 221). These interpretations suggest that the interpretations al-Dinawari attributes to the Christians were based on authentic Christian dreamlore, whether oral or written: it does not coincide with the material found in the Oneirocriticon, however, which does not mention dreaming of prostrating oneself in front of either Mary or Jesus.
A similar interpretation of Jesus is quoted in *al-Muntakhab: <sup>38</sup>*
~ ..:.IJL\_,...... tU.:. ~J <Lil~ b~)J ..::.JJ i'.)L...JI ~ ~ "51J :r-J ..;:.,IJ .:,) µJ ····i\_,\_L.JI *:ro* d.Jj ~J ~IF i~J ~I~ ~I .~~I ..;:.,..UJ J-.~ ~J i)l.....JI ~ r;..ro .J-7 ~ bl..rol
If someone dreams of Jesus, the son of Mary, his dream signifies that he is a useful and blessed man, who will be surrounded by blessings and travel extensively. He will be distinguished for his medical expertise or some other kind of knowledge .... It is said that if a pregnant woman dreams of Jesus, the son of Mary, may peace be upon him, she will give birth to a wise son.
And also in lbn Shahin (no. 334):
<L.J ~J ..;:.,lcWI ~ "5~J ~I <L.Jli....!.I ~ <t..:iµ ~ "51J :r-J "5µ1J J.Aj.JIJ bJL\_,...JI J.JJ-:' ~ "51J *:ro* µJ ,..;:.,l~I JUJ ~\_,.:;JI .:,~ *'i* ~ .\_,.J..JI ~ j.)J-:> ~JJ ...J~ L.... ~J bJU.........1 ..;:.,~ ~JJ . ....c...... 4..:iL..J ~
If someone dreams of Jesus, he will revive his neglected39 activities, will be strong in his pious deeds and will be granted success in carrying out good works. It is said that if someone dreams of Jesus he will granted devotion, abstinence and godliness. Possibly his travels will multiply and he will be safe from his fears. It is also possible that he will be endowed with knowledge of medicine to such a degree that there will be no one like him in his time.
The interpretations of Jesus given by al-Nabulusi are identical to those quoted above. <sup>40</sup>
Icons in Arabic dreambooks are a negative sign because Islam disapproves of them. Al-Dinawari states:<sup>41</sup>
<Lf.ll ....:ii .:,.l;J o....J ~ d+JI <Lil <L.J ~ *'il.'.:....* JIU...., JI bJ~ "51J .:,Ll *WojL* u......o.JIJ bJ\_,\_....::JI dL ~I~ L. u-11 J.bL:J~ '-:-'~<Liu b~ ~ bJJ--=!-~ .:,~ °iJ ~\_,...:; °iJ ~ *'i* -..JLL:; <Li.JI d...::.JJ .:,'i l~J--'!"JI ..::..WLS: . dJ.i.:i..J I
<sup>38</sup>*Al-Muntakhah,* chapter 2, pp. 30-31.
<sup>39</sup> Lit., "dead."
<sup>40</sup> Al-Nabulusi, vol. 2 , pp. 85-86, s.v. ~-
<sup>41</sup>Al-Dinawari,fae/ I, *bah* 3, *Esad Efendi* 1833, fol. 33b; BN *arahe* 2745, fol. 47b; repeated in *al-Muntakhah,* chapter I, p. 28.
If <someone> dreams of an image or a depiction or a picture and is told, "This is your God," and prostrates himself in front of it or thinks that this is his God and adores it, he will approach falsehood in proportion to how closely he linked himself to this image and depiction, whether incidentally or substantially, because the vision of God Almighty cannot be delineated, cannot be depicted and does not exist when one is awake.
Similarly, Ibn Shāhīn (no. 17) says: و من راى ان الله تـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ وهو بسجد لها فانه يفترى على الله تعالى (If someone dreams of an image of God Almighty and prostrates himself in front of it, he will fabricate lies against God Almighty).
The interpretations of Jesus and icons in the Oneirocriticon were clearly not copied from the equivalent interpretations of its Arabic sources, but neither were they invented by the Christian author. The Arabic interpretations of Godhead, the Prophet Muhammad, and the other prophets of Islam show a number of similarities to the contents of chapters 11 and 150. Both agree that the significance of dreaming about the prophets is almost equivalent to dreaming about angels. The first paragraph of chapter 11 was probably copied from an Arabic source without change, since Christianity and Islam do not disagree about the status of the prophets. Chapter 11 begins thus:42
Εάν τις ίδη κατ' όναρ προφήτην ή απόστολον ή διδάσκαλον ή μάρτυρα, και ούτοι δύναμιν ἔχουσι καθώσπερ ἐπὶ τῶν ἀγγέλων, μετριωτέρα δέ ἐστιν ή ἔκβασις των πραγμάτων, διότι όπτασίαν άγγέλων άμαρτωλός ἢ πονηρός οὐχ έώρακε, προφήτας δε και τους λοιπούς πάντες αμα έθεάσαντο.
If someone dreams of a prophet or an apostle or a teacher or a martyr, it has the same validity as <dreaming of> the angels. However, the outcome of events will be more moderate, because no sinner or evil person ever saw an angel, while everyone has seen the prophets and the rest <of the holy figures mentioned>.
A similar interpretation is quoted by Ibn Qutayba:43
ما اقرب ما بين الملايكة والانبياء فى التاويل الا فى الشهادة وجدها // فانها في رؤية الملايكة دون الانبياء لان الملايكة عند الله والشهداء عنده قال الله تعالى ان الذين عند ربك يعنى الملايكة وقـال في الشـهـداء عند ربهم يرزقون والذلك سمى شهيدا لانه يشهد ملكوت السماء.
How close is the interpretation of the angels to the interpretation of the prophets!
<sup>42</sup> Drexl 6, 17-21.
<sup>43</sup> Ibn Qutayba, bab 5, Yahuda ar. 196, fol. 26a-b.
#### CHAPTER SEVEN
Chapters 11 and 150 both interpret Christ as signifying victory and spiritual well-being. Chapter 11 connects this dream with worldly wealth; chapter 150 says that it is only dreamt by the very pious or the extremely sinful so that they may repent. According to the first paragraph of chapter 11:45
Εάν τις ΐδη κατ' όναρ τον κύριον και θεὸν ήμων Ίησουν Χριστὸν καὶ συνήση, ὅτι αὐτός ἐστιν ὁ Χριστός, οῦτος εὑρήσει σωτηρίαν ψυχῆς καὶ πλοῦτον κόσμου και νίκην μεγίστην.
If someone dreams of our Lord and God Jesus Christ and realizes that he is Christ, he will find salvation of his soul and worldly wealth and a very great victory.
The beginning of chapter 150 conveys more or less the same message, but with additional details and somewhat different phrasing: 46
Έαν Ίδη τις κατ΄ οναρ, ότι ἐφάνη αὐτφ ό κύριος καὶ θεὸς ήμων Ίησοῦς Χριστός -τουτο μόνον οί καθαροί και οί άγιοι ή οί βασιλείς πιστότατοι όρωσιν ώς ἐπὶ τὸ πλεῖστον ἢ οἱ καθ΄ ὑπερβολὴν ἀμαρτωλοί πρὸς ἐπιστροφήν-, καὶ ἐαν συνωμίλησεν αὐτῷ ἀπαραλλάκτως κρατείτω τὴν δμιλίαν, όποία αν εῖη· εἰ δ ούχ ώμίλησεν αὐτῷ, εἰ μέν ἐστιν βασιλεύς, εύρήσει χαράν καὶ νίκας καθολικὰς κατ΄ ἐχθρῶν, εἰ δὲ καθαρὸς καὶ ἄγιος καὶ εὐλαβής, στερέωμα καὶ δόξα καὶ αϊξησις τῆς κατὰ θεὸν προκοπῆς καὶ χαρᾶς αὐτοῦ ἔσται.
If someone dreams that our Lord and God Jesus Christ appeared to him (this is dreamt only by the pure and the saintly, or by the most pious emperors, or by the exceedingly sinful, so that they may repent), if he talked with Him, let him unfailingly hold on to the conversation, whatever it might have been; if he did not speak to him, if <the dreamer> is an emperor, he will find joy and overwhelming victories
<sup>44</sup> Cf. also Ibn Sirin, Ta'hīr al-ru'yā (Pakistani ed.), p. وامـار رؤيـة بـاقـى الانـبـيـاء فـي الـنـوم :20 (As for) فانهم مثل الملايكة ... غير انّه ليس فى رؤيتهم الشهادة كمـا فى تاويل رؤيـة الملايكة seeing the rest of the prophets in sleep, they are like the angels ... though in dreaming of them there is no martyrdom, as there is in dreaming of angels).
<sup>45</sup> Drex1 6, 22-24,
<sup>46</sup> Drexl 105, 13-21.
against his enemies; if he is a pure, saintly and pious man, he will have steadfastness, glory, and increase in his progress toward God and in his joy.
According to the Christian faith, Christ was both man and God. This doctrine means that the author of the *Oneirocriticon* had to avoid copying anything from the Muslim interpretation of Christ that refers to him as a prophet, and this may have led him to base his Christian interpretations of Christ on the Islamic chapters on the godhead. Indeed, a number of Arabic dreambooks associate God with victory. According to al-Dinawari,47
If someone sees Almighty <God> speaking words of blessing and praise at a certain place, His mercy and forgiveness will descend upon the people of that place. If they are at war, He will render these people victorious against their enemies ... and this will be the home of the martyrs and the blissful.
God is also interpreted as triumphing in war in Ibn Shahin (no. 13):
dJj ~-' ~JL.:.. \_,I ~.rA \_,I ~.l....a \_,I~)~ Jj...j *\_)W* <l.JI .JI .slJ .J-c-9 .• 1..i.c.'tl ~ ~~\_, .JLSll dJj JAi ~\_)LL; <l.JI .JI~ J~
Whoever sees that God Almighty descended to a land or a city or a village or a quarter of a city or some such place, it means that God Almighty will bring victory to the people of that place and grant them triumph over their enemies.
According to al-Nabulusi,48 dreaming of God signifies not only victory but also worldly success and salvation in the Hereafter. In addition, the sinners who dream of Him will repent, an effect also claimed for dreaming of Christ in chapter 150 of the *Oneirocriticon:*
~-' ~ ~ J:J..1~ ')'\_, ~ *'t\_,* ~ ~ oJ~\_, ~ oiJ <sup>~</sup> ~,J.) .s~I JL.....;:, i.IJ .JI\_, ... o4h ~ ~..1L)l.....\_,11~..1 ~ oJ ~J~ .<Gl..i.c.1 ~~I i\_,..L.k.... oiJ .JI\_, ~I
Whoever sees God in his power and glory, without attributes, human form, or <other> characteristics, this is a sign of goodness and glad tidings regarding his
<sup>.</sup>J~ C.rLi-o\_, ......::.....a.J .Ju ~-9-4 ~~\_,I~ oJ~ ~ i.IJ .Ju .IJ.t..!...ll ..!.11.4A ... ~\_,~ ~ 4-LJll ~ '-:-'~ ~ .JLS: .JI\_, ~\_,.11 dJj JAi . • I .i.a....J I \_,
<sup>47</sup> Al-Dinawari, *fa:;/* I, *biib* I, *Esad Efendi* 1833, fol. 32b; BN *arabe* 2745, fol. 46b.
<sup>48</sup> Al-Nabulusi, vol. I, p. 8, s.v. ~LA.:; <l.J I .
worldly affairs and the soundness of his faith <which will secure his well-being> in the Hereafter. If a sinner dreams of God, he will be rightly guided, and if the dreamer is oppressed, he will be victorious against his enemies.
Some of the Arabic interpretations of godhead are similar to those relating to the Prophet Mul).ammad. Since we do not know the exact contents of the Arabic texts that the author of the *Oneirocriticon* used for his compilation, it is conceivable that the Greek interpretations of Christ were based on the Arabic interpretations of Mul).ammad, possibly because Christ is the founder of the Christian religion, and Mul).ammad is the founder of Islam. In *al-Muntakhab,<sup>49</sup>* dreaming of Mul).ammad is interpreted as promising victory, as well as spiritual and worldly improvement:
,,1-> \_:,I\_, <lJI ,,L..L..:;, ~.r" ,,j-> .:,)\_, ~..1 <l.JI ~ \_:,-"=..1..-4 ,,j-> \_:,I <i....:.U ~~ ~)-,? (.Sj-> .:,)\_,~I~->-'~,,]->.:,)\_, <lJI ,,~ *'-:-'->L:........* ...J~ ~-P-,? \_,I Y..LL ~ J..1.:" r1WI ~ Ll.\_j .U ~.J-"-,? \_,I ~I .ilil .:\_....I
If someone who is in debt dreams of MuJ:iammad, God will release him from his debt; if he is sick, God will heal him; if he is at war, God will grant him victory; if he has never performed the pilgrimage to Mecca, he will manage to perform it. If MuJ:iammad is dreamt of in an arid land, it will be made fertile; and if he is dreamt of in a land where tyranny had spread it will change to justice, and if he is dreamt of at a place where there is fear, its people will be secure.
Mul).ammad is associated with worldly wealth, success, victory, and bliss in the Hereafter in Ibn Shahin (nos. 303 and 336):
04-JI\_, j..LJI ~ J..l..:' LlLl...,o lj 1->J.)-'-"-" l...::..\_;-A ~I (.SI->.:\_..... ..\_...:.L...~1 Ju\_, .\_;-il;JI\_,
Al-Kirmani said: "Whoever dreams of the Prophet as joyful and cheerful with a smile <on his face>, it signifies power, rank and victory."
ri.JI ..I....!..:> (\_;--ijl 4J ~ ~µ F-' ~ <l.JI ~~I (.SI\_).:\_.....\_, L...1\_, ,~~I\_, ~I ..;:,~I\_,.:. .b....::...J.\_, *J..::'-'""-,?* .:,LS.:,)\_, ... ~..1 ~-' o..1LL....... ~ J~ i)L..JI ~ ~j-> ~-' ... ~ ..11..1..J.= .....:iµ ~ .:,LS ljl <l.JI ol..Ll. ~..>-".:,LS.:,)\_, ·L'..l..LLI ~~~~.:,LS .:,Iµ\_, ·c..r.-WI .c)L....:.
<sup>49</sup> *Al-Muntakhah,* p. 32.
Whoever sees in his dream the Chosen one, may the peace and blessings of God be upon him, indeed joy after distress will come to him and his debts will be paid ... If he is in dire straits and in a poor financial situation, beneficence and goodness will fall amply upon him. If he is rich, his wealth will increase ... It is said that the vision of the Prophet, may peace be upon him, signifies bliss in the Hereafter. And it is said that if <the dreamer> was defeated, he will be rendered victorious over his enemies, and if he is sick, God will heal him.
Dreaming of the Prophet is also associated with victory in al-Nabulusi:10 وربما دلت رؤيته على نصر المؤمنين ودمار الكافرين... وان راه محرب نصره الله تحالى (Sometimes, the vision <of Muḥammad> signifies victory of the believers and destruction of the unbelievers ... If someone who is at war dreams of the Prophet, God will grant him victory).
The excerpt from chapter 150 stating that whatever is said by Christ in a dream will be fulfilled is repeated in chapter 11:35 έαν δε και όμιλήση αύτω, παν, ό τι άκούση παρ' αυτού, κρατείτω άναμφιβόλως (If he also converses with Him, let him without doubt hold on to everything that he heard from Him). Arabic dreambooks state that whatever is uttered by God in a dream will be fulfilled. Al-Dīnawarī says: وكذلك ان حكم عليـه فـي النـوم بحكم او امره بامر فه فى اليقظة كما حكم او امر به لقوله تعالى اليس الله باحكم الحكم الحكم الحكم الحاكمين (Likewise, if God decrees something for the dreamer in his sleep or orders him <to do> something, in his wakefulness things will happen according to God's decree or command, according to the saying of God Almighty: "Is not God the justest of judges?" [Qur'an 95:8]).
According to chapter 11 of the Oneirocriticon, whoever dreamt of Christ is blessed:33 άλλα και μακάριος ό τοιούτου θεάματος θεωρός (And the one who saw such a vision is blessed). Modern scholars have thought that this interpretation might have been inspired by the phrase from Luke 10:23:34 μακάριοι οί όφθαλμοι οι βλέποντες α βλέπετε (Blessed are the eyes which see the things that ye see), but Arabic dreambooks also claim that whoever dreams of the Prophet Muhammad is blessed. According to al-Muntakhab (p. 32), قـــال الاســــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ
<sup>50</sup> Al-Nābulusī, vol. 2 , p. 216, s.v. محمد .
<sup>51</sup> Drexl 6, 25-26.
<sup>52</sup> Al-Dinawari, fast 1, bab 1; Esad Efendi 1833, fol. 32b; BN arabe 2745, fol. 46a.
<sup>53</sup> Drexl 6, 26-27.
<sup>54</sup> Brackertz, Das Traumbuch des Achmet, n. 54.
محمدا صلى الله عليه وسلم رحمة للعالمين فطوبى لمن رأه فى حياته ه فاتبــعـه وطـوبـى لـن يـراه فـي مـنـامـهـ (Master Abū Sa'd, may God be pleased with him, said: "God sent Muhammad, may the peace and blessings of God be upon him, out of mercy for the world, and blessed is whoever saw him during his lifetime and followed him. Whoever will see him in his dream is also blessed").
According to the Oneirocriticon, dreaming of entering a house in order to join Christ may signify the dreamer's death:55
έαν δε τδη αυτον έν οϊκω τινι άγνωρίστω και είσέλθη έν τω οϊκω ό θεωρών, ουχ ύποστρέψη δέ, γινωσκέτω, ότι αποθανείται μέν ταχύ, σωθήσεται δέ: άλλα και οί κληρονόμοι αύτου άξιωμάτων και πλούτου πλησθήσονται.
If <the dreamer> sees Christ in an unknown house and enters the house and does not come back, he should know that he will soon die, but will attain salvation, and his heirs will be loaded with offices and wealth.
According to Arabic dream interpretation, under certain circumstances dreaming of God can foretell the dreamer's death. Ibn Shāhīn (no. 15) says: ومن
راى ان الله تعالى قالى قالى قالى قالى قال الى يدل على قرب اجله that God Almighty said to him "Come to me!" this signifies the approach of his death). The interpretation of the house as a dream symbol associated with death in the Oneirocriticon can also be traced to Arabic dream interpretation. The Arabic word dar can mean "a mansion" or "a large compound of structures," or, in a more general sense, "abode."56 The same word is also used to designate the "abode of the Hereafter" (al-dār al-ākhira). Arabic dreambooks therefore interpret an unknown dar as the abode of the Hereafter. According to Ibn الدار الجنهـولـة البـنـاء والـتـربـة والموضع والاهـل هـى دار الآخـرة ",Qutayba ولاسيما اذا راى بها موتى يعرفهم (A compound with unfamiliar structures, grounds, location, and people is the abode of the Hereafter, especially if <the dreamer> sees in it dead persons whom he recognizes). A more detailed version و مين دخـل دار of the same interpretation can be found in al-Nābulusī: و مـن دخـل دار مجهولة البناء والتربة والموضع والاهل منفردة عن الدور لا سيما ان راي
<sup>55</sup> Drexl 6, 27-7, 2.
<sup>56</sup> The word hayt (house) designates a room in the larger compound of a dar; in a Bedouin context, dar refers to a camp, bayt to an individual tent.
<sup>57</sup> Ibn Qutayba, hah 17, Yahuda ar. 196, fol. 35a.
<sup>58</sup> Al-Nābulusī, vol. 1 , p. 207, s.v. دخول الدار.
فيها موتى يعرفهم فهى الدار الآخرة فانه يموت with unfamiliar structures, grounds, location, and people and isolated from the <other> houses, especially if he sees in it dead persons whom he recognizes, this is the abode of the Hereafter and he will die).59
According to the excerpt from chapter 150, only the very pure or the very sinful can dream of Christ. This idea is then further developed in chapter 1 50.60
εί δε ό ίδών έστι καθ ύπερβολήν άμαρτωλός, επιστραφήσεται και έν μετανοία ἀποθανείται. όμοίως γὰρ μαρτυρεῖ τῷ λόγω καὶ βασιλεύς: ἐαν ἔχη ἄνθρωπον ἐν κινδύνω καὶ ὀργὴ καὶ προσκαλέσηται καὶ ἴδῃ αὐτόν, συγχωρεῖ, ὅσα ὕμαρτεν, αύτων διά τουτο και ό είρημένος ούτος άμαρτωλός τη θεωρία του κυρίου σωθήσεται. όμοίως και έκ των άποστόλων και προφητών και άγίων έαν ίδη τις έκ των είρημένων συνομιλούντα αύτφ, εί μέν έστι βασιλεύς, κρατείτω τὴν όμίλιαν ἀναμφιβόλως εἰ δ οὐκ ἐλάλησεν αὐτῶ, μόνον δὲ εἶδεν, εὐρήσει χαρὰν ἐλάτονα της του κυρίου όπτασίας εί δε του κοινού λαου ή, εύρήσει και είς τὴν πίστιν αύτου έκαστος και είς το έργον αύτου προκοπήν. διά τουτο και είκονίζομεν αύτούς.
If the dreamer is exceedingly sinful, he will repent and die in repentance. This interpretation is also evidenced by <the conduct of> emperors: if they hold someone liable <for something> or if they are angry at him, if they invite him and see him, <that means> they forgive him his wrongdoings. For the same reason, the aforementioned sinner will be saved through beholding his Lord. Likewise, if one of the aforementioned persons sees one of the apostles, prophets, or saints talking to him, if he is king, let him consider that whatever <they said will) undoubtedly <come to pass>. If <the dreamer> does not talk to <the holy figure> but only sees it, he will find joy, but less than in the vision of the Lord. If the dreamer is a commoner, he will make progress in his faith and in his work. This is why we represent <holy figures> on icons.
Al-Muntakhab' also states that both the righteous and the sinners dream about God. For the latter, such a dream is a warning sent to them in order to repent:60
<sup>59</sup> These interpretations can also be found, in an almost word-for-word correspondence, in chap. 146 of the Oneirocriticon (Drexl 100, 25 ff.): Έάν τις ίδη, ότι είσηλθεν είς οικον άγνωριστον έχοντα κτίσματα και τόπον και χουν και οικήτορας άγνωρίστους ή γνωρίμους μέν, προτελευτήσαντας δέ, νοείτω, ότι ό οίκος έκεινος ό αλλος κόσμος έστι και κρίνεται και ό ίδών ταχύ τελευτά (If someone dreams that he entered an unfamiliar estate that had structures, grounds, land and inhabitants he did not know or whom he knew, but who were already dead, let him understand that this estate indeed is, and is interpreted as, the hereafter and the dreamer will soon die).
<sup>60</sup> Drexl 105, 21-106, 4.
<sup>61</sup> Al-Muntakhah, chap. 1, p. 27.
<sup>62</sup> Repeated in BL Or. 6262, fol. 19b [19], which contains al-Khargushi's al-bishara wa-al-
#### CHAPTER SEVEN
قال الاستاذ ابو سعد رضى الله عنه من راى فى منامه كأنه قائم بين يدى الله تعالى والله تعالى ينظر اليه فإن كان الرائى من الصالمين فرؤياه رؤيا رحمة وإن لم كان من الصالحين فعليه بالمذر لقوله تعالى يوم يقوم الناس لرب العالمين.
Master Abu Sa'd, may God be pleased with him, said: "If someone dreams that he is standing in front of God Almighty and that God Almighty is looking at him, if the dreamer is one of the righteous, his dream is a dream of mercy. If he is not one of the righteous, let him be warned, according to the saying of God Almighty: "The day when mankind shall stand before the Lord of all Being" [Qur'an 83:6]].
A similar interpretation is quoted by Ibn Shāhīn (no. 2):
ومن رأه وهو قائم والله تعالى ينظر اليبه دائما <يدل> على ان هذا الـعبـد يسلم فى امر ويكون فى رحمة الله تعالى فإن كان مذنبا ينبغى ان يتوب.
Whoever dreams that he is standing in front of God Almighty and that He is constantly looking at him, <the dream signifies> that this servant of God will have peace in his affairs and be in the mercy of God Almighty. If he is a sinner, it is necessary that he repent.
Dreaming of the Prophet is also interpreted as repentance in al-Nābulusi:63 وان كان الرائى على بدعة وضلالة فليـتـق الله (If someone dreams of the Prophet>, if he is in heresy and in error, let him be afraid of God).
In effect, any prophet can signify glad tidings or a warning. According to al-Muntakhab, قال الاستاذ ابو سعد رحمه الله رؤيا الانبياء صلوات الله مليهم احد شيئين إما بشارة وإما الذارة وإما انذار mercy on him, said: "Dreaming of the prophets, may the blessings of God be upon them, is one of two things, either glad tidings or a warning"). Something similar is also mentioned in the dreambook of al-Nābulusī: وتدل رؤیتیے ة (The vision of the vision of the l الانبياء والمرسلين عليه السلام على الانذار والبيشارة prophets and the apostles, may peace be upon them, signifies <both> a warning
and glad tidings).
nidhara with an omission that obscures the meaning of the passage; the same omission is made in the repetition of this interpretation in Ibn Shāhīn, no. 28; both texts should be corrected on the basis of al-Muntakhab .
<sup>63</sup> Al-Nābulusī, vol. 2, p. 215, s.v. محمد
<sup>64</sup> Al-Muntakhah, chapter 2, p. 28.
<sup>6</sup> Al-Nābulusī, vol. 2, p. 287, s.v. نبى من الانبياء .
The passage in chapter 150 connecting Christ with earthly royalty conveys an idea that is, of course, well founded in Byzantine imperial ideology, but it is also not without parallel in Arabic dream interpretation; and the author of the Oneirocriticon must have found the coincidence very convenient. Ibn Qutayba only hints at a connection between the prophets and royalty, as the last section of his chapter on the prophets interprets the significance of dreaming about becoming a prophet or a king. Al-Nabulusi associates royalty with the رؤيتـهم تـدل على الملوك لانهم ملوك الدنيـا 64.prophets in a direct statement والآخرة (The vision of the prophets also signifies kings, because the prophets are the kings of this world and the next).
According to chapter 150, the more elevated the status of the holy figure dreamt about the more auspicious the dream. The idea is repeated in chapter 11:67
άλλα καί τινα των αποστόλων ή πατριαρχών έάν τις ίδη ώς τον Χριστόν βλέπων, ούτως έσται αύτω ή έκβασις του πράγματος, πλην μετριωτέρα και κατωτέρα.
And if someone dreams of any of the apostles or the patriarchs as if he were seeing Christ, the outcome of things will be the same, only more modest and inferior.
A similar way of thinking seems to be reflected in the following passage from al-Dīnawarī:68
فمن راى احدهم < = احد من اصحاب النبى؛ فى منامه فان رؤيتهم بركة من فضل بركة النبى صلى الله عليه وسلم على اقدارهم وهى كرامة الكرم الله تعالى بها صاحب الرؤيا.
If someone dreams of one of them [the companions of the Prophet], indeed, their vision is a blessing from the overflow of the Prophet's blessing-may the peace and blessings of God be upon him-analogous to the identity of the companions <that the dreamer dreams of> and an honor by which God Almighty distinguished the dreamer.
The second half of chapter 150 discusses icons that depict Christ and various holy figures. Dreaming of icons is given more or less the same interpretation as dreaming of the figures depicted on them. Icons of Christ are therefore
<sup>66</sup> Al-Nābulusī, vol. 2, p. 288, s.v. نبى من الانبيياء.
<sup>67</sup> DrexI 7, 2-5.
هه Al-Dīnawarī, faș! 4, bāb 1 اصحاب النبيى), Esad Efendi 1833, fol. 42a; BN arabe 2745, fol. 62a.
#### CHAPTER SEVEN
associated with victory, and whatever is uttered by an icon in a dream will be fulfilled:69
' Εάν τις ϊδη είκόνα του κυρίου ήμων Ιησού λριστού ἐσταυρωμένου, ὅτι προσεκύνησεν ή ήσπάσατο ή έδεήθη αυτής, εί μέν έστι βασιλεύς, και ούτος ευρήσει νίκας κατ΄ ἐχθρών καὶ χαράν ἐν τὴ δόξη αὐτοῦ, διότι ὁ σταυρὸς μετὰ τοῦ σταυρωθέντος έν αύτω κυρίου ήμων Ιησού νίκας μεγίστας σημαίνει και πάσης θλίψεως μεταβολήν και έαν ίδη είκόνα μόνην του κυρίου ήμων Ιησού Χριστού ανευ του σταυρού, έστιν ή χαρά έλαττοτέρα, όμοίως και ή νίκη. εί δε ίδη τις του κοινού λαου τουτο, εύρήσει χαράν και επιτυχίαν της δεήσεως αύτου. ἐαν ίδη τις εἰκόνα ἀποστόλων ή προφητών ἢ άγίων, εἰ μέν ἐστι βασιλεύς, μήνυμα νίκης μεγίστης δέξεται καὶ ἡ νίκη μετριωτέρα ἔσται ου γὰρ εἶδεν ἐκείνους αὐτούς, ἀλλὰ τὴν εἰκόνα· διὰ τοῦτο καὶ ἐλάττων ή νίκη. εί δέ έστι του κοινού λαου, εύρήσει χαράν και προκοπήν έν τοῖς έργοις αύτου. ἐάν τις ἴδη, ὅτι ώμίλησαν αὐτφ αί εἰκόνες, καὶ μνημονεύει τῆς ὁμιλίας, καὶ ταῦτα ὡς ἐπὶ τὸ πλεῖστον εἰς πέρας ἔρχονται εἰ δὲ ἢ βασιλεὺς ὁ τοῦτο ίδών, ἐλεύσεται αὐτῷ χαρὰ καὶ νίκη κατ᾽ ἐχθρῶν μετὰ θαύματος, καθ᾽ ὅσον τὸ θαύμα της όμιλίας της είκόνος εί δε του κοινού λαού, και ούτος μετά θαύματος πλουτήσει.
If someone sees an icon with the Crucifixion of our Lord Jesus Christ, and he prostrates himself or kisses it or prays in front of it, if he is emperor he will have victories against his enemies and joy in his glory, because the cross with our Lord Jesus Christ crucified on it signifies great victories and relief from every kind of sorrow. If the dreamer sees an icon of our Lord Jesus Christ only, without the cross, the joy and victory will be inferior. If a commoner dreams this, he will find joy and fulfillment of whatever his request is. If the icon represents the apostles or prophets or saints, if the dreamer is an emperor he will receive news of a great victory which will prove to be not so great. If he is a commoner, he will find joy and progress in his works. If someone dreams that the icons talked to him and he remembers their words, most of what they said will be fulfilled. If the dreamer is an emperor, he will receive joy and a victory against his enemies through a miracle commensurate to the miracle of the talking icon. If the dreamer is a commoner, he will become rich through a miracle.
Chapter 150 attaches to the icons of the prophets, apostles, and saints the same interpretations that have so far been mentioned for Christ: icons of holy figures signify victory and worldly success, and their utterances are truthful. The same interpretations are given for dreaming of the prophets in Arabic dreambooks. According to Ibn Outayba.70
<sup>69</sup> Drex1 106, 5-24.
<sup>70</sup> Ibn Outayba, bab 5, Yahuda ar. 196, fol, 26b.
فـمن راى النبـيـين والمرسلين فى المنام قــد رأهم وهو عــزه ... وإن ر ءى فى ارض جدبة اخصب اهلها او عند قوم مظلومين نصروا او قوم مغمومين فرج عنهم .
If someone dreams of the prophets and apostles, indeed he saw them, " and this signifies his power ... If they are seen in a arid land, its people will enjoy abundance, or if the people are wronged they will be victorious and if they are in sorrow they will be cheered.
In al-Muntakhab dreaming of the Prophets is also interpreted as worldly success and triumph over one's enemies: 2من رای> نبیا علی حالته وهیځته فذلك? f<) دلیل علی صلاح صاحب الرؤیا وعزه وجمال جاهه وظفره بمن عاداه. someone dreams of> a prophet in his proper condition and form, this signifies the dreamer's righteousness and power, as well as a good reputation and triumph over his enemies).
Similar interpretations are also quoted by Ibn Shāhīn (no. 302): قـــال اىن سيرين رؤيا اولى العزم من الرسل تدل على العز والشرف ورؤيا الرسل تدل على الظفر والنصر (Ibn Sīrīn said: "Dreaming of the powerful apostles means power and honor. Dreaming of the apostles means triumph and victory").
In addition, Arabic dream interpretations suggest that the utterances of prophets in a dream will be fulfilled. According to Ibn Qutayba, فــمـن راى النبيين والمرسلين فى النام قد راهم وهو عزة وإن كلموه ببرا اوخيرا ه فهو ما قالوه (If someone sees the prophets and apostles in his dream, indeed he saw them, and this signifies his might. If they talk to him using words of righteousness and goodness, whatever they tell him is <true>).
The last section of chapter 150 interprets icons with golden revetments and ordering an icon to be made: 74
όσας ἐαν Ίδη ἄνθρωπος κατ' ὅναρ εἰκόνας κεχρυσωμένας ἢ ἀχρυσώτους, ἐπὶ των άχρυσώτων μόνον χαριεστέρα ή έκβασις: ἐπὶ γὰρ τῶν χρυσέων πολλάκις και θλίψιν ύποδηλούσι διά τον χρυσόν.
Εαν ίδη τις, ότι διετάξατο γενέσθαι είκόνας άγίων, εί μέν έστι βασιλεύς, ἐπιτηδεύματα καὶ τρόπους κατ᾽ ἐχθρών ἐργάσεται· καὶ ἐαν ἐπιτύχῃ ἀκριβῶς ὁ των είκόνων τύπος, ἐπιτεύξεται καὶ τὰ ἐπιτηδεύματα αὐτοῦ, ἐἰ δὲ μή, ἀναλόγως
<sup>71</sup> The purpose of this otherwise incongruous phrase is probably to suggest that the only way truly to behold the long ago dead prophets and apostles is in dreams.
<sup>72</sup> Al-Muntakhah, chapter 2, p. 28.
<sup>73</sup> Ibn Qutayba, bah 5, Yahuda ar. 196, fol. 26b.
<sup>74</sup> Drexl 106, 24-107, 2.
#### 284 CHAPTER SEVEN
arcoBi\crE'tal au1tj). 6µoi~ Kat KOlVOU A.aou, EV £Kacr1cµ £pycµ auwu arcoBi\crE'tat Ka'tcl TIJV Em nixiav auwu.
If a man dreams of gilded or ungilded icons, the fulfillment of the dream is more auspicious for those that are ungilded, for the gilded ones often signify sorrow because of the gold.
If someone dreams that he ordered icons of saints to be made, if he is emperor he will devise stratagems and decoys against his enemies. If the representation in the icons succeeds with exactitude, his stratagems will also succeed. If not, the fulfillment of the dream will be proportional. Likewise, if the dreamer is a commoner, each of his activities will succeed according to the <degree of> success <in making the icons>.
The interpretations of chapters 11 and 150 were not invented by the author of the *Oneirocriticon,* but were based on material he found in his Arabic sources. This is as much as one can say with certainty. The Greek interpretations are similar to the Arabic interpretations of God, MuI:iammad, and the other prophets, and the Arabic interpretations of these three dream symbols resemble each other. It is therefore impossible to know exactly what Islamic concept the Greek author adapted into what Christian concept. Undoubtedly, the Greek author used the Arabic interpretations creatively, but his efforts to adapt the Muslim material to the needs of his Christian readers are still in agreement with the principles set forth in Arabic dream interpretation, whereby a dream interpreter is encouraged to infer the interpretation of a dream symbol that is missing from the written sources by drawing analogies with the existing interpretations sanctioned by the oneirocritic tradition. 75 Such an approach to the Arabic material would explain why the interpretations of Christ and other holy figures are essentially repeated in the interpretations of their icons. Since, according to Christian belief, an icon is only a depiction of the original and under no circumstances can it be treated or regarded as the original itself, the *Oneirocriticon* states that the fulfillment of dreaming of an icon is less intense than when dreaming of the actual holy figure the icon represents.
The placement of the chapter on icons in the *Oneirocriticon* is unexpected.
<sup>75</sup>Ibn Shahin says (introduction, p. 11: ~G. ~I ~ ~ L. ~ .:, *<sup>3</sup> \_,..,...J* I ~I \_,\_J *<sup>3</sup>* ~ .\_;ul:JI ~JJ ...J~l3 ~IF *.:,'i* .~I~ ~j.:j rJ ·~.~I .J-c. 13~ . •u--6Ll .U ~ (If dream interpreters were to rely specifically on what is set down in books, they would have been helpless regarding many things that are not mentioned in the books, because the science of dream interpretation and differentiation among the dreams of people is like a sea that does not have a shore); see also Fahd, "Les songes et leur interpretation selon !'Islam," p. 146.
It is not among its first religious chapters, but inserted much later, following a chapter on priests and monks (chapter 149), adapted from Muslim interpretations found in the religious section of Arabic dreambooks. In Arabic dreambooks this section is usually placed at the beginning of the work, where Muslim authors put the material on which the author of the *Oneirocriticon* based his interpretations of Christ, the holy figures and icons. Apparently the Greek author remembered to write a chapter on icons-without which no Christian dreambook would be complete-while writing on priests and monks and perusing the religious chapters of his sources for a second time. Christ and other holy figures are treated somewhat cursorily in chapter 11. Their treatment in chapter 150 is more extensive and evidently bolder in its adaptations, possibly because the Christian author, by then almost midway in his compilation, had become more familiar with the Arabic methods and approaches to dream interpretation and felt better able to improvise.
#### *Priests and Priestly Duties*
The *Oneirocriticon* interprets dreams about priests and priestly duties in four chapters (11, 12, 139 and 149). Chapter 11, "From the Indians on Prophets, Apostles and Teachers," discusses the significance of dreaming about being ordained patriarch, presbyter, or deacon and of reading to the people from the Holy Books. Such dreams generally signify power, sovereignty and justice exercised by the dreamer (Drexl 7, 6-20). Chapter 12, "From the Account of the Indians on Various Faiths," 76 states that dreaming of becoming a priest and praying for the people is interpreted as sovereignty, and reading aloud to people foretells wisdom (Drexl 9, 12-16). If a woman dreams of ordination as priest, however, it means that she will be dishonorably divorced from her husband (chapter 11, Drexl 7, 11-12). An anecdote recounted in chapter 139 (Drexl 92, 1-7), tells of a woman who had such a dream, and prayed on behalf of the people. Sereim interpreted it as future separation from her husband, fornication, and bearing an illegitimate child. Needless to say, his interpretation proved to be accurate.
Chapter 149 (Drexl 103, 25-105, 11), which also appears at a seemingly
<sup>76 &</sup>quot;£K *wii* Ahyou i:Oiv · Iv80iv itEpt Jtt<HEW~ 8ta<j>6pou."
#### 286 CHAPfERSEVEN
irrelevant place unconnected to the first, religious, chapters of the work, is the longest chapter on priests contained in the *Oneirocriticon.* It interprets dreaming that a priest arrived at a place, or slept in the dreamer's bed, died, had enlarged bodily members, abandoned the priesthood, walked in the dark or in the light, or lost an article that belonged to his priestly attire. It also repeats the interpretations of chapter 11 on the ordination of a lay person as priest. At the very end of the chapter, *Vat. gr.* 573 (fol. 162b) adds one further interpretation that is missing from Drexl' s critical edition: yuvfi £av lblJ on *£y£vew* 1epeur; Kat CTUV£A£t 'tOUpyn t£p£t, 'taXU ano8av£t 'tat Kai crco8iicr£mt (If a woman dreams that she became a priest and performed the liturgy together with a priest, she will die and be saved), which is at odds with the interpretation of a woman's ordination in chapters 11 and 139.
The interpretations of priests generate a number of questions. Were they invented by the Greek author, or copied or adapted from the Arabic model? How can the seemingly arbitrary positions of chapters 139 and 149 be accounted for and how can the disagreement between chapters 11 and 139, on the one hand, and chapter 149, on the other, be reconciled?
As there is no sacerdotal class in Islam, the interpretations of priests given in the *Oneirocriticon* cannot have been copied from an Arabic source. Though Arabic dreambooks do discuss dreaming of Christian priests and church prelates, the interpretations they provide clearly reflect a Muslim point of view and do not coincide with those given in the *Oneirocriticon.* Al-Dinawari interprets dreaming of the *katholikos,* of a priest, a monk, and a metropolitan as follows: 77 ~ .u1~1 J1 u~ ...91 ct.:.~~ J..l.::' ci....:.u ~4 *)-·"""* ct....:.L.S (..\$1\_\_, -:,u . ~ J 1 JjJ <.!.l)y\_J I (If someone dreams that he became a *katholikos,* this signifies that he will die or drown or will be on the brink of ruin and will lose allhiswealth). 78 D 4-L..l..:' '-:-"'""""'L....o ci....:.U ~I\_\_, ct....:.1 i.SI--> .J-o \_:,~I Ju *4....'.:.\_\_,.J-:* <Liu ~I\_) ...91 L......,j J~ <Li~ (.SI\_) .J-o (.S\_JL..::...:J I c.Ju...9 ~ ..b\_\_,.jl <sup>~</sup>*.. ...9* l..::. ~ •• Lu .-->...9 U. *\_)\_)* <L.......k .. ~...9 - • <t...:.Ll <L.......k .. ~ i...J-'-' • <I l.:............ ~ .. 1> \_\_, ..' • I (The Muslims said: "If someone dreams that he is a monk, he is a heretic <or an innovator> who has just exceeded the proper bounds." The Christians said: "If someone dreams that he became a priest or a monk, he will inherit something exquisite, though his circumstances will become straitened
<sup>77</sup> The same interpretations are repeated in *al-Muntakhah,* chapter 46, p. 331.
<sup>78</sup> Al-Dinawari, *fa#* 8, *hiih* 77; *Esad Efendi* 1833, fol. 65b; BN *arahe* 2745, fol. 144a.
and his daily sustenance will be insecure. Possibly, the fulfillment of this dream will benefit those around him" ).79
من راي انه مطران يخضع له فانه رجل صاحب سلطان يدعوا قوما الى بدعة فيجيبونه بقدر ما خضعوا له ويغلوا فيه بقدر ما علا من امره بين الناس فان دعي مطرانا وهو كــاره فـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ برىء ... فان راي انه راهپ اوحبر فانه صلحب بدعة وهو مفرط لقوله تعالى ورهيانية ابتدعوها الاية.
If someone dreams that he is a metropolitan obeyed <by the people>, he is a powerful person who will invite others to heresy and they will follow him by analogy to their obedience <in the dream> and will exceed the proper bounds regarding this by analogy to the loftiness of his rank among the people <in his dream>. If he summoned a metropolitan and <the metropolitan> was unwilling <to come>, he will blindly follow a heresy or a lie and will be accused of it, but will be found innocent ... If someone dreams that he is a monk or a bishop, he is a heretic and a prodigal person, according to the saying of God Almighty: "But monasticism they invented-We ordained it not for them-only seeking God's pleasure, and they observed it not with right observance. So We give those of them who believe their reward, but many of them are evil-livers" [Qur'an 57:27].80
Al-Dinawari's interpretation of a Christian sermon can be deduced from an anecdote: 81 A Christian dreamt that he had become a priest and was delivering a sermon to other priests, who were listening to him in silence. A Christian dream interpreter who interpreted the dream said that the dreamer would perpetrate slander three days later, as indeed happened.
The interpretation of the Holy Books in the Oneirocriticon is very different from the interpretation of the Gospel in al-Dīnawarī: مـن راى انــه يـتـلـو التوراة والانجيل فلم يعرفهما فانه رجل يذهب مذهب القدرية والجبرية . If someone dreams that he is reading aloud from the Torah and the Gospel and does not know them, he is a man who will follow the opinions of the Qadariyya and the Jabariyya,80
<sup>79</sup> Al-Dinawari, fasl 8, hāh 78; Esad Efendi 1833, fol. 65b, BN arahe 2745, fol. 144a .
<sup>80</sup> Al-Dinawarī, fași 8, hãb 80; Esad Efendi 1833, fol. 65b; BN arahe 2745, fol. 144a-b. The same interpretation of a metropolitan is repeated in al-Nābulusī, vol. 2, p. 267, s.v. مطران .
<sup>81</sup> Al-Dīnawarī, fasl 8, bāh 79; Esad Efendi 1833, fol. 65b; BN arahe 2745, fol. 144a.
<sup>82</sup> In the early theological debates in Islam, the Qadariyya were the advocates of free will, as opposed to the Jabariyya, the advocates of predestination. The theological implications of this interpretation quoted by al-Dinawari escape me. See W. Montgomery Watt, Free Will and Predestination in Early Islam (London, 1948).
#### 288 CHAPTER SEVEN
according to the saying of God Almighty: "And ye are readers of the Scripture! Have ye then no sense?" [Qur'an 2:44]).83 Al-Dinawari's chapter on reciting the Qur'an84 quotes a Christian interpretation for dreaming of a book: ..::.JLlJ ~ <t..:.Ll L:iL.:...S: ....J.JJ.=J ~..>7.J ......i=..L......:.ll ~ <t..:.L.S '-51,,; 0-o (.5,,;L....:....:JI . J'"'L.:J I ~ LIJ..i.......:. ~J .U~ (The Christians said: "If someone dreams that he put in order and bound the pages of a book85 and assembled a volume, his words will be favorably received and he will become <known for his> honesty among the people"). But even this passage does not appear among the interpretations of books in the *Oneirocriticon.*
Ibn Shahln's interpretation of the Christian Gospel (no. 847) also does not include anything close to the relevant passages in the *Oneirocriticon:* (.51,,; 0-o 0-oJ ''-5.J~I *µ* 0-o ~ ~ <t\_jlJ '-:-'~10-o ~'ti 1\_,....i: <t..:.I ·'-5,,;L....:...:J..J ~ 0~.J ~I~ J.b4Jl:it~ <t..:.µ '-:-'~ ~ 0-o 1~
(If someone dreams that he is reciting the Gospel from a book, he will receive benefits from the Christians. If he dreams that he is reciting it without a book, he will fail to see clearly what is right and what is wrong and will become a friend of the Christians).
Al-Nabulusi's chapter on the Gospel includes both positive and negative interpretations that reflect the Muslim view of Jesus, his miracles, and the written revelation that he brought to the world: <sup>86</sup>
~~I \_;;1\_, ..u.i\_).:j\_, 6..i~I ..i~ ~I <L1....o .:\_,I i~YI j.AI ..j..o lSI..> ..j..o ~ ~J.) -::..J..i l...,i.)J OJJ...c. ~Lil... .:\_,Ls: *.J!.J* Uj..t.JI\_, tU..i.:.YI\_, ~~\_)I\_, *L..Sk.....* .:\_,Ls: .:\_,I~~~ -,.J-C. l...,i..>J -.::...1..:.....o.::J.I ...J:U\_, .:\_,~I\_, '-:"':is.JI ..j..o F L.:...:.r-" .:\_,Ls: *.J).J* ~ y ~ ~ \_,1 .).Jj.Jt\_,. ..i+-..!. l..u1Ll .:\_,Ls: .:\_,I\_, F ~,La.WI ..:,...c. J.i.:JI \_,I L..w+JI ri' ~ ~J..> -::..J..i ~..>.J ~.r-" ·'-:"'.rh.JI\_, ,Lii.JI\_, ..r=~I *'-:"'t.\_,.)\_,* '-:"'l:&JI ~ ~J..> -::..J..i ~..>.J
If a Muslim dreams that he has the Gospel with him, he will renounce the Muslim religious observances, withdraw from the world and chose <to become> an anchorite and an ascetic and <live in> withdrawal and seclusion. If <the dreamer> is a king he will triumph over his enemies. Sometimes this dream signifies deception, slander, and defamation of women with unblemished reputations. Sometimes <the dreamer>
<sup>83</sup>Al-Dinawarl,fa,~/ 8. *hiih* \_81; *Esad Efendi* 1833, fols. 65b-66a; BN *arahe* 2645, fol. 144b.
<sup>84</sup> Al-Dinawari,fa0/ 8, *hiih* 20; *Esad Efendi* 1833, fol. 78b; BN *arahe* 2745, fols. 126a-127a.
<sup>85</sup>*Ma\$ii/:lif,* which I here render as "pages of a book," can also mean "copies of the Qur'an."
<sup>86</sup> Al-Nabulusi, vol. I, p. IO, s.v. ~I.
will triumph against his litigants, if there is a trial pending against him. If he is a witness, he will give false testimony or will report something that does not concern him. If he is sick he will be cured of his illness. Maybe this dream signifies the science of geometry or the conveyance of what <the dreamer> teaches among scholars. Possibly this dream <also> signifies the book and the masters of drawing, singing and musical entertainment.
The association of Christianity in general and the Gospel in particular with slander and falsification of the truth reflects the Muslim belief that the New Testament the Christians follow is incorrect, because its original text has suffered corruption *(tabrif)* in the hands of later editors. 87 The interpretation of the Gospel as signifying healing and geometry given by al-Nabulusi also seems to have been inspired by the miracles of Jesus described in the Qur'an.88 Finally, the association of the Gospel with triumph over one's enemies or litigants appears to have its roots in al-Nabulusi's interpretation of Christians, which is based on the identity of the radical consonants employed in the Arabic words for "Christian" *(na\$rdni)* and "victory" ( *nu\$ra)* :89 .U *\_:,\_l* ;;, ~ *.)* . o ., I 11 <sup>J</sup> d......o~ (Becoming Christian *(tana\$\$Ur)* signifies victory *(nu\$ra)* for whoever has *bukuma).* The word *bukuma* has multiple meanings; it can be translated as "judgment" and "sentence," in which case the dream signifies the triumph of someone facing litigation; it can also be translated as "authority" and "power of government," and then the dream signifies triumph for a ruler or a king.
These parallels make clear that the Christian interpretations of priests and their duties recorded in the *Oneirocriticon* were not copied from the relevant interpretations of Christian dream symbols in Arabic dreambooks, but may have resulted from the adaptation of equivalent Muslim dream symbols. The duties of a Christian priest in worship are partly carried out by an imam in Islam. The imam stands in front of his congregation, though on the same level as everyone else, and leads them in prayer. Mosques generally have an appointed imam who is supported by an endowment, but any suitable male can also lead the prayer (a woman can only lead the prayers for an exclusively female congregation).90 The imam is chosen from among the members of a congregation on the basis of his know ledge, particularly of the Qur'an, his age, and
<sup>87</sup> For an analysis of specific examples of corruption in the text of the Bible analyzing the Muslim point of view, see *Concise Encyclopedia of Islam,* s.v. "Bible."
<sup>88</sup> See Parrinder, *Jesus in the Qur'iin,* pp. 83-91.
<sup>89</sup> Al-Nabulusi, vol. 2, p. 301, s.v . .\_...:.I~.
<sup>90</sup> When men and women pray in the same space, women always stand behind men in order not to tempt and distract them from prayer.
#### CHAPTER SEVEN
his leading position in society. The imam may also deliver the khutba, a sermon that takes place on Fridays after the noon prayer, according to the example set by the Prophet. The khatib (preacher) stands on the minbar (pulpit)" and either reads his sermon from his written text or delivers it from memory. The interpretations of priestly duties in the Oneirocriticon imitate the interpretations that Arabic dreambooks give for similar duties usually performed by an imam.
The Oneirocriticon states that the dreamer's ordination "on a patriarchal throne" (in other words on an elevated place, such as the Muslim minbar) signifies that "his nation will rule." Ordination as presbyter signifies that the dreamer "will become lofty" ( wηλός έσται) (a metaphor that also has connotations of a physically elevated position), and will "rule over the royal people" (έξουσιάσει του βασιλικού λαου). It is unclear what the Greek author meant by the term "royal people." It could be a literal rendering of the Arabic ملكيون or ملكيية or ملكيية or ملكنية (lit. "royalists"), that is, the Melkites, the Christians under Muslim rule who remained loyal to the Byzantine emperor and Chalcedonian orthodoxy. 22 It could also have a much broader meaning. The term "royal people" has its origins in the Old Testament and indicates God's chosen people, the Jews.93 When it later occurs in the New Testament, it refers to the new chosen people of God, the community of Christians.94 The Islamic counterpart to this Judeo-Christian concept is the umma or community of Muslims, the "nation of Islam," in a definition that transcends ethnic and political boundaries. The Christian interpretations of ordination therefore are similar to the Muslim interpretations of the minbar. especially in the relevant chapter of al-Dinawari. The Greek interpretation of ordination is as follows: 95
Ε Ί τις Ίδη έαυτον κατ' όναρ χειροτονηθέντα ἐπὶ θρόνου πατριάρχου, βασιλεύσει το έθνος αύτου. έσται και περίβλεπτος έν άξιώματι και έξουσιάσει του βασιλικού λαου, άλλά και δίκαιος έσται του διακόνου πρός τον πρεσβύτερον. ἐαν δὲ γυνὴ τοῦτο θεάσηται, χωρίζεται ατίμως του ίδίου άνδρός.
<sup>91</sup> At the time of the Prophet, the minhar consisted of three steps but later became much higher.
<sup>92</sup> See G. Graf, Verzeichnis arabischer kirchlicher Termini (Louvain, 1954), p. 108.
<sup>93</sup> Exodus 19:6 and 23:22.
<sup>94</sup> Cf. 1 Peter, 2:9: ύμεῖς δὲ γένος ἐκλεκτόν, βασίλειον ίεράτευμα, ἔθνος ἄγιον, λαὸς είς περιποίησιν ... (But ye are a chosen generation, a royal priesthood, an holy nation, a peculiar people ... ).
<sup>95</sup> Chap. 11, Drexl 7, 6-12.
If someone dreams that he was ordained on the patriarchal throne, his race will rule. If he dreams that he was ordained a presbyter, he will become lofty, be admired by all for his office, rule over the royal people and be just. If he <was ordained> deacon, his honor will be inferior in proportion to a deacon's status as compared to a presbyter's. If a woman dreams this, she will be dishonorably divorced from her husband.
Al-Dinawari interprets the *minbar* thus: <sup>96</sup>
~ v1~ <Ll.JI ,,\_;;j i:?.:JI ~µ1 rilll\_, r~L..81 d..L4\_, .\_,.,~I .:\_,L.h..L..... ~I L.:.U:..L...~ ....:.U .11 ~ .1<·;. ... . I~ ....:.1 I .. <1.....iL::....S: . - - j-7-' r . *.--= .:r--'* ~ .\_,...- *i.>.)* ~ . -~"~~~~,) . ..ii ~I~ *,).S:* ,)I~\_;;¥..!~~
The *minbar* <signifies> the rule of the Arabs <over other peoples>.97 <It also signifies> the community of Muslims and the glorious station *(al-maqiim al-karlm)* which God Almighty mentions in his Book.98 If someone dreams that he is on the *minbar* and is talking in words of righteousness, he will attain honorable, lofty and eminent rulership, if he is worthy of the *minbar.* If he is not, this <dream> indicates that he is well known for his goodness.
Al-Dinawari's interpretation of the *khutba* (Friday sermon) mentions the possibility of a woman dreaming that she is delivering it; its interpretation is similar to that in the *Oneirocriticon,* in chapter 11 and in an anecdote in chapter 139: <sup>99</sup>
..::..\_,s....., rJ'J ~ ,:u~ \_\_,.,L:JI\_, ~ <L.a)\.S r-3-' ~I ~I\_,~ .:\_,Ll \* ~ d.J \_\_,.,L:JI ~ ~y\_, ~ ...:.u .J..,:.UI (~ ~ u~ ...:;).l...:. ~\_, ~ ~-' ~ ~µ d.....S.:.. ~ [+..)l.S *,).S:\_,* ~~I ;;1->-°I ..;:.,!..! .:\_,u ... . "L.....:J I J.U .J-o ~
And if <someone> delivers the sermon properly and completes his speech and the people pay attention to him by keeping silent, and if after the sermon his prayer is
<sup>99</sup>Al-Dinawari,fa~-/ 8, *bah* 45; *Esad Efendi* 1833, fol. 80b; BN *arahe* 2745, fol. 136a. Similar interpretations are repeated in al-Nabulusi, vol. 1, pp. 173-74, s.v. ~.
<sup>96</sup>Al-Dinawari,fa~/ 8, *bah* 46; *Esad Efendi* 1833, fol. 80b; BN *arahe* 2745, fol. 136b; repeated in al-Nabulusi, vol. 2, p. 232, s.v. ~ .
<sup>97</sup> The Arabic phrase ':"'~I .:,L.h.L... could also be translated as "a sultan of the Arabs." However, the use of the word *su/fan* to mean "ruler" is relatively late (for example, it does not occur in the Qur'an).
<sup>98</sup> The word *maqam* means "site, location, position" , but also "situation, station" and "rank, dignity," as well as "sacred place." The interpretation of the *minhar* as the "glorious *maqam"* obviously draws upon the multiple meanings of the word *maqam.* Qur'an 42:81 mentions the *maqdm maf:zmud,* translated as "glorious station," which is understood to be a special place in heaven reserved for MuJ:iammad.
#### CHAPTER SEVEN
also completed according to the precepts of Islam, he will exercise sovereign power and the people will obey him ... If a woman dreams that she delivered the sermon but her words are unwise, she will be publicly exposed and will become notorious for an abominable deed of the kind that is disapproved of for women to do.
Ibn Shāhīn (no. 1015) gives a similar interpretation: وان رأت المراة انهـــا If a woman dreams) تقرءا الخطبة وتتكلم بـالعلم و الحكمـة فـانـهـا تـفـتـضـح. that she is reading the sermon or is talking knowledgeably and wisely, she will be disgraced). Additional interpretations of the minbar and the sermon (nos. 923, 927, 1011) also agree with those on priestly ordination given in the Oneirocriticon:
من راى انه يخطب على المنبر وهو اهل لذلك يحصل له علو وقدر وعـز وجاه ... وان كان سلطانا مصلحا يدل على عدله وإنصافه ... وإن كان امراة فيفتضيع زوجها، وقيل يشتهر على رؤوس الاشهاد بكلام لا خير فيه، وقيل انها تتزوج وربما تطلق او تأتى بوكد من الزنا وعلى كل حال لا خير فيه.
If someone dreams that he is delivering the sermon from the minhar, if he is qualified, he will attain grandeur, power, might and glory .... If he is a righteous ruler, this signifies his fairness and equity ... If the dreamer is a woman, her husband will be publicly disgraced, and it is said that he will become notorious among the leaders of the witnesses for <uttering> words that have nothing good in them. It is also said that <the woman who dreams this> will marry and perhaps her husband will divorce her, or she will conceive a child from fornication. At any rate, nothing good will come of this dream.
ومن راي انه يعظ الناس وكان اهلا للولاية فإنه يتولى امرا يحكم فيه
If someone dreams that he is preaching to the people, if he is suited to rule he will exercise it and will do it well.
If someone dreams that he is standing on the minbar and is talking knowledgeably and wisely, or that he is delivering a sermon, if he is from the people of that place he will acquire from the imam, or whoever substitutes for him, grandeur, power and honor.
Priesthood and priestly duties in the Oneirocriticon are generally associated with rulership and authority, as well as fairness in exercising it. The interpretations from chapters 11 and 12 about reading from the Holy Books to the people and praying, as well as the additional interpretations of ordination
#### from chapter 149, repeat this pattern:
Έάν τις βιβλίον άναγνώ κατ' όναρ αποστόλων ή διδασκάλων ή προφητών είς ἐπήκοον λαου, μεγιστάνων προσώπων γενήσεται ὑπηρέτης, τιμώμενος ὑπὸ λαοῦ, ή άποσταλήσεται βασιλικός και δικαιοπραγήσει και επαινεθήσεται. έαν δε τουτο Ίδη βασιλεύς, ποθεινός γενήσεται τώ λαφ αύτου και δικαιοκρίτης. ἐαν δε το βιβλίον εὐαγγέλιον ή και θεάσηται αὐτό τις, ενδρήσει ύψος ἀξιώματος και έξουσίας, διότι, δ ουκ έχουσιν έξουσίαν οί λαϊκοί, αύτος καθ' ύπνους έποίησεν. 100
If someone reads a book in his dream written by any of the apostles or teachers or prophets, and the people are listening, he will serve powerful persons and will be honored by the people, or he will be sent on an imperial mission, will act justly and receive praise. If an emperor dreams this, he will be loved by his people and be a fair judge. If the book is the Gospel and somebody dreams <that he is reading it aloud and people are listening>, he will achieve high rank and authority, because in his sleep he did something that laymen have no authority to do.
' Εαν δέ τις ίδη έαυτόν, ότι πρεσβυτέρου του λαου ύπερεύχεται, λαου αρξει και άγαπηθήσεται παρά πάντων. 101
If someone dreams that he has the rank of a priest and prays on behalf of the people, he will rule over the people and be loved by everyone.
Έαν ϊδη τις, στι λαϊκός ών έχειροτονήθη ίερεύς, εί μέν έστι βασιλικός, μεγίστην και πρώτην άξίαν εύρήσει παρά του βασιλέως, εἰ δε του κοινού λαου ή πτωχός. μεγίστω άκολουθήσει και τιμήν και χάριν εύρήσει παρ' αύτου έαν Ίδη τις, ότι λαϊκός ών ἐχειροτονήθη διάκονος, καὶ οῦτος εὐρήσει χάριν καὶ τιμὴν ἢττονα τοῦ ίερέως. ἐαν ϊδη τις, ὅτι χειροτονηθείς ἀπέλιπε τὴν χειροτονίαν, ἐξουσίαν εύρήσει και χάριν και ταχύ ἀπολέσας αὐτὴν κινδυνεύσει. 102
If someone sees that, though he is a layman, he was ordained a priest, if he is a royal official, he will achieve exalted and high rank from the emperor. If he is a commoner or a pauper, he will follow a powerful person and will be given honor and favor from him. If someone sees that, though he is a lay person, he was ordained a deacon, he will also receive favor and honor, <though> less than in <dreaming of being ordained> priest. If someone sees that, after being ordained, he abandoned the priesthood, he will receive power and favor, but will soon lose it and be exposed to danger.
Arabic dreambooks similarly interpret dreaming about the functions of an imam as signifying rulership and wielding authority with fairness, since the word can be a honorific title applied to the head of a community or a group,
<sup>100</sup> Chap. 11, Drex1 7, 13-20.
<sup>101</sup> Chap. 12, Drexl 9, 12-14.
<sup>102</sup> Chap. 149, Drexl 104, 30-105,7.
and can also be used for the caliph. In particular, the interpretation of being a priest and praying for the people (chapter 12, Drexl 9, 12-14) has its counterpart in the Arabic interpretation of leading a congregation in prayer. Ibn Qutayba interprets this dream as follows: من ر اى انـه يـؤـم الـنـاس فـى الـصــلاة و لـي ولاية يعدل فيها ان استقامت قبلته وتمت صلاته وتمت صلاته leads the people in prayer will have sovereign power and will be just in exercising it if the direction towards which he recites his prayer is correct and his prayer is complete).
Al-Dinawari also interprets leading prayer as sovereignty:104
من راى انه امـام يصلى بـالنـاس لا يكون في اليــقظة امــامــا فــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ شريفة يطاع فيها كما يطاع الامام ويقتدى به من وراءه إن كان للولاية اهلا واذا استقامت قبلته وتمت فريضته وكان ركوعه وسجوده على منهاج الدين وشر ائع الاسلام فإنه يعدل فى ولايته على قدر صلاته وخشوعه.
If someone dreams that he is an imam who leads the people in prayer, if he is not an imam when he is awake, if he is suitable for leadership he will exercise illustrious sovereign power and will be obeyed to the extent that the people who are <praying> behind him obey him and follow his example. If the direction towards which he prays is correct, the obligatory procedure is performed to its completion, and his kneelings and prostrations are in accordance with the precepts of religion and the canon law of Islam, he will exercise his sovereignty with fairness commensurate with the validity of his prayer and his submission <to the rules of religion>. 105
Al-Muntakhab likewise associates prayer with sovereignty, as well as the dreamer's religious well-being: الاصل فى رؤيا الصـلاة فى المنام انهـا مــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ ودنيا وتدال على إدراك ولاية ونيل رسالة او قضاء او اداء امانة او اقامة فربضة من فرائض الله تعالى (The basic meaning of prayer in a dream is that of something good for one's religion and worldly pursuits. It signifies the achievement of rulership, or the granting of a mission or a judgship, or the fulfillment of a trust and the performance of one of the obligatory duties <of a believer> to God). 106 Al-Muntakhab then adds: 107
<sup>103</sup> Ibn Qutayba, Yahuda ar. 196, fol. 27a ff.
<sup>104</sup> Al-Dīnawarī, fașl 8, bāb 18, Esad Efendi 1833, fol. 76a; BN arabe 2745, fol. 125a.
<sup>105</sup> Al-Dinawari continues by explaining what kind of appointment the dreamer will receive depending on what kind of people comprised the congregation which he led in prayer.
<sup>108</sup> Al-Muntakhab, p. 48. Several passages on the interpretation of prayer that follow in al-Muntakhab are identical with phrases found in the relevant entry in al-Dinawari.
<sup>107</sup> Al-Muntakhab, Dp. 49-50.
#### ~I ~')'\_,J..J .:.,~\_, o~I..,? '-""L:JI (").;: <Lil.S ~I..,? f'L.l,. ~.:.,..... L>IJ .:.,Ll *..,?* J~ LL; o)l..... ("'+:'fa\_,~ I ~I("'+:' rl .:.,Ll LdJ ...... *JL..:.J* ~~ ~')' <sup>J</sup>JL..:. *.G.,:'i* J
If someone who is not an imam when he is awake dreams that he leads the people in prayer, if he is suitable for leadership he will receive exalted sovereign power and will be obeyed. If he leads them in prayer in the correct direction *(qibla)* and prays the complete prayer, he will exercise his sovereignty with fairness.
Ibo Shahin and al-Nabulusi also interpret prayer as sovereign power. Ibo Shahin (no. 407) quotes al-Kirmani: Lo~ i~ ....:.1 i..SI.> j.....a ~LoyS...11 JLlJ /)...>-"I jL..L,J *.>u* ~ <L..:.µ (al-Kirmani said: "If someone dreams that he is leading a congregation in prayer, this is grandeur, power and implementation of his orders"). According to al-Nabulusi: 108 ~ iW I ~ ~ \_riJ I ~~ UL......> JI L.....L:i *.>* JI ~ '.i J (The obligatory prayer 109 in a dream <signifies> rulership or leadership or a mission for the one who was praying).
The interpretation of reading from the Holy Books to the people also coincides with the interpretation of reading from the Qur'an in Arabic dreambooks. Al-Dinawari interprets such a dream as follows: <sup>110</sup>
I~ O...:..l lSIJ.:.,..... J::.JJ\* JJ..>-4"J ...J~\_, #J rl ~I..,? .:.,l\_rijl o.I~ .:\_,..c ~ \_, ...J \_, \_r->-ll,. J-"~ L........;uL;.. L.:....).... .:.,~ ... '"4- *<sup>J</sup>*<Li[.....S I J-'bl..J; .:., I \_;--i.J <sup>I</sup> .~'ti J..=.1..JI L..:.1 <l.JI ..::..~I.:.,~ .~\_:i j..c. d....1\_,.i.J \_pi
Reading the Qur'an from the pages of the book indicates a command, a prohibition, honor and joy. It is said that if someone dreams that he read the Qur'an in public, he is a man who ... will be faithful and submissive to what is fair and will forbid what is reprehensible, according to the saying of God Almighty: "They recite the revelations of God in the watches of the night and fall prostrate <before Him>" [Qur'an 3: 113].
<sup>108</sup> Al-Nabulusl, vol. 2, pp. 31 ff., s.v. ;;'.)L.....:.. Al-Nabulusi's entry on prayer begins like that in *al-Muntakhah.*
<sup>109</sup> In addition to the five obligatory daily prayers Cm/a) there is the *du'a* (lit. "calling"), an individual spontaneous prayer, and the *dhikr* (lit. "remembrance" ) or invocation to God.
<sup>110</sup> Al-Dlnawarl,fa\$18, *bah* 20; *Esad Efendi* 1833, fol. 78b; BN *arahe* 2745, fols. 126a-127a. Repeated in al-Nabulusl, vol. 2, p. 135, s.v. \_:,1\_,\_j.
<sup>111</sup> Al-Nabulusl, vol. 2, pp. 135-36, s.v. \_:,]\_,........The beginning of this chapter in al-Nabulusl is copied from al-Dlnawarl,.fa\$1 8. *bah* 20 (~).Interpretations other than those of al-Dlnawarl begin on p. 136.
(f someone dreams that he memorized the Qur'an but has not memorized it <in real life>, he will become a ruler ... . The recitation of the Qur'an signifies an abundance of good deeds and elevation of one's rank). And: من قرأ القران او شـيـتُـا مـنـه فى مـنـامـه نـال ر فـعـة وعـزا ... وان قـر أ الـقـر أن بـصـوت حـسن نال عزا ورفعة وشهرة (If someone recites the Qur'an or a passage from it in his dream, he will attain high rank and power ... And if he recites the Qur'an in a beautiful voice, he will attain power, high rank and a fine reputation).112
In the Oneirocriticon reading from books without further specification is interpreted as wisdom:13 ò αύτος έαν ίδη, ότι τω λαώ ύπαναγινώσκευ βίβλους, ούτος σοφός έσται. εί δε βασιλεύς τουτο ίδη, τροπαιούχος έσται και σοφός. (If the same person dreams that he is reading books to the people, he will become wise. If an emperor dreams this, he will be victorious and wise). Reading the Qur'an is interpreted as wisdom in Arabic dreambooks, القرآن حكمة فمن راى انه قرأ according to a passage from Ibn Qutayba:144 من المصحف او نَشَره فَإن ذلك حكمة ياتى بها او يلتمسها (The Qur'an is wisdom. If someone dreams that he is reading the Qur'an from a book, or that he is disseminating it, 115 this is indeed wisdom that he will achieve or that he will seek). The same interpretation is repeated in Ibn Shāhīn (no. 449) (قــبـل من راي انه يقرأ شـيـــَـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ بكورن حاكما ان كان لائقا (It is said that if someone dreams that he is reading something from the Qur'an, he will speak words of righteousness. Ibn Sirin said that he will become wise, if he is worthy of it).
Although chapters 11 and 139 agree that a woman's ordination as priest or
<sup>12</sup> Al-Nābulusī, vol. 2, p. 136, s.v. قراءة. Al-Nābulusīʾs interpretation of reciting the Qurʾān is the opposite of his interpretation of the Gospel indicates giving false testimony, as opposed to reciting the Qur'an which guarantees the truth (vol. 2, p. 136): من قــرأ الـقــرآن ... If someone recites the او شـيـئـا مـنـ ذو ات الـشـهـادات شـهـد بـالـمق... Qur'an or a passage from it in his dream ... , if he is about to serve as a witness, he will give accurate testimony).
<sup>113</sup> Chap. 12, Drexl 9, 14-16.
<sup>114</sup> Ibn Qutayba, hab 9, Yahuda ar. 196, fol. 27b.
<sup>115</sup> The verb nashara, here rendered as "disseminated," rarely has this meaning. Its primary meaning is "unrolled," but the word for "book" (mushaf) used in the same phrase refers to a codex, not a scroll.
performance of priestly duties foretells her public disgrace, chapter 149 in Vat. gr. 57316 says that a woman who dreamt she became priest (ἐγένετο ίερεύς) and performed the liturgy together with a priest (συνελειτούργει ispel) will die. The ultimate source of this interpretation is probably Artemidoros; 117 it appears in three of the five Arabic dreambooks examined. Al-Dinawari states: فان راي ان المراة تؤم بالناس مـاتت لان المراة لا تتـقـدم الناس الاعبد الموت) (If he dreams that a woman is leading the people in prayer she will die, because a woman never stands in front of people except at ان رات امــر اة : death). Al-Muntakhab (p. 50) quotes a similar interpretation a) كـأنـهـا تـؤم بـالـرجــال مــاتـت لان المراة لا تـتــدم الـرجــال الا فـى الموت woman dreams that she is leading men in prayer, she will die, because a woman never stands in front of men except at death). Al-Nābulusī repeats it as وان راي امراة انها تؤم الرجال فانها تموت لانها لا تصلح للامـامـة well: فلا يكون ذلك الّا عند الموت تتقدم امامهم وهم يصلون عليها (If someone dreams that a woman is leading men in prayer, she will die, because it is not proper for a woman to lead the prayer <of men> and such a thing never happens except at death, when she is placed before them and they pray for her). A possible explanation of this inconsistency is that the woman's ordination in chapter 11 was adapted from an Arabic passage interpreting the significance a woman standing on the minbar and delivering the khutba. On the other hand, the ordination of a priest and a woman officiating at the liturgy in chapter 149 must have been adapted from an Arabic interpretation of leading the prayer of a congregation.
An Arabic counterpart to the anecdote related in chapter 139 of the Oneirocriticon could not be found in any of the five Arabic dreambooks used for comparison with the Greek text. But a similar anecdote is related in a small Arabic dreambook of unknown date that is (possibly falsely) attributed to the Persian mathematician and astronomer 'Umar al-Khayyam (1048-1125), who is best known in the Western world for his collection of poetry called the
<sup>116</sup> Fols, 162v-163r.
<sup>117</sup> ii.30; Pack 153, 12-14: πασα δε ίερωσύνη και πασα άρχή, ής μὴ μέτεστι γυναικί, εὰν ύπολάβη γυνὴ ίερατεύειν ή αρχειν, θάνατον αυτή προαγορεύει (If a woman dreams that she is a priest or holds any priestly office or magistracy not open to a woman, it means that she will die). In the translation of Ḥunayn b. Isḥāq (Fahd, Artémidore d'Éphèse, p. 277, 14-15): ف . الكهانة وجميع الرياسات التى لا يصلح ان تكون للنسا فان المراة ان راتها دلّت على موتها
<sup>118</sup> Al-Dīnawarī, fasl 8, bāb 18, Esad Efendi 1833, fol. 76a; BN arahe 2745, fol. 125b.
<sup>119</sup> Al-Nābulusī, vol. 1, p. 20, s.v. امام الصلاة .
Ruba iyyat .120 Chapter 139 of the Oneirocriticon relates the following (Drexl 92, 1-7):
Γύναιόν ποτε πρός τον όνειροκρίτην Σηρείμ έλθον είπε: τί μοι έσται, ότι είδον κατ΄ όναρ ταύτη τη νυκτί έμαυτην ώς ίερέα ήτοι προσευχίτην προσευχομένην ύπερ του λαου; ό δε είπεν έν ποία ώρα τουτο έθεάσω; ή δε έφη ένάτη. και άπεκρίθη: ότι τρεῖς μῆνες ἔσονται και χωρισθήση του συζύγου σου καὶ ἐκ πορνείας συλλήψεις και τέξεις.
A woman once came to the dream interpreter Sereim and said: "What will happen to me, for tonight I dreamt of myself as a priest, that is, one who says prayers, praying on behalf of people." He said: "At what hour did you see this?" She said: "The ninth." He answered: "Three months will pass, and you will be separated from your husband and will conceive from fornication and give birth." And it happened so.
A corresponding anecdote is narrated in 'Umar al-Khayyam's chapter on the interpretation of "Elevated Places and High Locations" :121
واتى رجل إلى ابن سيـرين فـقـال رايـت امـراة تخطب على المنبـر والـنـاس حوله، فقال إن صدقت رؤياك لتفضحن هذه المراة على روؤس الناس.
A man came to Ibn Sirin and said: "I saw a woman who was delivering a sermon from the minbar and the people were around her." Ibn Sirin said: "If your dream is truthful, this woman will be disgraced in public."
As soon as we realize that the anecdote in chapter 139 is an adaptation of a similar Arabic anecdote about a woman delivering the sermon from the minbar, we can also account for the puzzling position of the Greek anecdote in the Oneirocriticon. The content of chapter 139 seems unrelated to the content of the chapters immediately surrounding it, as it is tucked between chapter 138, "From the Persians and the Egyptians on Emesis," and chapter 140, "From the Indians on Purgatives." Further on, chapter 141 is entitled "From the Persians and Egyptians on Purgatives," chapter 142 "From the Indians on the Ownership of Land and Houses," and chapter 143 "From the Persians and Egyptians on the Ownership of Lands and Houses." These chapters interpret owning high buildings, but also hills and mountains. 'Umar al-Khayyam's anecdote about the woman who delivered a sermon from the minbar was included in a chapter on "Elevated Places and High Locations," which interprets dreaming of mountains, steep roads, minbars and canopies and states that "every mountain is a
<sup>120 ·</sup> Umar al-Khayyām (attributed to), Ta bīr al-manām (Cairo, 1991),
<sup>121</sup> Ibid., p. 17.
*minbar"* (~ ~ JS). It is therefore possible that chapter 139 of the *Oneirocriticon* was originally attached to chapters 142 and 143 and was displaced by the Greek author, who either did not understand the logic of arranging the chapters or was careless about preserving the order of the interpretations found in his Arabic sources. Or the dislocation might have occurred in the course of the transmission of the Greek text, since the relationship of the anecdote to the interpretation of elevated places was no longer obvious, the *minbar* having disappeared from the Greek text.
Chapter 149, "On Priests," followed by chapter 150, "On Holy Icons," and 151, "On Trees and Plants," is found after the chapter called "From the Persians and the Egyptians on Buildings," the last paragraph of which discusses dreams about pagan temples. Chapter 149 is based on the Arabic interpretations of dreaming about an imam. Arabic dreambooks generally concentrate on dreams concerning leaders (caliphs, sultans, kings, imams, etc.) toward the end of the religious chapters at the beginning of the book, though they can also be found in the chapters on various human activities. 122 The *Oneirocriticon* has a few interpretations based on Muslim interpretations pertaining to imams in its religious chapters, but the interpretation of the significance of emperors in dreams (chapter 127) is placed near the chapters on activities associated with war, such as decapitation and slaughter, as is done in *al-Muntakhab.*
The appearance of the chapter on priests in a third section of the *Oneirocriticon* unconnected to the other two has two possible explanations. Either the Greek author used three different Arabic dreambooks (which would be consistent with his claim that he relied on three different kinds of sources), each with a different arrangement; or he inserted a chapter on priests after the interpretations of pagan temples on his own initiative. Perhaps the preceding discussion of pagan temples reminded him that he had neglected to dedicate a chapter to priests. Priests in tum reminded him to insert a chapter on icons, the material for which was furnished by the first, religious, chapters of his Arabic sources. Chapter 149 contains nothing that was invented by the author of the *Oneirocriticon,* since almost every paragraph on the interpretation of priests (Drexl 103, 25-105, 11) corresponds to an analogous paragraph found in the surviving Arabic dreambooks:
Dre xi 103, 26-104, 2:' Eav ni; 'io11 Ka't' ovap, on ttcriiA8£v tEpEili; EO"'toAtcrµ£voi; EV 1:07tq:>, OU OUK ~v Tlmoi; Eicr£pxrnem UUTOV, di; wui; KUTOlKOUVTai; EV T<\l
<sup>122</sup> Cf. chap. 30 in *al-Muntakhab,* which is between a chapter on clothes (29) and a chapter on war (31).
#### CHAPTER SEVEN
τόπω θλιψις και φόβος από έξουσιάζοντος διακρίνεται όμοίως και έξουσιαστὴς έαν είσηλθεν έν τόπω τινὴ παρὰ τύπον, είς θλιψιν καὶ φόβον τῶν οἰκητόρων κρίνεται, άλλά ἐλάττονα.
If someone sees in his dream that a priest in his priestly attire entered a place where it was not customary for him to enter, it is interpreted as sorrow and fear coming from a magistrate to the inhabitants of that place. Likewise, if a magistrate enters a place contrary to regular <custom>, it is interpreted as sorrow and fear for the inhabitants, but to a lesser degree.
Cf. Ibn Qutayba, Yahuda ar. 196, fol. 27a-b:
فان راي الامام دخل دارا او محلة او قرية ينكر دخوله // منها اصاب اهل ذلك الموضع مصيية عظيمة وكذلك السلطان دونه.
And if he dreams that an imam entered a house or a quarter of a city or a village, and his entry there is disapproved of, a great misfortune will befall the people of that place. And <the entry of> a sultan is interpreted likewise, but <the calamity> is less than <in the case of an imam>.
Cf. also Ibn Qutayba, Yahuda ar. 196, fol. 5b, 1. 2 (Ankara Is. Saib Sincer I, 4501, fol. 189b):
وكاللك يرى فى الملة او البلدة او الدار وقدرها يصغر عن قدره وينكر دخول مثله دعبي ذلك تالصحينية والذل بنال اهل ذلك الموضع لقوله ان الطوك اذا
... and likewise, <if> a king is dreamt of <as being> at a place or a land or a house the grandeur of which is inferior to his own grandeur, and the entry of someone like him <in this place> is frowned upon, <this> is interpreted as a calamity and disgrace that will befall the people of that location according to the saying of God, "Lo! kings, when they enter a township, ruin it and turn the honor of its people into shame" [Qur'an 27:34].123
Drexl 104, 2-5: έάν τις ίδη, ὅτι εἰσῆλθεν ὁ ίερεύς καὶ ὕπνωσεν εἰς τὴν στρωμνὴν αύτου, συμφιλιάσει τώ ίερει έκείνω και ό ίερεύς δολιευόμενος ἀναβήσεται τῆ γυναικί του ίδόντος, άλλα και κληρονομήσει αύτον.
If someone sees that the priest entered and slept in his bed, he will befriend that priest and the priest will treacherously mount the wife of the dreamer, but <the dreamer> will also receive an inheritance from <the priest>.
<sup>123</sup> The same interpretation is recorded in Ibn Shahin, no. 1127.
Cf. al-Dinawari,fa,1'19, *bah* 2; *Esad Efendi* 1833, fol. 67b; BN *arabe* 2745, fol. 148a:
~J il.~ iuJ *b\_;:....,....* ~ ~ ...Jl:J ~ il.~I t--° L...:,L..:, 4........i.:i ':?l.J .ju ~.UL.~ ~J ~ <~ ~> \*\* w~ iL.~1 ..b..J~ <t....:.Ll ~L..:. ~ ..JIJ \* ~L.. ~J .jL. .•. :dl b~J ljl ~IJ ~l.5: ~L:JI ..)~ ....:;L... JI....::.~ .j:.1~ ~ ~L..:, ....:,I lSl.J ..JU .... 4..JL.. ~ .JL...:. L...'..L. ~ ~J il..~I ~I JI bl\_\_r-ol ....:,J.l.:.,..... JI iL..~1.:.,..... ~ ....:,U UJ.J--'--0 .j:.1.;--i.JI ..JL.S:J iL..~1 .b~J .j:.1\_,.i.JI d.Jj .J.i..i..;> ~.J4
If someone dreams of himself sleeping under the same bedcovers as an imam without a curtain between them, and the dreamer gets up in reverence for the imam, but the imam remains <in bed> sleeping, the dreamer will become associated with the imam and <\*\*>124 and his money will end up in the dreamer's hands, either during the imam's lifetime or after his death, because someone who is asleep is like someone who is dead, and whenever people find someone dead they also find money. But if the imam is awake and the dreamer remains asleep, the imam will receive <the dreamer's> money. If someone dreams that he is sleeping in the bed of the imam, if the bed is known <to him>, he will receive from the imam or someone else a woman or a slave girl analogous to the bed and his own importance.
Drexl 104, 5-9: £av tOl:J *·w;,* on £n6mos \.i::pi::us arc£8avi::v, Eis amoA.i::tav -rf\s TCtCH£(J)s i:ou \.i::p£ros E<Hl v ii Kpi.crts. fov <sup>0</sup> l01] \.i::pfo, on VOC5£l, v6crov -rf\s rci.cr"t£(J)s aui:o\J KplVE"t(J) Kat <1>av£procrtv Kat µaKpoi;;roi.av aui:o\J Kat Uy£tav.
If someone dreams that a local priest died, the interpretation refers to the loss of the priest's faith. If he dreams that a priest is sick, let him interpret <it> as weakness in the priest's faith and disclosure <of this fact[?]>, as well as longevity and health.
Cf. *al-Muntakhab* p. 147, line 10:
~ti-: J\_.L;.. ...:.~J ~I dLl ~ ~ ~J <Lo.l.J:. ~.i iL.~1 ~..>-"J ~.J .J..:!.lJ ~.i ~J ~~J *b\_,.J* ~l:.LI ~ bl:!I J4\_rll J...=..J ~ ·()l..:. •4.) ~.:.,.....
Illness of an imam is an indication of his wrongdoing, but his body will be sound throughout that year. His death signifies something harmful that will occur in his
<sup>124</sup> One word is incomprehensible in the manuscripts. BN *arahe* 2745, fol. 148a, has ~ (from the verb .l.-.>-.7 =to be or become curly) which makes no sense in this context. Possibly, the intended verb was ~ ~ (in which case the sentence should be translated "he will become associated with him and the association will be pleasant for him" ).
#### CHAPTER SEVEN
kingdom, and the people with him will carry upon their shoulders the vehemence of his rule, and his religious faith will be weakened, while the faith of his flock from throughout the region will be sound.
Drexl 104, 10-21: Έαν ίδη τις, ότι ή κεφαλὴ του ίερέως ἐγένετο μείζων, ή ἐξουσία αὐτοῦ μείζων ἔσται κατὰ τῶν οἰκητόρων τῆς γῆς. εἰ δὲ τὸ πρόσωπον του ίερέως έγένετο μείζων, είς τους λόγους του ίερέως δύναμιν και φόβον κρινέτω. εί δε έν τφ τραχήλω αύτου πάχος και ίσχυν ίδη, έσται δυνατός είς τας θυσίας αύτου, εί δε είς τὴν χειρα αὐτου μῆκος καὶ ἰσχύν, εὐρήσει ό ίερεύς δούλους και βοηθούς δυνατούς. εί δὲ ϊδη, ὅτι ὁ ίερεύς μεγίστην κοιλίαν ἐκτήσατο, εὐρήσει ούτος ὁ ίερεύς φαμιλίας καὶ πλοῦτον πολὺν καὶ συνέλευσιν των συγγενών αύτου. εί δὲ οί πόδες αὐτου ἐπαχύνθησαν καὶ ἐστερεώθησαν, και ούτος εύρήσει χρυσίον και δούλους βοηθούς όμοίως. ἐαν Ίδη ταῦτα ἐπὶ τὸ χειρον, απ' έναντίας νοείτω τα πράγματα.
If someone dreams that the priest's head becomes larger, <the priest's> power over the inhabitants of the land will become greater. If the face of the priest becomes larger, let this be interpreted as power and awe in the priest's words. If <the dreamer> sees thickness and strength in the priest's neck, he will be powerful in his sacraments. If [the dreamer] sees length and strength in the priest's hand, the priest will acquire strong slaves and assistants. If he sees that the priest had a big belly, this priest will acquire a large household, much wealth and an assembly of his relatives. If his feet become fatter and steadier, he will find gold and likewise servants who will assist him. If he sees <all> these deteriorating, let him understand that things <will develop> in the opposite way.
Cf. Ibn Qutayba, Yahuda ar. 196, fol. 27b:
وما ر ءي من جوار حه من فضل فذلك قوته في سلطانه وما راي في بطنه من فضل وعظم فذلك زيادة فى ماله وولده واهل بيته والنقصان بحسب ذلك.
And whichever limb of the priest is seen as <becoming> excessive, this is the strength <of the priest> in his authority. And if the dreamer sees the priest's belly becoming bigger and stronger <it indicates> an increase in the priest's wealth. offspring, and <number of> people in his household. And the decrease <of the limbs is interpreted> by analogy to that.
Cf. also al-Dīnawarī, fasl 9, bāb 3; Esad Efendi 1833, fol. 99a; BN arabe 2745. fols. 150b-151a:
ان راى فى اعضاء الامام خاصة زيادة بقدر ما يمنعه عن منافع سائر جسده فهو خير وان راي فيها وهنا او نقصانا او کسرا فانه نقصان فيما ينسب اليـه ذلك الـعضـو وندامـة فـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ سلطانه... فان راى غلظا فى عنقه فهو قـوته فى عدله وانصافه وهزيمته لاعدائه... فان راى فى يديه سـمنا وقوة فـانـه قـوة دينـه و اسـلامـه ... فـان راى في بطنه عظما فهو زيادة في اهل قوة وباس فان راى ان فيبهما <=الجسد و البطن › عظما فهو زيادة مـالله... فــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ بکون کٿنر المال حسث ادرك
If someone dreams of a particular increase in the bodily parts of the imam to a degree that prevents the use of the rest of his body, it is a good thing. If he dreams of a weakness, decrease, or breaking <in any of the imam's bodily parts> this means a decrease in whatever refers to that member, as well as regret. If someone dreams that the head of the imam becomes bigger, this means leadership and strength in the imam's authority .... If he dreams that the neck of the imam becomes thicker, this means strength in his impartiality, fairness, and defeat of his enemies .... If he dreams that the imam's hands becomes fatter and stronger, this means strength in the imam's religious faith and submission to God (Islam) .... If he dreams that the imam's belly becomes bigger, this means an increase in his household, power, and fortitude. If he dreams of an increase in both the belly and the body of the imam, this means an increase in the imam's money ... If he dreams that the feet of the imam become made of lead, this represents the greatest possible increase in his money.
Drexl 104, 26-29: ἐαν ϊδη τις, ὅτι ὁ ίερευς περιπατεῖ ἐν τόπω σκοτεινῶ, νοείτω, ότι έν σκοτεινοίς άμαρτάνει πρός τον θεόν εί δε περιπατεί έν τόπω φωτεινώ και χλοάζοντι, έν φωτι εύαρεστεί τω θεώ και άγαθοεργεί.
If someone dreams that the priest is walking in a dark place, let him reckon that <the priest> is secretly sinning against God. If he is walking in a bright and verdant place, < the priest> is openly pleasing God and performing good deeds.
Cf. Ibn Qutayba, bab 46 ( ) النوادر ) = "rare interpretations" or "odd interpretations" ), fol. 61a:
Light in dream interpretation is true religion and correct religious guidance (huda), while darkness is straying from the truth and from the right path. God Almighty said: "God is the protecting friend of those who believe. He bringeth them out of darkness into light" [Qur'an 2:257], that is, from the wrong into the right path.
#### 304 CHAPTER SEVEN
#### *Monastic Tonsure and Monks*
The *Oneirocriticon* interprets monastic tonsure in two passages. The first is in the last paragraph of chapter 11, "From the Indians on Prophets, Apostles and Teachers." It states that an emperor who dreams that he has been tonsured and dressed as a monk will soon lose his imperial status and die; if a commoner "suffers such a fate in a dream" (o *wino* 8rncraµEvo~ 1mi na8ffiv), he will die in poverty and sorrow; a married woman will be widowed, while a celibate woman will die celibate. <sup>125</sup>Though it may be surprising to the reader of a Christian dreambook that the adoption of monasticism, the only social condition primarily devoted to God, is tainted with negative associations, the interpretations contained in the *Oneirocriticon* nonetheless reflect some Byzantine realities. Regardless of the degree to which monastic vows were kept, upon adoption of the habit a novice took vows of poverty. A number of individuals, both men and women, adopted the monastic habit after being widowed. Fallen emperors or disgraced aristocrats and their family members were often forced to withdraw to monasteries (instead of being executed or sent to prison) as a means of eliminating them from the political arena. <sup>126</sup>Seen from this point of view, negative interpretations of monastic tonsure are understandable.
Chapter 30, "From the Indians on Tonsure," combines a new element with a variant of chapter 11. It repeats the negative interpretations of monastic tonsure that apply to emperors and women, but for a commoner such a dream signifies repentance, relief from sorrow and sin, and embellishment of faith, 127 effects that one would normally expect to find in a Christian dreambook. In addition to monastic tonsure, a whole section of the *Oneirocriticon,* which is missing from Drexl's critical edition but exists in *Vat. gr.* 573 (fol. 163r), discusses dreaming of monks at the very end of chapter 149 ("On Priests"). It states that whatever a dreamer hears from a monk who appears with crosses on his habit is accurate. Being beaten by such a monk means receiving grace from a saint. An old monk without crosses on his habit is misfortune. Talking with him or receiving something from him indicates acquiring something moderately good. A young monk without crosses on his habit is interpreted as
<sup>125</sup> Drexl 7, 21-26.
<sup>126</sup> In the course of the 9th and 10th centuries, such was the fate, for example, of Emperor Michael I Rangabe (r. 811-13), who abdicated in favor of Leo V and became a monk on the Princes' Islands, where he died in 844; also of Emperor Romanos I Lekapenos, who became a monk after he was deposed in 944 and died as a monk in 945.
<sup>127</sup> Drexl 20, 20-26.
illness or an enemy. If the dreamer beats the monk, he will make a votive offering and address a request to a saint. These interpretations present the reader with a mixture of the predictable with the unexpected and raise the question of whether they are the product of the Greek author's imagination or were adapted from interpretations of analogous Muslim institutions found in his Arabic sources.
Islam expressly forbids monasticism, the invention of which the Qur'an attributes to Christians. Christian monks were a familiar sight in pre-Islamic Arabia, and at least one of them, the Nestorian Bahirā, whom the Prophet met on his trip to Syria, figures prominently in Muslim tradition because he is said to have been the first human who acknowledged Muhammad as a prophet, even before his prophetic mission began. Of course, Muslims were familiar with monasticism because monasteries existed and dhimmī monks actually lived in Muslim lands. Al-Dinawari interprets dreaming of Christian monks as follows:128
```
قال المسلمون من راي انه راهب فانه صاحب بدعة قد افرط فيها وقالت
النصاري من راي كانه تحول قسا او راهيا فانه يورثه ٿنا حسنا لکن يعسر
عليه حاله و يضيق عليه رزقه وربما يقع التاويل لغيره …
```
The Muslims said: "If someone dreams that he is a monk, he is a heretic who exceeded the proper bounds <and embraced> heresy." The Christians said: "If someone dreams that he became a priest or a monk, he will inherit something nice, though his circumstances will become straitened and his daily sustenance will be insecure. Possibly, the fulfillment of this dream will benefit those around him."
Further on, al-Dīnawarī adds:129
فان راي انه راهيا ومبر فانه صاحب بدعة وهو مفرط لقوله تعالى و رهيانية اتتدعو ها الانة.
If someone dreams that he is a monk or a pious person, he is a heretic and a transgressor, according to the saying of God Almighty: "But monasticism they invented-We ordained it not for them-only seeking God's pleasure, 130 and they observed it not with right observance. So We give those of them who believe their reward, but many of them are evil-livers" [Qur'an 57:27].
<sup>128</sup> Al-Dinawari, fasl 8, bah 78; Esad Efendi 1833, fol. 65b; BN arahe 2745, fol. 144a.
<sup>129</sup> Al-Dinawari, fasl 8, hāh 80; Esad Efendi 1833, fol. 65b; BN arahe 2745, fol. 144a-b.
<sup>130</sup> Cf. W. M. Watt, A Companion to the Qur'an (Oxford and Rockport, Mass., 1994), p. 256: " sc. they invented monasticism in their strong desire to be pleasing to God."
Similar interpretations are quoted in al-Muntakhab: 31
فان راى انه صار راهيا فانه مبتدع مفرط فى بدعته لقوله تعالى ورهبانية ابتدعوها وقيل ان صاحب هذا الرؤيا يضيق عليه معاشه وتتعسر عليه اموره ويصحبه في جميع الامور ذل وخوف ورهبة لا تزايله ويدل ايضا على انه مكار خداع كياد مبتدع داع إلى بدعته وبالله العياذ من ذلك.
If someone dreams that he had become a monk he is a heretic who exceeded the proper bounds <and embraced> heresy, according to the saying of God Almighty: "But monasticism they invented" [Qur'an 57:27]. It is said that the dreamer's livelihood will be straitened, his affairs will become adverse, and all his dealings will be associated with ignominy, fear and terror without end. < This dream> also signifies that the dreamer is an impostor, a crook, a deceiver, and a heretic who invites others to his heresy, may God protect us from such a thing.
Ibn Shāhīn quotes Abū Saʿīd al-Wāʿiz, and therefore partly coincides with al-Muntakhab, since Abū Sa'id al-Wa'iz was used as a source for both al-Muntakhab and Ibn Shahin's dreambook, but also adds (nos. 5420-5422):
( ... من راى انه صار راهبا… ربمادل الرؤيا على ارتكاب ما لا يجوز له واستمراره عليه. (١٤٩) وقال بعض المعبرين من راى انه صار راهبا وكان من التقات فإنه يؤول بكثرة الخشوع والخوف من الله تعالى لقوله عز وجل واضسم إليك جناحك من الرهب وهو خوف. (٥٤٢٢) وقـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ بعض الصالحين الراهب من رهب الله أى خافه، وقيل رؤيا الراهب برجل مكار خداع مبتد ع.
(5420) ... If someone dreams that he became a monk ... possibly the dream signifies his perpetration of something forbidden to him and his persistence in it. (5421) Some of the dream interpreters say: "If someone dreams that he became a monk and was one of those worthy of confidence, 132 this is interpreted as abundance of his submission to and fear of God, according to the saying of God Almighty: "and guard thy heart from fear" [Qur'an 28:32], and the monk <symbolizes> fear." (5422) Some of the righteous say that a monk is someone in awe and fear of God. It is also said that dreaming of a monk <indicates> an impostor and crook who is a heretic.
Al-Nabulusī repeats the interpretations of the previous dreambooks. 133 He also
<sup>131</sup> Al-Muntakhab, chapter 46, p. 331.
<sup>132 &</sup>quot;Worthy of confidence" (thiqa) is a term used in the study of the hadith, when referring to a reliable transmitter of these traditions.
<sup>133</sup> Al-Nabulusi, vol. 1 , p. 250, s.v. راهب.
interprets monasteries 134 and monastic cells 135 in separate entries:
$\omega\_{\vec{\omega}} : \vec{\omega} \perp \vec{\omega}\_{\vec{\omega}} \perp \vec{\omega}\_{\vec{\omega}} \perp \vec{\omega}\_{\vec{\omega}} \perp \vec{\omega}\_{\vec{\omega}}$ $\vec{\omega} \perp \vec{\omega}\_{\vec{\omega}} \perp \vec{\omega}\_{\vec{\omega}} \perp \vec{\omega}\_{\vec{\omega}} \perp \vec{\omega}\_{\vec{\omega}}$ $\vec{\omega} \perp \vec{\omega}\_{\vec{\omega}} \perp \vec{\omega}\_{\vec{\omega}} \perp \vec{\omega}\_{\vec{\omega}} \perp \vec{\omega}\_{\vec{\omega}}$
The monastery: its vision in a dream is like the vision of a church. Perhaps a monastery signifies the end of sorrows and adversities and redemption from afflictions. If the dreamer is sick he will die.
صومعة : هى فى المنام دالة على الخلوة وحسن السيرة والعزلة والانقطاع وقطع اللذات وطلاق الازواج وهجر الاخوان وربما دلت رؤيتها على الامراض وترك الشهوات من المأكول والمشروب وان كان الرائي مريضا مات او حصن خصوصا ان رآها ملك كذلك فى منامه وكان يقصد عدوا وربما دل ذلك على قهره وغلبه وتدل الصومعة على الغربة والوحشة والمقاطعة للاصحاب وتدل رؤيتها على الاخفاء والصومعة تدل على السلطان والرئيس ومن له ذكر فما اصابها او نزل بها من هدم او سقوط او غير ذلك عاد تاويله على من دلت عليه .
The monastic cell: dreaming of a monastic cell signifies seclusion, excellence of demeanor, retirement, separation, deprivation of pleasures, divorce from one's spouse and forsaking one's brethren. Possibly this dream signifies sickness and abandonment of the desire to eat and drink. If the dreamer is sick he will die or will be protected against attacks, especially if a king dreamt of a hermitage when he was about to meet his enemy. Possibly this dream signifies <that the king> will vanquish and conquer <his enemy>. The hermitage signifies exile, loneliness, and separation from one's friends. Dreaming of it also signifies concealment. The hermitage signifies a sultan, a leader, and somebody famous. The interpretation of whatever happens or befalls the hermitage, such as destruction or ruin or something like that, refers to whoever is signified by it.
In the first three Arabic dreambooks quoted, the predominantly negative interpretation of monasticism as heresy clearly reflects a Muslim point of view, while its association with terror derives from the fact that the Arabic words for "monk" (rahib) and "terror" (rahba) have the same radical consonants (r-h-b). Only al-Nabulusi's interpretation of monasteries, which are not discussed in the Oneirocriticon, associates monasticism with (among other things) death, as does the interpretation of monastic tonsure in the Greek dreambook. Evidently, verv little of what the Arabic dreambooks say is directly connected with the interpretations of tonsure and monks in the Oneirocriticon.
<sup>134</sup> Al-Nābulusī, vol. 1 , p. 205, s.v. د.
<sup>135</sup> Al-Nābulusī, vol. 2, p. 48, s.v. صومعة.
#### CHAPTER SEVEN
The negative interpretations of tonsure are not limited to the passages about cutting one's hair when one becomes a monk or nun. Tonsure is interpreted negatively whenever it is mentioned in the Oneirocriticon. Chapter 22, "From the Egyptians on the Significance of Hair," states that if a woman dreams that she cut her hair, her husband will die of a grave illness or in battle (xxx) } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } } Ίδη ταυτα γυνή <= στι έκοψεν έκ τών τριχών αύτης>, ό άνηρ αὐτῆς εἰς όξου νόσημα ή είς πόλεμον τελευτήσει).136 Chapter 31, "From the Persians and the Egyptians on Tonsure," gives additional interpretations:
' Εάν τις ίδη, ότι έκουρεύσατο τὴν κεφαλὴν αὐτου διόλου, εἰ μέν ἐστιν ὑπεξούσιος, άπολέσει τον ἐξουσιάζοντα αὐτὸν καὶ εἰς θλῖψιν μεγάλην ἐλεύσεται ἐἰ δέ ἐστιν αὐτεξούσιος, ἀτίμως θεατρισθήσεται καὶ ἐν πτωχεία τελευτήσει. ἐαν δὲ ϊδη, ότι ό κουρεύσας βλέπων αύτον ἐμπροσθίως ἐκούρευσεν, οῦτος καὶ προγνώσεται τον τρόπον της άπωλείας αύτου εί δὲ όπισθίως ἐκούρευσεν αὐτὸν, αδηλος έσται ή απώλεια αύτου. 137
If someone dreams that he has shaved his head completely, if he is under someone's authority, he will lose his leader and suffer great sorrow. If he is his own master, he will be publicly dishonored and will die in poverty. If he sees that the person who cut his hair is standing in front of him, the dreamer will know the manner of his ruin; but if he sheared off his hair from behind, <the manner of> the dreamer's ruin will be unknown,
Further on in the Oneirocriticon, in chapter 265, "From the Persians and the Egyptians on Female Hair," we read the following:
' Εαν Ίδη γυνή, ότι ἐκείρατο ἀπ΄ ἄλλου τὰς τῆς κεφαλῆς αὐτῆς, ἀποθνήσκει ό ἀνήρ αὐτῆς ἢ χωρίζεται αὐτὴ εἰ δὲ ἴδη, ὅτι ἐκουρεύθη μερικῶς, ἀνὰ μέσον αντης και του άνδρος μάχη έσται χωρισμου. εί δὲ ϊδη, ὅτι κρατεί τις γνώριμος τας τρίχας αὐτῆς καὶ κόπτει αὐτὸς ὅπισθεν, οῦτος βουλὴν ποιήσει μετὰ τοῦ άνδρος αύτης λάθρα είς τον χωρισμόν αύτου εί δε τδη, ὅτι κρατεί τας τρίχας αυτης και κόπτει έμπροσθίως, φανερός γενήσεται παρ' αύτου ό χωρισμός διά το έμπρόσθιον της κουράς.
Εαν ίδη τις, ότι τὴν ἰδίαν γυναϊκα ἐκούρευσεν αὐτός, προαγωγὸς ἔσται αυτης και έν βουλή αύτου μοιχεύσει και χωρισθήσεται αύτης. ώσαύτως έαν ίδη, ότι άλλοτρίαν ἐκούρευσεν, καὶ αὐτὸς ἐπίβουλος ἔσται αὐτῆς καὶ εἰς χωρισμόν του άνδρος αυτής κρίνεται.
Γυνὴ ἐαν νοσήση ὀξέως καὶ ἴδη, ὅτι ἐκουρεύθη, αὐτὴ ἀποθανεῖται. εἰ δὲ νοσῆ ὁ ἀνήρ αὐτῆς καὶ αὐτὴ τοῦτο ἴδη, ἀποθανεῖται ὁ ἀνήρ αὐτῆς· ή γὰρ κουρὰ είς χωρισμόν κρίνεται. 138
<sup>136</sup> Drexl 17, 19-20.
<sup>137</sup> Drexl 21, 1-8.
<sup>138</sup> Drex1 217, 8-26.
If a woman dreams that someone else cut the hair on her head, her husband will die or be divorced from her; if she dreams that her hair was only partly cut, a quarrel resulting in separation will occur between them. If she dreams that someone she knows was holding her by the hair and was cutting it from behind, that man will secretly assist her husband in planning a divorce from her; if he was cutting the hair from in front, the divorce through that man will happen openly because of the frontal cutting.
If someone dreams that he himself cut his wife's hair, he will become her pimp and deliberately cause her to commit adultery: later, he will divorce her. In a like manner, if he dreams that he cut the hair of somebody else's <wife>, he will plot against her, and <the dream is to be> interpreted as separation from her husband. If a woman who is very ill dreams that she had her hair cut, she will die; if her husband is ill and she has this dream, her husband will die: for the cropping of hair denotes separation.
In view of the above excerpts from the Oneirocriticon, the negative interpretations of monastic tonsure in chapters 11 and 30 could have been based on material in the Arabic sources independent of any discussion of Christian monasticism. Indeed, a closer look at the two passages on monastic tonsure gives the impression that the negative interpretations they contain were translated twice from the same Arabic text. Their wording in Greek is different, but what they say is identical. Proof of their close relationship is the fact that the same English translation renders accurately the interpretation of female tonsure from both chapter 11 and 30. According to chapter 11 ("From the Indians on Prophets, Apostles and Teachers"),
Έάν τις Ίδη, ότι έκάρη και ήμφιάσθη ώς μοναχός, εί μέν έστι βασιλεύς. άποβαλεί τὴν ἐξουσίαν τοῦ λαοῦ αὐτοῦ καὶ τὴν βασιλείαν καὶ μετὰ θλίψεως άποθανείται ταχύ. ἐαν δὲ τῶν ὑπηκόων τις ή ὁ τοῦτο θεασάμενος καὶ παθών, ἐν πτωχεία καὶ θλίψει ἀποθανεῖται. ἐὰν δὲ γυνὴ ὕπανδρος ἴδη τοῦτο, χηρεύσει αγαμος δε έαν ή, αγαμος ἀποθανεῖται. 139
If someone dreams that he was tonsured and dressed as a monk, if he is an emperor, he will lose his authority over his people and his kingship and will soon die in sorrow. If the one who dreamt and suffered this was one of his subjects, he will die in poverty and sorrow. If a married woman dreams this, she will be widowed. If she is unmarried, she will die unmarried.
According to chapter 30 ("From the Indians on Tonsure"),
<sup>139</sup> Drexl 7, 21-26.
#### 310 CHAPfERSEVEN
, E<iv nc; 'i8n, O'tl EKOUp£UO"£ TI,v K£<\JaAi]v autou *We;* EV rimcp µovaxou, µ£i:avoljcr£t o{noc; Kat anoPaA£l tac; 0At\jl£tc; Kat aµaptiac; Kat Kocrµljon TI,v 7tt<mV autou. £i. 8£ Ecr'tl pacnA£uc;, anoPaA£l TI,v pacnA.tiav autou Kat t£A£Utljcr£t EV 0At\jl£l. £t OE yuvi] touto 'ion fucavopoc;, XTJP£Ucr£t, £i. OE ayaµoc;, Kat ayaµoc; t£A£utljcra <sup>1</sup> 40
If someone dreams that he cut his hair in the manner of monks, <sup>141</sup>he will do penance, be relieved from his sorrows and sins, and embellish his faith. If he is emperor, he will lose his kingship and die in sorrow. If a married woman dreams this, she will be widowed. If she is unmarried, she will die unmarried.
The overwhelmingly negative interpretations of hair cuttingwere clearly copied from the Arabic sources of the *Oneirocriticon.* Arabic dreambooks interpret only one kind of tonsure positively: the tonsure of pilgrims to Mecca. The rites of the pilgrimage are performed in the i~rdm, the pilgrim's sacred robe, and the pilgrim should abstain from a number of activities, including cutting his or her hair. 142 When the rites of the pilgrimage have been completed, the i~rdm is abandoned by symbolically cutting a lock of hair or completely shaving the head. Tonsure therefore signifies that the person has carried out one of the most important religious duties of a Muslim, and dreaming of it understandably signifies the "embellishment" of one's faith:
Cf. **lbn** Qutayba, *bab* 25, *Yahuda ar.* 196, fol. 43a-b:
..:.,µLA~ ~I \_,I~ \_,I Lp\_,\_; ..;:..,L. Ll\_,.h...... ~j..> ~.ly-11 d..> ..:.,I\_, ~ 0::-7J 4-:..,,....; 1~3 ~ ..:.LJj ..:.,LS i~I ~~I~..,? LA~~ ~ rJ i~I ~~I..,? d.Jj ..:.,LS \_,I~~ J"'-;:' i~ ~\_Jyl..,? ..:.,LS ..:.,I\_, 4-'.il..>J ..>--LA~~ L.:.L.......:.1 ..:.,I ..;:\_,I..>..:.,µ\* ~..ill .L....::..:i ..:.,LS\_, ~I\_, jl:JI y\_...u....., *..>--* o~ ..:.,Ll d.J:i\_,. ~~ .L.........:JI ..>--LA~ \_)I~\_,\_; l\_,\_c."'-;:' <t....:.µ ... I \_r'lllJ; dJ <sup>j</sup>..:.,LS
If a woman dreams that her head was shaved, her husband or guardian will die or her veil will be torn apart. If she cuts her hair at a time other than the holy months, <sup>143</sup>a quarrel and discord <sup>144</sup><will come> between her and her guardian. If
<sup>140</sup>Drexl 20, 20-26.
<sup>141</sup>For the various types of monastic tonsure, see *ODB,* s.v. "Tonsure."
<sup>142</sup>The others are sexual intercourse, paring the nails, cutting down a tree, and killing an animal.
<sup>143</sup>The holy months include the period of the pilgrimage. They are Dhii al-Qa'da, Dhii al-I:Iijja (during which the pilgrimage takes place), MuJ:iarram, and Rajab. They are lunar months and their dates in relation to the solar months vary.
<sup>144</sup>The word *nashz* used by Ibn Qutayba for "discord" has the legal connotation of the husband's or wife's violation of marital duties, specifically a wife's disobedience toward her husband or a husband's brutal treatment of his wife.
words are spoken in the dream, it signifies something good. If the dream occurs during the holy months, shaving or cutting <her hair> signifies, not something harmful, but accomplishment of her religious duties. If she dreams that someone is pulling her hair from behind her, he will invite her husband secretly to exchange her for another woman. If he is pulling <her hair> from in front of her, this will happen openly.
Al-Dīnawarī, fașl 6, bāb 37 (فى حلق شــعـر الراس); Esad Efendi 1833, fol. 50b-51a; BN arabe 2745, fol. 76b-77a;
حلق الراس آداء الامــانـة والامـن من الضـوف بـ وكـذلك جــزه وحلقـــه فى الـــج قضاء دين لقوله تعالى لتدخلون المسجد الحرام ان شاء الله امين محلوقين رؤسكم ومقصرين لا تخاقون وينال مع ذلك فتحا \* والتقصير امان من الخوف # فإن حلقه في غير المج فهو دون ذلك فى الصلاح # فإن كان صاحبه فى كرب او دين فرج عنه وقيل ان حلق فى غير الموسم وكان رئيسا غنيا افتقر \* واإن كان مديونا قضى الله دينه وربما دل ذلك على تهتك ستره وعزل رئيسه عنه بمكروه او بتوبة ۾ وإن کان مما يلبس السلاح فانه يذهب بطشه وهيبته وإن كان غنيا نقص ذهب ماله وإن كان فقيرا او مديونا قضيت دينه فان راي انه محلوق الراس فإنه يظفر باعدائه وينال قوة وعز لان النبى صلى الله عليه وسلم راى في المنام انه مقصر شعره واصحابه محلقون ر ؤسهم فظفر باعدائه فان حلق راسه فارنه يودى امانه وقالت النصارى من رای کانه یقطع شعر راسه فانه یسقط من جاهه و مرتبه فان رای کانه یحلق فانه يمرض مرضا و قال ارطاهيدورس اذا راي الانسان کان راسه محلوق فإن ذلك صالح
Shaving the hair from one's head <signifies> the pursuit of security and safety from fear. Likewise, shearing off or shaving the hair during the hajj indicates excellence in religion, according to the saying of God Almighty: "Ye shall enter the Inviolable Place of Worship, if God will, secure, having your hair shaven and cut, not fearing" [Qur'an 48:27]. The dreamer will also obtain victory. 445 Cutting the hair shorter signifies safety from fear. If the dreamer <sees that he> shaved his head at a time other than the hajj, his dream is not as auspicious. If the dreamer is in sorrow or in debt he will be relieved. It is said that if a rich leader dreams that he shaved his head at a time other than the hajj season he will become poor. If the dreamer is in debt God will release him from it. Perhaps this dream signifies the shamelessness of his excuses and the deposition of his leader by reason of something reprehensible or <his> penance.
<sup>145</sup> This interpretation is obviously inspired from the remainder of verse 48:27, which is not quoted by al-Dinawarī: فعلم ما لم تعلموا فجعل من دون ذلك فتحا قريبا (But he knoweth that which ye know not, and hath given you a near victory beforehand). In addition, "Chapter of Victory" (surat al-fath) is the title by which the 48th chapter of the Qur'an is known.
Al-Dīnawarī, fașl 6, bāb 42 (فى حلق المراة شــعـرهـا); Esad Efendi 1833, fol. 50a; BN arabe 2745, fol. 77b-78a:
من راي راس امراته محلوقا طلقها او مات او يفرق بعضهما عن بعض قبل الموت // فان رات ان جوزها چز شعرها او حلق راسها فهو حجسه لها في منزله الا تري ان الطاير اذا قُص جناحـاه بقى في وكتره ۾ و قـيل انهـا اذا حلقته تهتك سترها فان كان حلقها له و قصرها اياه على حال صلاح فى دينها وكان معه كلام يستدل به على الخير كان ذلك قضاء دينها و اذا امانه فى يدها \* وان رات كل ذلك فى الحرم ۾ فان دعاها انسان الى جز شعرها فانه يدعى زوجها الى غيرها من النساء سرا منها ويقع بينها وبين من راي شغب و قالت الروم من راي ذوابة امراة مقطوعة لم تلد ولد ابدأ وقال جاماسب من قطع شعره نقص قوته.
If someone dreams that the head of his wife was shaved he will divorce her or die, or one will be separated from the other before dying. If <a woman> dreams that her husband sheared her hair or shaved her head he will confine her to his house; indeed, do you not see that the bird remains in its nest after its wings are clipped? And it is said that if she shaved it her veil will be torn apart. If the shaving and shortening of her hair was for no other reason than for righteousness in religion <i.e., for the completion of the hajj>, and with it there were also words guiding her to goodness, this dream signifies the accomplishment of her religious duties, if she saw all this in the sacred precincts <of Mecca and Medina>. If someone appealed to her to shear her hair, he will appeal to her husband secretly from her to exchange her for another woman and there will be quarreling between her and whomever she saw. The Rum said: "If someone sees that the locks of a woman's hair were cut, <this woman> will never have a son."146 Jāmāsb447 said: "If someone <dreams that he> cut his hair his power will diminish."
The remaining three Arabic dreambooks repeat the same interpretations. 448 In addition, Ibn Shāhīn quotes Ibn Sīrīn in a paragraph (no. 1284) that closely corresponds to the contents of chapter 30 in the Oneirocriticon:
قال ابن سيبرين من راى ان حلق راسه فى ايام الحج فاإنه صلاح فى الدين وكفارة للذنوب و ان كان فى الأشهر الحرم او فى بعضها فإنه قضاء فى دين وزوال هم و غم ، وقيل ان راى ذلك ذو منصب فليس بمحمود ، وإن رات
<sup>146</sup> Such an interpretation is given neither by Artemidoros, nor by any of the surviving Byzantine dream books.
<sup>147</sup> A semi-legendary Persian writer.
<sup>148</sup> Al-Dīnawarī's interpretations are repeated in al-Muntakhah, pp. 80-81; also in al-Nābulusī, vol. 1, pp. 144-46, s.v. طق شعر الراس وغيره and vol. 2, pp. 8-9, s.v. شعر الراس.
LAI~,) *d..!* .:,)\_i, 4--..J~ J...=>.I \_,I LAI.)~..:;..,\_,....~ J~ ...:.Ll .ill:i o1 \_;-ii . ~ J~ *µ\_,* LAl.J-P. t--° ~~ ~ J~ ...:.µ ~ \_,I <sup>~</sup>
Ibn Sirin said: "If someone dreams that he is shaving his head on the days of the *hajj,* it signifies righteousness in religion and penance for his sins. If the dream occurs during the sacred months or thereabout, this signifies the completion of his religious duties and the end of his sorrows and troubles. But it is said that if a high dignitary dreams this, it is not good. If a woman dreams this, it signifies the death of her husband or one of her kinsmen. And if she dreams that she cut her hair, or part of it, off, she will have a dispute with her husband. And it is said <that this signifies> the occurrence of a calamity.
The Arabic interpretations quoted above explain the discrepancy in the interpretation of monastic tonsure dreamt by a commoner in chapter 11 of the *Oneirocriticon,* where it is said to signify poverty and sorrow, and its interpretation in chapter 30, where it is said to indicate embellishment of the dreamer's faith. Apparently the negative interpretations of chapter 11 were inspired by the secular interpretation of tonsure, which, in Arabic dreambooks, invariably signifies something negative. The positive interpretations of chapter 30 were apparently inspired by the Muslim tonsure connected with the pilgrimage, since the tonsure of both a Muslim pilgrim and a Christian monk is evidence of their pursuit of God.
The provenence of the interpretations of monks given in *Vat. gr.* 573 is more difficult to trace. Though Islam forbids monasticism as an institution and enjoins the faithful to marry, 149 a number of pious Muslims chose to live in religious seclusion and some became celebrated saints *(awliy<Y,* lit. "friends of God"). An Islamic counterpart to Christian monks are the *murabitun* ("the bound ones"), who lived in remote outposts *(rubut)* in order to defend the frontiers of the *dtir al-Islam,* the Muslim *oikoumene.* Renunciation of worldly possessions and desires is also practiced by the followers of Sufism, and the respect accorded to monks in the Christian world is enjoyed in Islam by the scholars of religion « *ulamti')* and jurisprudence *(juqahii'),* both of whom are distinguishable by their attire, just as monks are immediately recognizable by their habit. Arabic dreambooks interpret dreams about saintly figures, as well as religious scholars and jurists, and it is therefore conceivable that these interpretations were the source for the Christian interpretations of monks found
<sup>149</sup>For Quranic references, see M. Ali, *A Manual ofHadith* (London, 1978), pp. 266-68. For references to traditions of the Prophet that recommend marriage, see Hughes, *Dictionary of Islam,* s.v. "Marriage."
#### CHAPTER SEVEN
in the Oneirocriticon, which according to the text of Vat. gr. 573 (fol. 163r) are as follows.
ἐάν τις ἴδη ὅτι συνάντησε τινὰ μοναχὸν φορῶντα σταυρὸν ὑπογεγραμμένον ὡς μεγαλόσχημον, εί μεν ώμίλησεν αύτω γλυκέα και χρηστά, ίδου κρατείτω ούτως εί δε μὴ. τοὐναντίον. εἰ δὲ ἔτυψεν αὐτὸν ὁ τοιοῦτος μοναχός, εὑρήσει χάριν ὁ ίδών, είς τὴν ψυχὴν αὐτου παρὰ ἀγίου. ὁ γὰρ μοναχὸς ὁ σεσημειωμένος εἰς ανθρωπον του θεού κρίνεται. και άπερ τυφθη ἐξ ἐκείνου, δωρεὰ ἐστὶ ἀχίου τινός. Εί δε έστι μοναχός έκτος σημείου σταυρού ύπογεγραμμένου ή μεγαλοσχήματος μόνου ρασοφόρου, εὐ μὲν ἔστι γέρων, εἰς καιρὸν κακὸν κρίνεται. και εί μὲν όμιλεῖ αὐτῷ καλῶς καὶ εὐτάκτως, ἢ ἐδίδου αὐτῶ τὶ δόμα, ὡς πρὸς τὸ δόμα μέλλει εύρεῖν, ἀναλόγως του δόματος καὶ τοῦ πράγματος μεμετρημένον καλόν. εί δὲ ἔστιν ὁ μοναχὸς νέος, καὶ ἐπολέμει καὶ αὐτὸς μετ᾽ αὐτοῦ, ὡς τὸν γέροντα είς νόσημα άνίατον μέλλει έμπεσειν. εί δὲ ἔκρουεν ούτος τὸν μεγαλόσχημον, μέλλει ούτος δώσειν θυσίαν και παράκλησιν πρός άγιον τινα ό ίδων άναλόγως του δαρμου. εί δε έδαιρε [sic] τον νέον, ζημιωθήσεται παρ ἐχθροῦ.
If someone dreams that he met a monk wearing the cross like monks of the higher order <do>, if <the monk> spoke to <the dreamer> in sweet and kind words, let the dreamer hold on to them [?]. If not, the opposite. If this monk beat him, the dreamer's soul will receive the grace of a saint, for a monk wearing the cross is interpreted as a man of God. And whatever beating the dreamer suffered from the monk is a spiritual benefit from a saint. If the monk did not wear the sign of the cross and did not have the long habit of monks of the higher order, but was only wearing his cassock, if he is old, he is interpreted as representing bad times. If he was talking to the dreamer in a nice and orderly way, or gave him something, the dreamer will obtain something moderately good, commensurate with the <monk's> gift. If the monk was young and fought with the dreamer, the dreamer will fall ill with an incurable sickness, as if he were an old man [?]. If the dreamer beat up the monk with the long habit of a higher order, he will make an offering and address a request to a saint commensurate with the blows. If he beat up the young <monk>, he will be harmed by an enemy.
Ibn Shahin gives an interpretation of receiving something from one of the early Muslims and hearing something from a religious figure similar to Vat. gr. 573's interpretation of hearing or receiving something from a monk. According to Ibn Shāhīn (no. 366): من راي احدا من التـابــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ ... (If someone dreamt that)كلمـه او خـالـطـه فــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ one of the Muslims of the generation after the Prophet150 gave him something or spoke to him or associated with him, in all cases this indicates the attainment
ISo The tabifun (lit., "followers" ) are important for religious scholarship because of their role in transmitting the hadith.
of something good). In his chapter on the interpretation of judges, religious scholars, jurists and martyrs, <sup>151</sup>which also discusses "Sufis and the like" (~~J ~~I), saints *(awliyif),* righteous people *(\$dlif:iun),* and dervishes *(ahdal, majadhzb,fuqard,),* Ibn Shahin states (no. 1123): ..J--o I ..l-...:1.1 (\$1.> ..J--oJ ~ .:,~ <t....:iµ ->-°~ b~IJ '-:-'l:J I ll.AI ~~.>\_,sill (If someone dreams of one of those mentioned in this chapter and <the person dreamt about> informs <the dreamer> of something, the dreamer will be an eyewitness to its fulfillment). In addition, al-Dinawari's chapter on the prophets, <sup>152</sup>a higher order of holy figure that is also discussed in chapters 11 and 150 of the *Oneirocriticon,* interprets dreaming of being beaten by a prophet along lines similar to the Greek interpretation of being beaten by a monk: ..J--o ~(.SI.> .:,LJ <L:i~]j b~..lj ~J ..>-"I .J-a b~ F <t....:.Ll 4-..;>~ i)l.....JI ~-~~I (If someone dreams that a prophet from among the prophets, may peace be upon them, is hitting him, <the dreamer> will attain what he wishes regarding his religious faith in this world in the Hereafter).
The chapters of the five Arabic dreambooks on the companions of the Prophet, the Muslims of the next generation, the martyrs, religious scholars, jurists, etc., do not discuss any other dreams comparable to those in *Vat. gr.* 573, 153 nor do they differentiate between young and old. Some of the interpretations on monks are not unique to this passage, but are repeated in other chapters of the *Oneirocriticon* as well. In a number of other passages the *Oneirocriticon* interprets an unknown young man as an enemy, and an unknown old man as the dreamer's luck, either good or bad, depending on the old man's appearance:
Chapter 18, "From the Indians on the Significance of Hair" (Drexl 14, 18-21): ti. 8£ VEcO'tEpOV ayvciJptO''tOV 'i81], £x9po~ Ecr'tl V 0 opciJµEVO~· EaV 8£ yEpOV'ta ayvciJ-
<sup>151</sup> Ibn Shahin, *bah* 14 (nos. 1098-1126).
<sup>152</sup> Al-Dinawari, *fas/* 2, *bah* 22, BN *arahe* 2745, fol. 55a; repeated in *al-Muntakhah,* p. 14; and in al-Nabulusi, vol. 2, pp. 287-89, s.v .• ~.YI .:,.... ~·
<sup>151</sup> Ibn Qutayba has a chapter on the imam and two on the *qai/.i* which are not comparable to the Greek interpretations of monks, though they are clearly related to the chapters of the *Oneirocriticon* on priests and judges. Al-Dinawari's chapters on the companions of the Prophet and various religious figures *(fa:;/* 4, *bah* 1-4; *Esad Efendi* 1833, fols. 42a-b; BN *arahe* 2745, fols. 62a-63b) have nothing comparable. Neither does *al-Muntakhah (bah* 4), pp. 37-38; among Ibn Shahin's interpretations only the two quoted above are comparable to those given in the Greek text; nor do al-Nabulusi's potentially relevant chapters (vol. 1, pp. 13-14, s.v. ~I .\_,.,~I ; vol. 2, p. 30, s.v . .:,\_rJL...:. ; vol. 2, p. 78, s.v. *(') \_ \_.\_,,.y* I .L....h .:,..... rJ~ ; vol.1, p. 116, s.v. ~)quote any comparable dreams.
#### CHAPTER SEVEN
ριστον ή γνώριμον, ό γέρων ή τύχη έστι του όρωντος. και εί ό γέρων ίσχυρός έστιν, τοιαύτη και ή τύχη <ἐστι του δρώντος>, εἰ δὲ μή, τουναντίον.
If <someone> dreams of an unknown young man, the person whom he saw is an enemy. If he dreams of an unknown or a familiar old man, the old man is the dreamer's fate. If the old man is strong, such is also the fate of the dreamer; if not, the opposite.
Chapter 32, "From the Indians on Cupping and Bleeding" (Drexl 21, 22-24): èàv δε ΐδη, ότι παρά γέροντος έσικυάσθη ή έσικύασε γέροντα, ό γέρων ἔσται ή τύχη του ίδόντος.
If <someone> dreams that he was cupped by an old man, or that he cupped an old man himself, the old man is the fate of the dreamer.
Chapter 125, "From the Persians and Egyptians on Nodding" (Drexl 74, 15-16): εί δε ένευσε γέροντι άγνώστω και συνήκε το νεύμα, ή τύχη του ίδόντος έπλ το κρείττον έσται.
If he nodded to an unknown old man, and the old man understood <the meaning of> the nod, the dreamer's luck will improve.
Chapter 127, "From the Indians on Women" (Drexl 76, 18-21): αλλά και ό χέρων ό ἀγνώριστος εὐδειδὴς ών ή τύχη του ἀνθρώπου ἐστί, καὶ ὁ γνώριμος ὡς ἐπὶ τὸ πολύ είς τύχην διακρίνεται του ίδόντος. εί δὲ νεώτερος ὁ ἀγνώριστος, εἰς πρόσωπον έχθρου κρίνεται διηνεκώς.
An unknown, handsome old man is a person's fate. A familiar <old man> is also interpreted as the dreamer's fate in most cases. If the unknown person was young, he is always interpreted as an enemy.
Arabic dreambooks also interpret an unknown young man as an enemy and an unknown old man as the dreamer's fate, which is analogous to the old man's appearance, as is evident from the examples quoted below:
Ibn Outayba, bab 14, Yahuda ar. 196, fol. 29a:
If <the man whom the dreamer saw> is unknown, if he is a youth, he is an enemy, and if he is old, he is <the dreamer's> ancestor or his destiny .
Al-Dīnawarī, fasl 6, bāb 11, Esad Efendi 1833, fol. 46b; BN arabe 2745, fol. الشاب عدو للرجل :69b A youth is a man's enemy.
Al-Dīnawarī, fașl 6, băb 14, Esad Efendi 1833, fol. 47b-48a; BN arabe 2745, fol. 71b:
الرجل الشيخ والكهل السمن الحسن جد الانسان ... فاذا كان الشيخ المجهول او الكهل المجهول قويا فهو قوته واذا كان ضعيفا فهو ضعفه وعلى اية حالة راه عليه فان جده يكون على تلك الحالة مذمومة كانت او محمودة.
The old or the middle-aged man who is corpulent and handsome is the luck of people ... If the unknown old or middle-aged man is strong, this is the strength of the dreamer's luck, and if he is weak, this is its weakness, and in whatever state the man was in the dream, the dreamer's faith will be analogous, whether bad or good.
#### Al-Muntakhab, chapter 21, p. 73:
ورؤيا الشيخ والكهل المجهولين تدل على جد صاحبها فاذا راهما او احدهما ضعيفا فهو ضعف جده واذا ر آهما او احدهما قويا فهو قوة جده … وشاب فى التاويل عدو الرجل.
Dreaming of an unknown old and an <unknown> middle-aged man signifies the luck of the dreamer. If <the dreamer> sees both or one of them weak, this is weakness in his fortune, and if he sees either both or one of them strong, this is strength in his fortune .... A youth in dream interpretation is an enemy of the dreamer.
Ibn Shāhīn (nos. 1181 and 1183) repeats similar interpretations and adds (no. 1184):
ومن راي جامعة مشايخ او شباب فهم رحمة خصوصا إذا جرى منهم كلام الببر ومن راي ان احدا منهم اعطاه شيئا فهو اجود خصوصا إذا كان صنف ذلك الشيء محبوبا وإن راى انه هو العاطى فإنه جيد ايضا ومن راى احدا منهم وهو ناقص فإن كان شيخا فالنقص فى جده و إن كان شابا فالنقص فى عدو ه .
If someone dreams of a gathering of older and younger men, this is kindness, especially if they spoke in words of righteousness. If someone dreams that one of them gave him something, this is even better, especially if the kind of thing given <to the dreamer> was desirable. If he dreams that the giver was himself, this is also good. If he dreams that one of the <men> is deficient <in some way>, if the man is old <the dream> signifies that the dreamer's luck will be deficient. If the man is young, an enemy of the dreamer will suffer a deficiency.154
<sup>14</sup> The interpretation of a youth as an enemy is repeated in al-Nābulusī, vol. 2, p. 46, s.v. صبى. Cf. also al-Nābulusī, vol. 1, p. 236, s.v. رجل. The interpretation of an old man as the dreamer's fortune and of a young man as an enemy is also repeated in al-Nabulusi, vol. 2, pp. 12-13, s.v. شس مجهول
The passage of Vat. gr. 573 on monks interprets being beaten by someone as receiving something beneficial from him: if the dreamer was beaten by a monk he will receive a saint's grace, and if he was beating the man he will make an offering to a saint. Beating is also interpreted as benefaction received from the beater in chapter 218 ("From the Persians and Egyptians on Hitting, That Is, Beating"):155
Εαν ίδη τις, ότι ετύφθη βουνεύροις παρά τινος γνωρίμου, ανάλογον του μέτρου κερδήσει δια χαράγματος χρυσίον εί δε ό τύψας έστιν έν ἐξουσία, καὶ ἐκ ταύτης δώσει αύτφ. ἐαν ίδη ὁ βασιλεύς, ὅτι ώρισε τυφθηναί τινα βουνεύροις, εύρήσει ό τυφθείς χαράν μεγάλην και πλούτον κατά τον αριθμον αύτων. όμοίως ἐὰν αὐτὸς ἐκεῖνος ὁ βασιλεὺς ἔτυψέ τινα μαγκλάβια ἢ χαρζάνια, εὑρήσει ἐξ αυτού περίφημον και μέγιστον άξίωμα ύπέρ καύχημα. εί δε αὐτοχειρι έτυψέ τινα δάβδω, εύρήσει ό τυφθείς παρά του πρωτεύοντος του βασιλέως είδήσει και γνώμη του βασιλέως χαράν και έξουσίαν.
If someone dreams that he was beaten with a strap of raw ox-hide by someone he knew, he will find gold in coins commensurate with the number of blows. If the one who beat him is a person of authority, he will give some of his authority to the dreamer. If the king dreams that he ordered somebody to be beaten with a strap, the person who was beaten will receive from the king great joy and wealth commensurate with the number of blows. Likewise if the king himself beat someone with a strap or whip, the person who was beaten will receive from him a distinguished and great office, greater than anything one could boast about. If <the king> held a staff with his own hand and beat someone with it, the person who was beaten will receive joy and power from the most powerful person after the king, with the king's knowledge and consent.
Similar interpretations are repeated in the Arabic dreambooks. The meaning of such dreams is given in a nutshell by al-Dīnawarī: اما الضرب فقد <قـال> 65 As for hitting, the) المسلمون هو معروف ينال المضروب على يدى الضارب Muslims said: "It signifies that the person who was hit will receive something good and beneficial from the hands of the one who hit him" ).
The evidence presented above proves that the interpretations of monks appended to chapter 149 ("On Priests") in Vat. gr. 573 were not invented by the author of the Oneirocriticon, but were based on material taken over from his Arabic sources. It is, however, difficult to reconstruct the manner in which he handled this material. It is possible that he composed his interpretations of
<sup>155</sup> Drext 171, 1-12 (σιη΄ Έκ τών Περσών και Αίγυπτίων περὶ τύψεως ἤτοι δαρμού).
<sup>156</sup> Al-Dinawari, fas! 10, bab 31, Esad Efendi 1833, fol. 109a; repeated in Ibn Shāhīn, no. 2316; repeated in al-Nabulusi, vol. 2, p. 54, s.v. ﻀﺮﺏ
monks based on the interpretations of secular men and beating that are repeated in other parts of the Greek dreambook. Or, he might have copied the interpretations of dreams that concerned Muslim figures whom he perceived as counterparts to Christian monks. It is impossible to state with certainty which of the two was in fact the case, but occasional awkward expressions found in the Greek text could be indications of a struggle in rendering the expressions in an Arabic text that did not lend themselves readily to translation.157 These instances make the second hypothesis more likely than the first. But even if we grant that the Greek author modeled his interpretations of monks after analogous Muslim interpretations (as he evidently did with other religious notions), it is impossible to say which Muslim figures in particular were the source of his inspiration.
#### Ringing the Sounding Board
#### The Oneirocriticon interprets ringing the sounding board as follows:
Έάν τις ϊδη καθ΄ ύπνους έαυτον σημαίνοντα το σημαντήριον, ούτος μέγιστος γενήσεται, άλλά και μεγαλόφωνος, ἐπισυνάχων λαοὺς πρὸς τὸ ἑαυτοῦ θέλημα έαν δε βασιλεύς ή ό τουτο ίδών, τους μεγιστάνους αύτου συγκαλέσει πρός εύβουλίαν, άλλά και πολυζώητος έσται και πολυχαρής διά τὴν ἐκ τοῦ σημαντῆρος έγγινομένην χαράν τε και ήδονήν. ἐαν δὲ οὐ σημαντήριον, άκούση δὲ ἑτέρου σημαίνοντος, ή αὐτὴ κρίσις καὶ λύσις ἐστὶ τοῦ ὀνείρου. έαν δε πτωχός ή ό τουτο ίδών, ύπηρέτης έσται μεγιστάνων και προκόψει πλείστα. ἐὰν δὲ γύναιον τοῦτο ἴδη, θέατρον αἰσχύνης ἔσται παντὶ τῶ λαώ.
If someone dreams that he is ringing the sounding board, he will become very great and also loud voiced, congregating people to carry out his will. If the dreamer is a king, he will summon his nobles to a prudent council, and he will be longlived and filled with joy because of the joy and pleasure that come from the sounding board. If he does not ring the sounding board himself, but hears somebody else ringing it, the interpretation and solution of the dream is still the same. If a poor person has this dream, he will serve nobles and succeed greatly. If a woman has this dream, she will be a spectacle of shame for all the people.
<sup>157</sup> For example, redundant repetition of a cognative: " } ἐδίδου αὐτώ τὶ δόμα, ώς πρός το δόμα, μέλλει εύρεῖν, ἀναλόγως του δόματος καὶ τοῦ πράγματος, μεμετρημένον καλόν" ; repetition of pronouns: "και ἐπολέμει καὶ αὐτοῦ" ; repetition of pronouns and redundant insertion of a participle: εί δε έκρουεν ούτος τον μεγαλόσχημον, μέλλει ούτος δώσειν θυσίαν και παράκλησιν πρός άγιον τινά ό ίδων αναλόγως του δαρμου.
#### 320 CHAPTER SEVEN
The sounding board was also used by Christian minorities living in Muslim lands. Therefore, when it is discussed in Arabic dreambooks, the interpretations they give are negative and clearly reflect a Muslim point of view. According to al-Dinawari, the Christian sounding board is to be interpreted as follows: <sup>158</sup>
<Liu ->"\_,.jL.:J I '-:""'~ <Lil I.SI J ~ ~ ~ *"'i* \_µI..:.... '-:"'l.i.S ~J ->"\_,.jL:J I <Liu~~ ->"\_,.:gL:JL; '-:""'~ 4-.'..I 1.S1J .>-o l>-1u\_, b~ )lbl; 1~ ~ ~I ->"\_,.jLll '-:"'~-'~I ~I *.:\_;"'i* <LjJ~J bl~\_,~ \_\_...k ~ .d......i~I
The sounding board is a man who is a liar and a hypocrite who has nothing good in him. If someone dreams that he is ringing the sounding board, he will spread and proclaim a false rumor. The <dream interpreters> said: "If someone dreams that he is ringing the sounding board in a church, he will take an oath regarding his activity in sales, purchases and commerce, because the church *(hra)* indicates the sale (hai'). The ring of the sounding board is a false oath."
The same interpretation is repeated in part in *al-Muntakhab, <sup>159</sup>*and similar interpretations are quoted by Ibn Shahin. <sup>160</sup>Al-Nabulusi's paragraph on the sounding board repeats al-Dinawari and adds the following: 161 ~ < JU~L.:J I> ~.) *6* .~.II~ J..i ~.JJ ..:\_,j)-a \_,I ..iY\_,I ...::...l.:i 4\_,j \_,I) ....... r~ d .,, \_ •. .:. 9 11 \_, (The sounding board in a dream is a broker or a wife with children or a muezzin. <sup>162</sup>Possibly it signifies notoriety and ignominy).
The *Oneirocriticon' s* passage on the sounding board does not copy the interpretations given in Arabic dreambooks, but neither is it independent of Arabic dream interpretation. Ringing the sounding board summons the Christian flock, and therefore congregating people to carry out the dreamer's will is an understandable interpretation of such a dream. That the dreamer should become "loud voiced," however, is somewhat out of place, until one remembers that the Muslim community is summoned with the call to prayer *(adhdn,* literally "an announcement") from the top of a minaret, which requires a loud and melodious voice. The Christian interpretation of ringing the sounding board is inspired by the Muslim *adhdn.* Al-Dinawari interprets the *adhdn* as
<sup>158</sup> Al-Dinawari,fa\$1 8, *hah* 76; *Esad Efendi* 1833, fol. 144a; BN *arahe* 2745, fol. 143b.
<sup>159</sup>*Al-Muntakhah,* p. 331. .
<sup>160</sup> Ibn Shahin, nos. 7385-86 bis.
<sup>161</sup> Al-Nabulusi, vol. 2, pp. 300-1, s.v. J"\_,\_:<L.:..
<sup>162</sup> Somebody who performs the call to prayer for Muslims.
sovereign power for those fit to exercise it and as a profitable profession for commoners; so does the *Oneirocriticon:* <sup>163</sup>
~~I ~Y\_,l.J .J~ .JI <t.:.U L..:.:i\_,\_., ~I~~ ~J c:i:i.J-: <Lil *..slJ* .JU JI *;;J4:>* ~<Wu ~I L.....l::.) ~ ~ .ju ... <t..:J\_,\_..., FL. *J..U..,.* ~y <sup>J</sup> ·0J~IJ •• l..P.-1 ~ U..P-
If someone dreams that he performed the call to prayer and he is not a muezzin when he is awake, he will rule as far away as his voice reached, if he comes from an appropriate family .... If he is unsuitable for leadership because of his family, he will do well in commerce or in a profession where his profits and the cubits of selling and buying <that is, the goods exchanged,> will be plentiful. <sup>164</sup>
Al-Nabulusi quotes further interpretations of the *adhiin.* Their positive overtones, and especially the connection of the *adhiin* with the sultan remind one of the *euboulia* (prudent council) to which the emperor's nobles are summoned in the *Oneirocriticon:* <sup>165</sup>
<Lil *..slJ* :.,-OJ ... ~I J\_.....JJ ..::...Ld~IJ ~IJ .Lc...iJI ~ .Jlj')'I J~ ..i....:;J .~..i .Jlj')'I .J~ ..UJ ... ~I .\_,Jj L.1\_,.jl J-L~ <Liµ~•·-,:- • r.,.J ~ c:i:i.J.= . .jlhl..uJI *µ* .j...a ..>-"I .\_,JI
The *adhan* signifies supplication, godliness, pious acts and good deeds .... If someone dreams that he performed the *adhan* among congregated people, he will invite people to the truth .... The *adhan* is indeed an invitation to a business matter in front of the sovereign.
A woman dreaming that she performed the *adhan* has a negative interpretation in Arabic dreambooks, as a woman dreaming that she rang a sounding board does in the *Oneirocriticon.* Ibn Shahin (no. 872) interprets such a dream as follows:;; *.':* ·o • ..::..JJ.....:>. ~ J~ <t....:.µ ~ JAi t-" .:,.:i.J7. <t....:.11.51.J .J-aJ . .:,.:i\_µ ~I ..::..1 .J lj) ~ l\_,..11 dJ l.S.J (If someone dreams that he performed the *adhiin* together with the members of his household, it signifies that a disaster will occur. It is the same if a woman dreams that she performed the *adhiin).*
<sup>163</sup>Al-Dinawari,fa~/ 8, *bah* 11, *Esad Efendi* 1833, fol. 74a.
<sup>164</sup>These interpretations are repeated almost verbatim in *al-Muntakhah* (p. 47) and in al-Nabulusi, vol. l,p.21,s.v . .:.,1:i1.
<sup>165</sup> Al-Nabulusi, vol. I, pp. 21-22, s.v . .:\_,l:il.
#### Building a Church
The final paragraph from chapter 12 of the Oneirocriticon associates the building of churches with marriage and with women (Drexl 9, 28-10, 4):
Εάν τις ίδη, δα εκτισεν έκκλησίαν, εί μεν ή βασιλεύς, γυναικα έκ νέας κοσμήσει και ύψωσει αναλόγως του μεγέθους της εκκλησίας. ἐαν δὲ τοῦτο του κοινοῦ λαού θεάσηταί τις, ούτος πλούτον εύρήσει από γυναικός προστασίας. έαν δὲ τουτο ίδη γυνή, εί μεν άγαμός έστιν, λαμβάνει ανδρα, εί δε ύπανδρος, χηρεύσει και αλλον λήψεται.
If someone dreams that he built a church, if he is the emperor he will adorn a young woman and elevate her by analogy to the size of the church. If a commoner dreams this, he will find wealth through the protection of a woman. If a woman dreams this, if she is celibate, she will take a husband, and if she is married, she will be widowed and marry someone else.
Christian churches are also discussed in Arabic dreambooks, but their interpretations differ greatly from those recorded in the Oneirocriticon. Al-Dinawari quotes the views of both Muslim and Christian dream interpreters, but even the Christian interpretations do not coincide with those of the Oneirocriticon: 166
من راي ان في منزله بيعة فان قوله بالقدر يضار ع قول النصارى وكذلك لو رائ ان منزله بيعة فان راى انه تحول بيعة فانه يخرج خارجى على رئيسه فان رای انه فی پیعة فان مذهبه مذهب النصاری ۾ فان رای انه نقب فی بيعة فانه يفتش عن بدعـة وقـالت النصـارى من راى انـه يدخل هيكـلا من الهياكل ويجعل فيه مرمة وبصلح بسبب من الاسباب فانه يجرى على يديه خير ويتني عليه الناس فان راى كانه يهدمه فانه يضيع ضيعة يعرف على يديه جماعة او يخرج من دينه فان راى انه دخل هيكل بعض الشهداء فانه يموت او يقتل ويشرف على الهلاك.
If someone dreams that there was a church in his house, his words will have the same power as the words of the Christians; it is the same if he dreams that his house was a church. If he dreams that it was transformed into a church he will revolt against his leader. If he dreams that he was in a church, his views are the views of the Christians. If he passes through a church he will inquire after heresy.
<sup>166</sup> Al-Dīnawarī, fași 8, băh 75; Esad Efendi 1833, fol. 144a; BN arabe 2745, fol. 143b. The first half of al-Dinawari's chapter is repeated in al-Nabulusi's paragraph on Jewish synagogues (vol. 1, p. 52, s.v. Wherever al-Dinawari has "Christians" al-Nābulusī substituted "Jews." The word bra can mean both "church" and "synagogue." Al-Dinawari's word for "synagogue" is kanīsat-al-yahūd. The word kanīsa in al-Nābulusī refers to the Christian churches.
The Christians said: "If someone dreams that he went to an altar and made repairs on it and mended something for some reason, he will practice goodness and people will speak of him in the most laudatory terms. If he dreams that he demolished it, he will let perish a landed estate that a group entrusted him with, or will abandon his religion. If he dreams that he went to the altar of some martyrs he will die or will be killed and will be honored upon his death.
Both *al-Muntakhab* and lbn Shahin regard Christian churches from a clearly Muslim point of view, which results in overwhelmingly negative interpretations. According to *al-Muntakhab: <sup>167</sup>*
pl )JJ ~I ..::..\_,..:.L.. ~J ~lj.ll *)J* ~J o~I ~ UIJ ~I J::>\_,\_.JI\_, Jl\_,...JI\_, (~I *)J* ~J ,L:..i.JI\_, ->-"\_)I\_, ...J\_JWI *)J* ~J t~I\_, .~I ~J "--:'..> ~ *.j..o )J* ~ ~J
A church signifies a graveyard, a brothel, a wine tavern, the abode of the infidels and of heresy, a place with music and singing, a place of lamentation, mourning and wailing, Hell and a God-forsaken place, as well as prison.
There follow another ten lines of interpretations in the same spirit.
lbn Shahin's interpretations of churches are as follows (nos. 1032-1035):
(1032) If someone dreams of a church or a monastery or something like that, it is interpreted as a man who is a liar and beguiles people with his deeds, and there is no result <from this dream>. (1033) If someone dreams that he did in a church something which offends its people <i.e., the Christians> but does not violate the *sharf'a* <i.e., Islamic law>, this is an offense committed by the aforementioned man, and it is said that this is good. (I 034) If someone dreams that he observed one of these things <that Christians do>, if he is a righteous person, this is a good <portent>, but if he is a sinner, there is nothing good in this <dream>. (1035) It is said that if someone dreams that he did inside a church what is suitable for Christians <to do>, this signifies the perpetration of sins.
<sup>167</sup> *Al-Muntakhah,* p. 270.
#### CHAPTER SEVEN
Al-Nabulusi's chapter on churches discusses this dream from both a Christian and a Muslim perspective, though he does not clearly state which of the interpretations should be applied to a Christian and which to a Muslim dreamer: 168
كنيسة: هي عند اهلها في المنام دالة على العلم والعمل والزهد والخشية والبكاء وربما دلت على الهم والنكد والكذب والبــهــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ الكنيسة على البدعة ودار الظلم واللهو واللهو واللعب واللعب والالفة على الخمر واللناجاسات وعلى الجوزة والامـة وتدل على المـاكم بـالجـور فـمن دخل اللى كنيـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ وخصوصا ان سجد للتماثيل او قبلها او شاركهم في قربان او كان في وسطه زنار .
The church: this is for its people <i.e., for Christians> in a dream a symbol of knowledge, activity, asceticism, fear, and lamentation. Possibly it signifies sorrow, misfortune, deceit, slander, and defamation. Possibly a church also signifies heresy, the abode of tyranny, entertainment, gatherings, games, wine parties and impurities, a wife or a slave girl. < In addition>, it signifies an unjust ruler. If a bachelor dreams that he entered a church he will get married or will beget a son or will err after pursuing the true religion, especially if he prostrates himself in front of the icons, or kisses them or shares an offering with them or has a zunnar.100 around his waist.
There follow another twelve lines of interpretations that largely repeat al-Dīnawarī and al-Muntakhab.
Al-Nabulusi's interpretations of a church are both negative and positive; the church can signify knowledge, pious deeds, marriage, and begetting a son, interpretations that coincide with the interpretations of mosques quoted in both al-Nabulusi and other Arabic dreambooks. They should evidently be applied only to Christian dreamers and appear to have been influenced by the significance of mosques for Muslim dreamers.170 The association of mosques with marriage and women is analogous to the association of churches with
<sup>168</sup> Al-Nābulusī, vol. 2, p. 187, s.v. كنــــــــة .
<sup>169</sup> The zunnar is a belt that Christians living in Muslim lands were obliged to wear.
<sup>170</sup> Ibn Qutayba (hah 9) associates building a mosque with the performance of good deeds. Al-Dinawarī (fașl 8, băb 56; Esad Efendi 1833, fol. 82a-b; BN arahe 2745, fol. 139b) interprets a mosque as representing a knowledgeable person (ر جل عــالـم), while building a mosque signifies ذخاب ) the treasures of religion or a pious deed and enhancement of <the dreamer's> religion" and praying in the milhrāh of a mosque signifies begetting a son (fas) 8, hab 58; Esad Efendi 1833, fol. 82b; BN arahe 2745, fol. 140a). Similar interpretations are repeated in al-Muntakhab, pp. 51-52; Ibn Shāhīn no. 984 (knowledge), no. 987 (pious deeds); al-Nābulusī, vol. 2, pp. 229-31, s.v. محراب and مسجد.
marriage and women in the *Oneirocriticon,* and can be found in all five Arabic dreambooks. lbn Qutayba states: <sup>171</sup>
~J\_;..:;JIJ i~\_,~I 4..L:. ~ dJj ,:.,LS: ~..JJ ~I J.U ~ L.,\_.,.. I~ l.:...,. .j..<>J .\_,.5::.i.JIJ o~I .\_.k ~µI ~J <Y'LJI .....i.J\_J..:i ~I 0~ dJj ot...,....;.IJ
Whoever builds a mosque will be prominent for his good deeds, and possibly this refers to a bond of kinship and marriage and such, because a mosque unites people and congregates those who are scattered in order to pray and remember the name of God.
Al-Dinawari also interprets the mosque as marriage: <sup>172</sup>~ ~1 lSIJ .:\_,L-..ii ~\_,I .UL ..:.,...a i\_,\_.:g ~\_,\_;:. \_,1 iL,,.J~l li...o ~ .ill.:i .:.,~ ~.) ... I I?'"' 0 .ill j (If <someone> dreams that he built a mosque . . . possibly this signifies a bond of kinship or marriage within his clan, or something like that).
The interpretation of mosques quoted in *al-Muntakhab* (p. 267) is close to that of churches in the dreams of commoners according to the *Oneirocriticon:* ~\_,\_;:.\_, (.~ ~ <J..l7.> '-:-'~~I~ ... iL.:...ll ~I~~~ L:-.:. ..J I\_, JU I '-:-')lb.J \_, (If someone saw in his dream that he built a mosque ... for those who are celibate it signifies wedlock and marriage, as well as pursuit of wealth and worldly acquisitions ). <sup>173</sup>
lbn Shahin also associates dreaming of mosques with women and wedlock (nos. 985 and 995): (J~ I' ?' .... ~ O....:.l lSI *J* ..:.,...a ~y--i..11 .J-:'4 JU J..i..i.J I~ ;;I J-"l:i J,;)..: ~I,; ... ~J ;;IJ-" I (Jiibir al-Maghribi said: "If someone dreamt that he erected a mosque, he will marry a devout woman .... The mosque is interpreted as a woman of lofty rank").
It is unlikely that the author of the *Oneirocriticon* relied on Muslim interpretations of Christian churches, even if they were recorded in his Arabic sources. More probably, he modeled his Greek interpretation of churches after the Arabic interpretation of mosques.
<sup>171</sup> Ibn Qutayba, *biib* 9, *Yahuda ar.* 196, fol. 27b.
<sup>172</sup> Al-Dlnawari,fa\$1 8, *biib* 56, *Esad Efendi* 1833, fol. 82b; BN *arabe* 2745, fol. 139b. The interpretations of al-Oinawari are repeated in al-Nabulusi, vol. 2, p. 229, s.v. ~.
<sup>173</sup> *Al-Muntakhab* also discusses mosques on pp. 51-52.
#### On Changing Faith and Idol Worshiping
In the first part of chapter 12 of the Oneirocriticon, "From the Account of the Indians on Changing Faith,"174 dreams about conversion from Christianity to another religion are interpreted. Conversion to Judaism and Islam is generally interpreted as lack of faith and blasphemy, but other interpretations dependent on the identity of the dreamer are also given (Drexl 8, 1-9):
Έάν τις ίδη κατ' όναρ, ότι χριστιανός ων έβράϊσεν, ουτος ούκ ήν τέλειος, άλλα δύσπιστος και βλάσφημος, και είς άπώλειαν καταντήσει. ἐαν δὲ βασιλεὺς τούτο Ίδη, αϊρεσιν νεωτερίσεται κατά του λαου αύτου. ἐαν δὲ τις του κοινοῦ λαου τουτο ίδη, είς γευδομαρτυρίαν αύτομολήσει: ἐαν δὲ γυνή, ἐπίβουλος έσται τω άνδρι αύτης έαν δε δούλος, κατά του ίδίου δεσπότου βουλεύσεται. ἐὰν δέ τις ἴδη κατ' ὅναρ, ὅτι ἐμαγάρισεν, όμοίως τῶν εἰρημένων ἀποβήσεται αύτω.
If someone dreams that, though he is a Christian, he converted to Judaism, he is not perfect, but unbelieving and blasphemous, and will end up damned. If a king dreams this, he will invent a heresy to the detriment of his people. If a commoner dreams this, he will resort to false witness. If a woman, she will be treacherous to her husband. If a slave, he will entertain thoughts against his own master. If someone dreams that he became a Muslim, the same thing will happen to him.
Similar interpretations are given in Arabic dreambooks. According to Ibn من راى انـه تحـول الـى جنـس مـن الكفر فى مـنـامـه فـذلك هـوى هو 23,"Qutayba مضاه عليه ذلك الحنس (If someone dreamt that he changed into one of the unbelieving races, this is a heretic tendency <of the dreamer> that is practiced by this race). Arabic dreambooks state that dreaming of converting to a religion other than Islam foretells that the dreamer will indeed convert to the religion he dreamt about, 76 but also that he will perpetrate grave sins, depravity, vileness.
<sup>174</sup> ιβ 'Ex του λόγου των Ινδών περί πίστεως διαφόρου.
<sup>175</sup> Ibn Qutayba, hab 7, Yahuda ar. 196, fol. 27a.
<sup>176</sup> Al-Dīnawarī (fasl 8, bāb 85; Esad Efendi 1833, fol. 66a; BN arabe 2745, fol. 145a) interprets conversion to polytheism, including Christianity (which, from the Muslim point of view, counts as من راءِ انه تحول من دار الاسلام : polytheism because of its doctrine of the Trinity), as follows: الى دار الشرك فانه يكفر ويخرج الى دار الشرك فانه يكفر ويخرج الى دار الشرك Islam to a polytheistic religion, he will abandon Islam for polytheism). The same interpretation is فان رای کانه تحول کافرا فان اعتقاده یوافق :( repeated in al-Muntakhab, which adds (p. 330): . اعتقاد ذلك الجنس من الكفار (If someone dreams that he became an unbeliever (kaffir), his faith will coincide with the faith of the race of unbelievers <that he dreamt of>).
وقـــيل مـن راي انـه يـهــودى فـــانــه ",and heresy. According to al-Dīnawarî مجترى على المعاصى يصير على الذنوب والكبائر وعلى ما نهاه الله تعالى عنه (It is said that if someone dreamt he was a Jew, he will be bold with sinning and will insist upon <perpetrating> offenses and atrocious crimes by doing what God Almighty has forbidden him <to do>). 10 Ibn Shāhīn states من رائي انـه تحول عن الاسلام الـي احد الاديـان الـبـاطــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ . فانه ار تكاب معاص وقيل ذلة وحقارة (If someone dreamt that he changed his religion from Islam to one of the false religions, this indicates the perpetration of a sin and, it is said, depravity and vileness). He then adds: و اما الليهود فمن ()) راى انه صار پهوديا فإنه ير تكب طريق البدعـة...و يكون على ضلالة. someone dreamt that he had become a Jew, he will take the path of heresy ... and will be in error).
The next paragraph in the Oneirocriticon interprets dreaming of becoming a magos (Drexl 8, 10-12): Εάν τις ϊδη, ότι μάγος ἐγένετο, ούτος φιλάργυρος και φιλόπλουτος γενήσεται οί γάρ μάγοι κοσμόφρονές είσι, μηδέν περὶ τῆς ἐκεῖθεν ἀνταποδόσεως λογιζόμενοι. (If someone dreams that he became a magos, he will be parsimonious and will love wealth, because the magoi are worldly minded, and think nothing of the retribution in the hereafter). The word magos in Greek means both "a magician" and "a Magian," that is, an adherent of the Zoroastrian religion.179 The Christian interpretation of such a dream as pursuit of worldly wealth brings to mind the story of Simon the sorcerer (Simon magos), which is related in Acts 8:9-25, and justifies the translation of the word magos as "magician." However, according to Arabic dreambooks, the pursuit of worldly wealth is also a characteristic of the Magians. Al-Dīnawarī]490 claims that somebody who becomes a Magian يطلب في will pursue worldly things without conflict with his religion, because the Magians are seekers of worldly
<sup>177</sup> Al-Dinawari, fast 8, bab 72; Esad Efendi 1833, fol. 64a; BN arahe 2745, fol. 143a. Al-Dinawari's interpretations of Jews are repeated in al-Nabulusī, vol. 2, pp. 348-49, s.v. يهودى
<sup>178</sup> Al-Dinawari justifies this interpretation by adducing the example of Qur'an 7:161-66.
<sup>179</sup> The German translator of the Oneirocriticon chose to render magos as "magician" (Brackertz, Traumbuch des Achmet, p. 27: "Hat er sich einer der Magie ergeben, wird er sein Herz an Geld und Reichtum hängen; Magier sinnen nämlich nur auf Irdisches und rechnen nicht mit der Vergeltung in Jenseits" ). The English translator rendered magos as "Magian," but signaled the problem in a footnote (Oberhelman, Oneirocriticon of Achmet, p. 90 and n. 19).
<sup>180</sup> Al-Dīnawarī, fasl 8, bāb 64; Esad Efendi 1833, fol. 63b; BN arahe 2745, fol. 142a.
<pleasures and possessions>). A similar interpretation is repeated by al-Nābu-ومن راي انه مجوسي فانه يطلب الدنيا لأن الجوس هم كلاب الدنيا lusī: 186 و اصحابها) (If someone dreams that he is a Magian he will seek worldly acquisitions, for the Magians are the dogs of the world, 180 as well as the lovers of temporal things and masters of worldly possessions).
The Oneirocriticon also discusses dreams of worshiping idols. The interpretation of such a dream depends on what the idol is made of (Drex1 8, 13-9, 5):
Έάν τις ϊδη, ότι προσεκύνησεν είδωλον ή ώμίλησεν ή ἐπάλαισεν ἢ ἔτυψεν ἢ ἐτύφθη παρ' αὐτοῦ, ὁ μὲν προσκυνήσας ψευήσεται εἰς θεὸν διὰ τὸ ψευδές του είδώλου, ό δε όμιλήσας ή παλαίσας ή τύψας ή τυφθείς είς θλίψιν ίσχυραν ἐμπεσειται, ψευδὴς δὲ ἔσται ή θλίψις. ἐἀν εἰδώλω ξυλίνω τις προσκυνήσει κατ' όναρ, ουτος ένδεὴς έσται τινὸς μεγιστάνου πονηρου καὶ οὐχ ὑπακουσθήσεται, περὶ ών ἀν ἐπιδεηθῆ. ἐὰν δέ τις προσκυνήση κατ' ὅναρ εἰδώλω έν σανίδι κατεσκευασμένω, αίρεσιάρχης λογομάχος έσται ψευδής. ἐαν δέ τις προσκυνήση είδώλω έξ άργύρου, ούτος πρός γυναϊκας έπίορκος και ψεύστης γενήσεται. ἐαν δὲ τὸ εἶδωλον χρυσοῦν ἢ, πάντλομος ἔσται καὶ παρὰ βασιλέως θλιβήσεται και τιμωρηθήσεται και κυκλώσει αύτον πλείστα κακά: έαν δε τούτο βασιλεύς ίδη, παρ' έναντίων θλιβήσεται, διότι, ώσπερ ό αργυρος είς γυναίκας διακρίνεται, ούτω και ό χρυσός, τοις άνδράσιν όσην χαράν ἐμποιεῖ έξω ύπνου, τοσουτον είς θλιψιν αγει τους όρωντας αὐτὸν καθ' ὕπνους. ἐαν δὲ το είδωλον χαλκούν ή σιδηρούν ή μολυβδούν ή. ό τουτο ίδαν ψευδόπλουτος έσται ταυτα γὰρ πάντα σκεύη του κόσμου εἰσίν.
If someone dreams that he worshiped an idol or talked to it or wrestled with it or beat it or was beaten by it, the one who worshiped it will be untruthful to God because of the falseness of the idol; he who talked to it or wrestled with it or beat it or was beaten by it will be afflicted by grave sorrow, but this sorrow will be fake. If someone in a dream worships a wooden idol, he will beseech an evil noble <for some favor> and will not be heard, whatever his request. If someone in a dream worships an idol made of planks, he will become a heresiarch and a quibbler filled with lies. If someone worships an idol made of silver, he will lie and swear falsely to women. If the idol is golden, he will be all-daring and will be distressed by the king and will be punished and numerous evils will surround him. If a king dreams this, he will be distressed by enemies, for just as silver is interpreted as women, gold causes to men who have dreamt of it as much sorrow as the joy it brings them when they are awake. If the idol is made of bronze or iron or lead, the one who dreams of it will obtain fake riches, because all these metals are objects of this world.
<sup>181</sup> Al-Nābulusī, vol. 2, pp. 627-28, s.v. مجوس.
<sup>182</sup> On the basis of al-Dinawari's text, al-Nābulusī's کلای الدنیا (dogs of the world) should probably be changed to طلاب الدند (pursuers of worldly pleasures and possessions).
Arabic dreambooks also connect the interpretion of the adoration of idols with the material used for the idol's construction. Their interpretations generally coincide with those given in the Oneirocriticon, as is evident in the examples below :
Al-Dīnawarī, fasl 8, bāb 87; Esad Efendi 1833, fol. 66a-b; BN arabe 2745, fol. 145a-b:183
قال المسلمون الصنم تمثال باطل مختلق وهو انسان غدار حسن الوجه سىء الخلق # فمن راي انه يعبد صنصا من دون الله تعالى فهو يكذب على الله تعالى بباطل اختلقه برايه فان كان الصنم من خشب فانه يتقرب بدينه الى رجل سلطان ظالم مذافق حسن لقوله تعالى كانهم خشب مسندة وان كان من حطب فانه يطلب بدينه الجدال والكلام فى الدين وان كان من فضة فانه ياتى فى دينه ما يتقرب به الى امراة او جارية اعجمية خيانة او بطرا وان كان من ذهب فانه ياتى فى دينه بامر مكروه ويتقرب الى رجل يبغضه الله تعالى وينال منه مكروه ويخاف على مال يغرمه او بزينة الدنيا وان كان من صفرا او حديد او رصاص فانه يطلب بدينه الدنيا ومتاعها وينسي ربه.
The Muslims say: "An idol is a false and fabricated image; it is a deceitful man with a beautiful face and an evil character. If someone dreams that he worshiped an idol rather than God Almighty, he will lie against God with a falsehood that he devised according to his own opinion. If the idol is made of wood he will approach by analogy to his faith <in the idol> a man of power who is unfair, hypocritical, and handsome, according to the saying of God Almighty: "And when thou seest them their figures please thee; and if they speak thou givest ear unto their speech. They are as though they were propped up blocks of wood" [Qur'an 63:4]. If the idol is made of faggots, the dreamer will pursue theological disputes and debates in his religion. If the idol is made of silver, the dreamer will go to a woman or a foreign slave girl in treachery or wantonness by using his religion. If the idol is made of gold, he will perpetrate something abominable in connection with his religion, approach a man whom God Almighty detests and suffer from him something abominable, and fear for money that he paid or for worldly ornaments. If the idol is made of brass, iron, or lead, he will pursue worldly pleasures and acquisitions through his religion and will forget his Lord.
Cf. also al-Muntakhab, chapter 46, pp. 329 ff:
المستحق للعبادة هو الله تعالى فمن عبد غيره فقد خاب وخسر فمن راي كانه يعبد غيره دل على انه مشتغل بباطل مؤثر لهوى نفسه على رضا ربه فان كان ذلك الصنم الذي عبده من ذهب فانه يتقرب الى رجل يبغضه الله
<sup>183</sup> The same interpretations are repeated in al-Nabulusī, vol. 2, pp. 48-49, s.v. صنح
تعالى ويصيبه منه ما يكره وتدل رؤياه على ذهاب ماله مع وهن دينه وان كان ذلك الصنم من فضة فانه يحصل له سبب يتوصل به الى امراة او جارية على وجه الخيانة والفساد فان كان ذلك الصنم من صغر او حديد او رصاص فانه يترك الدين لاجل الدنيا ومتاعها وينسى ربه وان كان ذلك الصنم من خشب فانه ينبذ دينه وراء ظهره ويصاحب واليا ظالما او رجلا منافقا ويكون متحليا بالدين لاجل امر من امور الدنيا لا من اجل الله تعالى.
The <only> one entitled to be worshiped is God Almighty. Whoever worships anything other than Him is wrong and has gone astray. If someone dreams that he worshiped anything other than God, it means that he is occupied with deception affecting his own love for what is approved by his Lord. If the idol which he worshiped is made of gold, he will approach a man who is odious to God Almighty and will suffer something horrible at his hands. His dream signifies the loss of his money together with weakness in his faith. And if this idol is made of silver, something regarding disloyalty and immorality will afflict the dreamer in connection with a woman or a slave girl. And if this idol is made of brass or iron or lead, he will renounce religion for the sake of worldly pleasures and will forget his Lord. If this idol is made of faggots, he will renounce his religion and leave it behind him, and will befriend an oppressive governor or a hypocritical man and will find religion pleasant, not for the sake of God, but for the sake of worldly things.
#### Cf. also Ibn Shahin, nos. 1073 and 1075:
( ۷۲) ومن راى انه يعبد صنما من خشب فإنه يقترب برجل باطل إلى رجل خبـيث مـنافق، وإن كـان مـن حطب مـشـبك فـإنـه يـطلب بذلك مـا ياتـى بـه مـن الجدال وما اشبه ذلك، وقيل انه يتقرب لاحد بنميمة، وإن كان الصنع من فضة فانه باتي الى امراة بما لا بلبق، وان كان من ذهب فانه يتقرب الى امر يكرهه ويحصل له من ذلك ضرر وإن كان من نحاس او حديد او رصاص او ما اشببه ذلك فانه يتقرب لطلب الدنيا وقيل انه يتقرب رجل متلصص... (۷۰) ومن راى انه يعبد صنما من الاصنام او كلمه او فعل معه فعل إنسان فى اليقظة فإنه يصحب من لا فائدة فى صحبته وربما يكون حصول ضرر من ذلك الصاحب، وقيل من ارتكاب معاص وحدوث امور له بسببها حتى انه يتعجب من ذلك غاية العجب ولا تكون خطرة بباله قط.
(1073) If someone dreams that he adored an idol made of wood, he will approach, through a good-for-nothing man, a devious and hypocritical individual. If the idol is made of woven twigs, he will pursue whatever disputes are brought to him, or something like that, and it is said that he will become associated with someone because of slander. If the idol is made of silver, he will go to a woman for something improper. If the idol is made of gold, he will become involved in an abominable situation and harm will befall him because of it. If the idol is made of
copper or iron or lead or something like that, he will become associated <with someone> because of worldly pursuits, and it is said that he will approach a man who behaves like a thief ...
(1075) If someone dreams that he worshiped an idol or talked to it or did with it something that people do when they are awake, he will befriend someone whose friendship will not be beneficial, and harm may befall him because of this friend. It is said that <this harm> will be the perpetration of a sin or something that will happen to <the dreamer unbeknownst to him> because of this <friendship> to the point that <the dreamer> will be greatly amazed <when he finds out>, and will have no idea at all about his situation.
The Oneirocriticon includes a paragraph on fire worship (Drexl 9, 6-11):
Έάν τις ίδη κατ' όναρ, ότι προσεκύνησε τφ πυρί, ούτός έστιν ἐπιζητῶν ἀξιώματα και έξουσίαν διά την του πυρός δύναμιν. ἐάν τις προσκυνήση ανθρακας, ήγουν πυρ άφλογον και άκαπνον και άναιθάλωτον, κτήτωρ έσται χρυσού απο αδικίας άναλόγως της ποσότητος ού είδε πυρός και άποκτήσεται τὴν πίστιν αύτου.
If someone dreams that he worshiped fire, he is seeking office and authority because of the power of fire. If someone worships coals, that is, fire that has no flame, smoke, or soot, he will acquire through wrongdoing gold in proportion to the amount of fire he saw and will lose his faith.
Similar interpretations are given in Arabic dreambooks. According to Ibn إن راى انـه يعــبـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ الشيطان او پطلب الصرب فان لم يكن للنار الهب فإنه حرام يطلبه بدينه . لان الحرام (If someone dreams that he worshiped fire, he will defy God Almighty by showing obedience to the devil or will seek war. If the fire had no flames, it represents something unlawful that he will seek with his religious life, because unlawfulness <is symbolized by> fire). Ibn Shāhīn (no. 1072) adds: من راي انه يعبد النار فانه يعيد النار فانه يعين السلطان، فان كانت النار خامدة . فـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ dreams that he worshiped fire, he will support the sultan, and if the fire abates, he will seek unlawful money. And it is said that the adoration of fire <indicates> serving an unjust king).
<sup>184</sup> Ibn Qutayba, hah 7, Yahuda ar. 196, fol. 27a; the interpretation is repeated verbatim in al-Muntakhah, chap. 46, p. 330; similar interpretations in al-Nābulusī, vol. 2, p. 290, s.v. . Lj.
#### Judges
In the Byzantine Empire, the term krites (judge) could mean any one of a number of high-ranking officials with judicial and administrative authority. 485 The significance of judges in dreams is discussed in chapter 15 of the Oneirocriticon, "From the Indians on Judges and Balances." The qualifications and functions of a Byzantine kritēs do not justify placing this discussion with the religious chapters; the author merely followed the order he found in his Arabic sources, 186 as is confirmed by the fact that almost every interpretation contained in chapter 15 can be matched to interpretations found in Arabic dreambooks:
Έάν τις ϊδη κατ΄ οναρ έαυτον κριτὴν γεγονότα λαοῦ καὶ οὐκ ήν ίκανὸς εἰς τὸ κρίνειν, ούτος, έαν έχη όδον ταξειδίου, ληστευθήσεται έαν δε όδον ούκ Έχη ταξειδίου, πτωχεύσει σφόδρα ἐλεεινὸς γενόμενος. 187
If someone dreams that he became a judge of people and he was not fit to judge, if he is traveling, he will be robbed. If he is not traveling, he will become poor and extremely pitiful.
Cf. Ibn Qutayba, bab 10 (القاضى), Yahuda ar. 196, fol. 27b:
و من راي انه صار قاضيـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ الطريق ان كان مسافرا وان لم يكن مسافرا ابتلي ببعض بلاء الدنيا.
If someone dreams that he became a judge among the people and was not suitable for it, he will be robbed if he is traveling, and if he is not traveling, he will be afflicted with some worldly tribulation. 188
' Εάν τις ΐδη, ότι ἐκρίθη παρὰ ἀγνωρίστου, οῦτος πᾶν, ὅ τι αν αὐτῶ ἐκρίθη, κρατείτω ὁ γὰρ κριτὴς εἰς θεοῦ πρόσωπον ἀνάγεται. 189
If someone dreams that he was judged by an unknown judge, let him hold onto whatever was decided in his case, because the judge is reckoned to be God.
186 In the Muslim lands, a judge (gādī) is appointed by the government on the basis of his superior knowledge of religious law, and by virtue of this connection is usually listed among the religious dream symbols of Arabic dreambooks.
188 The same interpretation is repeated in al-Dinawari, fas! 9, băb 19 (القــاضـي); Esad Efendi 1833, fol. 101a; BN arabe 2745, fol. 154a-b; also in al-Muntakhab, p. 148; Ibn Shāhīn, no. 1098; al-Nābulusī, vol. 2, p. 157, s.v. قاض .
189 Drexl 12, 10-12,
<sup>185</sup> For the meaning of the term in various texts and periods, see ODB, s.v. "Judge," Additional information and bibliography in S. Troianos, Hē thesē tou nomikou dikastē stē Byzantinē koinonia (Athens, 1993).
<sup>187</sup> Drexl 12, 1-6.
#### Cf. Ibn Qutayba, bab 10, Yahuda ar. 196, fol. 27b:
واذا كان مجهولا فانه فى التاويل الله عز وجل لقوله يقضى للحق وهو خير الفاصلين والقوله وقض ربك وهو يقضى بين عباده لانّ كل شىء بقضائه فان رای انه قضی له بامر فهو کما قضی.
And if the judge is unknown, this is interpreted as God, the omnipotent and exalted, according to His saying "He telleth the truth and He is the Best of Deciders" [Qur'an 6:57],190 and according to His saying "Your Lord has decreed" [Qur'an 17:23] and <because> He will judge His servants and because everything is done according to His decree. If he dreams that <the unknown judge> has pronounced a judgment about something in his favor, this will be done according to what was decreed. 191
Eav τις "δη κατ" όναρ ζύγιον ή το λεγόμενον καμπανον έν τόπω τινι σταθμιζόμενα, ταυτα είς πρόσωπον νοείτω κριτού. και έαν έχη ό ταυτα ίδών δίκην, εί μεν ίδη αυτά έξισούμενα έν τώ σταθμίζεσθαι, δικαιωθήσεται, εί δε μή, ου δικαιωθήσεται αί γαρ πλάστιγγες του ζυγού είς τα ώτα άνάγονται του κριτού, τα δε έν ταις πλάστιγξιν είδη οί λόγοι είσι των κρινομένων. ἐαν δὲ τὸ ζύγιον Ίδη ἐπιδέξιον καὶ καθαρόν, νοείτω, ὅτι ὁ κριτὴς του τόπου δίκαιός ἐστιν εἰ δὲ τὰς πλάστιγγας ίδη συνεστραμμένας ή κεκλασμένας, τὸν κριτὴν τοῦ τόπου, ἐν ᠪ ταθτα είδε, νοείτω άδικον. εί δὲ ἴδῃ κλασθέντα τὸν ῥυμὸν τοῦ ζυγίου, ἐν ὡ αί πλάστιγγες ἀπηώρηνται, ϊστω, ὅτι κινδυνεύει ὁ κριτὴς τοῦ τόπου ἢ ἀποθανείται.
Ομοίως και έπι του λεγομένου καμπανού. όμοίως τά τε μόδια και τα μέτρα τὴν αὐτὴν κρίσιν καὶ λύσιν ἔχουσιν ἀναλόγως, κατωτέροις δὲ προσώποις κριτών έφαρμόζονται. 192
If someone sees in his dream that a balance or a steelyard was used someplace for weighing, let him understand it to represent the person of a judge. If the dreamer is involved in a lawsuit and dreams that the scales were balanced in the weighing, he will be vindicated; if they were uneven, he will not be vindicated, because the scales of the balance are reckoned as the judge's ears, while the goods on the scales are the words of the litigants. If he sees that the balance was well adjusted and honest, let him know that the local judge is fair. If he sees the scales turned upside down or broken, let him understand that the judge of the place where he saw this is unfair. If he sees that the pole of the balance that the scales hang from is broken, he should know that the local judge is in danger or will die.
Likewise for the steelyard. And likewise, the modii193 and other measures have the same interpretation and solution, but are applied to judges of lower ranks.
<sup>(99</sup> The quotation reads slightly differently in standard editions of the Qur'ān: پقض الحق .
<sup>19</sup> The same interpretation is repeated in al-Dinawari, fas! 9, băb 19 (القــاضـي) 19 ); Esad Efendi 1833, fol. 101a; BN arabe 2745, fol. 154b. Also in al-Muntakhab, p. 149; Ibn Shahin, no.1100; al-Nābulusī, vol. 2, p. 158, s.v. ضاض .
<sup>192</sup> Drex1 12, 16-13, 3.
<sup>193</sup> The modios was a unit of measure for both grain and land. Cf. ODB, s.v. "modios."
Ibn Qutayba, bab 11, Yahuda ar. 196, fol. 28a:
الميزان قضاء القاضى فما راه فى الميزان من استقامة او ميل كان فى قضاء القاضى مثل ذلك... وكفة الميزان سمع القاضي والدراهم الخصومات فى هذا الموضع شبه اجتماع الخصومات فى سمع القاضى باجتماع الدراهم فى كفة الميزان والصحان العدل وعمود الليزان ولسانه القاضى نفسه والكيال مثل الميزان الا انه دونه.
The balance is the judgment of the judge, and whatever <a dreamer> saw regarding the balance, either straightness or tilt, <indicates that> the judgment of the judge will be analogous ... The scale of the balance is the hearing of the judge and the dirhams 14 are the litigations. In this case, the judge will hear the litigations <and decide> by analogy with the manner in which the dirhams are put on the scale of the balance. The dish is justice; the pole of the balance and its scale are the judge himself. The measure of grain is like the balance, only less <important>.
Al-Dīnawarī, fasl 13, bāb 33; Esad Efendi 1833, fol. 144a:
والميزان القائم الصحيح هو قاضى ذلك الموضع فى عدله وصحة عمله ... وقيل السنجات فقه القاضى... وعموده نفس القاضى... وكفتاه سمعه والسنجات عدله الذى يفصل به القضاء دو > كذلك يحمل العدل على الكلام كــمـــا يـحــمـل الـسنـجـــات على الدراهم فـي كــفــة الاخــرى \* والدراهم هـي الخاصو مات
The steadfast and correct balance signifies the local judge with respect to his fairness and the soundness of his knowledge .... And it is said that the weights <placed as counterpoise on the balance> is his knowledge of jurisprudence ... and the pole of the balance is the judge himself .... Its two scales are the hearing of the judge and the weights are the fairness with which he will make a decision. And likewise, justice will be brought to bear upon the dispute according to how the weights counter the dirhams on the other scale. The dirhams are the litigations.
Cf. also Ibn Shahin, no. 5217:
واما الليزان فانه يؤول بالقاضى... وكفة الليزان هى سمع القاضى والدراهم الذي بكفة الميزان خصومة عند القاضى وصنيج الميزان هي عدل القاضى بين
As for the balance, it is interpreted as the judge ... The scale of the balance is the hearing of the judge and the dirhams that are on the scale of the balance are the
<sup>194</sup> The dirham in this context is a measure of weight. The word can also refer to a monetary unit.
litigations brought before the judge. The brass plates of the balance represent the justice brought by the judge to the litigants. 195
lbn Shahin, no. 5218:
~ J~ <t..:iµ 4-,WL,.. -'-"-I\_})~ 4-:il ~l.i.:......I JL..:.. ~ ,)..:r.,..11 <.SI.) 0-oJ . <L.J ..IL J '-""°W I .\_jl....,. .. :,1
If someone dreams that a balance that was straight tilted to one side, this signifies the fairness and equity of the judge.
lbn Shahin, no. 5220:
'-""°Li WJ-" ~ J~ <Liµ ~I J,.j ,)~I ..l~ (\$1.) :\_,.... ~L..~1 JLi,J . .:,~I dJj
Al-Kirmanl said: "If someone dreams that the pole of a balance is broken, this is interpreted as the death of the local judge."
The interpretation of the Byzantine *kampanon* (steelyard) in the *Oneirocriticon* copies the interpretation of its Arabic counterpart, the *qahhan,* as is evident from the interpretations quoted by lbn Shahin (nos. 5221 bis and 5222):JLl 3 ~WI *µ <sup>3</sup>*~ J..W .:,,4-i-JI ~J.> y-:>.,;-..i-ll ..>-:'4(Jabiral-Maghribisaid: "The dream of a steelyard *(qabban)* signifies the proxy of the judge"). And: ~~WI *J.:.S <sup>3</sup>*.:,I~ J-'-: .c.µ ~.:\_,~I~~ <Lil l>I-> *.:,,..... <sup>3</sup>* • d -, *.!* . .; 9 ~ (If someone dreams that he was weighing something with a steelyard, this signifies that the proxy of the judge will appoint him to judge the litigations brought before <the proxy>).
#### *Hands and Fingers as Symbols of Prayer*
Chapter 72 of the *Oneirocriticon,* "From the Indians on Hands," interprets the fingers of the hand as follows:
<sup>013</sup>' ·EK 'toov' Ivowv ni::p\. xi::tpwv
Ai. XEtpi::i; Kat Ol OclK'tUAOl i::ii; 'tel £pya 'tiii; nicni::coi; 'tcOV av0pcimcov OtaKpivov'tat. £av ni; l01J Ka't' ovap, O'tl 1tAEiovai; OaK'tUAoui; EK'ttj<Ja'tO EV 'tD XElPl m'noii, npoa0Tj<JEl Eti; 'tai; dixai; m'noii Kat ouvmoi; Ea'tat EV 't1j 1tl<J'tEl aU'tOU. oµoicoi; £av ni; <sup>0</sup> l01J, on EK01tll EK 'tcOV OaK'tUACOV au'tOU di;, aµEA.Tj<JEl Kat U<j>EAEt EK 'tcOV
<sup>195</sup> This interpretation is repeated in al-Nabulusi, vol. 2, p. 256, s.v. ,)~ .
#### CHAPTER SEVEN
ευχών αύτου. και πρώτος ό μέγας τών δακτύλων κρίνεται είς την εύχὴν του όρθρου, δεύτερος ήτοι ό λιχανός είς τῆς τρίτης ώρας τὴν εὐχὴν κρίνεται, τρίτος ό μέσος της έκτης ώρας, ό απ' αύτου ήτοι ό τέταρτος της ένάτης ώρας, πέμπτος ό έλάχιστος της έσπερινής ευχής. και δ έαν πάθη τις έξ αυτών, είς τας είρημένας εύχας διακρίνεται. όμοίως έαν ίδη τις, στι έψύγη ή έκόπη τέλεον ή ἐχωρίσθη, εἰς μεγίστην ἀμαρτίαν ἐλεύσεται ἐὰν δὲ ἴδῃ, ὅτι ἀποκατέστη και ιάθη, είς μετάγνωσιν και έπιστροφήν ήξει.
#### 72. From the Indians on Hands
The hands and fingers signify a person's acts of faith. If someone dreams that he had more <than five> fingers on his hand, he will add to his prayers and will be strong in his faith. Likewise, if someone dreams that one of his fingers has been cut off, he will neglect and lessen the number of his prayers. The thumb signifies the morning prayer; the forefinger, the prayer of the third hour; the middle finger, the prayer of the sixth hour; the ring finger, the prayer of the ninth hour; the little finger, the evening prayer. And whatever happens to one of the fingers refers to the <corresponding> prayers just mentioned. Likewise, if someone dreams that <one of his hands> is frozen, or completely cut off, or severed, he will commit a very great sin. If he dreams that it is healed and became healthy, he will repent and reform.
In Christianity, prayer is the means by which the believer communicates with God, but there are no set times or numbers for prayer. The New Testament enjoins Christians to pray without ceasing. Monastic communities from the fourth century onward developed a schedule of daily prayer that evolved into the six (not five) liturgical hours contained in the horologia, the earliest surviving manuscript of which dates from the 9th century.196 In Islam, on the other hand, ritual prayer (salah) is one of the Five Pillars of the faith.'97 Its performance with prescribed movements and recitations at five set hours daily is obligatory from the age of seven. If it is impossible to perform the ritual prayer at the appointed time, it must be made up later. The five fingers representing five daily prayer in the Oneirocriticon is therefore much more appropriate for Islam, and in fact can be found in Arabic dreambooks and was copied from the Arabic source for the Greek text. Ibn Qutayba states it succinctly (bāb 14, fol. 30a): الاصـابـع عـن الـيـد فـهـى الصـلوات الخـمـس (The fingers on the hand are the five daily prayers). Al-Dinawari's relevant passage (fas! 6,
<sup>196</sup> See ODB, s.v. "Horologion," "Prayer," and "Hours, Liturgical."
<sup>197</sup> The Five Pillars of Islam (i.e., the fundamental tenets that all Muslims are required to observe) are the following; the bearing of witness (shahada), the five daily prayers (salah), almsgiving on a stipulated scale (zakāh), fasting during the month of Ramadān (sawm), and performing the pilgrimage to Mecca (hajj).
bab 90; Esad Efendi 1833, fol. 58a) is more detailed:
واصباع اليد اليمنى هى الصلوات الخمس ۾ فالابهام صلوة الفجر والسبابة صلوة الظهر والوسطى صلوة العصر والبنصر صلوة العشا والخنصر صلوة العقـمـي فــان راى ان اصابـعـه قـصــار فــانــه يـتـوافـى فـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ اصابعه اطول واحسن مما كانت فانه قيامه في الصلوة من اصابعه فانه سترك تلك الصلوة.
The fingers of the right hand are the five <daily> prayers: the thumb is the early morning prayer; the forefinger is the midday prayer; the middle finger is the afternoon prayer; the ring finger is the sunset prayer and the small finger is the night prayer. And if the dreamer dreams that his fingers are short, this refers to his prayers; and if he dreams that his fingers became longer and more beautiful than they were, this signifies the proper performance of his prayers. And if one of his fingers is missing, <it means> he is neglecting the prayer that corresponds to that finger. 198
#### The Color Green as Representing Piety and Godliness
In discussing various articles of clothing and ornamentation, the Oneirocriticon frequently refers to the special significance of color for the interpretation of a given object seen in a dream. Green, in particular, is consistently associated with religious faith. For example, five of the six times it is mentioned in the Oneirocriticon it involves religion. The six passages are the following: εί δέ ίδη, ότι έβαλε διπλόν νέον, νέαν γυναϊκαν εύρήσει ... εί δε <έστι> πράσινον, πιστή έσται τα πρός θεόν (If someone dreams he wore a new cloak, he will get a new wife and if it <is> green, she will be faithful in matters pertaining to God); '99 και το πράσινον χρώμα ἐπὶ παντὸς προσώπου είς πίστιν καλὴν διακρίνεται και ἐπὶ νεκροῦ καὶ ζῶντος εἰς σώσμα αυτού διακρίνεται (The color green is interpreted as good faith for every person. For both a dead and a living person it is interpreted as salvation):200 εί δε πράσινον <καμελαύκιον έβαλε> εύρήσει ύψος και ονομα μέγιστον είς τὴν πίστιν αὐτού (If someone puts on a green kamelaukion,201 he will
<sup>198</sup> Repeated verbatim in al-Muntakhab, p. 90, and Ibn Shāhīn, no.1527.
<sup>199</sup> Chap. 156, Drexl 115, 9.
<sup>200</sup> Chap. 156, Drexl 116, 24-26.
<sup>201</sup> A kamelaukion is a kind of cap; for details, see E. Piltz, Kamelaukion et mitra. Insignes byzantins impériaux et ecclésiastiques (Stockholm, 1977).
#### CHAPTER SEVEN
acquire a lofty and exalted reputation on account of his faith);200 εί δε ήν πράσινος <ή ζούπα>, εύρήσει πλούτον και έξουσίαν δια γυναικός (If his outer garment is green, he will find wealth and power through a woman); 200 εί δε ήν έκ πρασίνων λίθων <τὸ στέμμα>, εὑρήσει ὑπέρφημον ὅνομα ἐν πασιν είς τε την πίστιν και τὴν βασιλείαν αὐτου (If the crown was adorned with green stones, he will enjoy a tremendous reputation throughout the world for his faith and rulership);204 and <ei δε ίδη, ότι έλαβε δακτύλιον> εί δέ ἐστιν πράσινος ὁ λίθος <τοῦ δακτυλίου> εύρήσει ἐξουσίαν καὶ πίστιν αγαθην κατά την καθαρότητα του λίθου (If he dreams that he received a ring, if the stone of the ring is green, he will have power and good faith commensurate with the clarity of the stone).205
In the Byzantine world, green was also not devoid of social and political symbolism, since commerce in certain green items was controlled, and court ceremonial dictated its use on particular occasions.206 However, what its religious symbolism was is unclear. The green color of the unpainted relief icons in steatite, which were produced from the tenth century onwards, is associated with purity and is used as a metaphor for the Virgin's immaculate state. But it is uncertain whether this symbolism was understood by the Byzantines as universal and was applied to everything green.207 In the Muslim world, on the other hand, green is indisputably an auspicious color; its special significance is established in the Our'an. where the believers resting in Paradise are said to be dressed in robes of green silk (18:31; 36: 80; 76:21) and "reclining upon green cushions and lovely druggets" (55:76). Green is the color of both heaven and water. The Prophet's standard and the clothes of his son-in-law 'Ali were also green. To this day green turbans are the exclusive prerogative of the descendants of the Prophet through 'Alī, and green is the color of the ceremonial dress of religious scholars, the 'ulama'.
The special significance of green in the Muslim context is also evident in و الحضر ة the interpretations of Arabic dreambooks.2008 Ibn Qutayba states that
200 An analysis of the significance of green in Arabic dream interpretation based on al-Muntakhab,
<sup>202</sup> Chap. 216, Drexl 169, 7-8.
<sup>203</sup> Chap. 225, Drex1 177, 14-15.
<sup>204</sup> Chap. 245, Drexl 201, 19-20.
<sup>205</sup> Chap. 258, Drexl 212, 13-15.
<sup>206</sup> See Hunger, Reich der neuen Mitte, pp. 86-88; see also I. Kalavrezou, Byzantine Icons in Steatite, 2 vols. (Vienna, 1985), vol. 1, p. 84.
<sup>207</sup> To the best of my knowledge, the only discussion of the symbolism of green in Byzantine civilization is by Kalavrezou, Byzantine Icons in Steatie, vol. 1, esp. pp. 31-32 and 79-85.
فى اللقياب جيدة فى الدين لانها لباس اهل الجنة... excellence in religion, because green is the color of the clothes worn by the inhabitants of Paradise).209 Further on in his dreambook, Ibn Qutayba interprets a broken sandal as a journey, and its color as revealing the purpose of the فان کانت سـوداء کانت مـالا وسـؤددا وان کـانـت حـمـر اء کـانـت 042.journey لطلب سرور وان کانت خضراء کانت لطلب دين وان کانت صفراء کان , مرض و هم (If the sandal is black <the journey> is about money and sovereignty, if it is red it is in pursuit of joy, if it is green <the journey> is in pursuit of religion, and if it is yellow <the journey> is about sickness and sorrow). A few lines down, Ibn Qutayba states that a new sandal symbolizes a woman. Her qualities depend on the color of the sandal, according to the symbolism of colors just outlined.211
In his chapter on green clothes, al-Dinawari also associates green with religious piety:212
فان راى انه لبس لبس الاخضر فان الاخضر للحى دين وعبادة لقوله تعالى ويلبسون ٿيابا خضرا من سندس واستبرق وهو للميت حسن حال عند الله تعالى لان تُپاب اهل الجنة خضر وبالحرب ان يكون فارق الدنيا شهيدا ۾ وقبل من ليس خضرة اعطى ميراٿا...
If someone dreams that he wore green clothes, the color green for a living person signifies religion and submission to God, according to the saying of God Almighty: "They shall be robed in green garments of silk and brocade" [Qur'an 18:31]. Green for a dead person <indicates> that he is in a good position beside God Almighty, because the clothes of the inhabitants of Paradise are green.213 At war <a green garment> means that <the person wearing it in the dream> will depart from this world as a martyr. It is said that if someone is wearing green, he will receive an inheritance.
Ibn Shāhīn, and al-Nābulusī was made by I. Zilio Grandi, "Un esempio di interpretazione dei sogni nell'Islam: il colore verde," Annali di Ca' Foscari 26:3 (1987), pp. 53-66; the special significance of green in Islam is discussed on pp. 57-58.
<sup>209</sup> Ibn Qutayba, bah 24, Yahuda ar. 196, fol. 40b.
<sup>210</sup> Ibn Qutayba, bah 25, Yahuda ar. 196, fol. 41b.
<sup>211</sup> An identical interpretation of sandals and their color (omitting green, however) can be found in the Oneirocriticon, chap. 226, Drexl 178, 12-15. The same significance for each of the colors discussed by Ibn Qutayba can be found in various passages in the Oneirocriticon; see, for example, chapter 156, Drex1 115, 6-11 and 116, 15-26.
<sup>212</sup> Faşl 25, bâh 30 (في الخـضــرة من اللبـــاس); Esad Efendi 1833, fol. 251a; repeated in al-Nabulusī; see Zilio Grandi, "Un esempio di interpretazione dei sogni," p. 58.
<sup>213</sup> Cf. Drexl 116, 24-26 mentioned earlier.
Al-Dinawari's interpretations connect precious stones in shades of green with religious righteousness: <sup>214</sup>*t* ~I '-:-'l.+ll J4\_,...l 10--a ..l--~.\_.>-:'j...J 1...9 ..i->-"\_;J I -~ J)l::.. ~JUI~ ,:J\_,l..il-9 ~J t.JJ-9 07...J '-:-'""'"L..o J-=..i......::.J (The emerald and the chrysolite <signify>, among men, a cultured215 and valiant person and a truthful friend who is characterized by religious piety. <They also signify> godliness, high esteem, and attainment of lawful money).
Interpretations of green along the same lines can also be found in *al-Muntakhab,* Ibn Shahin, and al-Nabulusi.216 It is possible that the symbolism of green in Islam coincided with an analogous Christian notion in Byzantium, but if so it is not nearly as well attested and omnipresent as the Muslim one. <sup>217</sup> The evidence of the Arabic dreambooks suggests that the Greek interpretations of green were not prompted by associations this color might have had for the Byzantines, but by the interpretations found in the Arabic sources employed by the Greek author.
#### *Muslim Taboos*
Among the best known of Muslim religious observances is abstinence from alcohol and pork. Given the Islamic provenence of the interpretations contained in the *Oneirocriticon,* one might expect that these Muslim taboos would somehow be reflected in the Christian dreambook, but instead the *Oneirocriticon*
<sup>214</sup> Al-Dinawari,fa# 20, *bah* 32; *Esad Efendi* 1833, fol. 206a.
<sup>215</sup> The text is not vocalized. The word '-:-'Lt--- could mean either "well-mannered, cultured, educated" *(muhadhdhah)* or "teacher, educator" *(muhadhdhih* ).
<sup>216</sup> For references, translation and discussion of further examples from *al-Muntakhah,* Ibn Shahin and al-Nabulusi, see I. Zilio Grandi, "Un esempio di interpretazione dei sogni," esp. pp. 58 ff.
<sup>217</sup> In Revelation 21: 18 the foundation of the wall in New Jerusalem is of jasper, which is thought of as a dark green color. At least two Middle Byzantine descriptions of Paradise mention thrones of green stone. In the 9th-century Life of St. Philaretos the Merciful (who lived in the 8th century), God is said to be seated on an emerald stone; and in the I 0th-century Life of St. Basil the Younger, the saint is said to be seated in Paradise on a green throne. See M. H. Fourmy and M. Leroy, "La Vie de S. Philarete," *Byzantion* 9 (1934), p. 163; A. N. Veselovskij, "Razyskanija v oblasti russkogo duchovnogo sticha," *Shornik Otdelenija russkogo jazyka I slovesnosti lmperatorskoj Akademii nauk* 46 (1889-90), p. 44. However, a system of color symbolism does not seem to have existed in Byzantium; see L. James, *Light and Colour in Byzantine Art* (Oxford-New York, 1996), pp. 108-9.
consistently interprets both drunkenness and alcohol as wealth and authority;218 the eating of pork is also interpreted as representing wealth, and pigs and boars as signifying a strong enemy.219 Although these chapters have no religious overtones, this does not necessarily indicate that the author of the Oneirocriticon departed radically from his Arabic sources. In fact, he followed the Arabic interpretations on the same topics quite closely.
Most of the interpretations appearing in chapter 279, "From the Indians, Persians and Egyptians on Pigs,"220 can be matched with identical interpretations found in the five Arabic dreambooks:
Drexl 226, 14-20: Ο χοίρος είς πρόσωπον καταπεφρονημένου έχθρου δυνατού κρίνεται. ἐαν ϊδη τις, ὅτι ἐπολέμησεν χοῖρον ἄγριον, ἐχθρὸν δυνατὸν καὶ πολυεξόπλιστον πολεμήσει: <ει δε έφόνευσε τον χοιρον, τον έχθρον απωλέσει>. Τουτο δε έξουσιασταις μόνοις θεωρείται άλλος δε έαν Ίδη τουτο, είς πρόσωπον ἐκείνων κρίνεται αὐτό εἰ δὲ πτωχὸς ἢ τοῦ κοινοῦ λαοῦ ὁ ἰδών, εὑρήσει φόβον και θλίψιν και τιμωρίαν άπο έχθρών.
The pig is to be interpreted as a strong and contemptible enemy. If someone dreams that he fought a wild pig, he will fight a mighty and heavily armed enemy; <if he killed the pig, he will slay the enemy>. Such a dream is only dreamt by rulers. If someone else dreams it, it should be interpreted as referring to a ruler. If the dreamer is a pauper or a commoner, he will find fear, sorrow and punishment at the hands of his enemies.
Cf. Ibn Qutayba, bab 39, Yahuda ar. 196, fol. 53b-54a:
و الخنزير رجل شديد الشوكة بدني
The pig represents a corpulent221 man who is vehement in the fury of his fighting.
<sup>218</sup> Cf. chap. 113, "From the Indians on Inebriety," Drexl 66, 21-67, 4; chap. 114, "From the Persians and Egyptians on Inebriety," Drexl 67, 5-14; chap. 195, "From the Indians on Wine and Beverages," Drexl 150, 11-152, 7; chap. 196, "From the Persians and Egyptians on Wine and Beverages," Drexl 152, 8-153, 21. Four additional excerpts from other chapters also associate alcohol with wealth, authority, and figures in a position of power. Cf. chap. 130, "From the Persians and Egyptians on Weddings," Drexl 83, 1-3; chap. 183, "From the Indians, Persians and Egyptians on Warm Water," Drexl 143, 6-8; chap. 207, "From the Indians on Radishes and Carrots," Drexl 162, 22-25; chap. 244, "From the Persians and Egyptians on Cinnamon," Drexl 200, 12-14.
<sup>219</sup> Chap. 279, "From the Indians, Persians and Egyptians on Pigs," Drex1 226, 12-227, 14.
<sup>220</sup> Drexl 226, 14-227, 14.
دنىى Corpulent, fat); this can easily has بَدَنـىَىَ (corpulent, fat); this can easily be corrupted to اينى (contemptible), which occurs in other Arabic dreambooks, as well as in the Greek Oneirocriticon (as καταπεφρονημένος).
Al-Dinawari,fa,s/ 21, *bah* 50, *Esad Efendi* 1833, fol. 218a-b (repeated in al-Nabulusi, vol. 1, p. 198, s.v. ->-=~ ):
$$\text{באַטרייכטייטן די אַז אַז די אַז די אַז די אַז די אַז די אַז אַז די אַז די אַז די אַז די אַז די אַז די אַז די אַז די אַז די אַז די אַז די אַז די אַז די אַז אַז די אַז אַז די אַז אַז די אַז אַז די אַז אַז די אַז די אַז אַז די אַז אַז די אַז אַז די אַז אַז די אַ אַ אַז די אַ אַ אַז די אַ אַ אַז די אַ אַ אַז די אַ אַ אַ }$$
The Muslims said: "The pig is an abominable, strong, crafty, and ugly enemy who becomes apprehensive when there is a misfortune and does not keep his word."
*Al-Muntakhab, bah* 34, p. 191 (repeated in al-Nabulusi, vol. **1,** p. 199):
$$\text{Vže: } \epsilon \text{ -} \epsilon \text{ -} \epsilon \text{ -} \epsilon \text{ -} \epsilon \text{ -} \epsilon \text{ -} \epsilon \text{ -} \epsilon \text{ -} \epsilon \text{ -} \epsilon \text{ -} \epsilon \text{ -} \epsilon \text{ -} \epsilon \text{ -} \epsilon \text{ -} \epsilon \text{ -} \epsilon \text{ -} \epsilon$$
The pig signifies a man who is stout, wealthy, religiously corrupt, wicked in order to make profit, filthy, repulsive, 222 an unbeliever or a Christian, vehement in the fury of his fighting and contemptible.
**lbn** Shahin, no. 5933-5934:
$\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\downarrow\downarrow}$ $\epsilon\_{\downarrow\downarrow}$ $\epsilon\_{\downarrow\downarrow}$ $\epsilon\_{\downarrow\downarrow}$ $\epsilon\_{\downarrow\downarrow}$ $\epsilon\_{\downarrow\downarrow}$ $\epsilon\_{\downarrow\downarrow}$ $\epsilon\_{\downarrow\uparrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\downarrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$ $\epsilon\_{\uparrow\uparrow}$
If someone dreams that he fought a pig, he will fight a contemptible man who has no good in him. And if he defeated the pig, he will attain what he hopes for from his enemy.
Cf. al-Nabulusi, vol. **1,** p. 199:
If someone dreams that he has fought a pig he will triumph over a tyrannical enemy.
Drexl 226, 20-22: El 8£ t81J *w;,* O'tl £moxfrco xoip(J) aypi(J), av8pa µ£ytcr'tOV £x0pov Kma8ouA-ci>cr£t" Ei 8£ fonv 6 i8cl>v rc1rox6~. Ei~ u\j/o~ Kal. rcA-ouwv µ£yav ~~El.
If someone dreams that he rode on a boar, he will vanquish a great enemy. If the dreamer is a pauper, he will attain a lofty position and tremendous wealth.
<sup>222</sup> The word ~Jj does not exist in dictionaries; I decided its meaning extrapolating from the meaning of the verb Jlj =to drive away, to repell (Lane, *Arabic-English Lexicon)* and the participle .J..'..lj =propellens, repellens (Freytag, *Lexicon arabico-latinum).*
Cf. lbn Qutayba, *bab* 39, *Yahuda ar.* 196, fol. 54a (repeated verbatim in *al-Muntakhab,* p. 191; repeated in lbn Shahin, no. 5932):
If someone rides on a pig he will obtain sovereignty and victory over his enemy.
Cf. al-Dinawari,fa,S'l 21, *biib* 50, *EsadEfendi* 1833, fol. 218b; repeated without the Quranic passage in al-Nabulusi, vol. 1, p. 198, s.v. ->-=~:
i *,.r:-* WI \_\_,J~ <U\_,.i.J I~ ':iLo '-:-'L..ol <I->-=~>~.> .WI ..sl.> .:,u ~~li.JIJ41~
If someone dreams that he rode on a pig he will obtain a great deal of money, according to the saying of God Almighty: "He hath forbidden for you only carrion and blood and swineflesh and that which hath been immolated in the name of any other than Allah; but he who is driven thereto, neither craving nor transgressing, lo! then Allah is Forgiving, Merciful" [Qur'an 2: 173 and 16: 115223 ].
Cf. lbn Shahin, no. 5948:
$$\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}\text{•}$$
If someone dreams that he rode on a pig, this signifies the conquest of his enemies and the elevation of his worldly status.
Drexl 226, 23-227, 2: £i. 8£ 'i81] 6 ~acnA.cuc;, on auaypouc; ljvcy1rnv a1rc<\>, lj aui:oc; £8i]paa£ KUVTJYOOV, £x8pouc; OUVal:Ouc; avaA6yooc; 0£aµi]a£t Kat auvcil;Et' El OE lOl], on fiµ£pou.; xoi.pou.; ljvcyrnv aui:0, Ol aui:Ti Kpi.cn.; i:fic; avoo8£v Etpl]µEvl]c;, aU' £A.mi:ovooc; Kat µEi:pi.ooc;.
If the king dreams that boars were brought to him, or that he captured them while hunting, he will capture and bring together an analogous number of strong enemies. If he dreams that domesticated pigs were brought to him, the interpretation is the same as above, but to a lesser degree and more moderate.
Cf. lbn Shahin, no. 5931:
~..i.....'.u .\_ \_ \_:...i ~.; :.ra ~ <t..:i[-Q l\_r:~ '-:-'L..:.I <Lil L>I.; :.ra ~L..~I JuJ ~\_,..:;JI
<sup>223</sup> The two verses are identical; their connection with the interpretation that riding a pig signifies money is not clear.
#### CHAPTER SEVEN
If someone dreams that he obtained a pig he will gain power over a contemptible and strong man.
Drexl 227, 3-4: Εί δὲ ΐδη <τις>, ὅτι τρώγει χοίρειον κρέας, εύρήσει χρυσίον και πλούτον άνάλογον της βρώσεως.
If someone dreams that he ate pork, he will find wealth commensurate to the amount he ate.
Cf. Ibn Qutayba, bab 39, Yahuda ar. 196, fol. 53b-54a:
ولحمه <=لحم الخنزيز > وجلده وبطونه مال حرام دني
The meat, skin, and entrails of a pig represent unlawful and contemptible money.
Cf. al-Dinawari, fast 21, bab 50, Esad Efendi 1833, fol. 218b; repeated in al-Nābulusī, vol. 1, p. 198, s.v. خنزير:
فان راى انه اكل لحمه اكل حراما وهو يعلم \* فإن كان لحمه مطبوخا نال فى تجارته مالا من غير حله وكذلك المشوى.
If someone dreams that he ate <pig's> meat he knowingly ate something unlawful. If its meat was cooked, he will obtain money from commerce that he was not supposed to make. Likewise if its meat was roasted.
Cf. also al-Nabulusī, vol. 1, p. 199:
ومن راي انه پکل لحم خنزير فانه يصيب مالا حراما محضا.
If someone dreams that he ate the flesh of a pig, he will obtain exclusively unlawful money.
Al-Muntakhab, p. 191:
ولحمه وشحمه وبطنه وجلده مال حرام دنيء.
Its flesh, fat, belly, and skin represents unlawful and contemptible money.
Ibn Shahin, no. 5934:
ولحم الخنزير مال حرام.
The flesh of a pig represents unlawful money.
Drexl 227, 4-5: εί δε ίδη τις, ότι περιπατεί ώσπερ χοιρος, εύρήσει χαράν ταχελαν.
If someone dreams that he walked like a pig, he will soon find pleasure.
Al-Dinawarī, fasl 21, bab 50, Esad Efendi 1833, fol. 218b; repeated in al-Nābulusī, vol. 1, p. 198, s.v. ذنزير:
ومن راي انه يمشى كما مشى الخنزير اصاب قرة عين عاجلا
If someone dreams that he walked like pigs walk, he will soon find pleasure.
As these examples demonstrate, Arabic dream interpretation does reflect the Islamic abomination of the pig and of the consumption of its flesh. Although the flesh of any living being signifies money and wealth, 24 the money represented by unclean pork is unlawful (haram). This prejudice is much attenuated in the Oneirocriticon, where the wealth and the enemy225 signified by the pig are not given negative characteristics, although they must have been prominent in its Arabic sources. The Greek author appears to have "de-Islamicized" the chapter on the pig simply by removing the negative characteristics when he copied the Arabic interpretations.
Something similar seems to have happened with the interpretations of drunkenness and alcohol, which also lost many of their negative attributes in their transfer from Arabic to Greek. The majority of the interpretations found in the "Indian" chapters of the Oneirocriticon have parallel Arabic interpretations:
Drexl 66, 21-67, 4: ' Eάν τις ίδη, ότι ἐμεθύσθη οϊνω, εὐρήσει πλουτον καὶ ἐξουσίαν ἀνάλογον τῆς μέθης. ἐάν τις ἴδη, ὅτι ἐμεθύσθη ἄνευ οῖνου, εὐρήσει φόβον ανάλογον και τιμωρίαν από έξουσίας. ἐάν τις ἴδη, ὅτι ἐμεθύσθη ἀπὸ σκευαστού γλυκέος, δουλεύσει μεγίστω ανθρώπω και έξ αύτου εύρήσει πλούτον. ἐαν ίδη, ὅτι ἐμεθύσθη ἀπὸ ὕδατος, φανήσεται πλούσιος γευδὴς καὶ ἐπικαυχήσεται ἀλλοτρία ἐξουσία.
If someone dreams that he was intoxicated with wine, he will find wealth and power analogous to his intoxication. If he dreams that he was inebriated without <having drunk> wine, he will suffer an equal amount of fear and punishment at the hands of the authorities. If someone dreams that he became drunk from a sweet fermented beverage, he will serve a powerful person and receive wealth from him. If he dreams that he became drunk from water, his wealth will prove to be false and he will boast of power that is not his.
<sup>224</sup> This includes both human and animal flesh. Cf. Oneirocriticon, chap. 87 (Drexl 53, 10-18); chap. 272 (Drexl 223, 2-3); chap. 276 (Drexl 225, 5-7); chap. 277 (Drexl 225, 19-20); chap. 278 (Drexl 226, 6-7); chap. 281 (Drexl 229, 9-11); etc.
<sup>225</sup> It only briefly mentions that this enemy is "contemptible" (καταπεφρονημένος) and dwells more on his strength.
#### CHAPTER SEVEN
Drexl 67, 6-7: Η μέθη είς κόρον λογαρίου κρίνεται. ἐάν τις ϊδη, ὅτι ἐμεθύσθη άπο οίνου, πληθος χρυσίου κταται.
Drunkenness is interpreted as abundance of riches. If someone dreams that he was drunk with wine, he will acquire much wealth.
Similar interpretations of wine and drunkenness are repeated in chapter 195, "From the Indians on Wine and Beverages":
Drexl 150, 11-151, 2: Έαν ίδη τις, ότι έπιεν οινον ακρατον ή μεθ' ύδατος, εύρήσει χρυσίον ανάλογον και έξουσίαν. εί δε πιών έμεθύσθη, εύρήσει χρυσίον έξ άδικίας και άξίωμα μεῖζον άνάλογον τῆς μέθης εἰ δὲ ΐδη, ὅτι έμεθύσθη δίχα οίνου, τουτο έξουσία κρίνεται, άλλα ψευδής αποβήσεται. ό βασιλεύς έαν ίδη, ότι ἐμέθυσε, ἐξουσιαστῶν κυριεύσει καὶ χαρὰν μεγίστην ευρήσει δια τον οίνον. ἐαν ίδη, ὅτι ἐπιεν ἐκ τῶν φοινίκων οἶνον, εὐρήσει πλούτον έξ ἐθνῶν διὰ κολάσεως διὰ τὴν ἐκ πυρὸς κόλασιν τῆς ἐψήσεως τοῦ οίνου.
Εί δε ίδη, ότι έπιεν οινον έκ σαχάριτος το λεγόμενον ζουλάπιν, εί μεν ἐμεθύσθη, εὐρήσει πλουτον καὶ ἐξουσίαν μετὰ κόπου διὰ τὸ πῦρ· εἰ δὲ οὐκ έμεθύσθη, είς πλουτον και μόνον τουτο κρίνεται. εί δε ίδη τις, ότι έπιεν τον ἐκ σταφίδων οῖνον καὶ ἐμεθύσθη, εὑρήσει πλοῦτον καὶ ἐξουσίαν ἀπὸ γυναικών. ὁ οίνος ὁ καθαρὸς εἰς χρυσίου κτῆσιν ακοπον καὶ ἐξουσίαν ὑψηλὴν έκρίθη. και ταυτα τα ποτά τα άπο σκευασίας δια τουτό είσιν είς κόπον και ἐξουσίαν ὑπεύθυνον διὰ τὴν ἐκ τοῦ πυρὸς ἔψησιν.
If someone dreams that he drank wine, either pure or mixed with water, he will acquire a corresponding amount of wealth and power. If he drank it and became intoxicated, he will attain wealth through wrongdoing and a major office commensurate with his intoxication. If he dreams that he became inebriated without wine, this is interpreted as power, but it will turn out to be false. If the emperor dreams that he became inebriated, he will vanquish powerful people and have great joy because of the wine. If he dreams that he drank wine made from dates, he will acquire through punishment wealth from foreign races, because of the fire <used> in distillation <which symbolizes punishment>.
If he dreams that he drank wine made of sugar cane, the so-called julep, if he becomes inebriated he will find wealth and power through toil because of the fire <of the distillation>. If he does not get drunk, <the dream> is interpreted as wealth only. If someone dreams that he drank wine made from resins and became inebriated. he will be granted by women terrible wealth and power. Undiluted wine is interpreted as the effortless acquisition of wealth and lofty power. Distilled beverages are interpreted as toil and power with many responsibilities because of their preparation with fire 226
<sup>226</sup> On the preparation of julep and wine, see Paris, gr. 2419, fol. 154v (inventory of the recipes in CMAG, vol. 1, p. 63); on the various methods of preparing and serving wine and on various beverages consumed by the Byzantines, see Ph. Koukoules, Byzantinon bios kai politismos, vol. 5, pp. 121-35.
Ibn Qutayba, bab 20 (الاشربة), Yahuda ar. 196, fol. 37a-b:
الخمر مال حرام بلا نصب والسكر منها مال وسلطان و السكر من غير شراب خوف شديد لقوله عز وجل وتري الناس سكارى وما هم بسكارى ولكن عذاب الله شديد فيه نصب
Wine is unlawful money <acquired> without toil; intoxication from it is money and sovereignty. Intoxication without drinking represents great fear, according to the saying of God Almighty: "And thou shalt see mankind drunk, yet they are not drunk, but God's chastisement is terrible" [Qur'an 22:2]. Nabīdh22] is licit money; the wickedness <involved in acquiring it> will be analogous to how strong the nabidh was when drinking it, and the toil analogous to the fire needed for its preparation.
Cf. al-Dīnawarī, fasl 24, bāb 2 ( خمر ), Esad Efendi 1833, fol. 143a:
والخمر مال حرام ليس فيه نصب و لا كد.
Wine is unlawful money <acquired> without labor and toil.
Cf. al-Dīnawarī, fașl 24, bāb 7 (السكر ); Esad Efendi 1833, fol. 143b:
قال المسلمون كل الاشربة اللكروهة لا خير فيها والسكر اشدها وهو هم وزن < > لقوله تعالى لا تقريبوا الصلاة وانتم سكارى وقوله عز ذكره وترى الناس سكاري وما هم بسكاري من نبيذ او ما يستحل شربه فهو سلطان على كل حال الا ان اضعف فى المبلغ واوهن ... وهن راي انه شرب خصرا وسکر هنه فانه يصيب مالا هراما ... وكذلك السكر سلطان ومال اذا كان من شراب والسكر من غير شراب خوف شديد بمغلب السكر لقوله تعالى وترى الناس سكار.
The Muslims said: "Nothing good will come from <dreaming of> any of the reprehensible beverages. Drunkenness is worse than they are. It is trouble and ... 28 according to the saying of God Almighty: "Draw not near to prayer when you are drunken" [Qur'an 4:43] and according to His saying: "And thou shalt see mankind drunk, yet they are not drunk, but God's chastisement is terrible" [Qur'an 22:21. Drunkenness represents inexhaustible wealth combined with vanity. If the intoxication comes from nabidh or something <else> that is lawful to drink, this
<sup>221</sup> Nahīdh is an alcoholic beverage made from dates, raisins, or grains. Its use by Muslims is lawful (see Hughes, Dicitionary of Islam, s.v. "nabidh" ).
<sup>228</sup> Illegible word.
#### CHAPTER SEVEN
represents authority in every case, unless it is weak and ineffective in its results. If someone dreamt that he drank wine and was intoxicated by it, he will receive unlawful money ... Likewise, inebriation signifies authority and wealth, if it is the result of drinking. Becoming intoxicated without having drunk signifies great fear commensurate with the degree of <the dreamer's> intoxication, according to the saying of God Almighty: "And thou shalt see mankind drunk, yet they are not drunk, but God's chastisement is terrible" [Qur'an 22:2].
Al-Dīnawarī, fașl 24, bāb 4 (نبيذ من التمر والزبيب), Esad Efendi 1833, fol. 143b:
اما نبيذ التمر فان صرفه مال فيه شبهة... وقالوا النبيذ مال حلال فيه كد ونصب وغنى لما نالت منه النار وهو خيير الامـوال التى هى تـاويل الخـمـر لمال النبيد مع الاشربة # فاما شرب النبيد فمن راي انه شرب نبيدا او غيره مما يسكر فانه مال تحليله وتحريمه بقدر كراهة ذلك الشرب فى الدين واليقظة ويكون فى ذلك المال من النصب والكلام والعلاج له بقدر ما نالت النار منه.
As for nabidh <made> from dried dates, its unmixed state is money characterized by uncertainty as to its lawfulness .... < The dream interpreters> said: "Nabidh is lawful money combined with hard work, toil, and wealth commensurate with how much the fire affected it. In the case of nabidh and <other> drinks in its category, <the dream also signifies> the blessing of the wealth which is <generally> signified by wine. As for drinking nabidh, if someone dreams that he drank it, or something like it, from the category of drinks that cause intoxication, this indicates money which is either lawful or unlawful, depending on the status of the beverage with regard to religion and to life when one is awake. The money <signified by this dream> will be acquired through toil, dispute and effort commensurate with the fire needed for its preparation."229
Cf. Ibn Shāhīn, no. 4346:
ومن راي انه يشرب خمرا فسكر منه فإنه يصيب مالا حراما ويصيب من ذلك الملان سلطنة بقدر السكر، وان السكر بغير خمر فانه يصيب هم وخوف شديد لقوله تعالى وتراى الناس سكارى الآية.
If someone dreamt that he was drinking wine and got drunk from it, he will acquire unlawful money and, from this money, he will also attain power analogous to his drunkenness. If he became drunk without wine, sorrow and terrible fear will befall him, according to the saying of God Almighty: "And thou shalt see mankind drunk, yet they are not drunk, but God's chastisement is terrible" [Qur'ân 22:2].
<sup>229</sup> Ibn Qutayba's and al-Dinawari's interpretations are repeated in al-Muntakhah, pp. 134-35; al-Dinawarî's interpretations are repeated in al-Nābulusī, vol. 1, p. 309, s.v. سكر and vol. 1, p. 182, s.v. خمر.
Drexl 151, 3-12: Εαν ίδη τις, ότι διαμάχεταί τινι ένεκεν πόσεως οίνου έν ποτηρίω, εί μεν γνωρίζει αύτόν, μαχήσεται μετ' αὐτοῦ χάριν χρυσίου καὶ νικος και κέρδος εύρήσει ό νικήσας, εί δε ου γνωρίζει αύτόν, έχθρώ μαχήσεται ένεκεν τούτου εί δε ίδη, ότι έμαχήσατο διά το ποτήριον το ύέλινον, περι γυναικός μαχήσεται αύτφ. εί δε ΐδη, ότι έπιεν οινον συγκεκερασμένον, εύρήσει θλίψιν είς τον πλουτον αύτου άναλόγως της ζέσεως, εί δε ϊδη, ότι ἐκίρνα αὐτῷ έτερος γνώριμος τουτον τον οίνον, δια τον κιρνώντα έκεινον θλιβήσεται εί δέ αγνώριστος ήν, δι' έχθρου.
If someone dreams that he is quarreling with someone over drinking wine in a cup, if he knows him, he will quarrel with him over gold, and the winner will attain victory and profit. If the dreamer does not know the person he is quarreling with, he will fight an enemy for this reason. If he dreams that he fought over a glass goblet, he will fight his opponent over a woman. If he dreams that he drank wine mixed with warm water, he will grieve over his wealth by analogy to the warmth <of the drink>. If he dreams that this wine was mixed by someone he knew, <the dreamer> will grieve because of this man; if he did not know the man, he will grieve because of an enemy.
Ibn Qutayba, bab 20 (الاشربة), Yahuda ar. 196, fol. 37a-b:230
ومنازعة الكأس منازعة الخصوصة وكذلك منازعة الدلاء.
Fighting for a goblet <of wine> signifies fighting over litigation; fighting over buckets <of wine is interpreted> likewise.
Al-Dīnawarī, fasl 24, bāb 2 ( خمر ), Esad Efendi 1833, fol. 143a:
ومن راي انه يشربها وليس معه من ينازعه فيها فانه يصيب مالا حراما وقالوا بل مالا حلالا و الخاصومات والصخب بقدر ذلك بع فان كانوا ثلثة او رسادة فنقدر ذلك.
If someone dreams that he was drinking <wine> and there was no one with him to fight about it, <the dreamer> will acquire unlawful money; and even lawful money. If someone drank the wine and did not fight over the cup, he will fight disputes and litigation, and the uproar will be analogous to what he saw in his dream. And if there were three or more <persons in the dream, its interpretation> is by analogy to their number.
Drexl 151, 13-14: Έαν ίδη τις, ὅτι πατεί ἐν ληνώ καὶ ποιεῖ οἶνον, δουλεύσει έξουσία μεγίστη και καταπιστευθήσεται διοικήσεις μεγίστας.
<sup>230</sup> Repeated in al-Muntakhah, p. 134, 11. 15-17.
#### CHAPTER SEVEN
If someone dreams that he was walking in a vat and was making wine, he will serve a person of great authority and be entrusted with very important administrative matters.
Ibn Qutayba, bab 20 (الاشربة), Yahuda ar. 196, fol. 37a-b:231
ومن اعتصر خمرا خدم سلطانا واخصب وجرى على يده امور عظام.
If someone dreamt that he was pressing <grapes to make> wine, he will serve a person of authority and prosper and carry out important affairs.
Cf. al-Dinawarī, fasl 24, bāb 2, Esad Efendi 1833, fol. 143a:
فمن رای انه یعصر خمرا فانه یخدم السلطان ویجری علی یده امور عظام. If someone dreams that he was pressing <grapes to make> wine, he will serve a person of authority, 22 and will carry out important affairs.
Drexl 151, 15-21: Έαν ίδη τις ποταμόν έξ οινου ρέοντα, όπερ ανεπίδεκτόν ἐστιν ἑτέρῳ θεαθῆναι πλὴν βασιλέως ἢ μεγίστων, ἢ μὲν ἐξουσιάζει τὸν ποταμόν, εύρήσει ἐξουσίαν μείζονα καὶ πόλεμον κατ΄ ἐχθρῶν καὶ ἀνυπότακτος ἐσται πασιν εί δε έτέρων ό ποταμός έδεσπόζετο, είς τὴν δεσποτείαν ή κρίσις. έαν ίδη τις, ότι ήγγισε τώ τοιούτω ποταμώ ή ήρεν έξ αύτου, εύρήσει έξουσίαν ανάλογον ού ήρε ποσού έκ του δεσπότου του ποταμού.
If someone dreams of a river flowing with wine, which is impossible for anyone other than an emperor or nobleman to dream, if he is master over the river he will be given great authority, will wage war against his enemies and will not succumb to anyone. If others were masters over the river, the interpretation <of the dream> refers to these masters. If someone dreams that he approached this river or took wine from it, he will be given power by the master of the river by analogy to how much wine he takes.
Arabic dreambooks also interpret dreaming of a river flowing with wine as having some association with fitna, which means both "temptation" and "riot" or "civil strife". Given that the Oneirocriticon connects a river of wine with war against one's enemies, the second meaning of fitna should possibly be given preference:
<sup>231</sup> Repeated in al-Muntakhab, p. 134; and Ibn Shahin, no. 4349.
<sup>232</sup> The Greek phrase δουλεύσει ἐξουσία μεγίστη (lit., "he will serve an authority") has its exact parallel in the Arabic phrases خدم سلطانا from Ibn Qutayba and يخدم السلطان from al-Dinawari, since the Arabic word sultan in earlier texts (such as the Qur'an) means "authority, " while in later texts it also means "ruler" , that is the person in authority.
Ibn Qutayba, *bab* 20 (~~~I), *Yahuda ar.* 196, fol. 37b:
. ....:..... JL:.1 L... ..>.i..i.:> ~ cil:.....:...., '-:-'Lol-9 ~ ..:\_,..... I-\* lSI..> ..:\_,.......!
If someone dreams of a river of wine and took some from it, *fitna* will influence him by analogy to how much he took.
Cf. al-Dinawari,fa~/ 24, *biib* 2 (~), *Esad Efendi* 1833, fol. 143a:
~J d...1:..J ..JU o~J ~ ~ ~ <L.ju ~..:\_,.....I-\* '-:-'L.....:.1 <L.jl lSI..> ..JU . ....:.....JL.:.L....;.i..i.:>~~
If someone dreams that he reached a river of wine, *fitna* will befall him in his worldly life. If he stepped into the river, he will meet with233 *fitna* by analogy to how much he took from the river.
Cf. Ibn Shahin, no. 4350:
If someone dreams of a river of wine, it has a twofold interpretation: if he entered the river he will meet with *fitna* and harm. If he did not enter the river, this is interpreted as a change in his leaders.
Arabic dreambooks do include negative interpretations of wine. For example, al-Dinawari states (ja~l 24, *bab* 2, *Esad Efendi* 1833, fol. 143a): <t...:. I <.SI..; ~ I~ WI ... *.:* . *<sup>0</sup> .:* <t...:.U I~ '-:-' ~ (If someone dreams that he drank wine, he will fall into grave sin). Conceivably, such interpretations did exist in the Arabic sources of the *Oneirocriticon,* and the Greek author simply omitted them. There was plenty of neutral material he could draw from, and this is evidently what he did.
The interpretations of Christian dream symbols in the *Oneirocriticon* are based on equivalent interpretations of Muslim dream symbols found in Arabic dreambooks. Depending on the dream symbol, the Greek author either did not change anything (e.g., Resurrection, Paradise and Hell, Angels), or made minimal changes in order to transform Muslim notions into Christian ones (e.g., mosques become churches), including omitting details that would associate a
<sup>233</sup>~ ~J· *Esad Efendi* has ~ ~l.c...9 instead, which makes no sense. I corrected the text on the basis of identical phrases that occur in *al-Muntakhab* (chap. 18, p. 134) and al-Nabulusi, vol. I, p. 182, s.v. ~-
#### 352 CHAPTER SEVEN
dream symbol and its interpretation too closely with Islam (e.g., in his handling of well-known Muslim taboos). Though it is not always possible to extrapolate which Arabic interpretation inspired a given Greek one (as in the interpretation of monks), it is safe to conclude that the Greek author did not invent new interpretations, but kept as close as possible to his Muslim sources.
#### CHAPTER EIGHT
### REFLECTIONS OF THE OLD AND THE NEW TESTAMENT IN THE INTERPRETATIONS OF THE ONEIROCRITICON
Artemidoros, the earliest of the authors that we are examining, based a good number of his interpretations not only on language and figures of speech, but also on a body of literary texts, including Homer, Hesiod, Euripides and Menander, that must have enjoyed a wide readership in the ancient world. Many of the interpretations quoted in the Babylonian Talmud, on the other hand, are based on the Torah, a text of religious rather than literary significance. Arabic dream interpretation is based on both sacred and literary texts. Its interpretations are inspired by the Qur'an and the hadith, but also poetry, proverbs, etc. The Quranic passages are introduced with the standard phrase قال الله تعالى or قال الله تعالى ("God the Most High said" or "The Most High said"4); the passages from the hadith are also introduced with a standard phrase, قــال صليـــه وسلم or قـــال النبـــى صلى اللـه عليـــه وسلم , phrase ("The Prophet said, may the peace and blessings of God be upon him," or "He said, may the peace and blessings of God be upon him"), so they are immediately identifiable, even by someone who is not familiar with the original texts.
Since the Oneirocriticon is a Christian adaptation of Muslim material, it is important to ascertain whether it rejects or retains the Quranic contents. The interpretations based on Quranic passages were certainly repeated in the Oneirocriticon, without, however, translating the relevant verse. But does the author try to substitute for the Quranic passages similar passages from the Christian revelation or even from the Old Testament?
For example, Artemidoros cites Theognis (i.32 Pack 41, 5-6: i.66 Pack 71, 20-21), Homer (i.50 Pack 56, 2-3; iv.proem. Pack 239, 6-13; iv.59 Pack 285, 2-3), Menander (ii.4 Pack 106, 1-2; ii.12 Pack 125, 15), Euripides (ii.10 Pack 116, 10; iv.59 Pack 284, 15-22), Hesiod (iv.59 Pack 284, 12), among others.
<sup>4</sup> Contrary to the Christian Gospels, which were written by divine ispiration but by human beings, the Qur'an is considered to be the word of God itself. Other eulogistic formulas that accompany the name of God in Arabic dreambooks are: قـال الله عـز وجل، قـال الله ســــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ "God Almighty. و تعالى ، قال الله جل ذكره
<sup>3</sup> For examples, see the sections titled "Crucifixion and Cross," "Decapitation and Eating a Human Head," and "Elephant" in chapter 5; also the sections on "Resurrection of the Dead," "Paradise and Hell," and "Color Green as Symbol of Piety and Godliness" in chapter 7.
This is an open access chapter distributed under the terms of the CC BY 4.0 license.
#### CHAPTER EIGHT
The author says that he uses a scriptural quotation only twice. The first time, the quotation is incorporated into the introduction to the dreambook according to Syrbacham, the dream interpreter to the king of the Indians, to whom the Christian interpretations throughout the Oneirocriticon are ascribed. In this case the author of the Oneirocriticon in all likelihood quoted a passage from the Gospels to replace a Quranic passage with a Christian one.4 Both the Muslim and the Christian quotations convey a similar message. The second instance is provided to justify an interpretation in the same way as Quranic quotations are used in Arabic dreambooks:5
ἐάν Ίδη τις, ὅτι ό πάλαι νεκρὸς ἀνέζησε καὶ λέγει· "ὅτι οὐκ ἀπέθανον, ἀλλὰ ζώ" και θαυμάζει ό ίδών, τουτο είς σώσμα του νεκρού κρίνεται, πληροφορούμενον έκ της θείας γραφής λεγούσης, ότι ό θεός ούκ έστι νεκρών, άλλα ζώντων. νεκρός ούν ό βλεπόμενος, ότι ζη, σέσωσται.
If someone dreams that a person who is already dead returned to life and said "I did not die, but I am alive," at which the dreamer was amazed, this represents the dead man's salvation, which has been fully assured in the Holy Writ: "He is not the God of the dead, but of the living" [Matthew 22:32; Mark 12:27; Luke 20:38]. Thus, a dead man seen as being alive has been saved.
The same interpretation is given by Ibn Qutayba, but he justifies it with a Quranic passage: 6
ومن راي ميتا فاخبره انه حي فهو صلاح لماله لقول الله عز وجل و لا تحسين الذين قتلوا فى سبيل الله أمواتا بل احياء عند ربهم يرزقون
Whoever dreams that a dead person informs him that he is alive, this is goodness in <the dead man's> condition <in the Hereafter>, according to the saying of God Almighty, "Think not of those who are slain in the way of God, as dead. Nay, they are living. With their Lord they have provision" [Qur'an 3:169].
An identical interpretation justified with the same Quranic passage, is quoted by al-Dīnawarī:7
ومن راي ميتا قد عرفه واخبره بانه لم يمت فان ذلك صلاح ذلك الليت فى الاخرة لقوله تبارك وتعالى تحسبن الذين قتلوا فى سبيل الله أمـواتا بل
<sup>4</sup> See chapter 4, pp. 161-163.
<sup>5</sup> Drexl 83, 15-19,
<sup>6</sup> Ibn Qutayba, Yahuda ar . 196, fol. 32b.
<sup>7</sup> Al-Dīnawarī, fasl 29, bāb 16, Esad Efendi 1833, fol. 269a.
$$\text{אַר די קאַנען די קאַנען די קאַדיאַר } \vec{\nabla}$$
Whoever dreams of a dead person whom he knew and who informs him that he is not dead, this is goodness in the condition of this dead person in the Hereafter, according to the saying of God Almighty: "Think not of those who are slain in the way of God as dead. Nay, they are living. With their Lord they have provision" [ Qur'iin 3: 169]; and the dead do not speak lies. Likewise, if the dreamer thought in his sleep that he would never die, he will be slain in the way of God <i.e., he will die a martyr>.
Similar interpretations are also given in the remaining three Arabic dreambooks. 8 In both instances where a biblical passage is quoted in the *Oneirocriticon,* it replaced a Quranic passage quoted in its Arabic sources. Since there are numerous interpretations that Arabic dreambooks base on passages from the Qur'an, the two biblical passages quoted in the *Oneirocriticon* suggest that finding a quotation from the New Testament with a meaning analogous to a given Quranic passage is not easy, and considerable effort must have been spent in identifying the two New Testament passages that finally did appear. The first biblical passage in the introduction is important for the whole work, as it gives a Christian justification for dream interpretation in general; the effort expended by the author to locate it is therefore understandable. The second quotation inserted in order to support one of the many interpretations does not, at first sight, seem to be of special importance.
What is the particular meaning of Qur'an 3: 169, the verse that consistently appears in this place in the Arabic dreambooks and which the author of the *Oneirocriticon* apparently wanted to substitute with a passage from the New Testament? Muslim commentators of the Qur'an have sometimes interpreted
<sup>8</sup> The first section of al-Dinawari's interpretation is repeated in *al-Muntakhab,* chap. 16, p. 63; the second section is repeated verbatim in ibid., p. 59; Ibn Shahin's relevant entry (no. 2586) appears to be a combination of the two interpretations quoted by al-Dinawari: IJ..:>.I .:\_,I .sl.; 0-4.J ~ ~ .:.,~.; <l.J I ~ ,.\_,? *J.::..i=* -i..:.µ 1-'-:'I .::,,~ *'J* -i..:.t,. ·~~I ,.\_,? <...l\_,.:i ~.:.,...... ~'JI .~I J.,. L.::.l.J-Ai <l.JI ~ ,.\_,? l\_,.1:...l ~lll ~ *'i.;* .\_,JW <.J\_,.J.J dJj (Whoever dreams that one of those who, in waking life, is reliable in his words, informed him that he will never die, he will be slain in the way of God <i.e., he will die a martyr> and will always live after that, according to the saying of God Almighty: "Think not of those who are slain in the way of God as dead. Nay, they are living. With their Lord they have provision" [Qur'an 3: 169]). Al-Nabulusi gives an abbreviated version of the interpretations found in lbn Qutayba and al-Dinawari without the Quranic passage (vol. 2, p. 280, II. 5-6, s.v . .::,,.;-o ): ,.\_,?<Wu~ ,.\_J <i..:.I ~ ·~I 0-4.J • ~ "j I ~ F . I ..i.+..!....l I rll.... (When a dead person informs someone that he did not die, this dead person is in the place reserved for martyrs, luxuriating in the abode of the Hereafter).
#### 356 CHAPTER EIGHT
it as an elucidation of Qur'an 4: 157-58, which states that Jesus was not killed but raised up to God,9 but its meaning is most often connected with the special status attained in the Hereafter by the martyrs who die for the Muslim faith, and this is evidently the way it is understood in the dreambooks where it is cited. The martyrs of *jihad,* that is the victims of Holy War, are guaranteed entry into Paradise. Though a concept of a Christian holy war never developed in Byzantium in response to *thejihad, <sup>10</sup>*both the civil and the religious authorities were aware of its existence among their Muslim adversaries and Byzantine theologians, including Theodore Abu Qurra and Niketas of Byzantium, who lived earlier than the tenth century, launched arguments against it in their polemical writings. 11 It is therefore conceivable that the special effort expended by the author of the *Oneirocriticon* to substitute a passage from the New Testament for Qur'an 3: 169 might represent his Christian answer to the significance of this verse either in the context of Christology or of *jihad.*
In addition to these two quotations modern scholars have discerned other biblical influences in the *Oneirocriticon.* Both its editor and its two most recent translators believed that some of its interpretations were inspired by passages from the Bible. 12 The question is whether they represent innovations introduced by the Christian author into his Muslim material, or whether they
<sup>9</sup> See *Concise Encyclopedia of Islam,* s.v. "Martyrs"; also Parrinder, *Jesus in the Qur'iin,* p. 113.
<sup>10</sup>See A. Laiou, "On Just War in Byzantium," in *TcfEAA1)Vt1<6v. Studies in Honor of Speros Vryonis, Jr.,* ed. S. Reinert, J. Langdon, J. Allen, vol. 1 (New Rochelle, 1993), pp. 153-74; and N. Oikonomides, "The Concept of 'Holy War' and Two Tenth-Century Byzantine Ivories," in *Peace and War in Byzantium. Essays in Honor of George T. Dennis, SJ.,* ed. T. Miller and J. Nesbitt (Washington, D.C., 1995), pp. 62-88. For the opposite view, see A. Kolia-Dermitzake, "He idea tou hierou polemou sto Byzantio kata ton dekato aiona: he martyria ton taktikon kai ton demegorion," *Konstantinos* Z *ho Porphyrogennetos kai he epoche tou* (Athens, 1989), pp. 39-58. For a broader view, see A. Kolia-Dermitzake, *Ho Byzantinos "hieros polemos" : he ennoia kai he proho/e tou threskeutikou polemou sto Byzantio* (Athens, 1991 ). The author dates the origin of the idea of Christian holy war to the 4th century A.D. and follows its development until the I Ith. For a review by W. Kaegi that challenges many of Dermitzake's arguments, see *Speculum* 69:2 (1994), pp. 518-20.
<sup>11</sup>For an overview and references to sources, see Khoury, *Polemique byzantine contre /'Islam,* pp. 243-59.
<sup>12</sup> Drexl, *Achmetis Oneirocriticon,* p. 242, gives a list of *loci laudati* to his edition. Only two of these *loci* suggest that interpretations were actually based on the New Testament (the interpretation of the fruit of Paradise and of locusts); the rest simply signal biblical influences in the language and expression of the *Oneirocriticon.* As is evident from the footnotes in their respective translations, both Oberhelman and Brackertz agreed with him and especially Brackertz singled out further passages which they thought were inspired by either the Old or the New Testament.
copy interpretations already quoted in the Oneirocriticon's Arabic sources.13
#### Locusts
One interpretation thought to have been inspired by the New Testament is that of locusts representing a host of enemies: 4 καθ όλου ή άκρις ἀπαραλλάκτως είς πληθος έχθρών κρίνεται ούτως γάρ γέγραπται, ότι θείω κελεύσματι έκστρατεύονται είς απώλειαν τόπου (In general, a locust is unfailingly interpreted as a host of enemies; for so it is written, that by divine order they march out to destroy the land).
Drexl connected this interpretation to two passages in Revelation, 3 and not
14 See Drexl's testimonia in 238, 24-26. See also Brackertz, Traumbuch des Achmet, n. 552: and Oberhelman, The Oneirocriticon of Achmet, n. 975.
<sup>13</sup> Some have been dealt with in passing already: Brackertz, Traumbuch des Achmet, n. 36, correlates Drexl 2, 19-22 with Matthew 5:45 and Luke 6:35. However, for Arabic passages equivalent to Drexl 2, 19-22, see chapter 4, "Principles of Dream Interpretation." Brackertz (n. 42) follows Drexl and correlates the interpretation of the fruit of Paradise in Drexl 4, 26 ff. with John 14:17 and I John 4:6 ff; however, the interpretation of Paradise in Arabic dreambooks shows that this passage was copied from Arabic sources. Brackertz (n. 48) correlates the Greek interpretation of eunuchs with Matthew 19:12; but this passage was derived from Arabic dream interpretation as shown in chapter 7, "Angels and Eunuchs." Brackertz (n. 54) correlates the interpretation of Jesus with Luke 10:23, but see chapter 7, "Prophets, Apostles and Teachers, Jesus Christ and Icons." Brackertz (n. 407) comments on the differentiation between 2002 ("people" = God's people, the Byzantines) and ěθνος ("nation" = the foreign peoples) made in Drexl 163, 19-23; the same differentiation is made in Islam between the umma (the nation of Islam, the Muslims) and the kafirun (the unbelievers). Chapter 159 ("From the Persians and Egyptians on Fire", Drexl 121, 29-122, 10) gives the following interpretation: Ei δε ίδη, ότι έφερε πύρ ἐν ίματίω ἀκαύστως, οἰκειότητα εύρήσει καὶ μυστήρια δέξεται παρὰ τοῦ βασιλέως (If someone dreamt that he was carrying fire in his clothes without being burned, he will be a familiar of the king and will listen to his secrets). Brackertz (n. 321) correlates this with Proverbs 6:27: ἀποδήσει τις πυρ έν κόλπω, τα δε ίμάτια ου κατακαύσει; (Can a man take fire in his bosom, and his clothes not be burnt?). The biblical passage adduced by Brackertz does not concern the interpretation, but the dream symbol itself. Dreams of fire that does not burn are also discussed in Arabic dreambooks, e.g., Ibn Shāhīn, no. 5293: ومن راي ان احدا القاد في النار ولم يحـرقـه... : 5293 someone dreams that somebody <else> threw him into fire and he was not burned ...). Cf. also ومن راي ان بيده نارا مـشـتـعلـة فـانـه بــصــل لــه خـــر و مـنـفـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ السلطان (If someone dreams that he held in his hand fire that was ablaze, he will be granted goodness and beneficence from the ruler).
<sup>15</sup> Revelation 9:3: και έκ του καπνού ἐξῆλθον ἀκρίδες εἰς τὴν γῆν, καὶ ἐδόθη αὐταῖς έξουσία ώς έχουσιν οι σκορπίοι της γης (And there came out of the smoke locusts upon the earth: and unto them was given power, as the scorpions of the earth have power). Also Revelation 9:7: και τα δμοιώματα των άκρίδων όμοια ίπποις ήτοιμασμένοις είς πόλεμον (And the shapes of the locusts were like unto horses prepared unto battle).
to the plague of locusts in Egypt related in the Old Testament (Exodus 10: 1- 20).16 However, Arabic dreambooks also interpret locusts as hostile armies. <sup>17</sup> IbnQutaybastates, <sup>18</sup>Jl~~I ~I ~J ~~I Jl~I ~~~J (And as their saying about locusts that they are an army and about an army that it is locusts). <sup>19</sup>Al-Dinawari repeats this interpretation and explains it as follows: <sup>20</sup>r)L...JI ~ ~J--" ...::..il:l .J-.c *4 .. :d* v1Lw <l.JI ~J '-:"'ll..c. Jl~I (Locusts are a punishment and an army <sent> by God Almighty, for they were one of the signs of Moses, may peace be upon him). 21 A similar interpretation is quoted by lbn Shahin (nos. 6299 and 6300): <t...J[-.Q Jl~I ~I <sup>I</sup>\_;.Lb ..:,LS: I j *<sup>l</sup>*~ '-:> J J"'= (As for locusts, they are interpreted as an army, if they were flying) v1W..::. <l.JI 0-..c ~ Jl~I ~~~I JLl (al-Kirmani said: "Locusts are an army <sent> by God Almighty").
As al-Dinawari mentions, the Arabic interpretation of this dream refers to an incident described in the Qur'an; 22 a plague of locusts was sent by God to afflict the land of the pharaoh in order to persuade him that Moses is indeed a prophet. The same story is told in the Old Testament (Exodus 10:1-20) and is also implied in the *Oneirocriticon,* whose author saw no difficulty in incorporating this interpretation into his work, since the story of Moses is also part of Christian sacred history.
Three further interpretations that modern scholars considered as having been influenced by the Bible all refer to the Old Testament story of Joseph. Pressing
18 Ibn Qutayba, introduction, *Yahuda ar.* 196, fol. Sa; Ankara *Is. Saib Sine-er* I, 4501, fol. 192b.
<sup>19</sup>Repeated almost verbatim in *al-Muntakhab,* p. 11: 41 ~.J ~GI Jl~I ~ ~\_,\_..J
JI~ GI (And their saying about locusts that they are an army and about an army that it is locusts).
20AJ-Dinawari,fa~/ 21, *biib* 39, *Esad Efendi* 1833, fol. 214a.
<sup>21</sup>Repeated in al-Nabulusi, vol. I, p. 126, s.v. JI ..P.- .
<sup>16</sup>Brackertz, *Traumbuch des Achmet,* n. 552, adds the reference to Exodus 10:1-20 and Joel 1:2~2: 11.
<sup>17</sup> A different interpretation is given by Artemidoros, ii.21(Pack139, 16-19):' AKpioEs oE: Kat mipvoitEs Kat oi A.qoµEvot µcicn:aKEs yEUipyots µE:v mpopiav 11 <j>0opav tciiv Kapnciiv itpoayopEUoU<Jt. cri vovmt yap ta CT7tEpµma Tl ota<j>0Eipoucrt. tots OE AOlltOts ltOVT]poi>s avopas Tl yuvatKas CTT]µaivoucrt (Grasshopers, locusts, and the so-called jaw-locusts indicate barren soil or the devastation of their crops for farmers. For these insects damage seeds or destroy them completely. But for other men, they signify wicked men and women).
<sup>22</sup> Qur'an 7:133: "So We sent against them the flood and the locusts and the vermin and the frogs and the blood-a succession of clear signs. But they were arrogant and became a guilty folk."
grapes is interpreted as serving a person of authority, 3 and this reminds one of the interpretation that Joseph gave to the dream of pressing grapes into the pharaoh's cup that the pharaoh's butler had in prison.24 However, the Greek interpretation of pressing grapes was copied from the Arabic sources, as is evident from the passages regarding the interpretation of wine.25
Another two instances that bring to mind incidents from the story of Joseph are the interpretations of cows and chariots.
#### Cows
The Oneirocriticon interprets cows as representing years. Fat cows symbolize years of abundance; lean cows years of want. This interpretation evidently reflects the Old Testament dream of the pharaoh and its interpretation by Joseph (Genesis 41:1-32), but the Oneirocriticon makes no direct reference to the Bible:26
Τα βοΐδια ... τα θήλεια είς χρόνους <κρινέτω>. και εί μέν είσι πίονα, χρόνους εὐφόρους δηλούσιν, τα δὲ λεπτὰ καὶ πτωχά, χρόνους δυσφόρους …. ἐαν ἴδη τις, ότι εύρε βουν λιπαράν ή πολλάς, εύρήσει χρόνους καλούς και ευφόρους και προκοπὴν έκαστος της ἐργασίας αὐτου. ἐαν ϊδη τις, ὅτι εὐρε βοῦν ἢ ἀγέλην λεπτὴν καὶ ἐδέσποσεν αὐτῶν, εὑρήσει χρόνους ἀναλόγως του μέτρου αυτών κακούς και λιμόν και ζημίαν και πενίαν ἀνάλογον τῆς λεπτότητος.
Εαν ίδη ό βασιλεύς, ότι ήνεγκαν αύτφ βους λιπαράς, εύρήσει χρόνους άγαθούς και εὐτυχεῖς ἐν πολέμω καὶ πλοῦτον τῆς γῆς αὐτοῦ καὶ εὐφορίαν ἀναλόγως τῆς τῶν βοῶν πιάνσεως καὶ τοῦ πλήθους: τοῦ γὰρ βασιλέως τὸ ὅναρ ἐν τούτοις εἰς πάντας τοὺς ὑπὸ τὴν χεῖραν αὐτοῦ κρίνεται. όμοίως ἐαν ϊδη ό βασιλεύς, ότι ηνεγκαν αύτώ λεπτάς βούς και πτωχάς, εύρήσει χρόνους δυστυχείς των τε πολεμίων και της γης αυτού άναλόγως του μέτρου και της λεπτότητος έν τε έαυτώ και είς πάντας τους ύπ' αυτώ.
Let heifers ... be interpreted as years. If they are fat, they indicate fruitful years, if lean and meager, years that are hard to bear ... If someone dreams that he found one or many fat heifers, he will have either one or several years of good fortune and abundance, as well as progress in his work. If someone dreams that he found either one or a herd of lean heifers and owned them, he will have one or several
<sup>23</sup> Chap. 195, Drexl 151, 13-14.
<sup>24</sup> Genesis 40:1-13; Qur'an 12:36-41.
<sup>25</sup> See chapter 7, "Muslim Taboos."
<sup>26</sup> Chap. 237, Drexl 189, 25-190, 17. Cf. also Drexl 207, 12-13: είς ένιαυτον γάρ ή κρίσις tov Boóς (oxen are interpreted as years).
bad years, depending on their number, as well as famine, ruin and poverty by analogy with their leanness.
If the king dreams that they brought him fat cows, he will have good years with success in war, wealth from his land, and abundance in proportion to the fatness and number of cows. For the dream of the king, as far as such things are concerned, is interpreted with respect to everyone in his power.27 Likewise, if the king dreams that they brought him lean and meager cows, he will find years of misfortune regarding both his enemies and his land in proportion to the number and leanness <of the cows>, regarding both himself and everyone in his power.
The pharaoh's dream of seven fat and seven lean cows and its interpretation by Joseph are also recounted in Qur'an 12:43-49. Arabic dreambooks interpret cows in the same way that Joseph interpreted them in the biblical and the Quranic narrative. lbn Qutayba says,28 4-a~..9u\_,.:....,...~1\_, L..... *i.µ* 1\_, ~ *µI* ..:,...........\_, ... J 1,,..... I (One cow signifies one year, while many cows signify many years. Their flesh is interpreted as money .... The fatness of the cow is abundance). Al-Dinawari's interpretations of cows are also in the same spirit:<sup>29</sup>
.\_.JLL. 4.J\_,.:a..J ~\_, .;J\_,....... ~ 4...:........ .l\_,\_L...:JI b\_µI\_, .1..i~I b\_µI ~'1 41-=j[+.. ~.\_.JI ~I 4.\$.L *.Jl 4-=>L.............* 0-:~ .J..:>I JL..:i ~..rl=L.:JI ~ t7-'-" *l.S)* .).1 .\_.JLL. <L.J \_,\_:aJ '-:"~I \_,..:........ *J..=* j4J.1 \_, \* ~I \_,..:........ .:.,L.............. <sup>I</sup> ,,,.\_j4-e ~ ~L:> ..::..olfa
A black and a yellow cow <signifies> a year during which there is joy and abundance, according to the saying of God Almighty, "She shall be a golden cow, bright her color, gladdening the beholders" [Qur'an 2:69]. Ibn Sirin said: "The fat among them, for whoever owned them is preferable to the lean ones, because fat cows <indicate> a more abundant year. Lean cows <signify> a more barren year, according to the saying of God Almighty: "Lo! I saw in a dream seven fat kine which seven lean were eating" [Qur'an 12:43].
*Al-Muntakhab* repeats al-Dinawart:3°
• 11 .\_..:..I I .<I - • I A ·II • L......... J.;;. . . I . ~ ~ i. A ·II *i..r.* . "ii""""'-"..,,....~ ..\_, *..r-;:* <.>-:~ l>-7 ..\_, ..9 ~ L.:.:i.l '-:-'~ *u\_,.:....,...* J-:.:,411\_, ~ *u\_,.:....,... uL.........11 u'i* J-:.:,411 .J-o ... r)l....J I~ *...J.......-"'=*
<sup>27</sup> The expression *imo* 't'fiv XEtpav auwu corresponds to a word-for-word translation of the Arabic expression "-=""=~=under his hands= under his power.
<sup>28</sup> Ibn Qutayba, *biib* 35, *Yahuda ar.* 196, fol. 52b (cf. also Ibn Shahin, no. 5689).
<sup>29</sup>Al-Dinawari,fa~/21, *biib* 21, *Esad Efendi* 1833, fol. 2lla.
<sup>30</sup> *Al-Muntakhab,* p. 181; repeated in lbn Shahin, no. 5702.
A cow is a year. Ibn Sirin used to say: "Fat cows, for whoever owns them, are, according to me, preferable to the lean ones, because the fat cows represent years of abundance, while the lean cows years of barrenness, according to the story of Joseph, may peace be upon him."
Ibn Shahin's interpretation of cows (no. 5692) is also similar:
وقال ابن سيرين من راى بقرة وهو ملكها وكانت سمينة فإنها تدل على النعمة الكثيرة فى تلك السنة، وإن كانت مجهولة فإنها تدل على حصول النعمة لأهل ذلك المكان فى تلك السنة، وإن كانت مهزولة فتأويلها بضد ذلك.
Ibn Sirin said: "If someone dreams that he owns a cow, if she is fat <the dream> signifies plentiful blessings throughout that year. If she is unknown <as to who her owner is>, she signifies the attainment of blessings by the people of that place throughout that year. If she is lean, her interpretation is the opposite of that.
Similar interpretations of cows are also quoted in al-Nābulusī.31 Evidently, the interpretation of cows in the Oneirocriticon was copied from its Arabic sources. The Christian author did not need to make any changes in the Arabic interpretations in order to conform with Joseph's story in Genesis; the details of the equivalent Quranic narrative, the immediate source of the Arabic interpretations, are no different from the details found in the Old Testament.
#### Chariots
The Oneirocriticon makes its third and only direct reference to the story of Joseph in its interpretation of chariots:32
όσοις δὲ τὸ βασίλειον ὅχημα θεωρεῖται κατ᾽ ὅναρ, εἰς ἐγγύτητα βασιλείας ἐκρίθη, διότι καὶ τῶ Ίωσήφ, ὅτε ήλευθερώθη καὶ τῆς Αἰγύπτου ῆρξεν, καὶ τοις λοιποίς έκ του όχήματος έδηλώθη το της βασιλείας δόξης ύψος παραδεδομένον έκ των αρχαίων.
The royal chariot, for whoever dreams of it, is interpreted as closeness to kingship, because both for Joseph, after he was freed and ruled over Egypt, and for other people the loftiness of royal glory was made apparent by the royal chariot, according to a tradition handed down from the ancients.
<sup>31</sup> Al-Nābulusī, vol. 1, pp. 60-62, s.v. بقر .
<sup>32</sup> Chap. 238, "From the Egyptians on Carriages, Chariots and Wagons"; Drexl 191, 22-26.
This passage follows Genesis 41:43, where Joseph's accession to power is described: και άνεβίβασεν αύτον έπι το αρμα το δεύτερον των αύτου, καὶ ἐκήρυξεν ἔμπροσθεν αὐτοῦ κῆρυξ· καὶ κατέστησεν αὐτὸν ἐφ' ὅλης ync Alyontov (And he made him ride in the second chariot which he had; and they cried before him, "Bend the knee": and he made him ruler over all the land of Egypt).33 The Quranic narrative, which is far less detailed than the story in Genesis, does not mention a chariot. However, Muslims considered the Torah, which includes the first five books of the Old Testament from Genesis to Deuteronomy, as a divinely inspired text and used it to supplement their version of sacred history. The role played by the chariot in the story of Joseph was therefore known to al-Dinawari, who refers to it to justify his interpretation of chariots as sovereignty and power:34 امــا العــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ المسلمون انه كان من مراكب اللوك اللوك الاوائل وحمل يوسف عليه السلام حين اکےرم ونال ملکا عظيہ (As for the chariot, the Muslims said, "This is one of the vehicles used by the ancient kings, and it carried Joseph, may peace be upon him, when honors were bestowed upon him and he attained mighty rulership").
The other Arabic dreambooks interpret chariots in the same way as al-Dinawari and the Oneirocriticon, without, however, directly referring to من رای انه رکب عجلة اصاب سلطانا Joseph's chariot. Ibn Qutayba says:35 عـمــا و نــال شـر فـا و نــال شـــ فــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ will obtain a foreign dominion and will have glory and honor). Ibn Shahin also gives similar interpretations (nos. 7353, 7355, 7356):
(V٢٥٢) والعجلة تؤول على أوجه، ومن رأى أنه ركب عجلة وهى تسيير به فازنه يدل على حصول الشرف والرفعة. (۷۳۰۰) ومن راي أن ملكا أعطى له عجلة فإنه يصيب سلطانا بقدر كبرها. (٧٣٥٦) ومن راي يتبع عجلة فإنه يتبع صاحب السلطان.
(7353) A chariot is interpreted in several ways. Whoever dreams that he rode on a chariot and that it was led by him, it means that he will gain honor and a lofty
<sup>33</sup> Cf. also the Syriac dreambook from BL Or. 4434, the text of which was translated by Furlani, "Une clef des songes en syriaque," p. 139. The relevant passage in Furlani's translation reads as follows: "VIII. Chapitre sur les pouvoirs, les serviteurs et les chars: Celui qui (se voit) monté sur une charette trainée par des taureaux ou des chevaux, il se peut que royauté et puissance lui soient dévolues; le roi Pharaon montait sur un chariot." Cf. also Exodus 15:4: αρματα Φαραώ και τὴν δύναμιν αὐτοῦ ἔρριψεν είς θάλασσαν (Pharaoh's chariots and his hosts hath he cast into the sea).
<sup>34</sup> Al-Dīnawarī, fasl 13, hāh 22, Esad Efendi 1833, fol. 71a.
<sup>35</sup> Ibn Qutayba, hah 46, Yahuda ar. 196, fol. 67a.
position. (7355) Whoever dreams that a king offered him a chariot, he will gain power analogous to the size of the chariot. (7356) If someone dreams that he was following a chariot, he will follow a powerful person.
#### Al-Nabulusi also interprets chariots as sovereignty and power: 36
(عجلة) تدل فى منام على تدبير عيش صاحب الر ؤيا لانها مركبة من اشياء كثيرة وتحمل أشياء كثيرة وتنقلها من مكان الى مكان ومن راى انه راكب عجلة و تحت يد العجلة رجال فانه يدل على أن صاحب الرؤيا يسوس قوما كثيرين أو على انه يولد له اولاد خيار ... والعجلة عز من سلطان أعجمى لمن ركبها أو ادراك شرف وكرامة ومن راى انه متعلق بعجلة تحمله أو يتبعها فانه يتبع سلطانا ويستمكن منه بقدر استمكانه من العجلة.
The chariot in a dream is interpreted as the organization of the life of the dreamer, because a chariot is the vehicle for several things: it carries several things and transports them from one place to another. If someone dreams that he rode a chariot and that the chariot was drawn by men, it means that the dreamer will rule a numerous people or that he will give birth to good children ... A chariot signifies power from a foreign sovereign or the attainment of glory and honor for whoever rides on it. When someone dreams that he was clinging to a carriage that was carrying him, or that he was following it, it means that he will follow a ruler or he will win control over him commensurate to the control he had over the carriage.
It is evident that the author of the Oneirocriticon did not have to invent or adapt interpretations in order for the contents of his dreambook to conform to the interpretations given in Genesis. His Arabic sources already contained all the requisite details; he only had to copy them.
#### Eyes
The opening statement in the Oneirocriticon's discussion of eyes says:31 oi όφθαλμοί πίστις και δόξα και φώς είσι της γυχής (Eyes are the faith, glory and light of the soul39). Brackertz99 connected this interpretation with a
<sup>36</sup> Al-Nābulusī, vol. 2, p. 96, s.v. عحلة .
<sup>37</sup> Chap. 52, "From the Indians on Eyes"; Drexl 33, 4.
<sup>38</sup> The phrase could also be translated as "The eyes are the faith, glory and vision of the soul." The word of of was used in Greek in reference to the eyes as early as Homer; cf. Liddell-Scott, s.v. "φάος."
<sup>39</sup> Brackertz, Traumbuch des Achmet, p. 245, n. 130.
phrase from the New Testament: 6 λύχνος του σώματός έστιν ό όφθαλμός (the <lamp> light4 of the body is the eye). Three of the five Arabic dreambooks العين دين الرجل وبصــيـرتـه have interpretations that are very close to this التى يبـصــر بـه الـهـدى و الـهـدى و الـضـلالـهـدى و الضــلالـة vision by which he can distinguish the rightly guided path from the one leading him astray ).
#### Arms
Among the Oneirocriticon's interpretations of arms we read the following: 41
έαν ουν ίδη τις κατ΄ όναρ ότι ταυτα <οί βραχίονες και οί μύες αυτών> ἐμεγεθύνθησαν ἢ ἐνεδυναμώθησαν, εἰ μέν ἐστι βασιλεύς, ταῦτα εἰς τὴν δύναμιν αυτού διακρίνονται και μεγαλυνθήσεται ή δύναμις αύτού κατά των έχθρων αύτού.
If someone dreams that his arms and their muscles became larger or stronger, if he is king, they are interpreted as his power, and his power against his enemies will become greater.
Brackertz44 correlated this interpretation with a biblical passage that occurs both in the Old and the New Testament:45 κύριε, τις ἐπίστευσεν τῆ ἀκοῦ ήμων; και ό βραχίων κυρίου τίνι απεκαλύφθη; (Lord, who hath believed our report? and to whom hath the arm of the Lord been revealed?). The arm as a metaphor for strength in the Greek Bible originates in the literal translation of a Semitic expression which is found not only in Hebrew, the language from which the Old Testament was translated into Greek, but also in Arabic: taht
<sup>40</sup> Matthew 6:22 and Luke 11:34.
<sup>41</sup> λύχνος = lamp, light. See Lampe, s.v. "λύχνος." This is a different word than the Greek φως (light) employed in the Oneirocriticon.
<sup>42</sup> Al-Dinawari, fasl 6, bab 51, Esad Efendi 1833, fol. 52a; repeated in al-Muntakhab, chap. 22, p. 81; repeated in Ibn Shāhīn, no. 1388. In addition, Arabic dream interpretation generally correlates the eyes with religious faith, as is done in the Oneirocriticon. Cf. Ibn Qutayba, bab 14, Y ahuda ar. 196, fol. 30a: واشقار العين وقاية الدين (Redness in the eyes <signifies> precaution in one's religious faith); Ibn Shāhīn, no. 1386 bis: واما العدنان فسؤولان سالدين و غيره (As for the two eyes, they are interpreted as one's faith and the like).
<sup>43</sup> Chap. 70, "From the Indians on Arms and Their Muscles"; Drexl 44, 9-12.
<sup>44</sup> Brackertz, Traumbuch des Achmet, p. 248, n. 154.
<sup>45</sup> John 12:38 and Isaiah 53:1.
365
yadih= under his hand/arm46 = under his power. The Arabic interpretation of the lengthening and strengthening of one's hands/arms is based not only on the literal and metaphorical meaning of this expression, but also on the fact that the Arabic words for "length" (tul) and "might" or "power" (tawl) are spelled identically as طول . Ibn Qutayba interprets the lengthening of hands/arms as follows:43 و من راى فى يده طولا كان ذلك طولا على الناس (If someone dreams of length (tul) in his hand/arm, this is power (tawl) over men "). In addition, al-Dīnawarī states: فان رای ان یده طالت او قـویت ان کـان و الیـا فهو ظفره باعدائه وقوة اعوائه وقوة اعوانه وقواده وقواده وقواده became longer or stronger, if he is a governor (wālī), this represents his triumph over his enemies, and the power of his assistants and his chief bodyguard). Ibn Shāhīn's interpretation is also similar (no. 1498 bis): وطول الليــــديـن زيــادة مقدر ة (The lengthening of the two hands/arms means an increase in power).
#### Blood
The following passage from the Oneirocriticon (Drex1 60, 10-12) appears to have been inspired by the Bible: έαν δε έπιεν ή ἔφαγεν αΐμα, ενρήσει ἐφαμάρτως ἀλλότριον χρυσίον. οὐ γὰρ θεόθεν ώρισται ή μετάληψις τοῦ aluatos (If he drank or ate blood, he will sinfully acquire gold that belongs to somebody else. For the partaking of blood is not ordained by God). This interpretation has been connected to a number of passages in the Old Testament.50
<sup>46</sup> The Arabic yad is the exact equivalent of the Greek cheir: it means primarily "hand", but also signifies the whole arm, from the shoulder joint to the tips of the fingers.
<sup>47</sup> Ibn Qutayba, hāh 14, Yahuda ar. 196, fol. 30b.
<sup>48</sup> Ibn Qutayba further justifies this interpretation with an Arab proverb, which is repeated in al-Muntakhah, p. 10.
<sup>40</sup> Al-Dinawari, fasl 6, bah 82, Esad Efendi 1833, fol. 56a; repeated in al-Muntakhah, chap. 22, p. 89; al-Nābulusī, vol. 2, p. 340, s.v. د.
<sup>50</sup> Cf. Brackertz, Traumbuch des Achmet, p. 253, n. 193; also Oberhelman, Oneirocriticon of Achmet, n. 213; both cite Leviticus 17:10-14: και ανθρωπος τών νίων Ισραήλ ή των προσηλύτων των προσκειμένων έν ύμιν, δς αν φάγη παν αΐμα, καὶ ἐπιστήσω τὸ πρόσωπόν μου επι τὴν ψυχὴν τὴν ἔσθουσαν τὸ αἴμα καὶ ἀπολῶ αὐτὴν ἐκ τοῦ λαοῦ αὐτῆς ... διὰ τοῦτο εϊρηκα τοις νίοις Ισραήλ Πασα ψυχή έξ όμων ου φάγεται αίμα ... (And whatsoever man they be of the house of Israel, or of the strangers that sojourn among you, that eateth any manner of blood; I will even set my face against that soul that eateth blood, and will cut him off from among his people ... Therefore I said unto the children of Israel, No soul of you shall eat blood ... ); also Genesis 9:4; Deuteronomy 12:16.
In a more strictly Christian context, it appears to derive from the New Testament prohibition against eating blood and "things strangled."" However, the Qur'an also forbids the eating of blood, 32 and Arabic dreambooks interpret blood as unlawful money. 33 In addition, Ibn Shāhīn (no. 1777), quoting al-Kirmānī, interprets imbibing blood in the same way as the Oneirocriticon: قــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ الككرمانى ومن راى انه يشرب دمـا فـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ . بــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ means he will acquire unlawful money or <be guilty of> unjust bloodshed").
#### White Clothes
According to the Oneirocriticon, dreaming of wearing white clothes is very auspicious:4 όσα ίδη ανθρωπος ότι φορεί λευκα ίμάτια, εύρήσει χαράν και δόξαν είς τὴν πίστιν καὶ ἐν τῷ κόσμῳ καὶ κόσμιος ἔσται ἐν τῇ θέα ourov (If a man dreams that he is wearing white clothes, he will find joy and glory in his faith and in his worldly affairs, and his appearance will be decorous).
Brackertz correlates this interpretation with numerous passages from Revelation that refer to the white robes of the Christian martyrs,56 but Arabic
<sup>51</sup> Acts 15:19: και απέχεσθαι είδωλοθύτων και αϊματος και πνικτών και πορνείας(That ye abstain from meats offered to idols, and from blood, and from things strangled, and from fornication).
<sup>32</sup> Qur'an 2:173: "He hath forbidden you only carrion, and blood, and swineflesh, and that which hath been immolated to (the name of) any other than Allah. But he who is driven by necessity, neither craving nor transgressing, it is no sin for him"; 5:3: "Forbidden to you are carrion, blood"; 6:145: "Except if it be carrion or blood outpoured"; 16:115: "These things only has He forbidden you: carrion, blood ... "
<sup>53</sup> Cf. al-Muntakhab p. 104; cf. also al-Nabulusi, vol. 1, p. 216, s.v. دم.
<sup>54</sup> Chap. 156, "From the Indians on Clothes" ; Drexl 116, 15-7.
<sup>55</sup> Brackertz, Traumbuch des Achmet, p. 265, n. 306.
<sup>56</sup> Revelation 6:11: και έδόθη αύτοις έκάστω στολή λευκή και έρρέθη αυτοίς ϊνα άναπαύσονται έτι χρόνον μικρόν, έως πληρωθώσιν και οί σύνδουλοι αύτων και οί άδελφοι αύτων οί μέλλοντες άποκτέννεσθαι ώς και αύτοί (And white robes were given unto every one of them; and it was said unto them, that they should rest yet for a little season, until their fellowservants also and their brethren, that should be killed as they were, should be fulfilled); Revelation 7:9: Μετά ταύτα είδον, και ίδου οχλος όν αριθμήσαι αντον ούδείς ἐδύνατο, ἐκ παντὸς ἔθνους καὶ φυλῶν καὶ λαῶν καὶ γλωσσῶν ἐστῶτες ἐνώπιον τοῦ θρόνου και ένώπιον του αρνίου περιβεβλημένους στολάς λευκάς και φοίνικες έν ταις χερσίν αυτών (After this I beheld, and, lo, a great multitude, which no man could number, of all nations, and kindreds, and people, and tongues, stood before the throne, and before the Lamb, clothed with white robes, and palms in their hands); Revelation 7:13-14: και άπεκρίθη είς εκ τών πρεσβυτέρων
dreambooks interpret white clothes in the same way. Al-Dinawari states:57 ~..1 J--01 ~ e:JL....:. ~I~ i...il:UI 01 ~J .\_,.;:.L+:JI \_:,~I JLl Jl...a....:;-J i.~JJ (The Muslims said: "White clothes, for someone who wears them habitually when he is awake, signify goodness both in his religious faith and in his worldly affairs, as well as beauty"). *Al-Muntakhab* quotes a similar interpretation:58 0-:'J.l IJ ~JJ I~ J4 '-:-'4-=J I j-o .\_,.;:.~J (Whiteness of clothes signifies beauty in one's worldly affairs and religious faith). lbn Shahin (no. 4558) justifies this interpretation with a Quranic passage: ~L..~ I JuJ ~J ~..1 (~ ~ J~ ~U ~ ~~ ~4'.' ~ o....:.I 1.51\_; j-o ~ ~~J u-'LL::. <lJ I cLJ\_,\_iJ <L..oJ--"-'11 '-:-'LJ:ijJ .UG:.. (Al-Kirmani said: "If someone dreams that he was wearing clean white clothes, this signifies righteousness in his religious faith, well-being and relief from his troubles, according to the saying of God Almighty: "Thy robes purify" [Qur'an 74:4]). Al-Nabulusi's interpretation is vague, but in the same spirit as the interpretations quoted above: 59 iL.:..11 ~ ~ 01 ~ ~I '-:-'~I J (White clothes mean good fortune for whoever wears them in a dream).
#### *Earthquakes*
The *Oneirocriticon* quotes the following interpretation of earthquakes:60
£riv ni; '£811 Km' ovap EV 't07t(\), £v8a unapxn, crncrµ6v, 'tOU'tO VEoV 8tri1ayµa napel (3acrtAtwi; 011µaivi:::t 8poouv Ka\. cri:::tov 1:0ui; KmotKouv1ai;· Kat £av E:cntv 6 CT£tcrµoi; nanaxou, Ka8oAtKOV ECT'tat Kat 'tO 8tri1ayµa, £\. 8£ µi:::ptKoi; 6 crttcrµ6i;, µi:::ptKOV Kat 'tO 8ta1ayµa.
Myrov µot· ou-rnt oi 1t£ptj3£13A.riµ£vot 'tel<; cnoA.ai; 1:ai; AEUKai; 1:iv£i; Eicrtv Kat n60£v ~A.eov; Kat dpriKa m'.mji· KUplE µou, cru olOai;. Kat ct1tEV µot• OU'tOl ElCJtV Ot £pxoµEV0l EK 1:f\<; 0A.iljlEOO<; 1:f\i; µqaA.rii; Kat i:'nA.uvav 'tel<; CJ'tOAel<; aU1:WV Kat EAEUKavav au1:ai; EV nii a'iµan -mu apviou (And one of the elders answered, saying unto me, What are these which are arrayed in white robes? and whence came they? And I said unto him, Sir, thou knowest. And he said to me, These are they which came out of great tribulation, and have washed their robes, and made them white in the blood of the Lamb).
<sup>57</sup> Al-Dinawari, *fa#* 25, *bah* 32, *Esad Efendi* 1833, fol. 251 b.
<sup>58</sup>*A/-Muntakhab,* chap. 19, p. 127.
<sup>59</sup> Al-Nii.bulusi, vol. 1, p. 95, s.v. *'-:';:..*
<sup>60</sup> Chap. 145, "From the Indians, Persians and Egyptians on Earthquakes" ; Drexl 99, 16-18.
#### CHAPTER EIGHT
If someone dreams that in the place where he is there was an earthquake, this means a new edict from a king will shake and move the inhabitants. And if the earthquake occurred everywhere, the edict will also be universal. If the earthquake is partial, so too the edict.
Brackertz commented that the Christian interpretation of earthquakes as representing the wrath of kings is analogous to the imagery employed in the prophetic books of the Old Testament, where all kinds of punishment and misfortune are equated with an earthquake.61 Here again, however, the Arabic dreambooks turn up the same interpretation. According to Ibn Qutayba, 60 والزلزلزلة حدث في الناس من قبل الملك الاعظم الاعظم الاعظم <brought about> among the people by the most powerful king). Al-Dīnawarī 's interpretation is similar: قـــال المسلمـــون من راى ارضـــا زلـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ السلطان ينزل تلك الارض ويعذب اهلها (The Muslims said: "If someone dreams that a land was shaken by an earthquake, ... the ruler will come to that land and punish its people"). Al-Nabulusi's interpretation of earthquakes distinguishes between an earthquake that shakes a particular locale and one that is universal, just as the Oneirocriticon does:64
زلزلة: هى فى المنام خوف من سلطان وقيل الزلزلة فى المكان الخصوص تدل على نقلة وتحويل وقيل الزلزلة حادث يحدث من قبل آلملك الاعظم فان كانت عامة فالصادث عام والا فالموضم والبلدة التى خصصت بالزلزلة.
Earthquake: This in a dream is fear of authority. And it is said that the earthquake in a specific place signifies migration and dislocation. And it is said that an earthquake is an event <brought about> by the most powerful king. And if the earthquake is general, the event is also general; if not, <it will occur> only in the place or the land that was specified by the earthquake.
#### Wheat
The Oneirocriticon quotes the following interpretation of a land sown with wheat.65
<sup>61</sup> Brackertz, Traumbuch des Achmet, n. 330.
<sup>62</sup> Ibn Qutayba, hāh 17, Yahuda ar. 196, fol. 35a.
<sup>64</sup> Al-Dīnawarī, fașl 15, bāb 31, Esad Efendi 1833, fol. 158a; repeated verbatim in al-Muntakhab, chap. 38, p. 233; Ibn Shahin's interpretation on the sinking of the earth (no. 3084) copies the interpretation of earthquakes quoted by al-Dinawari (earthquakes and sinking of the earth are interpreted in the same chapter in al-Dinawari).
<sup>64</sup> Al-Nābulusī, vol. 1, p. 270, s.v. ز لز اللة.
<sup>65</sup> Chap. 208, "From the Indians on Fields and Seeds" ; Drexl 163, 3-7.
ἐάν τις ϊδη κατ΄ ὅναρ λογιζόμενος, ὅτι ἔσπειρε χώραν γνωρίμην σῖτον καὶ ὅτι ἐφύτρωσε καὶ ἦν καλὴ καὶ ἦν ὁ καιρὸς φυτρώσεως, εἰ μέν ἐστι ὁ ἰδών τῶν πιστών και εύλαβών ή και άναχωρητών, τουτο είς τας πράξεις αύτου διακρίνεται τας κατά θεόν.
If someone dreams that by his reckoning he sowed a familiar field with grain and that the grain sprouted and grew well and that it was the growing season, if he is pious and devout or even an anchorite, this represents his deeds before God.
This interpretation might appear to a Christian reader as inspired by the parable of the sower that is narrated in three of the four Gospels (Luke 8:4-48; Matthew 13:1-9; Mark 4:1-9).66 However, an examination of Arabic dream interpretation demonstrates that it was again copied from the Arabic sources of the Oneirocriticon. The relevant interpretation from Ibn Qutayba reads as follows:60
الارض المكلئة خص ومال وخير للعامة ، والزر ع اعمال بنى أدم اذا كان كقدر الزر ع فى طوله يقال فى المثل من يزر ع خيرا يحصد غبطة ومن يزر ع شرا يحصد ندامة
A land abounding with herbage is abundance, money, and goodness for everybody. Sowing represents the deeds of men; if it is a good <deed>, its places <i.e. where the good deed was done> will resemble the places of sowing <i.e., like the grain it will grow and bear fruit> and it will be like the size of the grain in its length. It is said in a proverb: He who sows goodness will harvest bliss and he who sows evil will harvest regrets.
ومن راى انه زر ع حنطة A similar interpretation is quoted by al-Dīnawarī عمل عمل عمل عملاته الله تعالى (Whoever dreams that he sowed wheat, he will do a deed that will please God). Yet another example comes from al-Muntakhah:69
ومن بذر بذرا فى وقته فانه قد عمل خيرا فان كان واليا اصاب سلطانا وإن كان تاجراً نال ربحا وإن كان سوقيا اصاب بلغة وان كان زاهدا نال ورعا.
If someone sows seeds in the appropriate season, he will do a good deed. And if he is a ruler he will gain power; if he is a merchant he will make a profit; if he is a a commoner, he will obtain sufficiency; and if he is an ascetic he will gain in piety .
<sup>66</sup> See Brackertz, Traumbuch des Achmet, p. 275, n. 405.
<sup>67</sup> Ibn Qutayba, chap. 22, fol. 38b; repeated in al-Muntakhah, p. 317.
<sup>68</sup> Al-Dīnawarī, fasl 19, bāh 32, Esad Efendi 1833, fol. 188a.
<sup>69</sup> Al-Muntakhah, p. 317.
Ibn Shāhīn (no. 4234) states: وربما دلت رؤيـا الـزر ع على اعــمــال الــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ فان كانت مخضرة فان الاعمال صالحة، وإن كان غير ذلك فتعبيره ضده (Dreaming of sowing possibly signifies the deeds of people. If <the land sown> is green, the deeds are righteous, and if it is not, the interpretation of the dream is the opposite). Al-Nabulusi also connects the sowing of wheat with good deeds:"من رااى انـه زر ع حنطة عـمل عـمـلا فـيـه لـلـه تـعـالـى ر ضـا (If someone dreams that he sowed wheat, he will do a deed that will please God).
#### Thorns
According to the Oneirocriticon, ' ή ρακις των άκανθών είς σφάλματα της πίστεως διακρίνεται (Thorns are interpreted as mistakes in one's faith). Brackertz"2 suggested that this interpretation was also inspired by the parable of the sower.13 In the words of Luke 8:14, το δέ είς τας ακάνθας πεσόν, ούτοί είσιν οί ακούσαντες, και ύπο μεριμνών και πλούτου και ήδονών του βίου πορευόμενοι συμπήγονται και ού τελεσφορούσι (And that <seed> which fell among thorns are they, which, when they have heard, go forth, and are choked with cares and riches and pleasures of this life, and bring no fruit to perfection).
The Arabic interpretations of this dream again demonstrate that the Oneirocriticon's passage is from an Arabic source. According to an interpretation quoted by al-Dīnawarī and repeated elsewhere, و من ناله من الشعواك ضدر نال من الدىن مـا يكرهـه تقدر مـا نالـهـمـن الشـواك (If someone was harmed from thorns, he will commit something reprehensible in his religious life, by analogy to the harm he suffered from the thorns).
<sup>70</sup> Al-Nabulusi, vol. 1, p. 149, s.v. حنطة.
<sup>71</sup> Chapter 212, "Likewise on Thorns" ; Drexl 166, 23.
<sup>72</sup> Brackertz, Traumbuch des Achmet, n. 410.
<sup>73</sup> Matthew 13:22; Mark 4:18; Luke 8:14.
<sup>74</sup> Al-Dinawari, fas! 19, bãh 74, Esad Efendi 1833, fol. 194a; repeated in al-Muntakhab, p. 316. Al-Nabulusī, vol. 2, p. 25, s.v. شارك (thorn) repeats that thorns signify harm, but omits the religious dimension that is evident in al-Dinawari.
#### Pearls
The Oneirocriticon consistently interprets pearls and precious stones as religious knowledge and wisdom: οί μαργαρίται και οί λίθοι των τιμίων είς θείους λόγους και θεογνωσίαν και θεοσοφίαν ώς ἐπὶ τὸ πολὺ διακρίvoveat. (Pearls and precious stones are generally interpreted as divine words, knowledge of God and divine wisdom); } ἐαν ἴδη τις, ὅτι εὐρεν ἢ ἔλαβε παρά τινος μαργαρίτας ή λίθους των τιμίων, θείων έφεύρη λαμπρότητα δογμάτων αναλόγως του μεγέθους και της λαμπρότητος (If someone dreams that he found or received from someone pearls or precious stones, he will devise a brilliant doctrine commensurate with the size and brilliance <of the gems>).76
Brackertz suggested that these interpretations were inspired by the parable quoted in the New Testament, whereby the kingdom of heaven is likened to a precious pearl, 77 but the direct source is once more Arabic dreambooks frequently interpret pearls as wisdom and knowledge, especially of the Qur'an. According to Ibn Qutayba, " for example, فيان اللؤلئؤ المنظوم كلام الله عـز وجل او ڪلام من کلام البر (Strung pearls are the words of God Almighty on words of goodness). Further on, Ibn Qutayba quotes the words of Ibn Sirin, which are authenticated by a preceding isnād:79
قال محمد بن سيرين اذا رايت اللؤلؤ فهو قرأن واذا رايت العقد فهو حكمة ... فان كان اللؤلؤ النتثورا غير منظوم... ربما كان كلاما حسنا لقول الناس في وصف ما يستحسنون من الكلام كانه لوّلؤ منتّور ...
Muhammad b. Sirin said: "If you see pearls, this is the Qur'an, and if you see a necklace it is wisdom ... The scattered pearls that are not strung ... are possibly beautiful words, according to the popular saying the description of beautiful words that they are like scattered pearls.
The interpretation of pearls as religious knowledge and wisdom is repeated by
<sup>75</sup> Chap. 245, "From the Indians on Crowns, Pearls and Precious Stones" ; Drexl 200, 22.
<sup>76</sup> Ibid., Drexl 201, 9-12,
<sup>77</sup> Brackertz. Traumbuch des Achmet, n. 479; Matthew 13:45-46: Πάλιν όμοία έστιν ή βασιλεία των ούρανων άνθρώπω έμπόρφ ζητουντι καλούς μαργαρίτας. εύρων δε ένα πολύτιμον μαργαρίτην απελθών πέπρακε πάντα όσα είχεν και αγόρασεν αυτόν (Again, the kingdom of heaven is like unto a merchant man, seeking goodly pearls: Who, when he had found one pearl of great price, went and sold all that he had, and bought it).
<sup>78</sup> Ibn Qutayba, hab 27, Yahuda ar. 196, fol. 44b.
<sup>79</sup> Ibn Qutayba, bab 27, Yahuda ar. 196, fol. 44b; repeated in al-Muntakhab, p. 280.
al-Dinawari: <sup>80</sup>rLJ IJ -:,1\_;--.i..J I i~I j.J\_jl.J I (Strung pearls are the Qur'an and knowledge) and lbn Shahin (no. 4798): JJ.J..= J~I j.Jjll I ~J.) µJ ... ~IJ rLJL...;i JJ.J...:> ct...:.µ L.\_,.J;....:..... -:,U l.'.llJ .-:,i\_r-i.Jl...;i (It is said that scattered pearls are interpreted as the Qur'an, and if they were strung they are interpreted as knowledge and wisdom ... ). <sup>81</sup>
It is unlikely that the agreement between the Arabic interpretations and the interpretations of the *Oneirocriticon* that appear to have been inspired by the Bible is the result of the Greek and Arab authors having independently used the Old and New Testament as a source and arriving at the same interpretations. Not only the passages excerpted above, but a whole series of additional passages in the *Oneirocriticon* on the same topics, coincide with passages found in Arabic dreambooks. This indicates that the Greek author, despite his occasional creativity in adapting the Islamic material, resorted to adaptation only when absolutely necessary; generally he chose to remain close to his Arabic source, especially when the Islamic interpretations not only did not clash with Christian beliefs, but even seemed to be supported by the Scriptures. The elements common to Christianity and Islam, including eschatology, sacred history, and even linguistic expression, facilitated the absorption of Muslim interpretations into a Christian text. The imagery and language of the Greek Bible, which permeates the writings of several medieval Greek authors, including the author of the *Oneirocriticon,* is interspersed with semitisms, because the Septuagint is a translation from Hebrew and the Greek New Testament was written mostly by authors whose mother tongue was Aramaic. 82 Since Arabic is also a Semitic language and the imagery and expression of its medieval authors is influenced by the Qur'an, which presents analogies with the imagery and expressions of the Bible, it shares with biblical Greek a number of expressions and imagery. The equivalents can sometimes be so close that a word-by-word translation of certain Arabic expressions results in irregular classical but regular biblical Greek. 83 The linguistic coincidences must have facilitated the work of the
<sup>80</sup> Al-Dinawari,fa# 20, *bah* 33, *Esad Efendi* 1833, fol. 206a. Al-Dinawari's interpretations of pearls are repeated in *al-Muntakhab,* pp. 279-80.
<sup>81</sup>Similar interpretations also quoted in Ibn Shahin, nos. 4804, 4808, 4809, 4814 bis; al-Nabulusi, vol. 2, p. 197, s.v. )J)J.
<sup>82</sup> For the introduction of Semiticisms into the Greek *koine,* see A. Debrunner, 0. Hoffmann and A. Scherer, *Geschichte der griechischen Sprache,* vol. 2 (Berlin, 1969), § § 147-51.
<sup>83</sup> E.g . .\_;,\_, y I ~J = itp6crwnov 1fi~ yfi~ ="surface of the earth."
author of the *Oneirocriticon,* who seems to have been familiar with the biblical language. Among the several Semitisms that can be identified in the text of the *Oneirocriticon,* one can distinguish Arabisms from biblical expression only when a given linguistic peculiarity does not occur in biblical Greek but can be translated back into correct Arabic. The closeness between Christian and Muslim eschatological beliefs, religious imagery, and linguistic expression is also the reason why the Greek interpretations of religious notions such as angels, Paradise and Hell, and the interpretations that seem to have been inspired by the language, imagery and content of the Old and New Testaments, could, in the eyes of the readers of the *Oneirocriticon,* be considered genuinely Christian for a good millennium, from the tenth to the twentieth century, even though they had all been copied almost verbatim from a Muslim source. The two instances where the *Oneirocriticon* actually does quote passages from the New Testament represent a deliberate effort to replace a passage from the Qur'an with one from the Christian scriptures.
If the author of the *Oneirocriticon* understood the role of the Quranic passages in Arabic dream interpretation, why did he not try to replace them with biblical passages more often? It is possible that he simply was not creative enough, but I suspect that the reason for the discrepancy is related rather to the nature of basic education in Byzantium and among the Muslims. For Muslims, education began with memorizing the Qur'an at a very young age, beginning possibly as early as four or five. A verse in the Qur'an itself encourages its memorization (54: 17), and completion of this task was a prerequisite for entering a school of higher study. Regardless of whether or not these students were able to remember all of the Qur'iin later in life without the discipline of reciting a portion of it every single day, 84 memorization of the Holy Book at a very young age enabled them at least to recognize a verse and recite the rest of it, if given only a few words. For that reason, Arabic dreambooks rarely quote a verse in its entirety; they usually quote a piece of it followed by the word ~~I ("the verse"), which suggests to the reader that he should supply the rest himself.
The author of the *Oneirocriticon* was probably not nearly as proficient with the Bible; he could not recite by heart innumerable appropriate passages at will. The study of the Bible in Byzantium was not nearly as systematic as the study of the Qur'an in the Muslim world. The New Testament and the Psalter could form part of elementary education, but memorizing them was not its
<sup>84</sup> This is how contemporary *J:iuffd?* (memorizers of the Qur'an) manage to retain their ability.
goal, though it is possible that individuals memorized passages from both, especially since excerpts from them were read during church services. The parts of the Bible used in the liturgy were obviously more familiar than others, and familiarity with the texts of the Old and New Testament lent a biblical coloring to the language of Byzantine authors, but having the entire text of the Old and the New Testament (which is much longer than the Qur'an) at one's fingertips is quite a different matter. Nor did a medieval author have a concordance to help him find a particular passage with the same ease that modern concordances afford.
#### CHAPTER NINE
#### THE THIRTEEN ANECDOTES QUOTED IN THE ONEIROCRITICON
In addition to providing the meanings of the various dream symbols, the Oneirocriticon recounts thirteen short anecdotes to illustrate the fine points of dream interpretation. Such stories are also standard features of Arabic dreambooks, and equivalents to at least seven of the thirteen repeated in the Oneirocriticon can be found in Arabic sources. They are usually sprinkled throughout a dreambook, accompanying the theoretical interpretation of the dream symbol that they discuss, but some collected them together either at the very end of the work, in a manner vaguely reminiscent of Artemidoros, or at its beginning, as is done by Ibn Qutayba. Eleven of the thirteen anecdotes of the Oneirocriticon are clearly appended to the pertinent theoretical interpretation of the dream symbol that forms their core (chapters 19, 20, 46, 96, 144, 147, 153, 176, 194, 199 and 264). The remaining two (chapters 36 and 139) at first seem out of place, but an explanation for their position can be found.3
3 The position of chapter 139 could be explained if we view it as the Christian adaptation of an Islamic dream about a woman preaching from the minhar; in at least one Arabic dreambook this dream belongs to a chapter on mountains and elevated places, which would explain why one finds it in an approximately equivalent position in the Oneirocriticon; see chapter 7, "Priests and Priestly Duties." It has been impossible to locate an Arabic narrative equivalent to the contents of chap. 36, which might have provided some insight regarding its position. Chap. 36 narrates a dream dreamt by one of al-Ma'mun's nobles who had a white grapevine at his house. He dreamt that this vine produced black grapes. Sirim's interpretation was that one of the dreamer's wives was pregnant by his Ethiopian servant, and did later gave birth to a black child. Chapter 36 is inserted between a chapter on beards (35) and two chapters on dying one's hair (37 and 38). Further chapters around it discuss members of the human body. The only conceivable relationship of the narrative with its surrounding chapters is its reference to the change in the color of the
I This is the approach in al-Dinawari, in the Pakistani edition of the dreambook attributed to Ibn Sirin, Ta'hir al-ru'ya, Arabic text and Urdu translation, and in the texts that formed the basis for its Italian translation (Ibn Sirin, Libro del sogno, trans. Zilio Grandi), as well as in the dreambook also called Ta hir al-ru'ya, attributed to 'Umar al-Khayyam.
<sup>2</sup> Al-Muntakhab quotes anecdotes throughout its chapters, and includes a final chapter that consists mostly of anecdotes preceded by an isnad, i.e., a list of the people who transmitted the story orally until it was written down. The same practice is followed by Ibn Shāhīn, and in the text (the Arabic version of which, either printed or manuscript, is not identified by the translator) used as the basis for the French translation of ps.-Ibn Sirin, Ta'hir al-ru'ya (Interprétation des rêves, trans. Penot). Artemidoros mentions dreamt and interpreted throughout his work, but the last of his five books is exclusively dedicated to this kind of practical exercise.
Four of the thirteen narratives refer to historical events allegedly predicted by interpreting dreams. The first one, chapter 46 (Drexl 29, 18-30, 10), tells of a dream dreamt by Caliph al-Ma'mun:
"Οναρ έθεάσατο ό πρωτοσύμβουλος Μαμούν τοιούτον, ότι εύρέθη είς τον μέγιστον ναόν είς το Μέκκε, ήγουν έν τη σκηνή του Αβραάμ και άπελθών είς το ανω μέρος, ἔνθα ὁ θρόνος ἐστί, πρὸς τὰς δύο γωνίας οὐρησεν ἔξω τοῦ ναοῦ ἐν αύταις, και θορυβηθείς ώς ἐπ' άμαρτία μεγίστη ἀπέστειλε πρὸς τὸν όνειροκρίτην Σηρείμ λάθρα οίκειωσάμενον το όναρ έρμηνευθηναι αυτφ. ό δε όνειροκρίτης Σηρείμ διήλεγξεν αύτον λέγων: ότι τουτο έσυ ούχ έώρακας: ου γαρ εί έκ των του πρωτοσυμβούλου αίμάτων. και διά τουτο ού μὴ κρίνω αὐτό. τὸτε δὲ αὐτοψεὶ καλέσας αὐτὸν ὁ πρωτοσύμβουλος ἐξηγόρευσεν ὁρκίσας αὐτὸν μὴ ἀποκρύψαι τὴν του όνείρατος κρίσιν. ὁ δὲ εἶπεν αὐτῶ· αί τέσσερες γωνίες τοῦ ναοῦ, ἐν αίς οϋρησας, τεσσαρα τέκνα σημαίνουσιν. τὰ μὲν δύο του ανου μέρους, ε̃νθα ό θρόνος, του θρόνου και της άρχης διάδοχοι έσονται τα δε δύο του κάτω μέρους γεννήσεις μέν, ούκ έπιλήψονται δε της αρχής και του θρόνου.
The caliph Mamoun had the following dream: he found himself in the most holy shrine in Mecca, that is the tent of Abraham. He went into the upper part where the throne stands and urinated in the two corners outside the temple. The caliph was troubled over this very great sin and sent a person to the dream interpreter Sereim to find out secretly the dream's meaning by pretending that he had seen it himself. But the dream interpreter refuted him, saying: " You did not dream this, for you are not a blood relative of the caliph. Therefore, I will not interpret the dream." The caliph then summoned Sereim into his presence and, having confessed, made Sereim swear not to conceal the interpretation of the dream. Sereim replied: " The four corners of the temple in which you urinated signify four children. The two of the upper part, where the "throne" is located, will become successors to your throne and empire. The two of the lower part will be born to you but will not gain possession of your throne and kingdom."
The anachronism of presenting Ibn Sirin (d. 728) and Caliph al-Ma'mun (r. 813-33) as contemporaries has been pointed out by modern scholars time and again. In addition, the events foretold by al-Ma'mun's dream-that he will be succeeded by two of his sons-are also historically wrong. Al-Ma'mun did not father any caliphs. He was followed on the throne by his brother al-Mu tasim who, in his turn, was succeeded by two of his sons, al-Wathiq and al-Mutawakkil.
grapes, which was possibly perceived as similar to the effect of the dyes on human hair. If the position of chapter 36 is not arbitrary but somehow reflects its placement in the Arabic sources used by him, it suggests that a number of Arabic interpretations of a change of color in various objects was omitted by the Greek author, which would be consistent with his statement, "I have extracted summarily' (έξ έκάστου τούτων ἐκλεξάμενος κεφαλαιωδώς) (Drexl 1, 9-10). According to Arabic dream interpretation, a tree in one's yard is to be interpreted as one's wife, while the products of the tree should be understood as her offspring. A similar but not identical anecdote to the one in chapter 139 is recounted by al-Dinawari, fasl 19, bab 71, Esad Efendi 1833, fol. 194a.
Al-Ma'miin's dream is not original. Similar dreams were purportedly dreamt by or about at least two other caliphs, both belonging to the previous Umayyad dynasty. One of the oldest surviving records of Islamic dream interpretation, the list of dreams interpreted by Sa'id b. al-Musayyab (d. 712) and mentioned in the biographical dictionary of Ibn Sa'd (d. 845), includes a dream signifying that Caliph 'Abd al-Malik (685-705) would beget four of his successors:
j.,>1 J.,,....r:.L........ul ~ r-'""WI .J.:> ~I~..\...:.. :Ju~ .J.:> ~ L.:i~I :Ju ~' ~ u.,w ~ J~ dill~ ::,u: ~iJ ~J Ju :Ju~ *er.'* <sup>~</sup>iu c:J~JJ ~~ .. :.,): Jw ~I .J.:> ~ d.Jj '..:.,~ii, *)\_r-0* ~) \_.w,;.~~I ~~
He said:" MuI:iammad b. 'Umar informed us and said that al-I:Iakam b. al-Qasim reported on the authority of Isma'il ibn Ab! I:Iaklm, who said: 'A man said: "I dreamt that 'Abd al-Malik b. Marwan urinated in the *qibla* of the Mosque of the Prophet <in Mecca> four times." I mentioned this to Sa'ld b. al-Musayyab who said: "If your dream is truthful, four caliphs from among his offspring will rise in it.".
'Abd al-Malik was in fact the father of al-Walld, Sulayman, Yazid II, and Hisham, who were the sixth, seventh, ninth, and tenth Umayyad caliphs respectively. A story quoted by al-Dlnawari says that a similar dream was dreamt by 'Abd al-Malik's father, Caliph Marwan I:4 J~ .c.L.S: ~10-7 ..:,l\_,J-A lSl.JJ ~UJ..:..J I .J..:. ~I JL.i\_g *'.I.!* '" 11 0-7 I \_.. ... ~ ~L:J.) ~ '-;-' 1.r=J-1 <sup>~</sup> (Marwan b. al-I:Iakam dreamt that he urinated on the mi~riib. <sup>5</sup>He narrated his dream to Sa'id b. al-Musayyab, who said: "Indeed, you will beget caliphs").6 Further on in his dreambook, al-Dinawari quotes a similar anecdote, though the identity of the dreamer is not mentioned and the outcome of the dream is less grand: <sup>7</sup>
dJ ..J-'7. Jl.Ll I~ ......:...c J~ ..,..,1~ ~ Jl,i ....:.Ls: <t....L...... ~ ~J 1..5IJ . "'-:> ':?'~ L.L.1 ~ i)l.L
<sup>4</sup>*Fa\$f* 7, *bah* 20; *Esad Efendi* 1833, fol. 69a, BN *arahe* 2745, fol. l 13a.
<sup>5</sup> The prayer niche of a mosque.
<sup>6</sup> The same anecdote is also recorded in the dreambook of al-Kharkiishi and in *al-Muntakhah.* Cf. Lamoreaux, "Dream Interpretation in the Early Medieval Near East," p. 321(table12).
<sup>7</sup> Al-Dinawari,fa# 8, *bah* 59; *Esad Efendi* 1833, fol. 62b, BN *arahe* 2745, fol. 140b.
#### 378 CHAPTER NINE
A man saw in his dream that he urinated on a mi~rdb and asked a dream interpreter about it. He said: "A boy will be born to you who will become an imam and his example will be followed."
Unusual urination as a symbol of begetting exceptional sons is an ancient topos of Near Eastern dream interpretation (but is unknown in Artemidoros). Similar dreams are quoted in the Assyrian dreambook of the seventh century B.c., though it does not attach them to a specific historical personage, and by Herodotos of Halicarnassos (fifth century B.c.), a Greek author who wrote on Persian history and mentions it in connection with the birth of the Persian king Cyrus the Great (r. 550-29 B.c.).8 Throughout the centuries, the symbolism of such a dream was adapted to reflect the particular circumstances to which it was applied. The example concerning Marwan I could have been intended as propaganda during his son's, or even his own, reign. 9 Both Marwan b. al-I:Iakam (Marwan I) and his son 'Abd al-Malik, who was named heir apparent shortly after his father's accession, had to struggle to consolidate their position. A dream foretelling that they would beget future caliphs would add supernatural legitimacy to their power. As for the dream concerning 'Abd al-Malik, which is quoted in the ninth century by Ibo Sa'd, the accuracy of its detail suggests that the narrative was invented after the fact, possibly in order to enhance the prestige of dream interpretation as an art. The dream of al-Ma'miln quoted in the *Oneirocriticon* represents yet another revision of the topos. <sup>10</sup>Three of the
<sup>8</sup> The passage in question is from Herodotus 1: 107 ff.; see Oppenheim, "The Interpretation of Dreams in the Ancient Near East," p. 265. Oppenheim says that a narrative about a similar dream dreamt, not by Cyrus's grandfather Astyages, but by his daughter Mandane, is given by Ktesias of Knidos (4th century B.c.). I have been unable to locate such a narrative in the fragments of Ktesias assembled and published by F. Jacoby, *Die Fragmente der griechischen Historiker,* (Berlin, 1923), no. 688. Unfortunately, Oppenheim does not give the source of his information.
<sup>9</sup> The accession of Marwan changed the branch of the clan of Umayyads who exercised power. After three descendants of Abii Sufyan (Mu'awiyah I, Yazid I and Mu'awiyah II) Marwan began a line of caliphs descended from al-f:lakam, in spite of the existence of Sufyanid descendants, because they were too young to be acceptable. Al-f:lakam's line eventually produced both the remaining Umayyad caliphs and the Umayyad rulers of Spain. For a history of the Umayyads, see H. Kennedy, *The Prophet and the Age of the Caliphates. The Islamic Near East from the Sixth to the Eleventh Century* (London and New York, 1986).
<sup>1</sup>° Chapter 46 repeats a second topos: a dream interpreter can deduce from the nature of a dream whether he is conversing with the dreamer or with someone who pretends to be the dreamer. In the latter case, he would either demand to know the identity of the dreamer (as is done in chapter 46 of the *Oneirocriticon)* or announce that he is aware he is dealing with an impersonator and then nonetheless interpret the dream correctly (chapters 20 and 176). The same topos can be found in the Arabic sources, beginning with the interpretations by Ibn al-Musayyab quoted by Ibn Sa'd. After a chain of authorities ( *isnad)* that is meant to guarantee the authenticity of the narrative, the events are narrated by an eyewitness: "'Umar b. f:labib b. Qali' said: 'I was sitting by Sa'id b.
caliphs who reigned between the seventh and the eleventh centuries fathered two of their successors and one of them, al-Walid I (705-15), was indeed a contemporary of Ibn Sirin; the other two were al-Mahdi (r. 775-85) and al-Mu tasim (r. 833-42). It is conceivable that the narrative of the Oneirocriticon was initially devised with al-Walid in mind and that al-Ma'mun's name was inserted later, most likely not by an Arab author, but by the Greek compiler of the Oneirocriticon.
The second anecdote that seems to refer to actual historical events appears in chapter 192 (Drex1 148, 18-25):
Έλθών 'Αμάμ, ό δεύτερος πολεμιστής του πρωτοσυμβούλου, ήρώτησε τον Σηρείμ λέγων είδον κατ' όναρ, ότι πάντες οί κέραμοι της πόλεως άνευ ύετου ἔρρεον ἔνθολον και πάντες ἐδέχοντο τὸ ὕδωρ ἄνευ ἐμοῦ καὶ τῶν ἐμῶν ἀνθρώπων. ό δε έφη επίβουλοι και άρπαγες έσονται του πρωτοσυμβούλου πολλοί, συ δέ ἐκεῖσε οὐ μὴ ἐγχειρισθῆς. καὶ μεθ' ήμέρας δέκα ἐγένετο σφαγὴ αὐτοῦ καὶ άρπαγὴ των αὐτου. και αὐτὸς ὁ ׳ Αμαμ οὐκ ἐκοινώνησεν ἐν αὐτοῖς, καθώς είπεν ό Σηρείμ.
Amam, the second most important warrior of the caliph Mamoun, came and consulted Sereim: "I dreamt that without rain all the clay pipes in the city were overflowing with turbid water and that everybody was receiving it except for me and my people." Sereim replied: "Many will plot against and rob the caliph, but you will have no hand in it." Ten days later the caliph was slaughtered and his possessions plundered. However, Amam did not participate with them, just as Sēreim had predicted.
The details of the story are both vague and inaccurate, obfuscated by the simultaneous presence of Ibn Sirin and Caliph al-Ma'mun and the fact that al-Ma'mun did not die in a revolt. Any effort to identify Amam with a historical
al-Musayyab one day and I was anguished by certain things and a debt was burdening me. So I was sitting by Ibn al-Musayyab not knowing where to go. And a man came to him and said: "Abu Muhammad, I had a dream." Ibn al-Musayyab said: "What was it?" The man said: "I saw myself seize 'Abd al-Malik b. Marwan and make him prostrate on the earth. Then I threw him on the ground and fixed four pegs in his back." Ibn al-Musayyab said: "It is not you who had this dream." The man said: "And yet I had it." Ibn al-Musayyab said: "I am not going to interpret this dream unless you tell me <who the dreamer was>." The man said: "Ibn al-Zubayr saw it, and it was he who sent me to you." Ibn al-Musayyab said: "Indeed, if his dream was truthful, 'Abd al-Malik b. Marwan is going to kill him. And from the line of 'Abd al-Malik four men are going to spring forth and all four of them are going to be caliphs." < Umar b. Habib> said: " Therefore I went to 'Abd al-Malik b. Marwan in Syria [al-Sham] and gave him the news from Sa'id b. al-Musayyab and delighted him. He asked me about Sa'id and about how he was and I told him. And he ordered that my debt be cleared on my behalf and I received beneficence from him." For the Arabic text, see Ibn Sa'd, Al-tahagat al-kubra, vol. 5 (Beirut, 1957), p. 123; see also Fahd. La divination arahe, p. 310. The same dream is related in pseudo-Ibn Sirin, Ta'bir al-ru'ya, chap. 9, Pakistani ed., p. 93; Interprétation des rêves, trans. Penot, p. 77.
figure is frustrated by the inadequacy of the rendering of his name in Greek; it might have been changed beyond recognition in the course of the Greek manuscript tradition. <sup>11</sup>
An equivalent Arabic anecdote quoted in al-Dinawari's dreambook <sup>12</sup>and from there copied to *al-Muntakhab13* is backed by solid and accurate historical details:
A man came to lbn Sirin and said: "I saw that the water was pouring into the clay pipes without rain, and I saw that people were taking from it." Ibn Sirin said: "You did not take from it." The man said: "Indeed, I did not do it, and I did not take any of it." <lbn Sirin> said: "Then your dream was a good omen." It did not take long before the revolt of lbn Muhallab [sic] broke out.
The revolt referred to is that of Yazid b. al-Muhallab, the son of the celebrated general Muhallab b. Abi ~ufra, which took place in 720. Ibn Sirin died in 728, and therefore his presence in the narrative is historically accurate. But when we compare al-Dinawari's story with the anecdote in the *Oneirocriticon,* the details do not tally. The *Oneirocriticon* describes a successful revolt that resulted in the reigning caliph's assassination; in reality, the revolt of Yazid b. al-Muhallab was crushed by General Maslama b. 'Abd al-Malik before the rebels could overthrow the regime. It is possible that the dream of overflowing pipes without rain was a topos that could be applied to a variety of historical situations, as is demonstrably the case with the dream of urinating in a mosque. However, the small number of its attestations (only two, and effectively one, since the narrative of *al-Muntakhab* is clearly copied from al-Dinawari) argue against such an explanation. The Greek narrative of chapter 192 could be manipulated to apply to the revolt oflbn al-Muhallab: if we make a slight textual emendation by inserting 'tIDV between cr<j>ayfi and amou in the penultimate period of the anecdote (making the Greek text read: µe0' fiµ£pm; <>E1rn E:y£vi::w mj>ayi)
<sup>11</sup> A confusion between µ (m) and v (n) is usual in Greek manuscripts in the minuscule. It is even conceivable that "Amam" is not a name, but conveys the Arabic title "imam."
<sup>12</sup>Al-Dinawari,fa,5/ 15, *bab* 41, *Esad Efendi* 1833, fol. 160b.
<sup>13</sup>*Al-Muntakhab,* p. 231, II. 11 ff. The version of *al-Muntakhab* is almost identical with the one quoted in al-Dinawari: ~ .'..>-" *J.:.........=.* ~L,...11 =:I; JW ~~ .J-!I ..,:;1 ~;.:,I ~.J ~.J J.UI ~ ..fil J..::-..)1 Jl...i..i....:..... .:i..:;,.L.:; *'i* ~~ .J-!I Jl...i..i <...:...... *.:,\_,.:i..:;,.''"=* ..>"L.:JI ..:...-:1;.J .rb-o . ..,..J+.11 .J-!I ~..::...:.LS:~ I~ 'ii~~ ~I .U Jl.li ~ <...:...... :i.:;..1
των αύτου και άρπαγή των αύτού) and understand the pronoun αύτου as referring to Amam and not the reigning caliph, the interpretation of Amam's dream could be as follows:
Sereim replied: "Many will plot against and rob the caliph, but you will not be touched by <their schemes>." Ten days later the members of Amam's household were slaughtered and his possessions plundered. However, Amam himself was not harmed, just as Sereim had predicted.
Al-Dinawari's anecdote does not explain how the dream was fulfilled and therefore cannot be used to corroborate either version of the Greek text. The historical accuracy of the Arabic anecdote and the inaccuracy of the Greek version suggest that the original was tampered with, possibly on the Greek side of the tradition.
A third anecdote in the Oneirocriticon suggests that Ibn Sirin and Caliph al-Ma'mun were contemporaries (chapter 144, Drexl 99, 6-13):
Εθεάσατό τις κατ` όναρ ἐκ τῶν μεγιστάνων Μαμουν ὅναρ καὶ ἐλθών ἐιπε τῶ όνειροκρίτη Σηρείμ είδον κατ΄ όναρ, ότι πεδίον εύρύχωρον περιείχέ με έστωτα, όπερ τα πρώτα μεν ήν δασύ και πολύχορτον, μετά ταυτα δε έχένετο ψιλόν και μαδαρόν. και άπεκρίθη ό όνειροκρίτης Σηρείμ: στι τουτον άπιστίαν και διγνωμίαν του ίδόντος σημαίνει. και καθώς είπεν, ούτως και έγένετο ἐν τῶ ίδόντι.
A certain noble of the Caliph Mamoun had a dream and came to the dream interpreter Sereim and said: "I dreamt that I was standing in the middle of a roomy and flat plain. At first, this plain was thickly wooded and very grassy, but then it became barren and treeless." The dream interpreter Sereim replied: "This signifies the infidelity and double-mindedness of the dreamer." And as he said, such was the outcome for the dreamer.
Though the Oneirocriticon comments that Sereim's interpretation was accurate, it does not explain in what way the dream was fulfilled.
The anecdote of chapter 144 again turns out to be an Arabic topos, whereby the fertile, wooded, and therefore shaded land represents Islam, and its abandonment foretells apostasy. The symbolism of this imagery is consistent with the psychology of the Arabs as a people of the desert, where a shaded and cool place is much desired, and is further exploited in the Arabic interpretation of gardens. 4 The earliest attestation of an Arabic anecdote equivalent to chapter 144 of the Oneirocriticon is found in the Tabagat of Ibn Sa'd:15
<sup>14</sup> See Zilio Grandi, "Un esempio di interpretazione dei sogni," pp. 53-66.
<sup>15</sup> Ibn Sa'd, Al-tahagāt al-kuhra, vol. 5, p. 125; Fahd, La divination arahe, p. 312, no. 13.
#### CHAPTER NINE
قال: اخبرنا محمد بن عمر قال: حدثنا ابن ابي ذئب عن مسلم الخياط عن ابن المسيب قال: الكبل فى النوم ثبات فى الدين. قال وقال له رجل: يا ابا محمد إنى رايت كانى جالس فى الظل فقمت الى الشمس. فقال ابن السيب: والله لئن صدقت رؤياك لتخرجن من الاسلام. قال: يا ابا محمد إنىى ارانى اخرجت حتى ادخلت فى الشمس فجلست. قال: تكره على الكفر . قـال فـخـر ج فى الزمان عبد الملك بن مروان فأسر فأكره على الكفر فرجع ثم قدم الدينة وكان نخبر هذا
He said that Muhammad b. 'Umar said that Ibn Abī Dhi'b told us on the authority of Muslim al-Khayyat on the authority of Ibn al-Musayyab that <Ibn al-Musayyab> said: "Fetters in a dream are firmness in religion." A man told him: "O Abū Muhammad, I dreamt that I was sitting in the shade and I stood up <in order to sit> in the sun." Ibn al-Musayyab said: "By God, if your dream is truthful, you will abandon Islam." The man said: "O Abu Muhammad, I saw that I was ousted until brought into the sun where I sat down." Ibn al-Musayyab said: "You will be forced to become an infidel." He went out <to fight> at the time of 'Abd al-Malik b. Marwan. He was taken prisoner and was forced to become an infidel. He later returned and went to Medina, where he used to tell this <story>.
A similar anecdote is repeated by al-Dinawari. This time the land abandoned by the dreamer is said not simply to be shaded, but covered with lush vegetation, which brings al-Dinawari's paradigm closer to chapter 144 of the Oneirocriticon. This time the dream is said to have been interpreted by Abu Bakr, a companion of the Prophet and first caliph of Islam, who, according to Muslim tradition, was also a gifted dream interpreter:16
قال رجل لابي بكر الصديق رضى الله عنه رايت كانى فى ارض مخصبة معشبة فخرجت منها الى الارض كالحة مجدبة قال ان صدقت رؤياك خرجت من الاسلام الى الشرك فكان كذلك.
A man said to Abu Bakr al-Siddiq, may God be pleased with him: "I dreamt that I was in a fertile and grassy land and I left it to go to a fallow and barren land." Abu Bakr said: "If your dream is truthful, you will abandon Islam for polytheism." And it happened thus.
The dreambook called Ta bir al-Ru'ya and attributed to Ibn Sirin (though definitely written later than the late eleventh century'') quotes a longer version
<sup>16</sup> Al-Dīnawarī, fașl 18, băh 4; Esad Efendi 1833, fol. 174b.
<sup>17</sup> The anecdote can be found in Ibn Sirin, Ta bir al-ru'yã (Pakistani ed.), p. 54; Libro del sogno, trans. Zilio Grandi, p. 34; Interprétation des rêves, trans. Penot, p. 45.
of the story and identifies the dreamer by name. This version also explains the insertion of fetters (كيل) into the interpretation that preceded the narrative of the same dream in Ibn Sa'd:
حُكَ ان ربيعة بن اميّة بن خلف جاء الى ابي بكر الصديق رضى الله تعالى عنه فقال يا خليفة رسول الله صلى الله عليه وسلم انّى رايت البارحة فى منامى كانّى فى ارض خضرة مخصبة وقد افضيتُ منها الى ارض مجدبة لا نبات فيها ورايتُك قد جمعت يداك وغلتا الى عنقك فقال له الامام ابو بكر الصديق رضى الله تعالى عنه إن صدقت رؤياك خرجت من دين الاسلام الى دين الكفر فامًا انا فقد جمعت لى اموري وغلت يداي عن حطام الدنيا قـال فلما كان فى ايام عمر ابن الخطاب رضى الله تعالى عنه خرج ربيــعـة من الدينة ولحق بارض الروم فتَنَصَّرَ عند قيصر ومات نصرانيًا والله اعلم
It is told that Rabi'ah b. Umayyah b. Khalaf'" went to Abu Bakr al-Siddiq, may God be pleased with him, and said: "O viceregent of the messenger of God, may the peace and blessings of God be upon him, indeed I saw yesterday in my sleep that I was in a green and fertile land and I abandoned it to go to a barren land with no vegetation. And I saw you with your hands bound and fettered to an iron collar around your neck. And the imam Abu Bakr al-Siddiq, may God be pleased with him, said: "If your dream is truthful, you will leave the religion of Islam for the religion of the infidels. As for me, my affairs are bound <to develop in a certain way> and my hands are fettered <so that I cannot touch> the vanities of this world." <The narrator> said: "In the days of 'Umar b. al-Khattab, may God be pleased with him, Rabi'ah left Medina and entered the land of the Byzantines and became a Christian before the emperor and died as a Christian. God knows best."
The abandonment of a green and fertile land as representing the abandonment of Islam is repeated in all the Arabic versions. It is evidently related to the significance of the color green and, by extension, of vegetation, in Muslim religious symbolism. It is therefore unlikely that the dream was interpreted simply as " infidelity and double-mindedness" (απιστία και διγνωμία) and not as apostasy from Islam in the original Arabic version used for compiling the Oneirocriticon. The down-playing of the dream's interpretation in the Oneirocriticon must be the work of the Christian author.
A fourth anecdote (chapter 153) purports to be set in historical time, as it mentions al-Ma'mun, Ibn Sīrīn and the outbreak of a revolt.19 An analogous
<sup>18</sup> The father of Rabia, Umayya b. Khalaf, was one of the earliest converts in Islam; see Ibn Sīrīn, Libro del sogno, trans. Zilio Grandi, p. 34.
<sup>19</sup> Drexl 111, 24s112, 4:"Αρχων λαού έλθών του πρωτοσυμβούλου Μαμούν ήρώτησε τφ
narrative could be found in the Arabic sources investigated and therefore neither the revolt nor the dreamer can be identified.
Though the Greek versions of the anecdotes purporting to refer to historical events and repeatedly mentioning Ibn Sirin and Caliph al-Ma'mun as if they were contemporaries are unhistorical, in their Arabic versions, whenever the names of both the dreamer and the dream interpreter are given, great care is taken to make the narrative historically accurate. Both the author and readers would have had biographical information about the people mentioned, since Abū Bakr, Ibn al-Musayyab and Ibn Sīrīn were important as transmitters of the Prophetic traditions. The need to authenticate these traditions spurred Muslims to compile biographical dictionaries with information on the transmitters, especially as to dates of birth and death.20 The gross error of presenting Ibn Sirin and Caliph al-Ma'mun as contemporaries would have been easy for an educated Muslim to avoid and for an educated Muslim reader to detect, but it could go completely unnoticed by a Greek. It is therefore more likely that the mistake was introduced by the Greek author of the Oneirocriticon, rather than by his Arabic source. Ibn Sirin was chosen as the protagonist for a number of Greek narratives, evidently because of his reputation as the quintessential Arab dream interpreter and the frequent references to him in Arabic dreambooks. Al-Ma'mun's name could have appeared in the Arabic sources of the Oneirocriticon as well-though not in connection with Ibn Sirin-but making him a figure in seven out of the thirteen Greek narratives21 is a more plausible
Σηρείμ λέγων… "ειδον κατ΄ όναρ φάραν ξανθόν άχρι μιλίου έποχούμενον κάμε ἐκεῖνον κάκεινον έμε". ό δε έφη "ποία ώρα ήν, ότε είδες τουτο;" και άπεκρίθη "ότι πρός την εύχὴν τῆς αὐγῆς ἤτοι τὸ περιόρθριον". καὶ εῖπεν "ἐπειδὴ ή όδός σου μετ᾽ αὐτοῦ μίλιον ήν, ού μὴ παρέλθη ό μὴν οῦτος καὶ σφαγήση ἐν πολέμῳ μούλτου". καὶ ἐχένετο μοῦλτον τῶν Αράβων και έσφάγη ό ίδών (One of the leaders of Caliph Mamoun came and asked Sereim saying: "I saw in my dream that a light brown horse was riding on me and that I was riding on it for about a mile." Sereim said, "What time was it, when you dreamt this?" He answered, "It was around the morning prayers, that is dawn." Sereim said, "Because your trip with the horse was a mile long, before this month goes by you will be slain in a rebellion." <Indeed,> a rebellion of the Arabs took place, and the dreamer was slain ).
<sup>20</sup> On the importance of these three figures for hadith literature and on the compilation of biographical dictionaries, see M. Zubayr Siddigi, Hadith Literature: Its Origin, Development and Special Features (Cambridge, 1993), esp. pp. 91-105. For an essay on the cultural and intellectual significance of biographical dictionaries in classical Islam, see W. al-Qadi, "Biographical Dictionaries: Inner Structure and Cultural Significance," in The Islamic World: The Written Word and Communication in the Middle East, ed. G. N. Atiyeh (New York, 1995), pp. 93-122.
<sup>21</sup> Chaps. 19 (Drexl 15, 18 ff.); 20 (Drexl 16, 1 ff.); 36 (Drexl 23, 23 ff.); 46 (Drexl 29, 28 ff.); 96 (Drexl 57, 11 ff.); 144 (Drexl 99, 6 ff.); 153 (Drexl 111, 24 ff.).
choice for the Byzantine author, since al-Ma'mun is cited in Byzantine sources as a caliph distinguished for his scientific interests.22
Two more of the Greek anecdotes repeat topoi frequently attested in the Arabic dreambooks. Chapter 147 (Drex1 102, 1-12) demonstrates through the use of an example that the structural elements of a house correspond to the members of the family that inhabits it:
Ηρώτησέ τις γυνή τον όνειροκρίτην Σηρείμ· ότι είδον κατ' όναρ, ότι ή ανω φλιά της οίκίας μου τη κάτω φλιά προσέπεσε και των δύο θυρών ή μεν έσω, ή δε έξω έπεσε. και άντηρώτησεν ό Σηρειμ τη γυναικί: έχεις ανδρα και τέκνα; ή δε έφη έχω ἐν τῆ ξένῃ δέ ἐστιν ὁ ἀνήρ μου μετὰ τοῦ νίοῦ μου, ή δὲ θυγάτηρ μου παρ' ἐμοί. ὁ δὲ Σηρεὶμ ἀπεκρίνατο· ταχὸ ήξει πρός σε ὁ ἀνὴρ διὰ τὴν ἄνω φλιάν μετά του νίου σου και της νύμφης αύτου διά την πρός τα ένδον πτώσιν ή δε θυγάτηρ σου έτέρω προσκολληθήσεται ανδρί και άποδημήσει, διότι πρὸς τὰ έξω ή μία θύρα έπεσεν. και καθώς ήρμήνευσεν, ούτως και έγένετο.
A woman consulted the dream interpreter Sirim: "I dreamt that the top door jamb of my house fell down onto the lower one, causing one door leaf to fall to the outside of the house, the other to the inside." Sirim asked the woman: "Do you have a husband and children?" She said: "I do, my husband is abroad with my son, and my daughter is with me." Sirim answered: "Your husband will soon return to you because of the fall of the top door jamb, together with your son and his bride, because one door fell to the inside. But your daughter will become attached to another man and go away, because one of the doors fell to the outside." And things happened exactly according to his interpretation.
The same anecdote is told by al-Dinawari, except that one of the details of the dream (the leaf of the door that fell toward the inside) remains unexplained. The Greek version therefore complements our understanding of al-Dinawari's narrative.23
جاءت امراة الى ابن سيرين فقالت رايت فى المنام اسكفة بابى العليا وقعت على اسكفة السفلى ورايت المصراعين سقطا فوقم احدهما داخل البــيت والاخر خرجه فقال هل لك زوج وولد غاشبان قالت نعم قال اما سقط الاسكفة العليا فزوجك يقدم عليك عاجلا واما المصراع الذي وقع خارجا فابنك
<sup>22</sup> See also the remarks in P. Magdalino, "The Road to Baghdad in the Thought World of Ninth-Century Byzantium," in Byzantium in the Ninth Century: Dead or Alive? ed. L. Brubaker (Hampshire, Eng., and Brookfield, Vt., 1998), pp. 195-213. Magdalino notes that in the Byzantine account of the invitation extended to Leo the Mathematician to visit the caliphal court, the name of the caliph was changed from al-Mu'tasim to al-Ma'mun, because of the latter's reputation among the Byzantines as a patron of letters and science (ibid., pp. 200-1).
<sup>21</sup> Al-Dinawari, fash 18, bab 51; Esad Efendi 1833, fol. 182b; repeated in al-Muntakhab, p. 252 (corrections to Esad Efendi introduced on the basis of al-Muntakhab).
$\vec{\text{L}}\text{-}\text{:}\mathsf{L}\text{-}\mathsf{L}\text{-}\mathsf{L}\text{-}\mathsf{L}\text{-}\mathsf{L}\text{-}\mathsf{L}\text{-}\mathsf{L}\bot\mathsf{L}\bot\mathsf{L}\bot\mathsf{L}\top\mathsf{L}\text{-}\mathsf{L}\mathsf{L}\text{-}\mathsf{L}\mathsf{L}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-}\mathsf{L}\mathsf{-$
A woman went to Ibn Sirin and said: "I saw in a dream that the upper lintel of my door fell on the lower one. And I saw that the two leaves of the door had fallen and one of them was lying toward the inside of the house, while the other toward the outside." He said: "Do you have a husband and son who are away?" She said, "Yes." He said: "The fall of the upper lintel <means that> your husband will return to you soon. Regarding the doorleaf that fell toward the outside, your son will marry a foreign woman." It did not take long before the husband and her son came back, and with him there was also a foreign woman.
Al-Dinawari quotes yet another anecdote with similar contents:24
جات امراة الى معبر واسقطت جلبابها متحديره وكانت غائبة من وطنها وسألت عن رؤيا راتها وهي انها رات كان بيتها انهدم على اولادها وطار كتاب منه الى وطنها الاول فتقال الحبير ابشرى فانه يصل اليك مال وسترجعين الى وطنك مكان الكتاب الطائر فلم يبرح مكانها حتى ورد عليها نعى عم لها تاجر بانه غرق في البحر وقد اتيت بالدواب فرجعت مع او لادها الى وطنها.
A woman went to a dream interpreter and was so disconcerted that her jilbab dropped. She had been away from her homeland and asked about a dream she had seen. This was that her house had fallen on her children and a written message had flown from it to her homeland. The dream interpreter said: "Be happy, because this tells you that <you will receive> money and will return to your homeland, the source of the flying message." She was still in her place when she received news that an uncle of hers who was a merchant had drowned at sea. She was brought beasts to ride and returned with her children to her homeland.
Interpreting a house's structural elements as standing for one of its inhabitants is mentioned both in Artemidoros (ii.10; Pack 115, 2-117, 2) and in the Babylonian Talmud.25 It is repeated in numerous anecdotes found in the Arabic dreambooks, some of which seek to make points beyond the equation between the symbol and its interpretation. Ibn Qutayba tells the following story:26
وحدثني احمد بن سعيد عن ابي عبيد في كتاب غريب الحديث ان امراة اتت رسول الله صلى الله عليه وسلم فقالت رايت كان حائزة بيتى انكسرت قال ﻳﻘﺪﻡ ﺯﻭﺟﻚ ...
<sup>24</sup> Al-Dīnawarī, fași 18, bāh 38; Esad Efendi 1833, fol. 180a.
<sup>25</sup> Kister, "Interpretation of Dreams, " p. 100.
<sup>26</sup> Ibn Outayba, Yahuda ar. 196, fol. 8a-b. Ankara fol. 194a-b.
AJ:imad b. Sa'id told me on the authority of Abu 'Ubayd in the book, *Rare Words in the Traditions,* that a woman went to the Prophet, may the peace and blessings of God be upon him, and said: "I saw that the beam of the roof of my house broke." He said: "Your husband will return."
Ibn Qutayba further relates that the woman dreamt the same thing a second time. She went to the Prophet again, but could only find Abu Bakr. He said that her dream meant that her husband would die, and so it happened. Ibn Qutayba comments that the same dream was given two different interpretations, either because the countenance of the woman changed or because the time that she dreamt it was different in each case. 27 A similar dream is narrated by al-Dinawari: <sup>28</sup>
~ <lJI ~ <lJI J\_,\_.....\_; \_}I u•~ J...~ ~J ~Jj (~ ~1->-"I .:\_,I~ J~\_; JW ~I~ Jj--"-C- .:\_,IJ ~)·~ u.i..lJ ~I ~I\_; ~I dl.LI FJ i..LI ~ u.i..l~ ~Jj ~..>-:!J <lJI .Ll .:.,) ~..u\_:; rhuJ ~ <l.JI ~ <l.JI JLLI FJ ~ <lJI ~~I \_}I u•l..\_;....j d.Jj *J.'.:......* d\_;J (~ ~ lp.Jj ~I ..::....:;u ~,J\_; d\_;J (~ ~ ~Jj J~ ~ d.J.:i.5: *.:\_,l..S....Q* JJ~I <i...J~ J.'.:.-. 4J L:..... -~ • I )L..J I <L...L: <L.::.. I . I . I . I dW d.....il.:; • I " • I . .. <L...L: <lJ I I. - i - . *J.)* ~ '-+' r- r-J - *c.r-* <l.J I~ ~I -::.WI~ ~Jj u~J ~..u\_:; 4J ~ ~~U ..:.l~.J>:> *.:\_,LS....Q* JUL.~ *J-A* JU rL' -:.JU .i....::..1 d.J ~ *JA* YJ JL.i...i FJ ~ .d.J.:i.s:
It is told that a pregnant woman whose husband was away went to the Messenger of God, may the peace and blessings of God be upon him, and said: "I dreamt that I gave birth to a girl and that the pillar of my house <sup>29</sup>broke." The Messenger of God, may the peace and blessings of God be upon him, said: "God willing, you will give birth and your husband will return." <Indeed,> she gave birth and her husband came back, and then left <again>. She had the same dream and went to the Prophet, who gave her the same interpretation as the first time, and it happened thus. Her husband came back and then left again, and she had the same dream. She went to the Prophet, may the peace and blessings of God be upon him, and did not find him. Some of his wives, may the peace and blessings of God be upon him, told her: "Relate your dream to us." She related it to them, and they said to her: "You will give birth and your husband will die." Then the Prophet
<sup>27</sup> Variations of this narrative are recorded in other Arabic sources and in the Talmud; see Kister, "Interpretation of Dreams," pp. 100-1.
<sup>28</sup> Al-Dinawarl, fa~/ 6, *bah* 3, *Esad Efendi* 1833, fol. 46a; BN *arahe* 2745, fol. 67a.
<sup>29</sup> The translation renders as "house" the Arabic word *hayt.* which could mean either a tent (especially in the context of mostly nomadic early Islamic Arabia) or a house erected as a permanent construction.
#### 388 CHAPTER NINE
came, and asked her: "Did someone interpret your dream <already>?" She said, "Yes." And he said "The dream will be fulfilled according to what he said," and so it happened.
Besides interpreting the pillar of a house as standing for its master, this anecdote illustrates a principle already recorded in the Talmud, whereby a dream is fulfilled according to its interpretation.30 The above narratives are representative not only of the agreements but also of the variety within the general consistency between the ancient Greek, Jewish, Arabic, and medieval Greek traditions of dream interpretation.
Chapter 199 of the *Oneirocriticon* (Drexl 156, 6-19) shows that the meaning of a dream symbol could change depending on the season of the year when it was dreamt. This principle is copied from the *Oneirocriticon'* s Arabic sources and is expounded in chapters 2 (Drexl 2, 15-16) and 301 (Drexl 240, 21-25). The story in chapter 199 reads as follows:
"A v8pom6i; ni; £A8wv i]pcirtTJcrE 1ov I.ripdµ f:voomov *1wA.A.rov A.tywv·* d8ov Km · ovap, on Etc; 8£v8pov, 0 AEYE'tat PEptKOKKia, civapai; ilcr8tov mu Kapnou. Kat E<j>T] au10· on EUpi\crEti; ciya8ov Kat EUJtOtfov cino civ8poi; JtAOUcriou. Kat µEff i)µf:pai; \, Kavai; f:A.8wv ETEpoi; av8pwnoi; EVOOJtlOV 'tOOV aU'tOOV i]pOOTTJO"EV aUTOV Atywv· on Etc; 8£v8pov, 0 AEYE'tat PEptKOKKia, civapai; ilcr8tov mu Kapnou. Kat E<j>T] au10· on EUpi\crEtc; 8At'lfEti; Kat Pcicravov. Kat 8tT]1t0pT]crav Ol cruyKa8i)µEVOl, nffii; £voe; OVmi; mu Opaµami; 8ta<j>OpOV TTJV KptCHV E1toti\crmo. Kat cl1tEKpt8T]· on, O'tE i]pOO'tTJO"EV 6 npromi;, Kapno<j>6poi; liv 6 Katp6i;, O'tE 8£ 6 8EU1:Epoi;, <j>8tv6nwpov liv. Kat Ota mum µETT]AAUYTJ Ti Kpicni;. Kat EpEuvi)crav1Ec; EUpov EV ciµ<j>o1f:poti;, Ka8c0i; dnEv 6 ovEtpoKpiTTJc;.
A man came and consulted the dream interpreter Sereim in front of many other people saying: "I saw in a dream that I climbed on an apricot tree and was eating its fruit." Sirim said to him that "you will receive goodness and benefaction from a wealthy man." A sufficient number of days later, a second man came and in front of the same people asked him the following question: "I climbed on an apricot tree and was eating its fruit." Sirim said to him that "you will have sorrow and troubles." His companions were surprised that, though the dream was the same, he gave a different interpretation <each time>, and he answered: " When the first man asked, it was the fruit-bearing season, but when the second man asked, it was autumn. This is why the interpretation changed." Upon investigation they found out that things had turned out for both of them the way the dream interpreter had foretold.
Jo See Kister, "Interpretation of Dreams," p. I 00.
The Arabic story closest to the one in chapter 199 is quoted in a little dreambook called Ta bir al-ru'ya attributed to 'Umar al-Khayyam:
واتاى رجل الى ابن سيرين فقال: ر إيت كأن في يدي رمانة أكل منها، فقال تصيب ولدا، وتصيب خيرا من جهته. فإتى رجل اللى أبن سيرين فقال: رأيت أكل الرمان في حينه، قال: وما عملك، قال: انا تاجر، قال له: اطرح مالك كله في التجارة، ففعل فأصاب خيرا. فراي بعد ذلك أنه ياكل الرمان فى غير حينه، ففعل كما فعل اولا ولم يسال ابن سيرين، فذهب ماله كله، فسال ابن سـيـرين عن ذلك فـقـال شـتـّان بـين أن تأكله فى غـيـر حيـنـه وبـين أن تأكله فى حسنه .
A man came to Ibn Sirin and said: "I dreamt that I had a pomegranate in my hand and that I was eating from it." Ibn Sirin said: "You will beget a son, and you will receive goodness from him." Another man came to Ibn Sirin and said: " I dreamt that I was eating pomegranates in season." Ibn Sirin said "What do you do <for a living>?" He said, "I am a merchant." Ibn Sirin said to him "Put all your money in your commerce." The man did this and made profit. After that. he dreamt that he was eating pomegranates out of season, and did as he had done the first time and did not consult Ibn Sirin, and lost all his money. He asked Ibn Sirin about that, and he said: "What a difference between eating it out of season and eating it in season!"
Analogous examples from the Arabic sources could easily be multiplied.34
Chapters 19 and 20 of the Oneirocriticon (Drex1 15, 18-16, 10) tell us that the length and density of body hair in a dream was interpreted as indicating wealth; 32 the same interpretation is illustrated with similar stories recounted
<sup>31</sup> Two of many examples are Ibn al-Musayyab apud Ibn Sa'd (Fahd, La divination arabe, p. 311, no. 10); al-Dīnawarī, fasl 19, bāh 26, Esad Efendi 1833, fol. 187b.
<sup>32</sup> Chap. 19 (Drex) 15, 17-24): Έλθών τις ανθρωπος ήρώτησέ μοι τω ΄Αχμείται το νίώ Σηρείμ, τω όνειροκρίτη του πρωτοσυμβούλου Μαμοῦν: είδον ἐν όράματι, ὅτι αί τρίχες τῶν σκελών μου έδασύνθησαν και ηύξήνθησαν, και ταύτας τη γαλίδι έκούρευον. καὶ ἀπεκρίθην αὐτώ ότι τὸ λογάριον σου καὶ ὁ πλοῦτός σου ἐπληθύνθησαν: καὶ ὅσον ἔκοψας τῶν τριχῶν σου, τοσούτον κακοδιοικείς αυτά. και εύρέθη ούτως το πραγμα (A certain man came and consulted me, Achmet, the son of Sereim, the dream interpreter of the caliph Mamoun: "I saw in my dream that the hair on my legs grew longer and thicker and that I was cutting it with scissors." I replied to him: "Your wealth and riches have multiplied, but you mismanage it by however much of your hair you cut off." Indeed, things turned out to be so). Chap. 20 (Drex1 16, 1-10):" Avθρωπός τις τών μεγιστάγων του πρωτοσυμβούλου Μαμούν είδε κατ΄ ζναρ, ότι έχρίσατο καθ΄ όλου του σώματος και του μεν σώματος όλου αί τρίχες απέλυσαν, αί δε των κρυπτών αύτου τρίχες μαλλον ηύξήνθησαν. και άπέστειλεν ανθρωπον αύτου οικειοποιησάμενον το πραγμα αναγγείλαι τω όνειροκρίτη και είπεν αύτο μέγας ἐθεάσατο, οὐχι δὲ σύ. άλλ' εἰς κίνδυνον έλεύσεται και ού καταλειφθήσεται αύτω έκ του πλούτου αύτου πλήν αί γυναικες αύτου. και μετ΄ όλίγας ήμέρας ἐγένετο, καθώς εἶπεν (One of caliph Mamoun's noblemen saw in his dream that he anointed his entire body and his bodily hair fell off, except for his pubic
by al-Dinawari. 33 In chapter 176 (Drexl 137, 22-29) drinking up all the water in the Tigris river signifies the death of the dreamer. 34 No similar narrative can be found in the Arabic dreambooks, but al-Dinawari says that the drying up of the Euphrates (not the Tigris) in a dream foretells the death of the caliph.35
Chapter 96 (Drexl 57, 11) repeats that the penis is to be interpreted as a man's progeny and suggests that being deprived of it signifies the death of the dreamer. 36 Though no parallel Arabic narrative could be found, the interpretation of the penis as progeny is shared by all five Arabic dreambooks investigated, and three among them, al-Dinawari, *al-Muntakhab* and lbn Shahin, also state that being deprived of it indicates death. 37 No Arabic narrative could be found
33 Al-Dinawarl,fa.)"/ 6, *bah* 32, *Esad Efendi* 1833, fol. 47a-b, BN *arahe* 2745, fol. *76a;fa.)"l* 6, *bah* 45, *Esad Efendi* 1833, fol. 51 b. BN *arahe* 2745, fol. *78b;fa.)"l* 6, *hiih* 73, *Esad Efendi* 1833, fol. 55a, BN *arahe* 2745, fol. 86b.
34 0iKct(l)craµ£v6<; n<; i:ov *'U5wv* 8£cr7t6i:T]v 7tctpal;(l)v TjA.0£ Kat T]pwi:T]cr£ i:ov 6vnpoKpi i:T]v LT]p£tµ AEY(l)V" doov Ka1: <sup>0</sup>ovap, on £mov 7tavi:a i:ov TiypT] 7toi:aµov. 0 OE E<l>TJ' *wuw* cru OUK dlic<;" aliUvmov yap. Kat ciJµoA.6yT]cr£V' i:in OU1:(1)<; EX£l' i':yciJ µEv OUK dliov, au· 6 m~µljla<; µ£ 0£0"7t01:T]<;. Kat ct7t£V 6 LT]pciµ- i:in a7tcA.0ciJv £Upiicrct<; aui:ov 0avm(l)0£vi:a. Kat Ka0w<; ct7t£V, oui:(l)<; Kat i':yi':vci:o (Someone wishing to test the dream interpreter Sereim pretended that he was his master and consulted him saying: "I saw in my dream that I drank all of the river Tigris." Sereim said: "You did not dream this, because it is impossible." He confessed that this was indeed the case: "I did not dream this, but my master who sent me." Sereim said: "You will leave here and will find him slain." Things happened as he said).
35 Al-Dinawari,fa.)"/ 17, *bah* 5, *Esad Efendi* 1833, fol. 166a.
36'Av0p(l)7tO<; n<; v£oyaµo<; £A.0cbv T]pwi:T]cr£ i:ov 6v£tpoKpii:T]v LT]pciµ wu Maµouv· dliov Kaff u7tvou<;, i:in a7t£K07tTJ £~ i':µou i:i':A.rnv 6 KauA.6<;. 6 OE ct7t£v· i:ii:£ *wuw* £0£acr(l), Jtoia Tiv i:wv wpwv; Kat a7t£Kpi0T] i:in EKi:T] Tjv. 6 OE ct7t£V' alto 1:0V1:T]<; i:f]<; ropa<; apieµT]O"OV µf]va<; E~ Kat n\i EKi:qi µT]vt i:£A£miicrct<; ai:£KVO<;. Kat i:£A.£t(l)0£vw<; wu EKWU µT]VO<; *£y£v£w* oui:(l)<; (A newly wedded man came and consulted Sereim, the dream interpreter of Mamoun: "I saw in my sleep that my penis was completely severed from me." The dream interpreter said: "Count six months since the time that you had the dream. By the sixth month you will die childless." When the sixth month expired it happened thus).
37 Ibn Qutayba, *bah* 14, *Yahuda ar.* 196, fol. 31a: ..J\_, ~ *,;-S:J* I.:.,... (..,...;.. ~ jS\_, (Anything that goes out of the penis is a son); al-Dinawari, *fa.)"16, bah* 126, *Esad Efendi* 1833, fol. 69a; BN *arahe* 2745, fol. 98b: ... ~ oJl:.J ~ '--::'-' oJl:.JJ • ..w\_,\_, ...,....L..:JI ~ ..\_\_.~\_, .\_;;;:i ~.)I \_;£.J....Q ..::...Lo.....:..... .:,L; .\_..::...:..~I o\_,;-S:i \_:,I (.SI\_; .:,Ll (The penis signifies a man's reputation and honor among the people, as well as his son. An increase of his penis signifies increase of those things .... If someone dreamt that his penis was cut until it was completely severed from him, he will die). The same interpretations are repeated in *a!-Muntakhah,* p. 95; Ibn Shahin, no.1237; al-Nabulusi, vol. l, p. 224, s.v. \_;;;:i.
hair which grew even longer. The nobleman sent one of his men to narrate it to the dream interpreter, pretending that he had dreamt it himself. The dream interpreter said: "A nobleman had this dream, not you. He will be in danger and nothing will be left to him except for his women.'" A few days later things happened the way the dream interpreter had foretold).
to parallel the contents of chapter 264 (Drex1 217, 1-7).38
Contrary to the innumerable interpretations recorded in the Oneirocriticon, the thirteen Greek anecdotes constitute a finite body of material that was possible to compare exhaustively with eight Arabic sources" and extract some statistics regarding the degree to which we can follow the translation (or adaptation) of the Greek interpretations from the Arabic: barely half the anecdotes (six-chapters 36, 139,144, 147, 194 and 199-out of thirteen) could be matched up with an Arabic narrative. One of the six (chapter 139) was found in a source other than the five Arabic dreambooks, suggesting that the number of identifiable Arabic passages corresponding to interpretations from the Oneirocriticon could increase, if the Greek text is compared with still other Arabic dreambooks. Anecdotes in five chapters (19, 20, 36, 96, and 176) have only vague correspondences to the contents of the Arabic dreambooks investigated. Two of the Greek narratives (chapters 153 and 164) cannot be matched at all. The comparison of the Oneirocriticon with Arabic dream interpretation is in some cases eye-opening and in others frustrating. Inevitably, some problems raised by the Greek text can never be answered.
<sup>39</sup> Γυνὴ ύπανδρος ἐλθοῦσα ήρώτησε τὸν ἀνειροκρίτην λέγουσα: εἶδον, ὅτι περιεπάτουν ασκέπαστος έν μέσω λαου. τι έσται έν εμοί; ό δέ φησιν έχεις άνδρα; και είπε ναι, έχω, αλλ' έν ξενία έστίν. και άπεκρίθη ότι ου θεάση αύτόν. και καθώς είπεν, ούτως και ἐγένετο καὶ ἀπέθανεν ὁ ἀνήρ αὐτῆς ἐν τῆ ξενιτεία (A married woman came and consulted the dream interpreter saying: "I saw that I was walking unveiled in public. What will happen to me?" He asked, "Do you have a husband?" She said, "Yes, I do, but he is abroad." He answered: "You will not see him <again>." Indeed things happened as he said, because her husband died abroad).
<sup>39</sup> In addition to the five Arabic dreambooks regularly cited I also investigated the narratives from the dreambooks attributed to Ibn-Sirin, Ta bir al-ru'ya and to 'Umar al-Khayyam, Ta'hir al-manam. Due to the nature of the material, it was also possible to use the anecdotes concerning Ibn al-Musayyab from the biographical dictionary of Ibn Sa'd (French trans, by Fahd, La divination arabe, pp. 310-12).
#### CHAPTER TEN
#### THE ONEIROCRITICON AND BYZANTINE INTELLECTUAL ACTIVITY IN THE NINTH AND TENTH CENTURIES
The Oneirocriticon is a Christian adaptation of Islamic dream interpretation, probably based on more than one Arabic dreambook. Though at times Islamic concepts are changed into Christian ones with a certain creativity, the author generally seems to have followed his Arabic sources closely. We can even go so far as to characterize the Oneirocriticon as a nonliteral translation from the Arabic. Byzantine literature is generally considered as unreceptive to texts originally written in other languages, which is an understandable phenomenon for a culture endowed with a long and powerful literary tradition of its own. By the tenth century, Greek literature was already eighteen centuries old and possessed a body of texts in all fields of human experience to which it could turn in order to renew itself from within.
The Oneirocriticon was composed during a period of intense intellectual activity which modern scholars have called "the Macedonian Renaissance" and which has been viewed as a Byzantine return to classical and late-antique models. To what degree does the transfer of the Oneirocriticon from Arabic into Greek in the course of the tenth century constitute an exceptional literary event? This question would best be answered if we first try to understand to what field of knowledge the Oneirocriticon belonged. The kinds of texts with which it is copied, when it does not constitute a volume by itself, is revealing: of the sixteen manuscripts in which the Oneirocriticon appears together with other texts, one (Zagora 89) combines it with the Physiologos, a kind of Christian zoology. Six (Paris. Suppl. gr. 690; Vat. gr. 573; Ambros. gr. 592 [O 94 sup]; Vindob. philos. et philol. gr. 162; Athos, Iviron 4285.165) contain a wide variety of subjects and authors: Paris. Suppl. gr. 690 includes excerpts from ninety-four different works: a group of gnomologia, questions on physics and physiology by Alexander of Aphrodisias, excerpts from the Alexander
<sup>1</sup> Only seven of the twenty-three manuscripts of the Oneirocriticon known to me contain this work alone: Paris. gr. 2538; Paris. gr. 2427; Borbon. gr. 356 (III.E.34); Berol. gr. 171 (Phil. gr. 1575); Leidens. Voss. 49; Vindob. philos. et philol. gr. 297; Vindob. philos. et philol. gr. 111.
Romance, poetry by pagans, such as Homer *(Batrachomyomachia)* and Phocylides, as well as Christians, like George of Pisidia, John Geometres, Symeon Metaphrastes, Christopher of Mytilene, Psellos, and a series of epigrams by late antique and Byzantine authors; excerpts from the Old Testament; patristic excerpts from the works of John of Damascus, Saint Athanasios, John Chrysostom, and Maximus the Confessor; a group of dreambooks; Lucian, Aesop, *Life of Aesop,* a legend on the building of Hagia Sophia and the fabulous lives of the prophets by St. Epiphanius.
*Vat. gr.* 573 was put together from pages that previously belonged to seven different manuscripts, which means that today the *Oneirocriticon* belongs to an even larger collection of texts than it originally did. Still, the original manuscript to which the *Oneirocriticon* belonged2 is a collection of texts with both religious and secular subjects: liturgical and patristic texts by authors such as John Chrysostom, Gregory of Nyssa, Saint Athanasios, and Gregory of Nazianzus co-exist with instructions on the correct format for letters written at the patriarchal chancery (no. 13), poems by Manuel Philes (nos. 16 and 23), chronology (no. 25), and excerpts from the agricultural compilation titled *Geoponica* (no. 26). As for the field of knowledge pertinent to prognostication, *Vat. gr.* 573 contains the *Oneirocriticon,* a lunar dreambook (no. 28) and a text on Chaldean oracles by Psellos (no. 24). *Ambros. gr.* 592 (0 94 sup) combines a collection of court poetry by the twelfth-century author Theodore Prodromos with excerpts from the alphabetical dream book of Pseudo-Nikephoros (here ascribed to Gregory of Nazianzus) and excerpts from the *Oneirocriticon.* This combination of texts is somewhat similar to what we have seen in *Paris. Suppl. gr.* 690 and *Vat. gr.* 573, though *Ambros. gr.* 592 is not as extensive. In the remaining two codices the *Oneirocriticon* is combined with religious texts: *Vindob. philos. et philol. gr.* 162 begins with the Akathist Hymn and continues with the *Oneirocriticon* and a collection of prophecies about Constantinople (a combination also found in *Vat. gr.* 573); in *Athas, lviron* 4285.165 the *Oneirocriticon* and the Alexander Romance (together also in *Paris. Suppl. gr.* 690) are sandwiched between biblical excerpts and commentaries, an arrangement possibly meant to disguise the works in the middle of the volume. This impression is strengthened by the layout of the pages in which the *Oneirocriticon* appears: part of the text is written around the margins, exactly in the way biblical commentaries are usually copied around
<sup>2</sup>Part V of the manuscript as it is bound now, fols. 46-214, text nos. 11-28 in the description of the manuscript by Devreesse, *Codices Vaticani Graeci,* vol. 2, pp. 469-77.
#### CHAPTER TEN
the text of the Bible on which they comment. BL Additicius 8240 includes, besides the Oneirocriticon, a rhetorical description (ekphrasis), two gnomologia, a lunar dreambook, the dreambook of Nikephoros, an exposition of the doctrinal differences between the Greek and the Latin churches in the form of a canon, and a treatise on ancient Greek meter.
The combination of texts found in these manuscripts seems at first glance arbitrary, but there might be an underlying logic, dictated by how the compiler or the scribe understood the definition of the sciences of Hermes Trismegistus.3 The manuscripts could therefore be understood as anthologies of a greater or smaller number of Hermaic sciences, in which the Christian excerpts should possibly be counted as philosophy.4 The majority, nine of the sixteen manuscripts, combine it with texts on the exact and occult sciences, such as alchemy, astronomy, astrology, various forms of divination, and medicine: Marc. gr. 299 with texts mostly on alchemy; Bononiensis (Bibl. Univ.) 3632 with texts on medicine, astrology, magic and divination; Petropolitanus Bibl. Acad. scient. graec. 161 with various texts on divination; Vindob. philos. et philol. gr. 287 with a collection of astrological texts; Cantabrig. (Trinity College) gr. 1386 with Actuarius's De Urinis; Paris. gr. 2419 with a collection of texts on astrology and divination; Hierosol. (of the Patriarchate) gr. 220 with a collection of texts on divination; Hierosol. (of St. Sabbas) gr. 555 with three brief brontologia; Paris. gr. 2511 with a collection of astrological texts. Though to a
5 The connection of medieval medicine with the occult sciences is well established, in spite of the difficulties a modern scholar might have in comprehending it. In the words of J. Scarborough, "Introduction [to the papers from the symposium on Byzantine Medicine]," DOP 38 (1984), pp. ix-x, "Although the modern physician will immediately reject any notion that magic and astrology might be useful in medical practice, many of our sources
<sup>3</sup> See Mutinensis (Bibliot. Estense di Modena) gr. 85 (III, C, 6), in CCAG, vol. 4, p. 117: αί Έρμαϊκαι έπιστημαι και τέχναι ήγουν προφητείαι, μαντείαι, φιλοσοφία, άστρονομία, γραμματικὴ, ἀριθμητικὴ, γεωμετρία καὶ τὰ τούτοις παραπλήσια (The Hermaic sciences and arts, that is oracles, divinations, philosophy, astronomy, grammar, rhetoric, arithmetic, geometry and similar things). For the source of this definition, see ibid., pp. 113-14.
<sup>4</sup> See F. Dölger, "Zur Bedeutung von φιλόσοφος and φιλοσοφία in byzantinischer Zeit," Byzanz und die europäische Staatenwelt (Ettal, 1953), pp. 127-208; also I. Sev\_ cenko, "The Definition of Philosophy in the Life of Saint Constantine," in For Roman Jakobson: Essays on the Occasion of His Sixtieth Birthday, 11 October, 1956, ed. M. Halle et al. (The Hague, 1956), pp. 449-57; rpt. in idem, Byzantium and the Slavs (Cambridge, Mass. and Naples, 1991), pp. 93-106. Ambros. gr. 592 contains only part of the Hermaic sciences, rhetoric (court poetry) and divination (dream interpretation).
modern mind dream interpretation belongs to the realm of superstition, to a medieval mind it belonged to the realm of science. 6
The history of Byzantine science is one of the least researched topics by contemporary scholars. The manuscript tradition of the relevant texts is complicated and the technical knowledge required of any philologist attempting to prepare a critical edition of a scientific work is considerable. It is not surprising that the connection between Byzantine and Arabic science has received less scholarly attention than it deserves. Despite the limited number of publications that can be cited, however, a case can be made that the *Oneirocriticon* was only one of several Arabic scientific texts made available to Greek readers in the course of the ninth and the tenth centuries. A number of Greek texts on empirical science that were written before the year 1000 are either avowed translations from Arabic originals or indirectly indicate earlier Byzantine contact with Arabic science. The following are some aspects of Byzantine science that should be further investigated.
6 This is also evident if we examine the Western European milieu that was interested in dreams and dreambooks in early modem times. Pierre Vattier (d. 1667), the earliest translator of an Arabic dreambook directly into a Western European vernacular language ('Abd al-Ral\_lman b. Na~r. *L'Onirocrite Mussulman, ou la doctrine et /'interpretation des songes selon /es arabes, par Gabdorrhachaman fits de Nasar. De la traduction de P. Vattier* ... *sur le manuscrit arabe du "recueil de ce qui se peut dire de meilleur sur !'interpretation des songes"* [Paris, 1664 ]), was the physician of the duke of Orleans who had learnt Arabic in order to read the works of Avicenna in the original. For the interest of other European physicians in dream interpretation, including Caspar Peucer (1525-1602), professor of mathematics and medicine in Wittenberg, the Italian physician, mathematician, physicist and philosopher Girolamo Cardano (1501-76), the Swiss physician Johann Jacob Huggelin (fl. 1560), and the physician Janus Comarius (1500-1558) who translated into Latin Artemidoros and several Greek medical authors, see Fahd, "Oniromancie orientale," pp. 347-74.
for Byzantine medicine show a strong influence of both, ranging from the religious medical cures of pilgrims' tokens ... to a continued employment of the medical astrology of the quasi-mythical Hermes Trismegistus and the firm belief in demonology. One should not, therefore, be surprised that one of the greatest Byzantine medical practitioners, Alexander of Tralles, sanctioned magic, particularly when the patient's belief aided a cure. Moreover, the student of Byzantine medicine must not attempt to impose a "modem view" that would excise the so-called non-rational elements of medical practice, since this would wrest medicine from its matrix, thereby warping the conclusions to suit modem preconceptions." Outside of Byzantium, the connection of medieval medicine with astrology and divination is evident, for example, in the contents of the medical and para-medical fragments in the Cambridge Geniza collection; see H. D. Isaacs, *Medical and Para-Medical Manuscripts in the Cambridge Genizah Collection* (Cambridge, 1995).
#### 396 CHAPTER TEN
*Astronomy-Astrology.* The investigation of the relationship between Byzantine and Arabic astronomy is more advanced than other areas, thanks to the publication of the *Catalogus Codicum Astrologorum Graecorum* during the first half of this century, and to the more recent work of Otto Neugebauer and his students (the most notable among whom is David Pingree) in the United States and of Joseph Mogenet and Anne Tihon in Belgium. The majority of Greek manuscripts that preserve texts on the empirical sciences belong to the late- or post-Byzantine periods. The translations from Arabic are therefore difficult to date, and the only *termini* that can be established are sometimes the *floruit* of the Arab author and the date of the Greek manuscript, and these can be several centuries apart.
Compared with the other empirical sciences, Byzantine astronomy and astrology have the advantage that inferences about the time and place in which a text was written, copied, or used can be made, based on the computations or examples of observations incorporated in the text or in the scholia of the Greek astronomical and astrological manuscripts. Historians of Byzantine science date the available evidence very conservatively. They agree that by the eleventh century Byzantine astronomers and astrologers had access to a large pool of Arabic texts, at least some of which had been adapted into Greek. <sup>7</sup>But a series of scholia by a thirteenth-century hand in a ninth-century copy of the Greek *Almagest (Vat. gr.* 1594), as well as other material from later Greek manuscripts, are based on calculations and observations by astronomers knowledgeable in Arabic astronomy that are datable to the ninth and tenth centuries. 8
The historians of science who published these scholia are reluctant to draw any conclusions at this point, 9 or are inclined to believe that their data are derived not from direct observations by Byzantine astronomers but from translations of Arabic originals. <sup>10</sup>However, this evidence was interpreted
<sup>7</sup> See A. Tihon, "Tables islamiques a Byzance," *Byzantion* 60 (1990), rpt. in eadem, *Etudes d'astronomie byzantine* (London,1994), no. 4, p. 415;
<sup>8</sup> For the relevant bibliography, see Magdalino, "Road to Baghdad," p. 209, nn. 50 and 51.
<sup>9</sup> See A. Tihon, "Le calcul de la longitude de Venus d'apres un texte anonyme du *Vat. gr.* 184," *Bulletin de l'institut historique beige de Rome* 39 (1968), rpt. in eadem, *Etudes d'astronomie byzantine,* no. 2, p. 82; eadem, "Le calcul de la longitude des planetes d'apres un texte anonyme du *Vat. gr.* 184," *Bulletin de l'institut historique beige de Rome* 52 (1982), rpt. in eadem, *Etudes d'astronomie byzantine,* no. 3, p. 26.
<sup>10</sup> J. Mogenet, "Sur quelques scolies de I' Almageste," in *le monde grec: Hommage a*
differently and correlated to the larger picture of the Arab-Byzantine intellectual contacts, not by a historian of science but by a historian. In a carefully built and entirely convincing discussion, Paul Magdalino notes that the datable material gleaned from these scholia comes from either 829-30 or 906-7, years that coincide with the embassies of John the Grammarian and Leo Choirosphaktes to the caliphate. At least one of the scholia pertains to an observation made in Damascus in the year 829 when al-Ma'miin lived there, which could be the reason why Byzantine ambassadors visited him there and not in Baghdad. They are dated according to the Byzantine emperor's regnal year, which points to a Byzantine source.
It is possible that the Byzantines were brought into contact with Arabic astrology without the mediation of a translation, through the works of Theophilos of Edessa <sup>11</sup>(ca. 695-785), a Maronite Christian historian and astrologer to Caliph al-Mahdi. <sup>12</sup>There does not seem to be a consensus as to whether Theophilos wrote in Greek, in Syriac, or in Arabic, 13 though it is certain that he knew Greek, since he translated Homer into Syriac. Many more Greek manuscripts contain the astrological works of Theophilos than Arabic or Syriac ones, because there is hardly a Greek astrological manuscript that does not excerpt a chapter or two from his works. 14 Whatever the original language of the treatises surviving in the Greek manuscripts, Theophilos was a non-Chalcedonian Christian living in Muslim lands, whose works seem to have
13 Breydy, *Geschichte der syro-arabischen Literatur der Maroniten,* p. 135, implies that Theophilos's astrological works were written in Arabic and Syriac and then translated into Greek, but he does not give any reasons for such a statement.
<sup>14</sup>Cumont in *CCAG,* vol. 5, p. 231; repeated by Breydy, *Geschichte der syro-arahischen Literatur der Maroniten,* p. 135 and n. 20.
*Claire Preaux,* ed. J. Bingen, et al. (Brussels, 1975), pp. 301-11.
<sup>11</sup> In Arabic, Taufil ibn Tuma or Tha'ufil ibn Thuma, sometimes erroneously written Naufil.
*<sup>12</sup>*On the life and works of Theophilos, see M. Breydy, *Geschichte der syro-arahischen Literatur der Maroniten vom VII. bis XVI. Jahrhundert* (Opladen, 1985), pp. 132-38; D. Pingree, "The Indian and Pseudo-Indian Passages in Greek and Latin Astronomical and Astrological Texts," *Viator* 7 (1976), pp. 148-49, enumerated three works by Theophilos that survive almost entirely in Greek and announced that he is preparing an edition of both the Greek and the Arabic works. See also idem, Review of W. Gundel and H.G. Gundel, *Astrologumena. Die astrologische Literatur in der Antike und ihre Geschichte, Gnomon* 40 (1968), p. 279. On the importance ofTheophilos for Byzantine astronomy, see idem, "Classical and Byzantine Astrology", pp. 236-37.
#### 398 CHAPTER TEN
circulated in the Greek-speaking world as early as the year 792. <sup>15</sup>Pingree attributes Byzantium's acquaintance with Arabic astronomy to the mediation of Stephen the Philosopher, a student of Theophilos who arrived in Constantinople from the caliphate at around that time. He believes that Stephen was the one who introduced the works ofTheophilos ofEdessa to the Greek-speaking world. <sup>16</sup>
Besides the practical application of Arabic astronomical and astrological principles by the Byzantines, the Arabic texts translated into Greek earlier than the year 1000 indicate that knowledge of Arabic astronomy and astrology did not reach the Byzantines solely through oral instruction possibly occasioned by diplomatic encounters, but also through texts. Based on the combined evidence of two manuscripts, *Laurent. Plut.* 28, 34 (copied around 1000) and *Vindob. phi!. gr.* 115 (early 13th century), Pingree asserted that their direct common ancestor was copied around the middle of the tenth century. Material from the *Vindobonensis* which does not appear in the *Laurentianus,* but which is in the part of the *Vindobonensis* that is closely associated with the *Laurentianus* (and is therefore likely to have existed in their common direct ancestor, the aforementioned mid-tenth-century manuscript), includes excerpts translated from the works of Abu Ma'shar (787-886), whom his contemporaries and
<sup>15</sup> To prove knowledge of Theophilos in Byzantium by this date Pingree adduced the following example ("Classical and Byzantine Astrology," p. 239): "That Theophilos' s work was known in Byzantium in the 790s ... is indicated by an incident recorded in Theophanes's *Chronographia* under the year 6284 (A.o. 792). In July of that year the emperor, Constantine VI, marched against the Bulgars. He built a fortress, Marcellae, on the border. On the twentieth of the month Kardam, the Bulgarian ruler, led his army across the frontier up to the fortifications. The emperor, being advised by his "pseudoprophet and astrologer," Pancratius, that victory would be his, sallied forth to ignominous defeat. Theophanes does not describe the astrological technique used by Pancratius, but it can be guessed from chapter 20 of Theophilos's IT6vot, where we read: "The Moon in Gemini with the aspect of the benefits indicates the unsuccessful withdrawal of the besieging troops" (Constantine was within the fortress, which was surrounded by the Bulgars). At about noon on 20 July 792 the Moon was in the twentieth degree of Gemini aspected in sextile by the benefic planet, Venus, which was in the twentieth degree of Leo."
<sup>16</sup>Pingree, "Classical and Byzantine Astrology," pp. 238-39; see also idem, *From Astral Omens to Astrology,* p. 64. A different interpretation of the data regarding Stephen the Philosopher is given in Dagron, "Les Diseurs d'evenements. Reflexions sur un 'theme astrologique' byzantin," *Melanges Georges Duby,* vol. 4, pp. 57-65; idem, "Formes et fonctions du luralisme linguistique a Byzance," p. 235; for a discussion of both views, see Magdalino, "Road to Baghdad," pp. 210-12.
later generations considered the father of Arabic astrology .<sup>17</sup>
According to Pingree, there was a massive transmission of Arabic texts on genethlialogical, catarchic and interrogational astrology in Greek translations that were made around the year 1000. The most recent Arabic author translated in this group of texts is 'AH b. Al.imad al-'Imrani ( d. 955). 18 In addition, there is ample securely dated evidence that throughout the eleventh century Byzantine astronomers knew the existence of and used several Arabic astronomical works originally written in the ninth and tenth centuries. The translation of these texts from Arabic into Greek could have been made at any time between their writing and the date of their attested use in Byzantium. By 1032 someone in Constantinople was informed of the *neoteroi* (the more recent ones), that is the astronomers of Caliph al-Ma'miin, and was able to use the tables of lbn al-A 'lam (d. 985). <sup>19</sup>The works of at least two more Arab astronomers of the ninth and tenth centuries were utilized in the composition of a Byzantine manual written in Constantinople in the years l 060-72: 20 the commentary of
<sup>18</sup>See Pingree, *From Astral Omens to Astrology,* p. 71.
<sup>19</sup>See J. Mogenet, "Une scolie inedite du *Vat. gr.* 1594 sur les rapports entre I'astronomie arabe et Byzance," *Osiris* 14 (1962), 198-221; also idem, "L'influence de I'astronomie arabe a Byzance du IXe au XIVe siecle," *Colloques d'Histoire des Sciences I ( 1972) et II ( 1973)* (Louvain, 1976), pp. 45-55 and Tihon, "Tables islamiques a Byzance," p. 402. On Ibn al-A'lam as mathematician and astronomer, see Sezgin, *GAS,* vol. 5, p. 309 and *GAS,* vol. 6, p. 215.
<sup>20</sup>*Methodoi psephophorias diaphor6n hypothese6n astronomikon: An Eleventh-Century Manual of Araho-Byzantine Astronomy,* ed. and trans. A. Jones (Amsterdam, 1987). According to Jones, the anonymous *Methods of Computing Various Astronomical Hypotheses* "is one of several documents that attest to a knowledge of Arabic astronomical writings among the Byzantine astronomers of the eleventh and twelfth centuries, and
<sup>17</sup>Pingree, *From Astral Omens to Astrology,* pp. 64-65. Abu Ma'shar's *Kittih ft abkiim tabtiwll sinl al-mawtilid,* the Greek title of which is postulated as nEpt iii<; 1wv £1wv £vaUayf]<; (On the Revolution of Nativities) survives in two manuscripts that were copied in the 14th and 15th centuries. Pingree, in his edition of the text, dates the translation to around the year 1000; see Abu Ma'shar, *Alhumasaris: De revolutionihus nativitatum,* ed. Pingree (Leipzig, 1968), p. viii. This work of Abu Ma'shar was later translated from Greek into Latin, as was the case with the *Oneirocriticon* (ibid., p. vi.). See also Pingree's article in *DSB,* s.v. "Abu Ma'shar al-Balkhi, Ja'far ibn Mu~ammad." *De revolutionihus* is number 19 in Pingree's enumeration of Abu Ma'shar's works. According to Pingree, three other works of Abu Ma'shar (or at least excerpts thereof) were also translated into Greek along with *De revolutionihus.* They survive in a compilation entitled Ta MucrTfipta wu' Anoµcicrap *(The Mysteries of Apomasar);* see Pingree, "Classical and Byzantine Astrology," p. 227, n. 2.
Ibn al-Muthanna (10th century) to al-Khwarizmi (ca. 840)21 and the tables of I:Iabash al-I:Iasib (9th century). 22 The Byzantine astrological compendium surviving in *Vat. gr.* 1056 (14th century) names about twenty Arab and Jewish astrologers. 23 The compilation itself was produced in the twelfth century,24 but the evidence gleaned from the dated horoscopes contained in it point to the eleventh century, and possibly the years around 1000, for the production of the Greek translations from the Arabic.25 This is as far back as the manuscript evidence allows one to go, but it is possible that these translations were made, not around 1000, but a few decades earlier, since at least one of the astrologers excerpted in *Vat. gr.1056,* Sahl b. Bishr (9th century)26 is said by a tenth-century Arabic source to have been highly esteemed in Byzantium,27 which suggests that one or more of his works are likely to have been translated into Greek by that time.
*Alchemy:* Research on Greek alchemy is just beginning, and not all the Greek alchemical manuscripts have even been published.28 The alchemical writings
21 Ibid., p. 12, n. 4.
22 Ibid., p. 13. I:Iabash al-I:Iasib was an astronomer in Baghdad in the time of al-Ma'miin and al-Mu'ta~im. When he died, around 912, he was approximately one hundred years old; see Sezgin, *GAS,* vol. 6, p. 173.
23 The compilation is described in *CCAG,* vol. 5:3, pp. 7-64. Further comments in Pingree, *From Astral Omens to Astrology,* pp. 68-71.
24 For the date of the compilation, see Tihon, "Tables islamiques a Byzance," pp. 405-13.
25 Pingree, *From Astral Omens to Astrology,* p. 71.
<sup>26</sup>See Sezgin, *GAS,* vol. 7, pp. 125-28; for an edition of his Greek excerpts, see *CCAG,* vol. 5:3, pp. 98-107.
27 lbn al-Nadim, *Fihrist,* trans. Dodge, vol. 2, pp. 651-52. See also Dagron, "Formes et fonctions du pluralisme linguistique a Byzance," p. 235.
28 An inventory of the Greek alchemical manuscripts can be found in the eight volumes of the *CMAG* (Brussels, 1924-39). These are the product of a collaborative effort directed by J. Bidez. The next phase would have been to publish all the extant Greek texts on alchemy. The project was interrupted by World War II and was never carried out. The
stands out among them for its size and for the advanced level of some of its chapters. Other texts indicate the transmission from Arabic sources of a few numerical parameters and simple trigonometric and spherical functions: here alone we find proof that a Constantinopolitan of the eleventh century had access to, and knew how to use, tables for solar, lunar and planetary motion derived from more than one Arabic astronomical treatise" (ibid., p. 7).
available to the larger scholarly community are those published by M. Berthelot,29 complemented to some degree by the texts that appeared in the series of the *Catalogue des manuscrits alchimiques grecs.* In the *Oxford Dictionary of Byzantium,* the entry on alchemy concludes:
Unlike astronomy, astrology, mathematics and medicine, Byzantine alchemy seems barely to have been enriched by translations from the Arabic, though there are some traces of eastern influence in the treatise from the fourteenth-century codex *Holkham gr.* 290, now in the Bodleian,30 and in the work of Kosmas. 31 The few treatises translated from the Latin texts influenced by the Arabic science were available only in Italy.
This statement has to be modified, since *Marc. gr.* 299, the oldest Greek alchemical codex that survives, copied in the tenth century ,32 contains two alchemical recipes on the tempering of copper and "Indian iron" that refer to some of the ingredients by their Arabic names, 33 indicating that at least elements
29 M. Berthelot, *Collection des anciens alchimistes grecs,* 3 vols. (Paris 1887-88).
30 Ed. 0. Lagercrantz, *CMAG,* vol. 3.
31 Kosmas the monk must have lived later than the I Ith century, since he postdates Psellos; see Berthelot, *Collection des anciens alchimistes grecs,* vol. 3, pp. 442 ff. (VI.xvi).
32 On dating *Marc. gr.* 299 to the 10th century, see chapter 3, n. 50.
33 Published in Berthelot, *Collection des anciens alchimistes grecs,* vol. 3, pp. 346-48. In the Greek text and French translation by Berthelot these terms are the following: p. 346, 10: 1ca't A.apwv arro TOU <j>OlVlK01ta<HlAAOU 'WU £pu8pou TOU A.cyoµ£vou vaTI']<j> *ev,* Apapou; (Puis prenant de la couleur de palmier, je dis du rouge appele *natef* chez !es Arabes). Oriental (Arabic?) alchemy is mentioned on p. 346, 18-19: Kat 01£ 8£A.tjcrrn; pa\j/m xaAKOV clVclHEpov OU KpEinmv OU parr1£ml *ev* 0Epcri8t (lorsque tu voudras teindre le cuivre precite, ainsi qu' on ne teint pas mieux en Perse ... ); p. 347, 11-15: Kat A.aPmv <j>AOlOV <j>OlVtKoPaA.avou TOU A.cyoµ£vou l':A.tA.Ey *ev* 'Apapou;, crm8µov µt' tE', Kat cr1a8µov µt" 8' PEAlAEY oµoimi; K£Ka8apµ£vou *ano* 10JV £n6i;, tjwt '"GOV <j>A.otov µ6vov, Kat aµpA.ay oµoimi; K£Ka8apµ£vou µt' 8', (Puis prenant de l'ecorce des fruits de palmier, nommee *elileg* chez !es Arabes, 15 parties en poids, et 4 parties en poids de *helileg,* pareillement nettoye a l'interieur, c' est-a-dire l'ecorce seule; ainsi que 4 parties *d'amhileg,* semblablement nettoye). For an analysis of the Greek and the Arabic terms, see ibid., p.
effort has recently been resumed under the direction of H. D. Saffrey and the patronage of the Union Academique Internationale. It is anticipated that the entire Greek alchemical literature will be published in twelve volumes in the Collection des Universites de France. To the best of my knowledge, only three volumes have so far appeared: R. Halleux, *Les alchimistes grecs I: Papyrus de Leyde, Papyrus de Stockholm. Fragments de recettes.* (Paris, 1981 ); M. Mertens, *Les alchimistes grecs* 4: *I: Memoires authentiques. Zosime de Panopolis* (Paris, 1995); and A. Coli net, *Les alchimistes grecs* I 0: *L'anonyme de Zuretti* (Paris, 2000).
of Arabic alchemy were known in Byzantium as early as the tenth century. <sup>34</sup>
Though the evidence of *Marc. gr.* 299 is so far the earliest indication of Byzantine know ledge of Arabic alchemy, two alchemical recipes that survive in Greek manuscripts of the thirteenth century and later indicate Byzantine contact with Arabic alchemy earlier than the fourteenth century. The first one is the short treatise, µ£8o8oi; 8t' ~i; anotEAEttat fj mjlatpoi::t8fii; xaA,al;a, Kataoxi::uacr8i::1cm napa tou £v ti::xvoupyi~ ni::pt~ofiwu 'Apa~oi; tou I:aA,µava (Method through which the sphere-shaped hail is brought to perfection, fabricated by the Arab Salmanas, famous in <the field of alchemy> ), 35 which explains how to make artificial pearls. P. Kraus identified the "famous Arab Salmanas" as Salm or Salim al-I:Iarrani, the director of the House of Wisdom36 in Baghdad during the reign of al-Ma'mun (813-33). <sup>37</sup>
35 Berthelot, *Collection des anciens alchimistes grecs,* vol. 3, pp. 364-67. The Greek text survives in at least five manuscripts, the earliest of which, *Paris. gr.* 2325, dates from the 13th century.
36 This is a famous library used by intellectuals frequenting the Abbasid court; for a discussion of its exact function, see D. Gutas, *Greek Thought, Arabic Culture. The Graeco-Arabic Translation Movement in Baghdad and Early Abbasid Society (2nd-4th/8th-10th centuries)* (London and New York, 1998), pp. 53-60.
37 P. Kraus, "Zu Ibn al-Muqaffa' ," *Rivista degli Studi Orientali* 14: 1 (1934), p. 11; rpt. in idem, *Alchemie, Ketzerei, Apokryphen imfriihen Islam. Gesammelte Aufsiitze* (Hildesheim, Zurich, New York, 1994 ), pp. 89-108. At least one Arabic alchemical *risiila* by al-Salim al-I:Iarrani survives; see Ullmann, *Die Natur- und Geheimwissenschaften im Islam,* pp. 216-17. The Arabic text has never been compared with the Greek.
<sup>331,</sup> n.1; ibid. p. 332, nn. 2-4.
<sup>34</sup> In addition, the term 8ou8ia (translated by Berthelot as "tutie" and explained as "la partie qui s'est sublimee a la partie superieure du fourneau: c'est surtout de l'oxyde du zinc") , which is used at the beginning of the first recipe, does not occur in any of the ancient alchemical texts (see ibid., pp. 330-31). The two formulas quoted in *Marc. gr.* 299, at least in their present form, must have been written later than the 7th century, though they purport to have been written "in the reign of Philip" (ypa<!Jctcra *ano* aPXii~ <I>tA.innou), to which some of the later manuscripts add "of the Macedonians" ( 1:0ii nov MaKEOOVON). This could possibly be Philip III Arrhidaios, who, upon the death of Alexander the Great in Babylon (323 B.c.) was elected regent of Alexander's empire. Either the core of the recipes dates to that time, or the claim is false. At least the second recipe admits to an oriental origin, as it ends with the following statement: afrn1 £cr1t v Ti npwn1 Kat PacrtAlKTJ E:pyacria, fiv E7tlTI]0£UoVWl cri]µcpov, £1; ~~Kat 'tel 8auµacrw si<!Jri 1:£K1:aivov1at. Hup£8ri 8£ lino nov' Iv8oov, Kat £1;£068ri TI£pcrm~, Kat nap· EK£ivwv ~A.8cv Et~ i]µii~ (Such is the first and royal operation, which they practice nowadays, by means of which the marvelous swords are also made. It was discovered by the Indians and was passed on to the Persians, and through them it came to us).
The *terminus post quern* for the Greek text attributed to him-evidently a translation from Arabic into Greek-is the early ninth century, which is the *floruit* of Salim al-I:Iarrani. It is possible that this recipe was translated from the Arabic together with a recipe for making artificial emeralds attributed to "the wise Ishmaelites" ( oi cro<jlot 't&v' Icrµarilcu&v ). 38 Even if the thirteenthcentury *terminus ante quern* for these two recipes is a late one, at least the evidence of the tenth-century *Marc. gr.* 299 indicates the possibility of Byzantium's acquaintance with Arabic alchemy by that time. 39 The problem becomes all the more interesting, if we consider that, during the ninth and tenth centuries, the embassies of each state tried to impress the others by various means, including the demonstration of alchemical procedures.40
All four recipes that indicate Byzantine contacts with Arabic alchemy are clearcut and follow a practical approach to their topic.41 Given that a number of alchemical formulas had mainly allegorical significance and virtually no practical application,42 we should keep in mind that the instructions given in the Greek formulas that indicate knowledge of Arabic alchemy stand out by their clarity.
39 On the connections between Arabic and Byzantine alchemy, suggesting the translation of texts from Arabic into Greek, see P. Kraus, *Jabir ibn ljayyan. Contribution a /'histoire des idees scientifiques dans /'Islam. Jabir et la science grecque.* Memoires presentes a l 'Institut d'Egypte 45 (Cairo, 1942) rpt. Paris, 1986, pp. 38-39. For a Byzantine alchemical text of unknown date (before 14 78) and its connection with Arabic alchemy, see A. Colinet, "Le *Travail des quatre elements* OU lorsqu'un alchimiste byzantin s'inspire de Jabir," in *Occident et Proche-Orient: Contacts scientifiques au temps des Croisades. Acres du colloque de Louvain-la-Neuve, 24 et 25 mars 1997,* ed. I. Draelants, A. Tihon, B. van den Abeele (Turnhout, 2000), pp. 165-190.
40 'Umara b. Hamza, secretary to Caliph al-Man~iir, was sent as ambassador to Constantinople during the reign of Constantine V (r. 741-75), where he saw the emperor change copper into silver and gold by means of a dry powder. The story is recorded by Ibn al-Faqlh al-Hamadhiinl, *Kitab al-buldan,* ed. de Goeje (Leiden, 1967), trans. by H. Masse, *Abrege du livre des pays* (Damascus, 1973), pp. 138-39; see also El Cheikh Saliba, "Byzantium Viewed by the Arabs," p. 59; further comments and bibliography in Gutas, *Greek Thought, Arabic Culture,* pp. 115-16.
<sup>41</sup>About the formula of Salmanas, Berthelot, *Collection des anciens alchimistes grecs,* vol. 3, p. 349, n. I, observes that it is purely technical and contains neither references to ancient authors, nor charlatan phrases, nor mystic pretensions.
42 See J. Ruska, *Turba philosophorum. Ein Beitrag zur Geschichte der Alchemie* (Berlin, 1931 ), pp. 279-87.
<sup>38</sup> Berthelot, *Collection des anciens alchimistes grecs,* vol. 3, p. 358, 25; the formula appears in the same manuscripts as the formula of Salmanas.
*Meteorology.* According to V. Christides, the naval rivalry between Byzantium and the Arabs reached its height between the ninth and the eleventh century and led to the extensive reorganization of the Byzantine and the Arab fleets. <sup>43</sup> Christides believes that an increased interest in meteorology developed at this time because of its relevance to navigation, and three anonymous texts from *Marc. gr.* 335 (fols. 420r-423v)44 can in fact be dated to the tenth century. <sup>45</sup> Direct textual influence from a comparable work in Arabic cannot be proved, but at least two of the three texts were clearly written from the geographical perspective, if not of the Arab lands, at least of the Byzantine Empire's eastern frontiers: they mention the preferred naval itineraries of the Mardaites, a group of Arabic-speaking Christians from the mountains of Lebanon,46 and of the Saracens when they sail from Egypt to Syria and along the Syrian coast, from Antioch to Tarsus.47 The last of the three texts refers to Antioch as *Antiocheia* but states that it is called *Entakeion* by the Arabs.48
A text on the prediction of the weather that was written prior to the eleventh century seems to have been influenced by Oriental wisdom. It is called ITepi
44 Published by S. Lampros, "Tria keimena symballonta eis ten historian tou nautikou para tois Byzantinois," *Neos Hellenomnemon* 9 (1912), pp. 162-77; Lampros (p. 162) says the manuscript is *Marc. f?r.* 135. The last two texts published by Lampros also appeared in *CCAG,* vol. 2, pp. 70 and 214. Comments on these texts can be found in V. Christides, *Conquest of Crete,* pp. 10-11. English translation and introduction in R. H. Dolley, "Meteorology in the Byzantine Navy," *The Mariner's Mirror* 37 (1951), pp. 5-16.
45 See also Dolley, "Meteorology in the Byzantine Navy," pp. 5-10.
46 Christi des, *Conquest of Crete,* p. 11; on the Mardaites, see also H. Antoniadis-Bibicou, *Etudes d'histoire maritime de Byzance; apropos du "theme des Caravisiens"* (Paris, 1966), pp. 29-33.
47 Text r, Lampros, "Tria keimena," pp. 175, 19-176, 3.
48 Lampros, "Tria keimena," p. 176, 32-35: apµEvil;ouow £~ Aiyunwu µ£XPt tou Kacrtpou tfi~ A VttoXEia~, 07t£p Ol LapaKT]VOl , EvtaKEtoV KUAOUcrt, toU votou.' *Ano* OE tOU Kacrtpou tfi~· AvttoXEia~, µEXPt tfi~ Tapcrou apµEvil;oum toU KOA7t0\J Kat 1:0U ~oppa ("They sail from Egypt with a southerly wind as far as the Castle of Antioch, which the Saracens call Entakeion. But from the Castle of Antioch, they sail with a northerly wind to the Gulf of Tarsos," trans. Dolley, "Meteorology in the Byzantine Navy," p. 15).
<sup>43</sup> V. Christides, *The Conquest of Crete by the Arabs (ca. 824). A Turning Point in the Struggle Between Byzantium and Islam* (Athens, 1984), p. 37; for an overview of the Byzantine and Arab naval rivalry in the course of the 11th century, see also W. Felix, *Byzanz und die islamische Welt im fruheren 11. Jahrhundert. Geschichte der politischen Beziehunf?en van 1001his1055* (Vienna, 1981).
U£1:c0V 061nµov *ano* <jlcovfj.; 'Lupou nvo.;, which could roughly be translated asA *Trustworthy [Treatise] on Rain, Based on the Lecture Notes of the Teaching by a Syrian* (?). 49 The text survives in at least three manuscripts, the earliest of which dates from the eleventh century. <sup>50</sup>
<sup>50</sup>*Laurent. Plut.* 28, 34 of the 11th century, *CCAG,* vol. I, p. 63; *Laurent. Plut.* 28, 13 of the 14th century, *CCAG* I, p. 10; *Taurinensis* C.VII.10 (B. VI.12) of the 14th century, *CCAG* 4, p. 7. The text is edited in *CCAG* I, p. 131. The question of what the provenance of this text might be is interesting, because the oldest manuscript that contains it, *Laurent. Plut.* 28, 34, is one of the most important pieces of evidence for the early influx of Arabic astrology into Byzantium. A direct ancestor of *Laurent. Plut.* 28, 34 was copied in the middle of the I 0th century; see the remarks by Pingree, *From Astral Omens to Astrology,* p. 65. One could possibly add to this list another text that seems to have been translated from Arabic into Greek, though its date is uncertain and the only *terminus ante quern* that can be established is the date of the manuscripts in which it survives, the earliest of which were copied in the 15th century; for the manuscripts, see *CCAG,* vol. 4, pp. 145-46; *CCAG,* vol. 12, p. 113; *CCAG,* vol. 4, p. 33; *CCAG,* vol. 7, p. 75. The text is a calendar that outlines the days in a year when the configuration of the stars will produce bad weather. It is attributed to "the Persian philosopher by the name Zanati," the foremost master of geomancy, another form of divination. He might have hailed from northwestern Africa, because in the geomantic treatises attributed to him the names of the figures are sometimes given in both Berber and Arabic, and the epithet Zanati seems to refer to the name of one of the Berber tribes that still reside in North Africa. The dates of his life are unknown, and the only *terminus ante quern* that we have is the year 1266, when a geomantic treatise attributed to him was translated into Greek. Al-Zanati's geomantic
<sup>49</sup> The editor of the text, A. Olivieri, believed its author was the same as the author of the text titled Ka0oAtK~ rrapayyi::A.ia rri::p't Km:acr'tTjµci'twv· LUpou (CCAG, vol. I, p. 171 ), which appears later in *Laurent. Plut.* 28, 34 (fol. 144r). Olivieri deduced that "Syros" is not an ethnic designation (Syrian), but a personal name, though he added that the identification of the author with Syrus to whom Ptolemy's *Tetrabiblos* is dedicated is pure conjecture (CCAG, vol. I, pp. 131-32, n. I). Between the 6th and 8th centuries the expression *apo phlines x* (lit., "from the voice of x") in the title of a work meant that it was the product of notes taken down by someone as x was orally explaining his methods; from the 9th century on, it meant "by x"; see M. Richard, "'Ano <j>wviji;," *Byzantion* 20 (1950), pp. 191-222. Deciding whether *apo phcines Syrou tinos* means "by a Syrian" or "according to the oral instruction of a Syrian" depends on the date one assigns to the composition of the text. Though the noun 86Ktµov in the title CTi::p't UE'tOOV 86Ktµov alto <j>wviji; LUpou nvoi; (according to other manuscripts "ooKiµwv"; see *CCAG,* vol. I, p. 132) does not occur in Greek dictionaries, in our case it should be understood as a method tested by the person who dictated it, as the text itself says: f:yw 8£ f:v rri::ip~ 1'.'.crxov ("I have experienced"; p. 132, 20-21 ), orri::p ouv f:yw f:ooKiµacra ("which I have tested"; p. 134, 7). Indeed, the adjective OOKtµoi; has, among other meanings, that of "approved by test, tried and true" (see W. Bauer, *A Greek-English lexicon of the New Testament,* s.v. "86Ktµoi;"; also Demetrakos, s.v. "86Ktµoi;").
*Medicine.* The earliest influences of Arabic medicine on contemporary Byzantine medicine can be detected in the eleventh century. In the current state of research, the works of Symeon Seth, who lived during the reign of Alexios Komnenos, are considered to be the earliest Greek medical texts that mention Arabic and Hindu spices and drugs. <sup>51</sup>The Greek translation of a work by Mul)ammad b. Zakariyya' al-Razi (d. 925), namely *Peri loimikes* (On Measles), 52 is considered to be a product of the eleventh century. The translation of a second work by al-Razi, the *Zad al-musafir,* in Greek *Ta ephodia tou apodemountos* (Supplies of a Traveler) was made (presumably by Constantine of Rhegion in Italy, also known as Constantine the African), toward the end of the eleventh century. 53
There are indications, however, that Byzantine doctors may have known something about contemporary Arabic medicine even earlier. 54 At least one medical treatise, Mercurius's *De pulsibusdoctrina,* is said to have been translated
<sup>51</sup>Scarborough, "Introduction [to the papers from the Symposium on Byzantine Medicine]," pp. xiii-xiv.
52 Sezgin, *GAS,* vol. 3, p. 276. Published in *Alexandri Tralliani Medici lib. XII, Rhazae de pestilentia libellus ex Syrorum lingua in Graecam translatus, Jacobi Goupyli in eosdem castigationes* (Paris, 1548).
53 See A. Kousis, "Quelques considerations sur Jes traductions en grec des oeuvres medicales orientales et principalement sur les deux manuscrits de la traduction d'un traite persan par Constantin Melitiniotis," *Akademia Athenon, Praktika* 14 ( 1939), pp. 205-20.
54 An indication that elements of Arabic medicine had been introduced into the Byzantine world earlier than the works of Symeon Seth is furnished by the *Oneirocriticon* itself. The concoction called *zoulapi* is generally thought to have been introduced into Byzantine medicine through Symeon Seth towards the end of the I Ith century; see G. Harig, "Von den arabischen Quellen des Simeon Seths," *Medizinhistorisches Journal* 2:3-4 ( 1967), p. 260. The presence of this word in the *Oneirocriticon* which, on the basis of its manuscript tradition, was undoubtedly put together before the middle of the 11th century, indicates that *zoulapi* was introduced to Byzantine medicine even earlier, especially because its Hellenized form is already stable and because in the *Oneirocriticon* it is talked about as if it is a well-known item that requires no further explanation.
treatise survives in a number of Greek manuscripts, including *Neapol.* 11.C.33, olim 34 (see *CCAG,* vol. 4, p. 51), which also contains al-Zanati's meteorological text. According to P. Kunitzsch, "Die "Unwettersterne" und die "Geomantie" des Zanati," *BZ* 60 ( 1967), pp. 309-17, the names of the stars in the Greek treatise attributed to al-Zanati are neither Arabic nor Berber. Kunitzsch deduced that this astronomical calendar of the weather had been translated from Syriac into Arabic and then into Greek and that the names of the stars must originally have been Syriac, while their rendering into Arabic and then into Greek as loan words made them unrecognizable.
from Arabic into Greek in the tenth century. 55 J. Sonderkamp, who studied the medical compilation written by the so-called Theophanes Nonnos56 and dedicated to Constantine VII Porphyrogennetos, concedes that Arabic medical works were being translated into Greek by the end of the tenth century. <sup>57</sup> However, as H. Hunger put it, our extremely fragmentary knowledge of medical texts that were translated from Syriac, Arabic, and Latin stands in the way of understanding the history of medicine in Byzantium.58
*Agriculture.* Pingree suggested that the fragments of the agricultural tradition attributed to Zoroaster in the tenth-century compilation known as the *Geoponika* could have originated in Byzantine and not late-antique translations of Oriental sources. 59 Pingree does not argue this point at length.
Two more categories of texts influenced by their Arabic counterparts also deserve to be mentioned here, though the evidence for dating them, or their parts, as early as the tenth century is at present lacking:
*Geomancy.* This form of divination was known in Byzantium by the name pciµnA.wv *(ramblion)* or pa~6A.tov *(rabolion),* a loan word reflecting the
<sup>55</sup> See Hunger, *Die hochsprachliche profane Literatur der Byzantiner,* vol. 2, p. 303, n. 64; Greek translation with updated bibliography in Hunger, *Byzantine logotechnia,* vol. 3, p. 130. Mercurius's text was published by S. Cyrillus, *De pulsibus doctrina* (Naples, 1812); I have been unable to locate a copy of this book in the United States.
<sup>56</sup> This Theophanes is the author of the *Introduction to Medicine* that C. Mango ("Greek Culture in Palestine after the Arab Conquest," in *Scritture, lihri e testi nelle aree provinciali di Bisanzio,* ed. G. Cavallo et al., vol. 1 [Spoleto, 1991], p. 160), seemed inclined to accept as a work of Theophanes Graptos, the iconodule saint who was born in the Moabite mountains ca. 778 and came to Constantinople from Palestine in 813. J. Sonderkamp discussed the various authors to whom the work is attributed in its manuscript tradition and chose to accept Theophanes Chrysobalantes (about whom nothing else is known) as the genuine one. See J. Sonderkamp, *Untersuchungen zur Uberlieferung der Schriften des Theophanes Chrysobalantes (sag. Theophanes Nonnos)* (Bonn, 1987), pp. 1-68; for the attribution of the work to "Theophanes Graptos" or "Theophanes the Monk," see ibid., pp. 27, 30, 43, 265; on Theophilos's work, see also T. Miller, *The Birth of the Hospital in the Byzantine Empire* (Baltimore and London, 1997), pp. 177-78.
<sup>57</sup> J. Sonderkamp, "Theophanes Nonnus: Medicine in the Circle of Constantine Porphyrogennitus," *DOP* 38 (1984), p. 40.
<sup>58</sup> Hunger, *Die hochsprachliche profane Literatur der Byzantiner,* vol. 2, p. 303; Greek trans. Hunger, *Byzantine logotechnia,* vol. 3, p. 130.
<sup>59</sup> Pingree, "Classical and Byzantine Astrology," pp. 236-37.
Arabic *ram!* (sand),60 though it is also called -ro A-a~i::mi]pwv wu nueayopa (The stone cutter's tool of Pythagoras). Of the plethora of Byzantine geomantic texts, only one has been published. <sup>61</sup>Some among them appear to have been translated from the Arabic, because they refer to the various geomantic figures by their Arabic names. At present no inventory of the Byzantine geomantic texts is available; C.O. Zuretti had started one before his death in 1931, but it has not been resumed since. 62 P. Tannery's study of Byzantine geomancy and its connection with its Arabic and Latin counterparts was also interrupted by death. <sup>63</sup>
It is unknown when Arabic treatises of geomancy were first translated into Greek. The only established date is 1266, the date for the translation from Arabic into Greek given in the title of a treatise that survives in at least two copies. 64 The beginnings of Arabic geomancy are themselves obscure. Geo-
60 The Arabic name for geomancy is *'ilm al-ram!* (lit., "science of the sand").
<sup>61</sup>A. Delatte and L. Delatte, "Un traite byzantin de geomancie (codex Parisinus 2419)," *Annuaire de l'lnstitut de philologie et d'histoire orientates et slaves* 4 ( 1936), pp. 575-658. This treatise appears to have been based, at least in part, on a Latin source (ibid., p. 589). The beginning pages from a second treatise, which survives in *Paris. gr.* 2424 and *Paris. gr.* 2419 and is based on an Arabic model, is published in P. Tannery, "Le rabolion. Traites de geomancie arabes, grecs et latins," *Memoires Scientifiques* 4 (Toulouse and Paris, 1920), pp. 359-72.
62 A. and L. Delatte were supposed to see Zuretti's project to completion, but confessed that the task was too difficult, due to the obscurity of the subject and the need to collate several manuscripts. The enumeration of the texts collected by Zuretti is to date the only available catalogue of Byzantine texts on geomancy (see A. Delatte and L. Delatte, "Un traite byzantin de geomancie," p. 587, n. 3); see also A. Delatte, *Anecdota Atheniensia,* vol. 1, pp. 388-96 and 557-61.
63 It was published posthumously accompanied by an essay on Arabic geomancy by Carra de Vaux in Tannery, *Memoires Scientifiques,* vol. 4, pp. 297-411.
64 I currently know of *Bononiensis (Bihl. Univ.)* 3632 *(CCAG* 4, p. 43), where the title of the work reads as follows: rruriµa IT£pcrou <l>tA.ocro<j>ou rnuvoµa ZavaTij µET£PA.u0t 8£ Ett; ljµai; Kma To t;\j/08 £Toi; rrapa 'ApcrEviou µovaxou KEAEUcrT\ Tiii; Kupiai; (The work of the Persian philosopher by the name of Zanati was translated into our own language in the year 1266 by the monk Arsenios at the instigation of the lady ... ); and *Neapol.* 11.C.33, olim 34 (CCAG, vol. 4, p. 51), where the title is somewhat more informative: IToiriµa IT£pcrou <j>tA.ocro<j>ou rnu Mai;ouvm:fj· a-uTot; yap E'tqvEucrm:o Kat Km:acro<j>icrm:o TI]v8E TlJV Texvriv 8ta \JHiµµou rrpoi; µna poA.ai; *[sic]* TE Kat rrotOTT\Tat; nov cicrT£pcov rnii; pouA.oµ£voti; A.£yEtv Ta cruµPriaoµEva ciyaea TE rnt <j>auA.a Ka0ooi; µETa µtKpov Ka06A.coi; 8riA.w0TicrE1:m.' Yrroµviicrw 8£ cipTiwi; orrwi; Kat rroTE µETEcrrn1xEtc00ri cirro crrn1xEiwv Kat yA.ooarii; ITEpm8oi; Ett; 'tljv Toov Pwµaiwv ypa<j>ljv Kat 8tciA.EKrnv rrapa 'ApcrEviou µovaxou. 'Ev yap TQl slJI08 (/; should be corrected to t;) ETEt cirro KTtCTEWt; Kocrµou £~ E1tl Tporrlji; Tllt;
mantic procedures are described by lbn al-'Arabi (763-844) and Abu Zayd al-An~ari (d. 830),65 and by the twelfth century the practice was well established in North Africa, Egypt and Syria. The Arabic tradition of geomancy is at least as vast as that of Arabic dream interpretation.66 It is possible that translations of geomantic treatises from Arabic into Greek began earlier than the thirteenth century. No firm conclusions, however, can be drawn before the Greek texts are collated and published. <sup>67</sup>
*Other Forms of Divination, Magic and Secret Alphabets.* Forms of divination both in Greek and in Arabic are virtually uncharted territory. There are striking similarities between texts on apocryphal wisdom in Greek, Arabic, and Latin, and many of them can be attributed to the fact that this domain of human knowledge in all three languages was based on foundations laid down in the Graeco-Roman world of late antiquity that survived throughout the Middle
66 Geomancy as a form of divination is still very popular in the Middle East, and new treatises on the subject were written as late as the beginning of the 20th century. For bibliography on the subject of Arabic geomancy, see *Oxford Encyclopedia of the Modern Islamic World,* s.v. "Geomancy."
67 The study of Byzantine geomancy could profit from further investigation of the Slavonic geomantic treatises called *rafli* which must have been translated from the Greek. For reference to editions of the Slavonic texts and a brief overview of the current state of research regarding Slavonic geomancy, see I. Sevcenko, "Remarks on the Diffusion of Byzantine Scientific and Pseudo-Scientific Literature among the Orthodox Slavs," *Slavonic and East European Review* 59:3 (1981 ), p. 342; rpt. in idem, *Byzantium and the Slavs,* p. 611. For a view tracing the immediate ancestry of Slavic geomancy, not to Byzantium but to western Europe, see R. Mathiesen, "Magic in Slavia Orthodoxa: the Written Tradition," in *Byzantine Magic,* ed. H. Maguire (Washington, D.C., 1995), p. 166.
i::ucri::Prni:rii:rii; Kupiai; E>rnowpai; Ti pipA.oi; aul:T] i:o1i; 'Pwµaio1i; £y£vi::i:o (The work of the Persian philosopher Mazounati. He devised and invented this art through <the use of> the sand and[?] the changes and the qualities of the stars[?] for <the use of> those who would wish to <pre>dict the good and bad things that will happen, as will soon be thoroughly demonstrated. Let me briefly mention how and when it was changed from the Persian letters and language into the Roman writing and tongue by Arsenios the monk. This book reached the Romans in the year 1266 thanks to the patronage of the most pious lady Theodora). The patroness of the translation was Theodora Doukaina, the wife of the future emperor Michael VIII Paleologos; see Pingree, *From Astral Omens to Astrology,* p. 77; also A.-M. Talbot, "The empress Theodora Palaiologina, Wife of Michael VIII," *DOP* 46 (1992), p. 30 I.
<sup>&</sup>quot;'See Fahd, *La divination arahe,* pp. 196-97.
#### 410 CHAPTER TEN
Ages and into the Renaissance. 68 But much of this material was translated and recycled between the three languages, and it is very difficult to trace the exact path trod by each and every element found in the sources on magic and divination, at least in the form in which they survive today.69 The Greek texts on apocryphal wisdom suggest influences from both East and West. <sup>70</sup>The indications of the Greek contact with Near Eastern lore include the (genuine or pretended) Hebrew names of the angels evoked during a divinatory or magical procedure, the use of Hebrew letters for secret alphabets71 and a "Persian" treatise on prognostication by observing pulsation. 72 Some Greek
69 A case illustrating the complicated travels of such works is the literature ascribed to Daniel in Greek, Latin, Arabic and other Near Eastern languages. In the 10th century Liutprand of Cremona remarked that both the Arabs and the Byzantines possessed prophetic texts attributed to Daniel. For references, as well as an analysis on how the specific vision of Daniel summarized by Liutprand was related to political considerations of the moment, see P. Alexander, *The Byzantine Apocalyptic Tradition* (Berkeley, Calif., 1985), pp. 96-122. Another type of text, also attributed to Daniel, this time not a revelation of events that will come to pass but rather an interpretation of celestial omens, was also among the literature recycled by the Byzantines and the Arabs. In the year 1245 a certain Alexios, while held captive by the Arabs as prisoner of war, translated a work on celestial omens from Arabic into Greek. In his preface he traced the history of the translated text as follows: it was originally written in Hebrew (presumably by Daniel), translated into Greek at the time of Ptolemy Philadelphos, and discovered during the reign of Constans II (641-68) by Caliph Mu'awiya (661-80) when he sacked the environs of Constantinople. He had it translated into Arabic, and Alexios translated it from Arabic back into Greek. The historical details given in this preface are accurate, but this does not necessarily prove the authenticity of the text's claimed itinerary. What is important to retain, however, is Alexios's awareness that the text had been recycled. It is published in *CCAG* 8:3, pp. 171-79; see also the remarks in Pingree, *From Astral Omens to Astrology,* p. 76.
70 See, for example, the incantations from *Paris. gr.* 2494 (15th century) that are clearly Greek transliterations of Latin texts, published in A. Delatte, *Anecdota Atheniensia,* vol. 1, pp. 588-89.
<sup>71</sup>Cf. the secret alphabet used in *Paris. gr.* 2419 (15th century, a manuscript copied by Georgios Midiates that also contains the *Oneirocriticon);* it was deciphered in A. Delatte, *Anecdota Atheniensia,* vol. 1, p. 446.
72 The work is called ~i.~A.oi; cro<j>i.ai; CTEpmiiv itaA.µud1; only its beginning survives in *Athen. Bibi. Nat.* 1493 (13th century), fol. 159; published in A. Delatte, *Anecdota Atheniensia,* vol. I, pp. 209-10. In Arabic this method of prognostication is called *'ilm al-ikhtiliij* and seems to have been based, at least in its present literary form, on ancient Greek literature on the subject; see Fahd, *La divination arabe,* pp. 397-402.
<sup>68</sup> Some remarks on the topic in D. A. M. Pielow, *Die Quellen der Weisheit* (Hildesheim, Zurich, New York, 1995), p. 4 ff.
and Arabic secret alphabets disguise the original letters (or numbers that stand for letters) by adding curled edges to their design, which results in a similarity of form between some Greek and Arabic secret characters, though in this case it is impossible to say whether this method traveled from east to west or from west to east. 73 Nor does the problem of dating the influx of Oriental elements into the Greek lore have an obvious solution, other than adopting the *terminus ante quern* established by the age of the relevant manuscripts. Unfortunately for our purposes, all of them are late.
In sum, we have evidence (often circumstantial, it is true) that a few of what we would today call pseudo-scientific texts were translated from Arabic into Greek in the course of the tenth century, though the existence of these translations is better documented for the eleventh century. This raises a number of questions, including the milieu and patronage that produced them. Dagron has discussed the categories of bilingual or multilingual individuals who lived in the Byzantine Empire between the ninth and the twelfth century, including speakers of Arabic, Latin, Slavic and Armenian. 74 They were missionaries, soldiers,75 frontiermen, prisoners of war, intellectuals and ordinary people. Indeed, as far as tenthcentury Constantinople is concerned, the legislative text known as the *Book of the Eparch* (published by the emperor Leo VI in 911-12) mentions a permanent colony in Constantinople of Syrian merchants specializing in textiles.76 The sixth book of the chronicle of Theophanes Continuatus, dealing in part with
75 For the terms and conditions under which Arab captives were settled in Byzantine territories, see [Constantine VII], *De cerimoniis,* ed. Reiske, pp. 694, 22-695, 14. English translation and commentary on this excerpt in McGeer, *Sowing the Dragon's Teeth,* pp. 365-67.
<sup>76</sup>*Book of the Eparch* §5.2:' H EtcrEpxoµ£vri rcpayµmEia 'tOOV rcpavlhorcpmoov o<jlEiAEl rcacra arcoi:i9Ecr9m EV i::v't OlK(j) tOOV µttcl'tCOV, cOO'tE rcavmi; Emcruvayoµ£voui; ErctµEpi1;Ecr9m tll'lJ'tTIV. oµoicoi; Kat lj arco LUpta<; EPXOµEvTI I:apaKTIVlKi] ... Kat ErctCTTI<; rcavta<; yivEcr9m Kai. ErctµEpit;rnem µEta Kai. 'tOOV arco I:upiai; oi.Kricravtcov Kai. OEKaE'tfj xp6vov EV ti] ~acrtlcEuoucr11 owvucravi:cov (The imported merchandise of the *prandiopratai* [textile merchants] ought to be deposited in one of the lodgings <where foreign merchants stay>, so that all can congregate and have a share in it. Likewise the Saracen merchandise coming from Syria ... should be equally divided <among the *prandiopratai>,* including those who came from Syria and have resided for ten years in Constantinople).
<sup>73</sup> Compare the secret alphabets from *Vindoh. phi/. gr.* 108, fols. 369v-370r (published in Delatte, *Anecdota Atheniensia,* vol. 1, p. 637), and especially the alphabet on lines 9-10 from the top, with the Arabic secret alphabet published by Ullmann, *Die Natur- und Geheimwissenschaften im Islam,* p. 3. On this type of secret alphabet, see also Pielow, *Die Quellen der Weisheit,* pp. 156-58.
<sup>74</sup> G. Dagron, "Formes et fonctions du pluralisme linguistique," pp. 219-40.
the reign of Leo VI, mentions in passing the presence of Saracens at the Praetorium(Bonn 372, 23-373, 2). Samonas, one of the most powerful ministers of Leo VI, was an Arab from Tarsus who had been brought to Constantinople as a captive. 77 As for intellectuals, most of Dagron's examples from the ninth to the twelfth century are proficient in Greek and Arabic. <sup>78</sup>
The prosopography of the Arabic-speaking intellectuals active in the Middle Byzantine period is very difficult to establish. The narrative sources are mostly silent on the subject, but unexpected data can be retrieved from other types of sources, including archaeological evidence. Given the kinds of texts translated, one would imagine the translators could have been astrologers or physicians, so turning our attention to the tools of their trade could provide some clues. The only surviving Byzantine astrolabe was made in 1062 for the *protospatharios* and *hypatos* (both terms being Byzantine court titles) Sergios, "from the race of the Persians." 79 An intriguing Middle Byzantine silver stamp has a Greek inscription declaring its owner to be an *iatros* (physician) named Ishmael.80
For the ninth century, we have the name of only one translator: St. Michael
79 For a description of this astrolabe, a full quotation of the two inscriptions identifying its owner and for further bibliography, see the catalogue of the exhibition *Byzantine Art: European Art* (Athens, 1964), no. 549. Interestingly, the star list used by the maker of this astrolabe is the same as the one in the last datable item in the astrological writings contained in *Laurent. Plut.* 28, 34 of the early 11th century, a list of astrolabe stars whose coordinates are given for the year 908; see Pingree, *From Astral Omens to Astrology,* p. 65, n. 7. As was noted earlier, the *Laurentianus* is one of the most important surviving Greek manuscripts attesting to the early influx of elements of Arabic astrology into Byzantium.
80 G. Vikan, "Art, Medicine and Magic in Early Byzantium," *DOP* 38 (1984), p. 65: "The silver doctor's stamp is unpublished, and in the Limbourg collection, Cologne. It is mid-Byzantine in date and bears the following inscription: "Lord, help Ishmael [the] doctor."
<sup>77</sup> For a summary of the diplomatic relations and a chronicle of the recorded exchanges of prisoners, see H. Kennedy, "Byzantine-Arab Diplomacy in the Near East from the Islamic Conquests to the Mid Eleventh Century," in *Byzantine Diplomacy. Papers from the Twenty-fourth Spring Symposium on Byzantine Studies, Cambridge, March 1990,* ed. J. Shepard and S. Franklin (London, 1992), pp. 133-44.
<sup>78</sup> To the intellectuals proficient in Greek and Arabic, one could add Cyril and Methodios, the Apostles of the Slavs, who were proficient in Greek and Slavic; Dagron discusses them under the separate heading "le missionaire" (Dagron, "Formes et fonctions du pluralisme linguistique a Byzance," pp. 223-28). One could further add the Italian intellectuals active in the Komnenian court, especially under Manuel I. For their names and an outline of their careers, see Haskins, *Studies in the History of Mediaeval Science,* pp. 194-222.
the Synkellos (ca. 760-846), who translated at least one work of Abu Qurra from Arabic into Greek. <sup>81</sup>He subsequently sought refuge in Constantinople to avoid the Muslim hostilities and the sack of Christian monasteries in the anarchy that followed the death ofHariin al-Rashid (r. 809-13). 82 A second individual is the anonymous translator of the Qur'an, whose Greek version of the Muslim holy book is quoted in the polemical treatise by Niketas of Byzantium, written during the reign of either Michael III (r. 842-67) or Basil I (r. 867-86).83 The known translations of both are limited to religious texts.
Crediting Christians under Muslim rule with contributing to the corpus of the early translations of Arabic scientific texts into Greek implies that Greek language and culture survived, at least among Melkite Christians, well after the Islamic conquest; in fact modern scholarship has not reached a consensus on the point. Greek was used as the language of administration until the end of the seventh century, which suggests that an elite of bureaucrats fluent in Greek existed up to two generations after the Islamic conquest. 84 John of Damascus (ca. 675-ca. 754), a member of this elite and the greatest Eastern systematizer of Christian dogma, wrote all of his works in Greek. 85 A number of scholars point out that when Caliph 'Abd al-Malik (r. 684-705) made Arabic the language
83 See K. Versteegh, "Greek Translations of the Qur'an in Christian Polemics (9th century A.D.)," *ZDMG* 141 (1991 ), pp. 52-68 (with bibliography). It is unknown whether a Greek translation of the full text of the Qur'an was available by the 9th century. Niketas discusses in detail only the first 18 of the 114 Quranic chapters. However, Versteegh (pp. 59-60) argues that Niketas seems to have had a Greek translation of the full text of the Qur'an at his disposal.
84 The exact date when the change from Greek to Arabic was introduced into chancery use is uncertain. A 10th-century Arabic source, the *Fihrist* of lbn al-Nadlm, dates the demise of Greek as the administrative language in Syria to the reign of either 'Abd al-Malik (r. 684-705), or his son, Hisham (r. 724-43); trans. in Rosenthal, *Classical Heritage in Islam,* pp. 45-50; see also the lbn al-Nadlm, *Fihrist,* trans. Dodge, vol. 2, pp. 581 ff.
85 See D. Sahas, *John of Damascus on Islam, the "Heresy of the /shmaelites"* (Leiden, 1972), pp. 17-48.
<sup>81</sup>S. Griffith, "Stephen of Ramlah and the Christian Kerygma in Arabic in Ninth Century Palestine," *Journal of' Ecclesiastical History* 36 (1985), pp. 34 ff; rpt. in idem, *Arabic Christianity in the Monasteries of Ninth-Century Palestine* (London, 1992), no. 7.
<sup>82</sup> S. Vailhe, "Saint Michel le syncelle et les deux freres graptoi, saint Theodore et saint Theophane," *Revue de !'Orient Chretien* 9 (1901), pp. 313-32, 610-42; I. - Sev, cenko, "Hagiography of the Iconoclast Period," in *Iconoclasm,* ed. A. Bryer and J. Herrin (Birmingham, 1977).
#### 414 CHAPTER TEN
of bureaucracy, the know ledge of Greek seriously declined, even among Melkite Christians. The emergence of Christian literature in Arabic towards the end of the eighth century indicates that by that time the Christian flock of Syria and Palestine had begun to be more comfortable writing and reading Arabic than Greek.
The literary career of Theodore Abu Qurra, the earliest Christian author writing in Arabic whose name is known to us, began around 772, and at least one recent scholar believes that he relied on the services of a translator for his Greek works. 86 Evidence gleaned from hagiographical texts suggests that, toward the end of the eighth century, the Greek language, in spite of its cultural prestige in the Melkite circles, was no longer understood. 87 On the other hand, Mango has discussed a number of examples that indicate the continuing vitality of Greek culture in Palestine during the eighth and early ninth centuries and argues that the Byzantine humanism of the ninth and tenth centuries was spurred by, among other things, the presence of Palestinian immigrants in Constantinople.88 In his concluding remarks Mango suggests that "the Byzantine revival was not a purely internal flowering; it was stimulated by a drawing in of resources from the periphery as also happened in the contemporary Carolingian Renaissance." 89 Although translations of Arabic scientific texts into Greek in that period can be seen in the context of that revival, Palestine
88 Mango, "Greek Culture in Palestine after the Arab Conquest," pp. 149-60. For the presence of Palestinian immigrants in Constantinople before the 10th century, see Sevcenko, "Hagiography of the Iconoclast Period"; Vailhe, "Saint Michel le syncelle et Jes deux freres graptoi"; J. Featherstone, "Theophanes of Caesarea, Encomium of Theodore Graptos," *Analecta Bollandiana* 98 ( 1980), pp. 93-150; F. Halkin, "Saint Antoine le jeune et Petronas le vainquer des arabes en 863," *Analecta Bollandiana* 62 ( 1944), pp. 187-225; J. Gouillard, "Un 'quartier' d'emigres palestiniens a Constantinople au IXe siecle?" *Revue des Etudes Sud-Est Europeenes* 7 ( 1969), pp. 73-76; M. F. Auzepy, "De la Palestine a Constantinople (VIIle-IXe siecles): Etienne le Saba'ite et Jean Damascene," *TM* 12 (1994), pp. 183-218.
<sup>86</sup> For the date of Abu Qurra's literary career, see S. Griffith, "Greek into Arabic: Life and Letters in the Monasteries of Palestine in the Ninth Century: the Example of the *Summa Theologiae Arabica," Byzantion* 56 (1986), p. 124; rpt. in idem, *Arabic Christianity in the Monasteries of Ninth-Century Palestine,* no. 8. For the opinion that he did not write anything in Greek, see idem, "Stephen of Ramlah," p. 34 ff.
<sup>87</sup> See I. Sevcenko, "Constantinople Viewed from the Eastern Provinces in the Middle Byzantine Period," *Eucharisterion: Essays Presented to Omeljan Pritsak* = *Harvard Ukrainian Studies* 3/4, pt. 2 (Cambridge, Mass., 1979-80), pp. 735 ff.; rpt. in idem, *Ideology, Letters and Culture in the Byzantine World* (London, 1982), no. 6.
<sup>89</sup> Mango, "Greek Culture in Palestine after the Arab Conquest," p. 160.
was not the only peripheral region of the Byzantine Empire to be in contact with Muslim culture. The Arabs first penetrated southern Italy and Sicily in 827, when the forces of the Tunisian Aghlabids set foot on the island. The conquest of Sicily from the Byzantines was completed in 878, and it remained under the rule first of Aghlabid and later of Fatimid governors (the Aghlabids were subdued by the Fatimids in 909), until its conquest by the Normans in the latter part of the eleventh century, an event that also marked the end of Byzantine rule in the area.90
Byzantine culture survived in southern Italy and Sicily well into the thirteenth century. 91 The coexistence of Arabs and Byzantines in the south of Italy resulted not only in Byzantine speakers of Arabic (and Latin) but also Arab speakers of Greek,92 as well as a fusion of the Greek and Arabic medical tradition in the activity and the writings produced by the school of Salerno, the existence of which is attested at least as early as the tenth century. <sup>93</sup>
Contact between Byzantine and Arabic medical science necessitating the mediation of bilingual individuals is also attested for tenth-century Umayyad Spain. Evidence can be found in the famous story about the Dioscorides manuscript recounted by lbn Juljul, a tenth-eleventh-century Spanish pharma-
92 For information on the bilinguals and trilinguals before and after the Norman conquest and their demographical distribution in the area, see V. von Falkenhausen, "I gruppi etnici nel regno di Ruggero II e la loro partecipazione al potere," *Societa, potere e popolo ne/l'eta di Ruggero II,* Atti delle terze giornate normanno-sveve (Bari, 1977), pp. 133-56; and eadem, "Friedrich II. und die Griechen im Konigreich Sizilien," in *Friedrich JI. Tagung des Deutschen Historischen Instituts in Rom im Gedenkjahr 1994.* ed. A. Esch and N. Kamp (Tilbingen, 1996), pp. 235-62.
91 Legend attributes the founding of the school of Salerno to four physicians, a Greek, an Arab, a Jew and a Latin. For an overview regarding the activity of the school of Salerno and further bibliography, see M. Pasca, "The Salerno School of Medicine," *American Journal of Nephrology* 4 ( 1994 ), pp. 478-82. For a summary of the most recent research contradicting the traditional view about the school of Salerno espoused by Pasca, see P. Skinner, *Health and Medicine in Early Medieval Southern Italy* (Leiden, 1997), pp. 127-36. According to this view, several medical texts were copied in the south of Italy very early, but the study and teaching of medicine was not institutionalized before the end of the 11th or the 12th century.
<sup>9</sup>° For a brief overview on the survival of the Greek language and culture in Italy from the 9th until the I Ith century, see P. Lemerle, *Le premier humanisme byzantin* (Paris, 1971); Greek trans. with updated bibliography in idem, *Ho protos hyzantinos houmanismos,* 2nd ed. (Athens, 1985), pp. 25-28.
<sup>91</sup>See G. Cavallo, "La cultura italo-greca nella produzione libraria," *I Bizantini in Italia* (Milan, 1982). pp. 495-612; M. Gigante, "La civilta letteraria," ibid., pp. 613-51.
#### 416 CHAPTER TEN
cologist. 94 In the mid tenth century, the Byzantine emperor95 corresponded with 'Abd al-Ral.iman b. Mul.iammad al-N~ir, the Umayyad ruler of Cordoba, and in 948-49 he sent gifts to Cordoba, which included a Greek illustrated Dioscorides and a Latin history by Orosius.96 Until this time the only available text of Dioscorides in Spain had been a ninth-century translation, where the Arabic rendering of terms such as plant names was problematic. Therefore, a new Dioscorides manuscript provided an opportunity to correct the older translation. The letter of the Byzantine emperor that accompanied the gifts stated that
... one could profit from Dioscurides's book only if there was someone who understood Greek well and knew the individual remedies. "If you have such a man in your country, 0 King, you can properly profit from the book. As far as the work of Orosius is concerned, you have surely in your country Latinists who are able to read Latin, and if you urge them to make the book available, they can translate it for you from Latin into Arabic .... " At that time there was no Spanish Christian in Cordoba who could read *ighriqf,* that is to say, old Greek. Thus the book of Dioscurides remained in Greek as it was, in the treasury of 'Abd al-Rai).miin al-Nii~ir and was not translated into Arabic. The book remained in Spain, but the translation of I~~ifiin, which had come from Baghdad, remained in use. Then, when al-Nii~ir addressed a written reply to King Miirlnus [!], he begged him to send him someone who spoke Greek *[ighriqf]* and Latin in order to teach slaves to translate for him. Thereupon King Armiiniyus sent a monk named Niqulii [Nicholas] to al-Nii~ir. He arrived in Cordoba in the year 340 H. / 951-52 A.o.97
According to Ibn Juljul, the Byzantine monk had an enormous impact on the advancement of pharmacology in Cordoba, because he explained the unknown names of drugs in the work of Dioscorides and was able to show others, among them the pharmacologist Abu 'Abd Allah the Sicilian, "who spoke
<sup>94</sup> The report was translated in Rosenthal, *Classical Heritage in Islam,* pp. 194-97 (with bibliography). See also M. M. Sadek, *The Arabic Materia Medica of Dioscorides* (St-Jean-Chrysostome, Quebec, 1983), p. 9.
<sup>95</sup> The Arabic versions of the Byzantine emperor's name that appear in Ibn Juljul's text correspond to the Greek "Romanos." However, in the years Ibn Juljul places the events (948-52) the emperor in Constantinople was Constantine VII (r. 944-59). Both his predecessor and his successor were named Romanos: Romanos I Lekapenos (r. 920-44) and Romanos II (r. 959-63).
<sup>96</sup> On the Byzantine embassy to the court of 'Abd al-Rai).miin b. Mui).ammad al-Nii~ir, see also F. Di:ilger, *Regesten der Kaiserkunden des Ostromischen Reiches von 595-1453,* vol. 1 (Munich and Berlin, 1924 ), p. 82, no. 657.
<sup>97</sup> Rosenthal, *Classical Heritage in ls/am,* p. 196.
Greek and knew the individual remedies," how to prepare medicines with improved ingredients. Nicholas remained in Cordoba until his death at the beginning of al-J:Iakam al-Mustan~ir's reign (r. 961-76).
lbn Juljul's account makes clear the role science played as a diplomatic tool in the contacts between the two dominant cultures in the Mediterranean at that time. Taken at face value, he also seems to say that there was no one in Constantinople proficient in Arabic in the middle of the tenth century, because the translator of Dioscorides provided by the Byzantine Empire was proficient in Greek and not Arabic but Latin. The narrative indicates that there were people proficient in Latin in Constantinople at that time, but it does not say anything about people proficient in Arabic. In addition, translating the text of Dioscorides requires more than proficiency in the languages involved. Though Ibn Juljul seems to suggest that the Cordobans could not understand Dioscorides because of their ignorance of classical Greek,98 the structure of Dioscorides's sentences is simple and easy for a speaker of medieval Greek to follow. The difficulty in translating Dioscorides lies in knowing the Greek names of the plants current in the first century A.D. and being able to convey them accurately in a second language. It therefore takes not just any bilingual but a bilingual pharmacologist to translate the text. In addition, the Umayyad caliph al-Na~ir specifically requested an individual proficient in Greek and Latin to teach, not high-brow intellectuals, but (presumably Christian) slaves to translate for him. The request of the caliph is important for the history of the transmission of science in Umayyad Spain, but no indication about knowledge of Arabic in tenth-century Constantinople can be extracted from lbn Juljul's report.
An idea about the kind of education that the translators from Arabic into Greek received can be gleaned from the Greek texts they produced. The *Oneirocriticon* avoids Arabic loan words and neologisms in its vocabulary. The astronomical and astrological texts translated before the twelfth century also use a technical vocabulary that is Hellenized and avoids Arabisms, in contrast to translations of later periods. 99 Compared with the Arabic translations of Greek texts, which can be obscure and even incomprehensible without the help of a commentary, the language of the Greek texts translated from the
<sup>98</sup> The Arabic text designates the language of Dioscorides as *"ighrfql.* that is *yunanf."* The term *yunanlyun* in Arabic sources refers to the ancient Greeks as opposed to the Byzantines, who are referred to as *Rum.*
<sup>99</sup>*Methodoi psephophorias,* ed. and trans. Jones. pp. 16-17; Tihon, "Tables islamiques a Byzance," pp. 404 and 417; Pingree, *From Astral Omens to Astrology,* pp. 71-74.
#### 418 CHAPTER TEN
Arabic is relatively smooth and, on the whole, clear. <sup>100</sup>The overall approach to translation was not literal but paraphrastic; or, in other words, the Greek translations from the Arabic were not text-oriented but reader-oriented, a remarkable fact, given that no tradition of translating from Arabic into Greek existed before the ninth century and that rendering a technical text from a Semitic into an Indoeuropean language must have presented the translator with numerous challenges.
Some idea of the difficulties involved can be obtained by looking at the process of translating in the opposite direction, from Greek into Arabic, about which we are much better informed. <sup>101</sup>In this case, problems were often circumvented through the use of a Syriac intermediary. A Syriac translation from the Greek often existed, but even when it did not one could be made with the express purpose of serving as a link between the Greek and the Arabic rendering of a work. Syriac scholars had been translating Greek texts since the fourth century; using a Syriac intermediary allowed the translator to take advantage of the accumulated experience in translating over centuries. Rendering the text from Syriac into Arabic was a much simpler process, given that both are Semitic languages. 102
A reader of the Greek texts known to have been translated from the Arabic before the eleventh century, namely the *Oneirocriticon* and the astrological texts pointed out by Pingree, 103 immediately realizes that scientific translations from Arabic into Greek are rendered in an idiom closer to the spoken Greek of the time, much like the translations of scientific texts from Greek into
<sup>102</sup>See S. Brock, "The Syriac Background to I:Iunayn's Translation Techniques," *Aram* 3: 1-2 (1991 ), pp. 139-62 (with further bibliography); idem, "Greek into Syriac and Syriac into Greek," *Journal of the Syriac Academy* 3 (1977), esp. pp. 3-6 (I \t-. ·n ); rpt. in idem, *Syriac Perspectives on Late Antiquity* (London, 1984). See also G. Strohmaier, "I:Iunain ibn Isl).al5.: An Arab Scholar Translating Into Syriac," *Aram* 3:1-2 (1991), pp. 163-70; rpt. in idem, *Von Demokrit his Dante* (Hildesheim, urich, New York, 1996), pp. 199-206. For references to medieval sources on Arabic translators and translations, see Hitti, *History of the Arabs,* pp. 309-16.
103 Only one of these translations has received critical edition in its entirety: Abu Ma'shar, *Alhumasaris: De revolutionibus nativitatum,* ed. D. Pingree. Excerpts from the rest have been published in various volumes of the *CCAG;* for references, see Pingree, *From Astral Omens to Astrology,* pp. 63-77.
<sup>100</sup>For comments on the approach to translation by translators from Greek into Arabic, see Gutas, *Greek Thought, Arabic Culture,* pp. 142 ff.
<sup>101</sup>Some of the translators from Greek into Arabic were notorious for their poor Arabic style; see Gutas, *Greek Thought, Arabic Culture,* pp. 136-41.
Arabic, which also employ a language closer to the vernacular than to classical Arabic. 104 Given the strong literary tradition of both languages and the "classical" style to which authors in each of them aspired, this use of the vernacular by translators provides food for thought. Although the style of any author reflects his level of education and therefore also his social background, 105 it is possible that in this case both the Greek and the Arabic translations were rendered into a more demotic idiom because they were meant to be utilitarian; practical requirements outweighed literary ambition.
It is too soon to assess the volume and influence of Greek translations from the Arabic. But even allowing for some future finds, it is likely that the total number of Arabic works translated into Greek was far smaller than that of Greek works translated into Arabic. This disparity can be explained in part by the fact that, for a number of political, ideological, and social reasons, the numerous translations from Greek into Arabic were commissioned by the
105 The problem is very complex, even regarding the translators from Greek into Arabic, about whom we are better informed. A number of them were Jewish or Christian, and several contemporary scholars agree that these religious communities wrote in a literary idiom that was closer to the vernacular than to classical Arabic, the literary standard to which Muslim authors aspired. The educational curriculum of a Christian intellectual living in the Arab lands around the 9th-10th century is, for the time being, beyond the grasp of scholars. The major proponent of the view that the various religious communities under Muslim rule used their own literary idiom is J. Blau, *A Grammar of Christian Arabic. Base Mainly on South-Palestinian Texts from the First Millennium* (Louvain, 1966); and idem, *The Emergence and Linguistic Background ofludaeo-Arabic: A Study of the Origins of Middle Arabic,* 2nd ed. (Jerusalem, 1981 ). For an analysis rejecting Blau's conclusion that Christian Arabs had their own literary language, different from that of their contemporary Muslims, see K. Samir, "Existe-t-il une grammaire Arabe chretienne?" *inActes du premier congres international d'hudes arabes chrhiennes. Orientalia Christiana Analecta* 218 (Rome, 1982), pp. 52-59. For the non-Muslim translators and their social and intellectual standing in the caliphate, see H. Kennedy, "The Melkite Church from the Islamic Conquest to the Crusades: Continuity and Adaptation in the Byzantine Legacy," *The Seventeenth International Byzantine Congress: Major Papers* (New Rochelle, NY, 1986), pp. 325-43; A. Fattal, *Le statut legal des non-musulmans en pays d'lslam* (Beirut, 1958); for the Dar al-Rum in the Christian quarter of Baghdad, see G. Le Strange, *Baghdad during the Abhasid Caliphate* (Oxford, 1900), pp. 207-10; especially for the earlier period, see R. Schick, *The Christian Communities of Palestine from Byzantine to Islamic Rule: A Historical and Archeological Study* (Princeton, N. J., 1995); for a synthesis of secondary literature on the topic, see W. Kallfelz, *Nichtmuslimische Untertanen im Islam* (Wiesbaden, 1995).
*w4* See G. Bergstrasser, *lfunain Ibn* ls~ii~ *und seine Schule* (Leiden, 1913), pp. 28 ff.; also Fahd, *Artemidore d'Ephese,* p. xviii.
Abbasid royalty and the aristocracy, as well as literate classes in Baghdadi society, and were paid for with enormous funds, both public and private, set aside for the purpose. ' 06 Byzantine translations of Arabic works into Greek, in contrast, were incidental, rather than the result of an institutionalized, sustained, and uninterrupted effort as in the Arab world. True, some translations of Arabic works into Greek that were made between the ninth and the fourteenth century were prepared at the request of important individuals, including an emperor; Michael Synkellos translated Abu Qurra at the request of Patriarch Thomas of Jerusalem (r. 807-21); Symeon Seth translated *Stephanites kai Ichnelates* (the Arabic *Kalila wa-Dimna)* at the request of Emperor Alexios I (r. 1081-1118); Michael Andreopoulos translated the *Book of Syntipas,* this time not from Arabic but from Syriac into Greek, for the Byzantine governor *(dux)* of Melitene, Gabriel (ca. 1100); in 1266 the monk Arsenios translated the geomantic treatise of al-Zanati at the request of the lady Theodora, who has been identified as Theodora Doukaina, the future empress of Michael VIII Paleologos. The work of al-Raz! on measles was translated (presumably in the eleventh century) at the command of an emperor, who remains unknown since he is not named in the introduction to the Greek translation. Only Alexios, the translator of Daniel's celestial omens in the year 1245, states that he decided to translate this work on his own initiative. But there never seems to have been extensive and systematic patronage for translations, possibly because the scientific texts already possessed by the Greek-speaking world covered most of what Byzantine intellectuals perceived to be their scientific requirements. In addition, the translations of a fair number of Greek texts had been assimilated into Arabic learning, forming a body of Arabic scientific treatises based on Greek knowledge that was of no interest to a Byzantine readership already conversant with the Greek originals. As Pingree has observed, though Arabic scientific texts were translated into Greek from the ninth to the fourteenth century and into Latin in the West only from the twelfth to the thirteenth century, more Latin translations were made in these two centuries than the Greeks made in six. According to Pingree, this disparity can be explained in part by the fact that the Byzantines had access to Greek scientific texts that were either superior or equal to those of the Arabs, but the Latins did not. '07
<sup>106</sup> The political, ideological and social reasons for the wave of translations from Greek into Arabic are briefly outlined in *E/2,* s.v. "tarQ.iama," and more extensively discussed in the study by Gutas, *Greek Thought, Arabic Culture.*
<sup>1117</sup> Pingree, "Indian and Pseudo-Indian Passages," p. 143.
If Byzantium had no pressing need for Arab learning, why translate at all? The reasons will become clearer as our understanding of developments in Byzantine science improves. But at least in the case of astrology, the difficulty, in spite of a very conscientious effort, to render every single Arabic astronomical term into Greek, evident already in the early stage of the translations that were in place by the eleventh century, shows how much Arabic astrology had departed from its Hellenistic antecedents. ws As for the *Oneirocriticon,* one reason for translating it, in spite of the availability of a Greek manual on dream interpretation as comprehensive as Artemidoros's was, must have been that by the tenth century Byzantine readers could no longer connect their own reality with the pagan deities and the civic and social institutions of late antiquity that so frequently figure in Artemidoros. 109 Muslim sources were much closer to home for the Byzantines, since they are addressed to the adherents of a monotheistic religion born out of a Judaeo-Christian background who lived at the same time in a society with institutions and an ideology that had its parallels in the Byzantine Empire. As for medicine, the preface to the translation of al-RazI's work on measles (Ili::pt AotµtKiji;) briefly discusses the reason for preparing a Greek version of the text, stating what both the translator and his imperial patron expected to find in the translated work. The preface explains that, though Galen had discussed in detail everything about medicine in his works, as is well known to those who had studied them carefully, he had not dealt with measles sufficiently. In the words of the translator, this was because almost no one in Galen's time had caught the disease in question. <sup>110</sup>Measles is a relatively new disease that first came to Europe from Asia in the sixth century A.o.
In sum, the astrological and oneiromantic translations imply and the medical translation expressly states that the Byzantines turned to Arab works to improve their knowledge about subjects that were not treated adequately in their own literature. That they were aware of these gaps in their know ledge and that they made a concerted effort to fill them indicates that Byzantine science was not an imitation, regurgitation, or reorganization of works inherited from late
<sup>108</sup> See Pingree, *From Astral Omens to Astrology,* pp. 72-73.
<sup>109</sup>It was exactly this difficulty that I:Iunayn b. Ist.iaq tried to bypass by not simply translating the ancient Greek text of Artemidoros into Arabic, but by adapting something written for a pagan late-antique society to the needs of a monotheistic medieval one.
<sup>110</sup> See al-Riizl, *De pestilentia* in *Alexandri Tralliani Medici Lib. XII, Rhazae de pestilentia libellus ex Syrorum lingua in Graecam translatus,* p. 243.
antiquity, but rather was constantly taking into account contemporary reality and trying to address new needs and concerns as they arose.
The translations from Arabic into Greek begun in the ninth and tenth centuries should therefore be viewed as part of the\_intense intellectual activity flourishing in Byzantium at that time, so that we can better understand whether the intellectual accomplishments of the caliphate had any bearing on the Byzantine revival of the ninth and tenth centuries and what role the classical tradition played in this revival.
The magisterial study by Paul Lemerle, *Le premier humanisme byzantin* (1971) considers the possibility that the Byzantine relationships with the caliphate might have played a role in accelerating Byzantine humanism, particularly through the rediscovery of classical and Hellenistic texts that could no longer be found in the Byzantine Empire, but concludes that they were of no consequence. The question has recently been reexamined by Gutas, who suggested that Greek manuscripts in the minuscule containing ancient Greek scientific texts might have been copied in response to the translation and study of the same texts in Baghdad, or because of specific Arab demand for manuscripts of these works. 111 Though more research is required before this issue can be resolved, the translations from Arabic into Greek made at that time indicate that at least some elements of Arabic learning, which were compatible with and complementary to the existing Byzantine tradition (since both the Arabs and the Byzantines were indebted to the classical and late-antique Greek heritage for their intellectual and scientific accomplishments) found their way to Byzantium.
The Byzantine intellectual activity of the ninth and tenth centuries, known as the Macedonian Renaissance is usually viewed as exclusively or primarily classicizing, but the translations from Arabic into Greek suggest that it was more complex than the appellation "Renaissance" implies. The *Oneirocriticon* and the astrological texts translated from the Arabic in the course of the tenth century are not high-style literature, though this does not mean that they were unknown to Byzantine high society or the imperial court, the circles who were supposed to have produced and consumed the intellectual products of the so-called Macedonian Renaissance. There is evidence to suggest that the *Oneirocriticon* was compiled for an imperial patron or at least used by an interpreter of imperial dreams. As for the utilization of classical and late-antique Greek sources, the example of the *Oneirocriticon* and its relationship with the work
<sup>111</sup> Gutas, *Greek Thought, Arabic Culture,* pp. 175-86.
of Artemidoros and other ancient Greek dreambooks that circulated in Byzantium during the tenth century is telling. Though these older Greek sources on dream interpretation were available, the compiler of the *Oneirocriticon* did not use them, but instead drew upon Arabic dreambooks. <sup>112</sup>
In several passages of the *Oneirocriticon* Artemidoros's wisdom returns to Byzantium by way of the Arabic sources used by the Byzantine compilation, a recycling of ancient tradition between Byzantium and Islam that is not unique. Pingree has argued that elements of ancient Greek astrology returned to Byzantium via translated Arabic texts at around the same time. 113 The example of dream interpretation shows that we should proceed with caution in drawing our conclusions, however. While it is certain that the author of the *Oneirocriticon* did not use Artemidoros's work, there is no reason to believe that he was not aware of its existence. Whatever the case, at least some Byzantine readers read both the *Oneirocriticon* and Artemidoros, as can be deduced from the excerpts from the *Oneirocriticon* copied by an eleventh-century hand on the margins of the earliest surviving Artemidoros manuscript. These readers may have regarded the agreement between Arab and ancient Greek lore as confirmation of the tradition's validity. It seems that the translation of Artemidoros into Arabic functioned in a similar way, that is, not to reject the indigenous Semitic lore on dream interpretation but to complement it and to confirm its validity. As Strohmaier has noted, the Arabic translation appears to have been couched in terms conforming to the tradition of dream interpretation. 114 All this may explain the popularity of the *Oneirocriticon* in the Byzantine world, as evidenced
<sup>112</sup>The earliest Greek manuscript containing Artemidoros's work belongs to the I Ith century. However, there are two pieces of evidence that his work was read by Byzantine intellectuals in the course of the 10th century: (1) numerous and extensive passages from at least four out of his five books were copied into a Byzantine encyclopedia of the 10th century, the *Suda Lexicon;* see *Suidae Lexicon,* ed. Adler, vol. 5, index auctorum (pp. 67-68), s.v. "Ap'ti::µUicopoi;." (2) Artemidoros is mentioned in the *Philopatris,* a ps.-Lucianic dialogue that appears to have been written in the 10th century, possibly during the reign of Nikephoros Phokas; see Ch. Aggelide, "He chronologia kai ho suggrapheas tou dialogou *Philopatris," Hellenika* 30 (1977-78), pp. 34-50; for other views on the date of the text, see *ODB,* s.v. "Philopatris." For the passage of the *Philopatris* referring to Artemidoros, see del Como, *Graecorum de re onirocritica,* p. 22. Besides Artemidoros, four more ancient Greek dream interpreters whose works do not survive today are mentioned in the *Suda,* which tells us that a number of ancient texts on dream interpretation were known and available to the Byzantines during the 10th century.
<sup>113</sup>Pingree, "Classical and Byzantine Astrology," pp. 227-39.
<sup>114</sup> Strohmaier, "Die griechische Gotter," pp. 149-50.
#### 424 CHAPTER TEN
by its extensive manuscript tradition, as well as the popularity of Artemidoros in the Arab world, as evidenced by the abundant quotations from his treatise found in later Arabic dreambooks. Both works were very popular because the new is always more readily accepted if some of its elements are already familiar.
The best evidence for Byzantine and Arabic intellectual contacts are the translations of scientific texts from Arabic into Greek. Byzantine narrative sources are silent on the subject. But, in Magdalino's words, "One has to recognize that rejection, whether expressed through adverse comment or through silence, may be a rhetorical attitude, which does not preclude reception and may actually be used to disguise it." 115 It is not hard to understand why Byzantine intellectuals, including those belonging to imperial circles, would have welcomed the translation of Arabic scientific texts in the ninth and tenth centuries. Diplomatic contacts provided the Byzantines and the Arabs with a chance to test and display each other's learning. Contemporary scholarship has time and again examined the Byzantine narrative sources pertaining to the diplomatic missions to the Arabs of four eminent Byzantine scholars, John the Grammarian, St. Co!1stantine/Cyril, Patriarch Photios and Leo Choirosphaktes, as well as the alleged invitation to Baghdad extended by Caliph al-Ma'miin to Leo the Mathematician. In his recent reexamination of these sources, Magdalino rightly emphasized the significance of the fact that the four Byzantine ambassadors to the caliphal court were individuals "highly distinguished in secular learning," which indicates that they were chosen because learned discussions would be part of the diplomatic agenda and because the Byzantine government was anxious to prove to its Muslim neighbors "that the wisdom of the Greeks was still, contrary to rumor, alive and well in Christian hands." <sup>116</sup>
Diplomatic dazzle aside, the Arab-Byzantine military hostilities that reached their height in the ninth and tenth centuries provided a practical reason for wanting to know about the scientific progress of the enemy. Advances in science and technology guarantee superiority in the battlefield: astronomy, meteorology and dream interpretation are all mentioned in the C *onstitutiones Tacticae* of Emperor Leo VI (r. 886-912) as arts that minister to the waging of war. Practical alchemy of the kind encountered in Middle Byzantine formulas shows knowledge of Arabic alchemy, and it too is useful for developing war technology. Two of those formulas pertained to the tempering of metals, and one was applied to the manufacture of swords. Alchemy also contributed to
<sup>115</sup>Magdalino, "Road to Baghdad," p. 196.
<sup>116</sup> Ibid., p. 206.
the development of incendiary concoctions such as Greek fire. <sup>117</sup>
The nature of the few known translations of Byzantine (as opposed to ancient Greek) texts into Arabic confirms the impression that translations from and into Greek and Arabic were made for reasons other than intellectual curiosity. Excerpts from an Arabic translation of Byzantine works on military strategy, including the *Strategikon* attributed to Emperor Maurice (r. 582-602), were incorporated in *al-siyasatu-l-<ammiyyah* (Universal Governance) which is probably the earliest Islamic *speculum principis.* Its author was in the entourage of Salim Abu-I-'Ala', the secretary to Caliph Hisham b. 'Abd al-Malik (r. 724-43). <sup>118</sup>The fourteenth-century Egyptian author MuI:iammad b. Mangli, in two of his fifteen treatises on military subjects, quotes parts of the *Constitutiones Tacticae* by Emperor Leo VI (r. 886-912). A tenth-century date has been suggested for the translation of this Byzantine manual into Arabic. <sup>119</sup> In addition, a number of Arabic alchemical texts seem to have been based on a Byzantine, and not an ancient Greek, model. One is the *Kitab* al-~abib (Book of the Beloved), which is based on the alchemical works of Stephen of Alexandria and the works attributed to the emperor Heraclius and to Marianas, the Byzantine monk who, according to the Arabic tradition, taught alchemy to the Umayyad prince Khalid b. Yazid b. Mu'awiya (d. 702120).
<sup>118</sup>See M. Grignaschi, "L'origine et Jes metamorphoses du *Sirr-al-asrar," AHDLMA* 43 (1976), p. 9.
<sup>119</sup>See V. Christides, "Naval Warfare in the Eastern Mediterranean (6th-14th Centuries): An Arabic Translation of Leo Vi's *Naumachica," Graeco-Arahica* 3 (I 984 ), pp. 137-48; N. Serikoff, "Leo VI Arabus? A Fragment of Arabic Translations from the *Tactica* by Leo VI the Wise (886-912) in the Mamluk Military Manual by Ibn Mankali (d. 1382)," *Macedonian Studies* 9:3-4 (199, p. 59, suggests the sixties of the 10th century as the most probable date for the Arabic translation of the *Tactica,* but does not support this proposition with hard evidence.
rn On Marianas and the Arabic tradition, see Ullmann, *Die Natur- und Geheimwissenschaften im Islam,* p. 192; however, Khalid b. Yazid's involvement in the translation of Greek alchemical texts into Arabic is shown to be a later fabrication in M. Ullmann, "Khalid b. Yazid und die Alchemie: Eine Legende," *Der Islam* 55 ( 1978), pp. 181-218. Mention of Byzantine Greek (as opposed to ancient Greek) texts translated into Arabic is also made by the 10th-century historian al-Mas'iidi, without, however, specifying what kind of works were translated; for references, see Gutas, *Greek Thought, Arabic Culture,* p. 30.
<sup>117</sup>For a summary of discussions and bibliography on the Greek fire until 1984, see Christides, *Conquest of Crete,* pp. 63-66; see also Th. K. Korres, *Hygron pyr. Hena hoplo tes hyzantines nautikes taktikes* (Thessaloniki, 1995).
#### CHAPTER TEN
The military importance of dream interpretation is acknowledged in at least one Byzantine military manual of the tenth century, "Οσα δεί γίγνεσθαι του μεγάλου και ύψηλου βασιλέως των Ρωμαίων μέλλοντος φοσσατευσαι (What Should Be Observed When the Great and High Emperor of the Romans Goes on Campaign), written during the reign of Constantine VII (945-59) on the basis of a lost work by Leo Katakylas, a high-ranking official under Leo VI, Constantine's father.121 It details the logistics of an imperial campaign in Anatolia and includes a list of books that the emperor should take with him:
βιβλία ή ακολουθία της έκκλησίας, βιβλία στρατηγικά, βιβλία μηχανικά, έλεπόλεις έχοντα, και βελοποιϊκά και έτερα άρμόδια τη ύποθέσει, ήγουν πρός πολέμους και καστρομαχίας: βιβλία ίστορικά, έξαιρέτως δε τον Πολύαινον και τον Συριανόν: βιβλίον τον όνειροκρίτην:122 βιβλίον συναντηματικόν: βιβλίον
121 Published as "Appendix ad librum I" in De cerimoniis aulae byzantinae, ed. Reiske, vol. I (Bonn, 1829), pp. 444-508; new ed. J. Haldon, Three Treatises on Expeditions, text (C), pp. 94-151; hereafter cited as "treatise On Imperial Expeditions." On the authorship, date and sources of this text, see Haldon's detailed introduction to the edition, as well as the ODB, s.v. "De ceremoniis."
122 Both commentators of the text, Reiske (p. 523) and Haldon (p. 211, n. C 199), believe that this dreambook is by Artemidoros. However, the dreambook of Artemidoros is always called ß1821x overookoruxó, both in the manuscripts where it survives and in references to it that can be identified in other works Moreover, no ancient dreambook that we know of is called Oneirokrites. My information on the title of Artemidoros is based on Rigault's edition of 1603, which repeats the editio princeps without changes, as well as on the apparatus of the two critical editions, by R. Hercher (Leipzig, 1864) and Pack (1963), and on the descriptions of Artemidoros's manuscripts published the relevant catalogues. For a complete inventory of the manuscripts of Artemidoros and references to the relevant library catalogues, see the introduction to Pack's edition. All references to Artemidoros in ancient and medieval literature have been collected by Pack, p. xxiv. All references to ancient dream interpretation in ancient and Byzantine authors were collected by del Corno, Graecorum de re onirocritica. The dreambooks recorded there are the following (the numbers in parenthesis refer to pages in del Corno): τέχνη όνειροκριτική and ύποθηκαι θεωρημάτων (24, 32 59), περὶ ὀνείρων (25, 34, 61), περὶ κρίσεως ἀνείρων (45), περὶ των όνειράτων και τών σημείων δι΄ ών ό Σεουήρος τὴν αύτοκράτορα άρχήν ήλπισε (51), περι μαντικής (53, 83, 84, 162, 165, 167), φυσικός λόγος (53, 84), περι χρησμών (57, 164), περι ψυχής (58, 161), όνειροκριτικά (66), περὶ ἐνυπνίων (67, 153), περὶ ύπνου και ένυπνίων (68, 148), περὶ τῆς καθ΄ ὕπνον μαντικής (148), Βίβλοι "Ωρου καὶ Ίσιδος (151), περί τεράτων και σημείων (152), περι χρηστηρίων (164). Eustathios of Thessaloniki (12th century) in his commentary to the Iliad (A, 63) writes of the ancient dreambooks: καί είσιν ξηγραφοί τινες τέχναι όνειροκριτικαί, ώσπερ ήσαν και οιωνιστικαι και θυτικαί και έτεραι (There exist some written dreambooks [technai oneirokritikail as there used to be books on the interpretation of the flight of birds and the portents of sacrifices); see Eustathios of Thessaloniki, Commentarii ad Homeri Iliadem
το περιέχον περι εύδείας και χειμώνος και ζάλης, ύετου τε και άστραπών και βροντών και άνέμων έπιφορας πρός τούτοις βροντολόγιον και σεισμολόγιον, και έτερα, όσα παρατηρουνται οί πλευστικοί. ίστέον δε, ότι τοιούτον βιβλίον ἐφιλοπονήθη καὶ ἐκ πολλών βιβλίων ήρανίσθη παρ᾽ ἐμοῦ Κωνσταντίνου ἐν Χριστώ βασιλεί αίωνίω βασιλέως Ρωμαίωνί 123
Books: the liturgy of the Church, military manuals, books on mechanics, including siege machinery and the production of missiles and other information relevant to the enterprise, that is to say, to wars and sieges; historical books, especially those of Polyaenos and Syrianos;14 an oneirocritical book; a book of chances and occurrences; a book dealing with good and bad weather and storms, rain and lightening and thunder and the vehemence of winds; and, in addition to these, a treatise on thunder and a treatise on earthquakes, and other books, such as those to which sailors are wont to refer. Note that such a book was researched and compiled from many books for myself, Constantine, emperor of the Romans in Christ the eternal King. 125
A further comment on the military importance of dreams can be found in yet another military manual of the tenth century, the Constitutiones Tacticae by
123 Constantine VII (attributed to), De cerimoniis, ed. Reiske, vol. 1, p. 467; new edition by Haldon, ed. and trans., Three Treatises on Expeditions, p. 106
124 Both Polyaenos and Syrianos wrote treatises on warfare. Their characterization as "historical books" is due to the fact that both manuals included illustrations drawn from older historical works, such as Herodotos and Thucydides; see Haldon, ed. and trans., Three Treatises on Expeditions, p. 210, n. (C) 199. Quoted from the translation by Haldon, ibid., p. 107.
125 Quoted from the translation by Haldon, ibid., p. 107.
pertinentes, ed. M. van der Valk, vol. 1 (Leiden, 1971), p. 78. Greek manuscripts apply the title Oneirokrites only to Byzantine dreambooks: όνειροκρίτης κατά Μανουήλ του Παλαιολόγου [sic] from Paris. gr. 2419; όνειροκρίτης όμέγας by Blasios the Athenian from Athen. Bibl. Nat. 1350; όνειροκρίτης κατά την σελήνην from Vat. gr. 342; όνειροκρίτης του κάθε ψηφίου του κατά άλφαβήτου from Athen. Bibl. Nat. 1275 and έτερος όνειροκρίτης της σελήνης from the 19th-century Athen. Bibl. Nat. 1350. One irokrites is likely to have been the original title of the Oneirocriticon, as is suggested by part of its manuscript tradition. The Oneirocriticon is by far the longest and most comprehensive among the surviving Byzantine dreambooks, as it covers 241 printed pages (compared with approximately 20-30 printed pages for the other surviving dreambooks) and, when not excerpted, it usually occupies a whole volume in the manuscripts. The only extant Greek dreambook of comparable length is Artemidoros. It is therefore conceivable that the dreambook mentioned in the Treatise on Imperial Expeditions is none other than our Oneirocriticon. Of course, the title Oneirokrites is generic, an it is not impossible that a second dreambook of substantial length by this title circulated in Byzantium, but if it did, no information about it has survived
Leo VI, who addresses the following admonitions to his generals.100
Βέβαιον μεν ούδέν μοι δοκεί των όνείρων πλάττεσθαι δε και πείθειν τους στρατιώτας, ώστε πιστεύειν τους σούς όνείρους νίκην έπαγγελομένους, έν καιρώ μάλιστα πολέμου χρήσιμόν έστιν και άναγκαιον. Δόξαντες γαρ ώς από Θεου χρήσιμον τον παρά σου άφηγούμενον δνειρον, θαρσαλέως και ἀνεπιστρόφως κατὰ τῶν πολεμίων ἐπιχειρήσουσι, καὶ τὴ προθυμία τὴν ἀνδρείαν διπλασιάσουσιν.
Nothing about dreams seems reliable to me. But, in time of war, it is useful and even necessary to fabricate <them> and to persuade the soldiers to believe your dreams that promise victory. For, thinking that the dream that you narrate is a portent from God, they will attack the enemy courageously and steadily, and their bravery will be doubled by their eagerness.
Practical considerations may have initiated the enthusiasm for translations from Arabic into Greek, but genuine intellectual curiosity must have both sustained it and broadened the scope of the Arabic sources chosen for translation. Given that a number of the translated texts pertain to the occult sciences and
<sup>126</sup> Leo VI, Constitutiones Tacticae (PG 107, col. 1061A). See also ch. 1, n. 237. Towards the end of his treatise Leo enumerates the fields of human knowledge that are important for waging war (Epilogus, PG 107, col. 1088B): νή Συνεργούσι δε τη φύσει του πολέμου αύται αί τέχναι, οίον όπλιτική, λογιστική, άρχιτεκτονική, άστρονομική, icoatin', iamorn (The following arts contribute to the nature of war: the art of using heavy arms, practical arithmetic, architecture, astronomy, priesthood and medicine). Works on at least two of the above arts, astronomy and medicine, are known to have been translated from Arabic into Greek as early as the 10th century. Further on, Leo explains the role of each one of these arts. His treatment of astronomy-astrology deserves some attention, as it focuses not only on the practical but also on the occult aspects of reading the stars (Epilogus, PG 107, 1089B-C): ξα΄ Της δε άστρονομίας είσίν, τους καιρούς του ένιαυτού προλέγειν, έν οις χειμώνων και καυμάτων μεταβολαί γίνονται, ή ύδάτων όμβρίων μεταφοραί, ή πνευμάτων ἐξαισίων κινήσεις, ἐξ ῶν στρατεύματα πολλάκις εἰς μεγίστους κινδύνους ἐνέπεσον. ώρας τῶν ἐπιθέσεων καὶ πρὸς τάς ἀναπαύσεις, ἐν αῖς άστοχουντες πολλάπκις αχρηστον τὴν δύναμίν τινες ἀπεργάζονται. Περί τε σεισμών και των άλλων σημείων τα μέλλοντα δηλοποιείν, ή και πρός το συμφέρον τας ἐπιφανείας ἐντέχνως μεταρυθμίζειν [sic]. Πάντα ταῦτα τῆς ἀστρολογίας εἰσίν (The role of astronomy is to predict the seasons of the year during which changes in cold and hot weather occur, or alterations in rainfall or extraordinary motions of winds, because of which armies have frequently fallen into the gravest danger. It is also to divide day time and night time accurately in order [to determine] periods of attack and periods of rest. Some [generals] who fail [to do that] frequently render their forces useless. In addition, its role is to reveal the future through earthquakes and other portents, or even artfully to adapt [the interpretation] of these phenomena according to what is advantageous. All of the above belong to the realm of astrology).
that the Orient seems to have been stereotyped in the Byzantine mind as the cradle of apocryphal wisdom, 127 a possible Byzantine demand for Arabic knowledge on such topics may also have played its part, if not in the preparation of these translations, at least in their being copied and excerpted in later manuscripts.
The *Oneirocriticon* is only one small part of the intellectual exchanges between the Byzantines and the Arabs that took place in the course of the ninth and tenth centuries, in spite, or rather because, of their military and political confrontation. The source material that survives can help modern scholarship to document these exchanges, but research in this direction has barely begun. <sup>128</sup>Obtaining a comprehensive picture of mutual influences between the Byzantine and the Islamic civilizations will require the effort of many individuals over many years; their findings will help us write a new chapter in the history of world civilization.
<sup>127</sup> Cf. the remarks on the ethnic origin of the astrologers, diviners, and sorcerers in the Byzantine empire in Ph. Koukoules, *Byzantinon bias kai politismos,* Vol. 1:2, pp. 136-39.
<sup>128</sup> The lack of reliable and comprehensive philological tools is acutely felt in Arabic studies. The dearth of critical editions renders recourse to the manuscripts themselves obligatory. The extensive holdings of libraries in the Middle East have not yet been properly catalogued, and the catalogues that do exist sometimes provide very poor or incorrect information. For an overview of the quantity and quality of the available catalogues of Arabic manuscripts, see R. S. Humphreys, *Islamic History: A Framework for Inquiry,* rev. ed. (Princeton, N. J., 1991), pp. 36-40. Especially on the problems of philological and other scholarly preparation regarding the study of Arabic astronomy, but also the rich promises for important future discoveries that the field holds, see G. Saliba, "Writing the History of Arabic Astronomy: Problems and Differing Perspectives," *JAOS* 116:4 ( 1996), pp. 709-18; and idem, *A History of Arabic Astronomy: Planetary Theories during the Golden A.ge of Islam* (New York, 1994), pp. 1-47. In addition, the difficulty of obtaining microfilms from, or even gaining access to, the manuscript collections of several eastern European and Middle Eastern libraries forces scholars to base their inquiries mainly on the holdings in western European and American collections. This situation warrants the hope that currently unknown texts relevant to the history of Arabic and Byzantine science will surface in the future.
This is an open access chapter distributed under the terms of the CC BY 4.0 license.
Interpretations Extant only in Paris. Suppl. Gr. 690
#### Pearls and Black Clothes
#### Paris. Suppl. gr. 690, fol. 125r, col. 1:
το γαρ μέλαν ίμάτιον εί και θλίψεως δηλωτικόν έκρίθη, άλλ' ουχι και τοις συνήθως έχουσι του τοιούτου ἐνδύματος, ώσαύτως καὶ οἱ μαργαρίται καὶ οἱ λίθοι, εί και είς σοφίαν και θεογνωσίαν ἐκρίθησαν, ἀλλά τοις κεχρημένοις αυτοίς έξ έθους, ουχί δε και τοις πένησι και τοις ίδιώταις. τα αυτά ἐν ταῖς κρίσεσι των όνειράτων έδηλώθη.
Though a black garment was interpreted as indicative of sorrow, <such an interpretation> does not <apply> to those who usually wear such clothing. Likewise, though pearls and gems were interpreted as wisdom and knowledge of God, this applies to those who use them habitually, not to the poor and common people. The same things have been indicated in the <part of this work that contains> dream interpretations.
This passage is inserted in the abridgment in a place that corresponds to Drexl 240, 21-22; black clothes are discussed in Drexl 167, 24-168, 12; pearls are discussed in Drexl 200, 23-203, 25 and 210, 25-211, 9 (as this last passage appears with minor modifications in Vat. gr. 573, fols.199v-200r, I will quote it from there). The phrase τα αύτα έν ταις κρίσεσι των όνειράτων έδηλώθη (the same things have been indicated in the <part of this work that contains> dream interpretations), which corresponds to the contents of the entry on gems in the extended version but not in the abridgment, was probably copied from the extended version unchanged, without checking to see that it corresponded to the contents of the abridgment.
Cf. Pascalis Romanus II, 1; Collin-Roset, "Le Liber Thesauri Occulti," 165. 22-166. 2:
Si vestimenta nigra angustiam et tribulationem significent, tamen non talia significabunt monacho vel qui utitur eis. Similiter gemme ac margarite quamvis sapientiam et Dei cognitionem presignent nobili vel sapienti aut illas gestanti. verumptamen rustico, pauperi aut idiote non illa significabunt; cuius enim precii erit margarita inter porcos (Matt. 7:6)? Nullius.
Though black garments signify straits and tribulation, they do not signify the same for a monk or someone who uses them <regularly>. Likewise for a gem and a pearl. Though for a nobleman or a wise man or someone who wears them <regularly> they indicate wisdom and knowledge of God, for a peasant, a poor man or a commoner they do not signify the same thing. For what value could a pearl have among swine (Matt. 7:6)? None.
Cf. Paris. Suppl. gr. 690, fol. 122r, col. 2:
ΜΑΡΓΑΡΟΙ ΤΙΜΙΟΙ ΛΙΘΟΙ ΟΙ μαργαρίται και οί τίμιοι λίθοι είς θείους λόγους και σοφίαν και θεογνωσίαν, ώς ἐπὶ τὸ πολὺ διακρίνονται. τινὲς δὲ δάκρυα καὶ κλαυθμον είπον, ή φόβον από έξουσίας.
PEARLS, GEMS: Pearls and precious gems are most of the times interpreted as divine words, wisdom and knowledge of God. Some said <that they signify> tears and wailing, or fear of authority.
Cf. Vat. gr. 573, fol. 199v-200r:
εί δε ίδη <ό βασιλεύς> ότι ήνεγκαν αύτφ λίθους και μαργαρίτας, αναλόγως του πλήθους, χαράν δέξεται. αλλος έαν ίδη ταυτα ότι εύπόρησεν, ευρήσει πλουτον ανάλογον και δόξαν. και θάνατον από βασιλέως. Ταυτα γάρ, βασιλεί μόνω άρμόζει.
If <the emperor> dreams that he was brought gems and pearls, he will receive joy commensurate to their number. If someone else dreams that he received such things, he will find commensurate wealth and glory and death from the emperor, for such things are approprate for the emperor alone.
Cf. Vat. gr. 573, fol. 202v:
οί μάργαροι γάρ, είς σοφίαν και χαραν προεκρίθησαν, αναλόγως του μεγέθους. εί δε ίδη τις ὅτι εύρε μικρούς μαργάρους πλείστους, εύρήσει χαραν καὶ σοφίαν ἐλάττονα. πολλάκις δὲ ταῦτα πάντα, εἰς φόβον ἀπὸ ἐξουσίας κρίνονται. ἐαν ίδη τις ότι έξ αυτών ήμφίασεν ... εί δε του κοινού λαου [έστι], εύρήσει φόβον και τρόμον από έξουσίας. Τουτο γάρ βασιλεύσιν και έξουσιάρχαις άρμόδιον.
Pearls were previously interpreted as wisdom and joy commensurate to their size. If someone dreams that he found several small pearls, he will find moderate joy and wisdom. All such things are frequently interpreted as fear of authority. If someone dreams that he wore any of them ... if he is a commoner, he will find fear and terror from the authorities. For such a thing is appropriate <only> for emperors and people in a position of authority.
*Outcome of a Dream in Twenty Units of Time*
*Paris. Suppl. gr.* 690, fol. 125r, col. I:
*ano* npcirtrii; wpai; VUK1Epl vfii; ECO<; 1pi i:rii;, El KOcraEi:iii; l':crn v Ti EKPacrti; Tl yap ElKOO"UETIJ<; l':crnv Ti Kpicrt<; il E'iKOO"l µrivoov il c'iKOO"l £P8oµa8cov il E'iKOO"l ljµEpoov Jl £iKOO"l WpOOV.
<If a dream was dreamt> between the first and the third hour of the night, its outcome will come in twenty years or its interpretation <will be fulfilled> in twenty years or in twenty months or in twenty weeks or in twenty days or in twenty hours.
In the extended version the passage reads (Drexl 241, 2-3):
*ano* npcirtrii; yap wpai; 1fii; vunoi; ECO<; 1pi 'tTl<; Kpi vnm µ£xpt El KOO"l:OU El:OU<; Ti EKPacrti;.
From the first until the third hour of the night the outcome <of a dream> is interpreted as <taking> up to twenty years <to be fulfilled>.
The Greek model of the *Liber Thesauri Occulti* included this phrase of the Paris abridgment. Cf. Pascalis Romanus II: 1; Collin-Roset, "Le *Liber Thesauri Occulti,"* p. 167, II. 1-2:
Sompnium a prima hora noctis usque ad ejusdem terciam visum in .XX. annis vel mensibus seu ebdomadibus vel diebus aut horis .XX. terminabitur.2
A dream dreamt between the first and third hour of the night will be fulfilled in twenty years or twenty months or weeks or days or twenty hours.
*Change of Religion*
*Paris. Suppl. gr.* 690, fol. 125r:
<sup>2</sup> The same addition occurs in the Anglo-Norman version from the end of the 13th century (Berlin Q. 968): Si devez savoir qe le songe qe avient al comencement de! nuyt jeques a la terce houre doit avenir al plus tard en .XX. aunz ou en .XX. mois ou en .XX. semaignes ou en .XX. jors. (Berriot, *Exposicions et significacions des songes,* p. 30 I), Apparently this version used Pascalis Romanus as a source. The addition does not occur in Leo Tuscus (at least not in the text contained in *Digby* I 03).
435
ΕΝΑΛΛΑΓΗ ΠΙΣΤΕΩΣ: ή της πίστεως έναλλαγὴ μετατροπήν και κακοήθειαν των τρόπων σημαίνει.
change of religion: Changing one's faith indicates the transformation and vileness of <the dreamer's> manners.
Cf. Drexl 8, 1 ff .; Gigli notes that the first general statement is absent from the other Greek manuscripts, but it may well have existed in the archetype in order to serve as an introduction to particular cases.
Cf. Ibn Shāhīn, no.1070:
من راى أنه تحول عن الاسلام إلى احد الاديان الباطلة فإنه ارتكاب معاص وقيل ذلة وحقارة.
If someone sees that he changed his religion from Islam into one of the false religions, indeed this is perpetration of sins ( معاصر ) and it is said <that it means> depravity and vileness (κακοήθειαν).
#### Icons
Paris. Suppl. gr. 690, fol. 126v, col. 2: εί μὴ ό τουτο είδώς έπίσταται (wypaduchy (<Icons in gold mean sorrow>, unless the dreamer is a painter); cf. Drexl 106, 24-27.
#### Pumpkin and Melon
Paris. Suppl. gr. 690, fol. 127r, col. 1:
< κολοκψηθη και πεπαν:> ώστε τα μεν εύθαλή τούτων εὐθύτητα γνώμης σημαίνει. τα δε άσθενη και ροώδη και μεμαραμμένα τουναντίον, άλλα και ή τούτων κρίσις άλλως έν έαρι και άλλως ἐν φθινοπώρω νοεῖται.
<PUMPKIN AND MELON: > those that are thriving indicate straightness of thought. The feeble, overripe and withered ones <indicate> the opposite, but even their interpretation is meant <to be> different in the spring and different in the fall.
For the influence of spring and fall on the truthfulness of dreams, cf. also Drexl 240, 24-25. This interpretation is missing from the chapter on pumpkins and melons in the extended version (Drex1 158, 2-3), but it corresponds almost verbatim to another passage of the extended version, the interpretation of the leaves on trees (Drex) 107, 6-10) : τα γαρ εὐθαλῆ καὶ εὐτραφῆ εὐθύτητα
#### APPENDIX 1
γνώμης διασημαίνουσιν, τα δε άσθενή και ροώδη και μεμαραμμένα γνώμης διακρίνουσιν ασθενείς τρόπους (The thriving and well-nourished ones indicate straightness of thought, while the feeble, overripe and withered ones point to a feeble manner of thought).
#### Rain
Paris. Suppl. gr. 690, fol. 127r (cf. Drexl 133, 4 ff.): πολλάκις δε και είς δακρύων κρίσιν, εί αρα έστι ταραχώδης (<Rain> is frequently interpreted as tears, if it happens to be tempestuous).
I could not find an interpretation of rain as tears (مطر=دمو ع) in the Arabic dreambooks, though torrential rain is interpreted as sorrow and distress (i.e., cause for tears), e.g., in Ibn Shahin, nos. 257-59:
(٢٥٧)... وان نزل الطر شديدا مثل الطوفان يلحق اهل ذلك الكان غم عظيم. (٢٥٨) وإن راى مريض أن… مطرا شديدا كدرا نزل على تواتر يهلك فى ذلك المرض. (۲۵۹) وقال ابن سيرين : من راي مطرا شديدا كدرا نزل على تواتر في وقته على دوام يلحق باهل ذلك المكان عسكر وداء وبلاء.
(257) ... If the rain is falling heavily, like a flood, great distress will overtake the people of that place. (258) If a sick person dreams ... that a heavy and turbid rain is falling incessantly he will perish from his sickness. (259) Ibn Sîrîn says: Whoever sees that a heavy and turbid rain is falling incessantly throughout the rainy season <it means that> an army, a disease or tribulation will overtake the people of that place.
Likewise Drexl 134, 3-5: όπου αν βρέχη, ή καθαρά βροχή είς άγαθον παντός προσώπου κρίνεται, ή δε ένθολος είς έναντίον (Wherever it is raining, clear rain is interpreted as a good thing for every person, while turbid rain <is interpreted> as the opposite).
#### Baths
Paris. Suppl. gr. 690, fol. 127v: το δε εύκρατον ού φαύλον άλλά καθάρσεως δηλωτικόν (A tempered <bath> is not bad; rather, it indicates purification).
A similar interpretation in different words in Drexl 143, 25-27: ei δε έλούσατο μεθ' ύδατος μὴ ὅντος λίαν ψυχρού καὶ ἀπελούθη, ἀποβαλεῖ 02.041.03 (If he bathes in water that was not very cold and was washed clean, he will thrust out a sorrow).
#### Bees
Par. Suppl. gr. 690, 128v, col. 2: αι μέλισσαι αί μεν ἐν τῶ ἰδίω κυψελίω πλούτον δηλούσι (Bees in their own beehive indicate wealth).
Artem. gr. ii.22, Pack 139, 3-14, gives other interpretations of bees. Cf. Ibn Shāhin, no. 6472: هن راي ان النحل لدغته فانه يؤول تحصول مـال حـلال (If someone dreams that the bees are in the beehive, it is interpreted as lawful money <acquired> through labor).
Interpretations in Paris. Suppl. Gr. 690 and Nikephoros
#### Pearls
Gigli observes that pearls are interpreted as tears in Nikephoros, ed. Drexl 184 (ed. Guidorizzi 68) and Germanos 138: μαργαρίται δηλούσι δακρύων ροήν (pearls indicate the flowing of tears). Pearls also represent tears in Islamic dream interpretation. The interpretation of pearls as φόβον από έξουσίας (fear of authority) is found in Vat. gr. 573, fol. 202v. Cf. also al-Nabulusī, vol. 2, p. 198 (s.v. وربما دلت رؤيته على الدموع الجارية من العين لأنهم :( لؤلئؤ شبهوا الدموع باللوّ (And perhaps his dream <of pearls> signifies the tears that flow from the eyes, because tears are like pearls).
#### Fleas and Lice
Paris. Suppl. gr. 690, fol. 129r, col. 1: Αί ψύλλαι είς έχθρους πτωχούς κρίνονται. Σσαύτως και οί όλιγοι φθείρες. Οί δε πολλοί πλουτον
<sup>3</sup> Contradicted, though, by an interpretation immediately preceding it (Drex1 143, 24-25): εί δε ίδη, ότι έλούσατο χλιαρφ ύδατι ἐν λουτρφ, εύρήσει θλιψιν ἐλάσσονα καὶ μέτριον (If he sees, that he bathed with lukewarm water in a bath, he will find minor and moderate sorrow).
δηλούσι (Fleas are interpreted as poor enemies. A few lice <are interpreted> likewise, while many <lice> indicate wealth).
For fleas, cf. Nikephoros, ed. Drexl 343: Ψύλλας θεωρών δυσμενείς έση βλέπων (If you dream of fleas you will see enemies). Cf. also Nikephoros, ed. Guidorizzi, Appendix I, 80: Ψύλλας θεωρών δυσμενείς όραν δόκει (If you dream of fleas, know that you will see enemies).
For lice, cf. Nikephoros, ed. Drexl 324-26 (ed. Guidorizzi, Appendix II, 114-15): Φθείρας ό κρατών καιρόν άργον ανύει (The one who has lice is not occupied <with anything>); Φθείρας συνάγων κέρδος έξεις οὐ μέγα (If you collect lice you will have a small profit); Φθειρας φέρειν [sic] φθόνον φέροις τοις πλουσίοις (If you have lice you are envious of the rich). Cf. also Nikephoros, ed. Guidorizzi 63: Φθείρας φέρειν συνούσι δηλοί τον δόλον (Having lice indicates treachery against one's acquaintances).
The distinction between a few and many lice exists in Artem. gr. iii, 7; Pack 207, 8 ff., but the interpretations given there are different from the ones in ( ٦٤٨١) و اما القمل : Paris. Suppl. gr. 690. Cf. Ibn Shāhīn, nos. 6487 فايانه يؤول بالمال... (٦٤٨٧)... قــال ابن ســيــريـن رؤيـا البـــراغــيـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . says: "The dream of fleas <signifies> weak enemies").
Interpretations in Paris. Suppl. Gr. 690 and Artemidoros
#### Salt
Paris. Suppl. gr. 690, fol. 129r, col. 1: ΑΛΑΣ Το δε άλας και αύτο είς πλουτον ήδυν έκρίθη (SALT: Salt is also interpreted as pleasant wealth). The same entry is repeated verbatim in Vat. gr. 573, fol. 213r. Cf. al-Nābulusī, vol. 2, p. 260, s.v. هو فى المنام مـــال بـلا تــعب : مـلـح (Salt: this in a dream means money without labor).
The interpretation of salt is a good example of the transformation that a dream interpretation can undergo in its rendering from Greek into Arabic and back into Greek. In this case the interpretation in Greek is the opposite of the interpretation in Arabic. In Artem. gr. i.71, Pack 78, 1-5:
Τάριχοι δὲ καὶ πάντα τὰ ἀλιστὰ κρέα [καὶ ἰχθύες] ἐν μὲν τοῖς προκειμένοις παρολκάς και άναβολάς σημαίνει: δια γαρ τών άλων έπι πολύ τηρείται ἐν δὲ τοις άλλοις τηκεδόνα και λύπην. πολλάκις δε και νόσον σημαίνει διά το ύπο τών άλών τετηχθαι.
All pickled and salted meats [and fish] mean delays and postponements in one's present affairs, for salt is used to preserve meats for a long time. But, in other connections, it signifies wasting away and grief. Frequently it also means sickness because meats are consumed by salt.
However, in the corresponding Arabic passage, Artem. ar., Fahd 143, 10 ff., we read:
الباب الحادي والستون في السمك المالح: متي راى الانسان كانه ياكل سمكا مـالحافانه يدل على خير ومنفعة تكون له فى ذلك الـوقت من مـعارفه، وذلك ان الملع سبب بقاء السمك. فاما بين سائر الناس فهو يدل على حزن ويدل اكثر ذلك على المرض، وذلك انه يقهره غيره كما ان السمك المالح كبس فى الللح .
وقال الللب .
Chapter sixty-one on salted fish: When a man dreams that he is eating a salted fish, this indicates goodness and benefit that he will have at this time because of his acquaintances [or: because of his knowledge]; this is because salt is the reason for the preservation of the fish. However, among the rest of the people this signifies sorrow. Most such <dreams> indicate sickness, because <sickness> subdues <man> and changes him, like the salted fish which is preserved in salt.
The Arabic translation is inaccurate in several places. It mistranslates παρολκάς кαι ἀναβολάς (delays and postponements) as خــر ومنفقة (goodness and benefit). The first, خير (goodness), is clearly derived from the omission of the first two letters in the word تاخبر (delay). It is more difficult to account for منفعفة (benefit). It was probably a gloss for خيير which was finally incorporated into the main text. The scribal error of omiting Land incorporating a gloss into the main text, which must have occurred very early on in the manuscript tradition of the Arabic translation, most probably accounts for the positive interpretation that salt can have in Arabic dream interpretation, since the significance of salt is very negative in Artem. gr. and can be negative in the Arabic tradition as well.
Cf. also Artem. gr. i.77, Pack 85, 20-22:
οί δε των άλων ή του θείου [στέφανοι] βαρηθήναι πρὸς τινων ὑπερεχόντων τὸν ίδόντα σημαίνουσιν είσι γαρ φύσει βαρεῖς και ούδεν έχοντες τερπνόν.
Garlands of salt or sulphur indicate that the dreamer will feel the pressure of
#### APPENDIX 1
certain superiors. For these substances are strong-smelling and unpleasant by nature.
The corresponding Arabic passage, Artem. ar., Fahd 158, 16-159, 2, is:
واما اللّلّة الملّع والكبريت فانها تدل على ان صاحب الرؤيا يحمل ثقلا ومؤونة من هو ارفع درجة منه، وذلك ان هذه الاشياَ هي ثقيلة بالطبع و<ليس> فيها لبّة.
As for a wreath of salt or sulphur, it indicates that the dreamer will lift the weight and burden from someone on a higher grade than he, because these things are heavy by nature and there is nothing pleasing about them.
The interpretations of salt in Artem. gr. are all negative, but a scribal error made half of them positive in the Arabic translation. The Oneirocriticon in Drexl's edition has no entry at all on salt. Paris. Suppl. gr. 690 has a brief entry, which disagrees with Artem. gr., but corresponds perfectly to interpretations of salt found in Arabic dream interpretation. Cf. e.g., Ibn Shāhīn, nos. 6541-6541 bis:
(١١) واما الملح فإنه يوؤل بالمال قال الكرماني من راى أنه اشترى ملكا او وهب له فإنه يؤول بدراهم، وإن كان مريضا شفاه الله تعالى لا جاء فيه إنه شفاء من انتين وسبعين داء، وقيل رؤيا الملح تؤول على خمسة اوجه إعراب وحسن واستقامة وأمر جلي وصحة الجار . (٦٥٤١م) قال جعفر الصادق رؤيا الللم الأبيض على خمسة اوجه دراهم وحياء وفعل خير ومال كثير وخادم حسن، الللح المر يؤول على خمسة اوجه دراهم مردودة وكلام سيء وحزن وغم ﻭﻋﺪﻡ ﺣﺮﻛﺔ.
(6541) As for salt, it is interpreted as money. Al-Kirmani says: "Whoever sees that he bought salt, or that salt was given to him, it is interpreted as dirhams. And if he is sick, God Almighty will restore him to health, since it is reported that it is the cure for seventy-two diseases. And it is said that dreaming of salt is interpreted in five ways: a pronouncement, beauty, righteousness, an evident affair and the veracity of one's neighbor. (6541 bis) Ja'far al-Sadiq said: the dream of white salt is interpreted in five ways: dirhams, shame, a good deed, lots of money, and a handsome servant. Bitter salt [?] is interpreted as five things: returned dirhams, a straight word, grief, sorrow and lack of movement.
Paris. Suppl. gr. 690, fol. 127r: ό πηλὸς μέριμνα καὶ θλῖψιν ἢ νόσον δηλοῖ (Mud indicates concerns and sorrow or sickness).
Drexl 134, 7-8 only has: εαν ίδη τις, ότι περιπατεί έν πηλώ, εύρήσει μέριμναν και θλιψιν αναλόγως του πλήθους (If someone sees that he is walking on mud, he will find concerns and sorrow commensurate to the volume <of mud>).
Cf. Artem. gr. iii.29, Pack 216, 18: Πηλός νόσον σημαίνει και υβριν (Mud signifies sickness and lewdness).
(١١ ٦٠ ( ١٤ ) و من ر اى انه يمشى في طـين :6522 16 ( 1 ) د ( ١٤ ) او ماء كدر فانه يدل على حصول هم و غم. (٦٥٢٢) قيل رؤيا الطين تؤول بالمرض. ([6514] Whoever dreams that he was walking in mud or in turbid water, it means that sorrow and distress will befall him. [6522] ... It is said that dreaming of mud is interpreted as sickness).
#### Horses
Paris. Suppl. gr. 690, fol. 127: Ο ϊππος ... είς τὴν προνοουμένην παρά τινος οικονομίαν κρίνεται (A horse ... is interpreted as the provisions for one's livelihood).
Gigli compares this with Artem. gr. i.56, Pack 64, 18-9: όμοιος δ΄ αν είη καὶ δεσπότῃ [καὶ] ἐργοδότη καὶ φίλῳ τρέφοντι καὶ παντὶ τῳ βαστάζοντι (The horse is also like a master who farms out work, a friend who takes care of someone, and every man who carries a burden).
For a better correspondence than Gigli's suggestion, cf. al-Nābulusī, vol. 1, p. 175 فـ من راى عنده فى النام خـيــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ someone dreams that he had a horse, it signifies the extensiveness of his means of living).
#### I akes
Between rivers (Drexl, pp. 135-37) and the sea (Drexl, pp. 138-39), an entry on lakes is inserted in the Paris abridgment (Paris. Suppl. gr. 690, fol. 127r, col. 2): ΛΙΜΝΑΙ το αύτο δε και λίμναι γυναίκες μεγιστάνες σημαίνει (LAKES: Likewise, lakes indicate noble women).
Cf. Artem. gr. ii.27, Pack 149, 24 ff.: σύμμετρος δε και μικρά λίμνη γυναίκα σημαίνει εύπορον <και> άφροδισίοις χαίρουσαν: δέχεται γαρ τους είσβαίνειν βουλομένους και ή λίμνη και ούκ άπείργει (A mediumsized or small lake signifies a wealthy woman who indulges in sexual pleasures. For the lake also receives those who wish to enter and does not ward them off).
Cf. Nikephoros, ed. Drexl, 177 (ed. Guidorizzi, Appendix II, 65): λέπρα λίμνη [τε] και φρέαρ πορνική ύβρις (Leprosy, lakes and wells indicate the insolence of prostitutes).
The interpretation of a lake as a rich woman can also be found in Arabic dream interpretation, where it was carried over from Artemidoros. Cf. al-Nābulusī, vol. 1, p. 46, s.v. بحيرة:
و البحيرة الصغيرة تدل على امراة غنية والبحيرة تدل على امراة ذات يسار تحب الباشرة لأن البحيرة واقفة لا تجرى وهي تقتل <تقبل> من وقع فيها و لا تدفعه والبحيرة امراة حربية.
A small lake signifies a wealthy woman. A lake signifies a woman who is wealthy and fond of intercourse because a lake stands still and does not run, and receives whoever is inside it and does not repel him. The lake is a belligerent woman.
The expression of Artem. gr. γυναίκα εύπορον (wealthy woman) is not very close to the phrasing in Paris. Suppl. gr. 690, yovaikes μεγιστάνες (noble women). It is more probable that the interpretation in Paris. Suppl. gr. 690 originated in Artem. ar., from where it passed into subsequent Arabic dream books, and then into the Oneirocriticon and its Paris abridgment.
#### Harbors
After the interpretation of the sea, in Paris. Suppl. gr. 690 a paragraph is inserted on ports (fol. 127r, col. 2): οι δε λιμένες φίλους εύεργέτας δηλούσι,
<sup>4</sup> Artemidoros, Interpretation of Dreams, trans. White rendered λίμνη as "marsh"; the rendering "lake" is mine.
<sup>5</sup> Al-Nābulusī's interpretation repeats almost verbatim Artem. ar., Fahd 270, 6-9. Al-Nābulusi apparently copied al-Dinawari; cf. Esad Effendi 1833, fol. 166a.
διο και το θαλάσσιον ύδωρ είς πλούτον άπο βασιλέως κρίνεται (Ports indicate benefactor friends. For this reason sea water is interpreted as wealth from a king).
Artem .gr. ii.23, Pack 140, 25 ff .: αει δε λιμένες φίλους και εὐεργέτας σημαίνουσι (Harbors and all moorings always signify friends and benefactors).
و ايضـــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ الاصدقاء والمصنين الينا فى المصنين الينا فى افعالهم بنا. friends and those who are beneficent to us through their deeds towards us).
The rendering of ευεργέτας (benefactors) in Artem. ar. is periphrastic (those who are beneficent to us through their deeds towards us). The phrasing of Artem. gr. is closer to Paris. Suppl. gr. 690 than it is to Artem. ar.
#### Ants
Paris. Suppl. gr. 690, fol. 129r, col. 1: εί [instead of οί] δε τα πτερά έχοντες είς όλεθρον και αποδημίαν επικίνδυνον (The ones that have wings are interpreted as destruction and a perilous journey).
Artem. gr. iii.6, Pack 206, 20-21: μύρμηκες τούς μέν πτερά έχοντας ουδαμώς ίδειν αγαθόν: όλεθρον γαρ προαγορεύουσι και άποδημίας έπικινδύνους (Seeing winged ants is not at all a propitious sign. For they indicate death and perilous voyages).
Artem. ar., Fahd 373, 7: النمل الطيـــار في الرؤيـا دليل ردي، وذلك انـه . يدل على مــوت او ســفــر مـع شــدة (The flying ant in a dream is a sign of destruction, because it indicates death or a journey with hardship).
و من ر ای نملا خــر ج من بــيــتــه وهـو يطـيـــر فـى:Cf. Ibn Shāhīn, no. 6452 الهواء فانه يوّول بسنفر عيـاله (If someone sees an ant leaving his house by flying through the air, it is interpreted as a journey for his family).
و من راه پطيـر من مكان فـيـه : Cf. al-Nābulusī, vol. 2, pp. 307-8, s.v. و مـن راه پـطـيـر مـن مـکان فـيــه مريض فانه يمورت (If someone sees them flying from a place where there is a sick person, the sick man will die).
Artem. gr. is closer to the expression in Paris. Suppl. gr. 690 than Artem. ar. or anything I could find in Arabic dreambooks is. Artem. gr. Öλεθρος is rendered in Artem. ar. as مو ت (θάνατος) and not (öλεθρος). The Arabic dreambooks sometimes give interpretations to ants that are the opposite of those offered in the Oneirocriticon according to Drexl's edition. Though the
examples examined above indicate that the compiler of the Paris abridgment closely followed his model of the *Oneirocriticon,* which differed in several points from Drexl' s critical edition of the text, his handling of the interpretations on ports and ants seem to imply that he might also have been familiar with the Greek text of Artemidoros.
### APPENDIX 2: INTERPRETATIONS FROM THE LIBER THESAURI OCCULTI THAT DO NOT OCCUR IN THE ONEIROCRITICON OR ARTEMIDOROS
The following are the eighteen interpretations from the Liber Thesauri Occulti that its editor could not match with interpretations found either in the Oneirocriticon or in Artemidoros. She suggested therefore that they originated in Arabic sources that had been translated into Byzantine Greek but are now lost. However, a comparison of these eighteen interpretations with the Greek texts of the most recent critical editions of the Oneirocriticon and Artemidoros (rather than their Latin translations found in their 1603 combined edition, which the editor of the Liber Thesauri Occulti used) turns up four (nos. 1, 2, 8 and 10) that can be matched with excerpts from Drexl's edition of the Oneirocriticon, and a fifth (no. 17) that corresponds to a passage in the Oneirocriticon which does not occur in Drex}'s text but can be found in Vat. gr. 573, indicating that the Greek text used by Pascalis Romanus was more extensive than Drexl's text. In addition, nos. 6, 9, 11 and 15 can be matched with passages from Pack's critical edition of Artemidoros. The source of the remaining interpretations that cannot be found either in Artemidoros or the extant versions of the Oneirocriticon is probably the Greek text of the Oneirocriticon consulted by Pascalis for the composition of the Liber Thesauri Occulti. However, in at least two cases (nos. 4 and 7) I cannot exclude the possibility that these interpretations represent Pascalis's understanding of his Greek sources and his effort to render them into clear Latin.
1. Lib. II, 19, p. 183, line 10:
Si portasti mortuum et non ad sepeliendum, pecuniam acquires cum injusticia et rubore; si vero ad sepeliendum, a magno viro invenies gratiam.
If you carry a dead man, though not for a funeral procession, you will acquire money combined with injustice and disgrace; but if <you carry him> for a funeral procession, you will meet with favor from a great man.
Collin-Roset matched the passage with an equivalent from the twelfth-century Arabic dreambook by 'Abd al-Rahman b. Nasr al-Shirāzī (or al-Shayzari) in
the 1664 French translation by Pierre Vattier. See, however, Drexl 84, 19-22:
ἐαν Ίδη τις, ὅτι νεκρὸν ῆρεν οὐκ ἐν τύπω ἐξοδίου, εὐρήσει πλοῦτον ἀπὸ άδικίας μετά αισχύνης εί δὲ ἐν τύπω ἐξοδίου, ἀκολουθήσει ἐξουσιάζοντι μεγίστω, ώτινι δουλεύσει, και εύρήσει χάριν έξ αύτου.
If someone dreams that he carried a dead man, though not in the manner of a funeral procession, he will find wealth through injustice with disgrace; if <he carried him> in the manner of a funeral procession, he will follow a great man of authority whom he will serve and will meet with favor from him.
2. Lib. III, 2, p. 198, 1. 11: Eodem modo ruina domus corruptionem alteruter significat (Likewise, the collapse of a house indicates some kind of ruin).
Collin-Roset matched the passage with Vattier, p. 66. See, however, Drexl 101, 22-25:
ἐὰν Ίδη τις, ὅτι συνέπεσεν ἐκ τῶν οἰκημάτων τοῖχος ἢ ἀγίς ἢ ἐκλάσθη ξύλον ή έξηλώθη θύρα ή ἐκάη, ταῦτα πάντα κρίνονται εἰς πτώσιν του οίκοδεσπότου ή των αύτου συγγενών.
If someone dreams that a wall or a terrace or an arch of a building collapses, or that a beam broke, or that a door was ripped off or burnt down, all these things are interpreted as the ruin of the homeowner or of his relatives.
3. Lib. II, 3, p. 167, l. 23: Si Deus vel sanctus adoraverit te, gratiam alicujus principis invenies; converso modo e contra (If God or a saint adored you, you will meet with the favor of some nobleman; if <things happened> the other way around, the opposite <will take place>).
4. Lib. II, 9, p. 173, 1. 19: Libros intelligere bonum est2 (Understanding <the meaning of> books is good).
5. Lib. II, 9, p. 174, 1. 3: Scribere et legere in libro vel in tabula sollicitudinem
l L'onirocrite mussulman, ou la doctrine et l'interprétation des songes selon les arabes, par Gabdorrhachaman fils de Nasar. De la traduction de P. Vattier (Paris, 1664), p. 124.
2 This phrase cannot be matched with anything from the Oneirocriticon or Artemidoros. In the Liber Thesauri Occulti it serves as a general introductory statement before proceeding to particular interpretations that correspond word for word to Drexl 7, 13-20. It is difficult to decide whether this phrase existed in Pascalis Romanus's model of the Oneirocriticon or whether it was an addition that he made, prompted by the positive interpretations of dreams about reading a book found in his source.
hujus mundi significat. Albas litteras scribere manifestationem misteriorum significat (Writing and reading in a book or a tablet indicates the anxiety of this world. Writing white letters indicates the revelation of a mystery).
6. Lib. II, 17, p. 181, I. 4: De inferno ascendere omnibus significat bonum et evasionem periculi (Ascending from the underworld indicates something good and escape from danger).
Artem. ii.55; Pack 184, 22-23: 'to 8£ ava~aivttv £~ "At8ou 8tmjrnyi]v n va 1t0ptcraµEVOV E~ foxcl'tou n voi; Kl v8uvou cr</lSEt 'tOV approcrwv (If a sick man dreams that he has found some means of escape and has ascended from Hades, it indicates that he will be saved from the utmost danger).
7. Lib. II, 18, p. 181, I. 15: Tamen si injusti et tyranni fuerint ibi, affligentur quidem modicum; verum de illis tyrannis cito sumetur vindicta3 (If unjust and tyrannical people were at that place, indeed they will soon be crushed and revenge will quickly be exacted from those tyrants).
*<sup>3</sup>*The context of this phrase in the *Uher Thesauri Occulti* is the following: "Resurrectio mortuorum in quocumque loco visa fuerit, ibi justicia erit, quoniam injusti punientur et justi salvabuntur, quoniam in resurrectione judex solus Deus erit; ideoque si visa fuerit in loco aliquo, carceris absolutionem, tribulationis liberationem, gratiam regis et habundantiam illi loco significat; significat enim illi loco pacem, reditum peregrinantium, divicias et sanitates et pregnantium mulierum facilem partum. Tamen si injuisti et tyranni fuerint ibi, affligentur quidem modicum; verum de illis tyrannis cito sumetur vindicta." The interpretations on the resurrection of the dead that precede this last phrase correspond to Drexl 3, 25---4, 14. It is possible that this phrase existed in the Greek model of the *Oneirocriticon* used by Pascalis Romanus. Cf. in Arabic lbn Shahin, no. 193: '-51..! :.raJ dJ.i j..All ~ .JLS: .J! d ~ J..l-:' J"L..:JI ~ J~I <lJI .b.........:iJ ..::.....Ll L~I ,J1 ... ;;\_,.......;..llJ ;;..i...:;JI ~WI~ \_\_.JW <l.JI .hL..u .J>--\_,..u;..... .JI.ill (If someone dreams that the dead are being resurrected and that God spreads justice among men, it means that if there were oppressed among the people of that place, God Almighty would inflict misfortune and harm to their oppressors). It is also possible that this additional phrase was added by Pascalis Romanus, in order further to analyze "injusti punientur et justi saJvabuntur" (£av yap WC:HV alitKOl, nµoopiav bcOCTO\JO'lV, Et 8£ cilitKOUµEVOl, 8tKmw6iicrovmt i:ax£w~. Drexl 3, 28---4, I). Cf. the analytical way that this Greek phrase is rendered in BN ji·an~ais 1317, Berriot, *Exposicions et significacions des songes,* p. 57 (this French version of the *Oneirocriticon* was apparently based only on Leo Tuscus's Latin translation): "Celui qui, par un songe, a veu mors resusciter, signifie que, ou leiu ou estoit la vision, sera faite justice: et se aucuns malfaiteurs y sont ou demeurent, ilz seront pugnis; et se ilz ont souffert injure et villenie, ilz, en brief temps, seront justifiez, car en la resurrection sera seul Dieu juge juste .... "
### 8. Lib. 11, 19, p. 183, I. 19:
Si quis amplexatus fuerit mortuum vel deosculatus sive Iocutus aut tetigerit eum, sanus non erit sed longevus erit et *nova etiam audiet insperata et bona,* et quicquid fuerit Iocutus vera erunt, *sed infirmanti mortem significat.*
If someone has taken hold of or kissed or talked with or touched a dead person, he will not be healthy, though he will live long. *Moreover, he will hear unexpected and good news,* and whatever [the dead man] said will come true. *However, to a sick man [this dream] indicates death.*
Collin-Roset could not find the Greek source for the italicized passages, but matched part of this quotation with Artem. ii.69, Pack 195, 22: en 'tcOv al;to1ttcmov Eicrt 1mi Ot vi:xpoi, E1ttl 1tclV'tCO<; aA.ri0fi A.eyoucrt (Furthermore, one must consider the dead to be persons worthy of credence, since they speak the truth in every case). The first half of the Latin passage can be matched word for word with an excerpt from the *Oneirocriticon* (Drexl 84, 22-24): Ei oc 1811, O'tt 1tap£A.a~E VEKpov fl roµiA.TJO'EV fl E<j>lATJO'EV fl ll'lfa'tO amou' µaKp6~to<; µi::v EO''tat, o~ &vocro<; & (If someone sees that he took hold of or talked with or kissed or touched a dead person, he will live long, but will not be healthy). Conceivably, the second half of the Latin passage also existed in the Greek version of the *Oneirocriticon* used by Pascalis Romanus, though it is missing from Drexl' s critical edition.
9. Lib. II, 19, p. 185, I. 1: Mors peregrini reditum significat (Death indicates the return of a traveler).
Artem. ii.49, Pack 181, 22-23): 'A1to0avEtv 8oKEtv ... 'tOV E1tt l;evri<; ov'ta Et<; TI]v oidav ayEt (To dream that one is dead ... signifies that a man who is abroad will return to his native land).
10. Lib. II, 19, p. 185, I. 2:
Cum noto mortuo concumbere alicui benefacere est; cum ignoto inimicum vincere. Si mortuus aliquis concumbierit cum uxore tua vel filia te vidente, ab heredibus mortui lucrum habebis.
Lying together with a familiar dead person is to confer a benefit on someone; lying with an unfamiliar dead person is vanquishing an enemy. If some dead person lay with your wife or daughter while you were watching, you will have profit from the heirs of this dead person.
Cf. Oneirocriticon (Drexl 85, 26-86, 2):
εί δε ίδη τις, ότι συνουσίασε νεκρώ γνωρίμφ, άγαθοποιήσει τους κληρονόμους αύτου, εί δε άγνωρίστω, καταπατήσει έχθρον έξουσιάζοντα μέγιστον. εί δέ τις ίδη, ότι νεκρός συνουσίασε τη γυναικι αύτου ή τη θυγατρι, ἐκ τῶν κληρονόμων του νεκρου χάριν και κέρδος λήψεται ό ιδών.
If someone dreams that he had intercourse with a familiar dead person, he will confer a benefit on his heirs; if <he had intercourse> with an unfamiliar <dead person>, he will vanquish a great and powerful enemy. If someone sees that the dead man had intercourse with his wife or daughter, the dreamer will receive favors and profit from the heirs of this dead person.
11. Lib. II, 20, p. 185, 1. 16: Mori, occidi, strangulari, suspendi, vinci tribulationem significant aliquam (Dying, being killed, strangled, hung, or defeated indicates some kind of tribulation).
Cf. Artem. ii.50, Pack 182, 17-18: απάγξεσθαι και έαυτον αναρτήσαι θλίψεις και στενοχωρίας σημαίνει (To be strangled or to hang oneself signifies oppression and distress).
12. Lib. II, 20, p. 186, 1. 16: Si quis interfecerit porcum vel lupum aut puerum, si est infirmus, cito morietur (If someone kills a pig or a wolf or a child, if he is sick, he will soon die).
13. Lib. II, 21, p. 187, 1. 18: In celum ascendere infirmanti periculum significat (To a sick person, ascending to heaven indicates danger).
14. Lib. III, 1, p. 191, 1. 16: Ignis spontanee accensus et extinctus dampnum significat (A fire lit and extinguished by itself indicates loss).
15. Lib. III, 1, p. 192, 1. 12: Major lucerna dominum, minor vero dominam domus significat (A big oil lamp indicates the master of a house, while a small one indicates [its] mistress).
Cf. Artem. i.74, Pack 80, 18-19: λυχνία <δὲ> γυναίκα σημαίνει, λύχνος δε τον της οικίας άρχοντα (A small lamp indicates a wife; a lamp, the lord of the house).
<sup>4</sup> Artemidoros, Interpretation of Dreams, trans. White rendered λυχνία as "lampstand," which is the meaning of the word in classical Greek; however, in later Greek, i.e., the kind of Greek used by Pascalis Romanus's Greek-speaking contemporaries, λυχνία can also be understood as a diminutive of λύχνος (lamp).
16. Lib. III, 1, p. 192, 1. 15: Fumus vero significat tristiciam (Smoke indicates unhappiness).
17.Lib.III, l,p.193,1.8:
Nam ardentes cerei unicuique significant bonum profectum in arte sua; extincti autem infortunium et mortem.
Lit candles indicate to someone something good achieved in his profession; however, extinguished candles <indicate> misfortune and death.
Cf. *Oneirocriticon (Vat. gr.* 573, fol. 209v):
oµoi~ EaV °i01J ni; On tUpE Kllpta Ci~Ecrta µtKpa Kat aitf]pEV UU'tcl, ytV(J)CiKEt(!) 0 'tOlOUTOi; on 8taCi'\lVTOµroi; µD.AEl anoeavEtV. El 8£ Eicrtv UJtTOµEva, EDpiicrtt xapav Kat ayaA.Aiacrtv avaA.6yroi; TWV K11pirov.
Likewise, if someone dreams that he found small, extinguished candles and took them, let such a person know that he will soon die. If they were lit, he will find joy and delight commensurate with the candles.
18. Lib. III, 2, p. 195, 1. 6:
Fulgur autem, obscurum et consuetum tempus, permutationes videntis significat. A fulgure percuti malum et manifestationem significat sceleris.
Lightning, during a dark and regular moment <for such an occurrence>, indicates changes in the dreamer's <circumstances>. Being persecuted by lightning indicates something bad and the uncovering of a crime.
#### APPENDIX 3: CHRISTIAN PASSAGES FROM AL-DĪNAWARĪ
I. The Time Required for a Dream to be Fulfilled
BN arabe 2745, fol. 38b-39a:
وقال المعبرون من نصاري والروم الرويا عند المغرب والعتملة لا تصح ولا تقبل ولا تصيير لأنها بدنية من الامـتـلا وفي ثلث الليل لأنهـا من البطنة و الففلة وفي نصف الليل ولم يكن صاحبها ممتليا تخرج بعد خمسة سنين وفى الثلث الاخير من الليل تصح من شهر الى سنة وعند طلوع الفجر الاول تخرج من شهر و عند طلوع الشمس تخرج في ذلك اليـوم وكـذلك في السـاعـات الاقـرب فالاقرب من النهار .
The dream interpreters of the Nasara and the Rum said: "The dream that is dreamt at sunset and dusk (عت ) is neither sound nor guaranteed nor does it come to pass, because it is a corporeal dream caused by the fullness of the stomach. < A dream dreamt> during the first third of the night <is likewise> unsound because it is caused by overeating and carelessness. < A dream dreamt> during the second third of the night, if the dreamer did not have a full stomach, is fulfilled after five years. < A dream dreamt> during the last third of the night is fulfilled within a month or a year. <A dream dreamt> at early dawn is fulfilled within a month or a week, at daybreak it is fulfilled within a day or a week, and at sunrise it is fulfilled that same day; <the time span for the fulfillment of dreams is> likewise for the nearest hours, that is, the hours nearest to daytime."
In the Oneirocriticon, the time spans for the fulfillment of a dream are given units of twenty with a twenty-year maximum (cf. Drexl, pp. 240-41). The units of twenty have a theological significance, as we read in al-Dinawari, BN arabe 2745, fol. وصحت رويا يوسف ومحمد صلى الله عليهما بعد :arabe 2745, fol. 29a . عــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ God be upon him, was fulfilled after twenty years2). This means that the time
<sup>1</sup> References are to BN arabe 2745. Whenever the text of this manuscript presented problems, I introduced corrections on the basis of Esad Efendi 1833. Passages from the Oneirocriticon that discuss the same dream symbol are quoted in full unless they were longer than five lines; longer passages are summarized and reference is made to the critical edition.
<sup>2</sup> In the Old Testament Joseph is said to have been seventeen years old before he had the two dreams indicating that he would rise in rank above his family and was subsequently sold in Egypt.
span for the fulfillment of a dream given in the Oneirocriticon definitely originated in Muslim sources.
II. Interpretations According to the Christians (Nasārā)
Dreaming of Christ (53b)
قالت النصارى من راء <عيـسـي> في منامـه فـانـه لا يصـيـبـه مكره فـي تـلك السنة وان طلب طلبا اصابه ومهر.
The Nașară said: "If someone sees Jesus in his dream, no harm will befall him throughout that year, and if he was trying to obtain something, he will achieve it and be dexterous."
The interpretations in the Oneirocriticon (Drexl 6, 22s7, 2; 105, 13-24; 106, 5-12) are different.
Dreaming of the angels (58a-b and 60a)
The Nasārā said: "If someone dreams that the archangel Michael3 laughed in his presence, he will gain profit, prosperity, and well-being."
Two years passed between Joseph's interpretation of the dreams of his co-prisoners, the pharaoh's cup-bearer and baker, and his interpretation of the pharaoh's dreams. Joseph is said to have been thirty years old when he stood before the pharaoh (Genesis 41:46). Seven years of abundance and two years of famine passed before his brothers came to Egypt. There are unspecified periods of time between these events, but twenty years is not mentioned and cannot be inferred as having passed between Joseph's dreams and their fulfillment. According to Jewish exegetical tradition represented by the Babylonian Talmud the time that elapsed between Joseph's dream and its fulfillment was twenty-two years; see Tractate Berakot, trans. A. Cohen, pp. 358-59 (fols. 55a-55b).
<sup>3</sup> Both BN arabe 2745 and Esad Efendi 1833 have ملكا من مــلائكـة المـيــاه (an angel from among the angels of the water), which is not a concept either in Christianity or Islam, instead of الباب ج من الفصل ج فى رؤيا the archangel Michael). However, the title of the bāh is) ميكائلا مــيكائل عليه (Paragraph 3 of chapter 3 on dreaming of the archangel Michael, may peace be upon him).
The Nasara said: "Whoever dreams of a spiritual angel, he will obtain power and blessings and profit and (?); however, at the end of his lifes his fortunes will decline and distress will befall him because of calumniators and slanderers."
The interpretations in the Oneirocriticon are different, with no reference to laughing angels; cf. Drexl 5, 18-6, 14.
Eating human brains (75b)
آ (The Nașārā) قال النصاری من رای انه یاکل دماغ انسان فانه یموت عاجلا said: "Whoever dreams that he is eating human brains will soon die.") The interpretation in the Oneirocriticon (Drexl 25, 12-16) is different:
Έάν τις ίδη κατ' οναρ, ότι έφαγε μυελον κεφαλής ἀνθρώπου γνωρίμου, εύρήσει τον πλούτον και το χρυσίον αύτου παν και κληρονομήσει αύτόν εί δε αγνωρίστου, εύρήσει πλούτον άπο μεγιστάνου τινὸς καὶ χρυσίον πλεῖστον άκόπως.
If someone dreams that he ate the brains from the head of a person he knew, he will gain that man's entire wealth and gold and will be his heir. If <he ate the brains> of someone unknown, he will receive without effort wealth and a large amount of gold from a nobleman.
Dreaming that your teeth broke (83a)
وقالت النصاري من راي كان اسنانه انكسرت فانه يموت احد اقربائه او اصدقائه وربما مرض ومات من ذلك المرض
The Nasara said: "If someone dreams that his teeth are broken, one of his relatives or friends will die, or perhaps will fall ill and die from that illness."
There is a similar interpretation in the Oneirocriticon (Drex1 38, 24-39, 5 and 35, 19-20). This interpretation is commonplace, since al-Dinawari also
<sup>4</sup> Problematic text. سترة = cover.
S BN arabe 2745, fol. 60a : امر ه : Esad Efendi 1833, fol. 35a, I. عمر ه
attributes it to the Muslim dream interpreters (BN arabe 2745, fol. 82a-b); it also occurs in Artemidoros i.31, Pack 37, 14-38, 6).
Dreaming that your upper arm broke (89b)
قالت النصاري من راي ان عضده انكسرت فهو موت صاحب الرؤيا او مصيبة من غم وشدة وبلايا ومصيبة.
The Nașără said: "If someone dreams that his upper arm broke, it indicates the death of the dreamer or a calamity of sorrow, distress, tribulations and misfortune."
The interpretation in Oneirocriticon (Drexl 44, 17-20) is different:
Έάν τις ϊδη κατ' όναρ, ότι οι βραχίονες αύτου ήδυνάτισαν ή έλεπτύνθησαν ή ἐκλάσθησαν, εἰ μέν ἐστι βασιλεύς, ἔσται τὸ πάθος αὐτοῦ εἰς τὸν στρατὸν αύτου και είς νόσημα και θλίψιν έλεύσεται ή ό νίος ή ό άδελφος αύτου.
If someone dreams that his arms became weaker or leaner or that they broke, if he is king, what he suffered <in the dream> refers to his army, while either his son or his brother will fall ill or will be distressed.
Also Drexl 44, 25-45, 4:
ἐάν τις ἴδη κατ' ὅναρ, ὅτι <οί βραχίονες αὐτοῦ> ἐλεπτύνθησαν ή ἐκλάσθησαν ή έψύγησαν ..., απολέσει δε τον άδελφον ή τον πρωτεύοντα δούλον αύτου.
If someone dreams that <his arms> became leaner or that they broke or that they froze ... , he will lose his brother or his most important servant.
Dreaming of milking a she-camel and drinking from her milk (104b)
قالت النصاري من راي انه حلب ناقة شـرب من لبنها فـانـهـ يـتـزوج امـراة صالحة فان كان الراى مستورا فانه ولد له غلام ويكون له فيه بركة.
The Nasārā said: "Whoever dreams that he was milking a she-camel and drinking from her milk will marry a righteous woman. And if the dreamer had a blameless record, a boy will be born to him, and there will be blessing because of him."
The interpretation in the Oneirocriticon (Drexl 186, 12-16) is different:
εί δε ίδη, ότι τῆς τοιαύτης γενεάς ἐκτήσατο θηλείας καμήλους καὶ ἤμελχεν αυτάς, εύρήσει ἐξουσίαν καὶ κυρείαν κατὰ τοῦ ἔθνους τῶν Σαρακηνῶν· ὁ βασιλεύς έαν ίδη τουτο, άλλότριον έθνος ύποτάξει άναλόγως της ύποταγής της καμήλου.
If he dreams that he acquired she-camels of such <i.e. Arabic> stock and was milking them, he will receive power and authority over the nation of the Saracens. If the king dreams this, he will subdue a foreign nation by analogy to the submissiveness of the camel.
Cf. also Drexl 187, 26:
ἐαν ϊδη τις, ὅτι γάλα καμήλου ἔπιεν, εὑρήσει ἀπὸ ἐξουσιαστοῦ πλουτον καὶ φόβον και νόσον.
If he dreams that he drank the milk of a camel, he will gain wealth, fear, and illness from a nobleman.
Drinking the milk of a lioness (104b)
قالت النصاري من شربه (=لن اللبوة) نال مالا من سلطان جدار او من كذ ﮓە
The Nasara said: "If someone dreams that he was drinking the milk of a lioness he will obtain money from an omnipotent king or from the toil of his hand."
A similar interpretation in the Oneirocriticon (Drexl 219, 19-20) is:
εί δε ίδη, ότι γάλα λέοντος έπιεν, εύρήσει πλούτον και χαράν από βασιλέως.
If he dreams that he drank the milk of a lion, he will find wealth from a king together with fear and joy.
This interpretation is commonplace, since al-Dinawari also attributes it to the Muslim dream interpreters (BN arabe 2745, fol. 104b):
فلبن اللبوة مال لشاربه وظفر بعدوه ومعاداة السلاطين والناس.
The milk of the lioness for whoever drinks it means money, triumph over his enemies, and the enmity of kings and commoners.
Drinking the milk of a bitch (105a)
وقالت النصارى من شربه (=لبن الكلبة) نال مقدرة ورياسة على اهل بلده.
The Naṣārā said: "If someone drinks <the milk of a bitch> he will attain power and authority over the people of his country."
The interpretation in Oneirocriticon (Drexl 225, 20) is different:
εί δε ΐδη, ότι έπιεν κύνειον γάλα, εύρήσει φόβον και νόσον μακράν (If he dreams that he drank the milk of a bitch, he will find fear and prolonged illness).
#### Eating clarified butter (106a)
The Nașară said: "fre Nașără said: "If someone eats or gathers clarified butter, a boy will be born to him.")
The interpretation in the Oneirocriticon (Drexl 190, 20) is different: ei δέ ϊδη, ότι τρώγει βούτυρον βοὸς θηλείας, εὑρήσει πλουτον ήδυν χρόνιον (If he dreams that he is eating butter <from the milk> of a cow, he will find pleasant and long-lasting wealth); cf. also Drexl 196, 20: εί δε [ήσθιεν] αρτον ψυχρόν μετά βουτύρου, εύρήσει πλούτον και καιρον άγαθον αναλόγως της βρώσεως (If <he eats> cold bread with butter, he will find wealth and pleasant times in proportion to how much he ate).
Nose bleeding (107b)
قالت النصارى من راى انفه راعفا نال كنزا ومالا عظيما. said: "Whoever dreamt that his nose was bleeding will obtain a treasure and a vast amount of money").
The interpretation in the Oneirocriticon (Drexl 61, 15-19) is different:
Τὸ αῖμα δύναμις καὶ ζωὴ τοῦ ἀνθρώπου ἐστίν. ἐαν ἴδη τις, ὅτι τὸ αἱμαι αὐτοῦ έκ της ρινός ή της κεφαλής έρρευσεν, είς τὴν κεφαλὴν ήτοι τὸν ὑπερέχοντα αύτου νοείτω τὴν ἐπιζήμιον ἔξοδον.
Blood is a man's strength and life. If someone dreams that blood flowed out of his nose or head, let him reckon this as an injurious expenditure® incurred by his leader, that is, the person who has authority over him.
Becoming a priest or a monk (144a)
وقالت النصاري من راي كانه تحول قساااو راهيا فانه يورث تمنا حسنا لكن يعسر عليه شانه ويضيق عليه رزقه وربما نفع التاويل لغيره.
The Naṣārā said: "Whoever dreams that he became a priest or a monk will inherit something valuable and beautiful; however, his affairs will bear down hard on him and his means of livelihood will shrink. It is possible that the interpretation <of this dream> will benefit someone around him."
The interpretation in the Oneirocriticon (Drexl 7, 6-12; 7, 21-26; 9, 11-16; 104, 30-105,7), is different: there becoming a priest means that the dreamer will obtain honor, power, and offices, whereas becoming a monk means he will lose his social standing and possibly die in poverty.
The position of a judge (154b)
قالت النصاري من راي وجه القاضي مستبشرا طلقا فانه ينال بشري وسرورا فان راي موضع قاض نال فزعا وخصومة وغلب.
The Nasara said: "Whoever dreams of the happy and cheerful countenance of a judge will receive good news and joy. If he dreams of the position of a judge, fear and a lawsuit will befall him, but he will win." ?
<sup>6</sup> The word "Egooo" in medieval Greek can also mean "outcome" or "death" (cf. Lampe and Sophocles, s.v. "E50805"); however, rendering it as "spending" is supported by the interpretation of blood as money in the Oneirocriticon (chap. 103, Drexl 61, 1-13) as well as by the phrase that immediately follows the one quoted above (Drexl 61, 19-20): ¿àv δè ἐκ τῆς σαρκός έρρευσε < τὸ αῖμα αὐτου> πλεῖστον πτωχεύσει καὶ όλιγόβιος ἔσται (If <his blood> flowed out of his flesh he will become poorer and will not live long).
<sup>7</sup> Depending on how one choses to vocalize the word غلب , its meaning could be either "to win" (ghalaba) in the active voice, or "to be overcome, to be conquered" (ghuliba) in the passive voice.
The interpretation in the Oneirocriticon (Drexl 13, 6-14) is different.
Eating the flesh of a youth (157a)
قالت النصارى من راى انه ياكل لحم صبى فانه ينال عطا من رجل عظيم ومرتبة ونعمة وشرفا.
The Nasara said: "Whoever dreams that he ate the flesh of a youth will receive a gift from a powerful man, and rank, prosperity and honor."
The interpretation in the Oneirocriticon (Drexl 53, 13-17) is different:
έάν τις ίδη, στι κρέατος ανθρώπου μετέλαβεν, έκ λοιδορίας πλουτήσει αναλόγως του πλήθους, ού έφαγεν. ἐάν τις ίδη, ὅτι όπτου κρέατος ἀνθρώπου μετέλαβεν, έκ τόκου αδικον πλούτον κταται. πτωχος έαν ίδη τουτο, έξ ύπερέχοντος τιμωρηθήσεται. γυνὴ ἐαν Ίδη τουτο, ἀπὸ κλοποσπορίας ἀσθενη τέξει τέκνα.
If someone dreams that he partook of human flesh, he will become wealthy by reviling <others> by analogy to the quantity that he ate. If someone dreams that he partook of cooked human flesh, he will acquire unjust wealth through usury. If a poor man dreams this, he will be punished by a superior. If a woman dreams this, she will give birth to sickly children through unlawful conception.
Holding a stretched bow (191b)
وقالت النصارى من راى في منامه كان بيده قوسا موترة فانه يولد له غلام قوى صاحب كتابة ورسالة.
The Nasara said: "Whoever sees in his dream that a stretched bow was in his hands, a boy will be born to him who will be strong, and will know how to write and compose epistles."
The interpretation in the Oneirocriticon (Drexl 204, 24-25) is different: εί δέ ευρε τόξον τεταμένον, εύρήσει όδον και ύποστροφήν χαρίεσσαν (If he finds a strung bow, he will go away on a trip and will have a happy return). Cf. also Drexl 114, 17: Το δε τόξον κατ΄ Αἰγυπτίων κρίσιν γυναικα ή θυγατέρα σημαίνει (A bow, according to the interpretation of the Egyptians, signifies a wife or a daughter).
#### Holding an arrow (192a)
. قالت النصارى من راى بيده سبهما فانه ينال ولاية وعزا ومالا . Naṣārā said: "Whoever dreams that he is holding an arrow in his hands will obtain sovereignty, power, and money").
The interpretation in the Oneirocriticon (Drexl 204, 7-8) is different.' Eàv ίδη τις, ότι συνάγει βέλη ήτοι σαγίττας, κατά άνθρώπων θάνατον μελετήσει (If someone dreams that he is collecting arrows, he will plan the death of men).
#### Holding a dagger (194a)
قالت النصارى من راى بيده خنجرا نال مالا وغنا
The Nasara said: "Whoever dreams that he was holding a dagger in his hands will obtain money and wealth."
The interpretation in Oneirocriticon (Drexl 205, 5-8) is similar:
Έαν ίδη τις, ότι είληφε παρά τινος σπάθην ή μάχαιραν ή έργον από σιδήρου, τέμνον ή μὴ τέμνον, ἢ σίδηρον ἀργόν, εἰ μέν ἐστι πλούσιος, πλέον πλουτήσει, εί δε πτωχός, πλουτήσει αναλόγως ου έλαβεν.
If someone dreams that he received from someone a sword or a knife or a metal object, whether sharp or not, or unwrought iron, if he is wealthy he will become even wealthier, and if he is poor he will become wealthy in proportion to what he received.
The interpretation of sharp metal objects as money is commonplace. Ibn Shāhīn attributes it to al-Kirmānī, the author of the oldest Arabic dreambook, and Ja'far al-Sadiq, the Twelfth Imam (nos. 5137 and 5140 bis).
Cuirass (195a)
وقالت النصاري من راي جوشنا فانه يتزوج امراة قوية حسنة ذات شهوة ومال يورڻه المتزوج بها اجمالا وان کان فقير استغني. ^
فانه يتزوج امراة قوية حسنا ذات شهوة وقد يورث اللتزوج : BN arabe 2745, fol. 195a فانه يتزو ج امراة قوية : Esad Efendi 1833, fol. 114bt فانه يتزو ج امراة فقيرا استخفني
#### APPENDIX 3
The Nasara said: "Whoever dreams of a cuirass will marry a powerful, beautiful and passionate woman with money that whoever marries her will inherit altogether; and if he was poor he will become rich."
The interpretation in the Oneirocriticon (Drexl 113, 27-114, 1) is different:
έαν Ίδη τις, ότι ένεδύσατο ή εύρε θώρακα ήτοι λωρίκιον, χαρήσεται έπλ απωλεία των έχθρών αυτού και πλουτήσει αναλόγως του βάρους του θώρακος.
If someone dreams that he put on or found a breastplate, he will rejoice over the destruction of his enemies and will become wealthy by analogy to the weight of the breastplate.
Cf. also Drexl 204, 21-22:
Έαν ίδη τις, ότι ευρεν ή εὐπόρησε θώρακα ήτοι λωρίκιον, εύρήσει χρυσίον και πλούτον μετά τόλμης αφόβως.
If someone dreams that he found or procured a breastplate, he will find through his daring gold and wealth without fear.
Coats of mail for arms = μανικέλια (195b)
قالت النصاري من راي ساعدين فانه يصحب رجلين قويين عظيمين وربما و قع التاويل على انن
The Nasara said: "Whoever dreams of mail armor for the arms will become closely associated with two powerful and mighty men; and probably the interpretation of this dream refers to his son or brother."
The interpretation in the Oneirocriticon (Drex1 114, 7-8) is:
εί δε <ένεδύσατο> τα λεγόμενα μανικέλια και περικνημιδας, εύρήσει χαραν και στερρότητα ἐν τοῖς δούλοις αὐτου.
If <he put on> a coat of mail for the arms and legs, he will find joy and steadfastness in his servants.
حسنة ذات شهوة ومال دورث التزويج لها وجمالا وان كان فقدرا استغنى.
#### Hemorrhage (203b)
```
قالت النصاري من راي انه يخرج الدم من جسده ورای جراحات بدنه فانه
يصيب صحة جسد وزيادة مال
```
The Nasara said: "If someone dreams that blood is flowing from his body and sees the wounds of his flesh, he will obtain good bodily health and increase his money."
The interpretations in the Oneirocriticon (Drexl 61, 3-5) in this case are similar:
έάν τις ίδη και όναρ, ότι από πληγής ή ρεύσεως καταρρεί αΐμα έρυθραίνον τα ίμάτια αύτου ή τας σάρκας, εύρήσει χρυσίον ανάλογον του καταρρνέντος αίματος.
If someone dreams that because of a wound or a discharge blood is flowing and reddening his clothes or flesh, he will find gold in proportion to the <amount of> blood that flowed.
The interpretation of blood as money is commonplace. In the same chapter, al-Dīnawarī also attributes it to the Muslim dream interpreters (BN arabe 2745, fol. 203b). The interpretation of blood as money occurs in Artemidoros as well (i.33, Pack 42, 6; i.61, Pack 67, 24).
#### Being crucified (205a)
قالت النصاري من راي انه مصلوب على سور الدينة والناس ينظرون الليه ينال رفعة وسلطانا ويصير الاقوياء والضعفاء تحت يديه فان سال منه الدم فان رعيته ينتقعون به ومن راي كانه ياكل لحم مصلوب فانه ينال مالا ومنفعة من جهة احد الروساء.
The Nasara said: "If someone dreams that he was crucified on the walls of a city and the people were looking at him, he will obtain a lofty position and sovereignty and both the powerful and the weak will be in his power. And if blood was streaming from him, his subjects will benefit from him. And if someone dreams that he was eating the flesh of a crucified person he will obtain money and profit from one of the leaders."
Interpretations in the Oneirocriticon (Drexl 54, 20-55, 3) that are similar are more or less commonplace, since the Christian interpretation of the crucifixion in al-Dinawari is similar to the Muslim one (BN arabe 2745, fol . 204b-204a).
The ultimate source of the interpretation of crucifixion as meaning a lofty position, glory, and wealth is in Artemidoros ii.53, Pack 183, 6-21, and especially iv.49. Pack 276, 6:
σημαίνοντος του δοκείν ἐσταυρώσθαι δόξαν καὶ εὐπορίαν δόξαν μὲν διὰ τὸ ύψηλότατον είναι τον έσταυρωμένον, εὐπορίαν δὲ διὰ τὸ πολλοὺς τρέφειν οίωνούς.
Crucifixion dreams signify honor and wealth; honor, because the crucified person is in a very high position, and wealth because he provides food for many birds of prey .
#### III. Interpretations According to the Byzantines (Rūm)
The hair of the dreamer's wife was cut (78a)
. وقالت الروم من راي ان ذوائب امراته مقطوعة لم تلد ولد ابدا (The Rūm said: "If someone dreams that locks of his wife's hair were cut off, she will never give birth to a child").
There is a different interpretation in the Oneirocriticon (Drexl 217, 10-26), where cutting a woman's hair is interpreted as divorce or death.
Suckling the milk of your wife (104a-b)
قــالت الروم من امــتــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ dreams that he was suckling milk from his wife will obtain money and profit). The interpretation in the Oneirocriticon (Drexl 79, 13-19) is similar:
εί δε ίδη, ότι ἐκράτει τῶν μασθών αὐτῆς καὶ ἐξέμαξε γάλα, χαρὰν μεγάλην κομιείται έν τω ένιαυτώ ἐκείνφ, ἀλλὰ καὶ ή γυνὴ ἐκείνη μεγάλως συγχαρήσεται αύτων εί δε τουτό τις του κοινού λαου ίδη, εύρήσει πλούτον έκαστος ἐν τῷ ἐπιτηδεύματι αὐτοῦ εἰ δὲ δοῦλος ἴδη τοῦτο, ἐλευθεροῦται ταχέως, εἰ δὲ πτωχός, πλουτήσει.
If he dreams that he held her breasts and squeezed out milk, he will have great joy during that year, and that woman will also have a great share in <the dreamer's>
<sup>&#</sup>x27; Esad Efendi 1833, fol. 65a: قالت الروم من امتص ابن امراه دامراته؟ < نال مالا وربحا. (The Rum said: "If someone dreams that he was suckling the son of his wife, he will obtain money and gain.").
joy. If a commoner dreams this, he will find wealth, each in his own trade, for one year. If a slave dreams this, he will soon be freed, and if a pauper, he will become wealthy.
The interpretation of a woman's milk as money is commonplace, since al-Dinawari also attributes it to the Muslim dream interpreters (BN arabe 2745, fol. 104a). It also occurs in Artemidoros i. 16, Pack 25, 4-6. Cf. also a different interpretation in Drexl 78, 10-11: ἐάν τις ἴδη, ὅτι θηλάξει ἄρσεν ανθρωπός τις ή αυτός, δέσμιος είς φυλακήν γενήσεται (If someone dreams that he or some other man was nursing a male child, he will be placed in bonds in prison).
Milking a cow and drinking her milk (104b)
قالت الروم من راي انه حلب بقرة و شرب لبنها فانه ان كان عبدا اعتق وان کان فقیرا استغنی.
The Rum said: "If someone dreams that he milked a cow and drank her milk, if he is a slave he will be freed and if he is poor he will become rich."
Similar interpretations in Oneirocriticon (Drexl 190, 22-28):
εί δε ίδη, ότι αμέλγει βουν και πίνει του γάλακτος, εί μέν έστι δούλος, έλευθερουται, και ϊσως κληρονομήσει τον δικον του δεσπότου ή τὴν δεσποτείαν αυτού λήψεται, εί δε του κοινού λαου, εύρήσει χαράν και άγαλλίασιν μείζονα ό βασιλεύς έαν ίδη τουτο, είς χαράν και άγαλλίασιν αύτου έσται τούτο.
If someone dreams that he milks a cow and drinks her milk, if he is a slave, he will be freed, and will possibly inherit the household of his master or receive his <master's> authority; if he is a commoner, he will find great joy and exultation; if a king dreams this, it will be <a sign of> joy and exultation.
This interpretation is commonplace. According to al-Dinawari, the Muslims also interpret the milk of the cow as money (BN arabe 2745, fol. 104b).10 A Muslim interpretation that is very close to the one in the Oneirocriticon can be found in Ibn Shahin, no. 5693:
<sup>10</sup> وقال المسلمون لبن البقرة خصب السنة ومال حلال وفطرة فى الدين (10 "The milk of the cow means abundance throughout the year, lawful money and a natural disposition for religion.")
#### APPENDIX 3
ومن راي انه يحلبها ويشرب من لبنها فإن كان عبدا يعتق ويتزوج بنت مولاه وان کان فقیرا یستغنی وان کان غنیا یزداد غنی وان کان حقیرا یصیر عز سز ا
And if someone dreams that he milks <a cow> and drinks her milk, if he is a slave he will be freed and will marry the daughter of his master. If he is poor he will become rich, and if he is rich his wealth will increase. If he is one of the lowly people, he will become one of the mighty. 1
Collecting cow dung (117b)
قالت الروم من راي كانه يجمع اختاء البقر و ارواڻ البعير فانه ينال مالا من رجل شريف فان التقط اختاء التور نال مسرة فان جلس على السرقين نال مالا من جهة بعض اقربائه.
The Rum said: "Whoever dreams that he collected the dung of cows and the droppings of goats will obtain money from an honorable man. And if he gathered the dung of bulls, he will obtain happiness. And if he sat on the manure he will receive money from some of his relatives.
The interpretation in the Oneirocriticon (Drex1 190, 28-29) is similar:
έαν Ίδη τις, ότι συνάγει κόπρον βοείαν, εύρήσει πλουτον άπο μεγιστάνων ανάλογον τῆς κόπρου.
If someone dreams, that he is collecting cow dung, he will find wealth from noblemen by analogy to the dung.
Al-Dinawari does not give Muslim interpretations for this particular dream. However, the interpretation of dung as money can be found in the dreambook of Ibn Qutayba (Yahuda ar. 196, fol. و البـقـر سـنـون ولــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ والعذرة مال الموال والت كلها اموال والعذرة مال (The cow is years and its flesh is money, and likewise its dung, and all kinds of manure is money, and human excrement is also money).
<sup>11</sup> In this paragraph, Ibn Shāhīn repeats whole phrases from the dreambook of Ibn Qutayba (d. 889). Cf. Yahuda ar. 196, fol. من راي انه يحلب بقرة ويشرب لبنها استغنى إن كان فقيرا وارتفع شآنه وان کان غنيا ازداد غنا الی غذاه وعزه milking a cow and drank her milk, he will become rich if he is poor, and his social standing will become loftier. If he is wealthy, he will add more riches to his wealth and power).
#### CHRISTIAN PASSAGES FROM AL-DĪNAWARĪ
#### IV. Interpretations According to Both the Christians (Nasārā) and the Byzantines (Rūm)
Adultery/fornication [zina] (168a)
وقالت النصاري والروم من راي انه زنى بزانية نال شرا اوفقنة."
The Nasara and the Rum said: "Whoever dreams that he was committing adultery with a harlot / an adultress, evil or temptation will befall him."
The interpretation in the Oneirocriticon (Drexl 77, 9-13) is different:
καὶ ὅσον συνουσιάζει τις κατ᾽ ὅναρ ἐταίραις γυναιξί, τοσούτον πλουτήσει φιλόκοσμος ών' ἐπὶ δὲ τῶν εὐλαβών καὶ ἀναχωρητῶν τοῦτο εἰς θλῖψιν κρίνεται. ἐαν ϊδη τις, ὅτι συνουσίασε πόρνη, αὐξήσει μὲν εἰς τὸν πλοῦτον αὐτοῦ, ἐξ άδικίας δέ.
In proportion to how much someone copulates with courtesans in his dream, he will become that wealthy, if he is a man of the world. But for pious men and ascetics this <dream> is interpreted as sorrow. If someone dreams that he copulated with a prostitute, his wealth will grow, but through injustice.
Deflowering a slave girl (181a)
$$\text{Lúl}\_{\text{"\raisebox{-0.5ex}\text{\raisebox{-0.5ex}\text{\raisebox{-0.5ex}}{\$}}}} $$
The Nasara and the Rum said: "Whoever dreams that he deflowered a virgin slave girl, harm and adversity will befall him throughout that year."
The interpretation in the Oneirocriticon (Drex1 79, 6-9) is different:
ἐάν τις ἴδη, ὅτι συνουσίασε παρθένω, εἰ μέν ἐστι βασιλεὺς, εὐρήσει χαρᾶς και ἐπιθυμίας τῆς πρώτης αὐτου τὸ πλήρωμα, εἰ δὲ τοῦ κοινοῦ λαοῦ, εὑρήσει πλούτον έμφοβον άπο έξουσίας.
If someone dreams that he copulated with a virgin, if he is king, he will find joy and fulfillment of his greatest desire; if he is a commoner, he will find wealth together with sorrow through <someone in> authority.
<sup>12</sup> I chose the reading of Esad Efendi 1833, fol. 97a, 1. 11, instead of the reading in BN arabe 2745: قـال سـرا وقـبـه [sic].
### APPENDIX 4: WORDS IN THE ONEIROCRITIC ON UNATTESTED IN OTHER GREEK TEXTS
This list contains the words marked by Drexl as unattested in Greek dictionaries in his Index Rerum et Verborum Potiorum appended to the critical edition (pp. 243-65). The explanations in Latin were copied from Drexl. The explanations in Greek were copied from Koukoules' review of the critical edition in Laographia 9 (1926), pp. 285-92. The explanations in English are mine. The following abbreviations were used:
I did not indicate that the word was included in K, A or T if the only reference given was the Oneirocriticon. A sometimes included the word without reference to a text, or gave a single reference to the Oneirocriticon but implied that the word also occurs elsewhere. In this case I indicated the word was included in the dictionary. Finally, I did not indicate that the word is included in either K or A whenever the meaning given by the dictionary did not coincide with the
meaning in the Oneirocriticon. The last 47 words of the list (from opxnto on) could not be checked with K, since the dictionary is still incomplete.
APPENDIX 4
δροσία (=δρόσος) 177, 20 KT K δωρεαστικώς 117, 19Α \*čykanvoc fumificus 119, 27 είσκόσμισις (nisi forte legendum est σκορπισμός seu σκόρπισις) = deminutio 126, 2. 129, 5. The text should be corrected to read either είσκόμισις (see K, s.v. "είσκομίζω") or είς σκορπισμόν, a reading which is supported by an equivalent Arabic passage. ἐκριζώνω 160, 6Α KA \*ἐμπήλωσις 138, 24 \*ἐνδοξέω 42, 9Α έξηλόω (cf. neogr. ξηλώνω) 101, 24· 212, 6 KA έπιλωβάομαι=lepra inficior 35,4 F έπτησις (όπτησις) 98, 6A \*εύαρέσκομαι 207, 21Α ευπόρησις 140, 17A △ \*ἐχθρομαχία 195, 28 \*ζατρικίζω schachis ludo 192, 28 ζούπα sagum 177, 1 Κ (Προσθήκες, τόμος 8) \*ζυγοστάθμιον = bilanx 161, 12 \*ήπατοπνεύμων = iecur et pulmo (cf. τα σκωτοφλέμονα, vocem Cretensibus adhuc usitatam, ortam ex o(v)kwn = ทั้กตุ et φλεμόνι = πλεμόνι = πνεμόνι = πνεύμων) 51, 16. 191, 6 \*θήλασις (θήλανσις) 76, 10Α ίδιοθανατέω (ίδιοθανατόω) 96, 11· 110, 2 T ίστούργημα 236, 10Α △ \* κάμβα (καμβά, καμβάν, καμα) vestimenti genus 114, 26A καυστηρία 128, 5A T \*καυστρία (καύτρια, καυτριά, καύστρια) = ustio 46,5 · 173, 27 κιτρινίζω 197, 11 ΚΛ \* κλοποσπορία = concubitus furtivus 53, 18 70, 1 κόχλα ("cocla" in versione Lat .; cf. κόχλαξ=calculus, lapillus) 33, 18 Koukoules, p. 291: "Κόχλα είναι ούχι calculus, lapillus, άλλά βαφικὴ ὕλη, κόνις θειούχου άντιμονίου μετά αίθάλης και συνικής μελάνης, δι' ής έβαφον τας όφρυς και τας βλεφαρίδας των όφθαλμών, άλλως χολλάς και κόχλος κατά τον μεσαίωνα λεγομένη." κρίσιμον (κρίσιμο, το)=decretum 91, 6 K \*λειποσωμασία (λειποσωματία) 67, 16 \*λεοντότεκνον 220, 6Α
\*λυπόπουλος (λυποπούλι) avis 235, 9. 235, 18.
Drexl adds in parenthesis: "σκώψ an κολοιός? Cf. Herwerden s.v. κολοιός et vocem κλαψοπούλι apud Hepitem, Lexicon Hellenogall." In the abridgment of the Oneirocriticon in Paris. Suppl. gr. 690 we read the following (fol. 129r, col. 1): O κόραξ γέροντα δηλοί παλαιόν, ξένον, πτωχὸν καὶ λαίμαργον, παραπλησίως δὲ και ό έποψ. This corresponds to Drexl 235, 10-18, which concludes with the interpretation of λυπόπουλος (235, 17-18). Based on this evidence, it is reasonable to suppose that λυπόπουλος is έποψ, a bird indeed known to have a mournful song).
\*όψησις=έψησις 159, 24Α \*όψητος =έψητος 159, 23Α. 206, 22Α \*παλουδάκιν (παλουδιν) dulcedo 198,4 \*περισκέλισμα tegumentum femorum interius 70, 3· 88, 8· 115, 17.20· 116, 19 \*περσικομανίκιον (περσικομανιάκιν)=manica Persico more confecta? 179, 1 Koukoules, p. 291: "Περσίκια ήσαν μεταξωτα ύφάσματα, ἐξ ών κατεσκευάζοντο χιτώνες χρυσοπέρσικος μάλιστα χιτωνίσκος άναφέρεται και ύπο Λέοντος του Γραμματικού (Χρον. 224.17)."Οταν λοιπόν ό ' Αχμέτ λέχη ότι "το περσικομανίκιον του ύποκαμίσου είς κατάστασιν και σύναξιν του βίου κρίνεται" έννοει όχι manicam Persico more confectam, άλλά το μανίκιον του περσικίου τούτου"). In later Greek astronomical texts the word περσίκιον means "pocket"; see Tihon, "Tables islamiques à Byzance," no. 6. πέτασις 122, 27Α T T πίανσις pinguedo 68, 7. 190, 11p. \*ποδόλωρον (ποδόλουρον) pedica 233, 1 \*ποδόρτιον soccus 176, 4 \* ποιόω intr. = qualitatem habeo 211, 24- 213, 10 sq. \*πολυεξόπλιστος 226, 16 \*πολυζωήμων 14, 6Α \*πολυζώητος 9, 21 · 132, 12 \*πολύοπλος 228, 22 πολυτεκνόω 49, 2A △ T προκόμισις exsequiae 83, 24
\*σωβάριον supparum? 115, 5A
\* τετραπέδιλος (τετραπέδουλος?) quattuor soleis ferreis munitus
χάσδιος (χάσδεος, χασδεός, χασείδιος) 115, 3· 170, 13· 177, 8· 204, 15 The translation suggested by Drexl is "textum sericum." However, the Oneirocriticon itself clarifies the meaning of the word twice: εί δε ίδη τις το τοιουτον <καβάδι>, ὅτι χάσδιον ήν, εύρήσει πλούτον ἐξ ἀνδρῶν πονηρῶν διὰ τὸ ἐς ἐρίου κυνοποτάμου είναι αύτο (115, 2-4); τα δε άπο έρίου κάστορος ή των αλλων ήτοι χάσδια λωρωτά ... (170, 12-13). The word khazz in Arabic has a number of meanings. Lane, Arabic-English Lexicon, gives: " خَنّ A certain kind of cloth ... , well known ... , woven of wool and silk: and also a kind of cloth entirely of silk; ... or it is the name of a certain beast [thought by Golius to be the beaver]: and afterwards applied to the cloth made of its fur ... " The last definition confirms the meaning of χάσδιος as textile έξ έρίου κυνοποτάμου and ἐξ ἐρίου κάστορος given in the Oneirocriticon. T △
yıxiov digiti extrema pars
(cf. vocem ή γίχα eadem significatione adhuc Cretensibus Mytilenaeis etc usitatam) 46, 22
A
#### Other unattested words not listed by Drexl
#### ούσίον 170, 14 ff.
Drexl gives the meaning of this word as "vestis talaris". I believe, however, that the word goes back to the Arabic washī (multi-colored ornamentation, embroidery; embroidered or painted fabric). This meaning is also warranted by the context (Drexl 170, 15-17).
ποιναλιστής=punitor 129, 18
#### BIBLIOGRAPHY
#### *Sources*
Albumasar. See Sources: Abu Ma'shar.
Alchemists. *Collection des anciens alchimistes grecs.* Ed. Marcellin Berthelot. 3 vols. Paris: G. Steinheil, 1887-88. Rpt. London: Holland Press, 1963.
Apomasar. See Sources: Abu Ma'shar.
255-307; English translation of the text in idem, "Kitabu ta'blr-ir-ru'ya of Abu 'Ali b. Sina." *lndo-lranica* 9:4 ( 1956): 43-57.
Ibn Sirin, Mui)ammad [attributed to]. *Ta/sir al-af:iliim al-kabir al-musammii muntakhab al-kaliim fi tafsir al-af:iliim.* Cairo: Maktaba w a-ma\ba'a Mui)ammad 'Ali S ubayi) w a-awladuh, 1963.
*]us Graecoromanum.* Ed. J. and P. Zepos. Aalen: Scientia, 1962.
al-Kharkushi, Abu Sa'id 'Abd al-Malik b. 'Uthman al-Wa'i;i:. *Al-bishiira wa al-nidhiirafi ta'bir a/-ru'yii wa -1-muriiqaba.* London, BL *Or* 6262.
al-Khayyam, 'Umar. *Ta'bir al-maniim.* Cairo: Dar al-rashad, 1991.
Georg,1893. Rpt. by Ivan Dujcev, *To eparchikon bib/ion.* London: Variorum Reprints, 1970.
Leo Tuscus. See Achmet [attributed to], *Achmetis Oneirocriticon.* Trans. Leo Tuscus.
Manuel Paleologos. "Epistle to Andreas Asan." Ed. Jean Franc;:ois Boissonade, *Anecdota Nova.* Paris: Dumont, 1844. Rpt. Hildesheim: Georg Olms Verlagsbuchhandlung, 1962, 239-246.
Nikephoros Gregoras. *Scholia in Synesium De insomniis. PG* 149, cols. 522-642.
#### BIBLIOGRAPHY 479
convivialium 8: 10." Trans. Lionel Pearson and F. H. Sandbach. *Plutarch" s Moralia in Fifteen Volumes.* Vol. 9. Loeb Classical Library. Cambridge, Mass: Harvard University Press, 1961, 202-211.
#### *Secondary Literature*
Ali, Mul).ammad. *A Manual of Hadith.* London: Curzon Press, 1978.
#### BIBLIOGRAPHY 481
die Kunde des Morgenlandes 19:2. Leipzig: Deutsche morgenliindische Gesellschaft, 1932.
*---.A History of Zoroastrianism.* Vol. 2. Leiden: E. J. Brill, 1982.
Breydy, Michael. *Geschichte der syro-arahischen Literatur der Maroniten vom Vil. his XVI. Jahrhundert.* Forschungsberichte des Landes Nordrhein-Westfalen 3194. Fachgruppe Geisteswissenschaften. Opladen: Westdeutscher Verlag, 1985.
Brinner, William M. [Review of Yehia Gouda, *Dreams and Their Meanings in the Old Arab Tradition.* New York: Vantage Press 1991]. *lnternationallournal of Middle East Studies* 28:2 ( 1996): 302-304.
Brock, Sebastian. "The Syriac Background to J-.lunayn's Translation Techniques." *Aram* 3:1-2 (1991): 139-162.
Browne, Gerald. "Ad Artemidorum Amburn." *Le Museon* 97 (1984): 207-220.
Calofonos, George. "Dream Interpretation: A Byzantinist Superstition?" *BMGS* 9 (1984-85): 215-220.
ta Anekdota tou Prokopiou." *Opsis enypniou. He chrese ton oneiron sten hellenike kai romaike archaioteta.* Ed. D. Kyrtatas. Herakleion: Panepistemiakes ekdoseis Kretes, 1993, 285-327.
Cameron, Averil. "The Eastern Provinces in the 7th Century A.o.: Hellenism and the Emergence of Islam." *Hellenismos. Quelques jalons pour un histoire de !' identite grecque. Actes du colloque de Strasbourg,* 25-27 *octobre I989.* Ed. S. Said. Leiden and New York: E. J. Brill, 1991, 287-313.
Colinet, Andree. See Sources: Alchemists. Ed. Andree Colinet.
#### 484 BIBLIOGRAPHY
*Actes du colloque de Louvain-/a-Neuve, 24 et 25 mars 1997,* Ed. Isabelle Draelants, Anne Tihon, Baudouin van den Abeele. Turnhout: Brepols, 2000, 165-190.
Collin-Roset, Simone. See Sources: Pascalis Romanus.
*de I' independance de la Beliique.* Publications de la Faculte de Philosophie et Lettres de l'Universite de Liege 94. Paris: Belles Lettres, 1930.
Dondaine, Antoine. "Hugues Etherien et Leon Toscan." *AHDLMA* 27 (1952): 67- 134.
-
-
Du Cange, Charles du Fresne, sieur. *Glossarium ad scriptores mediae et infimae Graecitatis.*
#### 486 BIBLIOORAPHY
Lyon: Anisson, J. Posuel, & C. Rigaud, 1688. Rpt. 2 vols. in one, Paris: College de France 1943.
Falco, V. de. [Review of *Achmetis Oneirocriticon,* ed. F. Drexl.] *BZ* 27 (1927): 113 ff.
Scriptorum Christianorum Orientalium 420, Subsidia 59. Louvain: Secretariat du CSCO, 1980.
#### BIBLIOGRAPHY 489
*bibliothecis latentium.* Zurich: Christophorus Froschouerus, 1545.
generalement tout ce qui regarde la conoissance des peuples de l'Orient. Paris: Compagnie des Libraires, 1697. Several rpts.
- --- Byzantinē logotechnia. Hē logia kosmikē grammateia tōn bvzantinõn. 3 vols. Athens: Morphōtiko Hidryma Ethnikēs Trapezēs, 1987-1994.
Kessels, Antonius Hendrik Maria. "Ancient Systems of Dream Classification." *Mnemosyne* ser. 4, 22 (1969): 389-424.
*---.Studies on the Dream in Greek Literature.* Utrecht: HES, 1978.
Kousis, A. "Quelques considerations sur Jes traductions en grec des oeuvres medicales
orientales et principalement sur Jes deux manuscrits de la traduction d'un traite persan par Constantin Melitiniotis." *Praktika Akademias Athenon* 14 (1939): 205-220. Kraus, Paul. "Zu lbn al-Muqaffa'." *Rivista degli Studi Orientali* 14: 1 (1934): 1-20.
33. *Papers presented at the Twelfth International Conference on Patristic Studies*
#### BIBLIOORAPHY 495
*held in Oxford, 1995.* Ed. E. Livingstone. Louvain: E. Peeters, 1997, 553-557.
Le Goff, Jacques. *L' imaginaire medieval.* Paris: Gallimard, 1985, 265-330.
- *Pour un mitre Mayen Age. Temps, travail et culture en Occident: 18 essais.* Paris: Gallimard, 1977, 299-306.
*Byzantium in the Ninth Century: Dead or Alive? Papers from the Thirtieth Spring Symposium of Byzantine Studies, Birmingham, March 1996.* Ed. Leslie Brubaker. Hampshire, Eng. and Brookfield, Vt.: Ashgate, 1998, 195-213.
Magdi, Cherifa. See Sources: I;>iyii' ad-Din al Djazlri.
#### 498 BIBLIOGRAPHY
Pingree, David. "Historical Horoscopes." *JAOS* 82 (1962): 487-502.
*---.The Historic Role of the Christian Arabs of Palestine.* Text in English and Arabic. Carreras Arab Lecture 1968. Harlow: Longman's for the University of Essex, 1970.
Sevcenko, Ihor."Hagiography of the Iconoclast Period." *Iconoclasm: Papers Given at the Ninth Spring Symposium of Byzantine Studies, University of Birmingham, March 1975,* eds. Antony Bryer and Judith Herrin. Birmingham: Centre for Byzantine Studies, 1977, 112-118.
Shahid, Irfan. *Byzantium and the Arabs in the Fourth Century.* Washington, D.C.: Dumbarton Oaks, 1984.
Sidarus, Adel Yussef. "Les lexiques onomasiologiques greco-copto-arabes du Moyen Age et leurs origines anciennes." *Lingua Restituta Orientalis: Festgabe fur Julius Assfalg.* Ed. Regine Schulz and Manfred Gorg. Aegypten und altes Testament 20. Wiesbaden: 0. Harassowitz, 1990, 348-359.
#### 504 BIBLIOGRAPHY
*Studien zum 8. und 9. Jahrhundert in Byzanz.* Ed. Helga Kopstein. Berliner byzantinische Arbeiten 51. Berlin: Akademie-Verlag, 1983, 179-183.
#### BIBLIOGRAPHY 505
*Etudes d' astronomie byzantine.* Aldershot and Brookfield, Vt.: Variorum Reprints, 1994.
Walzer, Richard. "Al-Farabi's Theory of Prophecy and Divination." *Journal of Hellenic Studies* 77:1(1957):142-148.
*---.A Companion to the Qur'an.* London: Allen and Unwin, 1967. Rpt. Oxford and Rockport, Mass.: Oneworld Publications, 1994.
Winkler, John J. *The Constraints of Desire. The Anthropology of Sex and Gender in Ancient Greece.* New York and London: Routledge, 1990, 23-44.
#### GENERAL INDEX
Aaron, physician 244 Aaron, brother of Moses 247 Abbasids 218, 420 'Abd al-Malik, caliph 377, 378, 413 'Abd al-Rahman b. Muhammad al-Nasir, Umayyad caliph of Cordoba 416 'Abd al-Rahman b. Nasr al-Shirazl (or al-Shayzarl) 113, 114, 395, 445 Abel 247 Abrahah 234 Abraham 247 abridgement of the *Oneirocriticon,* see MS *Paris. Suppl. gr.* 690 Abu 'Abd-Allah the Sicilian 416 Abu Bakr 220, 382, 383, 384, 387 Abu Makhlid b. Bukhushu' 243 Abu Ma'shar 6, 7, 8, 9, 10, 14,18, 34, 37, 40, 108, 110, 116, 253-254, 398, 399, 418 Abu Sa'ld al-Wa'iz 306 Abu Sa'ld Shadhan b. Bahr 7, 34 Abu Sahl Isa b. Yahya al-Maslhl al-FaylasUf 245 Abu Sufyan b. Harb 249 Abu Zayd al-Ansari 409 Abu-I-Faraj al-Isfahanl 249 Achilles Tattius 3 Achmet, son of Sereim (see also Sereim and lbn Slrln) 9, 10, 11, 15. 17, 19, 22, 32, 33, 34, 35, 36, 37, 38, 40, 64, 103, 106,113,119 Actuarius 111, 394 Adam and Eve 247 *adhan* (call to prayer) 147, 148, 320-321 adultery 465 Aesop 393 Aghlabids 415 agriculture 407 Ahrun the physician, see Aaron, physician 'A'isha 220
Akathist Hymn 393
al-Razl 67, 406, 420, 421 al-Salemi 13; see also al-Salimi al-Salimi, Abu 'Abd-Allah Muhammad b. 'Umar 30, 32, 133; see also al-Salemi al-Sijistanl, Abu al-'Abbas Ahmad b. Khalaf b. Ahmad 26, 134, 243 *al-siyasatu-1-'ammiyyah* 425 al-Tabar! 249 al-Walld, caliph 377 al-Zanan 405-406, 408-409 Albumasar, see Abu Ma'shar alchemy 52, 76, 92, 107, 245, 394, 400-403, 424 alcohol 256, 340-341, 345-351 Alexander of Aphrodisias 392 Alexander Romance 392-393 Alexios Comnenos, emperor 4, 14, 406, 420 Alexios, translator from Arabic into Greek 410, 420 'All b. Abl Talib 180 'All b. Ahmad al-'Imranl 399 Alice de Courtenay (Alice de Couty?) 117- 118 *Almagest* 396 alphabetical order 107, 142 Alphonso IV, king of Castille, Leon, and Galicia 24 Amam 379-380 amir 202 *anakaras* 70, 73, 101 analogy 187-188 *anaskolopismos* 175 Andreas Asan I anecdotes 33, 158, 174, 180, 256, 285, 287, 298, **375-391** angel Gabriel 266 angel lsrafll 247 angel Jibrll 247 angel Mika'll 247; see also archangel Michael angel of dreams 166 angels 21, 24, 145, 146, 148, 200, 201, 239, 247,264-269,351, 373,452 animal heads 194 animals 55
antimony, see *kochla* Antioch 404 ants 443-444 Aphrodite 139 *apo phones* 405 Apomasar, see Abu Ma'shar *Apomasaris apotelesmata* 108 apostles 21, 145, 146, 162, 269-285 apricot tree 388 Arabic dreambooks chosen for comparison with the *Oneirocriticon* 25 Arabic into Greek, problems in translating 227, 228; literal translations 75-76, 82- 86, 89-90, 360; method of translating 254, 309, 372-374, 417-419; peculiarities in grammar and syntax of the *Oneirocriticon* due to translation 237; transformation of dream symbols when rendered from Greek into Arabic and back into Greek 200-201, 438-440; see also *skorpismos, eiskosmisis* Arabic language 14, 418 arch 446 archangel Michael 452 *arche* 185 architecture 428 Aristotle I, 130, 143, 145, 239 arithemtic 428 Armenian language 238 arms 247, **364-365,** 454 arrangement of chapters 112, 124, 126, 142- 149, 284-285, 286, 298-299, 375,431 arrows 248, 459 Arsenios the monk 408-409, 420 Artemidoros 2, 3, 4, 5, 9,13, 16, 17, 19, 21, 28, 38,45,46,63,64, 68,94,95,96, 112, 113, 119, 126, **128-167,** 239, 243, 245, 248,267-268, 297,375, 378, 386, 395,421,423,426,437,438-440,441, 442,443,445,448,449 Artemidoros, son of Artemidoros 135 Ascari, Joseph 15 ascending from the underworld 447 Assyrian dreambook 129, 378 Astrampsychos 64, 119
astrolabe of Brescia 412 Black Stone 186 astrology 8, 43, 92, 225, 245, 250, 253, 394, Bland, N. 16-17, 22, 23, 64 396-400, 421, 423 bleeding 247, 456, 461 astrology and dream interpretation 1, 8, 9 blood 190, **365-366** astronomy 92, 394, 396-400, 424, 428 *Book of Signs* 240-241 Attic syntax 80-81 *Book of the Eparch* 73, 411 *augusta* 223 *Book of the Speech of Wisdom* 241 *awliya'* (friends of God) 313, 315 booty, see *ghanlma* Ahmad the Persian 8, 18 bows 248, 458 Babylonians 8 Bah1ra 305 balances **333-335** Bar Hebraeus 241, 245 Bar Sauma 245 Baram 32, 34, 36, 41, 43, **47,** 103 Bardas 218 Basil I, emperor 413 Basil of Caesarea 3, 241 Basil the Bastard 218 Basil the Younger 340 *basileopator* 218 *basileus* 59, 198, 209, 223 *Basilica* 67 baths 436-437 *bayt278,* 387 *bayt al-mal* 225 beating 99, 304-305, 314, 318 bees 100, 230, 437 belly 146, 302-303 benefactors 443 *Berakot* 129, 235,452 Berber language 405 Berthelot, Marcelin 401 Bessarion 107 Bible 50, 78, 80, 157, 256; see also Old Testament, Septuagint, and Torah *Bibliotheque orientale* 12 biographical dictionaries 384 birth 126 bishops 251 bitch 247, 455 black 339, 432
Athanasios of Alexandria 242, 393 *Book of the Beloved,* see *Kit ab al-hablb* Avicenna 395 books 247, 288, 446, 447; see also reading Brackertz 2, 20, 21, 68 brains 247 brass 329 bread 247 bridges 253-254 *brontologion* 394 bronze 328-331 building a temple 49 buildings 148 Bukhtishu' 244; see also Abu Makhlid b. Bukhtishu' buildings 299 Bulgars 398 bull 194 butter 247 Byzantine imperial ideology 182, 198, 200, 271, 281 *caesar* 218 caliph 57, 147, 148, 187, 202, 203, 218, 222,294,299,379 call to prayer, see *adhan* camels 21, 190, 194, 247, 454 Camerarius, Joachim 6, 8 *Candelabrum of the Sanctuary* 241 candles 450 capital 193; see also *kephalaion* and *ra's ma!* Cardano, Girolamo 118-119, 395 Carmathians 30 Carolingian Renaissance 414 Cassius Maximus 135, 136 Chaldean oraclees 393 changing faith, see conversion
chapter 19 of the Oneirocriticon 32, 33, 35 chariots 55, 164, 361-363 charistikē 4 charzanion 69 chasdion 65, 471 cheese 247 chess, see zatrikion, zatrikizo chestnuts 100 child 449 Christian adaptation 237 Christian dream symbols 256-352 Christian eschatology 258 Christian identity 12, 14, 17, 20, 21, 22, 23 Christian-Muslim religious polemics 147 Christopher of Mytilene 393 Chrysochoou, Gregory 106 chrysolites 340 church officials 21 churches 183, 239, 253-254, 322-325, 351 clarified butter 456 Cleopatra 20, 52 clouds 214 coats of mail 248, 460 coins 190, 191-192, 195, 318 Collin-Roset, Simone 113, 114, 445 concubines 21 Constans II, emperor 410 Constantine the African 11, 406 Constantine the Great 175, 183-184 Constantine V, emperor 403 Constantine VI, emperor 398 Constantine VII, emperor 61, 144, 407, 426, 427 Constitutiones Tacticae (by Leo Vi) 424, 475 conversion 326-328, 434-435 copper 328-331, 401 Coptic language 144 Cordoba 416-417 Cornarius. Janus 113. 395 cotton 21, 65, 67 courtship 149 cows 194, 248, 359-361, 464 critical edition of the Oneirocriticon 91-92. 93
cross 99, 170, 174-188, 270, 282, 314, 353 crown 338 crucifixion 140-141, 170, 174-188, 248, 270, 282, 353, 461 cuirass 248, 459 Cumont, Franz 18 cutting hair 146 Cyranides 113 Cyrus the Great 378 d'Herbelot, Barthélemy 12, 17, 24 daggers 248, 459 Dagron, Gilbert 23, 134, Gilbert 411-412 Daim. Abdel 141 Damascus 397 Daniel, prophet 5, 21, 29, 42, 64, 93, 118, 163, 239, 410, 420 dār 278-279 dār al-ākhira 278 darkness 303 David 247 De cerimoniis 73 de Falco 124 De pulsibus doctrina 406 De revolutionibus nativitatum, see Peri tēs tōn etōn enallagēs De Slane 18, 19 De Unirinis 394 deacon 145, 269, 285, 291 dead man 445-446, 448-449 death 149, 448-449 decapitation 188 deer 191 deflowering a slave girl 465 Demetrios of Lampe 115 dervishes 315 despotēs 12, 41, 59, 60, 61 desserts 231 dhikr 295 dhimmī 50, 251, 305 Dietrich 64 Diogenianos of Heraklea 143 Dioscorides 10, 416 dirham 334 divination 92, 131, 243, 250, 394, 409-411
#### ર્ટ | 0
doors 385 doumakin 71 dreams caused by daytime preoccupations 238 dreams caused by demons 238 Drexl, Franz 18, 63, 91, 93, 94, 109, 111, 113, 122, 124, 126, 127, 239, 445 Du Cange 11 du a 295 dung 248, 464 ear 247 earthquakes 367-368 eating human brains 453 eating stars 227 education in Byzantium 373 education in the Muslim world 373 egg 248 Egypt, ancient 51-58, 358 Egyptians 6, 7, 8, 20, 32, 33, 35, 36, 41, 44, five daily prayers 96, 46, 47, 58, 171, 175, 191 eiskosmisis 228, 468 ekphrasis 394 elephants 46, 170, 231-235, 353 elevated places 298 elevations of the earth 239 emeralds 340 emperor 199, 201, 207, 209, 270, 271, 274, fruits 158, 159, 160 282, 299, 304, 309 empiricism 188 Ephodia tou apodemountos, see Supplies of a Traveler Ephraim the Syrian 3, 238 eschatology 372 Ethikon 241 Etymologicum Magnum 76-77 etymology 165, 170, 185, 190, 192, 194, 197, 198, 205, 307, 365 eunuchs 268-269, 357 Euphrates 390 Euripides 129, 353 Euryviadēs 3 Eustathios of Thessaloniki 426 exodos 190. 457 Exodus 357
eyebrows 170-174 eyes 363-364 Fahd, Toufic 19, 23, 134, 136, 137 faith 145 false and truthful dreams 160, 166 false dreams 238 fālūdhaj, see paloudakin family of the dreamer 155 father 263 Fathers of the Church 91 Fatimids 415 feet 83, 248, 302-303 fetters 382, 383 fingers 96, 252, 335-337 fire 48, 51, 331, 357, 449 first person narrative 33 Fischer, Anton 23 fitna 350-351 Five Pillars of Islam 186, 336 fleas 101, 102, 437-438 flora 158 Florilegium Baroccianum 3, 15, 38, 40 flowers 158 folleis 4 four bodily humors 238 funeral procession 445-446 fugahā' (jurists) 313 Furlani 239 Gabriel, dux of Melitene 420 Galen 130, 137, 244, 421 Gemini 398 Geminus of Tyre 129 Genesis 157, 228, 238, 359, 361, 362 geomancy 405-406, 407-409 Geoponika 60, 78, 393, 407 George of Pisidia 393 Georgios Grammatikos 110 Georgios Midiates 110, 410 Georgius Elmacinus 12 Germanos, patriarch 95 Gessner, Konrad 10, 11 ghanam 194, 198, 236
*ghanima* 194, 198, 236 gibbet, see impalement and crucifixion Gigli, D. 93, 114 glass 349 *gnomologia* 394 gnosticism 30 goat 191, 248 God 146, 148, 161, 162, 166, 178, 181, 196, 200,201, 236,238, 239,258, 259,272- 273,275-276,277, 284, 303, 304, 330, 333, 365, 371,446 godhead 24, 273 gold 328-330 gold revetments in icons 270, 283 Gospel 44, 161, 239, 247, 251, 269, 287, 288, 289, 293, 296 grapevine 375 *grapheus* 225; see also scribe, *kiitib* Greek ethnicity 14, 20, 22 Greek fire 425 Greeks 245 green **337-340,** 353, 383 Gregory of Nazianzus 3, 241, 242, 393 Gregory of Nyssa 393 *gunaika tes charas* 219 Gutas, Dimitri 422 Habash al-Hasib 400 Hades 447 *hadith* (traditions of the Prophet MuJ:iammad) 16, 34, 131, 160, 165, 199, 250,251, 353,384 Hagia Sophia 393 hair 33, 125, 126, 146, 248, 389-390, 462 *hiijib* 200, 225 *hajj,* see pilgrimage to Mecca Hajji Khallfa 31 hands 146, 302-303, **335-337,** 449 *hapax legomena* 68, 69, 228 harbors 442-443, 444 Harlin al-Rashid 413 Haskins 116 *Hayiit al-hayawiin* 141 head 170, 188, 302-303, 353; see also *kephale, kephalaion, rii's, ra'is, riyiisa,*
*rii's miil* Heaven 239 heavenly bodies 170 Hebrew language 364, 410 Hell 21, 24, 145, 146, 148, **262-264,** 351, 353, 373 hemorrhage 248 Heraclius, emperor 246, 248, 249, 425 Hermaic sciences 394 Hermas 189 Hermes 139 Hermes Trismegistus 394 Herodotos of Halicarnassos 378, 427 Herophilus of Chalcedon 242 Hesiod 353 Hindu drugs 406 Hindus 8 Hippocrates 64, 239 Hisham b. 'Abd al-Malik, caliph 130, 377, 425 *Historia Saracenica* 12 historical events 375 Holy Books 287, 295 holy figures 269-285 Holy Spirit 247 Holy Writ 239; see also Bible, Old Testament, Gospel, New Testament Homer 129, 353, 393, 397 honey 170, **229-231** *horologion* 253, 336 horses 21, 55, 148, 170, **203-208,** 441 houses 148, 183, 278 *housion* 70, 71-72, 77-78, 471 Huggelin, Johann Jacob 395 Hugo Etherianus 10, 39, **115** *hukuma* 289 Hunayn b. Ishaq 17, 19, 45, 89, 131, 133, 136-142, 169, 172, 177, 185, 197, 200, 201, 243, 268, 297,421,439-440,443 Hunger, Herbert 407 *hypatos* 412 *hegemonikon* 74
Ianus Antionius Saracen us 10 *'lbiiriit al-ru'yii* 26, 133
Ibn al-Athir 24 lbn al-Bahlul **240-241,** 252 Ibn al-Faqih al-Hamadhani 403 lbn al-'Alam 399 Ibn al-'Arabi 409 Ibn al-Jazzar, Abu Ja'far Ahmad b. Ibrahim intestines 58 b. Abi Khalid 10-11 *Introduction to Astrology* 8 lbn al-Musayyab 131, 379, 382, 384, 389, iron 248, 328-331 391 *irtakaba* 205 lbn al-Muthanna 400 Ibn al-Nadim 17, 51, 136, 243, 400, 413 Ibn al-Qifn 244 Ibn Ghannam 144 Ishmael the doctor 412 lbn J uljul 415-417 Islamic dream interpretation 16, 19 lbn Khaldun 132 Islamic eschatology 259 lbn Qutayba, 'Abd Allah b. Muslim 25, 26, Islamic law 16, 147, 323 27, 96, 126, 130, 133, 134, 147, 153, Isma'll 247 240, 243, 252, 464; several instances in 150-167, 168-236, 256-352, 353-374, 375-391 lbn Sa'd 377, 378, 379, 381, 389, 391 Istifan b. Basil, translator from Greek into lbn Shahin, Ghars al-Din Khalll 22, 29, 59, Arabic 416 124, 134, 239, 435, 436, 437, 438, 440, Italy 231-232 441, 443, 447, 459, 463, 464; several instances in 150-167, 168-236, 256-352, 353-374, 375-391 Ibn Sina 142 lbn Slrln (see also Achmet, son of Sereim and Sereim)l3, 15, 20, 23, 24, 25, 28, 29, 34, 150, 155, 157, 159, 160, 180, 243, 256, 274, 296, 298, 312, 361, 371, 375, 376, 379, 381, 382, 383, 384, 386, 389,391 icons 2, 269-285, 299, 435 identity of the dreamer 151, 153 idols 48, 55, 148, **328-331** *ighriqi* (Greek language) 416, 417 Jewish dream interpretation 85, 129, 133, *ihram* 310 169, 235, 388 *'i/m al-ikhtilaj* 410 Jews 238, 243, 245, 290, 327 imam 75, 148, 187, 203, **289-290,** 293, 294, *jihad* 356 295, 299, 300, 315, 380 John Chrysostom 3, 393 impalement 176; see also crucifixion John Geometres 393 India 231-232 John Italos 1 Indian iron 401 Indian=Christian 44-45, 59
Indians 6, 7, 8, 14, 20, 32, 33, 35, 36, 41, 44, 46, 99, 149, 170, 175, 245, 258 Indians, Persians, Egyptians 93, 108, 100, 116, 145 inebriation 345-351 *irtakaba ma'siyatan* 236 Isaac 247 Ishmaelites 403 lsma'll b. al-Ash'ath 29 *isnad* (chain of authorities) 34, 199, 371, 375, 378 Jabariyya 287 Jabir al-Maghribi 29, 220 Jabir b. Hayyan 403 Jacob (Old Testament) 247 Jacob of Edessa 242 Ja'far al-Sadiq 29, 157, 440, 459 Jagaddeva 20, 45-46 *jalliid* 225 Jamasb 141, 245, 312 Janus Cornarius 4 Jesus Christ 145, 146, 183, 198, 199, 200, 247, 267, 269-285,288, 356, 357,452 John of Damascus 393, 413 John of the Ladder 3, 88
#### 514 GENERAL INDEX
John the Baptist 247 John the Grammarian 397, 424 Joseph (New Testament) 21, 42, 44, 161, 163, 238, 239 Joseph (Old Testament) 21, 53, 131, 154, 162, 163, 164, 167, 228, 236, 238, 247, 358-363, 451 Judaism 51, 241, 290, 326-327 judgements 146, 147 judges 125, 126, 146, 147, 148, 247, 332- 335, 457 julep 65, 77, 346, 406 Jupiter 9, 223, 224, 254 Ka'ba 75, 186, 187, 202, 203, 234 *Kalila wa-Dimna,* see *Stephanites kai Jchnelates* Kallinikos III, patriarch of Constantinople 106-107 *kamelaukion* 337 *kampanon* 335 *kara* 74 Karatzas, Nikolaos 103 *katholikos* 286 *klitib* 225 *kephalaion* 188-189, 190, 236 *kephale* 185, 188-189, 190, 192, 236 *Kethlibhli dhe-Pushshlik Helme* 245 Khalid b. Yazld b. Mu'awiya 425 *khatzb* (preacher) 290 *khutba* (Friday sermon) 147, 148, 290-292, 297 killing 149, 449 Kirnon 3 kings 57, 170, 182, 198-203, 247, 299 Kisra 167 *Kitlib al-aghlinl* 249 *Kitdb al-buldan* 403 *Kitdb al-Dala'il,* see *Book of Signs Kit ab al-dustur ft al-ta 'bir* 132 *Kitab al-fihrist* 17, 51, 136, 138, 243, 400, 413 *Kitab al-habib* 425 *Kitab al-ishdra ila 'ilm al-'ibdra* 133 *Kitdb al-ishdrdt fl 'ilm al-'ibdrdt* 29, 133
*Kit ab al-Qadirl ft al-ta 'bzr* 27, 133-134 *Kitdb al-ta'bir* 132 *Kitab muqarmat al-rn'ya* 30 *Kitab ta'bzr al-rn'ya* 24, 25 *Kitab tafslr al-mandmdt min al-nujum* 8 *Kitab tahawll sinl al-mawdlld,* see *Peri tes ton et6n enallages klibanion* 77 knife 248 known person 125 *kochla* 65, 67, 69, 77, 468 Kollar, Adam Frantisek (Colarius) 13-15, 17,24 Kosmas, alchemist 401 Koukoules 63, 68 Ktesias of Knidos 378 lakes 441-442 Lambeck, Peter 11-13 language of the *Oneirocriticon* 22, 63-90, 124 Latin translations of the *Oneirocriticon* 111-119 lead 328-331 leader, leadership, see *rli's, ra 'ls, riylisa, kephale* Lebanon 404 left side 238 legitimate wives 21 legs 83 Lemerle, Paul 422 Leo (sign of the Zodiac) 398 Leo Choirosphaktes 397, 424 Leo Katakylas 426 Leo the Mathematician 385, 424 Leo Tuscus 4, 10, 17, 39, 40, 94, 115, 121, 122, 125, 127, 434 Leo VI, emperor 4, 61, 62, 411-412, 424, 428 letters 76 lexicography, Arabic 144, 186 *Liber Thesauri Occulti* 38, 112-114, 116, 127, 445-450 lice 101, 102, 437-438 *Life of Aesop* 393
light 303 lightning 450 lioness 247, 455 liturgy 297 Liutprand of Cremona 410 liver 58 loan words 65, 67, 69-73 locusts **357-359** Loewenklau, Johann (Johannes Leunclavius) 5-10, 12, 17, 108, 110, 113,116,120 Lucian 393, 423 lunar dreambooks 242, 394 *lychnos* 364 *lypopoulos* 468 Macedonian Renaissance 5, 392, 422 Magdalino, Paul 397, 424 Magians 6 magic 394, 409-411 *magos* 327-328 *malak* 200 *malik* 200 Mandane 378 Mango, Cyril 414-415 *manikelia* 77 Manuel Komnenos, emperor 1, 109, 112, 115 Manuel Paleologos, emperor 142 Manuel Philes 393 manuscript illustrations 103 *maqiim* 291 Marcellae, fortress 398 Mardaites 404 Marianos the monk 425 marriage 64, 149 Mars 9, 223, 224 martyrdom 265, 267, 274, 339, 355, 356, 366 martyrs 21, 315 Marwan I, caliph 377 Mary, mother of Christ 42, 161, 163, 238, 271, 338 Maslama b. 'Abd al-Malik 380
Master of the Post 179 *maulii* 201 Maurice, emperor 425 *mawiili* 220 Maximus the Confessor 393 measles 406, 420, 421 measures 146, 147 Mecca 75, 186, 202, 234, 249, 266, 267, 276, 376 medicine 9-11, 14, 16, 73, 92, 112, 130, 243, 244, 245, 250, 394,406-407,412, 428, 435-436 Menander, poet 129, 198, 201, 353 Menander, priest of Zeus 180 Mercurius 406 Mercury 9, 223, 224 meteorology 404-405, 424 *Methodoi psephophorias diaphoron hypotheseon astronomikon* 399-400 metropolitans 286 Michael III, emperor 413 Michael VIII Paleologos, emperor 409 Michael Andreopoulos 420 Michael Attaleiates 4 Michael of Ephesos 1 Michael Synkellos 420 *mihriib* 186, 377 *miliaresia* 4 military manuals 426-428 milk 247, 248, 454, 455, 456, 463 milk of a woman 462-463 milking 247, 248, 463, 454 minaret 320 *minbar* (pulpit) 290, 291, 292, 297-299, 375 Modern Greek dialects 64 *modios* 333 Mogenet, Joseph 396 monasteries 307 monastic cells 307 monastic tonsure **304-319** monks 99, 247, 251, 253-254, 269, 285, 286, **304-319,** 457 moon (phases of the moon when a dream was dreamt) 153
MS *Vindab. phil. gr.* 108 411 MS *Vindab. phi!. gr.* 115 398 MS *Vindab. phi/as. et philal. gr.* 111 (R) 11, 35, 40, 93, 108, 122, 392 MS *Vindab. philas. et philal. gr.* 162 (S) 12, 35, 40, 104, 108, 392, 393 MS *Vindab. phi/as. et phi/al. gr.* 287 (T) 12, 37, 108, 121, 123, 124, 394 MS *Vindab. phi/as. et phi/al. gr.* 297 (V) 6, 7, 8, 9,12, 37, 40, 108, 110, 121, 392 MS Wolfenblittel *Guelpherb. lat.* 2917 39, 122 MS *Yahuda ar.* 196 23, 26, 96, 126, 153, 243; several instances in 150-167, 168- 236, 256-352, 353-374, 375-391 MS *Zagara* 89 37, 40, **105-107,** 145, 151, 392 *mu'adhdhin* 253 Mu'awiya, caliph 410 mud 441 Muhallab b. Abl Sufra 380 Muhammad b. Idris al-Shafi'l 179, 180 Muhammad b. Mangll 425 Muhammad b. Zakariyya al-Razl, see al-Razl Muhammad, Prophet 34, 57, 131, 147, 148, 202,220,234,243,248,249,250,252, 270-271, 273, **276-278,** 284, 305, 382, 387,451 *Muntakhab al-kaltimfi tafvir al-ahltim,* see *al-Muntakhab murtibitiin* (the bound ones) 313 muscles 76 Muslim 6 Muslim taboos **340-351,** 352 *Mysteria tau Apamasar* 7, 34 nape 197 338, 339, 340, 351, 353, 356, 357, 373 Nasara 245, 246, 248, 271 *paraikas* 4 378 115, 127, 431, 432-433, 434, 445 Nebuchadezzar 167 Nemesius 242
Neo-Syriac language 239 *ne6teroi* 399 Neugebauer, Otto 396 New Testament 3, 4, 68, 79, 80, 82, 88, 91, 259, 269, 277, 289, 290, 336,353-374; see also Gospel; corruptions in the New Testament, see *tahrif* Nicholas the monk 416-417 Nikephoros Gregoras I Nikephoros Phokas, emperor 78, 423 Nikephoros, patriarch 64, 94, 95, 119, 393, 394,431,438,442 Niketas of Byzantium 147, 356, 413 Nile 20, 52 Noah 247 *namismata* 4 Normans 415 North, Frederic I 03 nose 456 nostrils 90 nun 308 Oberhelman 3, 22, 168-170, 175, 202, 203- 204, 206, 217 *Odyssey* I oil lamp 449 old person 125, 126, 146, 315-317 Old Testament 3, 4, 53, 79, 82, 91, 164, 228, 236, 290, **353-374;** see also Bible and Torah Orosius 416 pagan Arabs 243 pagan deities 200, 201 *palaudakin* 70-71, 72, 78, 97 Pantaleon, metropolitan of Synnada 61 parables from the New Testament 369, 371 *nabldh* 347, 348 Paradise 7, 21, 24, 145, 146, 148, **258-262,** Near Eastern dream interpretation 129, 169, Pascalis Romanus 38, 39, 40, 96, **112,** 113, patriarch 145, 269, 285 patriarchal throne 290
pearls 95, 371-372, 432-433, 437 penis 390-391 perfumes 21, 85 Peri loimikës 406, 421 Peri tēs tōn etōn enallagēs 7 Persian language 16, 70, 71, 77 Persians 6, 7, 8, 12, 20, 32, 33, 35, 36, 38, 41, 43, 44, 46, 47, 55, 56, 57, 58, 170, 171, 175, 191 persikomanikion 470 Peucer, Caspar 395 pharaoh 99, 145, 147, 154, 167, 192, 357, 359, 362 pharmacology 92 phasganon 74 Philaretos the Merciful 340 Philopatris 423 philosophy 394 Phocylides 393 Phoebus of Antioch 129 Photios, patriarch of Constantinople 4, 424 phourkisis 175 physiognomists 243, 250 Physiologos 106 pigs 341-345. 449 pilgrimage to Mecca (hajj) 186, 196, 266, 267. 276. 310-313 Pingree, David 396, 398, 399, 407, 418, 420 planets 8, 48, 51, 96, 239 Pliny 143 poetry 130, 165, 209-210, 250, 353, 393 Poliorkētika 78 polo, see tzykanion, tzykanizo Polyaenos 427 polygamy 21, 50, 75, 96, 252 pomegranate 389 pork 256, 340-345 Poseidon 203 Praetorium 412 prayer 147, 247, 252, 285, 294, 295, 335-337; see also salā praying for the people 294 presbyter 285, 291 priesthood 428 priestly duties 285-303
priestly ordination 293 priests 75, 99, 139, 145, 147, 247, 251, 253-254, 269, 285-303, 457 proedros 218 prologues of the Oneirocriticon 32 ff. Prometheus 58 prophecies on Constantinople 393 prophecy 162 prophets 21, 24, 145, 146, 148, 162, 243, 247, 265, 269-285 Prosarhythmus, see rhythmical patterns in prose proskynēsis 55, 57, 271 prostration 247 prõtospatharios 412 proverbs 165, 369 Psellos, Michael 1, 393, 401 Ptolemy Philadelphos 410 pulsation 406, 410 pumpkins 435-436 Pythagoras 408 gabbān 335 Qadariyya 287 gādī 315. 332; see also judge ga id 214 Qatāda 180 qibla 186, 294, 295 Qisas al -anbiya3 157, 164, 167 qiyās, see analogy qualities of the dream interpreter 165 Qur'an 53, 56, 129, 131, 146, 147, 148, 157, 160, 162, 163, 164, 165, 178, 179, 181, 197, 228, 230, 234, 236, 250, 251, 257, 258, 259, 260, 261, 263, 264, 267, 270-271, 274, 277, 280, 287, 288, 289, 291, 295-296, 300, 303, 305, 306, 311, 313, 338, 339, 347, 348, 350, 353, 354, 355, 356, 357, 359, 360, 362, 366, 372, 373. 413 Qustā b. Lūgā 244 Raban al-Tabarī 243, 244 rabbis 251 Rabi ah b. Umayyah b. Khalaf 383
ર | 8
*rabolion,* see geomancy *rahba* (terror) 307 *rahib* (monk) 307 rain 379-381, 436 *ra'is* 192 *rakabafarasan* 205, 236 *ramblion,* see geomancy *ram[,* see geomancy *raqaba,* pl. *riqab* 197 *ra's* 185, 192, 193, 236 *ra's ma[* 236 *Rasa'il ikhwan al-safa'* 131-132 reading 145, 147, 269, 293, 295-296, 285, 447; see also books red 339 Resurrection of the Dead 7, 12, 21, 24, 49, 145, 148, **256-258,** 351, 353 Revelation 357 revelations and visions from God 238 revolt of the Arabs 12, 380, 383-384 rhythmical patterns in prose *(Prosarhythmus)* 86-88 Ricklin 112 Rigault (N. Rigaltius) 9, 10, 11, 63, 104, 110, 111, 113, 120, 121, 127, ring 338 *riyasa* 185, 192, 236 Rochefort, Gabriel 3, 93 Romanos I Lekapenos, emperor 416 Romanos II, emperor 416 roof beams 387, 446 Rosenthal, Franz 136-138 royal people 290 Ruelle, C.-E. 18, 120, 121 Rum245,246,248,312,417 Saadi, Abdul Massih 242 Saanisan, king of Persia 32, 41, 43, 47 Sabians 55-56 sacraments 239 *sahib al-barid* 225; see also Master of the Post *sabib al-'adhab* 225 *sahib al-jaysh* 225 *sahib al-kharaj* 225
*sahib al-shurta* 225 Sahl b. Bishr 400 Sa'ld b. al-Musayyab 377 saint 446 *sala* (prayer) 295, 336; see also prayer Salerno 415 Salim Abu-I-'Ala' 425 Salim al-Harranl, see Salmanas Salm al-Harranl, see Salmanas Salmanas 402-403 salt IOI, 102, 438-440 Sambucus, see Zsamboky Samonas 412 sandals 64, 66, 339 Saracens 412 Saturn 9, 223, 224 *Scala Paradisi* 241 sceptre 58 scientific literature 91-92, 394-411 scribe 223, 225; see also *katib* scriptural quotations in the *Oneirocriticon* 78, 161, 163, 259,354-356 scriptural substitution of a Quranic passage 258 season of the year in which dreams are dreamt 9, 42, 150, 153, 156, 159, 238, 239, 369, 388-389,435 secret alphabets 409-411 Selene 217 *semanterion,* see sounding board semitisms 372 Septuagint 372 Sereim (see also Achmet, son of Sereim and lbn Slrln) 285, 298, 379, 381, 388, 390,412 sermon, see *khutba* sex of the dreamer 150 sexual intercourse 55 shaving hair 146 sheep 77, 191, 194, 198, 236 shield 248 shoulder blades 75 Sicily 415-417 sick person 449 silver 328-330
Simon magos 327 sirat (bridge accross the infernal fire) 264 Sirhān 132 sklērochalinos 205, 236 skorpismos 228, 468; see also eiskosmisis Slavonic language 5, 238, 409 smoke 48, 450 snake 247 social position of the dreamer 9, 42, 60, 150, 154, 155 soldier 247 Solomon 247 Sonderkamp, Joseph 407 sounding board 64-65, 66, 145, 147, 319-321 southern Italy 415 Sozomenos 175 St. Constantine/Cyril 424 St. Epiphanius 393 St. Michael the Synkellos 412-413 stars 1, 49, 55, 76, 84, 210, 213, 215, 222-229 staurōsis 175 steelyard 333-335 Steinschneider, Moritz 17, 18, 23 stemma codicum 109, 122, 123 Stephanitēs kai Ichnēlatēs 14, 420 Stephen of Alexandria, see Stephen the Philosopher Stephen the Philosopher 398, 425 stones 247 strangling 449 Strategikon of Maurice 425 Strohmaier, Gotthardt 22, 138, 139, 200, 201, 423 strouthia 95 Stylianos Zaoutzes 218 Suda Lexicon 74, 423 Sufism 313, 315 sugar 65, 67 sugarcane 21 Sulayman, caliph 377 sultān (power) 199-200, 235, 291 sultans 148, 247, 299
sun 8, 48, 49, 51, 168, 208-217, 218, 223, 228 Supplies of a Traveler (Zād al-musāfīr, Ephodia tou apodemountos) 11, 406 swine 190 Symeon Metaphrastes 393 Symeon Seth 14-15, 406, 420 Synesios of Cyrene 1, 242 Syntipas 420 Syrbacham 32, 34, 36, 41, 42, 103, 149, 150, 152, 160, 161, 163, 164, 165 Syria 17, 64, 379 Syriac dreambook 237-240, 362 Syriac language 65, 81, 82, 86, 89, 144, 237 Syriac model of the Oneirocriticon 63-64 Syriac translation of Artemidoros 136-139 Syriacisms 137, 138 Syrian ethnicity 11, 64, 66 Syrianos 427 Tafsīr al-ahlām al-kabīr 28 tahrif 289 tail of a horse 204-205 Talmud 129, 235, 353, 386, 388, 452 Tannery, Paul 408 Tarphan 32, 34, 36, 40, 41, 53-54, 103, 105 Tarsus 404. 412 tawl (power) 365 teachers 21, 145, 146, 269-285 tears 436 technical terms, financial and administrative
র্বা teeth 247, 453 temples 49, 170, 198-203, 299 tent of Abraham 75 terrace 446 textual parallels between Artemidoros and the Oneirocriticon 168-169 Themistios 1 Theodora Doukaina, empress of Michael VI11 408-409, 420 Theodore Abū Qurra 147, 356, 413, 414, 420
Theodore Prodromos 393
#### GENERAL INDEX 521
Theognis 353 Theophanes Continuatus 411 Theophanes Graptos 407 Theophanes Nonnos 144, 407 Theophanes, *Chronographia* 398 Theophilos of Edessa 397-398 theoretical principles of dream interpretation 16, 238 Thomas, patriarch of Jerusalem 420 thorns 370 Thucydides 427 Tigris 390 Tihon, Anne 396 time required for the fulfillment of a dream 156, 158, 238, 239, 434, 451-452 time that a dream is dreamt 9, 151, 152, 158,238 Tollius, Jakob 121 Torah 129, 251, 353, 362 trees 55, 87, 88, 148, 158, 159 Tricassus Mantuanus 10 True Cross 183, 186 truthful dreams 238 truthfulness of dreams according to the day William of Aragon 118 of the week in which they are dreamt wine 346, 349 242 truthfulness of utterances by angels 267- 268 truthfulness of utterances by Christ 270, 277 truthfulness of utterances by holy figures 270 truthfulness of utterances by icons 282 truthfulness of utterances by prophets 283 *till* (length) 365 Turkish language 16 *Tuhfat al-mulilk* 134 *tzykanion, tzykaniz6* 70-71, 75, 124 *'ulama'* 313, 338 Ullmann, Manfred 138-139 'Umar al-Khayyam 297-298, 375, 389, 391 *yunani* (Greek language) 417
'Umar b. al-Khattab 383 Umayyad dynasty 377 *umma* (nation of !slam) 251, 290, 357
unknown person 126 urination 376-379 utterances of kings 198 Vattier, Pierre 113, 395, 446 Venancius of Moerbeke 118 Venus9,48,49,223,224,398 vernacular language (Arabic) 418 vernacular words and expressions 67, 68 virgin 465 vizier 187, 202, 218, 221, 222, 224 *Volksliteratur* 91 vomit 248 *wall* 214, 365 wall 446 Wansleben, Johann Michael 15 warfare 73 water 247, 345, 346, 379-381, 390 way of life of the dreamer 155 wheat 368-370 whips 318 white clothes 366-367 wine making 349-350, 358-359 wolf 247, 449 woman delivering a sermon 291-292 woman leading prayer 296-299 woman priest 285, 286, 296-299 woman ringing the sounding board 321 women 149, 217 wood 328-330 writing 447 Wiistenfeld, F. 16 Yazld II, caliph 130, 377
yellow 339 young person 125, 126, 146, 196, 247, 304, 315-317, 458
*Ztid al-mustifir,* see *Supplies of a Traveler zamara* 70, 73, I 0 I
*zatrikion, zatrikizo* 70, 73 Zayn al-Tabari, see Raban al-Tabari *zina* 465 Zoroaster 407 Zoroastrianism 47, **48-51,** 243, 245, 251, 327-328
*zaulapion, zoulapin,* see julep Zsamboky, Janos (Johannes Sambucus) 5, 7, 12, 108 *zunniir* 324 Zuretti, C. 0. 408
#### THE
## **MEDIEVAL MEDITERRANEAN**
#### PEOPLES, ECONOMIES AND CULTURES, 400-1453
#### *Editors:* Hugh Kennedy, Paul Magdalino (St. Andrews), David Abulafia (Cambridge), Benjamin Arbel (Tel Aviv), Mark Meyerson (Toronto), Larry]. Simon (Western Michigan University).
This series provides a forum for the publication of scholarly work relating to the interactions of peoples and cultures in the Mediterranean basin and the Black Sea area and is intended for readers with interest in late antiquity, the Middle Ages (Italy, Spain, the Latin East), Byzantium, Islam, the Balkans and the Black Sea. Manuscripts (in English, German and French) should be 60,000 to 120,000 words in length and may include illustrations. The editors would be particularly interested to receive proposals for monograph studies; studies with texts; editions with parallel translations of texts or collections of documents; or translations provided with full annotation.
| doab | 2025-04-07T04:13:02.999510 | 21-5-2022 06:02 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffa52fe3-258a-4af9-9030-900a85a489ae",
"url": "https://library.oapen.org/bitstream/20.500.12657/54647/1/9789004473461.pdf",
"author": "",
"title": "A Byzantine Book on Dream Interpretation",
"publisher": "Brill",
"isbn": "",
"section_idx": 0
} |
ffa90cf5-6ee0-48d5-8212-f65bf49c005f.0 | *Edited by Guy Huynh Thien Duc*
The book Immune Response Activation is aiming to analyse the multifaceted aspects of the immune response, treating a number of representative cases in which the immune response is, on one hand, activated against pathogens, and, on the other hand, involved in pathologic settings, leading to allograft rejection, allergy and autoimmunity. The regulatory mechanisms in which the immune response can be modulated for rendering its effector components more efficient and/or not harmful to the organism is also dissected in translational purposes in cancer immunotherapy, local immunity against bacteria and viruses, as well as in allergy and autoimmunity.
ISBN 978-953-51-1374-4
Immune Response Activation
Photo by selvanegra / iStock
| doab | 2025-04-07T04:13:03.363389 | 20-4-2021 17:49 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffa90cf5-6ee0-48d5-8212-f65bf49c005f",
"url": "https://mts.intechopen.com/storage/books/3811/authors_book/authors_book.pdf",
"author": "",
"title": "Immune Response Activation",
"publisher": "IntechOpen",
"isbn": "9789535113744",
"section_idx": 0
} |
ffa90cf5-6ee0-48d5-8212-f65bf49c005f.1 | Immune Response Activation
*Edited by Guy Huynh Thien Duc*
**IMMUNE RESPONSE**
Edited by **Guy Huynh Thien Duc**
**ACTIVATION**
| doab | 2025-04-07T04:13:03.363649 | 20-4-2021 17:49 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffa90cf5-6ee0-48d5-8212-f65bf49c005f",
"url": "https://mts.intechopen.com/storage/books/3811/authors_book/authors_book.pdf",
"author": "",
"title": "Immune Response Activation",
"publisher": "IntechOpen",
"isbn": "9789535113744",
"section_idx": 1
} |
ffa90cf5-6ee0-48d5-8212-f65bf49c005f.2 | **IMMUNE RESPONSE ACTIVATION**
Edited by **Guy Huynh Thien Duc**
#### **Immune Response Activation**
http://dx.doi.org/10.5772/57004 Edited by Guy Huynh Thien Duc
#### **Contributors**
Mohamed Labib Salem, Abdel-Aziz Zidan, Said Hammad, Karine Breckpot, Tatsuaki Tsuruyama, Jonathan Moorman, Jedidiah Griffin, Marie Moulton, Zhi Yao, Andrew Foey, Maria Ines Becker, Sergio Arancibia, Fabian Alberto Salazar, Miguel Del Campo, Alfredo De Ioannes, Louise Hafner, Danica Hickey, Trudi Collet, Ana Maria Carmona-Ribeiro, Takemi Otsuki, Suni Lee, Megumi Maeda, Hiroaki Hayashi, Hidenori Matsuzaki, Naoko Kumagai-Takei, Yasumitsu Nishimura, Kozo Urakami, Masayasu Kusaka
#### **© The Editor(s) and the Author(s) 2014**
The moral rights of the and the author(s) have been asserted.
All rights to the book as a whole are reserved by INTECH. The book as a whole (compilation) cannot be reproduced, distributed or used for commercial or non-commercial purposes without INTECH's written permission. Enquiries concerning the use of the book should be directed to INTECH rights and permissions department (permissions@intechopen.com).
Violations are liable to prosecution under the governing Copyright Law.
Individual chapters of this publication are distributed under the terms of the Creative Commons Attribution 3.0 Unported License which permits commercial use, distribution and reproduction of the individual chapters, provided the original author(s) and source publication are appropriately acknowledged. If so indicated, certain images may not be included under the Creative Commons license. In such cases users will need to obtain permission from the license holder to reproduce the material. More details and guidelines concerning content reuse and adaptation can be foundat http://www.intechopen.com/copyright-policy.html.
#### **Notice**
Statements and opinions expressed in the chapters are these of the individual contributors and not necessarily those of the editors or publisher. No responsibility is accepted for the accuracy of information contained in the published chapters. The publisher assumes no responsibility for any damage or injury to persons or property arising out of the use of any materials, instructions, methods or ideas contained in the book.
First published in Croatia, 2014 by INTECH d.o.o. eBook (PDF) Published by IN TECH d.o.o. Place and year of publication of eBook (PDF): Rijeka, 2019. IntechOpen is the global imprint of IN TECH d.o.o. Printed in Croatia
Legal deposit, Croatia: National and University Library in Zagreb
Additional hard and PDF copies can be obtained from orders@intechopen.com
Immune Response Activation Edited by Guy Huynh Thien Duc p. cm. ISBN 978-953-51-1374-4 eBook (PDF) ISBN 978-953-51-5394-8
| doab | 2025-04-07T04:13:03.363732 | 20-4-2021 17:49 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffa90cf5-6ee0-48d5-8212-f65bf49c005f",
"url": "https://mts.intechopen.com/storage/books/3811/authors_book/authors_book.pdf",
"author": "",
"title": "Immune Response Activation",
"publisher": "IntechOpen",
"isbn": "9789535113744",
"section_idx": 2
} |
ffa90cf5-6ee0-48d5-8212-f65bf49c005f.3 | We are IntechOpen, the world's leading publisher of Open Access books Built by scientists, for scientists
4,200+
Open access books available
116,000+
International authors and editors
125M+
Downloads
151 Countries delivered to Our authors are among the
Top 1%
most cited scientists
12.2% Contributors from top 500 universities
Selection of our books indexed in the Book Citation Index in Web of Science™ Core Collection (BKCI)
## Interested in publishing with us? Contact book.department@intechopen.com
Numbers displayed above are based on latest data collected. For more information visit www.intechopen.com
| doab | 2025-04-07T04:13:03.364092 | 20-4-2021 17:49 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffa90cf5-6ee0-48d5-8212-f65bf49c005f",
"url": "https://mts.intechopen.com/storage/books/3811/authors_book/authors_book.pdf",
"author": "",
"title": "Immune Response Activation",
"publisher": "IntechOpen",
"isbn": "9789535113744",
"section_idx": 3
} |
ffa90cf5-6ee0-48d5-8212-f65bf49c005f.4 | **Meet the editor**
Guy Huynh Thien Duc is Research Director Emeritus from the CNRS (Centre National de la Recherche Scientifique). Medicine student and Master's degree of Science, he started his career in the Pasteur Institute and the Sorbonne, Faculty of Science in Paris, where he prepared his Ph.D in the field of Immunopathology. Thereafter, as researcher in the CNRS, he has been mainly involved
in the fundamental aspects of Immunology, focusing on Transplantation Immunity and Immunomodulation. In the last two decades, his work at the University Paris XI, southern of Paris, presently inside the structure of INSERM U-1014, Groupe hospitalier Paul-Brousse-Villejuif, is essentially devoted to Immunology, its fundamental aspects and Cancer Immunotherapy.
Contents
**Preface VII**
**Section 2 Immunomodulation 73**
**Stimulation 75**
Karine Breckpot
Moorman
**and Deviation 121** Andrew D. Foey
Chapter 1 **Cationic Nanostructures for Vaccines 3** Ana Maria Carmona-Ribeiro
**Biomedical Applications 45**
Campo and Alfredo De Ioannes
**and Facilitating Infection 105**
Chapter 2 **Mollusk Hemocyanins as Natural Immunostimulants in**
María Inés Becker, Sergio Arancibia, Fabián Salazar, Miguel Del
Joeri J. Pen, Joeri. L. Aerts, Thérèse Liechtenstein, David Escors and
Jeddidiah Griffin, Marie Moulton, Rabab Elmezayen and Jonathan
Chapter 3 **Manipulating Immune Regulatory Pathways to Enhance T Cell**
Chapter 4 **Negative Immunomodulators – Blunting Immunostimulation**
Chapter 5 **Macrophages — Masters of Immune Activation, Suppression**
Chapter 6 **Immunostimulatory Effects of Triggering TLR3 Signaling**
**Pathway — Implication for Cancer Immunotherapy 151** Mohamed Labib Salem, Said M. Hammad, Mohamed R. Elshanshory, Mohamed A. Attia and Abdel-Aziz A. Zidan
**Section 1 Vaccines 1**
### Contents
#### **Preface XI**
Joeri J. Pen, Joeri. L. Aerts, Thérèse Liechtenstein, David Escors and Karine Breckpot
#### **Section 3 Local Immunity 175**
Preface
related to:
plants.
The immune system plays a central role against pathogens that is seen further, extend‐ ing beyond this context in recognizing alloantigens, modified transformed antigens ex‐ pressed on tumoral cells, but unfortunately also self-antigens. As such, the immune response is involved in a large number of clinical settings, among which, protection against pathogenic organisms (bacteria, viruses), transformed tumoral cells and, finally, recognition and reaction against allergens and autoantigens. Therefore, beside the as‐ pect relative to inducing generation of neutralizing protective antibodies and cytotoxic T lymphocytes against pathogens, there is need to consider the resistance developed by some pathogens to the traditional immune effector agents. Moreover, the immune re‐ sponse could also be acerbated in response to allergens and/or targeting the organism's own components leading to allergy and autoimmune diseases, respectively. Thus, along with its stimulation and activation, the immune response has to be tightly con‐ trolled and regulated. In this context, the multifaceted immune response is analyzed in the present book in several review chapters, treating a number of representative cases in which the immune response is, on one hand, activated against pathogens and, on the other hand, involved in pathologic settings, leading to allograft rejection, allergy and autoimmunity. The regulatory mechanisms in which the immune system can be modu‐ lated for rendering its effector components more efficient and not harmful to the organ‐ ism is also dissected in translational purposes in cancer immunotherapy, local
immunity against bacteria and viruses, as well as in allergy and autoimmunity.
naling Pathway — Implication for Cancer Immunotherapy.
In this regard, the book "Immune Response Activation", is divided into four sections
**Vaccines**, with chapters concerning: 1) Cationic Nanostructures for Vaccines, and 2) Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications.
**Immunomodulation**, with chapters: 1) Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation ; 2) Negative Immunomodulators – Blunting Immunosti‐ mulation and Facilitating Infection ; 3) Macrophages — Masters of Immune Activation, Suppression and Deviation, and 4) Immunostimulatory Effects of Triggering TLR3 Sig‐
**Local Immunity**, with chapters : 1) Immune Regulation of Chlamydia trachomatis In‐ fections of the Female Genital Tract, and 2) The Roles of Invariant NKT Cells in Bowel Immunity — Suppression of Tumor Progression and Rejection of Intestinal Trans‐
**Section 4 Autoimmune Disregulation 247**
#### Chapter 9 **Immunostimulation by Silica Particles and the Development of Autoimmune Dysregulation 249** Suni Lee, Hiroaki Hayashi, Megumi Maeda, Hidenori Matsuzaki, Naoko Kumagai-Takei, Ying Chen, Kozo Urakami, Masayasu Kusaka, Yasumitsu Nishimura and Takemi Otsuki
### Preface
**Section 3 Local Immunity 175**
**VI** Contents
**Female Genital Tract 177**
**Transplants 227**
**Section 4 Autoimmune Disregulation 247**
Chapter 7 **Immune Regulation of Chlamydia trachomatis Infections of the**
Louise M. Hafner, Trudi A. Collet and Danica K. Hickey
Chapter 9 **Immunostimulation by Silica Particles and the Development of**
Kusaka, Yasumitsu Nishimura and Takemi Otsuki
Suni Lee, Hiroaki Hayashi, Megumi Maeda, Hidenori Matsuzaki, Naoko Kumagai-Takei, Ying Chen, Kozo Urakami, Masayasu
**Suppression of Tumor Progression and Rejection of Intestinal**
Chapter 8 **The Roles of Invariant NKT Cells in Bowel Immunity —**
Tatsuaki Tsuruyama and Wulamujiang Aini
**Autoimmune Dysregulation 249**
The immune system plays a central role against pathogens that is seen further, extend‐ ing beyond this context in recognizing alloantigens, modified transformed antigens ex‐ pressed on tumoral cells, but unfortunately also self-antigens. As such, the immune response is involved in a large number of clinical settings, among which, protection against pathogenic organisms (bacteria, viruses), transformed tumoral cells and, finally, recognition and reaction against allergens and autoantigens. Therefore, beside the as‐ pect relative to inducing generation of neutralizing protective antibodies and cytotoxic T lymphocytes against pathogens, there is need to consider the resistance developed by some pathogens to the traditional immune effector agents. Moreover, the immune re‐ sponse could also be acerbated in response to allergens and/or targeting the organism's own components leading to allergy and autoimmune diseases, respectively. Thus, along with its stimulation and activation, the immune response has to be tightly con‐ trolled and regulated. In this context, the multifaceted immune response is analyzed in the present book in several review chapters, treating a number of representative cases in which the immune response is, on one hand, activated against pathogens and, on the other hand, involved in pathologic settings, leading to allograft rejection, allergy and autoimmunity. The regulatory mechanisms in which the immune system can be modu‐ lated for rendering its effector components more efficient and not harmful to the organ‐ ism is also dissected in translational purposes in cancer immunotherapy, local immunity against bacteria and viruses, as well as in allergy and autoimmunity.
In this regard, the book "Immune Response Activation", is divided into four sections related to:
**Vaccines**, with chapters concerning: 1) Cationic Nanostructures for Vaccines, and 2) Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications.
**Immunomodulation**, with chapters: 1) Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation ; 2) Negative Immunomodulators – Blunting Immunosti‐ mulation and Facilitating Infection ; 3) Macrophages — Masters of Immune Activation, Suppression and Deviation, and 4) Immunostimulatory Effects of Triggering TLR3 Sig‐ naling Pathway — Implication for Cancer Immunotherapy.
**Local Immunity**, with chapters : 1) Immune Regulation of Chlamydia trachomatis In‐ fections of the Female Genital Tract, and 2) The Roles of Invariant NKT Cells in Bowel Immunity — Suppression of Tumor Progression and Rejection of Intestinal Trans‐ plants.
#### XII Preface
**Autoimmune Disregulation**, with the chapter : Immunostimulation by Silica Particles and the Development of Autoimmune Dysregulation.
#### **Guy Huynh Thien Duc**
**Vaccines**
**Section 1**
Research Director Emeritus from the CNRS (Centre National de la Recherche Scientifique), University Paris XI, INSERM U-1014-Groupe Hospitalier Paul-Brousse, France
#### **Acknowledgments**
We thank Chaobin Zhu for his assistance.
**Section 1**
### **Vaccines**
**Autoimmune Disregulation**
VIII Preface
, with the chapter
and the Development of Autoimmune Dysregulation.
: Immunostimulation by Silica Particles
Research Director Emeritus from the CNRS (Centre National de la Recherche Scientifique),
We thank Chaobin Zhu for his assistance.
INSERM U-1014-Groupe Hospitalier Paul-Brousse,
**Guy Huynh Thien Duc**
University Paris XI,
**Acknowledgments**
France
**Chapter 1**
**Cationic Nanostructures for Vaccines**
Additional information is available at the end of the chapter
Cationic lipid bilayers, particles, polysaccharides and a variety of hybrid nanostructures provide adequate matrixes for supporting antigens such as peptides, proteins, DNA or oligonucleotides on model surfaces (latex, silica, silicon wafers, self-assembled monolayers, metals, polymers, insoluble drugs, biological cells and viruses).Particulate vaccines are currently an area receiving a high level of attention [1-3]. Particles deliver both antigen and adjuvant into the same population of antigen presenting cells limiting both the systemic distribution of the adjuvant and its potential toxicity [1]. Biological particlesrepresented by live or attenuated bacterial vacines, engineered biological vectors and virus-like particles are often less safe than synthetic particulates for which quality control and validation in vaccine development and production are more rapid [2]. While developing novel particulate vaccines, particle size and charge do matter [2]. Virus-sized particles (20–200 nm) are usually taken up by endocytosis via clathrin-coated vesicles, caveolae or their independent receptors [4], and preferentially ingested by dendritic cells (DC). Larger sized particles such as bacteria (500– 5000 nm) are predominantly taken up by phagocytosis, and primarily ingested by macro‐ phages. All particles used in vaccine formulations are consequently internalized efficiently by antigen presenting cells by one or a combination of the quoted mechanisms [5, 6]. Particles with diameters smaller than 500 nm, in particular the nanometric ones with sizes in the 40– 100 nm range are more eficient to promote CD8 and CD4 type 1 T cell responses than those with diameters above 500 nm, although the latter could induce good type 2 and antibody responses [5]. Cationic microparticles are effectively taken up both by macrophages and dendritic cells since electrostatic attraction promotes particle binding and subsequent inter‐ nalization. Cationized polymeric particles carrying antigen significantly enhanced both antibody production and cytotoxic T cells at low antigen dose [7] and induced maturation of dendritic cells [8, 9]. As a consequence particles for vaccines should be positive and available over a range of sizes. Cationic particles of aluminium compounds, identified as having
> © 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Ana Maria Carmona-Ribeiro
http://dx.doi.org/10.5772/57543
**1. Introduction**
#### **Chapter 1**
### **Cationic Nanostructures for Vaccines**
#### Ana Maria Carmona-Ribeiro
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/57543
#### **1. Introduction**
Cationic lipid bilayers, particles, polysaccharides and a variety of hybrid nanostructures provide adequate matrixes for supporting antigens such as peptides, proteins, DNA or oligonucleotides on model surfaces (latex, silica, silicon wafers, self-assembled monolayers, metals, polymers, insoluble drugs, biological cells and viruses).Particulate vaccines are currently an area receiving a high level of attention [1-3]. Particles deliver both antigen and adjuvant into the same population of antigen presenting cells limiting both the systemic distribution of the adjuvant and its potential toxicity [1]. Biological particlesrepresented by live or attenuated bacterial vacines, engineered biological vectors and virus-like particles are often less safe than synthetic particulates for which quality control and validation in vaccine development and production are more rapid [2]. While developing novel particulate vaccines, particle size and charge do matter [2]. Virus-sized particles (20–200 nm) are usually taken up by endocytosis via clathrin-coated vesicles, caveolae or their independent receptors [4], and preferentially ingested by dendritic cells (DC). Larger sized particles such as bacteria (500– 5000 nm) are predominantly taken up by phagocytosis, and primarily ingested by macro‐ phages. All particles used in vaccine formulations are consequently internalized efficiently by antigen presenting cells by one or a combination of the quoted mechanisms [5, 6]. Particles with diameters smaller than 500 nm, in particular the nanometric ones with sizes in the 40– 100 nm range are more eficient to promote CD8 and CD4 type 1 T cell responses than those with diameters above 500 nm, although the latter could induce good type 2 and antibody responses [5]. Cationic microparticles are effectively taken up both by macrophages and dendritic cells since electrostatic attraction promotes particle binding and subsequent inter‐ nalization. Cationized polymeric particles carrying antigen significantly enhanced both antibody production and cytotoxic T cells at low antigen dose [7] and induced maturation of dendritic cells [8, 9]. As a consequence particles for vaccines should be positive and available over a range of sizes. Cationic particles of aluminium compounds, identified as having
© 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
immunostimulatory properties more than 70 years ago, remain the only type of adjuvant licensed world-wide [3]. Oil-in-water emulsions or virosomes were licensed in some countries in compositions for influenza [10] or hepatitis B vaccines [11]. However, both of these adjuvants are characterised by inducing humoral immune response and are thus effective in elevating serum antibody titers whereas their ability to elicit cell-mediated immune response is limited [12]. For vaccines against influenza, malaria and HIV, the induction of a humoral response is insufficient and a substantial complementary cell-mediated immune response is necessary for adequate protection. For vaccines against tuberculosis, a cellular response seems to be the sole effector mechanism required for protection [13]. Among the cationic compounds used to produce cationic nanostructures to present antigens for vaccines formulations, some cationic compounds appear to be of special relevance.They are the cationic and synthetic lipid named dioctadecyldimethylammonium bromide (DODAB), the cationic polysaccharide named chitosan (CH) and its derivatives with pH-independent positive charge and the biocompatible cationic polymeric nanostructures.This chapter presents and critically evaluates these three types of cationic nanostructures.
BF-or are dispersions of cationic particles with controllable nature and size as obtained after covering silica or polystyrene sulfate latex (PSS) with a cationic DODAB bilayer (Figure 1).
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 5
**Figure 1.** Cationic nanostructures and particles based on the cationic lipid DODAB useful to present antigens (repro‐ duced with permission from [27]).The DODAB bilayer fragments are usually obtained at low ionic strength by sonica‐ tion with a macrotip.Particles acting as supports for the cationic bilayer can be organic or inorganic. Reprinted from Vaccine, 27/42, Nilton Lincopan,Noelí M. Espíndola,Adelaide J. Vaz,Maria Helena B. da Costa,Eliana Faquim-Mauro,Ana M. Carmona-Ribeiro, Novel immunoadjuvants based on cationic lipid: Preparation, characterization and
DODAB BF may interact with proteins both via the hydrophobic effect and the electrostatic attraction at low ionic strength. This interaction has been properly characterized by means of dynamic light scattering for sizing, zeta-potential analysis and evaluation of immunoadjuvant activity *in vivo*. The model antigens employed with DODAB-based cationic nanostructures were bovine serum albumin (BSA), purified 18 kDa/14 kDa antigens from *Taenia crassiceps* cysticerci (18/14-*Tcra*) or a recombinant heat-shock protein from *Mycobacterium leprae* [27]. Antigen choices were due to different reasons. BSA adsorption at interfaces [28] and, specifi‐ cally, onto large DODAB vesicles [29] is well described and has been useful to prevent nonspecific binding in immunoassays, biosensing and proteomics applications [28,30].The purified 18 kDa/14 kDa antigens from *Taenia crassiceps* cisticerci are proteins specific for this parasite found as circulating antigens and often employed in immunodiagnosis;they can be obtained from *in-vitro* cultures of *T. crassiceps* cysticerci in hybridoma media from vesicles budding from cysts contain the excretory/secretory (ES) antigens [31,32]. Human and pig infections by *T. solium* and *crassiceps*, respectively, represent an important health problem, with socio-economical repercussions affecting many countries in Latin America, Asia and Africa
activity *in vivo*, 5760-5771.Copyright 2009, with permission from Elsevier.
#### **2. Cationic nanostructures based on DODAB**
DODAB is a cationic lipid which can be dispersed ultrasonically in aqueous solution using a macrotip probe [14]. Thereby, bilayer vesicles are simultaneously obtained and disrupted yielding nano-sized bilayer disks or cationic bilayer fragments (BF) stabilized by electrostatic repulsion at low ionic strength [14,15].DODAB BF have previously been used as antimicrobial agents [16] or in the production of lipid-covered particles such as bilayer-coated silica [17] or latex [18]. These cationic, bilayer-covered latex or silica particles where the bilayer is solely composed of DODAB were recently employed to present antigens to the immune system with better results than alum as adjuvants for induction of cellular immune responses [19, 20]. The open DODAB BF in aqueous dispersions are different from their mother vesicles presenting the following features: (i) osmotic non-responsiveness of the dispersion indicative of absence of inner vesicle compartment [21]; (ii) discoidal shape with disks exhibiting one bilayer thickness as visualised by means of transmission electronic microscopy (TEM) after electronic staining the nanostructures [22]; (iii) cryo-TEM micrographs performed without any staining [23]; (iv) fluid and solid state coexistence and complex formation with oppositely charged surfactant [24]; (v) solubilization of hydrophobic drugs at the borders of DODAB bilayer fragments, which does not occur for DODAB closed bilayer vesicles [25]. These bilayer fragments have more fluid environments at their edges that are absent in closed bilayer systems such as vesicles or liposomes. Due to its cylindrical molecular shape, DODAB molecules selfassemble in aqueous solutions as bilayers instead of micelles as shown in the seventies by transmission electron microscopy [26]. Supramolecular assemblies of DODAB bilayer frag‐ ments by themselves or after interaction with supporting particles have been combined with different model antigens in separate and tested as immunoadjuvants. The cationic nanoadju‐ vants with DODAB BF are either reduced to a single-component, nanosized system-DODAB BF-or are dispersions of cationic particles with controllable nature and size as obtained after covering silica or polystyrene sulfate latex (PSS) with a cationic DODAB bilayer (Figure 1).
immunostimulatory properties more than 70 years ago, remain the only type of adjuvant licensed world-wide [3]. Oil-in-water emulsions or virosomes were licensed in some countries in compositions for influenza [10] or hepatitis B vaccines [11]. However, both of these adjuvants are characterised by inducing humoral immune response and are thus effective in elevating serum antibody titers whereas their ability to elicit cell-mediated immune response is limited [12]. For vaccines against influenza, malaria and HIV, the induction of a humoral response is insufficient and a substantial complementary cell-mediated immune response is necessary for adequate protection. For vaccines against tuberculosis, a cellular response seems to be the sole effector mechanism required for protection [13]. Among the cationic compounds used to produce cationic nanostructures to present antigens for vaccines formulations, some cationic compounds appear to be of special relevance.They are the cationic and synthetic lipid named dioctadecyldimethylammonium bromide (DODAB), the cationic polysaccharide named chitosan (CH) and its derivatives with pH-independent positive charge and the biocompatible cationic polymeric nanostructures.This chapter presents and critically evaluates these three
DODAB is a cationic lipid which can be dispersed ultrasonically in aqueous solution using a macrotip probe [14]. Thereby, bilayer vesicles are simultaneously obtained and disrupted yielding nano-sized bilayer disks or cationic bilayer fragments (BF) stabilized by electrostatic repulsion at low ionic strength [14,15].DODAB BF have previously been used as antimicrobial agents [16] or in the production of lipid-covered particles such as bilayer-coated silica [17] or latex [18]. These cationic, bilayer-covered latex or silica particles where the bilayer is solely composed of DODAB were recently employed to present antigens to the immune system with better results than alum as adjuvants for induction of cellular immune responses [19, 20]. The open DODAB BF in aqueous dispersions are different from their mother vesicles presenting the following features: (i) osmotic non-responsiveness of the dispersion indicative of absence of inner vesicle compartment [21]; (ii) discoidal shape with disks exhibiting one bilayer thickness as visualised by means of transmission electronic microscopy (TEM) after electronic staining the nanostructures [22]; (iii) cryo-TEM micrographs performed without any staining [23]; (iv) fluid and solid state coexistence and complex formation with oppositely charged surfactant [24]; (v) solubilization of hydrophobic drugs at the borders of DODAB bilayer fragments, which does not occur for DODAB closed bilayer vesicles [25]. These bilayer fragments have more fluid environments at their edges that are absent in closed bilayer systems such as vesicles or liposomes. Due to its cylindrical molecular shape, DODAB molecules selfassemble in aqueous solutions as bilayers instead of micelles as shown in the seventies by transmission electron microscopy [26]. Supramolecular assemblies of DODAB bilayer frag‐ ments by themselves or after interaction with supporting particles have been combined with different model antigens in separate and tested as immunoadjuvants. The cationic nanoadju‐ vants with DODAB BF are either reduced to a single-component, nanosized system-DODAB
types of cationic nanostructures.
4 Immune Response Activation
**2. Cationic nanostructures based on DODAB**
**Figure 1.** Cationic nanostructures and particles based on the cationic lipid DODAB useful to present antigens (repro‐ duced with permission from [27]).The DODAB bilayer fragments are usually obtained at low ionic strength by sonica‐ tion with a macrotip.Particles acting as supports for the cationic bilayer can be organic or inorganic. Reprinted from Vaccine, 27/42, Nilton Lincopan,Noelí M. Espíndola,Adelaide J. Vaz,Maria Helena B. da Costa,Eliana Faquim-Mauro,Ana M. Carmona-Ribeiro, Novel immunoadjuvants based on cationic lipid: Preparation, characterization and activity *in vivo*, 5760-5771.Copyright 2009, with permission from Elsevier.
DODAB BF may interact with proteins both via the hydrophobic effect and the electrostatic attraction at low ionic strength. This interaction has been properly characterized by means of dynamic light scattering for sizing, zeta-potential analysis and evaluation of immunoadjuvant activity *in vivo*. The model antigens employed with DODAB-based cationic nanostructures were bovine serum albumin (BSA), purified 18 kDa/14 kDa antigens from *Taenia crassiceps* cysticerci (18/14-*Tcra*) or a recombinant heat-shock protein from *Mycobacterium leprae* [27]. Antigen choices were due to different reasons. BSA adsorption at interfaces [28] and, specifi‐ cally, onto large DODAB vesicles [29] is well described and has been useful to prevent nonspecific binding in immunoassays, biosensing and proteomics applications [28,30].The purified 18 kDa/14 kDa antigens from *Taenia crassiceps* cisticerci are proteins specific for this parasite found as circulating antigens and often employed in immunodiagnosis;they can be obtained from *in-vitro* cultures of *T. crassiceps* cysticerci in hybridoma media from vesicles budding from cysts contain the excretory/secretory (ES) antigens [31,32]. Human and pig infections by *T. solium* and *crassiceps*, respectively, represent an important health problem, with socio-economical repercussions affecting many countries in Latin America, Asia and Africa [31]. The 18 kDa-hsp protein is a heat-shock protein of *M. leprae* displaying pronounced immunogenicity and considered able to induce proliferation of peripheral blood mononuclear cells and T-cell lines from *M. leprae* vaccinated subjects [33] andavailable at large amounts for studies on vaccine formulation; its overexpression and scaling-up in *Saccharomyces cerevisae* have already been described and the recombinant protein can be produced in large scale [34,35]. Nanostructured cationic adjuvant/antigen complexes based on DODAB were charac‐ terized over a range of adjuvant and antigen concentrations. Figure 2 shows the effect of antigen and adjuvant concentration on physical properties of the dispersions. Stable cationic nanostructures are available over a limited range of adjuvant and antigen concentrations which clearly depend on antigen nature and are different for different antigens.
The humoral and cellular immune responses induced by stable cationic adjuvant/antigen complexes were evaluated in mice from determination of antigen-specific-IgG antibody in serum by ELISA, delayed type hypersensitivity (DH) reactions from footpad swelling tests and cytokines analysis. The results evidence the good colloid stability of the complexes, complete absence of toxicity in mice (i.e. local or general reactions) and their potential utility to induce Th 1 immune response at reduced doses of cationic and toxic DODAB lipid. Possibly due to its chemical stability and low cost when compared to other natural or synthetic lipids, DODAB use as an immunoadjuvant started more than forty years ago [36,37]. This was well before the bilayer nature of DODAB self-assembly in water solution had been described [26]. DODAB as an effective immunoadjuvant has been intensively investigated aiming at subunit vaccine design [38-41]. A major problem of liposomal formulations based on DODAB has been the usually high concentration employed, namely 1-10 mM DODAB [38-41]. DODAB is a cationic lipid and as such, cytotoxic [14-16] requiring dose minimization for administration in vivo. Using the DODAB bilayer fragments to present antigens only 0.1 mM DODAB was employed [27]. The large cellular imune response achieved might be related to the total surface area available for antigen association, which is much larger in the BF dispersion than in closed, large and sometimes multibilayered liposomes. Moreover, the hydrophobic interaction possibly available from BF edges could be an additional and powerful driving force for antigen adsorption and presentation. The second advantage of the bilayer fragments besides their large surfaces area is their size. Depending on sonication power and time plus composition of the dispersing medium, which determine colloid stability, DODAB BF/antigens complexes may have their sizes reduced to and stabilized at a few tenths of nanometers acting similarly to solid inert beads of nanometric size (40-50 nm). These beads turned out to be effective for antigen delivery to antigen-presenting cells (APC), generating potent and combined humoral and CD8+T cell immunity [3]. They are also expected to be optimal for dendritic cells uptake since their size is inside the range of particle diameters (500 nm and below) for optimal dendritic cells uptake of antigens and elicitation of an adequate cellular response [2,5,42]. In reconstituted pig gastric mucus, sub-200 nm particulates from poly (D, L-lactic-co-glycolic) acid and DODAB condensing DNA on particles surface exhibited improved transport rates, stability in mucus, and ability to transfect cells [43]. Silica/DODAB, latex/DODAB and DODAB BF (Figure 1) are available over the sub-200 nm range of sizes thus presenting potential also for design of mucosal vaccines. The third advantage of the bilayer fragments is the absence of depots at the injection site [27]. These depots have been reported for other DODAB formula‐
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 7
tions at concentrations much higher than 0.1 mM [37,41,44].
Consistently with electrostatic forces between negatively charged antigens and positively charged DODAB BF, cationic bilayer fragments readily adsorb BSA, 18-14/*Tcra* and the hsp-18 kDa recombinant protein.Rapid and extensive adsorption of BSA, anti-BSA or ovalbumin onto large DODAB liposomes was indeed reported previously [29,45,46].On the other hand, bacteria, fungus and eukaryotic cells were also found to adsorb DODAB vesicles and bilayer fragments with high affinity at low ionic strength [14-16], so that cationic DODAB liposomes readily adsorb antigen, such as ovalbumin, and bind avidly to dendritic cells [41], thereby enhancing antigen uptake. Delivery of antigen to cells by immediate contact with the cell surface via electrostatic interaction followed by the induction of active uptake seems to be the mechanism behind the ability of DODAB liposomes or bilayer fragments to act as immunoad‐ juvants. Sizes and zeta-potentials for assemblies of antigen and cationic lipid based adjuvants
**Figure 2.** Effect of time and DODAB concentration on the zeta-average diameter of DODAB/BSA (A) or DODAB BF/ 18/14-Tcra complexes (B). In (A), the kinetics were obtained upon adding DODAB BF at a final concentration of 0.005 ((▷), 0.01 (△), 0.02 (◁), 0.05 (○), 0.5 (◇), 0.8 (□), and 1.0 mM DODAB (◁) to 0.5 mg/mL BSA. In (B), the kinetics were obtained upon adding DODAB BF at a final concentration of 0.0005 (□), 0.001 (○), 0.005 (▷), 0.01(▽), 0.05(◇), and 0.1 mM DODAB (△) to 0.05 mg/mL 18/14-Tcra. The effect of [DODAB] on zeta-potential of complexes with 0.5 mg/mL BSA (C) or 0.05 mg/mL 18/14-Tcra (D) were obtained after 1 h interaction at 25 °C in 1 mM NaCl.Reproduced with permission from reference [27].Reprinted from Vaccine, 27/42, Nilton Lincopan,Noelí M. Espíndola,Adelaide J. Vaz,Maria Helena B. da Costa,Eliana Faquim-Mauro,Ana M. Carmona-Ribeiro,Novel immunoadjuvants based on cati‐ onic lipid: Preparation, characterization and activity in vivo,5760-5771.Copyright 2009, with permission from Elsevier.
The humoral and cellular immune responses induced by stable cationic adjuvant/antigen complexes were evaluated in mice from determination of antigen-specific-IgG antibody in serum by ELISA, delayed type hypersensitivity (DH) reactions from footpad swelling tests and cytokines analysis. The results evidence the good colloid stability of the complexes, complete absence of toxicity in mice (i.e. local or general reactions) and their potential utility to induce Th 1 immune response at reduced doses of cationic and toxic DODAB lipid. Possibly due to its chemical stability and low cost when compared to other natural or synthetic lipids, DODAB use as an immunoadjuvant started more than forty years ago [36,37]. This was well before the bilayer nature of DODAB self-assembly in water solution had been described [26]. DODAB as an effective immunoadjuvant has been intensively investigated aiming at subunit vaccine design [38-41]. A major problem of liposomal formulations based on DODAB has been the usually high concentration employed, namely 1-10 mM DODAB [38-41]. DODAB is a cationic lipid and as such, cytotoxic [14-16] requiring dose minimization for administration in vivo. Using the DODAB bilayer fragments to present antigens only 0.1 mM DODAB was employed [27]. The large cellular imune response achieved might be related to the total surface area available for antigen association, which is much larger in the BF dispersion than in closed, large and sometimes multibilayered liposomes. Moreover, the hydrophobic interaction possibly available from BF edges could be an additional and powerful driving force for antigen adsorption and presentation. The second advantage of the bilayer fragments besides their large surfaces area is their size. Depending on sonication power and time plus composition of the dispersing medium, which determine colloid stability, DODAB BF/antigens complexes may have their sizes reduced to and stabilized at a few tenths of nanometers acting similarly to solid inert beads of nanometric size (40-50 nm). These beads turned out to be effective for antigen delivery to antigen-presenting cells (APC), generating potent and combined humoral and CD8+T cell immunity [3]. They are also expected to be optimal for dendritic cells uptake since their size is inside the range of particle diameters (500 nm and below) for optimal dendritic cells uptake of antigens and elicitation of an adequate cellular response [2,5,42]. In reconstituted pig gastric mucus, sub-200 nm particulates from poly (D, L-lactic-co-glycolic) acid and DODAB condensing DNA on particles surface exhibited improved transport rates, stability in mucus, and ability to transfect cells [43]. Silica/DODAB, latex/DODAB and DODAB BF (Figure 1) are available over the sub-200 nm range of sizes thus presenting potential also for design of mucosal vaccines. The third advantage of the bilayer fragments is the absence of depots at the injection site [27]. These depots have been reported for other DODAB formula‐ tions at concentrations much higher than 0.1 mM [37,41,44].
[31]. The 18 kDa-hsp protein is a heat-shock protein of *M. leprae* displaying pronounced immunogenicity and considered able to induce proliferation of peripheral blood mononuclear cells and T-cell lines from *M. leprae* vaccinated subjects [33] andavailable at large amounts for studies on vaccine formulation; its overexpression and scaling-up in *Saccharomyces cerevisae* have already been described and the recombinant protein can be produced in large scale [34,35]. Nanostructured cationic adjuvant/antigen complexes based on DODAB were charac‐ terized over a range of adjuvant and antigen concentrations. Figure 2 shows the effect of antigen and adjuvant concentration on physical properties of the dispersions. Stable cationic nanostructures are available over a limited range of adjuvant and antigen concentrations
**Figure 2.** Effect of time and DODAB concentration on the zeta-average diameter of DODAB/BSA (A) or DODAB BF/ 18/14-Tcra complexes (B). In (A), the kinetics were obtained upon adding DODAB BF at a final concentration of 0.005 ((▷), 0.01 (△), 0.02 (◁), 0.05 (○), 0.5 (◇), 0.8 (□), and 1.0 mM DODAB (◁) to 0.5 mg/mL BSA. In (B), the kinetics were obtained upon adding DODAB BF at a final concentration of 0.0005 (□), 0.001 (○), 0.005 (▷), 0.01(▽), 0.05(◇), and 0.1 mM DODAB (△) to 0.05 mg/mL 18/14-Tcra. The effect of [DODAB] on zeta-potential of complexes with 0.5 mg/mL BSA (C) or 0.05 mg/mL 18/14-Tcra (D) were obtained after 1 h interaction at 25 °C in 1 mM NaCl.Reproduced with permission from reference [27].Reprinted from Vaccine, 27/42, Nilton Lincopan,Noelí M. Espíndola,Adelaide J. Vaz,Maria Helena B. da Costa,Eliana Faquim-Mauro,Ana M. Carmona-Ribeiro,Novel immunoadjuvants based on cati‐ onic lipid: Preparation, characterization and activity in vivo,5760-5771.Copyright 2009, with permission from Elsevier.
which clearly depend on antigen nature and are different for different antigens.
6 Immune Response Activation
Consistently with electrostatic forces between negatively charged antigens and positively charged DODAB BF, cationic bilayer fragments readily adsorb BSA, 18-14/*Tcra* and the hsp-18 kDa recombinant protein.Rapid and extensive adsorption of BSA, anti-BSA or ovalbumin onto large DODAB liposomes was indeed reported previously [29,45,46].On the other hand, bacteria, fungus and eukaryotic cells were also found to adsorb DODAB vesicles and bilayer fragments with high affinity at low ionic strength [14-16], so that cationic DODAB liposomes readily adsorb antigen, such as ovalbumin, and bind avidly to dendritic cells [41], thereby enhancing antigen uptake. Delivery of antigen to cells by immediate contact with the cell surface via electrostatic interaction followed by the induction of active uptake seems to be the mechanism behind the ability of DODAB liposomes or bilayer fragments to act as immunoad‐ juvants. Sizes and zeta-potentials for assemblies of antigen and cationic lipid based adjuvants depend on cationic lipid and antigen concentrations. Adjuvant-antigen stability around sizes that are close to the one of adjuvants themselves indicates that the proteins readily adsorb and stabilize them. The adjuvants also stabilized the proteins acting as important dispersing nanocarriers able to induce remarkable degree of protein disaggregation by attaching the proteins either electrostatically or hydrophobically to their structure. At [DODAB] ≤ 0.1 mM and 0.001-0.05 mg/mL of antigen concentration, DODAB based adjuvant /antigen assemblies are cationic, well-dispersed, colloidally stable and immunogenic combining the advantages of low DODAB dose, low cost, controllable sizes for optimal dendritic cells uptake, high chemical stability, ability to incorporate multiple antigens and minimization of toxicity. Their perform‐ ance is remarkably superior to the one of alum as adjuvant regarding Th1 mediated responses. In contrast to alum or cationic liposomes at 1-10 mM of cationic lipid, local or systemic adverse effects in mice were completely absent at 0.1-0.01mM DODAB.
In viral infections, Il-12 enhances the cytotoxic activity of natural killer cells so that NK cellmediated killing of virus-infected cells eliminates the reservoir of infection. In this respect, vaccination against the dengue virus is urgently needed in tropical or neotropical regions of the planet and some recombinant DNA vaccines expressing membrane and envelope of viral proteins have been proposed [49]. Possibly the cationic adjuvants available from our group would properly enhance the required Th-1 response for a more effective vaccination against
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 9
Another possible application for the novel adjuvants might be in immunotherapy for tumors. This approach is based in augmentation of host immunity to tumors with tumor vaccines. Immune responses that are able of killing tumor cells consist of CTLs, NK cells, and activated macrophages and these may be actively enhanced by vaccination with tumor cells or antigens, administration of tumors modified to express high levels of cytokines that stimulate T cell proliferation and differentiation, and systemic administration of cytokines [49]. The induction of T cell responses in tumors depends on processing and presentation of tumor antigens to T cells by professional antigen-presenting cells (APCs) which might internalize the tumor antigen adsorbed onto the novel cationic adjuvants. These APCs may stimulate CD8+T cells and CD4+helper T lymphocytes to differentiate for recognition and killing of tumor cells.
Naïve CD4+T cells may differentiate into distinct subsets, such as Th1 and Th2 cells in response to different antigens. For example, the enhancement in production of IL-10 and Il-13 by lymphonode cells elicited by the antigens of *Taenia crassiceps* presented by the DODAB BF adjuvant can be appreciated in Figure 3 [27]. These cytokines are typically associated with responses to allergens and parasites such as helminths and mediate differentiation of CD4+-T cells into Th2 cells [50]. Consistently, low levels of these cytokines were elicited by the *M. leprae* antigen presented by the novel adjuvants (Figure 3). Responses were indeed different for the helminthes and the bacteria antigens and very antigen-specific as they should be [27]. The size, charge and hydrophobic features of DODAB BF led to novel applications in solubi‐ lization of hydrophobic drugs [25,51,52], production of biomimetic particles from bilayer coverage of silica [17] or polystyrene particles [53] and design of vaccines [27].Recently, BF was also combined with oligonucleotides [54]. Since synthetic oligonucleotides can inhibit the replication of the Rous sarcoma virus [55], antisense oligonucleotides have been considered a great promise as therapeutic agents and several oligonucleotide-based formulations have reached the clinical trial phase [56,57]. Antisense oligonucleotides have also been extensively used in research on gene expression and function [58-60], vaccine formulation [61], allergy [62] and cancer therapeutics [63].Major obstacles as their degradation by nucleases and poor delivery to the target cells [60,64] suggest the essential role of suitable carriers able to protect oligonucleotides in the biological milieu [60,63–65].There are peculiar features for the interac‐ tion between BF and oligonucleotides in comparison to other electrolytes.Effects of salt, dAMP or poly (dA) concentration on BF size and zeta-potential are shown on Figure 4 taken from reference [54]. From 0 to 0.25 mM salt, Dz and zeta-potentials decreased with salt concentration possibly due to massive phosphate anion binding. From 0.25 to 2.5 mM of divalent salt, Dz increased but zeta-potential remained approximately constant and low (Fig. 4A and D).Dz and zeta-potential decreased with dAMP concentration (0– 2.5 mM) (Fig. 4B and E). At 0.05 mM poly (dA) and 0.5 mM DODAB, extensive BF aggregation and/or fusion took place as depicted from large Dz (N 500 nm) (Fig. 4C) and zero of zeta-potential (Fig. 4F). The screening of DODAB
dengue.
An important component of the early innate immune response to viruses and bacteria is the secretion of cytokines, which mediate many of the effector functions of innate immunity. IL-10 is an inhibitor of activated macrophages and dendritic cells and is an example of negative feedback regulation because it is produced by macrophages to inhibit their function. This cytokine also inhibits the production of IL-12 and expression of class II major histocompati‐ bility (MHC) molecules. IL-12 is also secreted by macrophages and dendritic cells inducing T cells differentiation into Th1 and natural killer (NK) cells with increased IFN-gamma synthesis and cytotoxic activity. IL-12 and IFN-gamma are the most important cytokines in innate responses to intracellular bacteria such as *Mycobacterium leprae* or tuberculosis [13]. Figure 3 shows the high levels of IL-12 and IFN-gamma induced by the novel cationic adjuvants while presenting the hsp-18 kDa of *M. leprae* to lymphonode cells suggesting a possible application of the novel adjuvants for the design of subunit vaccines against intracelular bacteria. As in DH, adaptive immunity against intracellular bacteria is principally cell mediated and consists of activation of macrophages by CD4+T cells as well as killing of infected cells by CD8+cyto‐ toxic T lymphocytes (CTL).
On basis of IL-12 enhancement of IFN-gamma production and development of Th1 cells, this interleukin itself has been used as a vaccine adjuvant for many infections that are combated by cell-mediated immunity, e.g. leishmaniasis [47]. Subunit vaccines against protozoa that survive within macrophages require as principal defense mechanism cell-mediated immunity, particularly directed to macrophage activation by Th1 cell-derived cytokines. Leishmaniasis mucocutaneous and disseminated is caused by *Leishmania donovani* and CD4+Th1 cells are required to activate macrophages to kill phagocytosed parasites. Resistance to the infection is associated with activation of Leishmania-specific Th1 CD4+T cells which produce IFN-gamma and thereby activate macrophages to destroy intracellular parasites. Conversely, activation of Th2 cells by the protozoan results in increased parasite survival and exarcerbation of lesions because of the macrophage-suppressive actions of Th2 cytokines [47]. Other significant example is the protective role played by CD8+T cells in immunity to the hepatic stages of malaria. These effects may be mediated by direct killing of sporozoite-infected hepatocytes or indirectly by the secretion of IFN-gamma and activation of hepatocytes to produce nitric oxide and other agents that kill parasites. IL-12 induces resistance to sporozoite challenge in rodents and nonhuman primates, presumably by stimulating IFN-gamma production [48].
In viral infections, Il-12 enhances the cytotoxic activity of natural killer cells so that NK cellmediated killing of virus-infected cells eliminates the reservoir of infection. In this respect, vaccination against the dengue virus is urgently needed in tropical or neotropical regions of the planet and some recombinant DNA vaccines expressing membrane and envelope of viral proteins have been proposed [49]. Possibly the cationic adjuvants available from our group would properly enhance the required Th-1 response for a more effective vaccination against dengue.
depend on cationic lipid and antigen concentrations. Adjuvant-antigen stability around sizes that are close to the one of adjuvants themselves indicates that the proteins readily adsorb and stabilize them. The adjuvants also stabilized the proteins acting as important dispersing nanocarriers able to induce remarkable degree of protein disaggregation by attaching the proteins either electrostatically or hydrophobically to their structure. At [DODAB] ≤ 0.1 mM and 0.001-0.05 mg/mL of antigen concentration, DODAB based adjuvant /antigen assemblies are cationic, well-dispersed, colloidally stable and immunogenic combining the advantages of low DODAB dose, low cost, controllable sizes for optimal dendritic cells uptake, high chemical stability, ability to incorporate multiple antigens and minimization of toxicity. Their perform‐ ance is remarkably superior to the one of alum as adjuvant regarding Th1 mediated responses. In contrast to alum or cationic liposomes at 1-10 mM of cationic lipid, local or systemic adverse
An important component of the early innate immune response to viruses and bacteria is the secretion of cytokines, which mediate many of the effector functions of innate immunity. IL-10 is an inhibitor of activated macrophages and dendritic cells and is an example of negative feedback regulation because it is produced by macrophages to inhibit their function. This cytokine also inhibits the production of IL-12 and expression of class II major histocompati‐ bility (MHC) molecules. IL-12 is also secreted by macrophages and dendritic cells inducing T cells differentiation into Th1 and natural killer (NK) cells with increased IFN-gamma synthesis and cytotoxic activity. IL-12 and IFN-gamma are the most important cytokines in innate responses to intracellular bacteria such as *Mycobacterium leprae* or tuberculosis [13]. Figure 3 shows the high levels of IL-12 and IFN-gamma induced by the novel cationic adjuvants while presenting the hsp-18 kDa of *M. leprae* to lymphonode cells suggesting a possible application of the novel adjuvants for the design of subunit vaccines against intracelular bacteria. As in DH, adaptive immunity against intracellular bacteria is principally cell mediated and consists of activation of macrophages by CD4+T cells as well as killing of infected cells by CD8+cyto‐
On basis of IL-12 enhancement of IFN-gamma production and development of Th1 cells, this interleukin itself has been used as a vaccine adjuvant for many infections that are combated by cell-mediated immunity, e.g. leishmaniasis [47]. Subunit vaccines against protozoa that survive within macrophages require as principal defense mechanism cell-mediated immunity, particularly directed to macrophage activation by Th1 cell-derived cytokines. Leishmaniasis mucocutaneous and disseminated is caused by *Leishmania donovani* and CD4+Th1 cells are required to activate macrophages to kill phagocytosed parasites. Resistance to the infection is associated with activation of Leishmania-specific Th1 CD4+T cells which produce IFN-gamma and thereby activate macrophages to destroy intracellular parasites. Conversely, activation of Th2 cells by the protozoan results in increased parasite survival and exarcerbation of lesions because of the macrophage-suppressive actions of Th2 cytokines [47]. Other significant example is the protective role played by CD8+T cells in immunity to the hepatic stages of malaria. These effects may be mediated by direct killing of sporozoite-infected hepatocytes or indirectly by the secretion of IFN-gamma and activation of hepatocytes to produce nitric oxide and other agents that kill parasites. IL-12 induces resistance to sporozoite challenge in rodents
and nonhuman primates, presumably by stimulating IFN-gamma production [48].
effects in mice were completely absent at 0.1-0.01mM DODAB.
toxic T lymphocytes (CTL).
8 Immune Response Activation
Another possible application for the novel adjuvants might be in immunotherapy for tumors. This approach is based in augmentation of host immunity to tumors with tumor vaccines. Immune responses that are able of killing tumor cells consist of CTLs, NK cells, and activated macrophages and these may be actively enhanced by vaccination with tumor cells or antigens, administration of tumors modified to express high levels of cytokines that stimulate T cell proliferation and differentiation, and systemic administration of cytokines [49]. The induction of T cell responses in tumors depends on processing and presentation of tumor antigens to T cells by professional antigen-presenting cells (APCs) which might internalize the tumor antigen adsorbed onto the novel cationic adjuvants. These APCs may stimulate CD8+T cells and CD4+helper T lymphocytes to differentiate for recognition and killing of tumor cells.
Naïve CD4+T cells may differentiate into distinct subsets, such as Th1 and Th2 cells in response to different antigens. For example, the enhancement in production of IL-10 and Il-13 by lymphonode cells elicited by the antigens of *Taenia crassiceps* presented by the DODAB BF adjuvant can be appreciated in Figure 3 [27]. These cytokines are typically associated with responses to allergens and parasites such as helminths and mediate differentiation of CD4+-T cells into Th2 cells [50]. Consistently, low levels of these cytokines were elicited by the *M. leprae* antigen presented by the novel adjuvants (Figure 3). Responses were indeed different for the helminthes and the bacteria antigens and very antigen-specific as they should be [27].
The size, charge and hydrophobic features of DODAB BF led to novel applications in solubi‐ lization of hydrophobic drugs [25,51,52], production of biomimetic particles from bilayer coverage of silica [17] or polystyrene particles [53] and design of vaccines [27].Recently, BF was also combined with oligonucleotides [54]. Since synthetic oligonucleotides can inhibit the replication of the Rous sarcoma virus [55], antisense oligonucleotides have been considered a great promise as therapeutic agents and several oligonucleotide-based formulations have reached the clinical trial phase [56,57]. Antisense oligonucleotides have also been extensively used in research on gene expression and function [58-60], vaccine formulation [61], allergy [62] and cancer therapeutics [63].Major obstacles as their degradation by nucleases and poor delivery to the target cells [60,64] suggest the essential role of suitable carriers able to protect oligonucleotides in the biological milieu [60,63–65].There are peculiar features for the interac‐ tion between BF and oligonucleotides in comparison to other electrolytes.Effects of salt, dAMP or poly (dA) concentration on BF size and zeta-potential are shown on Figure 4 taken from reference [54]. From 0 to 0.25 mM salt, Dz and zeta-potentials decreased with salt concentration possibly due to massive phosphate anion binding. From 0.25 to 2.5 mM of divalent salt, Dz increased but zeta-potential remained approximately constant and low (Fig. 4A and D).Dz and zeta-potential decreased with dAMP concentration (0– 2.5 mM) (Fig. 4B and E). At 0.05 mM poly (dA) and 0.5 mM DODAB, extensive BF aggregation and/or fusion took place as depicted from large Dz (N 500 nm) (Fig. 4C) and zero of zeta-potential (Fig. 4F). The screening of DODAB
charges by Na2HPO4, followed by a decrease in electrostatic repulsion between fragments, could be responsible for DODAB BFs aggregation and/or fusion. DLS data also showed that diameters increase upon addition of Na2HPO4 concentrations above 0.5mM (Figure 4A), and this diameter increase up to 400 nm is related to a decrease in the zeta-potential of the fragments (Figure 4D). The charge screening of DODAB charged heads by Na2HPO4 explains the decrease in zeta-potential (Figure 4D) as well as the tighter bilayer packing represented by higher mean phase transition temperature for the bilayer [54]. Addition of dAMP leads to a decrease in diameter and zeta-potential of the assemblies (Figures 4B and E). It was shown that DODAB bilayer fragments are able to order dAMP molecules on their surface, causing the dAMP bases to stack [68]. In this case, bulky moieties of dAMP would be exposed at the fragments surface representing a steric hindrance to fragments aggregation and/or fusion which would also contribute to colloid stabilization in dispersion. For polynucleotides such as poly(dA), charge neutralization leads to flocculation whereas charge overcompensation upon increasing poly(dA) concentration leads to colloidal restabilization due to electrostatic repulsion (Figures 4C and F) [54]. This behavior has been often described for polyelectrolytes interacting with particles of opposite charge [67,68]. Beyond the neutralization point, the system regains stability due to charge overcompensation. This phenomenon was observed only for the electrolyte poly (dA) and not for the electrolytes Na2HPO4 and dAMP, which are unable to completely neutralize the bilayer [54], as shown in Figure 4. Colloid instability induced by oligonucleotide or salt could be associated with bilayer fusion [54]. In contrast, mononucleo‐ tide neither reduced colloid stability over the low range of concentrations tested nor caused
0.00 0.05 0.10 0.15 0.20 0.25
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 11
0.0 0.5 1.0 1.5 2.0 2.5
**Figure 4.** Effect of [Na2HPO4], [dAMP] or [poly (dA)] concentrations on the zeta-average diameter (A, B and C) and zeta-potential of DODAB BF at 0.5 mM DODAB (D, E and F).Reprinted from reference [54].Reprinted from Biochimica et Biophysica Acta (BBA)-Biomembranes, 1808/3, Julio H.K. Rozenfeld,Tiago R. Oliveira,M. Teresa Lamy,Ana M. Carmo‐ na-Ribeiro, Interaction of cationic bilayer fragments with a model oligonucleotide, 649-655.Copyright 2011, with per‐
[dAMP] (mM) [poly(dA)] (mM) [Na2
30
0.0 0.5 1.0 1.5 2.0 2.5
HPO4 ] (mM) 40
50
70
80
90
200
D E F
400
600
A B C
BF fusion [54].
mission from Elsevier.
z(mV)
100
200
Dz (nm) 300
400
**Figure 3.** Quantitative analysis of cytokines secreted by lymph node cells from DODAB/18/14-Tcra and Al(OH)3/18/14-Tcra immunized mice (on the left column, A-D) or from 18kDa-hsp, DODAB/18kDa-hsp, PSS/DODAB/ 18kDa-hsp, SiO2/DODAB/18 kDa-hsp and Al(OH)3/18kDa-hsp immunized mice (on the right column, E-F). Cells from lymph nodes of BALB/c mice previously immunized with 10 μg of 18/14-*Tcra* administered alone, in DODAB BF or in Al(OH)3 were *in vitro* stimulated with medium, 160 μg/ml of 18/14-*Tcra* or 2.5 µg/mL of ConA for 48 hours and the supernatants collected for cytokine analysis by sandwich kit enzyme-linked immunosorbent assay (ELISA). The results were expressed as mean of the cytokine concentration of two distinct assays ± standard deviation. Limits of detection are shown as horizontal dashed lines. Similarly, cells from lymph nodes of BALB-c mice previously immunized with 15 μg of 18 kDa-hsp from *M. leprae* administered alone or in DODAB BF, PSS/DODAB, silica/DODAB or Al (OH)3 were *in vitro* stimulated with medium, 250 μg/ml of 18 kDa-hsp or 2.5 µg/mL of ConA for 48 hours before following cytokines analysis as above (reproduced with permission from reference [27]). Reprinted from Vaccine, 27/42, Nilton Lincopan, Noelí M. Espíndola, Adelaide J. Vaz,Maria Helena B. da Costa, Eliana Faquim-Mauro, Ana M. Carmona-Ribeiro, Novel immunoadjuvants based on cationic lipid: Preparation, characterization and activity *in vivo*, 5760-5771.Copyright 2009, with permission from Elsevier.
charges by Na2HPO4, followed by a decrease in electrostatic repulsion between fragments, could be responsible for DODAB BFs aggregation and/or fusion. DLS data also showed that diameters increase upon addition of Na2HPO4 concentrations above 0.5mM (Figure 4A), and this diameter increase up to 400 nm is related to a decrease in the zeta-potential of the fragments (Figure 4D). The charge screening of DODAB charged heads by Na2HPO4 explains the decrease in zeta-potential (Figure 4D) as well as the tighter bilayer packing represented by higher mean phase transition temperature for the bilayer [54]. Addition of dAMP leads to a decrease in diameter and zeta-potential of the assemblies (Figures 4B and E). It was shown that DODAB bilayer fragments are able to order dAMP molecules on their surface, causing the dAMP bases to stack [68]. In this case, bulky moieties of dAMP would be exposed at the fragments surface representing a steric hindrance to fragments aggregation and/or fusion which would also contribute to colloid stabilization in dispersion. For polynucleotides such as poly(dA), charge neutralization leads to flocculation whereas charge overcompensation upon increasing poly(dA) concentration leads to colloidal restabilization due to electrostatic repulsion (Figures 4C and F) [54]. This behavior has been often described for polyelectrolytes interacting with particles of opposite charge [67,68]. Beyond the neutralization point, the system regains stability due to charge overcompensation. This phenomenon was observed only for the electrolyte poly (dA) and not for the electrolytes Na2HPO4 and dAMP, which are unable to completely neutralize the bilayer [54], as shown in Figure 4. Colloid instability induced by oligonucleotide or salt could be associated with bilayer fusion [54]. In contrast, mononucleo‐ tide neither reduced colloid stability over the low range of concentrations tested nor caused BF fusion [54].
**Figure 3.** Quantitative analysis of cytokines secreted by lymph node cells from DODAB/18/14-Tcra and Al(OH)3/18/14-Tcra immunized mice (on the left column, A-D) or from 18kDa-hsp, DODAB/18kDa-hsp, PSS/DODAB/ 18kDa-hsp, SiO2/DODAB/18 kDa-hsp and Al(OH)3/18kDa-hsp immunized mice (on the right column, E-F). Cells from lymph nodes of BALB/c mice previously immunized with 10 μg of 18/14-*Tcra* administered alone, in DODAB BF or in Al(OH)3 were *in vitro* stimulated with medium, 160 μg/ml of 18/14-*Tcra* or 2.5 µg/mL of ConA for 48 hours and the supernatants collected for cytokine analysis by sandwich kit enzyme-linked immunosorbent assay (ELISA). The results were expressed as mean of the cytokine concentration of two distinct assays ± standard deviation. Limits of detection are shown as horizontal dashed lines. Similarly, cells from lymph nodes of BALB-c mice previously immunized with 15 μg of 18 kDa-hsp from *M. leprae* administered alone or in DODAB BF, PSS/DODAB, silica/DODAB or Al (OH)3 were *in vitro* stimulated with medium, 250 μg/ml of 18 kDa-hsp or 2.5 µg/mL of ConA for 48 hours before following cytokines analysis as above (reproduced with permission from reference [27]). Reprinted from Vaccine, 27/42, Nilton Lincopan, Noelí M. Espíndola, Adelaide J. Vaz,Maria Helena B. da Costa, Eliana Faquim-Mauro, Ana M. Carmona-Ribeiro, Novel immunoadjuvants based on cationic lipid: Preparation, characterization and activity *in vivo*, 5760-5771.Copyright
2009, with permission from Elsevier.
10 Immune Response Activation
**Figure 4.** Effect of [Na2HPO4], [dAMP] or [poly (dA)] concentrations on the zeta-average diameter (A, B and C) and zeta-potential of DODAB BF at 0.5 mM DODAB (D, E and F).Reprinted from reference [54].Reprinted from Biochimica et Biophysica Acta (BBA)-Biomembranes, 1808/3, Julio H.K. Rozenfeld,Tiago R. Oliveira,M. Teresa Lamy,Ana M. Carmo‐ na-Ribeiro, Interaction of cationic bilayer fragments with a model oligonucleotide, 649-655.Copyright 2011, with per‐ mission from Elsevier.
Particles are finding a large variety of biomedical and pharmaceutical applications since their size scale can be similar to that of biomacromolecules (e.g., proteins, DNA) and structures (e.g., bacteria and viruses). Their utility for imaging, gene and drug delivery, and vaccine design is undeniable [69,70].Particulate systems are naturally targeted to antigen presenting cells (APC) so that particles deliver antigens to APC more efficiently than soluble antigen [71,72].Positively charged particles with diameters of 500 nm and below were shown to be optimal for dendritic cells uptake [42].DODAB bilayers electrostatically combine with a vast variety of negatively charged biomolecules or biological structures [14]. Silica [17], latex [21,73,74] or hydrophobic drug particles [51,75] have been coated with DODAB with optimal bilayer deposition on particles achieved by coalescence of bilayer fragments at an adequate ionic strength [17,76]. Figure 5 shows how DODAB can cover oppositely charged polystyrene nanoparticles modi‐ fying their charge as shown in reference [19]. These cationic nanoparticles contrast with alum regarding their small size and very low polydispersity as shown in Table 1 taken from reference [19]. The optimal bilayer coverage of polystyrene sulfate (PSS) nanoparticles with a DODAB bilayer produces homodisperse particles that successfully present a mixture of purified 18/14 *Taenia crassiceps* proteins (18/14-Tcra) to the immunological system [19]. presenting cells (APC) so that particles deliver antigens to APC more efficiently than soluble antigen [71,72].Positively charged particles with diameters of 500 nm and below were shown to be optimal for dendritic cells uptake [42].DODAB bilayers electrostatically combine with a vast variety of negatively charged biomolecules or biological structures [14]. Silica [17], latex [21,73,74] or hydrophobic drug particles [51,75] have been coated with DODAB with optimal bilayer deposition on particles achieved by coalescence of bilayer fragments at an adequate ionic strength [17,76]. Figure 5 shows how DODAB can cover oppositely charged polystyrene nanoparticles modifying their charge as shown in reference [19]. These cationic nanoparticles contrast with alum regarding their small size and very low polydispersity as shown in Table 1 taken from reference [19]. The optimal bilayer coverage of polystyrene sulfate (PSS) nanoparticles with a DODAB bilayer produces homodisperse particles that successfully present a mixture of purified 18/14 *Taenia crassiceps* proteins (18/14-Tcra) to the immunological system [19].
log [DODAB]
Mean diameter (nm)
Figure 5.Effect of [DODAB] (in mM) on mean z-average diameter (A) and zeta-potential (B) of PSS particles at 5 x 109 particles/mL, 25 C, in 1 mM NaCl. Bare particle diameter is 301 ± 2 nm. Regions I, II and III define particle charge, which is negative, zero and positive, respectively, from reference [19]. Reprinted from International Journal of Pharmaceutics, 340 / 1–2, N. Lincopan, N.M. Espíndola, A.J. Vaz, A.M. Carmona-Ribeiro, Cationic supported lipid bilayers for antigen presentation, 216-222.Copyright 2007, with permission from Elsevier. **Figure 5.** Effect of [DODAB] (in mM) on mean z-average diameter (A) and zeta-potential (B) of PSS particles at 5 x 109 particles/mL, 25 °C, in 1 mM NaCl. Bare particle diameter is 301 ± 2 nm. Regions I, II and III define particle charge, which is negative, zero and positive, respectively, from reference [19]. Reprinted from International Journal of Pharma‐ ceutics, 340 / 1–2, N. Lincopan, N.M. Espíndola, A.J. Vaz, A.M. Carmona-Ribeiro, Cationic supported lipid bilayers for antigen presentation, 216-222.Copyright 2007, with permission from Elsevier.
PSS - - 301 ± 2 -60 ± 1 0.064 ± 0.020 DODAB 2.00 - 81 ± 1 45 ± 2 0.230 ± 0.006 PSS/DODAB 0.01 - 309 ± 2 48 ± 2 0.040 ± 0.010 18/14-*Tcra* - 25 310 ± 5 -52 ± 1 0.214 ± 0.030 DODAB/18/14-*Tcra* 0.01 25 295 ± 3 6 ± 6 0.167 ± 0.023 PSS/DODAB/18/14-*Tcra* 0.01 25 328 ± 3 11 ± 8 0.060 ± 0.020 Al(OH)3 - - 883 ± 29 28 ± 3 0.381 ± 0.013 Al(OH)3/18/14-*Tcra* - 25 9574 ± 2361 -23 ± 1 0.525 ± 0.030
[Ag] (μg/mL)
Table 1.Physical properties of particles, DODAB dispersion, DODAB-covered particles, proteins and proteins/DODAB-
electron microscopy is 301±2 nm.The Al(OH)3 and PSS were tested at final concentration of 0.05 and 0.075 mg/mL, respectively,from reference [19]. Reprinted from International Journal of Pharmaceutics, 340 / 1–2, N. Lincopan, N.M. Espíndola, A.J. Vaz, A.M. Carmona-Ribeiro, Cationic supported lipid bilayers for antigen presentation, 216-222.
particles /mL; PSS particle diameter from transmission
Zeta-Potential
(mV) Polydispersity
**Dispersion [DODAB]**
Copyright 2007, with permission from Elsevier.
**(mM)**
**[Ag] (μg/mL)** **Mean diameter**
**Zeta-Potential**
**(mV) Polydispersity**
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 13
[80]. System‐
**(nm)**
**Table 1.** Physical properties of particles, DODAB dispersion, DODAB-covered particles, proteins and proteins/DODABcovered particles at 1 mM NaCl.Particles concentration is 5 x 109 particles /mL; PSS particle diameter from transmission electron microscopy is 301±2 nm.The Al(OH)3 and PSS were tested at final concentration of 0.05 and 0.075 mg/mL, respectively,from reference [19]. Reprinted from International Journal of Pharmaceutics, 340 / 1–2, N. Lincopan, N.M. Espíndola, A.J. Vaz, A.M. Carmona-Ribeiro, Cationic supported lipid bilayers for antigen presentation, 216-222.
Several cationic agents have been employed for DNA compaction such as cationic peptides and proteins [77], cationic lipids [78], cationic polyelectrolytes, cationic surfactants, or iron (III) [79].DNA compaction has also been used to model chromatin structure and its influence on gene expression. The self-assembled complex of basic histone proteins wrapped by approxi‐ mately two turns of DNA is a nucleosome, which is the building block in the chromatin structure where DNA of lengths on the order of meters suffers compaction into an∼10 μm diameter cell nucleous. Phage DNA, for example, is remarkable for its density of packing. In solution the 40 kbp T7 genome with its contour length of 13.6 μm might span a space several micrometers across and in an infected bacterium, ∼1 μm across. Thus, confinement to a 55 nm
atic studies on the physical chemistry of the association between cationic nanoparticles and DNA yield rather complex phase diagrams as a function of particle size and concentration [81,82].The way in which positively charged nanoparticles tie up DNA is not obvious, and mechanisms change dramatically with particle size [83]. Only the smallest (10 nm) particles allowed transcription to occur at intermediate loading densities. Larger particles shut tran‐ scription down rather abruptly [81]. Cationic nanoparticles have found many uses such as efficient cell transfection agents *in vitro* [84-86] and complexation with long-chained DNA as a simple model of chromatin for transcription studies [81, 87].The compaction of long duplex DNA by cationic nanoparticles (NP) used as a primary model of histone core particles has been systematically studied regarding the effect of salt concentration, particle size, and particle charge by means of single-molecule observations from fluorescence and transmission electron microscopy [87]. DNA compaction proceeds through the formation of beads-on-a-string structures of various morphologies with DNA adsorbed amount per particle depending weakly on NP concentration but increasing with particle size and being optimal at an inter‐
capsid represents a compaction marked by a density increase by a factor of ∼10<sup>4</sup>
PSS - - 301 ± 2 -60 ± 1 0.064 ± 0.020 DODAB 2.00 - 81 ± 1 45 ± 2 0.230 ± 0.006 PSS/DODAB 0.01 - 309 ± 2 48 ± 2 0.040 ± 0.010 18/14-*Tcra* - 25 310 ± 5 -52 ± 1 0.214 ± 0.030 DODAB/18/14-*Tcra* 0.01 25 295 ± 3 6 ± 6 0.167 ± 0.023 PSS/DODAB/18/14-*Tcra* 0.01 25 328 ± 3 11 ± 8 0.060 ± 0.020 Al(OH)3 - - 883 ± 29 28 ± 3 0.381 ± 0.013 Al(OH)3/18/14-*Tcra* - 25 9574 ± 2361 -23 ± 1 0.525 ± 0.030
covered particles at 1 mM NaCl.Particles concentration is 5 x 10<sup>9</sup>
(mM)
Copyright 2007, with permission from Elsevier.
Dispersion [DODAB]
Particles are finding a large variety of biomedical and pharmaceutical applications since their size scale can be similar to that of biomacromolecules (e.g., proteins, DNA) and structures (e.g., bacteria and viruses). Their utility for imaging, gene and drug delivery, and vaccine design is undeniable [69,70].Particulate systems are naturally targeted to antigen presenting cells (APC) so that particles deliver antigens to APC more efficiently than soluble antigen [71,72].Positively charged particles with diameters of 500 nm and below were shown to be optimal for dendritic cells uptake [42].DODAB bilayers electrostatically combine with a vast variety of negatively charged biomolecules or biological structures [14]. Silica [17], latex [21,73,74] or hydrophobic drug particles [51,75] have been coated with DODAB with optimal bilayer deposition on particles achieved by coalescence of bilayer fragments at an adequate ionic strength [17,76]. Figure 5 shows how DODAB can cover oppositely charged polystyrene nanoparticles modi‐ fying their charge as shown in reference [19]. These cationic nanoparticles contrast with alum regarding their small size and very low polydispersity as shown in Table 1 taken from reference [19]. The optimal bilayer coverage of polystyrene sulfate (PSS) nanoparticles with a DODAB bilayer produces homodisperse particles that successfully present a mixture of purified 18/14
presenting cells (APC) so that particles deliver antigens to APC more efficiently than soluble antigen [71,72].Positively charged particles with diameters of 500 nm and below were shown to be optimal for dendritic cells uptake [42].DODAB bilayers electrostatically combine with a vast variety of negatively charged biomolecules or biological structures [14]. Silica [17], latex [21,73,74] or hydrophobic drug particles [51,75] have been coated with DODAB with optimal bilayer deposition on particles achieved by coalescence of bilayer fragments at an adequate ionic strength [17,76]. Figure 5 shows how DODAB can cover oppositely charged polystyrene nanoparticles modifying their charge as shown in reference [19]. These cationic nanoparticles contrast with alum regarding their small size and very low polydispersity as shown in Table 1 taken from reference [19]. The optimal bilayer coverage of polystyrene sulfate (PSS) nanoparticles with a DODAB bilayer produces homodisperse particles that successfully present a mixture of purified 18/14 *Taenia*
I II III
**+** + **+ +** + **+** + + + **+ + +**
Figure 5.Effect of [DODAB] (in mM) on mean z-average diameter (A) and zeta-potential (B) of PSS particles at 5 x 109 particles/mL, 25 C, in 1 mM NaCl. Bare particle diameter is 301 ± 2 nm. Regions I, II and III define particle charge, which is negative, zero and positive, respectively, from reference [19]. Reprinted from International Journal of Pharmaceutics, 340 / 1–2, N. Lincopan, N.M. Espíndola, A.J. Vaz, A.M. Carmona-Ribeiro, Cationic supported lipid
**Figure 5.** Effect of [DODAB] (in mM) on mean z-average diameter (A) and zeta-potential (B) of PSS particles at 5 x 109 particles/mL, 25 °C, in 1 mM NaCl. Bare particle diameter is 301 ± 2 nm. Regions I, II and III define particle charge, which is negative, zero and positive, respectively, from reference [19]. Reprinted from International Journal of Pharma‐ ceutics, 340 / 1–2, N. Lincopan, N.M. Espíndola, A.J. Vaz, A.M. Carmona-Ribeiro, Cationic supported lipid bilayers for
log (DDA mM)
log [DODAB]
PSS - - 301 ± 2 -60 ± 1 0.064 ± 0.020 DODAB 2.00 - 81 ± 1 45 ± 2 0.230 ± 0.006 PSS/DODAB 0.01 - 309 ± 2 48 ± 2 0.040 ± 0.010 18/14-*Tcra* - 25 310 ± 5 -52 ± 1 0.214 ± 0.030 DODAB/18/14-*Tcra* 0.01 25 295 ± 3 6 ± 6 0.167 ± 0.023 PSS/DODAB/18/14-*Tcra* 0.01 25 328 ± 3 11 ± 8 0.060 ± 0.020 Al(OH)3 - - 883 ± 29 28 ± 3 0.381 ± 0.013 Al(OH)3/18/14-*Tcra* - 25 9574 ± 2361 -23 ± 1 0.525 ± 0.030
Mean diameter (nm)
Table 1.Physical properties of particles, DODAB dispersion, DODAB-covered particles, proteins and proteins/DODAB-
electron microscopy is 301±2 nm.The Al(OH)3 and PSS were tested at final concentration of 0.05 and 0.075 mg/mL, respectively,from reference [19]. Reprinted from International Journal of Pharmaceutics, 340 / 1–2, N. Lincopan, N.M. Espíndola, A.J. Vaz, A.M. Carmona-Ribeiro, Cationic supported lipid bilayers for antigen presentation, 216-222.
particles /mL; PSS particle diameter from transmission
Zeta-Potential
(mV) Polydispersity
bilayers for antigen presentation, 216-222.Copyright 2007, with permission from Elsevier.
antigen presentation, 216-222.Copyright 2007, with permission from Elsevier.
0
Zeta Potential (mV)
40
300
350
1000
Mean Diameter (nm)
1100
1200
B
[Ag] (μg/mL)
(mM)
covered particles at 1 mM NaCl.Particles concentration is 5 x 10<sup>9</sup>
Copyright 2007, with permission from Elsevier.
Dispersion [DODAB]
*Taenia crassiceps* proteins (18/14-Tcra) to the immunological system [19].
A
*crassiceps* proteins (18/14-Tcra) to the immunological system [19].
12 Immune Response Activation
**Table 1.** Physical properties of particles, DODAB dispersion, DODAB-covered particles, proteins and proteins/DODABcovered particles at 1 mM NaCl.Particles concentration is 5 x 109 particles /mL; PSS particle diameter from transmission electron microscopy is 301±2 nm.The Al(OH)3 and PSS were tested at final concentration of 0.05 and 0.075 mg/mL, respectively,from reference [19]. Reprinted from International Journal of Pharmaceutics, 340 / 1–2, N. Lincopan, N.M. Espíndola, A.J. Vaz, A.M. Carmona-Ribeiro, Cationic supported lipid bilayers for antigen presentation, 216-222. Copyright 2007, with permission from Elsevier.
Several cationic agents have been employed for DNA compaction such as cationic peptides and proteins [77], cationic lipids [78], cationic polyelectrolytes, cationic surfactants, or iron (III) [79].DNA compaction has also been used to model chromatin structure and its influence on gene expression. The self-assembled complex of basic histone proteins wrapped by approxi‐ mately two turns of DNA is a nucleosome, which is the building block in the chromatin structure where DNA of lengths on the order of meters suffers compaction into an∼10 μm diameter cell nucleous. Phage DNA, for example, is remarkable for its density of packing. In solution the 40 kbp T7 genome with its contour length of 13.6 μm might span a space several micrometers across and in an infected bacterium, ∼1 μm across. Thus, confinement to a 55 nm capsid represents a compaction marked by a density increase by a factor of ∼10<sup>4</sup> [80]. System‐ atic studies on the physical chemistry of the association between cationic nanoparticles and DNA yield rather complex phase diagrams as a function of particle size and concentration [81,82].The way in which positively charged nanoparticles tie up DNA is not obvious, and mechanisms change dramatically with particle size [83]. Only the smallest (10 nm) particles allowed transcription to occur at intermediate loading densities. Larger particles shut tran‐ scription down rather abruptly [81]. Cationic nanoparticles have found many uses such as efficient cell transfection agents *in vitro* [84-86] and complexation with long-chained DNA as a simple model of chromatin for transcription studies [81, 87].The compaction of long duplex DNA by cationic nanoparticles (NP) used as a primary model of histone core particles has been systematically studied regarding the effect of salt concentration, particle size, and particle charge by means of single-molecule observations from fluorescence and transmission electron microscopy [87]. DNA compaction proceeds through the formation of beads-on-a-string structures of various morphologies with DNA adsorbed amount per particle depending weakly on NP concentration but increasing with particle size and being optimal at an inter‐ mediate salt concentration [87]. Three different complexation mechanisms were proposed: free DNA adsorption onto NP surface, DNA wrapping around NP, and NP collection on DNA chain [87]. On the other hand, particle size has been recognized as an important parameter that determines the mechanism of particle entry into cells. Particles with a diameter of 200 nm or less enter cells almost exclusively via the clathrin-coated pathway whereas particles with a larger diameter penetrate cells via caveolae-mediated endocytosis [88,89].Cationic biomimetic particles produced from adsorption of dioctadecyldimethylammonium bromide (DODAB) bilayers onto polystyrene sulfate (PSS) microspheres have been described by our group since 1992 [18,73,76,90].These cationic bilayer-covered particles exhibit a narrow size distribution and can be produced at any desired size ranging from 70-500 nm of mean hydrodynamic diameter [90]. Polystyrene sulfate (PSS) particles with different sizes were covered by a dioctadecyldimethylammonium bromide (DODAB) bilayer yielding the so-called cationic biomimetic particles (PSS/DODAB). These cationic particles are highly organized, present a narrow size distribution and were obtained over a range of particle sizes [53,90].Thereafter, upon adding λ, T5 or T2-DNA to PSS/DODAB particles, supramolecular assemblies PSS/ DODAB/DNA were obtained and characterized over a range of DNA concentrations and particle sizes (80-700 nm). Over the low DNA concentration range, PSS/DODAB/DNA assemblies were cationic, colloidally stable with moderate polydispersity and high cytotoxicity against *E. coli*. From the DNA concentration corresponding to charge neutralization, neutral or anionic supramolecular assemblies PSS/DODAB/DNA exhibited low colloid stability, high polydispersity and moderate cytotoxicity [53]. Some nucleosome mimetic assemblies were observed by atomic force microscopy (AFM) at charge neutralization (zeta-potential equal to zero) [55].Figure 6 shows how cationic nanoparticles can induce DNA compaction [53].
DNA sequences containing unmethylated CpG dinucleotide are recognized as danger signals by the immune system since they are typical of bacteria and viruses but rare in vertebrates [91, 92]. Natural or synthetic sequences containing unmethylated CpG motifs activate cells that express Toll-like receptor 9 to induce an innate immune response characterized by the production of Th1 and proinflammatory cytokines [92]. Hence, CpG has been extensively used in the induction of cellular immune responses against cancer [63, 93], intracellular infections by pathogens [94, 95] and allergies [62,96]. Both CpG [92] and DODAB BF [27] were reported to improve Th1 responses against antigens when used separately. Recently, DODAB BF and CpG were combined in a single assembly aiming at the comparison between the small, stable and cationic DODAB BF carrying ovalbumin (OVA) and the small, stable and anionic DODAB BF/OVA/CpG assemblies of very similar sizes but opposite charges [97]. Both adjuvants produced similar enhanced Th1 immune responses despite their opposite charges emphasiz‐ ing the novel concept that particle charge does not matter. In comparison with the traditional alum, the size minimization for the cationic assemblies elicited different responses: alum drove the Th2 whereas DODAB BF/OVA/CpG and DODAB BF/OVA drove the Th1 response. The effects of DODAB BF or DODAB BF/CpG adjuvants with opposite charges but very similar sizes showed that the charge is not important but the size is [97]. Table 2 shows the physical properties of the assemblies. A comparison between mean hydrodynamic diameter (Dz), zetapotentials and polydispersities for DODAB BF/OVA/CpG and Al(OH)3/OVA/CpG as repro‐ duced from reference [97]. At 20 μM CpG and 0.1 mg/mL Al(OH)3, the dispersion was characterized by a low colloidal stability (large Dz, low zeta-potential (ζ) and high polydis‐ persity). OVA stabilized both Al(OH)3 or Al(OH)3/CpG to a certain extent (Table 2). However, sizes were still larger than those determined for assemblies based on DODAB BF (Table 2). In particular, polydispersity for DODAB BF/OVA/CpG was notably low (0.150) and well below the one obtained for DODAB BF only (0.251), suggesting that OVA/CpG induced stabilization
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 15
**Dispersion Dz** ± **δ (nm) ζ** ± **δ (mV) Polydispersity** ± **δ** DODAB BF 67 ± 0 47 ± 1 0.251 ± 0.006 DODAB BF/ OVA 274 ± 2 21 ± 0 0.291 ± 0.008 DODAB BF/ OVA/ CpG 245 ± 1 -26 ± 1 0.150 ± 0.020
Al(OH)3 3147 ± 197 16 ± 2 0.415 ± 0.018 Al(OH)3/ OVA 916 ± 17 -29 ± 1 0.231 ± 0.020 Al(OH)3/ CpG 3584 ± 74 9 ± 1 0.407 ± 0.026 Al(OH)3/ OVA/ CpG 570 ± 19 -35 ± 2 0.210 ± 0.020
**Table 2.** Physical properties of alum or DODAB BF dispersions combined with ovalbumin (OVA) and/or CpG
oligonucleotide. Concentrations are 0.1mg/mL OVA, 0.1mM DODAB BF, 0.1mg/mL Al(OH)3 and 20μM CpG. Reprinted from reference [97]. Reprinted from Journal of Controlled Release, 160/2, Julio H.K. Rozenfeld, Sandriana R. Silva, Priscila A. Ranéia, Eliana Faquim-Mauro, Ana M. Carmona-Ribeiro, Stable assemblies of cationic bilayer fragments and CpG oligonucleotide with enhanced immunoadjuvant activity *in vivo*, 367-373.Copyright 2012, with permission from
of the DODAB BF dispersion (Table 2).
Elsevier.
**Figure 6.** DNA compaction by biomimetic cationic particles of polystyrene microspheres covered by a DODAB cationic bilayer from reference [53].Adapted with permission from Rosa H, Petri DF, Carmona-Ribeiro AM.Interactions be‐ tween bacteriophage DNA and cationic biomimetic particles. J Phys Chem B. 2008; 112(51):16422-30.Copyright (2008) American Chemical Society.
DNA sequences containing unmethylated CpG dinucleotide are recognized as danger signals by the immune system since they are typical of bacteria and viruses but rare in vertebrates [91, 92]. Natural or synthetic sequences containing unmethylated CpG motifs activate cells that express Toll-like receptor 9 to induce an innate immune response characterized by the production of Th1 and proinflammatory cytokines [92]. Hence, CpG has been extensively used in the induction of cellular immune responses against cancer [63, 93], intracellular infections by pathogens [94, 95] and allergies [62,96]. Both CpG [92] and DODAB BF [27] were reported to improve Th1 responses against antigens when used separately. Recently, DODAB BF and CpG were combined in a single assembly aiming at the comparison between the small, stable and cationic DODAB BF carrying ovalbumin (OVA) and the small, stable and anionic DODAB BF/OVA/CpG assemblies of very similar sizes but opposite charges [97]. Both adjuvants produced similar enhanced Th1 immune responses despite their opposite charges emphasiz‐ ing the novel concept that particle charge does not matter. In comparison with the traditional alum, the size minimization for the cationic assemblies elicited different responses: alum drove the Th2 whereas DODAB BF/OVA/CpG and DODAB BF/OVA drove the Th1 response. The effects of DODAB BF or DODAB BF/CpG adjuvants with opposite charges but very similar sizes showed that the charge is not important but the size is [97]. Table 2 shows the physical properties of the assemblies. A comparison between mean hydrodynamic diameter (Dz), zetapotentials and polydispersities for DODAB BF/OVA/CpG and Al(OH)3/OVA/CpG as repro‐ duced from reference [97]. At 20 μM CpG and 0.1 mg/mL Al(OH)3, the dispersion was characterized by a low colloidal stability (large Dz, low zeta-potential (ζ) and high polydis‐ persity). OVA stabilized both Al(OH)3 or Al(OH)3/CpG to a certain extent (Table 2). However, sizes were still larger than those determined for assemblies based on DODAB BF (Table 2). In particular, polydispersity for DODAB BF/OVA/CpG was notably low (0.150) and well below the one obtained for DODAB BF only (0.251), suggesting that OVA/CpG induced stabilization of the DODAB BF dispersion (Table 2).
mediate salt concentration [87]. Three different complexation mechanisms were proposed: free DNA adsorption onto NP surface, DNA wrapping around NP, and NP collection on DNA chain [87]. On the other hand, particle size has been recognized as an important parameter that determines the mechanism of particle entry into cells. Particles with a diameter of 200 nm or less enter cells almost exclusively via the clathrin-coated pathway whereas particles with a larger diameter penetrate cells via caveolae-mediated endocytosis [88,89].Cationic biomimetic particles produced from adsorption of dioctadecyldimethylammonium bromide (DODAB) bilayers onto polystyrene sulfate (PSS) microspheres have been described by our group since 1992 [18,73,76,90].These cationic bilayer-covered particles exhibit a narrow size distribution and can be produced at any desired size ranging from 70-500 nm of mean hydrodynamic diameter [90]. Polystyrene sulfate (PSS) particles with different sizes were covered by a dioctadecyldimethylammonium bromide (DODAB) bilayer yielding the so-called cationic biomimetic particles (PSS/DODAB). These cationic particles are highly organized, present a narrow size distribution and were obtained over a range of particle sizes [53,90].Thereafter, upon adding λ, T5 or T2-DNA to PSS/DODAB particles, supramolecular assemblies PSS/ DODAB/DNA were obtained and characterized over a range of DNA concentrations and particle sizes (80-700 nm). Over the low DNA concentration range, PSS/DODAB/DNA assemblies were cationic, colloidally stable with moderate polydispersity and high cytotoxicity against *E. coli*. From the DNA concentration corresponding to charge neutralization, neutral or anionic supramolecular assemblies PSS/DODAB/DNA exhibited low colloid stability, high polydispersity and moderate cytotoxicity [53]. Some nucleosome mimetic assemblies were observed by atomic force microscopy (AFM) at charge neutralization (zeta-potential equal to zero) [55].Figure 6 shows how cationic nanoparticles can induce DNA compaction [53].
**Figure 6.** DNA compaction by biomimetic cationic particles of polystyrene microspheres covered by a DODAB cationic bilayer from reference [53].Adapted with permission from Rosa H, Petri DF, Carmona-Ribeiro AM.Interactions be‐ tween bacteriophage DNA and cationic biomimetic particles. J Phys Chem B. 2008; 112(51):16422-30.Copyright
(2008) American Chemical Society.
14 Immune Response Activation
**Table 2.** Physical properties of alum or DODAB BF dispersions combined with ovalbumin (OVA) and/or CpG oligonucleotide. Concentrations are 0.1mg/mL OVA, 0.1mM DODAB BF, 0.1mg/mL Al(OH)3 and 20μM CpG. Reprinted from reference [97]. Reprinted from Journal of Controlled Release, 160/2, Julio H.K. Rozenfeld, Sandriana R. Silva, Priscila A. Ranéia, Eliana Faquim-Mauro, Ana M. Carmona-Ribeiro, Stable assemblies of cationic bilayer fragments and CpG oligonucleotide with enhanced immunoadjuvant activity *in vivo*, 367-373.Copyright 2012, with permission from Elsevier.
At 0.1 mg/mL OVA, the dependence of DODAB BF/OVA size and zeta-potential on time and [DODAB] established 0.1 mM DODAB as suitable for obtaining stable and cationic DODAB BF/ OVA assemblies [97]. At 0.1 mM DODAB, 0.1 mg/mL OVA and 0.006 mM CpG, the zetapotential is zero showing charge neutralization [97]. At [CpG]> 0.006 mM, good colloidal stability for the anionic assemblies due to charge overcompensation was observed whereas at 0.020 mM CpG, these DODAB BF/OVA/CpG assemblies turned out to be highly effective *in vivo* generating responses similar to those elicited by the stable and cationic DODAB BF/OVA. The anti-OVA delayed-type hypersensitivity (DTH) reaction and the secretion of IFN-gamma and IL-12 resulted 6, 42 and 9 times larger for the DODAB BF/OVA/CpG-immunized mice than the same responses by OVA-immunized mice, respectively [97].Figure 7 A and B illustrate the colloidal stability of the assemblies over a range of DODAB concentrations.
the secretion of IFN-γ and IL-12 increased by 33 and 49 %, respectively when compared to DODAB BF/OVA-immunized mice group and 52% and 35% when compared to OVA/CpGgroup. In contrast, the highest secretion of IL-10 and IL-13 was observed in cultures of cells from mice that were immunized with Al(OH)3 /OVA whereas all other assemblies resulted in poor production of these cytokines [97]. Immune responses were similar for anionic DODAB BF/OVA/CpG and cationic DODAB BF/OVA of similar sizes showing that the charge is not important but the size is. The adsorption of antigen on the surface of DODAB large vesicles was shown to stimulate active antigen capture and presentation by dendritic cells (DCs) [41]. Administration of antigen adsorbed on DODAB large vesicles (LV) resulted in formation of an antigen depot at the site of injection which hampered the rapid clearance of antigen that takes place in absence of a carrier [100]. In contrast to DODAB LV, the small DODAB BF/ antigen assemblies did not result in any observable depot effect [27, 97]. Furthermore, since the depot was absent for DODAB BF/CpG/OVA the immunostimulatory effect must have occurred via direct effect of the assemblies on the lymphonode antigen presenting cells [97]. Only nanoparticles can specifically target lymph node-resident cells [101]. CpG combined with DODAB BF yielded improved cellular Th1 response [97] possibly due to the appropriate targeting of DODAB BF/CpG/antigen to DCs in charge of expressing endosomal toll like receptor 9 (TLR9) in the lymph nodes [102]. The enhanced Th1 response by anionic DODAB BF/OVA/CpG relative to OVA alone was evidenced by the 6 times increase in DTH reaction, the 42 times increase in IFN-γ secretion by lymph node cells in culture and the 9 times increase
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 17
in IL-12 secretion also by lymph node cells in culture (Figure 7) [97].
optimal physical properties for efficient antigen presentation [106].
deacetylated to yield chitosan as shown in Figure 8.
**polymers**
Other delivery systems based on anionic [103] or cationic lipid bilayer [41,44,100,104,105] have also been successful for improving Th1 response against important antigens such as those of influenza [103], hepatitis A and B [103,104], and fungal infections [105]. In general, cationic lipids are known for the production of a large inflammatory response [41, 44].However, for small cationic bilayer fragments as immunoadjuvants, this adverse reaction is absent [27, 97]. The small and anionic DODAB BF/OVA/CpG assemblies [97] also did not elicit adverse reactions similarly to other anionic assemblies [103]. There was no depot effect for these small assemblies [97] and their net negative charge ensured the absence of the adverse reactions observed previously for the large cationic liposomes and vesicles [100]. Recently, based on cross reactivity with *Neisseria lactamica* outer membrane vesicles (OMV) antigens, DODAB BF were combined with OMV to develop a vaccine against *Neisseria meningitidis* in young children [106]. Complexes of 25 μg of OMV in 0.1 mM of DODAB BF were colloidally stable, exhibiting
**3. Cationic nanostructures based on chitosan and other biocompatible**
Chitin is a long-chain polymer of N-acetylglucosamine, a derivative of glucose which can be
The delayed-type hypersensitivity reaction (DTH) is an important *in vivo* response mediated by cells that can be quantified from the footpad sweelling test [44]. Mice immunized with OVA alone or with DODAB BF/CpG or with OVA / Al(OH)3 exhibited a footpad swelling equal to the one observed for naive mice [97]. The largest increase in footpad swelling was observed for DODAB BF/OVA/CpG mice immunization, which was about 6 times larger than the one observed for naive mice and 1.3 times larger than the one observed for DODAB BF/OVA [97]. Figure 7 C shows the improved DTH response in mice induced by the assemblies as adapted from reference [97].
Since OVA isoelectric point is 4.5 [98], this protein is negatively charged at 6.3, the pH of water. Thus, OVA adsorption onto DODAB BF is initially electrostatically driven.The OVA titration with DODAB BF determined ranges of DODAB concentration for occurrence of stable DODAB BF/OVA assemblies as illustrated in Figure 7 A and B. Similar results had been previously described also for DODAB BF/ bovin serum albumin or DODAB BF/ purified antigens from *Taenia crassiceps* [27].When the net charge of the assemblies is zero (at charge neutralization), maximal aggregation was observed for the assemblies. Further increasing DODAB concen‐ tration, stabilized them to a certain extent (Figure 7 B). However, sizes and polydispersities were still higher than those of DODAB BF in absence of OVA. Optimal colloidal stability was only achieved upon CpG addition to the system yielding high and negative zeta-potentials plus remarkably small polydispersity [98]. Oligonucleotides with less than 20 nucleotide residues usually behave as rigid charged rods in solution [99]. CpG would also adsorb as rigid charged rods on vacant positive sites of DODAB BF/ OVA assemblies inducing charge overcompensation and recovery of colloidal stability as previously described for a model oligonucleotide above charge neutralization [56].
Figure 7 also illustrates the improvement in the cellular OVA-specific response from the analysis of cytokines secreted by lymph node cells of mice as adapted from reference [97]. IFNγ and IL-12 production is associated to the Th1 response whereas IL-10 and IL-13 production reflects the Th2 response. Levels of IFN-γ and IL-12 secretion observed in cell cultures of mice immunized with OVA or Al(OH)3/ OVA were low and close to the detection limit for these cytokine assays (Figure 7). For mice immunized with OVA/CpG or DODAB BF/OVA assem‐ blies, the secretion of IFN-γ and IL-12 substantially increased in comparison to secretion from cultured cells of OVA-immunized mice [97]. For the DODAB BF/OVA/CpG immunized mice, the secretion of IFN-γ and IL-12 increased by 33 and 49 %, respectively when compared to DODAB BF/OVA-immunized mice group and 52% and 35% when compared to OVA/CpGgroup. In contrast, the highest secretion of IL-10 and IL-13 was observed in cultures of cells from mice that were immunized with Al(OH)3 /OVA whereas all other assemblies resulted in poor production of these cytokines [97]. Immune responses were similar for anionic DODAB BF/OVA/CpG and cationic DODAB BF/OVA of similar sizes showing that the charge is not important but the size is. The adsorption of antigen on the surface of DODAB large vesicles was shown to stimulate active antigen capture and presentation by dendritic cells (DCs) [41]. Administration of antigen adsorbed on DODAB large vesicles (LV) resulted in formation of an antigen depot at the site of injection which hampered the rapid clearance of antigen that takes place in absence of a carrier [100]. In contrast to DODAB LV, the small DODAB BF/ antigen assemblies did not result in any observable depot effect [27, 97]. Furthermore, since the depot was absent for DODAB BF/CpG/OVA the immunostimulatory effect must have occurred via direct effect of the assemblies on the lymphonode antigen presenting cells [97]. Only nanoparticles can specifically target lymph node-resident cells [101]. CpG combined with DODAB BF yielded improved cellular Th1 response [97] possibly due to the appropriate targeting of DODAB BF/CpG/antigen to DCs in charge of expressing endosomal toll like receptor 9 (TLR9) in the lymph nodes [102]. The enhanced Th1 response by anionic DODAB BF/OVA/CpG relative to OVA alone was evidenced by the 6 times increase in DTH reaction, the 42 times increase in IFN-γ secretion by lymph node cells in culture and the 9 times increase in IL-12 secretion also by lymph node cells in culture (Figure 7) [97].
At 0.1 mg/mL OVA, the dependence of DODAB BF/OVA size and zeta-potential on time and [DODAB] established 0.1 mM DODAB as suitable for obtaining stable and cationic DODAB BF/ OVA assemblies [97]. At 0.1 mM DODAB, 0.1 mg/mL OVA and 0.006 mM CpG, the zetapotential is zero showing charge neutralization [97]. At [CpG]> 0.006 mM, good colloidal stability for the anionic assemblies due to charge overcompensation was observed whereas at 0.020 mM CpG, these DODAB BF/OVA/CpG assemblies turned out to be highly effective *in vivo* generating responses similar to those elicited by the stable and cationic DODAB BF/OVA. The anti-OVA delayed-type hypersensitivity (DTH) reaction and the secretion of IFN-gamma and IL-12 resulted 6, 42 and 9 times larger for the DODAB BF/OVA/CpG-immunized mice than the same responses by OVA-immunized mice, respectively [97].Figure 7 A and B illustrate
The delayed-type hypersensitivity reaction (DTH) is an important *in vivo* response mediated by cells that can be quantified from the footpad sweelling test [44]. Mice immunized with OVA alone or with DODAB BF/CpG or with OVA / Al(OH)3 exhibited a footpad swelling equal to the one observed for naive mice [97]. The largest increase in footpad swelling was observed for DODAB BF/OVA/CpG mice immunization, which was about 6 times larger than the one observed for naive mice and 1.3 times larger than the one observed for DODAB BF/OVA [97]. Figure 7 C shows the improved DTH response in mice induced by the assemblies as adapted
Since OVA isoelectric point is 4.5 [98], this protein is negatively charged at 6.3, the pH of water. Thus, OVA adsorption onto DODAB BF is initially electrostatically driven.The OVA titration with DODAB BF determined ranges of DODAB concentration for occurrence of stable DODAB BF/OVA assemblies as illustrated in Figure 7 A and B. Similar results had been previously described also for DODAB BF/ bovin serum albumin or DODAB BF/ purified antigens from *Taenia crassiceps* [27].When the net charge of the assemblies is zero (at charge neutralization), maximal aggregation was observed for the assemblies. Further increasing DODAB concen‐ tration, stabilized them to a certain extent (Figure 7 B). However, sizes and polydispersities were still higher than those of DODAB BF in absence of OVA. Optimal colloidal stability was only achieved upon CpG addition to the system yielding high and negative zeta-potentials plus remarkably small polydispersity [98]. Oligonucleotides with less than 20 nucleotide residues usually behave as rigid charged rods in solution [99]. CpG would also adsorb as rigid charged rods on vacant positive sites of DODAB BF/ OVA assemblies inducing charge overcompensation and recovery of colloidal stability as previously described for a model
Figure 7 also illustrates the improvement in the cellular OVA-specific response from the analysis of cytokines secreted by lymph node cells of mice as adapted from reference [97]. IFNγ and IL-12 production is associated to the Th1 response whereas IL-10 and IL-13 production reflects the Th2 response. Levels of IFN-γ and IL-12 secretion observed in cell cultures of mice immunized with OVA or Al(OH)3/ OVA were low and close to the detection limit for these cytokine assays (Figure 7). For mice immunized with OVA/CpG or DODAB BF/OVA assem‐ blies, the secretion of IFN-γ and IL-12 substantially increased in comparison to secretion from cultured cells of OVA-immunized mice [97]. For the DODAB BF/OVA/CpG immunized mice,
the colloidal stability of the assemblies over a range of DODAB concentrations.
from reference [97].
16 Immune Response Activation
oligonucleotide above charge neutralization [56].
Other delivery systems based on anionic [103] or cationic lipid bilayer [41,44,100,104,105] have also been successful for improving Th1 response against important antigens such as those of influenza [103], hepatitis A and B [103,104], and fungal infections [105]. In general, cationic lipids are known for the production of a large inflammatory response [41, 44].However, for small cationic bilayer fragments as immunoadjuvants, this adverse reaction is absent [27, 97]. The small and anionic DODAB BF/OVA/CpG assemblies [97] also did not elicit adverse reactions similarly to other anionic assemblies [103]. There was no depot effect for these small assemblies [97] and their net negative charge ensured the absence of the adverse reactions observed previously for the large cationic liposomes and vesicles [100]. Recently, based on cross reactivity with *Neisseria lactamica* outer membrane vesicles (OMV) antigens, DODAB BF were combined with OMV to develop a vaccine against *Neisseria meningitidis* in young children [106]. Complexes of 25 μg of OMV in 0.1 mM of DODAB BF were colloidally stable, exhibiting optimal physical properties for efficient antigen presentation [106].
### **3. Cationic nanostructures based on chitosan and other biocompatible polymers**
Chitin is a long-chain polymer of N-acetylglucosamine, a derivative of glucose which can be deacetylated to yield chitosan as shown in Figure 8.
node cells in culture (Figure 7) [97].
above charge neutralization [56].
chitosan molecules, a new water-soluble compound, glycated chitosan (GC), was synthesized [109,110]. GC is a non-toxic biodegradable product used in laser immunotherapy (LIT) which combines local laser irradiation and local administration of GC at primary or metastatic cancers where the tumoral antigens of the irradiated tumor cells combined with GC elicit a potent immune response against the cancer [111,112]. After the first step involving tumor irradiation with a laser beam that causes swelling and disruption of tumor cells by thermal effect, the local injection of GC as immunoadjuvant would lead to the capture of the tumoral antigens by dendritic cells and migration to the lymph nodes where the antigens would be presented to T cells, thus activating cytotoxic T-lymphocytes [113-115]. LIT using GC induced regression of primary and secondary tumours in rats and caused resistance to repeated challenges with tumours of the same type [116]. Furthermore, rats developed immunity could be adoptively
Since they are biocompatible, biodegradable by deacetylases, mucoadhesive, and nontoxic, with antimicrobial, antiviral, and adjuvant properties, chitin, chitosan and their derivatives have been widely applied in medicine, pharmacy and vaccine design [107, 108]. Chitosan is soluble in diluted acids but is insoluble in water due to deprotonation of its amino moiety [109]. The poor solubility of chitosan at the pH of water represents a serious limitation for its applications as an immunoadjuvant in the clinics [107]. Several chitosan derivatives have been obtained to circumvent this limitation.For example, by attaching galactose moieties to the chitosan molecules, a new water-soluble compound, glycated chitosan (GC), was synthesized [109,110]. GC is a non-toxic biodegradable product used in laser immunotherapy (LIT) which combines local laser irradiation and local administration of GC at primary or metastatic cancers where the tumoral antigens of the irradiated tumor cells combined with GC elicit a potent immune response against the cancer [111,112]. After the first step involving tumor irradiation with a laser beam that causes swelling and disruption of tumor cells by thermal effect, the local injection of GC as immunoadjuvant would lead to the capture of the tumoral antigens by dendritic cells and migration to the lymph nodes where the antigens would be presented to T cells, thus activating cytotoxic Tlymphocytes [113-115]. LIT using GC induced regression of primary and secondary tumours in rats and caused resistance to repeated challenges with tumours of the same type [116]. Furthermore, rats developed immunity could be
seen from the kinetics obtained after adding DODAB BF at a final concentration of 0.005 (); 0.01 (); 0.02 (); 0.05 (); 0.1 (); 0.2 (); 0.5 () and 1mM DODAB () to 0.1mg/mL OVA. In (B), kinetical data are detailed for the larger DODAB BF concentrations.Assemblies were prepared in 1mM NaCl.In (C), delayed-type hypersensitivity response for BALB/c mice immunized with OVA in different adjuvant formulations determined from the footpad swelling (nm) standard error of the mean. Final concentrations are 0.1mM DODAB BF, 20M CpG and 0.1mg/mL Al(OH)3 and 0.1mg/mL OVA. p< 0.05 compared to naive (), p< 0.05 compared to OVA/Al(OH)3 (#) and p< 0.05 compared to OVA/DODAB/CpG () from reference [97].Adapted from Journal of Controlled Release, 160/2, Julio H.K. Rozenfeld, Sandriana R. Silva, Priscila A. Ranéia, Eliana Faquim-Mauro, Ana M. Carmona-Ribeiro, Stable assemblies of cationic bilayer fragments and CpG oligonucleotide with enhanced immunoadjuvant activity *in vivo*, 367-373 .Copyright 2012,
Other delivery systems based on anionic [103] or cationic lipid bilayer [41,44,100,104,105] have also been successful for improving Th1 response against important antigens such as those of influenza [103], hepatitis A and B [103,104], and fungal infections [105]. In general, cationic lipids are known for the production of a large inflammatory response [41, 44].However, for small cationic bilayer fragments as immunoadjuvants, this adverse reaction is absent [27, 97]. The small and anionic DODAB BF/OVA/CpG assemblies [97] also did not elicit adverse reactions similarly to other anionic assemblies [103]. There was no depot effect for these small assemblies [97] and their net negative charge ensured the absence of the adverse reactions observed previously for the large cationic liposomes and vesicles [100]. Recently, based on cross reactivity with *Neisseria lactamica* outer membrane vesicles (OMV) antigens, DODAB BF were combined with OMV to develop a vaccine against *Neisseria meningitidis* in young children [106]. Complexes of 25 µg of OMV in 0.1 mM of DODAB BF were colloidally stable, exhibiting optimal physical properties for efficient antigen presentation [106].
Chitin is a long-chain polymer of N-acetylglucosamine, a derivative of glucose which can be deacetylated to yield
deacetylation
**3. Cationic nanostructures based on chitosan and other biocompatible polymers.**
Chitosan nanoparticles have been obtained by ionotropic gelation, complex coacervation, emulsion and microemulsion techniques, and self-assembly of hydrophobically modified chitosan [117]. Ionotropic gelation consists of the ionic crosslinking of chitosan with multiva‐ lent counter-ions such as sodium tripolyphosphate (TPP) by adding a dilute chitosan acid solution to a solution of TPP or vice versa, with stirring [118]. Chitosan particles of nanometric size were obtained for chitosan concentrations up to 2.8 g L−1 and TPP concentrations from 0.21 to 0.43 g L−1. The size and surface charge of particles can be modified by varying the ratio of chitosan and stabilizer. The main problematic aspects of the technique are the poor colloidal stability of the dispersion which may require the addition of stabilizers, and the need of using very dilute solutions which may be inconvenient when large amounts of nanoparticles are required [117]. Complex coacervation is achieved by mixing two oppositely charged polye‐ lectrolytes. The polyelectrolyte or coacervate complex is structured as nanoparticles. Chitosan– poly (acrylic acid) (PAA) nanoparticles carrying a positive charge with sizes from 50 to 400 nm were obtained by the dropwise addition of dilute chitosan solutions [119]. Carboxyme‐ thylcellulose and alginate have also been complexed with chitosan to prepare nanoparticles. Chitosan–carboxymethylcellulose nanoparticles were subsequently coated with plasmid DNA (pDNA) [120]. Chitosan–alginate nanoparticles were loaded with insulin [121]. Nanoparticles have also been obtained in which the polyanion is the active principle itself as for example heparin or DNA or even siRNA. Chitosan–heparin nanoparticles crosslinked with TPP were described [122]. Figure 9 shows on the left a scanning electron micrograph of chitosan– heparin nanoparticles adapted from reference [117] and, on the right, an atomic force micrograph of
particles can be modified by varying the ratio of chitosan and stabilizer. The main problematic aspects of the technique are the poor colloidal stability of the dispersion which may require the addition of stabilizers, and the need of using very dilute solutions which may be inconvenient when large amounts of nanoparticles are required [117]. Complex coacervation is achieved by mixing two oppositely charged polyelectrolytes. The polyelectrolyte or coacervate complex is structured as nanoparticles. Chitosan–poly (acrylic acid) (PAA) nanoparticles carrying a positive charge with sizes from
and TPP concentrations from 0.21 to 0.43 g L<sup>−</sup><sup>1</sup>
. The size and surface charge of
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 19
Chitosan nanoparticles have been obtained by ionotropic gelation, complex coacervation, emulsion and microemulsion techniques, and self-assembly of hydrophobically modified chitosan [117]. Ionotropic gelation consists of the ionic crosslinking of chitosan with multivalent counter-ions such as sodium tripolyphosphate (TPP) by adding a dilute chitosan acid solution to a solution of TPP or vice versa, with stirring [118]. Chitosan particles of nanometric size were obtained for
transferred [116].
adoptively transferred [116].
chitosan concentrations up to 2.8 g L<sup>−</sup><sup>1</sup>
with permission from Elsevier.
chitosan as shown in Figure 8.
Figure 8.Chitin yielding chitosan by deacetylation.
**Figure 8.** Chitin yielding chitosan by deacetylation.
inducing charge overcompensation and recovery of colloidal stability as previously described for a model oligonucleotide
Figure 7 also illustrates the improvement in the cellular OVA-specific response from the analysis of cytokines secreted by lymph node cells of mice as adapted from reference [97]. IFN-γ and IL-12 production is associated to the Th1 response whereas IL-10 and IL-13 production reflects the Th2 response. Levels of IFN- γ and IL-12 secretion observed in cell cultures of mice immunized with OVA or Al(OH)3/ OVA were low and close to the detection limit for these cytokine assays (Figure 7). For mice immunized with OVA/CpG or DODAB BF/OVA assemblies, the secretion of IFN- γ and IL-12 substantially increased in comparison to secretion from cultured cells of OVA-immunized mice [97]. For the DODAB BF/OVA/CpG immunized mice, the secretion of IFN- γ and IL-12 increased by 33 and 49 %, respectively when compared to DODAB BF/OVA-immunized mice group and 52% and 35% when compared to OVA/CpG-group. In contrast, the highest secretion of IL-10 and IL-13 was observed in cultures of cells from mice that were immunized with Al(OH)3 /OVA whereas all other assemblies resulted in poor production of these cytokines [97]. Immune responses were similar for anionic DODAB BF/OVA/CpG and cationic DODAB BF/OVA of similar sizes showing that the charge is not important but the size is. The adsorption of antigen on the surface of DODAB large vesicles was shown to stimulate active antigen capture and presentation by dendritic cells (DCs) [41]. Administration of antigen adsorbed on DODAB large vesicles (LV) resulted in formation of an antigen depot at the site of injection which hampered the rapid clearance of antigen that takes place in absence of a carrier [100]. In contrast to DODAB LV, the small DODAB BF/antigen assemblies did not result in any observable depot effect [27, 97]. Furthermore, since the depot was absent for DODAB BF/CpG/OVA the immunostimulatory effect must have occurred via direct effect of the assemblies on the lymphonode antigen presenting cells [97]. Only nanoparticles can specifically target lymph node -resident cells [101]. CpG combined with DODAB BF
anionic DODAB BF/OVA/CpG relative to OVA alone was evidenced by the 6 times increase in DTH reaction, the 42 times increase in IFN-γ secretion by lymph node cells in culture and the 9 times increase in IL-12 secretion also by lymph
*vivo*. In (A), the effect of time and DODAB concentration on zeta-average diameter of DODAB BF/OVA assemblies is **Figure 7.** Nanostructures of cationic bilayer fragments and CpG oligonucleotide with enhanced immunoadjuvant ac‐ tivity *in vivo*. In (A), the effect of time and DODAB concentration on zeta-average diameter of DODAB BF/OVA assem‐ blies is seen from the kinetics obtained after adding DODAB BF at a final concentration of 0.005 (∎); 0.01 (○); 0.02 (▲); 0.05 (▽); 0.1 (◆); 0.2 (□); 0.5 (●) and 1mM DODAB (△) to 0.1mg/mL OVA. In (B), kinetical data are detailed for the larger DODAB BF concentrations.Assemblies were prepared in 1mM NaCl.In (C), delayed-type hypersensitivity re‐ sponse for BALB/c mice immunized with OVA in different adjuvant formulations determined from the footpad swel‐ ling (nm) ± standard error of the mean. Final concentrations are 0.1mM DODAB BF, 20μM CpG and 0.1mg/mL Al(OH)<sup>3</sup> and 0.1mg/mL OVA. p< 0.05 compared to naive (○), p< 0.05 compared to OVA/Al(OH)3 (#) and p< 0.05 compared to OVA/DODAB/CpG (△) from reference [97].Adapted from Journal of Controlled Release, 160/2, Julio H.K. Rozenfeld, Sandriana R. Silva, Priscila A. Ranéia, Eliana Faquim-Mauro, Ana M. Carmona-Ribeiro, Stable assemblies of cationic bi‐ layer fragments and CpG oligonucleotide with enhanced immunoadjuvant activity *in vivo*, 367-373.Copyright 2012, with permission from Elsevier.
Figure 7.Nanostructures of cationic bilayer fragments and CpG oligonucleotide with enhanced immunoadjuvant activity *in*
Since they are biocompatible, biodegradable by deacetylases, mucoadhesive, and nontoxic, with antimicrobial, antiviral, and adjuvant properties, chitin, chitosan and their derivatives have been widely applied in medicine, pharmacy and vaccine design [107, 108]. Chitosan is soluble in diluted acids but is insoluble in water due to deprotonation of its amino moiety [109]. The poor solubility of chitosan at the pH of water represents a serious limitation for its applications as an immunoadjuvant in the clinics [107]. Several chitosan derivatives have been obtained to circumvent this limitation.For example, by attaching galactose moieties to the
Since they are biocompatible, biodegradable by deacetylases, mucoadhesive, and nontoxic, with antimicrobial, antiviral,
Chitin is a long-chain polymer of N-acetylglucosamine, a derivative of glucose which can be deacetylated to yield
seen from the kinetics obtained after adding DODAB BF at a final concentration of 0.005 (); 0.01 (); 0.02 (); 0.05 (); 0.1 (); 0.2 (); 0.5 () and 1mM DODAB () to 0.1mg/mL OVA. In (B), kinetical data are detailed for the larger DODAB BF concentrations.Assemblies were prepared in 1mM NaCl.In (C), delayed-type hypersensitivity response for BALB/c mice immunized with OVA in different adjuvant formulations determined from the footpad swelling (nm) standard error of the mean. Final concentrations are 0.1mM DODAB BF, 20M CpG and 0.1mg/mL Al(OH)3 and 0.1mg/mL OVA. p< 0.05 compared to naive (), p< 0.05 compared to OVA/Al(OH)3 (#) and p< 0.05 compared to OVA/DODAB/CpG () from reference [97].Adapted from Journal of Controlled Release, 160/2, Julio H.K. Rozenfeld, Sandriana R. Silva, Priscila A. Ranéia, Eliana Faquim-Mauro, Ana M. Carmona-Ribeiro, Stable assemblies of cationic bilayer fragments and CpG oligonucleotide with enhanced immunoadjuvant activity *in vivo*, 367-373 .Copyright 2012,
Other delivery systems based on anionic [103] or cationic lipid bilayer [41,44,100,104,105] have also been successful for improving Th1 response against important antigens such as those of influenza [103], hepatitis A and B [103,104], and fungal infections [105]. In general, cationic lipids are known for the production of a large inflammatory response [41, 44].However, for small cationic bilayer fragments as immunoadjuvants, this adverse reaction is absent [27, 97]. The small and anionic DODAB BF/OVA/CpG assemblies [97] also did not elicit adverse reactions similarly to other anionic assemblies [103]. There was no depot effect for these small assemblies [97] and their net negative charge ensured the absence of the adverse reactions observed previously for the large cationic liposomes and vesicles [100]. Recently, based on cross reactivity with *Neisseria lactamica* outer membrane vesicles (OMV) antigens, DODAB BF were combined with OMV to develop a vaccine against *Neisseria meningitidis* in young children [106]. Complexes of 25 µg of OMV in 0.1 mM of DODAB BF were colloidally stable, exhibiting optimal physical properties for efficient antigen presentation [106].
Figure 8.Chitin yielding chitosan by deacetylation. **Figure 8.** Chitin yielding chitosan by deacetylation.
with permission from Elsevier.
chitosan as shown in Figure 8.
Since they are biocompatible, biodegradable by deacetylases, mucoadhesive, and nontoxic, with antimicrobial, antiviral, and adjuvant properties, chitin, chitosan and their derivatives have been widely applied in medicine, pharmacy and vaccine design [107, 108]. Chitosan is soluble in diluted acids but is insoluble in water due to deprotonation of its amino moiety [109]. The poor solubility of chitosan at the pH of water represents a serious limitation for its applications as an immunoadjuvant in the clinics [107]. Several chitosan derivatives have been obtained to circumvent this limitation.For example, by attaching galactose moieties to the
Figure 7.Nanostructures of cationic bilayer fragments and CpG oligonucleotide with enhanced immunoadjuvant activity *in vivo*. In (A), the effect of time and DODAB concentration on zeta-average diameter of DODAB BF/OVA assemblies is
**Figure 7.** Nanostructures of cationic bilayer fragments and CpG oligonucleotide with enhanced immunoadjuvant ac‐ tivity *in vivo*. In (A), the effect of time and DODAB concentration on zeta-average diameter of DODAB BF/OVA assem‐ blies is seen from the kinetics obtained after adding DODAB BF at a final concentration of 0.005 (∎); 0.01 (○); 0.02 (▲); 0.05 (▽); 0.1 (◆); 0.2 (□); 0.5 (●) and 1mM DODAB (△) to 0.1mg/mL OVA. In (B), kinetical data are detailed for the larger DODAB BF concentrations.Assemblies were prepared in 1mM NaCl.In (C), delayed-type hypersensitivity re‐ sponse for BALB/c mice immunized with OVA in different adjuvant formulations determined from the footpad swel‐ ling (nm) ± standard error of the mean. Final concentrations are 0.1mM DODAB BF, 20μM CpG and 0.1mg/mL Al(OH)<sup>3</sup> and 0.1mg/mL OVA. p< 0.05 compared to naive (○), p< 0.05 compared to OVA/Al(OH)3 (#) and p< 0.05 compared to OVA/DODAB/CpG (△) from reference [97].Adapted from Journal of Controlled Release, 160/2, Julio H.K. Rozenfeld, Sandriana R. Silva, Priscila A. Ranéia, Eliana Faquim-Mauro, Ana M. Carmona-Ribeiro, Stable assemblies of cationic bi‐ layer fragments and CpG oligonucleotide with enhanced immunoadjuvant activity *in vivo*, 367-373.Copyright 2012,
0.0 0.2 0.4 0.6 0.8
OVA
OVA/CpG
DODAB/OVA
DODAB/OVA/CpG
Al(OH)3 /OVA
Al(OH)3 /OVA/CpG
OVA
OVA/CpG
DODAB/OVA
DODAB/OVA/CpG
Al(OH)3 /OVA
Al(OH)3 /OVA/CpG
01234
012
012
Cytokine concentration (ng/mL)
IL-10
ConA OVA Medium
IFN-
IL-13
IL-12
inducing charge overcompensation and recovery of colloidal stability as previously described for a model oligonucleotide
Figure 7 also illustrates the improvement in the cellular OVA-specific response from the analysis of cytokines secreted by lymph node cells of mice as adapted from reference [97]. IFN-γ and IL-12 production is associated to the Th1 response whereas IL-10 and IL-13 production reflects the Th2 response. Levels of IFN- γ and IL-12 secretion observed in cell cultures of mice immunized with OVA or Al(OH)3/ OVA were low and close to the detection limit for these cytokine assays (Figure 7). For mice immunized with OVA/CpG or DODAB BF/OVA assemblies, the secretion of IFN- γ and IL-12 substantially increased in comparison to secretion from cultured cells of OVA-immunized mice [97]. For the DODAB BF/OVA/CpG immunized mice, the secretion of IFN- γ and IL-12 increased by 33 and 49 %, respectively when compared to DODAB BF/OVA-immunized mice group and 52% and 35% when compared to OVA/CpG-group. In contrast, the highest secretion of IL-10 and IL-13 was observed in cultures of cells from mice that were immunized with Al(OH)3 /OVA whereas all other assemblies resulted in poor production of these cytokines [97]. Immune responses were similar for anionic DODAB BF/OVA/CpG and cationic DODAB BF/OVA of similar sizes showing that the charge is not important but the size is. The adsorption of antigen on the surface of DODAB large vesicles was shown to stimulate active antigen capture and presentation by dendritic cells (DCs) [41]. Administration of antigen adsorbed on DODAB large vesicles (LV) resulted in formation of an antigen depot at the site of injection which hampered the rapid clearance of antigen that takes place in absence of a carrier [100]. In contrast to DODAB LV, the small DODAB BF/antigen assemblies did not result in any observable depot effect [27, 97]. Furthermore, since the depot was absent for DODAB BF/CpG/OVA the immunostimulatory effect must have occurred via direct effect of the assemblies on the lymphonode antigen presenting cells [97]. Only nanoparticles can specifically target lymph node -resident cells [101]. CpG combined with DODAB BF yielded improved cellular Th1 response [97] possibly due to the appropriate targeting of DODAB BF/CpG/antigen to DCs in charge of expressing endosomal toll like receptor 9 (TLR9) in the lymph nodes [102]. The enhanced Th1 response by anionic DODAB BF/OVA/CpG relative to OVA alone was evidenced by the 6 times increase in DTH reaction, the 42 times increase in IFN-γ secretion by lymph node cells in culture and the 9 times increase in IL-12 secretion also by lymph
above charge neutralization [56].
node cells in culture (Figure 7) [97].
0 10 20 30 40 50 60
OVA
with permission from Elsevier.
t (min)
| doab | 2025-04-07T04:13:03.364226 | 20-4-2021 17:49 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffa90cf5-6ee0-48d5-8212-f65bf49c005f",
"url": "https://mts.intechopen.com/storage/books/3811/authors_book/authors_book.pdf",
"author": "",
"title": "Immune Response Activation",
"publisher": "IntechOpen",
"isbn": "9789535113744",
"section_idx": 4
} |
ffa90cf5-6ee0-48d5-8212-f65bf49c005f.5 | #
/OVA/CpG
Al(OH)3
3
/OVA
Al(OH)
OVA /CpG
180 1440
| doab | 2025-04-07T04:13:03.369683 | 20-4-2021 17:49 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffa90cf5-6ee0-48d5-8212-f65bf49c005f",
"url": "https://mts.intechopen.com/storage/books/3811/authors_book/authors_book.pdf",
"author": "",
"title": "Immune Response Activation",
"publisher": "IntechOpen",
"isbn": "9789535113744",
"section_idx": 5
} |
ffa90cf5-6ee0-48d5-8212-f65bf49c005f.6 | #
DODAB/OVA/CpG
DODAB/OVA
0
0.0 0.5 1.0 1.5
Naive
C
Footpad swelling (mm)
100
200
300
400
Dz (nm)
B
A
18 Immune Response Activation
chitosan molecules, a new water-soluble compound, glycated chitosan (GC), was synthesized [109,110]. GC is a non-toxic biodegradable product used in laser immunotherapy (LIT) which combines local laser irradiation and local administration of GC at primary or metastatic cancers where the tumoral antigens of the irradiated tumor cells combined with GC elicit a potent immune response against the cancer [111,112]. After the first step involving tumor irradiation with a laser beam that causes swelling and disruption of tumor cells by thermal effect, the local injection of GC as immunoadjuvant would lead to the capture of the tumoral antigens by dendritic cells and migration to the lymph nodes where the antigens would be presented to T cells, thus activating cytotoxic T-lymphocytes [113-115]. LIT using GC induced regression of primary and secondary tumours in rats and caused resistance to repeated challenges with tumours of the same type [116]. Furthermore, rats developed immunity could be adoptively transferred [116]. and adjuvant properties, chitin, chitosan and their derivatives have been widely applied in medicine, pharmacy and vaccine design [107, 108]. Chitosan is soluble in diluted acids but is insoluble in water due to deprotonation of its amino moiety [109]. The poor solubility of chitosan at the pH of water represents a serious limitation for its applications as an immunoadjuvant in the clinics [107]. Several chitosan derivatives have been obtained to circumvent this limitation.For example, by attaching galactose moieties to the chitosan molecules, a new water-soluble compound, glycated chitosan (GC), was synthesized [109,110]. GC is a non-toxic biodegradable product used in laser immunotherapy (LIT) which combines local laser irradiation and local administration of GC at primary or metastatic cancers where the tumoral antigens of the irradiated tumor cells combined with GC elicit a potent immune response against the cancer [111,112]. After the first step involving tumor irradiation with a laser beam that causes swelling and disruption of tumor cells by thermal effect, the local injection of GC as immunoadjuvant would lead to the capture of the tumoral antigens by dendritic cells and migration to the lymph nodes where the antigens would be presented to T cells, thus activating cytotoxic Tlymphocytes [113-115]. LIT using GC induced regression of primary and secondary tumours in rats and caused resistance to repeated challenges with tumours of the same type [116]. Furthermore, rats developed immunity could be adoptively transferred [116]. Chitosan nanoparticles have been obtained by ionotropic gelation, complex coacervation, emulsion and microemulsion techniques, and self-assembly of hydrophobically modified chitosan [117]. Ionotropic gelation consists of the ionic crosslinking of chitosan with multivalent counter-ions such as sodium tripolyphosphate (TPP) by adding a dilute chitosan acid solution to a solution of TPP or vice versa, with stirring [118]. Chitosan particles of nanometric size were obtained for chitosan concentrations up to 2.8 g L<sup>−</sup><sup>1</sup> and TPP concentrations from 0.21 to 0.43 g L<sup>−</sup><sup>1</sup> . The size and surface charge of
Chitosan nanoparticles have been obtained by ionotropic gelation, complex coacervation, emulsion and microemulsion techniques, and self-assembly of hydrophobically modified chitosan [117]. Ionotropic gelation consists of the ionic crosslinking of chitosan with multiva‐ lent counter-ions such as sodium tripolyphosphate (TPP) by adding a dilute chitosan acid solution to a solution of TPP or vice versa, with stirring [118]. Chitosan particles of nanometric size were obtained for chitosan concentrations up to 2.8 g L−1 and TPP concentrations from 0.21 to 0.43 g L−1. The size and surface charge of particles can be modified by varying the ratio of chitosan and stabilizer. The main problematic aspects of the technique are the poor colloidal stability of the dispersion which may require the addition of stabilizers, and the need of using very dilute solutions which may be inconvenient when large amounts of nanoparticles are required [117]. Complex coacervation is achieved by mixing two oppositely charged polye‐ lectrolytes. The polyelectrolyte or coacervate complex is structured as nanoparticles. Chitosan– poly (acrylic acid) (PAA) nanoparticles carrying a positive charge with sizes from 50 to 400 nm were obtained by the dropwise addition of dilute chitosan solutions [119]. Carboxyme‐ thylcellulose and alginate have also been complexed with chitosan to prepare nanoparticles. Chitosan–carboxymethylcellulose nanoparticles were subsequently coated with plasmid DNA (pDNA) [120]. Chitosan–alginate nanoparticles were loaded with insulin [121]. Nanoparticles have also been obtained in which the polyanion is the active principle itself as for example heparin or DNA or even siRNA. Chitosan–heparin nanoparticles crosslinked with TPP were described [122]. Figure 9 shows on the left a scanning electron micrograph of chitosan– heparin nanoparticles adapted from reference [117] and, on the right, an atomic force micrograph of particles can be modified by varying the ratio of chitosan and stabilizer. The main problematic aspects of the technique are the poor colloidal stability of the dispersion which may require the addition of stabilizers, and the need of using very dilute solutions which may be inconvenient when large amounts of nanoparticles are required [117]. Complex coacervation is achieved by mixing two oppositely charged polyelectrolytes. The polyelectrolyte or coacervate complex is structured as nanoparticles. Chitosan–poly (acrylic acid) (PAA) nanoparticles carrying a positive charge with sizes from therefore are potentially useful for vaccine design.
permission from Elsevier.
similar to those described for the non hybrid nanoparticles [128].
chitosan-carboxymethylcellulose nanoparticles [123] prepared by complex coacervation without using without using any organic solvent or crosslinker. These coacervates have the interesting property of combining with cisplatin [123] but can also be combined with peptides, proteins and DNA and therefore are potentially useful for vaccine design. example heparin or DNA or even siRNA. Chitosan–heparin nanoparticles crosslinked with TPP were described [122]. Figure 9 shows on the left a scanning electron micrograph of chitosan– heparin nanoparticles adapted from reference [117] and, on the right, an atomic force micrograph of chitosan-carboxymethylcellulose nanoparticles [123] prepared by complex coacervation without using without using any organic solvent or crosslinker. These coacervates have the interesting property of combining with cisplatin [123] but can also be combined with peptides, proteins and DNA and
50 to 400 nm were obtained by the dropwise addition of dilute chitosan solutions [119]. Carboxymethylcellulose and
nanoparticles were subsequently coated with plasmid DNA (pDNA) [120]. Chitosan–alginate nanoparticles were loaded with insulin [121]. Nanoparticles have also been obtained in which the polyanion is the active principle itself as for
charged MCC) without using any organic solvent or crosslinker yielded results similar to those
**Figure 10.** Uptake of chitosan (a), TMC (b) and MCC nanoparticles loaded with bovine serum albumin and labeled with a green fluorescent marker adapted from reference [124].Reprinted from International Journal of Pharmaceutics, 363/1–2, B. Sayın, S. Somavarapu, X.W. Li, M. Thanou, D. Sesardic, H.O.Alpar, S. Şenel, Mono-N-carboxymethyl chitosan (MCC) andN-trimethyl chitosan (TMC) nanoparticles for non-invasive vaccine delivery, 139-148. Copyright 2008, with
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 21
A very promising approach is the development of vaccines based on the initiation of immune response by transfecting dendritic cells (DC) with DNA encoding tumor-associated antigens or immunostimulatory molecules such as cytokines or chemokines [129].Although unmodi‐ fied chitosan may not be a good gene delivery carrier for DCs because of its low transfection efficiency, some modified chitosans showed improved behavior in delivering genes into DCs. IL-12 gene was delivered to DCs *in vivo* using mannosylated chitosan (MC) via mannose receptor-mediated endocytosis [130,131]. MC not only has good physicochemical properties and low cytotoxicity, but also transfect DCs with much higher efficiency than do the unmodi‐ fied chitosan particles *in vitro* [130]. *In vivo*, intratumoral injection of MC /plasmid encoding murine IL-12 complex into BALB/c mice bearing CT-26 carcinoma cells clearly suppressed tumor growth and angiogenesis, and significantly induced cell cycle arrest and apoptosis [131]. Mannose-bearing chitosan nanoparticles were also synthesized to entrap complexes of DNA with polyethyleneimine (PEI) and improve the delivery of DNA into antigen-presenting cells (APCs) after intramuscular (i.m.) injection [132]. Compared with the traditional chitosan microspheres, these nanoparticles targeted the DCs (which express a high density of mannose receptors when they are immature) and released their PEI/DNA cargo inside them. After i.m. immunization, the MC/PEI/DNA nanospheres induced significantly enhanced serum anti‐ body and cytotoxic T lymphocyte (CTL) responses in comparison to naked DNA [132]. Figure 11 shows the MC/PEI/DNA nanoparticles and superior cumulative DNA release *in vitro* of
In recent years, DC vaccines, especially DNA-based DC vaccines, have been the focus of attention in immunotherapy against cancer [129]. Genetically engineered DCs previously modified *in vitro* can then be implanted *ex vivo* and activate the tumor-specific CTL response for the killing of cancer cells [129]. Most of these immunotherapy approaches based on DCs genetic modification for the treatment of melanoma [133,134], renal carcinoma [135] and other malignant diseases [136] are in phase I/II clinical trials. However, in many of these clinical trials adenovirus vector is the carrier still employed for DCs modification [137].Since the major
described for the non hybrid nanoparticles [128].
permission from Elsevier.
these nanoparticles in PBS buffer as adapted from reference [132].
Figure 9.Nanoparticles obtained by complex coacervation of oppositely charged polyelectrolytes: chitosan-heparin on the left [117] and chitosan-carboxymethylcellulose on the right [123].On the left, adapted with permission from Peniche H, Peniche C. Polym Int. Chitosan nanoparticles: a contribution to nanomedicine.2011; 60:883–889. DOI: 10.1002/pi.3056, http://onlinelibrary.wiley.com/doi/10.1002/pi.3056/abstract. Copyright 2011, John Wiley and Sons.On the right, adapted with permission from Vieira DB, Kim V, Petri DFS, Menck CFM, Carmona-Ribeiro AM. Supramolecular assemblies of cisplatin and polyelectrolytes: preparation, characterization and activity against cancer cells. In: Laudon M, Romanowicz B. (eds.) Nanotech Conference & Expo 2012: An Interdisciplinary **Figure 9.** Nanoparticles obtained by complex coacervation of oppositely charged polyelectrolytes: chitosan-heparin on the left [117] and chitosan-carboxymethylcellulose on the right [123].On the left, adapted with permission from Peniche H, Peniche C. Polym Int. Chitosan nanoparticles: a contribution to nanomedicine.2011; 60:883–889. DOI: 10.1002/pi.3056, http://onlinelibrary.wiley.com/doi/10.1002/pi.3056/abstract. Copyright 2011, John Wiley and Sons.On the right, adapted with permission from Vieira DB, Kim V, Petri DFS, Menck CFM, Carmona-Ribeiro AM. Supra‐ molecular assemblies of cisplatin and polyelectrolytes: preparation, characterization and activity against cancer cells. In: Laudon M, Romanowicz B. (eds.) Nanotech Conference & Expo 2012: An Interdisciplinary Integrative Forum on Nanotechnology, Microtechnology, Biotechnology and Cleantechnology, Santa Clara, CA, United States, June 18-21, 2012 (2012), 3, 182-185. Publisher: (CRC Press, Boca Raton, FLA) ISBN 978-1-4665-6287-5.
Integrative Forum on Nanotechnology, Microtechnology, Biotechnology and Cleantechnology, Santa Clara, CA, United States, June 18-21, 2012 (2012), 3, 182-185. Publisher: (CRC Press, Boca Raton, FLA) ISBN 978-1-4665-6287-5. Other interesting chitosan derivatives have been synthesized and used for the preparation of nanoparticles such as Ntrimethylchitosan (TMC) and mono-N-carboxymethyl chitosan (MCC) [124]. In general, mucosal applications of antigens result in poor immune responses. Therefore, mucoadhesive adjuvants are required to enhance the immune response by improving both antigen protection and its cellular uptake. Nanoparticles of TMC or MCC were prepared using the ionic gelation method and loaded with tetanus toxoid (TT) exhibiting high loading efficacy (>90% m/m), sizes within the range of 40-400nm and negative or positive surface charge for MCC and TMC, respectively [124]. The structural integrity of the TT in the formulations was confirmed by SDS-PAGE electrophoresis and the intranasal application in mice elicited high serum IgG titres [124].Figure 10 illustrates the excellent uptake by macrophages of these nanoparticles labeled with a green fluorescent marker as adapted from [124]. Other interesting chitosan derivatives have been synthesized and used for the preparation of nanoparticles such as N-trimethylchitosan (TMC) and mono-N-carboxymethyl chitosan (MCC) [124]. In general, mucosal applications of antigens result in poor immune responses. Therefore, mucoadhesive adjuvants are required to enhance the immune response by improv‐ ing both antigen protection and its cellular uptake. Nanoparticles of TMC or MCC were prepared using the ionic gelation method and loaded with tetanus toxoid (TT) exhibiting high loading efficacy (>90% m/m), sizes within the range of 40-400nm and negative or positive surface charge for MCC and TMC, respectively [124]. The structural integrity of the TT in the formulations was confirmed by SDS-PAGE electrophoresis and the intranasal application in mice elicited high serum IgG titres [124].Figure 10 illustrates the excellent uptake by macro‐ phages of these nanoparticles labeled with a green fluorescent marker as adapted from [124].
> Due to the positive charge of chitosan and TMC, and the negative charge of TT, the antigen loading is driven by the electrostatic attraction. There is also a superior association and internalization of chitosan or chitosan derivatives nanoparticles with cells due to their mucoadhesive character [125-127]. CS nanoparticles have a higher association and internali‐ sation with gastrointestinal tissue cells due to electrostatic interactions compared to polystyr‐ ene nanoparticles [125].More recently, hybrid nanoparticles based on the complexation between oppositely charged chitosan derivatives (positively charged TMC and negatively
Figure 10.Uptake of chitosan (a), TMC (b) and MCC nanoparticles loaded with bovine serum albumin and labeled with a green fluorescent marker adapted from reference [124].Reprinted from International Journal of Pharmaceutics, 363/1–2, B. Sayn, S. Somavarapu, X.W. Li, M. Thanou, D. Sesardic, H.O.Alpar, S. Şenel, Mono-N-carboxymethyl chitosan (MCC) andN-trimethyl chitosan (TMC) nanoparticles for non-invasive vaccine delivery, 139-148. Copyright 2008, with
Due to the positive charge of chitosan and TMC, and the negative charge of TT, the antigen loading is driven by the electrostatic attraction. There is also a superior association and internalization of chitosan or chitosan derivatives nanoparticles with cells due to their mucoadhesive character [125-127]. CS nanoparticles have a higher association and internalisation with gastrointestinal tissue cells due to electrostatic interactions compared to polystyrene nanoparticles [125].More recently, hybrid nanoparticles based on the complexation between oppositely charged chitosan derivatives (positively charged TMC and negatively charged MCC) without using any organic solvent or crosslinker yielded results
chitosan-carboxymethylcellulose nanoparticles [123] prepared by complex coacervation without using without using any organic solvent or crosslinker. These coacervates have the interesting property of combining with cisplatin [123] but can also be combined with peptides,
Figure 9.Nanoparticles obtained by complex coacervation of oppositely charged polyelectrolytes: chitosan-heparin on the left [117] and chitosan-carboxymethylcellulose on the right [123].On the left, adapted with permission from Peniche H, Peniche C. Polym Int. Chitosan nanoparticles: a contribution to nanomedicine.2011; 60:883–889. DOI: 10.1002/pi.3056, http://onlinelibrary.wiley.com/doi/10.1002/pi.3056/abstract. Copyright 2011, John Wiley and Sons.On the right, adapted with permission from Vieira DB, Kim V, Petri DFS, Menck CFM, Carmona-Ribeiro AM. Supramolecular assemblies of cisplatin and polyelectrolytes: preparation, characterization and activity against cancer cells. In: Laudon M, Romanowicz B. (eds.) Nanotech Conference & Expo 2012: An Interdisciplinary Integrative Forum on Nanotechnology, Microtechnology, Biotechnology and Cleantechnology, Santa Clara, CA, United
**Figure 9.** Nanoparticles obtained by complex coacervation of oppositely charged polyelectrolytes: chitosan-heparin on the left [117] and chitosan-carboxymethylcellulose on the right [123].On the left, adapted with permission from Peniche H, Peniche C. Polym Int. Chitosan nanoparticles: a contribution to nanomedicine.2011; 60:883–889. DOI: 10.1002/pi.3056, http://onlinelibrary.wiley.com/doi/10.1002/pi.3056/abstract. Copyright 2011, John Wiley and Sons.On the right, adapted with permission from Vieira DB, Kim V, Petri DFS, Menck CFM, Carmona-Ribeiro AM. Supra‐ molecular assemblies of cisplatin and polyelectrolytes: preparation, characterization and activity against cancer cells. In: Laudon M, Romanowicz B. (eds.) Nanotech Conference & Expo 2012: An Interdisciplinary Integrative Forum on Nanotechnology, Microtechnology, Biotechnology and Cleantechnology, Santa Clara, CA, United States, June 18-21,
States, June 18-21, 2012 (2012), 3, 182-185. Publisher: (CRC Press, Boca Raton, FLA) ISBN 978-1-4665-6287-5.
Other interesting chitosan derivatives have been synthesized and used for the preparation of nanoparticles such as N-trimethylchitosan (TMC) and mono-N-carboxymethyl chitosan (MCC) [124]. In general, mucosal applications of antigens result in poor immune responses. Therefore, mucoadhesive adjuvants are required to enhance the immune response by improv‐ ing both antigen protection and its cellular uptake. Nanoparticles of TMC or MCC were prepared using the ionic gelation method and loaded with tetanus toxoid (TT) exhibiting high loading efficacy (>90% m/m), sizes within the range of 40-400nm and negative or positive surface charge for MCC and TMC, respectively [124]. The structural integrity of the TT in the formulations was confirmed by SDS-PAGE electrophoresis and the intranasal application in mice elicited high serum IgG titres [124].Figure 10 illustrates the excellent uptake by macro‐ phages of these nanoparticles labeled with a green fluorescent marker as adapted from [124].
2012 (2012), 3, 182-185. Publisher: (CRC Press, Boca Raton, FLA) ISBN 978-1-4665-6287-5.
Other interesting chitosan derivatives have been synthesized and used for the preparation of nanoparticles such as Ntrimethylchitosan (TMC) and mono-N-carboxymethyl chitosan (MCC) [124]. In general, mucosal applications of antigens result in poor immune responses. Therefore, mucoadhesive adjuvants are required to enhance the immune response by improving both antigen protection and its cellular uptake. Nanoparticles of TMC or MCC were prepared using the ionic gelation method and loaded with tetanus toxoid (TT) exhibiting high loading efficacy (>90% m/m), sizes within the range of 40-400nm and negative or positive surface charge for MCC and TMC, respectively [124]. The structural integrity of the TT in the formulations was confirmed by SDS-PAGE electrophoresis and the intranasal application in mice elicited high serum IgG titres [124].Figure 10 illustrates the excellent uptake by macrophages of these nanoparticles labeled with a
Figure 10.Uptake of chitosan (a), TMC (b) and MCC nanoparticles loaded with bovine serum albumin and labeled with a green fluorescent marker adapted from reference [124].Reprinted from International Journal of Pharmaceutics, 363/1–2, B. Sayn, S. Somavarapu, X.W. Li, M. Thanou, D. Sesardic, H.O.Alpar, S. Şenel, Mono-N-carboxymethyl chitosan (MCC) andN-trimethyl chitosan (TMC) nanoparticles for non-invasive vaccine delivery, 139-148. Copyright 2008, with
Due to the positive charge of chitosan and TMC, and the negative charge of TT, the antigen loading is driven by the electrostatic attraction. There is also a superior association and internalization of chitosan or chitosan derivatives nanoparticles with cells due to their mucoadhesive character [125-127]. CS nanoparticles have a higher association and internali‐ sation with gastrointestinal tissue cells due to electrostatic interactions compared to polystyr‐ ene nanoparticles [125].More recently, hybrid nanoparticles based on the complexation between oppositely charged chitosan derivatives (positively charged TMC and negatively
Due to the positive charge of chitosan and TMC, and the negative charge of TT, the antigen loading is driven by the electrostatic attraction. There is also a superior association and internalization of chitosan or chitosan derivatives nanoparticles with cells due to their mucoadhesive character [125-127]. CS nanoparticles have a higher association and internalisation with gastrointestinal tissue cells due to electrostatic interactions compared to polystyrene nanoparticles [125].More recently, hybrid nanoparticles based on the complexation between oppositely charged chitosan derivatives (positively charged TMC and negatively charged MCC) without using any organic solvent or crosslinker yielded results
50 to 400 nm were obtained by the dropwise addition of dilute chitosan solutions [119]. Carboxymethylcellulose and alginate have also been complexed with chitosan to prepare nanoparticles. Chitosan–carboxymethylcellulose nanoparticles were subsequently coated with plasmid DNA (pDNA) [120]. Chitosan–alginate nanoparticles were loaded with insulin [121]. Nanoparticles have also been obtained in which the polyanion is the active principle itself as for example heparin or DNA or even siRNA. Chitosan–heparin nanoparticles crosslinked with TPP were described [122]. Figure 9 shows on the left a scanning electron micrograph of chitosan– heparin nanoparticles adapted from reference [117] and, on the right, an atomic force micrograph of chitosan-carboxymethylcellulose nanoparticles [123] prepared by complex coacervation without using without using any organic solvent or crosslinker. These coacervates have the interesting property of combining with cisplatin [123] but can also be combined with peptides, proteins and DNA and
proteins and DNA and therefore are potentially useful for vaccine design.
therefore are potentially useful for vaccine design.
20 Immune Response Activation
green fluorescent marker as adapted from [124].
similar to those described for the non hybrid nanoparticles [128].
permission from Elsevier.
**Figure 10.** Uptake of chitosan (a), TMC (b) and MCC nanoparticles loaded with bovine serum albumin and labeled with a green fluorescent marker adapted from reference [124].Reprinted from International Journal of Pharmaceutics, 363/1–2, B. Sayın, S. Somavarapu, X.W. Li, M. Thanou, D. Sesardic, H.O.Alpar, S. Şenel, Mono-N-carboxymethyl chitosan (MCC) andN-trimethyl chitosan (TMC) nanoparticles for non-invasive vaccine delivery, 139-148. Copyright 2008, with permission from Elsevier.
charged MCC) without using any organic solvent or crosslinker yielded results similar to those described for the non hybrid nanoparticles [128].
A very promising approach is the development of vaccines based on the initiation of immune response by transfecting dendritic cells (DC) with DNA encoding tumor-associated antigens or immunostimulatory molecules such as cytokines or chemokines [129].Although unmodi‐ fied chitosan may not be a good gene delivery carrier for DCs because of its low transfection efficiency, some modified chitosans showed improved behavior in delivering genes into DCs. IL-12 gene was delivered to DCs *in vivo* using mannosylated chitosan (MC) via mannose receptor-mediated endocytosis [130,131]. MC not only has good physicochemical properties and low cytotoxicity, but also transfect DCs with much higher efficiency than do the unmodi‐ fied chitosan particles *in vitro* [130]. *In vivo*, intratumoral injection of MC /plasmid encoding murine IL-12 complex into BALB/c mice bearing CT-26 carcinoma cells clearly suppressed tumor growth and angiogenesis, and significantly induced cell cycle arrest and apoptosis [131]. Mannose-bearing chitosan nanoparticles were also synthesized to entrap complexes of DNA with polyethyleneimine (PEI) and improve the delivery of DNA into antigen-presenting cells (APCs) after intramuscular (i.m.) injection [132]. Compared with the traditional chitosan microspheres, these nanoparticles targeted the DCs (which express a high density of mannose receptors when they are immature) and released their PEI/DNA cargo inside them. After i.m. immunization, the MC/PEI/DNA nanospheres induced significantly enhanced serum anti‐ body and cytotoxic T lymphocyte (CTL) responses in comparison to naked DNA [132]. Figure 11 shows the MC/PEI/DNA nanoparticles and superior cumulative DNA release *in vitro* of these nanoparticles in PBS buffer as adapted from reference [132].
In recent years, DC vaccines, especially DNA-based DC vaccines, have been the focus of attention in immunotherapy against cancer [129]. Genetically engineered DCs previously modified *in vitro* can then be implanted *ex vivo* and activate the tumor-specific CTL response for the killing of cancer cells [129]. Most of these immunotherapy approaches based on DCs genetic modification for the treatment of melanoma [133,134], renal carcinoma [135] and other malignant diseases [136] are in phase I/II clinical trials. However, in many of these clinical trials adenovirus vector is the carrier still employed for DCs modification [137].Since the major
shows the MC/PEI/DNA nanoparticles and superior cumulative DNA release *in vitro* of these nanoparticles in PBS buffer
DNA, the NOD-like receptor 2 ligand muramyl dipeptide (MDP) and the GM1 ganglioside receptor ligand, cholera toxin B subunit (CTB). The TMC/OVA/immunopotentiator nanopar‐ ticles were characterised regarding their immunogenicity by determining the serum IgG, IgG1, IgG2a titres and secretory IgA levels in nasal washes after intradermal and nasal vaccination in mice. After nasal vaccination, TMC/OVA nanoparticles containing LPS or MDP elicited higher IgG, IgG1 and sIgA levels than non-adjuvanted TMC/OVA particles, whereas nano‐ particles containing CTB, PAM or CpG did not. After intradermal vaccination, the TMC/OVA/ CpG and TMC/OVA/LPS nanoparticles provoked higher IgG titres than plain TMC/OVA particles [141]. Co-encapsulation of the antigen and an additional immunopotentiator into
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 23
Nanoparticles of Fe3O4 coated with glutamic acid plus PEI were used to encapsulate DNA encoding a potent antigen of *Mycobacterium tuberculosis* (Ag85A-ESAT-6) plus IL-21 generating a strong immune response and marked growth inhibition of *M. tuberculosis* in mice [142]. More importantly, compared with using DNA vaccine Ag85A-ESAT-6-IL-21 alone, the nanoparticlebased DNA vaccine Ag85A-ESAT-6-IL-21 showed a statistically significant increase in the protective efficacy against *M. tuberculosis* infection in the immunized mice [142]. The expres‐ sion of both the fusion protein of Ag85A-ESAT-6 and secreted IL-21 protein was confirmed by western blot [142]. Cationic poly(lactide-co-glycolide) (PLGA) nanoparticles were obtained by coating the PLGA nanoparticles with chitosan and used as an intranasal delivery vehicle as a means of administering foot and mouth disease virus (FMDV) DNA vaccine encoding the FMDV capsid protein and the bovine IL-6 gene as a means of enhancing mucosal and systemic immune responses in animals [143]. Guinea pigs and rats were intranasally vaccinated with the respective chitosan-coated PLGA nano/microparticles-loaded FMDV DNA vaccine
formulations; the IL-6 gene was important as an additional immunoestimulator [143].
Cationized gelatin nanoparticles (GNPs) were used as carriers to improve delivery of immu‐ nostimulatory CpG oligonucleotides (CpG ODN) both *in vitro* and *in vivo* [144]. Gelatin nanoparticles were prepared according to the two step desolvation method including further optimization [145].Briefly, 1.25 g gelatin was dissolved in 25 ml highly purified water (5%w/w) at 50oC under constant stirring (700 rpm). The first desolvation step was initiated by quick addition of 25 ml acetone. After discarding the supernatant containing low molecular weight fractions of gelatin, the sediment was redissolved in 25 ml water under constant stirring at 50 oC. Depending on the desired particle size, the pH was then adjusted to a value between 2.3 (approx.resulting particle size: 150 nm) and 3.0 (approx. resulting particle size: 300 nm).Subsequently, the second desolvation step was initiated by drop-wise addition of 50 ml acetone (during constant stirring at 700 rpm) and resulted in the formation of nanoparticles. The nanoparticles were stabilized by cross-linking with glutaraldehyde (150 ml of a 25% solution) under stirring. After 12 h the particles were purified from non-reacted glutaralde‐ hyde by centrifugation and redispersion in water. Cationization of the nanoparticles was achieved through introduction of a quaternary amino group by covalent coupling of chola‐ minechloride hydrochloride onto the particles surface.The delivery by cationic gelatin nanoparticles strongly increased the imunostimulatory effects of CpG oligonucleotides [144].
TMC nanoparticles improved the immunogenicity of the vaccine [141].
A very promising approach is the development of vaccines based on the initiation of immune response by transfecting dendritic cells (DC) with DNA encoding tumor-associated antigens or immunostimulatory molecules such as cytokines or chemokines [129].Although unmodified chitosan may not be a good gene delivery carrier for DCs because of its low transfection efficiency, some modified chitosans showed improved behavior in delivering genes into DCs. IL-12 gene was delivered to DCs *in vivo* using mannosylated chitosan (MC) via mannose receptor-mediated endocytosis [130,131]. MC not only has good physicochemical properties and low cytotoxicity, but also transfect DCs with much higher efficiency than do the unmodified chitosan particles *in vitro* [130]. *In vivo*, intratumoral injection of MC /plasmid encoding murine IL-12 complex into BALB/c mice bearing CT-26 carcinoma cells clearly suppressed tumor growth and angiogenesis, and significantly induced cell cycle arrest and apoptosis [131]. Mannose-bearing chitosan nanoparticles were also synthesized to entrap complexes of DNA with polyethyleneimine (PEI) and improve the delivery of DNA into antigenpresenting cells (APCs) after intramuscular (i.m.) injection [132]. Compared with the traditional chitosan microspheres, these nanoparticles targeted the DCs (which express a high density of mannose receptors when they are immature) and
determined from 10 mg of nanoparticles suspended in 1 ml of PBS at pH 7.4 (120mM NaCl, 2.7mM KCl, 10mM PBS) at 37 °C under stirring. At predetermined time intervals, the suspension was centrifuged and replaced with the same volume of fresh medium. The DNA concentration in the supernatant was determined by absorbance at 260nm.Adapted with permission from reference [132]. Reprinted from Journal of Controlled Release, 121 /3, Xianfeng Zhou, Bin Liu, Xianghui Yu, Xizhen Zhang, Yu Chen, Xueyun Wang, Yinghuan Jin, Yongge Wu, Yue Chen, Yaming Shan, Yan Chen, Junqiu Liu, Wei Kong, Jiacong Shen, Controlled release of PEI/DNA complex from mannose-bearing chitosan microspheres as a potent delivery system to enhance immune response to HBV DNA vaccine, 200-207. Copyright 2007, with permission from Elsevier. **Figure 11.** Atomic force micrograph of MC nanoparticles. Release profiles of chitosan and MC nanoparticles as deter‐ mined from 10 mg of nanoparticles suspended in 1 ml of PBS at pH 7.4 (120mM NaCl, 2.7mM KCl, 10mM PBS) at 37 °C under stirring. At predetermined time intervals, the suspension was centrifuged and replaced with the same volume of fresh medium. The DNA concentration in the supernatant was determined by absorbance at 260nm.Adapted with permission from reference [132]. Reprinted from Journal of Controlled Release, 121 /3, Xianfeng Zhou, Bin Liu, Xian‐ ghui Yu, Xizhen Zhang, Yu Chen, Xueyun Wang, Yinghuan Jin, Yongge Wu, Yue Chen, Yaming Shan, Yan Chen, Jun‐ qiu Liu, Wei Kong, Jiacong Shen, Controlled release of PEI/DNA complex from mannose-bearing chitosan microspheres as a potent delivery system to enhance immune response to HBV DNA vaccine, 200-207. Copyright 2007, with permission from Elsevier.
In recent years, DC vaccines, especially DNA-based DC vaccines, have been the focus of attention in immunotherapy
Figure 11.Atomic force micrograph of MC nanoparticles. Release profiles of chitosan and MC nanoparticles as
barrier for the use of nonviral carriers is their low transfer efficiency compared with viral vectors, the design of novel and efficient nonviral carriers remains a most warranted area of research. Recently, CpG was combined with chitosan in order to enhance the Th1 and Th17 cell polarizing cytokines representing a convenient strategy to elicit cell-mediated immunity [138].Alum, aluminium phosphate and calcium phosphate adjuvants, as well as biodegradable PLG microparticles, inhibited the secretion of bioactive IL-12 by DCs.Thus many of the most widely used adjuvants in preclinical and clinical use exert a specific inhibitory effect on the secretion of IL-12 by DCs. In contrast, chitosan does not inhibit IL-12 secretion and CpGchitosan is a potent stimulus for the induction of pro-inflammatory T-cell responses [138]. CpG-chitosan [138] behaved similarly to DODAB BF/CpG [98]. against cancer [129]. Genetically engineered DCs previously modified *in vitro* can then be implanted *ex vivo* and activate the tumor-specific CTL response for the killing of cancer cells [129]. Most of these immunotherapy approaches based on DCs genetic modification for the treatment of melanoma [133,134], renal carcinoma [135] and other malignant diseases [136] are in phase I/II clinical trials. However, in many of these clinical trials adenovirus vector is the carrier still employed for DCs modification [137].Since the major barrier for the use of nonviral carriers is their low transfer efficiency compared with viral vectors, the design of novel and efficient nonviral carriers remains a most warranted area of research. Recently, CpG was combined with chitosan in order to enhance the Th1 and Th17-cell polarizing cytokines representing a convenient strategy to elicit cell-mediated immunity [138].Alum, aluminium phosphate and calcium phosphate adjuvants, as well as biodegradable PLG microparticles, inhibited the secretion of bioactive IL-12 by DCs.Thus many of the most widely used adjuvants in preclinical and clinical use exert a specific inhibitory effect on the secretion of IL-12 by DCs. In contrast, chitosan does not inhibit IL-12 secretion and CpG-chitosan is a potent stimulus for the induction of proinflammatory T-cell responses [138]. CpG-chitosan [138] behaved similarly to DODAB BF/CpG [98]. The design of DC-based vaccines depends strongly on the generation of effective DCs, so that technologies are needed
The design of DC-based vaccines depends strongly on the generation of effective DCs, so that technologies are needed that produce high antigen expression as a result of delivering DNA encoding antigen into the nucleus of DCs. Barriers to gene delivery into cells start with cell adhesion and uptake, escape from endosomes to the cytoplasm prior to fusion with the lysosomes, trafficking to the nucleus and uptake by the nucleus [139].In particular, large foreign molecules such as proteins and genes are mostly unable to enter the nucleus of nondividing cells. An interesting strategy has been the combination of cationic nanoparticles with a nuclear localization signal (NLS) which enhanced pDNA delivery to the nucleus of DCs when combined with electroporation [140]. that produce high antigen expression as a result of delivering DNA encoding antigen into the nucleus of DCs. Barriers to gene delivery into cells start with cell adhesion and uptake, escape from endosomes to the cytoplasm prior to fusion with the lysosomes, trafficking to the nucleus and uptake by the nucleus [139].In particular, large foreign molecules such as proteins and genes are mostly unable to enter the nucleus of non-dividing cells. An interesting strategy has been the combination of cationic nanoparticles with a nuclear localization signal (NLS) which enhanced pDNA delivery to the nucleus of DCs when combined with electroporation [140]. There is a consensus in the literature that additional immunostimulation provided by LPS, CpG and other molecules are required to improve the Th1 response of cationic chitosan or quaternized trimethylchitosan (TMC) [138,141].These nanoparticles containing ovalbumin as a model antigen (TMC/OVA nanoparticles) and an immunopotentiator (TMC/OVA/immunopotentiator nanoparticles) were recently evaluated [141]. The selection of immunopotentiators included Tolllike receptor (TLR) ligands lipopolysaccharide (LPS), PAM3CSK4 (PAM), CpG DNA, the NOD-like receptor 2 ligand muramyl dipeptide (MDP) and the GM1 ganglioside receptor ligand, cholera toxin B subunit (CTB). The
There is a consensus in the literature that additional immunostimulation provided by LPS, CpG and other molecules are required to improve the Th1 response of cationic chitosan or quaternized trimethylchitosan (TMC) [138,141].These nanoparticles containing ovalbumin as a model antigen (TMC/OVA nanoparticles) and an immunopotentiator (TMC/OVA/immuno‐ potentiator nanoparticles) were recently evaluated [141]. The selection of immunopotentiators included Tolllike receptor (TLR) ligands lipopolysaccharide (LPS), PAM3CSK4 (PAM), CpG TMC/OVA/immunopotentiator nanoparticles were characterised regarding their immunogenicity by determining the serum IgG, IgG1, IgG2a titres and secretory IgA levels in nasal washes after intradermal and nasal vaccination in mice. After nasal vaccination, TMC/OVA nanoparticles containing LPS or MDP elicited higher IgG, IgG1 and sIgA levels than DNA, the NOD-like receptor 2 ligand muramyl dipeptide (MDP) and the GM1 ganglioside receptor ligand, cholera toxin B subunit (CTB). The TMC/OVA/immunopotentiator nanopar‐ ticles were characterised regarding their immunogenicity by determining the serum IgG, IgG1, IgG2a titres and secretory IgA levels in nasal washes after intradermal and nasal vaccination in mice. After nasal vaccination, TMC/OVA nanoparticles containing LPS or MDP elicited higher IgG, IgG1 and sIgA levels than non-adjuvanted TMC/OVA particles, whereas nano‐ particles containing CTB, PAM or CpG did not. After intradermal vaccination, the TMC/OVA/ CpG and TMC/OVA/LPS nanoparticles provoked higher IgG titres than plain TMC/OVA particles [141]. Co-encapsulation of the antigen and an additional immunopotentiator into TMC nanoparticles improved the immunogenicity of the vaccine [141].
Nanoparticles of Fe3O4 coated with glutamic acid plus PEI were used to encapsulate DNA encoding a potent antigen of *Mycobacterium tuberculosis* (Ag85A-ESAT-6) plus IL-21 generating a strong immune response and marked growth inhibition of *M. tuberculosis* in mice [142]. More importantly, compared with using DNA vaccine Ag85A-ESAT-6-IL-21 alone, the nanoparticlebased DNA vaccine Ag85A-ESAT-6-IL-21 showed a statistically significant increase in the protective efficacy against *M. tuberculosis* infection in the immunized mice [142]. The expres‐ sion of both the fusion protein of Ag85A-ESAT-6 and secreted IL-21 protein was confirmed by western blot [142]. Cationic poly(lactide-co-glycolide) (PLGA) nanoparticles were obtained by coating the PLGA nanoparticles with chitosan and used as an intranasal delivery vehicle as a means of administering foot and mouth disease virus (FMDV) DNA vaccine encoding the FMDV capsid protein and the bovine IL-6 gene as a means of enhancing mucosal and systemic immune responses in animals [143]. Guinea pigs and rats were intranasally vaccinated with the respective chitosan-coated PLGA nano/microparticles-loaded FMDV DNA vaccine formulations; the IL-6 gene was important as an additional immunoestimulator [143].
barrier for the use of nonviral carriers is their low transfer efficiency compared with viral vectors, the design of novel and efficient nonviral carriers remains a most warranted area of research. Recently, CpG was combined with chitosan in order to enhance the Th1 and Th17 cell polarizing cytokines representing a convenient strategy to elicit cell-mediated immunity [138].Alum, aluminium phosphate and calcium phosphate adjuvants, as well as biodegradable PLG microparticles, inhibited the secretion of bioactive IL-12 by DCs.Thus many of the most widely used adjuvants in preclinical and clinical use exert a specific inhibitory effect on the secretion of IL-12 by DCs. In contrast, chitosan does not inhibit IL-12 secretion and CpGchitosan is a potent stimulus for the induction of pro-inflammatory T-cell responses [138].
In recent years, DC vaccines, especially DNA-based DC vaccines, have been the focus of attention in immunotherapy against cancer [129]. Genetically engineered DCs previously modified *in vitro* can then be implanted *ex vivo* and activate the tumor-specific CTL response for the killing of cancer cells [129]. Most of these immunotherapy approaches based on DCs genetic modification for the treatment of melanoma [133,134], renal carcinoma [135] and other malignant diseases [136] are in phase I/II clinical trials. However, in many of these clinical trials adenovirus vector is the carrier still employed for DCs modification [137].Since the major barrier for the use of nonviral carriers is their low transfer efficiency compared with viral vectors, the design of novel and efficient nonviral carriers remains a most warranted area of research. Recently, CpG was combined with chitosan in order to enhance the Th1 and Th17-cell polarizing cytokines representing a convenient strategy to elicit cell-mediated immunity [138].Alum, aluminium phosphate and calcium phosphate adjuvants, as well as biodegradable PLG microparticles, inhibited the secretion of bioactive IL-12 by DCs.Thus many of the most widely used adjuvants in preclinical and clinical use exert a specific inhibitory effect on the secretion of IL-12 by DCs. In contrast, chitosan does not inhibit IL-12 secretion and CpG-chitosan is a potent stimulus for the induction of pro-
Figure 11.Atomic force micrograph of MC nanoparticles. Release profiles of chitosan and MC nanoparticles as determined from 10 mg of nanoparticles suspended in 1 ml of PBS at pH 7.4 (120mM NaCl, 2.7mM KCl, 10mM PBS) at 37 °C under stirring. At predetermined time intervals, the suspension was centrifuged and replaced with the same volume of fresh medium. The DNA concentration in the supernatant was determined by absorbance at 260nm.Adapted with permission from reference [132]. Reprinted from Journal of Controlled Release, 121 /3, Xianfeng Zhou, Bin Liu, Xianghui Yu, Xizhen Zhang, Yu Chen, Xueyun Wang, Yinghuan Jin, Yongge Wu, Yue Chen, Yaming Shan, Yan Chen, Junqiu Liu, Wei Kong, Jiacong Shen, Controlled release of PEI/DNA complex from mannose-bearing chitosan microspheres as a potent delivery system to enhance immune response to HBV DNA vaccine, 200-207. Copyright 2007,
**Figure 11.** Atomic force micrograph of MC nanoparticles. Release profiles of chitosan and MC nanoparticles as deter‐ mined from 10 mg of nanoparticles suspended in 1 ml of PBS at pH 7.4 (120mM NaCl, 2.7mM KCl, 10mM PBS) at 37 °C under stirring. At predetermined time intervals, the suspension was centrifuged and replaced with the same volume of fresh medium. The DNA concentration in the supernatant was determined by absorbance at 260nm.Adapted with permission from reference [132]. Reprinted from Journal of Controlled Release, 121 /3, Xianfeng Zhou, Bin Liu, Xian‐ ghui Yu, Xizhen Zhang, Yu Chen, Xueyun Wang, Yinghuan Jin, Yongge Wu, Yue Chen, Yaming Shan, Yan Chen, Jun‐ qiu Liu, Wei Kong, Jiacong Shen, Controlled release of PEI/DNA complex from mannose-bearing chitosan microspheres as a potent delivery system to enhance immune response to HBV DNA vaccine, 200-207. Copyright
A very promising approach is the development of vaccines based on the initiation of immune response by transfecting dendritic cells (DC) with DNA encoding tumor-associated antigens or immunostimulatory molecules such as cytokines or chemokines [129].Although unmodified chitosan may not be a good gene delivery carrier for DCs because of its low transfection efficiency, some modified chitosans showed improved behavior in delivering genes into DCs. IL-12 gene was delivered to DCs *in vivo* using mannosylated chitosan (MC) via mannose receptor-mediated endocytosis [130,131]. MC not only has good physicochemical properties and low cytotoxicity, but also transfect DCs with much higher efficiency than do the unmodified chitosan particles *in vitro* [130]. *In vivo*, intratumoral injection of MC /plasmid encoding murine IL-12 complex into BALB/c mice bearing CT-26 carcinoma cells clearly suppressed tumor growth and angiogenesis, and significantly induced cell cycle arrest and apoptosis [131]. Mannose-bearing chitosan nanoparticles were also synthesized to entrap complexes of DNA with polyethyleneimine (PEI) and improve the delivery of DNA into antigenpresenting cells (APCs) after intramuscular (i.m.) injection [132]. Compared with the traditional chitosan microspheres, these nanoparticles targeted the DCs (which express a high density of mannose receptors when they are immature) and released their PEI/DNA cargo inside them. After i.m. immunization, the MC/PEI/DNA nanospheres induced significantly enhanced serum antibody and cytotoxic T lymphocyte (CTL) responses in comparison to naked DNA [132]. Figure 11 shows the MC/PEI/DNA nanoparticles and superior cumulative DNA release *in vitro* of these nanoparticles in PBS buffer
The design of DC-based vaccines depends strongly on the generation of effective DCs, so that technologies are needed that produce high antigen expression as a result of delivering DNA encoding antigen into the nucleus of DCs. Barriers to gene delivery into cells start with cell adhesion and uptake, escape from endosomes to the cytoplasm prior to fusion with the lysosomes, trafficking to the nucleus and uptake by the nucleus [139].In particular, large foreign molecules such as proteins and genes are mostly unable to enter the nucleus of nondividing cells. An interesting strategy has been the combination of cationic nanoparticles with a nuclear localization signal (NLS) which enhanced pDNA delivery to the nucleus of DCs when
There is a consensus in the literature that additional immunostimulation provided by LPS, CpG and other molecules are required to improve the Th1 response of cationic chitosan or quaternized trimethylchitosan (TMC) [138,141].These nanoparticles containing ovalbumin as a model antigen (TMC/OVA nanoparticles) and an immunopotentiator (TMC/OVA/immunopotentiator nanoparticles) were recently evaluated [141]. The selection of immunopotentiators included Tolllike receptor (TLR) ligands lipopolysaccharide (LPS), PAM3CSK4 (PAM), CpG DNA, the NOD-like receptor 2 ligand muramyl dipeptide (MDP) and the GM1 ganglioside receptor ligand, cholera toxin B subunit (CTB). The TMC/OVA/immunopotentiator nanoparticles were characterised regarding their immunogenicity by determining the serum IgG, IgG1, IgG2a titres and secretory IgA levels in nasal washes after intradermal and nasal vaccination in mice. After nasal vaccination, TMC/OVA nanoparticles containing LPS or MDP elicited higher IgG, IgG1 and sIgA levels than
The design of DC-based vaccines depends strongly on the generation of effective DCs, so that technologies are needed that produce high antigen expression as a result of delivering DNA encoding antigen into the nucleus of DCs. Barriers to gene delivery into cells start with cell adhesion and uptake, escape from endosomes to the cytoplasm prior to fusion with the lysosomes, trafficking to the nucleus and uptake by the nucleus [139].In particular, large foreign molecules such as proteins and genes are mostly unable to enter the nucleus of non-dividing cells. An interesting strategy has been the combination of cationic nanoparticles with a nuclear localization signal (NLS) which enhanced pDNA delivery to the
There is a consensus in the literature that additional immunostimulation provided by LPS, CpG and other molecules are required to improve the Th1 response of cationic chitosan or quaternized trimethylchitosan (TMC) [138,141].These nanoparticles containing ovalbumin as a model antigen (TMC/OVA nanoparticles) and an immunopotentiator (TMC/OVA/immuno‐ potentiator nanoparticles) were recently evaluated [141]. The selection of immunopotentiators included Tolllike receptor (TLR) ligands lipopolysaccharide (LPS), PAM3CSK4 (PAM), CpG
CpG-chitosan [138] behaved similarly to DODAB BF/CpG [98].
inflammatory T-cell responses [138]. CpG-chitosan [138] behaved similarly to DODAB BF/CpG [98].
combined with electroporation [140].
nucleus of DCs when combined with electroporation [140].
as adapted from reference [132].
22 Immune Response Activation
with permission from Elsevier.
2007, with permission from Elsevier.
Cationized gelatin nanoparticles (GNPs) were used as carriers to improve delivery of immu‐ nostimulatory CpG oligonucleotides (CpG ODN) both *in vitro* and *in vivo* [144]. Gelatin nanoparticles were prepared according to the two step desolvation method including further optimization [145].Briefly, 1.25 g gelatin was dissolved in 25 ml highly purified water (5%w/w) at 50oC under constant stirring (700 rpm). The first desolvation step was initiated by quick addition of 25 ml acetone. After discarding the supernatant containing low molecular weight fractions of gelatin, the sediment was redissolved in 25 ml water under constant stirring at 50 oC. Depending on the desired particle size, the pH was then adjusted to a value between 2.3 (approx.resulting particle size: 150 nm) and 3.0 (approx. resulting particle size: 300 nm).Subsequently, the second desolvation step was initiated by drop-wise addition of 50 ml acetone (during constant stirring at 700 rpm) and resulted in the formation of nanoparticles. The nanoparticles were stabilized by cross-linking with glutaraldehyde (150 ml of a 25% solution) under stirring. After 12 h the particles were purified from non-reacted glutaralde‐ hyde by centrifugation and redispersion in water. Cationization of the nanoparticles was achieved through introduction of a quaternary amino group by covalent coupling of chola‐ minechloride hydrochloride onto the particles surface.The delivery by cationic gelatin nanoparticles strongly increased the imunostimulatory effects of CpG oligonucleotides [144]. The effects of CpG are not restricted to modulation of immune responses in vaccine adjuvants. CpG itself has been considered for the treatment of asthma and other allergic diseases [62]. CpG oligonucleotides (CpG–ODN, resembling bacterial DNA) engage TLR-9 on B-cells, dendritic cells and other cell types, resulting in a cascade that includes induction of Th1-type and T-regulatory-type immune responses. Preclinical models of asthma have demonstrated that CpG–ODN are potent inhibitors of atopic responses, suppressing Th2 cytokine and, reducing airway eosinophilia, systemic levels of IgE, and bronchial hyperreactivity-in short the critical attributes of the asthmatic phenotype. In models of chronic allergen exposure, CpG– ODN are also effective at preventing the development of airway remodeling. In established asthma, CpG–ODN can reverse manifestations of disease, both when used alone or in combination with allergen immunotherapy. Early clinical trials have had mixed results, including a significant benefit when CpG–ODN were conjugated to ragweed allergen in an allergic rhinitis immunotherapy study, but only limited efficacy seen when administered prior to allergen challenge in asthmatics [145]. Further study of CpG–ODNs for the treatment of asthma and other atopic disorders is warranted by existing data.Figure 12 summarizes the cellular trafficking of CpG [62].
Polymer-based nanoscale assemblies based on cationic polymers have been intensively searched for biomedical applications [146-148]. Poly(2-aminoethyl methacrylate) (PAEM) cationic homopolymers with defined chain length and narrow molecular weight distribution were synthesized using atom transfer radical polymerization (ATRP), and PAEM/plasmid DNA polyplexes interaction with DCs showed its potential for DNA vaccines [149]. PAEM of different chain lengths (45, 75, and 150 repeating units) showed varying strength in condensing plasmid DNA into narrowly dispersed nanoparticles with very low cytotoxicity. Longer polymer chain length resulted in higher levels of overallcellular uptake and nuclear uptake of plasmid DNA, but shorter polymer chains favored intracellular and intranuclear release of free plasmid from the polyplexes [149]. When a model antigen-encoding ovalbumin plasmid was used, transfected DCs stimulated the activation of naïve CD8 (+) T cells to produce high levels of interferon-γ. The efficiency of transfection, DC maturation, and CD8 (+) T cell activation showed varying degrees of polymer chain-length dependence [149].These struc‐ turally defined cationic polymers may have much potential as efficient DNA vaccine carriers
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 25
Other strategies involve the production of hybrid nanoparticles. For example, biocompatible polymeric particles of poly caprolactone were coated with chitosan for effective immunization against influenza using recombinant Influenza A virus (A/California/07/2009) H1N1 hemag‐ glutinin (HA) protein, for the induction of humoral, cellular and mucosal immunity [150]. CS-PCL nanoparticles (cationic nanoparticles) produced humoral (both systemic and mucosal) and cellular immune responses upon nasal administration demonstrating high potential for their use as a carrier adjuvant for nasal administered influenza antigens [150]. Other example is represented by biocompatible core-shell cationic nanoparticles, composed of an inner hard core of poly(methylmethacrylate) (PMMA) and a hydrophilic tentacular shell bearing posi‐ tively charged groups and poly(ethyleneglycol) chains covalently bound to the core prepared by emulsion polymerization and characterized *in vitro* and *in vivo* for DNA vaccine applica‐ tions [151]. The nanoparticles reversibly adsorbed large amounts of DNA, mainly through electrostatic interactions, preserved its functional structure, efficiently delivered it intracellu‐ larly, and were not toxic *in vitro* or *in mice* [151]. Furthermore, two intramuscular (i.m.) immunizations (4 weeks apart) with a very low dose (1 microg) of the plasmid pCV-tat delivered by these nanoparticles followed by one or two protein boosts induced significant antigen-specific humoral and cellular responses and greatly increased Th1-type T cell re‐
Chitosan-based DNA vaccines against Japanese encephalitis virus using transdermal immu‐ nization were prepared by using the associative feature of cationic chitosan and anionic DNAs which formed nanoscale complexes [152]. The expression of green fluorescent protein (GFP) reporter gene was observable and traceable in epidermis and spleen over 3 days. The expres‐ sions of GFP and the activation of dendritic cells (DCs) were evident and co-localized in hair follicles and epidermis. Mice immunized with the developed chitosan-JEV DNA vaccines elicited desired JEV specific antibodies, whereby the mice maintained high survival rates against 50xLD(50) JEV challenge. The low-pressure-gene-gun mediated chitosan-based JEV
and immunoadjuvants.
sponses and CTLs against HIV-1 Tat [151].
**Figure 12.** CpG–ODNs are bound by CpG binding protein at the cell surface. The oligonucleotide then undergoes en‐ docytosis where it is bound by TLR9 in the endosome. TLR9 then interacts with MyD88 and the IKK complex and is translocated to the cell nucleus. Once in the cell nucleus, the CpG ODNs activate NF-kB and initiate the induction of cytokines and chemokines.Reproduced with permission from reference [62].Reprinted from Advanced Drug Delivery Reviews, 61 / 3, David E. Fonseca, Joel N. Kline, Use of CpG oligonucleotides in treatment of asthma and allergic dis‐ ease, 256-272. Copyright 2009, with permission from Elsevier.
Polymer-based nanoscale assemblies based on cationic polymers have been intensively searched for biomedical applications [146-148]. Poly(2-aminoethyl methacrylate) (PAEM) cationic homopolymers with defined chain length and narrow molecular weight distribution were synthesized using atom transfer radical polymerization (ATRP), and PAEM/plasmid DNA polyplexes interaction with DCs showed its potential for DNA vaccines [149]. PAEM of different chain lengths (45, 75, and 150 repeating units) showed varying strength in condensing plasmid DNA into narrowly dispersed nanoparticles with very low cytotoxicity. Longer polymer chain length resulted in higher levels of overallcellular uptake and nuclear uptake of plasmid DNA, but shorter polymer chains favored intracellular and intranuclear release of free plasmid from the polyplexes [149]. When a model antigen-encoding ovalbumin plasmid was used, transfected DCs stimulated the activation of naïve CD8 (+) T cells to produce high levels of interferon-γ. The efficiency of transfection, DC maturation, and CD8 (+) T cell activation showed varying degrees of polymer chain-length dependence [149].These struc‐ turally defined cationic polymers may have much potential as efficient DNA vaccine carriers and immunoadjuvants.
The effects of CpG are not restricted to modulation of immune responses in vaccine adjuvants. CpG itself has been considered for the treatment of asthma and other allergic diseases [62]. CpG oligonucleotides (CpG–ODN, resembling bacterial DNA) engage TLR-9 on B-cells, dendritic cells and other cell types, resulting in a cascade that includes induction of Th1-type and T-regulatory-type immune responses. Preclinical models of asthma have demonstrated that CpG–ODN are potent inhibitors of atopic responses, suppressing Th2 cytokine and, reducing airway eosinophilia, systemic levels of IgE, and bronchial hyperreactivity-in short the critical attributes of the asthmatic phenotype. In models of chronic allergen exposure, CpG– ODN are also effective at preventing the development of airway remodeling. In established asthma, CpG–ODN can reverse manifestations of disease, both when used alone or in combination with allergen immunotherapy. Early clinical trials have had mixed results, including a significant benefit when CpG–ODN were conjugated to ragweed allergen in an allergic rhinitis immunotherapy study, but only limited efficacy seen when administered prior to allergen challenge in asthmatics [145]. Further study of CpG–ODNs for the treatment of asthma and other atopic disorders is warranted by existing data.Figure 12 summarizes the
**Figure 12.** CpG–ODNs are bound by CpG binding protein at the cell surface. The oligonucleotide then undergoes en‐ docytosis where it is bound by TLR9 in the endosome. TLR9 then interacts with MyD88 and the IKK complex and is translocated to the cell nucleus. Once in the cell nucleus, the CpG ODNs activate NF-kB and initiate the induction of cytokines and chemokines.Reproduced with permission from reference [62].Reprinted from Advanced Drug Delivery Reviews, 61 / 3, David E. Fonseca, Joel N. Kline, Use of CpG oligonucleotides in treatment of asthma and allergic dis‐
ease, 256-272. Copyright 2009, with permission from Elsevier.
cellular trafficking of CpG [62].
24 Immune Response Activation
Other strategies involve the production of hybrid nanoparticles. For example, biocompatible polymeric particles of poly caprolactone were coated with chitosan for effective immunization against influenza using recombinant Influenza A virus (A/California/07/2009) H1N1 hemag‐ glutinin (HA) protein, for the induction of humoral, cellular and mucosal immunity [150]. CS-PCL nanoparticles (cationic nanoparticles) produced humoral (both systemic and mucosal) and cellular immune responses upon nasal administration demonstrating high potential for their use as a carrier adjuvant for nasal administered influenza antigens [150]. Other example is represented by biocompatible core-shell cationic nanoparticles, composed of an inner hard core of poly(methylmethacrylate) (PMMA) and a hydrophilic tentacular shell bearing posi‐ tively charged groups and poly(ethyleneglycol) chains covalently bound to the core prepared by emulsion polymerization and characterized *in vitro* and *in vivo* for DNA vaccine applica‐ tions [151]. The nanoparticles reversibly adsorbed large amounts of DNA, mainly through electrostatic interactions, preserved its functional structure, efficiently delivered it intracellu‐ larly, and were not toxic *in vitro* or *in mice* [151]. Furthermore, two intramuscular (i.m.) immunizations (4 weeks apart) with a very low dose (1 microg) of the plasmid pCV-tat delivered by these nanoparticles followed by one or two protein boosts induced significant antigen-specific humoral and cellular responses and greatly increased Th1-type T cell re‐ sponses and CTLs against HIV-1 Tat [151].
Chitosan-based DNA vaccines against Japanese encephalitis virus using transdermal immu‐ nization were prepared by using the associative feature of cationic chitosan and anionic DNAs which formed nanoscale complexes [152]. The expression of green fluorescent protein (GFP) reporter gene was observable and traceable in epidermis and spleen over 3 days. The expres‐ sions of GFP and the activation of dendritic cells (DCs) were evident and co-localized in hair follicles and epidermis. Mice immunized with the developed chitosan-JEV DNA vaccines elicited desired JEV specific antibodies, whereby the mice maintained high survival rates against 50xLD(50) JEV challenge. The low-pressure-gene-gun mediated chitosan-based JEV DNA vaccines have proven to be convenient and efficacious, thereby with high capacity in deployment for future prophylaxis against JEV outbreaks [152].
Similarly, cationic maltosylated PEI (mPEI) was electrostatically complexed to a plasmid encoding the human papillomavirus (HPV) type 16L1 protein (pHPV16L1), and further complexed to a maltose binding protein (MBP)-fused human papilloma virus type 16L1 fusion protein (HPV16L1-MBP) [159]. The intracellular co-delivery of protein and plasmid DNA vaccines was significantly higher for mPEI-based triple nanocomplexes than for a simple physical mixture of the proteins and DNA [160]. The high expression levels were comparable to those obtained using dual complexes of mPEI and the plasmid DNA. In vivo, co-immuni‐ zation of mice with HPV16L1-MBP/mPEI/pHPV16L1 nanocomplexes triggered the highest levels of humoral immune responses among various vaccination groups and there was a significant increase in the number of interferon-γ producing CD8 (+) T cells compared with the use of mixed proteins and plasmid DNA. This approach could enhance the immunoge‐
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 27
Platelet-derived growth factor-bb (PDGF-BB) and fibroblast growth factor-2(FGF-2) are known to induce chemotaxis, proliferation, differentiation, and matrix synthesis so that growth factor therapy is an emerging treatment modality that enhances tissue vascularization, promotes healing and regeneration and can treat a variety of inflammatory diseases or promote healing of chronic wounds [160]. Chitosan was combined with PDGF-BB and FGF-2 genes to induce the therapeutic expression of these genes in mice. The percentage of the residues that are glucosamine is called the degree of deacetylation (DDA). PDGF-BB and FGF-2 genes were amplified from human tissues by RT-PCR. To increase the secretion of FGF-2, a recombinant 4sFGF-2 was constructed bearing eight amino-acid residues of the signal peptide of FGF-4. PCR products were inserted into the expression vector pVax1 to produce recombinant plasmids pVax1-4sFGF2 and pVax1-PDGF-BB, which were then injected into BALB/C mice in the format of polyelectrolyte nanocomplexes with specific chitosans of controlled DDA and molecular weight, including 92-10, 80-10, and 80-80(DDA-number average molecular weight or M(n) in kDa) [160]. ELISA assays on mice sera showed that recombinant FGF-2 and PDGF-BB proteins were efficiently expressed and specific antibodies to these proteins could be identified in sera of injected mice, but with levels that were clearly dependent on the specific chitosan used [160]. High DDA low molecular weight chitosans were efficient protein expressors with minimal or no generation of neutralizing antibodies, while lowering DDA resulted in greater antibody levels and correspondingly lower levels of detected recombinant protein. Histological analyses corroborated these results by revealing greater inflammatory infiltrates in lower DDA chitosans, which produced higher antibody titers [160]. The subcu‐ taneous delivery of the plasmids was more efficient than the one by intramuscular injection [160]. Certain chitosans can be efficient, non-toxic and therapeutic protein delivery systems or vectors for DNA vaccines [160]. The recent advances that have led to a more rational design of chitosan polyplexes has recently been reviewed [161]. Recent progress in preparation and characterisation of chitosan based polyplexes has enabled coupling analysis of chitosans structural parameters that has led to increased transfection efficiency by tailoring of chitosan's
structure in accordance with a more rational design of chitosan polyplexes [161].
Genetic vaccination against leishmaniasis has been intensively pursued by Rafati´s group [162-166]. Leishmaniasis is a major health problem in many tropical and sub-tropical countries
nicity of HPV16L1 vaccines [159].
Pullulan, a fungal exopolysaccharide produced from starch is a water soluble, neutral linear biopolymer consisting of α-1,6-linked maltotriose residues. Recently, a cationic cholesteryl group-bearing pullulan (cCHP) derivative was used to obtain self-assembled nanogels for administering an intranasal vaccine containing a non-toxic subunit fragment of *Clostridium botulinum* type-A neurotoxin BoHc/A (cCHP-BoHc/A) which continuously adhered to the nasal epithelium and was effectively taken up by mucosal dendritic cells after its release from the cCHP nanogel. Vigorous botulinum-neurotoxin-A-neutralizing serum IgG and secretory IgA antibody responses were induced without co-administration of mucosal adjuvant. Importantly, intranasally administered cCHP-BoHc/A did not accumulate in the olfactory bulbs or brain. Moreover, intranasally immunized tetanus toxoid with cCHP nanogel induced strong tetanus-toxoid-specific systemic and mucosal immune responses suggesting that cCHP nanogel can be used as a universal protein-based antigen-delivery vehicle for adjuvant-free intranasal vaccination [153].Recently, the same system was used to deliver a vaccine against pneumonia [154]. This pneumococcal vaccine combined the advantages of pneumococcal surface protein A (PspA) with a nontoxic intranasal vaccine delivery system based on a nanometer-sized hydrogel cCHP. The efficacy of the nanogel-based PspA nasal vaccine (cCHP-PspA) was tested in murine pneumococcal airway infection models. Intranasal vaccination with cCHP-PspA provided protective immunity against lethal challenge with Streptococcus pneumoniae Xen10, reduced colonization and invasion by bacteria in the upper and lower respiratory tracts, and induced systemic and nasal mucosal Th17 responses, high levels of PspA-specific serum immunoglobulin G (IgG), and nasal and bronchial IgA antibody re‐ sponses. Moreover, there was no sign of PspA delivery by nanogel to either the olfactory bulbs or the central nervous system after intranasal administration showing the effectiveness and safety of the nanogel-based PspA nasal vaccine system as a universal mucosal vaccine against pneumococcal respiratory infection.
Lipid-biopolymer hybrid systems have long been described for incorporation of pDNA [155].Cationic liposomes mixed with pDNA that had been pre-condensed with a polycation (e.g.protamine) yielded the LPD formulation or lipid-polymer-DNA (DOTAP, protamine, pDNA) [156]. As compared to conventional lipoplexes, the double bilayer of LPD confers greater stability and structural homogeneity to the assemblies [156]. Strong antitumor immunity could be generated when a peptide antigen was incorporated into LPD (cationic liposome-polycation-pDNA) nanoparticles [155, 156]. In another study, both the cationic liposome and DNA were shown to be required for the full immunostimulation activity of LPD [157]. The unique ability of LPD to readily move into local lymphoid tissues and to activate antigen-presenting cells might be responsible for its strong immunostimulatory activity [157]. Moreover, cationic liposome stimulates the expression of CD80/CD86 on dendritic cells (DCs), but not the release of TNF-alpha from DCs, suggesting the existence of a NF-kappa Bindependent immunostimulation pathway for cationic lipids such as DOTAP [157].The LPD complexes were nanoparticles with less than 100 nm of mean diameter due to the compaction of DNA induced by the polycation [158].
Similarly, cationic maltosylated PEI (mPEI) was electrostatically complexed to a plasmid encoding the human papillomavirus (HPV) type 16L1 protein (pHPV16L1), and further complexed to a maltose binding protein (MBP)-fused human papilloma virus type 16L1 fusion protein (HPV16L1-MBP) [159]. The intracellular co-delivery of protein and plasmid DNA vaccines was significantly higher for mPEI-based triple nanocomplexes than for a simple physical mixture of the proteins and DNA [160]. The high expression levels were comparable to those obtained using dual complexes of mPEI and the plasmid DNA. In vivo, co-immuni‐ zation of mice with HPV16L1-MBP/mPEI/pHPV16L1 nanocomplexes triggered the highest levels of humoral immune responses among various vaccination groups and there was a significant increase in the number of interferon-γ producing CD8 (+) T cells compared with the use of mixed proteins and plasmid DNA. This approach could enhance the immunoge‐ nicity of HPV16L1 vaccines [159].
DNA vaccines have proven to be convenient and efficacious, thereby with high capacity in
Pullulan, a fungal exopolysaccharide produced from starch is a water soluble, neutral linear biopolymer consisting of α-1,6-linked maltotriose residues. Recently, a cationic cholesteryl group-bearing pullulan (cCHP) derivative was used to obtain self-assembled nanogels for administering an intranasal vaccine containing a non-toxic subunit fragment of *Clostridium botulinum* type-A neurotoxin BoHc/A (cCHP-BoHc/A) which continuously adhered to the nasal epithelium and was effectively taken up by mucosal dendritic cells after its release from the cCHP nanogel. Vigorous botulinum-neurotoxin-A-neutralizing serum IgG and secretory IgA antibody responses were induced without co-administration of mucosal adjuvant. Importantly, intranasally administered cCHP-BoHc/A did not accumulate in the olfactory bulbs or brain. Moreover, intranasally immunized tetanus toxoid with cCHP nanogel induced strong tetanus-toxoid-specific systemic and mucosal immune responses suggesting that cCHP nanogel can be used as a universal protein-based antigen-delivery vehicle for adjuvant-free intranasal vaccination [153].Recently, the same system was used to deliver a vaccine against pneumonia [154]. This pneumococcal vaccine combined the advantages of pneumococcal surface protein A (PspA) with a nontoxic intranasal vaccine delivery system based on a nanometer-sized hydrogel cCHP. The efficacy of the nanogel-based PspA nasal vaccine (cCHP-PspA) was tested in murine pneumococcal airway infection models. Intranasal vaccination with cCHP-PspA provided protective immunity against lethal challenge with Streptococcus pneumoniae Xen10, reduced colonization and invasion by bacteria in the upper and lower respiratory tracts, and induced systemic and nasal mucosal Th17 responses, high levels of PspA-specific serum immunoglobulin G (IgG), and nasal and bronchial IgA antibody re‐ sponses. Moreover, there was no sign of PspA delivery by nanogel to either the olfactory bulbs or the central nervous system after intranasal administration showing the effectiveness and safety of the nanogel-based PspA nasal vaccine system as a universal mucosal vaccine against
Lipid-biopolymer hybrid systems have long been described for incorporation of pDNA [155].Cationic liposomes mixed with pDNA that had been pre-condensed with a polycation (e.g.protamine) yielded the LPD formulation or lipid-polymer-DNA (DOTAP, protamine, pDNA) [156]. As compared to conventional lipoplexes, the double bilayer of LPD confers greater stability and structural homogeneity to the assemblies [156]. Strong antitumor immunity could be generated when a peptide antigen was incorporated into LPD (cationic liposome-polycation-pDNA) nanoparticles [155, 156]. In another study, both the cationic liposome and DNA were shown to be required for the full immunostimulation activity of LPD [157]. The unique ability of LPD to readily move into local lymphoid tissues and to activate antigen-presenting cells might be responsible for its strong immunostimulatory activity [157]. Moreover, cationic liposome stimulates the expression of CD80/CD86 on dendritic cells (DCs), but not the release of TNF-alpha from DCs, suggesting the existence of a NF-kappa Bindependent immunostimulation pathway for cationic lipids such as DOTAP [157].The LPD complexes were nanoparticles with less than 100 nm of mean diameter due to the compaction
deployment for future prophylaxis against JEV outbreaks [152].
26 Immune Response Activation
pneumococcal respiratory infection.
of DNA induced by the polycation [158].
Platelet-derived growth factor-bb (PDGF-BB) and fibroblast growth factor-2(FGF-2) are known to induce chemotaxis, proliferation, differentiation, and matrix synthesis so that growth factor therapy is an emerging treatment modality that enhances tissue vascularization, promotes healing and regeneration and can treat a variety of inflammatory diseases or promote healing of chronic wounds [160]. Chitosan was combined with PDGF-BB and FGF-2 genes to induce the therapeutic expression of these genes in mice. The percentage of the residues that are glucosamine is called the degree of deacetylation (DDA). PDGF-BB and FGF-2 genes were amplified from human tissues by RT-PCR. To increase the secretion of FGF-2, a recombinant 4sFGF-2 was constructed bearing eight amino-acid residues of the signal peptide of FGF-4. PCR products were inserted into the expression vector pVax1 to produce recombinant plasmids pVax1-4sFGF2 and pVax1-PDGF-BB, which were then injected into BALB/C mice in the format of polyelectrolyte nanocomplexes with specific chitosans of controlled DDA and molecular weight, including 92-10, 80-10, and 80-80(DDA-number average molecular weight or M(n) in kDa) [160]. ELISA assays on mice sera showed that recombinant FGF-2 and PDGF-BB proteins were efficiently expressed and specific antibodies to these proteins could be identified in sera of injected mice, but with levels that were clearly dependent on the specific chitosan used [160]. High DDA low molecular weight chitosans were efficient protein expressors with minimal or no generation of neutralizing antibodies, while lowering DDA resulted in greater antibody levels and correspondingly lower levels of detected recombinant protein. Histological analyses corroborated these results by revealing greater inflammatory infiltrates in lower DDA chitosans, which produced higher antibody titers [160]. The subcu‐ taneous delivery of the plasmids was more efficient than the one by intramuscular injection [160]. Certain chitosans can be efficient, non-toxic and therapeutic protein delivery systems or vectors for DNA vaccines [160]. The recent advances that have led to a more rational design of chitosan polyplexes has recently been reviewed [161]. Recent progress in preparation and characterisation of chitosan based polyplexes has enabled coupling analysis of chitosans structural parameters that has led to increased transfection efficiency by tailoring of chitosan's structure in accordance with a more rational design of chitosan polyplexes [161].
Genetic vaccination against leishmaniasis has been intensively pursued by Rafati´s group [162-166]. Leishmaniasis is a major health problem in many tropical and sub-tropical countries and development of a safe and easily-available vaccine has high priority. Several antigens potentially capable of inducing protective immunity have been studied like cathepsin L-like cysteine proteinases (CPs) and type Iand II CPs used in a heterologous prime-boost vaccination regime for experimental visceral leishmaniasis in dogs. Due to the promising results of the mentioned vaccination regime, cationic solid lipid nanoparticles (cSLNs) for in vitro delivery of these antigens were used as a cocktail DNA vaccine to deliver immunogenic CP genes efficiently. Briefly, the cationic lipid DOTAP (0.4% w/v) was dissolved in hot aqueous phase which was then added to the melted cetyl palmitate and cholesterol (5.1%w/v) phase contain‐ ing tween 80 as a nonionic surfactant at 3.2:1molar ratio. Emulsification was carried out by stirring the mixture at 2000 rpm for 10 min at 90o C. Samples were then homogenized using a high shear homogenizer at 18,000 g for 15 min and the cSLN dispersion was obtained by direct cooling of hot O/W microemulsion on an ice-bath while stirring at 1000 rpm. The cSLNs were washed by centrifugation (6000 g, 10 min, three times) using 100 kDa ultra centrifugal filters to purify the suspension from the excess of surfactant. Efficiency/cytotoxicity ratio of cSLNpDNAs formulations was comparable to linear PEI-25KD-pDNAs polyplexes while exhibiting significantly lower cytotoxicity [162].In addition, C-terminal domain deletion in CPs enhanced the protective activity of cpa/cpb loaded solid lipid nanoparticles against *Leishmania major* in BALB/c mice [163]. cSLN's were able to boost immuneresponse magnitude of cpa/cpb(-CTE) cocktail vaccination against leishmaniasis so that the average parasite inhibition percent could be increased significantly [163].The cSLNs were used to formulate three pDNAs encoding *L. major* cysteine proteinase type I (cpa), II (cpb) and III (cpc)\[164]. BALB/c mice were immu‐ nized twice with a 3-week interval, with SLN-pcDNA-cpa/b/c, pcDNA-cpa/b/c, SLN, SLNpcDNA and PBS.Footpad assessments, parasite burden, cytokine and antibody responses were evaluated. Mice vaccinated with SLN-pcDNA-cpa/b/c significantly (p<0.05) showed higher protection levels with specific Th1 immune response development compared to other groups [164]. This was the first report demonstrating cSLNs as a nanoscale vehicle boosting immune response quality and quantity [164].
Recently, the high-pressure extrusion method was used to obtain cationic liposomes entrap‐ ping pDNA over a range of sizes [168]. This is a well-known sizing method for liposomes, but had not been applied for liposomes that are already loaded with pDNA. Liposomes composed of egg PC, DOTAP, and DOPE with entrapped pDNA were prepared by the dehydrationrehydration method and subjected to various extrusion cycles, comparing different membrane pore sizes and extrusion frequencies [168]. At optimized extrusion conditions, liposome diameter and polydispersity were reduced from 560 nm and 0.56 to 150 nm and 0.14, respec‐ tively, and 35% of the pDNA was retained [168]. Importantly, gel electrophoresis and trans‐ fection experiments with pDNA extracted from these extruded liposomes demonstrated the preservation of the structural and functional integrity of the pDNA. The reduction in size resulted in enhanced transfection of HeLa cells, as detected by functional expression of the fluorescent protein, eGFP. In addition, these liposomes were able to stimulate Toll-like receptor 9, indicating efficient endosomal uptake and release of the included pDNA. In conclusion, high-pressure extrusion is a suitable technique to size cationic liposomes with entrapped pDNA and allows preparation of well-defined nanosized pDNA-liposomes, with preserved pDNA integrity [168]. Their improved transfection efficiency and ability to activate an important pattern-recognition receptor are favorable properties for DNA vaccine delivery
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 29
The formulation of DNA into both liposomal and polymeric cationic nanoparticles was found to block completely vaccination-induced antigen expression in mice and ex vivo human skin [169].This detrimental effect of cationic nanoparticle formulation was explained by immobili‐ zation of the nanoparticles in the extracellular matrix, caused by electrostatic interactions of the cationic nanoparticles with negatively charged extracellular matrix components [169]. Shielding the surface charge of the nanoparticles by PEGylation improved *in vivo* antigen expression more than 55 fold [169]. Furthermore, this shielding of cationic surface charge resulted in antigen-specific T cell responses that were similar as those induced by naked DNA for the two lipo-and polyplex DNA carrier systems tested [169]. Charge shielding should be
generally applied for the development of intradermal vaccine formulations [169].
Various aspects that could be decisive in the formulation of efficient and stable carrier system(s) for the development of malaria vaccine have recently been reviewed [170]. The capacity of multi subunit DNA vaccine encoding different stage *Plasmodium* antigens to induce CD8+cytotoxic T lymphocytes and interferon-γ responses in mice, monkeys and humans has been observed [170]. Moreover, genetic vaccination may be capable of eliciting both cell mediated and humoral immune responses. The cytotoxic T cell responses are categorically needed against intracellular hepatic stage and humoral response with antibodies targeted against antigens from all stages of malaria parasite life cycle. Therefore, the key to success for any DNA based vaccine is to design a vector able to serve as a safe and efficient delivery system. Also against malaria the development of non-viral DNA-mediated gene transfer techniques based on artificial vectors such as liposomes, virosomes, microspheres and nanoparticles is
vehicles [168].
strongly warranted [170].
Very fundamental questions regarding size and charge of nanoscale cationic assemblies and their role in adjuvanticity have seldom been asked and properly answered [167-169]. The role of surface charge density in cationic liposome-promoted dendritic cell maturation and vaccineinduced immune responses was recently evaluated for a series of DOTAP/DOPC cationic liposomes with different surface densities by incorporating varying amounts of DOPC (a neutrallipid) into DOTAP (a cationic lipid) [167]. The liposomes with a relatively high charge density, such as DOTAP/DOPC 5:0 and 4:1 liposomes, potently enhanced dendriticcell maturation, ROS generation, antigen uptake, as well as the production of OVA-specific IgG2a and IFN-γ [167]. In contrast, low-charge liposomes, such asDOTAP/DOPC 1:4 liposome, failed to promote immune responses even at highconcentrations, confirming that the immunoregu‐ latory effect of cationic liposomes is mostly attributable to their surface charge density. Moreover, the DOTAP/DOPC 1:4 liposome suppressed anti-OVA antibody responses *in vivo*. Overall, maintaining an appropriate surface charge is crucial for optimizing the adjuvant effect of cationic liposomes and enhancing the efficacy of liposome-based vaccines.
Recently, the high-pressure extrusion method was used to obtain cationic liposomes entrap‐ ping pDNA over a range of sizes [168]. This is a well-known sizing method for liposomes, but had not been applied for liposomes that are already loaded with pDNA. Liposomes composed of egg PC, DOTAP, and DOPE with entrapped pDNA were prepared by the dehydrationrehydration method and subjected to various extrusion cycles, comparing different membrane pore sizes and extrusion frequencies [168]. At optimized extrusion conditions, liposome diameter and polydispersity were reduced from 560 nm and 0.56 to 150 nm and 0.14, respec‐ tively, and 35% of the pDNA was retained [168]. Importantly, gel electrophoresis and trans‐ fection experiments with pDNA extracted from these extruded liposomes demonstrated the preservation of the structural and functional integrity of the pDNA. The reduction in size resulted in enhanced transfection of HeLa cells, as detected by functional expression of the fluorescent protein, eGFP. In addition, these liposomes were able to stimulate Toll-like receptor 9, indicating efficient endosomal uptake and release of the included pDNA. In conclusion, high-pressure extrusion is a suitable technique to size cationic liposomes with entrapped pDNA and allows preparation of well-defined nanosized pDNA-liposomes, with preserved pDNA integrity [168]. Their improved transfection efficiency and ability to activate an important pattern-recognition receptor are favorable properties for DNA vaccine delivery vehicles [168].
and development of a safe and easily-available vaccine has high priority. Several antigens potentially capable of inducing protective immunity have been studied like cathepsin L-like cysteine proteinases (CPs) and type Iand II CPs used in a heterologous prime-boost vaccination regime for experimental visceral leishmaniasis in dogs. Due to the promising results of the mentioned vaccination regime, cationic solid lipid nanoparticles (cSLNs) for in vitro delivery of these antigens were used as a cocktail DNA vaccine to deliver immunogenic CP genes efficiently. Briefly, the cationic lipid DOTAP (0.4% w/v) was dissolved in hot aqueous phase which was then added to the melted cetyl palmitate and cholesterol (5.1%w/v) phase contain‐ ing tween 80 as a nonionic surfactant at 3.2:1molar ratio. Emulsification was carried out by
high shear homogenizer at 18,000 g for 15 min and the cSLN dispersion was obtained by direct cooling of hot O/W microemulsion on an ice-bath while stirring at 1000 rpm. The cSLNs were washed by centrifugation (6000 g, 10 min, three times) using 100 kDa ultra centrifugal filters to purify the suspension from the excess of surfactant. Efficiency/cytotoxicity ratio of cSLNpDNAs formulations was comparable to linear PEI-25KD-pDNAs polyplexes while exhibiting significantly lower cytotoxicity [162].In addition, C-terminal domain deletion in CPs enhanced the protective activity of cpa/cpb loaded solid lipid nanoparticles against *Leishmania major* in BALB/c mice [163]. cSLN's were able to boost immuneresponse magnitude of cpa/cpb(-CTE) cocktail vaccination against leishmaniasis so that the average parasite inhibition percent could be increased significantly [163].The cSLNs were used to formulate three pDNAs encoding *L. major* cysteine proteinase type I (cpa), II (cpb) and III (cpc)\[164]. BALB/c mice were immu‐ nized twice with a 3-week interval, with SLN-pcDNA-cpa/b/c, pcDNA-cpa/b/c, SLN, SLNpcDNA and PBS.Footpad assessments, parasite burden, cytokine and antibody responses were evaluated. Mice vaccinated with SLN-pcDNA-cpa/b/c significantly (p<0.05) showed higher protection levels with specific Th1 immune response development compared to other groups [164]. This was the first report demonstrating cSLNs as a nanoscale vehicle boosting immune
Very fundamental questions regarding size and charge of nanoscale cationic assemblies and their role in adjuvanticity have seldom been asked and properly answered [167-169]. The role of surface charge density in cationic liposome-promoted dendritic cell maturation and vaccineinduced immune responses was recently evaluated for a series of DOTAP/DOPC cationic liposomes with different surface densities by incorporating varying amounts of DOPC (a neutrallipid) into DOTAP (a cationic lipid) [167]. The liposomes with a relatively high charge density, such as DOTAP/DOPC 5:0 and 4:1 liposomes, potently enhanced dendriticcell maturation, ROS generation, antigen uptake, as well as the production of OVA-specific IgG2a and IFN-γ [167]. In contrast, low-charge liposomes, such asDOTAP/DOPC 1:4 liposome, failed to promote immune responses even at highconcentrations, confirming that the immunoregu‐ latory effect of cationic liposomes is mostly attributable to their surface charge density. Moreover, the DOTAP/DOPC 1:4 liposome suppressed anti-OVA antibody responses *in vivo*. Overall, maintaining an appropriate surface charge is crucial for optimizing the adjuvant effect
of cationic liposomes and enhancing the efficacy of liposome-based vaccines.
C. Samples were then homogenized using a
stirring the mixture at 2000 rpm for 10 min at 90o
28 Immune Response Activation
response quality and quantity [164].
The formulation of DNA into both liposomal and polymeric cationic nanoparticles was found to block completely vaccination-induced antigen expression in mice and ex vivo human skin [169].This detrimental effect of cationic nanoparticle formulation was explained by immobili‐ zation of the nanoparticles in the extracellular matrix, caused by electrostatic interactions of the cationic nanoparticles with negatively charged extracellular matrix components [169]. Shielding the surface charge of the nanoparticles by PEGylation improved *in vivo* antigen expression more than 55 fold [169]. Furthermore, this shielding of cationic surface charge resulted in antigen-specific T cell responses that were similar as those induced by naked DNA for the two lipo-and polyplex DNA carrier systems tested [169]. Charge shielding should be generally applied for the development of intradermal vaccine formulations [169].
Various aspects that could be decisive in the formulation of efficient and stable carrier system(s) for the development of malaria vaccine have recently been reviewed [170]. The capacity of multi subunit DNA vaccine encoding different stage *Plasmodium* antigens to induce CD8+cytotoxic T lymphocytes and interferon-γ responses in mice, monkeys and humans has been observed [170]. Moreover, genetic vaccination may be capable of eliciting both cell mediated and humoral immune responses. The cytotoxic T cell responses are categorically needed against intracellular hepatic stage and humoral response with antibodies targeted against antigens from all stages of malaria parasite life cycle. Therefore, the key to success for any DNA based vaccine is to design a vector able to serve as a safe and efficient delivery system. Also against malaria the development of non-viral DNA-mediated gene transfer techniques based on artificial vectors such as liposomes, virosomes, microspheres and nanoparticles is strongly warranted [170].
#### **4. Conclusions**
Cationic nanostructures from lipids, biocompatible polymers, hybrid lipid-polymer or organic-inorganic systems efficiently combine with important biomolecules such as proteins, DNA, epitopes and enhancers of the immune response, most of them negatively charged. Their growing importance for vaccine design becomes evident from the examination of the works briefly described in this overview.
[5] Fifis T, Gamvrellis A, Crimeen-Irwin B, Pietersz GA, Li, J, Mottram PL, McKenzie I FC, Plebanski M. Size-Dependent Immunogenicity: Therapeutic and Protective Prop‐
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 31
[6] O'Hagan DT, MacKichan ML, Singh M.Recent developments in adjuvants for vac‐ cines against infectious diseases. Immunol Infectious Dis Biomol Eng 2001; 18:69-85.
[7] Singh M, Briones M, Ott G, O'Hagan D. Cationic microparticles: A potent delivery
[8] Thiele L, Rothen-Rutishauser B, Jilek S, Wunderli-Allenspach H, Merkle HP, Walter E. Evaluation of particle uptake in human blood monocyte-derived cells in vitro. Does phagocytosis activity of dendritic cells measure up with macrophages? J Con‐
[9] Little SR, Lynn DM, Ge Q, Anderson DG, Puram SV, Chen J, Eisen HN, Langer R.Poly-beta amino ester-containing microparticles enhance the activity of nonviral
[10] O'Hagan DT. MF59 is a safe and potent vaccine adjuvant that enhances protection
[11] Gluck R, Moser C, Metcalfe IC.Influenza virosomes as an efficient system for adju‐
[12] McMichael AJ, Gotch F, Cullen P, Askonas B, Webster RG.The human cytotoxic T cell response to influenza A vaccination.Clin Experim Immunol.1981; 43:276–84.
[13] North RJ, Jung Y-J.Immunity to tuberculosis.Annual Rev. Immunol.2004; 22:599-623.
[14] Carmona-Ribeiro AM.Lipid bilayer fragments and disks in drug delivery.Curr Med
[15] Carmona-Ribeiro AM.Biomimetic particles in drug and vaccine delivery.J Liposome
[16] Carmona-Ribeiro AM.Bilayer-forming synthetic lipids: drugs or carriers? Curr Med
[17] Moura SP, Carmona-Ribeiro AM.Cationic Bilayer Fragments on Silica at Low Ionic Strength: Competitive Adsorption and Colloid Stability.Langmuir.2003; 19:6664-67.
[18] Carmona-Ribeiro AM, Lessa MM.Interactions between bilayer vesicles and latex.Col‐
[19] Lincopan N, Espíndola NM, Vaz AJ, Carmona-Ribeiro AM.Cationic supported lipid
[20] Lincopan N, Santana MRA, Faquim-Mauro E, da Costa MHB, Carmona-Ribeiro AM.Silica-based cationic bilayers as immunoadjuvants. BMC Biotechnol.2009; 9:5.
bilayers for antigen presentation. Int J Pharm. 2007; 340(1-2):216-22.
erties of Nano-Vaccines against Tumors. J Immunol 2004; 173:3148-54.
system for DNA vaccines. Proc Natl Acad Sci U S A. 2000; 97(2):811-6.
genetic vaccines. Proc Natl Acad Sci U S A. 2004; 101(26):9534-9.
vanted vaccine delivery. Expert Opin Biol Ther 2007; 4:1139–45.
against influenza virus infection. Expert Rev Vaccines 2007; 6:699–710.
trol Release 2001; 76:59-71.
Chem.2006; 13:1359-70.
Chem.2003; 10:2415-46.
loids Surf A.1999; 153:355-61.
Res.2007; 17:165-72.
#### **Acknowledgements**
Financial support from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and research grant 2011/00046-5 to AMC-R from Fundação de Âmparo à Pesquisa do Estado de São Paulo (FAPESP) are gratefully acknowledged.
#### **Author details**
Ana Maria Carmona-Ribeiro\*
Address all correspondence to: amcr@usp.br
Biocolloids Laboratory, Instituto de Química, Universidade de São Paulo, São Paulo SP, Brazil
#### **References**
[5] Fifis T, Gamvrellis A, Crimeen-Irwin B, Pietersz GA, Li, J, Mottram PL, McKenzie I FC, Plebanski M. Size-Dependent Immunogenicity: Therapeutic and Protective Prop‐ erties of Nano-Vaccines against Tumors. J Immunol 2004; 173:3148-54.
**4. Conclusions**
30 Immune Response Activation
briefly described in this overview.
**Acknowledgements**
**Author details**
Brazil
**References**
Ana Maria Carmona-Ribeiro\*
Address all correspondence to: amcr@usp.br
pert Rev Vaccines.2003; 2:269-83.
ter? Methods 2006; 40:1-9.
Delivery 2008; 5:230-42.
Cationic nanostructures from lipids, biocompatible polymers, hybrid lipid-polymer or organic-inorganic systems efficiently combine with important biomolecules such as proteins, DNA, epitopes and enhancers of the immune response, most of them negatively charged. Their growing importance for vaccine design becomes evident from the examination of the works
Financial support from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and research grant 2011/00046-5 to AMC-R from Fundação de Âmparo à Pesquisa do
Biocolloids Laboratory, Instituto de Química, Universidade de São Paulo, São Paulo SP,
[1] O'Hagan DT, Singh M. Microparticles as vaccine adjuvants and delivery systems.Ex‐
[2] Xiang SD, Scholzen A, Minigo G, David C, Apostolopoulos V, Mottram PL, Pleban‐ ski M. Pathogen recognition and development of particulate vaccines: Does size mat‐
[3] Caputo A, Sparnacci K, Ensoli B, Tondelli, L. Functional Polymeric Nano/Microparti‐ cles for Surface Adsorption and Delivery of Protein and DNA Vaccines. Curr Drug
[4] Pelkmans L. Secrets of caveolae-and lipid raft-mediated endocytosis revealed by mammalian viruses. Biochim Biophys Acta, Mol Cell Res 2005; 1746:295-304.
Estado de São Paulo (FAPESP) are gratefully acknowledged.
rified from *Saccharomyces cerevisiae*; enhancement of delayed-type hypersensitivity af‐
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 33
[35] Costa MHB, Ueda C, Sato RA, Liberman C, Raw I. Procedures for scaling up the re‐ combinant 18 kDa-hsp lepra protein production. Biotechnol Tech. 1995; 9:527-32.
[36] Gall D.The adjuvant activity of aliphatic nitrogenous bases. Immunology.1966;
[37] Snippe H, Belder M, Willers JM. Dimethyl diotadecyl ammonium bromide as adju‐
[38] Klinguer C, Beck A, De-Lys P, Bussat MC, Blaecke A, Derouet F, Bonnefoy JY, Nguy‐ en TN, Corvaïa N, Velin D.Lipophilic quaternary ammonium salt acts as a mucosal adjuvant when co-administered by the nasal route with vaccine antigens. Vaccine
[39] Davidsen J, Rosenkrands I, Christensen D, Vangala A, Kirby D, Perrie Y, Agger EM, Andersen PA. Characterization of cationic liposomes based on dimethyldioctadecy‐ lammonium and synthetic cord factor from M. tuberculosis (trehalose 6,6V-dibehen‐ ate)—A novel adjuvant inducing both strong CMI and antibody responses.Biochim
[40] Vangala A, Kirby D, Rosenkrands I, Agger EM, Andersen P, Perrie Y. A comparative study of cationic liposome and niosome-based adjuvant systems for protein subunit vaccines: characterisation, environmental scanning electron microscopy and immuni‐
[41] Korsholm KS, Agger EM, Foged C, Christensen D, Dietrich J, Andersen CS, Geisler C, Andersen P.The adjuvant mechanism of cationic dimethyldioctadecylammonium
[42] Foged C, Brodin B, Frokjaer S, Sundblad A.Particle size and surface charge affect par‐ ticle uptake by human dendritic cells in an in vitro model. Int J Pharm.2005; 298:
[43] Dawson M, Krauland E, Wirtz D, Hanes J. Transport of polymeric nanoparticle gene
[44] Hilgers LA, Snippe H. DDA as an immunological adjuvant. Res Immunol.1992; 143:
[45] Tsuruta LR, Quintilio W, Costa, MHB, Carmona-Ribeiro AM. Interactions between cationic liposomes and an antigenic protein: the physical chemistry of the immu‐
[46] Brgles M, Jurasin D, Sikirić MD, Frkanec R, Tomasić J. Entrapment of ovalbumin into liposomes-Factors affecting entrapment efficiency, liposome size, and zeta potential.
sation studies in mice. J Pharm Pharmacol.2006; 58:787-99.
carriers in gastric mucus. Biotechnol Progress.2004; 20:851-57.
liposomes. Immunology.2007; 121:216–26.
noadjuvant action.J Lipid Res.1997; 38:2003–11.
J Liposome Res.2008; 18:235-48.
vant for delayed hypersensitivity in mice. Immunology.1997; 33:931-36.
ter gamma-irradiation. Int J Lepr Other Mycobact Dis.1995; 63:381-390.
11:369-86.
2001; 19:4236–44.
315-22.
494–503.
Biophys Acta. 2005; 1718:22-31.
rified from *Saccharomyces cerevisiae*; enhancement of delayed-type hypersensitivity af‐ ter gamma-irradiation. Int J Lepr Other Mycobact Dis.1995; 63:381-390.
[35] Costa MHB, Ueda C, Sato RA, Liberman C, Raw I. Procedures for scaling up the re‐ combinant 18 kDa-hsp lepra protein production. Biotechnol Tech. 1995; 9:527-32.
[21] Carmona-Ribeiro AM.Synthetic amphiphile vesicles.Chem Soc Rev.1992; 21:207-13. [22] Carmona-Ribeiro AM, Castuma CE, Sesso A, Schreier S.Bilayer structure and stabili‐ ty in dihexadecyl phosphate dispersions.J Phys Chem US 1991; 95:5361-6.
[23] Hammarstroem L, Velikian I, Karlsson G, Edwards K. Cryo-TEM evidence: sonica‐ tion of dihexadecyl phosphate does not produce closed bilayers with smooth curva‐
[24] Cocquyt J, Olsson U, Olofsson G, van der Meeren, P. Temperature quenched DO‐ DAB dispersions: fluid and solid state coexistence and complex formation with op‐
[25] Vieira DB, Carmona-Ribeiro AM.Synthetic bilayer fragments for solubilization of
[26] Kunitake T, Okahata Y, Tamaki K, Kumamaru F, Takayanagi M.Formation of the bi‐ layer membrane from a series of quaternary ammonium salts.Chem Lett.1977; 6:
[27] Lincopan N, Espíndola NM, Vaz AJ, da Costa MH, Faquim-Mauro E, Carmona-Ri‐ beiro AM. Novel immunoadjuvants based on cationic lipid: Preparation, characteri‐
[28] Gray JJ.The interaction of proteins with solid surfaces.Curr Opin Struct Biol.2004;
[29] Carvalho LA, Carmona-Ribeiro AM.Interactions between cationic vesicles and serum
[30] Bodzon-Kulakowska A, Bierczynska-Krzysik A, Dylag T, Drabik A, Suder P, Noga M, Jarzebinska J, Silberring J. Methods for samples preparation in proteomic re‐
[31] Espíndola NM, Vaz AJ, Pardini AX, Fernandes I. Excretory/secretory antigens (ES) from in-vitro cultures of *Taenia crassiceps* cysticerci, and use of an anti-ES monoclonal antibody for antigen detection in samples of cerebrospinal fluid from patients with
[32] Espindola NM, Iha AH, Fernandes I, Takayanagui OM, dos Ramos Machado L, Liv‐ ramento JA, Maia AAM, Peralta JM, Vaz AJ.Cysticercosis immunodiagnosis using 18-and 14-kilodalton proteins from *Taenia crassiceps* cysticercus antigens obtained by
[33] Mustafa AS, Lundin KE, Oftung F.Human T cells recognize mycobacterial heat shock proteins in the context of multiple HLA-DR molecules: studies with healthy subjects vaccinated with *Mycobacterium bovis* BCG and *Mycobacterium leprae*. Infect Immun.
[34] Pinho JR, Cardi BA, Andrade HF Jr, Barr PJ, Bathurst IC, Vicente EJ, Schenberg AC. Immunogenic properties of the *Mycobacterium leprae* recombinant 18-kDa antigen pu‐
search.J Chromatogr B Analyt Technol Biomed Life Sci.2007; 849:1-31.
neurocysticercosis.Ann Tropical Med Parasitol.2002; 96:361-68.
immunoaffinity chromatography. J Clin Microbiol.2005; 43:3178-84.
positely charged surfactant. Langmuir.2004; 20:3906-12.
amphotericin B. J Colloid Interface Sci.2001; 244:427-31.
zation and activity in vivo.Vaccine. 2009; 27(42):5760-71.
proteins.Langmuir.1998; 14:6077-81.
ture.Langmuir.1995; 11:408-10.
387-90.
32 Immune Response Activation
14:110-15.
1993; 61:5294-5301.
[47] Sacks D, Noben-Traut N. The immunology of susceptibility and resistance to Leish‐ mania major in mice. Nature Rev Immunol.2002; 2:845-58.
[61] Liu L, Zhou X, Liu H, Xiang L, Yuan Z. CpG motif acts as a 'danger signal' and pro‐ vides a T helper type 1-biased microenvironment for DNA vaccination.Immunology.
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 35
[62] Fonseca DE, Kline JN.Use of CpG oligonucleotides in treatment of asthma and aller‐
[63] Kuramoto Y, Nishikawa M, Hyoudou K, Yamashita F, Hashida M. Inhibition of peri‐ toneal dissemination of tumor cells by single dosing of phosphodiester CpG oligonu‐
[64] Shi F, Hoekstra D. Effective intracellular delivery of oligonucleotides in order to
[65] Felgner PL, Barenholz Y, Behr JP, Cheng SH, Cullis P, Huang L, Jessee JA, Seymour L, Szoka F, Thierry AR, Wagner E, Wu G. Nomenclature for synthetic gene delivery
[66] Nantes IL, Correia FM, Faljoni-Alario A, Kawanami AE, Ishiki HM, Amaral AT,Car‐ mona-Ribeiro AM.Nucleotide conformational change induced by cationic bilayers.
[67] Kabanov VA, Yaroslavov AA. What happens to negatively charged lipid vesicles upon interacting with polycation species? J Control Release. 2002; 78(1-3):267-71.
[68] Sybachin AV, Efimova AA, Litmanovich EA, Menger FM, Yaroslavov AA.Complexa‐ tion of polycations to anionic liposomes: composition and structure of the interfacial
[69] O'Hagan DT, Singh M, Ulmer JB.Microparticles for the delivery of DNA vaccines.
[70] El-Sayed IH, Huang X, El-Sayed MA. Surface plasmon resonance scattering and ab‐ sorption of anti-EGFR antibody conjugated gold nanoparticles in cancer diagnostics:
[71] Kovacsovics-Bankowski M, Clark K, Benacerraf B, Rock KL. Efficient major histo‐ compatibility complex class I presentation of exogenous antigen upon phagocytosis
[72] Vidard L, Kovacsovics-Bankowski M, Kraeft SK, Chen LB, Benacerraf B, Rock KL.Analysis of MHC class II presentation of particulate antigens of B lymphocytes. J
[73] Carmona-Ribeiro AM, Midmore BR.Synthetic bilayer adsorption onto polystyrene
[74] Carmona-Ribeiro AM.Bilayer vesicles and liposomes as interface agents.Chem Soc
cleotide/cationic liposome complex.J Control Release. 2006; 115(2):226-33.
make sense of antisense. J Control Release.2004; 97(2):189-209.
gic disease. Adv Drug Deliv Rev. 2009; 61(3):256-62.
systems. Hum Gene Ther. 1997; 8(5):511-2.
Arch Biochem Biophys. 2003; 416(1):25-30.
complexes. Langmuir. 2007; 23(20):10034-9.
applications in oral cancer. Nano Lett. 2005; 5(5):829-34.
by macrophages. Proc Natl Acad Sci U S A. 1993; 90(11):4942-6.
Immunol Rev. 2004; 199:191-200.
Immunol. 1996; 156(8):2809-18.
Rev.2001; 30:241-7.
microspheres. Langmuir.1992; 8:801-6.
2005; 115(2):223-30.
[61] Liu L, Zhou X, Liu H, Xiang L, Yuan Z. CpG motif acts as a 'danger signal' and pro‐ vides a T helper type 1-biased microenvironment for DNA vaccination.Immunology. 2005; 115(2):223-30.
[47] Sacks D, Noben-Traut N. The immunology of susceptibility and resistance to Leish‐
[48] Good MF, Xu H, Wykes M, Engwerda CR. Development and regulation of cell-medi‐ ated immune responses to the blood stages of malaria: implications for vaccine re‐
[49] De Paula SO, Lima DM, França RFO, Gomes-Ruiz AC, Fonseca BAL. A DNA vaccine candidate expressing dengue-3 virus prM and E proteins elicits neutralizing antibod‐
[50] Abbas AK, Lichtman AH, Pillai S. Cellular and Molecular Immunology 6th Ed. 2006,
[51] Pacheco LF, Carmona-Ribeiro AM.Effects of synthetic lipids on solubilization and colloid stability of hydrophobic drugs. J Colloid Interface Sci. 2003; 258(1):146-54. [52] Vieira DB, Pacheco LF, Carmona-Ribeiro AM.Assembly of a model hydrophobic drug into cationic bilayer fragments. J Colloid Interface Sci. 2006; 293(1):240-7.
[53] Rosa H, Petri DF, Carmona-Ribeiro AM.Interactions between bacteriophage DNA
[54] Rozenfeld JH, Oliveira TR, Lamy MT, Carmona-Ribeiro AM. Interaction of cationic bilayer fragments with a model oligonucleotide. Biochim Biophys Acta.2011; 1808(3):
[55] Zamecnik PC, Stephenson ML. Inhibition of Rous sarcoma virus replication and cell transformation by a specific oligodeoxynucleotide. Proc Natl Acad Sci U S A.1978;
[56] Flaherty KT, Stevenson JP, O'Dwyer PJ. Antisense therapeutics: lessons from early
[57] Tamm I, Dörken B, Hartmann G. Antisense therapy in oncology: new hope for an old
[58] De Backer MD, Nelissen B, Logghe M, Viaene J, Loonen I, Vandoninck S, de Hoogt R, Dewaele S, Simons FA, Verhasselt P, Vanhoof G, Contreras R, Luyten WH.An anti‐ sense-based functional genomics approach for identification of genes critical for
[59] Dean NM. Functional genomics and target validation approaches using antisense oli‐
[60] Juliano R, Alam MR, Dixit V, Kang H. Mechanisms and strategies for effective deliv‐ ery of antisense and siRNA oligonucleotides. Nucleic Acids Res. 2008; 36(12):4158-71.
clinical trials. Curr Opin Oncol. 2001; 13(6):499-505.
growth of *Candida albicans*. Nat Biotechnol. 2001; 19(3):235-41.
gonucleotide technology.Curr Opin Biotechnol. 2001; 12(6):622-5.
idea? Lancet.2001; 358(9280):489-97.
and cationic biomimetic particles. J Phys Chem B. 2008; 112(51):16422-30.
ies and protects mice against lethal challenge.Arch Virol.2008; 153:2215–23.
mania major in mice. Nature Rev Immunol.2002; 2:845-58.
search. Annual Rev Immunol.2005; 23:69-99.
Saunders Elsevier.
34 Immune Response Activation
649-55.
75(1):280-4.
[75] Lincopan N, Carmona-Ribeiro AM. Lipid-covered drug particles: combined action of dioctadecyldimethylammonium bromide and amphotericin B or miconazole. J Anti‐ microb Chemother.2006; 58(1):66-75.
[89] Hoekstra D, Rejman J, Wasungu L, Shi F, Zuhorn I. Gene delivery by cationic lipids:
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 37
[90] Tsuruta LR, Lessa MM, Carmona-Ribeiro AM. Effect of particle size on colloid stabil‐ ity of bilayer-covered polystyrene microspheres.J Colloid Interface Sci. 1995; 175(2):
[91] Krieg AM, Yi AK, Matson S, Waldschmidt TJ, Bishop GA, Teasdale R, Koretzky GA,Klinman DM. CpG motifs in bacterial DNA trigger direct B-cell activation. Na‐
[92] Vollmer J, Krieg AM.Immunotherapeutic applications of CpG oligodeoxynucleotide
[93] Manegold C, Gravenor D, Woytowitz D, Mezger J, Hirsh V, Albert G, Al-Adhami M, Readett D, Krieg AM, Leichman CG. Randomized phase II trial of a toll-like receptor 9 agonist oligodeoxynucleotide, PF-3512676, in combination with first-line taxane plus platinum chemotherapy for advanced-stage non-small-cell lung cancer.J Clin
[94] Zimmermann S, Egeter O, Hausmann S, Lipford GB, Röcken M, Wagner H, Heeg K. CpG oligodeoxynucleotides trigger protective and curative Th1 responses in lethal
[95] Thompson AJ, Colledge D, Rodgers S, Wilson R, Revill P, Desmond P, Mansell A, Visvanathan K, Locarnini S. Stimulation of the interleukin-1 receptor and Toll-like re‐ ceptor 2 inhibits hepatitis B virus replication in hepatoma cell lines *in vitro*. Antivir
[96] Simons FE, Shikishima Y, Van Nest G, Eiden JJ, HayGlass KT. Selective immune redi‐ rection in humans with ragweed allergy by injecting Amb a 1 linked to immunosti‐
[97] Rozenfeld JH, Silva SR, Ranéia PA, Faquim-Mauro E, Carmona-Ribeiro AM.Stable assemblies of cationic bilayer fragments and CpG oligonucleotide with enhanced im‐
[98] Beeley JA, Stevenson SM, Beeley JG. Polyacrylamide gel isoelectric focusing of pro‐ teins: determination of isoelectric points using an antimony electrode.Biochim Bio‐
[99] Walker HW, Grant SB. Influence of surface charge and particle size on the stabiliza‐ tion of colloidal particles by model polyelectrolytes.Colloids Surf A.1998;
[100] Henriksen-Lacey M, Bramwell VW, Christensen D, Agger EM, Andersen P, Perrie Y.Liposomes based on dimethyldioctadecylammonium promote a depot effect and enhance immunogenicity of soluble antigen.J Control Release. 2010; 142(2):180-6.
in and out of an endosome. Biochem Soc Trans. 2007; 35 (Pt 1):68-71.
TLR9 agonists. Adv Drug Deliv Rev. 2009; 61(3):195-204.
murine leishmaniasis.J Immunol. 1998; 160(8):3627-30.
mulatory DNA. J Allergy Clin Immunol. 2004; 113(6):1144-51.
munoadjuvant activity *in vivo*.J Control Release.2012; 160(2):367-73.
470-5.
ture.1995; 374(6522):546-9.
Oncol. 2008; 26(24):3979-86.
Ther. 2009; 14(6):797-808.
phys Acta. 1972; 285(2):293-300.
135,123-133.
[89] Hoekstra D, Rejman J, Wasungu L, Shi F, Zuhorn I. Gene delivery by cationic lipids: in and out of an endosome. Biochem Soc Trans. 2007; 35 (Pt 1):68-71.
[75] Lincopan N, Carmona-Ribeiro AM. Lipid-covered drug particles: combined action of dioctadecyldimethylammonium bromide and amphotericin B or miconazole. J Anti‐
[76] Pereira EMA, Vieira DB, Carmona-Ribeiro AM.Cationic bilayers on polymeric parti‐ cles: effect of low NaCl concentration on surface coverage. J Phys Chem B. 2004;
[77] Li W, Suez I, Szoka FC Jr. Reconstitution of the M13 major coat protein and its trans‐ membrane peptide segment on a DNA template.Biochemistry. 2007; 46(29):8579-91.
[78] Luo D, Saltzman WM. Synthetic DNA delivery systems. Nat Biotechnol. 2000; 18(1):
[79] Gaweda S, Morán MC, Pais AA, Dias RS, Schillén K, Lindman B, Miguel MG.Cation‐
[80] Cerritelli ME, Cheng N, Rosenberg AH, McPherson CE, Booy FP, Steven AC.Encap‐
[81] Zinchenko AA, Luckel F, Yoshikawa K. Transcription of giant DNA complexed with cationic nanoparticles as a simple model of chromatin. Biophys J. 2007; 92(4):1318-25.
[82] Zinchenko AA, Yoshikawa K, Baigl D. Compaction of single-chain DNA by histone-
[83] Lindsay S. Chromatin control of gene expression: the simplest model. Biophys J.
[84] Kneuer C, Sameti M, Bakowsky U, Schiestel T, Schirra H, Schmidt H, Lehr CM.A nonviral DNA delivery system based on surface modified silica-nanoparticles can ef‐
[85] Thomas M, Klibanov AM.Conjugation to gold nanoparticles enhances polyethyleni‐ mine's transfer of plasmid DNA into mammalian cells. Proc Natl Acad Sci U S A.
[86] Roy I, Ohulchanskyy TY, Bharali DJ, Pudavar HE, Mistretta RA, Kaur N, Prasad PN. Optical tracking of organically modified silica nanoparticles as DNA carriers: a non‐ viral, nanomedicine approach for gene delivery. Proc Natl Acad Sci U S A. 2005;
[87] Zinchenko AA, Sakaue T, Araki S, Yoshikawa K, Baigl D. Single-chain compaction of long duplex DNA by cationic nanoparticles: modes of interaction and comparison
[88] Rejman J, Oberle V, Zuhorn IS, Hoekstra D. Size-dependent internalization of parti‐ cles via the pathways of clathrin-and caveolae-mediated endocytosis.Biochem J.
ficiently transfect cells *in vitro*. Bioconjug Chem. 2000; 11(6):926-32.
ic agents for DNA compaction. J Colloid Interface Sci. 2008; 323(1):75-83.
sidated conformation of bacteriophage T7 DNA. Cell.1997; 91(2):271-80.
inspired nanoparticles.Phys Rev Lett. 2005; 95(22):228101.
with chromatin. J Phys Chem B. 2007; 111(11):3019-31.
microb Chemother.2006; 58(1):66-75.
108:11490-5.
2007; 92(4):1113.
2003; 100(16):9138-43.
2004;377(Pt 1):159-69.
102(2):279-84.
33-7.
36 Immune Response Activation
[101] Manolova V, Flace A, Bauer M, Schwarz K, Saudan P, Bachmann MF. Nanoparticles target distinct dendritic cell populations according to their size. Eur J Immunol. 2008; 38(5):1404-13.
[113] Chen WR, Korbelik M, Bartels KE, Liu H, Sun J, Nordquist RE.Enhancement of laser cancer treatment by a chitosan-derived immunoadjuvant. Photochem Photobiol.2005;
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 39
[114] Yoon TJ, Kim JY, Kim H, Hong C, Lee H, Lee CK, Lee KH, Hong S, Park SH.Antitumor immunostimulatory effect of heat-killed tumor cells. Exp Mol Med. 2008;
[115] Jäger E, Jäger D, Knuth A.Antigen-specific immunotherapy and cancer vaccines.Int J
[116] Wu F, Zhou L, Chen WR. Host antitumour immune responses to HIFU ablation. Int J
[117] Chen WR, Jeong SW, Lucroy MD, Wolf RF, Howard EW, Liu H, Nordquist RE.In‐ duced antitumor immunity against DMBA-4 metastatic mammary tumors in rats us‐
[118] Peniche H, Peniche C. Chitosan nanoparticles: a contribution to nanomedicine.Polym
[119] Calvo P, Remuñán-López C, Vila-Jato JL, Alonso MJ, Novel hydrophilic chitosan– polyethylene oxide nanoparticles as protein carriers.J Polym Sci.1997;63:125–132. [120] Hu Y, Jiang X, Ding Y, Ge H, Yuan Y, Yang C.Synthesis and characterization of chito‐
[121] Cui Z, Mumper RJ.Chitosan-based nanoparticles for topical genetic immunization.J
[122] Sarmento B, Ribeiro AJ, Veiga F, Ferreira DC, Neufeld RJ.Insulin-loaded nanoparti‐ cles are prepared by alginate ionotropic pre-gelation followed by chitosan polyelec‐
[123] Vila AI, Suárez S, Alonso MJ. Chitosan and heparin nanoparticles. Spanish Patent.
[124] Vieira DB, Kim V, Petri DFS, Menck CFM, Carmona-Ribeiro AM. Supramolecular as‐ semblies of cisplatin and polyelectrolytes: preparation, characterization and activity against cancer cells. In: Laudon M, Romanowicz B. (eds.) Nanotech Conference & Ex‐ po 2012: An Interdisciplinary Integrative Forum on Nanotechnology, Microtechnolo‐ gy, Biotechnology and Cleantechnology,18-21 June 2012, Santa Clara, CA, United
[125] Sayin B, Somavarapu S, Li XW, Thanou M, Sesardic D, Alpar HO, Senel S.Mono-Ncarboxymethyl chitosan (MCC) and N-trimethyl chitosan (TMC) nanoparticles for
san-poly(acrylic acid) nanoparticles.Biomaterials. 2002; 23(15):3193-201.
trolyte complexation.J Nanosci Nanotechnol.2007; 7(8):2833-41.
non-invasive vaccine delivery.Int J Pharm. 2008; 363(1-2):139-48.
ing laser immunotherapy. Int J Cancer. 2003; 107(6):1053-7.
81(1):190-5.
40(1):130-44.
Cancer.2003; 106(6):817-20.
Int. 2011; 60:883–889.
Hyperthermia. 2007; 23(2):165-71.
Control Release. 2001; 5(3):409-19.
States.Boca Raton, Fla: CRC Press, 2012.
2007; WO/2007/042572.
[113] Chen WR, Korbelik M, Bartels KE, Liu H, Sun J, Nordquist RE.Enhancement of laser cancer treatment by a chitosan-derived immunoadjuvant. Photochem Photobiol.2005; 81(1):190-5.
[101] Manolova V, Flace A, Bauer M, Schwarz K, Saudan P, Bachmann MF. Nanoparticles target distinct dendritic cell populations according to their size. Eur J Immunol. 2008;
[102] Krieg AM.Therapeutic potential of Toll-like receptor 9 activation.Nat Rev Drug Dis‐
[103] Joseph A, Louria-Hayon I, Plis-Finarov A, Zeira E, Zakay-Rones Z, Raz E, Hayashi T, Takabayashi K, Barenholz Y, Kedar E. Liposomal immunostimulatory DNA se‐ quence (ISS-ODN): an efficient parenteral and mucosal adjuvant for influenza and
[104] Mitchell LA, Joseph A, Kedar E, Barenholz Y, Galun E. Mucosal immunization against hepatitis A: antibody responses are enhanced by co-administration of syn‐ thetic oligodeoxynucleotides and a novel cationic lipid. Vaccine.2006; 24(25):5300-10.
[105] Mayorga O, Muñoz JE, Lincopan N, Teixeira AF, Ferreira LC, Travassos LR,Taborda CP.The role of adjuvants in therapeutic protection against paracoccidioidomycosis
[106] Gaspar EB, Rosetti AS, Lincopan N, De Gaspari E. *Neisseria lactamica* antigens com‐ plexed with a novel cationic adjuvant. Hum Vaccin Immunother.2013; 9(3). [Epub
[107] Li X, Min M, Du N, Gu Y, Hode T, Naylor M, Chen D, Nordquist RE, Chen WR. Chi‐ tin, chitosan, and glycated chitosan regulate immune responses: the novel adjuvants
[108] Jarmila V, Vavríková E.Chitosan derivatives with antimicrobial, antitumour and an‐
[109] Pillai CKS, Paul W, Sharma CP. Chitin and chitosan polymers: chemistry, solubility
[110] Chung YC, Yeh JY, Tsai CF. Antibacterial characteristics and activity of water-soluble chitosan derivatives prepared by the Maillard reaction. Molecules.2011; 16(10):
[111] Song S, Zhou F, Nordquist RE, Carubelli R, Liu H, Chen WR.Glycated chitosan as a new non-toxic immunological stimulant Glycated chitosan immunological stimu‐
[112] Li X, Ferrel GL, Guerra MC, Hode T, Lunn JA, Adalsteinsson O, Nordquist RE,Liu H, Chen WR.Preliminary safety and efficacy results of laser immunotherapy for the treatment of metastatic breast cancer patients.Photochem Photobiol Sci. 2011; 10(5):
tioxidant activities-a review. Curr Pharm Des.2011; 17(32):3596-607.
and fiber formation. Progress in Polymer Science.2009; 34(7):641–678.
lant.Immunopharmacol Immunotoxicol. 2009; 31(2):202–208.
after immunization with the P10 peptide. Front Microbiol.2012; 3:154.
hepatitis B vaccines. Vaccine. 2002; 20(27-28):3342-54.
ahead of print] PubMed PMID: 23296384.
for cancer vaccine. Clin Dev Immunol. 2013; 2013:387023.
38(5):1404-13.
38 Immune Response Activation
8504-14.
817-21.
cov.2006; 5(6):471-84.
[126] Behrens I, Pena AI, Alonso MJ, Kissel T. Comparative uptake studies of bioadhesive and non-bioadhesive nanoparticles in human intestinal cell lines and rats: the effect of mucus on particle adsorption and transport. Pharm Res. 2002; 19(8):1185-93.
p53 cancer vaccine with chemotherapy in patients with extensive stage small cell
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 41
[138] Xiong W, Candolfi M, Liu C, Muhammad AK, Yagiz K, Puntel M, Moore PF, Avalos J, Young JD, Khan D, Donelson R, Pluhar GE, Ohlfest JR, Wawrowsky K, Lowenstein PR, Castro MG. Human Flt3L generates dendritic cells from canine peripheral blood precursors: implications for a dog glioma clinical trial.PLoS One. 2010; 5(6):e11074.
[139] Mori A, Oleszycka E, Sharp FA, Coleman M, Ozasa Y, Singh M, O'Hagan DT, Tajber L, Corrigan OI, McNeela EA, Lavelle EC. The vaccine adjuvant alum inhibits IL-12 by promoting PI3 kinase signaling while chitosan does not inhibit IL-12 and enhan‐
[140] Yang Z, Sahay G, Sriadibhatla S, Kabanov AV. Amphiphilic block copolymers en‐ hance cellular uptake and nuclear entry of polyplex-delivered DNA. Bioconjug
[141] Kanazawa T, Takashima Y, Murakoshi M, Nakai Y, Okada H. Enhancement of gene transfection into human dendritic cells using cationic PLGA nanospheres with a syn‐
[142] Bal SM, Slütter B, Verheul R, Bouwstra JA, Jiskoot W. Adjuvanted, antigen loaded Ntrimethyl chitosan nanoparticles for nasal and intradermal vaccination: adjuvant-and
[143] Yu F, Wang J, Dou J, Yang H, He X, Xu W, Zhang Y, Hu K, Gu N.Nanoparticle-based adjuvant for enhanced protective efficacy of DNA vaccine Ag85A-ESAT-6-IL-21
against *Mycobacterium tuberculosis* infection. Nanomedicine.2012; 8(8):1337-44. [144] Wang G, Pan L, Zhang Y, Wang Y, Zhang Z, Lü J, Zhou P, Fang Y, Jiang S.Intranasal delivery of cationic PLGA nano/microparticles-loaded FMDV DNA vaccine encoding IL-6 elicited protective immunity against FMDV challenge. PLoS One.2011;
[145] Zwiorek K, Bourquin C, Battiany J, Winter G, Endres S, Hartmann G, Coester C.De‐ livery by cationic gelatin nanoparticles strongly increases the immunostimulatory ef‐
[146] Coester CJ, Langer K, van Briesen H, Kreuter J. Gelatin nanoparticles by two step desolvation--a new preparation method, surface modifications and cell uptake.J Mi‐
[147] Doll TAPF, Raman S, Dey R,Burkhard P. 2013 Nanoscale assemblies and their bio‐
[148] Carmona-Ribeiro AM, de Melo Carrasco LD.Cationic antimicrobial polymers and
fects of CpG oligonucleotides.Pharm Res. 2007; 25(3):551-62.
medical applications. J R Soc Interface 10: 20120740.
their assemblies. Int J Mol Sci. 2013; 14(5):9906-46.
site-dependent immunogenicity in mice. Eur J Pharm Sci. 2012; 45(4):475-81.
ces Th1 and Th17 responses. Eur J Immunol. 2012; 42(10):2709-19.
thesized nuclear localization signal. Int J Pharm. 2009; 379(1):187-95.
lung cancer. Clin Cancer Res. 2006; 12(3 Pt 1):878-87.
Chem. 2008; 19(10):1987-94.
6(11):e27605.
croencapsul. 2000; 17(2):187-93.
p53 cancer vaccine with chemotherapy in patients with extensive stage small cell lung cancer. Clin Cancer Res. 2006; 12(3 Pt 1):878-87.
[138] Xiong W, Candolfi M, Liu C, Muhammad AK, Yagiz K, Puntel M, Moore PF, Avalos J, Young JD, Khan D, Donelson R, Pluhar GE, Ohlfest JR, Wawrowsky K, Lowenstein PR, Castro MG. Human Flt3L generates dendritic cells from canine peripheral blood precursors: implications for a dog glioma clinical trial.PLoS One. 2010; 5(6):e11074.
[126] Behrens I, Pena AI, Alonso MJ, Kissel T. Comparative uptake studies of bioadhesive and non-bioadhesive nanoparticles in human intestinal cell lines and rats: the effect of mucus on particle adsorption and transport. Pharm Res. 2002; 19(8):1185-93.
[127] Huang M, Khor E, Lim LY.Uptake and cytotoxicity of chitosan molecules and nano‐ particles: effects of molecular weight and degree of deacetylation. Pharm Res. 2004;
[128] Chowdary KP, Rao YS.Mucoadhesive microspheres for controlled drug delivery. Biol
[129] Sayin B, Somavarapu S, Li XW, Sesardic D, Senel S, Alpar OH.TMC-MCC (N-tri‐ methyl chitosan-mono-N-carboxymethyl chitosan) nanocomplexes for mucosal de‐
[130] Chen YZ, Yao XL, Tabata Y, Nakagawa S, Gao JQ.Gene carriers and transfection sys‐ tems used in the recombination of dendritic cells for effective cancer immunothera‐
[131] Kim TH, Nah JW, Cho MH, Park TG, Cho CS. Receptor-mediated gene delivery into antigen presenting cells using mannosylated chitosan/DNA nanoparticles.J Nanosci
[132] Kim TH, Jin H, Kim HW, Cho MH, Cho CS. Mannosylated chitosan nanoparticlebased cytokine gene therapy suppressed cancer growth in BALB/c mice bearing
[133] Zhou X, Liu B, Yu X, Zha X, Zhang X, Chen Y, Wang X, Jin Y, Wu Y, Chen Y, Shan Y, Chen Y, Liu J, Kong W, Shen J. Controlled release of PEI/DNA complexes from man‐ nose-bearing chitosan microspheres as a potent delivery system to enhance immune
[134] Engell-Noerregaard L, Hansen TH, Andersen MH, Thor Straten P, Svane IM.Review of clinical studies on dendritic cell-based vaccination of patients with malignant mel‐ anoma: assessment of correlation between clinical response and vaccine parame‐
[135] Smits EL, Anguille S, Cools N, Berneman ZN, Van Tendeloo VF.Dendritic cell-based
[136] Frankenberger B, Regn S, Geiger C, Noessner E, Falk CS, Pohla H, Javorovic M, Sil‐ berzahn T, Wilde S, Buchner A, Siebels M, Oberneder R, Willimsky G, Pezzutto A, Blankenstein T, Schendel DJ. Cell-based vaccines for renal cell carcinoma: genetical‐ ly-engineered tumor cells and monocyte-derived dendritic cells. World J Urol. 2005;
[137] Antonia SJ, Mirza N, Fricke I, Chiappori A, Thompson P, Williams N, Bepler G, Si‐ mon G, Janssen W, Lee JH, Menander K, Chada S, Gabrilovich DI.Combination of
21(2):344-53.
40 Immune Response Activation
23(3):166-74.
Pharm Bull.2004; 27(11):1717-24.
livery of vaccines. Eur J Pharm Sci. 2009; 38(4):362-9.
CT-26 carcinoma cells.Mol Cancer Ther. 2006; 5(7):1723-32.
ters.Cancer Immunol Immunother. 2009; 58(1):1-14.
cancer gene therapy. Hum Gene Ther. 2009; 20(10):1106-18.
response to HBV DNA vaccine. J Control Release.2007;121(3):200-7.
py. Clin Dev Immunol. 2010; 2010:565643.
Nanotechnol. 2006; 6(9-10):2796-803.
[149] Carmona-Ribeiro AM. Biomimetic nanoparticles: preparation, characterization and biomedical applications. Int J Nanomedicine.2010; 5:249-59.
nant FGF-2 and PDGF-BB or generation of antibodies.Gene Ther. 2009; 16(9):
Cationic Nanostructures for Vaccines http://dx.doi.org/10.5772/57543 43
[162] Buschmann MD, Merzouki A, Lavertu M, Thibault M, Jean M, Darras V. Chitosans for delivery of nucleic acids. Adv Drug Deliv Rev. 2013;doi:pii:S0169-409X(13)00159-2. 10.1016/j.addr.2013.07.005. [Epub ahead of print]
[163] Doroud D, Vatanara A, Zahedifard F, Gholami E, Vahabpour R, Rouholamini Najafa‐ badi A, Rafati S. Cationic solid lipid nanoparticles loaded by cysteine proteinase genes as a novel anti-leishmaniasis DNA vaccine delivery system: characterization
[164] Doroud D, Zahedifard F, Vatanara A, Taslimi Y, Vahabpour R, Torkashvand F, Va‐ ziri B, Rouholamini Najafabadi A, Rafati S.C-terminal domain deletion enhances the protective activity of cpa/cpb loaded solid lipid nanoparticles against *Leishmania ma‐*
[165] Doroud D, Zahedifard F, Vatanara A, Najafabadi AR, Taslimi Y, Vahabpour R, Tor‐ kashvand F, Vaziri B, Rafati S.Delivery of a cocktail DNA vaccine encoding cysteine proteinases type I, II and III with solid lipid nanoparticles potentiate protective im‐ munity against Leishmania major infection.J Control Release. 2011; 153(2):154-62. [166] Saljoughian N, Zahedifard F, Doroud D, Doustdari F, Vasei M, Papadopoulou B, Ra‐ fati S. Cationic solid lipid nanoparticles are as efficient as electroporation in DNA vaccination against visceral leishmaniasis in mice. Parasite Immunol.2013.doi:
[167] Mohit E, Rafati S. Biological delivery approaches for gene therapy: Strategies to po‐ tentiate efficacy and enhance specificity.Mol Immunol. 2013; 56(4):599-611.doi:
[168] Ma Y, Zhuang Y, Xie X, Wang C, Wang F, Zhou D, Zeng J, Cai L.The role of surface charge density in cationic liposome-promoted dendritic cell maturation and vaccine-
[169] Carstens MG, van der Maaden K, van der Velden D, Ottenhoff TH, Melief CJ,Ossen‐ dorp F, Bouwstra JA, Jiskoot W.Evaluation of the high-pressure extrusion technique as a method for sizing plasmid DNA-containing cationic liposomes.J Liposome Res.
[170] van den Berg JH, Oosterhuis K, Hennink WE, Storm G, van der Aa LJ, Engbersen JF, Haanen JB, Beijnen JH, Schumacher TN, Nuijen B.Shielding the cationic charge of nanoparticle-formulated dermal DNA vaccines is essential for antigen expression
[171] Tyagi RK, Garg NK, Sahu T.Vaccination strategies against malaria: novel carrier(s)
and *in vitro* evaluations.J Pharm Pharm Sci. 2010;13(3):320-35.
*jor* in BALB/c mice. PLoS Negl Trop Dis.2011; 5(7):e1236.
induced immune responses.Nanoscale. 2011; 3(5):2307-14.
and immunogenicity.J Control Release. 2010; 141(2):234-40.
more than a tour de force.J Control Release. 2012; 162(1):242-54.
10.1111/pim.12042. [Epub ahead of print]
10.1016/j.molimm.2013.06.005.
2011;21(4):286-95.
1097-110.
nant FGF-2 and PDGF-BB or generation of antibodies.Gene Ther. 2009; 16(9): 1097-110.
[162] Buschmann MD, Merzouki A, Lavertu M, Thibault M, Jean M, Darras V. Chitosans for delivery of nucleic acids. Adv Drug Deliv Rev. 2013;doi:pii:S0169-409X(13)00159-2. 10.1016/j.addr.2013.07.005. [Epub ahead of print]
[149] Carmona-Ribeiro AM. Biomimetic nanoparticles: preparation, characterization and
[150] Ji W, Panus D, Palumbo RN, Tang R, Wang C.Poly(2-aminoethyl methacrylate) with well-defined chain length for DNA vaccine delivery to dendritic cells.Biomacromole‐
[151] Gupta NK, Tomar P, Sharma V, Dixit VK. Development and characterization of chi‐ tosan coated poly-ε-caprolactone nanoparticulate system for effective immunization
[152] Castaldello A, Brocca-Cofano E, Voltan R, Triulzi C, Altavilla G, Laus M,Sparnacci K, Ballestri M, Tondelli L, Fortini C, Gavioli R, Ensoli B, Caputo A. DNA prime and pro‐ tein boost immunization with innovative polymeric cationic core-shell nanoparticles elicits broad immune responses and strongly enhance cellular responses of HIV-1 tat
[153] Huang HN, Li TL, Chan YL, Chen CL, Wu CJ.Transdermal immunization with lowpressure-gene-gun mediated chitosan-based DNA vaccines against Japanese ence‐
[154] Nochi T, Yuki Y, Takahashi H, Sawada S, Mejima M, Kohda T, Harada N, Kong IG,Sato A, Kataoka N, Tokuhara D, Kurokawa S, Takahashi Y, Tsukada H, Kozaki S,Akiyoshi K, Kiyono H.Nanogel antigenic protein-delivery system for adjuvant-free
[155] Kong IG, Sato A, Yuki Y, Nochi T, Takahashi H, Sawada S, Mejima M, Kurokawa S,Okada K, Sato S, Briles DE, Kunisawa J, Inoue Y, Yamamoto M, Akiyoshi K, Kiyo‐ noH. Nanogel-based PspA intranasal vaccine prevents invasive disease and nasal
[156] Vangasseri DP, Han SJ, Huang L. Lipid-protamine-DNA-mediated antigen deliv‐
[157] Watson DS, Endsley AN, Huang L. Design considerations for liposomal vaccines: in‐ fluence of formulation parameters on antibody and cell-mediated immune responses
[158] Cui Z, Han SJ, Vangasseri DP, Huang L.Immunostimulation mechanism of LPD
[159] Gao X, Huang L. Potentiation of cationic liposome-mediated gene delivery by poly‐
[160] Cho HJ, Han SE, Im S, Lee Y, Kim YB, Chun T, Oh YK.Maltosylated polyethyleni‐ mine-based triple nanocomplexes of human papillomavirus 16L1 protein and DNA
[161] Jean M, Smaoui F, Lavertu M, Méthot S, Bouhdoud L, Buschmann MD, Merzouki A.Chitosan-plasmid nanoparticle formulations for IM and SC delivery of recombi‐
to liposome associated antigens.Vaccine. 2012; 30(13):2256-72.
nanoparticle as a vaccine carrier. Mol Pharm. 2005; 2(1):22-8.
as a vaccine co-delivery system.Biomaterials. 2011; 32(20):4621-9.
colonization by *Streptococcus pneumoniae*.Infect Immun. 2013; 81(5):1625-34.
biomedical applications. Int J Nanomedicine.2010; 5:249-59.
against influenza.Vaccine. 2011; 29(48):9026-37.
DNA vaccination.Vaccine. 2006; 24(29-30):5655-69.
phalitis virus.Biomaterials. 2009; 30(30):6017-25.
intranasal vaccines.Nat Mater. 2010;9(7):572-8.
ery.Curr Drug Deliv.2005; 2(4):401–6.
cations.Biochemistry.1996; 35(3):1027-36.
cules. 2011; 12(12):4373-85.
42 Immune Response Activation
**Chapter 2**
**Mollusk Hemocyanins as Natural Immunostimulants in**
Hemocyanins, the massive oxygen-transporting glycoproteins found freely dissolved in the hemolymph of some mollusks, are potent natural immunostimulants when inoculated in mammals, enhance the innate and adaptive immune response with beneficial clinical outcomes. Hemocyanins are easily purified and molecularly correspond to large multisubunit structures, some over 107 Da (van Holde and Miller, 1995). Currently, hemocya‐ nins are commonly used as carriers/adjuvants for producing antibodies against different antigens. These antigens include tumor-associated antigens (TAAs), such as glycolipid and glycoprotein (mucin-like) antigens in cutting-edge therapeutic vaccines for cancer, along with idiotypes, the most commonly used tumor antigen to prepare vaccines for nearly all non-Hodgkin lymphomas. Other therapeutic strategies using hemocyanins include their use as adjuvants to disrupt self-tolerance to tumor antigens in the generation of *ex vivo* autologous tumor cell lysate-loaded dendritic cells (DCs) to induce T-cell responses in cancer patients (Del Campo et al., 2011). Furthermore, hemocyanins can be used as non-specific immunostimulants during therapy for recurrent superficial bladder cancer after transure‐ thral surgical resection with negligible toxic side effects, thus making them ideal for long-
Biomedical interest in hemocyanins arose in the 1960s when their remarkable immunogenic‐ ity and immunostimulatory properties in experimental animals, and in human beings, were discovered. The most popular hemocyanin, due to being the earliest discovered and the first used in biomedical studies, is purified from the giant keyhole limpet gastropod *Megathura crenulata* and is commonly known as **KLH** (keyhole limpet hemocyanin) (Harris
> © 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
**Biomedical Applications**
http://dx.doi.org/10.5772/57552
**1. Introduction**
María Inés Becker, Sergio Arancibia, Fabián Salazar,
Miguel Del Campo and Alfredo De Ioannes
Additional information is available at the end of the chapter
term on going treatments (Arancibia et al., 2012b).
### **Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications**
María Inés Becker, Sergio Arancibia, Fabián Salazar, Miguel Del Campo and Alfredo De Ioannes
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/57552
#### **1. Introduction**
Hemocyanins, the massive oxygen-transporting glycoproteins found freely dissolved in the hemolymph of some mollusks, are potent natural immunostimulants when inoculated in mammals, enhance the innate and adaptive immune response with beneficial clinical outcomes. Hemocyanins are easily purified and molecularly correspond to large multisubunit structures, some over 107 Da (van Holde and Miller, 1995). Currently, hemocya‐ nins are commonly used as carriers/adjuvants for producing antibodies against different antigens. These antigens include tumor-associated antigens (TAAs), such as glycolipid and glycoprotein (mucin-like) antigens in cutting-edge therapeutic vaccines for cancer, along with idiotypes, the most commonly used tumor antigen to prepare vaccines for nearly all non-Hodgkin lymphomas. Other therapeutic strategies using hemocyanins include their use as adjuvants to disrupt self-tolerance to tumor antigens in the generation of *ex vivo* autologous tumor cell lysate-loaded dendritic cells (DCs) to induce T-cell responses in cancer patients (Del Campo et al., 2011). Furthermore, hemocyanins can be used as non-specific immunostimulants during therapy for recurrent superficial bladder cancer after transure‐ thral surgical resection with negligible toxic side effects, thus making them ideal for longterm on going treatments (Arancibia et al., 2012b).
Biomedical interest in hemocyanins arose in the 1960s when their remarkable immunogenic‐ ity and immunostimulatory properties in experimental animals, and in human beings, were discovered. The most popular hemocyanin, due to being the earliest discovered and the first used in biomedical studies, is purified from the giant keyhole limpet gastropod *Megathura crenulata* and is commonly known as **KLH** (keyhole limpet hemocyanin) (Harris
© 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
and Markl, 1999). However, the limited resource availability and the growing demand for KLH, in addition to the biodiversity of hemocyanins, have led to a growing interest in identifying new hemocyanin candidates with either similar or improved immunological and biochemical properties. Therefore, today, there are a variety of new gastropod hemocyanins that have been purified and studied biochemically and immunologically, including *Concholepas concholepas* (**CCH**) (De Ioannes et al., 2004; Moltedo et al., 2006), *Fissurella latimarginata* (**FLH**) (Arancibia et al., 2014), *Haliotis tuberculata* (**HtH**) (Keller et al., 1999; Markl et al., 2001) and *Rapana thomasiana* (**RtH**) hemocyanins (Idakieva et al., 1993; Tchorbanov et al., 2008).
In this chapter, we reviewed what is presently known about the uses of mollusk hemocyanins as carriers/adjuvants and non-specific immunostimulants in biomedicine, with a general description of hemocyanin structure and, finally, the current knowledge on the innate and adaptive mechanisms involved in the biomedical efficacy of these colossal proteins. Various authors explain the causes of immunogenicity and immunostimulatory effects of hemocyanins in mammals as due to their extremely xenogenic complex structures, which have abundant carbohydrates, and their symmetry, which resembles an icosahedral virus capsid (D5 point group symmetry, i.e., consisting of 10 equivalent subunits). However, this explanation does not account for the mechanisms involved in the effects of hemocyanin on the mammalian immune system; our investigations have attempted to elucidate such mechanisms.
**Figure 1.** Electron microscopy of negatively stained purified mollusk hemocyanin molecules showing their characteris‐ tic hollow cylinder form. A. Hemocyanin from the Keyhole limpet (*Megathura crenulata*), known as KLH. B. Hemocya‐ nin from *Concholepas concholepas,* known as CCH. C. Hemocyanin from *Fissurella latimarginata,* known as FLH. The images show the top (circles) and lateral (rectangles) views of the molecules. The side views show the characteristic didecameric forms of the protein, with subunits arranged in layers. The arrow shows decamers. The scale bar repre‐
Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications
http://dx.doi.org/10.5772/57552
47
KLH from *Megathura crenulata* has been used for more than 40 years for biomedical and biotechnological applications. In the search for new hemocyanins that might act with similar immunostimulatory properties, the hemocyanins from *Concholepas concholepas* (CCH), *Rapana thomasiana* (RtH) and *Haliotis tuberculata* (HtH) emerge as the most promising. Surprisingly, in this search, we identified a new hemocyanin from *Fissurella latimarginata* (FLH) with superior immunogenicity compared to any other hemocyanin known to date (Arancibia et al., 2014). Background on these hemocyanins and the experimental data that support their potential use
The hemocyanin from the giant keyhole limpet *Megathura crenulata*, which naturally lives on the coast of Southern California to Baja California, is a mixture of two immunologically different didecamers, named KLH1 and KLH2. In fact, diverse electrophoretic analyses have revealed that the KLH preparation is made up of two independent oligomeric isoforms that coexist in variable proportions in the animal's circulation, each composed of one type of subunit, KLH1 (350 KDa) and KLH2 (350 KDa) (Gebauer et al., 1999a; Swerdlow et al., 1996). Immunoelectrophoretic analyses using polyclonal antibodies provided insight into the antigenic individuality of these KLH isoforms, indicating that they are very different struc‐ turally (i.e., they do not display shared epitopes) (Harris and Markl, 1999). Although both KLH subunits are glycosylated, they differ in their oligosaccharide patterns: KLH1 (3.0% carbohy‐ drate, w/w) has higher mannose content than KLH2 (3.4% carbohydrate, w/w), which has more N-acetylgalactosamine than KLH1, in addition to the presence of an O-glycosylation site
as immunostimulants is provided below and summarized in Table 1.
sents 100 nm.
**3. Hemocyanins used in biomedicine**
**3.1.** *Megathura crenulata* **hemocyanin (KLH)**
#### **2. Structure of the mollusk hemocyanins**
Mollusk hemocyanins are enormous glycoproteins (4–8 MDa) formed by an intricate arrange‐ ment of 10 subunits, approximately 350-450 kDa each, that are self-assembled into hollow cylinders 35 nm in diameter and are referred to as **decamers**. This structure is easily observed by negative staining using transmission electron microscopy, as shown in Figure 1. Each subunit consists of 7 or 8 globular folded domains known as functional units (FUs), connected by linker peptide strands of 10–15 amino acid residues. FUs vary in size from 45–55 kDa, and each of them is capable of reversibly binding one oxygen molecule through a pair of copper atoms. In gastropod hemocyanins, the hemocyanins being described in this review, the decamers may associate in pairs to form truly immense molecules called **didecamers** with molecular masses of approximately 9 x 106 Da (Decker et al., 2007; van Holde and Miller, 1995; Markl, 2013).
An additional feature of hemocyanin structures are their carbohydrate contents, which play fundamental roles in their organization and their immunological efficacy (Paccagnella et al., 2004). Gastropod hemocyanins contain variable, heterogeneous N-and O-glycosylation sites, with as high as 9% (w/w) mannose as the most abundant oligosaccharide (Hall and Wood, 1976). Some hemocyanins possess the Thomsen-Friedenreich antigen disaccharide (T antigen disaccharide Galβ1-3GalNAcα1-Ser/Thr), which, together with mannose, may play a role in the immune-related properties of hemocyanins (Geyer et al., 2005; Harris and Markl, 1999), as discussed later.
**Figure 1.** Electron microscopy of negatively stained purified mollusk hemocyanin molecules showing their characteris‐ tic hollow cylinder form. A. Hemocyanin from the Keyhole limpet (*Megathura crenulata*), known as KLH. B. Hemocya‐ nin from *Concholepas concholepas,* known as CCH. C. Hemocyanin from *Fissurella latimarginata,* known as FLH. The images show the top (circles) and lateral (rectangles) views of the molecules. The side views show the characteristic didecameric forms of the protein, with subunits arranged in layers. The arrow shows decamers. The scale bar repre‐ sents 100 nm.
#### **3. Hemocyanins used in biomedicine**
and Markl, 1999). However, the limited resource availability and the growing demand for KLH, in addition to the biodiversity of hemocyanins, have led to a growing interest in identifying new hemocyanin candidates with either similar or improved immunological and biochemical properties. Therefore, today, there are a variety of new gastropod hemocyanins that have been purified and studied biochemically and immunologically, including *Concholepas concholepas* (**CCH**) (De Ioannes et al., 2004; Moltedo et al., 2006), *Fissurella latimarginata* (**FLH**) (Arancibia et al., 2014), *Haliotis tuberculata* (**HtH**) (Keller et al., 1999; Markl et al., 2001) and *Rapana thomasiana* (**RtH**) hemocyanins (Idakieva et al., 1993;
In this chapter, we reviewed what is presently known about the uses of mollusk hemocyanins as carriers/adjuvants and non-specific immunostimulants in biomedicine, with a general description of hemocyanin structure and, finally, the current knowledge on the innate and adaptive mechanisms involved in the biomedical efficacy of these colossal proteins. Various authors explain the causes of immunogenicity and immunostimulatory effects of hemocyanins in mammals as due to their extremely xenogenic complex structures, which have abundant carbohydrates, and their symmetry, which resembles an icosahedral virus capsid (D5 point group symmetry, i.e., consisting of 10 equivalent subunits). However, this explanation does not account for the mechanisms involved in the effects of hemocyanin on the mammalian
immune system; our investigations have attempted to elucidate such mechanisms.
Mollusk hemocyanins are enormous glycoproteins (4–8 MDa) formed by an intricate arrange‐ ment of 10 subunits, approximately 350-450 kDa each, that are self-assembled into hollow cylinders 35 nm in diameter and are referred to as **decamers**. This structure is easily observed by negative staining using transmission electron microscopy, as shown in Figure 1. Each subunit consists of 7 or 8 globular folded domains known as functional units (FUs), connected by linker peptide strands of 10–15 amino acid residues. FUs vary in size from 45–55 kDa, and each of them is capable of reversibly binding one oxygen molecule through a pair of copper atoms. In gastropod hemocyanins, the hemocyanins being described in this review, the decamers may associate in pairs to form truly immense molecules called **didecamers** with
An additional feature of hemocyanin structures are their carbohydrate contents, which play fundamental roles in their organization and their immunological efficacy (Paccagnella et al., 2004). Gastropod hemocyanins contain variable, heterogeneous N-and O-glycosylation sites, with as high as 9% (w/w) mannose as the most abundant oligosaccharide (Hall and Wood, 1976). Some hemocyanins possess the Thomsen-Friedenreich antigen disaccharide (T antigen disaccharide Galβ1-3GalNAcα1-Ser/Thr), which, together with mannose, may play a role in the immune-related properties of hemocyanins (Geyer et al., 2005; Harris and Markl, 1999), as
Da (Decker et al., 2007; van Holde and Miller, 1995;
**2. Structure of the mollusk hemocyanins**
molecular masses of approximately 9 x 106
Markl, 2013).
discussed later.
Tchorbanov et al., 2008).
46 Immune Response Activation
KLH from *Megathura crenulata* has been used for more than 40 years for biomedical and biotechnological applications. In the search for new hemocyanins that might act with similar immunostimulatory properties, the hemocyanins from *Concholepas concholepas* (CCH), *Rapana thomasiana* (RtH) and *Haliotis tuberculata* (HtH) emerge as the most promising. Surprisingly, in this search, we identified a new hemocyanin from *Fissurella latimarginata* (FLH) with superior immunogenicity compared to any other hemocyanin known to date (Arancibia et al., 2014). Background on these hemocyanins and the experimental data that support their potential use as immunostimulants is provided below and summarized in Table 1.
#### **3.1.** *Megathura crenulata* **hemocyanin (KLH)**
The hemocyanin from the giant keyhole limpet *Megathura crenulata*, which naturally lives on the coast of Southern California to Baja California, is a mixture of two immunologically different didecamers, named KLH1 and KLH2. In fact, diverse electrophoretic analyses have revealed that the KLH preparation is made up of two independent oligomeric isoforms that coexist in variable proportions in the animal's circulation, each composed of one type of subunit, KLH1 (350 KDa) and KLH2 (350 KDa) (Gebauer et al., 1999a; Swerdlow et al., 1996). Immunoelectrophoretic analyses using polyclonal antibodies provided insight into the antigenic individuality of these KLH isoforms, indicating that they are very different struc‐ turally (i.e., they do not display shared epitopes) (Harris and Markl, 1999). Although both KLH subunits are glycosylated, they differ in their oligosaccharide patterns: KLH1 (3.0% carbohy‐ drate, w/w) has higher mannose content than KLH2 (3.4% carbohydrate, w/w), which has more N-acetylgalactosamine than KLH1, in addition to the presence of an O-glycosylation site (Idakieva et al., 2004). Whether or not KLH1 and KLH2 have similar effects on the immune response has not been comprehensively analyzed. Currently, the variable proportions of KLH subunits in different preparations of the agent, in addition to its propensity to precipitate (Gathuru et al., 2005), can be problematic for some biomedical applications in which the components should be invariant and precisely defined. In addition, the KLH genes have been cloned and sequenced, and the complete amino acid sequences of its subunits are now known; however, until now, it has not been possible to express a heterologous KLH (Altenhein et al., 2002; Gatsogiannis and Markl, 2009; Markl et al., 2001).
et al., 2008; Matus et al., 2007; Mura et al., 2002). Additionally, CCH has been used as carrier of a gonadotropin-releasing hormone (GnRH) in a contraceptive vaccine to control reproduc‐ tion in deer, providing a longer-lasting contraceptive effect (Miller et al., 2008; Pilon et al., 2007). Furthermore, CCH has been pre-clinically evaluated in a murine experimental model of superficial bladder cancer and may be considered a safe alternative therapy to KLH (Atala,
Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications
http://dx.doi.org/10.5772/57552
49
Finally, the most important support for clinical attention to CCH in future biomedical developments includes the results of a recent study of its use as an adjuvant in a DC-based cancer vaccine for castration-resistant prostate cancer (CRPC) patients. This study demon‐ strated that CCH was able to induce an immune memory response, as measured by the delayed-type hypersensitivity (DTH) test, and did not produce toxic or allergic adverse reactions when administered subcutaneously in the patients. These results led to the conclu‐ sion that CCH may be considered as an alternative to KLH for providing safe and effective
The hemocyanin from the black keyhole limpet *Fissurella latimarginata* (FLH) was discovered most recently (Arancibia et al., 2014). The experimental data demonstrate that FLH didecamers are composed of a single type of polypeptide with a molecular mass of approximately 350 kDa. Although the total carbohydrate content has not yet been estimated, carbohydrate staining with specific lectins has shown that FLH has exposed N-and O-linked oligosaccharides, similar to KLH (Arancibia et al., 2014). The evaluation of the humoral immune responses in different mouse strains immunized with CCH, FLH and KLH indicated that FLH is intrinsically more immunogenic than CCH and KLH and reaches titers an order of magnitude higher than those of CCH and KLH. Moreover, FLH had potent *in vivo* anti-tumor activity against melanoma in the B16F10 mouse model, delaying tumor growth and increasing the survival of mice chal‐ lenged with these cells in a prophylactic setting (i.e., with the aim of preventing tumor growth). The most striking effect was observed in a therapeutic setting (specifically, therapy for established tumors in animals without previous FLH priming) (Arancibia et al., 2014). To elucidate the early immunologic mechanisms involved in this anti-tumor effect, we investi‐ gated the effect of FLH on murine DCs cultured *in vitro*. These studies demonstrated that FLH, but not CCH or KLH, is able to rapidly induce the secretion of certain pro-inflammatory cytokines, including interleukin (IL)-6, tumor necrosis factor (TNF)-α, IL-12p40 and IL-23α, a phenomenon that may explain its enhanced immunological activities (Arancibia et al., 2014).
The hemocyanin from the European abalone *Haliotis tuberculata* (**HtH**), a species under commercial aquaculture conditions, is composed of two types of didecamers that coexist in the animal's circulation, HtH1 and HtH2. Immunological analysis demonstrated that these two didecamers are closely related (Keller et al., 1999). These isoforms are formed by the HtH1 and HtH2 subunits, and each has a molecular mass of approximately 392 kDa (Altenhein et al., 2002; Lieb et al., 2000). The carbohydrate content of HtH is 4.5% (w/w); a highly heteroge‐ neous group of structures with appreciable amounts of 3-o-methyl-d-mannose and 3-o-
2006; Moltedo et al., 2006), as described later.
adjuvanticity in cancer vaccines (Reyes et al., 2013).
**3.3. FLH**
**3.4. HtH**
In the 1970s, KLH was introduced in the clinic as part of a test to evaluate immunocompetence in immunosuppressed patients (Curtis et al., 1971; Curtis et al., 1970); thus, its benefits for the treatment of recurrent superficial bladder cancer were discovered serendipitously (Olsson et al., 1974), as described later. In addition, at that time, through the coupling of small molecules (or haptens, defined as substances unable to induce antibodies by themselves) to KLH, it was possible to obtain specific antibodies against them. Thereby, KLH has become, and remains to this day, the most used carrier protein to produce polyclonal and monoclonal antibodies against small molecules, including peptides, drugs, hormones, toxins, antibiotics and countless chemicals. The optimal carrier properties of KLH add to its intrinsic immunostimulatory effects, leading to its current use in the development of several therapeutic vaccines for cancer through the generation of antibodies against TAAs; these antibodies act to eliminate cancer recurrence caused by circulating tumor cells and micrometastases. Moreover, KLH, when conjugated with idiotype antibodies, can induce strong anti-idiotypic antibody responses and cell-mediated responses to tumor antigen(s) *in vivo*, which has resulted in an objective outcome in patients with B-cell lymphoma (Nelson et al., 1996). In addition, the adjuvant/immunosti‐ mulatory properties of KLH have been amply supported through its use in vaccine studies as an immunological tracer protein due to its neo-antigen character. Thus, this protein serves as a strong "surrogate" antigen and an immunogenic "marker" for immunization studies using DC-based vaccines (Shimizu et al., 2001).
#### **3.2.** *Concholepas concholepas* **hemocyanin (CCH)**
The hemocyanin from the edible gastropod *Concholepas concholepas*, a specie distributed on the coasts of Chile and southern Perú, is structurally distinct from traditional KLH. The decamers are formed by two subunits, named CCHA (405 KDa) and CCHB (350 kDa), which are intermingled in the molecule and form heterodecamers. Consequently, their association in pairs results in heterodidecamers (De Ioannes et al., 2004). The carbohydrate content has been determined and has demonstrated that each subunit is differently glycosylated. Thus, CCHA (3.6% carbohydrate, w/w) has N-linked and O-linked glycans; CCHB (2.5% carbohydrate, w/w) has only sugar with N-linkages, while O-linked moieties are nearly absent (Becker et al., 2009). A feature that distinguishes CCH from the remaining hemocyanins is its great stability and solubility. In fact, in contrast with other hemocyanins, after its purification from the hemolymph, the stabilization of CCH does not require additional divalent cations, such as Ca2+and/or Mg2+, in the storage media to maintain its structure (De Ioannes et al., 2004). This quality facilitates chemical coupling reactions when CCH is used as a carrier protein. Thus, CCH has been successfully used as a carrier protein to generate antibodies against hapten molecules (Becker et al., 1998; Torres et al., 1999) and peptides (Duvillie et al., 2003; Grenegard et al., 2008; Matus et al., 2007; Mura et al., 2002). Additionally, CCH has been used as carrier of a gonadotropin-releasing hormone (GnRH) in a contraceptive vaccine to control reproduc‐ tion in deer, providing a longer-lasting contraceptive effect (Miller et al., 2008; Pilon et al., 2007). Furthermore, CCH has been pre-clinically evaluated in a murine experimental model of superficial bladder cancer and may be considered a safe alternative therapy to KLH (Atala, 2006; Moltedo et al., 2006), as described later.
Finally, the most important support for clinical attention to CCH in future biomedical developments includes the results of a recent study of its use as an adjuvant in a DC-based cancer vaccine for castration-resistant prostate cancer (CRPC) patients. This study demon‐ strated that CCH was able to induce an immune memory response, as measured by the delayed-type hypersensitivity (DTH) test, and did not produce toxic or allergic adverse reactions when administered subcutaneously in the patients. These results led to the conclu‐ sion that CCH may be considered as an alternative to KLH for providing safe and effective adjuvanticity in cancer vaccines (Reyes et al., 2013).
#### **3.3. FLH**
(Idakieva et al., 2004). Whether or not KLH1 and KLH2 have similar effects on the immune response has not been comprehensively analyzed. Currently, the variable proportions of KLH subunits in different preparations of the agent, in addition to its propensity to precipitate (Gathuru et al., 2005), can be problematic for some biomedical applications in which the components should be invariant and precisely defined. In addition, the KLH genes have been cloned and sequenced, and the complete amino acid sequences of its subunits are now known; however, until now, it has not been possible to express a heterologous KLH (Altenhein et al.,
In the 1970s, KLH was introduced in the clinic as part of a test to evaluate immunocompetence in immunosuppressed patients (Curtis et al., 1971; Curtis et al., 1970); thus, its benefits for the treatment of recurrent superficial bladder cancer were discovered serendipitously (Olsson et al., 1974), as described later. In addition, at that time, through the coupling of small molecules (or haptens, defined as substances unable to induce antibodies by themselves) to KLH, it was possible to obtain specific antibodies against them. Thereby, KLH has become, and remains to this day, the most used carrier protein to produce polyclonal and monoclonal antibodies against small molecules, including peptides, drugs, hormones, toxins, antibiotics and countless chemicals. The optimal carrier properties of KLH add to its intrinsic immunostimulatory effects, leading to its current use in the development of several therapeutic vaccines for cancer through the generation of antibodies against TAAs; these antibodies act to eliminate cancer recurrence caused by circulating tumor cells and micrometastases. Moreover, KLH, when conjugated with idiotype antibodies, can induce strong anti-idiotypic antibody responses and cell-mediated responses to tumor antigen(s) *in vivo*, which has resulted in an objective outcome in patients with B-cell lymphoma (Nelson et al., 1996). In addition, the adjuvant/immunosti‐ mulatory properties of KLH have been amply supported through its use in vaccine studies as an immunological tracer protein due to its neo-antigen character. Thus, this protein serves as a strong "surrogate" antigen and an immunogenic "marker" for immunization studies using
The hemocyanin from the edible gastropod *Concholepas concholepas*, a specie distributed on the coasts of Chile and southern Perú, is structurally distinct from traditional KLH. The decamers are formed by two subunits, named CCHA (405 KDa) and CCHB (350 kDa), which are intermingled in the molecule and form heterodecamers. Consequently, their association in pairs results in heterodidecamers (De Ioannes et al., 2004). The carbohydrate content has been determined and has demonstrated that each subunit is differently glycosylated. Thus, CCHA (3.6% carbohydrate, w/w) has N-linked and O-linked glycans; CCHB (2.5% carbohydrate, w/w) has only sugar with N-linkages, while O-linked moieties are nearly absent (Becker et al., 2009). A feature that distinguishes CCH from the remaining hemocyanins is its great stability and solubility. In fact, in contrast with other hemocyanins, after its purification from the hemolymph, the stabilization of CCH does not require additional divalent cations, such as Ca2+and/or Mg2+, in the storage media to maintain its structure (De Ioannes et al., 2004). This quality facilitates chemical coupling reactions when CCH is used as a carrier protein. Thus, CCH has been successfully used as a carrier protein to generate antibodies against hapten molecules (Becker et al., 1998; Torres et al., 1999) and peptides (Duvillie et al., 2003; Grenegard
2002; Gatsogiannis and Markl, 2009; Markl et al., 2001).
48 Immune Response Activation
DC-based vaccines (Shimizu et al., 2001).
**3.2.** *Concholepas concholepas* **hemocyanin (CCH)**
The hemocyanin from the black keyhole limpet *Fissurella latimarginata* (FLH) was discovered most recently (Arancibia et al., 2014). The experimental data demonstrate that FLH didecamers are composed of a single type of polypeptide with a molecular mass of approximately 350 kDa. Although the total carbohydrate content has not yet been estimated, carbohydrate staining with specific lectins has shown that FLH has exposed N-and O-linked oligosaccharides, similar to KLH (Arancibia et al., 2014). The evaluation of the humoral immune responses in different mouse strains immunized with CCH, FLH and KLH indicated that FLH is intrinsically more immunogenic than CCH and KLH and reaches titers an order of magnitude higher than those of CCH and KLH. Moreover, FLH had potent *in vivo* anti-tumor activity against melanoma in the B16F10 mouse model, delaying tumor growth and increasing the survival of mice chal‐ lenged with these cells in a prophylactic setting (i.e., with the aim of preventing tumor growth). The most striking effect was observed in a therapeutic setting (specifically, therapy for established tumors in animals without previous FLH priming) (Arancibia et al., 2014). To elucidate the early immunologic mechanisms involved in this anti-tumor effect, we investi‐ gated the effect of FLH on murine DCs cultured *in vitro*. These studies demonstrated that FLH, but not CCH or KLH, is able to rapidly induce the secretion of certain pro-inflammatory cytokines, including interleukin (IL)-6, tumor necrosis factor (TNF)-α, IL-12p40 and IL-23α, a phenomenon that may explain its enhanced immunological activities (Arancibia et al., 2014).
#### **3.4. HtH**
The hemocyanin from the European abalone *Haliotis tuberculata* (**HtH**), a species under commercial aquaculture conditions, is composed of two types of didecamers that coexist in the animal's circulation, HtH1 and HtH2. Immunological analysis demonstrated that these two didecamers are closely related (Keller et al., 1999). These isoforms are formed by the HtH1 and HtH2 subunits, and each has a molecular mass of approximately 392 kDa (Altenhein et al., 2002; Lieb et al., 2000). The carbohydrate content of HtH is 4.5% (w/w); a highly heteroge‐ neous group of structures with appreciable amounts of 3-o-methyl-d-mannose and 3-o-
ND: Not determinate
methyl-d-galactose has been identified in HtH (Idakieva et al., 2004; Velkova et al.). It is important to note that, although there are numerous studies on the biochemical characteristics of HtH, comprehensive studies about its immunologic and therapeutic properties are essen‐ tially absent. However, a stable preparation of abalone hemocyanin with high antiviral activity has recently been formulated and needs to be evaluated as therapeutic agent in future clinical
Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications
http://dx.doi.org/10.5772/57552
51
The hemocyanin of the marine snail *Rapana thomasiana* (RtH), a species native of the China Sea and transferred to the West coast of the Black Sea, was described in the 1990s (Idakieva et al., 1993). RtH is a mixture of two hemocyanin isoforms, RtH1 and RtH2, that coexist in the animal's circulation (Gebauer et al., 1999b). Both have carbohydrate contents of 2.6% (w/w), with very similar monosaccharide compositions featuring type N-glycosylation (Idakieva et al., 2004). The biomedical applications of RtH as an adjuvant have been explored and have demonstrated that mice immunized with influenza vaccine developed a specific humoral and cellular immune response characterized by the induction of specific antibodies to viral proteins and a cytotoxic response lasting at least 5 months (Gesheva et al., 2011). Furthermore, antimi‐ crobial activities against different strains of the genital herpes simplex virus have been reported
**4. Hemocyanins as non-specific immunostimulants in recurrent superficial**
A significant decrease in the recurrence of superficial bladder cancer in 10 patients with transitional cell carcinoma (TCC) treated with KLH without toxic side effects was reported early on in the field of hemocyanin research (Olsson et al., 1974). This study was followed by tens of papers, which collectively provided encouraging support for the use of KLH and CCH as alternatives in SBC immunotherapy. The most frequently used immunotherapeutic agent to prevent recurrence of TCC is the Bacillus Calmette-Guerin (BCG) vaccine for tuberculosis (Morales et al., 1976). This attenuated form of *Mycobacterium bovis* has also been used as a therapy in other types of cancers (Edwards and Whitwell, 1974; Milas and Withers, 1976). However, despite the successful results (Nseyo and Lamm, 1997), the severe side effects of the
Three different *in vivo* models have been primarily used to evaluate the therapeutic properties of hemocyanins in SBC, with the therapy being administered either subcutaneously, intraper‐ itoneally or intravesically according to the site where the tumor was established (Lamm, 2003; Schenkman and Lamm, 2004). The mouse bladder tumor-2 cell (MBT-2) transplantable murine model of SBC was the first of these models to be developed in 1981 (Lamm et al., 1981). Mice were pre-immunized with 200 μg of KLH after a previous subcutaneous inocula‐
BCG vaccine have led scientists to explore new immunostimulants (Lamm, 2003).
applications (Zanjani et al., 2013).
for the RtH2 subunit (Genova-Kalou et al., 2008).
**bladder cancer (SBC)**
**4.1. Experimental studies**
**3.5. RtH**
Table 1. Hemocyanins available today
**Table 1.** Hemocyanins available today
methyl-d-galactose has been identified in HtH (Idakieva et al., 2004; Velkova et al.). It is important to note that, although there are numerous studies on the biochemical characteristics of HtH, comprehensive studies about its immunologic and therapeutic properties are essen‐ tially absent. However, a stable preparation of abalone hemocyanin with high antiviral activity has recently been formulated and needs to be evaluated as therapeutic agent in future clinical applications (Zanjani et al., 2013).
#### **3.5. RtH**
The hemocyanin of the marine snail *Rapana thomasiana* (RtH), a species native of the China Sea and transferred to the West coast of the Black Sea, was described in the 1990s (Idakieva et al., 1993). RtH is a mixture of two hemocyanin isoforms, RtH1 and RtH2, that coexist in the animal's circulation (Gebauer et al., 1999b). Both have carbohydrate contents of 2.6% (w/w), with very similar monosaccharide compositions featuring type N-glycosylation (Idakieva et al., 2004). The biomedical applications of RtH as an adjuvant have been explored and have demonstrated that mice immunized with influenza vaccine developed a specific humoral and cellular immune response characterized by the induction of specific antibodies to viral proteins and a cytotoxic response lasting at least 5 months (Gesheva et al., 2011). Furthermore, antimi‐ crobial activities against different strains of the genital herpes simplex virus have been reported for the RtH2 subunit (Genova-Kalou et al., 2008).
### **4. Hemocyanins as non-specific immunostimulants in recurrent superficial bladder cancer (SBC)**
A significant decrease in the recurrence of superficial bladder cancer in 10 patients with transitional cell carcinoma (TCC) treated with KLH without toxic side effects was reported early on in the field of hemocyanin research (Olsson et al., 1974). This study was followed by tens of papers, which collectively provided encouraging support for the use of KLH and CCH as alternatives in SBC immunotherapy. The most frequently used immunotherapeutic agent to prevent recurrence of TCC is the Bacillus Calmette-Guerin (BCG) vaccine for tuberculosis (Morales et al., 1976). This attenuated form of *Mycobacterium bovis* has also been used as a therapy in other types of cancers (Edwards and Whitwell, 1974; Milas and Withers, 1976). However, despite the successful results (Nseyo and Lamm, 1997), the severe side effects of the BCG vaccine have led scientists to explore new immunostimulants (Lamm, 2003).
#### **4.1. Experimental studies**
Table 1. Hemocyanins available today
**Table 1.** Hemocyanins available today
**PROPERTIES FEATURE KLH CCH FLH HtH RtH**
50 Immune Response Activation
**Taxonomy** Phylum Mollusk Mollusk Mollusk Mollusk Mollusk
Order Gastropoda Gastropoda Gastropoda Gastropoda Gastropoda
Family Fissurellidae Muricidae Fissurellidae Haliotidae Muricidae
Genus and specie *Megathura crenulata Concholepas concholepas Fissurella latimarginata Haliotis tuberculata Rapana thomasiana*
Pacific coast of Perú and
Waters of the Channel
Japan sea , and the East
China Sea
Islands
Chile
**Distribution** Habitat Coast of California Pacific coast of Perú and
Chile
**Biochemical** Molecular weights ≈ 8 MDa ≈ 8 MDa ≈ 8 MDa ≈ 8 MDa ≈ 8 MDa
Size of molecule ≈ 360 Å Ø 400Å height ≈ 325Å Ø 392Å height ≈350 Å Ø 370 Å height ≈ 350 Å Ø 380Å height ≈ 350 Å Ø 380Å height
Subunit organization Homodecamer Heterodecamer Homodecamer Homodecamer Homodecamer
Required Ca+2 and Mg+2 Not required Ca+2 and
Required Ca+2 and Mg+2 Required Ca+2 and Mg+2 Required Ca+2 and Mg+2
Mg+2
Isoelectric point 5,7 6,0 ND ND ND
% Oligosaccharides 3,2% 3,1% ND 4,5% 2,6%
Type of Oligosaccharides N- and O-linked N- and O-linked N- and O-linked N- and O-linked N- and O-linked
**Immunological** Immunogenicity High High Very high ND ND
Yes Yes ND ND ND
Yes Yes Yes ND ND
Evaluated as hapten carrier Yes Yes Yes ND ND
Yes Yes ND ND ND
Yes No No No No
Antitumor effects in
animals models of SBC
Antitumor effects in
animals models of
melanoma
**Biomedical uses**
Used as adjuvant in DC
vaccines
Used as immunostimulant
in bladder cancer
ND: Not determinate
Stability structure in
solution
Three different *in vivo* models have been primarily used to evaluate the therapeutic properties of hemocyanins in SBC, with the therapy being administered either subcutaneously, intraper‐ itoneally or intravesically according to the site where the tumor was established (Lamm, 2003; Schenkman and Lamm, 2004). The mouse bladder tumor-2 cell (MBT-2) transplantable murine model of SBC was the first of these models to be developed in 1981 (Lamm et al., 1981). Mice were pre-immunized with 200 μg of KLH after a previous subcutaneous inocula‐ tion with MBT-2 and were then intralesionally immunized with 50 μg at one and seven days after implantation, leading to a significant decrease in tumor growth and a prolongation of animal survival (Lamm et al., 1981). Further studies showed that priming was essential for achieving a therapeutic effect (Lamm et al., 1993b; Lamm et al., 1982; Marsh et al., 1987). However, other researchers have studied KLH immunotherapy in the same model without promising results. The priming and transplantation of MBT-2 tumor cells, either subcutane‐ ously or into the bladder, with immunotherapy of 50 or 200 μg of KLH did not yield different results compared to controls (Walsh et al., 1983). In a later study, the administration routes of KLH were compared. Without priming, the intralesional route was more effective than intraperitoneal administration in terms of inhibiting tumor growth (Lau et al., 1986). After that, an additive effect was observed in animals that received KLH and interferon (IFN)-α intra‐ peritoneally without prior immunization, compared with KLH and IFN-α alone (Riggs et al., 1992). However, in a subsequent study, the presence of endotoxin in KLH preparations partially accounted for its anti-tumor effect (Lamm et al., 1993a).
KLH had significantly reduced tumor recurrence rates over a period of two years (Olsson et al., 1974). More than 10 years later, a controlled study corroborated that KLH was more effective than mitomycin C (Jurincic et al., 1988). Then, it was determined that the incidence of recurrence in patients with TCC related to urinary schistosomiasis was diminished after KLH treatment (Wishahi et al., 1995). However, studies with patients who were unresponsive to chemotherapeutic agents (Flamm et al., 1990) or who had carcinoma *in situ* (CIS) (Jurincic-Winkler et al., 1995c; Jurincic-Winkler et al., 2000) have not shown any significant effects after
Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications
http://dx.doi.org/10.5772/57552
53
A Phase III clinical trial was conducted by Intracel Resources (USA) to assess the safety and effectiveness of the KLH BCI-Immune Activator compared to doxorubicin in BCG-intolerant or refractory patients with CIS with or without resected SBC. Nevertheless, although the study was completed, no results have been posted. Recently, Biosyn (USA) evaluated the efficacy and adverse effects of IMMUCOTHEL®, a clinical-grade KLH preparation developed by Biosyn Arzneimittel GmbH (Germany). Their outcomes showed almost no KLH-induced adverse effects, and the efficacy of IMMUCOTHEL® was comparable to mitomycin C.
Due to some of the successful results using hemocyanins in SBC therapy, their therapeutic effects have been evaluated in other cancers, such as melanoma, breast, pancreatic, esophageal
The effectiveness of KLH has been demonstrated in diverse human cancer cell lines, such as estrogen-independent breast (ZR75-1), estrogen-dependent breast (MCF-7), esophagus (SEG-1 and BIC-1), pancreas (PANC-1), melanoma (HTB68 and HTB72) and prostate (DU145), through the inhibition of cellular growth in an apoptotic-dependent or independent manner (McFadden et al., 2003; Riggs et al., 2002; Riggs et al., 2005; Somasundar et al., 2005). In addition, KLH was shown to exert an additive effect with IL-2 and IFN-α in a combined therapy against melanoma, which encourages the use of this type of bivalent therapy as an effective treatment
Diverse hemocyanins have been evaluated as a therapy in murine models of melanoma. First, the combined effect of KLH with IL-2 and IFN-α was evaluated in an HTB68 mouse melanoma model. KLH augmented the effect of IFN-α, one of the most common immunotherapeutic agents against melanoma (Rizvi et al., 2007). Then, the effect of diverse glycoconjugates (GCs), which induce NK-cell mediated cytotoxicity *in vitro*, were evaluated with or without KLH in a B16F10 murine melanoma model. However, no synergism was observed, which was attributed to the common epitopes shared between the protein and the GCs (Hulikova et al.). Recently, it was demonstrated that an oxidated-modified CCH, which had enhanced structural stability and immunogenicity, has the same anti-tumor effect in the B16F10 melanoma model
against these types of aggressive disease (McFadden et al., 2007).
KLH treatment.
**4.3. Other models**
and prostatic.
*4.3.1. In vitro studies*
*4.3.2. In vivo studies*
The syngeneic orthotopic murine bladder cancer model MB-49 was evaluated in 1994 (Gunther et al., 1999; Swerdlow et al., 1994). Subcutaneous immunization with KLH-Immune Activator, a clinical-grade KLH preparation, two weeks prior to intravesical implantation of tumor cells, followed by intravesical administration of 10 or 100 μg of KLH, resulted in significantly decreased tumor growth (Swerdlow et al., 1994). Prior immunization was required, and no significant histopathological abnormalities were observed. A third model was developed in 1989 that consisted of the induction of a bladder carcinoma in Wistar rats using N-butyl-N-(4 hydroxybutyl) nitrosamine (BBN) (Arentsen et al., 2009). Subcutaneous sensitization with 1 mg of KLH, followed by intravesical and subcutaneous administration twice weekly, led to a decrease of BBN-induced bladder tumors (Recker and Rubben, 1989). A subsequent study demonstrated that subcutaneous administration was more effective than intravesical instilla‐ tion in terms of tumor growth and rats bearing tumors (Linn et al., 2000).
Finally, more recent studies have evaluated the effect of CCH in the MBT-2 murine model (Becker et al., 2009; Moltedo et al., 2006). Priming with CCH previous subcutaneous implan‐ tation of tumor cells, succeeded by subcutaneous immunization with 1 or 0.1 mg of CCH, led to decreases in the incidence and growth of tumors, prolonged survival and the absence of toxic effects (Moltedo et al., 2006). In a subsequent study of the contributions of CCH-A and CCH-B subunits, it was determined that each subunit alone has an anti-tumor effect. However, in terms of tumor incidence and animal survival, CCH-A had the maximum effect (Becker et al., 2009).
#### **4.2. Clinical studies**
As a result of complications of complete bladder removal and the possibility of cancer recurrence, some SBCs have been treated with intravesical administration of biological and chemotherapeutic agents in the initial stages of the disease to either treat an established tumor or avoid progression and recurrence after transurethral resection (Perabo and Muller, 2004). The first study in humans was performed in 1974, in which 10 immunocompetent patients subcutaneously primed with 5 mg of KLH and followed with immunization using 200 μg of KLH had significantly reduced tumor recurrence rates over a period of two years (Olsson et al., 1974). More than 10 years later, a controlled study corroborated that KLH was more effective than mitomycin C (Jurincic et al., 1988). Then, it was determined that the incidence of recurrence in patients with TCC related to urinary schistosomiasis was diminished after KLH treatment (Wishahi et al., 1995). However, studies with patients who were unresponsive to chemotherapeutic agents (Flamm et al., 1990) or who had carcinoma *in situ* (CIS) (Jurincic-Winkler et al., 1995c; Jurincic-Winkler et al., 2000) have not shown any significant effects after KLH treatment.
A Phase III clinical trial was conducted by Intracel Resources (USA) to assess the safety and effectiveness of the KLH BCI-Immune Activator compared to doxorubicin in BCG-intolerant or refractory patients with CIS with or without resected SBC. Nevertheless, although the study was completed, no results have been posted. Recently, Biosyn (USA) evaluated the efficacy and adverse effects of IMMUCOTHEL®, a clinical-grade KLH preparation developed by Biosyn Arzneimittel GmbH (Germany). Their outcomes showed almost no KLH-induced adverse effects, and the efficacy of IMMUCOTHEL® was comparable to mitomycin C.
#### **4.3. Other models**
tion with MBT-2 and were then intralesionally immunized with 50 μg at one and seven days after implantation, leading to a significant decrease in tumor growth and a prolongation of animal survival (Lamm et al., 1981). Further studies showed that priming was essential for achieving a therapeutic effect (Lamm et al., 1993b; Lamm et al., 1982; Marsh et al., 1987). However, other researchers have studied KLH immunotherapy in the same model without promising results. The priming and transplantation of MBT-2 tumor cells, either subcutane‐ ously or into the bladder, with immunotherapy of 50 or 200 μg of KLH did not yield different results compared to controls (Walsh et al., 1983). In a later study, the administration routes of KLH were compared. Without priming, the intralesional route was more effective than intraperitoneal administration in terms of inhibiting tumor growth (Lau et al., 1986). After that, an additive effect was observed in animals that received KLH and interferon (IFN)-α intra‐ peritoneally without prior immunization, compared with KLH and IFN-α alone (Riggs et al., 1992). However, in a subsequent study, the presence of endotoxin in KLH preparations
The syngeneic orthotopic murine bladder cancer model MB-49 was evaluated in 1994 (Gunther et al., 1999; Swerdlow et al., 1994). Subcutaneous immunization with KLH-Immune Activator, a clinical-grade KLH preparation, two weeks prior to intravesical implantation of tumor cells, followed by intravesical administration of 10 or 100 μg of KLH, resulted in significantly decreased tumor growth (Swerdlow et al., 1994). Prior immunization was required, and no significant histopathological abnormalities were observed. A third model was developed in 1989 that consisted of the induction of a bladder carcinoma in Wistar rats using N-butyl-N-(4 hydroxybutyl) nitrosamine (BBN) (Arentsen et al., 2009). Subcutaneous sensitization with 1 mg of KLH, followed by intravesical and subcutaneous administration twice weekly, led to a decrease of BBN-induced bladder tumors (Recker and Rubben, 1989). A subsequent study demonstrated that subcutaneous administration was more effective than intravesical instilla‐
Finally, more recent studies have evaluated the effect of CCH in the MBT-2 murine model (Becker et al., 2009; Moltedo et al., 2006). Priming with CCH previous subcutaneous implan‐ tation of tumor cells, succeeded by subcutaneous immunization with 1 or 0.1 mg of CCH, led to decreases in the incidence and growth of tumors, prolonged survival and the absence of toxic effects (Moltedo et al., 2006). In a subsequent study of the contributions of CCH-A and CCH-B subunits, it was determined that each subunit alone has an anti-tumor effect. However, in terms of tumor incidence and animal survival, CCH-A had the maximum effect (Becker et
As a result of complications of complete bladder removal and the possibility of cancer recurrence, some SBCs have been treated with intravesical administration of biological and chemotherapeutic agents in the initial stages of the disease to either treat an established tumor or avoid progression and recurrence after transurethral resection (Perabo and Muller, 2004). The first study in humans was performed in 1974, in which 10 immunocompetent patients subcutaneously primed with 5 mg of KLH and followed with immunization using 200 μg of
partially accounted for its anti-tumor effect (Lamm et al., 1993a).
tion in terms of tumor growth and rats bearing tumors (Linn et al., 2000).
al., 2009).
**4.2. Clinical studies**
52 Immune Response Activation
Due to some of the successful results using hemocyanins in SBC therapy, their therapeutic effects have been evaluated in other cancers, such as melanoma, breast, pancreatic, esophageal and prostatic.
#### *4.3.1. In vitro studies*
The effectiveness of KLH has been demonstrated in diverse human cancer cell lines, such as estrogen-independent breast (ZR75-1), estrogen-dependent breast (MCF-7), esophagus (SEG-1 and BIC-1), pancreas (PANC-1), melanoma (HTB68 and HTB72) and prostate (DU145), through the inhibition of cellular growth in an apoptotic-dependent or independent manner (McFadden et al., 2003; Riggs et al., 2002; Riggs et al., 2005; Somasundar et al., 2005). In addition, KLH was shown to exert an additive effect with IL-2 and IFN-α in a combined therapy against melanoma, which encourages the use of this type of bivalent therapy as an effective treatment against these types of aggressive disease (McFadden et al., 2007).
#### *4.3.2. In vivo studies*
Diverse hemocyanins have been evaluated as a therapy in murine models of melanoma. First, the combined effect of KLH with IL-2 and IFN-α was evaluated in an HTB68 mouse melanoma model. KLH augmented the effect of IFN-α, one of the most common immunotherapeutic agents against melanoma (Rizvi et al., 2007). Then, the effect of diverse glycoconjugates (GCs), which induce NK-cell mediated cytotoxicity *in vitro*, were evaluated with or without KLH in a B16F10 murine melanoma model. However, no synergism was observed, which was attributed to the common epitopes shared between the protein and the GCs (Hulikova et al.). Recently, it was demonstrated that an oxidated-modified CCH, which had enhanced structural stability and immunogenicity, has the same anti-tumor effect in the B16F10 melanoma model as its native counterpart (Arancibia et al., 2012a). In addition, we have recently demonstrated the outstanding immunomodulatory properties of FLH in a B16F10 murine melanoma model. FLH promotes significantly higher antibody titers than CCH and KLH, in addition to exhib‐ iting potent anti-tumor activity that delays the growth of B16F10 melanoma cells and prolongs murine survival (Arancibia et al., 2014).
mimetic peptide was coupled to KLH; immunized mice were able to inhibit *C. burnetii* infection
Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications
http://dx.doi.org/10.5772/57552
55
For papillomavirus, peptide fragments were used that mimicked B-cell epitopes of the capsid protein L1 of human papillomavirus (HPV) type 31 coupled to KLH, generating specific antipeptide antibodies in mice (Andreev et al., 2012). For influenza, a conjugate vaccine consisting of the peptide of the highly conserved M2 membrane protein coupled to KLH was highly immunogenic and able to confer protection against lethal challenge with either H1N1 or H3N1 virus in mice (Fan et al., 2004; Kiraly et al., 2011). For HIV-1, aiming to develop neutralizing antibodies in a caprine model, KLH was coupled with a synthetic peptide representative of the p17 functional epitope (AT20) derived from HIV-1 or with MPR peptide from the gp41 membrane proximal region to prevent transmission of the virus through colostrum. This conjugated vaccine effectively induced specific sIgA and IgG in the colostrum of a lactating
For *Cryptococcus neoformans,* which causes meningoencephalitis in AIDS patients, one compo‐ nent of the cryptococcal capsular polysaccharide was coupled to KLH, producing antibodies that were protective against the pathogen (May et al., 2003). However, a protective epitope of *Candida albicans* conjugated to KLH induced a Th1-type cytokine expression pattern in C57BL/
The antitumoral vaccines are intended to treat an existing cancer by strengthening the body's natural defenses against cancer and have emerged as alternatives to anti-proliferative treat‐ ments, such as chemo-or radio-therapy. The design of such vaccines is focused on the search for specific epitopes in tumor cells to raise antibodies against them. These epitopes are usually aberrant branches of polysaccharides located on the cell surface. Some of them are glycolipids, such as Globo H, Lewis Y, GM2, GD2, GD3 and fucosyl-GM1, and others are glycoproteins, such as MUC-1, Tn, sialyl-Tn and TF (Galonic and Gin, 2007). The aim of these vaccines is to transform these tumor antigens into immunogens powerful enough to achieve an immune response. Furthermore, hemocyanins are used as immunomodulatory agents in DC vaccines in which these autologous presenting cells are loaded with tumor antigen, whether as tumor lysate, recombinant antigen or transfecting cell tumor RNA, in the presence of hemocyanins (Steinman and Banchereau, 2007). In this case, the patient's responses are measured by DTH tests against hemocyanins or tumor extract (Escobar et al., 2005). In addition to the extensive use of KLH for this purpose, promising results have been obtained with CCH as an adjuvant in a DC vaccine administered to patients with CRPC (Reyes et al., 2013). Next, we review selected antitumor vaccine clinical trials with hemocyanins that have been approved by the
and to develop significant protection against *C. burnetii* challenge (Peng et al., 2012).
*5.1.2. Vaccines against viruses*
species.(Dorosko et al., 2008).
*5.1.3. Vaccines against fungi*
6J mice (Su et al., 2007).
**5.2. Antitumor vaccines**
US National Institutes of Health (NIH).
#### **5. Hemocyanins as carriers/adjuvants in vaccines**
Mollusk hemocyanins have been widely used in therapeutic and prophylactic vaccines, enhancing the immune response by building on hemocyanin's carrier properties and adjuvant capacities (Del Campo et al., 2011). In order to generate immunity against antigens, pathogens, TAAs and certain chemical substances, such as commonly abused drugs, it is necessary to chemically couple the compound to the hemocyanin and then immunize the patient. The hemocyanins can also be used as genuine adjuvants to further enhance T cell reactivity to tumors in DC vaccine therapy; in this case, hemocyanins contribute to the reversal of the DC tolerogenic profile in cancer patients toward an immunostimulatory profile (Presicce et al., 2008). Indeed, the generation of an optimal cytotoxic T cell (CTL) immune response requires the presence of T helper lymphocytes (CD4+) and the expression of both helper-and CTLdefined antigen determinants on the same DC (Shimizu et al., 2001; Timmerman and Levy, 2000). Thus, hemocyanins produce a bystander effect, inducing a potent specific memory T cell response associated with the secretion of cytokines that indirectly promote the specific cellular response against the antigen of interest, whether a tumor cell, a pathogen or a deleterious compound.
The following overview offers a brief look at the experimental vaccines that have used hemocyanins in their formulations.
#### **5.1. Vaccines against pathogens**
Pathogenic immunodominant peptides or polysaccharides coupled to hemocyanin have been tested in the development of experimental vaccines that are intended to prevent diseases caused by bacteria, viruses and fungi. We next present some examples.
#### *5.1.1. Vaccines against bacteria*
For *Pseudomonas aeruginosa*, which causes pulmonary infection, the use of synthetic peptides representing surface-exposed, linear B-cell epitopes of outer membrane protein F (Hughes and Gilleland, 1995) or mucoid exopolysaccharide (MEP) coupled to KLH demonstrated that significantly enhances its immunogenicity and the capacity to elicit opsonic antibodies in mice and rabbits (Theilacker et al., 2003). For *Streptococcus pneumonia,* which causes pneumococcal infection, administration of a DNA plasmid encoding the FLt3 ligand gene as a mucosal adjuvant plus phosphorylcholine (PC) conjugated to KLH demonstrated elicitation of PCspecific immune responses at the mucosal and serum levels (Baatarjav et al., 2011). For *Coxiella burnetii,* which causes acute and chronic Q fever in humans, a lipopolysaccharide (LPS) mimetic peptide was coupled to KLH; immunized mice were able to inhibit *C. burnetii* infection and to develop significant protection against *C. burnetii* challenge (Peng et al., 2012).
#### *5.1.2. Vaccines against viruses*
as its native counterpart (Arancibia et al., 2012a). In addition, we have recently demonstrated the outstanding immunomodulatory properties of FLH in a B16F10 murine melanoma model. FLH promotes significantly higher antibody titers than CCH and KLH, in addition to exhib‐ iting potent anti-tumor activity that delays the growth of B16F10 melanoma cells and prolongs
Mollusk hemocyanins have been widely used in therapeutic and prophylactic vaccines, enhancing the immune response by building on hemocyanin's carrier properties and adjuvant capacities (Del Campo et al., 2011). In order to generate immunity against antigens, pathogens, TAAs and certain chemical substances, such as commonly abused drugs, it is necessary to chemically couple the compound to the hemocyanin and then immunize the patient. The hemocyanins can also be used as genuine adjuvants to further enhance T cell reactivity to tumors in DC vaccine therapy; in this case, hemocyanins contribute to the reversal of the DC tolerogenic profile in cancer patients toward an immunostimulatory profile (Presicce et al., 2008). Indeed, the generation of an optimal cytotoxic T cell (CTL) immune response requires the presence of T helper lymphocytes (CD4+) and the expression of both helper-and CTLdefined antigen determinants on the same DC (Shimizu et al., 2001; Timmerman and Levy, 2000). Thus, hemocyanins produce a bystander effect, inducing a potent specific memory T cell response associated with the secretion of cytokines that indirectly promote the specific cellular response against the antigen of interest, whether a tumor cell, a pathogen or a
The following overview offers a brief look at the experimental vaccines that have used
Pathogenic immunodominant peptides or polysaccharides coupled to hemocyanin have been tested in the development of experimental vaccines that are intended to prevent diseases
For *Pseudomonas aeruginosa*, which causes pulmonary infection, the use of synthetic peptides representing surface-exposed, linear B-cell epitopes of outer membrane protein F (Hughes and Gilleland, 1995) or mucoid exopolysaccharide (MEP) coupled to KLH demonstrated that significantly enhances its immunogenicity and the capacity to elicit opsonic antibodies in mice and rabbits (Theilacker et al., 2003). For *Streptococcus pneumonia,* which causes pneumococcal infection, administration of a DNA plasmid encoding the FLt3 ligand gene as a mucosal adjuvant plus phosphorylcholine (PC) conjugated to KLH demonstrated elicitation of PCspecific immune responses at the mucosal and serum levels (Baatarjav et al., 2011). For *Coxiella burnetii,* which causes acute and chronic Q fever in humans, a lipopolysaccharide (LPS)
caused by bacteria, viruses and fungi. We next present some examples.
murine survival (Arancibia et al., 2014).
54 Immune Response Activation
deleterious compound.
hemocyanins in their formulations.
**5.1. Vaccines against pathogens**
*5.1.1. Vaccines against bacteria*
**5. Hemocyanins as carriers/adjuvants in vaccines**
For papillomavirus, peptide fragments were used that mimicked B-cell epitopes of the capsid protein L1 of human papillomavirus (HPV) type 31 coupled to KLH, generating specific antipeptide antibodies in mice (Andreev et al., 2012). For influenza, a conjugate vaccine consisting of the peptide of the highly conserved M2 membrane protein coupled to KLH was highly immunogenic and able to confer protection against lethal challenge with either H1N1 or H3N1 virus in mice (Fan et al., 2004; Kiraly et al., 2011). For HIV-1, aiming to develop neutralizing antibodies in a caprine model, KLH was coupled with a synthetic peptide representative of the p17 functional epitope (AT20) derived from HIV-1 or with MPR peptide from the gp41 membrane proximal region to prevent transmission of the virus through colostrum. This conjugated vaccine effectively induced specific sIgA and IgG in the colostrum of a lactating species.(Dorosko et al., 2008).
#### *5.1.3. Vaccines against fungi*
For *Cryptococcus neoformans,* which causes meningoencephalitis in AIDS patients, one compo‐ nent of the cryptococcal capsular polysaccharide was coupled to KLH, producing antibodies that were protective against the pathogen (May et al., 2003). However, a protective epitope of *Candida albicans* conjugated to KLH induced a Th1-type cytokine expression pattern in C57BL/ 6J mice (Su et al., 2007).
#### **5.2. Antitumor vaccines**
The antitumoral vaccines are intended to treat an existing cancer by strengthening the body's natural defenses against cancer and have emerged as alternatives to anti-proliferative treat‐ ments, such as chemo-or radio-therapy. The design of such vaccines is focused on the search for specific epitopes in tumor cells to raise antibodies against them. These epitopes are usually aberrant branches of polysaccharides located on the cell surface. Some of them are glycolipids, such as Globo H, Lewis Y, GM2, GD2, GD3 and fucosyl-GM1, and others are glycoproteins, such as MUC-1, Tn, sialyl-Tn and TF (Galonic and Gin, 2007). The aim of these vaccines is to transform these tumor antigens into immunogens powerful enough to achieve an immune response. Furthermore, hemocyanins are used as immunomodulatory agents in DC vaccines in which these autologous presenting cells are loaded with tumor antigen, whether as tumor lysate, recombinant antigen or transfecting cell tumor RNA, in the presence of hemocyanins (Steinman and Banchereau, 2007). In this case, the patient's responses are measured by DTH tests against hemocyanins or tumor extract (Escobar et al., 2005). In addition to the extensive use of KLH for this purpose, promising results have been obtained with CCH as an adjuvant in a DC vaccine administered to patients with CRPC (Reyes et al., 2013). Next, we review selected antitumor vaccine clinical trials with hemocyanins that have been approved by the US National Institutes of Health (NIH).
#### **5.3. Breast, ovary, fallopian tube and peritoneal cancers**
Sialyl-Tn (STn) is the tumor antigen most frequently expressed in breast, ovary, colon, rectal, stomach and pancreas adenocarcinomas. This antigen, conjugated to KLH, has been named Theratope and is capable of generating positive serological results, inducing significant titers of IgM and IgG antibodies against STn (Gilewski et al., 2007). The results of a Phase III multicenter clinical trial, which included 1,028 women with metastatic breast cancer, showed that the sialyl-Tn (STn)-KLH vaccine was well tolerated by the patients, but no overall benefit in time to progression or survival was observed (Miles et al., 2011). However, women who also received endocrine therapy and showed greater antibody response to the STn-KLH vaccine had significantly longer median overall survival than those who had lower antibody responses, demonstrating that this dual treatment may improve clinical outcomes in women with metastatic breast cancer (Miles et al., 2011). Other monovalent vaccines have been developed in which a single carbohydrate antigen, including Globo-H, fucosyl GM1 and Lewisy (Ley), were coupled to KLH with saponin fractions (QS21 or OPT-82142) added as adjuvants; these vaccines have exhibited varying degrees of promise in early clinical settings (Ragupathi et al., 2011; Zhu et al., 2009). Additionally, multivalent vaccines have been formulated with GM2, Globo-H, Lewis Y, Tn (c), STn (c), TF (c) and Tn-MUC1, each individ‐ ually conjugated to KLH. QS21 was added as an adjuvant to this cocktail of conjugates and was evaluated in patients with either epithelial ovarian, fallopian tube or peritoneal cancer in second or greater complete clinical remission, demonstrating that this heptavalent vaccine was safe and induced antibody responses against five of seven antigens (Sabbatini et al., 2007). Currently, more elaborate constructs have been designed in a new vaccine against ovarian cancer, based on clusters of a globotriaosylceramide (Gb3) antigen and incorporating a MUC5AC peptide epitope that mimics the surfaces of targeted tumor cells (Zhu et al., 2009).
vaccines and therapies, a quantitative enzyme-linked immunosorbent assay (ELISA) has been developed to measure KLH-specific antibodies as biomarkers in humans, demonstrating that it is possible to monitor the dynamics of both KLH-specific antibody levels and antibody class-
Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications
http://dx.doi.org/10.5772/57552
57
Numerous tumor antigens coupled to hemocyanins have been tested in monovalent, divalent or polyvalent vaccines (Zhu et al., 2009). To date, there are 10 clinical trials in various stages of development using primarily the antigen MUC-2 with KLH. When comparing preclinical studies between mono-and polyvalent vaccines, the controversy arises over which strategy is best and why, based on the observations that the monovalent vaccines induced high titers of antibodies, whereas the polyvalent vaccines reached a greater range of antibodies (Slovin et al., 2007). Recently, CCH, as an adjuvant in a cancer vaccine based on DCs loaded with prostate cancer cell lysates and TRIMEL, was used for the first time in CRPC patients. This study demonstrated that CCH is safe and is capable of inducing memory T cell responses, as determined *in vivo* by a DTH test (Reyes et al., 2013), thus indicating that CCH can complement
In the eradication of micrometastases in chemotherapy-resistant small cell lung cancer (SCLC), the ganglioside fucosyl-GM1 is used as a therapeutic target. Therefore, in a clinical study, 11 patients with SCLC were vaccinated by injection with this synthetic antigen coupled to KLH. Higher IgM titers against fucosyl-GM1 were detected in eight of the patients, and only one patient developed IgG with a similar titer; these results led to a Phase II study (Krug et al., 2004). Recently, vaccination with N-propionyl polysialic acid-KLH conjugate plus QS-21 in patients with SCLC has demonstrated that this vaccine is safe and results in consistent hightiter antibody responses. These results open up the possibility of incorporating polysialic acid into a polyvalent vaccine against SCLC with four glycolipid antigens that are also widely expressed in SCLC, such as GD2, GD3, fucosylated GM1 and globo H (Krug et al., 2012). In consideration of the fact that tumor angiogenesis is critical for tumor growth, infiltration and metastasis, another promising approach includes the generation of a mimotope of Bevacizu‐ mab (Avastin), a monoclonal antibody that reacts directly against vascular endothelial growth factor (VEGF) conjugated to KLH, which induces humoral immunity and results in the
For the non-Hodgkins lymphomas, the strategy of conjugating the variable regions of the immunoglobulin to the tumorigenic clone with KLH has been used. These variable regions act as tumor antigens. The principle of these vaccines is that the clonal immunoglobulin idiotype displayed on the surface of most malignant B cells is not only tumor-specific; it is also a patientspecific antigen that can be used for therapeutic vaccination (Bendandi, 2006; de Cerio et al., 2007). These idiotype vaccines have shown some degrees of both biological efficacy and
induction of anti-angiogenesis responses in mice (Li et al., 2013).
switching in individuals who are repeatedly exposed to KLH (Aarntzen et al., 2012).
**5.5. Prostate cancer**
**5.6. Lung cancer**
or substitute KLH in such therapies.
**5.7. Lymphomas and myelomas**
#### **5.4. Melanomas**
Vaccines were designed with the gangliosides GM2, GD2 and GD3 conjugated to KLH plus the adjuvant QS21. These vaccines produced high titers of IgM and IgG in patients with melanoma compared to vaccines with GM2 and BCG as adjuvants (Ragupathi et al., 2000). Although it was initially postulated that immunotherapy against GM2 was effective in reducing the risk of relapse in post-surgery patients with stage III melanoma, the vaccine was ineffective in a Phase III clinical trial in 1314 patients (Eggermont et al., 2013). Additionally, gangliosides GD2-L and GD3-L have been used as therapeutic targets, generating higher titers of IgM and IgG (Ragupathi et al., 2003).
Regarding cellular vaccines, the first clinical study included patients with advanced melanoma who were treated with DCs loaded with tumor lysate or a mixture of melanoma peptides in the presence of KLH. Patients tolerated the vaccine and did not develop autoimmunity (Nestle et al., 1998). Currently, other antigens, such as MART-1/MelanA, tyrosinase, MAGE-3 and gp100 (Fay et al., 2006; Tittarelli et al.; Tittarelli et al., 2012), in addition to the heat-shocked melanoma lysate named TRIMEL (Aguilera et al., 2011), with KLH as an adjuvant have been loaded on DCs, yielding promising results in terms of the prolongation of overall survival time. Interestingly, in the search for parameters of clinical effectiveness of KLH-containing vaccines and therapies, a quantitative enzyme-linked immunosorbent assay (ELISA) has been developed to measure KLH-specific antibodies as biomarkers in humans, demonstrating that it is possible to monitor the dynamics of both KLH-specific antibody levels and antibody classswitching in individuals who are repeatedly exposed to KLH (Aarntzen et al., 2012).
#### **5.5. Prostate cancer**
**5.3. Breast, ovary, fallopian tube and peritoneal cancers**
**5.4. Melanomas**
56 Immune Response Activation
of IgM and IgG (Ragupathi et al., 2003).
Sialyl-Tn (STn) is the tumor antigen most frequently expressed in breast, ovary, colon, rectal, stomach and pancreas adenocarcinomas. This antigen, conjugated to KLH, has been named Theratope and is capable of generating positive serological results, inducing significant titers of IgM and IgG antibodies against STn (Gilewski et al., 2007). The results of a Phase III multicenter clinical trial, which included 1,028 women with metastatic breast cancer, showed that the sialyl-Tn (STn)-KLH vaccine was well tolerated by the patients, but no overall benefit in time to progression or survival was observed (Miles et al., 2011). However, women who also received endocrine therapy and showed greater antibody response to the STn-KLH vaccine had significantly longer median overall survival than those who had lower antibody responses, demonstrating that this dual treatment may improve clinical outcomes in women with metastatic breast cancer (Miles et al., 2011). Other monovalent vaccines have been developed in which a single carbohydrate antigen, including Globo-H, fucosyl GM1 and Lewisy (Ley), were coupled to KLH with saponin fractions (QS21 or OPT-82142) added as adjuvants; these vaccines have exhibited varying degrees of promise in early clinical settings (Ragupathi et al., 2011; Zhu et al., 2009). Additionally, multivalent vaccines have been formulated with GM2, Globo-H, Lewis Y, Tn (c), STn (c), TF (c) and Tn-MUC1, each individ‐ ually conjugated to KLH. QS21 was added as an adjuvant to this cocktail of conjugates and was evaluated in patients with either epithelial ovarian, fallopian tube or peritoneal cancer in second or greater complete clinical remission, demonstrating that this heptavalent vaccine was safe and induced antibody responses against five of seven antigens (Sabbatini et al., 2007). Currently, more elaborate constructs have been designed in a new vaccine against ovarian cancer, based on clusters of a globotriaosylceramide (Gb3) antigen and incorporating a MUC5AC peptide epitope that mimics the surfaces of targeted tumor cells (Zhu et al., 2009).
Vaccines were designed with the gangliosides GM2, GD2 and GD3 conjugated to KLH plus the adjuvant QS21. These vaccines produced high titers of IgM and IgG in patients with melanoma compared to vaccines with GM2 and BCG as adjuvants (Ragupathi et al., 2000). Although it was initially postulated that immunotherapy against GM2 was effective in reducing the risk of relapse in post-surgery patients with stage III melanoma, the vaccine was ineffective in a Phase III clinical trial in 1314 patients (Eggermont et al., 2013). Additionally, gangliosides GD2-L and GD3-L have been used as therapeutic targets, generating higher titers
Regarding cellular vaccines, the first clinical study included patients with advanced melanoma who were treated with DCs loaded with tumor lysate or a mixture of melanoma peptides in the presence of KLH. Patients tolerated the vaccine and did not develop autoimmunity (Nestle et al., 1998). Currently, other antigens, such as MART-1/MelanA, tyrosinase, MAGE-3 and gp100 (Fay et al., 2006; Tittarelli et al.; Tittarelli et al., 2012), in addition to the heat-shocked melanoma lysate named TRIMEL (Aguilera et al., 2011), with KLH as an adjuvant have been loaded on DCs, yielding promising results in terms of the prolongation of overall survival time. Interestingly, in the search for parameters of clinical effectiveness of KLH-containing
Numerous tumor antigens coupled to hemocyanins have been tested in monovalent, divalent or polyvalent vaccines (Zhu et al., 2009). To date, there are 10 clinical trials in various stages of development using primarily the antigen MUC-2 with KLH. When comparing preclinical studies between mono-and polyvalent vaccines, the controversy arises over which strategy is best and why, based on the observations that the monovalent vaccines induced high titers of antibodies, whereas the polyvalent vaccines reached a greater range of antibodies (Slovin et al., 2007). Recently, CCH, as an adjuvant in a cancer vaccine based on DCs loaded with prostate cancer cell lysates and TRIMEL, was used for the first time in CRPC patients. This study demonstrated that CCH is safe and is capable of inducing memory T cell responses, as determined *in vivo* by a DTH test (Reyes et al., 2013), thus indicating that CCH can complement or substitute KLH in such therapies.
#### **5.6. Lung cancer**
In the eradication of micrometastases in chemotherapy-resistant small cell lung cancer (SCLC), the ganglioside fucosyl-GM1 is used as a therapeutic target. Therefore, in a clinical study, 11 patients with SCLC were vaccinated by injection with this synthetic antigen coupled to KLH. Higher IgM titers against fucosyl-GM1 were detected in eight of the patients, and only one patient developed IgG with a similar titer; these results led to a Phase II study (Krug et al., 2004). Recently, vaccination with N-propionyl polysialic acid-KLH conjugate plus QS-21 in patients with SCLC has demonstrated that this vaccine is safe and results in consistent hightiter antibody responses. These results open up the possibility of incorporating polysialic acid into a polyvalent vaccine against SCLC with four glycolipid antigens that are also widely expressed in SCLC, such as GD2, GD3, fucosylated GM1 and globo H (Krug et al., 2012). In consideration of the fact that tumor angiogenesis is critical for tumor growth, infiltration and metastasis, another promising approach includes the generation of a mimotope of Bevacizu‐ mab (Avastin), a monoclonal antibody that reacts directly against vascular endothelial growth factor (VEGF) conjugated to KLH, which induces humoral immunity and results in the induction of anti-angiogenesis responses in mice (Li et al., 2013).
#### **5.7. Lymphomas and myelomas**
For the non-Hodgkins lymphomas, the strategy of conjugating the variable regions of the immunoglobulin to the tumorigenic clone with KLH has been used. These variable regions act as tumor antigens. The principle of these vaccines is that the clonal immunoglobulin idiotype displayed on the surface of most malignant B cells is not only tumor-specific; it is also a patientspecific antigen that can be used for therapeutic vaccination (Bendandi, 2006; de Cerio et al., 2007). These idiotype vaccines have shown some degrees of both biological efficacy and improved clinical outcomes in lymphoma (Bendandi, 2009). The effect of this type of vaccine has been enhanced by incorporating granulocyte macrophage colony-stimulating factor (GM-CSF); showing DCs recruiment and activation of CD4+and CD8+T lymphocytes (Bendandi et al., 1999). These encouraging results led to the initiation of Phase III clinical trials to establish the clinical benefits induced by anti-idiotypic vaccines (Neelapu and Kwak, 2007). In addition, DC-based anti-idiotype vaccination using KLH as an adjuvant induces a specific cellular immune response in patients with stage 1 myeloma (Rollig et al., 2011).
oligosaccharides on mollusk hemocyanins has not been clearly demonstrated. However, it has been reported that KLH promotes the *in vitro* maturation of human DCs through its engage‐ ment of the mannose receptor, as assessed by the up-regulation of the cell surface expression of major histocompatibility complex (MHC) class II and co-stimulatory molecules (Presicce et al., 2008). It seems that this phenomenon has been described only for human DCs; in mice, neither KLH nor CCH induce DC maturation *in vitro* through canonical mechanisms (i.e., via up-regulation of MHCII and co-stimulatory molecules) during the early hours of *in vitro* culture (Arancibia et al., 2012a). Moreover, Teitz-Tennenbaum and collaborators (2008) demonstrated that murine DCs pulsed with KLH for 18 h *in vitro* did not undergo DC maturation, a result that is consistent with *in vivo* experiments performed by Moltedo et al. 2009 (Moltedo et al., 2009; Teitz-Tennenbaum et al., 2008). Recently, we observed that mouse DCs internalized this hemocyanin but did not mature within 72 h of culture *in vitro* with CCH. Remarkably, FLH, unlike KLH or CCH, promoted the high secretion of pro-inflammatory cytokines, such as IL-6, IL-12p40, TNF-α and IL-23, from murine DCs *in vitro* (Arancibia et al., 2014). Moreover, the secretion of cytokines was dependent on the presence of oligosaccharides on FLH, indicating a role for lectin-like receptors in the response. These results indicate the complexity of the immune response and demonstrate that different hemocyanins can activate diverse molecular and cellular pathways. In addition, the divergence regarding the role of DCs in hemocyanin recognition highlights the importance of studying those mechanisms in human
Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications
http://dx.doi.org/10.5772/57552
59
Once hemocyanins are internalized by APCs, the proteolytic machinery of the cell degrades them slowly in comparison to classical antigens, thus increasing antigen persistence. This process permits the presence of hemocyanins for longer periods of time, resulting in aug‐ mented immunogenicity, as demonstrated for CCH (Arancibia et al., 2012a). In fact, using electron microscopy, we have determined that CCH molecules are internalized by DCs mainly through macropinocytosis and are then localized intact in lysosomes for up to 24 h. Moreover, we compared the antigen-processing kinetics of DCs for CCH and ovalbumin, a widely used antigen model, demonstrating the persistence of a band of approximately 49 kDa in DC cultures with CCH for 72 h; in contrast, in cultures of DCs pulsed with ovalbumin, the antigen
Hemocyanins are thymus-dependent antigens; therefore, they require the interaction between T and B lymphocytes to initiate antibody production. In fact, it has been established that nonspecific immunotherapeutic effects of hemocyanins in superficial bladder cancer rely on adequate priming, emphasizing the importance of the adaptive immune response in this property. Considering that the antitumor effect is induced by hemocyanins themselves without adjuvant, we assume that KLH and CCH have common structural features that promote inflammation and maintain innate immunity, leading to the onset of an antitumor adaptive immune response. These common characteristics of hemocyanins may rely on carbohydrates, which may act as natural adjuvants. A proposed alternative mechanism is that CD4+T lymphocytes reacting to preserved xenogenic peptidic sequences stimulate T lympho‐
cells.
was completely degraded by this time.
**6.2. At the adaptive immunity level**
#### **5.8. Vaccines against abused drugs**
Experimental vaccines have been developed to curb recidivism in the consumption of addictive drugs, including cocaine and nicotine. The aim of these vaccines is to generate antibodies that reduce blood concentrations of the given drug and its effect on the central nervous system by preventing the drug from crossing the blood-brain barrier. The experi‐ mental strategies to induce protective and specific antibodies against cocaine included subcutaneous and intranasal immunization (Hrafnkelsdottir et al., 2007). Preclinical studies using KLH as a carrier of these drugs are promising (Moreno and Janda, 2009). However, the only clinical evidence with good results for the effectiveness of these vaccines used cholera toxin as a carrier of the cocaine-B subunit (Martell et al., 2009). Another interesting experi‐ mental approach is the generation of two complementary but mechanistically distinct active vaccines (i.e., non-catalytic and catalytic hapten designs) using a cocaine-like hapten, GNE, and a cocaine transition-state analogue, GNT, coupled to KLH. Although GNT-KLH induced higher titers of catalytic antibodies, these titers were not sufficient to provide protection in mice challenged with cocaine (Cai et al., 2013). Additionally, an anti-nicotine vaccine that used nicotine coupled to KLH, which was named Niccine®, was no better than the placebo (Gorelick, 2012).
#### **6. The mechanisms of hemocyanin immunostimulation**
#### **6.1. At the innate immunity level**
The effects induced by hemocyanins during the early phases of the immune response and the identities of the target cell type(s) have been scarcely studied. Several authors have suggested that the oligosaccharides may play a role in this process. As mentioned, the carbohydrate compositions of hemocyanins are quite diverse; however, the most abundant monosaccharide is mannose, which is often found in high-mannose or hybrid structures of oligosaccharides. The mammalian innate immune system has a variety of cell types that express several receptors on the cell surface. In this context, specialized professional antigen-presenting cells (APCs), such as macrophages and DCs, are key players in immune surveillance. These cells present a broad range of germ-line encoded pattern recognition receptors that recognize conserved pathogen-associated molecular patterns. Among these receptors, the C-type lectin family of receptors recognizes several types of oligosaccharides present on pathogens and foreign molecules (McGreal et al., 2006; van Vliet et al., 2008). The biological role of the high mannose oligosaccharides on mollusk hemocyanins has not been clearly demonstrated. However, it has been reported that KLH promotes the *in vitro* maturation of human DCs through its engage‐ ment of the mannose receptor, as assessed by the up-regulation of the cell surface expression of major histocompatibility complex (MHC) class II and co-stimulatory molecules (Presicce et al., 2008). It seems that this phenomenon has been described only for human DCs; in mice, neither KLH nor CCH induce DC maturation *in vitro* through canonical mechanisms (i.e., via up-regulation of MHCII and co-stimulatory molecules) during the early hours of *in vitro* culture (Arancibia et al., 2012a). Moreover, Teitz-Tennenbaum and collaborators (2008) demonstrated that murine DCs pulsed with KLH for 18 h *in vitro* did not undergo DC maturation, a result that is consistent with *in vivo* experiments performed by Moltedo et al. 2009 (Moltedo et al., 2009; Teitz-Tennenbaum et al., 2008). Recently, we observed that mouse DCs internalized this hemocyanin but did not mature within 72 h of culture *in vitro* with CCH. Remarkably, FLH, unlike KLH or CCH, promoted the high secretion of pro-inflammatory cytokines, such as IL-6, IL-12p40, TNF-α and IL-23, from murine DCs *in vitro* (Arancibia et al., 2014). Moreover, the secretion of cytokines was dependent on the presence of oligosaccharides on FLH, indicating a role for lectin-like receptors in the response. These results indicate the complexity of the immune response and demonstrate that different hemocyanins can activate diverse molecular and cellular pathways. In addition, the divergence regarding the role of DCs in hemocyanin recognition highlights the importance of studying those mechanisms in human cells.
Once hemocyanins are internalized by APCs, the proteolytic machinery of the cell degrades them slowly in comparison to classical antigens, thus increasing antigen persistence. This process permits the presence of hemocyanins for longer periods of time, resulting in aug‐ mented immunogenicity, as demonstrated for CCH (Arancibia et al., 2012a). In fact, using electron microscopy, we have determined that CCH molecules are internalized by DCs mainly through macropinocytosis and are then localized intact in lysosomes for up to 24 h. Moreover, we compared the antigen-processing kinetics of DCs for CCH and ovalbumin, a widely used antigen model, demonstrating the persistence of a band of approximately 49 kDa in DC cultures with CCH for 72 h; in contrast, in cultures of DCs pulsed with ovalbumin, the antigen was completely degraded by this time.
#### **6.2. At the adaptive immunity level**
improved clinical outcomes in lymphoma (Bendandi, 2009). The effect of this type of vaccine has been enhanced by incorporating granulocyte macrophage colony-stimulating factor (GM-CSF); showing DCs recruiment and activation of CD4+and CD8+T lymphocytes (Bendandi et al., 1999). These encouraging results led to the initiation of Phase III clinical trials to establish the clinical benefits induced by anti-idiotypic vaccines (Neelapu and Kwak, 2007). In addition, DC-based anti-idiotype vaccination using KLH as an adjuvant induces a specific cellular
Experimental vaccines have been developed to curb recidivism in the consumption of addictive drugs, including cocaine and nicotine. The aim of these vaccines is to generate antibodies that reduce blood concentrations of the given drug and its effect on the central nervous system by preventing the drug from crossing the blood-brain barrier. The experi‐ mental strategies to induce protective and specific antibodies against cocaine included subcutaneous and intranasal immunization (Hrafnkelsdottir et al., 2007). Preclinical studies using KLH as a carrier of these drugs are promising (Moreno and Janda, 2009). However, the only clinical evidence with good results for the effectiveness of these vaccines used cholera toxin as a carrier of the cocaine-B subunit (Martell et al., 2009). Another interesting experi‐ mental approach is the generation of two complementary but mechanistically distinct active vaccines (i.e., non-catalytic and catalytic hapten designs) using a cocaine-like hapten, GNE, and a cocaine transition-state analogue, GNT, coupled to KLH. Although GNT-KLH induced higher titers of catalytic antibodies, these titers were not sufficient to provide protection in mice challenged with cocaine (Cai et al., 2013). Additionally, an anti-nicotine vaccine that used nicotine coupled to KLH, which was named Niccine®, was no better than the placebo
The effects induced by hemocyanins during the early phases of the immune response and the identities of the target cell type(s) have been scarcely studied. Several authors have suggested that the oligosaccharides may play a role in this process. As mentioned, the carbohydrate compositions of hemocyanins are quite diverse; however, the most abundant monosaccharide is mannose, which is often found in high-mannose or hybrid structures of oligosaccharides. The mammalian innate immune system has a variety of cell types that express several receptors on the cell surface. In this context, specialized professional antigen-presenting cells (APCs), such as macrophages and DCs, are key players in immune surveillance. These cells present a broad range of germ-line encoded pattern recognition receptors that recognize conserved pathogen-associated molecular patterns. Among these receptors, the C-type lectin family of receptors recognizes several types of oligosaccharides present on pathogens and foreign molecules (McGreal et al., 2006; van Vliet et al., 2008). The biological role of the high mannose
immune response in patients with stage 1 myeloma (Rollig et al., 2011).
**6. The mechanisms of hemocyanin immunostimulation**
**5.8. Vaccines against abused drugs**
58 Immune Response Activation
(Gorelick, 2012).
**6.1. At the innate immunity level**
Hemocyanins are thymus-dependent antigens; therefore, they require the interaction between T and B lymphocytes to initiate antibody production. In fact, it has been established that nonspecific immunotherapeutic effects of hemocyanins in superficial bladder cancer rely on adequate priming, emphasizing the importance of the adaptive immune response in this property. Considering that the antitumor effect is induced by hemocyanins themselves without adjuvant, we assume that KLH and CCH have common structural features that promote inflammation and maintain innate immunity, leading to the onset of an antitumor adaptive immune response. These common characteristics of hemocyanins may rely on carbohydrates, which may act as natural adjuvants. A proposed alternative mechanism is that CD4+T lymphocytes reacting to preserved xenogenic peptidic sequences stimulate T lympho‐ cytes to secrete Th1 cytokines that, in turn, break tumor tolerance via a bystander effect, thus braking tolerance to tumor antigens and enhancing the immune response against the tumor. This hypothesis is supported by the high secretion of cytokines, such as IFNγ and IL-2, observed in the regional lymph nodes (Arancibia et al., 2012b) after hemocyanin challenge. Accordingly, mice that were primed with CCH or KLH, challenged with MBT-2 cells and then subjected to immunotherapy with the respective hemocyanin increased natural killer cell activity and exhibited cytokine environment polarization toward a Th1 response (i.e., IFNγ production increased significantly in mouse sera), which correlated with antibodies belonging to the Th1 isotypes (Moltedo et al., 2006). Moreover, in patients with superficial bladder cancer under intravesical KLH therapy, a significant increase of CD4+T lymphocytes in the submu‐ cosa and among urothelial cells was observed, in contrast to a slight increase in CD8+T lymphocytes (Jurincic-Winkler et al., 1995a; Jurincic-Winkler et al., 1995b). In addition, KLHconjugated vaccines against cancer using mucin-like or ganglioside epitopes have induced tumor-specific antibodies of the IgG1 and IgG3 isotypes (Musselli et al., 2001).
**Acknowledgements**
**Author details**
Alfredo De Ioannes2
**References**
María Inés Becker1,2, Sergio Arancibia1
2 Biosonda Corporation, Santiago, Chile
*Gene*. 301:53-60.
*Bioorg Khim.* 38:667-675.
*Cancer Immunol Immunother*. 61:2003-2011.
cally effective dendritic cells. *Clin Cancer Res.* 17:2474-2483.
Becker.
This work was partially supported by research grant from **FONDECYT 1110651** to María Inés
, Fabián Salazar1
[1] Aarntzen, E.H., I.J. de Vries, J.H. Goertz, M. Beldhuis-Valkis, H.M. Brouwers, M.W. van de Rakt, R.G. van der Molen, C.J. Punt, G.J. Adema, P.J. Tacken, I. Joosten, and J.F. Jacobs. 2012. Humoral anti-KLH responses in cancer patients treated with den‐ dritic cell-based immunotherapy are dictated by different vaccination parameters.
[2] Aguilera, R., C. Saffie, A. Tittarelli, F.E. Gonzalez, M. Ramirez, D. Reyes, C. Pereda, D. Hevia, T. Garcia, L. Salazar, A. Ferreira, M. Hermoso, A. Mendoza-Naranjo, C. Ferrada, P. Garrido, M.N. Lopez, and F. Salazar-Onfray. 2011. Heat-shock induction of tumor-derived danger signals mediates rapid monocyte differentiation into clini‐
[3] Altenhein, B., J. Markl, and B. Lieb. 2002. Gene structure and hemocyanin isoform HtH2 from the mollusc Haliotis tuberculata indicate early and late intron hot spots.
[4] Andreev, S.M., A.V. Giliazova, M.R. Khaitov, and D.G. Kozlov. 2012. [Anti-peptide antibodies specifically recognize the L1 protein of human papilloma virus type 31].
[5] Arancibia, S., M. Del Campo, E. Nova, F. Salazar, and M.I. Becker. 2012a. Enhanced structural stability of Concholepas hemocyanin increases its immunogenicity and maintains its non-specific immunostimulatory effects. *Eur J Immunol.* 42:688-699.
[6] Arancibia, S., C. Espinoza, F. Salazar, M. Del Campo, R. Tampe, T. Zhong, P. De Ioannes, B. Moltedo, J. Ferreira, E.C. Lavelle, A. Manubens, and M.I. Becker. 2014. A Novel Immunomodulatory Hemocyanin from the Limpet Fissurella latimarginata
1 Fundación Ciencia y Tecnología para el Desarrollo (FUCITED), Santiago, Chile
, Miguel Del Campo1
Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications
and
http://dx.doi.org/10.5772/57552
61
#### **7. Future prospects**
The information available shows that the mollusk hemocyanins presented herein – including KLH, the traditional and most utilized hemocyanin in biomedicine, and **CCH**, **RtH**, **HtH** and **FLH,** the most recently characterized hemocyanins – differ in their immunogenicity and nonspecific immunostimulatory properties, opening basic questions on the structural character‐ istics responsible for these differences. Moreover, it is important to note that although several subunits of these hemocyanins have been sequenced via cDNA and that the complete amino acid sequences for these subunits are currently available, until now, it has not been possible to express heterologous or synthetic hemocyanins, mainly because of their size, complex structures and glycosylations. Therefore, all of the hemocyanins mentioned in this review can be obtained only as purified biological products from their natural sources, emphasizing the importance of pre-clinically evaluating diverse sources of hemocyanins for biomedical purposes. In this respect, only CCH hemocyanin from *Concholepas,* which has been preclinically evaluated in a murine experimental model of superficial bladder cancer (Moltedo et al., 2006) and has been recently used as an adjuvant in DC immunotherapy of patients with prostate cancer, is considered a safe complement or alternative to KLH (Reyes et al., 2013; Salazar-Onfray et a. 2013).
Finally, from a basic science perspective, a detailed molecular understanding of the receptors and signaling pathways involved in hemocyanin-induced immunostimulatory processes in mammals will be required to determine the basis of the qualitative and quantitative differences in the effects of the different hemocyanins on tumor immunology. Undoubtedly, this knowl‐ edge will support the development of new and better biomedical applications of these proteins, as hemocyanins exert their effects as immunopotentiators without the unwanted inflammatory side effects of classical adjuvants that drive cell-mediated immune responses.
#### **Acknowledgements**
cytes to secrete Th1 cytokines that, in turn, break tumor tolerance via a bystander effect, thus braking tolerance to tumor antigens and enhancing the immune response against the tumor. This hypothesis is supported by the high secretion of cytokines, such as IFNγ and IL-2, observed in the regional lymph nodes (Arancibia et al., 2012b) after hemocyanin challenge. Accordingly, mice that were primed with CCH or KLH, challenged with MBT-2 cells and then subjected to immunotherapy with the respective hemocyanin increased natural killer cell activity and exhibited cytokine environment polarization toward a Th1 response (i.e., IFNγ production increased significantly in mouse sera), which correlated with antibodies belonging to the Th1 isotypes (Moltedo et al., 2006). Moreover, in patients with superficial bladder cancer under intravesical KLH therapy, a significant increase of CD4+T lymphocytes in the submu‐ cosa and among urothelial cells was observed, in contrast to a slight increase in CD8+T lymphocytes (Jurincic-Winkler et al., 1995a; Jurincic-Winkler et al., 1995b). In addition, KLHconjugated vaccines against cancer using mucin-like or ganglioside epitopes have induced
tumor-specific antibodies of the IgG1 and IgG3 isotypes (Musselli et al., 2001).
The information available shows that the mollusk hemocyanins presented herein – including KLH, the traditional and most utilized hemocyanin in biomedicine, and **CCH**, **RtH**, **HtH** and **FLH,** the most recently characterized hemocyanins – differ in their immunogenicity and nonspecific immunostimulatory properties, opening basic questions on the structural character‐ istics responsible for these differences. Moreover, it is important to note that although several subunits of these hemocyanins have been sequenced via cDNA and that the complete amino acid sequences for these subunits are currently available, until now, it has not been possible to express heterologous or synthetic hemocyanins, mainly because of their size, complex structures and glycosylations. Therefore, all of the hemocyanins mentioned in this review can be obtained only as purified biological products from their natural sources, emphasizing the importance of pre-clinically evaluating diverse sources of hemocyanins for biomedical purposes. In this respect, only CCH hemocyanin from *Concholepas,* which has been preclinically evaluated in a murine experimental model of superficial bladder cancer (Moltedo et al., 2006) and has been recently used as an adjuvant in DC immunotherapy of patients with prostate cancer, is considered a safe complement or alternative to KLH (Reyes et al., 2013;
Finally, from a basic science perspective, a detailed molecular understanding of the receptors and signaling pathways involved in hemocyanin-induced immunostimulatory processes in mammals will be required to determine the basis of the qualitative and quantitative differences in the effects of the different hemocyanins on tumor immunology. Undoubtedly, this knowl‐ edge will support the development of new and better biomedical applications of these proteins, as hemocyanins exert their effects as immunopotentiators without the unwanted inflammatory
side effects of classical adjuvants that drive cell-mediated immune responses.
**7. Future prospects**
60 Immune Response Activation
Salazar-Onfray et a. 2013).
This work was partially supported by research grant from **FONDECYT 1110651** to María Inés Becker.
#### **Author details**
María Inés Becker1,2, Sergio Arancibia1 , Fabián Salazar1 , Miguel Del Campo1 and Alfredo De Ioannes2
1 Fundación Ciencia y Tecnología para el Desarrollo (FUCITED), Santiago, Chile
2 Biosonda Corporation, Santiago, Chile
#### **References**
Promotes Potent Anti-tumor Activity in Melanoma. *Plos One.* doi:10.1371/jour‐ nal.pone.0087240
[18] de Cerio, A.L., N. Zabalegui, M. Rodriguez-Calvillo, S. Inoges, and M. Bendandi. 2007. Anti-idiotype antibodies in cancer treatment. *Oncogene*. 26:3594-3602.
Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications
http://dx.doi.org/10.5772/57552
63
[19] De Ioannes, P., B. Moltedo, H. Oliva, R. Pacheco, F. Faunes, A.E. De Ioannes, and M.I. Becker. 2004. Hemocyanin of the molluscan Concholepas concholepas exhibits an un‐
[20] Decker, H., N. Hellmann, E. Jaenicke, B. Lieb, U. Meissner, and J. Markl. 2007. Mini‐ review: Recent progress in hemocyanin research. *Integr Comp Biol*. 47:631-644.
[21] Del Campo, M., S. Arancibia, E. Nova, F. Salazar, A. Gonzalez, B. Moltedo, P. De Ioannes, J. Ferreira, A. Manubens, and M.I. Becker. 2011. [Hemocyanins as immunos‐
[22] Dorosko, S.M., S.L. Ayres, and R.I. Connor. 2008. Induction of HIV-1 MPR(649-684) specific IgA and IgG antibodies in caprine colostrum using a peptide-based vaccine.
[23] Duvillie, B., M. Attali, V. Aiello, E. Quemeneur, and R. Scharfmann. 2003. Label-re‐ taining cells in the rat pancreas: location and differentiation potential in vitro. *Diabe‐*
[24] Edwards, F.R., and F. Whitwell. 1974. Use of BCG as an immunostimulant in the sur‐
[25] Eggermont, A.M., S. Suciu, P. Rutkowski, J. Marsden, M. Santinami, P. Corrie, S. Aamdal, P.A. Ascierto, P.M. Patel, W.H. Kruit, L. Bastholt, L. Borgognoni, M.G. Ber‐ nengo, N. Davidson, L. Polders, M. Praet, and A. Spatz. 2013. Adjuvant ganglioside GM2-KLH/QS-21 vaccination versus observation after resection of primary tumor > 1.5 mm in Patients with stage II melanoma: results of the EORTC 18961 randomized
[26] Escobar, A., M. Lopez, A. Serrano, M. Ramirez, C. Perez, A. Aguirre, R. Gonzalez, J. Alfaro, M. Larrondo, M. Fodor, C. Ferrada, and F. Salazar-Onfray. 2005. Dendritic cell immunizations alone or combined with low doses of interleukin-2 induce specif‐
[27] Fan, J., X. Liang, M.S. Horton, H.C. Perry, M.P. Citron, G.J. Heidecker, T.M. Fu, J. Joyce, C.T. Przysiecki, P.M. Keller, V.M. Garsky, R. Ionescu, Y. Rippeon, L. Shi, M.A. Chastain, J.H. Condra, M.E. Davies, J. Liao, E.A. Emini, and J.W. Shiver. 2004. Pre‐ clinical study of influenza virus A M2 peptide conjugate vaccines in mice, ferrets,
[28] Fay, J.W., A.K. Palucka, S. Paczesny, M. Dhodapkar, D.A. Johnston, S. Burkeholder, H. Ueno, and J. Banchereau. 2006. Long-term outcomes in patients with metastatic melanoma vaccinated with melanoma peptide-pulsed CD34(+) progenitor-derived
ic immune responses in melanoma patients. *Clin Exp Immunol*. 142:555-568.
gical treatment of carcinoma of the lung. *Thorax*. 29:654-658.
phase III trial. *J Clinical Oncol.* 31:3831-3837.
and rhesus monkeys. *Vaccine*. 22:2993-3003.
dendritic cells. *Cancer Immunol Immunother*. 55:1209-1218.
usual heterodecameric array of subunits. *J Biol Chem*. 279:26134-26142.
timulants]. *Rev Med Chil*. 139:236-246.
*Vaccine*. 26:5416-5422.
*tes*. 52:2035-2042.
[18] de Cerio, A.L., N. Zabalegui, M. Rodriguez-Calvillo, S. Inoges, and M. Bendandi. 2007. Anti-idiotype antibodies in cancer treatment. *Oncogene*. 26:3594-3602.
Promotes Potent Anti-tumor Activity in Melanoma. *Plos One.* doi:10.1371/jour‐
[7] Arancibia, S., F. Salazar, and M.I. Becker. 2012b. Hemocyanins in the Immunothera‐ py of Superficial Bladder Cancer. *In* Bladder Cancer. Janeza, Croatia: INTECH. pp
[8] Arentsen, H.C., K. Hendricksen, E. Oosterwijk, and J.A. Witjes. 2009. Experimental
[10] Baatarjav, T., K. Kataoka, R.S. Gilbert, Y. Terao, M. Fukui, M. Goto, S. Kawabata, M. Yamamoto, K. Fujihashi, and H.O. Ito. 2011. Mucosal immune features to phosphor‐
[11] Becker, M.I., I. Carrasco, C. Beltran, M. Torres, B. Jaureguiberry, and A.E. De Ioannes. 1998. Development of monoclonal antibodies to gizzerosine, a toxic component
[12] Becker, M.I., A. Fuentes, M. Del Campo, A. Manubens, E. Nova, H. Oliva, F. Faunes, M.A. Valenzuela, M. Campos-Vallette, A. Aliaga, J. Ferreira, A.E. De Ioannes, P. De Ioannes, and B. Moltedo. 2009. Immunodominant role of CCHA subunit of Concho‐ lepas hemocyanin is associated with unique biochemical properties. *Int Immunophar‐*
[13] Bendandi, M. 2006. Clinical benefit of idiotype vaccines: too many trials for a clever
[14] Bendandi, M., C.D. Gocke, C.B. Kobrin, F.A. Benko, L.A. Sternas, R. Pennington, T.M. Watson, C.W. Reynolds, B.L. Gause, P.L. Duffey, E.S. Jaffe, S.P. Creekmore, D.L. Lon‐ go, and L.W. Kwak. 1999. Complete molecular remissions induced by patient-specific vaccination plus granulocyte-monocyte colony-stimulating factor against lymphoma.
[15] Cai, X., T. Whitfield, M.S. Hixon, Y. Grant, G.F. Koob, and K.D. Janda. 2013. Probing active cocaine vaccination performance through catalytic and noncatalytic hapten de‐
[16] Curtis, J.E., E.M. Hersh, W.T. Butler, and R.D. Rossen. 1971. Antigen dose in the hu‐ man immune response. Dose-relationships in the human immune response to Key‐
[17] Curtis, J.E., E.M. Hersh, J.E. Harris, C. McBride, and E.J. Freireich. 1970. The human primary immune response to keyhole limpet haemocyanin: interrelationships of de‐ layed hypersensitivity, antibody response and in vitro blast transformation. *Clin Exp*
ylcholine by nasal Flt3 ligand cDNA-based vaccination. *Vaccine*. 29:5747-5757.
rat bladder urothelial cell carcinoma models. *World J Urol*. 27:313-317. [9] Atala, A. 2006. This Month in Investigative Urology. *J Urol*. 176:2335-2336.
present in fish meal. *Hybridoma*. 17:373-381.
demonstration? *Rev Recent Clin Trials.* 1:67-74.
hole limpet hemocyanin. *J Lab Clin Med*. 78:61-69.
nal.pone.0087240
*macol*. 9:330-339.
*Nat Med*. 5:1171-1177.
*Immunol*. 6:473-491.
sign. *J Med Chem*. 56:3701-3709.
221-242.
62 Immune Response Activation
[29] Flamm, J., A. Bucher, W. Holtl, and W. Albrecht. 1990. Recurrent superficial transi‐ tional cell carcinoma of the bladder: adjuvant topical chemotherapy versus immuno‐ therapy. A prospective randomized trial. *J Urol*. 144:260-263.
[41] Gunther, J.H., A. Jurczok, T. Wulf, S. Brandau, I. Deinert, D. Jocham, and A. Bohle. 1999. Optimizing syngeneic orthotopic murine bladder cancer (MB49). *Cancer Res*.
Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications
http://dx.doi.org/10.5772/57552
65
[42] Hall, R.L., and E.J. Wood. 1976. The carbohydrate content of gastropod haemocya‐
[43] Harris, J.R., and J. Markl. 1999. Keyhole limpet hemocyanin (KLH): a biomedical re‐
[44] Hrafnkelsdottir, K., J. Valgeirsson, S. Bjarnadottir, S. Olafsdottir, K. Olafsdottir, S.T. Hedinsdottir, E.V. Magnusdottir, and S. Gizurarson. 2007. Immunization prevents
[45] Hughes, E.E., and H.E. Gilleland, Jr. 1995. Ability of synthetic peptides representing epitopes of outer membrane protein F of Pseudomonas aeruginosa to afford protec‐ tion against P. aeruginosa infection in a murine acute pneumonia model. *Vaccine*.
[46] Hulikova, K., V. Grobarova, R. Krivohlava, and A. Fiserova. Antitumor activity of Nacetyl-D-glucosamine-substituted glycoconjugates and combined therapy with key‐ hole limpet hemocyanin in B16F10 mouse melanoma model. *Folia Microbiol (Praha)*.
[47] Idakieva, K., S. Severov, I. Svendsen, N. Genov, S. Stoeva, M. Beltramini, G. Tognon, P. Di Muro, and B. Salvato. 1993. Structural properties of Rapana thomasiana grosse hemocyanin: isolation, characterization and N-terminal amino acid sequence of two
[48] Idakieva, K., S. Stoeva, W. Voelter, and C. Gielens. 2004. Glycosylation of Rapana thomasiana hemocyanin. Comparison with other prosobranch (gastropod) hemocya‐
[49] Jurincic-Winkler, C., K.A. Metz, J. Beuth, U. Engelmann, and K.F. Klippel. 1995a. Im‐ munohistological findings in patients with superficial bladder carcinoma after intra‐
[50] Jurincic-Winkler, C., K.A. Metz, J. Beuth, J. Sippel, and K.F. Klippel. 1995b. Effect of keyhole limpet hemocyanin (KLH) and bacillus Calmette-Guerin (BCG) instillation
[51] Jurincic-Winkler, C.D., H. Gallati, M. Alvarez-Mon, J. Sippel, J. Carballido, and K.F. Klippel. 1995c. Urinary interleukin-1 alpha levels are increased by intravesical instil‐ lation with keyhole limpet hemocyanin in patients with superficial transitional cell
[52] Jurincic-Winkler, C.D., K.A. Metz, J. Beuth, and K.F. Klippel. 2000. Keyhole limpet hemocyanin for carcinoma in situ of the bladder: a long-term follow-up study. *Eur*
vesical instillation of keyhole limpet haemocyanin. *Br J Urol*. 76:702-707.
on carcinoma in situ of the urinary bladder. *Anticancer Res*. 15:2771-2776.
DDT buildup in mouse tissues. *Int Immunopharmacol*. 7:1179-1184.
different dissociation products. *Comp Biochem Physiol B*. 106:53-59.
nins. *Comp Biochem Physiol B Biochem Mol Biol*. 138:221-228.
carcinoma of the bladder. *Eur Urol*. 28:334-339.
*Urol*. 37 Suppl 3:45-49.
59:2834-2837.
13:1750-1753.
55:528-532.
nins. *Biochem Soc Trans*. 4:307-309.
view. *Micron*. 30:597-623.
[41] Gunther, J.H., A. Jurczok, T. Wulf, S. Brandau, I. Deinert, D. Jocham, and A. Bohle. 1999. Optimizing syngeneic orthotopic murine bladder cancer (MB49). *Cancer Res*. 59:2834-2837.
[29] Flamm, J., A. Bucher, W. Holtl, and W. Albrecht. 1990. Recurrent superficial transi‐ tional cell carcinoma of the bladder: adjuvant topical chemotherapy versus immuno‐
[30] Galonic, D.P., and D.Y. Gin. 2007. Chemical glycosylation in the synthesis of glyco‐
[31] Gathuru, J.K., F. Koide, G. Ragupathi, J.L. Adams, R.T. Kerns, T.P. Coleman, and P.O. Livingston. 2005. Identification of DHBcAg as a potent carrier protein compara‐
[32] Gatsogiannis, C., and J. Markl. 2009. Keyhole limpet hemocyanin: 9-A CryoEM struc‐ ture and molecular model of the KLH1 didecamer reveal the interfaces and intricate
[33] Gebauer, W., J.R. Harris, G. Geisthardt, and J. Markl. 1999a. Keyhole limpet hemo‐ cyanin type 2 (KLH2): detection and immunolocalization of a labile functional unit h.
[34] Gebauer, W., S. Stoeva, W. Voelter, E. Dainese, B. Salvato, M. Beltramini, and J. Markl. 1999b. Hemocyanin subunit organization of the gastropod Rapana thomasi‐
[35] Genova-Kalou, P., D. Dundarova, K. Idakieva, A. Mohmmed, S. Dundarov, and R. Argirova. 2008. Anti-herpes effect of hemocyanin derived from the mollusk Rapana
[36] Gesheva, V., K. Idakieva, N. Kerekov, K. Nikolova, N. Mihaylova, L. Doumanova, and A. Tchorbanov. 2011. Marine gastropod hemocyanins as adjuvants of non-conju‐
[37] Geyer, H., M. Wuhrer, A. Resemann, and R. Geyer. 2005. Identification and charac‐ terization of keyhole limpet hemocyanin N-glycans mediating cross-reactivity with
[38] Gilewski, T.A., G. Ragupathi, M. Dickler, S. Powell, S. Bhuta, K. Panageas, R.R. Ko‐ ganty, J. Chin-Eng, C. Hudis, L. Norton, A.N. Houghton, and P.O. Livingston. 2007. Immunization of high-risk breast cancer patients with clustered sTn-KLH conjugate
[39] Gorelick, D.A. 2012. Pharmacokinetic strategies for treatment of drug overdose and
[40] Grenegard, M., K. Vretenbrant-Oberg, M. Nylander, S. Desilets, E.G. Lindstrom, A. Larsson, I. Ramstrom, S. Ramstrom, and T.L. Lindahl. 2008. The ATP-gated P2X1 re‐ ceptor plays a pivotal role in activation of aspirin-treated platelets by thrombin and
gated bacterial and viral proteins. *Fish Shellfish Immunol.* 30:135-142.
plus the immunologic adjuvant QS-21. *Clin Cancer Res*. 13:2977-2985.
ble to KLH for augmenting MUC1 antigenicity. *Vaccine*. 23:4727-4733.
therapy. A prospective randomized trial. *J Urol*. 144:260-263.
topology of the 160 functional units. *J Mol Biol*. 385:963-983.
*J Struct Biol*. 128:280-286.
64 Immune Response Activation
ana. *Arch Biochem Biophys*. 372:128-134.
thomasiana. *Z Naturforsch C*. 63:429-434.
addiction. *Future Med Chem.* 4:227-243.
epinephrine. *J Biol Chem*. 283:18493-18504.
Schistosoma mansoni. *J Biol Chem*. 280:40731-40748.
conjugate antitumour vaccines. *Nature*. 446:1000-1007.
[53] Jurincic, C.D., U. Engelmann, J. Gasch, and K.F. Klippel. 1988. Immunotherapy in bladder cancer with keyhole-limpet hemocyanin: a randomized study. *J Urol*. 139:723-726.
[65] Lieb, B., B. Altenhein, and J. Markl. 2000. The sequence of a gastropod hemocyanin
Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications
http://dx.doi.org/10.5772/57552
67
[66] Linn, J.F., P. Black, K. Derksen, H. Rubben, and J.W. Thuroff. 2000. Keyhole limpet haemocyanin in experimental bladder cancer: literature review and own results. *Eur*
[67] Markl, J., B. Lieb, W. Gebauer, B. Altenhein, U. Meissner, and J.R. Harris. 2001. Ma‐ rine tumor vaccine carriers: structure of the molluscan hemocyanins KLH and htH. *J*
[68] Markl, J. 2013. Evolution of molluscan hemocyanin structures. *Biochim Biophys Acta*.
[69] Marsh, C.L., R.R. Torrey, J.L. Woolley, G.R. Barker, and B.H. Lau. 1987. Superiority of intravesical immunotherapy with Corynebacterium parvum and Allium sativum in
[70] Martell, B.A., F.M. Orson, J. Poling, E. Mitchell, R.D. Rossen, T. Gardner, and T.R. Kosten. 2009. Cocaine vaccine for the treatment of cocaine dependence in metha‐ done-maintained patients: a randomized, double-blind, placebo-controlled efficacy
[71] Matus, S., P.V. Burgos, M. Bravo-Zehnder, R. Kraft, O.H. Porras, P. Farias, L.F. Bar‐ ros, F. Torrealba, L. Massardo, S. Jacobelli, and A. Gonzalez. 2007. Antiribosomal-P autoantibodies from psychiatric lupus target a novel neuronal surface protein caus‐
[72] May, R.J., D.O. Beenhouwer, and M.D. Scharff. 2003. Antibodies to keyhole limpet hemocyanin cross-react with an epitope on the polysaccharide capsule of Cryptococ‐ cus neoformans and other carbohydrates: implications for vaccine development. *J Im‐*
[73] McFadden, D.W., D.R. Riggs, B.J. Jackson, A. Ng, and C. Cunningham. 2007. Keyhole limpet hemocyanin potentiates standard immunotherapy for melanoma. *Am J Surg*.
[74] McFadden, D.W., D.R. Riggs, B.J. Jackson, and L. Vona-Davis. 2003. Keyhole limpet hemocyanin, a novel immune stimulant with promising anticancer activity in Bar‐
[75] McGreal, E.P., M. Rosas, G.D. Brown, S. Zamze, S.Y. Wong, S. Gordon, L. Martinez-Pomares, and P.R. Taylor. 2006. The carbohydrate-recognition domain of Dectin-2 is
[76] Milas, L., and H.R. Withers. 1976. Nonspecific immunotherapy of malignant tumors.
[77] Miles, D., H. Roche, M. Martin, T.J. Perren, D.A. Cameron, J. Glaspy, D. Dodwell, J. Parker, J. Mayordomo, A. Tres, J.L. Murray, N.K. Ibrahim, and G. Theratope Study.
a C-type lectin with specificity for high mannose. *Glycobiology*. 16:422-430.
(HtH1 from Haliotis tuberculata). *J Biol Chem*. 275:5675-5681.
*Urol*. 37 Suppl 3:34-40.
1834:1840-1852.
*munol*. 171:4905-4912.
*Radiology*. 118:211-218.
193:284-287.
*Cancer Res Clin Oncol*. 127 Suppl 2:R3-9.
trial. *Arch Gen Psychiatry*. 66:1116-1123.
control of murine bladder cancer. *J Urol*. 137:359-362.
ing calcium influx and apoptosis. *J Exp Med*. 204:3221-3234.
rett's esophageal adenocarcinoma. *Am J Surg*. 186:552-555.
[65] Lieb, B., B. Altenhein, and J. Markl. 2000. The sequence of a gastropod hemocyanin (HtH1 from Haliotis tuberculata). *J Biol Chem*. 275:5675-5681.
[53] Jurincic, C.D., U. Engelmann, J. Gasch, and K.F. Klippel. 1988. Immunotherapy in bladder cancer with keyhole-limpet hemocyanin: a randomized study. *J Urol*.
[54] Keller, H., B. Lieb, B. Altenhein, D. Gebauer, S. Richter, S. Stricker, and J. Markl. 1999. Abalone (Haliotis tuberculata) hemocyanin type 1 (HtH1). Organization of the ap‐ proximately 400 kDa subunit, and amino acid sequence of its functional units f, g and
[55] Kiraly, J., E. Vareckova, V. Mucha, and F. Kostolansky. 2011. Evaluation of anti-influ‐ enza efficiency of polyclonal IgG antibodies specific to the ectodomain of M2 protein
[56] Krug, L.M., G. Ragupathi, C. Hood, C. George, F. Hong, R. Shen, L. Abrey, H.J. Jen‐ nings, M.G. Kris, and P.O. Livingston. 2012. Immunization with N-propionyl poly‐ sialic acid-KLH conjugate in patients with small cell lung cancer is safe and induces IgM antibodies reactive with SCLC cells and bactericidal against group B meningo‐
[57] Krug, L.M., G. Ragupathi, K.K. Ng, C. Hood, H.J. Jennings, Z. Guo, M.G. Kris, V. Miller, B. Pizzo, L. Tyson, V. Baez, and P.O. Livingston. 2004. Vaccination of small cell lung cancer patients with polysialic acid or N-propionylated polysialic acid con‐
[58] Lamm, D.L. 2003. Intravesical therapy for superficial bladder cancer: slow but steady
[59] Lamm, D.L., J.I. DeHaven, D.R. Riggs, C. Delgra, and R. Burrell. 1993a. Keyhole lim‐ pet hemocyanin immunotherapy of murine bladder cancer. *Urol Res*. 21:33-37.
[60] Lamm, D.L., J.I. DeHaven, D.R. Riggs, and R.F. Ebert. 1993b. Immunotherapy of murine bladder cancer with keyhole limpet hemocyanin (KLH). *J Urol*. 149:648-652.
[61] Lamm, D.L., J.A. Reyna, and D.F. Reichert. 1981. Keyhole-limpet haemacyanin and immune ribonucleic acid immunotherapy of murine transitional cell carcinoma. *Urol*
[62] Lamm, D.L., D.E. Thor, V.D. Stogdill, and H.M. Radwin. 1982. Bladder cancer immu‐
[63] Lau, B.H., J.L. Woolley, C.L. Marsh, G.R. Barker, D.H. Koobs, and R.R. Torrey. 1986. Superiority of intralesional immunotherapy with Corynebacterium parvum and Alli‐ um sativum in control of murine transitional cell carcinoma. *J Urol*. 136:701-705. [64] Li, W., Y. Ran, M. Li, K. Zhang, X. Qin, X. Xue, C. Zhang, Q. Hao, W. Zhang, and Y. Zhang. 2013. Mimotope vaccination for epitope-specific induction of anti-VEGF anti‐
jugated to keyhole limpet hemocyanin. *Clin Cancer Res*. 10:916-923.
of influenza A virus by passive immunization of mice. *Acta Virol*. 55:261-265.
139:723-726.
66 Immune Response Activation
h. *Eur J Biochem*. 264:27-38.
cocci. *Cancer Immunol Immunother*. 61:9-18.
progress. *J Clin Oncol*. 21:4259-4260.
notherapy. *J Urol*. 128:931-935.
bodies. *BMC Biotechnol*. 13:77.
*Res*. 9:227-230.
2011. Phase III multicenter clinical trial of the sialyl-TN (STn)-keyhole limpet hemo‐ cyanin (KLH) vaccine for metastatic breast cancer. O*ncologist*. 16:1092-1100.
[91] Peng, Y., Y. Zhang, W.J. Mitchell, and G. Zhang. 2012. Development of a lipopolysac‐ charide-targeted peptide mimic vaccine against Q fever. *J Immunol*. 189:4909-4920.
Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications
http://dx.doi.org/10.5772/57552
69
[92] Perabo, F.G., and S.C. Muller. 2004. Current and new strategies in immunotherapy
[93] Pilon, J., C. Loiacono, D. Okeson, S. Lund, K. Vercauteren, J. Rhyan, and L. Miller. 2007. Anti-prion activity generated by a novel vaccine formulation. *Neurosci Lett*.
[94] Presicce, P., A. Taddeo, A. Conti, M.L. Villa, and S. Della Bella. 2008. Keyhole limpet hemocyanin induces the activation and maturation of human dendritic cells through
[95] Ragupathi, G., J.R. Gardner, P.O. Livingston, and D.Y. Gin. 2011. Natural and syn‐ thetic saponin adjuvant QS-21 for vaccines against cancer. *Expert Rev Vaccines*.
[96] Ragupathi, G., P.O. Livingston, C. Hood, J. Gathuru, S.E. Krown, P.B. Chapman, J.D. Wolchok, L.J. Williams, R.C. Oldfield, and W.J. Hwu. 2003. Consistent antibody re‐ sponse against ganglioside GD2 induced in patients with melanoma by a GD2 lac‐ tone-keyhole limpet hemocyanin conjugate vaccine plus immunological adjuvant
[97] Ragupathi, G., M. Meyers, S. Adluri, L. Howard, C. Musselli, and P.O. Livingston. 2000. Induction of antibodies against GD3 ganglioside in melanoma patients by vac‐ cination with GD3-lactone-KLH conjugate plus immunological adjuvant QS-21. *Int J*
[98] Recker, F., and H. Rubben. 1989. Variation of the immunosystem by ciclosporin and keyhole-limpet hemocyanin--are there effects on chemically induced bladder carcino‐
[99] Reyes, D., L. Salazar, E. Espinoza, C. Pereda, E. Castellon, R. Valdevenito, C. Huido‐ bro, M. Ines Becker, A. Lladser, M.N. Lopez, and F. Salazar-Onfray. 2013. Tumour cell lysate-loaded dendritic cell vaccine induces biochemical and memory immune response in castration-resistant prostate cancer patients. *Br J Cancer*. 109:1488-1497.
[100] Riggs, D.R., B. Jackson, L. Vona-Davis, and D. McFadden. 2002. In vitro anticancer effects of a novel immunostimulant: keyhole limpet hemocyanin. *J Surg Res*.
[101] Riggs, D.R., B.J. Jackson, L. Vona-Davis, A. Nigam, and D.W. McFadden. 2005. In vi‐ tro effects of keyhole limpet hemocyanin in breast and pancreatic cancer in regards
[102] Riggs, D.R., W.F. Tarry, J.I. DeHaven, J. Sosnowski, and D.L. Lamm. 1992. Immuno‐ therapy of murine transitional cell carcinoma of the bladder using alpha and gamma interferon in combination with other forms of immunotherapy. *J Urol*. 147:212-214.
to cell growth, cytokine production, and apoptosis. *Am J Surg*. 189:680-684.
the involvement of mannose receptor. *Mol Immunol*. 45:1136-1145.
for superficial bladder cancer. *Urology*. 64:409-421.
429:161-164.
10:463-470.
*Cancer*. 85:659-666.
ma? *Urol Int*. 44:77-80.
108:279-284.
QS-21. *Clin Cancer Res*. 9:5214-5220.
[91] Peng, Y., Y. Zhang, W.J. Mitchell, and G. Zhang. 2012. Development of a lipopolysac‐ charide-targeted peptide mimic vaccine against Q fever. *J Immunol*. 189:4909-4920.
2011. Phase III multicenter clinical trial of the sialyl-TN (STn)-keyhole limpet hemo‐
cyanin (KLH) vaccine for metastatic breast cancer. O*ncologist*. 16:1092-1100.
comparison of several GnRH preparations. *Am J Reprod Immunol*. 60:214-223.
mocyanins. *J Urol*. 176:2690-2695.
immunity. *J Immunol*. 183:3569-3573.
*Res Clin Oncol*. 127 Suppl 2:R20-26.
92:199-205.
68 Immune Response Activation
[78] Miller, L.A., J.P. Gionfriddo, K.A. Fagerstone, J.C. Rhyan, and G.J. Killian. 2008. The single-shot GnRH immunocontraceptive vaccine (GonaCon) in white-tailed deer:
[79] Moltedo, B., F. Faunes, D. Haussmann, P. De Ioannes, A.E. De Ioannes, J. Puente, and M.I. Becker. 2006. Immunotherapeutic effect of Concholepas hemocyanin in the mur‐ ine bladder cancer model: evidence for conserved antitumor properties among he‐
[80] Moltedo, B., C.B. Lopez, M. Pazos, M.I. Becker, T. Hermesh, and T.M. Moran. 2009. Cutting edge: stealth influenza virus replication precedes the initiation of adaptive
[81] Morales, A., D. Eidinger, and A.W. Bruce. 1976. Intracavitary Bacillus Calmette-Gue‐
[82] Moreno, A.Y., and K.D. Janda. 2009. Immunopharmacotherapy: vaccination strat‐ egies as a treatment for drug abuse and dependence. *Pharmacol Biochem Behav*.
[83] Mura, C.V., M.I. Becker, A. Orellana, and D. Wolff. 2002. Immunopurification of Gol‐
[84] Musselli, C., P.O. Livingston, and G. Ragupathi. 2001. Keyhole limpet hemocyanin conjugate vaccines against cancer: the Memorial Sloan Kettering experience. *J Cancer*
[85] Neelapu, S.S., and L.W. Kwak. 2007. Vaccine therapy for B-cell lymphomas: nextgeneration strategies. *Hematology Am Soc Hematol Educ Program*. 2007:243-249.
[86] Nelson, E.L., X. Li, F.J. Hsu, L.W. Kwak, R. Levy, C. Clayberger, and A.M. Krensky. 1996. Tumor-specific, cytotoxic T-lymphocyte response after idiotype vaccination for
[87] Nestle, F.O., S. Alijagic, M. Gilliet, Y. Sun, S. Grabbe, R. Dummer, G. Burg, and D. Schadendorf. 1998. Vaccination of melanoma patients with peptide- or tumor lysate-
[88] Nseyo, U.O., and D.L. Lamm. 1997. Immunotherapy of bladder cancer. *Semin Surg*
[89] Olsson, C.A., R. Chute, and C.N. Rao. 1974. Immunologic reduction of bladder can‐
[90] Paccagnella, M., L. Bologna, M. Beccaro, I. Micetic, P. Di Muro, and B. Salvato. 2004. Structural subunit organization of molluscan hemocyanins. *Micron*. 35:21-22.
rin in the treatment of superficial bladder tumors. *J Urol*. 116:180-183.
gi vesicles by magnetic sorting. *J Immunol Methods*. 260:263-271.
B-cell, non-Hodgkin's lymphoma. *Blood*. 88:580-589.
pulsed dendritic cells. *Nat Med*. 4:328-332.
cer recurrence rate. *J Urol*. 111:173-176.
*Oncol*. 13:342-349.
[103] Rizvi, I., D.R. Riggs, B.J. Jackson, and D.W. McFadden. 2007. Keyhole limpet hemo‐ cyanin: an effective adjunct against melanoma in vivo. *Am J Surg*. 194:628-632.
[115] Tchorbanov, A., K. Idakieva, N. Mihaylova, and L. Doumanova. 2008. Modulation of the immune response using Rapana thomasiana hemocyanin. *Int Immunopharmacol*.
Mollusk Hemocyanins as Natural Immunostimulants in Biomedical Applications
http://dx.doi.org/10.5772/57552
71
[116] Teitz-Tennenbaum, S., Q. Li, M.A. Davis, and A.E. Chang. 2008. Dendritic cells pulsed with keyhole limpet hemocyanin and cryopreserved maintain anti-tumor ac‐
[117] Theilacker, C., F.T. Coleman, S. Mueschenborn, N. Llosa, M. Grout, and G.B. Pier. 2003. Construction and characterization of a Pseudomonas aeruginosa mucoid exo‐
[118] Timmerman, J.M., and R. Levy. 2000. Linkage of foreign carrier protein to a self-tu‐ mor antigen enhances the immunogenicity of a pulsed dendritic cell vaccine. *J Immu‐*
[119] Tittarelli, A., F.E. Gonzalez, C. Pereda, G. Mora, L. Munoz, C. Saffie, T. Garcia, D. Di‐ az, C. Falcon, M. Hermoso, M.N. Lopez, and F. Salazar-Onfray. Toll-like receptor 4 gene polymorphism influences dendritic cell in vitro function and clinical outcomes
in vaccinated melanoma patients. *Cancer Immunol Immunother*. 61:2067-2077.
[120] Tittarelli, A., F.E. Gonzalez, C. Pereda, G. Mora, L. Munoz, C. Saffie, T. Garcia, D. Di‐ az, C. Falcon, M. Hermoso, M.N. Lopez, and F. Salazar-Onfray. 2012. Toll-like recep‐ tor 4 gene polymorphism influences dendritic cell in vitro function and clinical outcomes in vaccinated melanoma patients. *Cancer Immunol Immunother.*
[121] Torres, M., H. Manosalva, I. Carrasco, A.E. De Ioannes, and M.I. Becker. 1999. Proce‐ dure for radiolabeling gizzerosine and basis for a radioimmunoassay. *J Agric Food*
[123] van Vliet, S.J., J.J. Garcia-Vallejo, and Y. van Kooyk. 2008. Dendritic cells and C-type lectin receptors: coupling innate to adaptive immune responses. *Immunol Cell Biol*.
[124] Velkova, L., P. Dolashka, B. Lieb, A. Dolashki, W. Voelter, J. Van Beeumen, and B. Devreese. Glycan structures of the structural subunit (HtH1) of Haliotis tuberculata
[125] Walsh, W.G., P. Tomashefsky, C.A. Olsson, and R. deVere White. 1983. Keyhole-lim‐ pet haemocyanin (KLH) immunotherapy of murine transitional cell carcinoma. *Urol*
[126] Wishahi, M.M., I.M. Ismail, H. Ruebben, and T. Otto. 1995. Keyhole-limpet hemocya‐ nin immunotherapy in the bilharzial bladder: a new treatment modality? Phase II tri‐
[122] van Holde, K.E., and K.I. Miller. 1995. Hemocyanins. *Adv Protein Chem*. 47:1-81.
polysaccharide-alginate conjugate vaccine. *Infect Immun*. 71:3875-3884.
tivity in a murine melanoma model. *Clin Immunol*. 129:482-491.
8:1033-1038.
*nol*. 164:4797-4803.
61:2067-2077.
86:580-587.
*Res*. 11:263-265.
*Chem*. 47:4231-4236.
hemocyanin. *Glycoconj J*. 28:385-395.
al: superficial bladder cancer. *J Urol*. 153:926-928.
[115] Tchorbanov, A., K. Idakieva, N. Mihaylova, and L. Doumanova. 2008. Modulation of the immune response using Rapana thomasiana hemocyanin. *Int Immunopharmacol*. 8:1033-1038.
[103] Rizvi, I., D.R. Riggs, B.J. Jackson, and D.W. McFadden. 2007. Keyhole limpet hemo‐ cyanin: an effective adjunct against melanoma in vivo. *Am J Surg*. 194:628-632. [104] Rollig, C., C. Schmidt, M. Bornhauser, G. Ehninger, M. Schmitz, and S. Auffermann-Gretzinger. 2011. Induction of cellular immune responses in patients with stage-I multiple myeloma after vaccination with autologous idiotype-pulsed dendritic cells.
[105] Sabbatini, P.J., G. Ragupathi, C. Hood, C.A. Aghajanian, M. Juretzka, A. Iasonos, M.L. Hensley, M.K. Spassova, O. Ouerfelli, D.R. Spriggs, W.P. Tew, J. Konner, H. Clausen, N. Abu Rustum, S.J. Dansihefsky, and P.O. Livingston. 2007. Pilot study of a heptavalent vaccine-keyhole limpet hemocyanin conjugate plus QS21 in patients with epithelial ovarian, fallopian tube, or peritoneal cancer. *Clin Cancer Res*.
[106] Salazar-Onfray, F., Pereda, C., Reyes, D, and M.N, López. 2013. TAPCells, the Chil‐ ean dendritic cell vaccine against melanoma and prostate cancer. Biol Res. ;
[107] Schenkman, E., and D.L. Lamm. 2004. Superficial bladder cancer therapy. *Scientific‐*
[108] Shimizu, K., E.K. Thomas, M. Giedlin, and J.J. Mule. 2001. Enhancement of tumor ly‐ sate- and peptide-pulsed dendritic cell-based vaccines by the addition of foreign
[109] Slovin, S.F., G. Ragupathi, C. Fernandez, M. Diani, M.P. Jefferson, A. Wilton, W.K. Kelly, M. Morris, D. Solit, H. Clausen, P. Livingston, and H.I. Scher. 2007. A polyva‐ lent vaccine for high-risk prostate patients: "are more antigens better?". *Cancer Immu‐*
[110] Somasundar, P., D.R. Riggs, B.J. Jackson, and D.W. McFadden. 2005. Inhibition of melanoma growth by hemocyanin occurs via early apoptotic pathways. *Am J Surg*.
[111] Steinman, R.M., and J. Banchereau. 2007. Taking dendritic cells into medicine. *Nature*.
[112] Su, Q.P., D.Z. Wen, Q. Yang, Y.H. Zhang, C. Liu, and L. Wang. 2007. Comparison of phage pVIII and KLH as vector in inducing the production of cytokines in C57BL/6J
[113] Swerdlow, R.D., R.F. Ebert, P. Lee, C. Bonaventura, and K.I. Miller. 1996. Keyhole limpet hemocyanin: structural and functional characterization of two different subu‐
[114] Swerdlow, R.D., T.L. Ratliff, M. La Regina, J.K. Ritchey, and R.F. Ebert. 1994. Immu‐ notherapy with keyhole limpet hemocyanin: efficacy and safety in the MB-49 intra‐
nits and multimers. *Comp Biochem Physiol B Biochem Mol Biol*. 113:537-548.
vesical murine bladder tumor model. *J Urol*. 151:1718-1722.
*J Immunother*. 34:100-106.
*WorldJournal*. 4 Suppl 1:387-399.
*nol Immunother*. 56:1921-1930.
mice. *Vaccine*. 25:970-975.
helper protein. *Cancer Res*. 61:2618-2624.
13:4170-4177.
70 Immune Response Activation
46:431-440.
190:713-716.
449:419-426.
[127] Zanjani, N.T., F. Sairi, G. Marshall, M.M. Saksena, P. Valtchev, V.G. Gomes, A.L. Cunningham, and F. Dehghani. 2013. Formulation of abalone hemocyanin with high antiviral activity and stability. *Eur J Pharm Sci*. 53:77-85.
**Section 2**
**Immunomodulation**
[128] Zhu, J., Q. Wan, D. Lee, G. Yang, M.K. Spassova, O. Ouerfelli, G. Ragupathi, P. Dam‐ ani, P.O. Livingston, and S.J. Danishefsky. 2009. From synthesis to biologics: preclini‐ cal data on a chemistry derived anticancer vaccine. *J Am Chem Soc*. 131:9298-9303.
### **Immunomodulation**
[127] Zanjani, N.T., F. Sairi, G. Marshall, M.M. Saksena, P. Valtchev, V.G. Gomes, A.L. Cunningham, and F. Dehghani. 2013. Formulation of abalone hemocyanin with high
[128] Zhu, J., Q. Wan, D. Lee, G. Yang, M.K. Spassova, O. Ouerfelli, G. Ragupathi, P. Dam‐ ani, P.O. Livingston, and S.J. Danishefsky. 2009. From synthesis to biologics: preclini‐ cal data on a chemistry derived anticancer vaccine. *J Am Chem Soc*. 131:9298-9303.
antiviral activity and stability. *Eur J Pharm Sci*. 53:77-85.
72 Immune Response Activation
**Chapter 3**
**Manipulating Immune Regulatory Pathways to Enhance**
Cancer immunotherapy aspires to treat malignant disease by activating cancer specific immune responses. It is generally accepted that the latter can only be achieved by an approach in which tumor specific T cells are educated to recognize and kill tumor cells, whilst they are furthermore empowered to overcome immunosuppressive mechanisms present both at peripheral sites and in the tumor environment. Dendritic cells (DCs) have been extensively explored as a cellular vaccine for the stimulation of tumor specific T cells. Several strategies have been devised to manipulate these cells to become strongly activated tumor associated antigen (TAA) presenting cells. Our growing knowledge on the biology of DCs and the costimulatory as well as inhibitory molecules expressed by them, provides us with opportunities to generate DCs that are capable of hyper-activating cytotoxic T lymphocytes (CTLs) whilst they impact on regulatory T cells (Treg), which are now well established to be an important contributor to failure of cancer vaccines. In this chapter, we will focus on the cross talk between DCs and T cells mediated by the CD70/CD27 and PD-L1/PD-1 axis as these have been identified
Cancer is a disease in which cells divide abnormally and acquire the ability to invade other tissues. Given the fact that virtually any tissue in the body can become transformed from normal cells towards cancer cells, a large degree of heterogeneity exists. This implies that cancer is an overall name covering a group of over 100 different diseases, which have in
> © 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
**T Cell Stimulation**
http://dx.doi.org/10.5772/57539
**1. Introduction**
David Escors and Karine Breckpot
Joeri J. Pen, Joeri. L. Aerts, Thérèse Liechtenstein,
as critical pathways in the regulation of immunity versus tolerance.
**2. Cancer — A therapeutic challenge**
Additional information is available at the end of the chapter
### **Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation**
Joeri J. Pen, Joeri. L. Aerts, Thérèse Liechtenstein, David Escors and Karine Breckpot
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/57539
#### **1. Introduction**
Cancer immunotherapy aspires to treat malignant disease by activating cancer specific immune responses. It is generally accepted that the latter can only be achieved by an approach in which tumor specific T cells are educated to recognize and kill tumor cells, whilst they are furthermore empowered to overcome immunosuppressive mechanisms present both at peripheral sites and in the tumor environment. Dendritic cells (DCs) have been extensively explored as a cellular vaccine for the stimulation of tumor specific T cells. Several strategies have been devised to manipulate these cells to become strongly activated tumor associated antigen (TAA) presenting cells. Our growing knowledge on the biology of DCs and the costimulatory as well as inhibitory molecules expressed by them, provides us with opportunities to generate DCs that are capable of hyper-activating cytotoxic T lymphocytes (CTLs) whilst they impact on regulatory T cells (Treg), which are now well established to be an important contributor to failure of cancer vaccines. In this chapter, we will focus on the cross talk between DCs and T cells mediated by the CD70/CD27 and PD-L1/PD-1 axis as these have been identified as critical pathways in the regulation of immunity versus tolerance.
#### **2. Cancer — A therapeutic challenge**
Cancer is a disease in which cells divide abnormally and acquire the ability to invade other tissues. Given the fact that virtually any tissue in the body can become transformed from normal cells towards cancer cells, a large degree of heterogeneity exists. This implies that cancer is an overall name covering a group of over 100 different diseases, which have in
© 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
common the occurrence of abnormal cells, which grow out of control and can be present within different parts of the body. Therefore, it is not surprising that cancer is a frequently occurring disease, affecting about one in three people during their lifetime in developed countries [1].
One of the major hurdles in the fight against cancer is the presence of regulatory T cells (Treg) in the tumor environment as well as in the periphery [9, 10]. Recruitment, expansion and *de novo* generation of Treg is a common theme in cancer (Figure 1). For instance, melanoma cells can secrete transforming growth factor-β (TGF-β) and interleukin-10 (IL-10), factors that are
Another well-described mechanism is the expression of the enzyme indoleamine 2,3-dioxy‐ genase (IDO) by tumor cells and tumor associated antigen presenting cells, which mediates
theless, it highlights that not only tumor cells drive the stimulation of Treg but that DCs, in particular immature DCs, trapped within the tumor environment, also have an important role in this process. Furthermore, it is well known that Treg prevent full maturation of DCs and as such are at the basis of a vicious circle [14]. The suppressive activity of Treg is mediated by various mechanisms [13]. As mentioned, they hamper full DC maturation [14]. Moreover, Treg can hijack the co-stimulatory molecules CD80 and CD86 from the membrane of DCs through transendocytosis via interactions with cytotoxic T lymphocyte antigen-4 (CTLA-4 or CD152). As such they deprive effector T (Teff) cells from co-stimulatory signals [15]. It has furthermore been shown that Treg influence the character of T cell differentiation by selectively dampening TH1 responses and CTLs [16, 17]. Importantly, Treg not only hamper DCs in their ability to induce CTL mediated anti-tumor immune responses, they furthermore hamper the functionality of Teff cells or can cause Teff cell death through exocytosis of granzymes and perforin [18]. It has also been described that Treg, which express high levels of the IL-2 receptor CD25, are capable of scavenging IL-2, thus depriving Teff from this cytokine [19]. Finally, Treg can disrupt metabolic pathways through tryptophan conversion [20] or adenosine secretion [21], which leads to the production of suppressive metabolites. In this way, Treg are able to
T cells to Treg [12]. The expression of IDO is only one out of many mecha‐
T cells to inducible Treg [11].
http://dx.doi.org/10.5772/57539
77
T cell differentiation towards Treg [13]. None‐
Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation
linked to expansion of natural Treg and differentiation of naive CD4+
control the induction and/or presence of CTLs as well as their functionality.
Taken together, therapeutic vaccination against tumors takes place in the presence of a milieu that counteracts tumor specific T-cell responses. Therefore, the method of vaccination, *e.g.* DCbased vaccination, should fulfill several requirements. It should result in strong activation of tumor specific CTLs that are able to migrate to the tumor site, recognize and kill tumor cells. Furthermore, the activated CTLs should be refractory to the immunosuppressive mechanisms exerted for instance by Treg. Linked herewith, the vaccination strategy should avoid expansion of Treg or even drive the TGF-β characterized Treg signature into a more favorable IFN-γ
**3. The rationale behind therapeutic anti-cancer vaccination: tumor antigens**
Active therapeutic vaccination builds on the power and specificity of the patient's im‐ mune system to eradicate cancer cells. The latter is based on the ability of immune cells to discriminate healthy cells from tumor cells, mainly based on the expression of distinct antigens, the so called TAAs. It has to be noted that in contrast to environmentally induced or virus-related tumors (e.g. HPV induced cervix carcinoma, causing genomic mutations
nisms that are exploited by DCs to drive CD4+
conversion of CD4+
signature [22-24].
Despite the enormous amount of research and rapid developments in prevention, diagnosis and treatment, cancer remains one of the most difficult diseases to cure. Today, the first line treatment for most solid cancer types is surgical resection or radiotherapy to remove or necrose the primary tumor. Since these strategies do not tackle residual tumor cells or metastases, a commonly encountered problem is disease relapse. Therefore, the aforementioned treatment strategies are often combined with chemotherapy, a systemic but aggressive and non-selective approach to eradicate tumor cells. Despite major improvements in these conventional treat‐ ment strategies, the requirement to develop alternative treatment modalities that are tumor cell specific and provide better long-term protection to avoid recurrence of tumor cells remains high [2].
A very promising new approach that fulfills these criteria is immunotherapy. The main premise of cancer immunotherapy is to harness the patients' immune system to specifically recognize and kill tumor cells. This is based on the knowledge that our immune system can discriminate healthy cells from cancer cells as the latter express TAAs [3]. Several strategies have been developed to achieve destruction of TAA expressing cancer cells. These are extensively reviewed in [4]. Of these the exploitation of dendritic cells (DCs), the professional antigen presenting cells of our immune system, to present TAAs to CD4+ T helper 1 (TH1) cells and CD8+ cytotoxic T lymphocytes (CTLs) is considered to be very promising. Of note, in 2010 the FDA approved a first DC-vaccine, Sipuleucel-T (Provenge®), which was shown to be effective in metastatic, asymptomatic, hormone refractory prostate cancer [5, 6]. Thus, cancer immunotherapy could be a powerful new treatment strategy that oncologists can offer to patients.
It needs to be clarified that cancer development is a complex progressive process that involves a sequence of interactions between cancer cells and immune cells. As mentioned above, the immune system can specifically identify and eliminate cancer cells based on the expression of TAAs. This process was first described by Burnet and Thomas and is generally referred to as tumor immunosurveillance [7]. It occurs physiologically to protect the body from tumor formation. Nevertheless, this process is not perfect and tumors can develop despite tumor immunosurveillance. In fact, the manifestation of malignancy means that the disease has eventually prevailed over immunity. The latter can be explained by the more recently descri‐ bed notion of "tumor immunoediting", a process that comprises three phases [8]. The first phase is the elimination phase in which most immunogenic tumor cells are eliminated by CTLs and natural killer (NK) cells. This phase is followed by an equilibrium phase in which tumor cells that show a reduced immunogenicity are selected for. These tumor cell variants enable the third phase, the escape phase, as they are no longer sensitive to the host immune system. Nonetheless, T cells recognizing these tumor cell variants are often isolated from blood and tumor of cancer patients. Thus the question arises why these cells are unable to eradicate tumor cells?
One of the major hurdles in the fight against cancer is the presence of regulatory T cells (Treg) in the tumor environment as well as in the periphery [9, 10]. Recruitment, expansion and *de novo* generation of Treg is a common theme in cancer (Figure 1). For instance, melanoma cells can secrete transforming growth factor-β (TGF-β) and interleukin-10 (IL-10), factors that are linked to expansion of natural Treg and differentiation of naive CD4+ T cells to inducible Treg [11]. Another well-described mechanism is the expression of the enzyme indoleamine 2,3-dioxy‐ genase (IDO) by tumor cells and tumor associated antigen presenting cells, which mediates conversion of CD4+ T cells to Treg [12]. The expression of IDO is only one out of many mecha‐ nisms that are exploited by DCs to drive CD4+ T cell differentiation towards Treg [13]. None‐ theless, it highlights that not only tumor cells drive the stimulation of Treg but that DCs, in particular immature DCs, trapped within the tumor environment, also have an important role in this process. Furthermore, it is well known that Treg prevent full maturation of DCs and as such are at the basis of a vicious circle [14]. The suppressive activity of Treg is mediated by various mechanisms [13]. As mentioned, they hamper full DC maturation [14]. Moreover, Treg can hijack the co-stimulatory molecules CD80 and CD86 from the membrane of DCs through transendocytosis via interactions with cytotoxic T lymphocyte antigen-4 (CTLA-4 or CD152). As such they deprive effector T (Teff) cells from co-stimulatory signals [15]. It has furthermore been shown that Treg influence the character of T cell differentiation by selectively dampening TH1 responses and CTLs [16, 17]. Importantly, Treg not only hamper DCs in their ability to induce CTL mediated anti-tumor immune responses, they furthermore hamper the functionality of Teff cells or can cause Teff cell death through exocytosis of granzymes and perforin [18]. It has also been described that Treg, which express high levels of the IL-2 receptor CD25, are capable of scavenging IL-2, thus depriving Teff from this cytokine [19]. Finally, Treg can disrupt metabolic pathways through tryptophan conversion [20] or adenosine secretion [21], which leads to the production of suppressive metabolites. In this way, Treg are able to control the induction and/or presence of CTLs as well as their functionality.
common the occurrence of abnormal cells, which grow out of control and can be present within different parts of the body. Therefore, it is not surprising that cancer is a frequently occurring disease, affecting about one in three people during their lifetime in developed countries [1].
Despite the enormous amount of research and rapid developments in prevention, diagnosis and treatment, cancer remains one of the most difficult diseases to cure. Today, the first line treatment for most solid cancer types is surgical resection or radiotherapy to remove or necrose the primary tumor. Since these strategies do not tackle residual tumor cells or metastases, a commonly encountered problem is disease relapse. Therefore, the aforementioned treatment strategies are often combined with chemotherapy, a systemic but aggressive and non-selective approach to eradicate tumor cells. Despite major improvements in these conventional treat‐ ment strategies, the requirement to develop alternative treatment modalities that are tumor cell specific and provide better long-term protection to avoid recurrence of tumor cells remains
A very promising new approach that fulfills these criteria is immunotherapy. The main premise of cancer immunotherapy is to harness the patients' immune system to specifically recognize and kill tumor cells. This is based on the knowledge that our immune system can discriminate healthy cells from cancer cells as the latter express TAAs [3]. Several strategies have been developed to achieve destruction of TAA expressing cancer cells. These are extensively reviewed in [4]. Of these the exploitation of dendritic cells (DCs), the professional
the FDA approved a first DC-vaccine, Sipuleucel-T (Provenge®), which was shown to be effective in metastatic, asymptomatic, hormone refractory prostate cancer [5, 6]. Thus, cancer immunotherapy could be a powerful new treatment strategy that oncologists can offer to
It needs to be clarified that cancer development is a complex progressive process that involves a sequence of interactions between cancer cells and immune cells. As mentioned above, the immune system can specifically identify and eliminate cancer cells based on the expression of TAAs. This process was first described by Burnet and Thomas and is generally referred to as tumor immunosurveillance [7]. It occurs physiologically to protect the body from tumor formation. Nevertheless, this process is not perfect and tumors can develop despite tumor immunosurveillance. In fact, the manifestation of malignancy means that the disease has eventually prevailed over immunity. The latter can be explained by the more recently descri‐ bed notion of "tumor immunoediting", a process that comprises three phases [8]. The first phase is the elimination phase in which most immunogenic tumor cells are eliminated by CTLs and natural killer (NK) cells. This phase is followed by an equilibrium phase in which tumor cells that show a reduced immunogenicity are selected for. These tumor cell variants enable the third phase, the escape phase, as they are no longer sensitive to the host immune system. Nonetheless, T cells recognizing these tumor cell variants are often isolated from blood and tumor of cancer patients. Thus the question arises why these cells are unable to eradicate tumor
cytotoxic T lymphocytes (CTLs) is considered to be very promising. Of note, in 2010
T helper 1 (TH1) cells
antigen presenting cells of our immune system, to present TAAs to CD4+
high [2].
76 Immune Response Activation
and CD8+
patients.
cells?
Taken together, therapeutic vaccination against tumors takes place in the presence of a milieu that counteracts tumor specific T-cell responses. Therefore, the method of vaccination, *e.g.* DCbased vaccination, should fulfill several requirements. It should result in strong activation of tumor specific CTLs that are able to migrate to the tumor site, recognize and kill tumor cells. Furthermore, the activated CTLs should be refractory to the immunosuppressive mechanisms exerted for instance by Treg. Linked herewith, the vaccination strategy should avoid expansion of Treg or even drive the TGF-β characterized Treg signature into a more favorable IFN-γ signature [22-24].
#### **3. The rationale behind therapeutic anti-cancer vaccination: tumor antigens**
Active therapeutic vaccination builds on the power and specificity of the patient's im‐ mune system to eradicate cancer cells. The latter is based on the ability of immune cells to discriminate healthy cells from tumor cells, mainly based on the expression of distinct antigens, the so called TAAs. It has to be noted that in contrast to environmentally induced or virus-related tumors (e.g. HPV induced cervix carcinoma, causing genomic mutations
immunological control of cancer cells [30]. Initiation of CTL mediated immune responses requires antigen presentation in MHC I molecules that are present on antigen presenting cells,
that promote effective display of antigenic peptide/MHC I complexes. Importantly, DCs also
CTLs [31]. Although antigen presentation is a first prerequisite for stimulation of CTLs, additional signals, co-stimulation by surface expressed stimulatory molecules and an inflam‐ matory environment provided by the secretion of a plethora of cytokines, are required. Both
Over a hundred TAAs that are recognized by T cells have been identified to date [33]. The question which TAAs has the most potential as a target for therapeutic vaccination was recently addressed in a pilot study in which 75 TAAs were ranked according to predefined and weighted objective criteria, including its therapeutic function, immunogenicity and specificity. This study brought the Wilms' Tumor 1 (WT-1) antigen forward as an excellent, nearly
Over the years, several strategies have been developed to introduce TAAs to DCs. Generally these approaches are divided into viral and non-viral strategies. Examples hereof are lentiviral modification and mRNA electroporation of DCs, respectively [35-41]. Both strategies have been shown to be efficient for delivery of TAAs to DCs [42]. Of note, these approaches have later on been fine tuned to furthermore provide immune modulating signals to the DCs [32, 43-46]. A thorough review of the use of lentiviral vectors and mRNA for modification of DCs
Dendritic cells, the most potent antigen presenting cells of our immune system, are at the centre of our immune system. They bridge the innate and adaptive immune system, and orchestrate amongst others T-cell mediated immune responses. Therefore, it is not surprising that DCs are
Immature DCs have the intrinsic capacity to scan peripheral tissues for antigens (both selfantigens and foreign antigens) and capture these antigens using several complementary mechanisms [52]. Intrinsically, DCs have a poor T-cell activating capacity as they express low levels of MHC and co-stimulatory molecules. In order to potently stimulate antigen-specific T-cell responses, DCs need to be fully matured, which will be described below. It is however important to note that immature DCs are not immunologically quiescent as they induce T-cell tolerance against self-antigens through several mechanisms such as depletion of T cells, induction of T-cell anergy, Treg amplification or differentiation [53-55]. When immature DCs capture an antigen under pro-inflammatory circumstances, in the presence of pathogen associated molecular patterns or danger associated molecular patterns they undergo a maturation process [56]. This process is characterized by a decreased capacity for phagocytosis and an increased expression of MHC and co-stimulatory molecules, such as CD80 and CD86. In addition, the chemokine receptor CCR7 is up-regulated, allowing DCs to migrate to
**T cell activation**
frequently targeted to generate therapeutic immunity against cancer cells [51].
present peptides in the context of MHC II molecules to CD4+
is provided in references [47, 48] and [49, 50] respectively.
can be delivered by DCs when these are appropriately activated [32].
T lymphocytes. In particular, DCs are equipped with specialized machinery
T cells help in the induction, expansion and maintenance of
Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation
T cells, which through delicate
http://dx.doi.org/10.5772/57539
79
to the TCR of CD8+
universal TAA [34].
**3.1. Dendritic cells guide CD8+**
interactions with DCs and CD8+
**Figure 1.** Treg are a major hurdle in cancer. It is well known Treg can be found at elevated numbers in cancers. Three main mechanisms are at the basis of this phenomenon. Firstly, CD4+CD25+Foxp3+Treg or so-called naturally occurring Treg infiltrate the tumor. The latter is driven by the expression of the chemokine receptor CCR4 on the Treg and the ex‐ pression of its ligand CCL22 in the tumor environment. Secondly, several immune cells found in the tumor environ‐ ment, including MDSCs, DCs, macrophages and natural Treg, as well as tumor cells themselves drive the differentiation of CD4+Foxp3 cells into so-called inducible Treg (Foxp3+) and/or IL-10 producing Tr1 cells (CD4+IL-10+Foxp3- ) through the secretion of TGF-β and/or IL-10. Thirdly, it is well known that dysfunctional myeloid cells within the tumor environ‐ ment further stimulate Treg expansion. Importantly, there is an extensive cross talk between the different immune cells found within the tumor as well as between these immune cells and the tumor cells themselves. This cross talk ensures that the tumor environment remains a sanctuary for immunosuppresive cells, such as the Treg.
resulting in strongly immogenic antigens), for most tumors the etiology is unclear and that they are not particularly immunogenic. Nonetheless, several TAAs have been identified as immunological targets. TAAs can be classified in different classes, dependent on their expression pattern. The most important classes are the cancer-testis antigens (e.g. MAGEfamily), differentiation antigens (e.g. MART-1/Melan-A, CEA), mutated antigens (e.g. BRAF), antigens that are overexpressed in tumors (e.g. hTERT, WT-1). Examples hereof are carcinoembryonic antigen in colorectal carcinoma [25], α-fetoprotein in hepatocellular carcinoma [26] and members of the melanoma antigen gene (MAGE) family in melanoma [27].
The immune cells that are responsible for the specific recognition of TAAs are T cells. Peptides derived from a TAA can be presented in major histocompatibility class (MHC) molecules to the T-cell receptor (TCR) of T cells offering the opportunity to induce specific and long lasting immune reactivity against TAAs [28, 29]. As mentioned above, CTLs are critical in the immunological control of cancer cells [30]. Initiation of CTL mediated immune responses requires antigen presentation in MHC I molecules that are present on antigen presenting cells, to the TCR of CD8+ T lymphocytes. In particular, DCs are equipped with specialized machinery that promote effective display of antigenic peptide/MHC I complexes. Importantly, DCs also present peptides in the context of MHC II molecules to CD4+ T cells, which through delicate interactions with DCs and CD8+ T cells help in the induction, expansion and maintenance of CTLs [31]. Although antigen presentation is a first prerequisite for stimulation of CTLs, additional signals, co-stimulation by surface expressed stimulatory molecules and an inflam‐ matory environment provided by the secretion of a plethora of cytokines, are required. Both can be delivered by DCs when these are appropriately activated [32].
Over a hundred TAAs that are recognized by T cells have been identified to date [33]. The question which TAAs has the most potential as a target for therapeutic vaccination was recently addressed in a pilot study in which 75 TAAs were ranked according to predefined and weighted objective criteria, including its therapeutic function, immunogenicity and specificity. This study brought the Wilms' Tumor 1 (WT-1) antigen forward as an excellent, nearly universal TAA [34].
Over the years, several strategies have been developed to introduce TAAs to DCs. Generally these approaches are divided into viral and non-viral strategies. Examples hereof are lentiviral modification and mRNA electroporation of DCs, respectively [35-41]. Both strategies have been shown to be efficient for delivery of TAAs to DCs [42]. Of note, these approaches have later on been fine tuned to furthermore provide immune modulating signals to the DCs [32, 43-46]. A thorough review of the use of lentiviral vectors and mRNA for modification of DCs is provided in references [47, 48] and [49, 50] respectively.
#### **3.1. Dendritic cells guide CD8+ T cell activation**
resulting in strongly immogenic antigens), for most tumors the etiology is unclear and that they are not particularly immunogenic. Nonetheless, several TAAs have been identified as immunological targets. TAAs can be classified in different classes, dependent on their expression pattern. The most important classes are the cancer-testis antigens (e.g. MAGEfamily), differentiation antigens (e.g. MART-1/Melan-A, CEA), mutated antigens (e.g. BRAF), antigens that are overexpressed in tumors (e.g. hTERT, WT-1). Examples hereof are carcinoembryonic antigen in colorectal carcinoma [25], α-fetoprotein in hepatocellular carcinoma [26] and members of the melanoma antigen gene (MAGE) family in melanoma
that the tumor environment remains a sanctuary for immunosuppresive cells, such as the Treg.
**Figure 1.** Treg are a major hurdle in cancer. It is well known Treg can be found at elevated numbers in cancers. Three main mechanisms are at the basis of this phenomenon. Firstly, CD4+CD25+Foxp3+Treg or so-called naturally occurring Treg infiltrate the tumor. The latter is driven by the expression of the chemokine receptor CCR4 on the Treg and the ex‐ pression of its ligand CCL22 in the tumor environment. Secondly, several immune cells found in the tumor environ‐ ment, including MDSCs, DCs, macrophages and natural Treg, as well as tumor cells themselves drive the differentiation
cells into so-called inducible Treg (Foxp3+) and/or IL-10 producing Tr1 cells (CD4+IL-10+Foxp3-
the secretion of TGF-β and/or IL-10. Thirdly, it is well known that dysfunctional myeloid cells within the tumor environ‐ ment further stimulate Treg expansion. Importantly, there is an extensive cross talk between the different immune cells found within the tumor as well as between these immune cells and the tumor cells themselves. This cross talk ensures
) through
The immune cells that are responsible for the specific recognition of TAAs are T cells. Peptides derived from a TAA can be presented in major histocompatibility class (MHC) molecules to the T-cell receptor (TCR) of T cells offering the opportunity to induce specific and long lasting immune reactivity against TAAs [28, 29]. As mentioned above, CTLs are critical in the
[27].
of CD4+Foxp3-
78 Immune Response Activation
Dendritic cells, the most potent antigen presenting cells of our immune system, are at the centre of our immune system. They bridge the innate and adaptive immune system, and orchestrate amongst others T-cell mediated immune responses. Therefore, it is not surprising that DCs are frequently targeted to generate therapeutic immunity against cancer cells [51].
Immature DCs have the intrinsic capacity to scan peripheral tissues for antigens (both selfantigens and foreign antigens) and capture these antigens using several complementary mechanisms [52]. Intrinsically, DCs have a poor T-cell activating capacity as they express low levels of MHC and co-stimulatory molecules. In order to potently stimulate antigen-specific T-cell responses, DCs need to be fully matured, which will be described below. It is however important to note that immature DCs are not immunologically quiescent as they induce T-cell tolerance against self-antigens through several mechanisms such as depletion of T cells, induction of T-cell anergy, Treg amplification or differentiation [53-55]. When immature DCs capture an antigen under pro-inflammatory circumstances, in the presence of pathogen associated molecular patterns or danger associated molecular patterns they undergo a maturation process [56]. This process is characterized by a decreased capacity for phagocytosis and an increased expression of MHC and co-stimulatory molecules, such as CD80 and CD86. In addition, the chemokine receptor CCR7 is up-regulated, allowing DCs to migrate to lymphoid organs. While DCs migrate to lymph nodes, they process the captured antigens into peptides that primarily bind to MHC II molecules for presentation to CD4+ T cells. Importantly, exogenous antigens can enter the MHC I presentation pathway for presentation to CD8+ T cells through a process called cross-presentation. Although the mechanisms allowing exogenous antigens to enter the MHC I pathway are not fully understood, two main pathways have been described, leakage of antigens from phagosomes/endosomes to the cytosol and loading of peptides generated by lysosomal degradation onto MHC I molecules that are recycled [57]. The ability to cross-present exogenous antigens, such as TAAs, acquired from for instance dying tumor cells is of utmost importance in cancer immunosurveillance [58].
**3.2. Co-signaling by dendritic cells**
cell and T cells [69].
these groups [74, 75].
focus on the PD-L1/PD-1 pathway".
Mature DCs express a plethora of diverse co-signaling molecules. These bind to their partner‐ ing receptor on T cells and as such deliver various signals that contribute in their own way to the quality of the T cell response. Co-signaling molecules can act on particular aspects of T cell activation, such as survival, cell cycle progression and differentiation to either Teff cell or memory T cells [64]. It is furthermore important to note that intracellular signals may be transduced through both the co-signaling ligand and its receptor, a concept known as bidir‐ ectional signaling [65-68]. Consequently, co-signaling can affect both the antigen presenting
The function of a particular co-signaling molecule is strongly related to the timing of its action, as early co-signaling molecules are functionally distinct from those that act late during the Tcell response. It is for this reason that the expression of each co-signaling molecule and/or its
Another important issue is that the outcome of the signaling is not necessarily stimulatory hence co-signaling molecules can be categorized as being co-stimulatory or co-inhibitory based on the outcome of their signals. Linked herewith, it is worthwhile noting that co-signaling molecules are as essential in determining the expansion and function of Teff cells as they are
There is a growing list of co-stimulatory molecules, of which most belong to either the B7 or the TNF receptor superfamily, and are expressed on DCs and other antigen presenting cells [60, 72, 73]. Some co-stimulatory molecules however, exemplified by CD83 and the CD300 family, which is also expressed on mature DCs, cannot be classified in either of
The co-stimulatory function of B7.1 (CD80) and B7.2 (CD86), which are the founding members of the B7 family and which themselves belong to the immunoglobulin superfamily, have been studied most extensively. These co-stimulatory molecules and their receptor CD28 were extensively discussed in [76] and are therefore not addressed in this chapter. It needs to be highlighted that the B7 family now encompasses several other members, which were more recently described. These include inducible co-stimulator ligand (ICOS-L), programmed death 1 receptor-ligand 1 (PD-L1 or B7-H1), PD-L2 (B7-DC), B7-H3 and B7-H4. The corresponding ligands that are inducibly expressed on T cells are ICOS, programmed death 1 receptor (PD-1), and B and T Lymphocyte Attenuator (BTLA). Although their function in T-cell activation is not yet fully understood, it is clear that these B7 members are not innocent bystanders [77]. In this regard, we will discuss the PD-L1/PD-1 axis in more detail in the section "Co-inhibition:
The second group of co-stimulatory receptors expressed on T cells consists of the TNF receptor type family and includes CD27, OX40 (CD134), glucocorticoid induced TNF receptor family related protein (GITR, TNFRSF18; CD357) and 4-1BB (CD137). Their corresponding ligands CD70, OX40L (CD252), GITRL (TNFSF18) and 4-1BBL (TNFSF9) are expressed on DCs [73]. A
CD25+
Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation
http://dx.doi.org/10.5772/57539
81
Foxp3+
cells) [71].
receptor is tightly regulated and depends on the activation status of the cell [70].
for the development and function of Treg (defined as CD4+
**3.3. Co-stimulation: focus on the CD70/CD27 pathway**
In the lymph nodes, fully mature, antigen presenting DCs can activate naive CD4+ and CD8+ T cells. While naive CD8+ T cells can differentiate into CTLs, naive CD4+ T cells can differenti‐ ate into different TH cell types, of which TH1 cells are critical in the induction of potent antitumor immune responses. In this context, it is important to note that both CD4+ and CD8+ T cells require a minimum level of MHC/peptide-TCR interactions for proper activation. Too few MHC/peptide-TCR interactions can result in T-cell ignorance or tolerance while persistent interactions can lead to induction of T-cell anergy or deletion [59]. Although recognition of antigen, presented on DCs as peptide/MHC complexes is considered the first signal required for T-cell activation, it is not sufficient. Adequate activation of T cells at least requires co-stimulation, the so-called second signal, delivered by surface expressed molecules. Co-stimulation can promote more efficient engagement of TCR molecules in order to enhance the immunological synapse or can provide additional signals to pro‐ mote cell proliferation, augment cell survival or induce effector functions such as cyto‐ kine secretion or cytotoxicity [60]. Signaling via the TCR and additional co-stimulatory receptors initiates clonal expansion and acquisition of effector functions by primed T cells [61]. The requirement for a third signal, delivered as inflammatory cytokines such as IL-12, to stimulate antigen dependent proliferation is variable. When antigen levels are low, the third signal is critical to induce maximal T-cell proliferation. However, at high antigen levels extensive proliferation can occur in its absence. Importantly, in the absence of a third signal proliferating T cells often fail to develop full effector functions. Thus, proliferation and development of cytolytic function can be fully uncoupled and the absence of the third signal can render T cells functionally tolerant, in the sense that subsequent re-stimulation with a potent stimulus results in limited clonal expansion, impaired IFN-γ production and no cytolytic function. Thus, the presence or absence of the third signal appears to be a critical variable in determining whether antigenic challenge results in tolerance *versus* develop‐ ment of effector function and establishment of a responsive memory population [62]. It is important to note that activated CD8+ CTLs disseminate to the periphery (*e.g.* tumor site) and mediate efficient and rapid killing of target cells by various mechanisms, including secretion of cytokines such as tumor necrosis factor-α (TNF-α), exocytosis of lytic gran‐ ules containing granzymes and perforin, or by ligation of death receptors such as Fas on target cells [63].
#### **3.2. Co-signaling by dendritic cells**
lymphoid organs. While DCs migrate to lymph nodes, they process the captured antigens into
through a process called cross-presentation. Although the mechanisms allowing exogenous antigens to enter the MHC I pathway are not fully understood, two main pathways have been described, leakage of antigens from phagosomes/endosomes to the cytosol and loading of peptides generated by lysosomal degradation onto MHC I molecules that are recycled [57]. The ability to cross-present exogenous antigens, such as TAAs, acquired from for instance
T cells can differentiate into CTLs, naive CD4+
ate into different TH cell types, of which TH1 cells are critical in the induction of potent anti-
cells require a minimum level of MHC/peptide-TCR interactions for proper activation. Too few MHC/peptide-TCR interactions can result in T-cell ignorance or tolerance while persistent interactions can lead to induction of T-cell anergy or deletion [59]. Although recognition of antigen, presented on DCs as peptide/MHC complexes is considered the first signal required for T-cell activation, it is not sufficient. Adequate activation of T cells at least requires co-stimulation, the so-called second signal, delivered by surface expressed molecules. Co-stimulation can promote more efficient engagement of TCR molecules in order to enhance the immunological synapse or can provide additional signals to pro‐ mote cell proliferation, augment cell survival or induce effector functions such as cyto‐ kine secretion or cytotoxicity [60]. Signaling via the TCR and additional co-stimulatory receptors initiates clonal expansion and acquisition of effector functions by primed T cells [61]. The requirement for a third signal, delivered as inflammatory cytokines such as IL-12, to stimulate antigen dependent proliferation is variable. When antigen levels are low, the third signal is critical to induce maximal T-cell proliferation. However, at high antigen levels extensive proliferation can occur in its absence. Importantly, in the absence of a third signal proliferating T cells often fail to develop full effector functions. Thus, proliferation and development of cytolytic function can be fully uncoupled and the absence of the third signal can render T cells functionally tolerant, in the sense that subsequent re-stimulation with a potent stimulus results in limited clonal expansion, impaired IFN-γ production and no cytolytic function. Thus, the presence or absence of the third signal appears to be a critical variable in determining whether antigenic challenge results in tolerance *versus* develop‐ ment of effector function and establishment of a responsive memory population [62]. It is
and mediate efficient and rapid killing of target cells by various mechanisms, including secretion of cytokines such as tumor necrosis factor-α (TNF-α), exocytosis of lytic gran‐ ules containing granzymes and perforin, or by ligation of death receptors such as Fas on
exogenous antigens can enter the MHC I presentation pathway for presentation to CD8+
T cells. Importantly,
T cells
and CD8+
and CD8+
T cells can differenti‐
CTLs disseminate to the periphery (*e.g.* tumor site)
T
T
peptides that primarily bind to MHC II molecules for presentation to CD4+
dying tumor cells is of utmost importance in cancer immunosurveillance [58].
cells. While naive CD8+
80 Immune Response Activation
important to note that activated CD8+
target cells [63].
In the lymph nodes, fully mature, antigen presenting DCs can activate naive CD4+
tumor immune responses. In this context, it is important to note that both CD4+
Mature DCs express a plethora of diverse co-signaling molecules. These bind to their partner‐ ing receptor on T cells and as such deliver various signals that contribute in their own way to the quality of the T cell response. Co-signaling molecules can act on particular aspects of T cell activation, such as survival, cell cycle progression and differentiation to either Teff cell or memory T cells [64]. It is furthermore important to note that intracellular signals may be transduced through both the co-signaling ligand and its receptor, a concept known as bidir‐ ectional signaling [65-68]. Consequently, co-signaling can affect both the antigen presenting cell and T cells [69].
The function of a particular co-signaling molecule is strongly related to the timing of its action, as early co-signaling molecules are functionally distinct from those that act late during the Tcell response. It is for this reason that the expression of each co-signaling molecule and/or its receptor is tightly regulated and depends on the activation status of the cell [70].
Another important issue is that the outcome of the signaling is not necessarily stimulatory hence co-signaling molecules can be categorized as being co-stimulatory or co-inhibitory based on the outcome of their signals. Linked herewith, it is worthwhile noting that co-signaling molecules are as essential in determining the expansion and function of Teff cells as they are for the development and function of Treg (defined as CD4+ CD25+ Foxp3+ cells) [71].
#### **3.3. Co-stimulation: focus on the CD70/CD27 pathway**
There is a growing list of co-stimulatory molecules, of which most belong to either the B7 or the TNF receptor superfamily, and are expressed on DCs and other antigen presenting cells [60, 72, 73]. Some co-stimulatory molecules however, exemplified by CD83 and the CD300 family, which is also expressed on mature DCs, cannot be classified in either of these groups [74, 75].
The co-stimulatory function of B7.1 (CD80) and B7.2 (CD86), which are the founding members of the B7 family and which themselves belong to the immunoglobulin superfamily, have been studied most extensively. These co-stimulatory molecules and their receptor CD28 were extensively discussed in [76] and are therefore not addressed in this chapter. It needs to be highlighted that the B7 family now encompasses several other members, which were more recently described. These include inducible co-stimulator ligand (ICOS-L), programmed death 1 receptor-ligand 1 (PD-L1 or B7-H1), PD-L2 (B7-DC), B7-H3 and B7-H4. The corresponding ligands that are inducibly expressed on T cells are ICOS, programmed death 1 receptor (PD-1), and B and T Lymphocyte Attenuator (BTLA). Although their function in T-cell activation is not yet fully understood, it is clear that these B7 members are not innocent bystanders [77]. In this regard, we will discuss the PD-L1/PD-1 axis in more detail in the section "Co-inhibition: focus on the PD-L1/PD-1 pathway".
The second group of co-stimulatory receptors expressed on T cells consists of the TNF receptor type family and includes CD27, OX40 (CD134), glucocorticoid induced TNF receptor family related protein (GITR, TNFRSF18; CD357) and 4-1BB (CD137). Their corresponding ligands CD70, OX40L (CD252), GITRL (TNFSF18) and 4-1BBL (TNFSF9) are expressed on DCs [73]. A common theme for the members of this family is their coordinated expression at the interface between T cells and DCs, and their ability to shape the environment that sets the stage for the cross talk between T cells and DCs. In the remainder of this section we will focus on the CD70/ CD27 pathway as this pathway has been identified as a critical regulator of immunity *versus* tolerance.
Similar to the B7 receptor CD28, the receptor CD27 is expressed on naive T cells [78]. Although initially up-regulated upon T cell activation, its expression decreases to undetectable levels after T cells undergo several rounds of division [79]. However, CD27 expression has been reported on central memory T cells [80]. The expression of its ligand CD70 is highly restricted and activation dependent, as it is only transiently expressed on activated T and B cells as well as DCs [81-83]. Studies using CD27 deficient mice have provided insights into the role of CD27 in the generation of T cell immunity. It was indicated that CD27 is of importance in the stimulation of both primary and memory T-cell responses [84]. Although, CD27 seemingly was not required to induce T-cell proliferation, it was shown that CD27 enhances T-cell survival upon successive rounds of cell division [85]. Nonetheless, a role for CD27 in CD8+ T cell proliferation was shown in the absence of IL-2, although no functional differentiation was detected [86]. In this regard, CD27 has been linked to the induction of IL-2 and IFN-γ secretion by CD4+ T cells, as such instructing the latter to support CD8+ T-cell activation [87]. These findings were confirmed in mouse studies using blocking anti-CD70 antibodies [88, 89]. An observation that pinpointed the CD27/CD70 pathway as one of the most important costimulatory pathways known today, is the observation that due to their CD70 expression, DEC205+ spleen DCs are able to prime CD4+ T cells in the absence of a third signal, *e.g.* IL-12 [90]. Moreover, it was demonstrated in mice that stimulation of T cells by DCs that constitutively express CD70 is effective in the absence of signal 3, suggesting that TCR triggering and CD27/ CD70 interaction-at least in mice-is sufficient to promote CTL responses [91]. Many of these studies were performed in the context of infectious diseases. Also in cancer, the CD27/CD70 interaction has been described to help immune cell survival and induction of potent CTL responses [92-95]. Importantly, our group has developed a potent DC-based anti-cancer vaccine based on the use of mRNA to introduce CD70 in addition to two DC activating stimuli CD40 ligand and a constitutive active form of toll-like receptor 4, collectively referred to as TriMix (Figure 2) [43]. It has been demonstrated that these DCs are able to strongly activate allogeneic naive CD4+ CD45RA+ T cells and when loaded with TAA derived peptides are able to hyper-activate CD8+ T cells (>500-fold increase in Melan A-specific CD8+ T cells was observed when compared with immature DCs, and a >200-fold increase when compared with DCs activated with a classical cytokine cocktail). Simultaneous introduction of TAA and TriMix mRNA was shown to be feasible [96]. These data formed the basis for clinical evaluation of these so-called TriMix-DCs as a cellular vaccine for immunization of advanced stage melano‐ ma patients [97]. Extensive immunomonitoring of delayed type hypersensitivity reactions [98] and/or blood samples [99] obtained from these vaccinated patients, demonstrated priming of CD4+ and CD8+ T cell responses against multiple epitopes of the TAAs present within the DCvaccine. More importantly, in a small-scale (15 patients) phase IB clinical trial in advanced melanoma, it was shown that combined intradermal/intravenous vaccination with TriMix-DCs induces anti-tumor activity with durable disease control in a number of patients [100].
**Figure 2.** TriMix-DCs hyper-activate CTLs that are partially protected from Treg suppression and drive Treg towards a TH1 phenotype. Immature DCs can be artificially maturated using genetic modification. To obtain strong T-cell responses, DCs can be electroporated with a cocktail of mRNA molecules encoding constitutive active toll like receptor 4, CD40 ligand and CD70, not only maturing the DCs but also providing them with a powerful co-stimulatory molecule. These DCs are able to stimulate both naive CD4+and CD8+T cells. CD8+T cells will become activated to CTLs, while CD4+TH1 cells will facilitate this process. Moreover, it has recently been shown that these DCs are also able to thwart the func‐ tion of Treg, whereby their suppressive effect on DCs, CD4+and CD8+T cells is largely incapacitated. As a consequence, the stimulatory potential of these DCs is enhanced because they do not only provide co-stimulation, but also counter
Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation
http://dx.doi.org/10.5772/57539
83
As mentioned co-stimulatory molecules can also affect Treg. For instance, it was described that triggering OX40 (human and mouse), GITR (mouse) or 4-1BB leads to a reduced Treg sup‐
Given its important role in T-cell stimulation as described above, it is worthwhile discussing the role of the co-stimulatory CD70/CD27 pathway in the context of Treg suppression. Although it is well described that Treg express high levels of CD27, the effect of CD27/CD70 signaling on the function of Treg has not been studied in detail in the context of cancer [107-109]. Importantly, it is worthwhile to note that different effects of CD70/CD27 signaling on Treg have been described and that these effects seem to be dependent on the context of the CD70 expression. For instance, Claus and colleagues demonstrated that CD70/CD27 interaction in the chroni‐ cally inflamed tumor environment increased the frequency of Treg, resulting in reduced tumor specific T cell responses and the promotion of tumor growth. The latter was shown to be due
IL-2 a well-known survival factor for Treg [107, 110]. It has been suggested by Moser and colleagues that in a pro-inflammatory environment, such as secondary lymphoid organs, priming of TH1 cells (a necessary component for efficient priming of CTLs) is driven by IL-12 or CD70/CD27 signaling, depending on the function of Treg. They observed that enhanced TH1
Teff cells to produce
to an enhanced survival of Treg, which was linked to the stimulation of CD4+
immune inhibition.
pression [45, 101-106].
Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation http://dx.doi.org/10.5772/57539 83
common theme for the members of this family is their coordinated expression at the interface between T cells and DCs, and their ability to shape the environment that sets the stage for the cross talk between T cells and DCs. In the remainder of this section we will focus on the CD70/ CD27 pathway as this pathway has been identified as a critical regulator of immunity *versus*
Similar to the B7 receptor CD28, the receptor CD27 is expressed on naive T cells [78]. Although initially up-regulated upon T cell activation, its expression decreases to undetectable levels after T cells undergo several rounds of division [79]. However, CD27 expression has been reported on central memory T cells [80]. The expression of its ligand CD70 is highly restricted and activation dependent, as it is only transiently expressed on activated T and B cells as well as DCs [81-83]. Studies using CD27 deficient mice have provided insights into the role of CD27 in the generation of T cell immunity. It was indicated that CD27 is of importance in the stimulation of both primary and memory T-cell responses [84]. Although, CD27 seemingly was not required to induce T-cell proliferation, it was shown that CD27 enhances T-cell survival upon successive rounds of cell division [85]. Nonetheless, a role for CD27 in CD8+
cell proliferation was shown in the absence of IL-2, although no functional differentiation was detected [86]. In this regard, CD27 has been linked to the induction of IL-2 and IFN-γ secretion
findings were confirmed in mouse studies using blocking anti-CD70 antibodies [88, 89]. An observation that pinpointed the CD27/CD70 pathway as one of the most important costimulatory pathways known today, is the observation that due to their CD70 expression,
Moreover, it was demonstrated in mice that stimulation of T cells by DCs that constitutively express CD70 is effective in the absence of signal 3, suggesting that TCR triggering and CD27/ CD70 interaction-at least in mice-is sufficient to promote CTL responses [91]. Many of these studies were performed in the context of infectious diseases. Also in cancer, the CD27/CD70 interaction has been described to help immune cell survival and induction of potent CTL responses [92-95]. Importantly, our group has developed a potent DC-based anti-cancer vaccine based on the use of mRNA to introduce CD70 in addition to two DC activating stimuli CD40 ligand and a constitutive active form of toll-like receptor 4, collectively referred to as TriMix (Figure 2) [43]. It has been demonstrated that these DCs are able to strongly activate
T cells (>500-fold increase in Melan A-specific CD8+
T cell responses against multiple epitopes of the TAAs present within the DC-
when compared with immature DCs, and a >200-fold increase when compared with DCs activated with a classical cytokine cocktail). Simultaneous introduction of TAA and TriMix mRNA was shown to be feasible [96]. These data formed the basis for clinical evaluation of these so-called TriMix-DCs as a cellular vaccine for immunization of advanced stage melano‐ ma patients [97]. Extensive immunomonitoring of delayed type hypersensitivity reactions [98] and/or blood samples [99] obtained from these vaccinated patients, demonstrated priming of
vaccine. More importantly, in a small-scale (15 patients) phase IB clinical trial in advanced melanoma, it was shown that combined intradermal/intravenous vaccination with TriMix-DCs induces anti-tumor activity with durable disease control in a number of patients [100].
T cells, as such instructing the latter to support CD8+
spleen DCs are able to prime CD4+
CD45RA+
T
T-cell activation [87]. These
T cells was observed
T cells in the absence of a third signal, *e.g.* IL-12 [90].
T cells and when loaded with TAA derived peptides are able
tolerance.
82 Immune Response Activation
by CD4+
DEC205+
allogeneic naive CD4+
to hyper-activate CD8+
and CD8+
CD4+
**Figure 2.** TriMix-DCs hyper-activate CTLs that are partially protected from Treg suppression and drive Treg towards a TH1 phenotype. Immature DCs can be artificially maturated using genetic modification. To obtain strong T-cell responses, DCs can be electroporated with a cocktail of mRNA molecules encoding constitutive active toll like receptor 4, CD40 ligand and CD70, not only maturing the DCs but also providing them with a powerful co-stimulatory molecule. These DCs are able to stimulate both naive CD4+and CD8+T cells. CD8+T cells will become activated to CTLs, while CD4+TH1 cells will facilitate this process. Moreover, it has recently been shown that these DCs are also able to thwart the func‐ tion of Treg, whereby their suppressive effect on DCs, CD4+and CD8+T cells is largely incapacitated. As a consequence, the stimulatory potential of these DCs is enhanced because they do not only provide co-stimulation, but also counter immune inhibition.
As mentioned co-stimulatory molecules can also affect Treg. For instance, it was described that triggering OX40 (human and mouse), GITR (mouse) or 4-1BB leads to a reduced Treg sup‐ pression [45, 101-106].
Given its important role in T-cell stimulation as described above, it is worthwhile discussing the role of the co-stimulatory CD70/CD27 pathway in the context of Treg suppression. Although it is well described that Treg express high levels of CD27, the effect of CD27/CD70 signaling on the function of Treg has not been studied in detail in the context of cancer [107-109]. Importantly, it is worthwhile to note that different effects of CD70/CD27 signaling on Treg have been described and that these effects seem to be dependent on the context of the CD70 expression. For instance, Claus and colleagues demonstrated that CD70/CD27 interaction in the chroni‐ cally inflamed tumor environment increased the frequency of Treg, resulting in reduced tumor specific T cell responses and the promotion of tumor growth. The latter was shown to be due to an enhanced survival of Treg, which was linked to the stimulation of CD4+ Teff cells to produce IL-2 a well-known survival factor for Treg [107, 110]. It has been suggested by Moser and colleagues that in a pro-inflammatory environment, such as secondary lymphoid organs, priming of TH1 cells (a necessary component for efficient priming of CTLs) is driven by IL-12 or CD70/CD27 signaling, depending on the function of Treg. They observed that enhanced TH1 priming in the absence of Treg is CD70 dependent, whereas IL-12 is required for TH1 develop‐ ment when Treg are at place. The fact that CD70 can exert its co-stimulatory function for TH1 cells in the absence of Treg, might be explained by their observation that the expression of CD70 on DCs was abrogated by Treg *in vitro* [16]. More recently, we extensively investigated the influence of CD70 expressing TriMix-DCs on Treg (Figure 2). We demonstrated that these DCs do not induce Treg starting from CD4+ CD25- T cells. More importantly, the stimulation of CD8+ T cells by TriMix-DCs was only marginally influenced by the presence of Treg. In addition, we showed that CD8+ T cells that were cultured with TriMix-DCs were partially protected from subsequent Treg suppression. Besides desensitization of CD8+ T cells to Treg, we further showed that Treg co-cultured in the presence of TriMix-DCs partially lost their suppressive capacity, a phenomenon that was accompanied by a decrease in CD27 and CD25 expression on these Treg, as well as an increase in the expression of the transcription factor T-bet and secretion of cytokines linked to a TH1 phenotype. Although the assays were performed *in vitro* and their outcome cannot solely be attributed to CD70, these data further underline the notion that the effect of CD70/CD27 signaling on Treg might depend on its context. Importantly, these data suggest that the potency of CD70 expressing TriMix-DCs is linked to their ability to hyperactivate CTLs that are refractory to Treg as well as on their ability to suppress the function of Treg, even reprogram them to TH1 cells under certain circumstances [108].
through the addition of cytokines [116]. Importantly, its expression is not limited to activated T cells, as it is also expressed on activated B cells and myeloid cells [117, 118]. Although PD-1 is not present on the surface of freshly isolated Treg, it can be found at high levels intracellularly.
Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation
http://dx.doi.org/10.5772/57539
85
The ligands for PD-1 are B7-H1 (PD-L1) and B7-DC (PD-L2). These show a distinct expression pattern. Whereas expression of B7-H1 can be induced on several different cell types, including monocytes, DCs, mast cells, T and B cells, epithelial, endothelial and muscle cells, the expres‐ sion of B7-DC is more restricted to DCs, macrophages and mast cells [120-122]. Importantly, B7-H1 expression was also found on several human and mouse tumors [123, 124]. Of note, to emphasize that both B7-H1 and B7-DC bind PD-1, they were renamed as PD-L1 and PD-L2, respectively. However, PD-L1 is polyamorous, it can bind to PD-1 as well as B7.1. Therefore, this nomenclature does not reflect the complexity of the signaling pathways that involve PD-L1 [125, 126]. For instance, it was shown that B7.1 (CD80) expressed on T cells transduces an inhibitory signal to T cells following its ligation with B7-H1 *in vitro* [125, 127]. Conversely, B7- H1 expressed on T cells transduced an inhibitory signal to T cells after its interaction with B7.1 (CD80) [125]. So far, it has not been demonstrated that similar T cell-T cell interactions occur *in vivo*. Moreover, it remains to be clarified whether interaction of B7.1 (CD80) expressed on myeloid cells, *e.g.* DCs and myeloid derived suppressor cells (MDSCs), with B7-H1 on T cells
The role of the receptor PD-1 is predominantly described as being inhibitory for immune responses (Figure 3), a notion that is supported by the phenotype of PD-1 deficient mice [128]. There is little doubt that PD-L1 is the main partner in this inhibitory function as experiments utilizing strategies to either interfere with PD-1 or PD-L1 (*e.g.* antibodies, genetic downregulation, *etc.*) often have a similar outcome [70]. Of note, whereas PD-1 deficient mice develop severe autoimmunity, this is not the case in PD-L1 deficient mice [129]. In addition, it needs to be mentioned that contradictory results have been obtained *in vitro* showing both negative and positive signaling of B7-H1 on T cell proliferation and cytokine production. These
With regard to suppression of immune responses, it has been described that PD-1 plays a critical role in maintaining peripheral tolerance. In this regard, it has been shown that the co-
Treg [130]. Moreover, PD-1 is up-regulated on Treg upon TCR stimulation and has been suggested to enhance their proliferation [131]. It is furthermore thought that the interaction of PD-L1 with PD-1 is responsible for the lack of T-cell responsiveness in settings of persistent antigen stimulation such as those encountered in cancer and chronic infectious diseases [132-135]. Several modes of action that result in the corruption of the Teff cell functionality have been described, including inhibition of kinases involved in T cell activation [118], induction of cell death [115] and modulation of the duration of T cell-DC or T cell-target cell contact by down-regulation of the TCR [136-138]. The physiological importance of TCR down-regulation is yet unclear [139], however most of the experimental evidence points out that TCR downregulation seems to act at different levels to prevent autoimmunity during the onset of immune
T cells to induced
observations caution us to conclude that PD-L1 is exclusively suppressive.
inhibitory PD-1 receptor plays a role in the conversion of naive CD4+
Here it is stored for export to the cell surface upon TCR stimulation [119].
would result in a similar inhibitory signal.
Foxp3+
responses [140].
#### **3.4. Co-inhibition: focus on the PD-L1/PD-1 pathway**
Over the years, two negative stimulatory (co-inhibitory) receptors on T cells have been extensively studied in the context of cancer. These are CTLA-4 and PD-1, which are both members of the B7 receptor family. Moreover, they are both considered to be critical immune checkpoints that can be exploited to enhance anti-tumor T-cell responses.
The role of CTLA-4 in dampening T cell mediated immune responses has been extensively reviewed in [111, 112]. Briefly, CTLA-4 is up-regulated on activated T cells. At that time CTLA-4 interacts with the same B7 molecules as the CD28 receptor but outcompetes the CD28 receptor due to its higher affinity for these B7 molecules. The net effect of this B7.1 or B7.2/ CTLA-4 signaling is a decrease in T-cell activity hence dampening of the immune response. Several mechanisms have been described by which CTLA-4 inhibits effector T-cell functions, including induction of reduced contact period between antigen presenting cells and T cells [113] and by transendocytosis of CD80/CD86, thus compromising the T-cell stimulatory capacity of antigen presenting cells [15]. Furthermore, CTLA-4 is constitutively expressed on Treg and has been shown to be essential for their suppressive function [14]. Today, it is evidenced that unleashing this immunological break through delivery of blocking anti-CTLA-4 antibodies (*e.g.* ipilimumab) is therapeutically beneficial in the fight against cancer [114].
One of the more recently identified B7 receptor family members PD-1 is a distant homologue of CTLA-4, which was originally identified as a gene that was highly expressed by cells undergoing programmed cell death hence its name programmed death 1 receptor [115]. Similar to CTLA-4, PD-1 is up-regulated upon stimulation of T cells, although it has been shown on human T cells that PD-1 can be up-regulated in the absence of TCR triggering through the addition of cytokines [116]. Importantly, its expression is not limited to activated T cells, as it is also expressed on activated B cells and myeloid cells [117, 118]. Although PD-1 is not present on the surface of freshly isolated Treg, it can be found at high levels intracellularly. Here it is stored for export to the cell surface upon TCR stimulation [119].
priming in the absence of Treg is CD70 dependent, whereas IL-12 is required for TH1 develop‐ ment when Treg are at place. The fact that CD70 can exert its co-stimulatory function for TH1 cells in the absence of Treg, might be explained by their observation that the expression of CD70 on DCs was abrogated by Treg *in vitro* [16]. More recently, we extensively investigated the influence of CD70 expressing TriMix-DCs on Treg (Figure 2). We demonstrated that these DCs
cells by TriMix-DCs was only marginally influenced by the presence of Treg. In addition, we
that Treg co-cultured in the presence of TriMix-DCs partially lost their suppressive capacity, a phenomenon that was accompanied by a decrease in CD27 and CD25 expression on these Treg, as well as an increase in the expression of the transcription factor T-bet and secretion of cytokines linked to a TH1 phenotype. Although the assays were performed *in vitro* and their outcome cannot solely be attributed to CD70, these data further underline the notion that the effect of CD70/CD27 signaling on Treg might depend on its context. Importantly, these data suggest that the potency of CD70 expressing TriMix-DCs is linked to their ability to hyperactivate CTLs that are refractory to Treg as well as on their ability to suppress the function of
Over the years, two negative stimulatory (co-inhibitory) receptors on T cells have been extensively studied in the context of cancer. These are CTLA-4 and PD-1, which are both members of the B7 receptor family. Moreover, they are both considered to be critical immune
The role of CTLA-4 in dampening T cell mediated immune responses has been extensively reviewed in [111, 112]. Briefly, CTLA-4 is up-regulated on activated T cells. At that time CTLA-4 interacts with the same B7 molecules as the CD28 receptor but outcompetes the CD28 receptor due to its higher affinity for these B7 molecules. The net effect of this B7.1 or B7.2/ CTLA-4 signaling is a decrease in T-cell activity hence dampening of the immune response. Several mechanisms have been described by which CTLA-4 inhibits effector T-cell functions, including induction of reduced contact period between antigen presenting cells and T cells [113] and by transendocytosis of CD80/CD86, thus compromising the T-cell stimulatory capacity of antigen presenting cells [15]. Furthermore, CTLA-4 is constitutively expressed on Treg and has been shown to be essential for their suppressive function [14]. Today, it is evidenced that unleashing this immunological break through delivery of blocking anti-CTLA-4 antibodies (*e.g.* ipilimumab) is therapeutically beneficial in the fight against cancer
One of the more recently identified B7 receptor family members PD-1 is a distant homologue of CTLA-4, which was originally identified as a gene that was highly expressed by cells undergoing programmed cell death hence its name programmed death 1 receptor [115]. Similar to CTLA-4, PD-1 is up-regulated upon stimulation of T cells, although it has been shown on human T cells that PD-1 can be up-regulated in the absence of TCR triggering
T cells that were cultured with TriMix-DCs were partially protected from
T cells. More importantly, the stimulation of CD8+
T cells to Treg, we further showed
T
CD25-
Treg, even reprogram them to TH1 cells under certain circumstances [108].
checkpoints that can be exploited to enhance anti-tumor T-cell responses.
subsequent Treg suppression. Besides desensitization of CD8+
**3.4. Co-inhibition: focus on the PD-L1/PD-1 pathway**
do not induce Treg starting from CD4+
showed that CD8+
84 Immune Response Activation
[114].
The ligands for PD-1 are B7-H1 (PD-L1) and B7-DC (PD-L2). These show a distinct expression pattern. Whereas expression of B7-H1 can be induced on several different cell types, including monocytes, DCs, mast cells, T and B cells, epithelial, endothelial and muscle cells, the expres‐ sion of B7-DC is more restricted to DCs, macrophages and mast cells [120-122]. Importantly, B7-H1 expression was also found on several human and mouse tumors [123, 124]. Of note, to emphasize that both B7-H1 and B7-DC bind PD-1, they were renamed as PD-L1 and PD-L2, respectively. However, PD-L1 is polyamorous, it can bind to PD-1 as well as B7.1. Therefore, this nomenclature does not reflect the complexity of the signaling pathways that involve PD-L1 [125, 126]. For instance, it was shown that B7.1 (CD80) expressed on T cells transduces an inhibitory signal to T cells following its ligation with B7-H1 *in vitro* [125, 127]. Conversely, B7- H1 expressed on T cells transduced an inhibitory signal to T cells after its interaction with B7.1 (CD80) [125]. So far, it has not been demonstrated that similar T cell-T cell interactions occur *in vivo*. Moreover, it remains to be clarified whether interaction of B7.1 (CD80) expressed on myeloid cells, *e.g.* DCs and myeloid derived suppressor cells (MDSCs), with B7-H1 on T cells would result in a similar inhibitory signal.
The role of the receptor PD-1 is predominantly described as being inhibitory for immune responses (Figure 3), a notion that is supported by the phenotype of PD-1 deficient mice [128]. There is little doubt that PD-L1 is the main partner in this inhibitory function as experiments utilizing strategies to either interfere with PD-1 or PD-L1 (*e.g.* antibodies, genetic downregulation, *etc.*) often have a similar outcome [70]. Of note, whereas PD-1 deficient mice develop severe autoimmunity, this is not the case in PD-L1 deficient mice [129]. In addition, it needs to be mentioned that contradictory results have been obtained *in vitro* showing both negative and positive signaling of B7-H1 on T cell proliferation and cytokine production. These observations caution us to conclude that PD-L1 is exclusively suppressive.
With regard to suppression of immune responses, it has been described that PD-1 plays a critical role in maintaining peripheral tolerance. In this regard, it has been shown that the coinhibitory PD-1 receptor plays a role in the conversion of naive CD4+ T cells to induced Foxp3+ Treg [130]. Moreover, PD-1 is up-regulated on Treg upon TCR stimulation and has been suggested to enhance their proliferation [131]. It is furthermore thought that the interaction of PD-L1 with PD-1 is responsible for the lack of T-cell responsiveness in settings of persistent antigen stimulation such as those encountered in cancer and chronic infectious diseases [132-135]. Several modes of action that result in the corruption of the Teff cell functionality have been described, including inhibition of kinases involved in T cell activation [118], induction of cell death [115] and modulation of the duration of T cell-DC or T cell-target cell contact by down-regulation of the TCR [136-138]. The physiological importance of TCR down-regulation is yet unclear [139], however most of the experimental evidence points out that TCR downregulation seems to act at different levels to prevent autoimmunity during the onset of immune responses [140].
higher number of multifunctional T cells that showed increased cytokine secretion, while it
Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation
http://dx.doi.org/10.5772/57539
87
Besides antigen presenting cells, tumor cells use the PD-1/PD-L1 pathway to control anti-tumor immunity. Many human cancers have been shown to aberrantly express B7-H1. Up-regulation of B7-H1 is associated to local inflammation (predominantly the presence of IFN-γ) [148]. Importantly, expression of B7-H1 is a poor prognostic factor in several cancer types, including ovarian cancer [149], renal cancer [150, 151], pancreatic cancer [152] and breast cancer [153]. The latter can be explained by three different mechanisms. First, it has been described that binding of PD-1 (T cells) to B7-H1 on cancer cells provides anti-apoptotic stimuli to these cancer cells [65]. Secondly, it was shown in experiments in which cancer cells were modified to express B7-H1, that these cells become refractory to CTL mediated killing [154, 155]. In another model it was shown that this phenomenon was reversible, when the binding of B7-H1 was prevented by blocking antibodies [156]. Thirdly, tumor associated antigen presenting cells utilize the PD-1/PD-L1 (B7-H1) pathway to control anti-tumor T cell responses using the mechanisms that were described above. This results in the stimulation of Teff cells that show high expression of PD-1, are impaired in their function and resemble the exhausted antigen specific T cells that can be found during chronic viral infection. Interaction between PD-L1 expressed on tumor cells and PD-1 expressed on effector T cells was also shown to directly induce apoptosis of T
The knowledge that blockade of PD-1/PD-L1 signaling aids the priming of multifunctional Teff cells, avoids expansion of Treg and facilitates anti-tumor immune responses at the tumor site, has instigated research in which stimulation of T cell responses in tumor bearing mice was combined with the administration of anti-PD-1 or anti-PD-L1 blocking antibodies, demonstrating a synergistic effect of both strategies [152, 154, 156, 158, 159]. Of note, treatment of mice with blocking antibodies without other treatment modalities only had marginal effects. So far, four humanized antibodies have been generated and tested in clinical trials to block the PD-1/PD-L1 pathway. Three of them are blocking anti-PD-1 antibodies, CT-011 [160, 161], MDX-1106 [162] and BMS-936558 [163, 164], whereas the fourth antibody is a blocking anti-PD-L1 antibody called BMS-936559 [165]. The antibody CT-011 was used in a phase I clinical trial in 17 patients with advanced hematological malignancies, demonstrating clinical benefit in about 33% of patients of which one complete response. There was no development of autoimmunity reported [160]. The antibody MDX-1106 was tested in 39 patients with advanced solid cancer, including patients with colorectal cancer, melanoma, prostate cancer, non small cell lung carcinoma and renal cell carcinoma. Manageable toxicity, such as anemia, lympho‐ penia, colitis and arthritis was reported. Clinical benefit was demonstrate in about 10% of patients of which one complete response, two partial responses and two mixed responses [162]. The same team conducted a large-scale study in a similar patient population (over 260 patients) using the anti-PD-1 antibody, BMS-936558. Objective responses were described in approximately one in four to one in five patients with non-small-cell lung cancer, melanoma or renal cell cancer. Again, the adverse event profile did not appear to preclude the use of the tested antibody. Moreover, a relationship between PD-L1 expression on tumor cells and objective response was suggested in this trial [163, 164]. Also the anti-PD-L1 antibody,
did not induce Treg.
cells [157].
**Figure 3.** The PD-1/PD-L1 interaction has an inhibitory effect on T cells, while it simultaneously provides survival sig‐ nals to cancer cells. DCs can be manipulated to stimulate both naive CD4+and CD8+T cells. Immune stimulation of these T cells leads to up-regulation of the receptor PD-1, interacting with PD-L1 on DCs, to terminate the immune re‐ sponse before causing harm to the host. In the tumor environment, however, PD-1 up-regulation will incapacitate T cells before they are able to tackle tumor cells. Moreover, the tumor itself, can up-regulate PD-L1 to avoid T-cell medi‐ ated killing. Treg can be induced from naive CD4+T cells. The role of PD-1 in Treg is less documented. Nonetheless it is supposed that Treg are able to express PD-1 and can expand upon PD-1 triggering.
Most studies that have demonstrated negative co-signaling through interaction of PD-1 with PD-L1 during antigen presentation by DCs, have predominantly focused on expansion of T cells and secretion of single cytokines [141]. However, there is a growing appreciation that poly-functional T cells might reflect the effectiveness of the T cells to mediate tumor cell eradication, as was described in the eradication of virally infected cells [142]. In this regard, several studies have found enhanced numbers of tumor-specific poly-functional T cells in patients responding favorably to various forms of immunotherapy, including anti-CTLA-4 antibody therapy [143] and tumor vaccines [144]. In addition, the studies performed by Imai and colleagues [145] suggest that poly-functionality of adoptively transferred CD8+ T cells is a sensitive immune correlate for successful immunotherapy against malignancy. Therefore, we recently evaluated whether interference of the PD-1/PD-L1 pathway during antigen presen‐ tation by DCs impacted on the ability of CD4+ and CD8+ T cells to perform multiple functions at the same time [146]. To that end, we genetically engineered monocyte-derived DCs to secrete soluble PD-1 or soluble PD-L1. The latter approach was based on the observation that patients with chronic lymphocytic leukemia that have a splice variants of PD-1, which gives rise to a soluble form of PD-1, have a prolonged survival [147]. Interestingly, we observed that modification of DCs with soluble PD-1 or soluble PD-L1 mRNA resulted in increased levels of the co-stimulatory molecule CD80 and a distinct cytokine profile, characterized by the secretion of IL-10 and TNF-α, respectively. Importantly, these DCs stimulated a significantly higher number of multifunctional T cells that showed increased cytokine secretion, while it did not induce Treg.
Besides antigen presenting cells, tumor cells use the PD-1/PD-L1 pathway to control anti-tumor immunity. Many human cancers have been shown to aberrantly express B7-H1. Up-regulation of B7-H1 is associated to local inflammation (predominantly the presence of IFN-γ) [148]. Importantly, expression of B7-H1 is a poor prognostic factor in several cancer types, including ovarian cancer [149], renal cancer [150, 151], pancreatic cancer [152] and breast cancer [153]. The latter can be explained by three different mechanisms. First, it has been described that binding of PD-1 (T cells) to B7-H1 on cancer cells provides anti-apoptotic stimuli to these cancer cells [65]. Secondly, it was shown in experiments in which cancer cells were modified to express B7-H1, that these cells become refractory to CTL mediated killing [154, 155]. In another model it was shown that this phenomenon was reversible, when the binding of B7-H1 was prevented by blocking antibodies [156]. Thirdly, tumor associated antigen presenting cells utilize the PD-1/PD-L1 (B7-H1) pathway to control anti-tumor T cell responses using the mechanisms that were described above. This results in the stimulation of Teff cells that show high expression of PD-1, are impaired in their function and resemble the exhausted antigen specific T cells that can be found during chronic viral infection. Interaction between PD-L1 expressed on tumor cells and PD-1 expressed on effector T cells was also shown to directly induce apoptosis of T cells [157].
The knowledge that blockade of PD-1/PD-L1 signaling aids the priming of multifunctional Teff cells, avoids expansion of Treg and facilitates anti-tumor immune responses at the tumor site, has instigated research in which stimulation of T cell responses in tumor bearing mice was combined with the administration of anti-PD-1 or anti-PD-L1 blocking antibodies, demonstrating a synergistic effect of both strategies [152, 154, 156, 158, 159]. Of note, treatment of mice with blocking antibodies without other treatment modalities only had marginal effects. So far, four humanized antibodies have been generated and tested in clinical trials to block the PD-1/PD-L1 pathway. Three of them are blocking anti-PD-1 antibodies, CT-011 [160, 161], MDX-1106 [162] and BMS-936558 [163, 164], whereas the fourth antibody is a blocking anti-PD-L1 antibody called BMS-936559 [165]. The antibody CT-011 was used in a phase I clinical trial in 17 patients with advanced hematological malignancies, demonstrating clinical benefit in about 33% of patients of which one complete response. There was no development of autoimmunity reported [160]. The antibody MDX-1106 was tested in 39 patients with advanced solid cancer, including patients with colorectal cancer, melanoma, prostate cancer, non small cell lung carcinoma and renal cell carcinoma. Manageable toxicity, such as anemia, lympho‐ penia, colitis and arthritis was reported. Clinical benefit was demonstrate in about 10% of patients of which one complete response, two partial responses and two mixed responses [162]. The same team conducted a large-scale study in a similar patient population (over 260 patients) using the anti-PD-1 antibody, BMS-936558. Objective responses were described in approximately one in four to one in five patients with non-small-cell lung cancer, melanoma or renal cell cancer. Again, the adverse event profile did not appear to preclude the use of the tested antibody. Moreover, a relationship between PD-L1 expression on tumor cells and objective response was suggested in this trial [163, 164]. Also the anti-PD-L1 antibody,
Most studies that have demonstrated negative co-signaling through interaction of PD-1 with PD-L1 during antigen presentation by DCs, have predominantly focused on expansion of T cells and secretion of single cytokines [141]. However, there is a growing appreciation that poly-functional T cells might reflect the effectiveness of the T cells to mediate tumor cell eradication, as was described in the eradication of virally infected cells [142]. In this regard, several studies have found enhanced numbers of tumor-specific poly-functional T cells in patients responding favorably to various forms of immunotherapy, including anti-CTLA-4 antibody therapy [143] and tumor vaccines [144]. In addition, the studies performed by Imai
supposed that Treg are able to express PD-1 and can expand upon PD-1 triggering.
**Figure 3.** The PD-1/PD-L1 interaction has an inhibitory effect on T cells, while it simultaneously provides survival sig‐ nals to cancer cells. DCs can be manipulated to stimulate both naive CD4+and CD8+T cells. Immune stimulation of these T cells leads to up-regulation of the receptor PD-1, interacting with PD-L1 on DCs, to terminate the immune re‐ sponse before causing harm to the host. In the tumor environment, however, PD-1 up-regulation will incapacitate T cells before they are able to tackle tumor cells. Moreover, the tumor itself, can up-regulate PD-L1 to avoid T-cell medi‐ ated killing. Treg can be induced from naive CD4+T cells. The role of PD-1 in Treg is less documented. Nonetheless it is
and colleagues [145] suggest that poly-functionality of adoptively transferred CD8+
tation by DCs impacted on the ability of CD4+
86 Immune Response Activation
sensitive immune correlate for successful immunotherapy against malignancy. Therefore, we recently evaluated whether interference of the PD-1/PD-L1 pathway during antigen presen‐
at the same time [146]. To that end, we genetically engineered monocyte-derived DCs to secrete soluble PD-1 or soluble PD-L1. The latter approach was based on the observation that patients with chronic lymphocytic leukemia that have a splice variants of PD-1, which gives rise to a soluble form of PD-1, have a prolonged survival [147]. Interestingly, we observed that modification of DCs with soluble PD-1 or soluble PD-L1 mRNA resulted in increased levels of the co-stimulatory molecule CD80 and a distinct cytokine profile, characterized by the secretion of IL-10 and TNF-α, respectively. Importantly, these DCs stimulated a significantly
and CD8+
T cells is a
T cells to perform multiple functions
BMS-936559, was evaluated by this group in a phase I trial including over 200 patients suffering from the aforementioned cancers. From this study, it was concluded that antibody mediated blockade of PD-L1 induced durable tumor regression (objective response rate of 6 to 17%) and prolonged stabilization of disease (rates of 12 to 41% at 24 weeks) in patients with advanced cancers, including non–small cell lung cancer, melanoma and renal cell cancer [165].
**Author details**
, Joeri. L. Aerts1
\*Address all correspondence to: kbreckpo@vub.ac.be
Service of Navarre, Pamplona, Navarre, Spain
statistics. *CA Cancer J Clin* 61, 69-90.
, Thérèse Liechtenstein2
1 Laboratory of Molecular and Cellular Therapy, Vrije Universiteit Brussel (VUB), Brussels,
3 Biomedical Research Centre Navarrabiomed-Fundacion Miguel Servet, National Health
[1] Bruno, T.C., French, J.D., Jordan, K.R., Ramirez, O., Sippel, T.R., Borges, V.F., Hau‐ gen, B.R., McCarter, M.D., Waziri, A., Slansky, J.E. (2013) Influence of human im‐ mune cells on cancer: studies at the University of Colorado. *Immunol Res* 55, 22-33.
[2] Jemal, A., Bray, F., Center, M.M., Ferlay, J., Ward, E., Forman, D. (2011) Global cancer
[3] Boon, T., Coulie, P.G., Van den Eynde, B.J., van der Bruggen, P. (2006) Human T cell
[4] Lizee, G., Overwijk, W.W., Radvanyi, L., Gao, J., Sharma, P., Hwu, P. (2013) Harness‐ ing the power of the immune system to target cancer. *Annu Rev Med* 64, 71-90.
[5] Gardner, T.A., Elzey, B.D., Hahn, N.M. (2012) Sipuleucel-T (Provenge) autologous vaccine approved for treatment of men with asymptomatic or minimally symptomat‐ ic castrate-resistant metastatic prostate cancer. *Hum Vaccin Immunother* 8, 534-9.
[6] Kantoff, P.W., Higano, C.S., Small, E.J., Whitmore, J.B., Frohlich, M.W., Schellham‐ mer, P.F. (2012) Re: interdisciplinary critique of sipuleucel-T as immunotherapy in castration-resistant prostate cancer. *J Natl Cancer Inst* 104, 1107-9; author reply
[7] Burnet, F.M. (1970) The concept of immunological surveillance. *Prog Exp Tumor Res*
[8] Urosevic, M., Dummer, R. (2008) Human leukocyte antigen-G and cancer immunoe‐
responses against melanoma. *Annu Rev Immunol* 24, 175-208.
2 Division of Infection and Immunity, University College London (UCL), London, UK
, David Escors2,3 and Karine Breckpot1
http://dx.doi.org/10.5772/57539
89
Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation
Joeri J. Pen1
Belgium
**References**
1109-12.
13, 1-27.
diting. *Cancer Res* 68, 627-30.
In conclusion, experimental and early clinical results indicate that manipulation of the B7-H1 (PD-L1)/PD-1 pathway provides novel opportunities in anti-tumor therapy. In particular the combination of PD-L1/PD-1 blockade with vaccination strategies might synergize to initially enhance the stimulation of Teff cells, whilst enabling these Teff cells to exert their function in the tumor environment.
#### **4. Discussion**
Progress in our understanding on how immune cells interact with each other and how they are affected by cancer cells has led to the development of immunotherapeutic strategies that aim to harness the potential of CTLs in order to improve the outcome for cancer patients. Although CTLs are critical to fight cancer cells, there is a general consensus that stimulation of CTLs will only be successful when they are able to overcome the immunosuppression exerted by cancer cells and cancer associated immune cells, such as Treg. To achieve the latter, research has focused on two approaches. One approach is designing DCs that are able to stimulate CTLs that are no longer sensitive to immunosuppression and to use these DCs as a cellular vaccine. Another approach is to develop strategies to modulate immunosuppressive mechanisms and use these either alone or in combination with vaccination. Our growing knowledge on the expression pattern and function of individual co-signaling molecules and how these educate T cells to respond or not, has been invaluable for the development of both approaches.
Co-signaling molecules start a signal cascade in T cells, which in turn allows them to sense environmental conditions and consequently respond accordingly. There is a growing list of co-signaling molecules that can either transduce stimulatory or inhibitory signals and therefore are often categorized as such. However, whether co-signaling will be stimulatory or inhibitory is dependent on the level and timing of expression of these molecules hence in reality (when multiple co-signaling molecules are expressed on the antigen presenting cell, be it DCs or cancer cells) the situation is much more complex. Nonetheless, major discoveries on new signaling molecules, their interactions and multilevel functionality have had a tremendous impact on cancer immunotherapy over the years. In this regard, important milestones have been the description of the CD70/CD27 and PD-L1/PD-1 pathways. Although it's early days in their clinical exploitation, the pre-clinical and clinical results described in this chapter demonstrate that manipulation of these pathways is promising. In particular, clinical results such as those obtained with improved DC vaccines, e.g. TriMix-DCs or blocking anti-PD-1 or PD-L1 antibodies, re-fuel the enthusiasm that immunotherapy strategies will find their place as an effective fourth treatment modality for cancer.
#### **Author details**
BMS-936559, was evaluated by this group in a phase I trial including over 200 patients suffering from the aforementioned cancers. From this study, it was concluded that antibody mediated blockade of PD-L1 induced durable tumor regression (objective response rate of 6 to 17%) and prolonged stabilization of disease (rates of 12 to 41% at 24 weeks) in patients with advanced
In conclusion, experimental and early clinical results indicate that manipulation of the B7-H1 (PD-L1)/PD-1 pathway provides novel opportunities in anti-tumor therapy. In particular the combination of PD-L1/PD-1 blockade with vaccination strategies might synergize to initially enhance the stimulation of Teff cells, whilst enabling these Teff cells to exert their function in the
Progress in our understanding on how immune cells interact with each other and how they are affected by cancer cells has led to the development of immunotherapeutic strategies that aim to harness the potential of CTLs in order to improve the outcome for cancer patients. Although CTLs are critical to fight cancer cells, there is a general consensus that stimulation of CTLs will only be successful when they are able to overcome the immunosuppression exerted by cancer cells and cancer associated immune cells, such as Treg. To achieve the latter, research has focused on two approaches. One approach is designing DCs that are able to stimulate CTLs that are no longer sensitive to immunosuppression and to use these DCs as a cellular vaccine. Another approach is to develop strategies to modulate immunosuppressive mechanisms and use these either alone or in combination with vaccination. Our growing knowledge on the expression pattern and function of individual co-signaling molecules and how these educate T cells to respond or not, has been invaluable for the development of both
Co-signaling molecules start a signal cascade in T cells, which in turn allows them to sense environmental conditions and consequently respond accordingly. There is a growing list of co-signaling molecules that can either transduce stimulatory or inhibitory signals and therefore are often categorized as such. However, whether co-signaling will be stimulatory or inhibitory is dependent on the level and timing of expression of these molecules hence in reality (when multiple co-signaling molecules are expressed on the antigen presenting cell, be it DCs or cancer cells) the situation is much more complex. Nonetheless, major discoveries on new signaling molecules, their interactions and multilevel functionality have had a tremendous impact on cancer immunotherapy over the years. In this regard, important milestones have been the description of the CD70/CD27 and PD-L1/PD-1 pathways. Although it's early days in their clinical exploitation, the pre-clinical and clinical results described in this chapter demonstrate that manipulation of these pathways is promising. In particular, clinical results such as those obtained with improved DC vaccines, e.g. TriMix-DCs or blocking anti-PD-1 or PD-L1 antibodies, re-fuel the enthusiasm that immunotherapy strategies will find their place
as an effective fourth treatment modality for cancer.
cancers, including non–small cell lung cancer, melanoma and renal cell cancer [165].
tumor environment.
88 Immune Response Activation
**4. Discussion**
approaches.
Joeri J. Pen1 , Joeri. L. Aerts1 , Thérèse Liechtenstein2 , David Escors2,3 and Karine Breckpot1
\*Address all correspondence to: kbreckpo@vub.ac.be
1 Laboratory of Molecular and Cellular Therapy, Vrije Universiteit Brussel (VUB), Brussels, Belgium
2 Division of Infection and Immunity, University College London (UCL), London, UK
3 Biomedical Research Centre Navarrabiomed-Fundacion Miguel Servet, National Health Service of Navarre, Pamplona, Navarre, Spain
#### **References**
[9] Jacobs, J.F., Nierkens, S., Figdor, C.G., de Vries, I.J., Adema, G.J. (2012) Regulatory T cells in melanoma: the final hurdle towards effective immunotherapy? *Lancet Oncol* 13, e32-42.
[22] McClymont, S.A., Putnam, A.L., Lee, M.R., Esensten, J.H., Liu, W., Hulme, M.A., Hoffmuller, U., Baron, U., Olek, S., Bluestone, J.A., Brusko, T.M. (2011) Plasticity of human regulatory T cells in healthy subjects and patients with type 1 diabetes. *J Im‐*
Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation
http://dx.doi.org/10.5772/57539
91
[23] Oldenhove, G., Bouladoux, N., Wohlfert, E.A., Hall, J.A., Chou, D., Dos Santos, L., O'Brien, S., Blank, R., Lamb, E., Natarajan, S., Kastenmayer, R., Hunter, C., Grigg, M.E., Belkaid, Y. (2009) Decrease of Foxp3+Treg cell number and acquisition of effec‐
[24] Wei, G., Wei, L., Zhu, J., Zang, C., Hu-Li, J., Yao, Z., Cui, K., Kanno, Y., Roh, T.Y., Watford, W.T., Schones, D.E., Peng, W., Sun, H.W., Paul, W.E., O'Shea, J.J., Zhao, K. (2009) Global mapping of H3K4me3 and H3K27me3 reveals specificity and plasticity in lineage fate determination of differentiating CD4+T cells. *Immunity* 30, 155-67. [25] Gold, P., Freedman, S.O. (1965) Demonstration of Tumor-Specific Antigens in Hu‐ man Colonic Carcinomata by Immunological Tolerance and Absorption Techniques.
[26] Abelev, G.I., Tsvetkov, V.S., Biriulina, T.I., El'gort, D.A., Olovnikov, A.M. (1971) [Evaluation of the use of highly sensitive methods of determining alpha-fetoprotein for the diagnosis of hepatocellular cancer and teratoblastoma]. *Biull Eksp Biol Med* 71,
[27] Knuth, A., Wolfel, T., Klehmann, E., Boon, T., Meyer zum Buschenfelde, K.H. (1989) Cytolytic T-cell clones against an autologous human melanoma: specificity study and definition of three antigens by immunoselection. *Proc Natl Acad Sci U S A* 86, 2804-8.
[28] Boon, T., van der Bruggen, P. (1996) Human tumor antigens recognized by T lym‐
[29] Van den Eynde, B.J., Boon, T. (1997) Tumor antigens recognized by T lymphocytes.
[30] Foged, C., Hansen, J., Agger, E.M. (2012) License to kill: Formulation requirements for optimal priming of CD8(+) CTL responses with particulate vaccine delivery sys‐
[31] Bonehill, A., Heirman, C., Thielemans, K. (2005) Genetic approaches for the induc‐ tion of a CD4+T cell response in cancer immunotherapy. *J Gene Med* 7, 686-95.
[32] Breckpot, K., Escors, D. (2009) Dendritic cells for active anti-cancer immunotherapy: targeting activation pathways through genetic modification. *Endocr Metab Immune*
[33] Novellino, L., Castelli, C., Parmiani, G. (2005) A listing of human tumor antigens rec‐ ognized by T cells: March 2004 update. *Cancer Immunol Immunother* 54, 187-207. [34] Cheever, M.A., Allison, J.P., Ferris, A.S., Finn, O.J., Hastings, B.M., Hecht, T.T., Mell‐ man, I., Prindiville, S.A., Viner, J.L., Weiner, L.M., Matrisian, L.M. (2009) The prioriti‐
tor cell phenotype during lethal infection. *Immunity* 31, 772-86.
*munol* 186, 3918-26.
*J Exp Med* 121, 439-62.
phocytes. *J Exp Med* 183, 725-9.
tems. *Eur J Pharm Sci* 45, 482-91.
*Disord Drug Targets* 9, 328-43.
*Int J Clin Lab Res* 27, 81-6.
75-81.
[22] McClymont, S.A., Putnam, A.L., Lee, M.R., Esensten, J.H., Liu, W., Hulme, M.A., Hoffmuller, U., Baron, U., Olek, S., Bluestone, J.A., Brusko, T.M. (2011) Plasticity of human regulatory T cells in healthy subjects and patients with type 1 diabetes. *J Im‐ munol* 186, 3918-26.
[9] Jacobs, J.F., Nierkens, S., Figdor, C.G., de Vries, I.J., Adema, G.J. (2012) Regulatory T cells in melanoma: the final hurdle towards effective immunotherapy? *Lancet Oncol*
[10] Nishikawa, H., Sakaguchi, S. (2010) Regulatory T cells in tumor immunity. *Int J Can‐*
[11] Nizar, S., Meyer, B., Galustian, C., Kumar, D., Dalgleish, A. (2010) T regulatory cells, the evolution of targeted immunotherapy. *Biochim Biophys Acta* 1806, 7-17.
[12] Munn, D.H., Mellor, A.L. (2007) Indoleamine 2,3-dioxygenase and tumor-induced
[13] Emeagi, P.U., Goyvaerts, C., Maenhout, S., Pen, J., Thielemans, K., Breckpot, K. (2013) Lentiviral vectors: a versatile tool to fight cancer. *Curr Mol Med* 13, 602-25.
[14] Wing, K., Onishi, Y., Prieto-Martin, P., Yamaguchi, T., Miyara, M., Fehervari, Z., No‐ mura, T., Sakaguchi, S. (2008) CTLA-4 control over Foxp3+regulatory T cell function.
[15] Qureshi, O.S., Zheng, Y., Nakamura, K., Attridge, K., Manzotti, C., Schmidt, E.M., Baker, J., Jeffery, L.E., Kaur, S., Briggs, Z., Hou, T.Z., Futter, C.E., Anderson, G., Walker, L.S., Sansom, D.M. (2011) Trans-endocytosis of CD80 and CD86: a molecular
[16] Denoeud, J., Moser, M. (2011) Role of CD27/CD70 pathway of activation in immunity
[17] Oldenhove, G., de Heusch, M., Urbain-Vansanten, G., Urbain, J., Maliszewski, C., Leo, O., Moser, M. (2003) CD4+CD25+regulatory T cells control T helper cell type 1 responses to foreign antigens induced by mature dendritic cells in vivo. *J Exp Med*
[18] Cao, X., Cai, S.F., Fehniger, T.A., Song, J., Collins, L.I., Piwnica-Worms, D.R., Ley, T.J. (2007) Granzyme B and perforin are important for regulatory T cell-mediated sup‐
[19] Barthlott, T., Moncrieffe, H., Veldhoen, M., Atkins, C.J., Christensen, J., O'Garra, A., Stockinger, B. (2005) CD25+CD4+T cells compete with naive CD4+T cells for IL-2 and
[20] Fallarino, F., Grohmann, U., Hwang, K.W., Orabona, C., Vacca, C., Bianchi, R., Bella‐ donna, M.L., Fioretti, M.C., Alegre, M.L., Puccetti, P. (2003) Modulation of trypto‐
[21] Kobie, J.J., Shah, P.R., Yang, L., Rebhahn, J.A., Fowell, D.J., Mosmann, T.R. (2006) T regulatory and primed uncommitted CD4 T cells express CD73, which suppresses ef‐ fector CD4 T cells by converting 5'-adenosine monophosphate to adenosine. *J Immu‐*
exploit it for the induction of IL-10 production. *Int Immunol* 17, 279-88.
phan catabolism by regulatory T cells. *Nat Immunol* 4, 1206-12.
basis for the cell-extrinsic function of CTLA-4. *Science* 332, 600-3.
13, e32-42.
90 Immune Response Activation
*cer* 127, 759-67.
*Science* 322, 271-5.
198, 259-66.
*nol* 177, 6780-6.
tolerance. *J Clin Invest* 117, 1147-54.
and tolerance. *J Leukoc Biol* 89, 195-203.
pression of tumor clearance. *Immunity* 27, 635-46.
zation of cancer antigens: a national cancer institute pilot project for the acceleration of translational research. *Clin Cancer Res* 15, 5323-37.
[45] De Keersmaecker, B., Heirman, C., Corthals, J., Empsen, C., van Grunsven, L.A., Al‐ lard, S.D., Pen, J., Lacor, P., Thielemans, K., Aerts, J.L. (2011) The combination of 4-1BBL and CD40L strongly enhances the capacity of dendritic cells to stimulate
Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation
http://dx.doi.org/10.5772/57539
93
[46] Escors, D., Breckpot, K. (2010) Lentiviral vectors in gene therapy: their current status
[47] Breckpot, K., Aerts, J.L., Thielemans, K. (2007) Lentiviral vectors for cancer immuno‐ therapy: transforming infectious particles into therapeutics. *Gene Ther* 14, 847-62. [48] Breckpot, K., Emeagi, P.U., Thielemans, K. (2008) Lentiviral vectors for anti-tumor
[49] Boudreau, J.E., Bonehill, A., Thielemans, K., Wan, Y. (2011) Engineering dendritic
[50] Van Lint, S., Heirman, C., Thielemans, K., Breckpot, K. (2013) mRNA: From a chemi‐ cal blueprint for protein production to an off-the-shelf therapeutic. *Hum Vaccin Im‐*
[51] Steinman, R.M. (2007) Dendritic cells: understanding immunogenicity. *Eur J Immunol*
[52] Trombetta, E.S., Mellman, I. (2005) Cell biology of antigen processing in vitro and in
[53] Cools, N., Ponsaerts, P., Van Tendeloo, V.F., Berneman, Z.N. (2007) Balancing be‐ tween immunity and tolerance: an interplay between dendritic cells, regulatory T
[54] Mahnke, K., Schmitt, E., Bonifaz, L., Enk, A.H., Jonuleit, H. (2002) Immature, but not inactive: the tolerogenic function of immature dendritic cells. *Immunol Cell Biol* 80,
[55] Yamazaki, S., Inaba, K., Tarbell, K.V., Steinman, R.M. (2006) Dendritic cells expand antigen-specific Foxp3+CD25+CD4+regulatory T cells including suppressors of allor‐
[56] Steinman, R.M. (2003) Some interfaces of dendritic cell biology. *Apmis* 111, 675-97.
[57] Lin, H.H., Ray, S., Tongchusak, S., Reinherz, E.L., Brusic, V. (2008) Evaluation of MHC class I peptide binding prediction servers: applications for vaccine research.
[58] Kurts, C., Wagner, H. (2011) Grand challenges in molecular antigen-presenting cell-
[59] Nierkens, S., den Brok, M.H., Garcia, Z., Togher, S., Wagenaars, J., Wassink, M., Boon, L., Ruers, T.J., Figdor, C.G., Schoenberger, S.P., Adema, G.J., Janssen, E.M. (2011) Immune adjuvant efficacy of CpG oligonucleotide in cancer treatment is
HIV-specific T cell responses. *J Leukoc Biol* 89, 989-99.
immunotherapy. *Curr Gene Ther* 8, 438-48.
vivo. *Annu Rev Immunol* 23, 975-1028.
eactivity. *Immunol Rev* 212, 314-29.
*BMC Immunol* 9, 8.
biology. *Front Immunol* 2, 8.
cells, and effector T cells. *J Leukoc Biol* 82, 1365-74.
*munother* 9.
477-83.
37 Suppl 1, S53-60.
and future potential. *Arch Immunol Ther Exp (Warsz)* 58, 107-19.
cells to enhance cancer immunotherapy. *Mol Ther* 19, 841-53.
[45] De Keersmaecker, B., Heirman, C., Corthals, J., Empsen, C., van Grunsven, L.A., Al‐ lard, S.D., Pen, J., Lacor, P., Thielemans, K., Aerts, J.L. (2011) The combination of 4-1BBL and CD40L strongly enhances the capacity of dendritic cells to stimulate HIV-specific T cell responses. *J Leukoc Biol* 89, 989-99.
zation of cancer antigens: a national cancer institute pilot project for the acceleration
[35] Breckpot, K., Dullaers, M., Bonehill, A., van Meirvenne, S., Heirman, C., de Greef, C., van der Bruggen, P., Thielemans, K. (2003) Lentivirally transduced dendritic cells as
[36] Dyall, J., Latouche, J.B., Schnell, S., Sadelain, M. (2001) Lentivirus-transduced human monocyte-derived dendritic cells efficiently stimulate antigen-specific cytotoxic T
[37] Michiels, A., Tuyaerts, S., Bonehill, A., Heirman, C., Corthals, J., Thielemans, K. (2008) Delivery of tumor-antigen-encoding mRNA into dendritic cells for vaccina‐
[38] Ponsaerts, P., Van den Bosch, G., Cools, N., Van Driessche, A., Nijs, G., Lenjou, M., Lardon, F., Van Broeckhoven, C., Van Bockstaele, D.R., Berneman, Z.N., Van Tende‐ loo, V.F. (2002) Messenger RNA electroporation of human monocytes, followed by rapid in vitro differentiation, leads to highly stimulatory antigen-loaded mature den‐
[39] Schroers, R., Sinha, I., Segall, H., Schmidt-Wolf, I.G., Rooney, C.M., Brenner, M.K., Sutton, R.E., Chen, S.Y. (2000) Transduction of human PBMC-derived dendritic cells and macrophages by an HIV-1-based lentiviral vector system. *Mol Ther* 1, 171-9.
[40] Tuyaerts, S., Noppe, S.M., Corthals, J., Breckpot, K., Heirman, C., De Greef, C., Van Riet, I., Thielemans, K. (2002) Generation of large numbers of dendritic cells in a
[41] Van Tendeloo, V.F., Ponsaerts, P., Lardon, F., Nijs, G., Lenjou, M., Van Broeckhoven, C., Van Bockstaele, D.R., Berneman, Z.N. (2001) Highly efficient gene delivery by mRNA electroporation in human hematopoietic cells: superiority to lipofection and passive pulsing of mRNA and to electroporation of plasmid cDNA for tumor antigen
[42] Dullaers, M., Breckpot, K., Van Meirvenne, S., Bonehill, A., Tuyaerts, S., Michiels, A., Straetman, L., Heirman, C., De Greef, C., Van Der Bruggen, P., Thielemans, K. (2004) Side-by-side comparison of lentivirally transduced and mRNA-electroporated den‐ dritic cells: implications for cancer immunotherapy protocols. *Mol Ther* 10, 768-79.
[43] Bonehill, A., Tuyaerts, S., Van Nuffel, A.M., Heirman, C., Bos, T.J., Fostier, K., Neyns, B., Thielemans, K. (2008) Enhancing the T-cell stimulatory capacity of human den‐ dritic cells by co-electroporation with CD40L, CD70 and constitutively active TLR4
[44] Breckpot, K., Heirman, C., Neyns, B., Thielemans, K. (2004) Exploiting dendritic cells for cancer immunotherapy: genetic modification of dendritic cells. *J Gene Med* 6,
closed system using Cell Factories. *J Immunol Methods* 264, 135-51.
of translational research. *Clin Cancer Res* 15, 5323-37.
a tool for cancer immunotherapy. *J Gene Med* 5, 654-67.
lymphocytes. *Blood* 97, 114-21.
92 Immune Response Activation
tion. *Methods Mol Biol* 423, 155-63.
dritic cells. *J Immunol* 169, 1669-75.
loading of dendritic cells. *Blood* 98, 49-56.
encoding mRNA. *Mol Ther* 16, 1170-80.
1175-88.
founded specifically upon TLR9 function in plasmacytoid dendritic cells. *Cancer Res* 71, 6428-37.
[73] Watts, T.H. (2005) TNF/TNFR family members in costimulation of T cell responses.
Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation
http://dx.doi.org/10.5772/57539
95
[74] Ju, X., Zenke, M., Hart, D.N., Clark, G.J. (2008) CD300a/c regulate type I interferon and TNF-alpha secretion by human plasmacytoid dendritic cells stimulated with
[75] Lechmann, M., Berchtold, S., Hauber, J., Steinkasserer, A. (2002) CD83 on dendritic
[76] Sharpe, A.H., Freeman, G.J. (2002) The B7-CD28 superfamily. *Nat Rev Immunol* 2,
[77] Wang, S., Chen, L. (2004) Co-signaling molecules of the B7-CD28 family in positive and negative regulation of T lymphocyte responses. *Microbes Infect* 6, 759-66.
[78] de Jong, D.M., Pate, J.L., Kirkland, T.N., Taylor, C.E., Baker, P.J., Takayama, K. (1991) Lipopolysaccharidelike immunological properties of cell wall glycoproteins isolated
[79] Baars, P.A., Sierro, S., Arens, R., Tesselaar, K., Hooibrink, B., Klenerman, P., van Lier, R.A. (2005) Properties of murine (CD8+)CD27-T cells. *Eur J Immunol* 35, 3131-41.
[80] Wherry, E.J., Teichgraber, V., Becker, T.C., Masopust, D., Kaech, S.M., Antia, R., von Andrian, U.H., Ahmed, R. (2003) Lineage relationship and protective immunity of
[81] Bowman, M.R., Crimmins, M.A., Yetz-Aldape, J., Kriz, R., Kelleher, K., Herrmann, S. (1994) The cloning of CD70 and its identification as the ligand for CD27. *J Immunol*
[82] Hintzen, R.Q., Lens, S.M., Beckmann, M.P., Goodwin, R.G., Lynch, D., van Lier, R.A. (1994) Characterization of the human CD27 ligand, a novel member of the TNF gene
[83] Tesselaar, K., Gravestein, L.A., van Schijndel, G.M., Borst, J., van Lier, R.A. (1997) Characterization of murine CD70, the ligand of the TNF receptor family member
[84] Hendriks, J., Gravestein, L.A., Tesselaar, K., van Lier, R.A., Schumacher, T.N., Borst, J. (2000) CD27 is required for generation and long-term maintenance of T cell im‐
[85] Hendriks, J., Xiao, Y., Borst, J. (2003) CD27 promotes survival of activated T cells and complements CD28 in generation and establishment of the effector T cell pool. *J Exp*
[86] Carr, J.M., Carrasco, M.J., Thaventhiran, J.E., Bambrough, P.J., Kraman, M., Edwards, A.D., Al-Shamkhani, A., Fearon, D.T. (2006) CD27 mediates interleukin-2-independ‐
cells: more than just a marker for maturation. *Trends Immunol* 23, 273-5.
*Annu Rev Immunol* 23, 23-68.
116-26.
152, 1756-61.
family. *J Immunol* 152, 1762-73.
CD27. *J Immunol* 159, 4959-65.
munity. *Nat Immunol* 1, 433-40.
*Med* 198, 1369-80.
TLR7 and TLR9 ligands. *Blood* 112, 1184-94.
from Cytophaga johnsonae. *Infect Immun* 59, 2631-7.
memory CD8 T cell subsets. *Nat Immunol* 4, 225-34.
[73] Watts, T.H. (2005) TNF/TNFR family members in costimulation of T cell responses. *Annu Rev Immunol* 23, 23-68.
founded specifically upon TLR9 function in plasmacytoid dendritic cells. *Cancer Res*
[60] Croft, M. (2003) Co-stimulatory members of the TNFR family: keys to effective T-cell
[61] Villadangos, J.A., Schnorrer, P. (2007) Intrinsic and cooperative antigen-presenting
[62] Curtsinger, J.M., Lins, D.C., Mescher, M.F. (2003) Signal 3 determines tolerance ver‐ sus full activation of naive CD8 T cells: dissociating proliferation and development of
[64] Nolte, M.A., van Olffen, R.W., van Gisbergen, K.P., van Lier, R.A. (2009) Timing and tuning of CD27-CD70 interactions: the impact of signal strength in setting the bal‐ ance between adaptive responses and immunopathology. *Immunol Rev* 229, 216-31.
[65] Azuma, T., Yao, S., Zhu, G., Flies, A.S., Flies, S.J., Chen, L. (2008) B7-H1 is a ubiqui‐
[66] Grohmann, U., Volpi, C., Fallarino, F., Bozza, S., Bianchi, R., Vacca, C., Orabona, C., Belladonna, M.L., Ayroldi, E., Nocentini, G., Boon, L., Bistoni, F., Fioretti, M.C., Ro‐ mani, L., Riccardi, C., Puccetti, P. (2007) Reverse signaling through GITR ligand ena‐
[67] Orabona, C., Grohmann, U., Belladonna, M.L., Fallarino, F., Vacca, C., Bianchi, R., Bozza, S., Volpi, C., Salomon, B.L., Fioretti, M.C., Romani, L., Puccetti, P. (2004) CD28 induces immunostimulatory signals in dendritic cells via CD80 and CD86. *Nat Immu‐*
[68] Zhao, F., Cannons, J.L., Dutta, M., Griffiths, G.M., Schwartzberg, P.L. (2009) Positive and negative signaling through SLAM receptors regulate synapse organization and
[69] Aerts-Toegaert, C., Heirman, C., Tuyaerts, S., Corthals, J., Aerts, J.L., Bonehill, A., Thielemans, K., Breckpot, K. (2007) CD83 expression on dendritic cells and T cells:
[70] Chen, L., Flies, D.B. (2013) Molecular mechanisms of T cell co-stimulation and co-in‐
[71] Bour-Jordan, H., Bluestone, J.A. (2009) Regulating the regulators: costimulatory sig‐ nals control the homeostasis and function of regulatory T cells. *Immunol Rev* 229,
[72] Carreno, B.M., Collins, M. (2002) The B7 family of ligands and its receptors: new pathways for costimulation and inhibition of immune responses. *Annu Rev Immunol*
correlation with effective immune responses. *Eur J Immunol* 37, 686-95.
tous antiapoptotic receptor on cancer cells. *Blood* 111, 3635-43.
bles dexamethasone to activate IDO in allergy. *Nat Med* 13, 579-86.
[63] Griffiths, G.M. (1995) The cell biology of CTL killing. *Curr Opin Immunol* 7, 343-8.
functions of dendritic-cell subsets in vivo. *Nat Rev Immunol* 7, 543-55.
71, 6428-37.
94 Immune Response Activation
*nol* 5, 1134-42.
41-66.
20, 29-53.
immunity? *Nat Rev Immunol* 3, 609-20.
effector function. *J Exp Med* 197, 1141-51.
thresholds of cytolysis. *Immunity* 36, 1003-16.
hibition. *Nat Rev Immunol* 13, 227-42.
ent clonal expansion of the CD8+T cell without effector differentiation. *Proc Natl Acad Sci U S A* 103, 19454-9.
Neyns, B. (2011) Therapeutic vaccination with an autologous mRNA electroporated dendritic cell vaccine in patients with advanced melanoma. *J Immunother* 34, 448-56. [98] Van Nuffel, A.M., Tuyaerts, S., Benteyn, D., Wilgenhof, S., Corthals, J., Heirman, C., Neyns, B., Thielemans, K., Bonehill, A. (2012) Epitope and HLA-type independent monitoring of antigen-specific T-cells after treatment with dendritic cells presenting
Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation
http://dx.doi.org/10.5772/57539
97
[99] Benteyn, D., Van Nuffel, A.M., Wilgenhof, S., Corthals, J., Heirman, C., Neyns, B., Thielemans, K., Bonehill, A. (2013) Characterization of CD8+T-cell responses in the peripheral blood and skin injection sites of melanoma patients treated with mRNA electroporated autologous dendritic cells (TriMixDC-MEL). *Biomed Res Int* 2013,
[100] Wilgenhof, S., Van Nuffel, A.M., Benteyn, D., Corthals, J., Aerts, C., Heirman, C., Van Riet, I., Bonehill, A., Thielemans, K., Neyns, B. (2013) A phase IB study on intrave‐ nous synthetic mRNA electroporated dendritic cell immunotherapy in pretreated ad‐
[101] Valzasina, B., Guiducci, C., Dislich, H., Killeen, N., Weinberg, A.D., Colombo, M.P. (2005) Triggering of OX40 (CD134) on CD4(+)CD25+T cells blocks their inhibitory ac‐ tivity: a novel regulatory role for OX40 and its comparison with GITR. *Blood* 105,
[102] Piconese, S., Valzasina, B., Colombo, M.P. (2008) OX40 triggering blocks suppression by regulatory T cells and facilitates tumor rejection. *J Exp Med* 205, 825-39.
[103] Shimizu, J., Yamazaki, S., Takahashi, T., Ishida, Y., Sakaguchi, S. (2002) Stimulation of CD25(+)CD4(+) regulatory T cells through GITR breaks immunological self-toler‐
[104] Jame, D.W., West, J.M., Dooley, P.C., Stephenson, D.G. (2004) Differential effects of arginine, glutamate and phosphoarginine on Ca(2+)-activation properties of muscle
[105] Stephens, G.L., McHugh, R.S., Whitters, M.J., Young, D.A., Luxenberg, D., Carreno, B.M., Collins, M., Shevach, E.M. (2004) Engagement of glucocorticoid-induced TNFR family-related receptor on effector T cells by its ligand mediates resistance to sup‐
[106] Choi, B.K., Bae, J.S., Choi, E.M., Kang, W.J., Sakaguchi, S., Vinay, D.S., Kwon, B.S. (2004) 4-1BB-dependent inhibition of immunosuppression by activated CD4+CD25+T
[107] Claus, C., Riether, C., Schurch, C., Matter, M.S., Hilmenyuk, T., Ochsenbein, A.F. (2012) CD27 signaling increases the frequency of regulatory T cells and promotes tu‐
[108] Pen, J.J., De Keersmaecker, B., Maenhout, S.K., Van Nuffel, A.M., Heirman, C., Cor‐ thals, J., Escors, D., Bonehill, A., Thielemans, K., Breckpot, K., Aerts, J.L. (2013) Mod‐
fibres from crayfish and rat. *J Muscle Res Cell Motil* 25, 497-508.
pression by CD4+CD25+T cells. *J Immunol* 173, 5008-20.
full-length tumor antigens. *J Immunol Methods* 377, 23-36.
vanced melanoma patients. *Ann Oncol*.
ance. *Nat Immunol* 3, 135-42.
cells. *J Leukoc Biol* 75, 785-91.
mor growth. *Cancer Res* 72, 3664-76.
976383.
2845-51.
Neyns, B. (2011) Therapeutic vaccination with an autologous mRNA electroporated dendritic cell vaccine in patients with advanced melanoma. *J Immunother* 34, 448-56.
[98] Van Nuffel, A.M., Tuyaerts, S., Benteyn, D., Wilgenhof, S., Corthals, J., Heirman, C., Neyns, B., Thielemans, K., Bonehill, A. (2012) Epitope and HLA-type independent monitoring of antigen-specific T-cells after treatment with dendritic cells presenting full-length tumor antigens. *J Immunol Methods* 377, 23-36.
ent clonal expansion of the CD8+T cell without effector differentiation. *Proc Natl Acad*
[87] Xiao, Y., Peperzak, V., Keller, A.M., Borst, J. (2008) CD27 instructs CD4+T cells to provide help for the memory CD8+T cell response after protein immunization. *J Im‐*
[88] Dolfi, D.V., Boesteanu, A.C., Petrovas, C., Xia, D., Butz, E.A., Katsikis, P.D. (2008) Late signals from CD27 prevent Fas-dependent apoptosis of primary CD8+T cells. *J*
[89] Taraban, V.Y., Rowley, T.F., Tough, D.F., Al-Shamkhani, A. (2006) Requirement for CD70 in CD4+Th cell-dependent and innate receptor-mediated CD8+T cell priming. *J*
[90] Soares, H., Waechter, H., Glaichenhaus, N., Mougneau, E., Yagita, H., Mizenina, O., Dudziak, D., Nussenzweig, M.C., Steinman, R.M. (2007) A subset of dendritic cells induces CD4+T cells to produce IFN-gamma by an IL-12-independent but CD70-de‐
[91] Keller, A.M., Schildknecht, A., Xiao, Y., van den Broek, M., Borst, J. (2008) Expression of costimulatory ligand CD70 on steady-state dendritic cells breaks CD8+T cell toler‐
[92] Arens, R., Schepers, K., Nolte, M.A., van Oosterwijk, M.F., van Lier, R.A., Schumach‐ er, T.N., van Oers, M.H. (2004) Tumor rejection induced by CD70-mediated quantita‐ tive and qualitative effects on effector CD8+T cell formation. *J Exp Med* 199, 1595-605.
[93] Aulwurm, S., Wischhusen, J., Friese, M., Borst, J., Weller, M. (2006) Immune stimula‐ tory effects of CD70 override CD70-mediated immune cell apoptosis in rodent glio‐ ma models and confer long-lasting antiglioma immunity in vivo. *Int J Cancer* 118,
[94] French, R.R., Taraban, V.Y., Crowther, G.R., Rowley, T.F., Gray, J.C., Johnson, P.W., Tutt, A.L., Al-Shamkhani, A., Glennie, M.J. (2007) Eradication of lymphoma by CD8 T cells following anti-CD40 monoclonal antibody therapy is critically dependent on
[95] Roberts, D.J., Franklin, N.A., Kingeter, L.M., Yagita, H., Tutt, A.L., Glennie, M.J., Bul‐ lock, T.N. (2010) Control of established melanoma by CD27 stimulation is associated with enhanced effector function and persistence, and reduced PD-1 expression of tu‐
[96] Bonehill, A., Van Nuffel, A.M., Corthals, J., Tuyaerts, S., Heirman, C., Francois, V., Colau, D., van der Bruggen, P., Neyns, B., Thielemans, K. (2009) Single-step antigen loading and activation of dendritic cells by mRNA electroporation for the purpose of
[97] Wilgenhof, S., Van Nuffel, A.M., Corthals, J., Heirman, C., Tuyaerts, S., Benteyn, D., De Coninck, A., Van Riet, I., Verfaillie, G., Vandeloo, J., Bonehill, A., Thielemans, K.,
therapeutic vaccination in melanoma patients. *Clin Cancer Res* 15, 3366-75.
pendent mechanism in vivo. *J Exp Med* 204, 1095-106.
CD27 costimulation. *Blood* 109, 4810-5.
mor infiltrating CD8(+) T cells. *J Immunother* 33, 769-79.
ance and permits effective immunity. *Immunity* 29, 934-46.
*Sci U S A* 103, 19454-9.
96 Immune Response Activation
*munol* 181, 1071-82.
*Immunol* 180, 2912-21.
*Immunol* 177, 2969-75.
1728-35.
ulation of Regulatory T Cell Function by Monocyte-Derived Dendritic Cells Matured through Electroporation with mRNA Encoding CD40 Ligand, Constitutively Active TLR4, and CD70. *J Immunol* 191, 1976-83.
[120] Dong, H., Zhu, G., Tamada, K., Chen, L. (1999) B7-H1, a third member of the B7 fami‐ ly, co-stimulates T-cell proliferation and interleukin-10 secretion. *Nat Med* 5, 1365-9.
Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation
http://dx.doi.org/10.5772/57539
99
[121] Tamura, H., Dong, H., Zhu, G., Sica, G.L., Flies, D.B., Tamada, K., Chen, L. (2001) B7- H1 costimulation preferentially enhances CD28-independent T-helper cell function.
[122] Tseng, S.Y., Otsuji, M., Gorski, K., Huang, X., Slansky, J.E., Pai, S.I., Shalabi, A., Shin, T., Pardoll, D.M., Tsuchiya, H. (2001) B7-DC, a new dendritic cell molecule with po‐
[123] Blank, C., Gajewski, T.F., Mackensen, A. (2005) Interaction of PD-L1 on tumor cells with PD-1 on tumor-specific T cells as a mechanism of immune evasion: implications
[124] Blank, C., Kuball, J., Voelkl, S., Wiendl, H., Becker, B., Walter, B., Majdic, O., Gajew‐ ski, T.F., Theobald, M., Andreesen, R., Mackensen, A. (2006) Blockade of PD-L1 (B7- H1) augments human tumor-specific T cell responses in vitro. *Int J Cancer* 119,
[125] Butte, M.J., Keir, M.E., Phamduy, T.B., Sharpe, A.H., Freeman, G.J. (2007) Program‐ med death-1 ligand 1 interacts specifically with the B7-1 costimulatory molecule to
[126] Butte, M.J., Pena-Cruz, V., Kim, M.J., Freeman, G.J., Sharpe, A.H. (2008) Interaction of
[127] Park, J.J., Omiya, R., Matsumura, Y., Sakoda, Y., Kuramasu, A., Augustine, M.M., Yao, S., Tsushima, F., Narazaki, H., Anand, S., Liu, Y., Strome, S.E., Chen, L., Tama‐ da, K. (2010) B7-H1/CD80 interaction is required for the induction and maintenance
[128] Okazaki, T., Honjo, T. (2006) The PD-1-PD-L pathway in immunological tolerance.
[129] Dong, H., Zhu, G., Tamada, K., Flies, D.B., van Deursen, J.M., Chen, L. (2004) B7-H1 determines accumulation and deletion of intrahepatic CD8(+) T lymphocytes. *Immun‐*
[130] Wang, L., Pino-Lagos, K., de Vries, V.C., Guleria, I., Sayegh, M.H., Noelle, R.J. (2008) Programmed death 1 ligand signaling regulates the generation of adaptive
[131] Francisco, L.M., Salinas, V.H., Brown, K.E., Vanguri, V.K., Freeman, G.J., Kuchroo, V.K., Sharpe, A.H. (2009) PD-L1 regulates the development, maintenance, and func‐
[132] Ahmadzadeh, M., Johnson, L.A., Heemskerk, B., Wunderlich, J.R., Dudley, M.E., White, D.E., Rosenberg, S.A. (2009) Tumor antigen-specific CD8 T cells infiltrating
Foxp3+CD4+regulatory T cells. *Proc Natl Acad Sci U S A* 105, 9331-6.
tion of induced regulatory T cells. *J Exp Med* 206, 3015-29.
tent costimulatory properties for T cells. *J Exp Med* 193, 839-46.
for tumor immunotherapy. *Cancer Immunol Immunother* 54, 307-14.
inhibit T cell responses. *Immunity* 27, 111-22.
human PD-L1 and B7-1. *Mol Immunol* 45, 3567-72.
of peripheral T-cell tolerance. *Blood* 116, 1291-8.
*Trends Immunol* 27, 195-201.
*ity* 20, 327-36.
*Blood* 97, 1809-16.
317-27.
[120] Dong, H., Zhu, G., Tamada, K., Chen, L. (1999) B7-H1, a third member of the B7 fami‐ ly, co-stimulates T-cell proliferation and interleukin-10 secretion. *Nat Med* 5, 1365-9.
ulation of Regulatory T Cell Function by Monocyte-Derived Dendritic Cells Matured through Electroporation with mRNA Encoding CD40 Ligand, Constitutively Active
[109] Pepe, C.A., Ricci, R., Cortelazzi, C., Mori, G., Zambito, F., Campanini, N., Aimi, F., Santoro, S., Lombardi, M., Santini, M., de Panfilis, G. (2011) The vast majority of lym‐ phocytes infiltrating primary cutaneous melanoma express the CD27 costimulatory
[110] Riether, C., Schurch, C., Ochsenbein, A.F. (2012) Modulating CD27 signaling to treat
[111] Keilholz, U. (2008) CTLA-4: negative regulator of the immune response and a target
[112] Pentcheva-Hoang, T., Corse, E., Allison, J.P. (2009) Negative regulators of T-cell acti‐ vation: potential targets for therapeutic intervention in cancer, autoimmune disease,
[113] Schneider, H., Downey, J., Smith, A., Zinselmeyer, B.H., Rush, C., Brewer, J.M., Wei, B., Hogg, N., Garside, P., Rudd, C.E. (2006) Reversal of the TCR stop signal by
[114] Dequen, P., Lorigan, P., Jansen, J.P., van Baardewijk, M., Ouwens, M.J., Kotapati, S. (2012) Systematic review and network meta-analysis of overall survival comparing 3 mg/kg ipilimumab with alternative therapies in the management of pretreated pa‐
[115] Ishida, Y., Agata, Y., Shibahara, K., Honjo, T. (1992) Induced expression of PD-1, a novel member of the immunoglobulin gene superfamily, upon programmed cell
[116] Kinter, A.L., Godbout, E.J., McNally, J.P., Sereti, I., Roby, G.A., O'Shea, M.A., Fauci, A.S. (2008) The common gamma-chain cytokines IL-2, IL-7, IL-15, and IL-21 induce the expression of programmed death-1 and its ligands. *J Immunol* 181, 6738-46. [117] Agata, Y., Kawasaki, A., Nishimura, H., Ishida, Y., Tsubata, T., Yagita, H., Honjo, T. (1996) Expression of the PD-1 antigen on the surface of stimulated mouse T and B
[118] Freeman, G.J., Long, A.J., Iwai, Y., Bourque, K., Chernova, T., Nishimura, H., Fitz, L.J., Malenkovich, N., Okazaki, T., Byrne, M.C., Horton, H.F., Fouser, L., Carter, L., Ling, V., Bowman, M.R., Carreno, B.M., Collins, M., Wood, C.R., Honjo, T. (2000) En‐ gagement of the PD-1 immunoinhibitory receptor by a novel B7 family member leads
[119] Raimondi, G., Shufesky, W.J., Tokita, D., Morelli, A.E., Thomson, A.W. (2006) Regu‐ lated compartmentalization of programmed cell death-1 discriminates CD4+CD25+resting regulatory T cells from activated T cells. *J Immunol* 176, 2808-16.
to negative regulation of lymphocyte activation. *J Exp Med* 192, 1027-34.
tients with unresectable stage III or IV melanoma. *Oncologist* 17, 1376-85.
receptor: implications for melanoma progression. *Eur J Dermatol* 21, 178-83.
TLR4, and CD70. *J Immunol* 191, 1976-83.
98 Immune Response Activation
cancer. *Oncoimmunology* 1, 1604-1606.
CTLA-4. *Science* 313, 1972-5.
death. *Embo J* 11, 3887-95.
lymphocytes. *Int Immunol* 8, 765-72.
for cancer therapy. *J Immunother* 31, 431-9.
and persistent infections. *Immunol Rev* 229, 67-87.
the tumor express high levels of PD-1 and are functionally impaired. *Blood* 114, 1537-44.
metastatic melanoma patients with clinical benefit. *Proc Natl Acad Sci U S A* 105,
Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation
http://dx.doi.org/10.5772/57539
101
[144] Perales, M.A., Yuan, J., Powel, S., Gallardo, H.F., Rasalan, T.S., Gonzalez, C., Manu‐ kian, G., Wang, J., Zhang, Y., Chapman, P.B., Krown, S.E., Livingston, P.O., Ejadi, S., Panageas, K.S., Engelhorn, M.E., Terzulli, S.L., Houghton, A.N., Wolchok, J.D. (2008) Phase I/II study of GM-CSF DNA as an adjuvant for a multipeptide cancer vaccine in
[145] Imai, N., Ikeda, H., Tawara, I., Shiku, H. (2009) Tumor progression inhibits the in‐ duction of multifunctionality in adoptively transferred tumor-specific CD8+T cells.
[146] Pen, J. J., B. D. Keersmaecker, C. Heirman, J. Corthals, T. Liechtenstein, D. Escors, K. Thielemans, and K. Breckpot. 2014. Interference with PD-L1/PD-1 co-stimulation during antigen presentation enhances the multifunctionality of antigen-specific T
[147] Grzywnowicz, M., Zaleska, J., Mertens, D., Tomczak, W., Wlasiuk, P., Kosior, K., Piechnik, A., Bojarska-Junak, A., Dmoszynska, A., Giannopoulos, K. (2012) Program‐ med death-1 and its ligand are novel immunotolerant molecules expressed on leuke‐
[148] Keir, M.E., Butte, M.J., Freeman, G.J., Sharpe, A.H. (2008) PD-1 and its ligands in tol‐
[149] Hamanishi, J., Mandai, M., Iwasaki, M., Okazaki, T., Tanaka, Y., Yamaguchi, K., Hi‐ guchi, T., Yagi, H., Takakura, K., Minato, N., Honjo, T., Fujii, S. (2007) Programmed cell death 1 ligand 1 and tumor-infiltrating CD8+T lymphocytes are prognostic fac‐
[150] Thompson, R.H., Gillett, M.D., Cheville, J.C., Lohse, C.M., Dong, H., Webster, W.S., Krejci, K.G., Lobo, J.R., Sengupta, S., Chen, L., Zincke, H., Blute, M.L., Strome, S.E., Leibovich, B.C., Kwon, E.D. (2004) Costimulatory B7-H1 in renal cell carcinoma pa‐ tients: Indicator of tumor aggressiveness and potential therapeutic target. *Proc Natl*
[151] Thompson, R.H., Kuntz, S.M., Leibovich, B.C., Dong, H., Lohse, C.M., Webster, W.S., Sengupta, S., Frank, I., Parker, A.S., Zincke, H., Blute, M.L., Sebo, T.J., Cheville, J.C., Kwon, E.D. (2006) Tumor B7-H1 is associated with poor prognosis in renal cell carci‐
[152] Nomi, T., Sho, M., Akahori, T., Hamada, K., Kubo, A., Kanehiro, H., Nakamura, S., Enomoto, K., Yagita, H., Azuma, M., Nakajima, Y. (2007) Clinical significance and therapeutic potential of the programmed death-1 ligand/programmed death-1 path‐
[153] Ghebeh, H., Mohammed, S., Al-Omair, A., Qattan, A., Lehe, C., Al-Qudaihi, G., El‐ kum, N., Alshabanah, M., Bin Amer, S., Tulbah, A., Ajarim, D., Al-Tweigeri, T., Der‐
mic B cells in chronic lymphocytic leukemia. *PLoS One* 7, e35178.
tors of human ovarian cancer. *Proc Natl Acad Sci U S A* 104, 3360-5.
noma patients with long-term follow-up. *Cancer Res* 66, 3381-5.
way in human pancreatic cancer. *Clin Cancer Res* 13, 2151-7.
erance and immunity. *Annu Rev Immunol* 26, 677-704.
patients with advanced melanoma. *Mol Ther* 16, 2022-9.
20410-5.
*Eur J Immunol* 39, 241-53.
cells. *Gene Ther* 21:262-271.
*Acad Sci U S A* 101, 17174-9.
metastatic melanoma patients with clinical benefit. *Proc Natl Acad Sci U S A* 105, 20410-5.
[144] Perales, M.A., Yuan, J., Powel, S., Gallardo, H.F., Rasalan, T.S., Gonzalez, C., Manu‐ kian, G., Wang, J., Zhang, Y., Chapman, P.B., Krown, S.E., Livingston, P.O., Ejadi, S., Panageas, K.S., Engelhorn, M.E., Terzulli, S.L., Houghton, A.N., Wolchok, J.D. (2008) Phase I/II study of GM-CSF DNA as an adjuvant for a multipeptide cancer vaccine in patients with advanced melanoma. *Mol Ther* 16, 2022-9.
the tumor express high levels of PD-1 and are functionally impaired. *Blood* 114,
[133] Francisco, L.M., Sage, P.T., Sharpe, A.H. (2010) The PD-1 pathway in tolerance and
[134] Radziewicz, H., Ibegbu, C.C., Fernandez, M.L., Workowski, K.A., Obideen, K., Weh‐ bi, M., Hanson, H.L., Steinberg, J.P., Masopust, D., Wherry, E.J., Altman, J.D., Rouse, B.T., Freeman, G.J., Ahmed, R., Grakoui, A. (2007) Liver-infiltrating lymphocytes in chronic human hepatitis C virus infection display an exhausted phenotype with high
[135] Rutebemberwa, A., Ray, S.C., Astemborski, J., Levine, J., Liu, L., Dowd, K.A., Clute, S., Wang, C., Korman, A., Sette, A., Sidney, J., Pardoll, D.M., Cox, A.L. (2008) Highprogrammed death-1 levels on hepatitis C virus-specific T cells during acute infec‐ tion are associated with viral persistence and require preservation of cognate antigen
[136] Fife, B.T., Pauken, K.E., Eagar, T.N., Obu, T., Wu, J., Tang, Q., Azuma, M., Krummel, M.F., Bluestone, J.A. (2009) Interactions between PD-1 and PD-L1 promote tolerance
[137] Karwacz, K., Arce, F., Bricogne, C., Kochan, G., Escors, D. (2012) PD-L1 co-stimula‐ tion, ligand-induced TCR down-modulation and anti-tumor immunotherapy. *On‐*
[138] Karwacz, K., Bricogne, C., MacDonald, D., Arce, F., Bennett, C.L., Collins, M., Escors, D. (2011) PD-L1 co-stimulation contributes to ligand-induced T cell receptor down-
[139] Cai, Z., Kishimoto, H., Brunmark, A., Jackson, M.R., Peterson, P.A., Sprent, J. (1997) Requirements for peptide-induced T cell receptor downregulation on naive CD8+T
[140] Escors, D., Bricogne, C., Arce, F., Kochan, G., Karwacz, K. (2011) On the Mechanism of T cell receptor down-modulation and its physiological significance. *J Biosci Med* 1.
[141] Liechtenstein, T., Dufait, I., Bricogne, C., Lanna, A., Pen, J., Breckpot, K., Escors, D. (2012) PD-L1/PD-1 Co-Stimulation, a Brake for T cell Activation and a T cell Differen‐
[142] Seder, R.A., Darrah, P.A., Roederer, M. (2008) T-cell quality in memory and protec‐
[143] Yuan, J., Gnjatic, S., Li, H., Powel, S., Gallardo, H.F., Ritter, E., Ku, G.Y., Jungbluth, A.A., Segal, N.H., Rasalan, T.S., Manukian, G., Xu, Y., Roman, R.A., Terzulli, S.L., Heywood, M., Pogoriler, E., Ritter, G., Old, L.J., Allison, J.P., Wolchok, J.D. (2008) CTLA-4 blockade enhances polyfunctional NY-ESO-1 specific T cell responses in
tion: implications for vaccine design. *Nat Rev Immunol* 8, 247-58.
levels of PD-1 and low levels of CD127 expression. *J Virol* 81, 2545-53.
by blocking the TCR-induced stop signal. *Nat Immunol* 10, 1185-92.
1537-44.
100 Immune Response Activation
autoimmunity. *Immunol Rev* 236, 219-42.
during chronic infection. *J Immunol* 181, 8215-25.
modulation on CD8+T cells. *EMBO Mol Med* 3, 581-92.
*coimmunology* 1, 86-88.
cells. *J Exp Med* 185, 641-51.
tiation Signal. *J Clin Cell Immunol* S12.
mime, S. (2006) The B7-H1 (PD-L1) T lymphocyte-inhibitory molecule is expressed in breast cancer patients with infiltrating ductal carcinoma: correlation with important high-risk prognostic factors. *Neoplasia* 8, 190-8.
fractory solid tumors: safety, clinical activity, pharmacodynamics, and immunologic
Manipulating Immune Regulatory Pathways to Enhance T Cell Stimulation
http://dx.doi.org/10.5772/57539
103
[163] Lipson, E.J., Sharfman, W.H., Drake, C.G., Wollner, I., Taube, J.M., Anders, R.A., Xu, H., Yao, S., Pons, A., Chen, L., Pardoll, D.M., Brahmer, J.R., Topalian, S.L. (2013) Du‐ rable cancer regression off-treatment and effective reinduction therapy with an anti-
[164] Topalian, S.L., Hodi, F.S., Brahmer, J.R., Gettinger, S.N., Smith, D.C., McDermott, D.F., Powderly, J.D., Carvajal, R.D., Sosman, J.A., Atkins, M.B., Leming, P.D., Spigel, D.R., Antonia, S.J., Horn, L., Drake, C.G., Pardoll, D.M., Chen, L., Sharfman, W.H., Anders, R.A., Taube, J.M., McMiller, T.L., Xu, H., Korman, A.J., Jure-Kunkel, M., Agrawal, S., McDonald, D., Kollia, G.D., Gupta, A., Wigginton, J.M., Sznol, M. (2012) Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. *N Engl J Med*
[165] Brahmer, J.R., Tykodi, S.S., Chow, L.Q., Hwu, W.J., Topalian, S.L., Hwu, P., Drake, C.G., Camacho, L.H., Kauh, J., Odunsi, K., Pitot, H.C., Hamid, O., Bhatia, S., Martins, R., Eaton, K., Chen, S., Salay, T.M., Alaparthy, S., Grosso, J.F., Korman, A.J., Parker, S.M., Agrawal, S., Goldberg, S.M., Pardoll, D.M., Gupta, A., Wigginton, J.M. (2012) Safety and activity of anti-PD-L1 antibody in patients with advanced cancer. N Engl J
correlates. *J Clin Oncol* 28, 3167-75.
366, 2443-54.
Med 366, 2455-65.
PD-1 antibody. *Clin Cancer Res* 19, 462-8.
fractory solid tumors: safety, clinical activity, pharmacodynamics, and immunologic correlates. *J Clin Oncol* 28, 3167-75.
[163] Lipson, E.J., Sharfman, W.H., Drake, C.G., Wollner, I., Taube, J.M., Anders, R.A., Xu, H., Yao, S., Pons, A., Chen, L., Pardoll, D.M., Brahmer, J.R., Topalian, S.L. (2013) Du‐ rable cancer regression off-treatment and effective reinduction therapy with an anti-PD-1 antibody. *Clin Cancer Res* 19, 462-8.
mime, S. (2006) The B7-H1 (PD-L1) T lymphocyte-inhibitory molecule is expressed in breast cancer patients with infiltrating ductal carcinoma: correlation with important
[154] Hirano, F., Kaneko, K., Tamura, H., Dong, H., Wang, S., Ichikawa, M., Rietz, C., Flies, D.B., Lau, J.S., Zhu, G., Tamada, K., Chen, L. (2005) Blockade of B7-H1 and PD-1 by monoclonal antibodies potentiates cancer therapeutic immunity. *Cancer Res* 65,
[155] Iwai, Y., Ishida, M., Tanaka, Y., Okazaki, T., Honjo, T., Minato, N. (2002) Involvement of PD-L1 on tumor cells in the escape from host immune system and tumor immuno‐
[156] Strome, S.E., Dong, H., Tamura, H., Voss, S.G., Flies, D.B., Tamada, K., Salomao, D., Cheville, J., Hirano, F., Lin, W., Kasperbauer, J.L., Ballman, K.V., Chen, L. (2003) B7- H1 blockade augments adoptive T-cell immunotherapy for squamous cell carcinoma.
[157] Dong, H., Strome, S.E., Salomao, D.R., Tamura, H., Hirano, F., Flies, D.B., Roche, P.C., Lu, J., Zhu, G., Tamada, K., Lennon, V.A., Celis, E., Chen, L. (2002) Tumor-asso‐ ciated B7-H1 promotes T-cell apoptosis: a potential mechanism of immune evasion.
[158] Curiel, T.J., Wei, S., Dong, H., Alvarez, X., Cheng, P., Mottram, P., Krzysiek, R., Knut‐ son, K.L., Daniel, B., Zimmermann, M.C., David, O., Burow, M., Gordon, A., Dhur‐ andhar, N., Myers, L., Berggren, R., Hemminki, A., Alvarez, R.D., Emilie, D., Curiel, D.T., Chen, L., Zou, W. (2003) Blockade of B7-H1 improves myeloid dendritic cell-
[159] Webster, W.S., Thompson, R.H., Harris, K.J., Frigola, X., Kuntz, S., Inman, B.A., Dong, H. (2007) Targeting molecular and cellular inhibitory mechanisms for im‐ provement of antitumor memory responses reactivated by tumor cell vaccine. *J Im‐*
[160] Berger, R., Rotem-Yehudar, R., Slama, G., Landes, S., Kneller, A., Leiba, M., Koren-Michowitz, M., Shimoni, A., Nagler, A. (2008) Phase I safety and pharmacokinetic study of CT-011, a humanized antibody interacting with PD-1, in patients with ad‐
[161] Wong, R.M., Scotland, R.R., Lau, R.L., Wang, C., Korman, A.J., Kast, W.M., Weber, J.S. (2007) Programmed death-1 blockade enhances expansion and functional capaci‐
[162] Brahmer, J.R., Drake, C.G., Wollner, I., Powderly, J.D., Picus, J., Sharfman, W.H., Stankevich, E., Pons, A., Salay, T.M., McMiller, T.L., Gilson, M.M., Wang, C., Selby, M., Taube, J.M., Anders, R., Chen, L., Korman, A.J., Pardoll, D.M., Lowy, I., Topalian, S.L. (2010) Phase I study of single-agent anti-programmed death-1 (MDX-1106) in re‐
ty of human melanoma antigen-specific CTLs. *Int Immunol* 19, 1223-34.
vanced hematologic malignancies. *Clin Cancer Res* 14, 3044-51.
therapy by PD-L1 blockade. *Proc Natl Acad Sci U S A* 99, 12293-7.
high-risk prognostic factors. *Neoplasia* 8, 190-8.
mediated antitumor immunity. *Nat Med* 9, 562-7.
1089-96.
102 Immune Response Activation
*Cancer Res* 63, 6501-5.
*Nat Med* 8, 793-800.
*munol* 179, 2860-9.
**Chapter 4**
**Negative Immunomodulators –**
Rabab Elmezayen and Jonathan Moorman
Additional information is available at the end of the chapter
The human immune system requires a delicate balance of positive and negative regulators to provide appropriate responses to different stimuli. It is important to control the extent of the response and limit the response at the conclusion of an immune event. The immune system has many negative immunomodulators that assist in this function, including such recently described molecules as PD-1, KLRG1, SOCS1, and Tim-3. These immunomodulators, amongst others, are under direct or indirect control of a group of small non-coding RNAs called microRNAs (miRNAs) that regulate gene expression via degradation or translational repres‐ sion of their target mRNAs. While these modulators perform a vital function during normal immune responses, they have also been implicated in the pathology of chronic infections, including hepatitis C virus (HCV), herpes simplex virus (HSV) and human immunodeficiency virus (HIV) (Table 1). It is believed that these viruses alter the expression of negative immu‐ nomodulators to blunt appropriate immune responses, facilitating persistent viral infection. Because of the role of these modulators and the involved miRNAs, they are being investigated as possible treatments for chronic infections. This review is not exhaustive but highlights several lines of research in regard to negative immunomodulators in innate and adaptive
> © 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
**Blunting Immunostimulation**
**and Facilitating Infection**
Jeddidiah Griffin, Marie Moulton,
http://dx.doi.org/10.5772/57540
immune responses to stimulation.
**1. Introduction**
**Chapter 4**
## **Negative Immunomodulators – Blunting Immunostimulation and Facilitating Infection**
Jeddidiah Griffin, Marie Moulton, Rabab Elmezayen and Jonathan Moorman
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/57540
#### **1. Introduction**
The human immune system requires a delicate balance of positive and negative regulators to provide appropriate responses to different stimuli. It is important to control the extent of the response and limit the response at the conclusion of an immune event. The immune system has many negative immunomodulators that assist in this function, including such recently described molecules as PD-1, KLRG1, SOCS1, and Tim-3. These immunomodulators, amongst others, are under direct or indirect control of a group of small non-coding RNAs called microRNAs (miRNAs) that regulate gene expression via degradation or translational repres‐ sion of their target mRNAs. While these modulators perform a vital function during normal immune responses, they have also been implicated in the pathology of chronic infections, including hepatitis C virus (HCV), herpes simplex virus (HSV) and human immunodeficiency virus (HIV) (Table 1). It is believed that these viruses alter the expression of negative immu‐ nomodulators to blunt appropriate immune responses, facilitating persistent viral infection. Because of the role of these modulators and the involved miRNAs, they are being investigated as possible treatments for chronic infections. This review is not exhaustive but highlights several lines of research in regard to negative immunomodulators in innate and adaptive immune responses to stimulation.
© 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
lating Tim-3 is associated with an upregulation of the proinflammatory cytokine IL-12, but also IL-10, an anti-inflammatory cytokine, and PD-1 has been inversely correlated with Tim-3 expression. Due to the complexity of Tim-3 interactions, research into the extent and mecha‐ nism of the functions of Tim-3 is still ongoing. However, current data suggest Tim-3 is a key
Negative Immunomodulators – Blunting Immunostimulation and Facilitating Infection
http://dx.doi.org/10.5772/57540
107
Killer cell lectin like receptor G1 (KLRG1) is found on NK cells and activated T cells [14, 15]. As cells mature and differentiate, they express KLRG1 in higher numbers. KLRG1 also has an ITIM domain, suggesting an inhibitory activity. When bound with its ligand, E-cadherin, KLRG1 has been shown to inhibit degranulation and IFN-γ secretion by NK cells. Furthermore, KLRG1 also inhibits CD4 T cell activity by decreasing CD95 mediated lysis [14, 15]. Henson *et al*. [14, 15] have shown in CD8 T cells that blocking KLRG1 upregulates cyclin D and E and downregulates cyclin inhibitor p27. To prevent T cell activation, both KLRG1 and CD3/TLR need to be stimulated proximally [16], suggesting that both cell signals need to be provided by the same target cell. The KLRG1 ITIM recruits SHIP-1 and SHP-2 phosphatases, facilitating the conversion of PIP3 to PIP2 [17] and thus blocking the actions of PI3K and downstream phosphorylation of Akt [18]. When Akt is not phosphorylated, proliferation is decreased, perhaps by a lack of cyclin D and E and cyclin inhibitor p27 remaining at levels sufficient for
SOCS1 is a member of a family called suppressors of cytokine signaling [19]. SOCS members display a wide spectrum of negative cytokine regulation and have a part in preventing harmful cytokine effects in certain organs [20]. SOCS1 is a negative immunomodulator for both cytokine and TLR receptor mediated signaling, and the members of the SOCS family have two key motifs: a SOCS box and a SH2 domain [19]. In addition, SOCS 1 also possesses a kinaseinhibitory region (KIR). The SH2 domain allows SOCS to associate with cytokine receptors and tyrosine kinases, inhibiting cytokine signaling by dampening kinases or making adaptor molecules for receptor sites unavailable [19]. The SOCS box recruits proteins that perform such functions as regulating degradation of proteins bound with SOCS [21]. SOCS1 prefers to use its SH2 domain to bind JAK kinases [20]. SOCS1's inhibitory function comes from reducing JAK kinase activity through the KIR domain and using the SOCS box to degrade JAK kinases. Cytokines such as IL-4, TNF-γ, and some TLR ligands can cause the upregulation of SOCS1 expression [22]. SOCS1 also directly inhibits TLR signaling [21, 22]. The mechanism has not
been conclusively demonstrated, but several proposals exist [23, 24].
**3. Negative immunomodulators in viral infection — Hepatitis C**
Over 180 million people throughout world are infected with Hepatitis C Virus [25]. Of those with an acute HCV infection, 85% become chronic [26]. HCV has developed ways to evade the
regulator in both innate and adaptive immune responses.
**2.3. KLRG1**
inhibition [15].
**2.4. SOCS1**
**Table 1.** Negative Immunomodulators implicated in Chronic Infection
#### **2. Negative immunomodulators — The players**
#### **2.1. PD-1**
The gene for Programmed Death-1 (PD-1) was first discovered in 1992 during experimentation with T cell hybridomas [1]; PD-1 was found to be upregulated in cells undergoing apoptosis, and is a type 1transmembrane protein 288 amino acids long [2]. It contains an intracellular domain, a transmembrane domain, a stalk approximately 20 aa in length, and an immuno‐ globulin superfamily domain that contains an immunoreceptor tyrosine-based switch motif (ITSM) and an immunoreceptor tyrosine-based inhibitory motif (ITIM). PD-1 is expressed on T cells, B cells, dendritic cells (DC), natural killer T cells (NKT cells), and monocytes that have been activated, and two ligands, PD-L1 and PD-L2, have been described. PD-1 expression leads to inhibition of cytokine production, including TNF-α, IFN-γ, IL-2, and possibly the cell survival factor Bcl-xL [2, 3]. PD-1 signals have an inverse relationship with TCR stimulation; at lower levels of TCR stimulation, PD-1 signals are stronger; CD28 signaling, however, can overpower the inhibitory effects of PD-1 [4].When PD-1 binds to its ligand, TCR signaling is decreased, highlighting a function of PD-1 in adaptive immunity. On the cell surface, PD-1 activation induces the phosphorylation of cytoplasmic tyrosines [5, 6]. SHP-2 associates with the PD-1 ITSM, dephosphorylating signals via the P13K pathway and Akt and ultimately decreasing the expression of cytokines such as IL-2, TNF-α, and IFN-γ [2, 5]. Decreases in IL-2 have been associated with cell death, revealing a possible mechanism of function of PD-1 in apoptosis.
#### **2.2. Tim-3**
Tim-3 (T cell immunoglobulin and mucin domain 3) was initially found to be expressed on Th1 cells but has since been identified on monocytes, DC, mast cells, and microglia [7-10]. Tim-3 has an IgV domain, a mucin domain, a transmembrane domain, and a cytoplasmic tail [11]. Studies indicate that only the IgV domain binds to the ligand [10]. The most studied ligand of Tim-3 is galectin-9 (Gal-9) [12]. Tim-3 is thought to play a role in both innate and adaptive immune responses. In inactive immune cells, Tim-3 is expressed on DC and functions through Toll-like receptors (TLRs) to facilitate inflammation [13]. However, when Th1 cells are stimulated, Tim-3 expression on those cells is increased and, in the presence of Gal-9, impedes their responses. Tim-3 has also been shown to decrease macrophage activity [13]. Downregu‐ lating Tim-3 is associated with an upregulation of the proinflammatory cytokine IL-12, but also IL-10, an anti-inflammatory cytokine, and PD-1 has been inversely correlated with Tim-3 expression. Due to the complexity of Tim-3 interactions, research into the extent and mecha‐ nism of the functions of Tim-3 is still ongoing. However, current data suggest Tim-3 is a key regulator in both innate and adaptive immune responses.
#### **2.3. KLRG1**
**Molecule Ligand Expression**
KLRG1 E-cadherin NK cells, T cells SOCS1 JAK 2, cytokine receptors Multiple somatic cells
**Table 1.** Negative Immunomodulators implicated in Chronic Infection
**2.1. PD-1**
106 Immune Response Activation
apoptosis.
**2.2. Tim-3**
**2. Negative immunomodulators — The players**
PD-1 PD-L1/PD-L2 B cells, T cells, DC, NK T cells, Treg,
Tim-3 Gal-9 T cells, Treg, monocytes, macrophages, DC
The gene for Programmed Death-1 (PD-1) was first discovered in 1992 during experimentation with T cell hybridomas [1]; PD-1 was found to be upregulated in cells undergoing apoptosis, and is a type 1transmembrane protein 288 amino acids long [2]. It contains an intracellular domain, a transmembrane domain, a stalk approximately 20 aa in length, and an immuno‐ globulin superfamily domain that contains an immunoreceptor tyrosine-based switch motif (ITSM) and an immunoreceptor tyrosine-based inhibitory motif (ITIM). PD-1 is expressed on T cells, B cells, dendritic cells (DC), natural killer T cells (NKT cells), and monocytes that have been activated, and two ligands, PD-L1 and PD-L2, have been described. PD-1 expression leads to inhibition of cytokine production, including TNF-α, IFN-γ, IL-2, and possibly the cell survival factor Bcl-xL [2, 3]. PD-1 signals have an inverse relationship with TCR stimulation; at lower levels of TCR stimulation, PD-1 signals are stronger; CD28 signaling, however, can overpower the inhibitory effects of PD-1 [4].When PD-1 binds to its ligand, TCR signaling is decreased, highlighting a function of PD-1 in adaptive immunity. On the cell surface, PD-1 activation induces the phosphorylation of cytoplasmic tyrosines [5, 6]. SHP-2 associates with the PD-1 ITSM, dephosphorylating signals via the P13K pathway and Akt and ultimately decreasing the expression of cytokines such as IL-2, TNF-α, and IFN-γ [2, 5]. Decreases in IL-2 have been associated with cell death, revealing a possible mechanism of function of PD-1 in
Tim-3 (T cell immunoglobulin and mucin domain 3) was initially found to be expressed on Th1 cells but has since been identified on monocytes, DC, mast cells, and microglia [7-10]. Tim-3 has an IgV domain, a mucin domain, a transmembrane domain, and a cytoplasmic tail [11]. Studies indicate that only the IgV domain binds to the ligand [10]. The most studied ligand of Tim-3 is galectin-9 (Gal-9) [12]. Tim-3 is thought to play a role in both innate and adaptive immune responses. In inactive immune cells, Tim-3 is expressed on DC and functions through Toll-like receptors (TLRs) to facilitate inflammation [13]. However, when Th1 cells are stimulated, Tim-3 expression on those cells is increased and, in the presence of Gal-9, impedes their responses. Tim-3 has also been shown to decrease macrophage activity [13]. Downregu‐
monocytes
Killer cell lectin like receptor G1 (KLRG1) is found on NK cells and activated T cells [14, 15]. As cells mature and differentiate, they express KLRG1 in higher numbers. KLRG1 also has an ITIM domain, suggesting an inhibitory activity. When bound with its ligand, E-cadherin, KLRG1 has been shown to inhibit degranulation and IFN-γ secretion by NK cells. Furthermore, KLRG1 also inhibits CD4 T cell activity by decreasing CD95 mediated lysis [14, 15]. Henson *et al*. [14, 15] have shown in CD8 T cells that blocking KLRG1 upregulates cyclin D and E and downregulates cyclin inhibitor p27. To prevent T cell activation, both KLRG1 and CD3/TLR need to be stimulated proximally [16], suggesting that both cell signals need to be provided by the same target cell. The KLRG1 ITIM recruits SHIP-1 and SHP-2 phosphatases, facilitating the conversion of PIP3 to PIP2 [17] and thus blocking the actions of PI3K and downstream phosphorylation of Akt [18]. When Akt is not phosphorylated, proliferation is decreased, perhaps by a lack of cyclin D and E and cyclin inhibitor p27 remaining at levels sufficient for inhibition [15].
#### **2.4. SOCS1**
SOCS1 is a member of a family called suppressors of cytokine signaling [19]. SOCS members display a wide spectrum of negative cytokine regulation and have a part in preventing harmful cytokine effects in certain organs [20]. SOCS1 is a negative immunomodulator for both cytokine and TLR receptor mediated signaling, and the members of the SOCS family have two key motifs: a SOCS box and a SH2 domain [19]. In addition, SOCS 1 also possesses a kinaseinhibitory region (KIR). The SH2 domain allows SOCS to associate with cytokine receptors and tyrosine kinases, inhibiting cytokine signaling by dampening kinases or making adaptor molecules for receptor sites unavailable [19]. The SOCS box recruits proteins that perform such functions as regulating degradation of proteins bound with SOCS [21]. SOCS1 prefers to use its SH2 domain to bind JAK kinases [20]. SOCS1's inhibitory function comes from reducing JAK kinase activity through the KIR domain and using the SOCS box to degrade JAK kinases. Cytokines such as IL-4, TNF-γ, and some TLR ligands can cause the upregulation of SOCS1 expression [22]. SOCS1 also directly inhibits TLR signaling [21, 22]. The mechanism has not been conclusively demonstrated, but several proposals exist [23, 24].
#### **3. Negative immunomodulators in viral infection — Hepatitis C**
Over 180 million people throughout world are infected with Hepatitis C Virus [25]. Of those with an acute HCV infection, 85% become chronic [26]. HCV has developed ways to evade the host immune system, allowing the virus to persist for prolonged periods of time. HCV is a single-stranded RNA virus with a genome of approximately 9.5 kb. A member of the *Flavivi‐ rus* family, this virus's genome codes for a single polyprotein precursor that is cleaved into the proteins necessary for viral propagation. The HCV core protein is of particular interest to researchers, as it is implicated in playing a major role in evading host immune responses [27].
using siRNAs increased IL-12 expression and inhibited PD-1 up-regulation, suggesting
Negative Immunomodulators – Blunting Immunostimulation and Facilitating Infection
KLRG1 has been shown to be upregulated in acute hepatitis C infection in T cells but has a relatively low expression in chronic HCV patients [48]. These investigators suggested that, as the number of TCR-triggering events increases, so does the probability that a T cell will express KLRG1. This implies that TCRs are more frequently stimulated in acute HCV infections than chronic infections. Recent studies in NK cells demonstrated that individuals with chronic HCV infection display a reduction in NK cell number and function concomitant with a significant increase in KLRG1 expression. KLRG1 expression was associated with increased NK cell apoptosis and a diminished capacity to undergo cell activation and produce IFNγ, phenom‐
Tim-3 has been studied in individuals with HCV infection and found to have a role in both innate and adaptive immune responses. Patients who have a high level of Tim-3 early in infection are more likely to become chronically infected than those with normal levels of Tim-3. The individuals exhibit a higher percentage of CTLs that express both Tim-3 and PD-1 [50]. Tim-3 blockade restores proliferation, cytotoxicity, and killing of HCV-expressing hepato‐ cytes, and blocking PD-1 in the same cells has an even greater effect in most subjects [50].
HCV-infected hepatocytes expressed higher levels of Gal-9 and TGF-β, and drove conven‐
dependent manner and blocking Tim-3/Gal-9 ligations abrogated HCV-mediated Treg
In innate immune responses, Tim-3 expression was up-regulated on monocytes in HCVinfected individuals and this expression was inversely correlated with IL-12 expression [13, 53]. Tim-3 blockade significantly improved HCV-mediated IL-12 suppression, reduced HCV core-mediated expression of the negative immunoinhibitors PD-1 and SOCS-1, and increased STAT-1 phosphorylation in monocyte macrophages in these studies. Figure 1 represents a potential model for the interplay between the negative immunomodulators during HCV
HIV is a retrovirus with an envelope and is a single-stranded RNA virus. Three enzymes, reverse transcriptase, protease, and integrase, are present in the core of the virus. These proteins are the target of many antiretroviral drugs used for HIV therapy. The HIV pandemic has spurred much of the research into negative immunomodulators, which appear to be
Tregulatory cells (Tregs), which could be
Tregs that express TGF-β and IL-10 [51, 52]. Additionally,
T cells into Foxp3+
T cells with HCV-expressing hepatocytes.
http://dx.doi.org/10.5772/57540
109
Tregs in a dose-
crosstalk between the immunomodulators [43].
enon; this could be reversed by KLRG1 blockade [49].
Tim-3 expression is also up-regulated on Foxp3+
T cells into CD25+
**Immunodeficiency Virus (HIV)**
tional CD4+
induction [52].
infection.
induced *in vitro* by incubating purified healthy CD4+
Foxp3+
**4. Negative immunomodulators in viral infection— Human**
involved in many aspects of HIV-mediated immune dysregulation.
recombinant Gal-9 protein transformed TCR-activated CD4+
Virus-specific cytotoxic T lymphocytes (CTL) are critical for the elimination of the virusinfected cells. Strong viral specific CTL responses have been demonstrated in those individuals who spontaneously resolve HCV infection [28-30]. In chronically HCV-infected patients, however, virus-specific CTL responses are significantly diminished both in the liver and peripheral blood [31, 32]. Numerous additional studies have reported that impaired virusspecific CD4+ and CD8+ T cell functions are associated with chronic HCV infection [33-36]. This downregulation of cell-mediated T cell responses may contribute to chronic infection; Th CD4+ T cell activation and differentiation may be impaired by HCV [42]. It has been reported that CD4+ and CD8+ T cell responses, including virus-specific interferon gamma (IFN-γ) production, are severely suppressed in chronically HCV-infected patients through the PD-1 inhibitory pathway [37-39]. Several other investigators have supported these studies, includ‐ ing data suggesting PD-1 blockade can restore T cell functions [40, 41] and that HCV core protein may induce PD-1 upregulation and T cell dysfunction [36].
Additional data on the role of PD-1 in HCV infection has focused on innate immune responses. PD-1 appears to be upregulated in monocytes derived from HCV-infected individuals, and these monocytes are functionally impaired in terms of their ability to express IL-12 [42, 43]. Additionally, both "successful HCV treatment with pegylated interferon/ribavirin as well as blocking PD-1/PDL-1 engagement *ex vivo* were associated with reduced PD-1 expression and improved IL-12 production by monocytes". Studies have also demonstrated that monocyte/ macrophages isolated from both healthy and HCV-infected individuals and treated with HCV core protein displayed increased PD-1 expression and decreased IL-12 expression [42, 43].
SOCS1 seems to play an interesting role in chronic HCV infection. Miyoshi *et al*. [44] reported that HCV core protein inhibited SOCS1 expression and Yoshida *et al*. [45] reported SOCS1 gene methylation in hepatitis C patients. These experiments were however performed with hepatocellular carcinomas (HCCs). SOCS1 has been demonstrated to suppress liver cell proliferation [46, 47] and has been implicated as a tumor suppressor, consistent with findings that SOCS1 is downregulated in HCCs [19]. Frazier *et al*. [25] have performed experiments with healthy T cells and found SOCS1 to be upregulated in the presence of HCV core protein. These results suggest HCV core protein increases expression of SOCS1 in T cell populations, decreasing immune responses through the pathways described above; moreover, SOCS1 appears to be downregulated in HCCs in the presence of HCV core protein, altering its role in tumor suppression and liver cell regeneration and further contributing to liver pathology during infection. In innate immune responses, Monocytes isolated from both healthy and HCV-infected individuals and treated with HCV core protein displayed increased SOCS-1 expression and decreased IL-12 expression [43]. PD-1 and SOCS-1 were found to associate by co-immunoprecipitation studies, and either blocking PD-1 or silencing SOCS-1 in M/MØ led to activation of STAT-1 during TLR-stimulated IL-12 production. Silencing SOCS-1 expression using siRNAs increased IL-12 expression and inhibited PD-1 up-regulation, suggesting crosstalk between the immunomodulators [43].
host immune system, allowing the virus to persist for prolonged periods of time. HCV is a single-stranded RNA virus with a genome of approximately 9.5 kb. A member of the *Flavivi‐ rus* family, this virus's genome codes for a single polyprotein precursor that is cleaved into the proteins necessary for viral propagation. The HCV core protein is of particular interest to researchers, as it is implicated in playing a major role in evading host immune responses [27].
Virus-specific cytotoxic T lymphocytes (CTL) are critical for the elimination of the virusinfected cells. Strong viral specific CTL responses have been demonstrated in those individuals who spontaneously resolve HCV infection [28-30]. In chronically HCV-infected patients, however, virus-specific CTL responses are significantly diminished both in the liver and peripheral blood [31, 32]. Numerous additional studies have reported that impaired virus-
downregulation of cell-mediated T cell responses may contribute to chronic infection; Th
production, are severely suppressed in chronically HCV-infected patients through the PD-1 inhibitory pathway [37-39]. Several other investigators have supported these studies, includ‐ ing data suggesting PD-1 blockade can restore T cell functions [40, 41] and that HCV core
Additional data on the role of PD-1 in HCV infection has focused on innate immune responses. PD-1 appears to be upregulated in monocytes derived from HCV-infected individuals, and these monocytes are functionally impaired in terms of their ability to express IL-12 [42, 43]. Additionally, both "successful HCV treatment with pegylated interferon/ribavirin as well as blocking PD-1/PDL-1 engagement *ex vivo* were associated with reduced PD-1 expression and improved IL-12 production by monocytes". Studies have also demonstrated that monocyte/ macrophages isolated from both healthy and HCV-infected individuals and treated with HCV core protein displayed increased PD-1 expression and decreased IL-12 expression [42, 43].
SOCS1 seems to play an interesting role in chronic HCV infection. Miyoshi *et al*. [44] reported that HCV core protein inhibited SOCS1 expression and Yoshida *et al*. [45] reported SOCS1 gene methylation in hepatitis C patients. These experiments were however performed with hepatocellular carcinomas (HCCs). SOCS1 has been demonstrated to suppress liver cell proliferation [46, 47] and has been implicated as a tumor suppressor, consistent with findings that SOCS1 is downregulated in HCCs [19]. Frazier *et al*. [25] have performed experiments with healthy T cells and found SOCS1 to be upregulated in the presence of HCV core protein. These results suggest HCV core protein increases expression of SOCS1 in T cell populations, decreasing immune responses through the pathways described above; moreover, SOCS1 appears to be downregulated in HCCs in the presence of HCV core protein, altering its role in tumor suppression and liver cell regeneration and further contributing to liver pathology during infection. In innate immune responses, Monocytes isolated from both healthy and HCV-infected individuals and treated with HCV core protein displayed increased SOCS-1 expression and decreased IL-12 expression [43]. PD-1 and SOCS-1 were found to associate by co-immunoprecipitation studies, and either blocking PD-1 or silencing SOCS-1 in M/MØ led to activation of STAT-1 during TLR-stimulated IL-12 production. Silencing SOCS-1 expression
protein may induce PD-1 upregulation and T cell dysfunction [36].
T cell activation and differentiation may be impaired by HCV [42]. It has been reported
T cell functions are associated with chronic HCV infection [33-36]. This
T cell responses, including virus-specific interferon gamma (IFN-γ)
specific CD4+
108 Immune Response Activation
CD4+
that CD4+
and CD8+
and CD8+
KLRG1 has been shown to be upregulated in acute hepatitis C infection in T cells but has a relatively low expression in chronic HCV patients [48]. These investigators suggested that, as the number of TCR-triggering events increases, so does the probability that a T cell will express KLRG1. This implies that TCRs are more frequently stimulated in acute HCV infections than chronic infections. Recent studies in NK cells demonstrated that individuals with chronic HCV infection display a reduction in NK cell number and function concomitant with a significant increase in KLRG1 expression. KLRG1 expression was associated with increased NK cell apoptosis and a diminished capacity to undergo cell activation and produce IFNγ, phenom‐ enon; this could be reversed by KLRG1 blockade [49].
Tim-3 has been studied in individuals with HCV infection and found to have a role in both innate and adaptive immune responses. Patients who have a high level of Tim-3 early in infection are more likely to become chronically infected than those with normal levels of Tim-3. The individuals exhibit a higher percentage of CTLs that express both Tim-3 and PD-1 [50]. Tim-3 blockade restores proliferation, cytotoxicity, and killing of HCV-expressing hepato‐ cytes, and blocking PD-1 in the same cells has an even greater effect in most subjects [50].
Tim-3 expression is also up-regulated on Foxp3+ Tregulatory cells (Tregs), which could be induced *in vitro* by incubating purified healthy CD4+ T cells with HCV-expressing hepatocytes. HCV-infected hepatocytes expressed higher levels of Gal-9 and TGF-β, and drove conven‐ tional CD4+ T cells into CD25+ Foxp3+ Tregs that express TGF-β and IL-10 [51, 52]. Additionally, recombinant Gal-9 protein transformed TCR-activated CD4+ T cells into Foxp3+ Tregs in a dosedependent manner and blocking Tim-3/Gal-9 ligations abrogated HCV-mediated Treg induction [52].
In innate immune responses, Tim-3 expression was up-regulated on monocytes in HCVinfected individuals and this expression was inversely correlated with IL-12 expression [13, 53]. Tim-3 blockade significantly improved HCV-mediated IL-12 suppression, reduced HCV core-mediated expression of the negative immunoinhibitors PD-1 and SOCS-1, and increased STAT-1 phosphorylation in monocyte macrophages in these studies. Figure 1 represents a potential model for the interplay between the negative immunomodulators during HCV infection.
### **4. Negative immunomodulators in viral infection— Human Immunodeficiency Virus (HIV)**
HIV is a retrovirus with an envelope and is a single-stranded RNA virus. Three enzymes, reverse transcriptase, protease, and integrase, are present in the core of the virus. These proteins are the target of many antiretroviral drugs used for HIV therapy. The HIV pandemic has spurred much of the research into negative immunomodulators, which appear to be involved in many aspects of HIV-mediated immune dysregulation.
tional treatment leads to a decrease in PD-1 expression [61]. Blocking the binding of PD-L1 to PD-1 in CTLs has led to an increase in CTLs, which may contribute to a stronger adaptive immune response. These data suggest PD-1 could be a potent target for therapies for HIV. However, PD-1 may protect the vascular system from severe immunological responses during acute infections, with potential difficulties with blocking PD-1 pathways *in vivo* [62, 63].
Negative Immunomodulators – Blunting Immunostimulation and Facilitating Infection
In HIV-infected patients, Tim-3 expression is increased on T cells [64-66]. The binding of Gal-9 to Tim-3 has been demonstrated to render them less susceptible to HIV-1 infection and replication. The Gal-9/Tim-3 interaction on activated CD4 T cells leads to down-regulation of HIV-1 coreceptors and up-regulation of the cyclin-dependent kinase inhibitor p21 [67]. During the chronic stage of HIV, this interaction can lead to TH1 and TC1 cell exhaustion [68]. Thus, T cells expressing Tim-3 during chronic infection have an impaired ability to function properly
been shown to increase perforin release, increasing cytotoxic ability and the acquired immune response [68]. These data suggest blocking Tim-3 may prove to be a potent therapeutic
**5. Negative immunomodulators in viral infection — Herpes simplex virus**
The viral family *Herpesviridae* contains linear dsDNA, and HSV is a member of the subfamily *Alphaherpesvirinae*. These viruses lyse infected cells and grow rapidly; frequently, herpes viruses establish chronic subclinical infections in nerve ganglia that can develop into recurrent acute infections. Chronic infection by HSV leads to up-regulation of negative immunomodu‐ lators in a manner similar to HCV and HIV chronic infections. Maheller *et al*. [69] noted that SOCS1 is up-regulated during infection by HSV used in gene therapy experimentation, and this up-regulation of SOCS1 aided in viral replication. Because of the role SOCS1 plays in the JAK/STAT pathway, inhibition of this pathway may be what facilitates viral replication. It has also been demonstrated that SOCS3 is required for efficient viral replication of some strains of HSV [70]. This evidence suggests SOCS family cytokines have an important role in viral
expressing increased levels of PD-1 have been shown to produce less cytokines, such as IL-2, INF-γ, and TNF-α, than T cells that express less PD-1. After murine HSV infection, PD-1's ligand, PD-L1, has been found to be upregulated on dendritic cells [71]. It is thought PD-L1 on
infections. Some studies support the possibility of blocking PD-1/PD-L1 interactions to increase immune responses to latent HSV infection [71]. Bryant-Hudson and Carr [72], however, reported an increase in viral load in ganglia and murine cases of keratitis when PD-1/ PD-L1 interactions were blocked during acute infections. The pathology associated with keratitis may result from innate immune responses, such as inflammation due to cytokines
CD8+
cells that are Tim-3 negative have been shown to have a
cells. Blocking Tim-3 pathways in Tim-3+
T cells to decrease the immune response in chronic viral
cells has
111
http://dx.doi.org/10.5772/57540
T cells [71], perhaps due to
T cells infected with HSV and
and to increase in number [64]. CD8+
better ability to degranulate than Tim-3+
technique in the treatment of chronic HIV infection.
replication and they warrant further research.
DCs interacts with PD-1 on CD8+
During chronic HSV infections, PD-1 levels are elevated on CD8+
frequent antigen stimulation leading to exhaustion. Murine CD8+
**Figure 1.** Model for the role of negative immunomodulators during chronic HCV infection.
KLRG1 has been shown by Ibegbu *et al*. [54] to be upregulated on CD8+ cells in HIV positive patients. KLRG1 and its ligand, E-cadhedrin, decreased the cytotoxicity of HIV+ CD8+ cells [55]. As noted previously, KLRG1 has also been shown to be co-expressed with other inhibitory molecules such as PD1 in chronic infections such as in HCV infected cells [55]. Chronic latent infections such as Epstein-Barr virus and cytomegalovirus also increase KLRG1 expression, while acute viral infections such as influenza do not [30]. This suggests the persistent antigen stimulation in chronic infections may lead to immune senescence through molecules such as KLRG1, making this cell signal a potential target for therapeutic treatments.
SOCS1 expression is increased in HIV infected cells [56]. SOCS1 may aid in HIV particle maturation and infectivity by interacting with the HIV-1 p55 Gag protein. Viral protease cleaves Gag into domains key to viral replication; SOCS1 promotes the stability of Gag, and the SH2 domain may be of key importance in HIV-1 Gag binding [56]. Other SOCS proteins such as SOCS3 do not increase viral replication, suggesting the role of SOCS1 is unique amongst the proteins with a SOCS box domain.
HIV exposures causes increased expression of PD-1 [57], and PD-1 expression on HIV specific CD8+ cells correlates with HIV viral load [58], implying antigen signal strength leads to expression of PD-1 and exhaustion [59]. An HIV protein, Nef, was shown to use a MAPKdependent pathway to increase the expression of PD-1 [60]. PD-1 has been demonstrated to impair the function of cytotoxic T lymphocytes [59], contributing to the decreased immune response characteristic of HIV infection. However, controlling the viral load through tradi‐ tional treatment leads to a decrease in PD-1 expression [61]. Blocking the binding of PD-L1 to PD-1 in CTLs has led to an increase in CTLs, which may contribute to a stronger adaptive immune response. These data suggest PD-1 could be a potent target for therapies for HIV. However, PD-1 may protect the vascular system from severe immunological responses during acute infections, with potential difficulties with blocking PD-1 pathways *in vivo* [62, 63].
In HIV-infected patients, Tim-3 expression is increased on T cells [64-66]. The binding of Gal-9 to Tim-3 has been demonstrated to render them less susceptible to HIV-1 infection and replication. The Gal-9/Tim-3 interaction on activated CD4 T cells leads to down-regulation of HIV-1 coreceptors and up-regulation of the cyclin-dependent kinase inhibitor p21 [67]. During the chronic stage of HIV, this interaction can lead to TH1 and TC1 cell exhaustion [68]. Thus, T cells expressing Tim-3 during chronic infection have an impaired ability to function properly and to increase in number [64]. CD8+ cells that are Tim-3 negative have been shown to have a better ability to degranulate than Tim-3+ CD8+ cells. Blocking Tim-3 pathways in Tim-3+ cells has been shown to increase perforin release, increasing cytotoxic ability and the acquired immune response [68]. These data suggest blocking Tim-3 may prove to be a potent therapeutic technique in the treatment of chronic HIV infection.
#### **5. Negative immunomodulators in viral infection — Herpes simplex virus**
KLRG1 has been shown by Ibegbu *et al*. [54] to be upregulated on CD8+
**Figure 1.** Model for the role of negative immunomodulators during chronic HCV infection.
patients. KLRG1 and its ligand, E-cadhedrin, decreased the cytotoxicity of HIV+
KLRG1, making this cell signal a potential target for therapeutic treatments.
amongst the proteins with a SOCS box domain.
CD8+
110 Immune Response Activation
As noted previously, KLRG1 has also been shown to be co-expressed with other inhibitory molecules such as PD1 in chronic infections such as in HCV infected cells [55]. Chronic latent infections such as Epstein-Barr virus and cytomegalovirus also increase KLRG1 expression, while acute viral infections such as influenza do not [30]. This suggests the persistent antigen stimulation in chronic infections may lead to immune senescence through molecules such as
SOCS1 expression is increased in HIV infected cells [56]. SOCS1 may aid in HIV particle maturation and infectivity by interacting with the HIV-1 p55 Gag protein. Viral protease cleaves Gag into domains key to viral replication; SOCS1 promotes the stability of Gag, and the SH2 domain may be of key importance in HIV-1 Gag binding [56]. Other SOCS proteins such as SOCS3 do not increase viral replication, suggesting the role of SOCS1 is unique
HIV exposures causes increased expression of PD-1 [57], and PD-1 expression on HIV specific
cells correlates with HIV viral load [58], implying antigen signal strength leads to expression of PD-1 and exhaustion [59]. An HIV protein, Nef, was shown to use a MAPKdependent pathway to increase the expression of PD-1 [60]. PD-1 has been demonstrated to impair the function of cytotoxic T lymphocytes [59], contributing to the decreased immune response characteristic of HIV infection. However, controlling the viral load through tradi‐
cells in HIV positive
cells [55].
CD8+
The viral family *Herpesviridae* contains linear dsDNA, and HSV is a member of the subfamily *Alphaherpesvirinae*. These viruses lyse infected cells and grow rapidly; frequently, herpes viruses establish chronic subclinical infections in nerve ganglia that can develop into recurrent acute infections. Chronic infection by HSV leads to up-regulation of negative immunomodu‐ lators in a manner similar to HCV and HIV chronic infections. Maheller *et al*. [69] noted that SOCS1 is up-regulated during infection by HSV used in gene therapy experimentation, and this up-regulation of SOCS1 aided in viral replication. Because of the role SOCS1 plays in the JAK/STAT pathway, inhibition of this pathway may be what facilitates viral replication. It has also been demonstrated that SOCS3 is required for efficient viral replication of some strains of HSV [70]. This evidence suggests SOCS family cytokines have an important role in viral replication and they warrant further research.
During chronic HSV infections, PD-1 levels are elevated on CD8+ T cells [71], perhaps due to frequent antigen stimulation leading to exhaustion. Murine CD8+ T cells infected with HSV and expressing increased levels of PD-1 have been shown to produce less cytokines, such as IL-2, INF-γ, and TNF-α, than T cells that express less PD-1. After murine HSV infection, PD-1's ligand, PD-L1, has been found to be upregulated on dendritic cells [71]. It is thought PD-L1 on DCs interacts with PD-1 on CD8+ T cells to decrease the immune response in chronic viral infections. Some studies support the possibility of blocking PD-1/PD-L1 interactions to increase immune responses to latent HSV infection [71]. Bryant-Hudson and Carr [72], however, reported an increase in viral load in ganglia and murine cases of keratitis when PD-1/ PD-L1 interactions were blocked during acute infections. The pathology associated with keratitis may result from innate immune responses, such as inflammation due to cytokines such as IF-γ [73, 74]. Bryant-Hudson and Carr also hypothesized that PD-L1 may disrupt proper antigen presentation through the down-regulation of costimulatory molecules such as CD80/CD86. The different findings in acute versus latent infections highlight the complex role of many immunomodulators and the difficulties faced in manipulating these molecules in therapeutic settings.
**6.2. miRNA in chronic viral infection — Human Immunodeficiency Virus (HIV)**
**6.3. miRNA in chronic viral infection — Herpes Simplex Virus (HSV)**
in productive HSV-1 replication and reactivation from latency [91].
between apoptosis and cell survival.
**7. Conclusions**
**Acknowledgements**
There are two functional miRNAs, miR-TAR-5p and miR-TAR-3p, which form the transacti‐ vating response (TAR) element of human immunodeficiency virus type 1 (HIV-1). Ouellet et al. [87] reported recently that TAR miRNAs derived from HIV-1 can incorporate into host effector Argonaute protein complexes, which is required if these miRNAs are to regulate host mRNA expression. Bioinformatic predictions and reporter gene activity assays identified regulatory elements complementary and responsive to miRTAR-5p and miR-TAR-3p in the 3' untranslated region (UTR) of several candidate genes involved in apoptosis and cell survival. These include Caspase 8, Aiolos, Ikaros and Nucleophosmin (NPM)/B23. Analyses of Jurkat cells that stably expressed HIV-1 TAR or contained a full-length latent HIV provirus suggested that HIV-1 TAR miRNAs could regulate the expression of genes in T cells that affect the balance
Negative Immunomodulators – Blunting Immunostimulation and Facilitating Infection
http://dx.doi.org/10.5772/57540
113
One of the characteristics of Herpesviruses is their ability to maintain life-long latent infections in their hosts. As reported by Umbach et al. [88], herpes simplex virus 1 (HSV-1) establishes latency in neurons of sensory ganglia, where the only abundant viral gene product is a noncoding RNA. HSV-1 expresses at least two primary miRNA precursors in latently infected neurons that may facilitate the establishment and maintenance of viral latency by posttranscriptionally regulating viral gene expression. These include the latency associated transcript (LAT) [89, 90], which functions as a primary miRNA precursor that encodes four distinct miRNAs in HSV-1 infected cells. One of these miRNAs, miR-H2 3p, is transcribed antisense to ICP0, a viral immediate-early transcriptional activator thought to play a key role
The negative immunomodulators described herein are the targets of researchers attempting to find treatments for chronic diseases. However, the multi-faceted role of many of these molecules makes it difficult to assign specific functions and these immunomodulators likely function in a complex and often interconnected molecular system. The delicate balance of "on" signals and "off" signals within the immune response leads to challenges for this area of
This work was supported by NIH NIAID and an NIH NIDDK grants to JPM. This publication is the result of work supported with resources and the use of facilities at the James H. Quillen Veterans Affairs Medical Center. The contents in this publication do not represent the views
research, but the recent strides made in negative immunomodulation are promising.
Gal-9 is up-regulated in the presence of both latent and acute HSV infection [75], and the majority of CD8+ T cells in neural ganglia infected by HSV express Tim-3. This suggests a role for Tim-3 expression in HSV infection. Predictably, it has been demonstrated that Gal-9/Tim-3 interactions decrease the efficiency of CD8+ T cells in HSV infected neural ganglia [76]. Furthermore, Gal-9:/Tim-3 interaction on CD8+ T cells increases their apoptosis [77]. Reddy *et al*. [75] have supporting evidence that Tim-3 up-regulation facilitates HSV latency. However, in murine HSV keratitis, the immunoinflammatory-mediated pathology of the cornea is increased when Gal-9 is blocked with mAb [76]. In these studies, at least 50% of T cells associated with murine keratitis express Tim-3. Interestingly, additional Gal-9 decreases the severity of ocular lesions, presumably through decreasing the inflammatory response.
#### **6. MicroRNAs in negative immune regulation**
Micro RNAs (miRNAs) have been under extensive research as molecular regulators of negative immunomodulation, with well-described roles in hematopoiesis, inflammation, cancer and many other pathological processes [78, 79]. MicroRNAs are a relatively newly discovered class of non-coding RNAs that function as regulators of gene expression, either by transcriptional inhibition or mRNA degradation. They demonstrate effects either through their loss of function or overexpression, leading to pathologic responses that affect diverse cellular processes including cell differentiation, proliferation, and apoptosis. The effective recognition of viral infection and subsequent triggering of antiviral innate immune responses is controlled by multiple regulators, including miRNAs. A panel of miRNAs has been identified, including miR-124, miR-125, miR-223, miR-155, and miR-146, amongst others, and the role of individual miRNAs has been described in relation to specific pathologic process.
#### **6.1. miRNAs in chronic viral infection — Hepatitis C (HCV)**
miR-146a was initially identified during efforts to find miRNAs that are important in the innate immune response to microbial infection. These initial studies confirmed miR-146a as a transcriptional target of NF-kB, and it was thought to have a role as a feedback inhibitor of NF-kB activation. It is highly expressed in mature immune cells, including DC, peritoneal macrophages, granulocytes and splenic B and T cells [80]. Recent studies revealed that Tim-3 expression on T cells is regulated by the Th1 lineage determining transcription factor T-bet (Anderson et al., 2010), which is translationally regulated by miR-146a. Several studies link dysregulation of expression of several miRNAs to HCV and/or its related immunologic phenomenon [81-86].
#### **6.2. miRNA in chronic viral infection — Human Immunodeficiency Virus (HIV)**
There are two functional miRNAs, miR-TAR-5p and miR-TAR-3p, which form the transacti‐ vating response (TAR) element of human immunodeficiency virus type 1 (HIV-1). Ouellet et al. [87] reported recently that TAR miRNAs derived from HIV-1 can incorporate into host effector Argonaute protein complexes, which is required if these miRNAs are to regulate host mRNA expression. Bioinformatic predictions and reporter gene activity assays identified regulatory elements complementary and responsive to miRTAR-5p and miR-TAR-3p in the 3' untranslated region (UTR) of several candidate genes involved in apoptosis and cell survival. These include Caspase 8, Aiolos, Ikaros and Nucleophosmin (NPM)/B23. Analyses of Jurkat cells that stably expressed HIV-1 TAR or contained a full-length latent HIV provirus suggested that HIV-1 TAR miRNAs could regulate the expression of genes in T cells that affect the balance between apoptosis and cell survival.
#### **6.3. miRNA in chronic viral infection — Herpes Simplex Virus (HSV)**
One of the characteristics of Herpesviruses is their ability to maintain life-long latent infections in their hosts. As reported by Umbach et al. [88], herpes simplex virus 1 (HSV-1) establishes latency in neurons of sensory ganglia, where the only abundant viral gene product is a noncoding RNA. HSV-1 expresses at least two primary miRNA precursors in latently infected neurons that may facilitate the establishment and maintenance of viral latency by posttranscriptionally regulating viral gene expression. These include the latency associated transcript (LAT) [89, 90], which functions as a primary miRNA precursor that encodes four distinct miRNAs in HSV-1 infected cells. One of these miRNAs, miR-H2 3p, is transcribed antisense to ICP0, a viral immediate-early transcriptional activator thought to play a key role in productive HSV-1 replication and reactivation from latency [91].
#### **7. Conclusions**
such as IF-γ [73, 74]. Bryant-Hudson and Carr also hypothesized that PD-L1 may disrupt proper antigen presentation through the down-regulation of costimulatory molecules such as CD80/CD86. The different findings in acute versus latent infections highlight the complex role of many immunomodulators and the difficulties faced in manipulating these molecules in
Gal-9 is up-regulated in the presence of both latent and acute HSV infection [75], and the
for Tim-3 expression in HSV infection. Predictably, it has been demonstrated that Gal-9/Tim-3
*al*. [75] have supporting evidence that Tim-3 up-regulation facilitates HSV latency. However, in murine HSV keratitis, the immunoinflammatory-mediated pathology of the cornea is increased when Gal-9 is blocked with mAb [76]. In these studies, at least 50% of T cells associated with murine keratitis express Tim-3. Interestingly, additional Gal-9 decreases the severity of ocular lesions, presumably through decreasing the inflammatory response.
Micro RNAs (miRNAs) have been under extensive research as molecular regulators of negative immunomodulation, with well-described roles in hematopoiesis, inflammation, cancer and many other pathological processes [78, 79]. MicroRNAs are a relatively newly discovered class of non-coding RNAs that function as regulators of gene expression, either by transcriptional inhibition or mRNA degradation. They demonstrate effects either through their loss of function or overexpression, leading to pathologic responses that affect diverse cellular processes including cell differentiation, proliferation, and apoptosis. The effective recognition of viral infection and subsequent triggering of antiviral innate immune responses is controlled by multiple regulators, including miRNAs. A panel of miRNAs has been identified, including miR-124, miR-125, miR-223, miR-155, and miR-146, amongst others, and the role of individual
miR-146a was initially identified during efforts to find miRNAs that are important in the innate immune response to microbial infection. These initial studies confirmed miR-146a as a transcriptional target of NF-kB, and it was thought to have a role as a feedback inhibitor of NF-kB activation. It is highly expressed in mature immune cells, including DC, peritoneal macrophages, granulocytes and splenic B and T cells [80]. Recent studies revealed that Tim-3 expression on T cells is regulated by the Th1 lineage determining transcription factor T-bet (Anderson et al., 2010), which is translationally regulated by miR-146a. Several studies link dysregulation of expression of several miRNAs to HCV and/or its related immunologic
T cells in neural ganglia infected by HSV express Tim-3. This suggests a role
T cells in HSV infected neural ganglia [76].
T cells increases their apoptosis [77]. Reddy *et*
therapeutic settings.
112 Immune Response Activation
phenomenon [81-86].
interactions decrease the efficiency of CD8+
Furthermore, Gal-9:/Tim-3 interaction on CD8+
**6. MicroRNAs in negative immune regulation**
miRNAs has been described in relation to specific pathologic process.
**6.1. miRNAs in chronic viral infection — Hepatitis C (HCV)**
majority of CD8+
The negative immunomodulators described herein are the targets of researchers attempting to find treatments for chronic diseases. However, the multi-faceted role of many of these molecules makes it difficult to assign specific functions and these immunomodulators likely function in a complex and often interconnected molecular system. The delicate balance of "on" signals and "off" signals within the immune response leads to challenges for this area of research, but the recent strides made in negative immunomodulation are promising.
#### **Acknowledgements**
This work was supported by NIH NIAID and an NIH NIDDK grants to JPM. This publication is the result of work supported with resources and the use of facilities at the James H. Quillen Veterans Affairs Medical Center. The contents in this publication do not represent the views of the Department of Veterans Affairs or the United States Government. The authors declare no competing financial interests.
[9] Nakae, S., et al., *TIM-1 and TIM-3 enhancement of Th2 cytokine production by mast cells.*
Negative Immunomodulators – Blunting Immunostimulation and Facilitating Infection
http://dx.doi.org/10.5772/57540
115
[10] Cao, E., et al., *T cell immunoglobulin mucin-3 crystal structure reveals a galectin-9-inde‐*
[11] Su, E.W., J.Y. Lin, and L.P. Kane, *TIM-1 and TIM-3 proteins in immune regulation.* Cyto‐
[12] Zhu, C., et al., *The Tim-3 ligand galectin-9 negatively regulates T helper type 1 immunity.*
[13] Zhang, Y., et al., *Tim-3 regulates pro-and anti-inflammatory cytokine expression in human*
[14] Henson, S.M. and A.N. Akbar, *KLRG1--more than a marker for T cell senescence.* Age
[15] Henson, S.M., et al., *KLRG1 signaling induces defective Akt (ser473) phosphorylation and proliferative dysfunction of highly differentiated CD8+T cells.* Blood, 2009. 113(26): p.
[16] Rosshart, S., et al., *Interaction of KLRG1 with E-cadherin: new functional and structural*
[17] Tessmer, M.S., et al., *KLRG1 binds cadherins and preferentially associates with SHIP-1.* Int
[18] Donahue, A.C. and D.A. Fruman, *PI3K signaling controls cell fate at many points in B lymphocyte development and activation.* Semin Cell Dev Biol, 2004. 15(2): p. 183-97.
[19] Fujimoto, M. and T. Naka, *SOCS1, a Negative Regulator of Cytokine Signals and TLR Re‐*
[20] Yoshimura, A., T. Naka, and M. Kubo, *SOCS proteins, cytokine signalling and immune*
[21] Cassel, S.L. and P.B. Rothman, *Chapter 3: Role of SOCS in allergic and innate immune*
[22] Naka, T., et al., *Negative regulation of cytokine and TLR signalings by SOCS and others.*
[23] Gingras, S., et al., *Re-examination of the role of suppressor of cytokine signaling 1 (SOCS1) in the regulation of toll-like receptor signaling.* J Biol Chem, 2004. 279(52): p. 54702-7.
[24] Prele, C.M., et al., *SOCS1 regulates the IFN but not NFkappaB pathway in TLR-stimulated*
[25] Frazier, A.D., et al., *Programmed death-1 affects suppressor of cytokine signaling-1 expres‐ sion in T cells during hepatitis C infection.* Viral Immunol, 2010. 23(5): p. 487-95.
*human monocytes and macrophages.* J Immunol, 2008. 181(11): p. 8018-26.
*sponses, in Human Liver Diseases.* Gastroenterol Res Pract, 2010. 2010.
*pendent ligand-binding surface.* Immunity, 2007. 26(3): p. 311-21.
*CD14+monocytes.* J Leukoc Biol, 2012. 91(2): p. 189-96.
*insights.* Eur J Immunol, 2008. 38(12): p. 3354-64.
*regulation.* Nat Rev Immunol, 2007. 7(6): p. 454-65.
*responses.* Adv Immunol, 2009. 103: p. 49-76.
Adv Immunol, 2005. 87: p. 61-122.
Blood, 2007. 110(7): p. 2565-8.
kine, 2008. 44(1): p. 9-13.
Nat Immunol, 2005. 6(12): p. 1245-52.
(Dordr), 2009. 31(4): p. 285-91.
Immunol, 2007. 19(4): p. 391-400.
6619-28.
#### **Author details**
Jeddidiah Griffin, Marie Moulton1 , Rabab Elmezayen1 and Jonathan Moorman1,2\*
\*Address all correspondence to: moorman@etsu.edu
1 Department of Internal Medicine, USA
2 James H. Quillen College of Medicine, East Tennessee State University, Johnson City, TN, USA
#### **References**
[9] Nakae, S., et al., *TIM-1 and TIM-3 enhancement of Th2 cytokine production by mast cells.* Blood, 2007. 110(7): p. 2565-8.
of the Department of Veterans Affairs or the United States Government. The authors declare
, Rabab Elmezayen1
2 James H. Quillen College of Medicine, East Tennessee State University, Johnson City, TN,
[1] Ishida, Y., et al., *Induced expression of PD-1, a novel member of the immunoglobulin gene*
[2] Keir, M.E., et al., *PD-1 and its ligands in tolerance and immunity.* Annu Rev Immunol,
[3] Chemnitz, J.M., et al., *SHP-1 and SHP-2 associate with immunoreceptor tyrosine-based switch motif of programmed death 1 upon primary human T cell stimulation, but only recep‐*
[4] Freeman, G.J., et al., *Engagement of the PD-1 immunoinhibitory receptor by a novel B7 family member leads to negative regulation of lymphocyte activation.* J Exp Med, 2000.
[5] Okazaki, T., et al., *PD-1 immunoreceptor inhibits B cell receptor-mediated signaling by re‐ cruiting src homology 2-domain-containing tyrosine phosphatase 2 to phosphotyrosine.* Proc
[6] Sheppard, K.A., et al., *PD-1 inhibits T-cell receptor induced phosphorylation of the ZAP70/ CD3zeta signalosome and downstream signaling to PKCtheta.* FEBS Lett, 2004. 574(1-3): p.
[7] Anderson AC, Lord GM, Dardalhon V, Lee DH, Sabatos-Peyton CA, Glimcher LH, Kuchroo VK. *T-bet, a Th1 transcription factor regulates the expression of Tim-3.* Eur J Im‐
[8] Monney, L., et al., *Th1-specific cell surface protein Tim-3 regulates macrophage activation*
*and severity of an autoimmune disease.* Nature, 2002. 415(6871): p. 536-41.
*superfamily, upon programmed cell death.* EMBO J, 1992. 11(11): p. 3887-95.
*tor ligation prevents T cell activation.* J Immunol, 2004. 173(2): p. 945-54.
Natl Acad Sci U S A, 2001. 98(24): p. 13866-71.
munol. 2010 Mar;40(3):859-66. doi: 10.1002/eji.200939842.
and Jonathan Moorman1,2\*
no competing financial interests.
Jeddidiah Griffin, Marie Moulton1
2008. 26: p. 677-704.
192(7): p. 1027-34.
37-41.
1 Department of Internal Medicine, USA
\*Address all correspondence to: moorman@etsu.edu
**Author details**
114 Immune Response Activation
USA
**References**
[26] Alter, H.J., *HCV natural history: the retrospective and prospective in perspective.* J Hepatol, 2005. 43(4): p. 550-2.
[41] Nakamoto, N., et al., *Functional restoration of HCV-specific CD8 T cells by PD-1 blockade is defined by PD-1 expression and compartmentalization.* Gastroenterology, 2008. 134(7):
Negative Immunomodulators – Blunting Immunostimulation and Facilitating Infection
http://dx.doi.org/10.5772/57540
117
[42] Ma, C.J., et al., *PD-1 negatively regulates interleukin-12 expression by limiting STAT-1 phosphorylation in monocytes/macrophages during chronic hepatitis C virus infection.* Im‐
[43] Zhang, Y., et al., *Cross-talk between programmed death-1 and suppressor of cytokine signal‐ ing-1 in inhibition of IL-12 production by monocytes/macrophages in hepatitis C virus infec‐*
[44] Miyoshi, H., et al., *Hepatitis C virus core protein exerts an inhibitory effect on suppressor of cytokine signaling (SOCS)-1 gene expression.* J Hepatol, 2005. 43(5): p. 757-63.
[45] Yoshida, T., et al., *SOCS1 is a suppressor of liver fibrosis and hepatitis-induced carcinogen‐*
[46] Brand, S., et al., *IL-22-mediated liver cell regeneration is abrogated by SOCS-1/3 overex‐ pression in vitro.* Am J Physiol Gastrointest Liver Physiol, 2007. 292(4): p. G1019-28. [47] Seki, E., et al., *Demonstration of cooperative contribution of MET-and EGFR-mediated STAT3 phosphorylation to liver regeneration by exogenous suppressor of cytokine signalings.*
[48] Bengsch, B., et al., *Analysis of CD127 and KLRG1 expression on hepatitis C virus-specific CD8+T cells reveals the existence of different memory T-cell subsets in the peripheral blood*
[49] Wang, J.M., et al., *KLRG1 negatively regulates natural killer (NK) cell functions through*
[50] McMahan, R.H., et al., *Tim-3 expression on PD-1+HCV-specific human CTLs is associated with viral persistence, and its blockade restores hepatocyte-directed in vitro cytotoxicity.* J
[51] Moorman, J.P., et al., *Tim-3 pathway controls regulatory and effector T cell balance during*
[52] Ji, X.J., et al., *HCV-infected hepatocytes drive CD4+CD25+Foxp3+regulatory T-cell develop‐ ment through the Tim-3/Gal-9 pathway.* Eur J Immunol, 2013. 43(2): p. 458-67.
[53] Zhang, Y., et al., *Tim-3 negatively regulates IL-12 expression by monocytes in HCV infec‐*
[54] Ibegbu, C.C., et al., *Expression of killer cell lectin-like receptor G1 on antigen-specific hu‐ man CD8+T lymphocytes during active, latent, and resolved infection and its relation with*
*Akt pathway in individuals with chronic hepatitis C.* J Virol, 2013.
*hepatitis C virus infection.* J Immunol, 2012. 189(2): p. 755-66.
p. 1927-37, 1937 e1-2.
munology, 2011. 132(3): p. 421-31.
*tion.* J Immunol, 2011. 186(5): p. 3093-103.
*esis.* J Exp Med, 2004. 199(12): p. 1701-7.
J Hepatol, 2008. 48(2): p. 237-45.
*and liver.* J Virol, 2007. 81(2): p. 945-53.
Clin Invest, 2010. 120(12): p. 4546-57.
*tion.* PLoS One, 2011. 6(5): p. e19664.
*CD57.* J Immunol, 2005. 174(10): p. 6088-94.
[41] Nakamoto, N., et al., *Functional restoration of HCV-specific CD8 T cells by PD-1 blockade is defined by PD-1 expression and compartmentalization.* Gastroenterology, 2008. 134(7): p. 1927-37, 1937 e1-2.
[26] Alter, H.J., *HCV natural history: the retrospective and prospective in perspective.* J Hepatol,
[27] Moorman, J.P., M. Joo, and Y.S. Hahn, *Evasion of host immune surveillance by hepatitis C virus: potential roles in viral persistence.* Arch Immunol Ther Exp (Warsz), 2001. 49(3):
[28] Cooper, S., et al., *Analysis of a successful immune response against hepatitis C virus.* Im‐
[29] Lechner, F., et al., *Analysis of successful immune responses in persons infected with hepati‐*
[30] Thimme, R., et al., *Increased expression of the NK cell receptor KLRG1 by virus-specific CD8 T cells during persistent antigen stimulation.* J Virol, 2005. 79(18): p. 12112-6.
[31] Lechmann, M., et al., *Decreased frequency of HCV core-specific peripheral blood mononu‐ clear cells with type 1 cytokine secretion in chronic hepatitis C.* J Hepatol, 1999. 31(6): p.
[32] Rehermann, B., et al., *Quantitative analysis of the peripheral blood cytotoxic T lymphocyte response in patients with chronic hepatitis C virus infection.* J Clin Invest, 1996. 98(6): p.
[33] King, E., et al., *Hepatitis C: the complications of immune dysfunction.* Expert Rev Clin Im‐
[34] Yao, Z.Q., et al., *HCV core/gC1qR interaction arrests T cell cycle progression through stabi‐*
[35] Yao, Z.Q., et al., *Direct binding of hepatitis C virus core to gC1qR on CD4+and CD8+T cells leads to impaired activation of Lck and Akt.* J Virol, 2004. 78(12): p. 6409-19.
[36] Yao, Z.Q., et al., *T cell dysfunction by hepatitis C virus core protein involves PD-1/PDL-1*
[37] Golden-Mason, L., et al., *Upregulation of PD-1 expression on circulating and intrahepatic hepatitis C virus-specific CD8+T cells associated with reversible immune dysfunction.* J Vi‐
[38] Nakamoto, N., et al., *Synergistic reversal of intrahepatic HCV-specific CD8 T cell exhaus‐ tion by combined PD-1/CTLA-4 blockade.* PLoS Pathog, 2009. 5(2): p. e1000313.
[39] Urbani, S., et al., *PD-1 expression in acute hepatitis C virus (HCV) infection is associated*
[40] Urbani, S., et al., *Restoration of HCV-specific T cell functions by PD-1/PD-L1 blockade in HCV infection: effect of viremia levels and antiviral treatment.* J Hepatol, 2008. 48(4): p.
*with HCV-specific CD8 exhaustion.* J Virol, 2006. 80(22): p. 11398-403.
*lization of the cell cycle inhibitor p27Kip1.* Virology, 2003. 314(1): p. 271-82.
2005. 43(4): p. 550-2.
munity, 1999. 10(4): p. 439-49.
munol, 2007. 3(2): p. 145-57.
rol, 2007. 81(17): p. 9249-58.
*tis C virus.* J Exp Med, 2000. 191(9): p. 1499-512.
*signaling.* Viral Immunol, 2007. 20(2): p. 276-87.
p. 189-94.
116 Immune Response Activation
971-8.
1432-40.
548-58.
[55] Bengsch, B., et al., *Coexpression of PD-1, 2B4, CD160 and KLRG1 on exhausted HCV-spe‐ cific CD8+T cells is linked to antigen recognition and T cell differentiation.* PLoS Pathog, 2010. 6(6): p. e1000947.
[69] Mahller, Y.Y., et al., *Molecular analysis of human cancer cells infected by an oncolytic HSV-1 reveals multiple upregulated cellular genes and a role for SOCS1 in virus replication.*
Negative Immunomodulators – Blunting Immunostimulation and Facilitating Infection
http://dx.doi.org/10.5772/57540
119
[70] Yokota, S., et al., *Induction of suppressor of cytokine signaling-3 by herpes simplex virus*
[71] Channappanavar, R., B.S. Twardy, and S. Suvas, *Blocking of PDL-1 interaction enhances primary and secondary CD8 T cell response to herpes simplex virus-1 infection.* PLoS One,
[72] Bryant-Hudson, K.M. and D.J. Carr, *PD-L1-expressing dendritic cells contribute to viral resistance during acute HSV-1 infection.* Clin Dev Immunol, 2012. 2012: p. 924619.
[73] Carr, D.J., P. Harle, and B.M. Gebhardt, *The immune response to ocular herpes simplex*
[74] Wickham, S. and D.J. Carr, *Molecular mimicry versus bystander activation: herpetic stro‐*
[75] Reddy, P.B., et al., *Influence of galectin-9/Tim-3 interaction on herpes simplex virus-1 la‐*
[76] Sehrawat, S., et al., *Role of Tim-3/galectin-9 inhibitory interaction in viral-induced immu‐ nopathology: shifting the balance toward regulators.* J Immunol, 2009. 182(5): p. 3191-201.
[77] Sehrawat, S., et al., *Galectin-9/TIM-3 interaction regulates virus-specific primary and*
[78] Kincaid, R.P. and C.S. Sullivan, *Virus-encoded microRNAs: an overview and a look to the*
[79] Roberts, A.P. and C.L. Jopling, *Targeting viral infection by microRNA inhibition.* Ge‐
[80] Boldin, M.P., et al., *miR-146a is a significant brake on autoimmunity, myeloproliferation,*
[81] Bala, S., et al., *Increased microRNA-155 expression in the serum and peripheral monocytes*
[82] El-Ekiaby, N., et al., *Repressed induction of interferon-related microRNAs miR-146a and miR-155 in peripheral blood mononuclear cells infected with HCV genotype 4.* FEBS Open
[83] Fognani, E., et al., *Role of microRNA profile modifications in hepatitis C virus-related*
[84] Bhanja Chowdhury, J., et al., *Hepatitis C virus infection modulates expression of interferon stimulatory gene IFITM1 by upregulating miR-130A.* J Virol, 2012. 86(18): p. 10221-5.
*memory CD8 T cell response.* PLoS Pathog, 2010. 6(5): p. e1000882.
*virus type 1 infection.* Exp Biol Med (Maywood), 2001. 226(5): p. 353-66.
*mal keratitis.* Autoimmunity, 2004. 37(5): p. 393-7.
*tency.* J Immunol, 2011. 187(11): p. 5745-55.
*future.* PLoS Pathog, 2012. 8(12): p. e1003018.
*and cancer in mice.* J Exp Med, 2011. 208(6): p. 1189-201.
*in chronic HCV infection.* J Transl Med, 2012. 10: p. 151.
*mixed cryoglobulinemia.* PLoS One, 2013. 8(5): p. e62965.
nome Biol, 2010. 11(1): p. 201.
Bio, 2012. 2: p. 179-86.
*type 1 confers efficient viral replication.* Virology, 2005. 338(1): p. 173-81.
Cancer Gene Ther, 2008. 15(11): p. 733-41.
2012. 7(7): p. e39757.
[69] Mahller, Y.Y., et al., *Molecular analysis of human cancer cells infected by an oncolytic HSV-1 reveals multiple upregulated cellular genes and a role for SOCS1 in virus replication.* Cancer Gene Ther, 2008. 15(11): p. 733-41.
[55] Bengsch, B., et al., *Coexpression of PD-1, 2B4, CD160 and KLRG1 on exhausted HCV-spe‐ cific CD8+T cells is linked to antigen recognition and T cell differentiation.* PLoS Pathog,
[56] Ryo, A., et al., *SOCS1 is an inducible host factor during HIV-1 infection and regulates the intracellular trafficking and stability of HIV-1 Gag.* Proc Natl Acad Sci U S A, 2008.
[57] Wang, X., et al., *B7-H1 up-regulation impairs myeloid DC and correlates with disease pro‐*
[58] Day, C.L., et al., *PD-1 expression on HIV-specific T cells is associated with T-cell exhaus‐*
[59] Trautmann, L., et al., *Upregulation of PD-1 expression on HIV-specific CD8+T cells leads*
[60] Muthumani, K., et al., *Human immunodeficiency virus type 1 Nef induces programmed death 1 expression through a p38 mitogen-activated protein kinase-dependent mechanism.* J
[61] Kaufmann, D.E. and B.D. Walker, *PD-1 and CTLA-4 inhibitory cosignaling pathways in HIV infection and the potential for therapeutic intervention.* J Immunol, 2009. 182(10): p.
[62] Freeman, G.J., et al., *Reinvigorating exhausted HIV-specific T cells via PD-1-PD-1 ligand*
[63] Okazaki, T. and T. Honjo, *The PD-1-PD-L pathway in immunological tolerance.* Trends
[64] Jones, R.B., et al., *Tim-3 expression defines a novel population of dysfunctional T cells with highly elevated frequencies in progressive HIV-1 infection.* J Exp Med, 2008. 205(12): p.
[65] Sakhdari, A., et al., *Tim-3 negatively regulates cytotoxicity in exhausted CD8+T cells in*
[66] Vali, B., et al., *HCV-specific T cells in HCV/HIV co-infection show elevated frequencies of dual Tim-3/PD-1 expression that correlate with liver disease progression.* Eur J Immunol,
[67] Elahi, S., et al., *Galectin-9 binding to Tim-3 renders activated human CD4+T cells less sus‐*
[68] Rahman, A.K.M.N., et al., *TIM-3 and its Immunoregulatory Role in HIV Infection.* J Clin
*gression in chronic HIV-1 infection.* Eur J Immunol, 2008. 38(11): p. 3226-36.
*tion and disease progression.* Nature, 2006. 443(7109): p. 350-4.
*to reversible immune dysfunction.* Nat Med, 2006. 12(10): p. 1198-202.
2010. 6(6): p. e1000947.
Virol, 2008. 82(23): p. 11536-44.
Immunol, 2006. 27(4): p. 195-201.
2010. 40(9): p. 2493-505.
Cell Immunol, 2012. S7: p. 007.
*blockade.* J Exp Med, 2006. 203(10): p. 2223-7.
*HIV infection.* PLoS One, 2012. 7(7): p. e40146.
*ceptible to HIV-1 infection.* Blood, 2012. 119(18): p. 4192-204.
105(1): p. 294-9.
118 Immune Response Activation
5891-7.
2763-79.
[85] Gupta, A., et al., *MicroRNAs, hepatitis C virus, and HCV/HIV-1 co-infection: new insights in pathogenesis and therapy.* Viruses, 2012. 4(11): p. 2485-513.
**Chapter 5**
**Macrophages — Masters of Immune Activation,**
Immunostimulation is fundamental to efficient host responses to pathogens; a comprehensive understanding of which will inform the development of the next generation of vaccines. Integral to vaccine development is the mode of delivery to the appropriate host tissue. Mucosal tissue, with its ability to both tolerise and activate host immune responses, represents an ideal site for vaccine delivery. The continued study of such mucosal tissue with respect to its ability to exert immune activation, suppression and in deed deviation, will be of benefit to the design
Mucosal immunity is governed by the interaction of the barrier epithelial cells of the mucosal surface with the underlying immune cells. The most important immune cells, which are responsible for determining immune responsiveness at these surfaces, are those of the innate immune system; antigen presenting cells (APCs) such as dendritic cells (DCs) and macro‐ phages (Mϕs). DCs play an important role in antigen processing and presentation, thus priming antigen-specific effector T cell responses by homing from the challenge site to the local lymph node. Mϕs, on the other hand, stay within the mucosal tissue at the forefront of the challenge site and play a major role in immune fate decision: activation, deviation or suppres‐ sion. Under certain genetic, environmental and immunological conditions, these resident tissue macrophages are rendered dysfunctional and play a pivotal role in establishment and persistence of pathology. Taking into consideration their tissue residence and their potential to prime a wide range of immune responses, this review will focus on immunostimulation in the context of mucosal Mϕ function in both homeostasis and pathology in the oral mucosa and gastrointestinal tract (GIT). In addition, immunostimulation via Mϕ functionality can only fully be appreciated by considering the relative relationship of activation with both immune-
deviation and suppression; such functionality will also be considered in this review.
© 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
**Suppression and Deviation**
Additional information is available at the end of the chapter
Andrew D. Foey
**1. Introduction**
http://dx.doi.org/10.5772/57541
of next-generation vaccines.
### **Macrophages — Masters of Immune Activation, Suppression and Deviation**
Andrew D. Foey
[85] Gupta, A., et al., *MicroRNAs, hepatitis C virus, and HCV/HIV-1 co-infection: new insights*
[86] Thibault, P.A., et al., *MicroRNA-122-dependent and-independent replication of Hepatitis C*
[87] Ouellet, D.L., et al., *Regulation of host gene expression by HIV-1 TAR microRNAs.* Retro‐
[88] Umbach, J.L., et al., *MicroRNAs expressed by herpes simplex virus 1 during latent infec‐*
[89] Bloom, D.C., *HSV LAT and neuronal survival.* Int Rev Immunol, 2004. 23(1-2): p.
[90] Stevens, J.G., et al., *Prominence of the herpes simplex virus latency-associated transcript in trigeminal ganglia from seropositive hum*ans. J Infect Dis, 1988. 158(1): p. 117-23.
[91] Everett, R.D., ICP0, a regulator of herpes simplex virus during lytic and latent infec‐
*Virus in Hep3B human hepatoma cells.* Virology, 2013. 436(1): p. 179-90.
*in pathogenesis and therapy.* Viruses, 2012. 4(11): p. 2485-513.
*tion regulate viral mRNAs.* Nature, 2008. 454(7205): p. 780-3.
virology, 2013. 10(1): p. 86.
tion. Bioessays, 2000. 22(8): p. 761-70.
187-98.
120 Immune Response Activation
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/57541
#### **1. Introduction**
Immunostimulation is fundamental to efficient host responses to pathogens; a comprehensive understanding of which will inform the development of the next generation of vaccines. Integral to vaccine development is the mode of delivery to the appropriate host tissue. Mucosal tissue, with its ability to both tolerise and activate host immune responses, represents an ideal site for vaccine delivery. The continued study of such mucosal tissue with respect to its ability to exert immune activation, suppression and in deed deviation, will be of benefit to the design of next-generation vaccines.
Mucosal immunity is governed by the interaction of the barrier epithelial cells of the mucosal surface with the underlying immune cells. The most important immune cells, which are responsible for determining immune responsiveness at these surfaces, are those of the innate immune system; antigen presenting cells (APCs) such as dendritic cells (DCs) and macro‐ phages (Mϕs). DCs play an important role in antigen processing and presentation, thus priming antigen-specific effector T cell responses by homing from the challenge site to the local lymph node. Mϕs, on the other hand, stay within the mucosal tissue at the forefront of the challenge site and play a major role in immune fate decision: activation, deviation or suppres‐ sion. Under certain genetic, environmental and immunological conditions, these resident tissue macrophages are rendered dysfunctional and play a pivotal role in establishment and persistence of pathology. Taking into consideration their tissue residence and their potential to prime a wide range of immune responses, this review will focus on immunostimulation in the context of mucosal Mϕ function in both homeostasis and pathology in the oral mucosa and gastrointestinal tract (GIT). In addition, immunostimulation via Mϕ functionality can only fully be appreciated by considering the relative relationship of activation with both immunedeviation and suppression; such functionality will also be considered in this review.
© 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
#### **2. Macrophages**
Macrophages have long since been recognised to play a central role in the immune response. They are mononuclear phagocytic cells, which have been demonstrated to be involved in pathogen recognition, clearance, antigen processing and presentation, inflammation and tissue repair as well as pro-and anti-tumoral responses. Tissue Mϕs perform a range of important homeostatic functions and exist in many different tissue-specific forms: microglia in the brain, alveolar Mϕs in the lung, Kupfer cell in the liver, osteoclast in the bone, as well as splenic, intestinal, bone marrow and lymph node-associated subcapsular sinusoidal/ medullary Mϕs. Thus, Mϕs are central to innate immune inflammatory mechanisms and the priming of adaptive responses to both intracellular and extracellular pathogens as well as immune regulation and tumour surveillance; such diversity of effector functionality would suggest mucosal macrophages to be ideal targets for vaccine therapy.
the presence of GM-CSF [14,15] and further polarising signals such as anoxic environments [16], β-chemokines [17] and phorbol myristate acetate, PMA [18,19]. The presence and activation of pro-inflammatory M1s would have to be tightly controlled, as over-activation or dysregulation of M1s may result in uncontrolled pro-inflammatory pathology. The involve‐ ment of such a distinct Mϕ subset has been suggested to be central to the tissue-destructive pathology associated with the chronic inflammatory bowel disease, Crohn's disease [20].
Macrophages — Masters of Immune Activation, Suppression and Deviation
http://dx.doi.org/10.5772/57541
123
In addition to the profile of factors listed in the paragraph above, the general profile of the M1 effector phenotype is IL-10lo IL-12hi IL-23hi [21]. Associated with the M1-polarising stimuli, IFNγ and LPS, are distinct expression and activation profiles of intracellular signalling components. M1 polarisation is transduced via the transcription factors STAT-1 and NFkB, driving the expression of immune mediators characteristic of this subset, and via the activity of SOCS3 [22]. Finally, the M1-specific transcription factor, STAT-1, has been demonstrated to inhibit STAT-6 activation; this factor is required for M2 polarisation [23], thus an M1-associated transcription factor cross-regulates/suppresses M2 polarisation/activation. These observations alone were suggestive that there was a level of plasticity that existed between M1 pro-
M2 macrophages, which are also referred to as alternatively activated Mϕs, were found to exhibit a functionally distinct phenotype to that of M1s, originally via the ability of IL-4 to induce MR expression, followed by another Th2 cytokine, IL-13 [24,25]. M2 functionality was generally described as an anti-inflammatory/immunoregulatory phenotype, which also mediates tissue remodelling and repair, resistance to parasites and tumour promotion, all of which was dependent on the activation/differentiation signals encountered [6,26]. Extensive research has found a wide range of M2 polarising signals to exist: these priming signals include IL4/IL13, IL-10, TGFβ, M-CSF, Vitamin D3, immune complexes and the Th2-derived cytokine, IL-21 [27] and reviewed in [28]. The versatility in polarising signals and effector functionality of M2 Mϕs resulted in the proposal that M2s exist in a variety of subsets: M2a (tissue repair), M2b (B cell IgG production) and M2c (anti-inflammatory/scavenging mechanisms) (reviewed
M2 polarisation is induced by a wealth of exogenous factors, presented above, which drive distinct signalling/transcription factor pathways as well as signature gene expression. There is an ever-increasing understanding of these endogenous molecular pathways initiated by polarising factors such as IL-4. Whereas p65 NFκB subunit expression is associated with M1 polarisation, M2 polarisation has been described to be orchestrated by p50 NFκB [30, 31], whereby p50 NFκB inhibits NFκB-dependent M1 polarisation. In addition, p50 deficient mice exhibit exacerbated M1-driven inflammatory responses with a concomitant suppression of M2-driven responses (allergy and anti-helminth immunity). In addition to p50 NFκB, M2-like differentiation/ polarisation has been demonstrated to be induced by agonists of PPARγ [32], CREB-controlled C/EBPβ [33] and epigenetic regulation of M2 defining gene products in mice (arginase-1, Ym1, FIZZ1 and mannose receptor) via the IL-4- STAT6-dependent histone demethylation of H3 by the histone demethylase, jmjd3 [34, 35]. Finally, signalling components not only positively regulate Mϕ differentiation and activation; the generation of alternatively activated (M2) Mϕs has been reported to be
inflammatory Mϕs and the anti-inflammatory/regulatory M2 Mϕs.
in [7,26]) (refer to figure 1 below).
#### **2.1. Mϕs exhibit distinct functional subsets**
In general, Mϕ functions are conferred by environmental stimuli, pre-programming or the route of differentiation. As such, Mϕs exist in several subsets, which exhibit discrete functional and phenotypic traits. A certain level of confusion exists however, as to whether the Mϕ subset has been generated by the use of distinct activation signals or as a consequence of longer-term culture/differentiation, or is even pre-programmed in the monocytes. There is some evidence to suggest that monocyte pre-programming is of some functional relevance, where there is a level of heterogeneity, and monocytes potentially exist as a classical (CD14++CD16- HLA-DR+ ) subset and a pro-inflammatory subset (CD14+ CD16+ HLA-DR++) [1,2] biased towards high level TNFα production and present in high numbers in inflammatory pathologies [3,4]. In its simplest guise, Mϕ subsets have been described to exist as two functionally distinct forms, in particular associated with Th1 or Th2 adaptive responses: a pro-inflammatory subset (classi‐ cally activated, type I or M1) and an anti-inflammatory/regulatory subset (alternative activa‐ tion, type II or M2) [5,6], although further studies have suggested several subdivisions may exist within the M2 subset (reviewed in [7]).
M1 macrophages, also referred to as classically activated Mϕs, are generally pro-inflammatory Mϕs, which exhibit a functional phenotype that is anti-microbial (mediating resistance to intracellular pathogens), tissue destructive and anti-tumoral [8]. Upon activation, these Mϕs express a plethora of pro-inflammatory cytokines (TNFα, IL-1β, IL-6, IL-12, IL-18, IL-23), chemokines (CXCL1, 2, 3, 5, 8, 9, 10, CCL2, 3, 4, 5, 11, 17 and 22)[9], proteases and ROS/RNS. As a consequence of this wide array of cytokines expressed, M1s play a pivotal role, not only in innate responses, but also adaptive, antigen-specific responses. As M1s secrete a high level of IL-12, they promote Th1 differentiation, hence cell mediated immune mechanisms for the defence against intracellular pathogens [10,11]. In addition, Holscher et al., 2001, described the protective role of IL-12p40; this subunit is also common to IL-23, which is produced by M1s and is important in the differentiation and activation of Th17 cells [12,13]. Mϕs can be polarised to the M1 phenotype in the presence of IFNγ and in combination with inflammatory stimuli such as LPS or TNFα. In addition, this M1 phenotype can also result from differentiation in the presence of GM-CSF [14,15] and further polarising signals such as anoxic environments [16], β-chemokines [17] and phorbol myristate acetate, PMA [18,19]. The presence and activation of pro-inflammatory M1s would have to be tightly controlled, as over-activation or dysregulation of M1s may result in uncontrolled pro-inflammatory pathology. The involve‐ ment of such a distinct Mϕ subset has been suggested to be central to the tissue-destructive pathology associated with the chronic inflammatory bowel disease, Crohn's disease [20].
**2. Macrophages**
122 Immune Response Activation
Macrophages have long since been recognised to play a central role in the immune response. They are mononuclear phagocytic cells, which have been demonstrated to be involved in pathogen recognition, clearance, antigen processing and presentation, inflammation and tissue repair as well as pro-and anti-tumoral responses. Tissue Mϕs perform a range of important homeostatic functions and exist in many different tissue-specific forms: microglia in the brain, alveolar Mϕs in the lung, Kupfer cell in the liver, osteoclast in the bone, as well as splenic, intestinal, bone marrow and lymph node-associated subcapsular sinusoidal/ medullary Mϕs. Thus, Mϕs are central to innate immune inflammatory mechanisms and the priming of adaptive responses to both intracellular and extracellular pathogens as well as immune regulation and tumour surveillance; such diversity of effector functionality would
In general, Mϕ functions are conferred by environmental stimuli, pre-programming or the route of differentiation. As such, Mϕs exist in several subsets, which exhibit discrete functional and phenotypic traits. A certain level of confusion exists however, as to whether the Mϕ subset has been generated by the use of distinct activation signals or as a consequence of longer-term culture/differentiation, or is even pre-programmed in the monocytes. There is some evidence to suggest that monocyte pre-programming is of some functional relevance, where there is a
CD16+
TNFα production and present in high numbers in inflammatory pathologies [3,4]. In its simplest guise, Mϕ subsets have been described to exist as two functionally distinct forms, in particular associated with Th1 or Th2 adaptive responses: a pro-inflammatory subset (classi‐ cally activated, type I or M1) and an anti-inflammatory/regulatory subset (alternative activa‐ tion, type II or M2) [5,6], although further studies have suggested several subdivisions may
M1 macrophages, also referred to as classically activated Mϕs, are generally pro-inflammatory Mϕs, which exhibit a functional phenotype that is anti-microbial (mediating resistance to intracellular pathogens), tissue destructive and anti-tumoral [8]. Upon activation, these Mϕs express a plethora of pro-inflammatory cytokines (TNFα, IL-1β, IL-6, IL-12, IL-18, IL-23), chemokines (CXCL1, 2, 3, 5, 8, 9, 10, CCL2, 3, 4, 5, 11, 17 and 22)[9], proteases and ROS/RNS. As a consequence of this wide array of cytokines expressed, M1s play a pivotal role, not only in innate responses, but also adaptive, antigen-specific responses. As M1s secrete a high level of IL-12, they promote Th1 differentiation, hence cell mediated immune mechanisms for the defence against intracellular pathogens [10,11]. In addition, Holscher et al., 2001, described the protective role of IL-12p40; this subunit is also common to IL-23, which is produced by M1s and is important in the differentiation and activation of Th17 cells [12,13]. Mϕs can be polarised to the M1 phenotype in the presence of IFNγ and in combination with inflammatory stimuli such as LPS or TNFα. In addition, this M1 phenotype can also result from differentiation in
HLA-DR+
HLA-DR++) [1,2] biased towards high level
)
level of heterogeneity, and monocytes potentially exist as a classical (CD14++CD16-
suggest mucosal macrophages to be ideal targets for vaccine therapy.
**2.1. Mϕs exhibit distinct functional subsets**
subset and a pro-inflammatory subset (CD14+
exist within the M2 subset (reviewed in [7]).
In addition to the profile of factors listed in the paragraph above, the general profile of the M1 effector phenotype is IL-10lo IL-12hi IL-23hi [21]. Associated with the M1-polarising stimuli, IFNγ and LPS, are distinct expression and activation profiles of intracellular signalling components. M1 polarisation is transduced via the transcription factors STAT-1 and NFkB, driving the expression of immune mediators characteristic of this subset, and via the activity of SOCS3 [22]. Finally, the M1-specific transcription factor, STAT-1, has been demonstrated to inhibit STAT-6 activation; this factor is required for M2 polarisation [23], thus an M1-associated transcription factor cross-regulates/suppresses M2 polarisation/activation. These observations alone were suggestive that there was a level of plasticity that existed between M1 proinflammatory Mϕs and the anti-inflammatory/regulatory M2 Mϕs.
M2 macrophages, which are also referred to as alternatively activated Mϕs, were found to exhibit a functionally distinct phenotype to that of M1s, originally via the ability of IL-4 to induce MR expression, followed by another Th2 cytokine, IL-13 [24,25]. M2 functionality was generally described as an anti-inflammatory/immunoregulatory phenotype, which also mediates tissue remodelling and repair, resistance to parasites and tumour promotion, all of which was dependent on the activation/differentiation signals encountered [6,26]. Extensive research has found a wide range of M2 polarising signals to exist: these priming signals include IL4/IL13, IL-10, TGFβ, M-CSF, Vitamin D3, immune complexes and the Th2-derived cytokine, IL-21 [27] and reviewed in [28]. The versatility in polarising signals and effector functionality of M2 Mϕs resulted in the proposal that M2s exist in a variety of subsets: M2a (tissue repair), M2b (B cell IgG production) and M2c (anti-inflammatory/scavenging mechanisms) (reviewed in [7,26]) (refer to figure 1 below).
M2 polarisation is induced by a wealth of exogenous factors, presented above, which drive distinct signalling/transcription factor pathways as well as signature gene expression. There is an ever-increasing understanding of these endogenous molecular pathways initiated by polarising factors such as IL-4. Whereas p65 NFκB subunit expression is associated with M1 polarisation, M2 polarisation has been described to be orchestrated by p50 NFκB [30, 31], whereby p50 NFκB inhibits NFκB-dependent M1 polarisation. In addition, p50 deficient mice exhibit exacerbated M1-driven inflammatory responses with a concomitant suppression of M2-driven responses (allergy and anti-helminth immunity). In addition to p50 NFκB, M2-like differentiation/ polarisation has been demonstrated to be induced by agonists of PPARγ [32], CREB-controlled C/EBPβ [33] and epigenetic regulation of M2 defining gene products in mice (arginase-1, Ym1, FIZZ1 and mannose receptor) via the IL-4- STAT6-dependent histone demethylation of H3 by the histone demethylase, jmjd3 [34, 35]. Finally, signalling components not only positively regulate Mϕ differentiation and activation; the generation of alternatively activated (M2) Mϕs has been reported to be
repressed by the negative regulator, src homology 2-containing inositol-5'-phosphatase,
Macrophages — Masters of Immune Activation, Suppression and Deviation
http://dx.doi.org/10.5772/57541
125
Tissue-resident macrophages can thus express a wide variety of functional phenotypes, which is governed by the manner in which they are activated, differentiated or pre-programmed. This results in potentially two distinct opposing types; pro-inflammatory M1 Mϕ and antiinflammatory/regulatory M2 Mϕ. If these subsets represent a terminally differentiated phenotype, then it is probable that Mϕs are incapable of functional plasticity. What is becoming clear however, is that M1 and M2 Mϕs do exhibit a level of plasticity and that tissue Mϕs are more likely represented on an intermediate sliding scale between pro-inflammatory/immune activatory and anti-inflammatory/regulatory functionality. Indeed, this functional plasticity and manipulation of Mϕ subset has been observed in many studies. A study focussing on bone marrow-derived Mϕs, observed that cytokine functional phenotypes could be altered/ reversed in the presence of Th1 or Th2 cytokines and that this plasticity in function was determined by changes to tissue microenvironment and whether these cytokines were encountered before or concomitantly with an activation stimulus [37]. This plasticity or reversal of Mϕ cytokine phenotype has recently been reinforced by the demonstration of a reversible, bi-directional differentiation of pro-inflammatory (M1) and anti-inflammatory (M2) monocyte-derived Mϕs in response to GM-CSF and M-CSF [38]. In addition, the control of these M-CSF and GM-CSF Mϕ and the invading monocyte populations has been demon‐ strated to be controllable via the use of neutralising antibodies to these differentiation factors in both homeostasis and inflammatory pathology of antigen-induced peritonitis and lung inflammation [39]. This study suggested that M-CSF-derived Mϕs played a non-reductant (M2-like regulatory) role whereas GM-CSF Mϕs maintained the inflammatory response (M1 like pro-inflammatory), contributing to prolonged monocyte and neutrophil infiltration. This is suggestive of the possibility of *in vivo* control of Mϕ effector phenotype through reduction
Macrophage plasticity is not only determined by the differentiation factors but also by the local environmental stimulation factors, the intracellular signalling molecules activated and by the effector cytokines expressed/secreted by the macrophages. These environmental factors not only modulate Mϕ functional plasticity, the Mϕs themselves are able to cross-regulate each other thus moving or consolidating Mϕ effector response along this sliding scale of function‐ ality between pro-inflammatory and anti-inflammatory/regulatory responses. These exoge‐ nous regulators include M1 and M2-associated cytokines (IFNγ, IL-12, IL-10, IL-4) which crossregulate each other and immune complexes (ICs) which, through binding of their respective FcγR, can revert activation status or suppress/modulate cytokine expression, such as suppress IL-12 and induce IL-10 to an M2-like phenotype [40,41]. There is a wide range of intracellular signalling molecules, which play distinct roles in this cross-regulation and plasticity between M1 and M2 subsets. Pro-M1 responses can be mediated by SHIP-1 [42], p65 NFκB and IRF5 [43]; all of which, suppress M2s. On the flip-side, pro-M2 responses have been demonstrated to be mediated by tpl-2, p50 NFκB [31], STAT-6 [34], C/EBPβ and c-maf [33]; again, these
SHIP, a potent negative regulator of the PI3K pathway [36].
**2.2. Functional and phenotypic plasticity of macrophages**
in numbers and manipulation of plasticity.
**Figure 1.** Activation and differentiation drive a range of functional macrophage phenotypes Mϕ effector subsets exist as pro-inflammatory M1 and regulatory or anti-inflammatory M2 phenotypes. M1s are elicited through activation by LPS and IFNγ and differentiated by GM-CSF, whereas M2s can be subdivided into M2a, M2b and M2c according to stimulation/differentiation and functional phenotype. Functional phenotype is characterised by a series of molecular markers listed in the incorporated table under the categories of (1.) signalling molecules, (2.) cytokine expression pro‐ file, (3.) chemokine profile, (4.) scavenger receptor expression and (5.) tryptophan metabolism (inducible nitric oxide synthase, iNOS, and arginase, Arg) and the poorly defined effector molecules, FIZZ-1 and Ym-1. The overall function of these Mϕ subsets is summarised in the lower segment of the table, ranging from pro-inflammatory, anti-tumour re‐ sponses to anti-inflammatory, regulatory and tissue reparative responses. This figure is adapted from [28] and infor‐ mation presented in [7, 9, 23, 26, 29].
repressed by the negative regulator, src homology 2-containing inositol-5'-phosphatase, SHIP, a potent negative regulator of the PI3K pathway [36].
#### **2.2. Functional and phenotypic plasticity of macrophages**
Tissue-resident macrophages can thus express a wide variety of functional phenotypes, which is governed by the manner in which they are activated, differentiated or pre-programmed. This results in potentially two distinct opposing types; pro-inflammatory M1 Mϕ and antiinflammatory/regulatory M2 Mϕ. If these subsets represent a terminally differentiated phenotype, then it is probable that Mϕs are incapable of functional plasticity. What is becoming clear however, is that M1 and M2 Mϕs do exhibit a level of plasticity and that tissue Mϕs are more likely represented on an intermediate sliding scale between pro-inflammatory/immune activatory and anti-inflammatory/regulatory functionality. Indeed, this functional plasticity and manipulation of Mϕ subset has been observed in many studies. A study focussing on bone marrow-derived Mϕs, observed that cytokine functional phenotypes could be altered/ reversed in the presence of Th1 or Th2 cytokines and that this plasticity in function was determined by changes to tissue microenvironment and whether these cytokines were encountered before or concomitantly with an activation stimulus [37]. This plasticity or reversal of Mϕ cytokine phenotype has recently been reinforced by the demonstration of a reversible, bi-directional differentiation of pro-inflammatory (M1) and anti-inflammatory (M2) monocyte-derived Mϕs in response to GM-CSF and M-CSF [38]. In addition, the control of these M-CSF and GM-CSF Mϕ and the invading monocyte populations has been demon‐ strated to be controllable via the use of neutralising antibodies to these differentiation factors in both homeostasis and inflammatory pathology of antigen-induced peritonitis and lung inflammation [39]. This study suggested that M-CSF-derived Mϕs played a non-reductant (M2-like regulatory) role whereas GM-CSF Mϕs maintained the inflammatory response (M1 like pro-inflammatory), contributing to prolonged monocyte and neutrophil infiltration. This is suggestive of the possibility of *in vivo* control of Mϕ effector phenotype through reduction in numbers and manipulation of plasticity.
Macrophage plasticity is not only determined by the differentiation factors but also by the local environmental stimulation factors, the intracellular signalling molecules activated and by the effector cytokines expressed/secreted by the macrophages. These environmental factors not only modulate Mϕ functional plasticity, the Mϕs themselves are able to cross-regulate each other thus moving or consolidating Mϕ effector response along this sliding scale of function‐ ality between pro-inflammatory and anti-inflammatory/regulatory responses. These exoge‐ nous regulators include M1 and M2-associated cytokines (IFNγ, IL-12, IL-10, IL-4) which crossregulate each other and immune complexes (ICs) which, through binding of their respective FcγR, can revert activation status or suppress/modulate cytokine expression, such as suppress IL-12 and induce IL-10 to an M2-like phenotype [40,41]. There is a wide range of intracellular signalling molecules, which play distinct roles in this cross-regulation and plasticity between M1 and M2 subsets. Pro-M1 responses can be mediated by SHIP-1 [42], p65 NFκB and IRF5 [43]; all of which, suppress M2s. On the flip-side, pro-M2 responses have been demonstrated to be mediated by tpl-2, p50 NFκB [31], STAT-6 [34], C/EBPβ and c-maf [33]; again, these
**Figure 1.** Activation and differentiation drive a range of functional macrophage phenotypes Mϕ effector subsets exist as pro-inflammatory M1 and regulatory or anti-inflammatory M2 phenotypes. M1s are elicited through activation by LPS and IFNγ and differentiated by GM-CSF, whereas M2s can be subdivided into M2a, M2b and M2c according to stimulation/differentiation and functional phenotype. Functional phenotype is characterised by a series of molecular markers listed in the incorporated table under the categories of (1.) signalling molecules, (2.) cytokine expression pro‐ file, (3.) chemokine profile, (4.) scavenger receptor expression and (5.) tryptophan metabolism (inducible nitric oxide synthase, iNOS, and arginase, Arg) and the poorly defined effector molecules, FIZZ-1 and Ym-1. The overall function of these Mϕ subsets is summarised in the lower segment of the table, ranging from pro-inflammatory, anti-tumour re‐ sponses to anti-inflammatory, regulatory and tissue reparative responses. This figure is adapted from [28] and infor‐
mation presented in [7, 9, 23, 26, 29].
124 Immune Response Activation
molecules suppress the development of the other canonical Mϕ subset, M1s (reviewed in [44]). Indeed, the importance of the NFkB signalling pathway has been further consolidated by a study demonstrating that IKKβ played an anti-inflammatory role by suppressing classical Mϕ activation/M1 Mϕ phenotype via negative cross-talk with STAT-1, which is involved in the expression of IL-12, iNOS and MHCII [45] and via the maintenance of the IL-10hi/IL12lo functional phenotype [46].
for a wide repertoire of gene expression and immune functionality. Indeed, activation of TLR4 by LPS has been observed to activate ERK-1/2, JNK and p38 MAPKs as well as NFκB and IRF3; these signalling pathways give rise to a multitude of immune mediators which include TNFα, IL-1β, IL-6, IL-8, IL-12, iNOS, IL-10, MHC II. Macrophage activation can thus exhibit a wide range of functions, which include pro-inflammatory, anti-microbial responses, anti-inflam‐ matory/regulatory responses and priming of adaptive immune responses through their
Macrophages — Masters of Immune Activation, Suppression and Deviation
http://dx.doi.org/10.5772/57541
127
Taking into consideration the previous sections, LPS and other PAMPs may be able to induce a wide variety of immune mediators; the expression of which will also be determined by other activatory stimuli, differentiation factors and both exogenous and endogenous suppressive molecules. TLR signals can be negatively regulated by a multitude of molecules (reviewed in [48]) which include; Myd88s, IRAK-M, IRF-4 [49,50], TREM-2, ST2 [51], Tollip, TRIAD3A, SOCS-1, SOCS-3, SHP1, SHP2, SIGIRR [52]. In addition to endogenous suppressors of TLR activation signals, it is becoming evident that TLRs can cross-regulate each other, resulting in suppression of distinct TLR signalling pathway or endotoxin tolerisation of Mϕs through homo-or hetero-tolerisation of distinct TLRs and their ensuing effector functionality (reviewed in [53]). Such mechanisms of tolerisation have been demonstrated to be differentially utilised, dependent on Mϕ subset as well as activation signal [54]. Such complex, inter-twined signal‐ ling pathways enable a level of innate "specificity" of anti-microbial activation response as well as a way in which these activatory signals are fine-tuned, suppressed or modulated.
**NOD-like receptors (NLRs);** due to their intracellular expression, originally thought to specifically respond to intracellular pathogens and their PAMPs, it has become apparent that NLRs are integral to the innate detection of both extracellular and intracellular-resident pathogens. The superfamily of NLRs consists of NOD1 and NOD2 which are intracellular receptors for peptidoglycan (D-γ-glutamyl-meso-DAP and muramyl dipeptide, respectively) [55]. NOD1 and NOD2 activate NFκB through the recruitment and activation of RIP2 and resulting activation of the IκB kinase complex [56]. As a result, NOD-signalling can induce pro-inflammatory, anti-microbial defences via the expression of a wide array of NF-κB-
In addition to the NFκB-activating NODs, the NLR family is also made up of a collection of receptors involved in the activation of caspases [58], specifically caspases 1 and 5 that are integral for the inflammasome construction, activation and processing of the pro-cytokines pro-IL-1β and pro-IL-18 into the mature and secretable IL-1β and IL-18. The inflammasome is activated by extracellular ATP and hypotonic stress, both representative of endogenous danger signals or DAMPs. ASC is an adaptor protein which recruits caspase-1 into an inflammasome activation complex; knock-out mice have demonstrated ASC to be essential for LPS-induced activation of caspase-1, where these mice failed to activate pro-caspase-1 or produce IL-1β and IL-18, even in the event of stimulation by both LPS and ATP (activation of the P2X7 receptor) [59]. This requirement for inflammasome-mediated caspase-1 activation, hence IL-1β matu‐ ration and secretion, in response to LPS (a TLR4 ligand), emphasises a collaborative role for TLR and NLR pathways in the production of IL-1β and the ensuing inflammatory response. Other NLR family members, NALP1, NALP3 and Ipaf, have all been demonstrated to be
capacity to process and present antigen.
dependent gene products (reviewed in [57]).
Considering Mϕ subset phenotype, its environmental differentiation factors, activatory stimuli and effector cytokine profile; macrophage responses can be modulated by a wide range of factors which includes both positive and negative regulators. As such, the potential involvement of macrophages as an immunotherapeutic target for several pathologies can only be appreciated by a full understanding of both activatory and regulatory signals: immunosti‐ mulation, immunosuppression and immune deviation.
#### **3. Stimulation and suppression of Mϕ effector responses**
Macrophages exhibit a wide variety of receptors, which transduce a multitude of immunosti‐ mulatory signals. These activatory receptors are centred on responsiveness to conserved broad-spectrum molecular patterns associated with immune responsiveness to pathogens or endogenous, self-associated, danger signals. Pattern recognition receptors (PRRs) which stimulate or co-activate immune responses include the toll-like receptors (TLRs), NOD-like receptors (NLRs), C-type lectin receptors (CLRs) and scavenger receptors (SRs). In addition to these PRRs, Mϕs can also be activated by immune complex recognition via FcRs and cytokine stimulation through their cognate receptors.
**Toll-like receptors** respond to a wide repertoire of PAMPs, expressed on bacterial, viral, fungal and parasitic pathogens, and DAMPs; generally activating innate inflammatory-or anti-viral responses through the involvement of discriminatory utilisation of receptors, adaptor proteins and either NFκB-or IRF-dependent signalling pathways [47]. Indeed, the utilisation of such signalling pathways, in Mϕ polarisation, make it impossible to totally devolve consideration of polarisation, activation, cross-regulation and suppression. TLRs are expressed according to the localisation of pathogens; TLR2 (CD282), TLR1 (CD281), TLR6 (CD286), TLR4 (CD284) and TLR5 (CD285) being expressed on extracellular membranes, recognise extracellular pathogens, whereas TLR3 (CD283), TLR7 (CD287), TLR8 (CD288), and TLR9 (CD289) are expressed intracellularly, associated with endosomal membranes and recognition of intracellular resident pathogens. The extracellular facing TLRs were found to recognise a range of conserved PAMPs expressed by bacteria, including lipopeptides, LPS and flagellin, whereas TLRs 3, 7, 8 and 9 recognise dsRNA, dsDNA and hypomethylated CpG DNA motifs, respectively, expressed by viruses and bacteria.
TLRs activate a range of important signalling pathways and transcription factors, where their PAMPs activate MAPKs (ERK-1/2, p38 MAPK and JNK), which enhance gene expression and, in the case of p38, mRNA stability via AU-rich repeats present in the 3'UTRs. Activation of MAPK-dependent transcription factors along with other transcription factors are responsible for a wide repertoire of gene expression and immune functionality. Indeed, activation of TLR4 by LPS has been observed to activate ERK-1/2, JNK and p38 MAPKs as well as NFκB and IRF3; these signalling pathways give rise to a multitude of immune mediators which include TNFα, IL-1β, IL-6, IL-8, IL-12, iNOS, IL-10, MHC II. Macrophage activation can thus exhibit a wide range of functions, which include pro-inflammatory, anti-microbial responses, anti-inflam‐ matory/regulatory responses and priming of adaptive immune responses through their capacity to process and present antigen.
molecules suppress the development of the other canonical Mϕ subset, M1s (reviewed in [44]). Indeed, the importance of the NFkB signalling pathway has been further consolidated by a study demonstrating that IKKβ played an anti-inflammatory role by suppressing classical Mϕ activation/M1 Mϕ phenotype via negative cross-talk with STAT-1, which is involved in the expression of IL-12, iNOS and MHCII [45] and via the maintenance of the IL-10hi/IL12lo
Considering Mϕ subset phenotype, its environmental differentiation factors, activatory stimuli and effector cytokine profile; macrophage responses can be modulated by a wide range of factors which includes both positive and negative regulators. As such, the potential involvement of macrophages as an immunotherapeutic target for several pathologies can only be appreciated by a full understanding of both activatory and regulatory signals: immunosti‐
Macrophages exhibit a wide variety of receptors, which transduce a multitude of immunosti‐ mulatory signals. These activatory receptors are centred on responsiveness to conserved broad-spectrum molecular patterns associated with immune responsiveness to pathogens or endogenous, self-associated, danger signals. Pattern recognition receptors (PRRs) which stimulate or co-activate immune responses include the toll-like receptors (TLRs), NOD-like receptors (NLRs), C-type lectin receptors (CLRs) and scavenger receptors (SRs). In addition to these PRRs, Mϕs can also be activated by immune complex recognition via FcRs and cytokine
**Toll-like receptors** respond to a wide repertoire of PAMPs, expressed on bacterial, viral, fungal and parasitic pathogens, and DAMPs; generally activating innate inflammatory-or anti-viral responses through the involvement of discriminatory utilisation of receptors, adaptor proteins and either NFκB-or IRF-dependent signalling pathways [47]. Indeed, the utilisation of such signalling pathways, in Mϕ polarisation, make it impossible to totally devolve consideration of polarisation, activation, cross-regulation and suppression. TLRs are expressed according to the localisation of pathogens; TLR2 (CD282), TLR1 (CD281), TLR6 (CD286), TLR4 (CD284) and TLR5 (CD285) being expressed on extracellular membranes, recognise extracellular pathogens, whereas TLR3 (CD283), TLR7 (CD287), TLR8 (CD288), and TLR9 (CD289) are expressed intracellularly, associated with endosomal membranes and recognition of intracellular resident pathogens. The extracellular facing TLRs were found to recognise a range of conserved PAMPs expressed by bacteria, including lipopeptides, LPS and flagellin, whereas TLRs 3, 7, 8 and 9 recognise dsRNA, dsDNA and hypomethylated CpG DNA motifs, respectively,
TLRs activate a range of important signalling pathways and transcription factors, where their PAMPs activate MAPKs (ERK-1/2, p38 MAPK and JNK), which enhance gene expression and, in the case of p38, mRNA stability via AU-rich repeats present in the 3'UTRs. Activation of MAPK-dependent transcription factors along with other transcription factors are responsible
functional phenotype [46].
126 Immune Response Activation
mulation, immunosuppression and immune deviation.
stimulation through their cognate receptors.
expressed by viruses and bacteria.
**3. Stimulation and suppression of Mϕ effector responses**
Taking into consideration the previous sections, LPS and other PAMPs may be able to induce a wide variety of immune mediators; the expression of which will also be determined by other activatory stimuli, differentiation factors and both exogenous and endogenous suppressive molecules. TLR signals can be negatively regulated by a multitude of molecules (reviewed in [48]) which include; Myd88s, IRAK-M, IRF-4 [49,50], TREM-2, ST2 [51], Tollip, TRIAD3A, SOCS-1, SOCS-3, SHP1, SHP2, SIGIRR [52]. In addition to endogenous suppressors of TLR activation signals, it is becoming evident that TLRs can cross-regulate each other, resulting in suppression of distinct TLR signalling pathway or endotoxin tolerisation of Mϕs through homo-or hetero-tolerisation of distinct TLRs and their ensuing effector functionality (reviewed in [53]). Such mechanisms of tolerisation have been demonstrated to be differentially utilised, dependent on Mϕ subset as well as activation signal [54]. Such complex, inter-twined signal‐ ling pathways enable a level of innate "specificity" of anti-microbial activation response as well as a way in which these activatory signals are fine-tuned, suppressed or modulated.
**NOD-like receptors (NLRs);** due to their intracellular expression, originally thought to specifically respond to intracellular pathogens and their PAMPs, it has become apparent that NLRs are integral to the innate detection of both extracellular and intracellular-resident pathogens. The superfamily of NLRs consists of NOD1 and NOD2 which are intracellular receptors for peptidoglycan (D-γ-glutamyl-meso-DAP and muramyl dipeptide, respectively) [55]. NOD1 and NOD2 activate NFκB through the recruitment and activation of RIP2 and resulting activation of the IκB kinase complex [56]. As a result, NOD-signalling can induce pro-inflammatory, anti-microbial defences via the expression of a wide array of NF-κBdependent gene products (reviewed in [57]).
In addition to the NFκB-activating NODs, the NLR family is also made up of a collection of receptors involved in the activation of caspases [58], specifically caspases 1 and 5 that are integral for the inflammasome construction, activation and processing of the pro-cytokines pro-IL-1β and pro-IL-18 into the mature and secretable IL-1β and IL-18. The inflammasome is activated by extracellular ATP and hypotonic stress, both representative of endogenous danger signals or DAMPs. ASC is an adaptor protein which recruits caspase-1 into an inflammasome activation complex; knock-out mice have demonstrated ASC to be essential for LPS-induced activation of caspase-1, where these mice failed to activate pro-caspase-1 or produce IL-1β and IL-18, even in the event of stimulation by both LPS and ATP (activation of the P2X7 receptor) [59]. This requirement for inflammasome-mediated caspase-1 activation, hence IL-1β matu‐ ration and secretion, in response to LPS (a TLR4 ligand), emphasises a collaborative role for TLR and NLR pathways in the production of IL-1β and the ensuing inflammatory response. Other NLR family members, NALP1, NALP3 and Ipaf, have all been demonstrated to be associated with ASC in the inflammasome, playing an integral role in caspase-1-dependent activation of pro-inflammatory cytokines [60]. NLRs have also been reported to suppress both NFκB and caspase activation pathways. With respect to inhibition of NFκB, NALP10 and NALP12 display a suppressive activity, whereas NALP7 and NALP10 are suppressive for caspase activation (reviewed in [61]). Again, it is imperative to the direction of Mϕ functional responses that the expression profile of NLRs be characterised, along with potential crosscommunicating PRRs; this total profile gives scientists an insight of potential control of Mϕ functionality as being activatory, suppressive or even being able to deviate/modify the cellular response.
SR-E (LOX-1) as well as other SR-A-like class members (SR-CLI – a potential CLR on the basis of C-type lectin domain expression), SR-D (macrosialin). In general, these scavenger receptors recognise a wide range of PAMPs including LTA, LPS, bacterial DNA, whole bacteria, sugar moieties and DAMPs, thus clearance of dead and dying host tissue, as well as pathogenic material (reviewed in [70]). The majority of the signalling scavenger receptors (Dectin-1,
Macrophages — Masters of Immune Activation, Suppression and Deviation
http://dx.doi.org/10.5772/57541
129
CLRs play a major role in recognition and activation of anti-microbial immunity against a diverse range of pathogens, which include fungi, bacteria, viruses and parasites. They are receptors which contain one or more carbohydrate recognition domains (CRDs) which bind a wide variety of carbohydrate ligand moieties. Generally, CLRs can be categorised as either having a built-in endogenous signalling activity or requiring interaction with a partner adaptor protein which possesses signalling activity (reviewed in [71]). CLRs signal via integral ITAMs or partner ITAM-bearing adaptors such as FcγRs. Integral ITAM-bearing CLRs generally signal through a syk/CARD9 pathway which ultimately results in NFκB activation; such signalling CLRs include Dectin-1, Dectin-2 and Mincle. These CLRs are particularly prominent in mounting anti-fungal responses; Dectin-1 responds to β-glucans expressed in the fungal cell wall by initiating a variety of cellular responses which include phagocytosis, respiratory burst, activation of inflammasomes, cytokine production and the polarisation of adaptive responses to cell mediated immunity mediated by Th1 and Th17 cells. In addition, Dectin-1 exhibits crosstalk with TLR2 and other myd88-dependent TLRs, thus again, co-operation of PRRs is critical for tailoring appropriate immune responses (reviewed in [72]). Dectin-2 and Mincle both recognise fungal α-mannans with Dectin-2 being associated with innate recognition and induction of Th17 responses. Of particular importance to the recognition and signal activation in response to fungal infection is the observation that Dectin-2 (signalling) can heterodimerise with Dectin-3, thus CLRs demonstrate a level of co-operativity with each other and fine-tuning
Bacterial sensing utilises Mincle, mannose receptor (MR) and DC-SIGN. MR has been shown to be involved in the recognition and responsiveness to *Klebsiella* and *Streptococcus* sp but so far, has not been recognised to contain a signalling motif. To initiate antibacterial responses, it is likely that the MR co-operates with other PRRs. Mincle however signals through the pathway detailed above, in order to activate NFκB-mediated responses required for innate and consequent cell mediated immune responses to Mycobacterial infection through recog‐ nition and binding of mycobacterial cord factor, trehalose-6,6'-dimycolate, TDM. Mycobacte‐ ria are also recognised and a response initiated through ManLam (mannosylated lipoarabinomannan) binding to DC-SIGN which is expressed by DCs; whether this CLR is expressed by Mϕs is yet to be definitively characterised [71]. DC-SIGN has also been described as being involved in viral recognition by binding to the gp120 envelope protein of HIV-1; this mechanism however may be more representative of a mode of viral transmission rather than anti-viral immunity. CLRs such as CLEC9A (DNGR-1) has been shown to be protective, by facilitating antigen cross-presentation, hence controlling infection of vaccinia virus and herpes simplex virus. Finally, both Dectin-2 and MR are involved in immunity to parasites, whereby Dectin-2 recognises Schistosoma mansonii egg antigen SEA resulting in the activation of the
Dectin-2, Mϕ Mannose Receptor) are CLRs and will be covered in the next section.
of their responses to the fungal pathogen [71].
**RIG-like receptors (RLRs):** are key intracellular innate immune sensors of viral infection. Whereas the predominant viral PRRs in DCs (particularly pDCs) are TLR3, TLR7, TLR8 and TLR9, RLR viral sensing and expression is predominant in Mϕs [62]. Currently, the main RLRs are retinoic acid inducible gene-1 (RIG-1) and melanoma differentiation-associated gene (MDA)5; both of which detect dsRNA [63,64]. Similar to the viral-sensing TLRs, RLRs activate NFκB and IRF3 signalling pathways, resulting in gene expression of pro-inflammatory cytokine genes and anti-viral type I interferons (IFNα and IFNβ). Upon viral infection, these interferons exhibit a positive feedback loop which up-regulates RLR expression. MDA5 has been demonstrated to be important in the detection of picornovirus and polyinosine polycy‐ tidylic acid, poly(I:C)-induced IFNα production. On the other hand, the importance of RLRs to macrophage anti-viral responses is enforced by observations that RNA helicases eg. Lgp2, inhibits RIG1 signalling [65] and that an evasion mechanism utilised by hepatitis C virus is the production of RLR signalling inhibitory proteins [66,67]. This overlap in structure/functional relationships between TLRs and RLRs is suggestive that these receptor types function in tandem to provide ubiquitous anti-viral protection. Additionally, it is unclear as to whether RLRs also interact and mediate inflammasome activity, as a study by Johnston et al in 2005, has shown that a pox virus pyrin domain protein interacts with ASC-1 (NLR family member) and inhibits inflammatory and apoptotic inflammasome activity [68]. This viral escape mechanism, targeting NLR-mediated inflammasome activity, is also suggestive that RLRs cross-communicate with NLRs; thus efficient anti-viral protection is truly a trinity of innate reception between TLRs, NLRs and RLRs (reviewed in [61]).
**Scavenger receptors (SRs) and C-type Lectin receptors (CLRs):** this is a large group of receptors widely expressed on macrophage cells [69] which scavenge pathogenic material and cellular debris, initiating clearing responses through receptor-mediated endocytosis, phago‐ cytosis or signals inflammatory responses. Responses elicited are dependent on patterns being recognised and combinations of PRRs being employed. The CLRs, being lectin receptors are generally sugar receptors, whereas SRs recognise and respond to a wider range of molecular patterns; however, there is a level of overlap, hence confusion between receptors of these families. Scavenger receptors are a broad family of PRRs, which involve both opsoninmediated phagocytosis and opsonin-independent PRRs; the latter even including TLRs, lectins, complement receptors (CRs), CD14, CD36, scavengers of ACAMPs (see later section). Scavenger receptors include the SR-A class of receptors (SR-AI, SR-AII, MARCO), SR-B (CD36, SR-BI and Croquemort, CD163 – haemoglobin scavenger receptor), SR-C class (dSR-CI) and SR-E (LOX-1) as well as other SR-A-like class members (SR-CLI – a potential CLR on the basis of C-type lectin domain expression), SR-D (macrosialin). In general, these scavenger receptors recognise a wide range of PAMPs including LTA, LPS, bacterial DNA, whole bacteria, sugar moieties and DAMPs, thus clearance of dead and dying host tissue, as well as pathogenic material (reviewed in [70]). The majority of the signalling scavenger receptors (Dectin-1, Dectin-2, Mϕ Mannose Receptor) are CLRs and will be covered in the next section.
associated with ASC in the inflammasome, playing an integral role in caspase-1-dependent activation of pro-inflammatory cytokines [60]. NLRs have also been reported to suppress both NFκB and caspase activation pathways. With respect to inhibition of NFκB, NALP10 and NALP12 display a suppressive activity, whereas NALP7 and NALP10 are suppressive for caspase activation (reviewed in [61]). Again, it is imperative to the direction of Mϕ functional responses that the expression profile of NLRs be characterised, along with potential crosscommunicating PRRs; this total profile gives scientists an insight of potential control of Mϕ functionality as being activatory, suppressive or even being able to deviate/modify the cellular
**RIG-like receptors (RLRs):** are key intracellular innate immune sensors of viral infection. Whereas the predominant viral PRRs in DCs (particularly pDCs) are TLR3, TLR7, TLR8 and TLR9, RLR viral sensing and expression is predominant in Mϕs [62]. Currently, the main RLRs are retinoic acid inducible gene-1 (RIG-1) and melanoma differentiation-associated gene (MDA)5; both of which detect dsRNA [63,64]. Similar to the viral-sensing TLRs, RLRs activate NFκB and IRF3 signalling pathways, resulting in gene expression of pro-inflammatory cytokine genes and anti-viral type I interferons (IFNα and IFNβ). Upon viral infection, these interferons exhibit a positive feedback loop which up-regulates RLR expression. MDA5 has been demonstrated to be important in the detection of picornovirus and polyinosine polycy‐ tidylic acid, poly(I:C)-induced IFNα production. On the other hand, the importance of RLRs to macrophage anti-viral responses is enforced by observations that RNA helicases eg. Lgp2, inhibits RIG1 signalling [65] and that an evasion mechanism utilised by hepatitis C virus is the production of RLR signalling inhibitory proteins [66,67]. This overlap in structure/functional relationships between TLRs and RLRs is suggestive that these receptor types function in tandem to provide ubiquitous anti-viral protection. Additionally, it is unclear as to whether RLRs also interact and mediate inflammasome activity, as a study by Johnston et al in 2005, has shown that a pox virus pyrin domain protein interacts with ASC-1 (NLR family member) and inhibits inflammatory and apoptotic inflammasome activity [68]. This viral escape mechanism, targeting NLR-mediated inflammasome activity, is also suggestive that RLRs cross-communicate with NLRs; thus efficient anti-viral protection is truly a trinity of innate
**Scavenger receptors (SRs) and C-type Lectin receptors (CLRs):** this is a large group of receptors widely expressed on macrophage cells [69] which scavenge pathogenic material and cellular debris, initiating clearing responses through receptor-mediated endocytosis, phago‐ cytosis or signals inflammatory responses. Responses elicited are dependent on patterns being recognised and combinations of PRRs being employed. The CLRs, being lectin receptors are generally sugar receptors, whereas SRs recognise and respond to a wider range of molecular patterns; however, there is a level of overlap, hence confusion between receptors of these families. Scavenger receptors are a broad family of PRRs, which involve both opsoninmediated phagocytosis and opsonin-independent PRRs; the latter even including TLRs, lectins, complement receptors (CRs), CD14, CD36, scavengers of ACAMPs (see later section). Scavenger receptors include the SR-A class of receptors (SR-AI, SR-AII, MARCO), SR-B (CD36, SR-BI and Croquemort, CD163 – haemoglobin scavenger receptor), SR-C class (dSR-CI) and
reception between TLRs, NLRs and RLRs (reviewed in [61]).
response.
128 Immune Response Activation
CLRs play a major role in recognition and activation of anti-microbial immunity against a diverse range of pathogens, which include fungi, bacteria, viruses and parasites. They are receptors which contain one or more carbohydrate recognition domains (CRDs) which bind a wide variety of carbohydrate ligand moieties. Generally, CLRs can be categorised as either having a built-in endogenous signalling activity or requiring interaction with a partner adaptor protein which possesses signalling activity (reviewed in [71]). CLRs signal via integral ITAMs or partner ITAM-bearing adaptors such as FcγRs. Integral ITAM-bearing CLRs generally signal through a syk/CARD9 pathway which ultimately results in NFκB activation; such signalling CLRs include Dectin-1, Dectin-2 and Mincle. These CLRs are particularly prominent in mounting anti-fungal responses; Dectin-1 responds to β-glucans expressed in the fungal cell wall by initiating a variety of cellular responses which include phagocytosis, respiratory burst, activation of inflammasomes, cytokine production and the polarisation of adaptive responses to cell mediated immunity mediated by Th1 and Th17 cells. In addition, Dectin-1 exhibits crosstalk with TLR2 and other myd88-dependent TLRs, thus again, co-operation of PRRs is critical for tailoring appropriate immune responses (reviewed in [72]). Dectin-2 and Mincle both recognise fungal α-mannans with Dectin-2 being associated with innate recognition and induction of Th17 responses. Of particular importance to the recognition and signal activation in response to fungal infection is the observation that Dectin-2 (signalling) can heterodimerise with Dectin-3, thus CLRs demonstrate a level of co-operativity with each other and fine-tuning of their responses to the fungal pathogen [71].
Bacterial sensing utilises Mincle, mannose receptor (MR) and DC-SIGN. MR has been shown to be involved in the recognition and responsiveness to *Klebsiella* and *Streptococcus* sp but so far, has not been recognised to contain a signalling motif. To initiate antibacterial responses, it is likely that the MR co-operates with other PRRs. Mincle however signals through the pathway detailed above, in order to activate NFκB-mediated responses required for innate and consequent cell mediated immune responses to Mycobacterial infection through recog‐ nition and binding of mycobacterial cord factor, trehalose-6,6'-dimycolate, TDM. Mycobacte‐ ria are also recognised and a response initiated through ManLam (mannosylated lipoarabinomannan) binding to DC-SIGN which is expressed by DCs; whether this CLR is expressed by Mϕs is yet to be definitively characterised [71]. DC-SIGN has also been described as being involved in viral recognition by binding to the gp120 envelope protein of HIV-1; this mechanism however may be more representative of a mode of viral transmission rather than anti-viral immunity. CLRs such as CLEC9A (DNGR-1) has been shown to be protective, by facilitating antigen cross-presentation, hence controlling infection of vaccinia virus and herpes simplex virus. Finally, both Dectin-2 and MR are involved in immunity to parasites, whereby Dectin-2 recognises Schistosoma mansonii egg antigen SEA resulting in the activation of the NLRP3 inflammasome and alleviating Th2 pathology. Of particular interest is the fact that Dectin-2 can also promote Th2responsiveness, as opposed to suppressing it, when responding to the common allergen of house dust mite [71]. MR drives protective Th2 responses during S.mansonii infection whereas enhances uptake of infectious circariae. These last few observa‐ tions with respect to Dectin-2 and MR and their ability to drive Th2 responses would appear to conform to the previously observed study of Mills et al, whereby M2 macrophages were associated with Th2 immunity [5]; indeed, MR at least, has been assigned as an M2 marker. On the flip-side, Dectin-1 has been associated with initiation of Th1 and Th17 anti-fungal responses; it would appear that any potential activation and modulation of specific immune responses would have to consider the role of pathogen associated carbohydrate expression and their respective CLRs expressed on host macrophages.
**Fc**γ**Rs:** are a family of receptors which bind the Fc region of IgG/IgG immune complexes and are responsible for mediating both activatory and suppressive responses by transducing their ligand-binding signal through either an endogenous cytoplasmic ITAM or ITIM motif, respectively. Three classes of FcγRs are expressed on immune cells such as Mϕs, these include the high affinity FcγRI (CD64) which binds monomeric IgG and the low affinity receptors FcγRII (CD32) and FcγRIII (CD16) that bind IgG immune complexes. Activatory receptors which possess cytoplasmic ITAM sequences include FcγRIIIa, FcγRI and FcγRIIa whereas the CD32 subtype, FcγRIIb is suppressive, containing a cytoplasmic ITIM sequence (reviewed in [79]). Thus recognition and clearance of IgG immune complexes may induce either an activatory or inhibitory signal which is dependent on expression profile of these FcRs and potential cross-talk with other PRRs. Ligation of FcγRIIIa induces the expression and secretion of Mϕ-derived TNFα, the mechanism of production and activation of which has been found to be associated with pathological inflammation [80]. Ligation of FcγRI however, was found to reverse Mϕ pro-inflammatory responses [81]. Indeed, taking into consideration the format of presentation of IgG, the immune response cell and their FcγR expression profile, Mϕs can be activated or inhibited with respect to inflammatory responses; differential FcγR-driven induction and suppression in collagen-induced arthritis model have resulted in this being considered as a realistic therapeutic regimen in the treatment of RA [82]. The final thing to consider with respect to FcγR-induced responses in Mϕs is the fact that these responses may differ between M1 and M2 subsets and that these responses may be deviated by manipulation of FcγR signalling or indeed exhibit suppression of inflammatory responses or manipulation of functional plasticity. Manipulation of these Mϕ receptors would represent a realistic option when considering their association with alternative activation, M2-like anti-inflammatory
Macrophages — Masters of Immune Activation, Suppression and Deviation
http://dx.doi.org/10.5772/57541
131
**Other macrophage suppressive receptors:** Two further immunosuppressive molecules that control Mϕ functionality which are worthy of consideration are CD200R and SIRPα (CD172a); both members of the immunoglobulin superfamily (IgSF). CD172a interacts with a broadly expressed ligand, CD47, which results in the suppression of Mϕs through interaction with the tyrosine phosphatases, SHP1 and SHP2 [83] and reviewed in [84]). In addition, this interaction is made more complex by the observation of bi-directionality of signalling, where CD47 ligation has been found to selectively suppress monocyte IL-12 production, whereas TNFα, IL-1 and IL-6 was not suppressed. In contrast, the immunosup‐ pressive cytokines, TGFβ and IL-10, were not altered by CD47 ligation; suppression of IL-12 was independent of TGFβ and IL-10 [85]. Thus, molecules such as CD172a and CD47 represent an attractive target couple for multi-mechanistic immune suppression. CD200R ligation, on the other hand, transduces a simpler immunological signal. CD200 (OX-2) is a cell surface glycoprotein containing Ig-like domains that interacts with CD200R expressed on Mϕs [86]. The human homologue of CD200R was characterised by Wright and collea‐ gues who found that its cytoplasmic domain contained known signalling motifs [87]. The role of CD200R ligation was appreciated to be potentially suppressive where CD200 deficient mice were shown to display increased susceptibility to autoimmune disease models affecting joints. In accordance with this study, CD200Fc fusion protein was found to exhibit beneficial immunomodulation of arthritis and allograft rejection [88,89]. Indeed,
Mϕs.
**Siglecs:** sialic acid-binding Ig-like lectins are members of the immunoglobulin superfamily involved in the recognition and reception of sialylated glycoconjugates, involved in endocy‐ tosis, positive and negative activation signals. Sialic acids are a family of nine-carbon sugars which are derivatives of neuraminic acid or keto deoxynonulosonic acid and appearing in different glycosidic linkages, typically at the exposed, non-reducing ends of oligosaccharide chains attached to proteins and lipids. It is thought that sialic acid residues either, mask subterminal sugars, prevent non-specific cell-cell interactions or act as ligands for modulating selective cell-cell interactions [73]. Currently, there are 13 Siglec family members expressed in humans and classified into two distinct subsets; siglecs 1 (CD169), 2 (CD22) and 4 are activatory receptors whereas siglecs 3 (CD33) and 5 (CD170) to 11, 14 and 16 (CD33-related siglecs) exhibit suppressive activity through two conserved ITIMs (immunereceptor tyrosine-based inhibition motif) in the cytoplasmic domain which are responsible for the recruitment of the signal suppressive phosphatase enzymes SHIP (lipid phosphatise), SHP-1 and SHP-2 (protein tyrosine phosphatises), which result in raised activation thresholds and antagonism of ITAMdependent activation. More recently, tyrosine phosphorylation of CD33 and siglec-7 can recruit SOCS3 which leads to ubiquitination and proteosomal degredation of the Siglec; regulation of the negative regulator! [74,75]. Thus, expression and activation of distinct expression profiles of siglecs can either positively or negatively regulate immune and inflam‐ matory responses. Monocytic cells have been shown to express Siglec-3 (CD33) and Si‐ glecs-5,-7,-9 and-10, whereas generally, Mϕs express sialoadhesin (Siglec-1) and Siglecs-3 and-5 (reviewed in [76]), yet differentiation of monocytes to pro-and anti-inflammatory Mϕs using GM-CSF and M-CSF, respectively, showed retention of the monocyte Siglec phenotype [77]. This wide array of CD33-related siglec expression on monocytes and macrophages is suggestive of a degree of functional redundancy exists between these siglecs however, the specific expression profiles between cell types is indicative of specificity of function. More recently, research has uncovered more functional information regarding CD33-related siglecs; they may not just exert inhibitory responses through suppression of ITAM-dependent signals, as over-expression of siglec-9 in Mϕs both inhibits the TLR-induced expression of the proinflammatory cytokine, TNFα, but also induced the anti-inflammatory cytokine, IL-10 [78]. Again, as with other Mϕ receptors, the importance of a balance between positive and negative signals within the immune system is paramount.
**Fc**γ**Rs:** are a family of receptors which bind the Fc region of IgG/IgG immune complexes and are responsible for mediating both activatory and suppressive responses by transducing their ligand-binding signal through either an endogenous cytoplasmic ITAM or ITIM motif, respectively. Three classes of FcγRs are expressed on immune cells such as Mϕs, these include the high affinity FcγRI (CD64) which binds monomeric IgG and the low affinity receptors FcγRII (CD32) and FcγRIII (CD16) that bind IgG immune complexes. Activatory receptors which possess cytoplasmic ITAM sequences include FcγRIIIa, FcγRI and FcγRIIa whereas the CD32 subtype, FcγRIIb is suppressive, containing a cytoplasmic ITIM sequence (reviewed in [79]). Thus recognition and clearance of IgG immune complexes may induce either an activatory or inhibitory signal which is dependent on expression profile of these FcRs and potential cross-talk with other PRRs. Ligation of FcγRIIIa induces the expression and secretion of Mϕ-derived TNFα, the mechanism of production and activation of which has been found to be associated with pathological inflammation [80]. Ligation of FcγRI however, was found to reverse Mϕ pro-inflammatory responses [81]. Indeed, taking into consideration the format of presentation of IgG, the immune response cell and their FcγR expression profile, Mϕs can be activated or inhibited with respect to inflammatory responses; differential FcγR-driven induction and suppression in collagen-induced arthritis model have resulted in this being considered as a realistic therapeutic regimen in the treatment of RA [82]. The final thing to consider with respect to FcγR-induced responses in Mϕs is the fact that these responses may differ between M1 and M2 subsets and that these responses may be deviated by manipulation of FcγR signalling or indeed exhibit suppression of inflammatory responses or manipulation of functional plasticity. Manipulation of these Mϕ receptors would represent a realistic option when considering their association with alternative activation, M2-like anti-inflammatory Mϕs.
NLRP3 inflammasome and alleviating Th2 pathology. Of particular interest is the fact that Dectin-2 can also promote Th2responsiveness, as opposed to suppressing it, when responding to the common allergen of house dust mite [71]. MR drives protective Th2 responses during S.mansonii infection whereas enhances uptake of infectious circariae. These last few observa‐ tions with respect to Dectin-2 and MR and their ability to drive Th2 responses would appear to conform to the previously observed study of Mills et al, whereby M2 macrophages were associated with Th2 immunity [5]; indeed, MR at least, has been assigned as an M2 marker. On the flip-side, Dectin-1 has been associated with initiation of Th1 and Th17 anti-fungal responses; it would appear that any potential activation and modulation of specific immune responses would have to consider the role of pathogen associated carbohydrate expression and their
**Siglecs:** sialic acid-binding Ig-like lectins are members of the immunoglobulin superfamily involved in the recognition and reception of sialylated glycoconjugates, involved in endocy‐ tosis, positive and negative activation signals. Sialic acids are a family of nine-carbon sugars which are derivatives of neuraminic acid or keto deoxynonulosonic acid and appearing in different glycosidic linkages, typically at the exposed, non-reducing ends of oligosaccharide chains attached to proteins and lipids. It is thought that sialic acid residues either, mask subterminal sugars, prevent non-specific cell-cell interactions or act as ligands for modulating selective cell-cell interactions [73]. Currently, there are 13 Siglec family members expressed in humans and classified into two distinct subsets; siglecs 1 (CD169), 2 (CD22) and 4 are activatory receptors whereas siglecs 3 (CD33) and 5 (CD170) to 11, 14 and 16 (CD33-related siglecs) exhibit suppressive activity through two conserved ITIMs (immunereceptor tyrosine-based inhibition motif) in the cytoplasmic domain which are responsible for the recruitment of the signal suppressive phosphatase enzymes SHIP (lipid phosphatise), SHP-1 and SHP-2 (protein tyrosine phosphatises), which result in raised activation thresholds and antagonism of ITAMdependent activation. More recently, tyrosine phosphorylation of CD33 and siglec-7 can recruit SOCS3 which leads to ubiquitination and proteosomal degredation of the Siglec; regulation of the negative regulator! [74,75]. Thus, expression and activation of distinct expression profiles of siglecs can either positively or negatively regulate immune and inflam‐ matory responses. Monocytic cells have been shown to express Siglec-3 (CD33) and Si‐ glecs-5,-7,-9 and-10, whereas generally, Mϕs express sialoadhesin (Siglec-1) and Siglecs-3 and-5 (reviewed in [76]), yet differentiation of monocytes to pro-and anti-inflammatory Mϕs using GM-CSF and M-CSF, respectively, showed retention of the monocyte Siglec phenotype [77]. This wide array of CD33-related siglec expression on monocytes and macrophages is suggestive of a degree of functional redundancy exists between these siglecs however, the specific expression profiles between cell types is indicative of specificity of function. More recently, research has uncovered more functional information regarding CD33-related siglecs; they may not just exert inhibitory responses through suppression of ITAM-dependent signals, as over-expression of siglec-9 in Mϕs both inhibits the TLR-induced expression of the proinflammatory cytokine, TNFα, but also induced the anti-inflammatory cytokine, IL-10 [78]. Again, as with other Mϕ receptors, the importance of a balance between positive and negative
respective CLRs expressed on host macrophages.
130 Immune Response Activation
signals within the immune system is paramount.
**Other macrophage suppressive receptors:** Two further immunosuppressive molecules that control Mϕ functionality which are worthy of consideration are CD200R and SIRPα (CD172a); both members of the immunoglobulin superfamily (IgSF). CD172a interacts with a broadly expressed ligand, CD47, which results in the suppression of Mϕs through interaction with the tyrosine phosphatases, SHP1 and SHP2 [83] and reviewed in [84]). In addition, this interaction is made more complex by the observation of bi-directionality of signalling, where CD47 ligation has been found to selectively suppress monocyte IL-12 production, whereas TNFα, IL-1 and IL-6 was not suppressed. In contrast, the immunosup‐ pressive cytokines, TGFβ and IL-10, were not altered by CD47 ligation; suppression of IL-12 was independent of TGFβ and IL-10 [85]. Thus, molecules such as CD172a and CD47 represent an attractive target couple for multi-mechanistic immune suppression. CD200R ligation, on the other hand, transduces a simpler immunological signal. CD200 (OX-2) is a cell surface glycoprotein containing Ig-like domains that interacts with CD200R expressed on Mϕs [86]. The human homologue of CD200R was characterised by Wright and collea‐ gues who found that its cytoplasmic domain contained known signalling motifs [87]. The role of CD200R ligation was appreciated to be potentially suppressive where CD200 deficient mice were shown to display increased susceptibility to autoimmune disease models affecting joints. In accordance with this study, CD200Fc fusion protein was found to exhibit beneficial immunomodulation of arthritis and allograft rejection [88,89]. Indeed, the CD200Fc anti-arthritic response targeted pro-inflammatory cytokines in murine CIA, suppressing mRNA gene expression of TNFα, IL-1β, MMP13 but also IL-10 [90]. It is probable that CD200R represents a family of closely-related receptors; four homologues exist in the mouse, two of which, however, may exhibit activatory properties rather than suppressive, as they have been shown to pair with an ITAM-containing adaptor protein DAP12 [87]. Considering these early observations, it is likely that such IgSF members represent realistic therapeutic targets for immune suppression, activation and deviation of Mϕ-mediated pathological mechanisms associated with cancer, transplantation and inflammation.
down-regulation of the expression of MHC II, B7 costimulation (suppressing APC function) and pro-inflammatory cytokine production, with an indirect effect through cross-regulation of M1-derived cytokines and functionality. In contrast to this TGFβ-mediated suppression of Mϕ-driven functionality, pro-inflammatory cytokines such as TNFα modulate M1-driven responses, favouring anti-microbial killing, CMI and tissue-destructive pro-inflammatory responses. TNFα activates Mϕs following IFNγ priming, which results in a strong activation of NFκB and enhanced cell migration to inflammatory sites and consequent iNOS-mediated anti-microbial killing [97]. It also primes a sustained inflammatory response [98] and Mϕ survival, important to inflammation and to the pathology of sepsis [99]. TNFα and other proinflammatory cytokines such as IL-1β, not only play a pivotal role in the initiation and maintenance of the inflammatory response, but also modulate immunosuppressive mecha‐ nisms through the process of Mϕ endotoxin tolerance. Thus, the state of Mϕ activation, polarisation of subset functionality and the relative contribution of activatory, suppressive and immune-deviation signals associated with Mϕs, is of paramount importance in determining an appropriate immune response or pathological response as a consequence on Mϕ immune dysfunction (refer to figure 2 below). Finally, one last consideration regarding Mϕ functional responses to cytokines is that responsiveness has been demonstrated to be dysregulated, especially, with regards anti-inflammatory IL-10 signal transduction by IL-10 receptors in a chronic inflammatory environment [100]. Consideration of cytokine responsiveness of Mϕs can only be realistically appreciated when considering the local environment; homeostatic or
Macrophages — Masters of Immune Activation, Suppression and Deviation
http://dx.doi.org/10.5772/57541
133
**Macrophage functionality and inflammatory pathology:** Mϕs are particularly abundant in mucosal tissues such as the oral cavity and the gastro-intestinal tract. Dysregulation of Mϕ functionality in such mucosal tissue has been shown to have grave impacts on mucosal function and to result in Mϕ-driven pathology. This section will focus on Mϕ dysregulation and impact on host tissue in the context of M1-and M2-driven responses associated with oral (chronic periodontitis and oral squamous cell carcinoma) and the GIT (Crohn's disease, ulcerative colitis and colorectal cancer) mucosal pathology. These pathologies display pathological mechanisms aligned to M1-or M2-like functional phenotypes; to fully appreciate these dysregulated responses, it is imperative to appreciate the functional phenotype of healthy, homeostatic mucosal macrophages. From studies focussing on the characterisation of gut mucosal Mϕs, in general mucosal Mϕs exhibit a functional phenotype similar to M2 Mϕs. Intestinal Mϕs express phagocytic scavenger receptors (CD33, CD36, CD68), HLA-DR, immunoregulatory receptors (CD200R and TGFβRI/RII) and immunoregulatory cytokines (IL-10, TGFβ) whereas fail to express the co-stimulatory molecules (CD40, CD80, CD86), pattern recognition receptors (CD14, TLRs), TREM-1, FcRs, CRs and the pro-inflammatory cytokines (TNFα, IL-1β, IL-6, IL-18). On the other hand, in general, inflammatory Mϕs express a wide variety of functional markers which include: PRRs (CD14, TLR2, TLR4, TLR5), FcRs (CD16, CD32, CD64, CD89), CRs, HLA-DR, co-stimulatory molecules, chemokine receptors (CCR5, CXCR4) and pro-inflammatory cytokines/markers (TREM-1, TNFα, IL-1β, IL-6, IL-18,
pathogenic!
CCL20) [28,101-103].
**Cytokine expression and responsiveness:** Dysregulation and aberrant functionality of Mϕs is widely associated with pathology. Macrophages are central to phagocytosis and the modulation of inflammatory responses, through the induction and regulation of cytokine expression. Mϕs not only phagocytose pathogens and particulate material, they are important for the clearance of dead and dying host cells, utilising a wealth of scavenger receptors such as SRA, CD36, CD68, integrins, CD14, complement components and the phosphatidyl serine receptor (reviewed in [91]). Clearance of necrotic host cell material is generally associated with the induction of inflammatory responses, where necrotic cells are recognised by Mϕs and result in the production of TNFα and the driving of Th1-mediated immune responses. In contrast, Mϕs recognise apoptotic cells [92], resulting in an anti-inflammatory environment, characterised by the production of TGFβ, PGE2 [93] and IL-10 [94]. In fact, the defective clearance of apoptotic cells has been linked with autoimmunity [95]. Recognition of apoptotic cells is thus important in immunoregulation modulated by Mϕs. Mϕs recognise and distin‐ guish between apoptotic and necrotic cells via danger-associated molecular patterns (DAMPs) and apoptotic cell-associated molecular patterns (ACAMPs). This dichotomy in responsive‐ ness is reflective of Mϕ subset functionality; indeed M2-like, alternatively activated Mϕs express a wide array of receptors for ACAMPs; it is likely that recognition of apoptotic cells will activate and programme M2 cells towards an anti-inflammatory and regulatory pheno‐ type, whereas DAMPs and necrotic cells drive an immune activatory, pro-inflammatory response (reviewed in [96]).
Cytokines are not only induced by phagocytic recognition and clearance of dead and dying cells. They are produced by Mϕs and a wide variety of other immune cells as part of immune activatory or regulatory responses. As such, Mϕs are modulated by a variety of cytokines; the predominant Mϕ-modulatory cytokines include TGFβ, IL-10, IL-12, IL-4/IL-13, IFNγ and TNFα. IL-4/IL-13 and TGFβ/IL-10 have been described above to be associated with priming M2 Mϕ subset polarisation (M2a and M2c, respectively); whereas IL-12 and IFNγ play an important role in the polarisation and functionality of M1 cells and concomitantly, suppressing development and activation/functionality of M2s. The relative expression of these cytokines is thus imperative to subset functionality and cross-regulation.
TGFβ and IL-10 modulate Mϕ polarisation and functional plasticity to that of an M2c subset which exhibits a characteristic cytokine phenotype of IL-10hi, IL-12lo, IL-23lo and TGFβ<sup>+</sup> which is associated with anti-inflammatory responses, scavenging, immunoregulation, tissue repair and tumour promotion. Both TGFβ and IL-10 directly suppress immune activation via the down-regulation of the expression of MHC II, B7 costimulation (suppressing APC function) and pro-inflammatory cytokine production, with an indirect effect through cross-regulation of M1-derived cytokines and functionality. In contrast to this TGFβ-mediated suppression of Mϕ-driven functionality, pro-inflammatory cytokines such as TNFα modulate M1-driven responses, favouring anti-microbial killing, CMI and tissue-destructive pro-inflammatory responses. TNFα activates Mϕs following IFNγ priming, which results in a strong activation of NFκB and enhanced cell migration to inflammatory sites and consequent iNOS-mediated anti-microbial killing [97]. It also primes a sustained inflammatory response [98] and Mϕ survival, important to inflammation and to the pathology of sepsis [99]. TNFα and other proinflammatory cytokines such as IL-1β, not only play a pivotal role in the initiation and maintenance of the inflammatory response, but also modulate immunosuppressive mecha‐ nisms through the process of Mϕ endotoxin tolerance. Thus, the state of Mϕ activation, polarisation of subset functionality and the relative contribution of activatory, suppressive and immune-deviation signals associated with Mϕs, is of paramount importance in determining an appropriate immune response or pathological response as a consequence on Mϕ immune dysfunction (refer to figure 2 below). Finally, one last consideration regarding Mϕ functional responses to cytokines is that responsiveness has been demonstrated to be dysregulated, especially, with regards anti-inflammatory IL-10 signal transduction by IL-10 receptors in a chronic inflammatory environment [100]. Consideration of cytokine responsiveness of Mϕs can only be realistically appreciated when considering the local environment; homeostatic or pathogenic!
the CD200Fc anti-arthritic response targeted pro-inflammatory cytokines in murine CIA, suppressing mRNA gene expression of TNFα, IL-1β, MMP13 but also IL-10 [90]. It is probable that CD200R represents a family of closely-related receptors; four homologues exist in the mouse, two of which, however, may exhibit activatory properties rather than suppressive, as they have been shown to pair with an ITAM-containing adaptor protein DAP12 [87]. Considering these early observations, it is likely that such IgSF members represent realistic therapeutic targets for immune suppression, activation and deviation of Mϕ-mediated pathological mechanisms associated with cancer, transplantation and
**Cytokine expression and responsiveness:** Dysregulation and aberrant functionality of Mϕs is widely associated with pathology. Macrophages are central to phagocytosis and the modulation of inflammatory responses, through the induction and regulation of cytokine expression. Mϕs not only phagocytose pathogens and particulate material, they are important for the clearance of dead and dying host cells, utilising a wealth of scavenger receptors such as SRA, CD36, CD68, integrins, CD14, complement components and the phosphatidyl serine receptor (reviewed in [91]). Clearance of necrotic host cell material is generally associated with the induction of inflammatory responses, where necrotic cells are recognised by Mϕs and result in the production of TNFα and the driving of Th1-mediated immune responses. In contrast, Mϕs recognise apoptotic cells [92], resulting in an anti-inflammatory environment, characterised by the production of TGFβ, PGE2 [93] and IL-10 [94]. In fact, the defective clearance of apoptotic cells has been linked with autoimmunity [95]. Recognition of apoptotic cells is thus important in immunoregulation modulated by Mϕs. Mϕs recognise and distin‐ guish between apoptotic and necrotic cells via danger-associated molecular patterns (DAMPs) and apoptotic cell-associated molecular patterns (ACAMPs). This dichotomy in responsive‐ ness is reflective of Mϕ subset functionality; indeed M2-like, alternatively activated Mϕs express a wide array of receptors for ACAMPs; it is likely that recognition of apoptotic cells will activate and programme M2 cells towards an anti-inflammatory and regulatory pheno‐ type, whereas DAMPs and necrotic cells drive an immune activatory, pro-inflammatory
Cytokines are not only induced by phagocytic recognition and clearance of dead and dying cells. They are produced by Mϕs and a wide variety of other immune cells as part of immune activatory or regulatory responses. As such, Mϕs are modulated by a variety of cytokines; the predominant Mϕ-modulatory cytokines include TGFβ, IL-10, IL-12, IL-4/IL-13, IFNγ and TNFα. IL-4/IL-13 and TGFβ/IL-10 have been described above to be associated with priming M2 Mϕ subset polarisation (M2a and M2c, respectively); whereas IL-12 and IFNγ play an important role in the polarisation and functionality of M1 cells and concomitantly, suppressing development and activation/functionality of M2s. The relative expression of these cytokines
TGFβ and IL-10 modulate Mϕ polarisation and functional plasticity to that of an M2c subset which exhibits a characteristic cytokine phenotype of IL-10hi, IL-12lo, IL-23lo and TGFβ<sup>+</sup>
is associated with anti-inflammatory responses, scavenging, immunoregulation, tissue repair and tumour promotion. Both TGFβ and IL-10 directly suppress immune activation via the
which
is thus imperative to subset functionality and cross-regulation.
inflammation.
132 Immune Response Activation
response (reviewed in [96]).
**Macrophage functionality and inflammatory pathology:** Mϕs are particularly abundant in mucosal tissues such as the oral cavity and the gastro-intestinal tract. Dysregulation of Mϕ functionality in such mucosal tissue has been shown to have grave impacts on mucosal function and to result in Mϕ-driven pathology. This section will focus on Mϕ dysregulation and impact on host tissue in the context of M1-and M2-driven responses associated with oral (chronic periodontitis and oral squamous cell carcinoma) and the GIT (Crohn's disease, ulcerative colitis and colorectal cancer) mucosal pathology. These pathologies display pathological mechanisms aligned to M1-or M2-like functional phenotypes; to fully appreciate these dysregulated responses, it is imperative to appreciate the functional phenotype of healthy, homeostatic mucosal macrophages. From studies focussing on the characterisation of gut mucosal Mϕs, in general mucosal Mϕs exhibit a functional phenotype similar to M2 Mϕs. Intestinal Mϕs express phagocytic scavenger receptors (CD33, CD36, CD68), HLA-DR, immunoregulatory receptors (CD200R and TGFβRI/RII) and immunoregulatory cytokines (IL-10, TGFβ) whereas fail to express the co-stimulatory molecules (CD40, CD80, CD86), pattern recognition receptors (CD14, TLRs), TREM-1, FcRs, CRs and the pro-inflammatory cytokines (TNFα, IL-1β, IL-6, IL-18). On the other hand, in general, inflammatory Mϕs express a wide variety of functional markers which include: PRRs (CD14, TLR2, TLR4, TLR5), FcRs (CD16, CD32, CD64, CD89), CRs, HLA-DR, co-stimulatory molecules, chemokine receptors (CCR5, CXCR4) and pro-inflammatory cytokines/markers (TREM-1, TNFα, IL-1β, IL-6, IL-18, CCL20) [28,101-103].
inflammatory signals induced through TLR2 [106]. In addition, NOD2 mutation also results in impaired recruitment of monocytes to mucosal tissue and their differentiation to a homeo‐ static intestinal Mϕ phenotype as well as impairment of Treg activity [107]. Thus, the combi‐ nation of a dysregulated innate response in favour of pro-inflammatory cytokines, aberrant anti-inflammatory mechanisms, coupled with adaptive immune system bias towards a Th1/ Th17 cell mediated inflammatory response, result in the breaking of mucosal immune tolerance in favour of the tissue-destructive inflammatory mechanisms observed in the pathology of Crohn's disease and its perpetuation by both pathogenic and commensal luminal contents (reviewed in [108]). This loss in regulation results in a mucosal environment that is IL-10low IL-12high, which induces M1-like Mϕ activation/differentiation and the up-regulation of
Macrophages — Masters of Immune Activation, Suppression and Deviation
http://dx.doi.org/10.5772/57541
135
The oral mucosa plays a significant role in tolerance induction; it is thought that dysregulation of mucosal Mϕ functionality contributes to the breaking of this tolerance and drives oral inflammatory pathologies such as oral lichen planus, recurrent aphthous stomatitis and chronic periodontitis [110]. Chronic periodontitis (CP) is a persistent inflammatory condition of the periodontal tissue, characterised by bouts of relapse and remission that, if left untreated, could result in destruction of periodontium and tooth loss. CP results from activation of host inflammatory response as a consequence of constant microbial challenge mounted to a dysbiotic microbial population present in the oral cavity, of which *Porphyromonas gingivalis* (PG) is a prominent member [111,112]. PG is an intracellular oral pathogen, which invades oral epithelial cells, dendritic cells and macrophages; clearance of such would require a Th1 dependent cell mediated response [113,114]. PG-LPS subverts both host innate and adaptive responses away from clearing responses by low endotoxin activity along with preferential utilisation of TLR2 rather than TLR4 and by inducing Th2-dependent humoral responses, respectively [115,116]. There is a dense population of Mϕs in the oral mucosa which react to PG by producing large amounts of pro-inflammatory cytokines (TNFα, IL-1α, IL-1β, IL-6, IL-8, IL-12, IL-18, IL-32, MCP-1) at the expense of anti-inflammatory cytokines (IL-10) [117]; displaying an M1 phenotype. What is a cause for concern, is that PG induces an M1-like response whereas deviates the adaptive response to one of a non-clearing Th2-humoral response. This, coupled with the fact that PG differentially induces endotoxin tolerance in M1 and M2-like Mϕs [54], presents scientists with real problems with regards utilising Mϕs as
Less complicated is oral lichen planus (OLP); an inflammatory condition presenting as white striations or plaques in the buccal mucosa, tongue and gingiva [118]. This pathology is characterised by a dense T cell infiltrate with a corresponding destruction of basal membrane and recruited monocytes, which display an M1 cytokine expression phenotype, with high levels of GM-CSF, TNFα and IFNγ produced at the site of inflammation [118,119]. T cells associated with OLP have been described to produce IFNγ [120], as such, IFNγ can activate
cytotoxic T cells, NK cells and can feedback to activate M1 Mϕs and perpetuate their inflammatory effector response. These Mϕs are situated close to the basal membrane and can contribute to its inflammatory destruction, through the action of TNFα produced [121]. Therefore, M1 Mϕs drive the progression of OLP by activating a Th1-mediated inflammatory
expression of pro-inflammatory cytokines and co-stimulatory molecules [109].
therapeutic targets for the treatment of CP.
CD8+
**Figure 2.** Macrophage effector functionality is determined by a variety of activation, suppression and deviation. Mϕ effector responses can be positively initiated by a range of receptors (yellow box) which include CLRs, TLRs, NLRs and complement receptors (CRs) which induce a range of activatory signalling pathways (p38, ERK and JNK mitogen acti‐ vated protein kinases, MAPKs, and NFκB), resulting in specific gene expression. These positive signals can be sup‐ pressed (purple box) by a range of exogenous (SIGIRR, ST2, Siglecs, FcγRIIb, CD200R, SIRP-1α, TGFβR, IL-10R) and endogenous signals (SOCS3, MKP-1, PI3K, ITIM-associated SHP-1, SHP-2, SHIP-1). Finally, Mϕ functionality can deviate between M1 and M2 phenotype (white box), determined by exogenous factors (IL-4, IL-13, M-CSF, IFNγ, GM-CSF, TGFβ, IL-10) and internal signalling molecules (STAT1, STAT6, p65 NFκB, p50 NFκB).
#### **3.1. M1-driven pathology**
Crohn's disease (CD) is an inflammatory bowel disease (IBD) characterised by tissue-destruc‐ tive transmural inflammatory skip lesions which present along the whole length of the GI tract. CD is characterised by dysfunctional innate immune responses, which result in a Th1/Th17 axis of inflammatory destruction. This granulomatous inflammation is characterised by high expression of IL-12/IL-23, resulting in the predominance of Th1/Th17 cells that produce IFNγ, TNFα and IL-17 [104]. This dysregulated pro-inflammatory response is believed to be associ‐ ated with mutations to PRRs and pathogen sensing ability of the innate immune system. Genetic mutations in the CARD15 gene which encodes NOD2, an intracellular sensing receptor for bacterial peptidoglycan, are associated with CD Mϕs and have been demonstrated to modulate NFκB activation and up-regulate expression of the pro-inflammatory cytokines TNFα, IL-1β and IL-12 [105], whereas NOD2 activation would normally mediate antiinflammatory signals induced through TLR2 [106]. In addition, NOD2 mutation also results in impaired recruitment of monocytes to mucosal tissue and their differentiation to a homeo‐ static intestinal Mϕ phenotype as well as impairment of Treg activity [107]. Thus, the combi‐ nation of a dysregulated innate response in favour of pro-inflammatory cytokines, aberrant anti-inflammatory mechanisms, coupled with adaptive immune system bias towards a Th1/ Th17 cell mediated inflammatory response, result in the breaking of mucosal immune tolerance in favour of the tissue-destructive inflammatory mechanisms observed in the pathology of Crohn's disease and its perpetuation by both pathogenic and commensal luminal contents (reviewed in [108]). This loss in regulation results in a mucosal environment that is IL-10low IL-12high, which induces M1-like Mϕ activation/differentiation and the up-regulation of expression of pro-inflammatory cytokines and co-stimulatory molecules [109].
The oral mucosa plays a significant role in tolerance induction; it is thought that dysregulation of mucosal Mϕ functionality contributes to the breaking of this tolerance and drives oral inflammatory pathologies such as oral lichen planus, recurrent aphthous stomatitis and chronic periodontitis [110]. Chronic periodontitis (CP) is a persistent inflammatory condition of the periodontal tissue, characterised by bouts of relapse and remission that, if left untreated, could result in destruction of periodontium and tooth loss. CP results from activation of host inflammatory response as a consequence of constant microbial challenge mounted to a dysbiotic microbial population present in the oral cavity, of which *Porphyromonas gingivalis* (PG) is a prominent member [111,112]. PG is an intracellular oral pathogen, which invades oral epithelial cells, dendritic cells and macrophages; clearance of such would require a Th1 dependent cell mediated response [113,114]. PG-LPS subverts both host innate and adaptive responses away from clearing responses by low endotoxin activity along with preferential utilisation of TLR2 rather than TLR4 and by inducing Th2-dependent humoral responses, respectively [115,116]. There is a dense population of Mϕs in the oral mucosa which react to PG by producing large amounts of pro-inflammatory cytokines (TNFα, IL-1α, IL-1β, IL-6, IL-8, IL-12, IL-18, IL-32, MCP-1) at the expense of anti-inflammatory cytokines (IL-10) [117]; displaying an M1 phenotype. What is a cause for concern, is that PG induces an M1-like response whereas deviates the adaptive response to one of a non-clearing Th2-humoral response. This, coupled with the fact that PG differentially induces endotoxin tolerance in M1 and M2-like Mϕs [54], presents scientists with real problems with regards utilising Mϕs as therapeutic targets for the treatment of CP.
Less complicated is oral lichen planus (OLP); an inflammatory condition presenting as white striations or plaques in the buccal mucosa, tongue and gingiva [118]. This pathology is characterised by a dense T cell infiltrate with a corresponding destruction of basal membrane and recruited monocytes, which display an M1 cytokine expression phenotype, with high levels of GM-CSF, TNFα and IFNγ produced at the site of inflammation [118,119]. T cells associated with OLP have been described to produce IFNγ [120], as such, IFNγ can activate CD8+ cytotoxic T cells, NK cells and can feedback to activate M1 Mϕs and perpetuate their inflammatory effector response. These Mϕs are situated close to the basal membrane and can contribute to its inflammatory destruction, through the action of TNFα produced [121]. Therefore, M1 Mϕs drive the progression of OLP by activating a Th1-mediated inflammatory
**3.1. M1-driven pathology**
134 Immune Response Activation
Crohn's disease (CD) is an inflammatory bowel disease (IBD) characterised by tissue-destruc‐ tive transmural inflammatory skip lesions which present along the whole length of the GI tract. CD is characterised by dysfunctional innate immune responses, which result in a Th1/Th17 axis of inflammatory destruction. This granulomatous inflammation is characterised by high expression of IL-12/IL-23, resulting in the predominance of Th1/Th17 cells that produce IFNγ, TNFα and IL-17 [104]. This dysregulated pro-inflammatory response is believed to be associ‐ ated with mutations to PRRs and pathogen sensing ability of the innate immune system. Genetic mutations in the CARD15 gene which encodes NOD2, an intracellular sensing receptor for bacterial peptidoglycan, are associated with CD Mϕs and have been demonstrated to modulate NFκB activation and up-regulate expression of the pro-inflammatory cytokines TNFα, IL-1β and IL-12 [105], whereas NOD2 activation would normally mediate anti-
TGFβ, IL-10) and internal signalling molecules (STAT1, STAT6, p65 NFκB, p50 NFκB).
**Figure 2.** Macrophage effector functionality is determined by a variety of activation, suppression and deviation. Mϕ effector responses can be positively initiated by a range of receptors (yellow box) which include CLRs, TLRs, NLRs and complement receptors (CRs) which induce a range of activatory signalling pathways (p38, ERK and JNK mitogen acti‐ vated protein kinases, MAPKs, and NFκB), resulting in specific gene expression. These positive signals can be sup‐ pressed (purple box) by a range of exogenous (SIGIRR, ST2, Siglecs, FcγRIIb, CD200R, SIRP-1α, TGFβR, IL-10R) and endogenous signals (SOCS3, MKP-1, PI3K, ITIM-associated SHP-1, SHP-2, SHIP-1). Finally, Mϕ functionality can deviate between M1 and M2 phenotype (white box), determined by exogenous factors (IL-4, IL-13, M-CSF, IFNγ, GM-CSF,
response and perpetuation/exacerbation of this immunopathological response through a positive feedback mechanism.
pathways (p65 versus p50 NFκB) exhibit functional plasticity where TAMs display proinflammatory, anti-tumoral M1-like responses or, conversely, anti-inflammatory, pro-tumoral
Macrophages — Masters of Immune Activation, Suppression and Deviation
http://dx.doi.org/10.5772/57541
137
In the case of oral mucosa, M1-driven chronic inflammatory pathology such as OLP and the resulting pro-inflammatory and anti-microbial environment (ROS/RNS) can induce mutagen‐ esis and malignant transformation of oral epithelial cells resulting in oral cancer. One such oral cancer is oral squamous cell carcinoma (OSCC), characterised by a significant leukocyte infiltration of lymphocytes and particularly monocytes, recruited by MCP-1, which become TAMs [129]. These TAMs, facilitated by priming factors, such as M-CSF and IL-10, secreted by cancer cells, develop a pro-tumoral M2 phenotype (IL-10, EGF, FGF, PDGF, VEGF)[130] resulting in advanced stages of tumour progression [131,132]. This M2-like effector phenotype of TAMs results in an immunosuppressive, anti-inflammatory and tissue reparative environ‐ ment, which benefits tumour growth. In addition to suppressing pro-inflammatory cytokines, activation of anti-tumour T cell responses through APC function is suppressed as the expres‐ sion of MHC molecules is down-regulated and inhibitory co-stimulatory molecules, such as B7-H4, are up-regulated [133-135]. Finally, the M2 cytokines, IL-10 and TGFβ, can induce Tregs, which further strengthen the suppression of host anti-tumour responses [136]. Thus, TAMs modulate cancer cell survival/growth whereas the environment presented by the tumour cells influences the functional plasticity of the TAMs and that, through the bidirectionality of cellular interactions, it may be possible for the TAM plasticity to be harnessed
in the control of tumour cell growth and even initiate tumour resolution.
Macrophages offer attractive drug targets for the treatment of a wealth of Mϕ-mediated diseases of the oral and gastrointestinal mucosa. The role of Mϕs in driving inflammato‐ ry diseases such as CD, OLP, CP, UC, and cancer (CRC and OSCC) is being extensively researched. With their role in the immunopathogenic mechanisms of these diseases, Mϕs are now a realistic candidate for "re-education" of their functional responses; either by immune activation, suppression or deviation. Mϕs exhibit pathogenic responses via M1 functionality, as is the case in CD, OLP and CP, whereas M2 effector phenotype drives UC, CRC and OSCC. These pathogenic immune cells express a wealth of immunoactivatory receptors (TLRs, NLRs, RLRs, CLRs, FcRs, SRs and cytokine receptors) and, as such, can be induced to exhibit pro-inflammatory responses, anti-inflammatory responses, antimicrobial killing responses as well as antigen processing and presentation, hence priming humoral and CMI adaptive responses against extracellular and intracellular pathogens. In addition, they also express immunosuppressive molecules (exogenous and endogenous PRR inhibitors, FcγRIIb, CD200R, SIRP-1α, Siglecs, NLR splice variants, IL-10, TGFβ) and may exhibit a sensitivity to endotoxin tolerance. Finally, Mϕs also possess receptors/molecules by which immune responses may be modulated or deviated; through FcRs, IL-4/IL-13/M-CSF/IL-10 and IFNγ/GM-CSF to the benefit of the host. Future aims of manipulating Mϕ effector functionality would focus on activation of M1 Mϕs to produce anti-tumour and
**4. Conclusions and future perspectives**
M2-like behaviour [128].
#### **3.2. M2-driven pathology**
In contrast to CD, ulcerative colitis (UC or idiopathic proctocolitis) is an IBD where its site of involvement is confined to the large bowel (colon and rectum) and characterised by ulcerations in the mucosal lining layer. Patients with UC present with abdominal pain, bleeding anaemia and dietary restriction, resulting as a consequence of inflammatory destruction of the superficial mucosal layer, which is possibly driven by dysbiosis or inappropriate recognition and responsiveness to commensal gut organisms [122]. This pathology is driven by a M2/Th2-mediated pathogenic mechanism resulting in a predomi‐ nance of Th2-derived IL-4, IL-5 and IL-13 cytokine production. These cytokines induce B cell class switching and the production of IgE with the consequent activation of mast cells and eosonophils, which degranulate toxic components upon cross-linkage of IgE and initiation of an ADCC response. As with CD, there is an increase in monocyte recruit‐ ment to the gut mucosa, where there is a high proportion of Mϕs expressing a less mature phenotype that are CX3CR1+ TLR2+ CD33+ [123]. These CX3CR1+ Mϕs have been described to suppress the severity of DSS-induced colitis and may represent a beneficial Mϕ subset with respect to inflammatory pathology [124]. Manipulation of this Mϕ phenotype may represent a realistic approach in controlling this M2/Th2-driven inflammation. Thus there is a real need to characterise the inflammatory Mϕ subset which is involved in driving this disease. One observation that informs the scientific community with respect to Mϕ involvement in UC is the fact that CD1d-restricted type II NKTs induce colitis in mice [125]; it is proba‐ ble that CD1d+ cells, where CD1d expression has been reported to be restricted to Mϕs, initiate or perpetuate this inflammatory pathology in humans. In addition to expression of IL-4, IL-13 and TNFα, serum levels of IL-23 have been associated with severity of disease in UC patients [126]; whereas CD pathology involved an immune axis of M1/Th1/Th17, the link with IL-23 production is suggestive of an M2/Th2/Th17 axis driving UC immunopatho‐ logical inflammation. In order for successful regimens to be adopted for the treatment of UC, it is imperative that this immunopathological axis is further characterised with respect to Mϕ functional phenotype and how these Mϕs can be modulated by expression analy‐ sis of both cell-associated immune activatory and suppressive molecules.
Macrophages have been realised to play a significant role in the immunopathology associated with tumorigenesis and the progression of solid tumours. Tumour initiation and development has been linked with NFκB activation and inflammation; indeed, it is now well established that tumour associated macrophages (TAMs) exhibit a pro-tumoral, anti-inflammatory and immunoregulatory M2 phenotype. As inflammation contributes to the progression of UC, links have been suggested between UC (an M2/Th2 pathology) and the development of colorectal cancer (CRC). Comparisons exist between UC and CRC with respect to the upregulation of both pro-inflammatory cytokines and the immunoregulatory cytokine, IL-10 [127], thus both pro-and anti-inflammatory mechanisms contribute to cancer progression. With respect to Mϕ functionality, NFκB activation through classical or alternative activation pathways (p65 versus p50 NFκB) exhibit functional plasticity where TAMs display proinflammatory, anti-tumoral M1-like responses or, conversely, anti-inflammatory, pro-tumoral M2-like behaviour [128].
In the case of oral mucosa, M1-driven chronic inflammatory pathology such as OLP and the resulting pro-inflammatory and anti-microbial environment (ROS/RNS) can induce mutagen‐ esis and malignant transformation of oral epithelial cells resulting in oral cancer. One such oral cancer is oral squamous cell carcinoma (OSCC), characterised by a significant leukocyte infiltration of lymphocytes and particularly monocytes, recruited by MCP-1, which become TAMs [129]. These TAMs, facilitated by priming factors, such as M-CSF and IL-10, secreted by cancer cells, develop a pro-tumoral M2 phenotype (IL-10, EGF, FGF, PDGF, VEGF)[130] resulting in advanced stages of tumour progression [131,132]. This M2-like effector phenotype of TAMs results in an immunosuppressive, anti-inflammatory and tissue reparative environ‐ ment, which benefits tumour growth. In addition to suppressing pro-inflammatory cytokines, activation of anti-tumour T cell responses through APC function is suppressed as the expres‐ sion of MHC molecules is down-regulated and inhibitory co-stimulatory molecules, such as B7-H4, are up-regulated [133-135]. Finally, the M2 cytokines, IL-10 and TGFβ, can induce Tregs, which further strengthen the suppression of host anti-tumour responses [136]. Thus, TAMs modulate cancer cell survival/growth whereas the environment presented by the tumour cells influences the functional plasticity of the TAMs and that, through the bidirectionality of cellular interactions, it may be possible for the TAM plasticity to be harnessed in the control of tumour cell growth and even initiate tumour resolution.
#### **4. Conclusions and future perspectives**
response and perpetuation/exacerbation of this immunopathological response through a
In contrast to CD, ulcerative colitis (UC or idiopathic proctocolitis) is an IBD where its site of involvement is confined to the large bowel (colon and rectum) and characterised by ulcerations in the mucosal lining layer. Patients with UC present with abdominal pain, bleeding anaemia and dietary restriction, resulting as a consequence of inflammatory destruction of the superficial mucosal layer, which is possibly driven by dysbiosis or inappropriate recognition and responsiveness to commensal gut organisms [122]. This pathology is driven by a M2/Th2-mediated pathogenic mechanism resulting in a predomi‐ nance of Th2-derived IL-4, IL-5 and IL-13 cytokine production. These cytokines induce B cell class switching and the production of IgE with the consequent activation of mast cells and eosonophils, which degranulate toxic components upon cross-linkage of IgE and initiation of an ADCC response. As with CD, there is an increase in monocyte recruit‐ ment to the gut mucosa, where there is a high proportion of Mϕs expressing a less mature
[123]. These CX3CR1+
cells, where CD1d expression has been reported to be restricted to Mϕs,
suppress the severity of DSS-induced colitis and may represent a beneficial Mϕ subset with respect to inflammatory pathology [124]. Manipulation of this Mϕ phenotype may represent a realistic approach in controlling this M2/Th2-driven inflammation. Thus there is a real need to characterise the inflammatory Mϕ subset which is involved in driving this disease. One observation that informs the scientific community with respect to Mϕ involvement in UC is the fact that CD1d-restricted type II NKTs induce colitis in mice [125]; it is proba‐
initiate or perpetuate this inflammatory pathology in humans. In addition to expression of IL-4, IL-13 and TNFα, serum levels of IL-23 have been associated with severity of disease in UC patients [126]; whereas CD pathology involved an immune axis of M1/Th1/Th17, the link with IL-23 production is suggestive of an M2/Th2/Th17 axis driving UC immunopatho‐ logical inflammation. In order for successful regimens to be adopted for the treatment of UC, it is imperative that this immunopathological axis is further characterised with respect to Mϕ functional phenotype and how these Mϕs can be modulated by expression analy‐
Macrophages have been realised to play a significant role in the immunopathology associated with tumorigenesis and the progression of solid tumours. Tumour initiation and development has been linked with NFκB activation and inflammation; indeed, it is now well established that tumour associated macrophages (TAMs) exhibit a pro-tumoral, anti-inflammatory and immunoregulatory M2 phenotype. As inflammation contributes to the progression of UC, links have been suggested between UC (an M2/Th2 pathology) and the development of colorectal cancer (CRC). Comparisons exist between UC and CRC with respect to the upregulation of both pro-inflammatory cytokines and the immunoregulatory cytokine, IL-10 [127], thus both pro-and anti-inflammatory mechanisms contribute to cancer progression. With respect to Mϕ functionality, NFκB activation through classical or alternative activation
Mϕs have been described to
positive feedback mechanism.
phenotype that are CX3CR1+
ble that CD1d+
TLR2+
CD33+
sis of both cell-associated immune activatory and suppressive molecules.
**3.2. M2-driven pathology**
136 Immune Response Activation
Macrophages offer attractive drug targets for the treatment of a wealth of Mϕ-mediated diseases of the oral and gastrointestinal mucosa. The role of Mϕs in driving inflammato‐ ry diseases such as CD, OLP, CP, UC, and cancer (CRC and OSCC) is being extensively researched. With their role in the immunopathogenic mechanisms of these diseases, Mϕs are now a realistic candidate for "re-education" of their functional responses; either by immune activation, suppression or deviation. Mϕs exhibit pathogenic responses via M1 functionality, as is the case in CD, OLP and CP, whereas M2 effector phenotype drives UC, CRC and OSCC. These pathogenic immune cells express a wealth of immunoactivatory receptors (TLRs, NLRs, RLRs, CLRs, FcRs, SRs and cytokine receptors) and, as such, can be induced to exhibit pro-inflammatory responses, anti-inflammatory responses, antimicrobial killing responses as well as antigen processing and presentation, hence priming humoral and CMI adaptive responses against extracellular and intracellular pathogens. In addition, they also express immunosuppressive molecules (exogenous and endogenous PRR inhibitors, FcγRIIb, CD200R, SIRP-1α, Siglecs, NLR splice variants, IL-10, TGFβ) and may exhibit a sensitivity to endotoxin tolerance. Finally, Mϕs also possess receptors/molecules by which immune responses may be modulated or deviated; through FcRs, IL-4/IL-13/M-CSF/IL-10 and IFNγ/GM-CSF to the benefit of the host. Future aims of manipulating Mϕ effector functionality would focus on activation of M1 Mϕs to produce anti-tumour and pro-inflammatory cytokines, hence fostering a hostile anti-tumour innate environment as well as priming adaptive anti-tumour responses for the treatment of colorectal cancer and oral squamous cell carcinoma. In addition to immune activation of anti-tumour respons‐ es, harnessing immune receptors for immune deviation would appear to be a potentially effective approach; whereby the functional plasticity of M2-like TAMs was manipulated and re-programmed to a more anti-tumour setting by encouraging deviation of response to be M1-directed [137]. Although translatability of these approaches has proved, as yet, unsuccessful, this activation and re-programming approach has been appreciated for quite some time [138,139].
**References**
CD14+
2578-2586.
2004; 25(12):677-686.
2003; 73(2):209-212.
2006; 211:487-501.
Immunology 1998; 10:459-464.
Immunology 2001; 167:6957-6966.
producing T cells. Immunity 2006; 24:179-189.
Opinion in Immunology 2007; 19 :281-286.
+
CD16+
B, Pforte A. The novel subset of CD14+
[1] Zeigler-Heitbrock HWL, Fingerle G, Strobel M, Schraut W, Stetler F, Schutt C, Passlick
subpopulation. Immunology Today 1996; 17(9):424-428.
monocytes are a major source of TNF. Journal of Immunology 2002; 168:3536-3542.
[4] Kawanaka N, Yamamura M, Aita T, Morita Y, Okamoto A, Kawashima M, Iwahashi M, Ueno A, Ohmoto Y, Makino H. CD14+CD16+blood monocytes and joint inflammation in rheumatoid arthritis. Arthritis and Rheumatism 2002; 46(10):
[5] Mills CD, Kincaid K, Alt JM, Heilman MJ, Hill AM. M-1/M-2 macrophages and the Th1/
[6] Anderson CF, Mosser DM. A novel phenotype for an activated macrophage: the type 2 activated macrophage. Journal of Leukocyte Biology 2002; 72(1):101-106.
[7] Mantovani A, Sica A, Sozzani S, Allavena P, Vecchi A, Locati M. The chemokine system in diverse forms of macrophage activation and polarisation. Trends in Immunology
[8] Mosser DM. The many faces of macrophage activation. Journal of Leukocyte Biology
[9] Van Ginderachter JA, Movahedi K, Hassanzadeh Ghassabeh G, Meerschaut S, Beschin A, Raes G, DeBaetselier P. Classical and alternative activation of mononuclear phagocytes: picking the best of both worlds for tumour promotion. Immunobiology
[10] Louis J, Himmelrich H, Parra-Lopez C, Tacchini-Cottier F, Launois P. Regulation of protective immunity against Leishmania major in mice. Current Opinion in
[11] Holscher C, Atkinson RA, Arendse B, Brown N, Myburgh E, Alber G, Brombacher F. A protective and agonistic function of IL-12p40 in mycobacterial infection. Journal of
[12] Veldhoen M, Hocking RJ, Atkins CJ, Locksley RM, Stockinger B. TGF-β in the context of an inflammatory cytokine milieu supports de novo differentiation of IL-17-
[13] Stockinger B, Veldhoen M. Differentiation and function of Th17 T cells. Current
tissue macrophages. European Journal of Immunology 1993; 23(9):2053-2058.
[3] Belge K-U, Dayyani F, Horelt A, Siedlar M, Frankenberger M, Frankenberger B,
[2] Zeigler-Heitbrock HWL. Heterogeneity of human blood monocytes: the
Espevik T, Ziegler-Heitbrock L. The pro-inflammatory CD14+
Th2 paradigm. Journal of Immunology 2000; 164:6166-6173.
/CD16+
blood monocytes exhibits features of
Macrophages — Masters of Immune Activation, Suppression and Deviation
CD16+
HLA-DR+
http://dx.doi.org/10.5772/57541
139
With the similarity and connectivity between cancer and inflammation, it would appear reasonable to use similar Mϕ-modulation approaches to the treatment of mucosal inflamma‐ tory pathology. With respect to CD, OLP and CP, the M1-driven pathologies, Mϕ-based immunotherapy is likely to focus on approaches that either suppresses, induce tolerance or deviate a response by encouraging a redressing of the balance between M1/M2 by immunedeviation towards an M2-driven response. Such a change towards M2 functionality can potentially be activated through M2-biassed signals through scavenger receptors or deviation via IC-FcγR, M-CSF or IL-4/IL-13 polarisation signals. More recently, this immune deviation can be encouraged by the introduction of products derived from M2-skewing extracellular pathogens such as helminth worms; the potential for medicine to travel full circle back to a more sophisticated look at old therapies! Finally, these destructive M1 driven inflammatory pathologies may warrant suppression. This can be achieved by manipulating mechanisms and molecules associated with endotoxin tolerance through chronic reception of pro-inflammatory cytokines or PAMPs through TLRs and through the induction of a range of intracellular negative regulators such as Tollip, SHP-1, SHIP1, p50 NFκB, PI3K, as well as extracellular/membrane-bound regulators (CD200R, SIRP-1α, ST-2, SIGIRR, IL-10R and TGFβR).
Through the presentation of the complex field of Mϕ biology with respect to Mϕ subsets; priming of monocytes through defined differentiation pathways, Mϕ effector phenotype and functional plasticity can be potentially manipulated in our favour for the treatment of M1 driven and M2-driven pathology. The future therapeutic regimens which focus on macro‐ phage functionality in CD, OLP, CP, UC, CRC and OSCC can only hope to be successful from considering total Mϕ functionality with respect to immune activation, suppression and deviation.
#### **Author details**
Andrew D. Foey
School of Biomedical & Healthcare Sciences, University of Plymouth, Drake Circus, Plymouth, UK
#### **References**
pro-inflammatory cytokines, hence fostering a hostile anti-tumour innate environment as well as priming adaptive anti-tumour responses for the treatment of colorectal cancer and oral squamous cell carcinoma. In addition to immune activation of anti-tumour respons‐ es, harnessing immune receptors for immune deviation would appear to be a potentially effective approach; whereby the functional plasticity of M2-like TAMs was manipulated and re-programmed to a more anti-tumour setting by encouraging deviation of response to be M1-directed [137]. Although translatability of these approaches has proved, as yet, unsuccessful, this activation and re-programming approach has been appreciated for quite
With the similarity and connectivity between cancer and inflammation, it would appear reasonable to use similar Mϕ-modulation approaches to the treatment of mucosal inflamma‐ tory pathology. With respect to CD, OLP and CP, the M1-driven pathologies, Mϕ-based immunotherapy is likely to focus on approaches that either suppresses, induce tolerance or deviate a response by encouraging a redressing of the balance between M1/M2 by immunedeviation towards an M2-driven response. Such a change towards M2 functionality can potentially be activated through M2-biassed signals through scavenger receptors or deviation via IC-FcγR, M-CSF or IL-4/IL-13 polarisation signals. More recently, this immune deviation can be encouraged by the introduction of products derived from M2-skewing extracellular pathogens such as helminth worms; the potential for medicine to travel full circle back to a more sophisticated look at old therapies! Finally, these destructive M1 driven inflammatory pathologies may warrant suppression. This can be achieved by manipulating mechanisms and molecules associated with endotoxin tolerance through chronic reception of pro-inflammatory cytokines or PAMPs through TLRs and through the induction of a range of intracellular negative regulators such as Tollip, SHP-1, SHIP1, p50 NFκB, PI3K, as well as extracellular/membrane-bound regulators (CD200R, SIRP-1α, ST-2,
Through the presentation of the complex field of Mϕ biology with respect to Mϕ subsets; priming of monocytes through defined differentiation pathways, Mϕ effector phenotype and functional plasticity can be potentially manipulated in our favour for the treatment of M1 driven and M2-driven pathology. The future therapeutic regimens which focus on macro‐ phage functionality in CD, OLP, CP, UC, CRC and OSCC can only hope to be successful from considering total Mϕ functionality with respect to immune activation, suppression and
School of Biomedical & Healthcare Sciences, University of Plymouth, Drake Circus, Plymouth,
some time [138,139].
138 Immune Response Activation
SIGIRR, IL-10R and TGFβR).
deviation.
**Author details**
Andrew D. Foey
UK
[14] Finnin M, Hamilton JA, Moss ST. Characterisation of a CSF-induced proliferating subpopulation of human peripheral blood monocytes by surface marker expression and cytokine production. Journal of Leukocyte Biology 1999; 66(6):953-960.
[27] Pesce J, Kaviratne M, Ramalingam TR, Thompson RW, Urban JF, Cheever AW, Young DA, Collins M, Grusby MJ, Wynn TA. The IL-21 receptor augments Th2 effector function and alternative macrophage activation. Journal of Clinical Invesitgation 2006;
Macrophages — Masters of Immune Activation, Suppression and Deviation
http://dx.doi.org/10.5772/57541
141
[28] Foey AD. Mucosal macrophages: phenotype and functionality in homeostasis and pathology. 2012. In Handbook of macrophages: life cycle, functions and diseases. Takahashi, R., Kai, H., Eds. Nova Science Publishers Inc., New York, USA. Chapter 4,
[29] Martinez FO, Sica A, Mantovani A, Locati M. Macrophage activation and polarisation.
[30] Biswas SK, Gangi L, Paul S, Schioppa T, Saccani A, Sironi M, Bottazzi B, Doni A, Bronte V, Pasqualini F, Vago L, Nebuloni M, Mantovani A, Sica A. A distinct and unique transcriptional programme expressed by tumour-associated macrophages (defective NF-kappaB and enhanced IRF-3/STAT1 activation). Blood 2006; 107(5):2112-2122.
[31] Porta C, Rimoldi M, Raes G, Brys L, Ghezzi P, Di Liberto D, Dieli F, Ghisletti S, Natoli G, De Baetselier P, Mantovani A, Sica A. Tolerance and M2 (alternative) macrophage polarisation are related processes orchestrated by p50 nuclear factor kappaB. Proc.
[32] Bouhlel MA, Derudas B, Rigamonti E, Dievart R, Brozek J, Harlon S, Zawadzki C, Jude B, Torpier G, Marx N, Staels B, Chinetti-Gbaguichi G. PPAR-gamma activation primes
[33] Ruffell D, Mourkioti F, Gambardella A, Kirstetter P, Lopez RG, Rosenthal N, Nerlov C. A CREB-C/EBPbeta cascade induces M2 macrophage-specific gene expression and promotes muscle injury repair. Proceedings of the National Academy of Science USA
human monocytes into alternative M2 macrophages with anti-inflammatory
[34] Welch JS, Escoubet-Lozach L, Sykes DB, Liddiard K, Greaves DR, Glass CK. Th2
dependent mechanism. Journal of Biological Chemistry 2002; 277:42821-42829.
[35] Ishii M, Wen H, Corsa CAS, Liu T, Coelho AL, Allen RM, Carson WF, Cavassani KA, Li X, Lukacs NW, Hogaboam CM, Dou Y, Kunkel SL. Epigenetic regulation of the alternatively activated macrophage phenotype. Blood 2009; 114(15):3244-3254.
[36] Rauh MJ, Ho V, Pereira C, Sham A, Sly LM, Lam V, Huxham L, Minchinton AL, Mui A, Krystal G. SHIP represses the generation of alternatively activated macrophages.
[37] Stout RD, Jiang C, Matta B, Tietzel I, Watkins SK, Suttles J. Macrophages sequentially change their functional phenotype in response to changes in microenvironmental
influences. Journal of Immunology 2005; 175:342-349.
cytokines and allergic challenge induce Ym-1 expression in macrophages by a STAT-6-
116(7):2044-2055.
Frontiers in Bioscience 2008; 13:453-461.
Natl. Acad. Sci. USA 2009; 106(35):14978-14983.
properties. Cell Metabolism 2007; 6(2):137-143.
2009; 106:17475-17480.
Immunity 2005; 23:361-374.
pp121-146.
[27] Pesce J, Kaviratne M, Ramalingam TR, Thompson RW, Urban JF, Cheever AW, Young DA, Collins M, Grusby MJ, Wynn TA. The IL-21 receptor augments Th2 effector function and alternative macrophage activation. Journal of Clinical Invesitgation 2006; 116(7):2044-2055.
[14] Finnin M, Hamilton JA, Moss ST. Characterisation of a CSF-induced proliferating
and cytokine production. Journal of Leukocyte Biology 1999; 66(6):953-960. [15] Verreck FAW, de Boer T, Langenberg DML, van der Zanden L, Ottenhoff THM. Phenotypic and functional profiling of human pro-inflammatory type-1 and antiinflammatory type-2 macrophages in response to microbial antigens and IFNγ-and CD40L-mediated co-stimulation. Journal of Leukocyte Biology 2006; 79:285-293.
[16] Albina J, Henry WL, Mastrofrancesca B, Martin BA, Reichner JS. Macrophage activation by culture in an anoxic environment. Journal of Immunology 1995;
*Trypanosoma cruzi*. Infection and Immunity 1999; 67:4819-4826.
Current Drug Targets Inflammation Allergy 2003; 2:155-160.
activation. European Journal of Immunology 2007; 37:14-16.
[17] Aliberti J, Machado FS, Souto JT, Campanelli AP, Teixeira MM, Gazzinelli RT, Silva JS. β-chemokines enhance parasite uptake and promote nitric oxide-dependent microbiostatic activity in murine inflammatory macrophages infected with
[18] Kielian M, Cohn Z. Phorbol myristate acetate stimulates phagosome – lysosome fusion in mouse macrophages. Journal of Experimental Medicine 1981; 154:101-111. [19] Green S, Phillips W. Activation of the macrophage respiratory burst by phorbol
myristate acetate: evidence for both tyrosine-kinase-dependent and –independent pathways. International Journal of Biochemistry, Biophysics and Molecular Biology
[20] Grip O, Janciauskiene S, Lindgren S. Macrophages in inflammatory bowel disease.
[21] Mantovani A, Sica A, Locati M. New vistas on macrophage differentiation and
[22] Liu Y, Stewart KN, Bishop E, Marek CJ, Kluth DC, Rees AJ, Wilson HM. Unique
[24] Stein M, Keshaw S, Harris N, Gordon S. Interleukin 4 potently enhances murine macrophage mannose receptor activity: a marker of alternative immunologic macrophage activation. Journal of Experimental Medicine 1992 ; 176:287-292.
development. Journal of Clinical Investigation 2007; 117:1155-1166.
IFNgamma or IL-10. Journal of Immunology 1993; 151:6370-6381.
expression of suppressor of cytokine signalling 3 is essential for classical macrophage activation in rodents in vitro and in vivo. Journal of Immunology 2008; 180:6270-6278.
[23] Sica A, Bronte V. Altered macrophage differentiation and immune function in tumour
[25] de Waal Malefyt R, Figdor CG, Huijbens R, Mohan-Peterson S, Bennett B, Culpepper J, Dang W, Zurawski G, de Vries JE. Effects of IL-13 on phenotype, cytokine production and cytotoxic function of human monocytes. Comparison with IL-4 and modulation by
[26] Gordon S. Alternative activation of macrophages. Nature Reviews in Immunology
155:4391-4396.
140 Immune Response Activation
1994; 1222:241-248.
2003; 3:23-35.
subpopulation of human peripheral blood monocytes by surface marker expression
[38] Xu W, Zhao X, Daha MR, van Kooten C. Reversible differentiation of pro-and antiinflammatory macrophages. Molecular Immunology 2013; 53:179-186.
[52] Wald D, Qin J, Zhan Z, Qian Y, Naramura M, Tian L, Towne J, Sims JE, Stark GR, Li X. SIGIRR, a negative regulator of Toll-like receptor-interleukin 1 receptor signalling.
Macrophages — Masters of Immune Activation, Suppression and Deviation
http://dx.doi.org/10.5772/57541
143
[53] Biswas SK, Lopez-Collazo E. Endotoxin tolerance: new mechanisms, molecules and
[55] Philpott DJ, Girardin S. The role of Toll-like receptors and Nod proteins in bacterial
[56] Abbott DW, Wilkins A, Asara JM, Cantley LC. The Crohn's disease protein, NOD2, requires RIP2 in order to induce ubiquitination of a novel site on NEMO. Current
[57] Strober W, Murray PJ, Kitani A, Watanabe T. Signalling pathways and molecular interactions of NOD1 and NOD2. Nature Reviews in Immunology 2006; 6:9-20.
[58] Chamaillard M, Girardin SE, Viala J, Philpott DJ. Nods, Nalps and Naip: intracellular regulators of bacterial-induced inflammation. Cellular Microbiology 2003; 5(9):
[59] Mariathasan S, Newton K, Monack DM, Vucic D, French DM, Lee WP, Roose-Girma M, Erickson S, Dixit VM. Differential activation of the inflammasome by caspase-1
[60] Martinon F, Tschopp J. Inflammatory caspases: linking an intracellular innate immune
[61] Creagh EM, O'Neill LAJ. TLRs, NLRs and RLRs: a trinity of pathogen sensors that co-
[62] Kato H, Sato S, Yoneyama M, Yamamoto M, Uematsu S, Matsui K, Tsujimura T, Takeda K, Fujita T, Takeuchi O, Akira S. Cell-type specific involvement of RIG-1 in antiviral
[63] Yoneyama M, Kikuchi M, Natsukawa T, Shinobu N, Imaizumi T, Miyagishi M, Taira K, Akira S, Fujita T. The RNA helicase RIG-1 has an essential function in double stranded RNA-induced innate anti-viral responses. Nature Immunology 2004;
[64] Kang DC, Gopalkrishnam RV, Wu Q, Jankowsky E, Pyle AM, Fisher PB. Mda-5: an interferon-inducible putative RNA helicase with double stranded RNA-dependent ATPase activity and melanoma growth suppressive properties. Proceedings of the
[65] Rothenfusser S, Goutagny N, DiPerna G, Gong M, Monks BG, Schoenemeyer A, Yamamoto M, Akira S, Fitzgerald KA. The RNA helicase Lgp2 inhibits TLR-
operate in innate immunity. Trends in Immunology 2006; 27(8):352-357.
[54] Foey AD, Crean SJ. Macrophage subset sensitivity to endotoxin tolerisation by
clinical significance. Trends in Immunology 2009; 30(10):475-487.
*Porphyromonas gingivalis*. PLoS ONE 2013; 8(7):e67955.
infection. Molecular Immunology 2004; 41(11):1099-1108.
adaptors ASC and Ipaf. Nature 2004; 430:213-218.
National Academy of Science USA 2002; 99:637-642.
response. Immunity 2005; 23:19-28.
system to autoinflammatory diseases. Cell 2004; 117:561-574.
Nature Immunology 2003; 4(9):920-927.
Biology 2004; 14:2217-2227.
581-592.
5:730-737.
[38] Xu W, Zhao X, Daha MR, van Kooten C. Reversible differentiation of pro-and anti-
[39] Lenzo JC, Turner AL, Cook AD, Vlahos R, Anderson GP, Reynolds EC, Hamilton JA. Control of macrophage lineage populations by CSF-1 receptor and GM-CSF in homeostasis and inflammation. Immunology and Cell Biology 2011; 90(4):429-440.
[40] Grazio Cappiello M, Sutterwala FS, Trinchieri G, Mosser DM, Ma X. Suppression of IL-12 transcription in macrophages following Fc gamma receptor ligation. Journal of
macrophage Fc gamma receptors. Journal of Immunology 2001; 166(11):6861-6868. [42] Weisser SB, McLarren KW, Voglmaier N, van Netten-Thomas CJ, Antov A, Flavell RA, Sly LM. Alternative activation of macrophages by IL-4 requires SHIP degradation.
Feldmann M, Udalova IA. IRF5 promotes inflammatory macrophage polarisation and
Lawrence T. An anti-inflammatory role for IKKβ through the inhibition of "classical" macrophage activation. Journal of Experimental Medicine 2008; 205(6):1269-1276. [46] Hagemann T, Lawrence T, McNeish I, Charles KA, Kulbe H, Thompson RG, Robinson SC, Balkwill FR. "Re-educating" tumour-associated macrophages by targeting NF-
[44] Martinez FO. Regulators of macrophage activation. European Journal of Immunology
[47] O'Neill LAJ, Fitzgerald KA, Bowie AG. The Toll-IL-1 receptor adaptor family grows to
[48] O'Neill LAJ. When signalling pathways collide: positive and negative regulation of
[49] Negishi H, Ohba Y, Yanai H, Takaoka A, Honma K, Yui K, Matsuyama T, Taniguchi T, Honda K. Negative regulation of Toll-like receptor signalling by IRF-4. Proceedings of
[50] Honma K, Udono H, Kohno T, Yamamoto K, Ogawa A, Takemori T, Kumatori A,
Suzuki S, Matsuyama T, Yui K. Interferon regulatory factor 4 negatively regulates the production of proinflammatory cytokines by macrophages in response to LPS. Proceedings of the National Academy of Science USA 2005; 102(44):16001-16006. [51] Brint EK, Xu D, Liu H, Dunne A, McKenzie AN, O'Neill LA, Liew FY. ST2 is an inhibitor of interleukin 1 receptor and Toll-like receptor 4 signalling and maintains endotoxin
inflammatory macrophages. Molecular Immunology 2013; 53:179-186.
[41] Gerber JS, Mosser DM. Reversing lipopolysaccharide toxicity by ligating the
[43] Krausgruber T, Blazek T, Smallie T, Alzabin S, Lockstone H, Sahgal N, Hussell T,
[45] Fong CHY, Bebien M, Didierlaurent A, Nebauer R, Hussell T, Broide D, Karin M,
European Journal of Immunology 2011; 41:1742-1753.
Th1-Th17 responses. Nature Immunology 2011; 12:231-238.
kappaB. Journal of Experimental Medicine 2008; 205:1261-1268.
Toll-like receptor signal transduction. Immunity 2008; 29:12-20.
the National Academy of Science USA 2005; 102(44):15989-15994.
tolerance. Nature Immunology 2004; 5(4):373-379.
five members. Trends in Immunology 2003; 24(6):286-289.
Immunology 2001; 166(7):4498-4506.
2011; 41:1531-1534.
142 Immune Response Activation
dependent sensing of viral replication by RIG-1. Journal of Immunology 2005; 175:5260-5268.
[79] Ravetch JV, Bolland S. IgG Fc receptors. Annual Reviews in Immunology 2001;
[80] Abrahams VM, Cambridge G, Lydyard PM, Edwards JCW. Induction of tumour
type IIIa in rheumatoid arthritis. Arthritis and Rheumatism 2000; 43:608-616.
[82] Kleinau S, Martinsson P, Heyman B. Induction and suppression of collagen-induced arthritis is dependent on distinct Fcγ receptors. Journal of Experimental Medicine 2000;
[83] Kharitonenkov A, Chen Z, Sures I, Wang H, Schilling J, Ullrich A. A family of proteins that inhibit signalling through tyrosine kinase receptors. Nature 1997; 386:181-186.
[85] Armant M, Avice MN, Hermann P, Rubio M, Kiniwa M, Delespesse G, Sarfati M. CD47 ligation selectively downregulates human interleukin 12 production. Journal of
[86] Wright GJ, Puklavec MJ, Willis AC, Hoek RM, Sedgwick JD, Brown MH, Barclay AN. Lymphoid/neuronal cell surface OX2 glycoprotein recognises a novel receptor on macrophages implicated in the control of their function. Immunity 2000; 13:233-242. [87] Wright GJ, Cherwinski H, Foster-Cuevas M, Brooke G, Puklavec MJ, Bigler M, Song Y, Jenmalm M, Gorman D, McClanahan T, Liu M-R, Brown MH, Sedgwick JD, Phillips JH, Barclay AN. Characterisation of the CD200 receptor family in mice and humans and
their interactions with CD200. Journal of Immunology 2003; 171:3034-3046.
prolongs allo-and xenograft survival. Journal of Immunology 1999; 163:1654. [89] Gorczynski RM, Chen Z, Yu K, Hu J. CD200 immunoadhesin suppresses collagen-
[90] Simelyte E, Criado G, Essex D, Uger RA, Feldmann M, Williams RO. CD200-Fc, a novel antiarthritic biologic agent that targets pro-inflammatory cytokine expression in the joints of mice with collagen-induced arthritis. Arthritis and Rheumatism 2008; 58(4):
[91] Savill J, Dransfield I, Gregory C, Haslett C. A blast from the past: clearance of apoptotic cells regulates immune responses. Nature Reviews in Immunology 2002; 2:965-975.
[92] Savill J, Wyllie AH, Henson JE, Walport MJ, Henson PM, Haslett C. Macrophage phagocytosis of aging neutrophils in inflammation. Programmed cell death in the
immunoadhesin incorporating the molecule OX-2 is a potent immunosuppressant that
[88] Gorczynski RM, Cattrall MS, Chen Z, Hu J, Lei J, Min WP, Yu G, Ni J. An
induced arthritis in mice. Clinical Immunology 2001; 101:328-334.
interactions in the control of myeloid cells. Trends in Immunology 2002; 23(6):285-290.
[84] Barclay AN, Wright GJ, Brooke G, Brown MH. CD200 and membrane protein
[81] Sutterwala FS, Noel GJ, Salgame P, Mosser DM. Reversal of proinflammatory responses by ligating the macrophage Fcgamma receptor type I. Journal of
Experimental Medicine 1998; 188:217-222.
Experimental Medicine 1999; 190(8):1175-1182.
necrosis factor a production by adhered human monocytes. A key role for Fcg receptor
Macrophages — Masters of Immune Activation, Suppression and Deviation
http://dx.doi.org/10.5772/57541
145
19:275-290.
191:1611-1616.
1038-1043.
[79] Ravetch JV, Bolland S. IgG Fc receptors. Annual Reviews in Immunology 2001; 19:275-290.
dependent sensing of viral replication by RIG-1. Journal of Immunology 2005;
[66] Meylan E, Curran J, Hofmann K, Moradpour D, Binder M, Bartenschlager R, Tschopp J. Cardif is an adaptor protein in the RIG-1 antiviral pathway and is targeted by
[67] Li XD, Sun L, Seth RB, Pineda G, Chen ZJ. Hepatitis C virus protease NS3/4A cleaves mitochondrial antiviral signalling protein off the mitochondria to evade innate immunity. Proceedings of the National Academy of Sciences USA 2005;
[68] Johnston JB, Barrett JW, Nazarian SH, Goodwin M, Riccioto D, Wang G, McFadden G. A poxvirus-encoded pyrin domain protein interacts with ASC-1 to inhibit host inflammatory and apoptotic responses to infection. Immunity 2005; 23:587-598.
[69] Peiser L, Gordon S. The function of scavenger receptors expressed by macrophages and their role in the regulation of inflammation. Microbes and Infection 2001; 3(2):149-159.
[70] Peiser L, Mukhopadhyay S, Gordon S. Scavenger receptors in innate immunity.
[72] Brown GD. Dectin-1: a signalling non-TLR pattern recognition receptor. Nature
[73] Kelm S, Schauer R. Sialic acids in molecular and cellular interactions. International
[74] Orr SJ, Morgan NM, Elliott J, Burrows JF, Scott CJ, McVicar DW, Johnston JA. CD33
[75] Orr SJ, Morgan NM, Buick RJ, Boyd CR, Elliott J, Burrows JF, Jefferies CA, Crocker PR, Johnston JA. SOCS3 targets siglec-7 for proteosomal degradation and blocks siglec-7-
[76] Crocker PR, Varki A. Siglecs, sialic acids and innate immunity. Trends in Immunology
macrophages via tyrosine-based motifs. Biochem. Biophys. Res. Commun. 2008;
human mononuclear phagocytes, monocyte-derived dendritic cells and plasmacytoid
mediated responses. Journal of Biological Chemistry 2007b; 282(6):3418-3422.
[77] Lock K, Zhang J, Lu J, Lee SH, Crocker PR. Expression of CD33-related siglecs on
[78] Ando M, Tu W, Nishijima K, Lijima S. Siglec 9 enhances IL-10 production in
dendritic cells. Immunobiology 2004; 209(1-2):199-207.
responses are blocked by SOCS3 through accelerated proteosomal mediated turnover.
[71] Hardison SE, Brown GD. C-Type lectin receptors orchestrate anti-fungal immunity.
175:5260-5268.
144 Immune Response Activation
102:17717-17722.
hepatitis C virus. Nature 2005; 437:1167-1172.
Current Opinion in Immunology 2002; 14:123-128.
Nature Immunology 2013; 13(9):817-822.
Reviews in Immunology 2006; 6:33-43.
Reviews in Cytology 1997; 175:137-240.
Blood 2007a; 109:1061-1068.
2001; 22(6):337-342.
369:878-883.
neutrophil leads to its recognition by macrophages. Journal of Clinical Investigation 1989; 83(3):865-875.
[104] Strober W, Zhang F, Kitani A, Fuss I, Fichtner-Feigl S. Proinflammatory cytokines underlying the inflammation of Crohn's disease. Current Opinions in
Macrophages — Masters of Immune Activation, Suppression and Deviation
http://dx.doi.org/10.5772/57541
147
[105] Maeda S, Hsu LC, Liu H, Bankston LA, Limura M, Kagnoff MF, Eckmann L, Karin M. Nod2 mutation in Crohn's disease potentiates NF-κB activity and IL-1β processing.
[106] Netea MG, Kullberg BJ, de Jong DJ, Franke B, Sprong T, Naber THJ, Drenth JPH, van der Meer JW. NOD2 mediates anti-inflammatory signals induced by TLR2 ligands: Implications for Crohn's disease. European Journal of Immunology 2004; 34:2052-2059.
immunomodulation of gut mucosal defences: homeostasis and immunopathology.
pathology: a macrophage account. British Journal of Oral and Maxillofacial Surgery
[109] Bouma G, Strober W. The immunological and genetic basis of inflammatory bowel
[110] Merry R, Belfield L, McArdle P, McLennan A, Crean SJ, Foey AD. Oral health and
[111] Socransky SS, Hafferjee AD, Cugini MA, Smith C, Kent RL. Microbial complexes in subgingival plaque. Journal of Clinical Periodontology 1998; 25:134-144. [112] Ezzo P, Cutler CW. Microorganisms as risk indicators for periodontal disease.
[113] Yilmaz O, Watanabe K, Lamont RJ. Involvement of integrins in fimbrae-mediated binding and invasion by *Porphyromonas gingivalis*. Cellular Microbiology 2002;
[114] Jotwani R, Cutler CW. Fimbriated *Porphyromonas gingivalis* is more efficient than
[115] Pulendran B, Kumar P, Cutler CW, Mohamadzadeh M, van Dyke T, Bancereau J. Lipopolysaccharide from distinct pathogens induce different classes of immune
[116] Darveau RP, Pham TT, Lemley K, Reife RA, Bainbridge BW, Coats SR, Howald WN, Way SS, Hajjar AM. *Porphyromonas gingivalis* lipopolysaccharide contains multiple lipid A species that functionally interact with both Toll-like receptors 2 and 4. Infection
[117] Barksby HE, Nile CJ, Jaedicke KM, Taylor JJ, Preshaw PM. Differential expression of immunoregulatory genes in monocytes in response to *Porphyromonas gingivalis* and
response in vivo. Journal of Immunology 2001; 167:5067-5076.
fimbrae-deficient *P. gingivalis* in entering human dendritic cells *in vitro* and induces an inflammatory Th1 effector response. Infection and Immunity 2004; 72:1725-1732.
[107] Schenk M, Mueller C. Adaptations of intestinal macrophages to an antigen-rich
environment. Seminars in Immunology 2007; 19:84-93.
disease. Nature Reviews in Immunology 2003; 3:521-533.
[108] Hardy H, Harris J, Lyon E, Beal J, Foey AD. Probiotics, prebiotics and
Gastroenterology 2010; 26:301-307.
Science 2006; 307:734-738.
Nutrients 2013; 5:1869-1912.
Periodontology 2000 2003; 32:24-35.
and Immunity 2004; 72:5041-5051.
2012; 50:2-7.
4:305-314.
[104] Strober W, Zhang F, Kitani A, Fuss I, Fichtner-Feigl S. Proinflammatory cytokines underlying the inflammation of Crohn's disease. Current Opinions in Gastroenterology 2010; 26:301-307.
neutrophil leads to its recognition by macrophages. Journal of Clinical Investigation
producing macrophages preferentially clear early apoptotic cells. Blood 2006; 107(12):
[95] Taylor PR, Carugati A, Fadok VA, Cook HT, Andrews M, Carroll MC, Savill JS, Henson PM, Botto M, Walport MJ. A hierarchical role for classical pathway complement proteins in the clearance of apoptotic cells. Journal of Experimental Medicine 2000;
[96] Albert ML. Death-defying immunity: do apoptotic cells influence antigen processing
[97] Wesemann DR, Benveniste EN. STAT-1alpha and IFN-gamma as modulators of TNFalpha signalling in macrophages: regulation and functional implications of the TNF receptor 1:STAT-1 alpha complex. Journal of Immunology 2003; 171(10):5313-5319.
dependent autocrine loop leading to sustained expression of chemokines and STAT1 dependent type I interferon-response genes. Nature Immunology 2008; 9(4):378-387.
[98] Yarilina A, Park-Min KH, Antoniv T, Hu X, Ivashkiv LB. TNF activates an IRF1-
[99] Conte D, Holcik M, Lefebvre CA, Lacasse E, Picketts DJ, Wright KE, Korneluk RG. Inhibitor of apoptosis protein cIAP2 is essential for lipopolysaccharide-induced macrophage survival. Molecular and Cellular Biology 2006; 26(2):699-708.
dysfunction in macrophages during chronic inflammation. Journal of Leukocyte
Scholmerich J, Gross V. T cell co-stimulatory molecules are up-regulated on intestinal
[100] Avdiushko R, Hongo D, Lake-Bullock H, Kaplan A, Cohen D. IL-10 receptor
[101] Rogler G, Hausmann M, Spottl T, Vogl D, Aschenbrenner E, Andus T, Falk W,
Gastroenterology and Hepatology 1999; 11(10):1105-1111.
Clinical Investigation 2005; 115(1):66-75.
macrophages from inflammatory bowel disease mucosa. European Journal of
[102] Smith PD, Smythies LE, Mosteller-Barnum M, Sibley DA, Russell MW, Merger M, Sellers MT, Orenstein JM, Shimada T, Graham MF, Kubagawa H. Intestinal macrophages lack CD14 and CD89 and consequently down-regulated for LPS-and
[103] Smythies LE, Sellers MT, Clements RH, Mosteller-Barnum M, Meng G, Benjamin WH, Orenstein JM, Smith PD. Human intestinal macrophages display profound
inflammatory anergy despite avid phagocytic and bacteriocidal activity. Journal of
IgA-mediated activities. Journal of Immunology 2001; 167(5):2651-2656.
and presentation? Nature Reviews in Immunology 2004;4:223-231.
[93] Fadok VA, Bratton DL, Konowal A, Freed PW, Westcott JY, Henson PM. Macrophages that have ingested apoptotic cells *in vitro* inhibit proinflammatory cytokine production through autocrine/paracrine mechanisms involving TGF-β, PGE2 and PAF. Journal of
[94] Xu W, Roos A, Schlagwein N, Woltman AM, Daha MR, van Kooten C. IL-10-
1989; 83(3):865-875.
146 Immune Response Activation
4930-4937.
192:359-366.
Biology 2001; 70:624-632.
Clinical Investigation 1998; 101(4):890-898.
*Escherichia coli* lipopolysaccharide. Clinical and Experimental Immunology 2009; 156:479-487.
[129] Koide N, Nishio A, Sato T, Sugiyama A, Miyagawa S. Significance of macrophage
carcinoma of the oesophagus. American Journal of Gastroenterology 2004;
[130] Lalla RV, Boisoneau DS, Spiro JD, Kreutzer DL. Expression of vascular endothelial
Archives in Otolaryngology and Head and Neck Surgery 2003; 129:882-888.
[131] Sica A, Allavena P, Mantovani A. Cancer related inflammation: the macrophage
Macrophage colony-stimulating factor as a tumour marker for squamous cell
[133] Kundu N, Fulton AM. Interleukin-10 inhibits tumour metastasis, downregulates MHC
[134] de Waal Malefyt R, Haanem J, Spits H, Roncarolo MG, te Velde A, Figdor C, Johnson K, Kastelein R, Yssel H, de Vries JE. Interleukin 10 (IL-10) and viral IL-10 strongly reduce antigen-specific human T cell proliferation by diminishing the antigen-
histocompatibility complex expression. Journal of Experimental Medicine 1991;
[135] Kryczek I, Zou L, Rodriguez P, Zhu G, Wei S, Mottram P, Brumlik M, Cheng P, Curiel T, Myers L, Lackner A, Alvarez X, Ochoa A, Chen L, Zou W. B7-H4 expression identifies a novel suppressive macrophage population in human ovarian carcinoma.
[136] Levings MK, Bacchetta R, Schulz U, Roncarolo MG. The role of IL-10 and TGF-beta in the differentiation and effector function of T regulatory cells. International Archives of
[137] Mantovani A, Sica A. Macrophages, innate immunity and cancer: balance, tolerance
[138] Hennemann B, Beckmann G, Eichelmann A, Rehm A, Andreesen R. Phase I trial of adoptive immunotherapy of cancer patients using monocyte-derived macrophages activated with interferon gamma and lipopolysaccharide. Cancer Immunology and
[139] Guiducci C, Vicari AP, Sangaletti S, Trinchieri G, Columbo MP. Redirecting in vivo
elicited tumour infiltrating macrophages and dendritic cells towards tumour rejection.
and diversity. Current Opinion in Immunology 2010; 22:231-237.
[132] Kuropkat C, Dunne AA, Plehn S, Ossendorf M, Herz U, Renz H, Werner JA.
carcinoma of the head and neck. Tumour Biology 2003; 24:236-240.
class I and enhances NK lysis. Cellular Immunology 1997; 180:55-61.
presenting capacity of monocytes via downregulation of class II major
Journal of Experimental Medicine 2006; 203:871-881.
Allergy and Immunology 2002; 129:263-276.
Immunotherapy 1998; 45(5):250-256.
Cancer Research 2005; 65(8):3437-3446.
connection. Cancer Letters 2008; 267:204-215.
99:1667-1674.
174:915-924.
chemoattractant protein-1 expression and macrophage infiltration in squamous cell
Macrophages — Masters of Immune Activation, Suppression and Deviation
http://dx.doi.org/10.5772/57541
149
growth factor receptors on tumour cells in head and neck squamous cell carcinoma.
[129] Koide N, Nishio A, Sato T, Sugiyama A, Miyagawa S. Significance of macrophage chemoattractant protein-1 expression and macrophage infiltration in squamous cell carcinoma of the oesophagus. American Journal of Gastroenterology 2004; 99:1667-1674.
*Escherichia coli* lipopolysaccharide. Clinical and Experimental Immunology 2009;
[118] Sugerman PB, Savage NW, Walsh LJ, Zhao ZZ, Zhou XJ, Khan A, Seymour GJ, Bigby M. The pathogenesis of oral lichen planus. Critical Reviews in Oral Biology and
[119] Yamamoto T, Osaki T, Yoneda K, Ueta E. Cytokine production by keratinocytes and
[120] Khan A, Farah CS, Savage NW, Walsh LJ, Harbrow DJ, Sugerman PB. Th1 cytokines in oral lichen planus. Journal of Oral Pathology and Medicine 2003; 32:77-83.
[121] Kono T, Tanii T, Furukawa M, Mizuno N, Taniguchi S, Ishii M, Hamada T. Effects of human recombinant tumour necrosis factor-alpha (TNF-alpha) on the proliferative potential of human keratinocytes cultured in serum-free medium. Journal of
[122] Gophna U, Sommerfeld K, Gophna S, Doolittle WF, Veldhuyzen van Zanten SJ.
[123] Candia E, Diaz-Jiminez D, Langjahr P, Nunez LE, de la Fuente M, Farfan N, Lopez-
[124] Medina-Contreras O, Geem D, Laur O, Williams IR, Lira SA, Nusrat A, Parkos CA, Denning TL. CX3CR1 regulates intestinal macrophage homeostasis, bacterial
[125] Liao CM, Zimmer MI, Shanmuganad S, Yu HT, Cardell SL, Wang CR. Dysregulation of CD1d-restricted type II natural killer T cells leads to spontaneous development of
[126] Mohammadi M, Hayatbakhsh MM, Zahedi MJ, Jalapour MR, Pakgohar A. Serum
[127] Szaradkiewicz A, Marciniak R, Chudzicka-Strgala I, Wasilewska A, Drews M,
[128] Karin M, Greten FR. NF-kappaB: linking inflammation and immunity to cancer development and progression. Nature Reviews in Immunology 2005; 5:749-759.
interleukin-23 levels in patients with ulcerative colitis. Iran Journal of Immunology
Majewski P, Karpinski T, Zwozdziak B. Proinflammatory cytokines and IL-10 in inflammatory bowel disease and colorectal cancer patients. Archives in Immunology
colitis in mice. Gastroenterology 2012; 142(2):326-334.
and Therapeutic Experimentation 2009; 57:291-294.
Differences between tissue-associated intestinal microfloras of patients with Crohn's disease and ulcerative colitis. Journal of Clinical Microbiology 2006; 44:4136-4141.
Kostner F, Abedrapo M, Alvarez-Lobos M, Pinedo G, Beltran CJ, Gonzales C, Gonzales M-J, Quera R, Hermoso MA. Increased production of soluble TLR2 by lamina propria mononuclear cells from ulcerative colitis patients. Immunobiology 2012; 217(6):
translocation and colitogenic Th17 responses in mice. Journal of Clinical Investigation
mononuclear infiltrates in oral lichen planus. Journal of Oral Pathology and Medicine
156:479-487.
148 Immune Response Activation
1994; 23:309-315.
634-642.
2011; 121(12):4787-4795.
2011; 8(3):183-188.
Medicine 2002; 13:350-365.
Dermatology 1990; 17:409-413.
**Chapter 6**
**Immunostimulatory Effects of Triggering TLR3 Signaling**
Immunostimulation based therapies hold promise of disease-specific interventions without the toxicity and other side effects, which are associated with traditional therapeutic modalities. Unfortunately, the exact manifestation of their therapeutic benefits and the mechanisms of their action are still pending for most immunotherapeutic strategies. Virtually, such kinds of treatments are ideal to establish standards of care and secure, as well as improve therapeutics outcomes. A growing interest for studying and understanding immunostimulation based therapeutics' strategies has increased since, it could stimulate different components of innate
These immunostimulation based therapies can be divided into two categories, nonspecific and specific immunotherapies. Nonspecific immunotherapy describes therapies that are designed to enhance the immune response without deliberately seeking to modulate the response to a particular antigen. These therapies include three main categories. The first depends on administration of certain agents that induces secretion of inflammatory cytokines that nonspecifically activate the immune system [4]. The second depends on administration of certain type of cytokine that can activate specific populations of immune cells. The third category include the administration of antibodies that block immune checkpoints or other suppressive pathways that are known to shut down the activity of immune system [5]. These nonspecific activation pathways are crucial for the initiation of fully functional immune
responses encompassing both the innate and adaptive immune compartments [4].
© 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
**Pathway — Implication for Cancer Immunotherapy**
Mohamed Labib Salem, Said M. Hammad,
Additional information is available at the end of the chapter
Abdel-Aziz A. Zidan
**1. Introduction**
http://dx.doi.org/10.5772/58575
Mohamed R. El-shanshory, Mohamed A. Attia and
immunity and consequently stimulation of adaptive immunity [1-3].
**Chapter 6**
### **Immunostimulatory Effects of Triggering TLR3 Signaling Pathway — Implication for Cancer Immunotherapy**
Mohamed Labib Salem, Said M. Hammad, Mohamed R. El-shanshory, Mohamed A. Attia and Abdel-Aziz A. Zidan
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/58575
#### **1. Introduction**
Immunostimulation based therapies hold promise of disease-specific interventions without the toxicity and other side effects, which are associated with traditional therapeutic modalities. Unfortunately, the exact manifestation of their therapeutic benefits and the mechanisms of their action are still pending for most immunotherapeutic strategies. Virtually, such kinds of treatments are ideal to establish standards of care and secure, as well as improve therapeutics outcomes. A growing interest for studying and understanding immunostimulation based therapeutics' strategies has increased since, it could stimulate different components of innate immunity and consequently stimulation of adaptive immunity [1-3].
These immunostimulation based therapies can be divided into two categories, nonspecific and specific immunotherapies. Nonspecific immunotherapy describes therapies that are designed to enhance the immune response without deliberately seeking to modulate the response to a particular antigen. These therapies include three main categories. The first depends on administration of certain agents that induces secretion of inflammatory cytokines that nonspecifically activate the immune system [4]. The second depends on administration of certain type of cytokine that can activate specific populations of immune cells. The third category include the administration of antibodies that block immune checkpoints or other suppressive pathways that are known to shut down the activity of immune system [5]. These nonspecific activation pathways are crucial for the initiation of fully functional immune responses encompassing both the innate and adaptive immune compartments [4].
© 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Toll like receptors (TLRs) are a group of receptors that are expressed mainly in innate immune cells, including macrophages, monocytes, dendritic cells, natural killer cells and mast cells [6]. These TLRs specifically recognize microbes as well as different microbial components called TLR ligands. Interaction of TLR and TLR ligands trigger signaling pathways that lead to activation of innate immune cells through secretion of a plethora of inflammatory mediators including cytokines and chemokines [7, 8]. As such, different TLR ligands have been designed to be used as adjuvant system,
cells, mast cells, and antigen-presenting cells (APCs), such as macrophages (MΦ) or dendritic cells (DCs). As such induction of TLR signaling activates APCs to provoke innate immunity and to induce adaptive immunity [6, 7, 20]. More recently, TLR have been found to be expressed on endothelial, epithelial cells and tumor cells, including melanoma [21]. TLRs are fundamen‐ tally located on the plasma membrane except TLR3, TLR7 and TLR9 that are located in the endoplasmic reticulum (ER) [22-24]. Mammalian TLRs include a large family consisting of ten to thirteen different types of TLRs that named simply TLR1 to TLR13 and each one of them
Immunostimulatory Effects of Triggering TLR3 Signaling Pathway — Implication for Cancer Immunotherapy
http://dx.doi.org/10.5772/58575
153
To date, ten human and thirteen murine TLRs have been identified; TLR1 to TLR9 are conserved between the human and mice. However, there are TLRs found in humans and not present in all mammals. For example, TLR10 in humans is present in mice. On the other hand, TLR11, TLR12 and TLR13 in mice are functional, but there is a stop codon in the human TLR11 TLR12 and TLR13 genes, which results in a lack of production of human TLR11 TLR12 and TLR13 [26, 27]. Signaling pathways of these TLRs can be triggered by their specific ligands. TLR ligands (TLRLs) are agents that bind to and activate TLRs. They are encoded in different types of organisms at the cell surface or at the internal cell compartments and they are
TLRs/TLRLs bind results in a cascade of intra-cellular signaling pathways that induce the production of inflammatory cytokines and immune response [7, 8]. TLRLs are recognized by specific TLRs that are expressed on the surface of cells, for example, TLR2 is essential for the recognition of bacterial lipoproteins, lipomannans and lipoteichoic acids [30], while TLR5 detects bacterial flagellin [31]. TLR4 has been reported to recognize bacterial cell wall compo‐ nent lipopolysaccharide (LPS) [30], and TLR11 which recognize a profilin-like protein of *Toxoplasma gondii* [32] and uropathogenic *E.coli* [33]. In contrast, TLRs located within the endoplasmic reticulum (ER) can detect microbial nucleic acids, for example, TLR3 is required for response to virus-derived double-stranded RNA [34] and TLR9 recognize unmethylated CpG motifs [35]. Whereas, TLR7 and TLR8 could recognize small synthetic antiviral molecules
Recent accumulating studies showed that TLRs signaling pathways can be divided into two signaling pathways, the MyD88-dependent pathway which is common to all the TLRs that resulted in the secretion of inflammatory cytokines, and TRIF-dependent pathway that is specific for TLR3 and TLR4 which associated with the stimulation of IFN-β and the maturation
The myeloid differentiation factor 88 (MyD88) dependent responses utilized by all TLRs except TLR3 and its occurring on hetero-or homo-dimerization of the TLRs receptor, upon activation by PAMPs or DAMPs. This dimerization induces the recruitment of adaptor proteins via the
expressed by different types of leucocytes and other cell types [28, 29].
recognize specific microbial components [25].
[36], and single-stranded RNA [37].
**3. TLRs Signaling pathways**
of dendritic cells [38-40].
**3.1. MyD88-dependent pathway**
The inclusion of TLR ligands as a potential new class of adjuvants candidates has enabled the development of clinical effective vaccination strategies against many diseases [9, 10]. Polyi‐ nosinic-polycytidylic acid [Poly(I:C)] and its clinical grade poly-L-lysine (Poly-ICLC; Hilto‐ nol®) represent important members of these immunostimulatory vaccine adjuvants as has been shown in several preclinical and clinical studies [11, 12]. In this review, we will introduce a brief description of the TLR family followed by a description of the signaling pathway of TLR and its role as a linker between innate and adaptive immunity. The structure, expression and function of Poly(I:C) will be reviewed along with its potential application in cancer vaccination, adoptive immunotherapy and chemo-immunotherapy in preclinical and clinical trials.
#### **2. Toll-like receptors (TLRs) and their agonists**
TLRs are a class of transmembrane signaling proteins that play a critical role in initiation and acceleration of innate and adaptive immune responses against different pathogen by recog‐ nizing these pathogens themselves or their products, including proteins, carbohydrates, lipids, and nucleic acids (single-and double-stranded RNA and DNA) [13]. TLRs were discovered in 1985 by Christiane Nüsslein-Volhard as factors involved in the embryonic development and resistance of the fly *Drosophila* to bacterial and fungal infection [14-16]. TLRs are patternrecognition receptors (PRR) with an extracellular leucine-rich repeats (LRRs) domain and a conserved cytoplasmic domain homologous to that of the interleukin-1 receptor (IL-1R), termed the Toll/IL-1R homology (TIR) domain [17, 18]. Such structure of TLRs recognizes pathogen-associated microbial patterns (PAMPs) encoded in various pathogens. These products include lipopolysaccharide (LPS), peptidoglycan (PGN), flagellin, bacterial DNA, viral single and double stranded RNA and synthetic double-stranded RNA such as poly(I:C). TLRs also recognize danger-associated molecular patterns (DAMPs) that are endogenous molecules including intracellular proteins such as heat shock proteins (HSPs) and protein fragments from the extracellular matrix (HMGB1) released from necrotic or dead cells [10, 19]. Given this capability of TLRs to recognize a wide range of microbial products and their synthetic mimic, modulation of these receptors can significantly play a central role in shaping the outcomes of the anti-microbial immunity. Since cancer originates from self-cells resulting in a poor anti-tumor immunity if not tolerance, triggering TLR signaling pathways during cancer immunotherapy by their specific agonists can accentuate the resultant anti-tumor immunity.
TLRs expression profiles differ among tissues and cell types of the innate and adaptive immune system. TLRs are predominantly expressed on T cells, B cells, neutrophils, monocytes, NK cells, mast cells, and antigen-presenting cells (APCs), such as macrophages (MΦ) or dendritic cells (DCs). As such induction of TLR signaling activates APCs to provoke innate immunity and to induce adaptive immunity [6, 7, 20]. More recently, TLR have been found to be expressed on endothelial, epithelial cells and tumor cells, including melanoma [21]. TLRs are fundamen‐ tally located on the plasma membrane except TLR3, TLR7 and TLR9 that are located in the endoplasmic reticulum (ER) [22-24]. Mammalian TLRs include a large family consisting of ten to thirteen different types of TLRs that named simply TLR1 to TLR13 and each one of them recognize specific microbial components [25].
To date, ten human and thirteen murine TLRs have been identified; TLR1 to TLR9 are conserved between the human and mice. However, there are TLRs found in humans and not present in all mammals. For example, TLR10 in humans is present in mice. On the other hand, TLR11, TLR12 and TLR13 in mice are functional, but there is a stop codon in the human TLR11 TLR12 and TLR13 genes, which results in a lack of production of human TLR11 TLR12 and TLR13 [26, 27]. Signaling pathways of these TLRs can be triggered by their specific ligands. TLR ligands (TLRLs) are agents that bind to and activate TLRs. They are encoded in different types of organisms at the cell surface or at the internal cell compartments and they are expressed by different types of leucocytes and other cell types [28, 29].
TLRs/TLRLs bind results in a cascade of intra-cellular signaling pathways that induce the production of inflammatory cytokines and immune response [7, 8]. TLRLs are recognized by specific TLRs that are expressed on the surface of cells, for example, TLR2 is essential for the recognition of bacterial lipoproteins, lipomannans and lipoteichoic acids [30], while TLR5 detects bacterial flagellin [31]. TLR4 has been reported to recognize bacterial cell wall compo‐ nent lipopolysaccharide (LPS) [30], and TLR11 which recognize a profilin-like protein of *Toxoplasma gondii* [32] and uropathogenic *E.coli* [33]. In contrast, TLRs located within the endoplasmic reticulum (ER) can detect microbial nucleic acids, for example, TLR3 is required for response to virus-derived double-stranded RNA [34] and TLR9 recognize unmethylated CpG motifs [35]. Whereas, TLR7 and TLR8 could recognize small synthetic antiviral molecules [36], and single-stranded RNA [37].
### **3. TLRs Signaling pathways**
Toll like receptors (TLRs) are a group of receptors that are expressed mainly in innate immune cells, including macrophages, monocytes, dendritic cells, natural killer cells and mast cells [6]. These TLRs specifically recognize microbes as well as different microbial components called TLR ligands. Interaction of TLR and TLR ligands trigger signaling pathways that lead to activation of innate immune cells through secretion of a plethora of inflammatory mediators including cytokines and chemokines [7, 8]. As such, different TLR ligands have been designed
The inclusion of TLR ligands as a potential new class of adjuvants candidates has enabled the development of clinical effective vaccination strategies against many diseases [9, 10]. Polyi‐ nosinic-polycytidylic acid [Poly(I:C)] and its clinical grade poly-L-lysine (Poly-ICLC; Hilto‐ nol®) represent important members of these immunostimulatory vaccine adjuvants as has been shown in several preclinical and clinical studies [11, 12]. In this review, we will introduce a brief description of the TLR family followed by a description of the signaling pathway of TLR and its role as a linker between innate and adaptive immunity. The structure, expression and function of Poly(I:C) will be reviewed along with its potential application in cancer vaccination, adoptive immunotherapy and chemo-immunotherapy in preclinical and clinical
TLRs are a class of transmembrane signaling proteins that play a critical role in initiation and acceleration of innate and adaptive immune responses against different pathogen by recog‐ nizing these pathogens themselves or their products, including proteins, carbohydrates, lipids, and nucleic acids (single-and double-stranded RNA and DNA) [13]. TLRs were discovered in 1985 by Christiane Nüsslein-Volhard as factors involved in the embryonic development and resistance of the fly *Drosophila* to bacterial and fungal infection [14-16]. TLRs are patternrecognition receptors (PRR) with an extracellular leucine-rich repeats (LRRs) domain and a conserved cytoplasmic domain homologous to that of the interleukin-1 receptor (IL-1R), termed the Toll/IL-1R homology (TIR) domain [17, 18]. Such structure of TLRs recognizes pathogen-associated microbial patterns (PAMPs) encoded in various pathogens. These products include lipopolysaccharide (LPS), peptidoglycan (PGN), flagellin, bacterial DNA, viral single and double stranded RNA and synthetic double-stranded RNA such as poly(I:C). TLRs also recognize danger-associated molecular patterns (DAMPs) that are endogenous molecules including intracellular proteins such as heat shock proteins (HSPs) and protein fragments from the extracellular matrix (HMGB1) released from necrotic or dead cells [10, 19]. Given this capability of TLRs to recognize a wide range of microbial products and their synthetic mimic, modulation of these receptors can significantly play a central role in shaping the outcomes of the anti-microbial immunity. Since cancer originates from self-cells resulting in a poor anti-tumor immunity if not tolerance, triggering TLR signaling pathways during cancer immunotherapy by their specific agonists can accentuate the resultant anti-tumor
TLRs expression profiles differ among tissues and cell types of the innate and adaptive immune system. TLRs are predominantly expressed on T cells, B cells, neutrophils, monocytes, NK
to be used as adjuvant system,
152 Immune Response Activation
**2. Toll-like receptors (TLRs) and their agonists**
trials.
immunity.
Recent accumulating studies showed that TLRs signaling pathways can be divided into two signaling pathways, the MyD88-dependent pathway which is common to all the TLRs that resulted in the secretion of inflammatory cytokines, and TRIF-dependent pathway that is specific for TLR3 and TLR4 which associated with the stimulation of IFN-β and the maturation of dendritic cells [38-40].
#### **3.1. MyD88-dependent pathway**
The myeloid differentiation factor 88 (MyD88) dependent responses utilized by all TLRs except TLR3 and its occurring on hetero-or homo-dimerization of the TLRs receptor, upon activation by PAMPs or DAMPs. This dimerization induces the recruitment of adaptor proteins via the cytoplasmic TIR domain. These adaptor proteins include the TIR-domain containing proteins, TIRAP (TIR-associated protein), TRIF (TIR domain-containing adaptor protein-inducing IFNβ) and TRAM (TRIF-related adaptor molecule), MyD88, Mal (MyD88 adaptor-like protein), which triggers the TLR-mediated signaling pathways [41-45].
Different signal transducer and activator of transcription (STAT) family members can be activated by interferons (IFNs). As a response to both type I IFNs and type II IFN (IFNγ) stimulation, STAT1-STAT1 homodimers can be formed. These homodimers bind to IFNγactivated site (GAS) enhancer elements in the promoters of IFN-stimulated genes, and these results in the induction of genes encoding pro-inflammatory cytokines and apoptotic factors. STAT1 and STAT2 heterodimers, which are activated by binding with IRF9, which in turn migrates to the nucleus to bind to IFN-stimulated response elements (ISREs) and activate
Immunostimulatory Effects of Triggering TLR3 Signaling Pathway — Implication for Cancer Immunotherapy
http://dx.doi.org/10.5772/58575
155
Although, several previous studies have been extensively studied poly(I:C) for over 35 years in humans [50, 51], it is receiving new interest as a crucial component in many new immu‐ nostimulatory combination therapeutic strategies. Poly(I:C) is a synthetic double-stranded RNA (dsRNA) that has recently been identified as a specific ligand for TLR3 [52-55]. dsRNA is a viral product produced by most viruses during their replication cycle, Both viral dsRNA and its artificial mimic, poly(I:C), are potent inducers of type I interferons (IFN-α/β) [56, 57], which is a crucial cytokines that exert anti-viral and immunostimulatory activities for both T and B lymphocytes [58, 59], DCs [11, 60] and activate monocytes to produce CSF, IL-1β, IL-12,
Several studies have been shown that Poly(I:C) induces a strong innate immune response initiated by two types of PRRs; the TLRs and the RIG-I-like receptors (RLRs) which is a family of cytoplasmic RNA helicases that includes RIG-I and MDA-5 [63-67]. In addition, several studies have shown that Poly(I:C) transduces signals which activate the NF-κB and the IFNβ promoter [53, 54] and activates several nuclear and cytoplasmic enzyme systems such as oligoadenylate synthetase (OAS), the dsRNA dependent protein kinase R (PKR), RIG-I Helicase, and MDA5 (melanoma differentiation associated gene) which are implicated in
Poly(I:C) interact with TLR3, that is the specific intracellular recognition system that responds and signals to the intracellular presence of dsRNA and RNA virus infection [68]. TLR3 is mainly expressed on a broad range of antigen-processing cells (APCs), dendritic cell (DC) subsets, fibroblasts, intestinal epithelial [54, 69, 70] monocytes, macrophage (MΦ), mast cells,
has potent pleiotropic immunostimulatory effects on several kinds of immune cells [74-76]. Poly(I:C), a synthetic dsRNA mimic copolymer, is a specific TLR3 agonist [13, 77]. Recently, we and others have reported that in vivo administration of the TLR3 agonist poly(I:C) into naïve mice induced rapid increases in the frequencies of NK cells mainly in the liver and that these effects associate with better antigen specific responses and antitumor efficacy [6, 78-81]. Consistent with earlier studies [81], we have demonstrated that the adjuvant effects of poly(I:C) to the antigen-specific CD8+T cell responses are partially dependent on NK cells through creation of a rapid beneficial cytokine milieu [82, 83]. Furthermore, in vitro treatment of highly
T cells [6, 73]. Several prior studies have shown that poly(I:C)
antiviral and antibacterial genes.
and PGE2 [61, 62].
NK cells [71, 72], CD4<sup>+</sup>
antiviral and antitumor host defenses [7].
and CD8+
**4. Polyinosinic-polycytidylic acid [Poly(I:C)]**
The transcription factor IRF7 is critical regulator for the production of IFN-α/β via MyD88 pathway. IRF7 presents in DCs is very low, and therefore TLR7, TLR8 and TLR9 operate mainly through the NF-κB−IRF5 pathway [10, 46, 47]. It is worth noting that the transducing ability of TLR7, TLR8 and TLR9 may change in favor of the IRF7 pathway. Whereas, TLR4 is located on the cell membrane, activation of TLR4 signal transduction through MyD88/TIRAP and TRAM/TRIF pathways leads to activation of innate immune responses, primarily through IRS serine phosphorylationI
#### **3.2. TRIF-dependent pathway**
The TLR3 ligand (dsRNA) and the TLR4 ligand (LPS) activate the TRIF-dependent pathway in which IRF-3 plays a key role. For TLR3 (dsRNA), activation of MyD88-independent pathways occurs via TRIF (TIR domain-containing adaptor protein-inducing IFN-β). Upon dsRNA binds within endosomes, TLR3 recruits the adaptor TRIF through a TIR–TIR interac‐ tion. TRIF, in turn, recruits RIP1 to activate NF-κB, via TRAFs and the IKK (IκB kinase) complex. TRIF also recruits protein kinases TBK1/IKKε and TRAF3 to activate IRF3/7. Since, IRF-3 can phosphorylate by TBK-1 and IKKε on C-terminal serines, leading to its dimerization and translocation into the nucleus. Active IRF-3 mediates transcription of genes from the IFNβ promoter. The secreted IFN-β binds to the IFN receptor (IFNR), and thus activates tran‐ scription of ISGs, such as IRF-7 [9, 48]. On the other hand, the intracellular dsRNA is recognized by the RNA helicase RIG-I (or MDA-5). Activation of the latter activates TBK-1 and IKKε via its CARD domain and activates IRF-3 as well. IRF-7 further stimulates transcription from the IFN-α and-β promoters in a positive feedback loop. Type I IFNs bind also to their receptors to trigger the JAK–STAT signaling pathway. Whereas, to induce inflammatory cytokines production. TLR4 activation is requires both the MyD88-dependent and independent path‐ ways. This occurs via TIR domain containing adaptors TIRAP, MyD88, TRAM and TRIF, which in turn initiates activation of NFκB, MAPK, and IRF3 allowing its nuclear localization and production of IFN-β [40, 49].
The signaling pathway of TLRs is summarized in Figure 1. TLR7, TLR8 and TLR9 are present in endosomes, which are stimulated by viral ssRNA. Upon activation, TLR7, TLR8 and TLR9 signal through the adaptor MyD88, which in turn leads to phosphorylation and activation of IRF3. Following its activation, TLR3 signals through its adaptor TRIF, and thus activates noncanonical IKK kinases (TBK1/IKKε), which subsequently phosphorylate both IRF3 and IRF7. Active IRF-3 induces transcription from the IFN-β promoter. NF-κ β is also activated by TRIF mediated signaling through canonical IKK kinases (IKKα, β, and γ). RIG-I and MDA5 are located in cytoplasm and can recognize both ssRNA and dsRNA. Intracellular ssRNA and dsRNA are recognized by the RNA helicase RIG-I (or mda-5), which are expressed in most cells, activates TBK-1 and IKKε via its CARD domain through interaction with the mitochond‐ rially located adaptor MAVS.
Different signal transducer and activator of transcription (STAT) family members can be activated by interferons (IFNs). As a response to both type I IFNs and type II IFN (IFNγ) stimulation, STAT1-STAT1 homodimers can be formed. These homodimers bind to IFNγactivated site (GAS) enhancer elements in the promoters of IFN-stimulated genes, and these results in the induction of genes encoding pro-inflammatory cytokines and apoptotic factors. STAT1 and STAT2 heterodimers, which are activated by binding with IRF9, which in turn migrates to the nucleus to bind to IFN-stimulated response elements (ISREs) and activate antiviral and antibacterial genes.
### **4. Polyinosinic-polycytidylic acid [Poly(I:C)]**
cytoplasmic TIR domain. These adaptor proteins include the TIR-domain containing proteins, TIRAP (TIR-associated protein), TRIF (TIR domain-containing adaptor protein-inducing IFNβ) and TRAM (TRIF-related adaptor molecule), MyD88, Mal (MyD88 adaptor-like protein),
The transcription factor IRF7 is critical regulator for the production of IFN-α/β via MyD88 pathway. IRF7 presents in DCs is very low, and therefore TLR7, TLR8 and TLR9 operate mainly through the NF-κB−IRF5 pathway [10, 46, 47]. It is worth noting that the transducing ability of TLR7, TLR8 and TLR9 may change in favor of the IRF7 pathway. Whereas, TLR4 is located on the cell membrane, activation of TLR4 signal transduction through MyD88/TIRAP and TRAM/TRIF pathways leads to activation of innate immune responses, primarily through IRS
The TLR3 ligand (dsRNA) and the TLR4 ligand (LPS) activate the TRIF-dependent pathway in which IRF-3 plays a key role. For TLR3 (dsRNA), activation of MyD88-independent pathways occurs via TRIF (TIR domain-containing adaptor protein-inducing IFN-β). Upon dsRNA binds within endosomes, TLR3 recruits the adaptor TRIF through a TIR–TIR interac‐ tion. TRIF, in turn, recruits RIP1 to activate NF-κB, via TRAFs and the IKK (IκB kinase) complex. TRIF also recruits protein kinases TBK1/IKKε and TRAF3 to activate IRF3/7. Since, IRF-3 can phosphorylate by TBK-1 and IKKε on C-terminal serines, leading to its dimerization and translocation into the nucleus. Active IRF-3 mediates transcription of genes from the IFNβ promoter. The secreted IFN-β binds to the IFN receptor (IFNR), and thus activates tran‐ scription of ISGs, such as IRF-7 [9, 48]. On the other hand, the intracellular dsRNA is recognized by the RNA helicase RIG-I (or MDA-5). Activation of the latter activates TBK-1 and IKKε via its CARD domain and activates IRF-3 as well. IRF-7 further stimulates transcription from the IFN-α and-β promoters in a positive feedback loop. Type I IFNs bind also to their receptors to trigger the JAK–STAT signaling pathway. Whereas, to induce inflammatory cytokines production. TLR4 activation is requires both the MyD88-dependent and independent path‐ ways. This occurs via TIR domain containing adaptors TIRAP, MyD88, TRAM and TRIF, which in turn initiates activation of NFκB, MAPK, and IRF3 allowing its nuclear localization and
The signaling pathway of TLRs is summarized in Figure 1. TLR7, TLR8 and TLR9 are present in endosomes, which are stimulated by viral ssRNA. Upon activation, TLR7, TLR8 and TLR9 signal through the adaptor MyD88, which in turn leads to phosphorylation and activation of IRF3. Following its activation, TLR3 signals through its adaptor TRIF, and thus activates noncanonical IKK kinases (TBK1/IKKε), which subsequently phosphorylate both IRF3 and IRF7. Active IRF-3 induces transcription from the IFN-β promoter. NF-κ β is also activated by TRIF mediated signaling through canonical IKK kinases (IKKα, β, and γ). RIG-I and MDA5 are located in cytoplasm and can recognize both ssRNA and dsRNA. Intracellular ssRNA and dsRNA are recognized by the RNA helicase RIG-I (or mda-5), which are expressed in most cells, activates TBK-1 and IKKε via its CARD domain through interaction with the mitochond‐
which triggers the TLR-mediated signaling pathways [41-45].
serine phosphorylationI
154 Immune Response Activation
**3.2. TRIF-dependent pathway**
production of IFN-β [40, 49].
rially located adaptor MAVS.
Although, several previous studies have been extensively studied poly(I:C) for over 35 years in humans [50, 51], it is receiving new interest as a crucial component in many new immu‐ nostimulatory combination therapeutic strategies. Poly(I:C) is a synthetic double-stranded RNA (dsRNA) that has recently been identified as a specific ligand for TLR3 [52-55]. dsRNA is a viral product produced by most viruses during their replication cycle, Both viral dsRNA and its artificial mimic, poly(I:C), are potent inducers of type I interferons (IFN-α/β) [56, 57], which is a crucial cytokines that exert anti-viral and immunostimulatory activities for both T and B lymphocytes [58, 59], DCs [11, 60] and activate monocytes to produce CSF, IL-1β, IL-12, and PGE2 [61, 62].
Several studies have been shown that Poly(I:C) induces a strong innate immune response initiated by two types of PRRs; the TLRs and the RIG-I-like receptors (RLRs) which is a family of cytoplasmic RNA helicases that includes RIG-I and MDA-5 [63-67]. In addition, several studies have shown that Poly(I:C) transduces signals which activate the NF-κB and the IFNβ promoter [53, 54] and activates several nuclear and cytoplasmic enzyme systems such as oligoadenylate synthetase (OAS), the dsRNA dependent protein kinase R (PKR), RIG-I Helicase, and MDA5 (melanoma differentiation associated gene) which are implicated in antiviral and antitumor host defenses [7].
Poly(I:C) interact with TLR3, that is the specific intracellular recognition system that responds and signals to the intracellular presence of dsRNA and RNA virus infection [68]. TLR3 is mainly expressed on a broad range of antigen-processing cells (APCs), dendritic cell (DC) subsets, fibroblasts, intestinal epithelial [54, 69, 70] monocytes, macrophage (MΦ), mast cells, NK cells [71, 72], CD4<sup>+</sup> and CD8+ T cells [6, 73]. Several prior studies have shown that poly(I:C) has potent pleiotropic immunostimulatory effects on several kinds of immune cells [74-76].
Poly(I:C), a synthetic dsRNA mimic copolymer, is a specific TLR3 agonist [13, 77]. Recently, we and others have reported that in vivo administration of the TLR3 agonist poly(I:C) into naïve mice induced rapid increases in the frequencies of NK cells mainly in the liver and that these effects associate with better antigen specific responses and antitumor efficacy [6, 78-81]. Consistent with earlier studies [81], we have demonstrated that the adjuvant effects of poly(I:C) to the antigen-specific CD8+T cell responses are partially dependent on NK cells through creation of a rapid beneficial cytokine milieu [82, 83]. Furthermore, in vitro treatment of highly purified NK cells with poly(I:C) significantly augmented NK cell-mediated cytotoxicity and up-regulated their expression of the activation marker CD69 [71]. Of particular interest, similar to viral infection, poly(I:C) treatment has been found to preferentially induce recruitment and activation of hepatic NK cells [79] and their trafficking to spleen coincided with enhanced cytokine expression [80]. These studies may explain the increased antitumor effects against metastasis observed in tumor models after poly(I:C) treatment [84].
recently reported a rapid induction of type I IFNs and activation of DCs with decreases in the levels of Treg cells after CTX treatment; resulting in augmented post vaccination responses which were further improved by addition of poly(I:C) to vaccination [111]. We have found also a marked expansion of immature DCs during the recovery phase after treatment with the anticancer drug cyclophosphamide (CTX), indicating to the mobilizing effect of CTX for DCs. Since DCs are equipped with TLRs that sense different microbial products such as poly(I:C), activation of these immature DCs in vivo may augment memory T cell responses. [85]. Recent studies, including ours have demonstrated the adjuvant effects of TLRLs to ACT in lympho‐ depleted hosts [112-114]. These studies suggest that combinatorial treatments with chemo‐ therapy/immunotherapy and ACT can markedly improve memory T cell responses. Our observation of expansion of DCs post CTX therapy is a suitable model to dissect the optimal timing of the adjuvant effects of TLRLs to post vaccination responses of ACT in a lymphode‐
Immunostimulatory Effects of Triggering TLR3 Signaling Pathway — Implication for Cancer Immunotherapy
Specifically, our recent studies showed that treatment of a recipient host with the non mye‐
with a quick activation of DCs in the lymphopenic phase [114]. We also found that CTX induces increases in the numbers of immature DCs during the recovery phase from days 9-16, peaking on day 12 [115]. These DCs demonstrated normal phagocytic ability in vitro and antigen uptake in vivo. Administration of poly(I:C) at the peak of DC expansion resulted in induction of rapid inflammatory milieu which associated with significant increases in the numbers of activated
tumor gp100 antigen on B16 melanoma, we demonstrated that vaccination of CTX-lympho‐ depleted mice with the MHC class-I melanoma gp100 peptide and poly(I:C) during the lymphopenic phase (day 2) and at the peak of DC expansion (day 12) resulted in significant augmentation in the number of activated DCs in lymph nodes with a temporal increase in the expansion of pmel-1 cells. Conditional depletion studies of DCs at day 12 before revaccination of CTX-treated recipient mice revealed that DCs are required for the optimal pmel-1 cell responses. Importantly, the therapeutic anti-tumor effects of the enhanced pmel-1 cell responses were demonstrated toward an advanced tumor of the poorly immunogenic B16 melanoma, indicating the clinical significance of our observation. Furthermore, we have found improvement in DCs expansion when CTX treatment was followed by a daily treatment with 5μg/mouse G-CSF for 5 days [6], without altering the efficacy of our prime-boost vaccination
Treg cells express different TLRs [117] and can be activated by certain TLR agonists such as TLR4 [118-120], and TLR5 [121], and can be inactivated by TLR2 and TLR7/8 agonists [122, 123]. Indeed, total body irradiation and CTX-preconditioned hosts demonstrated that inter‐ ference with Treg cell activities can enhance tumor-specific T cell responses and anti-tumor immunity [124], [125, 126]. Our preliminary studies and previous studies [105, 125-130], however, showed that although Treg cells decrease in numbers during the lymphopenic phase after CTX treatment they recover to their normal levels at the recovery phase. These recovered Treg cells may interfere with the efficacy of vaccination at recovery phase unless the vaccination
T cell responses, which were associated
http://dx.doi.org/10.5772/58575
157
T cells can recognize the natural
loablative dose of CTX augments post vaccination CD8+
Using the pmel-1 TCR transgenic mouse model, in which CD8+
(CCR7highCD40high) DCs in lymph nodes.
with hgp100/poly(I:C) [116].
pleted host.
#### **5. Poly(I:C) and anti-tumor Immunity**
A fundamental difference between tumor and microbes is that only the latter encode products (signatures) that are recognized as "danger signals" by pathogen recognition receptors (PPRs) expressed in the innate immune cells. [85] In most cases, immune system can mount vigorous immune responses against microbes, but not against cancer. Therefore, the challenge in cancer immunotherapy is how to manipulate the body's own immune system to fight cancer. [86] Mimicking the anti-microbial immunity, recent preclinical and clinical studies have establish‐ ed that provision of TLRL adjuvant system systemically or into tumor environment itself profoundly awaken the cross talk between innate and adaptive immunity, driving generation of efficacious anti-tumor immunity. [87] For instance, the addition of CpG DNA (typical TLR9L) to a melanoma vaccine resulted in effective cytolytic responses. [88] Imiquimod, a synthetic TLR7/8L, has been successfully used in the treatment of basal cell carcinoma [89, 90] and to enhance the immunogenicity of vaccine containing Flt3 ligand and a melanoma peptide. [91] MPL, a TLR4L, has been used as an adjuvant in clinical trials of vaccines against melanoma, glioma and pancreatic and colorectal carcinoma, inducing substantial tumorspecific immunity in response to vaccination [92, 93]. We have reported recently that coadministration of the TLR3L poly(I:C) in vivo with peptide vaccination established functional effector/memory responses [94] by induction of NK-driven beneficial cytokine milieu and DC activation. [81] Although the impact of both of TLRLs at the time of Ag priming has been the focus of several studies, their impacts at the time of Ag recall on the tumor-specific memory responses have not been actively investigated.
Since, cancer cells do not encode danger signals like virus, generation of anti-tumor memory T cells in the tumor setting would require a large pool of highly activated DCs. In response to TLRLs, resident immature DCs at the site of vaccination undergo a maturation program and migrate to the draining lymph nodes (LNs) [72, 95, 96]. Therefore, utilizing TLRLs is a potential approach to induce activation of immature DCs and maximize their contribution to memory cell responses. In line with this notion, recent studies have established the adjuvant effects of several TLRLs, in particular, TLR3L (poly(I:C), TLR7/8L (Imiquimod), and TLR9L (CpG) to the anti-tumor CD8+ T cell responses [81, 94, 97-101]. However, most of these studies tested the adjuvant effects of TLRLs in lymphodepleted hosts, and those few studies that utilized lymphodepleted hosts did not use adoptive cell therapy (ACT) system [98, 102-107].
Because the frequency of DCs in steady state condition is low, growth factors in particular Flt3L and G-CSF have been used to mobilize DCs in vivo [108-110]. With this regard, we have recently reported a rapid induction of type I IFNs and activation of DCs with decreases in the levels of Treg cells after CTX treatment; resulting in augmented post vaccination responses which were further improved by addition of poly(I:C) to vaccination [111]. We have found also a marked expansion of immature DCs during the recovery phase after treatment with the anticancer drug cyclophosphamide (CTX), indicating to the mobilizing effect of CTX for DCs. Since DCs are equipped with TLRs that sense different microbial products such as poly(I:C), activation of these immature DCs in vivo may augment memory T cell responses. [85]. Recent studies, including ours have demonstrated the adjuvant effects of TLRLs to ACT in lympho‐ depleted hosts [112-114]. These studies suggest that combinatorial treatments with chemo‐ therapy/immunotherapy and ACT can markedly improve memory T cell responses. Our observation of expansion of DCs post CTX therapy is a suitable model to dissect the optimal timing of the adjuvant effects of TLRLs to post vaccination responses of ACT in a lymphode‐ pleted host.
purified NK cells with poly(I:C) significantly augmented NK cell-mediated cytotoxicity and up-regulated their expression of the activation marker CD69 [71]. Of particular interest, similar to viral infection, poly(I:C) treatment has been found to preferentially induce recruitment and activation of hepatic NK cells [79] and their trafficking to spleen coincided with enhanced cytokine expression [80]. These studies may explain the increased antitumor effects against
A fundamental difference between tumor and microbes is that only the latter encode products (signatures) that are recognized as "danger signals" by pathogen recognition receptors (PPRs) expressed in the innate immune cells. [85] In most cases, immune system can mount vigorous immune responses against microbes, but not against cancer. Therefore, the challenge in cancer immunotherapy is how to manipulate the body's own immune system to fight cancer. [86] Mimicking the anti-microbial immunity, recent preclinical and clinical studies have establish‐ ed that provision of TLRL adjuvant system systemically or into tumor environment itself profoundly awaken the cross talk between innate and adaptive immunity, driving generation of efficacious anti-tumor immunity. [87] For instance, the addition of CpG DNA (typical TLR9L) to a melanoma vaccine resulted in effective cytolytic responses. [88] Imiquimod, a synthetic TLR7/8L, has been successfully used in the treatment of basal cell carcinoma [89, 90] and to enhance the immunogenicity of vaccine containing Flt3 ligand and a melanoma peptide. [91] MPL, a TLR4L, has been used as an adjuvant in clinical trials of vaccines against melanoma, glioma and pancreatic and colorectal carcinoma, inducing substantial tumorspecific immunity in response to vaccination [92, 93]. We have reported recently that coadministration of the TLR3L poly(I:C) in vivo with peptide vaccination established functional effector/memory responses [94] by induction of NK-driven beneficial cytokine milieu and DC activation. [81] Although the impact of both of TLRLs at the time of Ag priming has been the focus of several studies, their impacts at the time of Ag recall on the tumor-specific memory
Since, cancer cells do not encode danger signals like virus, generation of anti-tumor memory T cells in the tumor setting would require a large pool of highly activated DCs. In response to TLRLs, resident immature DCs at the site of vaccination undergo a maturation program and migrate to the draining lymph nodes (LNs) [72, 95, 96]. Therefore, utilizing TLRLs is a potential approach to induce activation of immature DCs and maximize their contribution to memory cell responses. In line with this notion, recent studies have established the adjuvant effects of several TLRLs, in particular, TLR3L (poly(I:C), TLR7/8L (Imiquimod), and TLR9L (CpG) to the
adjuvant effects of TLRLs in lymphodepleted hosts, and those few studies that utilized
Because the frequency of DCs in steady state condition is low, growth factors in particular Flt3L and G-CSF have been used to mobilize DCs in vivo [108-110]. With this regard, we have
lymphodepleted hosts did not use adoptive cell therapy (ACT) system [98, 102-107].
T cell responses [81, 94, 97-101]. However, most of these studies tested the
metastasis observed in tumor models after poly(I:C) treatment [84].
**5. Poly(I:C) and anti-tumor Immunity**
156 Immune Response Activation
responses have not been actively investigated.
anti-tumor CD8+
Specifically, our recent studies showed that treatment of a recipient host with the non mye‐ loablative dose of CTX augments post vaccination CD8+ T cell responses, which were associated with a quick activation of DCs in the lymphopenic phase [114]. We also found that CTX induces increases in the numbers of immature DCs during the recovery phase from days 9-16, peaking on day 12 [115]. These DCs demonstrated normal phagocytic ability in vitro and antigen uptake in vivo. Administration of poly(I:C) at the peak of DC expansion resulted in induction of rapid inflammatory milieu which associated with significant increases in the numbers of activated (CCR7highCD40high) DCs in lymph nodes.
Using the pmel-1 TCR transgenic mouse model, in which CD8+ T cells can recognize the natural tumor gp100 antigen on B16 melanoma, we demonstrated that vaccination of CTX-lympho‐ depleted mice with the MHC class-I melanoma gp100 peptide and poly(I:C) during the lymphopenic phase (day 2) and at the peak of DC expansion (day 12) resulted in significant augmentation in the number of activated DCs in lymph nodes with a temporal increase in the expansion of pmel-1 cells. Conditional depletion studies of DCs at day 12 before revaccination of CTX-treated recipient mice revealed that DCs are required for the optimal pmel-1 cell responses. Importantly, the therapeutic anti-tumor effects of the enhanced pmel-1 cell responses were demonstrated toward an advanced tumor of the poorly immunogenic B16 melanoma, indicating the clinical significance of our observation. Furthermore, we have found improvement in DCs expansion when CTX treatment was followed by a daily treatment with 5μg/mouse G-CSF for 5 days [6], without altering the efficacy of our prime-boost vaccination with hgp100/poly(I:C) [116].
Treg cells express different TLRs [117] and can be activated by certain TLR agonists such as TLR4 [118-120], and TLR5 [121], and can be inactivated by TLR2 and TLR7/8 agonists [122, 123]. Indeed, total body irradiation and CTX-preconditioned hosts demonstrated that inter‐ ference with Treg cell activities can enhance tumor-specific T cell responses and anti-tumor immunity [124], [125, 126]. Our preliminary studies and previous studies [105, 125-130], however, showed that although Treg cells decrease in numbers during the lymphopenic phase after CTX treatment they recover to their normal levels at the recovery phase. These recovered Treg cells may interfere with the efficacy of vaccination at recovery phase unless the vaccination is co-administered with a potent adjuvant that can block Treg cell function. Indeed, induction of maturation of DCs by certain TLR agonists has been found to control Treg cell function in vivo [122, 131, 132]. Therefore, it is possible that the enhanced pmel-1 cell expansion and antitumor effects of pmel-1 cells to peptide/poly(I:C) revaccination on day 12 post CTX treatment involve blocking of Treg cell.
tion of the TLR3 receptor which in turn induces the activation of NF-κB and the production of type I interferons that induces strong antiviral and antineoplastic effects accompanied by activation of CD8+T and NK cells [53, 147]. In addition, poly(I:C) is shown to serve as an
Immunostimulatory Effects of Triggering TLR3 Signaling Pathway — Implication for Cancer Immunotherapy
Hiltonol® or Poly-ICLC is a synthetic, nuclease-resistant, hydrophilic complex of polyino‐ sinic and polycytidylic acid, stabilized with poly-L-lysine and carboxymethyl cellulose. Poly-ICLC is IFN-β inducer, which is known to steady the blood-brain-barrier (BBB) and minimize cellular infiltration into damaged brain regions following stroke [149, 150]. Several clinical trials were conducted to utilize poly-ICLC as an IFN inducer in cancer patients, the first clinical trial (phase 1) were conducted to determine the maximum tolerated dose (MTD)
patients [51]. Another study was conducted to determine the immunomodulatory effect of Poly-ICLC in cancer patients and it showed that there was no detectable serum IFN in patients that received 1 mg/m2 Poly-ICLC by IM injection [151]. In contrast, IFN was detectable in the serum of patients that received 4 mg/m2 Poly-ICLC by IV injection. Then, phase I and II clinical trials were conducted in patients with many types of cancers including leukemia, lymphoma, brain tumors, myeloma, juvenile laryngeal papillomatosis, renal cell
In most of these early clinical trials, about 6 mg/m2 Poly-ICLC was generally used intrave‐ nously. Fever, often with temperatures greater than 40°C, was a universal adverse event in the trials and was the primary dose-limiting factor. Other common adverse events reported in these trials included nausea, vomiting, hypotension, thrombocytopenia, leukopenia, arthral‐ gia, myalgia, and fatigue. Additionally, very few objective responses were reported in these clinical trials. Due to its toxicity and relative ineffectiveness, utilizing Poly-ICLC was deserted after the availability of recombinant IFN. It was subsequently determined that lower doses (10 to 50 mg/kg) of Poly-ICLC resulted in a broader host defense stimulation, a potent adjuvant effect, and a specific antiretroviral effect mediated by the 2'5'OAS and PKR nuclear enzyme systems [161]. Consequently, Poly-ICLC is currently being developed for use only at doses up to 50 mg/kg. The poly-ICLC treatment prior to exposure to oxygen-glucose deprivation maintained the paracellular and transcellular transport across the endothelium and attenuated the drop in transendothelial electric resistance by enhancing IFN-β mRNA expression in astrocytes and microglia [162]. Recent study utilized poly-ICLC in activation of TLR3 signal‐ ing, which exerts a beneficial effect on NK cells, resulting in the increased cetuximab-depend‐
), which induced effective response in cancer
http://dx.doi.org/10.5772/58575
159
adjuvant to induce protective CD4+T cell responses against HIV [148].
carcinoma, breast cancer, ovarian cancer, and melanoma [152-160].
**7. Poly(I:C) and clinical trials**
and it found that the MTD was (12 mg/m2
ent lysis of head and neck cancer (HNC) cells [163].
**7.1. Hiltonol®**
The potent anti-tumor efficacy of prime-boost vaccination with tumor antigen/poly(I:C) at precise time points post CTX therapy would lead to its potential application in the clinical setting. We would envision two treatment protocols for the clinical application of this primeboost vaccination with tumor peptide and poly(I:C) on cancer (e.g. melanoma) patients. 1) Patients can be with a lymphodepletion dose of CTX followed by G-CSF to correct leucopenia, and vaccinate the patient with a candidate tumor Ag (gp100, MART, or TRP2) with co-administration of the clinical grade form of poly(I:C) such as Hiltonol®. Vaccina‐ tion and Hiltonol® treatments can be repeated during the recovery from leucopenia (after 10-15 days of chemotherapy) when the numbers of DCs are increased. 2) The same as above but with adoptive transfer of peripheral blood mononuclear cells harvested from the cancer patient prior chemotherapy. These cells can be stimulated in vitro with 1ng/mL of the tumor antigens in the presence of 10ng/mL IL-12 for 3-5 days as we recently described in our preclinical model [133].
#### **6. Poly(I:C) and antiviral immune response**
The innate immune response is the first barrier against the invading pathogens and viruses and it responds through activating inflammatory and antiviral defense mechanisms by inducing IFNs-α/β against the viruses [134]. Several in vitro and in vivo studies have demon‐ strated that the main effects of poly(I:C) is the induction of IFNs-α/β, which play a crucial role in innate anti-viral response [135, 136].Several studies have been extensively demonstrated that poly(I:C) triggers the activation of PKR and other kinases which followed by the phos‐ phorylation of the substrates of these enzymes that results in the subsequent translocation of transcription factors, NF-ҚB and IRF-3, to the nucleus where then they bind to the IFN-β promoter to form a transcription complex that induce IFN-β production [137-140]. It has been thought that the proximal inducer of IFNs-α/β is intracellular dsRNA generated as an intermediate during viral replication [53].
Recently, it has been shown that dsRNA and ssRNA molecules are recognized as intermediate by TLRs that is expressed on DCs, NK cells, MΦ and epithelium during virus replication [141]. Several studies have been extensively demonstrated that Poly(I:C) plays a critical role in inducing innate immune response against many viruses, such as influenza virus [142], human respiratory syncytial virus (RSV) [143], herpes simplex virus 2 (HSV-2) [144], and murine cytomegalovirus [25]. Several in vitro and in vivo studies have been shown that retinoic-acidinducible gene I (RIG-I) and melanoma-differentiation-associated gene 5 (MDA-5), are the key in the detection of viral dsRNA and Poly(I:C) in the cytosol and subsequent eradication of the replicating viral genomes [66, 145, 146]. Treatment of animals with poly(I:C) results inactiva‐ tion of the TLR3 receptor which in turn induces the activation of NF-κB and the production of type I interferons that induces strong antiviral and antineoplastic effects accompanied by activation of CD8+T and NK cells [53, 147]. In addition, poly(I:C) is shown to serve as an adjuvant to induce protective CD4+T cell responses against HIV [148].
#### **7. Poly(I:C) and clinical trials**
#### **7.1. Hiltonol®**
is co-administered with a potent adjuvant that can block Treg cell function. Indeed, induction of maturation of DCs by certain TLR agonists has been found to control Treg cell function in vivo [122, 131, 132]. Therefore, it is possible that the enhanced pmel-1 cell expansion and antitumor effects of pmel-1 cells to peptide/poly(I:C) revaccination on day 12 post CTX treatment
The potent anti-tumor efficacy of prime-boost vaccination with tumor antigen/poly(I:C) at precise time points post CTX therapy would lead to its potential application in the clinical setting. We would envision two treatment protocols for the clinical application of this primeboost vaccination with tumor peptide and poly(I:C) on cancer (e.g. melanoma) patients. 1) Patients can be with a lymphodepletion dose of CTX followed by G-CSF to correct leucopenia, and vaccinate the patient with a candidate tumor Ag (gp100, MART, or TRP2) with co-administration of the clinical grade form of poly(I:C) such as Hiltonol®. Vaccina‐ tion and Hiltonol® treatments can be repeated during the recovery from leucopenia (after 10-15 days of chemotherapy) when the numbers of DCs are increased. 2) The same as above but with adoptive transfer of peripheral blood mononuclear cells harvested from the cancer patient prior chemotherapy. These cells can be stimulated in vitro with 1ng/mL of the tumor antigens in the presence of 10ng/mL IL-12 for 3-5 days as we recently described in our
The innate immune response is the first barrier against the invading pathogens and viruses and it responds through activating inflammatory and antiviral defense mechanisms by inducing IFNs-α/β against the viruses [134]. Several in vitro and in vivo studies have demon‐ strated that the main effects of poly(I:C) is the induction of IFNs-α/β, which play a crucial role in innate anti-viral response [135, 136].Several studies have been extensively demonstrated that poly(I:C) triggers the activation of PKR and other kinases which followed by the phos‐ phorylation of the substrates of these enzymes that results in the subsequent translocation of transcription factors, NF-ҚB and IRF-3, to the nucleus where then they bind to the IFN-β promoter to form a transcription complex that induce IFN-β production [137-140]. It has been thought that the proximal inducer of IFNs-α/β is intracellular dsRNA generated as an
Recently, it has been shown that dsRNA and ssRNA molecules are recognized as intermediate by TLRs that is expressed on DCs, NK cells, MΦ and epithelium during virus replication [141]. Several studies have been extensively demonstrated that Poly(I:C) plays a critical role in inducing innate immune response against many viruses, such as influenza virus [142], human respiratory syncytial virus (RSV) [143], herpes simplex virus 2 (HSV-2) [144], and murine cytomegalovirus [25]. Several in vitro and in vivo studies have been shown that retinoic-acidinducible gene I (RIG-I) and melanoma-differentiation-associated gene 5 (MDA-5), are the key in the detection of viral dsRNA and Poly(I:C) in the cytosol and subsequent eradication of the replicating viral genomes [66, 145, 146]. Treatment of animals with poly(I:C) results inactiva‐
involve blocking of Treg cell.
158 Immune Response Activation
preclinical model [133].
**6. Poly(I:C) and antiviral immune response**
intermediate during viral replication [53].
Hiltonol® or Poly-ICLC is a synthetic, nuclease-resistant, hydrophilic complex of polyino‐ sinic and polycytidylic acid, stabilized with poly-L-lysine and carboxymethyl cellulose. Poly-ICLC is IFN-β inducer, which is known to steady the blood-brain-barrier (BBB) and minimize cellular infiltration into damaged brain regions following stroke [149, 150]. Several clinical trials were conducted to utilize poly-ICLC as an IFN inducer in cancer patients, the first clinical trial (phase 1) were conducted to determine the maximum tolerated dose (MTD) and it found that the MTD was (12 mg/m2 ), which induced effective response in cancer patients [51]. Another study was conducted to determine the immunomodulatory effect of Poly-ICLC in cancer patients and it showed that there was no detectable serum IFN in patients that received 1 mg/m2 Poly-ICLC by IM injection [151]. In contrast, IFN was detectable in the serum of patients that received 4 mg/m2 Poly-ICLC by IV injection. Then, phase I and II clinical trials were conducted in patients with many types of cancers including leukemia, lymphoma, brain tumors, myeloma, juvenile laryngeal papillomatosis, renal cell carcinoma, breast cancer, ovarian cancer, and melanoma [152-160].
In most of these early clinical trials, about 6 mg/m2 Poly-ICLC was generally used intrave‐ nously. Fever, often with temperatures greater than 40°C, was a universal adverse event in the trials and was the primary dose-limiting factor. Other common adverse events reported in these trials included nausea, vomiting, hypotension, thrombocytopenia, leukopenia, arthral‐ gia, myalgia, and fatigue. Additionally, very few objective responses were reported in these clinical trials. Due to its toxicity and relative ineffectiveness, utilizing Poly-ICLC was deserted after the availability of recombinant IFN. It was subsequently determined that lower doses (10 to 50 mg/kg) of Poly-ICLC resulted in a broader host defense stimulation, a potent adjuvant effect, and a specific antiretroviral effect mediated by the 2'5'OAS and PKR nuclear enzyme systems [161]. Consequently, Poly-ICLC is currently being developed for use only at doses up to 50 mg/kg. The poly-ICLC treatment prior to exposure to oxygen-glucose deprivation maintained the paracellular and transcellular transport across the endothelium and attenuated the drop in transendothelial electric resistance by enhancing IFN-β mRNA expression in astrocytes and microglia [162]. Recent study utilized poly-ICLC in activation of TLR3 signal‐ ing, which exerts a beneficial effect on NK cells, resulting in the increased cetuximab-depend‐ ent lysis of head and neck cancer (HNC) cells [163].
#### **7.2. Ampligen®**
Ampligen® is known as [poly(I)-poly(C12U)] that is composed of poly(IC) with a U mismatch at every 12th base of the C strand. The main effect of Ampligen® is the induction of Th1 response. This has been demonstrated in delayed-type hypersensitivity reactions and in current clinical studies with human immunodeficiency virus-infected patients. Other cells targeted are NK cells, cytotoxic CD8+ T cells, and LAK-NK cells. Despite initial trials in the 1990s, little is known about Ampligen®, since; it is not routinely used in clinical practice. However, due to its notable and potent antiviral effects, Ampligen® has been generally well tolerated in randomized clinical trials [164, 165]. Since, Ampligen® did not consistent‐ ly produce antiviral effects, but its antiviral activity was always seen after therapeutic treatment. Therefore, it was assumed that Ampligen have exquisite specificity for TLR3 [166, 167]. Early studies showed that treatment of chronically HBV-infected chimpanzees with polyICLC transiently reduced the levels of serum viral DNA, HBsAg and HBeAg [168]. In two independent studies, transient suppression of DHBV in ducks by Ampligen® was also observed [169, 170]. In vitro studies, Ampligen and zidovudine have combined: to act in synergy against HIV-1-infected cell lines [171, 172] and enabling restoration of the immune function in HIV-infected individuals. Wither through maintaining a stable or raising CD4 count, increasing delayed-type hypersensitivity reaction, and reducing rate of progression to AIDS [173].
mouse TLR3 is highly expressed in macrophages, but in both species, its expression is induced
Immunostimulatory Effects of Triggering TLR3 Signaling Pathway — Implication for Cancer Immunotherapy
http://dx.doi.org/10.5772/58575
161
Unlike other TLRs, TLR3 is unique, since, it is not required MyD88 for signaling, and instead, TRIF is the critical adaptor protein for its signaling. Whereas TLR3 displays a distinct intra‐ cellular localization compared with other TLRs, TLR7 and TLR9 trigger a quick, but short IFN response, whereas TLR3 might be more important for a prolonged response and the initiation of the adaptive immune response. Poly(I:C) interact with TLR3 that responds and signals to the intracellular presence of dsRNA and RNA virus infection. Poly(I:C), is strong inducers of IFN-α/β, which is a crucial cytokines that exert anti-viral and immunostimulatory activities for both T and B lymphocytes, DCs and activate monocytes to produce CSF, IL-1β, IL-12, and PGE2. It's also transduces signals which activate the NF-κB and the IFN-β promoter and activates several nuclear and cytoplasmic enzyme systems such as OAS, PKR, RIG-I Helicase,
The above mentioned pleotropic effects of the TLR3 agonist Poly(I:C) makes it as a candidate adjuvant system for anti-viral and anti-tumor immune responses, in particular its clinical forms Hiltonol® and Ampligen® induce similar immunomodulatory effects with accepted side
The authors would like to thank Prof. Karolin K. Abdul-Aziz and Dr. Muobarak J. Tuorkey (Zoology Department, Faculty of Science, Damanhour University, Egypt) for their kind help
1 Center of Excellence in Cancer Research (CECR), Tanta University Educational Hospital,
2 Immunology and Biotechnology Unit, Zoology Department, Faculty of Science, Tanta Uni‐
4 Pediatric Hematology and Oncology Unit, Faculty of Medicine, Tanta University, Egypt
and MDA5 which are implicated in antiviral and antitumor host defenses.
Mohamed Labib Salem1,2\*, Said M. Hammad1,3, Mohamed R. El-shanshory1,4,
3 Clinical Pathology Department, Faculty of Medicine, Tanta University, Egypt
\*Address all correspondence to: mohamed.labib@science.tanta.edu.eg
by IFN.
effects.
**Acknowledgements**
during writing this chapter.
Mohamed A. Attia1,3 and Abdel-Aziz A. Zidan1,2
Medical Campus, Tanta University, Egypt
**Author details**
versity, Egypt
More recently, Nicodemus et al (2010) evaluated polyIC12U and confirmed the potential of TLR3 stimulation with Ampligen® in enhancement bioactivity of cancer immunotherapies [166]. Their results revealed that Ampligen® is a potent inducer for the dendritic cell matura‐ tion and local cytokines producing in culture systems including IL-4, IL-6, IL-12p70, IFN g, MIP-1a and TNF α. Interestingly, treatment of wild-type mice with Ampligen® 24 hours following infection with a lethal viral inoculation was fully protective, whereas dosing 4 h prior to inoculation is ineffective in preventing mortality [167]. Intranasal Ampligen® administered to mice in conjunction with prototype avian flu vaccines, greatly enhances the cellular and humoral immunity achievable with the vaccine and also protects mice from lethal infection [174].
#### **8. Conclusion**
Since the discovery of the important role of TLRs in innate immunity and the initiation of an immune response that follows the activation of antigen-specific acquired immunity, rapid progress has been made on our understanding of the molecular mechanisms of TLRs. The inclusion of TLR ligands as a potential new class of adjuvants candidates has enabled the development of clinical effective vaccination strategies against many diseases. Poly(I:C) and Poly-ICLC represent two important member of these immunostimulatory vaccine adjuvants in mice, nonhuman primates and in human. TLR3 is specifically expressed in quite cell-type and species; it is expressed in specific myeloid cells, vascular endothelial cells and airway epithelial cells among others. Human TLR3 is highly expressed in immature DCs whereas mouse TLR3 is highly expressed in macrophages, but in both species, its expression is induced by IFN.
Unlike other TLRs, TLR3 is unique, since, it is not required MyD88 for signaling, and instead, TRIF is the critical adaptor protein for its signaling. Whereas TLR3 displays a distinct intra‐ cellular localization compared with other TLRs, TLR7 and TLR9 trigger a quick, but short IFN response, whereas TLR3 might be more important for a prolonged response and the initiation of the adaptive immune response. Poly(I:C) interact with TLR3 that responds and signals to the intracellular presence of dsRNA and RNA virus infection. Poly(I:C), is strong inducers of IFN-α/β, which is a crucial cytokines that exert anti-viral and immunostimulatory activities for both T and B lymphocytes, DCs and activate monocytes to produce CSF, IL-1β, IL-12, and PGE2. It's also transduces signals which activate the NF-κB and the IFN-β promoter and activates several nuclear and cytoplasmic enzyme systems such as OAS, PKR, RIG-I Helicase, and MDA5 which are implicated in antiviral and antitumor host defenses.
The above mentioned pleotropic effects of the TLR3 agonist Poly(I:C) makes it as a candidate adjuvant system for anti-viral and anti-tumor immune responses, in particular its clinical forms Hiltonol® and Ampligen® induce similar immunomodulatory effects with accepted side effects.
#### **Acknowledgements**
**7.2. Ampligen®**
160 Immune Response Activation
targeted are NK cells, cytotoxic CD8+
progression to AIDS [173].
infection [174].
**8. Conclusion**
Ampligen® is known as [poly(I)-poly(C12U)] that is composed of poly(IC) with a U mismatch at every 12th base of the C strand. The main effect of Ampligen® is the induction of Th1 response. This has been demonstrated in delayed-type hypersensitivity reactions and in current clinical studies with human immunodeficiency virus-infected patients. Other cells
1990s, little is known about Ampligen®, since; it is not routinely used in clinical practice. However, due to its notable and potent antiviral effects, Ampligen® has been generally well tolerated in randomized clinical trials [164, 165]. Since, Ampligen® did not consistent‐ ly produce antiviral effects, but its antiviral activity was always seen after therapeutic treatment. Therefore, it was assumed that Ampligen have exquisite specificity for TLR3 [166, 167]. Early studies showed that treatment of chronically HBV-infected chimpanzees with polyICLC transiently reduced the levels of serum viral DNA, HBsAg and HBeAg [168]. In two independent studies, transient suppression of DHBV in ducks by Ampligen® was also observed [169, 170]. In vitro studies, Ampligen and zidovudine have combined: to act in synergy against HIV-1-infected cell lines [171, 172] and enabling restoration of the immune function in HIV-infected individuals. Wither through maintaining a stable or raising CD4 count, increasing delayed-type hypersensitivity reaction, and reducing rate of
More recently, Nicodemus et al (2010) evaluated polyIC12U and confirmed the potential of TLR3 stimulation with Ampligen® in enhancement bioactivity of cancer immunotherapies [166]. Their results revealed that Ampligen® is a potent inducer for the dendritic cell matura‐ tion and local cytokines producing in culture systems including IL-4, IL-6, IL-12p70, IFN g, MIP-1a and TNF α. Interestingly, treatment of wild-type mice with Ampligen® 24 hours following infection with a lethal viral inoculation was fully protective, whereas dosing 4 h prior to inoculation is ineffective in preventing mortality [167]. Intranasal Ampligen® administered to mice in conjunction with prototype avian flu vaccines, greatly enhances the cellular and humoral immunity achievable with the vaccine and also protects mice from lethal
Since the discovery of the important role of TLRs in innate immunity and the initiation of an immune response that follows the activation of antigen-specific acquired immunity, rapid progress has been made on our understanding of the molecular mechanisms of TLRs. The inclusion of TLR ligands as a potential new class of adjuvants candidates has enabled the development of clinical effective vaccination strategies against many diseases. Poly(I:C) and Poly-ICLC represent two important member of these immunostimulatory vaccine adjuvants in mice, nonhuman primates and in human. TLR3 is specifically expressed in quite cell-type and species; it is expressed in specific myeloid cells, vascular endothelial cells and airway epithelial cells among others. Human TLR3 is highly expressed in immature DCs whereas
T cells, and LAK-NK cells. Despite initial trials in the
The authors would like to thank Prof. Karolin K. Abdul-Aziz and Dr. Muobarak J. Tuorkey (Zoology Department, Faculty of Science, Damanhour University, Egypt) for their kind help during writing this chapter.
#### **Author details**
Mohamed Labib Salem1,2\*, Said M. Hammad1,3, Mohamed R. El-shanshory1,4, Mohamed A. Attia1,3 and Abdel-Aziz A. Zidan1,2
\*Address all correspondence to: mohamed.labib@science.tanta.edu.eg
1 Center of Excellence in Cancer Research (CECR), Tanta University Educational Hospital, Medical Campus, Tanta University, Egypt
2 Immunology and Biotechnology Unit, Zoology Department, Faculty of Science, Tanta Uni‐ versity, Egypt
3 Clinical Pathology Department, Faculty of Medicine, Tanta University, Egypt
4 Pediatric Hematology and Oncology Unit, Faculty of Medicine, Tanta University, Egypt
#### **References**
[1] Pulendran, B. and R. Ahmed, *Translating innate immunity into immunological memory: implications for vaccine development.* Cell, 2006. 124(4): p. 849-63.
[16] Lemaitre, B., J.M. Reichhart, and J.A. Hoffmann, *Drosophila host defense: differential in‐ duction of antimicrobial peptide genes after infection by various classes of microorganisms.*
Immunostimulatory Effects of Triggering TLR3 Signaling Pathway — Implication for Cancer Immunotherapy
http://dx.doi.org/10.5772/58575
163
[17] Bowie, A. and L.A. O'Neill, *The interleukin-1 receptor/Toll-like receptor superfamily: sig‐ nal generators for pro-inflammatory interleukins and microbial products.* J Leukoc Biol,
[18] Thompson, M.R., et al., *Pattern recognition receptors and the innate immune response to*
[19] Takahashi, K. and T. Kawai, *[Pathogen recognition by Toll-like receptor].* Nihon Rinsho,
[20] Uematsu, S. and S. Akira, *[Innate immune recognition of viral infection].* Uirusu, 2006.
[21] Salaun, B., et al., *Toll-like receptor 3 expressed by melanoma cells as a target for therapy?*
[22] Nishiya, T. and A.L. DeFranco, *Ligand-regulated chimeric receptor approach reveals dis‐ tinctive subcellular localization and signaling properties of the Toll-like receptors.* J Biol
[23] He, X., et al., *Recognition of pathogen-associated nucleic acids by endosomal nucleic acidsensing toll-like receptors.* Acta Biochim Biophys Sin (Shanghai), 2013. 45(4): p. 241-58.
[24] Akira, S., K. Takeda, and T. Kaisho, *Toll-like receptors: critical proteins linking innate and*
[25] Tabeta, K., et al., *Toll-like receptors 9 and 3 as essential components of innate immune de‐ fense against mouse cytomegalovirus infection.* Proc Natl Acad Sci U S A, 2004. 101(10):
[26] Ariffin, J.K. and M.J. Sweet, *Differences in the repertoire, regulation and function of Tolllike Receptors and inflammasome-forming Nod-like Receptors between human and mouse.*
[27] Roach, J.C., et al., *The evolution of vertebrate Toll-like receptors.* Proc Natl Acad Sci U S
[28] Pulendran, B., *Modulating vaccine responses with dendritic cells and Toll-like receptors.*
[29] Takeda, K. and S. Akira, *Toll-like receptors.* Curr Protoc Immunol, 2007. Chapter 14: p.
[30] Takeuchi, O., et al., *Differential roles of TLR2 and TLR4 in recognition of gram-negative and gram-positive bacterial cell wall components.* Immunity, 1999. 11(4): p. 443-51.
Proc Natl Acad Sci U S A, 1997. 94(26): p. 14614-9.
*viral infection.* Viruses, 2011. 3(6): p. 920-40.
Clin Cancer Res, 2007. 13(15 Pt 1): p. 4565-74.
*acquired immunity.* Nat Immunol, 2001. 2(8): p. 675-80.
2000. 67(4): p. 508-14.
56(1): p. 1-8.
p. 3516-21.
Unit 14 12.
Curr Opin Microbiol, 2013.
A, 2005. 102(27): p. 9577-82.
Immunol Rev, 2004. 199: p. 227-50.
2007. 65 Suppl 2 Pt. 1: p. 53-7.
Chem, 2004. 279(18): p. 19008-17.
[16] Lemaitre, B., J.M. Reichhart, and J.A. Hoffmann, *Drosophila host defense: differential in‐ duction of antimicrobial peptide genes after infection by various classes of microorganisms.* Proc Natl Acad Sci U S A, 1997. 94(26): p. 14614-9.
**References**
162 Immune Response Activation
2009. 30(1): p. 23-32.
2009. 27(4): p. 549-57.
2004. 430(6996): p. 257-63.
PLoS Pathog, 2009. 5(4): p. e1000373.
1985. 42(3): p. 779-89.
1985. 42(3): p. 791-8.
7(2): p. 131-7.
p. 370-83.
1589-602.
*work.* Immunity, 2010. 33(4): p. 492-503.
*immunity.* Trends Immunol, 2002. 23(4): p. 201-8.
[1] Pulendran, B. and R. Ahmed, *Translating innate immunity into immunological memory:*
[2] Reed, S.G., et al., *New horizons in adjuvants for vaccine development.* Trends Immunol,
[3] Coffman, R.L., A. Sher, and R.A. Seder, *Vaccine adjuvants: putting innate immunity to*
[4] Johnson, C.B., et al., *Tumours: Immunotherapy*, in *eLS*. 2001, John Wiley & Sons, Ltd.
[5] Belardelli, F. and M. Ferrantini, *Cytokines as a link between innate and adaptive antitumor*
[6] Salem, M.L., et al., *The TLR3 agonist poly(I:C) targets CD8+T cells and augments their an‐ tigen-specific responses upon their adoptive transfer into naive recipient mice.* Vaccine,
[7] Kawai, T. and S. Akira, *Innate immune recognition of viral infection.* Nat Immunol, 2006.
[8] Pichlmair, A. and C. Reis e Sousa, *Innate recognition of viruses.* Immunity, 2007. 27(3):
[9] Beutler, B., *Inferences, questions and possibilities in Toll-like receptor signalling.* Nature,
[10] Kawai, T. and S. Akira, *The role of pattern-recognition receptors in innate immunity: up‐*
[11] Longhi, M.P., et al., *Dendritic cells require a systemic type I interferon response to mature and induce CD4+Th1 immunity with poly IC as adjuvant.* J Exp Med, 2009. 206(7): p.
[12] Stahl-Hennig, C., et al., *Synthetic double-stranded RNAs are adjuvants for the induction of T helper 1 and humoral immune responses to human papillomavirus in rhesus macaques.*
[13] Matsumoto, M. and T. Seya, *TLR3: interferon induction by double-stranded RNA includ‐*
[14] Anderson, K.V., G. Jurgens, and C. Nusslein-Volhard, *Establishment of dorsal-ventral polarity in the Drosophila embryo: genetic studies on the role of the Toll gene product.* Cell,
[15] Anderson, K.V., L. Bokla, and C. Nusslein-Volhard, *Establishment of dorsal-ventral po‐ larity in the Drosophila embryo: the induction of polarity by the Toll gene product.* Cell,
*date on Toll-like receptors.* Nat Immunol, 2010. 11(5): p. 373-84.
*ing poly(I:C).* Adv Drug Deliv Rev, 2008. 60(7): p. 805-12.
*implications for vaccine development.* Cell, 2006. 124(4): p. 849-63.
[31] Smith, K.D. and A. Ozinsky, *Toll-like receptor-5 and the innate immune response to bacte‐ rial flagellin.* Curr Top Microbiol Immunol, 2002. 270: p. 93-108.
[47] Kanayama, A., et al., *TAB2 and TAB3 activate the NF-kappaB pathway through binding to*
Immunostimulatory Effects of Triggering TLR3 Signaling Pathway — Implication for Cancer Immunotherapy
http://dx.doi.org/10.5772/58575
165
[48] Takeda, K. and S. Akira, *TLR signaling pathways.* Semin Immunol, 2004. 16(1): p. 3-9.
[49] Kawai, T. and S. Akira, *Signaling to NF-kappaB by Toll-like receptors.* Trends Mol Med,
[50] Freeman, A.I., et al., *Pharmacologic effects of polyinosinic-polycytidylic acid in man.* J Med
[51] Levine, A.S., et al., *Initial clinical trials in cancer patients of polyriboinosinic-polyribocyti‐ dylic acid stabilized with poly-L-lysine, in carboxymethylcellulose [poly(ICLC)], a highly ef‐*
[52] Cella, M., et al., *Plasmacytoid dendritic cells activated by influenza virus and CD40L drive*
[53] Alexopoulou, L., et al., *Recognition of double-stranded RNA and activation of NF-kappaB*
[54] Matsumoto, M., et al., *Establishment of a monoclonal antibody against human Toll-like re‐ ceptor 3 that blocks double-stranded RNA-mediated signaling.* Biochem Biophys Res Com‐
[55] Meusel, T.R., K.E. Kehoe, and F. Imani, *Protein kinase R regulates double-stranded RNA induction of TNF-alpha but not IL-1 beta mRNA in human epithelial cells.* J Immunol,
[56] Levy, H.B., *Induction of interferon in vivo and in vitro by polynucleotides and derivatives,*
[57] Meurs, E., et al., *Molecular cloning and characterization of the human double-stranded RNA-activated protein kinase induced by interferon.* Cell, 1990. 62(2): p. 379-90.
[58] Kolumam, G.A., et al., *Type I interferons act directly on CD8 T cells to allow clonal expan‐ sion and memory formation in response to viral infection.* J Exp Med, 2005. 202(5): p.
[59] Le Bon, A., et al., *Direct stimulation of T cells by type I IFN enhances the CD8+T cell re‐*
[60] Le Bon, A., et al., *Type i interferons potently enhance humoral immunity and can promote isotype switching by stimulating dendritic cells in vivo.* Immunity, 2001. 14(4): p. 461-70.
[61] Manetti, R., et al., *Polyinosinic acid: polycytidylic acid promotes T helper type 1-specific im‐ mune responses by stimulating macrophage production of interferon-alpha and interleu‐*
[62] Akiyama, Y., et al., *Differential ability of human blood monocyte subsets to release various*
*sponse during cross-priming.* J Immunol, 2006. 176(8): p. 4682-9.
*kin-12.* Eur J Immunol, 1995. 25(9): p. 2656-60.
*cytokines.* J Leukoc Biol, 1985. 37(5): p. 519-30.
*and preparation of derivatives.* Methods Enzymol, 1981. 78(Pt A): p. 242-51.
*polyubiquitin chains.* Mol Cell, 2004. 15(4): p. 535-48.
*fective interferon inducer.* Cancer Res, 1979. 39(5): p. 1645-50.
*a potent TH1 polarization.* Nat Immunol, 2000. 1(4): p. 305-10.
*by Toll-like receptor 3.* Nature, 2001. 413(6857): p. 732-8.
2007. 13(11): p. 460-9.
Virol, 1977. 1(2): p. 79-93.
mun, 2002. 293(5): p. 1364-9.
2002. 168(12): p. 6429-35.
637-50.
[47] Kanayama, A., et al., *TAB2 and TAB3 activate the NF-kappaB pathway through binding to polyubiquitin chains.* Mol Cell, 2004. 15(4): p. 535-48.
[31] Smith, K.D. and A. Ozinsky, *Toll-like receptor-5 and the innate immune response to bacte‐*
[32] Lauw, F.N., D.R. Caffrey, and D.T. Golenbock, *Of mice and man: TLR11 (finally) finds*
[33] Zhang, D.M., et al., *[Down-regulation of Toll-like receptor 4 is related to the tolerance to lipopolysaccharide in rat alveolar macrophages].* Zhonghua Jie He He Hu Xi Za Zhi, 2003.
[34] Schulz, O., et al., *Toll-like receptor 3 promotes cross-priming to virus-infected cells.* Nature,
[35] Sivori, S., et al., *CpG and double-stranded RNA trigger human NK cells by Toll-like recep‐ tors: induction of cytokine release and cytotoxicity against tumors and dendritic cells.* Proc
[36] Jurk, M., et al., *Human TLR7 or TLR8 independently confer responsiveness to the antiviral*
[37] Heil, F., et al., *Species-specific recognition of single-stranded RNA via toll-like receptor 7*
[38] Akira, S. and H. Hemmi, *Recognition of pathogen-associated molecular patterns by TLR*
[39] Adachi, O., et al., *Targeted disruption of the MyD88 gene results in loss of IL-1-and IL-18-*
[41] Horng, T., et al., *The adaptor molecule TIRAP provides signalling specificity for Toll-like*
[42] Fitzgerald, K.A., et al., *Mal (MyD88-adapter-like) is required for Toll-like receptor-4 signal*
[43] Yamamoto, M., et al., *Cutting edge: a novel Toll/IL-1 receptor domain-containing adapter that preferentially activates the IFN-beta promoter in the Toll-like receptor signaling.* J Im‐
[44] Oshiumi, H., et al., *TICAM-1, an adaptor molecule that participates in Toll-like receptor 3-*
[45] Oshiumi, H., et al., *TIR-containing adapter molecule (TICAM)-2, a bridging adapter re‐ cruiting to toll-like receptor 4 TICAM-1 that induces interferon-beta.* J Biol Chem, 2003.
[46] Keating, S.E., et al., *IRAK-2 participates in multiple toll-like receptor signaling pathways to NFkappaB via activation of TRAF6 ubiquitination.* J Biol Chem, 2007. 282(46): p.
*mediated interferon-beta induction.* Nat Immunol, 2003. 4(2): p. 161-7.
[40] Kawai, T. and S. Akira, *TLR signaling.* Semin Immunol, 2007. 19(1): p. 24-32.
*rial flagellin.* Curr Top Microbiol Immunol, 2002. 270: p. 93-108.
*profilin.* Trends Immunol, 2005. 26(10): p. 509-11.
Natl Acad Sci U S A, 2004. 101(27): p. 10116-21.
*compound R-848.* Nat Immunol, 2002. 3(6): p. 499.
*mediated function.* Immunity, 1998. 9(1): p. 143-50.
*and 8.* Science, 2004. 303(5663): p. 1526-9.
*family.* Immunol Lett, 2003. 85(2): p. 85-95.
*receptors.* Nature, 2002. 420(6913): p. 329-33.
*transduction.* Nature, 2001. 413(6851): p. 78-83.
munol, 2002. 169(12): p. 6668-72.
278(50): p. 49751-62.
33435-43.
26(7): p. 411-4.
164 Immune Response Activation
2005. 433(7028): p. 887-92.
[63] Kok, K.H., et al., *The double-stranded RNA-binding protein PACT functions as a cellular activator of RIG-I to facilitate innate antiviral response.* Cell Host Microbe, 2011. 9(4): p. 299-309.
[78] Wang, H., E. Rayburn, and R. Zhang, *Synthetic oligodeoxynucleotides containing deoxy‐ cytidyl-deoxyguanosine dinucleotides (CpG ODNs) and modified analogs as novel anticancer*
Immunostimulatory Effects of Triggering TLR3 Signaling Pathway — Implication for Cancer Immunotherapy
http://dx.doi.org/10.5772/58575
167
[79] Dong, Z., et al., *Involvement of natural killer cells in PolyI:C-induced liver injury.* J Hepa‐
[80] Salazar-Mather, T.P., R. Ishikawa, and C.A. Biron, *NK cell trafficking and cytokine ex‐ pression in splenic compartments after IFN induction and viral infection.* J Immunol, 1996.
[81] Salem, M.L., et al., *The adjuvant effects of the toll-like receptor 3 ligand polyinosinic-cyti‐ dylic acid poly (I:C) on antigen-specific CD8+T cell responses are partially dependent on NK cells with the induction of a beneficial cytokine milieu.* Vaccine, 2006. 24(24): p. 5119-32.
[82] Salem, M.L., et al., *Vaccination using Toll-like receptor ligand (TLRL)-primed dendritic cells and TLR3L poly I:C administration leads to an augmentation of antigen-specific memo‐ ry CD8+T cell responses..* 97th Annual Meeting for American Association of Cancer
[83] Alenzi, F.Q., et al., *Cellular and molecular aspects of Goodpasture syndrome.* Iran J Kidney
[85] Kowalczyk, D.W., *Tumors and the danger model.* Acta Biochim Pol, 2002. 49(2): p.
[86] Seya, T., et al., *Role of toll-like receptors and their adaptors in adjuvant immunotherapy for*
[87] Okamoto, M. and M. Sato, *Toll-like receptor signaling in anti-cancer immunity.* J Med In‐
[88] Speiser, D.E., et al., *Rapid and strong human CD8+T cell responses to vaccination with pep‐ tide, IFA, and CpG oligodeoxynucleotide 7909.* J Clin Invest, 2005. 115(3): p. 739-46.
[89] Smith, K.J., S. Hamza, and H. Skelton, *The imidazoquinolines and their place in the thera‐*
[90] Stockfleth, E., et al., *The use of Toll-like receptor-7 agonist in the treatment of basal cell car‐*
[91] Shackleton, M., et al., *The impact of imiquimod, a Toll-like receptor-7 ligand (TLR7L), on the immunogenicity of melanoma peptide vaccination with adjuvant Flt3 ligand.* Cancer Im‐
[92] Baldridge, J.R., et al., *Taking a Toll on human disease: Toll-like receptor 4 agonists as vac‐ cine adjuvants and monotherapeutic agents.* Expert Opin Biol Ther, 2004. 4(7): p. 1129-38.
*py of cutaneous disease.* Expert Opin Pharmacother, 2003. 4(7): p. 1105-19.
*cinoma: an overview.* Br J Dermatol, 2003. 149 Suppl 66: p. 53-6.
[84] Yang, Q., et al., *Antitumor activity of NK cells.* Immunol Res, 2006. 36(1-3): p. 13-25.
*therapeutics.* Curr Pharm Des, 2005. 11(22): p. 2889-907.
Research (AACR), Washington, DC., 2006.
*cancer.* Anticancer Res, 2003. 23(6a): p. 4369-76.
tol, 2004. 41(6): p. 966-73.
157(7): p. 3054-64.
Dis, 2012. 6(1): p. 1-8.
vest, 2003. 50(1-2): p. 9-24.
mun, 2004. 4: p. 9.
295-302.
[78] Wang, H., E. Rayburn, and R. Zhang, *Synthetic oligodeoxynucleotides containing deoxy‐ cytidyl-deoxyguanosine dinucleotides (CpG ODNs) and modified analogs as novel anticancer therapeutics.* Curr Pharm Des, 2005. 11(22): p. 2889-907.
[63] Kok, K.H., et al., *The double-stranded RNA-binding protein PACT functions as a cellular activator of RIG-I to facilitate innate antiviral response.* Cell Host Microbe, 2011. 9(4): p.
[64] Kawai, T. and S. Akira, *Antiviral signaling through pattern recognition receptors.* J Bio‐
[65] Koyama, S., et al., *Differential role of TLR-and RLR-signaling in the immune responses to influenza A virus infection and vaccination.* J Immunol, 2007. 179(7): p. 4711-20.
[66] Gitlin, L., et al., *Essential role of mda-5 in type I IFN responses to polyriboinosinic:polyribo‐ cytidylic acid and encephalomyocarditis picornavirus.* Proc Natl Acad Sci U S A, 2006.
[67] Kato, H., et al., *Cell type-specific involvement of RIG-I in antiviral response.* Immunity,
[68] Matsumoto, M., et al., *Subcellular localization of Toll-like receptor 3 in human dendritic*
[69] Muzio, M., et al., *Differential expression and regulation of toll-like receptors (TLR) in hu‐ man leukocytes: selective expression of TLR3 in dendritic cells.* J Immunol, 2000. 164(11): p.
[70] Cario, E. and D.K. Podolsky, *Differential alteration in intestinal epithelial cell expression of toll-like receptor 3 (TLR3) and TLR4 in inflammatory bowel disease.* Infect Immun, 2000.
[71] Schmidt, K.N., et al., *APC-independent activation of NK cells by the Toll-like receptor 3 ag‐*
[72] Banchereau, J., et al., *Dendritic cells: controllers of the immune system and a new promise for immunotherapy.* Novartis Found Symp, 2003. 252: p. 226-35; discussion 235-8,
[73] Gelman, A.E., et al., *Toll-like receptor ligands directly promote activated CD4+T cell sur‐*
[74] Cella, M., et al., *Maturation, activation, and protection of dendritic cells induced by double-*
[75] Haines, D.S., K.I. Strauss, and D.H. Gillespie, *Cellular response to double-stranded RNA.*
[76] Manetti, R., et al., *Polyinosinic acid: polycytidylic acid promotes T helper type 1-specific im‐ mune responses by stimulating macrophage production of interferon-alpha and interleu‐*
[77] Zhou, Y., et al., *Hepatitis C virus E2 envelope protein induces dendritic cell maturation.* J
*onist double-stranded RNA.* J Immunol, 2004. 172(1): p. 138-43.
299-309.
166 Immune Response Activation
chem, 2007. 141(2): p. 137-45.
103(22): p. 8459-64.
2005. 23(1): p. 19-28.
5998-6004.
257-67.
68(12): p. 7010-7.
*cells.* J Immunol, 2003. 171(6): p. 3154-62.
*vival.* J Immunol, 2004. 172(10): p. 6065-73.
J Cell Biochem, 1991. 46(1): p. 9-20.
Viral Hepat, 2007. 14(12): p. 849-58.
*stranded RNA.* J Exp Med, 1999. 189(5): p. 821-9.
*kin-12.* Eur J Immunol, 1995. 25(9): p. 2656-60.
[93] Stover, A.G., et al., *Structure-activity relationship of synthetic toll-like receptor 4 agonists.* J Biol Chem, 2004. 279(6): p. 4440-9.
[107] Adams, M., et al., *The rationale for combined chemo/immunotherapy using a Toll-like recep‐ tor 3 (TLR3) agonist and tumour-derived exosomes in advanced ovarian cancer.* Vaccine,
Immunostimulatory Effects of Triggering TLR3 Signaling Pathway — Implication for Cancer Immunotherapy
http://dx.doi.org/10.5772/58575
169
[108] Berhanu, A., et al., *Combinational FLt3 ligand and granulocyte macrophage colony-stimu‐ lating factor treatment promotes enhanced tumor infiltration by dendritic cells and antitumor CD8(+) T-cell cross-priming but is ineffective as a therapy.* Cancer Res, 2006. 66(9): p.
[109] Koopman, G., et al., *Increase in plasmacytoid and myeloid dendritic cells by progenipoie‐ tin-1, a chimeric Flt-3 and G-CSF receptor agonist, in SIV-Infected rhesus macaques.* Hum
[110] Maraskovsky, E., et al., *Dramatic numerical increase of functionally mature dendritic cells*
[111] Salem, M.L., et al., *Defining the Ability of Cyclophosphamide Preconditioning to Enhance the Antigen-specific CD8+T-cell Response to Peptide Vaccination: Creation of a Beneficial Host Microenvironment Involving Type I IFNs and Myeloid Cells.* J Immunother, 2007.
[112] Paulos, C.M., et al., *Toll-like receptors in tumor immunotherapy.* Clin Cancer Res, 2007.
[113] Paulos, C.M., et al., *Microbial translocation augments the function of adoptively transferred self/tumor-specific CD8 T cells via TLR4 signaling.* J Clin Invest, 2007. 117(8): p.
[114] Salem, M.L., et al., *Defining the ability of cyclophosphamide preconditioning to enhance the antigen-specific CD8+T-cell response to peptide vaccination: creation of a beneficial host mi‐ croenvironment involving type I IFNs and myeloid cells.* J Immunother (1997), 2007. 30(1):
[115] Salem, M.L., et al., *Recovery from cyclophosphamide-induced lymphopenia results in expan‐ sion of immature dendritic cells which can mediate enhanced prime-boost vaccination antitu‐ mor responses in vivo when stimulated with the TLR3 agonist poly(I:C).* J Immunol, 2009.
[116] Rubinstein, M.P., et al., *Loss of T cell-mediated antitumor immunity after construct-specific downregulation of retrovirally encoded T-cell receptor expression in vivo.* Cancer Gene
[117] Kabelitz, D., *Expression and function of Toll-like receptors in T lymphocytes.* Curr Opin
[118] Caramalho, I., et al., *Regulatory T cells selectively express toll-like receptors and are activat‐*
*ed by lipopolysaccharide.* J Exp Med, 2003. 197(4): p. 403-11.
*in FLT3 ligand-treated mice.* Adv Exp Med Biol, 1997. 417: p. 33-40.
2005. 23(17-18): p. 2374-8.
Immunol, 2004. 65(4): p. 303-16.
4895-903.
30(1): p. 40-53.
2197-204.
p. 40-53.
182(4): p. 2030-40.
Ther, 2009. 16(2): p. 171-83.
Immunol, 2007. 19(1): p. 39-45.
13(18 Pt 1): p. 5280-9.
[107] Adams, M., et al., *The rationale for combined chemo/immunotherapy using a Toll-like recep‐ tor 3 (TLR3) agonist and tumour-derived exosomes in advanced ovarian cancer.* Vaccine, 2005. 23(17-18): p. 2374-8.
[93] Stover, A.G., et al., *Structure-activity relationship of synthetic toll-like receptor 4 agonists.* J
[94] Salem, M.L., et al., *Defining the antigen-specific T-cell response to vaccination and poly(I:C)/TLR3 signaling: evidence of enhanced primary and memory CD8 T-cell responses*
[95] Bachmann, M.F., M. Kopf, and B.J. Marsland, *Chemokines: more than just road signs.*
[96] Shortman, K. and Y.J. Liu, *Mouse and human dendritic cell subtypes.* Nat Rev Immunol,
[97] Prins, R.M., et al., *The TLR-7 agonist, imiquimod, enhances dendritic cell survival and pro‐ motes tumor antigen-specific T cell priming: relation to central nervous system antitumor*
[98] Weigel, B.J., et al., *CpG oligodeoxynucleotides potentiate the antitumor effects of chemother‐ apy or tumor resection in an orthotopic murine model of rhabdomyosarcoma.* Clin Cancer
[99] Badovinac, V.P., et al., *Accelerated CD8+T-cell memory and prime-boost response after*
[100] Davila, E. and E. Celis, *Repeated administration of cytosine-phosphorothiolated guaninecontaining oligonucleotides together with peptide/protein immunization results in enhanced*
[101] Davila, E., R. Kennedy, and E. Celis, *Generation of antitumor immunity by cytotoxic T lymphocyte epitope peptide vaccination, CpG-oligodeoxynucleotide adjuvant, and CTLA-4*
[102] Craft, N., et al., *The TLR7 agonist imiquimod enhances the anti-melanoma effects of a re‐ combinant Listeria monocytogenes vaccine.* J Immunol, 2005. 175(3): p. 1983-90.
[103] Nowak, M., et al., *LPS-induced liver injury in D-galactosamine-sensitized mice requires se‐ creted TNF-alpha and the TNF-p55 receptor.* Am J Physiol Regul Integr Comp Physiol,
[104] Berraondo, P., et al., *Eradication of large tumors in mice by a tritherapy targeting the in‐ nate, adaptive, and regulatory components of the immune system.* Cancer Res, 2007. 67(18):
[105] Taieb, J., et al., *Chemoimmunotherapy of tumors: cyclophosphamide synergizes with exo‐*
[106] Garay, R.P., et al., *Cancer relapse under chemotherapy: why TLR2/4 receptor agonists can*
*CTL responses with anti-tumor activity.* J Immunol, 2000. 165(1): p. 539-47.
*and antitumor immunity.* J Immunother, 2005. 28(3): p. 220-8.
Biol Chem, 2004. 279(6): p. 4440-9.
Nat Rev Immunol, 2006. 6(2): p. 159-64.
*immunity.* J Immunol, 2006. 176(1): p. 157-64.
*blockade.* Cancer Res, 2003. 63(12): p. 3281-8.
*some based vaccines.* J Immunol, 2006. 176(5): p. 2722-9.
*help.* Eur J Pharmacol, 2007. 563(1-3): p. 1-17.
*dendritic-cell vaccination.* Nat Med, 2005. 11(7): p. 748-56.
2002. 2(3): p. 151-61.
168 Immune Response Activation
Res, 2003. 9(8): p. 3105-14.
2000. 278(5): p. R1202-9.
p. 8847-55.
[119] Lewkowicz, P., et al., *Lipopolysaccharide-activated CD4+CD25+T regulatory cells inhibit neutrophil function and promote their apoptosis and death.* J Immunol, 2006. 177(10): p. 7155-63.
[133] Diaz-Montero, C.M., et al., *Priming of naive CD8+T cells in the presence of IL-12 selective‐ ly enhances the survival of CD8+CD62Lhi cells and results in superior anti-tumor activity in a tolerogenic murine model.* Cancer Immunol Immunother, 2008. 57(4): p. 563-72.
Immunostimulatory Effects of Triggering TLR3 Signaling Pathway — Implication for Cancer Immunotherapy
http://dx.doi.org/10.5772/58575
171
[134] Le Bon, A. and D.F. Tough, *Links between innate and adaptive immunity via type I inter‐*
[135] Koyama, S., et al., *Innate immune response to viral infection.* Cytokine, 2008. 43(3): p.
[136] Kumar, A., J. Zhang, and F.S. Yu, *Toll-like receptor 3 agonist poly(I:C)-induced antiviral response in human corneal epithelial cells.* Immunology, 2006. 117(1): p. 11-21.
[137] Yoneyama, M., et al., *Direct triggering of the type I interferon system by virus infection: activation of a transcription factor complex containing IRF-3 and CBP/p300.* EMBO J, 1998.
[138] Wathelet, M.G., et al., *Virus infection induces the assembly of coordinately activated tran‐ scription factors on the IFN-beta enhancer in vivo.* Mol Cell, 1998. 1(4): p. 507-18.
[139] Lin, R., et al., *Virus-dependent phosphorylation of the IRF-3 transcription factor regulates nuclear translocation, transactivation potential, and proteasome-mediated degradation.* Mol
[140] Juang, Y.T., et al., *Primary activation of interferon A and interferon B gene transcription by interferon regulatory factor 3.* Proc Natl Acad Sci U S A, 1998. 95(17): p. 9837-42.
[141] Sen, G.C. and S.N. Sarkar, *Transcriptional signaling by double-stranded RNA: role of*
[142] Guillot, L., et al., *Involvement of toll-like receptor 3 in the immune response of lung epithe‐ lial cells to double-stranded RNA and influenza A virus.* J Biol Chem, 2005. 280(7): p.
[143] Rudd, B.D., et al., *Differential role for TLR3 in respiratory syncytial virus-induced chemo‐*
[144] Ashkar, A.A., et al., *Toll-like receptor (TLR)-3, but not TLR4, agonist protects against geni‐ tal herpes infection in the absence of inflammation seen with CpG DNA.* J Infect Dis, 2004.
[145] Yoneyama, M., et al., *The RNA helicase RIG-I has an essential function in double-stranded*
[146] Yoneyama, M., et al., *Shared and unique functions of the DExD/H-box helicases RIG-I, MDA5, and LGP2 in antiviral innate immunity.* J Immunol, 2005. 175(5): p. 2851-8.
[147] Akazawa, T., et al., *Antitumor NK activation induced by the Toll-like receptor 3-TICAM-1 (TRIF) pathway in myeloid dendritic cells.* Proc Natl Acad Sci U S A, 2007. 104(1): p.
*RNA-induced innate antiviral responses.* Nat Immunol, 2004. 5(7): p. 730-7.
*TLR3.* Cytokine Growth Factor Rev, 2005. 16(1): p. 1-14.
*kine expression.* J Virol, 2005. 79(6): p. 3350-7.
*feron.* Curr Opin Immunol, 2002. 14(4): p. 432-6.
336-41.
17(4): p. 1087-95.
5571-80.
252-7.
190(10): p. 1841-9.
Cell Biol, 1998. 18(5): p. 2986-96.
[133] Diaz-Montero, C.M., et al., *Priming of naive CD8+T cells in the presence of IL-12 selective‐ ly enhances the survival of CD8+CD62Lhi cells and results in superior anti-tumor activity in a tolerogenic murine model.* Cancer Immunol Immunother, 2008. 57(4): p. 563-72.
[119] Lewkowicz, P., et al., *Lipopolysaccharide-activated CD4+CD25+T regulatory cells inhibit neutrophil function and promote their apoptosis and death.* J Immunol, 2006. 177(10): p.
[120] den Haan, J.M., G. Kraal, and M.J. Bevan, *Cutting edge: Lipopolysaccharide induces IL-10-producing regulatory CD4+T cells that suppress the CD8+T cell response.* J Immunol,
[121] Crellin, N.K., et al., *Human CD4+T cells express TLR5 and its ligand flagellin enhances the suppressive capacity and expression of FOXP3 in CD4+CD25+T regulatory cells.* J Immu‐
[122] Sutmuller, R.P., et al., *Toll-like receptor 2 controls expansion and function of regulatory T*
[123] Liu, P.T., et al., *Toll-like receptor triggering of a vitamin D-mediated human antimicrobial*
[124] Antony, P.A., et al., *CD8+T cell immunity against a tumor/self-antigen is augmented by CD4+T helper cells and hindered by naturally occurring T regulatory cells.* J Immunol,
[125] Wada, S., et al., *Cyclophosphamide augments antitumor immunity: studies in an autochtho‐*
[126] Hirschhorn-Cymerman, D., et al., *OX40 engagement and chemotherapy combination pro‐ vides potent antitumor immunity with concomitant regulatory T cell apoptosis.* J Exp Med,
[127] Ghiringhelli, F., et al., *CD4+CD25+regulatory T cells suppress tumor immunity but are sensitive to cyclophosphamide which allows immunotherapy of established tumors to be cura‐*
[128] Hillebrands, J.L., et al., *A regulatory CD4+T cell subset in the BB rat model of autoimmune diabetes expresses neither CD25 nor Foxp3.* J Immunol, 2006. 177(11): p. 7820-32.
[129] Huleatt, J.W., et al., *Vaccination with recombinant fusion proteins incorporating Toll-like receptor ligands induces rapid cellular and humoral immunity.* Vaccine, 2007. 25(4): p.
[130] Teitz-Tennenbaum, S., et al., *Radiotherapy combined with intratumoral dendritic cell vac‐ cination enhances the therapeutic efficacy of adoptive T-cell transfer.* J Immunother, 2009.
[131] Fehervari, Z. and S. Sakaguchi, *A paragon of self-tolerance: CD25+CD4+regulatory T cells and the control of immune responses.* Arthritis Res Ther, 2004. 6(1): p. 19-25.
[132] Netea, M.G., et al., *Toll-like receptor 2 suppresses immunity against Candida albicans through induction of IL-10 and regulatory T cells.* J Immunol, 2004. 172(6): p. 3712-8.
*nous prostate cancer model.* Cancer Res, 2009. 69(10): p. 4309-18.
7155-63.
170 Immune Response Activation
2007. 178(9): p. 5429-33.
nol, 2005. 175(12): p. 8051-9.
2005. 174(5): p. 2591-601.
2009. 206(5): p. 1103-16.
763-75.
32(6): p. 602-12.
*cells.* J Clin Invest, 2006. 116(2): p. 485-94.
*response.* Science, 2006. 311(5768): p. 1770-3.
*tive.* Eur J Immunol, 2004. 34(2): p. 336-44.
[148] Trumpfheller, C., et al., *The microbial mimic poly IC induces durable and protective CD4+T cell immunity together with a dendritic cell targeted vaccine.* Proc Natl Acad Sci U S A, 2008. 105(7): p. 2574-9.
[162] Gesuete, R., et al., *Poly-ICLC preconditioning protects the blood-brain barrier against ische‐ mic injury in vitro through type I interferon signaling.* J Neurochem, 2012. 123 Suppl 2: p.
Immunostimulatory Effects of Triggering TLR3 Signaling Pathway — Implication for Cancer Immunotherapy
http://dx.doi.org/10.5772/58575
173
[163] Ming Lim, C., et al., *TLR3 agonists improve the immunostimulatory potential of cetuximab against EGFR head and neck cancer cells.* Oncoimmunology, 2013. 2(6): p. e24677.
[164] Strayer, D.R., et al., *A controlled clinical trial with a specifically configured RNA drug, poly(I).poly(C12U), in chronic fatigue syndrome.* Clin Infect Dis, 1994. 18 Suppl 1: p.
[165] Thompson, K.A., et al., *Results of a double-blind placebo-controlled study of the doublestranded RNA drug polyI:polyC12U in the treatment of HIV infection.* Eur J Clin Microbiol
[166] Nicodemus, C.F. and J.S. Berek, *TLR3 agonists as immunotherapeutic agents.* Immuno‐
[167] Gowen, B.B., et al., *TLR3 is essential for the induction of protective immunity against Pun‐ ta Toro Virus infection by the double-stranded RNA (dsRNA), poly(I:C12U), but not Poly(I:C): differential recognition of synthetic dsRNA molecules.* J Immunol, 2007. 178(8):
[168] Purcell, R.H., et al., *Modification of chronic hepatitis-B virus infection in chimpanzees by*
[169] Ijichi, K., et al., *In vivo antiviral effects of mismatched double-stranded RNA on duck hepati‐*
[170] Niu, J., et al., *The use of ampligen alone and in combination with ganciclovir and coumermy‐ cin A1 for the treatment of ducks congenitally-infected with duck hepatitis B virus.* Antiviral
[171] O'Marro, S.D., et al., *The effect of combinations of ampligen and zidovudine or dideoxyino‐ sine against human immunodeficiency viruses in vitro.* Antiviral Res, 1992. 17(2): p.
[172] Gillespie, D., et al., *Synergistic inhibition of AZT-resistant HIV by AZT combined with poly(I):poly(C12U), without synergistic toxicity to bone marrow progenitor cell elements.* In
[173] Carter, W.A., et al., *Mismatched double-stranded RNA, Ampligen (poly(I): poly(C12U), demonstrat*es antiviral and immunostimulatory activities in HIV disease. Int J Immu‐
[174] Ichinohe, T., et al., Intranasal immunization with H5N1 vaccine plus Poly I:Poly C12U, a Toll-like receptor agonist, protects mice against homologous and heterolo‐
gous virus challenge. Microbes Infect, 2007. 9(11): p. 1333-40.
*administration of an interferon inducer.* Lancet, 1976. 2(7989): p. 757-61.
75-85.
S88-95.
p. 5200-8.
169-77.
Infect Dis, 1996. 15(7): p. 580-7.
therapy, 2010. 2(2): p. 137-40.
Res, 1993. 21(2): p. 155-71.
Vivo, 1994. 8(3): p. 375-81.
nopharmacol, 1991. 13 Suppl 1: p. 69-76.
*tis B virus.* J Med Virol, 1994. 43(2): p. 161-5.
[162] Gesuete, R., et al., *Poly-ICLC preconditioning protects the blood-brain barrier against ische‐ mic injury in vitro through type I interferon signaling.* J Neurochem, 2012. 123 Suppl 2: p. 75-85.
[148] Trumpfheller, C., et al., *The microbial mimic poly IC induces durable and protective CD4+T cell immunity together with a dendritic cell targeted vaccine.* Proc Natl Acad Sci U
[149] Veldhuis, W.B., et al., *Interferon-beta blocks infiltration of inflammatory cells and reduces infarct volume after ischemic stroke in the rat.* J Cereb Blood Flow Metab, 2003. 23(9): p.
[150] Kraus, J., et al., *Interferon-beta stabilizes barrier characteristics of brain endothelial cells in*
[151] Maluish, A.E., et al., *Immunomodulatory effects of poly(I,C)-LC in cancer patients.* J Biol
[152] Droller, M.J., *Immunotherapy of metastatic renal cell carcinoma with polyinosinic-polycyti‐*
[153] Hawkins, M.J., M. Levin, and E.C. Borden, *An Eastern Cooperative Oncology Group phase I-II pilot study of polyriboinosinic-polyribocytidylic acid poly-L-lysine complex in pa‐ tients with metastatic malignant melanoma.* J Biol Response Mod, 1985. 4(6): p. 664-8.
[154] Krown, S.E., et al., *Phase I trials of poly(I,C) complexes in advanced cancer.* J Biol Re‐
[155] Lampkin, B.C., et al., *Phase II trial of a complex polyriboinosinic-polyribocytidylic acid with poly-L-lysine and carboxymethyl cellulose in the treatment of children with acute leukemia and neuroblastoma: a report from the Children's Cancer Study Group.* Cancer Res, 1985.
[156] Levine, A.S. and H.B. Levy, *Phase I-II trials of poly IC stabilized with poly-L-lysine.* Can‐
[157] Nakamura, O., et al., *Phase I-II trials of poly(ICLC) in malignant brain tumor patients.* J
[158] Rettenmaier, M.A., M.L. Berman, and P.J. DiSaia, *Treatment of advanced ovarian cancer with polyinosinic-polycytidylic lysine carboxymethylcellulose (poly(ICLC)).* Gynecol Oncol,
[159] Stevenson, H.C., et al., *A phase I evaluation of poly(I,C)-LC in cancer patients.* J Biol Re‐
[160] Theriault, R.L., et al., *Evaluation of polyinosinic-polycytidylic and poly-L-lysine in meta‐*
[161] Ewel, C.H., et al., *Polyinosinic-polycytidylic acid complexed with poly-L-lysine and carbox‐ ymethylcellulose in combination with interleukin 2 in patients with cancer: clinical and im‐*
*static breast cancer.* Cancer Treat Rep, 1986. 70(11): p. 1341-2.
*munological effects.* Cancer Res, 1992. 52(11): p. 3005-10.
S A, 2008. 105(7): p. 2574-9.
*vitro.* Ann Neurol, 2004. 56(2): p. 192-205.
Response Mod, 1985. 4(6): p. 656-63.
*dylic acid.* J Urol, 1987. 137(2): p. 202-6.
sponse Mod, 1985. 4(6): p. 640-9.
cer Treat Rep, 1978. 62(11): p. 1907-12.
Interferon Res, 1982. 2(1): p. 1-4.
sponse Mod, 1985. 4(6): p. 650-5.
45(11 Pt 2): p. 5904-9.
1986. 24(3): p. 359-61.
1029-39.
172 Immune Response Activation
**Section 3**
**Local Immunity**
**Section 3**
### **Local Immunity**
**Chapter 7**
**Immune Regulation of** *Chlamydia trachomatis* **Infections**
Tubal infertility and reproductive damage, preeclampsia and preterm births may potentially occur in many of the 50 million women who are annually infected with *Chlamydia trachoma‐ tis* [reviewed in 1; 2- 4]. *C.trachomatis* infections of the genital tract are also reported to increase the risk of human immunodeficiency virus type 1(HIV-1) susceptibility and viral shedding in the genital tract either directly by pro-inflammatory signalling pathways or indirectly by their effects on genital epithelial cells [5]. A dramatic increase in the number of HIV-1 chemokine co-receptor CXCR4- and CCR5-positive T cell targets has also been reported in the endocervix
Gynecological cancer development is also suggested to be linked with Chlamydial genital tract infections. Epidemiological reports have proposed that *C.trachomatis* may act as a co-factor in the development of human papilloma virus (HPV)-induced squamous cell carcinomas (SCC), with an association between distinct serovariants of *C. trachomatis* (B, D, E, G, I, and J) and SCC being reported [7,8]. It is also likely that *C. trachomatis* co-infections of the lower genital tract (LGT) play a role in carcinogenesis of the female upper genital tract (UGT), particularly epithelial ovarian and type II carcinomas [9,10]. To explain the epidemiological associations of Chlamydial infections cervical and ovarian cancer development it has been hypothesised that the pathogen perhaps triggers epigenetic changes in host chromatin and impairs host DNA repair pathways [11] or that it changes host cell survival pathways by disrupting DNA
The majority (70-80%) of genital tract infections in women are asymptomatic and are not diagnosed or treated and persistent or "silent" infections are reported as occurring years after
> © 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
damage signalling pathways associated with tumorigenesis [12].
**of the Female Genital Tract**
Louise M. Hafner, Trudi A. Collet and
Additional information is available at the end of the chapter
Danica K. Hickey
**1. Introduction**
http://dx.doi.org/10.5772/57542
of *C. trachomatis*-positive women [6].
### **Immune Regulation of** *Chlamydia trachomatis* **Infections of the Female Genital Tract**
Louise M. Hafner, Trudi A. Collet and Danica K. Hickey
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/57542
#### **1. Introduction**
Tubal infertility and reproductive damage, preeclampsia and preterm births may potentially occur in many of the 50 million women who are annually infected with *Chlamydia trachoma‐ tis* [reviewed in 1; 2- 4]. *C.trachomatis* infections of the genital tract are also reported to increase the risk of human immunodeficiency virus type 1(HIV-1) susceptibility and viral shedding in the genital tract either directly by pro-inflammatory signalling pathways or indirectly by their effects on genital epithelial cells [5]. A dramatic increase in the number of HIV-1 chemokine co-receptor CXCR4- and CCR5-positive T cell targets has also been reported in the endocervix of *C. trachomatis*-positive women [6].
Gynecological cancer development is also suggested to be linked with Chlamydial genital tract infections. Epidemiological reports have proposed that *C.trachomatis* may act as a co-factor in the development of human papilloma virus (HPV)-induced squamous cell carcinomas (SCC), with an association between distinct serovariants of *C. trachomatis* (B, D, E, G, I, and J) and SCC being reported [7,8]. It is also likely that *C. trachomatis* co-infections of the lower genital tract (LGT) play a role in carcinogenesis of the female upper genital tract (UGT), particularly epithelial ovarian and type II carcinomas [9,10]. To explain the epidemiological associations of Chlamydial infections cervical and ovarian cancer development it has been hypothesised that the pathogen perhaps triggers epigenetic changes in host chromatin and impairs host DNA repair pathways [11] or that it changes host cell survival pathways by disrupting DNA damage signalling pathways associated with tumorigenesis [12].
The majority (70-80%) of genital tract infections in women are asymptomatic and are not diagnosed or treated and persistent or "silent" infections are reported as occurring years after
© 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
initial acquisition of *Chlamydia* with 50% of women continuing to shed the organism one year after infection is diagnosed [13]. Left untreated the organisms can ascend to the uterus and Fallopian tubes to cause pelvic inflammatory disease (PID). Inflammation of the Fallopian tubes can lead to scarring and fibrosis which can then lead to ectopic pregnancy (EP), subfertility and infertility with repeated or multiple infections increasing the chances of disease sequelae [14-17]. Additionally, infections during pregnancy may result in peripartum trans‐ mission of *Chlamydia* to newborns potentially resulting in neonatal conjunctivitis and lateonset pneumonia [18, 19].
**1.2. Developmental cycle**
inclusion.
rounds of infection.
**1.3. Intracellular survival strategies**
responses recently have been summarized [58].
*Chlamydia* undergoes a unique biphasic developmental cycle that encompasses two forms of reproduction; an intracellular phase of non-infectious metabolically active and replicative forms of reticulate bodies (RBs) and the extracellular phase of infectious elementary bodies
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
179
Chlamydial infection of superficial columnar cervical cells in the FGT is initiated by the attachment and endocytosis of the EB into these mucosal epithelial cells. EB containing endosomes then are internalised within host-derived membranes where the pathogen then hijacks both vesicular and non-vesicular mediated pathways of the host to obtain host-derived membrane lipids, including sphingomyelin (SM) and cholesterol, for intracellular growth and development [reviewed in 51]. Once resources such as sphingomyelin are obtained from the host cell and the vesicle expands, the EBs differentiate into the metabolically active, noninfectious and larger (0.8μm) reticulate bodies (RB) in the normal developmental cycle. Transformation of the metabolically-inactive EBs into the metabolically-active RBs occurs at approximately 3 h post-infection (pi), and replication by binary fission over a 24-48h time period then is initiated [52]. The RB containing endosome expands and produces a micro‐ scopically-visible Chlamydial microcolony referred to as an "inclusion" [53]. RBs then use host cell ATP and other metabolites to grow and replicate by binary fission within the Chlamydial
Many inducers of persistence (including Interferon gamma (IFN-g), amino acid starvation and antibiotics) will result in deviation from the 'normal developmental cycle', resulting in a viable but non-cultivable growth stage and the formation of large, abnormal developmental forms termed aberrant bodies (ABs) also known as aberrant RB, or persistent bodies [54]. In normal development, RBs will continue to divide undergoing between 8-12 rounds of replication. Later in the infectious process, the RBs asynchronously begin to differentiate back into EBs, there is evidence of intermediate bodies (IBs) at this stage and the EBs accumulate within the inclusion. Depending upon the Chlamydial species, at 30-80h post-infection the infectious EB's that have matured from the RBs are released either by host cell lysis or by a packaged-release mechanism called extrusion, allowing the EBs to infect neighbouring cells for successive
*C. trachomatis* is proficient at maintaining enduring relationships with its host by modulating and evading the immune system using a multitude of mechanisms [55]. These include inhibiting IFN-g–inducible major histocompatibility complex (MHC) class II expression [56] that in human endocervical cells is known to be mediated by direct and indirect (soluble) factors [57]. The multiple potential mechanisms used by *Chlamydia* to dampen immune
To modulate host signal transduction cascades and allow their developmental cycles to take place Chlamydiae are known to secrete up to sixteen anti-host (effector) proteins (reviewed in 59; 60). They also recruit proteins from the host and the first identified host proteins to be
(EBs) that attach to and invade epithelial cells [47; reviewed in 48, and 49 and 50].
Globally it is estimated that over 100 million adults were infected with *C. trachomatis* in 2008 [20]. The annual direct medical cost of sexually transmitted diseases (including *Chlamydia* and those caused by HIV) has been estimated at \$US16.9 billion [21]. Despite continued improve‐ ments in diagnostic and screening procedures for *Chlamydia* the incidence of infection contin‐ ues to increase with a significant increase, particularly in the rates of EP [22]. Even though azithromycin is available to treat uncomplicated lower genital tract infections [23], antibiotic therapy will not always be efficacious and one hypothesis is that *Chlamydia* treatments may in fact attenuate protective immunity in some patients [24]. Recent evidence has shown that although spontaneous resolution of infection before treatment occurred in 22% of naturally infected women, following azithromycion treatment, 16.5% of women were re-infected and this rate was higher (19.9%) in those women with persistent infection when compared with the rate (5%) in women who spontaneously cleared the infection [25]. Repeat infections are thought to be one of the primary causes of ongoing transmission of *Chlamydia*. Development of a vaccine that can either prevent infection or disease is therefore of vital importance for *Chlamydia* [26, 27].
#### **1.1. The organism:** *C. trachomatis*
*Chlamydia*e are small (0.5μm and genomes of 1.05Mb) gram negative intracellular bacteria and the human pathobiotype, *C. trachomatis* comprises non-invasive biovars that infect squamo‐ columnar epithelial cells of the female genital tract [28-30]. *C.trachomatis*replicates within nonacidic vacuoles of these genital tract cells, acquiring essential lipids from the host cell during infection to survive and replicate [31]. Based on serologic differences elicited by variable domain (VD) regions of the Chlamydial major outer membrane protein A (OmpA), *C.tracho‐ matis* biovars currently are classified into 19 serovars with serovars D-K causing urogenital infections [32-37].
Urogenital isolates of *C.trachomatis* may also contain a highly conserved 7.5 kB non-integrative plasmid that has a role in the pathobiology of these organisms [38, 39]. The plasmid functions as a significant virulence factor in the murine model of genital tract disease [40] with plasmid loss resulting in failure to induce upper genital tract pathology in this model [41]. Within clinical isolates of *C.trachomatis*, carriage of the 7.5 kB plasmid is almost universal, although a few naturally-occurring, plasmid deficient strains have been reported that are not associated with increased disease severity [42-45]. Recently it was reported that naturally-occurring plasmid-free urogenital isolates of *C.trachomatis* serovar F also had reduced infectivity and virulence in a murine model [46].
#### **1.2. Developmental cycle**
initial acquisition of *Chlamydia* with 50% of women continuing to shed the organism one year after infection is diagnosed [13]. Left untreated the organisms can ascend to the uterus and Fallopian tubes to cause pelvic inflammatory disease (PID). Inflammation of the Fallopian tubes can lead to scarring and fibrosis which can then lead to ectopic pregnancy (EP), subfertility and infertility with repeated or multiple infections increasing the chances of disease sequelae [14-17]. Additionally, infections during pregnancy may result in peripartum trans‐ mission of *Chlamydia* to newborns potentially resulting in neonatal conjunctivitis and late-
Globally it is estimated that over 100 million adults were infected with *C. trachomatis* in 2008 [20]. The annual direct medical cost of sexually transmitted diseases (including *Chlamydia* and those caused by HIV) has been estimated at \$US16.9 billion [21]. Despite continued improve‐ ments in diagnostic and screening procedures for *Chlamydia* the incidence of infection contin‐ ues to increase with a significant increase, particularly in the rates of EP [22]. Even though azithromycin is available to treat uncomplicated lower genital tract infections [23], antibiotic therapy will not always be efficacious and one hypothesis is that *Chlamydia* treatments may in fact attenuate protective immunity in some patients [24]. Recent evidence has shown that although spontaneous resolution of infection before treatment occurred in 22% of naturally infected women, following azithromycion treatment, 16.5% of women were re-infected and this rate was higher (19.9%) in those women with persistent infection when compared with the rate (5%) in women who spontaneously cleared the infection [25]. Repeat infections are thought to be one of the primary causes of ongoing transmission of *Chlamydia*. Development of a vaccine that can either prevent infection or disease is therefore of vital importance for
*Chlamydia*e are small (0.5μm and genomes of 1.05Mb) gram negative intracellular bacteria and the human pathobiotype, *C. trachomatis* comprises non-invasive biovars that infect squamo‐ columnar epithelial cells of the female genital tract [28-30]. *C.trachomatis*replicates within nonacidic vacuoles of these genital tract cells, acquiring essential lipids from the host cell during infection to survive and replicate [31]. Based on serologic differences elicited by variable domain (VD) regions of the Chlamydial major outer membrane protein A (OmpA), *C.tracho‐ matis* biovars currently are classified into 19 serovars with serovars D-K causing urogenital
Urogenital isolates of *C.trachomatis* may also contain a highly conserved 7.5 kB non-integrative plasmid that has a role in the pathobiology of these organisms [38, 39]. The plasmid functions as a significant virulence factor in the murine model of genital tract disease [40] with plasmid loss resulting in failure to induce upper genital tract pathology in this model [41]. Within clinical isolates of *C.trachomatis*, carriage of the 7.5 kB plasmid is almost universal, although a few naturally-occurring, plasmid deficient strains have been reported that are not associated with increased disease severity [42-45]. Recently it was reported that naturally-occurring plasmid-free urogenital isolates of *C.trachomatis* serovar F also had reduced infectivity and
onset pneumonia [18, 19].
178 Immune Response Activation
*Chlamydia* [26, 27].
infections [32-37].
**1.1. The organism:** *C. trachomatis*
virulence in a murine model [46].
*Chlamydia* undergoes a unique biphasic developmental cycle that encompasses two forms of reproduction; an intracellular phase of non-infectious metabolically active and replicative forms of reticulate bodies (RBs) and the extracellular phase of infectious elementary bodies (EBs) that attach to and invade epithelial cells [47; reviewed in 48, and 49 and 50].
Chlamydial infection of superficial columnar cervical cells in the FGT is initiated by the attachment and endocytosis of the EB into these mucosal epithelial cells. EB containing endosomes then are internalised within host-derived membranes where the pathogen then hijacks both vesicular and non-vesicular mediated pathways of the host to obtain host-derived membrane lipids, including sphingomyelin (SM) and cholesterol, for intracellular growth and development [reviewed in 51]. Once resources such as sphingomyelin are obtained from the host cell and the vesicle expands, the EBs differentiate into the metabolically active, noninfectious and larger (0.8μm) reticulate bodies (RB) in the normal developmental cycle. Transformation of the metabolically-inactive EBs into the metabolically-active RBs occurs at approximately 3 h post-infection (pi), and replication by binary fission over a 24-48h time period then is initiated [52]. The RB containing endosome expands and produces a micro‐ scopically-visible Chlamydial microcolony referred to as an "inclusion" [53]. RBs then use host cell ATP and other metabolites to grow and replicate by binary fission within the Chlamydial inclusion.
Many inducers of persistence (including Interferon gamma (IFN-g), amino acid starvation and antibiotics) will result in deviation from the 'normal developmental cycle', resulting in a viable but non-cultivable growth stage and the formation of large, abnormal developmental forms termed aberrant bodies (ABs) also known as aberrant RB, or persistent bodies [54]. In normal development, RBs will continue to divide undergoing between 8-12 rounds of replication. Later in the infectious process, the RBs asynchronously begin to differentiate back into EBs, there is evidence of intermediate bodies (IBs) at this stage and the EBs accumulate within the inclusion. Depending upon the Chlamydial species, at 30-80h post-infection the infectious EB's that have matured from the RBs are released either by host cell lysis or by a packaged-release mechanism called extrusion, allowing the EBs to infect neighbouring cells for successive rounds of infection.
#### **1.3. Intracellular survival strategies**
*C. trachomatis* is proficient at maintaining enduring relationships with its host by modulating and evading the immune system using a multitude of mechanisms [55]. These include inhibiting IFN-g–inducible major histocompatibility complex (MHC) class II expression [56] that in human endocervical cells is known to be mediated by direct and indirect (soluble) factors [57]. The multiple potential mechanisms used by *Chlamydia* to dampen immune responses recently have been summarized [58].
To modulate host signal transduction cascades and allow their developmental cycles to take place Chlamydiae are known to secrete up to sixteen anti-host (effector) proteins (reviewed in 59; 60). They also recruit proteins from the host and the first identified host proteins to be recruited into the Chlamydial inclusion include the eukaryotic nuclear protein ZNF23, a proapoptotic factor that potentially allows *Chlamydia* to inhibit host cell apoptosis thus enabling bacterial survival in the host [61].
*1.4.2. Ophthalmia neonatorum and infant pneumonia*
symptoms within the first 60 days of life [96].
infants in India below six months of age [99].
**1.5. Immunopathogenesis and immunoprotection**
mother and her sex partner(s).
The high prevalence of Chlamydial infections of the cervix in women of child-bearing age results in annual exposure of an estimated 100,000 neonates to *Chlamydia* infections [87] and neonatal Chlamydial infections are generally acquired during passage through an infected birth canal. A Chlamydial cause should be considered for all infants aged ≤30 days who have conjunctivitis and, if positively diagnosed, appropriate treatment initiated for the infant,
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
181
*Chlamydia* infections of neonates have been reported to be at 18 percent for ophthalmia neonatorum and 16 percent for pneumonia for infants exposed to *C. trachomatis* in vaginal secretions in the birth canal, with up to 60% of these infants showing serological evidence of Chlamydial infection [88]. Using data pooled from studies since 1977 it has been estimated that the incidence of ophthalmia neonatorum caused by exposure of the neonate to *C.tracho‐ matis* at birth is actually around 15 percent and that of pneumonia is around 7 percent [89].
Neonatal infection with *C. trachomatis* involves the mucous membranes of the neonatal eye, oropharynx, urogenital tract, and rectum and when symptomatic can result in a number of diseases in the first few months of life such as inclusion conjunctivitis, pneumonia or both [90,91; reviewed in 92; 93). Ophthalmia neonatorum has been reported as developing from 5-14 days post-delivery in up to 50% of infants of *C.trachomatis* infected mothers [94] and perhaps even as late as 42 days after birth [94] with between 5-30% of perinatally-infected infants also having nasopharyngeal infection [reviewed in 95]. A recent prospective study from Brazil estimated the prevalence of rate of *Chlamydia* among pregnant women (<30 years) in their third trimester as 72.7% and reported that 50% of 16 newborn babies had respiratory
Afebrile pneumonia may be seen in infants colonised in the respiratory tract with *C.trachoma‐ tis* and this may/may not be associated with inclusion conjunctivitis. After three or more weeks following Chlamydial infection of the infants a subacute, afebrile pneumonia with onset at ages 1–3 months develops in approximately seven percent of these cases [97, 92, 19]. Clinical characteristics have been described for infant pneumonia caused by *C.trachomatis* in infants with nasopharyngeal shedding and in around half of the infected infants conjunctivitis was recorded, with middle ear abnormalities recorded in more than half of the *Chlamydia*-positive infants and wheezing an unusual finding in these infants [98]. A more recent study has also highlighted that *C. trachomatis* is an important cause of lower respiratory tract infection in
The host inflammatory response to *C.trachomatis* is involved in immunity but also the pathol‐ ogy that leads to serious morbidities of chronic pelvic pain, EP and tubal infertility following female genital tract infection. Protective responses to *C.trachomatis* infections of the female genital tract are elicited to clear primary infection and to resist re-infection and include both innate and adaptive (particularly CD4+ T cell) cells and immune responses. Cells of both the innate and adaptive immune systems are found in the FGT in Fallopian tubes, uterus, cervix
#### **1.4. Chlamydial infections and disease sequelae**
#### *1.4.1. Infections in women*
Significant reproductive morbidity can occur in women following infection of the LGT with *C. trachomatis*. In the estimated 20% of women with symptoms, *C.trachomatis* infection of the columnar epithelial cells of the endocervix can result in cervicitis [62-64; reviewed in 65] and urethritis [66]. Severe complications of untreated LGT Chlamydial infections in women, whether symptomatic or asymptomatic can result in ascension of the organism from the lower to the UGT. Endometritis represents an early stage in the continuum from LGT infection through to salpingitis, a Fallopian tube disease also confirmed to be associated with *C.tracho‐ matis* infection. Since a critical source of Fallopian tube inflammation is infection with *C.trachomatis,* it has been hypothesized that *C.trachomatis* infection may be involved in chronic tubal inflammation and subsequent fimbrial carcinogenesis and hence could perhaps be the origin of some ovarian cancers [67]. Interestingly, a recent retrospective analysis of 34 cases of serous pelvic carcinomas (which are the most frequent histologic type of ovarian cancers) reported that 70% of cases did in fact involve the Fallopian tubes [68].
Chlamydial infection of the pelvis can lead to symptomatic pelvic inflammatory disease (PID) that can result in Fallopian tube inflammation [69] and salpingitis is reported to occur in 20-40% of women with untreated or repeated Chlamydial infections [70,71]. A recent model predicts that annual screening would prevent 61% of *C.trachomatis*-related PID in women who became infected in their LGTs with *Chlamydia* [72]. In addition to salpingitis the long-term consequen‐ ces of PID can also include chronic pelvic pain and ectopic pregnancy (EP). It has been proposed that since *C.trachomatis* infection increases tubal expression of prokineticin receptor (PROKR2) mRNA this expression then predisposes to ectopic implantation in the Fallopian tube [73]. Recent papers have reported that one third of EPs could be attributable to Chlamydial infection [74] and have also reviewed the pathogenesis of *Chlamydia*-induced tubal EP [75].
PID is caused by infection of the female genital tract with many microorganisms including *C*. *trachomatis* [reviewed in 16] and following symptomatic PID of polymicrobial etiology up to 18% of women may develop tubal factor infertility (TFI), which is amongst the leading causes of infertility accounting for 7-9.8% of all female factor infertilities [76-80]. Testing for serum antibody to the Chlamydial 60kDa Heat shock protein (i.e.cHSP60 antibody) is an accurate means for predicting *Chlamydia*-associated TFI [81] and elevated levels of anti-Chlamydial caseinolytic protease P (ClpP) antibodies were recently identified in 21 TFI patients [82]. Prolonged exposure to *Chlamydia* due to a chronic persistent infection or frequent re-infection has also been associated with chronic inflammation and TFI [83-85]. It has been reported that 9.5% women developed PID over 12 months in women who tested positive for *Chlamydia* at baseline [86].
#### *1.4.2. Ophthalmia neonatorum and infant pneumonia*
recruited into the Chlamydial inclusion include the eukaryotic nuclear protein ZNF23, a proapoptotic factor that potentially allows *Chlamydia* to inhibit host cell apoptosis thus enabling
Significant reproductive morbidity can occur in women following infection of the LGT with *C. trachomatis*. In the estimated 20% of women with symptoms, *C.trachomatis* infection of the columnar epithelial cells of the endocervix can result in cervicitis [62-64; reviewed in 65] and urethritis [66]. Severe complications of untreated LGT Chlamydial infections in women, whether symptomatic or asymptomatic can result in ascension of the organism from the lower to the UGT. Endometritis represents an early stage in the continuum from LGT infection through to salpingitis, a Fallopian tube disease also confirmed to be associated with *C.tracho‐ matis* infection. Since a critical source of Fallopian tube inflammation is infection with *C.trachomatis,* it has been hypothesized that *C.trachomatis* infection may be involved in chronic tubal inflammation and subsequent fimbrial carcinogenesis and hence could perhaps be the origin of some ovarian cancers [67]. Interestingly, a recent retrospective analysis of 34 cases of serous pelvic carcinomas (which are the most frequent histologic type of ovarian cancers)
Chlamydial infection of the pelvis can lead to symptomatic pelvic inflammatory disease (PID) that can result in Fallopian tube inflammation [69] and salpingitis is reported to occur in 20-40% of women with untreated or repeated Chlamydial infections [70,71]. A recent model predicts that annual screening would prevent 61% of *C.trachomatis*-related PID in women who became infected in their LGTs with *Chlamydia* [72]. In addition to salpingitis the long-term consequen‐ ces of PID can also include chronic pelvic pain and ectopic pregnancy (EP). It has been proposed that since *C.trachomatis* infection increases tubal expression of prokineticin receptor (PROKR2) mRNA this expression then predisposes to ectopic implantation in the Fallopian tube [73]. Recent papers have reported that one third of EPs could be attributable to Chlamydial infection [74] and have also reviewed the pathogenesis of *Chlamydia*-induced tubal EP [75].
PID is caused by infection of the female genital tract with many microorganisms including *C*. *trachomatis* [reviewed in 16] and following symptomatic PID of polymicrobial etiology up to 18% of women may develop tubal factor infertility (TFI), which is amongst the leading causes of infertility accounting for 7-9.8% of all female factor infertilities [76-80]. Testing for serum antibody to the Chlamydial 60kDa Heat shock protein (i.e.cHSP60 antibody) is an accurate means for predicting *Chlamydia*-associated TFI [81] and elevated levels of anti-Chlamydial caseinolytic protease P (ClpP) antibodies were recently identified in 21 TFI patients [82]. Prolonged exposure to *Chlamydia* due to a chronic persistent infection or frequent re-infection has also been associated with chronic inflammation and TFI [83-85]. It has been reported that 9.5% women developed PID over 12 months in women who tested positive for *Chlamydia* at
reported that 70% of cases did in fact involve the Fallopian tubes [68].
bacterial survival in the host [61].
*1.4.1. Infections in women*
180 Immune Response Activation
baseline [86].
**1.4. Chlamydial infections and disease sequelae**
The high prevalence of Chlamydial infections of the cervix in women of child-bearing age results in annual exposure of an estimated 100,000 neonates to *Chlamydia* infections [87] and neonatal Chlamydial infections are generally acquired during passage through an infected birth canal. A Chlamydial cause should be considered for all infants aged ≤30 days who have conjunctivitis and, if positively diagnosed, appropriate treatment initiated for the infant, mother and her sex partner(s).
*Chlamydia* infections of neonates have been reported to be at 18 percent for ophthalmia neonatorum and 16 percent for pneumonia for infants exposed to *C. trachomatis* in vaginal secretions in the birth canal, with up to 60% of these infants showing serological evidence of Chlamydial infection [88]. Using data pooled from studies since 1977 it has been estimated that the incidence of ophthalmia neonatorum caused by exposure of the neonate to *C.tracho‐ matis* at birth is actually around 15 percent and that of pneumonia is around 7 percent [89].
Neonatal infection with *C. trachomatis* involves the mucous membranes of the neonatal eye, oropharynx, urogenital tract, and rectum and when symptomatic can result in a number of diseases in the first few months of life such as inclusion conjunctivitis, pneumonia or both [90,91; reviewed in 92; 93). Ophthalmia neonatorum has been reported as developing from 5-14 days post-delivery in up to 50% of infants of *C.trachomatis* infected mothers [94] and perhaps even as late as 42 days after birth [94] with between 5-30% of perinatally-infected infants also having nasopharyngeal infection [reviewed in 95]. A recent prospective study from Brazil estimated the prevalence of rate of *Chlamydia* among pregnant women (<30 years) in their third trimester as 72.7% and reported that 50% of 16 newborn babies had respiratory symptoms within the first 60 days of life [96].
Afebrile pneumonia may be seen in infants colonised in the respiratory tract with *C.trachoma‐ tis* and this may/may not be associated with inclusion conjunctivitis. After three or more weeks following Chlamydial infection of the infants a subacute, afebrile pneumonia with onset at ages 1–3 months develops in approximately seven percent of these cases [97, 92, 19]. Clinical characteristics have been described for infant pneumonia caused by *C.trachomatis* in infants with nasopharyngeal shedding and in around half of the infected infants conjunctivitis was recorded, with middle ear abnormalities recorded in more than half of the *Chlamydia*-positive infants and wheezing an unusual finding in these infants [98]. A more recent study has also highlighted that *C. trachomatis* is an important cause of lower respiratory tract infection in infants in India below six months of age [99].
#### **1.5. Immunopathogenesis and immunoprotection**
The host inflammatory response to *C.trachomatis* is involved in immunity but also the pathol‐ ogy that leads to serious morbidities of chronic pelvic pain, EP and tubal infertility following female genital tract infection. Protective responses to *C.trachomatis* infections of the female genital tract are elicited to clear primary infection and to resist re-infection and include both innate and adaptive (particularly CD4+ T cell) cells and immune responses. Cells of both the innate and adaptive immune systems are found in the FGT in Fallopian tubes, uterus, cervix and vagina [100] and we have recently reviewed the localisation of T lymphocyte populations in the female upper (UGT) and lower (LGT) genital tracts over the menstrual cycle [101].
During active primary genital infections in women, serum and genital mucosal IgA and IgG antibodies to Chlamydial EBs and specific Chlamydial proteins including heat-shock (HSP) and plasmid proteins are usually detected [118]. AntiChlamydial antibodies are not sufficient
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
183
Once *Chlamydia* has established intracellular infection, cells of the adaptive immune system, and particularly T helper 1 (Th1) type CD4+ T cells secreting IFN-g are required for clearance of primary infection and to protect from re-infection [120]. It has been observed from human data that MHC Class II-restricted CD4+T cells of the Th1 phenotype are critical for the recovery from primary Chlamydial infection and also have a role in protecting from disease sequelae [121,122]. Other major lineages of activated CD4+ T cells that play critical roles in Chlamydial infections include the Th2 cells, Th17 cells (producing IL-17 and IL-23) and Tregulatory (Treg) cells [123,124]. Additional T cells involved in adaptive immunity include intraepithelial lymphocytes (γδ T cells) and cytotoxic T lymphocytes (CD8+ T cells) that are known to induce
Disease outcomes are dependent upon the complex interactions between virulence factors of, and evasion strategies used by, *C.trachomatis* and host immune responses including variants in genetic markers associated with infection and inflammation. Host factors and cellular immune responses associated with susceptibility to, or protection from Chlamydial infection
**2. Chlamydial genital tract infection and immune regulation by sex**
The FGT comprises several immune compartments found in the UGT (endocervix, ovaries, Fallopian tubes and uterus) and LGT (ectocervix, and vagina). The UGT is lined with a single layer of columnar epithelium and the LGT with is lined with stratified squamous epithelium [100]. The transitional zone where squamous epithelium changes to columnar epithelium is
The human FGT contains components of the innate and adaptive mucosal immune responses that are found in various distributions at different sites throughout the tract. In the FGT the major lymphocyte components are Natural Killer (NK) cells and T lymphocytes including CD3+ T lymphocytes that are present in all tissues of the tract. In the LGT the CD8+ and CD4+ are dispersed throughout the stroma whilst lymphoid aggregates of these cells are formed in the uterus [128]; granulocytes are present and these are principally located in the Fallopian tubes. Finally, relative to T lymphocytes, smaller numbers of monocytes and B lymphocytes
Host protection of the FGT is afforded by two arms of defense comprising innate (non-specific) and acquired (specific) immunity. The cellular components of the specific immune responses at this site are T lymphocytes expressing αβ- and γδ-T cell receptors (TCR) and immunoglo‐ bulins (Ig) producing B lymphocytes [130]. Effectors of non-specific responses include
to protect infected females from re-infections [119].
apoptosis of infected Chlamydial cells [125].
and/or diseases have recently been reviewed [126].
the most immunologically active site of the FGT [127].
are found throughout all tissues of the FGT [129].
**hormones**
Effector cells of the innate (non-specific) immune response found in the FGT include epithelial cells, monocytes, macrophages, granulocytes (neutrophils, eosinophils and basophils), dendritic cells (DCs) and Natural Killer (NK) cells conferring protection through chemokines, cytokines and phagocytosis. Innate (non-specific) host defense particularly in the lower genital tract (LGT) is also provided by the vaginal microbiota maintained in a healthy equilibrium (reviewed in [102] as well as by the myriad of antimicrobial peptides (AMPs) and protease inhibitors of cervicovaginal fluid [103] including elafin, lysozyme and cathelicidins, amongst others and reviewed in (Wira and Fahey, 2004). The AMPs and pathogen recognition receptors (PRRs) such as toll-like receptors (TLRs) are key intermediaries of innate immunity in the FGT [104-107]. The innate immune cells (especially neutrophils) elicit inflammation to protect against infections however these inflammatory responses are usually inefficient and repeated Chlamydial infections are common [108].
The rapid innate immune response is the first line of defence against a *C. trachomatis* infection of columnar epithelial cells. PRRs recognise the pathogen and TLRs play a vital role in the host immune system by recognizing pathogenic components (pathogen associated molecular patterns (PAMPs) and danger associated molecular patterns (DAMPs)) and inducing a protective immune response to the pathogen [109]. Data indicate that host genetic factors may contribute up to 40% to the variation in clinical course of *Chlamydia* infection [110]. Host genetic variants in TLRs have been found to influence *Chlamydia* LGT infections in women [111]. For example protection against tubal disease following a *C.trachomatis* infection has been report‐ edly associated with single nucleotide polymorphisms (SNPs) in the innate immune (TLR) gene, TLR2 [112]. SNPs in TLR4 have also been reported to play a role in making women more prone to subfertility (i.e. no pregnancy after 1 year) as a late complication of *Chlamydia* infection [113] and worldwide estimates are that 10-15% of all couples are subfertile [114]. SNPs in other pathogen recognition receptors including nucleotide-binding oligomerization domain receptors (NODs) have been reported to influence susceptibility and severity of *Chlamydia* infections [115]. Taken together these observations have lead to the recent suggested potential translational and clinical value of adding diagnostic host genetic marker profiles on the basis of infection and inflammation to the current clinical management of subfertility [116].
The more gradual adaptive immune response to Chlamydial infection is mediated through antibodies and/or T cells and is a very specific recognition and response to foreign antigens. Cervical cell production of IFN-g and interleukin-12 (IL-12) in response to stimulation with *C.trachomatis* EBs has a positive correlation with fertility in *C.trachomatis* seropositive patients [117]. Recruitment and activation of B cell (humoral) and T cell (cell-mediated/adaptive) immunity is also coordinated by the release of these factors. The initiation of this adaptive response may lead to spontaneous resolution of natural infection in some populations [25] although chronic infections in women indicate that rarely does this response to *C.trachomatis* result in clearance of the infection; nor does it adequately protect against re-infection. The cellular components of this specific immunity include T lymphocytes expressing αβ- and γδ-T cell receptors (TCR) and immunoglobulin (Ig) producing B lymphocytes [reviewed in 101]. During active primary genital infections in women, serum and genital mucosal IgA and IgG antibodies to Chlamydial EBs and specific Chlamydial proteins including heat-shock (HSP) and plasmid proteins are usually detected [118]. AntiChlamydial antibodies are not sufficient to protect infected females from re-infections [119].
and vagina [100] and we have recently reviewed the localisation of T lymphocyte populations in the female upper (UGT) and lower (LGT) genital tracts over the menstrual cycle [101].
Effector cells of the innate (non-specific) immune response found in the FGT include epithelial cells, monocytes, macrophages, granulocytes (neutrophils, eosinophils and basophils), dendritic cells (DCs) and Natural Killer (NK) cells conferring protection through chemokines, cytokines and phagocytosis. Innate (non-specific) host defense particularly in the lower genital tract (LGT) is also provided by the vaginal microbiota maintained in a healthy equilibrium (reviewed in [102] as well as by the myriad of antimicrobial peptides (AMPs) and protease inhibitors of cervicovaginal fluid [103] including elafin, lysozyme and cathelicidins, amongst others and reviewed in (Wira and Fahey, 2004). The AMPs and pathogen recognition receptors (PRRs) such as toll-like receptors (TLRs) are key intermediaries of innate immunity in the FGT [104-107]. The innate immune cells (especially neutrophils) elicit inflammation to protect against infections however these inflammatory responses are usually inefficient and repeated
The rapid innate immune response is the first line of defence against a *C. trachomatis* infection of columnar epithelial cells. PRRs recognise the pathogen and TLRs play a vital role in the host immune system by recognizing pathogenic components (pathogen associated molecular patterns (PAMPs) and danger associated molecular patterns (DAMPs)) and inducing a protective immune response to the pathogen [109]. Data indicate that host genetic factors may contribute up to 40% to the variation in clinical course of *Chlamydia* infection [110]. Host genetic variants in TLRs have been found to influence *Chlamydia* LGT infections in women [111]. For example protection against tubal disease following a *C.trachomatis* infection has been report‐ edly associated with single nucleotide polymorphisms (SNPs) in the innate immune (TLR) gene, TLR2 [112]. SNPs in TLR4 have also been reported to play a role in making women more prone to subfertility (i.e. no pregnancy after 1 year) as a late complication of *Chlamydia* infection [113] and worldwide estimates are that 10-15% of all couples are subfertile [114]. SNPs in other pathogen recognition receptors including nucleotide-binding oligomerization domain receptors (NODs) have been reported to influence susceptibility and severity of *Chlamydia* infections [115]. Taken together these observations have lead to the recent suggested potential translational and clinical value of adding diagnostic host genetic marker profiles on the basis of infection and inflammation to the current clinical management of subfertility [116].
The more gradual adaptive immune response to Chlamydial infection is mediated through antibodies and/or T cells and is a very specific recognition and response to foreign antigens. Cervical cell production of IFN-g and interleukin-12 (IL-12) in response to stimulation with *C.trachomatis* EBs has a positive correlation with fertility in *C.trachomatis* seropositive patients [117]. Recruitment and activation of B cell (humoral) and T cell (cell-mediated/adaptive) immunity is also coordinated by the release of these factors. The initiation of this adaptive response may lead to spontaneous resolution of natural infection in some populations [25] although chronic infections in women indicate that rarely does this response to *C.trachomatis* result in clearance of the infection; nor does it adequately protect against re-infection. The cellular components of this specific immunity include T lymphocytes expressing αβ- and γδ-T cell receptors (TCR) and immunoglobulin (Ig) producing B lymphocytes [reviewed in 101].
Chlamydial infections are common [108].
182 Immune Response Activation
Once *Chlamydia* has established intracellular infection, cells of the adaptive immune system, and particularly T helper 1 (Th1) type CD4+ T cells secreting IFN-g are required for clearance of primary infection and to protect from re-infection [120]. It has been observed from human data that MHC Class II-restricted CD4+T cells of the Th1 phenotype are critical for the recovery from primary Chlamydial infection and also have a role in protecting from disease sequelae [121,122]. Other major lineages of activated CD4+ T cells that play critical roles in Chlamydial infections include the Th2 cells, Th17 cells (producing IL-17 and IL-23) and Tregulatory (Treg) cells [123,124]. Additional T cells involved in adaptive immunity include intraepithelial lymphocytes (γδ T cells) and cytotoxic T lymphocytes (CD8+ T cells) that are known to induce apoptosis of infected Chlamydial cells [125].
Disease outcomes are dependent upon the complex interactions between virulence factors of, and evasion strategies used by, *C.trachomatis* and host immune responses including variants in genetic markers associated with infection and inflammation. Host factors and cellular immune responses associated with susceptibility to, or protection from Chlamydial infection and/or diseases have recently been reviewed [126].
### **2. Chlamydial genital tract infection and immune regulation by sex hormones**
The FGT comprises several immune compartments found in the UGT (endocervix, ovaries, Fallopian tubes and uterus) and LGT (ectocervix, and vagina). The UGT is lined with a single layer of columnar epithelium and the LGT with is lined with stratified squamous epithelium [100]. The transitional zone where squamous epithelium changes to columnar epithelium is the most immunologically active site of the FGT [127].
The human FGT contains components of the innate and adaptive mucosal immune responses that are found in various distributions at different sites throughout the tract. In the FGT the major lymphocyte components are Natural Killer (NK) cells and T lymphocytes including CD3+ T lymphocytes that are present in all tissues of the tract. In the LGT the CD8+ and CD4+ are dispersed throughout the stroma whilst lymphoid aggregates of these cells are formed in the uterus [128]; granulocytes are present and these are principally located in the Fallopian tubes. Finally, relative to T lymphocytes, smaller numbers of monocytes and B lymphocytes are found throughout all tissues of the FGT [129].
Host protection of the FGT is afforded by two arms of defense comprising innate (non-specific) and acquired (specific) immunity. The cellular components of the specific immune responses at this site are T lymphocytes expressing αβ- and γδ-T cell receptors (TCR) and immunoglo‐ bulins (Ig) producing B lymphocytes [130]. Effectors of non-specific responses include epithelial cells, monocytes, macrophages, granulocytes (neutrophils, eosinophils and baso‐ phils), dendritic cells (DCs) and Natural Killer (NK) cells that confer protection due to chemokines, cytokines and phagocytosis. Cells of both the innate and adaptive immune systems are found in Fallopian tubes, uterus, cervix and vagina [131].
proteoglycan motif produced by *C. trachomatis* and *C.muridarum*) [147]. Expression of TLR genes 2,3,4,5,6,9 and 10 is reportedly significantly higher in human endometrial tissue during the secretory phase (when P4 is the predominant hormone) compared to all other
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
185
Changes in the populations of DC and macrophage cells in the murine FGT occur due to varying levels of ovarian steroid hormones at this site in the model [149], and in women, changes in these cell populations occur due to the varying stages of the menstrual cycle, noting that the frequencies of macrophages and immature DCs are higher in the menstru‐ al phase when compared to all other phases of the cycle [150, 151]. CD16-expressing macrophages and monocytes are also key components of the innate immune system of the FGT and are cells that are capable of producing cytokines and chemokines upon activa‐ tion. These cells are also known to be profoundly affected by E2 [reviewed in 152]. Studies have shown that E2 uses both estrogen receptor (ER) alpha (ERα) and ER beta (ERβ) to decrease CD16 expression on monocytes and hence to alter monocytic cytokine release
T lymphocytes present in the FGT, specifically CD4+ and CD8+ T cells, and immune mediators of cell-mediated immunity are found distributed within distinct regions of the female LGT. The normal vaginal mucosa contains T cells and APCs with the few immune cells present consisting primarily of CD1a+ DCs and CD8+ intraepithelial lymphocytes (IELs) [154]. The tissues in the UGT have been found to express a greater number Th1-associated chemokines than the tissues in the LGT comprising the cervical-vaginal region [153]. The menstrual cycle and menopause have been reported to have no apparent effect on cellular localization or abundance of T lymphocyte subsets and APCs in any of the lower genital tract tissues [127]. In humans, expansion of human CD4+CD25+ regulatory T cells (Tregs) has been reported to occur in the late follicular (proliferative) phases of the menstrual cycles of fertile women [124].
Levels of IgG and IgA antibodies in FGT fluids are also known to be modulated by the stage of the menstrual cycle in women. E2 up-regulates the *in vitro* expression in the FGT of secretory component (SC) increasing transport of IgA into the lumen [154] and transport of polymeric IgA (pIgA) into FGT tissues is significantly decreased by ovariectomy, although this decline can be reversed by addition of E2 but not P4 [155]. The effects of ovarian hormones on immunoglobulin secreting cells (ISC) function in healthy women were also investigated and it was reported that ISC frequency in peripheral blood mononuclear cells (PMBC) was highest
It is becoming increasingly well documented that sex hormones E2 and P4 regulate host susceptibility, as well as innate and adaptive immune responses to the sexually transmitted
Data from several animal models of genital tract Chlamydial infection have shown that Chlamydial infection may be influenced by timing of exposure and the steroid status of the mucosal epithelium. In the murine model of *C.muridarum* infection pre-exposure of the animals
pathogen *Chlamydia* infecting the mucosal surfaces of the FGT [reviewed in 101].
during the peri-ovulatory stage of the menstrual cycle [156].
**2.1. Sex hormones and chlamydial infection**
phases of the menstrual cycle [148].
following CD16 receptor activation [153].
Both innate and adaptive mucosal immunity in the FGT are regulated by the female sex hormones estrogen (E2) and progesterone (pregn-4-ene3,20-dione) or (P4) and have been reviewed [130, 132, reviewed in 101]. Hormones regulate the transport of Igs, levels of cytokines, expression of Toll-like receptor (TLR) genes [133,134] and the distribution of immune cells and antigen presentation in the genital tissues during the reproductive cycle. Estradiol regulates the structure and function of the FGT via the classic model of steroid action ('i.e. genomic regulation) which is mediated by activated nuclear receptors including estrogen receptors (ER)-α and ER-β [131,135]. Estradiol can also exert rapid "non-classic" steroid effects on diverse signalling pathways and second messenger systems that occur independently of transcriptional or classic genomic regulation [136]. These rapid responses are often referred to as 'non-genomic' steroid effects and they can affect genital tract cells and innate and humoral immunity at this site which may alter the pathogens ability to infect the genital tract [137].
The recruitment and function of immune components of the FGT are precisely regulated by hormonal changes during the menstrual cycle by the sex steroids that co-ordinate cell traf‐ ficking and immune activation at this mucosal site. Female sex hormones E2, progestins and androgens are produced by ovarian cells with levels of estradiol (17β-estradiol) and P4 varying in accordance with fluctuations in the menstrual cycle [130]. In the menstrual/ovarian cycle of humans the endometrium develops and follicles grow until ovulation in the Proliferative/ Follicular phase. The Secretory/Luteal phase is characterised by high levels of progesterone from the corpus luteum to maintain the endometrium. Finally, corpus luteum regression and menstruation occurs during the menstrual phase At menstruation, serum estradiol levels measure typically <50pg/ml. Gradually increasing amounts of estradiol are found in the follicular (proliferative) phase of the adult female menstrual cycle with serum levels at a preovulatory stage (day 14) ranging from 110-410 pg/ml and dropping briefly at ovulation. Levels of estradiol (20-160pg/mol) and P4(>5ng/ml) are present during the luteal (secretory) phase from days 14-28 of the menstrual cycle typically peaking at day 21 of the cycle. At the end of the secretory phase estradiol levels return to their menstrual levels.
The innate response is triggered by PRRs (including TLRs and non-TLRs such as NODlike receptors) expressed predominantly by innate immune effector cells such as macrophag‐ es, neutrophils and DCs that are found in the genital mucosa. Human oviduct epithelial cells (ECs) express Toll-like receptor 3 and E2 has been reported to suppress the TLR-3 induced cytokine and chemokine production in human endometrial epithelial cells [138].The effects of E2 on the differentiation and function of antigen-presenting cells (APCs) *in vitro* has been reviewed (see [139; 132]. Much of the current evidence shows that E2 can activate dendritic cells and differentially regulate adaptive immune responses through direct effects on DC functions [140] via E2 receptor (ER) ligands [141]. These include TLR4 and its coreceptor Cluster of Differentiation (CD) 14 (Chlamydial LPS and Chlamydial HSP60) [142; 143; 144] TLR2 (Chlamydial HSP60) [145,146] and NOD protein, Nod2 (rudimentary proteoglycan motif produced by *C. trachomatis* and *C.muridarum*) [147]. Expression of TLR genes 2,3,4,5,6,9 and 10 is reportedly significantly higher in human endometrial tissue during the secretory phase (when P4 is the predominant hormone) compared to all other phases of the menstrual cycle [148].
Changes in the populations of DC and macrophage cells in the murine FGT occur due to varying levels of ovarian steroid hormones at this site in the model [149], and in women, changes in these cell populations occur due to the varying stages of the menstrual cycle, noting that the frequencies of macrophages and immature DCs are higher in the menstru‐ al phase when compared to all other phases of the cycle [150, 151]. CD16-expressing macrophages and monocytes are also key components of the innate immune system of the FGT and are cells that are capable of producing cytokines and chemokines upon activa‐ tion. These cells are also known to be profoundly affected by E2 [reviewed in 152]. Studies have shown that E2 uses both estrogen receptor (ER) alpha (ERα) and ER beta (ERβ) to decrease CD16 expression on monocytes and hence to alter monocytic cytokine release following CD16 receptor activation [153].
T lymphocytes present in the FGT, specifically CD4+ and CD8+ T cells, and immune mediators of cell-mediated immunity are found distributed within distinct regions of the female LGT. The normal vaginal mucosa contains T cells and APCs with the few immune cells present consisting primarily of CD1a+ DCs and CD8+ intraepithelial lymphocytes (IELs) [154]. The tissues in the UGT have been found to express a greater number Th1-associated chemokines than the tissues in the LGT comprising the cervical-vaginal region [153]. The menstrual cycle and menopause have been reported to have no apparent effect on cellular localization or abundance of T lymphocyte subsets and APCs in any of the lower genital tract tissues [127]. In humans, expansion of human CD4+CD25+ regulatory T cells (Tregs) has been reported to occur in the late follicular (proliferative) phases of the menstrual cycles of fertile women [124].
Levels of IgG and IgA antibodies in FGT fluids are also known to be modulated by the stage of the menstrual cycle in women. E2 up-regulates the *in vitro* expression in the FGT of secretory component (SC) increasing transport of IgA into the lumen [154] and transport of polymeric IgA (pIgA) into FGT tissues is significantly decreased by ovariectomy, although this decline can be reversed by addition of E2 but not P4 [155]. The effects of ovarian hormones on immunoglobulin secreting cells (ISC) function in healthy women were also investigated and it was reported that ISC frequency in peripheral blood mononuclear cells (PMBC) was highest during the peri-ovulatory stage of the menstrual cycle [156].
#### **2.1. Sex hormones and chlamydial infection**
epithelial cells, monocytes, macrophages, granulocytes (neutrophils, eosinophils and baso‐ phils), dendritic cells (DCs) and Natural Killer (NK) cells that confer protection due to chemokines, cytokines and phagocytosis. Cells of both the innate and adaptive immune
Both innate and adaptive mucosal immunity in the FGT are regulated by the female sex hormones estrogen (E2) and progesterone (pregn-4-ene3,20-dione) or (P4) and have been reviewed [130, 132, reviewed in 101]. Hormones regulate the transport of Igs, levels of cytokines, expression of Toll-like receptor (TLR) genes [133,134] and the distribution of immune cells and antigen presentation in the genital tissues during the reproductive cycle. Estradiol regulates the structure and function of the FGT via the classic model of steroid action ('i.e. genomic regulation) which is mediated by activated nuclear receptors including estrogen receptors (ER)-α and ER-β [131,135]. Estradiol can also exert rapid "non-classic" steroid effects on diverse signalling pathways and second messenger systems that occur independently of transcriptional or classic genomic regulation [136]. These rapid responses are often referred to as 'non-genomic' steroid effects and they can affect genital tract cells and innate and humoral immunity at this site which may alter the pathogens ability to infect the genital tract [137].
The recruitment and function of immune components of the FGT are precisely regulated by hormonal changes during the menstrual cycle by the sex steroids that co-ordinate cell traf‐ ficking and immune activation at this mucosal site. Female sex hormones E2, progestins and androgens are produced by ovarian cells with levels of estradiol (17β-estradiol) and P4 varying in accordance with fluctuations in the menstrual cycle [130]. In the menstrual/ovarian cycle of humans the endometrium develops and follicles grow until ovulation in the Proliferative/ Follicular phase. The Secretory/Luteal phase is characterised by high levels of progesterone from the corpus luteum to maintain the endometrium. Finally, corpus luteum regression and menstruation occurs during the menstrual phase At menstruation, serum estradiol levels measure typically <50pg/ml. Gradually increasing amounts of estradiol are found in the follicular (proliferative) phase of the adult female menstrual cycle with serum levels at a preovulatory stage (day 14) ranging from 110-410 pg/ml and dropping briefly at ovulation. Levels of estradiol (20-160pg/mol) and P4(>5ng/ml) are present during the luteal (secretory) phase from days 14-28 of the menstrual cycle typically peaking at day 21 of the cycle. At the end of
The innate response is triggered by PRRs (including TLRs and non-TLRs such as NODlike receptors) expressed predominantly by innate immune effector cells such as macrophag‐ es, neutrophils and DCs that are found in the genital mucosa. Human oviduct epithelial cells (ECs) express Toll-like receptor 3 and E2 has been reported to suppress the TLR-3 induced cytokine and chemokine production in human endometrial epithelial cells [138].The effects of E2 on the differentiation and function of antigen-presenting cells (APCs) *in vitro* has been reviewed (see [139; 132]. Much of the current evidence shows that E2 can activate dendritic cells and differentially regulate adaptive immune responses through direct effects on DC functions [140] via E2 receptor (ER) ligands [141]. These include TLR4 and its coreceptor Cluster of Differentiation (CD) 14 (Chlamydial LPS and Chlamydial HSP60) [142; 143; 144] TLR2 (Chlamydial HSP60) [145,146] and NOD protein, Nod2 (rudimentary
systems are found in Fallopian tubes, uterus, cervix and vagina [131].
184 Immune Response Activation
the secretory phase estradiol levels return to their menstrual levels.
It is becoming increasingly well documented that sex hormones E2 and P4 regulate host susceptibility, as well as innate and adaptive immune responses to the sexually transmitted pathogen *Chlamydia* infecting the mucosal surfaces of the FGT [reviewed in 101].
Data from several animal models of genital tract Chlamydial infection have shown that Chlamydial infection may be influenced by timing of exposure and the steroid status of the mucosal epithelium. In the murine model of *C.muridarum* infection pre-exposure of the animals to P4 is required to achieve maximal infection outcomes in the animals [157]. However, treatment of female guinea pigs with E2 resulted in *Chlamydia caviae* infections of greater intensity and longer duration [157-160] an effect that it also observed following oral contra‐ ceptive administration to these animals [161]. In the female reproductive tract rat model it was reported that no Chlamydial inclusions were detected in either the uterus or vagina in rats that had been infected with *C. trachomatis* (mouse pneumonitis strain MoPn) at estrus and diestrus without prior P4 priming [162]. In rat studies it has also been reported that when ovariectom‐ ized animals were administered estradiol, progesterone, or a combination of both and infected with *Chlamydia trachomatis* via the intrauterine route no Chlamydiae were found vaginal secretions of E2-treated animals nor were there any signs of uterine or vaginal inflammation in these animals [163].
Studies of *C.trachomatis* infections of the FGT have reported many effects of hormones on infection and immunity to the pathogen at this mucosal site. For example significant increases in the sensitivities of cervical epithelial cells to infection with *C.trachomatis* occur in the later stages of the female menstrual cycle [174]. It has been reported that Chlamydial infection occurring in the early E2-dominant phases of the cycle was a significant predictor for devel‐ opment of salpingitis [175]. Clinical findings have also reported on the enhancing effects of E2 on Chlamydial infection and disease sequelae in the infected female genital tract. It has been reported that women are more susceptible to Chlamydial infection under the influence of E2 [165]. In women with fertility disorders (n= 115) and in women with *C.trachomatis* mucopur‐ ulent cervicitis (n=86) a significantly positive correlation was recorded between Chlamydial
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
187
Chemokine receptors direct T lymphocytes to the site of mucosal infection, including infection of genital tract epithelial cells with *C.trachomatis* and these receptors can be modulated by E2. In a study examining samples from human patients it was reported that CCR5 (chemokine (C-C motif) receptor 5) expression increased followed *C. trachomatis* genital infections [125].
Reproductive hormones may potentially act with cytokines to regulate immune responses in the FGT to Chlamydial genital tract infections *in vivo*. Significant negative correlation has been reported between E2 levels in women with primary Chlamydial genital infections and cervical wash concentrations of IL-10, IL-1beta and IL-6 cytokines. Significant negative correlations were also recorded between IL1-beta cytokines and P4 levels in women with recurrent Chlamydial infections [176]. Thus it would appear that Chlamydial infection of the human FGT is modulated in part by the combined actions of cytokines with E2 and P4. More recent investigations reveal that levels of E2 were significantly higher in *Chlamydia*-positive women with fertility disorders when compared to fertile women suggesting that this sex hormone contributes to development of disease sequelae to *C.trachomatis* infection of the FGT[165].
E2 has been reported to exacerbate Chlamydial infections in women following administration of oral contraceptives [177] an observation also noted for female guinea pigs [161]. A prospec‐ tive study on oral contraceptive pills or depo medroxyprogesterone acetate and sexually transmitted disease acquisition reported that women using oral contraceptive pills were at increased risk for acquisition of *Chlamydia* [178]. Hormonal contraceptive use in women, particularly depo medroxyprogesterone (DMPA), has also been reported in observational studies to be associated with a four-fold increase in Chlamydial genital tract infections [179].
**3. Immunopathogenesis — chlamydial infections, tubal factor infertility,**
A primary Chlamydial infection of the FGT can elicit host immune responses that are involved both in immunity to infection and in pathology of disease. Innate immune responses constitute the first line of defence against pathogens and include TLRs that are expressed in the FGT. Further, although the host immune response affords protection against infection *via* means of
load and E2 levels [117].
**ectopic pregnancy and pid**
Additionally, the use of *in vitro* cell lines to investigate *C. trachomatis* infection have produced data showing that the hormonal status of the epithelium can influence the outcome of Chlamydial infections in these models. It was shown that E2 treatment of HeLa cells preinfection with *C.trachomatis* (serovar K or L1) enhanced both of adherence of EBs to the cells as well as Chlamydial inclusion formation; post-infection, the presence of E2 was required for the enhancement of inclusion formation [164]. Using a hormone responsive human endome‐ trial cell line an association between repeated Chlamydial infections in women and high levels of E2 has been reported [165]. More recently it has been reported that the sex hormones E2 and P4 can directly modulate Chlamydial genes associated with persistence (the viable but nonculturable state in the Chlamydial developmental cycle) in the host. It is likely, however, that the major effects of these hormones on Chlamydial infection are indirect via the host cells [166]. The outer membrane complex B (*omcB)* and tryptophan B (*trpB*) genes are two of several reliable markers for Chlamydial persistence [167,168]. It was reported that in E2-supplemented cultures of *C.trachomatis* serovar D infections of a hormone-responsive human endometrial cell line, an up-regulation of the *omcB* and *trpB* genes was observed suggesting a stress response indicative of Chlamydial persistence; morphological changes (aberrant, enlarged RBs) were also observed in these cultures consistent with a persistence response [166]. P4-supplemented cultures resulted in a general up-regulation of genes encoding amino acid and carbohydrate metabolism pathways [166].
*Chlamydia trachomatis* infection of the female LGT stimulates innate immune cells by activation of TLR2/TLR4 [reviewed in 169]. TLRs 2 and 4 have been found at the highest expression levels in endometrium and Fallopian tube tissue [170] with the predominant expression of TLR4 and its co-receptor CD14 in the Fallopian tubes proposed to play an important role in the innate host defence mechanism against ascending *C. trachomatis* infections [169,171]. In a recent study *C.trachomatis*-infected women have been reported to express higher TLR2, TLR4 and inducible nitric oxide synthase (iNOS) in their cervical monocytes compared to controls. It is thought that TLR4 initiates the innate immune response to Chlamydial infection, activation of TLR2 leads to expression of inflammatory cytokines whilst iNOS contributes to Chlamydial clearance from this mucosal site [172]. Recently it has been reported that E2 reduces TLR4 expression of human monocyte derived Dendritic cells and suggested that this may increase host susceptibility to *C. trachomatis* infections [173].
Studies of *C.trachomatis* infections of the FGT have reported many effects of hormones on infection and immunity to the pathogen at this mucosal site. For example significant increases in the sensitivities of cervical epithelial cells to infection with *C.trachomatis* occur in the later stages of the female menstrual cycle [174]. It has been reported that Chlamydial infection occurring in the early E2-dominant phases of the cycle was a significant predictor for devel‐ opment of salpingitis [175]. Clinical findings have also reported on the enhancing effects of E2 on Chlamydial infection and disease sequelae in the infected female genital tract. It has been reported that women are more susceptible to Chlamydial infection under the influence of E2 [165]. In women with fertility disorders (n= 115) and in women with *C.trachomatis* mucopur‐ ulent cervicitis (n=86) a significantly positive correlation was recorded between Chlamydial load and E2 levels [117].
to P4 is required to achieve maximal infection outcomes in the animals [157]. However, treatment of female guinea pigs with E2 resulted in *Chlamydia caviae* infections of greater intensity and longer duration [157-160] an effect that it also observed following oral contra‐ ceptive administration to these animals [161]. In the female reproductive tract rat model it was reported that no Chlamydial inclusions were detected in either the uterus or vagina in rats that had been infected with *C. trachomatis* (mouse pneumonitis strain MoPn) at estrus and diestrus without prior P4 priming [162]. In rat studies it has also been reported that when ovariectom‐ ized animals were administered estradiol, progesterone, or a combination of both and infected with *Chlamydia trachomatis* via the intrauterine route no Chlamydiae were found vaginal secretions of E2-treated animals nor were there any signs of uterine or vaginal inflammation
Additionally, the use of *in vitro* cell lines to investigate *C. trachomatis* infection have produced data showing that the hormonal status of the epithelium can influence the outcome of Chlamydial infections in these models. It was shown that E2 treatment of HeLa cells preinfection with *C.trachomatis* (serovar K or L1) enhanced both of adherence of EBs to the cells as well as Chlamydial inclusion formation; post-infection, the presence of E2 was required for the enhancement of inclusion formation [164]. Using a hormone responsive human endome‐ trial cell line an association between repeated Chlamydial infections in women and high levels of E2 has been reported [165]. More recently it has been reported that the sex hormones E2 and P4 can directly modulate Chlamydial genes associated with persistence (the viable but nonculturable state in the Chlamydial developmental cycle) in the host. It is likely, however, that the major effects of these hormones on Chlamydial infection are indirect via the host cells [166]. The outer membrane complex B (*omcB)* and tryptophan B (*trpB*) genes are two of several reliable markers for Chlamydial persistence [167,168]. It was reported that in E2-supplemented cultures of *C.trachomatis* serovar D infections of a hormone-responsive human endometrial cell line, an up-regulation of the *omcB* and *trpB* genes was observed suggesting a stress response indicative of Chlamydial persistence; morphological changes (aberrant, enlarged RBs) were also observed in these cultures consistent with a persistence response [166]. P4-supplemented cultures resulted in a general up-regulation of genes encoding amino acid and carbohydrate
*Chlamydia trachomatis* infection of the female LGT stimulates innate immune cells by activation of TLR2/TLR4 [reviewed in 169]. TLRs 2 and 4 have been found at the highest expression levels in endometrium and Fallopian tube tissue [170] with the predominant expression of TLR4 and its co-receptor CD14 in the Fallopian tubes proposed to play an important role in the innate host defence mechanism against ascending *C. trachomatis* infections [169,171]. In a recent study *C.trachomatis*-infected women have been reported to express higher TLR2, TLR4 and inducible nitric oxide synthase (iNOS) in their cervical monocytes compared to controls. It is thought that TLR4 initiates the innate immune response to Chlamydial infection, activation of TLR2 leads to expression of inflammatory cytokines whilst iNOS contributes to Chlamydial clearance from this mucosal site [172]. Recently it has been reported that E2 reduces TLR4 expression of human monocyte derived Dendritic cells and suggested that this may increase
in these animals [163].
186 Immune Response Activation
metabolism pathways [166].
host susceptibility to *C. trachomatis* infections [173].
Chemokine receptors direct T lymphocytes to the site of mucosal infection, including infection of genital tract epithelial cells with *C.trachomatis* and these receptors can be modulated by E2. In a study examining samples from human patients it was reported that CCR5 (chemokine (C-C motif) receptor 5) expression increased followed *C. trachomatis* genital infections [125].
Reproductive hormones may potentially act with cytokines to regulate immune responses in the FGT to Chlamydial genital tract infections *in vivo*. Significant negative correlation has been reported between E2 levels in women with primary Chlamydial genital infections and cervical wash concentrations of IL-10, IL-1beta and IL-6 cytokines. Significant negative correlations were also recorded between IL1-beta cytokines and P4 levels in women with recurrent Chlamydial infections [176]. Thus it would appear that Chlamydial infection of the human FGT is modulated in part by the combined actions of cytokines with E2 and P4. More recent investigations reveal that levels of E2 were significantly higher in *Chlamydia*-positive women with fertility disorders when compared to fertile women suggesting that this sex hormone contributes to development of disease sequelae to *C.trachomatis* infection of the FGT[165].
E2 has been reported to exacerbate Chlamydial infections in women following administration of oral contraceptives [177] an observation also noted for female guinea pigs [161]. A prospec‐ tive study on oral contraceptive pills or depo medroxyprogesterone acetate and sexually transmitted disease acquisition reported that women using oral contraceptive pills were at increased risk for acquisition of *Chlamydia* [178]. Hormonal contraceptive use in women, particularly depo medroxyprogesterone (DMPA), has also been reported in observational studies to be associated with a four-fold increase in Chlamydial genital tract infections [179].
### **3. Immunopathogenesis — chlamydial infections, tubal factor infertility, ectopic pregnancy and pid**
A primary Chlamydial infection of the FGT can elicit host immune responses that are involved both in immunity to infection and in pathology of disease. Innate immune responses constitute the first line of defence against pathogens and include TLRs that are expressed in the FGT. Further, although the host immune response affords protection against infection *via* means of rapid and efficient bacterial destruction, resultant inflammatory responses may lead to tissue damage and/or persistent infection.
nomodulators may in fact, induce and sustain tissue damage and host inflammatory responses. Compounded with repeated infections, the continued production of cytokines may cause serious disease sequelae such as PID, TFI, chronic pelvic pain or EP. Chlamydial Hsp60 (cHsp60), a known potent inducer of host inflammatory responses, may protect against Chlamydial infections [186]; paradoxically, however, cHsp60 is also able to induce proinflammatory responses, which can lead to tissue damage in women infected with *C. tracho‐*
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
189
Recent data from a prospective study of 30 women diagnosed with a ruptured EP demon‐ strated that anti-cHsp60 immunity confers a higher risk of EP [190]. Furthermore, in those women who were positive for anti-cHsp60 antibodies and suffered an EP, all had significantly lower IL-1β levels than Chlamydial sero-negative EP cases [190]. Importantly, high serum concentrations of IL-1β in the Fallopian tubes correlated with successful uterine implantation of embryos [191-192]. Data obtained from a prospective cohort of women at high risk for *C. trachomatis* infection demonstrated that a Th1 response i.e. the release of IFN-g to cHSP60 was associated with protection against the bacterial infection [186]. In contrast, it was also shown that an IL-10 response to the same Chlamydial antigen increased infection risk. In a study that investigated cytokine responses of cervical mononuclear cells obtained from 153 women all serologically positive for *C. trachomatis*, it was found that exposure to cHsp60 and cHsp10 may drastically affect the host's mucosal immune function by increasing the production and subsequent release of IFN-g, IL-10 and tumour necrosis factor-α (TNF-α) [193] thereby
TLRs are responsible for initiation of the adaptive immune system, thereby leading to the ultimate destruction of foreign microbes. Moreover, since they recognise structurally con‐ served molecules from various bacteria, they play a vital role in host defence. Chlamydialinduced PID was investigated and it was reported that among women clinically diagnosed with PID, those who carried variants in TLR 1 and TLR 4 genes had significantly increased odds of *C. trachomatis* infection [191]. In a more recent study, it was concluded that racial variation in TLR variants amongst women with PID may alter signalling pathways subsequent to microbial recognition, thus causing an increase in associated inflammation [194]. An investigation of pregnant women positive for *C. trachomatis* showed that those with serological evidence of prenatal *C. trachomatis* infections were more likely to develop pre-eclampsia [192].
Polymorphic membrane proteins (Pmps) represent 13.6% of *C. trachomatis'* entire coding genome [195], therefore inferring a potential role in Chlamydial biology and virulence. In a group of 40 *C. trachomatis-*infected women with clinically diagnosed mild to moderate PID, antibody levels of PmpA, PmpD and PmpI were evaluated. Interestingly, those women with PmpA antibodies were less likely to fall pregnant (p=0.042) and achieve a live birth (p=0.005) compared with those sero-negative for PmpA [196] suggesting that the immunopathological damage may indeed be a result of PmpA virulence. Further, it was also reported that women who were sero-positive for PmpI antibodies had a higher incidence of upper genital tract infection (p=0.026). However, antibodies to PmpD and PmpI did not alter the risk of FGT
potentiating further tissue damage and disease sequelae.
*matis* [187-189].
inflammation or sequelae.
In reaction to Chlamydial entry in host cells, innate immune responses occur at the infected mucosal site 1-2 days post infection [180]. Intense inflammation and predominant mucosal infiltration of phagocytes, neutrophils and macrophages ingest and enzymatically destroy the bacteria. In conjunction with the production and subsequent activation of phagocytic cells, others, crucial to infection resolution are also activated [181]. Natural killer (NK) cells are activated in direct response to *C. trachomatis* infection or indirectly by macrophagic stimulation of IL-12, a potent NK cell-triggering cytokine. Once activated, NK cells produce IFN-g which stimulate macrophages, further promoting phagocytosis. Innate immunity may impede infection, however, the inaccessibility of *Chlamydia* to circulating antibodies whose sole purpose is elimination necessitates adaptive immune effector mechanisms of both humoral and cell-mediated immunity.
Activation of cell-mediated immunity post infection is demonstrated by the proliferative response ratio of peripheral blood lymphocytes to Chlamydial EBs. T cells, which functionally differentiate into Th1 or Th2 subtypes based upon cytokine secretion profiles, also amass at the infection site [182]. Further, interaction of *C. trachomatis* with the cytokine network is a key component for resolution of infection and one in which a cytokine response is elicited either through direct infection of epithelial cells or by host cell interaction. In concert with the inflammatory process, tissue repair, comprising the removal of dead cells and the ingrowth of fibroblasts, leads to scar formation and potential impairment of Fallopian tube function [183]. Moreover, all three tubal layers i.e. interstitial, muscular and serosal may be implicated in scar development. Infection of the serosal layer elicits an inflammatory exudate, present on the peritoneal surface, thereby adhering all adjacent surfaces and structures. Although the innate and adaptive immune responses systematically co-ordinate their efforts to ultimately resolve infection, the immune response to Chlamydial infection can cause deciliation of tubal epithe‐ lium, intralumineal adhesions, tubal occlusion and peritubal adhesions. As a result, such gross anatomical and pathological changes may lead to infertility, EP and possible spontaneous abortion [184].
Primary Chlamydial infections of the FGT thus induce both innate and adaptive host immune responses that contribute to clearance of the infections. However, results from a 5-year study of natural genital Chlamydial infections in women strongly suggested that *C. trachomatis* infection was cleared from the cervix epithelium in only 54% of the women at 12 months [185]. This low clearance rate is due, in part, to the organism's ability to utilise various mechanisms with which to evade the host immune system. In fact rarely do host immune responses to *Chlamydia* induce a total clearance of genital infections. Indeed, the women remaining infected in the genital tract with *C. trachomatis* are unable to mount strong adaptive immune responses to clear the pathogen hence they become asymptomatic and remain undiagnosed with a persistent Chlamydial infection. Chronic inflammation resultant from a persistent *C. tracho‐ matis* infection can lead to a wide spectrum of disease manifestations. In persistently-infected *Chlamydia*l host cells, inflammatory cytokines are released in response to the latent infection to aid in the eradication of the pathogen. Unfortunately, these cytokines, which act as immu‐ nomodulators may in fact, induce and sustain tissue damage and host inflammatory responses. Compounded with repeated infections, the continued production of cytokines may cause serious disease sequelae such as PID, TFI, chronic pelvic pain or EP. Chlamydial Hsp60 (cHsp60), a known potent inducer of host inflammatory responses, may protect against Chlamydial infections [186]; paradoxically, however, cHsp60 is also able to induce proinflammatory responses, which can lead to tissue damage in women infected with *C. tracho‐ matis* [187-189].
rapid and efficient bacterial destruction, resultant inflammatory responses may lead to tissue
In reaction to Chlamydial entry in host cells, innate immune responses occur at the infected mucosal site 1-2 days post infection [180]. Intense inflammation and predominant mucosal infiltration of phagocytes, neutrophils and macrophages ingest and enzymatically destroy the bacteria. In conjunction with the production and subsequent activation of phagocytic cells, others, crucial to infection resolution are also activated [181]. Natural killer (NK) cells are activated in direct response to *C. trachomatis* infection or indirectly by macrophagic stimulation of IL-12, a potent NK cell-triggering cytokine. Once activated, NK cells produce IFN-g which stimulate macrophages, further promoting phagocytosis. Innate immunity may impede infection, however, the inaccessibility of *Chlamydia* to circulating antibodies whose sole purpose is elimination necessitates adaptive immune effector mechanisms of both humoral
Activation of cell-mediated immunity post infection is demonstrated by the proliferative response ratio of peripheral blood lymphocytes to Chlamydial EBs. T cells, which functionally differentiate into Th1 or Th2 subtypes based upon cytokine secretion profiles, also amass at the infection site [182]. Further, interaction of *C. trachomatis* with the cytokine network is a key component for resolution of infection and one in which a cytokine response is elicited either through direct infection of epithelial cells or by host cell interaction. In concert with the inflammatory process, tissue repair, comprising the removal of dead cells and the ingrowth of fibroblasts, leads to scar formation and potential impairment of Fallopian tube function [183]. Moreover, all three tubal layers i.e. interstitial, muscular and serosal may be implicated in scar development. Infection of the serosal layer elicits an inflammatory exudate, present on the peritoneal surface, thereby adhering all adjacent surfaces and structures. Although the innate and adaptive immune responses systematically co-ordinate their efforts to ultimately resolve infection, the immune response to Chlamydial infection can cause deciliation of tubal epithe‐ lium, intralumineal adhesions, tubal occlusion and peritubal adhesions. As a result, such gross anatomical and pathological changes may lead to infertility, EP and possible spontaneous
Primary Chlamydial infections of the FGT thus induce both innate and adaptive host immune responses that contribute to clearance of the infections. However, results from a 5-year study of natural genital Chlamydial infections in women strongly suggested that *C. trachomatis* infection was cleared from the cervix epithelium in only 54% of the women at 12 months [185]. This low clearance rate is due, in part, to the organism's ability to utilise various mechanisms with which to evade the host immune system. In fact rarely do host immune responses to *Chlamydia* induce a total clearance of genital infections. Indeed, the women remaining infected in the genital tract with *C. trachomatis* are unable to mount strong adaptive immune responses to clear the pathogen hence they become asymptomatic and remain undiagnosed with a persistent Chlamydial infection. Chronic inflammation resultant from a persistent *C. tracho‐ matis* infection can lead to a wide spectrum of disease manifestations. In persistently-infected *Chlamydia*l host cells, inflammatory cytokines are released in response to the latent infection to aid in the eradication of the pathogen. Unfortunately, these cytokines, which act as immu‐
damage and/or persistent infection.
188 Immune Response Activation
and cell-mediated immunity.
abortion [184].
Recent data from a prospective study of 30 women diagnosed with a ruptured EP demon‐ strated that anti-cHsp60 immunity confers a higher risk of EP [190]. Furthermore, in those women who were positive for anti-cHsp60 antibodies and suffered an EP, all had significantly lower IL-1β levels than Chlamydial sero-negative EP cases [190]. Importantly, high serum concentrations of IL-1β in the Fallopian tubes correlated with successful uterine implantation of embryos [191-192]. Data obtained from a prospective cohort of women at high risk for *C. trachomatis* infection demonstrated that a Th1 response i.e. the release of IFN-g to cHSP60 was associated with protection against the bacterial infection [186]. In contrast, it was also shown that an IL-10 response to the same Chlamydial antigen increased infection risk. In a study that investigated cytokine responses of cervical mononuclear cells obtained from 153 women all serologically positive for *C. trachomatis*, it was found that exposure to cHsp60 and cHsp10 may drastically affect the host's mucosal immune function by increasing the production and subsequent release of IFN-g, IL-10 and tumour necrosis factor-α (TNF-α) [193] thereby potentiating further tissue damage and disease sequelae.
TLRs are responsible for initiation of the adaptive immune system, thereby leading to the ultimate destruction of foreign microbes. Moreover, since they recognise structurally con‐ served molecules from various bacteria, they play a vital role in host defence. Chlamydialinduced PID was investigated and it was reported that among women clinically diagnosed with PID, those who carried variants in TLR 1 and TLR 4 genes had significantly increased odds of *C. trachomatis* infection [191]. In a more recent study, it was concluded that racial variation in TLR variants amongst women with PID may alter signalling pathways subsequent to microbial recognition, thus causing an increase in associated inflammation [194]. An investigation of pregnant women positive for *C. trachomatis* showed that those with serological evidence of prenatal *C. trachomatis* infections were more likely to develop pre-eclampsia [192].
Polymorphic membrane proteins (Pmps) represent 13.6% of *C. trachomatis'* entire coding genome [195], therefore inferring a potential role in Chlamydial biology and virulence. In a group of 40 *C. trachomatis-*infected women with clinically diagnosed mild to moderate PID, antibody levels of PmpA, PmpD and PmpI were evaluated. Interestingly, those women with PmpA antibodies were less likely to fall pregnant (p=0.042) and achieve a live birth (p=0.005) compared with those sero-negative for PmpA [196] suggesting that the immunopathological damage may indeed be a result of PmpA virulence. Further, it was also reported that women who were sero-positive for PmpI antibodies had a higher incidence of upper genital tract infection (p=0.026). However, antibodies to PmpD and PmpI did not alter the risk of FGT inflammation or sequelae.
One of the most important and common long-term complications of Chlamydial PID is TFI. Numerous studies have demonstrated serological evidence of prior *C. trachomatis* infection is associated with TFI [197-201]. An investigation conducted by the World Health Organisation (1995) of TFI rates as a consequence of *C. trachomatis* or *N. gonorrhoeae*-induced PID showed that *C. trachomatis*-induced PID was associated with a higher incidence of TFI when compared to PID caused by *N. gonorrhoeae*. A retrospective study of 1844 women all laparascopically diagnosed with PID due to *C. trachomatis* showed that 209 (16%) failed to conceive [202]. Of these, confirmed TFI was established in 141 patients with PID. Importantly, the rate of infertility was directly associated with the number and severity of PID infections. Specifically, every subsequent episode of PID approximately doubled the rate of TFI i.e. 8% upon one *C. trachomatis* infection, to 19.5% from two exposures resulting in infection and an increase to 40% resultant from three or more exposures.
sperm) when deposited into the vaginal vault, rapidly ascend up into the uterus within 2 mins suggests as the general flow of fluids within the FGT may facilitate ascending infection [210, 211. For most women natural immunity appears to take a long time to acquire - if at all. As previously noted in a study by Molano and colleagues, it was demonstrated that 50% of women continue to shed the organism one year after infection is first documented, and 5 to 10% continue shedding even 3 years later [13]. There is limited evidence to suggest that the generation of natural immunity occurs in some individuals. A number of studies have reported that in around 20% of cases, spontaneous resolution of infection occurs [212,213]. As mentioned previously, Giesler and colleagues showed that women with spontaneously resolved genital infections were less likely to be re-infected within twelve months [25]. The host's immune factors associated with the spontaneous resolution of infection are yet to be elucidated. Unfortunately, any natural immunity induced following primary infection appears to be shortlived and serovar-specific resulting in many re-infections [214-217]. Animal models of Chlamydial genital tract infection have shown that the generation of Th1 cell mediated immunity is characterized by a strong IFNγ and TNFα cytokine production and that this strong response is essential for clearance of primary Chlamydial genital infections [218-220]. The observation that B cell deficient mice can resolve a primary infection suggests that antibodies play a non-critical role in protection [100]. However, B cell responses are important in resolving secondary infections. The generation of *Chlamydia* specific antibodies that are able to neutralize bacteria preventing infection have been reported in various cell lines *in vitro* [221-223] and *in vivo* [224]. CD4+ T helper cells induced during primary infection mediate affinity maturation of antibodies, Ig class switching, and B cell memory responses during subsequent infection. Therefore, B cell memory and antibody production in the presence or absence of CD4+
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
191
critical for the prevention of *Chlamydia* reinfection [225-227] In addition, results from murine studies show the infiltration of immune cells, including CD8+ T cells, B cells, neutrophils and dendritic cells is associated with clearance both of intracellular and of extracellular Chlamydial bodies [228, 117, 229,230]. Using cytobrush sampling of women, an increase in neutrophils, dendritic cells and lymphocytes (CD4+ and CD8+) was reported during *Chlamydia* infections
*Chlamydia* infection of endocervical cells results in an increase in pro-inflammatory cytokine and chemokine production such as IL-1, IL-8, IL-12, IL-6[232, 233, 209]. Live *C. trachomatis* is innately recognized by pattern recognition receptors (PRRs) including Toll-like receptors and nucleotide-binding oligomerization domain (NOD)-like receptors. Both TLR2 and TLR4 are highly expressed in female genital tissues and have shown to be activated by live *Chlamydia* [234]. The myeloid differentiation primary response gene (MyD)88 is essential for nuclear factor–κ-β (NFκβ) signalling and transcription of pro-inflammatory cytokines and along with TLR2, MyD88 has been shown to co-localize within the intracellular Chlamydial inclusion during infection [235]. Mouse TLR2 knockout animals produce lower TNFα and macrophage inflammatory protein-2 (MIP-2) and report a significant decrease on oviduct pathology and hydrosaphinx, suggesting a major role for TLR2 in pathogenesis [236]. The direct activation of another PPR NOD-1 by *Chlamydia* results in the production of IL-8, which is produced at high concentrations during infection and recruit neutrophils to the genital tract [237, 238]. Overall, cytokine production during infection mediates the recruitment of innate immune cells that
in these woemn [125,231].
cells is
In a more recent study that evaluated cHsp60-specific antibody and cell-mediated responses as predictors for TFI [201] it was reported that *C. trachomatis*-specific IgG antibodies were more common in TFI patients (43.2%) when compared to the control group (13.5%), thus suggesting *C. trachomatis* may play a significant pathogenic role in the development of TFI. To elucidate the clinical variability of TFI caused by *C. trachomatis*, functional polymorphisms in various cytokines were assessed [203]. The study showed that allelic variation in IL-10 and TNF-α increased the risk of severe tubal damage in women diagnosed with infertility caused by *C. trachomatis*. Similarly, a more recent study demonstrated that variation in the IL-12B gene was associated with increased susceptibility and TFI severity [204]. Interestingly, polymorphic changes in IL-10 and IFN-g appear to be linked with a more intensive lymphocytic proliferation in response to *C. trachomatis* antigens [205]. Antibody production following cHSP60 exposure was found to be significantly higher in those women (n=21) clinically diagnosed with TFI [206]. Furthermore, elevated antibody levels to caseinolytic protease P, a proteolytic subunit of the ATP-dependent Clp protease complex involved in the degradation of aberrant proteins, was also demonstrated in TFI patients. Serum levels of IgG1 and IgG3 antibodies against chlamy‐ dial major outer membrane protein (MOMP) and cHSP60 were shown to be elevated in women (n=70) with TFI compared with the control group (n=92; normal Fallopian tubes and seronegative for *Chlamydia*) (p=<0.001) [207]. A study that evaluated the potential association of anti-Chlamydial IgM antibodies and TFI showed that of the 50 women assessed, 60% were sero-positive [208]. Of these, 52% presented with bilateral tubal blockage, which was more commonly detected in the ampullary portion (36%) of the Fallopian tube.
#### **4. Immune responses, protection and pathology of chlamydial genital tract infections**
*Chlamydia* (serovars D-K) infects epithelial cells lining the FGT with primary infection estab‐ lishing in the endocervix [209]. If not controlled, primary infection can ascend the reproductive tract leading to the establishment of infection within the endometrium and Fallopian tubes. The kinetics of ascension of *Chlamydia* to the URT is unclear. Although *Chlamydia* can be transported via attachment onto sperm, the demonstration that small particles (the size of sperm) when deposited into the vaginal vault, rapidly ascend up into the uterus within 2 mins suggests as the general flow of fluids within the FGT may facilitate ascending infection [210, 211. For most women natural immunity appears to take a long time to acquire - if at all. As previously noted in a study by Molano and colleagues, it was demonstrated that 50% of women continue to shed the organism one year after infection is first documented, and 5 to 10% continue shedding even 3 years later [13]. There is limited evidence to suggest that the generation of natural immunity occurs in some individuals. A number of studies have reported that in around 20% of cases, spontaneous resolution of infection occurs [212,213]. As mentioned previously, Giesler and colleagues showed that women with spontaneously resolved genital infections were less likely to be re-infected within twelve months [25]. The host's immune factors associated with the spontaneous resolution of infection are yet to be elucidated. Unfortunately, any natural immunity induced following primary infection appears to be shortlived and serovar-specific resulting in many re-infections [214-217]. Animal models of Chlamydial genital tract infection have shown that the generation of Th1 cell mediated immunity is characterized by a strong IFNγ and TNFα cytokine production and that this strong response is essential for clearance of primary Chlamydial genital infections [218-220]. The observation that B cell deficient mice can resolve a primary infection suggests that antibodies play a non-critical role in protection [100]. However, B cell responses are important in resolving secondary infections. The generation of *Chlamydia* specific antibodies that are able to neutralize bacteria preventing infection have been reported in various cell lines *in vitro* [221-223] and *in vivo* [224]. CD4+ T helper cells induced during primary infection mediate affinity maturation of antibodies, Ig class switching, and B cell memory responses during subsequent infection. Therefore, B cell memory and antibody production in the presence or absence of CD4+ cells is critical for the prevention of *Chlamydia* reinfection [225-227] In addition, results from murine studies show the infiltration of immune cells, including CD8+ T cells, B cells, neutrophils and dendritic cells is associated with clearance both of intracellular and of extracellular Chlamydial bodies [228, 117, 229,230]. Using cytobrush sampling of women, an increase in neutrophils, dendritic cells and lymphocytes (CD4+ and CD8+) was reported during *Chlamydia* infections in these woemn [125,231].
One of the most important and common long-term complications of Chlamydial PID is TFI. Numerous studies have demonstrated serological evidence of prior *C. trachomatis* infection is associated with TFI [197-201]. An investigation conducted by the World Health Organisation (1995) of TFI rates as a consequence of *C. trachomatis* or *N. gonorrhoeae*-induced PID showed that *C. trachomatis*-induced PID was associated with a higher incidence of TFI when compared to PID caused by *N. gonorrhoeae*. A retrospective study of 1844 women all laparascopically diagnosed with PID due to *C. trachomatis* showed that 209 (16%) failed to conceive [202]. Of these, confirmed TFI was established in 141 patients with PID. Importantly, the rate of infertility was directly associated with the number and severity of PID infections. Specifically, every subsequent episode of PID approximately doubled the rate of TFI i.e. 8% upon one *C. trachomatis* infection, to 19.5% from two exposures resulting in infection and an increase to 40%
In a more recent study that evaluated cHsp60-specific antibody and cell-mediated responses as predictors for TFI [201] it was reported that *C. trachomatis*-specific IgG antibodies were more common in TFI patients (43.2%) when compared to the control group (13.5%), thus suggesting *C. trachomatis* may play a significant pathogenic role in the development of TFI. To elucidate the clinical variability of TFI caused by *C. trachomatis*, functional polymorphisms in various cytokines were assessed [203]. The study showed that allelic variation in IL-10 and TNF-α increased the risk of severe tubal damage in women diagnosed with infertility caused by *C. trachomatis*. Similarly, a more recent study demonstrated that variation in the IL-12B gene was associated with increased susceptibility and TFI severity [204]. Interestingly, polymorphic changes in IL-10 and IFN-g appear to be linked with a more intensive lymphocytic proliferation in response to *C. trachomatis* antigens [205]. Antibody production following cHSP60 exposure was found to be significantly higher in those women (n=21) clinically diagnosed with TFI [206]. Furthermore, elevated antibody levels to caseinolytic protease P, a proteolytic subunit of the ATP-dependent Clp protease complex involved in the degradation of aberrant proteins, was also demonstrated in TFI patients. Serum levels of IgG1 and IgG3 antibodies against chlamy‐ dial major outer membrane protein (MOMP) and cHSP60 were shown to be elevated in women (n=70) with TFI compared with the control group (n=92; normal Fallopian tubes and seronegative for *Chlamydia*) (p=<0.001) [207]. A study that evaluated the potential association of anti-Chlamydial IgM antibodies and TFI showed that of the 50 women assessed, 60% were sero-positive [208]. Of these, 52% presented with bilateral tubal blockage, which was more
commonly detected in the ampullary portion (36%) of the Fallopian tube.
**4. Immune responses, protection and pathology of chlamydial genital tract**
*Chlamydia* (serovars D-K) infects epithelial cells lining the FGT with primary infection estab‐ lishing in the endocervix [209]. If not controlled, primary infection can ascend the reproductive tract leading to the establishment of infection within the endometrium and Fallopian tubes. The kinetics of ascension of *Chlamydia* to the URT is unclear. Although *Chlamydia* can be transported via attachment onto sperm, the demonstration that small particles (the size of
resultant from three or more exposures.
190 Immune Response Activation
**infections**
*Chlamydia* infection of endocervical cells results in an increase in pro-inflammatory cytokine and chemokine production such as IL-1, IL-8, IL-12, IL-6[232, 233, 209]. Live *C. trachomatis* is innately recognized by pattern recognition receptors (PRRs) including Toll-like receptors and nucleotide-binding oligomerization domain (NOD)-like receptors. Both TLR2 and TLR4 are highly expressed in female genital tissues and have shown to be activated by live *Chlamydia* [234]. The myeloid differentiation primary response gene (MyD)88 is essential for nuclear factor–κ-β (NFκβ) signalling and transcription of pro-inflammatory cytokines and along with TLR2, MyD88 has been shown to co-localize within the intracellular Chlamydial inclusion during infection [235]. Mouse TLR2 knockout animals produce lower TNFα and macrophage inflammatory protein-2 (MIP-2) and report a significant decrease on oviduct pathology and hydrosaphinx, suggesting a major role for TLR2 in pathogenesis [236]. The direct activation of another PPR NOD-1 by *Chlamydia* results in the production of IL-8, which is produced at high concentrations during infection and recruit neutrophils to the genital tract [237, 238]. Overall, cytokine production during infection mediates the recruitment of innate immune cells that secrete IFNγ and TNFα. These include natural killer (NK), Dendritic cells and neutrophils [239, 229]. The recruitment of innate immune cells and cytokine production is linked to upper reproductive tract pathology and scaring leading to the development of tubal infertility and ectopic pregnancy. In particular, Hvid and colleagues reported that during a *C.trachomatis* infection the production of IL-1 from Fallopian tubes biopsies led to the destruction ciliated epithelium. Furthermore IL-1 production additionally induced IL-8 production from Fallopian tube epithelial cells thereby potentially meditating the recruitment of neutrophils to this site [232]. The development of a chronic or persistent *Chlamydia* infection would induce continued production of innate immune mediators perpetuating cellular recruitment leading to epithelial damage, scaring and disease. The innate immune cell generation of immunopathogenesis and disease is termed the cellular paradigm [182].
Recent evidence has shown that the Chlamydial plasmid may play a role in generating pathology during infection. As mentioned earlier, almost all *C. trachomatis* isolates contain a 7.5 kB cryptic plasmid although plasmid-free isolates have been shown in genital serotypes L2, D and E [43-45]. During infection of mice with plasmid-cured *C. muridarum* strains a strong Th-1 cell response is induced however immune pathology and tissue damge is not observed [41] Further, plasmid cured *C. muridarum or C. trachomatis* strains do not stimulate the pro‐ duction of cytokines though TLR2-dependent activation *in vivo* and *in vitro* respectively [251]. In contrast, Frazer *et al* showed that in the guinea pig model, a plasmid cured *C. caviae* strain both signals through TLR-2 and induced post infection pathology suggesting that the associ‐ ation of the *Chlamydia*l plasmid with virulence is not universally conserved among Chlamydial species [252]. Importantly, murine studies have shown that previous infection with plasmidfree is protective against the development of immunopathology with subsequent infection with a virulent strain [41]. This has lead to the use of plasmid-cured strains as attenuated
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
193
Overall, the natural history of Chlamydial infections in women varies in which some untreated individuals with untreated infections persist for long periods asymptomatically and progress to cause pathology and disease, while others spontaneously resolve infection. Although the mechanisms are yet to be fully elucidated, there are currently two proposed mechanisms for the development of pathology in the female reproductive tract; the cellular and the immuno‐ logical paridigms that are associated with innate and adaptive responses respectively. Further, these mechanisms are not mutually exclusive and may both play a role in the development of upper reproductive tract pathology and disease during primary and subsequent infections. The recent identification of plasmid-free strains associated with induced protection against immunopathology may provide a mechanism for immune stimulation leading to pathology
**5. Past and present treatments, how to improve local immunity and future**
*Chlamydia* infection of the FGT is largely asymptomatic with approximately 80% of women observing no reported symptoms [254]. Whilst there are no specific genital symptoms associated with Chlamydial cervical infection, 37% of women who develop cervicitis observe mucopurulent discharge and in 19%, hypertrophic ectopy (Distinctive oedema of the columnar epithelium in the female endocervix) [29, 62, 255]. Due to the asymptomatic nature of infection in the reproductive tract many infections remain undetected and therefore, untreated. Chlamydial infections, if symptomatic, can be treated through the administration of antibiot‐ ics. The current recommended antibiotic for treatment are either a single dose (1g) azithro‐ mycin or 7 day treatment with doxycycline (2x100mg per day)[256]. Other drug treatment programs used to treat *Chlamydia* infections in humans include erythromycin base, (4x500mg per day for 7 days), or erythromycin ethylsuccinate (4x800mg per day for seven days).
Chlamydial vaccines in mouse and primate models of infection [41, 253].
and disease.
**vaccine developments**
**5.1. Antibiotic treatment**
Following clearance of a primary *Chlamydia* infection, re-infections are common in women; moreover, re-infections are strongly associated with development of pathology in these women. Results from animal models of Chlamydial infection have shown that during reinfection T cells are present within the URT at a higher magnitude than in primary infection [240]. *Chlamydia* has shown to act as both an immunizing and sensitizing infection. Non-human primate vaccine trials have shown that immunization with whole cell organisms led to hypersensitivity and an increase immunopathogical response leading to greater scaring in an ocular model. This suggests a specific role for adaptive responses in pathology and this is termed the immunological paradigm. Although the specific mechanism or causative antigen in unknown, it is hypothesized that during re-exposure pathogenesis is associated with either delayed-type hypersentitivity (DTH) or molecular mimicry causing autoimmunity [241]. DTH is hypothesized to be associated with persistent Chlamydial infection. During latent infections, low level antigen specific immune stimulation is believed to contribute to chronic inflamma‐ tory cell infiltration [242]. A role for antigenic sensitivity associated with DTH has been observed in animal models. Guinea pigs sensitized with Triton-X-100 soluble EBs and monkeys sensitized through immunization with a whole Chlamydial vaccines both reported the generation greater inflammation and DTH during re-exposure compared to un-sensitized controls [243,244]. In additon to DTH, autoimmunity may play a role in pathogenesis through the mechanism of molecular mimicry. Chlamydial HSP-60 (cHSP60) shares similar homology to self HSP-60 with recent evidence showing that C. trachomatis HSP-60 contains four T-cell epitopes that display identity with human HSP-60 [245, 246].T cell stimulation by self HSP-60 when pre-immunized with cHSP60 is characterized by the production of pro-inflammatory IFN-g. In contrast, pre-immunization and and re-exposure with self HSP60 results in the production of the anti-inflammatory cytokine IL-10 [247] although previous studies have shown that T cells isolated from patients with PID and TFI respond to cHSP-60 stimulation, and the presence of cHSP-60 specific antibodies correlate with the severity of PID and TFI pathology [248,188]. A recent study by Ness *et al.,* showed that increased cHSP60 antibody titres are not associated with the development of PID [249]. Further the presence of strong cHSP60 T cell IFN-g has more recently been shown to predict protection rather than pathology during reinfection [250]. The conflicting association of cHSP60 with protection or pathology is indicative of a balance between specific cell mediated responses leading to protection or pathogy.
Recent evidence has shown that the Chlamydial plasmid may play a role in generating pathology during infection. As mentioned earlier, almost all *C. trachomatis* isolates contain a 7.5 kB cryptic plasmid although plasmid-free isolates have been shown in genital serotypes L2, D and E [43-45]. During infection of mice with plasmid-cured *C. muridarum* strains a strong Th-1 cell response is induced however immune pathology and tissue damge is not observed [41] Further, plasmid cured *C. muridarum or C. trachomatis* strains do not stimulate the pro‐ duction of cytokines though TLR2-dependent activation *in vivo* and *in vitro* respectively [251]. In contrast, Frazer *et al* showed that in the guinea pig model, a plasmid cured *C. caviae* strain both signals through TLR-2 and induced post infection pathology suggesting that the associ‐ ation of the *Chlamydia*l plasmid with virulence is not universally conserved among Chlamydial species [252]. Importantly, murine studies have shown that previous infection with plasmidfree is protective against the development of immunopathology with subsequent infection with a virulent strain [41]. This has lead to the use of plasmid-cured strains as attenuated Chlamydial vaccines in mouse and primate models of infection [41, 253].
Overall, the natural history of Chlamydial infections in women varies in which some untreated individuals with untreated infections persist for long periods asymptomatically and progress to cause pathology and disease, while others spontaneously resolve infection. Although the mechanisms are yet to be fully elucidated, there are currently two proposed mechanisms for the development of pathology in the female reproductive tract; the cellular and the immuno‐ logical paridigms that are associated with innate and adaptive responses respectively. Further, these mechanisms are not mutually exclusive and may both play a role in the development of upper reproductive tract pathology and disease during primary and subsequent infections. The recent identification of plasmid-free strains associated with induced protection against immunopathology may provide a mechanism for immune stimulation leading to pathology and disease.
#### **5. Past and present treatments, how to improve local immunity and future vaccine developments**
#### **5.1. Antibiotic treatment**
secrete IFNγ and TNFα. These include natural killer (NK), Dendritic cells and neutrophils [239, 229]. The recruitment of innate immune cells and cytokine production is linked to upper reproductive tract pathology and scaring leading to the development of tubal infertility and ectopic pregnancy. In particular, Hvid and colleagues reported that during a *C.trachomatis* infection the production of IL-1 from Fallopian tubes biopsies led to the destruction ciliated epithelium. Furthermore IL-1 production additionally induced IL-8 production from Fallopian tube epithelial cells thereby potentially meditating the recruitment of neutrophils to this site [232]. The development of a chronic or persistent *Chlamydia* infection would induce continued production of innate immune mediators perpetuating cellular recruitment leading to epithelial damage, scaring and disease. The innate immune cell generation of immunopathogenesis and
Following clearance of a primary *Chlamydia* infection, re-infections are common in women; moreover, re-infections are strongly associated with development of pathology in these women. Results from animal models of Chlamydial infection have shown that during reinfection T cells are present within the URT at a higher magnitude than in primary infection [240]. *Chlamydia* has shown to act as both an immunizing and sensitizing infection. Non-human primate vaccine trials have shown that immunization with whole cell organisms led to hypersensitivity and an increase immunopathogical response leading to greater scaring in an ocular model. This suggests a specific role for adaptive responses in pathology and this is termed the immunological paradigm. Although the specific mechanism or causative antigen in unknown, it is hypothesized that during re-exposure pathogenesis is associated with either delayed-type hypersentitivity (DTH) or molecular mimicry causing autoimmunity [241]. DTH is hypothesized to be associated with persistent Chlamydial infection. During latent infections, low level antigen specific immune stimulation is believed to contribute to chronic inflamma‐ tory cell infiltration [242]. A role for antigenic sensitivity associated with DTH has been observed in animal models. Guinea pigs sensitized with Triton-X-100 soluble EBs and monkeys sensitized through immunization with a whole Chlamydial vaccines both reported the generation greater inflammation and DTH during re-exposure compared to un-sensitized controls [243,244]. In additon to DTH, autoimmunity may play a role in pathogenesis through the mechanism of molecular mimicry. Chlamydial HSP-60 (cHSP60) shares similar homology to self HSP-60 with recent evidence showing that C. trachomatis HSP-60 contains four T-cell epitopes that display identity with human HSP-60 [245, 246].T cell stimulation by self HSP-60 when pre-immunized with cHSP60 is characterized by the production of pro-inflammatory IFN-g. In contrast, pre-immunization and and re-exposure with self HSP60 results in the production of the anti-inflammatory cytokine IL-10 [247] although previous studies have shown that T cells isolated from patients with PID and TFI respond to cHSP-60 stimulation, and the presence of cHSP-60 specific antibodies correlate with the severity of PID and TFI pathology [248,188]. A recent study by Ness *et al.,* showed that increased cHSP60 antibody titres are not associated with the development of PID [249]. Further the presence of strong cHSP60 T cell IFN-g has more recently been shown to predict protection rather than pathology during reinfection [250]. The conflicting association of cHSP60 with protection or pathology is indicative of a balance between specific cell mediated responses leading to protection or
disease is termed the cellular paradigm [182].
192 Immune Response Activation
pathogy.
*Chlamydia* infection of the FGT is largely asymptomatic with approximately 80% of women observing no reported symptoms [254]. Whilst there are no specific genital symptoms associated with Chlamydial cervical infection, 37% of women who develop cervicitis observe mucopurulent discharge and in 19%, hypertrophic ectopy (Distinctive oedema of the columnar epithelium in the female endocervix) [29, 62, 255]. Due to the asymptomatic nature of infection in the reproductive tract many infections remain undetected and therefore, untreated. Chlamydial infections, if symptomatic, can be treated through the administration of antibiot‐ ics. The current recommended antibiotic for treatment are either a single dose (1g) azithro‐ mycin or 7 day treatment with doxycycline (2x100mg per day)[256]. Other drug treatment programs used to treat *Chlamydia* infections in humans include erythromycin base, (4x500mg per day for 7 days), or erythromycin ethylsuccinate (4x800mg per day for seven days). Erythromycin is effective against *Chlamydia* infection in the FRT however its use is associated side effects that may limit compliance [256]. Likewise, Levofloxacin (1x500mg per day for 7 days) and ofloxacin (2x300mg per day for 7 days) are effective for treating *Chlamydia* infections but are expensive compared to other alternatives [256]. Following antibiotic treatment it is recommended that treated persons abstain from sexual intercourse for 7 days to prevent spreading the infection to partners which may lead to subsequent re-infections [41]. Partner notification and partner treatment are also important targets to halt re-infections.
However, latent bodies can revert to infectious organisms leading to a recurrent infection. *In vivo* persistent infections in women are supported by the observation that some positive samples obtained using *Chlamydia*-specific DNA or antigen specific testing do not report the presence live infectious EBs via cell culture [268-270]. Furthermore, alternating infectious and persistent phases of Chlamydial growth correlate with acute and chronic infections *in vivo* [271]. During Chlamydial infection in women, a large variation in bacterial load has been observed in the genital tract and antibiotic failure due to heterotrophic resistance is associated with high bacterial loads [272, 273]. West *et al* (2005) showed that 91% of individuals with a low load at baseline observed no infection at follow up (2 months), whereas 74% percent of individuals with high bacterial load reported no follow up infection [274]. The development of persistent bodies *in vivo* during high bacterial load may be an important mechanism driving antibiotic failure, as these slow growing aberrant forms of *Chlamydia* may be less susceptible to antibiotic therapy. *In vitro* studies have shown that antibiotic treatment directly induces the presence of non-cultivable *Chlamydia* that can be reactivated with the withdrawal of treatment. *In vivo* the re-emergence of the active infectious form from persistence may take a number of weeks or months in which previous short term (<3 weeks) follow up testing via culture methods would provide a false negative result. Likewise false positive results may arise posttreatment due to persistent DNA detected through NAAT testing [275]. As such, one recom‐ mendation is that follow up screening > 5weeks post treatment is implemented for screening programs to determine treatment failure [259]. Estimating a true rate of treatment failures is complicated as many re-current infections remain unreported due to the asymptomatic nature of infections coupled with a lack of follow up testing in many screening and treatment
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
195
**5.2. Are screening and treatment programs leading to arrested** *Chlamydia* **immunity?**
In many western countries, screening and antibiotic treatment programs for early detection of *Chlamydia* have been implemented over the past ten years. The goal of these programs is to reduce *Chlamydia* transmission leading to the long-term prevention of disease sequelae, including PID and infertility. *Chlamydia* infections are most common among young people accounting for approximately 80% of all infections. The United States Center for Disease Control (US CDC) estimates that in the United States alone, one in 15 sexually active females aged 14-19 years has *Chlamydia*. Due to this high burden, the US CDC recommends annual screening for all sexually-active women under 26 years of age [276]. As such, most programs have been specifically targeted towards young women with a focus to identify and treat asymptomatic infections. In addition, these programs seek to promote partner notification and treatment to prevent subsequent re-infection. Although screening and treatment programs coupled with aggressive education and health programs have been implemented in many countries for over a decade, *Chlamydia* rates have continued to increase. For example, the US CDC reported an increase of 9.7% of *Chlamydia*l infections between 2007 and 2008 [277]. This increasing trend has also be reported in many countries where long term screening and treatment programs have been in place including Sweden, Canada, the United Kingdom and Australia [278-281. Interpreting this increased prevalence or incidence of *Chlamydia* in the population requires caution and some argue that increases are directly associated with
programs.
Although antibiotics are an effective treatment option, they have been shown to be less effective in cases of well-established chronic Chlamydial infections. Early reports determined that the current treatment of azithromycin and doxycycline are more that 95% effective at clearing *Chlamydia* genital tract infection [257, 259]. However, recent strong evidence suggests that antibiotic treatment failure could be greater than the 5% previously reported when accounting for re-infection [260]. A major hurdle in measuring the rates of antibiotic treatment failure is the inability to distinguish between recurrent infections and re-infection from infected untreated or new sexual partners [261]. Discrepancies in the proportion of treatment failures have been directly associated with testing methods used. Most of the published work inves‐ tigating treatment failure involved tissue culture as a means to detect infection, which has now been predominately replaced with nucleic acid amplification testing (NAAT). Tissue culture is less sensitive and has shown to be less effective at identifying small numbers of persistent bacteria leading to an over estimation of successful *Chlamydia* eradication following treatment [257,262]. In a recent study by Batteiger *et al* (2010) investigating post treatment infection in adolescent girls, it was observed that 7.9% of those sampled developed recurrent infections following azithromycin treatment which was not attributed to reinfection [263]. Similarly, Goldern et al (2005) study in women without the risk for re-infection showed a recurring infection rate of 8% following azithromycin at 3-19 week follow-up representing treatment failure.[264]. Current studies argue that previous rates of antibiotic treatment failure of less than 5% were underestimated due to the available information at the time and that a true rate of treatment failure of may be greater than 8%. One major concern associated with increasing antibiotic failure is the development of antibiotic resistance. However, unlike other organisms such as *Staphylococcus*, to date there is no evidence of natural and stable antibiotic homotypic resistance (genetically inherited) *in vivo* in Chlamydial strains collected from human genital tract infections [265].
Contributing to antibiotic treatment failure in *Chlamydia* infections is the phenomenon called heterotrophic antimicrobial resistance. Heterotrophic resistance refers to the replication of heterogeneous population containing both resistant and susceptible bacteria from a subculture of a single resistant organism propagated on antimicrobial-containing medium, that is not genetically inherited [266]. It is hypothesized that at high bacterial loads a small population of *Chlamydia* organisms survive, potentially due to an innate ability to establish a latent/persistent infection [266]. Importantly these isolates at low loads do not show altered antibiotic suscept‐ ibility or increased resistance due to genetic changes [267]. Latent, non-replicating, noninfectious aberrant RBs associated with persistence can survive within host cells for extended periods of time at which infectious organism can not be isolated or able to be cultured. However, latent bodies can revert to infectious organisms leading to a recurrent infection. *In vivo* persistent infections in women are supported by the observation that some positive samples obtained using *Chlamydia*-specific DNA or antigen specific testing do not report the presence live infectious EBs via cell culture [268-270]. Furthermore, alternating infectious and persistent phases of Chlamydial growth correlate with acute and chronic infections *in vivo* [271]. During Chlamydial infection in women, a large variation in bacterial load has been observed in the genital tract and antibiotic failure due to heterotrophic resistance is associated with high bacterial loads [272, 273]. West *et al* (2005) showed that 91% of individuals with a low load at baseline observed no infection at follow up (2 months), whereas 74% percent of individuals with high bacterial load reported no follow up infection [274]. The development of persistent bodies *in vivo* during high bacterial load may be an important mechanism driving antibiotic failure, as these slow growing aberrant forms of *Chlamydia* may be less susceptible to antibiotic therapy. *In vitro* studies have shown that antibiotic treatment directly induces the presence of non-cultivable *Chlamydia* that can be reactivated with the withdrawal of treatment. *In vivo* the re-emergence of the active infectious form from persistence may take a number of weeks or months in which previous short term (<3 weeks) follow up testing via culture methods would provide a false negative result. Likewise false positive results may arise posttreatment due to persistent DNA detected through NAAT testing [275]. As such, one recom‐ mendation is that follow up screening > 5weeks post treatment is implemented for screening programs to determine treatment failure [259]. Estimating a true rate of treatment failures is complicated as many re-current infections remain unreported due to the asymptomatic nature of infections coupled with a lack of follow up testing in many screening and treatment programs.
Erythromycin is effective against *Chlamydia* infection in the FRT however its use is associated side effects that may limit compliance [256]. Likewise, Levofloxacin (1x500mg per day for 7 days) and ofloxacin (2x300mg per day for 7 days) are effective for treating *Chlamydia* infections but are expensive compared to other alternatives [256]. Following antibiotic treatment it is recommended that treated persons abstain from sexual intercourse for 7 days to prevent spreading the infection to partners which may lead to subsequent re-infections [41]. Partner
Although antibiotics are an effective treatment option, they have been shown to be less effective in cases of well-established chronic Chlamydial infections. Early reports determined that the current treatment of azithromycin and doxycycline are more that 95% effective at clearing *Chlamydia* genital tract infection [257, 259]. However, recent strong evidence suggests that antibiotic treatment failure could be greater than the 5% previously reported when accounting for re-infection [260]. A major hurdle in measuring the rates of antibiotic treatment failure is the inability to distinguish between recurrent infections and re-infection from infected untreated or new sexual partners [261]. Discrepancies in the proportion of treatment failures have been directly associated with testing methods used. Most of the published work inves‐ tigating treatment failure involved tissue culture as a means to detect infection, which has now been predominately replaced with nucleic acid amplification testing (NAAT). Tissue culture is less sensitive and has shown to be less effective at identifying small numbers of persistent bacteria leading to an over estimation of successful *Chlamydia* eradication following treatment [257,262]. In a recent study by Batteiger *et al* (2010) investigating post treatment infection in adolescent girls, it was observed that 7.9% of those sampled developed recurrent infections following azithromycin treatment which was not attributed to reinfection [263]. Similarly, Goldern et al (2005) study in women without the risk for re-infection showed a recurring infection rate of 8% following azithromycin at 3-19 week follow-up representing treatment failure.[264]. Current studies argue that previous rates of antibiotic treatment failure of less than 5% were underestimated due to the available information at the time and that a true rate of treatment failure of may be greater than 8%. One major concern associated with increasing antibiotic failure is the development of antibiotic resistance. However, unlike other organisms such as *Staphylococcus*, to date there is no evidence of natural and stable antibiotic homotypic resistance (genetically inherited) *in vivo* in Chlamydial strains collected from human genital
Contributing to antibiotic treatment failure in *Chlamydia* infections is the phenomenon called heterotrophic antimicrobial resistance. Heterotrophic resistance refers to the replication of heterogeneous population containing both resistant and susceptible bacteria from a subculture of a single resistant organism propagated on antimicrobial-containing medium, that is not genetically inherited [266]. It is hypothesized that at high bacterial loads a small population of *Chlamydia* organisms survive, potentially due to an innate ability to establish a latent/persistent infection [266]. Importantly these isolates at low loads do not show altered antibiotic suscept‐ ibility or increased resistance due to genetic changes [267]. Latent, non-replicating, noninfectious aberrant RBs associated with persistence can survive within host cells for extended periods of time at which infectious organism can not be isolated or able to be cultured.
notification and partner treatment are also important targets to halt re-infections.
tract infections [265].
194 Immune Response Activation
#### **5.2. Are screening and treatment programs leading to arrested** *Chlamydia* **immunity?**
In many western countries, screening and antibiotic treatment programs for early detection of *Chlamydia* have been implemented over the past ten years. The goal of these programs is to reduce *Chlamydia* transmission leading to the long-term prevention of disease sequelae, including PID and infertility. *Chlamydia* infections are most common among young people accounting for approximately 80% of all infections. The United States Center for Disease Control (US CDC) estimates that in the United States alone, one in 15 sexually active females aged 14-19 years has *Chlamydia*. Due to this high burden, the US CDC recommends annual screening for all sexually-active women under 26 years of age [276]. As such, most programs have been specifically targeted towards young women with a focus to identify and treat asymptomatic infections. In addition, these programs seek to promote partner notification and treatment to prevent subsequent re-infection. Although screening and treatment programs coupled with aggressive education and health programs have been implemented in many countries for over a decade, *Chlamydia* rates have continued to increase. For example, the US CDC reported an increase of 9.7% of *Chlamydia*l infections between 2007 and 2008 [277]. This increasing trend has also be reported in many countries where long term screening and treatment programs have been in place including Sweden, Canada, the United Kingdom and Australia [278-281. Interpreting this increased prevalence or incidence of *Chlamydia* in the population requires caution and some argue that increases are directly associated with increased testing rates and detection associate with easy home sampling and the development of the more sensitive nuclear acid amplification test in the 1990's.
[286]. This work suggested that specific antigens within the *Chlamydia* proteome are associated with protection while others mediate immunopathology that may be shared between Chla‐ mydial strains. The recent development of genomic and proteomic technology has provided an unbiased approach to antigen discovery for the generation of a sub-unit *Chlamydia* vaccine [287]. Antigen discovery approaches have included; (1) 2D gel elctrophoresis combined with immunoblotting or radio-immunprecipitation for the discovery of antibody specific antigens [288-290]; (2) Genome-wide protein expression to discover both antibody and T cell antigens [291]; (3) Antigen discovery using of T cell clone lines [292] and; (4) Immunoproteomic approach to identify peptides presented to CD4+ T cells on class II MHC molecules [293].
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
197
Sub-unit vaccines provide safer alternatives to whole cell preparations and a number of *Chlamydia* antigen candidates have been investigated [294]. A number of antigens have been used unsuccessfully in animal trials that include cHsp60 and Chlamydial lipopolysaccharide (cLPS). Chlamydial Hsp60 is an important pathogenicity factor for the development of human Chlamydial infection-associated disease sequelae [295]. Although *Chlamydia* specific immun‐ ity is generated, cHsp60 as an antigen was associated with the induction of inflammation limiting the potential as a vaccine antigen. Chlamydial LPS was a prime antigen candidate due to its localization to the outer cell membrane. Unfortunately, immunization with LPS does not induce the production either of post-immunization or of post-infection antibodies and does not confer protection following live bacterial challenge in primate ocular infection models [296]. Immunization with surface exposed membrane proteins have proved more successful in animal models. These include the Chlamydial major outer membrane protein (MOMP) and the polymorphic membrane protein family (Pmps). The 40kDa MOMP constitutes 50-60% of the outer membrane of the bacterium [297-299]. To date, Chlamydial MOMP has been the leading vaccine antigen candidate investigated in animal models, including non-human primate models. Within the many serovars of the *C. trachomatis* species a large amount of primary structural homology is observed, however, antigenic properties of MOMP are related to the serovar specific surface exposed variable domain (VD) regions. For example, serological studies have shown cross reactivity between the Chlamydial L2 and D serovars, but the G, H and I serovars differ significantly [300]. Therefore the use of Chlamydial MOMP from one serovar results in only serovar-specific immunity and this alone is not sufficient for an effective, protective, multi-serovar vaccine. Combining multiple MOMP proteins from specific serovars in a multi sub-unit vaccine may potentially protect against the most common human genital tract infections. Recently with the aid of immunproteomics Brunham and colleagues identified the expression of several Pmps which when used in mouse *C. muridarum* infection model led to the production of neutralizing antibodies *in vitro* and partial protective immunity *in vivo* [301, 302]. Pmp antigens have also been reported to be immunogenic in human infections [303]. In contrast to EB membrane-bound antigens, Chlamydial secreted proteins have been used as antigen targets. During intracellular *Chlamydia* replication, proteins are secreted from the inclusion body into the host cytosol and may therefore be available for processing and packaging on MHC molecules, thus specifically inducing T cell immunity. The secreted proteins known as Inclusion proteins (Inc) and Chlamydial proteasome/protease-like activity factor (CPAF) have been shown to be potential vaccine targets in mouse models [304, 305]. During Chlamydial infection of humans, both CPAF and Inc antigen specific responses are
However, it has been hypothesized that an increase in *Chlamydia* prevalence worldwide at a time of aggressive screening and testing programs is associated with 'arrested immune hypothesis' [24]. The 'arrested immune hypothesis' states that early detection and treatment of *Chlamydia* actually impairs the development of natural immunity in these antibiotic treated individuals. A recent study by Giesler *et al* (2013) supports the concept of antibiotic treatment may attenuate protective immunity against reinfection. This work showed that women with spontaneous resolution of *Chlamydia* infection were four times less likely to be re-infected compared to women with infections cleared by azithromycin treatment [25]. The potential hampering of protective immunity development through antibiotic treatment is believed to be the associated an increased *Chlamydia* re-infections. In the province of British Columbia, Canada, *Chlamydia* screening and treatment programs have been in place since 1991. Here they have reported a 5% increase in the re-infection rate over the course of the program [282]. Moreover, re-infection rates were reported to be greatest in young women and potential increased risk of development Chlamydial disease sequelae. The emergence of antibiotic resistant organisms, development of persistence, poor patient compliance and high reinfection rates has highlighted the need for an efficacious vaccine to halt the spread of infection throughout the population.
#### **5.3. Chlamydia vaccine development**
The development of an efficacious vaccine is the greatest potential to prevent infection and the subsequent development of disease. According to the World Health Organization a vaccine for the common STI "would have a significant impact on the spread of the disease". An ideal prophylactic Chlamydial vaccine would target immune responses to each of the two stages of the Chlamydial life cycle; (1) Antibodies directed against infectious extracellular elementary bodies preventing cell-adhesion; and (2) T cell mediated immune responses targeting infection cells during early to mid Chlamydial replication. Both responses require the activation of CD4- T cells, which have been shown through mouse models to be essential for clearing *Chlamydia* in the genital tract [283]. A number of considerations are required when designing vaccines; choice of antigen and adjuvant, administration route, the potential target of the vaccine (who to vaccinate?) and the purpose of the vaccine, prophylactic or therapeutic.
Early attempts in the 1950's and 1960's to develop a Chlamydial vaccine were based on the Pasteurian principle of isolate, inactivate and inject [284].However, although this has been a useful principle in the development of vaccine for veterinary field leading to a commercial vaccine for feline Chlamydial infections [285] a human whole-cell vaccine has been unsuc‐ cessful. Although crude whole cell vaccine trials in against ocular Trachoma in humans had observed short-term protection in 70% of vaccinated individuals, the development of strong pro-inflammatory immune responses to bacterial components led to adverse side effects in vaccinated individuals. Furthermore, vaccine trials using primates reported the generation of hypersensitivity leading to increased inflammation and extensive ocular scarring in vaccine recipients therefore whole *Chlamdyia* vaccine trials in humans subsequently were abandoned [286]. This work suggested that specific antigens within the *Chlamydia* proteome are associated with protection while others mediate immunopathology that may be shared between Chla‐ mydial strains. The recent development of genomic and proteomic technology has provided an unbiased approach to antigen discovery for the generation of a sub-unit *Chlamydia* vaccine [287]. Antigen discovery approaches have included; (1) 2D gel elctrophoresis combined with immunoblotting or radio-immunprecipitation for the discovery of antibody specific antigens [288-290]; (2) Genome-wide protein expression to discover both antibody and T cell antigens [291]; (3) Antigen discovery using of T cell clone lines [292] and; (4) Immunoproteomic approach to identify peptides presented to CD4+ T cells on class II MHC molecules [293].
increased testing rates and detection associate with easy home sampling and the development
However, it has been hypothesized that an increase in *Chlamydia* prevalence worldwide at a time of aggressive screening and testing programs is associated with 'arrested immune hypothesis' [24]. The 'arrested immune hypothesis' states that early detection and treatment of *Chlamydia* actually impairs the development of natural immunity in these antibiotic treated individuals. A recent study by Giesler *et al* (2013) supports the concept of antibiotic treatment may attenuate protective immunity against reinfection. This work showed that women with spontaneous resolution of *Chlamydia* infection were four times less likely to be re-infected compared to women with infections cleared by azithromycin treatment [25]. The potential hampering of protective immunity development through antibiotic treatment is believed to be the associated an increased *Chlamydia* re-infections. In the province of British Columbia, Canada, *Chlamydia* screening and treatment programs have been in place since 1991. Here they have reported a 5% increase in the re-infection rate over the course of the program [282]. Moreover, re-infection rates were reported to be greatest in young women and potential increased risk of development Chlamydial disease sequelae. The emergence of antibiotic resistant organisms, development of persistence, poor patient compliance and high reinfection rates has highlighted the need for an efficacious vaccine to halt the spread of infection
The development of an efficacious vaccine is the greatest potential to prevent infection and the subsequent development of disease. According to the World Health Organization a vaccine for the common STI "would have a significant impact on the spread of the disease". An ideal prophylactic Chlamydial vaccine would target immune responses to each of the two stages of the Chlamydial life cycle; (1) Antibodies directed against infectious extracellular elementary bodies preventing cell-adhesion; and (2) T cell mediated immune responses targeting infection cells during early to mid Chlamydial replication. Both responses require the activation of CD4- T cells, which have been shown through mouse models to be essential for clearing *Chlamydia* in the genital tract [283]. A number of considerations are required when designing vaccines; choice of antigen and adjuvant, administration route, the potential target of the vaccine (who
Early attempts in the 1950's and 1960's to develop a Chlamydial vaccine were based on the Pasteurian principle of isolate, inactivate and inject [284].However, although this has been a useful principle in the development of vaccine for veterinary field leading to a commercial vaccine for feline Chlamydial infections [285] a human whole-cell vaccine has been unsuc‐ cessful. Although crude whole cell vaccine trials in against ocular Trachoma in humans had observed short-term protection in 70% of vaccinated individuals, the development of strong pro-inflammatory immune responses to bacterial components led to adverse side effects in vaccinated individuals. Furthermore, vaccine trials using primates reported the generation of hypersensitivity leading to increased inflammation and extensive ocular scarring in vaccine recipients therefore whole *Chlamdyia* vaccine trials in humans subsequently were abandoned
to vaccinate?) and the purpose of the vaccine, prophylactic or therapeutic.
of the more sensitive nuclear acid amplification test in the 1990's.
throughout the population.
196 Immune Response Activation
**5.3. Chlamydia vaccine development**
Sub-unit vaccines provide safer alternatives to whole cell preparations and a number of *Chlamydia* antigen candidates have been investigated [294]. A number of antigens have been used unsuccessfully in animal trials that include cHsp60 and Chlamydial lipopolysaccharide (cLPS). Chlamydial Hsp60 is an important pathogenicity factor for the development of human Chlamydial infection-associated disease sequelae [295]. Although *Chlamydia* specific immun‐ ity is generated, cHsp60 as an antigen was associated with the induction of inflammation limiting the potential as a vaccine antigen. Chlamydial LPS was a prime antigen candidate due to its localization to the outer cell membrane. Unfortunately, immunization with LPS does not induce the production either of post-immunization or of post-infection antibodies and does not confer protection following live bacterial challenge in primate ocular infection models [296]. Immunization with surface exposed membrane proteins have proved more successful in animal models. These include the Chlamydial major outer membrane protein (MOMP) and the polymorphic membrane protein family (Pmps). The 40kDa MOMP constitutes 50-60% of the outer membrane of the bacterium [297-299]. To date, Chlamydial MOMP has been the leading vaccine antigen candidate investigated in animal models, including non-human primate models. Within the many serovars of the *C. trachomatis* species a large amount of primary structural homology is observed, however, antigenic properties of MOMP are related to the serovar specific surface exposed variable domain (VD) regions. For example, serological studies have shown cross reactivity between the Chlamydial L2 and D serovars, but the G, H and I serovars differ significantly [300]. Therefore the use of Chlamydial MOMP from one serovar results in only serovar-specific immunity and this alone is not sufficient for an effective, protective, multi-serovar vaccine. Combining multiple MOMP proteins from specific serovars in a multi sub-unit vaccine may potentially protect against the most common human genital tract infections. Recently with the aid of immunproteomics Brunham and colleagues identified the expression of several Pmps which when used in mouse *C. muridarum* infection model led to the production of neutralizing antibodies *in vitro* and partial protective immunity *in vivo* [301, 302]. Pmp antigens have also been reported to be immunogenic in human infections [303]. In contrast to EB membrane-bound antigens, Chlamydial secreted proteins have been used as antigen targets. During intracellular *Chlamydia* replication, proteins are secreted from the inclusion body into the host cytosol and may therefore be available for processing and packaging on MHC molecules, thus specifically inducing T cell immunity. The secreted proteins known as Inclusion proteins (Inc) and Chlamydial proteasome/protease-like activity factor (CPAF) have been shown to be potential vaccine targets in mouse models [304, 305]. During Chlamydial infection of humans, both CPAF and Inc antigen specific responses are observed supporting their potential as antigens in a human vaccine [305, 306] [305]. A number of Chlamydial antigens trialled as vaccine candidates have recently been reviewed by Hafner et *al.* [27].
could protect individuals by raising the infectiousness threshold and secondary reduce the peak load and the duration of the infection in vaccinated individuals. Animal models have shown that some experimental vaccines, although not producing sterilizing immunity in these models, may reduce the development of pathology e.g. hydrosalphinx in the murine model. This suggests that a therapeutic vaccine could potentially be given to women a past or even current infection to prevent the development of pathology. Although little work has been undertaken in this area, Carey et al. reported that in the mouse *C. muridarum* model vaccination during or after infection reduction in the strength of the immune response was observed. This work demonstrates that a therapeutic vaccine could be used to limit an uncontrolled host response leading to immunopathology during re-infection [320]. In light of the fact that there currently is no vaccine to prevent Chlamydial genital tract infections in women we believe that a vaccine designed to reduce bacterial burden and prevent disease may be a more rational
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
199
In this Chapter we have presented current knowledge of immune regulation during infections of the female genital tract caused by the intracellular bacterial pathogen, *Chlamydia trachoma‐ tis*. Our review has highlighted the following information regarding Chlamydial infections of
**•** Tubal infertility and reproductive damage, preeclampsia and preterm births may poten‐ tially occur in many of the 50 million women who are annually infected with *Chlamydia*
**•** Gynecological cancer development is also suggested to be linked with Chlamydial genital
**•** The annual direct medical cost of sexually transmitted diseases (including *Chlamydia* and
**•** Urogenital isolates of *C.trachomatis* may also contain a highly conserved 7.5kB non-integra‐
**•** *Chlamydia* infections of neonates have been reported to be at 18 percent for ophthalmia neonatorum and 16 percent for pneumonia for infants exposed to *C. trachomatis* in vaginal
**•** The host inflammatory response to *C.trachomatis* is involved in immunity but also the pathology that leads to serious morbidities of chronic pelvic pain, EP and tubal infertility
**•** Disease outcomes are dependent upon the complex interactions between virulence factors of, and evasion strategies used by, *C.trachomatis* and host immune responses including
**•** The majority (70-80%) of genital tract infections in women are asymptomatic
those caused by HIV) has been estimated at \$US16.9 billion
tive plasmid that has a role in the pathobiology of these organisms
variants in genetic markers associated with infection and inflammation
target for future research.
the FGT:
*trachomatis*
tract infections.
secretions in the birth canal
following female genital tract infection
**6. Summary/conclusions**
Although sub-unit vaccines provide safer alternatives to the use of live or attenuated organ‐ isms, they are poorly immunogenic. Adjuvants are required to direct the immune response to a co-administrated antigen. For *Chlamydia*, a balance of Th2 driven neutralizing antibodies and Th1 cytokine production characterized by strong IFN-g within the FGT mucosae is required. A number of adjuvants have been developed for use experimentally in trials for Chlamydial vaccines. These include live viral vectors [307], immunostimulating complexes such as DDA and ISCOMs [308, 309, liposomes [310], detergent/surfactant-based adjuvants [311, Lipid formulations [312,313, *Vibrio cholerae* ghosts [314, unmethylated 1.7.4.1CPG-ODN bacterial motifs [308,222] and cytokines [315]. In addition to antigen and adjuvant combinations, the route of immunization is an important component of vaccination strategies. Historically most vaccines are delivered through injection that is either sub-cutaneous (S.C) or intra-muscular (I.M). Although effective in generating strong systemic immunity, the generation of mucosal immunity has shown to be varied and limited when using systemic immunization routes. However, these methods have shown in animal models to generate partial protection against *Chlamydia* infections. A recent study by Eko et al. (2011) using I.M. vaccination of *Chlamydia* antigens expressed by Vibrio Cholerae Ghosts (VVG) has shown to be effective in reducing the duration of infection in mice; however, sterilizing immunity was not observed in this trial [316]. Another approach is targeting the mucosal immune system directly with needle free mucosal routes. Not only are mucosal immunization regimes able to generate immunity and multiple mucosal surfaces, the needle free nature of delivery has the advantage of being safe and inexpensive to deliver. Immunization with candidate Chlamydial vaccines via mucosal routes such as oral [312], intranasal [223,317], intravaginal [318 and transcutaneous [222,223,313] have shown to target protective immunity in the female reproductive tract; again, however sterilizing immunity has not yet been achieved using these vaccine approaches.
Over the past 60 years since the first crude whole cell Chlamydial vaccines were investigated, a number of technologies such as gemomic sequencing have enhanced the forward movement of the development of a protective Chlamydial vaccine. In light of the continued increased infection and re-infection rates, even with aggressive screening and treatment programs, the need for an efficacious Chlamydial vaccine is essential. Mathematical modeling that simulates transmission of *Chlamydia* in a heterosexual population has been developed to determine the impact of an efficacious vaccine regime [319]. The model tracks the infection time course, disease progression, and dynamic infectiousness of infected individuals and the transmission to others. This model determined that a fully protective vaccine administered prior to sexual debut could eradicate Chlamydial epidemics within 20 years. Furthermore, the specific targeting of women (100% vaccine coverage) would have an increased impact in reducing epidemiology than vaccination 50% of the male and female population together. However, currently no vaccine has been shown to generate sterilizing immunity. Modelling reports that in the absence of a sterilizing vaccine, a *Chlamydia* vaccine effective for at least 10 years would provide order to lead to population-level eradication. Furthermore, a non-sterilizing vaccine could protect individuals by raising the infectiousness threshold and secondary reduce the peak load and the duration of the infection in vaccinated individuals. Animal models have shown that some experimental vaccines, although not producing sterilizing immunity in these models, may reduce the development of pathology e.g. hydrosalphinx in the murine model. This suggests that a therapeutic vaccine could potentially be given to women a past or even current infection to prevent the development of pathology. Although little work has been undertaken in this area, Carey et al. reported that in the mouse *C. muridarum* model vaccination during or after infection reduction in the strength of the immune response was observed. This work demonstrates that a therapeutic vaccine could be used to limit an uncontrolled host response leading to immunopathology during re-infection [320]. In light of the fact that there currently is no vaccine to prevent Chlamydial genital tract infections in women we believe that a vaccine designed to reduce bacterial burden and prevent disease may be a more rational target for future research.
#### **6. Summary/conclusions**
observed supporting their potential as antigens in a human vaccine [305, 306] [305]. A number of Chlamydial antigens trialled as vaccine candidates have recently been reviewed by Hafner
Although sub-unit vaccines provide safer alternatives to the use of live or attenuated organ‐ isms, they are poorly immunogenic. Adjuvants are required to direct the immune response to a co-administrated antigen. For *Chlamydia*, a balance of Th2 driven neutralizing antibodies and Th1 cytokine production characterized by strong IFN-g within the FGT mucosae is required. A number of adjuvants have been developed for use experimentally in trials for Chlamydial vaccines. These include live viral vectors [307], immunostimulating complexes such as DDA and ISCOMs [308, 309, liposomes [310], detergent/surfactant-based adjuvants [311, Lipid formulations [312,313, *Vibrio cholerae* ghosts [314, unmethylated 1.7.4.1CPG-ODN bacterial motifs [308,222] and cytokines [315]. In addition to antigen and adjuvant combinations, the route of immunization is an important component of vaccination strategies. Historically most vaccines are delivered through injection that is either sub-cutaneous (S.C) or intra-muscular (I.M). Although effective in generating strong systemic immunity, the generation of mucosal immunity has shown to be varied and limited when using systemic immunization routes. However, these methods have shown in animal models to generate partial protection against *Chlamydia* infections. A recent study by Eko et al. (2011) using I.M. vaccination of *Chlamydia* antigens expressed by Vibrio Cholerae Ghosts (VVG) has shown to be effective in reducing the duration of infection in mice; however, sterilizing immunity was not observed in this trial [316]. Another approach is targeting the mucosal immune system directly with needle free mucosal routes. Not only are mucosal immunization regimes able to generate immunity and multiple mucosal surfaces, the needle free nature of delivery has the advantage of being safe and inexpensive to deliver. Immunization with candidate Chlamydial vaccines via mucosal routes such as oral [312], intranasal [223,317], intravaginal [318 and transcutaneous [222,223,313] have shown to target protective immunity in the female reproductive tract; again, however sterilizing immunity has not yet been achieved using these vaccine approaches.
Over the past 60 years since the first crude whole cell Chlamydial vaccines were investigated, a number of technologies such as gemomic sequencing have enhanced the forward movement of the development of a protective Chlamydial vaccine. In light of the continued increased infection and re-infection rates, even with aggressive screening and treatment programs, the need for an efficacious Chlamydial vaccine is essential. Mathematical modeling that simulates transmission of *Chlamydia* in a heterosexual population has been developed to determine the impact of an efficacious vaccine regime [319]. The model tracks the infection time course, disease progression, and dynamic infectiousness of infected individuals and the transmission to others. This model determined that a fully protective vaccine administered prior to sexual debut could eradicate Chlamydial epidemics within 20 years. Furthermore, the specific targeting of women (100% vaccine coverage) would have an increased impact in reducing epidemiology than vaccination 50% of the male and female population together. However, currently no vaccine has been shown to generate sterilizing immunity. Modelling reports that in the absence of a sterilizing vaccine, a *Chlamydia* vaccine effective for at least 10 years would provide order to lead to population-level eradication. Furthermore, a non-sterilizing vaccine
et *al.* [27].
198 Immune Response Activation
In this Chapter we have presented current knowledge of immune regulation during infections of the female genital tract caused by the intracellular bacterial pathogen, *Chlamydia trachoma‐ tis*. Our review has highlighted the following information regarding Chlamydial infections of the FGT:
**•** It is becoming increasingly well documented that sex hormones E2 and P4 regulate host susceptibility, as well as innate and adaptive immune responses to the sexually transmitted pathogen *Chlamydia* infecting the mucosal surfaces of the FGT
[7] Madeleine, M., Anttila, T., Schwartz, S. et al. Risk of cervical cancer associated with *Chlamydia trachomatis* antibodies by histology, HPV type and HPV cofactors Int J
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
201
[8] Deluca GD, Basiletti J, Schelover E, Vásquez ND, Alonso JM, Marín HM, Lucero RH, Picconi MA. *Chlamydia trachomatis* as a probable cofactor in human papillomavirus infection in aboriginal women from northeastern Argentina. Braz J Infect Dis. 2011;
[9] Shanmughapriya S, Senthilkumar G, Vinodhini K, Das BC, Vasanthi N, Natarajasee‐ nivasan K. Viral and bacterial aetiologies of epithelial ovarian cancer. Eur J Clin Mi‐
[10] Idahl A, Lundin E, Jurstrand M, Kumlin U, Elgh F, Ohlson N, Ottander U. *Chlamydia trachomatis* and *Mycoplasma genitalium* plasma antibodies in relation to epithelial ovarian tumors. Infect Dis Obstet Gynecol. 2011; Volume 2011, Article ID 824627, 10
[11] Chumduri C, Gurumurthy RK, Zadora PK, Mi Y, Meyer TF. *Chlamydia* infection pro‐ motes host DNA damage and proliferation but impairs the DNA damage response.
[12] Padberg I, Janßen S, Meyer TF. *Chlamydia trachomatis* inhibits telomeric DNA damage signaling via transient hTERT upregulation. I Int J Med Microbiol. 2013 pii:
[13] Molano M, Meijer CJ, Weiderpass E, Arslan A, Posso H, Franceschi S, et al. The natu‐ ral course of *Chlamydia trachomatis* infection in asymptomatic Colombian women: a 5-
[14] Stamm WE. *Chlamydia trachomatis* infections: progress and problems. J Infect Dis.
[15] Stamm WE. Lymphogranuloma venereum. In: Holmes KK, Sparling FP, Mardh PA, (editors). Sexually Transmitted Diseases. 4th ed: McGraw Hill Professional; 2008. p.
[16] Hafner L. M. and Pelzer, E.S.Tubal Damage, Infertility and Tubal Ectopic Pregnancy: *Chlamydia trachomatis* and Other Microbial Aetiologies, Ectopic Pregnancy - Modern Diagnosis and Management, Dr. Michael Kamrava (Ed.), ISBN: 978-953-307-648-5, Ri‐ jeka:InTech, 2011 DOI: 10.5772/21555. Available from: http://www.intechopen.com/ books/ectopic-pregnancy-modern-diagnosis-and-management/tubal-damage-infertil‐ ity-and-tubal-ectopic-pregnancy-*Chlamydia*-*trachomatis*-and-other-microbial-aetio
[17] Haggerty CL, Gottlieb SL, Taylor BD, Low N, Xu F, Ness RB. Risk of sequelae after *Chlamydia trachomatis* genital infection in women. J Infect Dis. 2010 ;201 Suppl
Cancer 2007; 120(3): 650-655.
crobiol Infect Dis. 2012;31(9):2311-7.
pages. doi:10.1155/2011/824627.
1999;179 Suppl 2:S380-3.
(acessed 14 September, 2013)
505-605.
2:S134-55.
Cell Host Microbe. 2013;13(6):746-758.
S1438-4221(13)00080-5. doi: 10.1016/j.ijmm.2013.06.001.
year follow-up study. J Infect Dis 2005;191:907-16.
15(6):567-72.
#### **Author details**
Louise M. Hafner, Trudi A. Collet and Danica K. Hickey
Institute of Health and Biomedical Innovation, (IHBI), Queensland University of Technology (QUT), Australia
#### **References**
[7] Madeleine, M., Anttila, T., Schwartz, S. et al. Risk of cervical cancer associated with *Chlamydia trachomatis* antibodies by histology, HPV type and HPV cofactors Int J Cancer 2007; 120(3): 650-655.
**•** It is becoming increasingly well documented that sex hormones E2 and P4 regulate host susceptibility, as well as innate and adaptive immune responses to the sexually transmitted
**•** In light of the continued increase in infection and re-infection rates even with aggressive
**•** Finally, since there currently is no vaccine to prevent Chlamydial genital tract infections we consider that a vaccine designed to reduce bacterial burden and prevent disease pathology
Institute of Health and Biomedical Innovation, (IHBI), Queensland University of Technology
[1] Stamm WE. *Chlamydia trachomatis* infections: progress and problems.J Infect Dis.
[2] Starnbach MN, Roan NR. Conquering sexually transmitted diseases. Nature Reviews
[3] Haggerty, C.L., Panum, I., Uldum, S.A., Bass, D.C., Olsen, J., Darville, T., Eastman, J.M., Simhan, H.N., Roberts, J.M., Ness, R.B. *Chlamydia trachomatis* infection may in‐ crease the risk of preeclampsia. Pregnancy Hypertension: An International Journal of
[4] Rours GI, Duijts L, Moll HA, Arends LR, de Groot R, Jaddoe VW, Hofman A, Steeg‐ ers EA, Mackenbach JP, Ott A, Willemse HF, van der Zwaan EA, Verkooijen RP, Ver‐ brugh HA. *Chlamydia trachomatis* infection during pregnancy associated with preterm delivery: a population-based prospective cohort study. Eur J Epidemiol. 2011a;26(6):
[5] Ferreira, V.H., Nazli, A., Khan, G., et al. Endometrial epithelial cell responses to coinfecting viral and bacterial pathogens in the genital tract can activate the HIV-1 LTR in an NF{kappa}B-and AP1-dependent manner. J Infect Dis. 2011; 204(2):299-308.
[6] Schust, D.J., Ibana, JA., Buckner, LR., Ficarra, M., Sugimoto, J., Amedee, AM and Quayle AJ Potential mechanisms for increased HIV-1 transmission across the endo‐ cervical epithelium during *C.trachomatis* infection. Curr HIV Res, 2012; 10(3): 218-27.
screening and treatment programs, a need for an efficacious vaccine is essential.
pathogen *Chlamydia* infecting the mucosal surfaces of the FGT
may be a rational target for future Chlamydial research.
Louise M. Hafner, Trudi A. Collet and Danica K. Hickey
**Author details**
200 Immune Response Activation
(QUT), Australia
**References**
493-502.
1999;179 Suppl 2:S380-3.
Immunology 2008; 8(4):313-317.
Women's Cardiovascular Health, 2013; 3:28-33.
[18] RoursIG, Hammerschlag, MR., Ott,A., DeFaber, TJ, VerbrughHA, deGroot R et al.*Chlamydia trachomatis* as a cause of neonatal conjunctivitis in Dutch infants. Pedia‐ trics 2008; 121:e321-326.
[30] Collingro, A., Tischler, P., Weinmaier, T., Penz, T., Heinz, E., Brunham, R.C., Read, TD., Bavoil, PM., Saschse, K., Kahane, S., et al., Unity in variety-the pan-genome of
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
203
[31] Hammerschlag, M.R. The intracellular life of Chlamydiae. Semin Pediatr Infect Dis.
[32] Stephens, R.S., Wagar, E.A. and Schoolnik, G.K. High-resolution mapping of serovarspecific and common antigenic determinants of the major outer membrane protein of
[33] Stephens, R.S., Sanchez-Pescador, R., Wager, E., et al. Diversity of the major outer membrane proteins of *Chlamydia trachomatis*. J. Bacteriol.,1987;169:3879-3885.
[34] Stephens, R.S., Myers, G., Eppinger, M. et al.Divergence without difference: phyloge‐ netics and taxonomy of *Chlamydia* resolved. FEMS immunology and medical micro‐
[35] Bush, R. M. and Everett, K.D.Molecular evolution of Chlamydiaceae. Int J Syst Evol
[36] Schachter, J., Stephens, R.S, Timms, P. et al. Radical changes to Chlamydial taxono‐
[37] Caldwell, H.D., Wood, H., Crane, D., et al. Polymorphisms in *Chlamydia trachomatis* tryptophan synthase genes differentiate between genital and ocular isolates. J Clin
[38] Thomas NS, Lusher M, Storey CC & Clarke IN. Plasmid diversity in *Chlamydia*.Mi‐
[39] Rockey, D.D Unravelling the basic biology and clinical significance of the Chlamy‐
[40] Russell, M., T. Darville, K. Chandra-Kuntal, B. Smith, C.W. Andrews Jr., and C.M. O'Connell. Infectivity acts as in *vivo* selection for maintenance of the Chlamydial
[41] O'Connell, C.M., Ingalls, R.R., Andrews, C.W. Jr, et al. Plasmid deficient *Chlamydia muridarum* fails to induce immune pathology and protect against oviduct disease. J
[42] Ripa, T., Nilsson, P. A variant of *Chlamydia trachomatis* with deletion in cryptic plas‐ mid: implications for use of PCR diagnostic tests. Eur Sureveill 2006; 11: E061109. [43] Farencena, A., M. Comanducci, M. Donati, G. Ratti, and R. Cevenini. Characteriza‐ tion of a new isolate of *Chlamydia trachomatis* which lacks the common plasmid and
has properties of biovar trachoma. Infect. Immun.1997; 65:2965-2969.
my are not necessary just yet. Int J Syst Evol Microbiol 2001; 51:251-253,
the Chlamydiae. Molec. Biol Evol*.* 2011; 28: 3253-3270.
*Chlamydia trachomatis*. J.Exp.Med.1988; 167:817-831.
2002; 13(4):239-48.
biology 2009; 55: 115-119.
Microbiol. 2001; 51:203-220
Invest.2003; 111:1757–69.
crobiol.1997; 143 ( Pt 6): 1847-1854.
Immunol.2007; 179: 4027–4034.
dial plasmid J. Exp. Med.2011; 208(11): 2159-2162.
cryptic plasmid. Infect. Immun.2011; 79:98–107.
[30] Collingro, A., Tischler, P., Weinmaier, T., Penz, T., Heinz, E., Brunham, R.C., Read, TD., Bavoil, PM., Saschse, K., Kahane, S., et al., Unity in variety-the pan-genome of the Chlamydiae. Molec. Biol Evol*.* 2011; 28: 3253-3270.
[18] RoursIG, Hammerschlag, MR., Ott,A., DeFaber, TJ, VerbrughHA, deGroot R et al.*Chlamydia trachomatis* as a cause of neonatal conjunctivitis in Dutch infants. Pedia‐
[19] Rours, GI, Hammerschlag, MR, Van Doornum, GJ, Hop WC, de Groot, R, Willemse HF et al., *Chlamydia* trachomais respiratory infection in Dutch infants. Arch Dis
[20] World Health Organization. Global prevalence and incidence of selected curable sex‐
[21] Chesson, H. W., Gift, T. L., Owusu-Edusei, K. Jr., et al. A Brief Review of the Estimat‐ ed Economic Burden of Sexually Transmitted Diseases in the United States: Inflation-Adjusted Updates of Previously Published Cost Studies Sexually Transmitted
[22] Reckart,M.L., Gilbert, M., Meza, R., Kim, P.H., Chang, M., Money, D.M. and Brun‐ ham, RC. *Chlamydia* public health programs and the epidemiology of pelvic inflam‐
[23] Centers for Disease Control (CDC) Sexually disease treatment guidelines. MMWR
[24] Brunham, RC and Reckart ML. The arrested immunity hypothesis and the epidemi‐
[25] Geisler, W.M., Lensing, S.Y., Press, C.G., and Hook, E.W. Spontaneous resolution of genital *Chlamydia trachomatis* infection in women and protection from re-infection. J.
[26] World Health Organisation (WHO) Initiative for Vaccine Research. SexuallyTrans‐ mitted Diseases. *Chlamydia trachomatis.*WorldHealthOrganisation.Available at http:// www.who.int/vaccine\_research/diseases/soa\_std/en/index1.html (accessed Septem‐
[27] Hafner LM, Wilson DP, Timms P. Development status and future prospects for a vaccine against *Chlamydia trachomatis* infection. Vaccine. 2013b doi:pii:
[28] Elwell, C. A., Jiang, S., Kim, J.H., et al. *Chlamydia trachomatis* Co-opts GBF1 and CERT to Acquire Host Sphingomyelin for Distinct Roles during Intracellular Development
[29] Stamm, W. E. *Chlamydia trachomatis* infections in adults. In: Holmes KK, Sparling PF,
edition. New York, NY
ually transmitted infections. Geneva: World Health Organization,2008.
matory disease and ectopic pregnancy. J.Infect.Dis., 2013; 207: 30-8.
ology of *Chlamydia* control. Sex. Transm. Dis.2008; 35: 53-54.
S0264-410X(13)01111-0. 10.1016/j.vaccine.2013.08.020.
Mardh, P-A et al., (editors). Sexually Transmitted Diseases. 4th
trics 2008; 121:e321-326.
202 Immune Response Activation
Child. 2009; 94: 705-707.
Diseases 2011; 38(10): 889-891.
Infect. Dis.2013; 207: 1850-6.
PLoS Pathog. 2011: 7(9):e1002198.
McGraw-Hill Companies Inc. 2008. 575-594.
ber 1, 2013).
Moribid Mortal Wkly Rep. 2006 51:1-78.
[44] Peterson, E. M., B. A. Markoff, J. Schachter, and L. M. De La Maza. The 7.5-kbplas‐ mid present in *Chlamydia trachomatis* is not essential for the growth of this microor‐ ganism. Plasmid 1990; 23:144-148
patibility complex class I expression in *Chlamydia*-infected cells. J Exp Med.
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
205
[57] Ibana, JA, Schust, DJ, Sugimato, J., Nagamatsu, T., Greene, SJ. and Quayle, AJ *Chla‐ mydia trachomatis* Immune Evasion via Downregulation of MHC Class I Surface Ex‐ pression Involves Direct and Indirect Mechanisms 2011 Inf Dis. Obstet. Gynecol.
[58] Bastidas, R.J., Elwell,C.A., Joanne N. Engel J.N., and Raphael H. Valdivia, R.H. Chla‐ mydial Intracellular Survival Strategies Cold Spring Harb Perspect Med. 2013: doi:
[59] Betts, H.J., Wolf, K., and Fields, K.A. Effector protein modulation of host cells:exam‐ ples in the *Chlamydia* spp. Arsenal. Curr. Opin. Microbiol. 2009; 12: 81-87.
[60] Dunn, J.D. and Valdivia, R.H. (2010).Uncivil engineers: *Chlamydia*, Salmonella and Shigella alter cytoskeleton architecture to invade epithelial cells.Fut. Microbiol.2010;
[61] Soupene, E., Rothschild, J., Kuypers, F.A. and Dean, D. Eukaryotic protein recruit‐ ment into the *Chlamydia* inclusion: Implications for survival and growth. PLoS ONE
[62] .Brunham, R., Paavonen, J., Stevens, C.E. et al. Mucopurulent cervicitis – the ignored counterpart in women of urethritis in men. New Eng. J. Med.1984; 311: 1-6.
[63] Marrazzo, J.M., and Martin DH. Management of women with cervicitis. Clin Infect
[64] Falk, L. The overall agreement of proposed definitions of mucopurulent cervicitis in women at high risk of Chlamydial infection Acta Derm Venereol, 2010; 90: 506-511.
[65] Lusk, M. J. and Konecny, P. Cervicitis: a review.Curr.Opin.Infect.Dis.2008; 21: 49-55.
[66] Falk, L., Fredlund, H., and Jensen, J.S. Signs and symptoms of urethritis andcervicitis among women with or without Mycoplasma genitalium or *Chlamydia trachomatis* in‐
[67] Carvalho JP, Carvalho FM. Is *Chlamydia*-infected tubal fimbria the origin of ovarian
[68] Diniz PM, Carvalho JP, Baracat EC, Carvalho FM. Fallopian tube origin of supposed ovarian high-grade serous carcinomas. Clinics (Sao Paulo). 2011; 66(1):73-6.
[69] Taylor-Robinson D, Stacey CM, Jensen JS, Thomas BJ, Munday PE. Further observa‐ tions, mainly serological, on a cohort of women with or without pelvic inflammatory
[70] World Health Organisation (WHO). Sexually transmitted infections, Fact sheet No.
Volume 2011, Article ID 420905, 8 pages doi:10.1155/2011/420905
2012; 7(5): : e36843. Doi: 10.1371/journal.pone.0036843
2000;191(9):1525-34.
10.1101/cshperspect.a010256.
Dis.2007; 44 (Suppl 3):S102-10.
fection. Sex. Trans. Infect. 2005; 81:73-78.
cancer? Med Hypotheses. 2008; 71(5):690-3.
disease. Int J STD AIDS. 2009 (10):712-8.
110, 2011
5(8):1219-1232
patibility complex class I expression in *Chlamydia*-infected cells. J Exp Med. 2000;191(9):1525-34.
[57] Ibana, JA, Schust, DJ, Sugimato, J., Nagamatsu, T., Greene, SJ. and Quayle, AJ *Chla‐ mydia trachomatis* Immune Evasion via Downregulation of MHC Class I Surface Ex‐ pression Involves Direct and Indirect Mechanisms 2011 Inf Dis. Obstet. Gynecol. Volume 2011, Article ID 420905, 8 pages doi:10.1155/2011/420905
[44] Peterson, E. M., B. A. Markoff, J. Schachter, and L. M. De La Maza. The 7.5-kbplas‐ mid present in *Chlamydia trachomatis* is not essential for the growth of this microor‐
[45] Stothard, D. R., J. A. Williams, B. Van Der Pol, and R. B. Jones. Identification of a *Chlamydia trachomatis* serovar E urogenital isolate which lacks the cryptic plasmid. In‐
[46] Sigar,I.M., Justin H. Schripsema, Yibing Wang, Ian N. Clarke, Lesley T. Cutcliffe, Helena M.B. Seth-Smith, Nicholas R. Thomson, Carina Bjartling, Magnus Unemo, Kenneth Persson and Kyle H. Ramsey. Plasmid deficiency in urogenital isolates of *Chlamydia trachomatis* reduces infectivity and virulence in a mouse model. Pathogens and Disease. Accepted manuscript online: 23 AUG 2013 09:02AM EST | DOI:
[47] Hatch, TP Developmental biology. In: RS Stephens, (ed.) *Chlamydia*: Intracellular biol‐ ogy, pathogenesis, and immunity. American Society for Microbiology, Washington,
[48] Schachter, J. and Stephens, R.S. Biology of *Chlamydia trachomatis*. In: Sexually Trans‐
[49] Schoborg, R.V.*Chlamydia* persistence – a tool to dissect *Chlamydia*-host interactions.
[50] Hafner, L.M. and Timms, P. *Chlamydia* In: Stanberry, L.R. and Rosenthal, S. L (eds.) Sexually Transmitted Diseases, 2nd edition, Elsevier (Academic Press) Ltd., Wal‐
[51] Saka, H.A., and Valdivia, R.H.Acquisition of nutrients by Chlamydiae: unique chal‐ lenges of living in an intracellular compartment. Curr Opin Microbiol.2010; 13(1):
[52] Abdelrahman, Y.M. and Belland, R.J.The Chlamydial developmental cycle. FEMS
[53] Hackstadt, T., Fischer, ER., Scidmore MA, Rockey DD, Heinzen RA. Origins and functions of the Chlamydial inclusion. Trends in Microbiol.1997; 5: 288-293.
[54] Wyrick, P.B. *Chlamydia trachomatis* persistence *in vitro*: an overview.J Infect Dis.
[55] Brunham, RC and Rey-Ladino,J. Immunology of *Chlamydia* infection: implications for
[56] Zhong G, Liu L, Fan T, Fan P, Ji H. Degradation of transcription factor RFX5 during the inhibition of both constitutive and interferon gamma-inducible major histocom‐
a *Chlamydia trachomatis* vaccine. Nature Rev. Immunol. 2005; 5(2):149–61.
Wasserheit JN, Corey L, Cohen MS, Watts C. The McGraw-Hill Companies.New
edition. Edited by: Holmes KK, Sparling PF, Stamm WE, Piot P,
ganism. Plasmid 1990; 23:144-148
204 Immune Response Activation
fect. Immun. 1998; 66:6010-6013.
10.1111/2049-632X.12086
DC. 1999 p 29-68.
mitted Diseases. 4th
York; 2008 p555-574.
4-10.
Microb. and Infect.2011; 13: 649-662.
tham, MA, USA, 2013: p369-410.
Microbiol.Rev.2005; 29, 949-959.
2010;201 Suppl 2:S88-95.
[71] Bakken IJ and Ghaderi S. Incidence of pelvic inflammatory disease in a large cohort of women tested for *Chlamydia trachomatis*: a historical follow-up study. BMC Infect Dis.2009; 9:130 http://www.biomedcentral.com/1471-2334/9/130
[83] Brunham RC, and Peeling RW. *Chlamydia trachomatis* antigens: role in immunity and
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
207
[84] Mårdh PA. Tubal factor infertility, with special regard to Chlamydial salpingitis.Curr
[85] Ness RB, Soper DE, Richter HE, Randall H, Peipert JF, Nelson DB, Schubeck D, McNeeley SG, Trout W, Bass DC, Hutchison K, Kip K, Brunham RC. *Chlamydia* anti‐ bodies, *Chlamydia* heat shock protein, and adverse sequelae after pelvic inflammatory disease: the PID Evaluation and Clinical Health (PEACH) Study.Sex Transm Dis.
[86] Oakeshott P, Kerry S, Aghaizu A, Atherton H, Hay S, Taylor-Robinson D, Simms I, Hay P. Randomised controlled trial of screening for *Chlamydia trachomatis* to prevent pelvic inflammatory disease: the POPI (prevention of pelvic infection) trial. BMJ.
[87] Institute of Medicine (US) Committee to Study Priorities for Vaccine Development (2000).Vaccines for the 21st Century. Washington, DC: National Academy Press
[88] Schachter, J., Grossman, M., Sweet, R.L., et al. Prospective Study of Perinatal Trans‐
[89] Rosenman, M.B., Mahon, B.E., Downs, S.M., et al. Oral erythromycin prophylaxis vs watchful waiting in caring for newborns exposed to *Chlamydia trachomatis*. Arch. Pe‐
[90] Brocklehurst P, Rooney G. Interventions for treating genital *Chlamydia trachomatis* in‐
[91] Beem, M.O., and Saxon, E.M. Respiratory-tract colonization and a distinctive pneu‐ monia syndrome in infants infected with *Chlamydia trachomatis*. N Engl J Med.1977;
[92] Darville, T.*Chlamydia trachomatis* infections in neonates and young children. Sem‐
[93] Darville T. Recognition and treatment of Chlamydial infections from birth to adoles‐
[94] Chandler JW, Alexander ER, Pheiffer TA, Wang SP, Holmes KK, English M. Ophthal‐ mia neonatorum associated with maternal Chlamydial infections. Trans Sect Oph‐
[95] Chandran, L. and Boykan, R. Chlamydial infections in children and adolescents.Pe‐
[96] de Borborema-Alfaia AP, de Lima Freitas NS, Filho SA, Borborema-Santos CM. *Chla‐ mydia trachomatis* infection in a sample of northern Brazilian pregnant women: preva‐
thalmol Am Acad Ophthalmol Otolaryngol. 1977;83(2):302-8.
fection in pregnancy. Cochrane Database Syst Rev. 2000; (2):CD000054
pathogenesis. Infect Agents Dis. 1994 Oct; 3(5):218-33.
2010 ;340:c1642. doi: 10.1136/bmj.c1642. Neonatal infections
mission of *Chlamydia trachomatis* JAMA.1986; 255(24):3374-3377.
diatr. Adolesc. Med.2003; 157(6):565-71.
in.Pediatr. Infect Dis.2005; 16(4):235-44.
diatr Rev. 2009; 30(7):243-50.
cence. Adv Exp Med Biol. 2013; 764:109-22.
Opin Infect Dis. 2004;17(1):49-52.
2008;35(2):129-35.
296(6):306-10.
[83] Brunham RC, and Peeling RW. *Chlamydia trachomatis* antigens: role in immunity and pathogenesis. Infect Agents Dis. 1994 Oct; 3(5):218-33.
[71] Bakken IJ and Ghaderi S. Incidence of pelvic inflammatory disease in a large cohort of women tested for *Chlamydia trachomatis*: a historical follow-up study. BMC Infect
[72] Price MJ, Ades AE, De Angelis D, Welton NJ, Macleod J, Soldan K, Simms I, Turner K, Horner PJ. Risk of Pelvic Inflammatory Disease Following *Chlamydia trachomatis* Infection: Analysis of Prospective Studies With a Multistate Model. Am J Epidemiol.
[73] Shaw JL, Wills GS, Lee KF, Horner PJ, McClure MO, Abrahams VM, Wheelhouse N, Jabbour HN, Critchley HO, Entrican G, Horne AW. *Chlamydia trachomatis* infection increases fallopian tube PROKR2 via TLR2 and NFκB activation resulting in a micro‐ environment predisposed to ectopic pregnancy. Am J Pathol. 2011;178(1):253-60. [74] Bebear C, de Barbeyrac B. Genital *Chlamydia trachomatis* infections. Clin Microbiol In‐
[75] Shao R, Wang X, Wang W, Stener-Victorin E, Mallard C, Brännström M, Billig H.From mice to women and back again: causalities and clues for *Chlamydia*-induced
[76] Ness RB, Smith KJ, Chang CC, Schisterman EF, Bass DC; Gynecologic Infection Fol‐ low-Through, GIFT, Investigators. Prediction of pelvic inflammatory disease among
[77] Haggerty CL, Gottlieb SL, Taylor BD, Low N, Xu F, Ness RB. Risk of sequelae after *Chlamydia trachomatis* genital infection in women. J Infect Dis. 2010 15;201 Suppl
[78] Hillis SD, Owens LM, Marchbanks PA, Amsterdam LF, Mac Kenzie WR. Chow JM, Yonekura, ML Recurrent Chlamydial infections increase the risks of hospitalization for ectopic pregnancy and pelvic inflammatory disease. Am J Obstet Gynecol.
[79] Chow JM, Yonekura ML, Richwald GA, Greenland S, Sweet RL, Schachter J. The as‐ sociation between *Chlamydia trachomatis* and ectopic pregnancy. A matched-pair,
[80] Bakken, IJ, Skjeldestad, FE and Nordbe, SA. *Chlamydia trachomatis* infections increase the risk for ectopic pregnancy: a population-based, nested case-control study. Sex.
[81] Claman P, Honey L, Peeling RW, Jessamine P, Toye B. The presence of serum anti‐ body to the Chlamydial heat shock protein (CHSP60) as a diagnostic test for tubal
[82] Rodgers AK, Wang J, Zhang Y, Holden A, Berryhill B, Budrys NM, Schenken RS, Zhong G. Association of tubal factor infertility with elevated antibodies to *Chlamydia trachomatis* caseinolytic protease P. Am J Obstet Gynecol. 2010 ;203(5):494.e7-494.e14.
young, single, sexually-active women. Sex Transm Dis. 2006; 33(3):137-42.
Dis.2009; 9:130 http://www.biomedcentral.com/1471-2334/9/130
tubal ectopic pregnancy. Fertil Steril. 2012;98(5):1175-85.
case-control study.JAMA. 1990;263(23):3164-7.
factor infertility. Fertil. Steril. 1997; 67:501-4.
2013;178(3):484-92.
206 Immune Response Activation
fect.2009; 15:4-10.
2:S134-55.
1997;176(1 Pt 1):103-7.
Trans. Dis.2007; 34: 166-169.
lence and prenatal importance. Braz J Infect Dis. 2013; pii: S1413-8670(13)00142-6. doi: 10.1016/j.bjid.2013.01.014.
*chomatis* :heritability estimate from a Gambian twin study. Drugs Today (Barc) 2009;
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
209
[111] Malogajski J, Brankovic I, Verweij SP, Ambrosino E, van Agtmael MA, Brand A, Ou‐ burg S, Morré SA. Translational potential into health care of basic genomic and ge‐ netic findings for human immunodeficiency virus, *Chlamydia trachomatis,* and human
[112] Karimi, O., Ouburg, S. De Vries, HJ., Pena, AS, Pliejster, J., Land, JA et al., TLR2 hap‐ lotypes in the susceptibility to and severity of *Chlamydia trachomatis* infections in
[113] Den Hartog JE, Lyons JM, Ouburg S, et al. TLR4 in *Chlamydia trachomatis* infections: knockout mice, STD patients and women with tubal factor subfertility. Drugs of To‐
[114] Broeze KA, Opmeer BC, Van Geloven N,Coppus SF, Collins JA, Den Hartog JE, Van‐ der Linden PJ, Marianowski P, Ng EH, Vander Steeg JW, Steures P, Strandell A, Van derVeen F, Mol BW: Are patient characteristics associated with the accuracy of hys‐ terosalpingography in diagnosing tubal pathology? An individual patient data meta-
[115] Morré SA, Karimi O, Ouburg S. Review *Chlamydia trachomatis*: identification of sus‐ ceptibility markers for ocular and sexually transmitted infection by immunogenetics.
[116] Lal JA, Malogajski J, Verweij SP, de Boer P, Ambrosino E, Brand A, Ouburg S, Morré SA. *Chlamydia trachomatis* infections and subfertility: opportunities to translate host pathogen genomic data into public health. Public Health Genomics.2013;16(1-2):
[117] Agrawal, T., Vats, V., Salhan, S., and Mittal, A. The mucosal immune response to *Chlamydia trachomatis* infection of the reproductive tract in women. J. Reprod. Immu‐
[118] Agrawal, T., Vats, V., Salhan, S., and Mittal, A. Mucosal and peripheral immune re‐ sponses to Chlamydial heat shock proteins in women infected with *Chlamydia tracho‐*
[119] Cohen CR., Koochesfahani, KM., Meier, AS, Shen, C., Karunakaran, K., Ondondo, B et al Immunoepidemiologic profile of *Chlamydia trachomatis* infection: importance of
heat shock protein 60 and interferon-gamma. J. Infect. Dis.2005; 192: 591-599.
[120] Roan, N.R. and Starnbach., M.N. Immune-mediated control of *Chlamydia* infection.
[121] Kimani, J., Maclean IW, Bwayo, JJ., MacDonald, K., Oyugi, J., Maitha, GM., et al., Risk factors for *Chlamydia trachomatis* pelvic inflammatory disease among sex work‐
papilloma virus. Biomed Res Int. 2013:892106. doi: 10.1155/2013/892106.
Dutch women. Drugs Today (Barc)2009; 45 (Suppl B): 67-74.
analysis. HumReprod Update 2011; 17: 293–300.
*matis*. Clin. Exp. Immunol. 2007a ; 148(3):461-468.
ers in Nairobi, Kenya J.Infect.Dis.1996; 173: 1437-1444.
FEMS Immunol Med Microbiol. 2009 Mar; 55(2):140-53.
45(supplB):45–50.
day B. 2009;45: 75–82.
50-61.
nol.2009; 83(1-2), 173-178.
Cell.Microbiol.2008;10: 9-19.
*chomatis* :heritability estimate from a Gambian twin study. Drugs Today (Barc) 2009; 45(supplB):45–50.
[111] Malogajski J, Brankovic I, Verweij SP, Ambrosino E, van Agtmael MA, Brand A, Ou‐ burg S, Morré SA. Translational potential into health care of basic genomic and ge‐ netic findings for human immunodeficiency virus, *Chlamydia trachomatis,* and human papilloma virus. Biomed Res Int. 2013:892106. doi: 10.1155/2013/892106.
lence and prenatal importance. Braz J Infect Dis. 2013; pii: S1413-8670(13)00142-6.
[97] Schachter J, Lum L, Gooding CA, Ostler B. Pneumonitis following inclusion blennor‐
[98] Tipple MA, Beem MO, Saxon EM. Clinical characteristics of the afebrile pneumonia associated with *Chlamydia trachomatis* infection in infants less than 6 months of age.
[99] Mishra KN, Bhardwaj P, Mishra A, Kaushik A. Acute *Chlamydia trachomatis* respirato‐
[100] Johansson, M and Lycke, N.Y. Immunology of the human genital tract. Curr. Opin.
[101] Hafner LM, Cunningham K, Beagley KW. Ovarian steroid hormones: effects on im‐ mune responses and *Chlamydia trachomatis* infections of the female genital tract. Mu‐
[102] Farage M A, Miller, K.W. and Sobel, J.D. Dynamics of the vaginal ecosystem - hor‐ monal influences.Infectious Diseases Research and Treatment 2010; 3: 1-15.
[103] Cole, A M. Innate host defense of human vaginal and cervical mucosae. Curr Top
[104] Wira, C.R., and Fahey, J.V. The innate immune system: gatekeeper to the female re‐
[105] Wira, C.R., Grant-Tschudy, K.S., Crane-Godreau, M,A. Epithelial cells in the female reproductive tract: a central role as sentinels of immune protection. Am. J. Reprod.
[106] Horne AW, Stock SJ, King AE. Epithelial cells in the female reproductive tract: a cen‐ tral role as sentinels of immune protection. Reprod. J, 2008; 135(6):739-49.
[107] Ochiel, D.O., Fahey, J.V., Ghosh, M., Haddad, S.N., Wira, C.R. Innate immunity in th female reproductive tract: Role of sex hormones in regulating uterine epithelial cell protection against pathogens.Curr. Womens Health Reviews, 2008;4(2): 102-117.
[108] Hosenfeld, CB, Workowski KA, Berman,S et al., Repeat infection with *Chlamydia* and gonorrhoea among females: a systematic review of the literature. Sex.Transm. Dis.
[109] Kumar H, Kawai T, Akira S. Pathogen recognition by the innate immune system. Int
[110] Bailey RL, Natividad-Sancho A, Fowler A,Peeling RW, Mabey DC, Whittle HC, Jep‐ son AP: Host genetic contribution to the cellular immune response to *Chlamydia tra‐*
ry infection in infants. J Glob Infect Dis. 2011;3(3):216-20.
doi: 10.1016/j.bjid.2013.01.014.
208 Immune Response Activation
rhea.J Pediatr1975;87(5):779-80.
Pediatrics.1979; 63(2):192-7.
Infect. Dis.2003; 16: 43-49.
Immunol.2005;, 53: 65-76.
2009; 36(8): 478-489.
Rev Immun.2011; 30(1):16-34.
cosal Immunol. 2013a;6(5):859-75.
Microbiol Immunol., 2006;306:199-230.
productive tract. Immunol., 2004; 111: 13-15.
[122] Brunham, RC, Kimani, J, Bwayo, J., Maitha, G., Maclean,I., Yang, C et al. The epi‐ demiology of *Chlamydia trachomatis* within a sexually transmitted core group. J. In‐ fect. Dis.1996; 173: 950-956.
[135] Wira, C.R., Grant-Tschudy, K.S., Crane-Godreau, M,A. Epithelial cells in the female reproductive tract: a central role as sentinels of immune protection. Am. J Reprod.
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
211
[136] Kang, L, Zhang, X, Xie, Y, Tu, Y, Wang, D, Liu, ZWang, ZY, Tsai, MJ Involvement of estrogen receptor variant ER-alpha36, not GPR30, in non-genomic estrogen signal‐
[137] Solar, P and Velasquez, T: Consequences of non-genomic actions of estradiol on pathogenic genital tract response. Journal of Molecular Signaling 2013 8:1. doi:
[138] Lesmeister, M.J., Jorgensen, R.L., Young, S.L., and Misfeldt, M.L. 17 beta-estradiol suppresses TLR3-induced cytokine and chemokine production in endometrial epi‐
[139] Nalbandian, G and Kovats, S. Understanding sex biases in immunity. Immunologic
[140] Hughes GC, Clark EA. Regulation of dendritic cells by female sex steroids: relevance
[141] Kovats, S., Carreras, E. Regulation of dendritic cell differentiation and function by es‐
[142] Heine H, Müller-Loennies S, Brade L, Lindner B, Brade H. Endotoxic activity and chemical structure of lipopolysaccharides from *Chlamydia trachomatis* serotypes E and
[143] Da Costa CU, Wantia N, Kirschning CJ, Busch DH, Rodriguez N, Wagner H, Miethke T. Heat shock protein 60 from *Chlamydia pneumoniae* elicits an unusual set of inflam‐ matory responses via Toll-like receptor 2 and 4 in vivo. Eur J Immunol.2004; 34(10):
[144] Derbigny WA, Kerr MS, Johnson RM. Pattern recognition molecules activated by *Chlamydia muridarum* infection of cloned murine oviduct epithelial cell lines. J Immu‐
[145] Vabulas RM, Ahmad-Nejad P, da Costa C, Miethke T, Kirschning CJ, Häcker H, Wagner H. Endocytosed HSP60s use toll-like receptor 2 (TLR2) and TLR4 to activate the toll/interleukin-1 receptor signaling pathway in innate immune cells. J Biol
[146] Prebeck S, Kirschning C, Dürr S, da Costa C, Donath B, Brand K, Redecke V, Wagner H, Miethke T. Predominant role of toll-like receptor 2 versus 4 in *Chlamydia pneumo‐*
[147] Welter-Stahl L, Ojcius DM, Viala J, Girardin S, Liu W, Delarbre C, Philpott D, Kelly KA, Darville T. Stimulation of the cytosolic receptor for peptidoglycan, Nod1, by in‐
*niae*-induced activation of dendritic cells. J Immunol. 2001;167(6):3316-23.
to immunity and autoimmunity. Autoimmunity2007; 40(6):470-81.
L2 and *Chlamydophila psittaci* 6BC. Eur J Biochem.2003; 270(3):440-50.
trogen receptor ligands. Cell. Immunol.,2008; 252: 81-90.
Immunol.2005;, 53: 65-76.
10.1186/1750-2187-8-1.
Research, 2005; 31(2): 91-106.
2874-84.
nol.2005;175(9):6065-75.
Chem.2001; 276(33):31332-9.
ling. Mol Endocrinol. 2010; 24(4): 709-730.
thelial cells. Reprod. Biol. Endocrinol.2005;, 3:74-89.
[135] Wira, C.R., Grant-Tschudy, K.S., Crane-Godreau, M,A. Epithelial cells in the female reproductive tract: a central role as sentinels of immune protection. Am. J Reprod. Immunol.2005;, 53: 65-76.
[122] Brunham, RC, Kimani, J, Bwayo, J., Maitha, G., Maclean,I., Yang, C et al. The epi‐ demiology of *Chlamydia trachomatis* within a sexually transmitted core group. J. In‐
[123] Kaldensjö,T.,Petersson,P.,Tolf,A.,Morgan,G.,Broliden,K., Hirbod T. Detection of in‐ traepithelial and stromal Langerin and CCR5 positive cells in the human endometri‐
[124] Arrivito, L., Sanz, M., Banham, A.H., Fainboim, L. Expansion of CD4+CD25+ and FOXP3+ regulatory T cells during the follicular phase of the menstrual cycle: implica‐
[125] Ficarra M, Ibana JS, Poretta C, Ma L, Myers L, Taylor SN, Greene S, Smith B, Hagen‐ see M, Martin DH, Quayle AJ. A distinct cellular profile is seen in the human endo‐ cervix during *Chlamydia trachomatis* infection. A J Reprod Immunol, 2008; 60:415-425.
[126] Mascellino MT, Boccia P, Oliva A. Immunopathogenesis in *Chlamydia trachomatis* In‐ fected Women. SRN Obstet Gynecol. 2011:436936. doi: 10.5402/2011/436936.
[127] Pudney J, Quayle AJ, Anderson DJ. Immunological microenvironments in the human vagina and cervix: mediators of cellular immunity are concentrated in the cervical
[128] Yeaman, G.R Guyre, P.M., Fanger, M.W., Collins, J.E., White, H.D., Rathbun, W., et al. Unique CD8+ T cell rich lymphoid aggregates in human uterine endometrium.
[129] Wira, C.R., Fahey, J.V.,Ghosh,M., Patel,V., Hickey D.K. and Ochiel D.O. Sex Hor‐ mone Regulation of Innate Immunity in the Female Reproductive Tract: The Role of Epithelial Cells in Balancing Reproductive Potential with Protection against Sexually
[130] Bouman A, Heineman MJ, Faas MM. Sex hormones and the immune response in hu‐
[131] Wira C, Fahey J, Wallace P, Yeaman G. Effect of the menstrual cycle on immunologi‐ cal parameters in the human female reproductive tract. J Acquir Immune Defic
[132] Butts, C.L., Sternberg, E.M. Neuroendocrine factors alter host defense by modulating
[133] Aflatoonian R, Tuckerman E, Elliott SL, Bruce C, Aflatoonian A, Li TC, Fazeli A. Menstrual cycle-dependent changes of Toll-like receptors in endometrium. Hum Re‐
[134] Aflatoonian R, Fazeli A. Toll-like receptors in female reproductive tract and their menstrual cycle dependent expression. J Reprod Immunol. 2008;77(1):7-13.
Transmitted Pathogens Am J Reprod. Immunol. 2010; 63: 544–565.
immune function. Cellular Immunology,2008; 252(1-2):7-15.
um: potential targets for HIV infection. PLoS One. 2011; 6(6): e21344.
tions for human reproduction.J. Immunol.2007; 178: 2572-2578.
transformation zone.Biol Reprod.2005; 73(6):1253-63.
mans Hum Reprod Update.2005; 11(4):411-23.
fect. Dis.1996; 173: 950-956.
210 Immune Response Activation
PMCID1997; : 9103229.
Syndr.2005;38 Suppl 1:S34-6.
prod.2007; 22(2):586-93.
fection with *Chlamydia trachomatis* or *Chlamydia muridarum.*Cell Microbiol.2006; 8(6): 1047-57.
[161] Barron AL, Pasley JN, Rank RG, White HJ, Mrak RE. Chlamydial salpingitis in fe‐ male guinea pigs receiving oral contraceptives. Sex Transm Dis1988; 15:169–173. [162] Kaushic C, Murdin AD, Underdown BJ, Wira CR. *Chlamydia trachomatis* infection in the female reproductive tract of the rat: influence of progesterone on infectivity and
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
213
[163] Kaushic C, Zhou F, Murdin AD, Wira CR. Effects of estradiol and progesterone on susceptibility and early immune responses to *Chlamydia trachomatis* infection in the
[164] Bose SK, Goswami PC. Enhancement of adherence and growth of *Chlamydia tracho‐ matis* by estrogen treatment of HeLa cells. Infect Immun. 1986;53(3):646-50.
[165] Agrawal T, Vats V, Wallace PK, Salhan S, Mittal A. Role of cervical dendritic cell sub‐ sets, co-stimulatory molecules, cytokine secretion profile and beta-estradiol in devel‐ opment of sequalae to *Chlamydia trachomatis* infection. Reprod Biol Endocrinol.2008;
[166] Amirshahi A, Wan C, Beagley K, Latter J, Symonds I, Timms P. Modulation of the *Chlamydia trachomatis* in vitro transcriptome response by the sex hormones estradiol and progesterone. BMC Microbiol. 2011;25;11:150. doi: 10.1186/1471-2180-11-150. [167] Hogan RJ, Mathews SA, Mukhopadhyay S, Summersgill JT, Timms P. Chlamydial persistence: beyond the biphasic paradigm. Infect Immun. 2004;72(4):1843-55
[168] Akers JC, Tan M. Molecular mechanism of tryptophan-dependent transcriptional
[169] Schaefer TM, Desouza K, Fahey JV, Beagley KW, Wira CR. Toll-like receptor (TLR) expression and TLR-mediated cytokine/chemokine production by human uterine ep‐
[170] Pioli, PA, Amiel E, Schaefer TM, Connolly JE, Wira CR, Guyre PM. Differential ex‐ pression of Toll-like receptors 2 and 4 in tissues of the human female reproductive
[171] Fazeli A, Bruce C, Anumba DO. Characterization of Toll-like receptors in the female
[172] Agrawal T, Bhengraj AR, Vats V, Salhan S, Mittal A. Expression of TLR 2, TLR 4 and iNOS in Cervical Monocytes of *Chlamydia trachomatis*-infected Women and Their Role
[173] Agrawal T, Bhengraj AR, Vats V, Mittal A. *Chlamydia trachomatis*: TLR4-mediated rec‐ ognition by human dendritic cells is impaired following oestradiol treatment. Br J Bi‐
in Host Immune Response. Am J Reprod Immunol. 2011 ; 66 (6):534-43.
reproductive tract in humans. Hum Reprod.2005; 20(5):1372-8.
regulation in *Chlamydia trachomatis*. J Bacteriol. 2006;188 (12):4236-43.
immune response.Infect Immun.1998 Mar;66(3):893-8.
6:46. doi: 10.1186/1477-7827-6-46.
ithelial cells. Immunol.2004; 112(3):428-36.
tract. Infect Immun., 2004;72(10):5799-806.
omed Sci. 2013;70(2):51-7.
female reproductive tract. Infect Immun. 2000;68(7):4207-16.
[161] Barron AL, Pasley JN, Rank RG, White HJ, Mrak RE. Chlamydial salpingitis in fe‐ male guinea pigs receiving oral contraceptives. Sex Transm Dis1988; 15:169–173.
fection with *Chlamydia trachomatis* or *Chlamydia muridarum.*Cell Microbiol.2006; 8(6):
[148] Kaushic, C. The role of the local microenvironment in regulating susceptibility and immune responses to sexually transmitted viruses in the female genital tract. J. Re‐
[149] Keenihan, S.N., and Roberston, A.A. Diversity in phenotype and steroid hormone dependence in dendritic cells and macrophages in the mouse uterus. Biol.Reprod.
[150] Patton, D.L., Thinn, S.S., Meier, A., Hooton, T.M. Stapleton, A.E., Eschenbach, D.A. Epithelial cell layer thickness and immune cell populations in the normal human vagina at different stages of the menstrual cycle. Am. J. Obstet. Gynecol.2000;183(4):
[151] Kaul R, Pettengell C, Sheth PM, Sunderji S, Biringer A, MacDonald K, Walmsley S, Rebbapragada A. The genital tract immune milieu: an important determinant of HIV susceptibility and secondary transmission. J Reprod Immunol.2008; 77(1):32-40. [152] Fish, E.N. The X-files in immunity: sex-based differences predispose immune re‐
[153] Kelly, K.A. T-lymphocyte trafficking to the female reproductive tract mucosae. In: Editor. P.B. Wyrick (ed.). *Chlamydia*: Genomics and Pathogenesis,. Norfolk, UK.: Ho‐
[154] Menge AC, Mestecky J. Surface expression of secretory component and HLA class II DR antigen on glandular epithelial cells from human endometrium and two endome‐
[155] Wang Y, Ben K, Cao X, Wang Y. Transport of anti-sperm monoclonal IgA and IgG into murine male and female genital tracts from blood. Effect of sex hormones. J Im‐
[156] Lu, F.X., Ma, Z., Moser, S., Evans, T.G., Miller, C.J. Effects of ovarian steroids on im‐ munoglobulin-secreting cell function in healthy women.Clin. and Diag. Lab. Immu‐
[157] Rank RG. Animal models for urogenital infections. Methods Enzymol.1994;235:83-93. [158] Rank RG, White HJ, Hough AJ Jr, Pasley JN, Barron AL. Effect of estradiol on Chla‐ mydial genital infection of female guinea pigs. Infect Immun. 1982;38(2):699-705. [159] Pasley JN, Rank RG, Hough AJ, Jr, Cohen C, Barron AL. Effects of various doses of estradiol on Chlamydial genital infection in ovariectomized guinea pigs. Sex.
[160] Rank RG, Barron AL.Specific effect of estradiol on the genital mucosal antibody re‐ sponse in Chlamydial ocular and genital infections. Infect Immun.1987; 55:2317–2319.
trial adenocarcinoma cell lines.J Clin Immunol. 1993; 13(4):259-64.
1047-57.
212 Immune Response Activation
967-73.
prod. Immunol.2009;, 83:168-72.
sponses. Nat. Rev. Immunol.2008; 8 :737-744.
rizon Bioscience, 2006. p. 413-435.
munol.1996;156(3):1014-9.
nol.2003; 10(5):944-949.
Transm. Dis.1985; 12:8–13.
2004; 70:1562-1572.
[174] Mahmoud EA, Hamad EE, Olsson SE, Mårdh PA. Anti-Chlamydial activity of cervi‐ cal secretion in different phases of the menstrual cycle and influence of hormonal contraceptives. Contraception.1994; 49(3):265-74.
[187] Morrison, R.P., et al., 1989. Chlamydial disease pathogenesis. The 57-kDa Chlamy‐ dial hypersensitivity antigen is a stress response protein. J. Exp.Med. 170, 1271–1283.
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
215
[188] Ault, K.A., et al., 1998. Antibodies to the Chlamydial 60 kilodalton heat shock protein in women with tubal factor infertility. Infect. Dis. Obstet. Gynecol. 6, 163–167 [189] Bax, C.J., et al., 2004. *Chlamydia trachomatis* heat shock protein 60 (cHsp60) antibodies in women without and with tubal pathology using a new commercially available as‐
[190] Dapontea,A, Pournarasb,, S., Deligeoroglouc,E., Skentoua,H., Messinisa, I.E. Serum interleukin-1\_, interleukin-8 and anti-heat shock 60 *Chlamydia trachomatis* antibodies
[191] Taylor BD, Darville T, Ferrell RE, Kammerer CM, Ness RB, Haggerty CL. Variants in toll-like receptor 1 and 4 genes are associated with *Chlamydia trachomatis* among women with pelvic inflammatory disease. J Infect Dis. 2012 Feb 15;205(4):603-9. doi:
[192] Haggerty, C.L., Panum, I., Uldum, S., et al., *Chlamydia trachomatis* infection may in‐ crease the risk of preeclampsia Pregnancy Hypertension: An international journal of
[193] Srivastava P, Jha R, Bas S, Salhan S, Mittal A. In infertile women, cells from *Chlamydia trachomatis* infected sites release higher levels of interferon-gamma, interleukin-10 and tumor necrosis factor-alpha upon heat-shock-protein stimulation than fertile women. Reprod Biol Endocrinol. 2008 May 20;6:20. doi: 10.1186/1477-7827-6-20. [194] Taylor BD, Darville T, Ferrell RE, Ness RB, Haggerty CL. Racial variation in toll-like receptor variants among women with pelvic inflammatory disease. J Infect Dis. 2013
[195] Stephens R, Kalman S, Lammel C, et al. Genome sequence of an obligate intracellular pathogen of humans: *Chlamydia trachomatis*. *Science*. 1998;282(5389):754–759
[196] Taylor BD, Darville T, Tan C, Bavoil PM, Ness RB, Haggerty CL. The role of *Chlamy‐ dia trachomatis* polymorphic membrane proteins in inflammation and sequelae among women with pelvic inflammatory disease. Infect Dis Obstet Gynecol.
[197] Gijsen, A.P., Land, J.A., Goossens, V.J., Slobbe, M.E., Bruggeman, C.A. *Chlamydia* an‐ tibody testing in screening for tubal factor subfertility: the significance of IgG anti‐
[198] Veenemans, L.M., van der Linden, P.J. The value of *Chlamydia trachomatis* antibody testing in predicting tubal factor infertility. Hum Reprod. 2002 17:695-98
Mar 15;207(6):940-6. doi: 10.1093/infdis/jis922. Epub 2012 Dec 18
2011;2011:989762. doi: 10.1155/2011/989762. Epub 2011 Oct 19.
body decline over time. Hum Reprod. 2002 17:699-703
as markers of ectopic pregnancy. J.Reprod. Immunol. 2012; 93: 102-108
say. Sex. Transm. Infect. 80, 415–416.
10.1093/infdis/jir822. Epub 2012 Jan 11
womens cardiovascular health 3 (2013) 28-33.
[187] Morrison, R.P., et al., 1989. Chlamydial disease pathogenesis. The 57-kDa Chlamy‐ dial hypersensitivity antigen is a stress response protein. J. Exp.Med. 170, 1271–1283.
[174] Mahmoud EA, Hamad EE, Olsson SE, Mårdh PA. Anti-Chlamydial activity of cervi‐ cal secretion in different phases of the menstrual cycle and influence of hormonal
[175] Sweet RL, Landers DV, Walker C, Schachter J. *Chlamydia trachomatis* infection and
[176] Agrawal, T., Vats, V., Wallace, P.K., Salhan, S., and Mittal, A. Cervical cytokine re‐ ponses in women with primary or recurrent Chlamydial infection. Journal of Inter‐
[177] Washington AE, Gove S, Schachter J, Sweet RL. Oral contraceptives, *Chlamydia tra‐ chomatis* infection, and pelvic inflammatory disease. A word of caution about protec‐
[178] Baeten JM, Nyange PM, Richardson BA, Lavreys L, Chohan B, Martin HL Jr, Manda‐ liya K, Ndinya-Achola JO, Bwayo JJ, Kreiss JK. Hormonal contraception and risk of sexually transmitted disease acquisition: results from a prospective study. Am J Ob‐
[179] Morrison CS, Bright P, Wong EL, Kwok C, Yacobson I, Gaydos CA, Tucker HT, Blu‐ menthal PD. Hormonal contraceptive use, cervical ectopy, and the acquisition of cer‐
[180] Williams DM, Schachter J, Drutz DJ, Sumaya CV. Pneumonia due to *Chlamydia tra‐ chomatis* in the immunocompromised (nude) mouse. *J Infect Dis.* 1981 143: 238-41.
[181] Igietseme JU, Eko FO, He Q, Black CM. Antibody regulation of Tcell immunity: im‐ plications for vaccine strategies against intracellular pathogens. *Expert Rev Vaccines*.
[182] Stephens RS. The cellular paradigm of Chlamydial pathogenesis. *Trends Microbiol*.
[183] Rice PA, Westrom LV. Pathogenesis and inflammatory response in pelvic inflamma‐ tory disease. In: Berger GS, Westrom LV. eds. *Pelvic Inflammatory Disease*. New York:
[184] Oakeshott P, Hay P, Hay S, Steinke F, Rink E, Kerry S. Association between bacterial vaginosis or Chlamydial infection and miscarriage before 16 weeks' gestation: pro‐
[185] Molano, M., Meijer, CJ, Weiderpass E et al., The natural course of *Chlamydia trachoma‐ tis* infection in asymptomatic Colombian women: a 5-year follow-up study. J. Infect.
[186] Cohen, C.R., Koochesfahani, A.S., Meoer et al., Immunoepidemiologic profile of *Chla‐ mydia trachomatis* infection: importance of heat shock protein 60 and interferon-γ.
spective community based cohort study. *BMJ.* 2002 325: 1334.
contraceptives. Contraception.1994; 49(3):265-74.
tion. J Am Med Assoc.1985; 253:2246–2250.
vical infections. Sex Transm Dis. 2004, 31(9):561-7.
stet Gynecol. 2001;185(2):380-5.
2004 3: 23-34.
214 Immune Response Activation
2003 11: 44-51.
Raven Press. 1992 35-47.
Dis., 2005, 191: 907-16
J.Infect.Dis.;192(4):591-99
pregnancy outcome. Am J Obstet Gynecol.1987;156(4):824-33.
feron and Cytokine Research,2007b 27: 221-226. contraceptives
[199] den Hartog, J.E., Land, J.A., Stassen, F.R., Kessels, A.G., Bruggeman, C.A. Serological markers of persistent *C. trachomatis* infections in women with tubal subfertility. Hum Reprod. 2005 20:986-90
[210] Krause W, Bohring C: Male infertility and genital Chlamydial infection: victim or
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
217
[211] Zervomanolakis I, Ott W, Hadziomerovic D, Mattle V, Seeber B, Virgolini I, Heute D, Kissler S, Leyendecker G, Wildt L: Physiology of upward transport in the human fe‐
[212] Geisler WM, Wang C, Morrison SG, Black CM, Bandea CI, Hook EW 3rd. The natural history of untreated *Chlamydia trachomatis* infection in the interval between screening
[213] Joyner JL, Douglas JM Jr, Foster M, Judson FN. Persistence of *Chlamydia trachomatis* infection detected by polymerase chain reaction in untreated patients. Sex Transm
[214] Kelly K. Cellular immunity and *Chlamydia* genital infection: induction, recruitment
[215] Hosenfeld CB, Workowski KA, Berman S, et al. Repeat infection with *Chlamydia* and gonorrhea among females: a systematic review of the literature. Sex Transm Dis 2009;
[216] Xu F, Stoner BP, Taylor SN, et al. Use of home-obtained vaginal swabs to facilitate rescreening for *Chlamydia trachomatis* infections: two randomized controlled trials.
[217] Dunne EF, Chapin JB, Rietmeijer CA, et al. Rate and predictors of repeat *Chlamydia*
[218] Yang, X., K.T. HayGlass, and R.C. Brunham, Genetically determined differences in IL-10 and IFN-gamma responses correlate with clearance of *Chlamydia trachomatis*
[219] Yang, X., Role of cytokines in *Chlamydia trachomatis* protective immunity and immu‐
[220] Su H, Caldwell HD., CD4+ T cells play a significant role in adoptive immunity to *Chlamydia trachomatis* infection of the mouse genital tract. Infect Immun., 1995 63:
[221] Caldwell, H.D. and L.J. Perry, Neutralization of *Chlamydia trachomatis* infectivity with antibodies to the major outer membrane protein, in Infect Immun. 1982. p. 745-54.
[222] Berry LJ, Hickey DK, Skelding KA, Bao S, Rendina AM, et al. (2004) Transcutaneous immunization with combined cholera toxin and CpG adjuvant protects against *Chla‐*
[223] Skelding KA, Hickey DK, Horvat JC, Bao S, Roberts KG, et al. (2006) Comparison of intranasal and transcutaneous immunization for induction of protective immunity
against *Chlamydia* muridarum respiratory tract infection. Vaccine 24: 355–366
*mydia* muridarum genital tract infection. Infect Immun 72: 1019–1028
*trachomatis* infection among men. Sex Transm Dis 2008; 35: S40–4.
mouse pneumonitis infection. J Immunol, 1996. 156(11): p. 4338-44.
nopathology. Curr Pharm Des, 2003. 9(1): p. 67-73.
perpetrator? Andrologia 2003; 35:209–216.
Dis 2002; 29:196–200.
36:478–89)
3302–3308.
male genital tract. Ann NY Acad Sci 2007; 1101:1–20.
and returning for treatment. Sex Transm Dis 2008; 35:119–23.
and effector mechanisms. Int Rev Immunol 2003;22:2-41.
Obstet Gynecol 2011; 118(2 Pt 1):231–9.
[210] Krause W, Bohring C: Male infertility and genital Chlamydial infection: victim or perpetrator? Andrologia 2003; 35:209–216.
[199] den Hartog, J.E., Land, J.A., Stassen, F.R., Kessels, A.G., Bruggeman, C.A. Serological markers of persistent *C. trachomatis* infections in women with tubal subfertility. Hum
[200] den Hartog, J.E., Morre, S.A., Land, J.A. *Chlamydia trachomatis*-associated tubal factor subfertility: Immunogenetic aspects and serological screening. Hum Reprod Update.
[201] Tiitinen, A., Surcel, H.M., Hlttunen, M., Birkeland, S., Bloigu, A., Christiansen, G., Koskela, P., Morrison, S.G., Morrison, R.P., Paavonen, J. *Chlamydia trachomatis* an Chlamydial heat shock protein 60-specific antibody and cell-mediated responses pre‐
[202] Westrom L, Joesoef R, Reynolds G, Hagdu A, Thompson SE. Pelvic inflammatory disease and fertility. A cohort study of 1,844 women with laparoscopically verified disease and 657 control women with normal laparoscopic results. *Sex Transm Dis.*
[203] Ohman H, Tiitinen A, Halttunen M, Lehtinen M, Paavonen J, Surcel HM. Cytokine polymorphisms and severity of tubal damage in women with *Chlamydia*-associated
[204] Ohman H, Bailey R, Natividad A, Ragoussis J, Johnson LL, Tiitinen A, Halttunen M, Paavonen J, Surcel HM. Effect of IL12A and IL12B polymorphisms on the risk of *Chlamydia trachomatis*-induced tubal factor infertility and disease severity. Hum Re‐
prod. 2012 Jul;27(7):2217-23. doi: 10.1093/humrep/des136. Epub 2012 May 11.
Mar;72(3):278-82. doi: 10.1016/j.humimm.2010.12.012. Epub 2011 Jan 4.
494.e7-494.e14. doi: 10.1016/j.ajog.2010.06.005.
10.1093/humrep/der167. Epub 2011 Jun 4.
JCDR/2012/3771.2632. Epub 2012 Sep 12.
[205] Ohman H, Tiitinen A, Halttunen M, Paavonen J, Surcel HM. Cytokine gene polymor‐ phism and *Chlamydia trachomatis*-specific immune responses. Hum Immunol. 2011
[206] Rodgers AK, Wang J, Zhang Y, Holden A, Berryhill B, Budrys NM, Schenken RS, Zhong G. Association of tubal factor infertility with elevated antibodies to *Chlamydia trachomatis* caseinolytic protease P. Am J Obstet Gynecol. 2010 Nov;203(5):
[207] Hjelholt A, Christiansen G, Johannesson TG, Ingerslev HJ, Birkelund S. Tubal factor infertility is associated with antibodies against *Chlamydia trachomatis* heat shock pro‐ tein 60 (HSP60) but not human HSP60. Hum Reprod. 2011 Aug;26(8):2069-76. doi:
[208] Surana A, Rastogi V, Nirwan PS. Association of the Serum Anti-Chlamydial Anti‐ bodies with Tubal Infertility. J Clin Diagn Res. 2012 Dec;6(10):1692-4. doi: 10.7860/
[209] Brunham RC, Rey-Ladino. Immunology of *Chlamydia* infection: implications for a
*Chlamydia trachomatis* vaccine. Nat Rev Immunol, 2005 5: 149–16.
infertility. J Infect Dis. 2009 May 1;199(9):1353-9. doi: 10.1086/597620.
dict tubal factor infertility. Hum Reprod. 2006 21:151-60
Reprod. 2005 20:986-90
2006 12:719-30
216 Immune Response Activation
1992 19: 185-192.
[224] Pal, S., et al., Monoclonal immunoglobulin A antibody to the major outer membrane protein of the *Chlamydia trachomatis* mouse pneumonitis biovar protects mice against a Chlamydial genital challenge Vaccine. 1997;15(5): 575-82.
[236] Darville T, O'Neill JM, Andrews CW Jr, Nagarajan UM, Stahl L, Ojcius DM: Toll-like receptor-2, but not toll-like receptor-4, is essential for development of oviduct pathol‐
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
219
[237] Buchholz, K.R. and Stephens, R.S. The Cytosolic Pattern Recognition Receptor NOD1
[238] Buchholz, K. R. and Stephens, R.S. Activation of the host cell proinflammatory inter‐ leukin-8 response by *Chlamydia trachomatis*. Cell. Microbiol.2006; 8:1768–1779.
[239] Tseng C-TK, Rank RG: Role of NK cells in early host response to Chlamydial genital
[240] Rank RG, Bowlin AK, Kelly KA. Characterization of lymphocyte re- sponse in the fe‐ male genital tract during ascending Chlamydial genital infection in the guinea pig
[241] Carey, A. J., & Beagley, K. W. (2010). *Chlamydia trachomatis*, a hidden epidemic: ef‐ fects on female reproduction and options for treatment. *American journal of reproduc‐*
[242] Patton DL, Sweeney YT, Kuo CC: Demonstration of delayed hypersensitivity in *Chla‐ mydia trachomatis* salpingitis in monkeys: a pathogenic mechanism of tubal damage. J
[243] Watkins NG Hadlow WJ, Moos AB, Caldwell HD: Ocular delayed hypersensitivity: a pathogenetic mechanism of Chlamydial-conjunctivitis in guinea pigs. Proc Natl Acad
[244] Taylor H, Johnson S, Schachter J, Caldwell H, Prendergast R: Pathogenesis of tracho‐
[245] Kinnunen A, Paavonen J, Surcel HM: Heat shock protein 60 specific T-cell response
[246] Campanella C, Marino Gammazza A, Mularoni L, Cappello F, Zummo G, Di Felice V: A comparative analysis of the products of GROEL-1 gene from *Chlamydia tracho‐ matis* serovar D and the HSP60 var1 transcript from Homo sapiens suggests a possi‐
[247] Yi Y, Yang X, Brunham RC: Autoimmunity to heat shock protein 60 and antigen-spe‐
[248] Peeling RW, Kimani J, Plummer F, Maclean I, Cheang M, Bwayo J, Brunham RC: An‐ tibody to Chlamydial hsp60 predicts an increased risk for Chlamydial pelvic inflam‐
[249] Ness,RB, Soper, DE,Richter, HE,et al. *Chlamydia* antibodies, *Chlamydia* heat shock protein, and adverse sequelae after pelvic inflammatory disease: the PID Evaluation
and Clinical Health (PEACH) Study. Sex Transm Dis 2008;35(2):129–135
ma: the stimulus for inflammation. J Immunol 1987;138:3023–3027.
in Chlamydial infections. Scand J Immunol 2001; 54:76–81.
ble autoimmune response. Int J Immunogenet 2009; 36:73–78
matory disease. J Infect Dis 1997; 175:1153–1158)
cific production of interleukin-10. Infect Immun 1997; 65:1669–1674.
ogy in Chlamydial genital tract infection. J Immunol 2003; 171:6187–6197.
Induces Inflammatory.Infection and Immunity, 2008; 76(7):3150–3155.
infection. Infect Immun.1998; 66:5867–5875.
model. Infect Immun 2000;68(9): 5293–5298).
Infect Dis 1994; 169:680–683
Sci USA 1986; 83: 7480–7484
*tive immunology (New York, N.Y. : 1989)*, *63*(6), 576–586.
[236] Darville T, O'Neill JM, Andrews CW Jr, Nagarajan UM, Stahl L, Ojcius DM: Toll-like receptor-2, but not toll-like receptor-4, is essential for development of oviduct pathol‐ ogy in Chlamydial genital tract infection. J Immunol 2003; 171:6187–6197.
[224] Pal, S., et al., Monoclonal immunoglobulin A antibody to the major outer membrane protein of the *Chlamydia trachomatis* mouse pneumonitis biovar protects mice against
[225] Morrison, S.G. and R.P. Morrison, Resolution of secondary *Chlamydia trachomatis* gen‐ ital tract infection in immune mice with depletion of both CD4+ and CD8+ T cells, in
[226] (Morrison, S.G. and R.P. Morrison, A predominant role for antibody in acquired im‐ munity to Chlamydial genital tract reinfection, in J Immunol. 2005. p. 7536-42. [227] (Morrison, S.G., et al., Immunity to murine *Chlamydia trachomatis* genital tract reinfec‐ tion involves B cells and CD4(+) T cells but not CD8(+) T cells, in Infect Immun. 2000.
[228] Igietseme JU, Magee DM, Williams DM, Rank RG: Role for CD8+ T cells in antiChla‐ mydial immunity defined by *Chlamydia*-specific T-lymphocyte clones. Infect Immun.
[229] Barteneva N, Theodor I, Peterson E, de la Maza L: Role of neutrophils in controlling early stages of a *Chlamydia trachomatis* infection. Infect Immun 1996; 64:4830–4833.
[230] Zhang D, Yang X, Lu H, Zhong G, Brunham RC: Immunity to *Chlamydia trachomatis* mouse pneumonitis induced by vaccination with live organisms correlates with early granulocyte-macrophage colony-stimulating factor and interleukin-12 production
[231] Agrawal T, Vats V, Wallace PK, Singh A, Salhan S, Mittal A: Recruitment of myeloid and plasmacytoid dendritic cells in cervical mucosa during *Chlamydia trachomatis* in‐
[232] Hvid M, Baczynska A, Deleuran B, Fedder J, Knudsen HJ, Christiansen G, Birkelund S: Interleukin-1 is the initiator of fallopian tube destruction during *Chlamydia tracho‐*
[233] Rasmussen SJ, Eckmann L, Quayle AJ, Shen L, Zhang Y-X, Anderson DJ, Fierer J, Ste‐ phens RS, Kagnoff MF: Secretion of proinflammatory cytokines by epithelial cells in response to *Chlamydia* infection suggests a central role for epithelial cells in Chlamy‐
[234] Horne AW, Stock SJ, King AE. Epithelial cells in the female reproductive tract: a cen‐ tral role as sentinels of immune protection. Reproduction J, 2008; 135(6):739-49. [235] O'Connell CM, Ionova IA, Quayle AJ, Visintin A, Ingalls RR. Localization of TLR2 and MyD88 to *Chlamydia trachomatis* inclusions. Evidence for signaling by intracellu‐ lar TLR2 during infection with an obligate intracellular pathogen. J Biol Chem. 2006;
and with dendritic cell-like maturation. Infect Immun1999; 67:1606–1613.
fection. Clin Microbiol Infect 2009; 15:50–59.
*matis* infection. Cell Microbiol 2007; 9:2795–2803.
dial pathogenesis. J Clin Invest 1997; 99:77–87.
a Chlamydial genital challenge Vaccine. 1997;15(5): 575-82.
Infect Immun. 2001. p. 2643-9.
p. 6979-87.
218 Immune Response Activation
1994; 62:5195–5197
281(3):1652-9.
[250] Cohen CR, Koochesfahani KM, Meier AS, et al. Immunoepidemiologic profile of *Chlamydia trachomatis* infection: importance of heat-shock protein 60 and interferongamma. J Infect Dis 2005; 192(4):591–599
[263] Batteiger BE, Tu W, Ofner S, et al. Repeated *Chlamydia trachomatis* genital infections
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
221
[264] Golden MR, Whittington WL, Handsfield HH, Hughes JP, Stamm WE, Hogben M, Clark A, Malinski C, Helmers JR, Thomas KK, Holmes KK. Effect of expedited treat‐ ment of sex partners on recurrent or persistent gonorrhea or Chlamydial infection N
[265] Kelsi M Sandoz and Daniel D Rockey, Antibiotic resistance in Chlamydiae, Future
[266] Horner PJ (2012) Azithromycin antimicrobial resistance and genital *Chlamydia tracho‐ matis* infection: duration of therapy may be the key to improving efficacy. Sex
[267] Drummond F, Ryder N, Wand H, et al. Is azithromycin adequate treatment for
[268] Thejls H, Gnarpe J, Lundkvist O, Heimer G, Larsson G, Arne V. Diagnosis and preva‐ lence of persistent *Chlamydia* infection in infer- tile women: tissue culture, direct anti‐
[269] Campbell LA, Patton DL, Moore DE, Cappuccio AL, Mueller BA, Wang SP. Detec‐ tion of *Chlamydia trachomatis* deoxyribonucleic acid in women with tubal infertility.
[270] Patton DL, Askienazy-Elbhar BD, Henry-Suchet J, et al. Detection of *Chlamydia tra‐ chomatis* in fallopian tube tissue in women with post infectious tubal infertility. Am J
[271] Beatty WL, Byrne GI, Morrison RP (1994) Repeated and persistent infection with
[272] Wiggins R, Graf S, Low N, et al; *Chlamydia* Screening Studies (ClaSS) Study Group. Real-time quantitative PCR to determine Chlamydial load in men and women in a
[273] Michel CE, Sonnex C, Carne CA, et al. *Chlamydia trachomatis* load at matched anatom‐ ic sites: implications for screening strategies. J Clin Microbiol 2007;45:1395e402 [274] West ES, Munoz B, Mkocha H, et al. Mass treatment and the effect on the load of *Chlamydia trachomatis* infection in a trachoma-hyperendemic community.Invest Oph‐
[275] Westrom LV. Sexually transmitted diseases and infertility. Sex Transm Dis
[276] Centers for Disease control and Prevention, 2011 Sexually transmitted Disease sur‐
veillance. Atlanta, GA: US Department of Health and Human Services
*Chlamydia* and the development of chronic inflammation and disease.
asymptomatic rectal *Chlamydia*? Int J STD AIDS 2011;22:478e80.
gen detection, and serology. Fertil Steril 1991; 55:304–310
community setting. J Clin Microbiol 2009;47:1824e9
in adolescent women. J Infect Dis 2010; 201:42–51
Engl J Med. 2005 Feb 17;352(7):676-85.
Transm Infect 88: 154–156
Fertil Steril 1993; 59:45–50
Trends Microbiol 2: 94–98
thalmol Vis Sci 2005;46:83–7
1994;21:S32–7
Obstet Gynecol 1994; 171:95– 101
Microbiology. 2010 September; 5(9): 1427-1442
[263] Batteiger BE, Tu W, Ofner S, et al. Repeated *Chlamydia trachomatis* genital infections in adolescent women. J Infect Dis 2010; 201:42–51
[250] Cohen CR, Koochesfahani KM, Meier AS, et al. Immunoepidemiologic profile of *Chlamydia trachomatis* infection: importance of heat-shock protein 60 and interferon-
[251] O'Connell, C. M., AbdelRahman, Y. M., Green, E., Darville, H. K., Saira, K., Smith, B., et al. (2011). Toll-like receptor 2 activation by *Chlamydia trachomatis* is plasmid de‐ pendent, and plasmid-responsive chromosomal loci are coordinately regulated in re‐ sponse to glucose limitation by C. trachomatis but not by C. muridarum. *Infection and*
[252] Frazer, L. C., Darville, T., Chandra-Kuntal, K., Andrews, C. W., Zurenski, M., Mintus, M., et al. (2012). Plasmid-cured *Chlamydia* caviae activates TLR2-dependent signaling and retains virulence in the guinea pig model of genital tract infection. *PloS one*, *7*(1),
[253] Kari, L., M.M. Goheen, L.B. Randall, L.D. Taylor, J.H. Carlson, W.M. Whitmire, D. Vi‐ rok, K. Rajaram, V. Endresz, G. McClarty, et al. 2011. Generation of targeted *Chlamy‐*
[254] Zimmerman HL, Potterat JJ, Dukes RL, Muth JB, Zimmerman HP, Fogle JS, Pratts CI. Epidemiologic differences between *Chlamydia* and gonorrhea. Am J Public Health.
[255] Stamm WE, Wagner KF, Amsel R, Alexander ER, Turck M, Counts GW, Holmes KK. Causes of the acute urethral syndrome in women. N Engl J Med.;303(8):409-15(1980)
[256] Centers for Disease Control and Prevention. Sexually transmitted Disease Treatment
[257] Lau CY, Qureshi AK. Azithromycin versus doxycycline for genital Chlamydial infec‐ tions: a meta-analysis of randomized clinical trials. Sex Transm Dis ;29:497e502 (2002)
[258] Centers for Disease Control and Prevention. Sexually Transmitted Diseases Treat‐ ment Guidelines. 2010. http://www.cdc.gov/std/treatment/2010/STD- Treat‐
[259] Horner P, Boag F. 2006 UK National Guideline for the Management of Genital Tract Infection with *Chlamydia trachomatis*. 2006. http://www.bashh.org/guidelines
[260] Handsfield HH. Questioning Azithromycin for Chlamydial Infection. Sex Transm
[261] Hosenfeld CB, Workowski KA, Berman S, Zaidi A, Dyson J, Mosure D, Bolan G, Ba‐ uer HM. Repeat infection with *Chlamydia* and gonorrhea among females: a systematic review of the literature.Sex Transm Dis. 2009 Aug;36(8):478-89. doi: 10.1097/OLQ.
[262] Gaydos CA. Nucleic acid amplification tests for gonorrhea and *Chlamydia*: Practice
and applications. Infect Dis Clin North Am 2005; 19:367–386
*dia trachomatis* null mutants. *Proc. Natl. Acad. Sci. USA.* 108:7189–7193
Guideline.Morbid Mortal Wkly Rep.;55:79-85 (2010)
gamma. J Infect Dis 2005; 192(4):591–599
*immunity*, *79*(3), 1044–1056.
80(11):1338-42 (1990)
ment-2010-RR5912.pdf
Dis 2011;38: 1028–1029.
0b013e3181a2a933.
e30747.
220 Immune Response Activation
[277] Centers for Disease control and Prevention, Sexually transmitted Disease surveil‐ lance 2008. Atlanta, GA: US Department of Health and Human Services 2009
[292] Starnbach MN, Loomis WP, Ovendale P, Regan D, Hess B, Alderson MR, Fling SP. An inclusion membrane protein from *Chlamydia trachomatis* enters the MHC class I pathway and stimulates a CD8+ T cell response. J Immunol. 2003 Nov 1;171(9):4742-9
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
223
[293] Karunakaran KP, Rey-Ladino J, Stoynov N, Berg K, Shen C, Jiang X, Gabel BR, Yu H, Foster LJ, Brunham RC. Immunoproteomic discovery of novel T cell antigens from the obligate intracellular pathogen *Chlamydia*. J Immunol. 2008 Feb 15;180(4):2459-65
[294] Brunham, R.C. and R.W. Peeling, *Chlamydia trachomatis* antigens: role in immunity
[295] Sanchez-Campillo, M., et al., Identification of immunoreactive proteins of *Chlamydia trachomatis* by Western blot analysis of a two-dimensional electrophoresis map with
[296] Taylor, H.R. and R.A. Prendergast, Attempted oral immunization with Chlamydial lipopolysaccharide subunit vaccine., in Invest Ophthalmol Vis Sci. 1987.
[297] Pal S, Theodor I, Peterson EM, de la Maza LM. Immunization with the *Chlamydia tra‐ chomatis* mouse pneumonitis major outer membrane protein can elicit a protective immune response against a genital challenge. Infect Immun, 2001. 69(10): p. 6240-7.
[298] Pal, S., E.M. Peterson, and L.M. de la Maza, Vaccination with the *Chlamydia trachoma‐ tis* major outer membrane protein can elicit an immune response as protective as that resulting from inoculation with live bacteria. Infect Immun, 2005. 73(12): p. 8153-60.
[299] Caldwell, H.D. and R.C. Judd, Structural analysis of Chlamydial major outer mem‐
[300] Caldwell, H.D. and J. Schachter, Antigenic analysis of the major outer membrane
[301] Crane DD, Carlson JH, Fischer ER, Bavoil P, Hsia RC, Tan C, et al. *Chlamydia tracho‐ matis* polymorphic membrane protein D is a species-common pan-neutralizing anti‐
[302] Yu H, Jiang X, Shen C, Karunakaran KP, Brunham RC. Novel *Chlamydia* muridarum T cell antigens induce protective immunity against lung and genital tract infection in
[303] Tan C, Hsia RC, Shou H, Haggerty CL, Ness RB, Gaydos CA, et al. *Chlamydia tracho‐ matis*-infected patients display variable antibody profiles against the nine-member polymorphic membrane protein family. Infect Immun 2009;77 (August (8)):3218–26
[304] Murphey C, Murthy AK, Meier PA, Neal Guentzel M, Zhong G, Arulanandam BP The protective efficacy of Chlamydial protease-like activity factor vaccination is de‐ pendent upon CD4+ T cells. Cell Immunol. 2006 Aug;242(2):110-7. Epub 2006 Nov 20.
[305] Meoni E, Faenzi E, Frigimelica, Zedda L,Skibinski Giovinazzi S,Bonci A,Petracca R,1 Bartolini E, Galli G, Agnusdei M, Nardelli F, Buricchi F, Norais N, Ferlenghi I, Dona‐
and pathogenesis, in Infectious agents and disease. 1994. p. 218-33.
patient sera, in Electrophoresis. 1999. p. 2269-79.
brane proteins. Infect Immun, 1982. 38(3): p. 960-8.
protein of *Chlamydia* spp, in Infect Immun. 1982. p. 1024-31.
gen. Proc Natl Acad Sci U S A 2006;103(February (6)):1894–9
murine models. J Immunol 2009;182(February (3)):1602–8
[292] Starnbach MN, Loomis WP, Ovendale P, Regan D, Hess B, Alderson MR, Fling SP. An inclusion membrane protein from *Chlamydia trachomatis* enters the MHC class I pathway and stimulates a CD8+ T cell response. J Immunol. 2003 Nov 1;171(9):4742-9
[277] Centers for Disease control and Prevention, Sexually transmitted Disease surveil‐ lance 2008. Atlanta, GA: US Department of Health and Human Services 2009
[278] Health protection agency, Sexually transmitted infections and young people in the
[279] European Centre for Disease Prevention and Control. *Chlamydia* Control in Europe.
[280] British Columbia Centre for Disease Control. STI Prevention & Control: Annual Re‐
[281] 2008 Annual Surveillance Report: HIV/AIDS, Viral Hepatitis and Sexually Transmis‐
[282] Brunham RC, Pourbohloul B, Mak S, White R, Rekart ML (2005) The unexpected im‐ pact of a *Chlamydia trachomatis* infection control program on susceptibility to reinfec‐
[283] Farris CM, Morrison RP.Vaccination against *Chlamydia* genital infection utilizing the murine C. muridarum model. Infect Immun. 2011 Mar;79(3):986-96. doi: 10.1128/IAI.
[284] Longbottom. Chlmaydial vaccine development. J Med Microbiology 2003;52(July (Pt
[285] Mitzel, Vaccination against feline pneumonitis. Am J Vet Res.1977 Sep;38(9):1361-3
[286] Grayston JT, Woolridge RL, Wang S. Trachoma Vaccine studies in Taiwan. Ann. N.Y.
[287] Poland GA, Kennedy RB, Ovsyannikova IG (2011) Vaccinomics and Personalized Vaccinology: Is Science Leading Us Toward a New Path of Directed Vaccine Devel‐
[288] Caldwell HD, Ku CC, Kenny GE. Antigenic analysis of Chlamydiae by two-dimen‐ sional immunoelectrophoresis. II A trachoma-LGV-specific antgen. Journal of Immu‐
[289] Sanchez-Campillo M, Bini L, ComanducciM et al. Identificationof immunoreactive proteins of a two dimensional electroporesis map with patient sera. Electrophoresis
[290] Murphey C, Murthy AK, Meier PA, Neal Guentzel M, Zhong G, Arulanandam BP. The protective efficacy of Chlamydial protease-like activity factor for vaccination is
[291] Stemke-Hale K, Kaltenboeck B, Degraves FJ at al. Screening the whole genomeof a pthogen in vivo for individual protective antigens (2005) Vaccine 23(23), 3016-3025
opment and Discovery? PLoS Pathog 7(12): e1002344
dependent on CD4+ T cells. (2006).242, 110-117
united kingdom: 2008 Report, July 2008
ports 1997 – 2011, March 2013
sible Infections in Australia
tion. J Infect Dis 192: 1836–1844
00881-10. Epub 2010 Nov 15
Acad. Sci. (1962) 98, 352-367
nology (1975). 115, 969-975
(1999). 20(11), 2269-2279
7)):537-40
Stockholm 2009
222 Immune Response Activation
ti, Cevenini R, Finco O,Grandi G, and Grifantini R. CT043, a Protective Antigen That Induces a CD4+ Th1 Response during *Chlamydia trachomatis* Infection in Mice and Humans. Infect Immun. 2009 September; 77(9): 4168–4176
cross-species immunity against genital *Chlamydia* muridarum challenge in mice.
Immune Regulation of *Chlamydia trachomatis* Infections of the Female Genital Tract
http://dx.doi.org/10.5772/57542
225
[318] Marks E, Helgeby A, Andersson JO, Schön K, Lycke NY (2011) CD4⁺ T-cell immunity in the female genital tract is critically dependent on local mucosal immunization. Eur
[319] R. T. Gray, K. W. Beagley, P. Timms, and D. P. Wilson, "Modeling the impact of po‐ tential vaccines on epidemics of sexually transmitted *Chlamydia trachomatis* infec‐
[320] Carey A, Cunningham K, Andrew D, Hafner L, Timms P, Beagley K. A comparison of the effects of a Chlamydial vaccine administered during or after a C. muridarum
tion," Journal of Infectious Diseases, vol. 199, no. 11, pp. 1680–1688, 2009
urogenital infection of female mice. Vaccine 2011;29: 6505–13
PLoS One. 2013 May 31;8(5):e64917. doi: 10.1371/journal.pone.0064917.
J Immunol 41: 2642–2653)
cross-species immunity against genital *Chlamydia* muridarum challenge in mice. PLoS One. 2013 May 31;8(5):e64917. doi: 10.1371/journal.pone.0064917.
[318] Marks E, Helgeby A, Andersson JO, Schön K, Lycke NY (2011) CD4⁺ T-cell immunity in the female genital tract is critically dependent on local mucosal immunization. Eur J Immunol 41: 2642–2653)
ti, Cevenini R, Finco O,Grandi G, and Grifantini R. CT043, a Protective Antigen That Induces a CD4+ Th1 Response during *Chlamydia trachomatis* Infection in Mice and
[306] Sharma J, Bosnic AM, Piper JM, Zhong G. Human Antibody Responses to a *Chlamy‐ dia*-Secreted Protease Factor. Infection and Immunity, Dec. 2004, p. 7164–7171
[307] He Q, Martinez-Sobrido L, Eko FO, Palese P, Garcia-Sastre A, Lyn D, et al. Live-atte‐ nuated influenza viruses as delivery vectors for *Chlamydia* vaccines. Immunology
[308] Karunakaran KP, Yu H, Foster LJ, Brunham RC. Development of a *Chlamydia tracho‐*
[309] Dong-Ji Z, Yang X, Shen C, Lu H, Murdin A, Brunham RC. Priming with *Chlamydia trachomatis* major outer membrane protein (MOMP) DNA fol- lowed by MOMP IS‐ COM boosting enhances protection and is associated with increased immunoglobu‐
[310] Hansen J, Jensen KT, Follmann F, Agger EM, Theisen M, Andersen P. Liposome de‐ livery of *Chlamydia* muridarum major outer membrane protein primes a Th1 re‐ sponse that protects against genital Chlamydial infection in a mouse model. J Infect
[311] Pal S, Peterson EM, Rappuoli R, Ratti G, de la Maza LM. Immunization with the *Chlamydia trachomatis* major outer membrane protein, using adjuvants developed for human vaccines, can induce partial protection in a mouse model against a genital
[312] Hickey DK, Aldwell FE, Beagley KW (2010) Oral immunization with a novel lipidbased adjuvant protects against genital *Chlamydia* infection. Vaccine 28: 1668–1672
[313] Hickey DK, Aldwell FE, Beagley KW (2009) Transcutaneous immunization with a novel lipid-based adjuvant protects against *Chlamydia* genital and respiratory infec‐
[314] Macmillan L, Ifere GO, He Q, Igietseme JU, Kellar KL, Okenu DM, et al. A recombi‐ nant multivalent combination vaccine protects against *Chlamydia* and genital herpes.
[315] Hickey DK, Bao S, Ikeda LT, Carey AJ, Beagley KW (2005) Induction of anti-Chlamy‐ dial mucosal immunity by transcutaneous immunization is enhanced by topical ap‐
[316] Eko FO, Ekong E, He Q, Black CM, Igietseme JU (2011) Induction of immune memo‐
[317] Manam S, Chaganty BK, Evani SJ, Zafiratos MT, Ramasubramanian AK, Arulanan‐ dam BP, Murthy AK. Intranasal vaccination with *Chlamydia* pneumoniae induces
ry by a multisubunit Chlamydial vaccine. Vaccine 29: 1472–1480
lin A and Th1 cellular immune responses. Infect Immun 2000;68:3074–8
Humans. Infect Immun. 2009 September; 77(9): 4168–4176
*matis* T cell vaccine. Hum Vaccin 2010;6:676–80
2007;122:28–37
224 Immune Response Activation
Dis 2008;198:758–67
challenge. Vaccine 2006;24:766–75
tions. Vaccine 27: 6217–6225
FEMS Immunol Med Microbiol 2007;49:46–55
plication of GM-CSF. Curr Mol Med 5: 599–605).
**Chapter 8**
**The Roles of Invariant NKT Cells in Bowel Immunity —**
The roles of natural killer (NK)T cells in intestinal immunity have not been sufficiently investigated. The bowel possesses its own unique mucosal immune system, the gut-associated lymphoepithelial tract (GALT). In this review, we focused on CD1d-independent invariant type NKT (iNKT) cells as modulators of GALT. iNKT cells have a restricted invariant T-cell receptor (TCR) Vα24 chain paired with Vβ11 chain. In addition to T helper (Th)1 cytokines, such as IFN-γ, iNKT cells can produce anti-inflammatory Th2 cytokines, such as interleukin
Although invariant NKT cells are rare in normal small-intestine mucosa, they have been observed in intestinal allografts during rejection. Infiltrating iNKT cells release IL-4 and IL-5, Th2-related cytokines that antagonize the Th1 responses that induce acute cellular rejection. Also, iNKT cells form an immunological barrier against parasite infection. We found that CD1d +cells are actually localized in the lamina propria of the villi in the human intestine and they
On the other hand, a small number of invariant NKT cells are resident in the normal colorectal mucosa. iNKT cells are involved in defense against colorectal tumor progression and meta‐ stasis through the apoptosis-inducing molecule Fas ligand (FasL). Their numbers increase markedly in colorectal carcinomas. Increased iNKT cell infiltration in colorectal carcinomas is an independent favorable prognostic factor. iNKT cells probably play an important role in the
> © 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
**Suppression of Tumor Progression and Rejection of**
**Intestinal Transplants**
http://dx.doi.org/10.5772/57588
**1. Introduction**
(IL)-4.
Tatsuaki Tsuruyama and Wulamujiang Aini
Additional information is available at the end of the chapter
may present the antigen for the recruited iNKT cells.
Thus, iNKT cells are modulators of bowel mucosal immunity.
pathogenesis of ulcerative colitis.
## **The Roles of Invariant NKT Cells in Bowel Immunity — Suppression of Tumor Progression and Rejection of Intestinal Transplants**
Tatsuaki Tsuruyama and Wulamujiang Aini
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/57588
#### **1. Introduction**
The roles of natural killer (NK)T cells in intestinal immunity have not been sufficiently investigated. The bowel possesses its own unique mucosal immune system, the gut-associated lymphoepithelial tract (GALT). In this review, we focused on CD1d-independent invariant type NKT (iNKT) cells as modulators of GALT. iNKT cells have a restricted invariant T-cell receptor (TCR) Vα24 chain paired with Vβ11 chain. In addition to T helper (Th)1 cytokines, such as IFN-γ, iNKT cells can produce anti-inflammatory Th2 cytokines, such as interleukin (IL)-4.
Although invariant NKT cells are rare in normal small-intestine mucosa, they have been observed in intestinal allografts during rejection. Infiltrating iNKT cells release IL-4 and IL-5, Th2-related cytokines that antagonize the Th1 responses that induce acute cellular rejection. Also, iNKT cells form an immunological barrier against parasite infection. We found that CD1d +cells are actually localized in the lamina propria of the villi in the human intestine and they may present the antigen for the recruited iNKT cells.
On the other hand, a small number of invariant NKT cells are resident in the normal colorectal mucosa. iNKT cells are involved in defense against colorectal tumor progression and meta‐ stasis through the apoptosis-inducing molecule Fas ligand (FasL). Their numbers increase markedly in colorectal carcinomas. Increased iNKT cell infiltration in colorectal carcinomas is an independent favorable prognostic factor. iNKT cells probably play an important role in the pathogenesis of ulcerative colitis.
Thus, iNKT cells are modulators of bowel mucosal immunity.
© 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
### **2. Background**
#### **2.1. Gut-associated lymphoepithelial tract (GALT)**
The bowel possesses a distinct mucosal immune system, known as the gut-associated lym‐ phoepithelial tract (GALT), which is comprised of high endothelial venules (HEVs) located in the inter-follicular region (IFR) around the Peyer's patches (PPs). GALT-associated immunity involves an exquisite balance between activation and suppression mediated through the release of cytokines [1] (Figure 1). The GALT contains the largest collection of lymphocytes in the body. Physiologic and barrier functions of GALT against bacterial infection are associated with mucosal surfaces and are essential; therefore, only limited damage can be tolerated. The mucosa is continuously exposed to a vast array of antigens, necessitating the distinguishing of self from non-self/infectious antigens. Infectious agents, malignant tumors such as adeno‐ carcinomas, and immune and inflammatory bowel diseases of the gut all contribute to morbidity and mortality.
The lymphocytes in the bowel are believed to play pivotal roles in bowel immunity. Two types of lymphocytes are found in the bowel: intra-epithelial lymphocytes (IELs) and lamina propria (LP) lymphocytes (LPLs). In addition, the PPs are comprised of B cells and T cells, unique lymphocyte subsets that interact with M epithelial cells. Antigens are taken up by the PPs through the M-cell layer, inducing an immune reaction (Figure 1).
In addition, lymphocytes from the GALT circulate in the common mucosal immune system, including the mesenteric lymph nodes.
LPLs play important roles in immune homeostasis. Although these cells function in mucosal defense against pathogens, they ignore non-pathogenic bacteria and foods, a phenomenon known as oral tolerance. LPLs also play a supporting role in barrier defenses and are involved in the pathogenesis of inflammatory intestinal diseases. At the onset of Crohn's disease, mucosal T cells appear to mount a typical T helper (Th)1 response that resembles an acute infectious process, but this response is lost with progression to later stages of Crohn's disease. Such polarization of bowel cytokines is believed to be responsible for inflammatory bowel diseases (IBD).
**2.3. Invariant natural killer T (iNKT) cells**
the lamina propria by release of cytokines such as IL-4, IL-5, IFN-γ.
immunological environment.
rare in the small intestine [10].
Among these NK cells, iNKT cells have a very restricted TCR repertoire, which in human peripheral blood consists of an invariant Vα24-Ja18 chain (formerly Vα24-JaQ) paired with a Vβ11 chain [4, 6]. Therefore, this classical type is termed "invariant" NKT (iNKT) cells (Figure 2). Human iNKT cells can be activated by glycolipid antigens such as beta-galactosylceramide and iGb3, which are presented by CD1d expressed on dendritic cells [7]. When activated, iNKT cells immediately produce pro-inflammatory Th1 cytokines such as IFN-γ and tumor necrosis factor (TNF)-α as well as anti-inflammatory Th2 cytokines, such as IL-4, IL-10, and IL-13. Thus, iNKT cells are capable of bidirectional differentiation, which is probably directed by the
**Figure 1.** GALT and iNKT cells. M cells and mucosal epithelial cells cover the lamina propria of the bowel. HEV: high endothelial venules, Th1 & Th2, helper T cell Type 1 & 2. iNKT cells relate to the development of B cells and T cells in
The Roles of Invariant NKT Cells in Bowel Immunity — Suppression of Tumor Progression and…
http://dx.doi.org/10.5772/57588
229
iNKT cells are localized primarily in the thymus, bone marrow, liver, and spleen, and they are rare in the lymph nodes [8]. Bannai et al. [9] reported that mouse colon contains a considerable number of NKT cells. More recently, O'Keeffe et al. reported that T cells expressing CD161+(a phenotypic marker of NK cells) are the major natural killer receptor–positive cell population in the intestine. Although the intestines harbor a diverse array of NKT cells, iNKT cells are
#### **2.2. Natural killer T cells**
NKT cells are associated with both innate and adaptive immunity due to their ability to interact with a broad spectrum of antigens.
NKT cells are currently classified into three types classical NKT cells (non-classical NKT cells); and NKT-like cells, or CD1d-independent NK1.1+T cells [2]. Classical NKT cells express CD161 [2, 3]. Human classical NKT cells express an invariant TCR, Vα24, whereas mouse NKT cells express TCRVα14 [4, 5]. In contrast, the T-cell receptor V regions expressed by non-classical NKT cells and NKT-like cells are highly diverse. CD56-NKT cells are believed to function as effector or regulatory cells in innate immunity.
The Roles of Invariant NKT Cells in Bowel Immunity — Suppression of Tumor Progression and… http://dx.doi.org/10.5772/57588 229
**Figure 1.** GALT and iNKT cells. M cells and mucosal epithelial cells cover the lamina propria of the bowel. HEV: high endothelial venules, Th1 & Th2, helper T cell Type 1 & 2. iNKT cells relate to the development of B cells and T cells in the lamina propria by release of cytokines such as IL-4, IL-5, IFN-γ.
#### **2.3. Invariant natural killer T (iNKT) cells**
**2. Background**
228 Immune Response Activation
morbidity and mortality.
diseases (IBD).
**2.2. Natural killer T cells**
with a broad spectrum of antigens.
effector or regulatory cells in innate immunity.
including the mesenteric lymph nodes.
**2.1. Gut-associated lymphoepithelial tract (GALT)**
The bowel possesses a distinct mucosal immune system, known as the gut-associated lym‐ phoepithelial tract (GALT), which is comprised of high endothelial venules (HEVs) located in the inter-follicular region (IFR) around the Peyer's patches (PPs). GALT-associated immunity involves an exquisite balance between activation and suppression mediated through the release of cytokines [1] (Figure 1). The GALT contains the largest collection of lymphocytes in the body. Physiologic and barrier functions of GALT against bacterial infection are associated with mucosal surfaces and are essential; therefore, only limited damage can be tolerated. The mucosa is continuously exposed to a vast array of antigens, necessitating the distinguishing of self from non-self/infectious antigens. Infectious agents, malignant tumors such as adeno‐ carcinomas, and immune and inflammatory bowel diseases of the gut all contribute to
The lymphocytes in the bowel are believed to play pivotal roles in bowel immunity. Two types of lymphocytes are found in the bowel: intra-epithelial lymphocytes (IELs) and lamina propria (LP) lymphocytes (LPLs). In addition, the PPs are comprised of B cells and T cells, unique lymphocyte subsets that interact with M epithelial cells. Antigens are taken up by the PPs
In addition, lymphocytes from the GALT circulate in the common mucosal immune system,
LPLs play important roles in immune homeostasis. Although these cells function in mucosal defense against pathogens, they ignore non-pathogenic bacteria and foods, a phenomenon known as oral tolerance. LPLs also play a supporting role in barrier defenses and are involved in the pathogenesis of inflammatory intestinal diseases. At the onset of Crohn's disease, mucosal T cells appear to mount a typical T helper (Th)1 response that resembles an acute infectious process, but this response is lost with progression to later stages of Crohn's disease. Such polarization of bowel cytokines is believed to be responsible for inflammatory bowel
NKT cells are associated with both innate and adaptive immunity due to their ability to interact
NKT cells are currently classified into three types classical NKT cells (non-classical NKT cells); and NKT-like cells, or CD1d-independent NK1.1+T cells [2]. Classical NKT cells express CD161 [2, 3]. Human classical NKT cells express an invariant TCR, Vα24, whereas mouse NKT cells express TCRVα14 [4, 5]. In contrast, the T-cell receptor V regions expressed by non-classical NKT cells and NKT-like cells are highly diverse. CD56-NKT cells are believed to function as
through the M-cell layer, inducing an immune reaction (Figure 1).
Among these NK cells, iNKT cells have a very restricted TCR repertoire, which in human peripheral blood consists of an invariant Vα24-Ja18 chain (formerly Vα24-JaQ) paired with a Vβ11 chain [4, 6]. Therefore, this classical type is termed "invariant" NKT (iNKT) cells (Figure 2). Human iNKT cells can be activated by glycolipid antigens such as beta-galactosylceramide and iGb3, which are presented by CD1d expressed on dendritic cells [7]. When activated, iNKT cells immediately produce pro-inflammatory Th1 cytokines such as IFN-γ and tumor necrosis factor (TNF)-α as well as anti-inflammatory Th2 cytokines, such as IL-4, IL-10, and IL-13. Thus, iNKT cells are capable of bidirectional differentiation, which is probably directed by the immunological environment.
iNKT cells are localized primarily in the thymus, bone marrow, liver, and spleen, and they are rare in the lymph nodes [8]. Bannai et al. [9] reported that mouse colon contains a considerable number of NKT cells. More recently, O'Keeffe et al. reported that T cells expressing CD161+(a phenotypic marker of NK cells) are the major natural killer receptor–positive cell population in the intestine. Although the intestines harbor a diverse array of NKT cells, iNKT cells are rare in the small intestine [10].
The LP, located beneath the mucosa, includes various types of immune cells that form the GALT. The PP is a predominant immune apparatus that is covered by M cells and is capable of controlling mucosal immunity. In addition to the lymph apparatus and lymphocytes, the intestinal immune system contains dendritic cells (DCs) that interact with goblet cells. Mucus sensing enhances DC tolerogenic functions, including the release of TGF-β and suppressive IL-10, which are factors that induce FoxP3+regulatory T (Treg) cell production. Populations of CD3+TCR-γδ cells are also observed [29]. CD56+NK cells are classically viewed as lymphocytes that provide innate immunity against virus-infected cells and tumor cells through the release
The Roles of Invariant NKT Cells in Bowel Immunity — Suppression of Tumor Progression and…
http://dx.doi.org/10.5772/57588
231
Intestinal iNKT cells are rare in the normal status [10], but small population is observed in the villi of the small intestine [30]. Although they have not been considered as minor participants in the control of GALT. However, iNKT cells are found to decrease in the celiac disease without affecting development of classical T cells. Coeliac disease is one of the autoimmune disorders of the small intestine. This disease occurs in genetically predisposed people of all ages from middle infancy. Symptoms include pain, chronic constipation and diarrhea in children. Diagnoses are occasionally being made in asymptomatic persons in the screening examination. Importantly, coeliac disease is not allergic disease, but caused by a reaction to a gluten protein in wheat. When the small intestine is exposed to gluten, the tissue transglutaminase modifies the protein, and the GALT autoimmunologically cross-reacts with the small intestinal tissue, causing an inflammatory reaction. As a result, villous atrophy is observed histologically and endoscopically. The absorption of nutrients is significantly affected. The effective treatment is a lifelong gluten-free diet. The deficiency of Vα24+ cells in coeliac disease was independent of age, gluten status of diet or duration of gluten‐free diet. Probably, T cells are selected by
Intestinal transplantation is a commonly accepted standard therapy for patients with irrever‐ sible parenteral nutrition complications associated with short-bowel syndrome, [31-33], Hirschsprung and related diseases, chronic idiopathic intestinal pseudo-obstruction syn‐ drome, and megacystis-microcolon-intestinal hypoperistalsis syndrome [34, 35]. Acute cellular rejection (ACR) remains the major cause of intestinal graft failure after transplantation. Diagnosis of ACR requires detection by histopathologic evaluation [35-38] as well as clinical
Pathologists have diagnosed acute rejection based on the finding of crypt apoptosis in the intestinal mucosa upon histologic examination of the graft [36-38], rather than lymphocytic
of cytolytic mediators and IFN-γ.
**3.2. iNKT cells in the small intestine and coeliac disease**
glycolipids on CD1d expressing CD4+CD8+thymocytes [30].
*3.3.2. Histologic criteria for diagnosis of ACR of small-intestine transplants*
**3.3. Small-intestine allograft immunity**
examination and laboratory data [38].
*3.3.1. Background on intestinal transplantation*
**Figure 2.** T, invariant natural killer (NK)T, and NK cells. The phenotypic surface antigens are shown. The T cell receptor (TCR) is variable in T cells; the TCR on the invariant type of human NKT cells is restricted to the Vα24β11 type. NK cells do not carry a TCR and express CD161, CD56, or CD57 NK markers, which are also expressed by NKT cells.
#### **2.4. Invariant natural killer T (iNKT) cells in diseases**
iNKT cells have been implicated as playing roles in infection control [11, 12], inflammatory bowel diseases [13], abortion [14], transplantation tolerance [15, 16], suppression of autoim‐ mune diseases [17-20], psoriasis and atopic dermatitis [21], and the regulation of allergic disorders [22, 23]. iNKT cells have the potential to activate or suppresses immune system by release of variable cytokines including IFN-γ, IL-4, IL-10, and IL-17.
Moreover, a number of reports have suggested that iNKT cells play protective as well as tolerogenic roles in tumor immunity [24-29]. In addition, infiltration of iNKT cells is a prog‐ nostic indicator for colorectal cancer metastasis [29].
#### **3. iNKT cells in the small intestine**
#### **3.1. Mucosal immunity of the small intestine**
The mucosal immune system of the small intestine is unique. Mucosal barrier damage and inflammation have been associated with high levels of flagellin protein in the lumen. One of the roles of intestinal immunity is defense against pathogens within the bacterial flora. The mucosa is covered by mucin, which is one of the central determinants of gut immune specificity and immune tolerance.
The LP, located beneath the mucosa, includes various types of immune cells that form the GALT. The PP is a predominant immune apparatus that is covered by M cells and is capable of controlling mucosal immunity. In addition to the lymph apparatus and lymphocytes, the intestinal immune system contains dendritic cells (DCs) that interact with goblet cells. Mucus sensing enhances DC tolerogenic functions, including the release of TGF-β and suppressive IL-10, which are factors that induce FoxP3+regulatory T (Treg) cell production. Populations of CD3+TCR-γδ cells are also observed [29]. CD56+NK cells are classically viewed as lymphocytes that provide innate immunity against virus-infected cells and tumor cells through the release of cytolytic mediators and IFN-γ.
#### **3.2. iNKT cells in the small intestine and coeliac disease**
Intestinal iNKT cells are rare in the normal status [10], but small population is observed in the villi of the small intestine [30]. Although they have not been considered as minor participants in the control of GALT. However, iNKT cells are found to decrease in the celiac disease without affecting development of classical T cells. Coeliac disease is one of the autoimmune disorders of the small intestine. This disease occurs in genetically predisposed people of all ages from middle infancy. Symptoms include pain, chronic constipation and diarrhea in children. Diagnoses are occasionally being made in asymptomatic persons in the screening examination. Importantly, coeliac disease is not allergic disease, but caused by a reaction to a gluten protein in wheat. When the small intestine is exposed to gluten, the tissue transglutaminase modifies the protein, and the GALT autoimmunologically cross-reacts with the small intestinal tissue, causing an inflammatory reaction. As a result, villous atrophy is observed histologically and endoscopically. The absorption of nutrients is significantly affected. The effective treatment is a lifelong gluten-free diet. The deficiency of Vα24+ cells in coeliac disease was independent of age, gluten status of diet or duration of gluten‐free diet. Probably, T cells are selected by glycolipids on CD1d expressing CD4+CD8+thymocytes [30].
#### **3.3. Small-intestine allograft immunity**
**Figure 2.** T, invariant natural killer (NK)T, and NK cells. The phenotypic surface antigens are shown. The T cell receptor (TCR) is variable in T cells; the TCR on the invariant type of human NKT cells is restricted to the Vα24β11 type. NK cells
iNKT cells have been implicated as playing roles in infection control [11, 12], inflammatory bowel diseases [13], abortion [14], transplantation tolerance [15, 16], suppression of autoim‐ mune diseases [17-20], psoriasis and atopic dermatitis [21], and the regulation of allergic disorders [22, 23]. iNKT cells have the potential to activate or suppresses immune system by
Moreover, a number of reports have suggested that iNKT cells play protective as well as tolerogenic roles in tumor immunity [24-29]. In addition, infiltration of iNKT cells is a prog‐
The mucosal immune system of the small intestine is unique. Mucosal barrier damage and inflammation have been associated with high levels of flagellin protein in the lumen. One of the roles of intestinal immunity is defense against pathogens within the bacterial flora. The mucosa is covered by mucin, which is one of the central determinants of gut immune specificity
do not carry a TCR and express CD161, CD56, or CD57 NK markers, which are also expressed by NKT cells.
**2.4. Invariant natural killer T (iNKT) cells in diseases**
nostic indicator for colorectal cancer metastasis [29].
**3. iNKT cells in the small intestine**
**3.1. Mucosal immunity of the small intestine**
and immune tolerance.
230 Immune Response Activation
release of variable cytokines including IFN-γ, IL-4, IL-10, and IL-17.
#### *3.3.1. Background on intestinal transplantation*
Intestinal transplantation is a commonly accepted standard therapy for patients with irrever‐ sible parenteral nutrition complications associated with short-bowel syndrome, [31-33], Hirschsprung and related diseases, chronic idiopathic intestinal pseudo-obstruction syn‐ drome, and megacystis-microcolon-intestinal hypoperistalsis syndrome [34, 35]. Acute cellular rejection (ACR) remains the major cause of intestinal graft failure after transplantation. Diagnosis of ACR requires detection by histopathologic evaluation [35-38] as well as clinical examination and laboratory data [38].
#### *3.3.2. Histologic criteria for diagnosis of ACR of small-intestine transplants*
Pathologists have diagnosed acute rejection based on the finding of crypt apoptosis in the intestinal mucosa upon histologic examination of the graft [36-38], rather than lymphocytic infiltrates. The number of apoptotic bodies is certainly augmented during the acute rejection; however, the mechanism underlying this phenomenon has not been determined.
Mucosal damage cannot be prevented once crypt apoptosis has become evident [38]. It is possible that apoptosis of crypt cells is induced by direct attack by cytotoxic T-lymphocytes (CTLs), but such a finding has not been reported. Therefore, it remains unclear whether CTL infiltration is necessary for ACR. Humoral factors may also play a role in crypt apoptosis.
Although there are reports describing T-cell apoptosis in the intestine in ongoing ACR, its clinical significance has not been determined [39]. T cells interact with Fas on other T cells to induce their apoptosis [40]. The apoptotic T cells are then phagocytosed by macrophages in the small-intestine allografts. Similar findings in a liver allograft has been reported by us [41]. Thus, these phagocytic findings may be common to various allografts in ACR. The phagocy‐ tosing macrophages form granulomas. Furthermore, apoptotic bodies are also positive for Fas, which is a surface antigen of activated T lymphocytes. Treatment with a steroid pulse signif‐ icantly diminishes the apoptotic and phagocytotic responses that are indicative of ACR [42]. Although it is possible that both recipient and donor lymphocytes interact with each other via Fas and FasL molecules to induce apoptosis, a study using animal models reported that TNFα (rather than interaction between Fas and FasL) is the primary factor for induction of lymphocytic apoptosis [39].
#### *3.3.3. iNKT cells infiltrate during acute rejection of the allograft*
Using systematic immunohistochemical analysis of an intestinal allograft, we assessed the behavior of iNKT cells during acute rejection of the intestinal allograft [38]. In practice, the diagnosis of intestinal rejection is difficult, particularly due to the complicated interactions between lymphocytes and other immunological cells resulting from the transient coexistence of donor-derived and recipient-derived cells in the graft.
iNKT cells are transiently recruited to the intestinal LP (Figure 3), but immunosuppressive therapy significantly decreased iNKT cells in number (Figure 4&5). The infiltrating intestinal iNKT cells have the potential to produce IL-4, which antagonizes IFN-γ and contributes to the development of Th2 cells. Higher levels of IL-4 prior to and shortly after kidney transplantation have been reported, and IL-4 may have a protective effect on renal graft survival [43]. Indeed, iNKT cells have been implicated in tolerance in experimental mouse models, in induction of chimerism in allogenic cardiac transplant models [44], and in acceptance of rat-islet xenografts in mice [16]. IFN-γ production remains low relative to normal donor intestine and does not change during the course of ACR development. Therefore, infiltration of iNKT cells implies that they are involved in the response to rejection. Asaoka et al. reported the activation of CTLs in granzyme B/perforin-mediated graft injury [45]. Therefore, release of Th2-related cytokines by iNKT cells may antagonize this activation.
encounters recipient-derived cells. Although, few studies have examined injury to the PPs during rejection, histology of PP showed the hyperplastic change of PP with an increase in
**Figure 4.** Time-course analyses using allograft tissues showed a significant decrease in the number of iNKT cells after immunosuppressant administration. The left photo illustrates the significant decrease in the number of iNKT cells after administration. Immunostaining of iNKT cells at the onset of ACR (left) and 48 hrs after administration of steroids (middle) is shown. Brown cells are iNKT cells in the allograft mucosa. Counterstaining was performed using methylgreen. iNKT cells underwent apoptosis in the graft. The right photo below shows TUNEL-positive (yellow) iNKT cells. Non-iNKT cells were stained red (phycoerythrin-labeled). This photo is the serial section of the left section [39].
(A) (B) (C) (D)
The Roles of Invariant NKT Cells in Bowel Immunity — Suppression of Tumor Progression and…
http://dx.doi.org/10.5772/57588
233
**Figure 3.** Immunohistochemistry of iNKT cells: (A) TCRVα24 (left) and (B) TCRVβ11 stained cells in brown. Fluorescent
immunostaining of TCRVα24 (C), IL-4 (D).
PPs consist primarily of B cells, T cells and DCs. In general, B cells carry out a variety of immune functions, including immunoglobulin production, cytokine release, antigen presentation, regulation of DC activity, and participation in the induction of immune tolerance [46]. A large number of T cells and B cells are present in the intestinal LP. The majority of T cells in the LP express the TCRαβ, and the proportion of CD8+T cells to CD4+T cells ranges from 2:1 to 3:1. The production of IL-4 and IL-5 by CD4+T cells in the LP promotes IgA production in the intestinal tract. Furthermore, CD8+T cells in the LP are responsible for damage by recognizing and activating macrophages. The primary antibodies secreted into the intestinal tract mucus are IgA. A number of IgA-producing cells distribute in the LP, and IgA molecules that are secreted into the intestinal tract are transported to the gut luminal side by binding to multi‐
expression of CD20, a mature B cell marker at the onset of rejection (Figure 6).
meric antibody receptors that are retained on the intestinal epithelial cells.
#### *3.3.4. PPs and iNKT cells in small-intestine allografts*
After transplantation, recipient-derived T cells traffic into the allograft across the HEVs located in the IFR around the PPs. Therefore, the PP serves as the interface through which the allograft The Roles of Invariant NKT Cells in Bowel Immunity — Suppression of Tumor Progression and… http://dx.doi.org/10.5772/57588 233
infiltrates. The number of apoptotic bodies is certainly augmented during the acute rejection;
Mucosal damage cannot be prevented once crypt apoptosis has become evident [38]. It is possible that apoptosis of crypt cells is induced by direct attack by cytotoxic T-lymphocytes (CTLs), but such a finding has not been reported. Therefore, it remains unclear whether CTL infiltration is necessary for ACR. Humoral factors may also play a role in crypt apoptosis.
Although there are reports describing T-cell apoptosis in the intestine in ongoing ACR, its clinical significance has not been determined [39]. T cells interact with Fas on other T cells to induce their apoptosis [40]. The apoptotic T cells are then phagocytosed by macrophages in the small-intestine allografts. Similar findings in a liver allograft has been reported by us [41]. Thus, these phagocytic findings may be common to various allografts in ACR. The phagocy‐ tosing macrophages form granulomas. Furthermore, apoptotic bodies are also positive for Fas, which is a surface antigen of activated T lymphocytes. Treatment with a steroid pulse signif‐ icantly diminishes the apoptotic and phagocytotic responses that are indicative of ACR [42]. Although it is possible that both recipient and donor lymphocytes interact with each other via Fas and FasL molecules to induce apoptosis, a study using animal models reported that TNFα (rather than interaction between Fas and FasL) is the primary factor for induction of
Using systematic immunohistochemical analysis of an intestinal allograft, we assessed the behavior of iNKT cells during acute rejection of the intestinal allograft [38]. In practice, the diagnosis of intestinal rejection is difficult, particularly due to the complicated interactions between lymphocytes and other immunological cells resulting from the transient coexistence
iNKT cells are transiently recruited to the intestinal LP (Figure 3), but immunosuppressive therapy significantly decreased iNKT cells in number (Figure 4&5). The infiltrating intestinal iNKT cells have the potential to produce IL-4, which antagonizes IFN-γ and contributes to the development of Th2 cells. Higher levels of IL-4 prior to and shortly after kidney transplantation have been reported, and IL-4 may have a protective effect on renal graft survival [43]. Indeed, iNKT cells have been implicated in tolerance in experimental mouse models, in induction of chimerism in allogenic cardiac transplant models [44], and in acceptance of rat-islet xenografts in mice [16]. IFN-γ production remains low relative to normal donor intestine and does not change during the course of ACR development. Therefore, infiltration of iNKT cells implies that they are involved in the response to rejection. Asaoka et al. reported the activation of CTLs in granzyme B/perforin-mediated graft injury [45]. Therefore, release of Th2-related cytokines
After transplantation, recipient-derived T cells traffic into the allograft across the HEVs located in the IFR around the PPs. Therefore, the PP serves as the interface through which the allograft
however, the mechanism underlying this phenomenon has not been determined.
lymphocytic apoptosis [39].
232 Immune Response Activation
*3.3.3. iNKT cells infiltrate during acute rejection of the allograft*
of donor-derived and recipient-derived cells in the graft.
by iNKT cells may antagonize this activation.
*3.3.4. PPs and iNKT cells in small-intestine allografts*
**Figure 3.** Immunohistochemistry of iNKT cells: (A) TCRVα24 (left) and (B) TCRVβ11 stained cells in brown. Fluorescent immunostaining of TCRVα24 (C), IL-4 (D).
**Figure 4.** Time-course analyses using allograft tissues showed a significant decrease in the number of iNKT cells after immunosuppressant administration. The left photo illustrates the significant decrease in the number of iNKT cells after administration. Immunostaining of iNKT cells at the onset of ACR (left) and 48 hrs after administration of steroids (middle) is shown. Brown cells are iNKT cells in the allograft mucosa. Counterstaining was performed using methylgreen. iNKT cells underwent apoptosis in the graft. The right photo below shows TUNEL-positive (yellow) iNKT cells. Non-iNKT cells were stained red (phycoerythrin-labeled). This photo is the serial section of the left section [39].
encounters recipient-derived cells. Although, few studies have examined injury to the PPs during rejection, histology of PP showed the hyperplastic change of PP with an increase in expression of CD20, a mature B cell marker at the onset of rejection (Figure 6).
PPs consist primarily of B cells, T cells and DCs. In general, B cells carry out a variety of immune functions, including immunoglobulin production, cytokine release, antigen presentation, regulation of DC activity, and participation in the induction of immune tolerance [46]. A large number of T cells and B cells are present in the intestinal LP. The majority of T cells in the LP express the TCRαβ, and the proportion of CD8+T cells to CD4+T cells ranges from 2:1 to 3:1. The production of IL-4 and IL-5 by CD4+T cells in the LP promotes IgA production in the intestinal tract. Furthermore, CD8+T cells in the LP are responsible for damage by recognizing and activating macrophages. The primary antibodies secreted into the intestinal tract mucus are IgA. A number of IgA-producing cells distribute in the LP, and IgA molecules that are secreted into the intestinal tract are transported to the gut luminal side by binding to multi‐ meric antibody receptors that are retained on the intestinal epithelial cells.
**Figure 5.** Distribution of iNKT cells in six intestinal allografts. The count of iNKT cells was evaluated by their number per 1mm2 of a section. PP, Peyer's patch; IFR, inter-follicular region; SED, mucosal sub-epithelial dome; LPV, lamina propria of the villi. Individual bars of intact, onset, and post-administration display the TCRVα24+T cell counts in the intact mucosa, in the mucosa with rejection, and in the mucosa 48 hours after >30mg of steroid administration (poststeroid therapy).
Both B-1 and B-2 cells produce IgA. B-2 cells are derived from PP B cells and exhibit a CD23+, CD11b-, CD5-, and CD19low phenotype. Follicular B cells in the PPs are stimulated by antigen in the intestinal lumen and differentiate from IgM+to IgA+cells by class switching. IgA+cells in the PPs circulate throughout the body via the thoracic duct and differentiate into IgAproducing cells by the effect of IL-6 produced by intestinal epithelial cells. B-1 cells, in contrast, are derived from B cells of the abdominal cavity and exhibit a CD19+, CD23-, and CD11b+ (B-1a, CD5+; B-1b, CD5-) phenotype. Whereas B-2 cell-derived secretory IgA recognizes foreign antigens with high affinity, B-1 cell-derived secretory IgA recognizes common bacterial phosphorylcholine, lipopolysaccharide, and self-antigens with low affinity.
**Figure 7.** Cytokine concentration in the histologic normal allograft, at the onset of acute cellular rejection (ACR) and
**Figure 6.** Peyer's patch (PP) hyperplasia in acute rejection of small intestinal allograft. Left photo shows an H&E stained hyperplastic PP in the graft at the onset of rejection. Right photo shows the CD20+ stained hyperplastic PP.
The Roles of Invariant NKT Cells in Bowel Immunity — Suppression of Tumor Progression and…
http://dx.doi.org/10.5772/57588
235
Cytokine production by iNKT cells was observed in grafts and total cytokine concentration in sampled tissue was measured. Graphs show plots of the concentrations of IL-4, IL-5, IL-10, and IFN-γ (pg/mg in graft tissue) versus the mean number of IL-4+-, IL-5+-, IFN-γ+-, and IL-10+-iNKT cells (per 10 high-power foci [HPF]). Linear regression lines and correlation
*3.3.6. Activated iNKT cells may damage the small-intestine allograft by release of IL-4/IL-5*
72 hours after steroid pulse. Error bars indicate S.D. [38].
) values are also shown (Figure 8).
coefficient (*R2*
#### *3.3.5. Cytokine production in the intestinal allograft*
Cytokine measurements have suggested that the levels of a variety of cytokines are increased at the onset of ACR and that TNF-α contributes to mucosal damage in graft-versus-host disease [47]. The graph below shows the production of various cytokines in graft mucosal tissues without PPs before transplantation, during ACR, and 72 hrs after administration of a steroid pulse. The data show that production of both IL-4 and IL-5 increase significantly relative to other cytokines in the allograft tissue during ACR, suggesting that ACR involves hypercyto‐ kinemia (Figure 7)[38].
The Roles of Invariant NKT Cells in Bowel Immunity — Suppression of Tumor Progression and… http://dx.doi.org/10.5772/57588 235
**Figure 6.** Peyer's patch (PP) hyperplasia in acute rejection of small intestinal allograft. Left photo shows an H&E stained hyperplastic PP in the graft at the onset of rejection. Right photo shows the CD20+ stained hyperplastic PP.
Both B-1 and B-2 cells produce IgA. B-2 cells are derived from PP B cells and exhibit a CD23+, CD11b-, CD5-, and CD19low phenotype. Follicular B cells in the PPs are stimulated by antigen in the intestinal lumen and differentiate from IgM+to IgA+cells by class switching. IgA+cells in the PPs circulate throughout the body via the thoracic duct and differentiate into IgAproducing cells by the effect of IL-6 produced by intestinal epithelial cells. B-1 cells, in contrast, are derived from B cells of the abdominal cavity and exhibit a CD19+, CD23-, and CD11b+ (B-1a, CD5+; B-1b, CD5-) phenotype. Whereas B-2 cell-derived secretory IgA recognizes foreign antigens with high affinity, B-1 cell-derived secretory IgA recognizes common bacterial
**Figure 5.** Distribution of iNKT cells in six intestinal allografts. The count of iNKT cells was evaluated by their number per 1mm2 of a section. PP, Peyer's patch; IFR, inter-follicular region; SED, mucosal sub-epithelial dome; LPV, lamina propria of the villi. Individual bars of intact, onset, and post-administration display the TCRVα24+T cell counts in the intact mucosa, in the mucosa with rejection, and in the mucosa 48 hours after >30mg of steroid administration (post-
Cytokine measurements have suggested that the levels of a variety of cytokines are increased at the onset of ACR and that TNF-α contributes to mucosal damage in graft-versus-host disease [47]. The graph below shows the production of various cytokines in graft mucosal tissues without PPs before transplantation, during ACR, and 72 hrs after administration of a steroid pulse. The data show that production of both IL-4 and IL-5 increase significantly relative to other cytokines in the allograft tissue during ACR, suggesting that ACR involves hypercyto‐
phosphorylcholine, lipopolysaccharide, and self-antigens with low affinity.
*3.3.5. Cytokine production in the intestinal allograft*
kinemia (Figure 7)[38].
steroid therapy).
234 Immune Response Activation
**Figure 7.** Cytokine concentration in the histologic normal allograft, at the onset of acute cellular rejection (ACR) and 72 hours after steroid pulse. Error bars indicate S.D. [38].
#### *3.3.6. Activated iNKT cells may damage the small-intestine allograft by release of IL-4/IL-5*
Cytokine production by iNKT cells was observed in grafts and total cytokine concentration in sampled tissue was measured. Graphs show plots of the concentrations of IL-4, IL-5, IL-10, and IFN-γ (pg/mg in graft tissue) versus the mean number of IL-4+-, IL-5+-, IFN-γ+-, and IL-10+-iNKT cells (per 10 high-power foci [HPF]). Linear regression lines and correlation coefficient (*R2* ) values are also shown (Figure 8).
**4. iNKT cells in large-intestine immunity**
Unlike the small intestine, a considerable number of NKT cells reside intra-epithelially in the colon of mice [9]. CD161+T cells are the major natural killer receptor-positive cell population in the intestine [10]. On the other hand, few reports have been published regarding human
The Roles of Invariant NKT Cells in Bowel Immunity — Suppression of Tumor Progression and…
http://dx.doi.org/10.5772/57588
237
Our previous study was the first report the presence of TCRVα24+ iNKT cells in the colon [29]. In that study, a small number of Vα24+NKT cells were observed in the normal colorectal mucosa (2.6 ± 3.7 cells/5 HPF), unlike the small intestine, which contained few resident NKT
Furthermore, we investigated intra-tumor TCRVα24-positive cells in 103 primary colorectal carcinoma samples [29]. The observation that intra-tumor Vα24+T cells invariably coexpress TCRVβ11 and other NK cell markers indicates that they are iNKT cells. The density of iNKT cells in colorectal carcinomas increased significantly. iNKT cells were actually found in the luminal space of tumor-infiltrating vessels. iNKT cells could migrate from the peripheral blood to the tumors. A higher fraction of iNKT cells express the activation marker CD69 in colorectal carcinomas than in normal mucosa, and these cells are probably involved in local cytotoxicity against tumor cells through secretion of IFN-γ and apoptosis-inducing molecules such as FasL, perforin, and granzyme B. In this way, iNKT cells may play a role in the primary defense against colorectal carcinoma by induction of tumor cell death. Alternatively, the increase in the number of NKT cells in tumors may represent a surrogate marker of antitumor activity. Coca et al. reported that colorectal carcinomas with a higher degree of CD57+NK cell infiltra‐ tion have better prognosis [51]. Because the CD57+cell population includes NKT cells, the present observation agrees in part with their conclusion. Another interesting feature of our study was the finding that iNKT cells play a role in the inhibition of lymph node metastasis of colorectal carcinomas [29]. Although iNKT cells are normally rare in lymph nodes, a large number of these cells appear in the metastasis-free swollen lymph nodes of colorectal carci‐ noma patients. The precise mechanism through which iNKT cells inhibit lymph node meta‐ stasis remains unclear. Activated iNKT cells may migrate to regional lymph nodes, where they may prevent metastasis by inducing apoptosis of the colorectal carcinoma cells. A recent study reported that chemokine CXCL16 suppresses liver metastasis of colorectal cancer via aug‐
Patients with a higher degree of intra-tumor NKT-cell infiltration show significantly higher rates of overall and disease-free survival. This is consistent with the fact that these patients exhibit much less lymph node metastasis. The degree of NKT cell infiltration in tumors was evaluated as low degree (<7 NKT cells/five x400 foci) or high degree (≥7 NKT cells/ five x400 foci). The number of iNKT cells increased markedly in colorectal carcinomas, and a majority of these cells showed a phenotype of activation expressing FasL or CD69. A higher degree of
NKT cells, particularly iNKT cells, in the normal and neoplastic colon.
mentation of tumor-infiltrating iNKT cells in a murine model [52].
**4.1. NKT cells in the large intestine**
**4.2. Intra-colorectal tumor iNKT cells**
cells.
**Figure 8.** Regression analysis of cytokine concentration and counts of iNKT cells. IL-4 and IL-5 iNKT cell count correlate with total concentration of IL-4 and IL-5 in the graft.
It is likely that as yet unknown humoral factors recruit iNKT cells to the graft mucosa in order to suppress allograft rejection. iNKT cells have the ability to release IL-4 and antagonize Th1 and CTL responses [48]. However, because the released IL-4 and IL-5 may damage the allograft via eosinophilic enteritis, appropriate immunosuppression is necessary for recovery. We therefore investigated humoral rejection. iNKT cells may modulate such ACR via the release of cytokines.
#### **3.4. NKT cells in intestinal inflammation caused by parasites**
iNKT cells can also modulate mucosal immunity through the release of IL-4 and IL-5 in response to parasite infection. A previous study showed that NKT cells play a critical role in the lethal ileitis induced in C57BL/6 mice after infection with *Toxoplasma gondii* [49]. This intestinal inflammation is caused by overproduction of IFN-γ by iNKT cells in the LP. This detrimental activity of iNKT cells can be blocked by treatment with alpha-galactosylceramide, which induces a shift in cytokine production by iNKT cells toward a Th2 profile. iNKT cells participate in the clearance of parasites by shifting the cytokine profile toward a Th1 pattern. Another study using CD1d-deficient mice lacking iNKT cells showed that iNKT cells are critical for the elimination of *T. gondii*[50]. Thus, iNKT cells may form an immunological barrier against parasite infection.
#### **4. iNKT cells in large-intestine immunity**
#### **4.1. NKT cells in the large intestine**
Unlike the small intestine, a considerable number of NKT cells reside intra-epithelially in the colon of mice [9]. CD161+T cells are the major natural killer receptor-positive cell population in the intestine [10]. On the other hand, few reports have been published regarding human NKT cells, particularly iNKT cells, in the normal and neoplastic colon.
Our previous study was the first report the presence of TCRVα24+ iNKT cells in the colon [29]. In that study, a small number of Vα24+NKT cells were observed in the normal colorectal mucosa (2.6 ± 3.7 cells/5 HPF), unlike the small intestine, which contained few resident NKT cells.
#### **4.2. Intra-colorectal tumor iNKT cells**
**Figure 8.** Regression analysis of cytokine concentration and counts of iNKT cells. IL-4 and IL-5 iNKT cell count correlate
It is likely that as yet unknown humoral factors recruit iNKT cells to the graft mucosa in order to suppress allograft rejection. iNKT cells have the ability to release IL-4 and antagonize Th1 and CTL responses [48]. However, because the released IL-4 and IL-5 may damage the allograft via eosinophilic enteritis, appropriate immunosuppression is necessary for recovery. We therefore investigated humoral rejection. iNKT cells may modulate such ACR via the release
iNKT cells can also modulate mucosal immunity through the release of IL-4 and IL-5 in response to parasite infection. A previous study showed that NKT cells play a critical role in the lethal ileitis induced in C57BL/6 mice after infection with *Toxoplasma gondii* [49]. This intestinal inflammation is caused by overproduction of IFN-γ by iNKT cells in the LP. This detrimental activity of iNKT cells can be blocked by treatment with alpha-galactosylceramide, which induces a shift in cytokine production by iNKT cells toward a Th2 profile. iNKT cells participate in the clearance of parasites by shifting the cytokine profile toward a Th1 pattern. Another study using CD1d-deficient mice lacking iNKT cells showed that iNKT cells are critical for the elimination of *T. gondii*[50]. Thus, iNKT cells may form an immunological barrier
with total concentration of IL-4 and IL-5 in the graft.
**3.4. NKT cells in intestinal inflammation caused by parasites**
of cytokines.
236 Immune Response Activation
against parasite infection.
Furthermore, we investigated intra-tumor TCRVα24-positive cells in 103 primary colorectal carcinoma samples [29]. The observation that intra-tumor Vα24+T cells invariably coexpress TCRVβ11 and other NK cell markers indicates that they are iNKT cells. The density of iNKT cells in colorectal carcinomas increased significantly. iNKT cells were actually found in the luminal space of tumor-infiltrating vessels. iNKT cells could migrate from the peripheral blood to the tumors. A higher fraction of iNKT cells express the activation marker CD69 in colorectal carcinomas than in normal mucosa, and these cells are probably involved in local cytotoxicity against tumor cells through secretion of IFN-γ and apoptosis-inducing molecules such as FasL, perforin, and granzyme B. In this way, iNKT cells may play a role in the primary defense against colorectal carcinoma by induction of tumor cell death. Alternatively, the increase in the number of NKT cells in tumors may represent a surrogate marker of antitumor activity.
Coca et al. reported that colorectal carcinomas with a higher degree of CD57+NK cell infiltra‐ tion have better prognosis [51]. Because the CD57+cell population includes NKT cells, the present observation agrees in part with their conclusion. Another interesting feature of our study was the finding that iNKT cells play a role in the inhibition of lymph node metastasis of colorectal carcinomas [29]. Although iNKT cells are normally rare in lymph nodes, a large number of these cells appear in the metastasis-free swollen lymph nodes of colorectal carci‐ noma patients. The precise mechanism through which iNKT cells inhibit lymph node meta‐ stasis remains unclear. Activated iNKT cells may migrate to regional lymph nodes, where they may prevent metastasis by inducing apoptosis of the colorectal carcinoma cells. A recent study reported that chemokine CXCL16 suppresses liver metastasis of colorectal cancer via aug‐ mentation of tumor-infiltrating iNKT cells in a murine model [52].
Patients with a higher degree of intra-tumor NKT-cell infiltration show significantly higher rates of overall and disease-free survival. This is consistent with the fact that these patients exhibit much less lymph node metastasis. The degree of NKT cell infiltration in tumors was evaluated as low degree (<7 NKT cells/five x400 foci) or high degree (≥7 NKT cells/ five x400 foci). The number of iNKT cells increased markedly in colorectal carcinomas, and a majority of these cells showed a phenotype of activation expressing FasL or CD69. A higher degree of iNKT cell infiltration correlated with lower lymph node metastasis (P=0.042). Patients with a high degree of iNKT cell infiltration showed higher overall (P=0.018) as well as disease-free (P=0.0006) survival rates. Intra-tumor iNKT cell infiltration was an independent prognostic factor for overall (P=0.033) and disease-free (P=0.0064) survival. A higher degree of iNKT infiltration in colorectal carcinomas is therefore an independent indicator of a favorable prognosis [29] (Figure 9).
absence of early exposure, exposure later in life to factors that stimulate these cells may induce
The Roles of Invariant NKT Cells in Bowel Immunity — Suppression of Tumor Progression and…
http://dx.doi.org/10.5772/57588
239
In a related report, a profound loss of CD56 expression by all lymphocyte populations was noted in the coeliac gut. Adult coeliac disease expanded populations of CD3+TCRγδ cells and
The relevance of iNKT cells against microbial infection has also been examined. CD1ddependent antigen presentation and microbial killing by iNKT cells are critical for host defense [55]. The host defense against microbial infection depends upon cargo trafficking into lyso‐ somes. The mechanism has been reported [56]. Arf-like GTPase Arl8b is one of the critical regulators of cargo delivery to lysosomes. The formation of CD1 antigen-presenting complexes in lysosomes and phagosome-lysosome fusion depends on Arl8b. Subsequently the delivered CD1 antigen-presenting complexes to the plasma membrane activate iNKT cells in microbial
The quantitative morphometry method to study iNKT cells was applied (Figure 10). Multi‐ variate morphometric analysis images were obtained by microscopy (PROVIS-AX80; Olym‐ pus, Tokyo, Japan) and saved as TIF files. Data were analyzed using CELAVIEW software (Olympus), with individual cell signals expressed in terms of relative fluorescent unit. Morphometric data were displayed according to stained area and circularity. Individual cells were represented by the pixel intensities of the nucleus and cytoplasm. Data sets for individual cells on a single slide were represented by two-dimensional scatter plots based on flow cytometry. Gated cells were further analyzed using antibodies for surface antigens. Plots were further gated according to fluorescence intensity, size, and circularity on the scattergram. Circularity was calculated based on the Heywood circularity factor using the following
an auto-inflammatory response.
killing [56].
**5. Methods**
formula:
**5.1. Quantitative morphometry**
**4.4. Microbial infection and iNKT cells**
decreased populations of NK, NKT, and iNKT cells [30].
**Figure 9.** iNKT cells and apoptosis of tumor cells. FasL, perforin, and granzyme B are apoptosis-inducing molecules. iNKT cells may contribute to apoptosis of the cancer cell by release of these molecules and prevent the metastasis to the lymph node.
#### **4.3. IBD and iNKT cells**
The relevance of iNKT cells in IBD has also been examined. iNKT cells probably play an important role in the pathogenesis of ulcerative colitis [13, 22, 52]. Although our immunohis‐ tochemical data indicated that iNKT cells certainly contribute to inflammation, their influence is limited in the initial phase of IBD (data not shown). Several reports have indicated that iNKT cells secrete abundant amounts of IL-4 upon activation [53]. There has been a report that oxazolone colitis, a Th2 colitis model resembling ulcerative colitis, is mediated by IL-13 producing NK-T cells [54].
Important data regarding the role of iNKT cells in IBD were recently reported [31]. iNKT cells accumulate in the colonic lamina propria in germ-free mice, resulting in increased morbidity in models of IBD as compared with specific pathogen-free mice. The chemokine ligand CXCL16 is associated with an increase in the number of mucosal iNKT cells and consequently an increase in susceptibility to tissue inflammation. Age-sensitive contact with commensal microbes is critical for establishing mucosal iNKT cell tolerance to later environmental exposures. Microbial exposure early in life elicits long-lasting effects on iNKT cells, and in the absence of early exposure, exposure later in life to factors that stimulate these cells may induce an auto-inflammatory response.
In a related report, a profound loss of CD56 expression by all lymphocyte populations was noted in the coeliac gut. Adult coeliac disease expanded populations of CD3+TCRγδ cells and decreased populations of NK, NKT, and iNKT cells [30].
#### **4.4. Microbial infection and iNKT cells**
The relevance of iNKT cells against microbial infection has also been examined. CD1ddependent antigen presentation and microbial killing by iNKT cells are critical for host defense [55]. The host defense against microbial infection depends upon cargo trafficking into lyso‐ somes. The mechanism has been reported [56]. Arf-like GTPase Arl8b is one of the critical regulators of cargo delivery to lysosomes. The formation of CD1 antigen-presenting complexes in lysosomes and phagosome-lysosome fusion depends on Arl8b. Subsequently the delivered CD1 antigen-presenting complexes to the plasma membrane activate iNKT cells in microbial killing [56].
#### **5. Methods**
iNKT cell infiltration correlated with lower lymph node metastasis (P=0.042). Patients with a high degree of iNKT cell infiltration showed higher overall (P=0.018) as well as disease-free (P=0.0006) survival rates. Intra-tumor iNKT cell infiltration was an independent prognostic factor for overall (P=0.033) and disease-free (P=0.0064) survival. A higher degree of iNKT infiltration in colorectal carcinomas is therefore an independent indicator of a favorable
**Figure 9.** iNKT cells and apoptosis of tumor cells. FasL, perforin, and granzyme B are apoptosis-inducing molecules. iNKT cells may contribute to apoptosis of the cancer cell by release of these molecules and prevent the metastasis to
The relevance of iNKT cells in IBD has also been examined. iNKT cells probably play an important role in the pathogenesis of ulcerative colitis [13, 22, 52]. Although our immunohis‐ tochemical data indicated that iNKT cells certainly contribute to inflammation, their influence is limited in the initial phase of IBD (data not shown). Several reports have indicated that iNKT cells secrete abundant amounts of IL-4 upon activation [53]. There has been a report that oxazolone colitis, a Th2 colitis model resembling ulcerative colitis, is mediated by IL-13-
Important data regarding the role of iNKT cells in IBD were recently reported [31]. iNKT cells accumulate in the colonic lamina propria in germ-free mice, resulting in increased morbidity in models of IBD as compared with specific pathogen-free mice. The chemokine ligand CXCL16 is associated with an increase in the number of mucosal iNKT cells and consequently an increase in susceptibility to tissue inflammation. Age-sensitive contact with commensal microbes is critical for establishing mucosal iNKT cell tolerance to later environmental exposures. Microbial exposure early in life elicits long-lasting effects on iNKT cells, and in the
prognosis [29] (Figure 9).
238 Immune Response Activation
the lymph node.
**4.3. IBD and iNKT cells**
producing NK-T cells [54].
#### **5.1. Quantitative morphometry**
The quantitative morphometry method to study iNKT cells was applied (Figure 10). Multi‐ variate morphometric analysis images were obtained by microscopy (PROVIS-AX80; Olym‐ pus, Tokyo, Japan) and saved as TIF files. Data were analyzed using CELAVIEW software (Olympus), with individual cell signals expressed in terms of relative fluorescent unit. Morphometric data were displayed according to stained area and circularity. Individual cells were represented by the pixel intensities of the nucleus and cytoplasm. Data sets for individual cells on a single slide were represented by two-dimensional scatter plots based on flow cytometry. Gated cells were further analyzed using antibodies for surface antigens. Plots were further gated according to fluorescence intensity, size, and circularity on the scattergram. Circularity was calculated based on the Heywood circularity factor using the following formula:
where the perimeter (*P*) of individual cells was divided by the circumference of a circle with the same area (*A*), with the real circle factor being equivalent to 1.0. As the boundary of a binary image is composed of discrete pixels, IMAQ Vision (National Instruments Corporation, Austin, TX, USA) was used to sub-sample the boundary [21, 39]. The scheme of quantitative morphometry is shown below.
immune response through the release of opposite-acting cytokines. iNKT cells thus play a
The Roles of Invariant NKT Cells in Bowel Immunity — Suppression of Tumor Progression and…
http://dx.doi.org/10.5772/57588
241
This work was supported by a Grant-in-Aid (Synergy of Fluctuation and Structure: Quest for Universal Laws in Non-Equilibrium Systems 2013-2017 Grant-in-Aid for Scientific Research on Innovative Areas, MEXT, Japan; http://sfs-dynamics.jp/eng/index.html). The funding agencies played no role in study design, data collection and analysis, in the decision to publish,
We thank Professor Shinji Uemoto and Dr. Shinya Okamoto of the Departments of Surgery and Pediatric Surgery at Kyoto University Hospital, respectively. Also we thank for Dr. Hisashi Onodera of the Department of Surgery at St. Luke's International Hospital of Tokyo, Japan. We dedicate this chapter to our great teacher, Professor Hirohiko Yamabe, Department of Diagnostic Pathology, Kyoto University Hospital. It was Professor Yamabe who established
Department of Diagnostic Pathology, Graduate School of Medicine, Kyoto University
[1] Neurath, M.F.; Finotto, S.; Glimcher, L.H. (2002). The role of Th1/Th2 polarization in
[2] Godfrey, D.I.; MacDonald, H.R.; Kronenberg M.; et al. (2004). NKT cells: what's in a
[3] Exley, M.; Porcelli, S.; Furman, M.; et al. (1998). CD161 (NKRP1A) costimulation of CD1d-dependent activation of human cells expressing invariant Va24JaQT-cell
[4] Taniguchi, M.; Harada, M.; Kojo, S.; Nakayama, T.; Wakao, H. (2003). The regulatory role of Valpha14 NKT cells in innate and acquired immune response. Annu Rev
pivotal role in modulating mucosal immunity in the intestine.
the pathologic diagnosis of organ transplantation in Japan.
Tatsuaki Tsuruyama and Wulamujiang Aini
mucosal immunity. Nat Med 8:567-573.
receptor a chains. J Exp Med 188:867-876.
name? Nat Rev Immunol 4:231-237.
Immunol 21:483-513.
**Acknowledgements**
**Author details**
**References**
Hospital, Kyoto City, Japan
or in preparation of the manuscript.
Lymphocytes are oval or circular and the circulatory factor is nearly to 1.0. Therefore, the subjects of which circulatory factor is nearly to to 1.0 is gated for subsequent analysis using immunostaining. iNKT cells were actually gated for subsequent analysis using antibodies for CD56 and TCRVα24. In the first gating, the area was also analyzed [42] (Figure 10).
**Figure 10.** A scheme of two dimensional plots of the morphometric analysis for analysis of surface antigens CD56 and TCRVα24 on the iNKT cells. The gating shown in the left graph shows an example of selection of the analyzed lympho‐ cytes for analysis of the surface antigens shown in the right graph. The similarity to the circle is evaluated by the circu‐ latory factor near 1.0. The right figures the factor values of various shapes.
#### **5.2. Immunohistochemistry**
The fluorescent staining method for iNKT cells was as reported using the CSA system (DAKO, Gostrup, Denmark). Frozen section samples and paraffin embedded specimens were used. For primary reagents, we used monoclonal antibody C15 (TCRVα24; Immunotech SA, Marseilles, France). Antibodies against CD3, CD4 and CD8 were purchased from DAKO, and DAB staining was performed for signal visualization. Negative controls for primary antibodies were IgG1, IG2a and IgG2b (Cat No. X0931, X0943 and X0944; DAKO) [29, 38].
#### **6. Conclusion**
The mucosa is exposed to an extremely wide variety of antigens in food, and iNKT cells within the mucosa are responsible for mounting immune responses to these antigens. The cytokines released during these responses may damage the small intestine. In the case of small-intestine transplantation, the response of iNKT cells during ACR indicates that iNKT cells may be recruited to the transplant to repress rejection. iNKT cells probably contribute to the suppres‐ sion of ACR by releasing IL-4, an antagonist of IFN-γ. On the other hand, iNKT cells normally reside within the large intestine, where they also play a pivotal role in immunity. iNKT cells can be activated to express apoptosis-inducing molecules and may induce cytolysis of tumor cells. Thus, iNKT cells have two roles; they can upregulate or downregulate the mucosal immune response through the release of opposite-acting cytokines. iNKT cells thus play a pivotal role in modulating mucosal immunity in the intestine.
#### **Acknowledgements**
where the perimeter (*P*) of individual cells was divided by the circumference of a circle with the same area (*A*), with the real circle factor being equivalent to 1.0. As the boundary of a binary image is composed of discrete pixels, IMAQ Vision (National Instruments Corporation, Austin, TX, USA) was used to sub-sample the boundary [21, 39]. The scheme of quantitative
Lymphocytes are oval or circular and the circulatory factor is nearly to 1.0. Therefore, the subjects of which circulatory factor is nearly to to 1.0 is gated for subsequent analysis using immunostaining. iNKT cells were actually gated for subsequent analysis using antibodies for
**Figure 10.** A scheme of two dimensional plots of the morphometric analysis for analysis of surface antigens CD56 and TCRVα24 on the iNKT cells. The gating shown in the left graph shows an example of selection of the analyzed lympho‐ cytes for analysis of the surface antigens shown in the right graph. The similarity to the circle is evaluated by the circu‐
The fluorescent staining method for iNKT cells was as reported using the CSA system (DAKO, Gostrup, Denmark). Frozen section samples and paraffin embedded specimens were used. For primary reagents, we used monoclonal antibody C15 (TCRVα24; Immunotech SA, Marseilles, France). Antibodies against CD3, CD4 and CD8 were purchased from DAKO, and DAB staining was performed for signal visualization. Negative controls for primary antibodies were
The mucosa is exposed to an extremely wide variety of antigens in food, and iNKT cells within the mucosa are responsible for mounting immune responses to these antigens. The cytokines released during these responses may damage the small intestine. In the case of small-intestine transplantation, the response of iNKT cells during ACR indicates that iNKT cells may be recruited to the transplant to repress rejection. iNKT cells probably contribute to the suppres‐ sion of ACR by releasing IL-4, an antagonist of IFN-γ. On the other hand, iNKT cells normally reside within the large intestine, where they also play a pivotal role in immunity. iNKT cells can be activated to express apoptosis-inducing molecules and may induce cytolysis of tumor cells. Thus, iNKT cells have two roles; they can upregulate or downregulate the mucosal
latory factor near 1.0. The right figures the factor values of various shapes.
IgG1, IG2a and IgG2b (Cat No. X0931, X0943 and X0944; DAKO) [29, 38].
CD56 and TCRVα24. In the first gating, the area was also analyzed [42] (Figure 10).
morphometry is shown below.
240 Immune Response Activation
**5.2. Immunohistochemistry**
**6. Conclusion**
This work was supported by a Grant-in-Aid (Synergy of Fluctuation and Structure: Quest for Universal Laws in Non-Equilibrium Systems 2013-2017 Grant-in-Aid for Scientific Research on Innovative Areas, MEXT, Japan; http://sfs-dynamics.jp/eng/index.html). The funding agencies played no role in study design, data collection and analysis, in the decision to publish, or in preparation of the manuscript.
We thank Professor Shinji Uemoto and Dr. Shinya Okamoto of the Departments of Surgery and Pediatric Surgery at Kyoto University Hospital, respectively. Also we thank for Dr. Hisashi Onodera of the Department of Surgery at St. Luke's International Hospital of Tokyo, Japan. We dedicate this chapter to our great teacher, Professor Hirohiko Yamabe, Department of Diagnostic Pathology, Kyoto University Hospital. It was Professor Yamabe who established the pathologic diagnosis of organ transplantation in Japan.
#### **Author details**
Tatsuaki Tsuruyama and Wulamujiang Aini
Department of Diagnostic Pathology, Graduate School of Medicine, Kyoto University Hospital, Kyoto City, Japan
#### **References**
[5] Cui, J.; Shin, T.; Kawano, T.; et al. (1997). Requirement for Valpha14 NKT cells in IL-12 mediated rejection of tumors. Science 278: 1623-1626.
[19] Hong, S.; Wilson, M.T.; Serizawa, I.; et al. (2001). The natural killer T-cell ligand agalactosylceramide prevents autoimmune diabetes in non-obese diabetic mice. Nat
The Roles of Invariant NKT Cells in Bowel Immunity — Suppression of Tumor Progression and…
http://dx.doi.org/10.5772/57588
243
[20] Kita, H.; Naidenko, O.V.; Kronenberg, M.; et al. (2002). Quantitation and phenotypic analysis of natural killer T cells in primary biliary cirrhosis using a human CD1d
[21] Kono, F.; Honda, T.; Aini, W.; Manabe, T.; Haga, H.; Tsuruyama T. (2013) IFN-γ/ CCR5 expression in invariant NKT cells and CCL5 expression in capillary veins of dermal papillae correlate with development of psoriasis vulgaris. Br J Dermatol. doi:
[22] Akbari, O.; Stock, P.; Meyer, E.; et al. (2003). Essential role of NKT cells producing IL-4 and IL-13 in the development of allergen-induced airway hyperreactivity. Nat Med
[23] Araujo, L.M.; Lefort, J.; Nahori, M.A.; et al. (2004). Exacerbated Th2-mediated airway inflammation and hyperresponsiveness in autoimmune diabetes-prone NOD mice: a
[24] Terabe, M.; Matsui, S.; Noben-Trauth, N.; et al. (2000). NKT cell-mediated repression of tumor immunosurveillance by IL-13 and the IL-4R-STAT6 pathway. Nat Immunol
[25] Smyth, M.J.; Crowe, N.Y.; Hayakawa, Y.; et al. (2002). NKT cells: conductors of tumor
[26] [26]Tahir, S.M.A.; Cheng, O.; Shaulov, A. (2001). Loss of IFN-gamma production by
[28] Crowe, N.Y.; Smyth, M.J.; Godfrey, D.I.; (2002). A critical role for natural killer T cells in immunosurveillance of methylcholanthrene-induced sarcomas. J Exp Med
Valpha24-positive natural killer T cells: a prognostic factor for primary colorectal
[30] Grose, R.H., Cummins, A.G., Thompson, F.M.; Deficiency of invariant natural killer T
[31] Avitzur, Y.; Grant, D. (2010). Intestine transplantation in children: update 2010. Pediatr
[32] Goulet, O.; Sauvat, F. (2006). Short bowel syndrome and intestinal transplantation in
[27] Kikuchi, A.; Nieda, M.; Schmidt, C.; et al. (2001). In vitro antitumor activity of agalactosylceramide-stimulated human invariant Valpha24+ NKT cells against
[29] Tachibana, T.; Onodera, H.; Tsuruyama, T.; et al. (2005). Increased intratumor
invariant NK T cells in advanced cancer. J Immunol 167:4046-4050.
critical role for CD1d-dependent NKT cells. Eur J Immunol 34:327-335.
Med 7:1052-1056.
9:582-588.
1:515-520.
196:119-127.
tetramer. Gastroenterology 123: 1031-1043.
immunity? Curr Opin Immunol 14:165-171.
carcinomas. Clin Cancer Res 11: 7322–7327.
cells in coeliac disease. Gut. (2007); 790-795.
children. Curr Opin Clin Nutr Metab Care 9:304–313.
Clin North Am 57:415–431.
melanoma. Br J Cancer 85:741-746.
10.1111/bjd.12812 (in press).
[19] Hong, S.; Wilson, M.T.; Serizawa, I.; et al. (2001). The natural killer T-cell ligand agalactosylceramide prevents autoimmune diabetes in non-obese diabetic mice. Nat Med 7:1052-1056.
[5] Cui, J.; Shin, T.; Kawano, T.; et al. (1997). Requirement for Valpha14 NKT cells in IL-12-
[6] Lee, P.T.; Benlagha, K.; Teyton, L.; Bendelac, A. (2002). Distinct functional lineages of
[7] Kawakami, K.; Yamamoto, N.; Kinjo, Y.; et al. (2003). Critical role of Valpha14+ natural killer T cells in the innate phase of host protection against *Streptococcus pneumonia*
[8] Kronenberg, M.; Gapin, L. (2002). The unconventional lifestyle of NKT cells. Nat Rev
[9] Bannai, M.; Kawamura, T.; Naito T.; et al. (2001). Abundance of unconventional CD8+
[10] O'Keeffe, J.; Doherty, D.G.; Kenna, T.; et al. (2004). Diverse populations of T cells with NK cell receptors accumulate in the healthy human intestine and in colorectal cancer.
[11] Kawakami, K.; Kinjo, Y.; Uezu, K.; et al. (2001). Monocyte chemoattractant protein-1 dependent increase of Valpha14 NKT cells in lungs and their roles in Th1 response and
[12] Grubor-Bauk, B.; Simmons, A.; Mayrhofer, G.; Speck, P.G. (2003). Impaired clearance of herpes simplex virus type 1 from mice lacking CD1d or NKT cells expressing the
[13] Fuss, I.J.; Heller, F.; Boirivant, M.; et al. (2004). Nonclassical CD1d-restricted NK T cells that produce IL-13 characterize an atypical Th2 response in ulcerative colitis. J Clin
[14] Ito, K.; Karasawa, M.; Kawano, T.; et al. (2000). Involvement of decidual Valpha14 NKT
[15] Seino, K.; Fukao, K.; Muramoto, K.; et al. (2001). Requirement for natural killer T (NKT) cells in the induction of allograft tolerance. Proc Natl Acad Sci USA 98:2577-2581.
[16] Ikehara, Y.; Yasunami, Y.; Kodama, S.; et al. (2000). CD4+ Valpha14 natural killer T cells are essential for acceptance of rat islet xenografts in mice. J Clin Invest 105:1761-1767.
[17] Hammond, K.J.L.; Godfrey, D.I. (2002). NKT cells: Potential targets for autoimmune
[18] Beaudoin, L.; Lalouz, V.; Novak, J.; et al. (2002). NKT cells inhibit the onset of diabetes by impairing the development of pathogenic T cells specific for pancreatic B cells.
natural killer T cells in the large intestine. Eur J Immunol 31:3361-3369.
host defense in cryptococcal infection. J Immunol 167: 6525-6532.
semivariant Valpha14-Ja281 TCR. J Immunol 170:1430-1434.
cells in abortion. Proc Natl Acad Sci USA 97:740-744.
Kugathasan, S.; Saubermann, L.J.; Smith, L.; et al. (2007).
disease therapy? Tissue Antigens 59:353-363.
mediated rejection of tumors. Science 278: 1623-1626.
infection. Eur J Immunol 33:3322-3330.
Immunol 2:557-568.
242 Immune Response Activation
Invest 113:149.
Immunity 17:725-736.
Eur J Immunol 34:2110-2118.
human Va24 natural killer T cells. J Exp Med 195:637-641.
[33] Pirenne, J.; Hoffman, I.; Miserez, M.; et al. (2006). Selection criteria and outcome of patients referred to intestinal transplantation: a European center experience. Transplant Proc 38:1671–1672.
rejection in formalin-fixed paraffin-embedded mucosal biopsies. Transpl Int
The Roles of Invariant NKT Cells in Bowel Immunity — Suppression of Tumor Progression and…
http://dx.doi.org/10.5772/57588
245
[46] Gonnella, PC.; Waldner, H.P.; Weiner, H.L. (2001). B cell-deficient (mMT) mice have alterations in the cytokine microenvironment of the gut-associated lymphoid tissue (GALT) and a defect in the low dose mechanism of oral tolerance. J Immunol
[47] Stuber, E.; Buschenfeld, A.; von Freier, A.; et al. (1999). Intestinal crypt cell apoptosis in murine acute graft versus host disease is mediated by tumour necrosis factor alpha and not by the FasL-Fas interaction: effect of pentoxifylline on the development of mucosal
[48] Kim, C.H.; Johnston, B.; Butcher, E.C. (2002). Trafficking machinery of NKT cells: shared and differential chemokine receptor expression among V alpha 24(+)V beta 11(+) NKT cell subsets with distinct cytokine-producing capacity. Blood 100:11-16.
[49] Ronet, C.; Darche, S.; Leite de Moraes, M.; Miyake, S.; Yamamura, T.; Louis, J.A.; Kasper, L.H.; Buzoni-Gatel, D. (2005). NKT cells are critical for the initiation of an inflammatory bowel response against Toxoplasma gondii. J Immunol 15:175:899-908.
[51] Coca, S.; Perez-Piqueras, J.; Martinez, D. (1997). The prognostic significance of
[52] Kee, J.Y.; Ito, A.; Hojo, S.; Hashimoto, I.; Igarashi, Y.; Tsukada, K.; Irimura, T.;
paradoxes. *Annu Rev Immunol* 26:877-900.
producing NK-T cells. Immunity 17: 629-638.
and NKT cells in microbial immunity. *Adv Immunol* 102:1-94.
[50] Olszak, T.; An, D.; Zeissig, S.; Vera, M.P.; Richter, J.; Franke A.; Glickman, J.N.; Siebert, R.; Baron R.M.; Kasper, D.L.; Blumberg, R.S. (2012). Microbial exposure during early life has persistent effects on natural killer T cell function. Science; 336:489-493.
intratumoral natural killer cells in patients with colorectal carcinoma. Cancer 79:
Shibahara, N.; Nakayama, T.; Yoshie, O.; Sakurai, H.; Saiki, I.; Koizumi, K. (2013). Chemokine CXCL16 suppresses liver metastasis of colorectal cancer via augmentation of tumor-infiltrating natural killer T cells in a murine model. Oncol Rep 29:975-982.
[53] Kronenberg, M. (2005). Toward an understanding of NKT cell biology: Progress and
[54] Heller, F.; Fuss, I.J.; Nieuwenhuis, E.E.; Blumberg, R.S.; Strober, W. (2002). Oxazolone colitis, a Th2 colitis model resembling ulcerative colitis, is mediated by IL-13-
[55] Cohen, N.R.; Garg, S.; Brenner, M.B. (2009). Antigen Presentation by CD1 lipids, T cells,
[56] Garg S, Sharma M, Ung C, Tuli A, Barral DC, Hava DL, Veerapen N, Besra GS, Hacohen N, Brenner MB. Lysosomal trafficking, antigen presentation, and microbial killing are
controlled by the Arf-like GTPase Arl8b. Immunity (2011) 35; 182-193.
24:697-707.
166:4456-4464.
2320-2328.
atrophy. Gut 45:229–235.
rejection in formalin-fixed paraffin-embedded mucosal biopsies. Transpl Int 24:697-707.
[33] Pirenne, J.; Hoffman, I.; Miserez, M.; et al. (2006). Selection criteria and outcome of patients referred to intestinal transplantation: a European center experience.
aganglionosis and intestinal pseudo-obstruction. Semin Pediatr Surg 13:286–292.
[35] Sharif, K.; Beath, S.V.; Kelly, D.A.; et al. (2003). New perspective for the management of near-total or total intestinal aganglionosis in infants. J Pediatr Surg 38:25–28.
[36] Ruiz, P.; Takahashi, H.; Delacruz, V.; et al. (2010). International grading scheme for acute cellular rejection in small-bowel transplantation: single-center experience.
[37] Wu, T.; Abu-Elmagd, K.; Bond, G.; et al. (2003). A schema for histologic grading of small
[38] Tsuruyama, T.; Fujimoto, Y.; Yonekawa, Y.; et al. (2012). Invariant natural killer T cells infiltrate intestinal allografts undergoing acute cellular rejection. Transpl Int 25: 537–
[39] Merger, M.; Viney, J.L.; Borojevic, R.; et al. (2002). Defining the roles of perforin, Fas/ FasL, and tumour necrosis factor alpha in T cell induced mucosal damage in the mouse
[40] [40]Suda, T.; Hashimoto, H.; Tanaka, M.; et al. (1997). Membrane Fas ligand kills
human peripheral blood T lymphocytes, and soluble Fas ligand blocks the killing. J Exp
[41] Miyagawa-Hayashino, A.; Tsuruyama, T.; Egawa, H.; et al. (2007). FasL expression in hepatic antigen-presenting cells and phagocytosis of apoptotic T cells by FasL+ Kupffer cells are indicators of rejection activity in human liver allografts. Am J Pathol 171:1499–
[42] Tsuruyama, T.; Okamoto, S.; Fujimoto, Y.; Yoshizawa, A.; Yoshitoshi, E.; Egawa, H.; Nakase, H.; Aini, W.; Miyao, M.; Tamaki, K.; Uemoto, S.; Haga, H. (2013). Histology of intestinal allografts: lymphocyte apoptosis and phagocytosis of lymphocytic apoptotic
bodies are diagnostic findings of acute rejection in addition to crypt apoptosis.
[43] Karczewski, J.; Karczewski, M.; Glyda, M.; Wiktorowicz, K. (2008). Role of TH1/TH2
Kuppireddi S, Wang CR, Kumar V, Alegre ML. (2006). Role of natural killer cell subsets
cytokines in kidney allograft rejection. Transplant Proc 40:3390-3392.
in cardiac allograft rejection. Am J Transplant. 6: 505-513.
[44] McNerney M,E.; Lee KM;Zhou P; Molinero L; Mashayekhi M; Guzior, Sattar H,
[45] Asaoka, T.; Island, E.R.; Tryphonopoulos, P.; et al. (2011). Characteristic immune, apoptosis and inflammatory gene profiles associated with intestinal acute cellular
[34] Bond, G.J.; Reyes, J.D.; (2004). Intestinal transplantation for total/near-total
intestine allograft acute rejection. Transplantation 75: 1241–1248.
Transplant Proc 38:1671–1672.
Transplant Proc 42:47–53.
intestine. Gut 51:155–163.
American J Surg Pathol 37:178-184.
Med 186: 2045–2050.
544.
244 Immune Response Activation
1508.
**Section 4**
**Autoimmune Disregulation**
### **Autoimmune Disregulation**
**Chapter 9**
**Immunostimulation by Silica Particles and the**
**Development of Autoimmune Dysregulation**
Immunostimulation by environmental and occupational factors has been shown to cause various human diseases such as allergy and autoimmune diseases [1,2]. For example, solvents such as vinyl chloride have been linked to the development of scleroderma (SSc) [3-5], and previous studies reported the relationship between exposure to solvents and rheumatoid arthritis (RA) [6] or multiple sclerosis [7,8]. Another typical substance is silica. Patients exposed to silica particles were shown to develop not only pulmonary fibrosis, known as silicosis [9,10], one of the typical forms of pneumoconiosis, but also various autoimmune diseases [11,12] such as RA (historically known as Caplan syndrome) [13,14], SSc [15-17], systemic lupus erythem‐ atous (SLE) [18,19], and anti-neutrophil cytoplasmic antibody (ANCA)-related vasculitis/ nephritis [20-22]. Many epidemiological reports have demonstrated that silica exposure is a
The mechanism of silica-induced autoimmune dysregulation has been attributed to silica acting as an adjuvant [23,24]. However, silica particles may directly stimulate circulating peripheral immune cells and cause certain alterations in the cellular or molecular functions of these cells because these particles may be retained in pulmonary lesions as well as the lymph nodes after they are inhaled into the body [9-12]. Therefore, if the direct effects of these particles change the characteristics of immune cells leading to the dysregulation of immune tolerance, clarifying these cellular and molecular mechanisms may be useful in preventing immune
> © 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Suni Lee, Hiroaki Hayashi, Megumi Maeda, Hidenori Matsuzaki, Naoko Kumagai-Takei, Ying Chen, Kozo Urakami, Masayasu Kusaka,
Yasumitsu Nishimura and Takemi Otsuki
risk factor of autoimmune diseases [11,12].
http://dx.doi.org/10.5772/57544
**1. Introduction**
Additional information is available at the end of the chapter
### **Immunostimulation by Silica Particles and the Development of Autoimmune Dysregulation**
Suni Lee, Hiroaki Hayashi, Megumi Maeda, Hidenori Matsuzaki, Naoko Kumagai-Takei, Ying Chen, Kozo Urakami, Masayasu Kusaka, Yasumitsu Nishimura and Takemi Otsuki
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/57544
#### **1. Introduction**
Immunostimulation by environmental and occupational factors has been shown to cause various human diseases such as allergy and autoimmune diseases [1,2]. For example, solvents such as vinyl chloride have been linked to the development of scleroderma (SSc) [3-5], and previous studies reported the relationship between exposure to solvents and rheumatoid arthritis (RA) [6] or multiple sclerosis [7,8]. Another typical substance is silica. Patients exposed to silica particles were shown to develop not only pulmonary fibrosis, known as silicosis [9,10], one of the typical forms of pneumoconiosis, but also various autoimmune diseases [11,12] such as RA (historically known as Caplan syndrome) [13,14], SSc [15-17], systemic lupus erythem‐ atous (SLE) [18,19], and anti-neutrophil cytoplasmic antibody (ANCA)-related vasculitis/ nephritis [20-22]. Many epidemiological reports have demonstrated that silica exposure is a risk factor of autoimmune diseases [11,12].
The mechanism of silica-induced autoimmune dysregulation has been attributed to silica acting as an adjuvant [23,24]. However, silica particles may directly stimulate circulating peripheral immune cells and cause certain alterations in the cellular or molecular functions of these cells because these particles may be retained in pulmonary lesions as well as the lymph nodes after they are inhaled into the body [9-12]. Therefore, if the direct effects of these particles change the characteristics of immune cells leading to the dysregulation of immune tolerance, clarifying these cellular and molecular mechanisms may be useful in preventing immune
© 2014 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
disorders in silicosis patients as well as providing an insight into the etiology of various autoimmune diseases.
**3. Schematic summary of the chronic activation of T cells by silica particles**
Immunostimulation by Silica Particles and the Development of Autoimmune Dysregulation
http://dx.doi.org/10.5772/57544
251
A summary of the findings obtained and considerations are schematically presented in Figure 1. Silica particles were shown to chronically activate various T cells. Previous studies reported that effector T cells expressed various activation markers such as PD-1 and CD25 and produced many molecular markers for chronic T cell activation such as soluble Fas (sFas), decoy receptor
On the other hand, silica particles were also shown to activate CD4+CD25+FoxP3+regulatory T cells (Treg). However, this chronic activation caused the enhanced expression of Fas/CD95 on the surface of Treg, which induced early apoptosis [41]. Therefore, Treg may be lost from the peripheral blood, with the resulting imbalance between Treg and effector T cells subse‐
Freshly isolated peripheral blood mononuclear cells (PBMCs) obtained from healthy donors were incubated with silica particles. The CD69 expression on the membrane surface was examined as a marker, demonstrating gradual activation of T lymphocytes during 10-day
Other activation markers were examined in serum derived from silicosis patients and com‐ pared with those from healthy donors [43]. The results showed that sIL-2R levels were slightly higher in silicosis patients than in healthy donors. sIL-2R levels were also examined in serum derived from SSc patients, and correlations between sIL-2R levels and the immunological status of healthy donors, silicosis patients, and SSc patients as 1, 2 and 3 for continuous variables were analyzed. The correlation coefficient was shown to be 0.575 with p=0.0008, which indicated that, from the viewpoint of immunological alterations based on serum sIL-2R levels, silicosis patients were located between healthy donors and SSc patients. Elevated sIL-2R levels may be a pathophysiological marker for hematological malignancies such as human T lymphotropic retrovirus type-1 (HTLV-1) associated with adult T cell leukemia (ATL) and hairy cell leukemia, which reflects the increased production of cells leading to an elevation in serum titers [44-47]. Elevated sIL-2R levels have recently been reported in various autoimmune or inflammatory diseases, suggesting that the immune response is activated by chronic stimulation of T cells with an auto-or foreign antigen [48-51]. Therefore, the moderate, chronic
3 (DcR3), and soluble interleukin (IL)-2 receptor (sIL-2R) [39,40].
Detailed explanations of these findings are presented below.
**4. Immune stimulation of effector T cells by silica particles**
activation status of the immune system may play a role in silicosis.
quently leading to autoimmune dysregulation.
incubation [42].
We previously investigated the immunological effects of silica using human peripheral blood immune cells derived from healthy donors and silicosis patients [25-27]. In this review, we summarized our findings, in which silica was shown to be an environmental immunostimu‐ lator and the chronic activation of immune cells by recurrent and chronic exposure to silica was demonstrated to cause an imbalance in the regulation of T cell responses.
### **2. Autoantibodies detected in silicosis patients and their relationship with HLA phenotypes**
We previously measured autoantibodies in silicosis patients, all of whom were Japanese brickyard workers in Bizen city, Okayama prefecture, Japan. All patients were diagnosed with silicosis based on their radiological findings in accordance with the ILO 2000 guidelines. The amount of free silica inhaled by these patients was estimated to be from 40 to 60%, as deter‐ mined by their work environment. No subjects exhibited any symptoms of autoimmune diseases such as sclerotic skin, Raynaud's phenomenon, facial erythema, or cancer.
We demonstrated that the overall titer of anti-nuclear antibodies (ANA) was higher in these patients than in healthy volunteers [28]. In addition, silicosis patients without bullous diseases tested positive for anti-desmoglein autoantibodies and the frequencies of HLA-DRB1\*04, HLA-DQB1\*03, \*0303, and \*05, and HLA-DPB1\*0402 and \*0501 alleles were higher in these patients than in a healthy Japanese control population in the literature [29,30]. Moreover, the relationship between the autoantibodies in silicosis patients and HLA phenotypes was also analyzed in silicosis patients with anti-topoisomerase I autoantibodies [30-34], and the results obtained revealed that the allelic frequency of HLA-DQB1\*0402 was significantly higher in anti-topoisomerase I positive Japanese silicosis patients than in anti-topoisomerase I negative patients or healthy donors [30-34].
We also assessed autoantibodies against Fas/CD95 [35], the cell death receptor, which plays an important role in the apoptosis of lymphocytes, and caspase-8 [36,37]. This anti-Fas autoantibody, in particular, was shown to induce apoptosis in Fas-expressing cells [35].
Even silicosis patients without any clinical symptoms of autoimmune diseases have various inapparent alterations in self-tolerance depending on individual factors, such as HLA pheno‐ types. In addition, when both respiratory and immunological factors were analyzed using factor analysis, this immunological progression was not concomitant with the development of respiratory disease [38]. Respiratory and immunological factors were shown to deteriorate to varying degrees in more than half of silicosis patients; however, a subpopulation was classified as a better respiratory and worse immunological group, while the opposite group was also reported [38].
Therefore, we attempted to confirm whether silica particles directly stimulated human immune cells, particularly T cells, with experimental evidence.
#### **3. Schematic summary of the chronic activation of T cells by silica particles**
A summary of the findings obtained and considerations are schematically presented in Figure 1. Silica particles were shown to chronically activate various T cells. Previous studies reported that effector T cells expressed various activation markers such as PD-1 and CD25 and produced many molecular markers for chronic T cell activation such as soluble Fas (sFas), decoy receptor 3 (DcR3), and soluble interleukin (IL)-2 receptor (sIL-2R) [39,40].
On the other hand, silica particles were also shown to activate CD4+CD25+FoxP3+regulatory T cells (Treg). However, this chronic activation caused the enhanced expression of Fas/CD95 on the surface of Treg, which induced early apoptosis [41]. Therefore, Treg may be lost from the peripheral blood, with the resulting imbalance between Treg and effector T cells subse‐ quently leading to autoimmune dysregulation.
Detailed explanations of these findings are presented below.
disorders in silicosis patients as well as providing an insight into the etiology of various
We previously investigated the immunological effects of silica using human peripheral blood immune cells derived from healthy donors and silicosis patients [25-27]. In this review, we summarized our findings, in which silica was shown to be an environmental immunostimu‐ lator and the chronic activation of immune cells by recurrent and chronic exposure to silica
**2. Autoantibodies detected in silicosis patients and their relationship with**
We previously measured autoantibodies in silicosis patients, all of whom were Japanese brickyard workers in Bizen city, Okayama prefecture, Japan. All patients were diagnosed with silicosis based on their radiological findings in accordance with the ILO 2000 guidelines. The amount of free silica inhaled by these patients was estimated to be from 40 to 60%, as deter‐ mined by their work environment. No subjects exhibited any symptoms of autoimmune
We demonstrated that the overall titer of anti-nuclear antibodies (ANA) was higher in these patients than in healthy volunteers [28]. In addition, silicosis patients without bullous diseases tested positive for anti-desmoglein autoantibodies and the frequencies of HLA-DRB1\*04, HLA-DQB1\*03, \*0303, and \*05, and HLA-DPB1\*0402 and \*0501 alleles were higher in these patients than in a healthy Japanese control population in the literature [29,30]. Moreover, the relationship between the autoantibodies in silicosis patients and HLA phenotypes was also analyzed in silicosis patients with anti-topoisomerase I autoantibodies [30-34], and the results obtained revealed that the allelic frequency of HLA-DQB1\*0402 was significantly higher in anti-topoisomerase I positive Japanese silicosis patients than in anti-topoisomerase I negative
We also assessed autoantibodies against Fas/CD95 [35], the cell death receptor, which plays an important role in the apoptosis of lymphocytes, and caspase-8 [36,37]. This anti-Fas autoantibody, in particular, was shown to induce apoptosis in Fas-expressing cells [35].
Even silicosis patients without any clinical symptoms of autoimmune diseases have various inapparent alterations in self-tolerance depending on individual factors, such as HLA pheno‐ types. In addition, when both respiratory and immunological factors were analyzed using factor analysis, this immunological progression was not concomitant with the development of respiratory disease [38]. Respiratory and immunological factors were shown to deteriorate to varying degrees in more than half of silicosis patients; however, a subpopulation was classified as a better respiratory and worse immunological group, while the opposite group was also
Therefore, we attempted to confirm whether silica particles directly stimulated human
immune cells, particularly T cells, with experimental evidence.
diseases such as sclerotic skin, Raynaud's phenomenon, facial erythema, or cancer.
was demonstrated to cause an imbalance in the regulation of T cell responses.
autoimmune diseases.
250 Immune Response Activation
**HLA phenotypes**
patients or healthy donors [30-34].
reported [38].
#### **4. Immune stimulation of effector T cells by silica particles**
Freshly isolated peripheral blood mononuclear cells (PBMCs) obtained from healthy donors were incubated with silica particles. The CD69 expression on the membrane surface was examined as a marker, demonstrating gradual activation of T lymphocytes during 10-day incubation [42].
Other activation markers were examined in serum derived from silicosis patients and com‐ pared with those from healthy donors [43]. The results showed that sIL-2R levels were slightly higher in silicosis patients than in healthy donors. sIL-2R levels were also examined in serum derived from SSc patients, and correlations between sIL-2R levels and the immunological status of healthy donors, silicosis patients, and SSc patients as 1, 2 and 3 for continuous variables were analyzed. The correlation coefficient was shown to be 0.575 with p=0.0008, which indicated that, from the viewpoint of immunological alterations based on serum sIL-2R levels, silicosis patients were located between healthy donors and SSc patients. Elevated sIL-2R levels may be a pathophysiological marker for hematological malignancies such as human T lymphotropic retrovirus type-1 (HTLV-1) associated with adult T cell leukemia (ATL) and hairy cell leukemia, which reflects the increased production of cells leading to an elevation in serum titers [44-47]. Elevated sIL-2R levels have recently been reported in various autoimmune or inflammatory diseases, suggesting that the immune response is activated by chronic stimulation of T cells with an auto-or foreign antigen [48-51]. Therefore, the moderate, chronic activation status of the immune system may play a role in silicosis.
Similar to sIL-2R, DcR3 has been identified as a chronic activation marker for the human immune response [52]. DcR3 was initially discovered in malignant cells such as lung and colon cancers [53], and its role was considered to be that of a protective molecule binding with the TNF-related apoptosis-inducing ligand (TRAIL) or the Fas ligand secreted from tumorattacking T cells [54]. These functions are similar to the soluble Fas molecule, which is an alternative splicing form of wild-type membrane Fas secreted from lymphocytes due to the absence of a transmembrane domain. Soluble Fas has also been shown to bind to the Fas ligand in extracellular areas, which prevented Fas ligand-inducing and Fas–mediated apoptosis in T
Immunostimulation by Silica Particles and the Development of Autoimmune Dysregulation
http://dx.doi.org/10.5772/57544
253
Elevated DcR3 levels were recently reported in the serum or pathological lesions of patients with various autoimmune diseases such as RA and SLE, and these findings indicated that the production of high levels of DcR3 may reflect chronic activation of the immune system
We previously demonstrated that the expression of DcR3 mRNA in PBMCs was higher in silicosis patients than in healthy donors [63]. Although the expression of DcR3 mRNA was only examined in whole PBMCs including lymphocytes and monocytes, taken together with recent findings showing elevated DcR3 levels in autoimmune diseases, these results suggest that examining serum DcR3 levels in silicosis patients is of importance. We have started this
sFas has been shown to have a similar role to that of lymphocytes. Although its molecular function is to prevent apoptosis, elevations in sFas levels have been reported in serum from patients with various autoimmune diseases [62-64] as well as silicosis patients [67]. Using PBMCs, the sFas and membrane (wild-type) Fas message expression ratio was also shown to be higher in silicosis patients than in healthy donors [68]. Our findings also revealed that peripheral T cells, which produce lower levels of surface membrane Fas, were the producers/ expressers of sFas, whereas normal (relatively higher) surface membrane Fas-expressing T cells produced lower levels of sFas [25]. Fas-mediated apoptosis may proceed more easily in the latter fraction due to various stimuli by self-or foreign antigens or the anti-Fas autoanti‐ body, which we discovered in the serum of silicosis patients, and repeated recruitment from the bone marrow. Peripheral T cells derived from silicosis patients were shown to be the dominant sFas producer with a smaller fraction of apoptosis-prone T cells than that from healthy donors [25-27]. The sFas-producing fraction may survive longer and retain a chroni‐ cally activated status. Thus, this fraction may be stimulated and respond to autoantigens.
Similar to DcR3, the higher expression and production of sFas suggest that the peripheral blood of silicosis patients frequently includes a chronically activated and self-antigen recognizing T
cells [55-58].
cell fraction.
These findings are summarized in Figure 2.
[52,59-62], particularly antigen-recognizing T cells.
analysis and will report our findings in the near future.
**Figure 1.** Schematic summary of the chronic activation of effector T cells, regulatory T cells (Treg; CD4+25+FoxP3+), and Th17 caused by exposure to silica particles. The chronic activation of effector T cells caused these cells to express activation surface markers such as CD25 and PD-1 with a decrease in the expression of the Fas/CD95 molecule. In‐ stead of a reduction in membrane Fas, these cells produced soluble Fas, an alternative spliced form of wild-type Fas, and DcR3 as well as the soluble IL-2 receptor (sIL-2R). sFas and DcR3 prevented fas-ligand/Trail-induced apoptosis, leading to longer survival. sIL-2R and DcR3 as well as sFas were markers for the chronic activation of effector T cells. These might be explained by autoantigen-recognizing cells contaminating the long-term surviving cells. Treg en‐ hanced expression of the surface Fas receptor and become prone to Fas-mediated apoptosis as a result of chronic acti‐ vation. Treg may die quickly in silicosis patients (even when repeatedly recruited from the bone marrow). The overall balance between Treg and reactive T cells moves toward decreased Treg, resulting in the subsequent aberrant regula‐ tion of autoimmunity.
Similar to sIL-2R, DcR3 has been identified as a chronic activation marker for the human immune response [52]. DcR3 was initially discovered in malignant cells such as lung and colon cancers [53], and its role was considered to be that of a protective molecule binding with the TNF-related apoptosis-inducing ligand (TRAIL) or the Fas ligand secreted from tumorattacking T cells [54]. These functions are similar to the soluble Fas molecule, which is an alternative splicing form of wild-type membrane Fas secreted from lymphocytes due to the absence of a transmembrane domain. Soluble Fas has also been shown to bind to the Fas ligand in extracellular areas, which prevented Fas ligand-inducing and Fas–mediated apoptosis in T cells [55-58].
Elevated DcR3 levels were recently reported in the serum or pathological lesions of patients with various autoimmune diseases such as RA and SLE, and these findings indicated that the production of high levels of DcR3 may reflect chronic activation of the immune system [52,59-62], particularly antigen-recognizing T cells.
We previously demonstrated that the expression of DcR3 mRNA in PBMCs was higher in silicosis patients than in healthy donors [63]. Although the expression of DcR3 mRNA was only examined in whole PBMCs including lymphocytes and monocytes, taken together with recent findings showing elevated DcR3 levels in autoimmune diseases, these results suggest that examining serum DcR3 levels in silicosis patients is of importance. We have started this analysis and will report our findings in the near future.
sFas has been shown to have a similar role to that of lymphocytes. Although its molecular function is to prevent apoptosis, elevations in sFas levels have been reported in serum from patients with various autoimmune diseases [62-64] as well as silicosis patients [67]. Using PBMCs, the sFas and membrane (wild-type) Fas message expression ratio was also shown to be higher in silicosis patients than in healthy donors [68]. Our findings also revealed that peripheral T cells, which produce lower levels of surface membrane Fas, were the producers/ expressers of sFas, whereas normal (relatively higher) surface membrane Fas-expressing T cells produced lower levels of sFas [25]. Fas-mediated apoptosis may proceed more easily in the latter fraction due to various stimuli by self-or foreign antigens or the anti-Fas autoanti‐ body, which we discovered in the serum of silicosis patients, and repeated recruitment from the bone marrow. Peripheral T cells derived from silicosis patients were shown to be the dominant sFas producer with a smaller fraction of apoptosis-prone T cells than that from healthy donors [25-27]. The sFas-producing fraction may survive longer and retain a chroni‐ cally activated status. Thus, this fraction may be stimulated and respond to autoantigens.
Similar to DcR3, the higher expression and production of sFas suggest that the peripheral blood of silicosis patients frequently includes a chronically activated and self-antigen recognizing T cell fraction.
These findings are summarized in Figure 2.
**Figure 1.** Schematic summary of the chronic activation of effector T cells, regulatory T cells (Treg; CD4+25+FoxP3+), and Th17 caused by exposure to silica particles. The chronic activation of effector T cells caused these cells to express activation surface markers such as CD25 and PD-1 with a decrease in the expression of the Fas/CD95 molecule. In‐ stead of a reduction in membrane Fas, these cells produced soluble Fas, an alternative spliced form of wild-type Fas, and DcR3 as well as the soluble IL-2 receptor (sIL-2R). sFas and DcR3 prevented fas-ligand/Trail-induced apoptosis, leading to longer survival. sIL-2R and DcR3 as well as sFas were markers for the chronic activation of effector T cells. These might be explained by autoantigen-recognizing cells contaminating the long-term surviving cells. Treg en‐ hanced expression of the surface Fas receptor and become prone to Fas-mediated apoptosis as a result of chronic acti‐ vation. Treg may die quickly in silicosis patients (even when repeatedly recruited from the bone marrow). The overall balance between Treg and reactive T cells moves toward decreased Treg, resulting in the subsequent aberrant regula‐
tion of autoimmunity.
252 Immune Response Activation
However, FoxP3-positive cells cannot be used in the experiment, since the use of the collected cells in the subsequent cell biological experiment precludes the permeabilization of the cell membrane because of the staining of nuclear molecules, such as FoxP3. Thus, it is unknown whether the CD4+25+ fraction used in the experiment was pure Treg or a mixture of chroni‐ cally-activated reactive T cells. In other words, it is unknown whether the reduced functions of the Treg fraction with peripheral CD4+25+ in the silicosis patients was caused by the impurity of the Treg cells or the contamination of chronically-activated CD4+25+reactive T
Immunostimulation by Silica Particles and the Development of Autoimmune Dysregulation
http://dx.doi.org/10.5772/57544
255
Therefore, we examined the expression of surface Fas in peripheral CD4+FoxP+ T cells derived from both silicosis patients and healthy donors [41], as shown in Figure 3. The results obtained revealed that the expression of Fas was higher in Treg from silicosis patients than in those from healthy donors. Since when Treg is stimulated, Fas expression was shown to be one of the markers for activated Treg; therefore, Treg may be a self-limited inhibitor for the immune response [69,70] and should be terminated by activation-induced cell death. Taken together, these results indicate that exposure to silica may activate Treg as well as effector T cells and
PBMCs from healthy donors and silicosis patients were incubated with silica particles for four days and the percentage of CD4+FoxP3+ cells was then measured [41]. As shown in Figure 3 and reported previously, the frequency of apoptosis-induced Treg cell death during cultivation with silica particles was higher in PBMCs from silicosis patients than in that from healthy
These results demonstrated that silica activated Treg, which then produced higher levels of surface Fas. Apoptosis then occurred in activated Treg. This cascaded reaction can continue for a long time in silicosis because of recurrent encounters between silica particles and T cells. The early loss of Treg may cause T cell recruitment. However, the overall balance between long-surviving reactive T and Treg cells will move toward predominance of reactive T cells
The mechanism by which silica affects Th17 cells has not yet been established. Th17 cells are considered to play an important role in the autoimmune reaction and increases in the Th17 cell population or typical cytokines produced by Th17 cells, including IL-17A, may be related to autoimmune reactions [71,72]. However, the microenvironment surrounding the develop‐ ment of T lymphocytes, defined by cytokine profiles such as IL-6 and TGF-β, may affect the developmental direction of both Treg and Th17 cells. Therefore, importance needs to be placed on investigations of how silica particles cause changes in the cellular and molecular charac‐ teristics of Th17 cells, and what the link is between these alterations and autoimmune dysre‐
cells.
donors.
[41].
gulation in silicosis patients.
induce the higher production of Fas by Treg.
**5. Current issues in immune stimulation by silica**
**Figure 2.** Evidence of chronically-activated effector T cells in silicosis patients. Serum soluble Fas, the soluble/ membrane Fas mRNA expression ratio in PBMCs, and DcR3 expression and serum sIL-2R levels were higher in silicosis patients than in healthy donors. In addition, immunological abnormality levels in silicosis patients were determined to be 1, 2, and 3 (normal individuals and silicosis and SSc patients) as continuous variables between healthy donors and SSc patients. Then, the serum sIL-2R levels were positively correlated with these immunological scores. All of these findings indicate that effector T cells chronically and recurrently exposed become activated and their cellular features proceed to autoimmune dysregulation.
#### **5. Immune stimulation of regulatory T cells by silica particle**
From the beginning of Treg analysis in silicosis patients, the CD4+CD25+ fraction from PBMCs derived from these patients was shown to be less functional than that from healthy donors [28]. However, FoxP3-positive cells cannot be used in the experiment, since the use of the collected cells in the subsequent cell biological experiment precludes the permeabilization of the cell membrane because of the staining of nuclear molecules, such as FoxP3. Thus, it is unknown whether the CD4+25+ fraction used in the experiment was pure Treg or a mixture of chroni‐ cally-activated reactive T cells. In other words, it is unknown whether the reduced functions of the Treg fraction with peripheral CD4+25+ in the silicosis patients was caused by the impurity of the Treg cells or the contamination of chronically-activated CD4+25+reactive T cells.
Therefore, we examined the expression of surface Fas in peripheral CD4+FoxP+ T cells derived from both silicosis patients and healthy donors [41], as shown in Figure 3. The results obtained revealed that the expression of Fas was higher in Treg from silicosis patients than in those from healthy donors. Since when Treg is stimulated, Fas expression was shown to be one of the markers for activated Treg; therefore, Treg may be a self-limited inhibitor for the immune response [69,70] and should be terminated by activation-induced cell death. Taken together, these results indicate that exposure to silica may activate Treg as well as effector T cells and induce the higher production of Fas by Treg.
PBMCs from healthy donors and silicosis patients were incubated with silica particles for four days and the percentage of CD4+FoxP3+ cells was then measured [41]. As shown in Figure 3 and reported previously, the frequency of apoptosis-induced Treg cell death during cultivation with silica particles was higher in PBMCs from silicosis patients than in that from healthy donors.
These results demonstrated that silica activated Treg, which then produced higher levels of surface Fas. Apoptosis then occurred in activated Treg. This cascaded reaction can continue for a long time in silicosis because of recurrent encounters between silica particles and T cells. The early loss of Treg may cause T cell recruitment. However, the overall balance between long-surviving reactive T and Treg cells will move toward predominance of reactive T cells [41].
#### **5. Current issues in immune stimulation by silica**
**Figure 2.** Evidence of chronically-activated effector T cells in silicosis patients. Serum soluble Fas, the soluble/ membrane Fas mRNA expression ratio in PBMCs, and DcR3 expression and serum sIL-2R levels were higher in silicosis patients than in healthy donors. In addition, immunological abnormality levels in silicosis patients were determined to be 1, 2, and 3 (normal individuals and silicosis and SSc patients) as continuous variables between healthy donors and SSc patients. Then, the serum sIL-2R levels were positively correlated with these immunological scores. All of these findings indicate that effector T cells chronically and recurrently exposed become activated and their cellular features
From the beginning of Treg analysis in silicosis patients, the CD4+CD25+ fraction from PBMCs derived from these patients was shown to be less functional than that from healthy donors [28].
**5. Immune stimulation of regulatory T cells by silica particle**
proceed to autoimmune dysregulation.
254 Immune Response Activation
The mechanism by which silica affects Th17 cells has not yet been established. Th17 cells are considered to play an important role in the autoimmune reaction and increases in the Th17 cell population or typical cytokines produced by Th17 cells, including IL-17A, may be related to autoimmune reactions [71,72]. However, the microenvironment surrounding the develop‐ ment of T lymphocytes, defined by cytokine profiles such as IL-6 and TGF-β, may affect the developmental direction of both Treg and Th17 cells. Therefore, importance needs to be placed on investigations of how silica particles cause changes in the cellular and molecular charac‐ teristics of Th17 cells, and what the link is between these alterations and autoimmune dysre‐ gulation in silicosis patients.
#### **6. Immunological effects of asbestos, a mineral silicate**
Asbestos is a mineral silicate, in which the chemical core atom is Si and O, and various metals such as iron, magnesium or calcium have been shown to bind to asbestos to chemically form asbestos fibers [73]. However, the physical properties of silica and asbestos are different. The former is particulate matter while the latter is a fibrous mineral. Although silica has been shown to have an effect on the human immune system, as mentioned above, asbestos fibers may also have immunological effects on human lymphocytes, which may alter human antitumor immunity because the most important clinical complication in patients exposed to asbestos is malignant tumors.
Although reports of autoimmune diseases in asbestos-exposed patients are rare [74,75], the main complication noted in these patients is malignancies such as mesothelioma and lung cancer [76,77]. In addition, the incidence of cancer in the larynx, GI-tract, and bladder was
Immunostimulation by Silica Particles and the Development of Autoimmune Dysregulation
http://dx.doi.org/10.5772/57544
257
We previously examined the immunological effects of asbestos [78,79], and demonstrated that temporary and relatively high-dose exposure to asbestos caused apoptosis in T cells as well as alveolar epithelial cells and mesothelial cells, whereas low-dose and continuous exposure to asbestos caused the acquisition of resistance to asbestos-caused apoptosis in T cells with the activation of Scr-family kinase, IL-10, l STAT3, and Bcl-2 [80]. The expression of CXCR3, one of the chemokine receptors related to the anti-tumor immunity, as well as IFNγ was also reduced in these cells [81,82]. Asbestos exposure also induced the impaired differentiation and proliferation of CD8+ cytotoxic T cells [83], and the reduced expression of NKp46 activating receptor in NK cells [84,85]. Taken together, these findings indicate that asbestos causes a
Even if core chemical substances, Si and O2, are identical, the immunological effects of silica seem to be completely opposite to those of asbestos. Silica is a chronic stimulator of T cells, with chronic exposure leading to autoimmune dysregulation due to the chronic activation of responder T cells as well as Treg, resulting in an imbalance in the regulation of T cell responses. On the other hand, asbestos reduces anti-tumor immunity. Therefore, asbestos is not a
The authors express special thanks to Prof. Ayako Ueki, former professor of the Department of Hygiene, Kawasaki Medical School, and Drs. Fuminori Hyodoh, Akiko Takata-Tomokuni, Yoshie Miura, and Shuko Murakami, Ping Wu, Chen Ying, former members, who assisted in achieving the presented experimental findings. In addition, the authors wish thank to Ms. Tamayo Hatayama, Shoko Yamamoto, Haruko Sakaguchi, Satomi Hatada, Yumika Isozaki, Yoshiko Yamashita, Keiko Kimura, Tomoko Sueishi, Misao Kuroki, Minako Katoh, and Naomi Miyahara, present and former technical assistants. The experimental studies performed by the authors were supported in part by Special Coordination Funds for Promoting Science and Technology (H18-1-3-3-1, Comprehensive approach on asbestos-related diseases), KAKENHI grants (18390186, 19659153 and 20390178), Kawasaki Medical School Project Grants (23S5, 23S6, 23B66, 23P3, 24S4,24S6, 24B39, 25B41, 25B65, 25B67 and 25B82), a Sumitomo Foundation Grant (053027), a Yasuda Memorial Foundation Grant (H18), and funding from the Takeda
shown to be high in asbestos-exposed patients [76,77].
stimulator, but can alter the function of immune cells.
reduction in anti-tumor immunity.
**7. Conclusions**
**Acknowledgements**
Science Foundation (I-2008).
**Figure 3.** Surface Fas expression in peripheral CD4+FoxP+ T cells (Treg) derived from both silicosis patients and healthy donors. The expression of Fas was higher in the Treg of silicosis patients than in those of healthy donors. PBMCs from healthy donors and silicosis patients were incubated with silica particles for four days and the percentage of CD4+FoxP3+ cells was then measured. The frequency of apoptosis-induced Treg cell death during cultivation with sili‐ ca particles was higher in PBMCs from silicosis patients than in that from healthy donors.
Although reports of autoimmune diseases in asbestos-exposed patients are rare [74,75], the main complication noted in these patients is malignancies such as mesothelioma and lung cancer [76,77]. In addition, the incidence of cancer in the larynx, GI-tract, and bladder was shown to be high in asbestos-exposed patients [76,77].
We previously examined the immunological effects of asbestos [78,79], and demonstrated that temporary and relatively high-dose exposure to asbestos caused apoptosis in T cells as well as alveolar epithelial cells and mesothelial cells, whereas low-dose and continuous exposure to asbestos caused the acquisition of resistance to asbestos-caused apoptosis in T cells with the activation of Scr-family kinase, IL-10, l STAT3, and Bcl-2 [80]. The expression of CXCR3, one of the chemokine receptors related to the anti-tumor immunity, as well as IFNγ was also reduced in these cells [81,82]. Asbestos exposure also induced the impaired differentiation and proliferation of CD8+ cytotoxic T cells [83], and the reduced expression of NKp46 activating receptor in NK cells [84,85]. Taken together, these findings indicate that asbestos causes a reduction in anti-tumor immunity.
#### **7. Conclusions**
**6. Immunological effects of asbestos, a mineral silicate**
asbestos is malignant tumors.
256 Immune Response Activation
Asbestos is a mineral silicate, in which the chemical core atom is Si and O, and various metals such as iron, magnesium or calcium have been shown to bind to asbestos to chemically form asbestos fibers [73]. However, the physical properties of silica and asbestos are different. The former is particulate matter while the latter is a fibrous mineral. Although silica has been shown to have an effect on the human immune system, as mentioned above, asbestos fibers may also have immunological effects on human lymphocytes, which may alter human antitumor immunity because the most important clinical complication in patients exposed to
**Figure 3.** Surface Fas expression in peripheral CD4+FoxP+ T cells (Treg) derived from both silicosis patients and healthy donors. The expression of Fas was higher in the Treg of silicosis patients than in those of healthy donors. PBMCs from healthy donors and silicosis patients were incubated with silica particles for four days and the percentage of CD4+FoxP3+ cells was then measured. The frequency of apoptosis-induced Treg cell death during cultivation with sili‐
ca particles was higher in PBMCs from silicosis patients than in that from healthy donors.
Even if core chemical substances, Si and O2, are identical, the immunological effects of silica seem to be completely opposite to those of asbestos. Silica is a chronic stimulator of T cells, with chronic exposure leading to autoimmune dysregulation due to the chronic activation of responder T cells as well as Treg, resulting in an imbalance in the regulation of T cell responses. On the other hand, asbestos reduces anti-tumor immunity. Therefore, asbestos is not a stimulator, but can alter the function of immune cells.
#### **Acknowledgements**
The authors express special thanks to Prof. Ayako Ueki, former professor of the Department of Hygiene, Kawasaki Medical School, and Drs. Fuminori Hyodoh, Akiko Takata-Tomokuni, Yoshie Miura, and Shuko Murakami, Ping Wu, Chen Ying, former members, who assisted in achieving the presented experimental findings. In addition, the authors wish thank to Ms. Tamayo Hatayama, Shoko Yamamoto, Haruko Sakaguchi, Satomi Hatada, Yumika Isozaki, Yoshiko Yamashita, Keiko Kimura, Tomoko Sueishi, Misao Kuroki, Minako Katoh, and Naomi Miyahara, present and former technical assistants. The experimental studies performed by the authors were supported in part by Special Coordination Funds for Promoting Science and Technology (H18-1-3-3-1, Comprehensive approach on asbestos-related diseases), KAKENHI grants (18390186, 19659153 and 20390178), Kawasaki Medical School Project Grants (23S5, 23S6, 23B66, 23P3, 24S4,24S6, 24B39, 25B41, 25B65, 25B67 and 25B82), a Sumitomo Foundation Grant (053027), a Yasuda Memorial Foundation Grant (H18), and funding from the Takeda Science Foundation (I-2008).
### **Author details**
Suni Lee1 , Hiroaki Hayashi2 , Megumi Maeda3 , Hidenori Matsuzaki1 , Naoko Kumagai-Takei1 , Ying Chen4 , Kozo Urakami5 , Masayasu Kusaka6 , Yasumitsu Nishimura1 and Takemi Otsuki1\* [8] Mortensen JT, Brønnum-Hansen H, Rasmussen K. Multiple sclerosis and organic sol‐
Immunostimulation by Silica Particles and the Development of Autoimmune Dysregulation
http://dx.doi.org/10.5772/57544
259
[9] World Health Organization. Silica, Some Silicates, Coal Dust and Para-Aramid Fi‐
[11] Uber CL, McReynolds RA. Immunotoxicology of silica. Critical Reviews in Toxicolo‐
[12] Steenland K, Goldsmith DF. Silica exposure and autoimmune diseases. American
[13] Caplan A. Rheumatoid disease and pneumoconiosis (Caplan's syndrome). Proceed‐
[14] Sluis-Cremer GK, Hessel PA, Hnizdo E Churchill AR. Relationship between silicosis
[15] Sluis-Cremer, G.K., Hessel, P.A., Nizdo, E.H., Churchill, A.R. and Zeiss, E.A. Silica, silicosis, and progressive systemic sclerosis. British Journal of Industrial Medicine
[16] Barnadas MA, Tuneu A, Rajmil HO, Abud O, de Moragas JM. Impotence in silicosisassociated scleroderma. Journal of the American Academy of Dermatology
[17] Cowie RL. Silica-dust-exposed mine workers with scleroderma (systemic sclerosis).
[18] Sanchez-Roman J, Wichmann I, Salaberri J, Varela JM, Nuñez-Roldan A. Multiple clinical and biological autoimmune manifestations in 50 workers after occupational
[19] Brown LM, Gridley G, Olsen JH, Mellemkjaer L, Linet MS, Fraumeni JF. Jr. Cancer risk and mortality patterns among silicotic men in Sweden and Denmark. Journal of
[20] Gregorini G, Ferioli A, Donato F, Tira P, Morassi L, Tardanico R, Lancini L, Maiorca R. Association between silica exposure and necrotizing crescentic glomerulonephritis with p-ANCA and anti-MPO antibodies: a hospital-based case-control study. Advan‐
[21] Gregorini G, Tira P, Frizza J, D'Haese PC, Elseviers MM, Nuyts G, Maiorca R, De Broe ME. ANCA-associated diseases and silica exposure. Clinical Reviews in Allergy
exposure to silica. Annals of the Rheumatic Diseases 1993;52(7) 534-8.
Occupational and Environmental Medicine 1997;39(7) 633-8.
ces in Experimental Medicine and Biology 1993;336 435-40.
and Immunology 1997;15(1) 21-40.
brils (IARC Monographs No. 68), 1997. ISBN: 978-9-2832126-83. [10] Leung CC, Yu IT, Chen W. Silicosis. Lancet 2012;379(9830) 2008-18.
vents. Epidemiology 1998;9(2) 168-71.
Journal of Industrial Medicine 1995;28(5) 603-8.
ings of the Royal Society of Medicine 1959;52 1111-3.
and rheumatoid arthritis. Thorax 1986;41(8) 596-601.
gy 1982;10(4) 303-19.
1985;42(12) 838-43.
1986;15(6) 1294-6.
Chest 1987;92(2) 260-2.
\*Address all correspondence to: takemi@med.kawasaki-m.ac.jp
3 Department of Biofunctional Chemistry, Division of Bioscience, Okayama University Graduate School of Natural Science and Technology, Okayama, Japan
4 Division of Pneumoconiosis, School of Public Health, China Medical University, She‐ nyang, China
#### **References**
**Author details**
258 Immune Response Activation
, Hiroaki Hayashi2
5 Hinase Urakami Iin, Bizen, Japan
6 Kusaka Hospital, Bizen, Japan
, Kozo Urakami5
, Megumi Maeda3
1 Department of Hygiene, Kawasaki Medical School, Kurashiki, Japan
Graduate School of Natural Science and Technology, Okayama, Japan
2 Department of Dermatology, Kawasaki Medical School, Kurashiki, Japan
3 Department of Biofunctional Chemistry, Division of Bioscience, Okayama University
4 Division of Pneumoconiosis, School of Public Health, China Medical University, She‐
[1] Cooper GS, Miller FW, Germolec DR. Occupational exposures and autoimmune dis‐
[2] Pollard KM. Gender differences in autoimmunity associated with exposure to envi‐
[3] Haustein UF, Ziegler V. Environmentally induced systemic sclerosis-like disorders.
[4] Fox RI, Kang HI. Genetic and environmental factors in systemic sclerosis. Current
[5] Nietert PJ, Silver RM. Systemic sclerosis: environmental and occupational risk fac‐
[6] Lundberg I, Alfredsson L, Plato N, Sverdrup B, Klareskog L, Kleinau S. Occupation, occupational exposure to chemicals and rheumatological disease. A register based
[7] Landtblom AM, Flodin U, Söderfeldt B, Wolfson C, Axelson O. Organic solvents and multiple sclerosis: a synthesis of the current evidence. Epidemiology 1996;7(4)
cohort study. Scandinavian Journal of Rheumatology 1994;23(6) 305-10.
eases. International Immunopharmacology 2002;2(2-3) 303-13.
ronmental factors. Journal of Autoimmunity 2012;38(2-3) J177-86.
International Journal of Dermatology 1985; 24(3) 147-51.
tors. Current Opinion in Rheumatology 2000;12(6) 520-6.
Opinion in Rheumatology 1992;4(6) 857-61.
\*Address all correspondence to: takemi@med.kawasaki-m.ac.jp
, Masayasu Kusaka6
, Hidenori Matsuzaki1
, Yasumitsu Nishimura1
, Naoko Kumagai-Takei1
and Takemi Otsuki1\*
,
Suni Lee1
Ying Chen4
nyang, China
**References**
429-33.
[22] Wichmann I, Sanchez-Roman J, Morales J, Castillo MJ, Ocaña C, Nuñez-Roldan A. Antimyeloperoxidase antibodies in individuals with occupational exposure to silica. Annals of the Rheumatic Diseases 1996;55(3) 205-7.
[32] Ueki A, Isozaki Y, Tomokuni A, Otsuki T, Hydoh F, Sakaguchi H, Tanaka S, Kusaka M. Is the anti-topoisomerase I autoantibody response associated with a distinct ami‐ no acid sequence in the HLA-DQbeta1 domain? Arthritis and Rheumatism 2001;44(2)
Immunostimulation by Silica Particles and the Development of Autoimmune Dysregulation
http://dx.doi.org/10.5772/57544
261
[33] Ueki A, Isozaki Y, Tomokuni A, Ueki H, Kusaka M, Tanaka S, Otsuki T, Sakaguchi H, Hyodoh F. Different distribution of HLA class II alleles in anti-topoisomerase I autoantibody responders between silicosis and systemic sclerosis patients, with a common distinct amino acid sequence in the HLA-DQB1 domain. Immunobiology
[34] Tomokuni A, Otsuki T, Sakaguchi H, Isozaki Y, Hyodoh F, Kusaka M, Ueki A. Detec‐ tion of anti-topoisomerase I autoantibody in patients with silicosis. Environmental
[35] Takata-Tomokuni A, Ueki A, Shiwa M, Isozaki Y, Hatayama T, Katsuyama H, Hyo‐ doh F, Fujimoto W, Ueki H, Kusaka M, Arikuni H, Otsuki T. Detection, epitope-map‐ ping and function of anti-Fas autoantibody in patients with silicosis. Immunology
[36] Ueki A, Isozaki Y, Tomokuni A, Hatayama T, Ueki H, Kusaka M, Shiwa M, Arikuni H, Takeshita T, Morimoto K. Intramolecular epitope spreading among anti-caspase-8 autoantibodies in patients with silicosis, systemic sclerosis and systemic lupus eryth‐ ematosus, as well as in healthy individuals. Clinical and Experimental Immunology
[37] Ueki A, Isozaki Y, Kusaka M. Anti-caspase-8 autoantibody response in silicosis pa‐ tients is associated with HLA-DRB1, DQB1 and DPB1 alleles. Journal of Occupational
[38] Otsuki T, Ichihara K, Tomokuni A, Sakaguchi H, Aikoh T, Matsuki T, Isozaki Y, Hyo‐ doh F, Kusaka M, Kita S, Ueki A. Evaluation of cases with silicosis using the parame‐ ters related to Fas-mediated apoptosis. International Journal of Molecular Medicine
[39] Maeda M, Nishimura Y, Kumagai N, Hayashi H, Hatayama T, Katoh M, Miyahara N, Yamamoto S, Hiratsuka J, Otsuki T. Dysregulation of the immune system caused
[40] Lee S, Hayashi H, Maeda M, Chen Y, Matsuzaki H, Takei-Kumagai N, Nishimura Y, Fujimoto W, Otsuki T. Environmental factors producing autoimmune dysregula‐ tion--chronic activation of T cells caused by silica exposure.. Immunobiology
[41] Hayashi H, Miura Y, Maeda M, Murakami S, Kumagai N, Nishimura Y, Kusaka M, Urakami K, Fujimoto W, Otsuki T. Reductive alteration of the regulatory function of the CD4(+)CD25(+) T cell fraction in silicosis patients. International Journal of Immu‐
by silica and asbestos. Journal of Immunotoxicology 2010;7(4) 268-78.
nopathology and Pharmacology 2010;23(4) 1099-109.
491-2.
2001;204(4) 458-65.
2005;116(1) 21-9.
2002;129(3) 556-61.
Health 2005;47(1) 61-7.
1999;4(4) 407-11.
2012;217(7) 743-8.
Health and Preventive Medicine 2002;7(1) 7-10.
[32] Ueki A, Isozaki Y, Tomokuni A, Otsuki T, Hydoh F, Sakaguchi H, Tanaka S, Kusaka M. Is the anti-topoisomerase I autoantibody response associated with a distinct ami‐ no acid sequence in the HLA-DQbeta1 domain? Arthritis and Rheumatism 2001;44(2) 491-2.
[22] Wichmann I, Sanchez-Roman J, Morales J, Castillo MJ, Ocaña C, Nuñez-Roldan A. Antimyeloperoxidase antibodies in individuals with occupational exposure to silica.
[23] Stratta P, Messuerotti A, Canavese C, Coen M, Luccoli L, Bussolati B, Giorda L, Mala‐ venda P, Cacciabue M, Bugiani M, Bo M, Ventura M, Camussi G, Fubini B. The role of metals in autoimmune vasculitis: epidemiological and pathogenic study. Science
[24] Cooper GS, Gilbert KM, Greidinger EL, James JA, Pfau JC, Reinlib L, Richardson BC, Rose NR. Recent advances and opportunities in research on lupus: environmental in‐ fluences and mechanisms of disease. Environmental Health Perspectives 2008;116(6)
[25] Otsuki T, Miura Y, Nishimura Y, Hyodoh F, Takata A, Kusaka M, Katsuyama H, To‐ mita M, Ueki A, Kishimoto T. Alterations of Fas and Fas-related molecules in pa‐ tients with silicosis. Experimental Biology and Medicine (Maywood) 2006;231(5)
[26] Otsuki., Maeda M, Murakami S, Hayashi H, Miura Y, Kusaka M, Nakano T, Fukuoka K, Kishimoto T, Hyodoh F, Ueki A, Nishimura Y. Immunological effects of silica and
[27] Kumagai N, Hayashi H, Maeda M, Miura Y, Nishimura Y, Matsuzaki H, Lee S, Fuji‐ moto W, Otsuki T. Immunological effects of silica and related dysregulation of auto‐ immunity. In Autoimmune Disorder Pathogenetic Aspects. (ed: Mavragani, C.P.) Intech Open Access Publisher, Rijeka: InTech; 2011 http://www.intechopen.com/ books/autoimmune-disorders-pathogenetic-aspects/immunological-effects-of-silica-
and-related-dysregulation-of-autoimmunity (accessed 22 November 2013).
[28] Wu P, Miura Y, Hyodoh F, Nishimura Y, Hatayama T, Hatada S, Sakaguchi H, Kusa‐ ka M, Katsuyama H, Tomita M, Otsuki T. Reduced function of CD4+25+regulatory T cell fraction in silicosis patients. International Journal of Immunopathology and
[29] Akaza T, Imanishi T, Fujiwara K, Tokunaga K, Yashiki S, Fujiyoshi T, Sonoda S, Tsuji T. HLA alleles and haplotypes of Japanese population, MHC IRS 1994;1 (Suppli)
[30] Ueki H, Kohda M, Nobutoh T, Yamaguchi M, Omori K, Miyashita Y, Hashimoto T, Komai A, Tomokuni A, Ueki A. Antidesmoglein autoantibodies in silicosis patients
[31] Ueki A, Isozaki Y, Tomokuni A, Tanaka S, Otsuki T, Kishimoto T, Kusaka M, Aikoh T, Sakaguchi H, Hydoh F. Autoantibodies detectable in the sera of silicosis patients. The relationship between the anti-topoisomerase I antibody response and HLA-DQB1\*0402 allele in Japanese silicosis patients. Science of the Total Environment
with no bullous diseases. Dermatology 2001:202(1) 16-21.
asbestos. Cellular and Molecular Immunology 2007;4(4) 261-8.
Annals of the Rheumatic Diseases 1996;55(3) 205-7.
of the Total Environment 2001;270(1-3) 179-90.
Pharmacology 2006;19(2) 357-68.
695-702.
260 Immune Response Activation
522-33.
219-26.
2001;270(1-3) 141-8.
[42] Wu P, Hyodoh F, Hatayama T, Sakaguchi H, Hatada S, Miura Y, Takata-Tomokuni A, Katsuyama H, Otsuki T. Induction of CD69 antigen expression in peripheral blood mononuclear cells on exposure to silica, but not by asbestos/chrysotile-A. Im‐ munology Letters 2005;98(1) 145-52.
Cohen RL, Goddard AD, Botstein D, Ashkenazi A. Genomic amplification of a decoy receptor for Fas ligand in lung and colon cancer. Nature 1998;396(6712) 699-703.
Immunostimulation by Silica Particles and the Development of Autoimmune Dysregulation
http://dx.doi.org/10.5772/57544
263
[55] Cheng J, Zhou T, Liu C, Shapiro JP, Brauer MJ, Kiefer MC, Barr PJ, Mountz JD. Pro‐ tection from Fas-mediated apoptosis by a soluble form of the Fas molecule. Science
[56] Orlinick JR, Chao MV. TNF-related ligands and their receptors. Cellular Signalling
[57] Nagata, S. (1994) Fas and Fas ligand: a death factor and its receptor. Adv Immunol.
[59] Chen CC, Yang YH, Lin YT, Hsieh SL, Chiang BL. Soluble decoy receptor 3: in‐ creased levels in atopic patients. Journal of Allergy and Clinical Immunology
[60] Hayashi S, Miura Y, Nishiyama T, Mitani M, Tateishi K, Sakai Y, Hashiramoto A, Kurosaka M, Shiozawa S, Doita M. Decoy receptor 3 expressed in rheumatoid syno‐ vial fibroblasts protects the cells against Fas-induced apoptosis. Arthritis and Rheu‐
[61] Lee CS, Hu CY, Tsai HF, Wu CS, Hsieh SL, Liu LC, Hsu PN. Elevated serum decoy receptor 3 with enhanced T cell activation in systemic lupus erythematosus. Clinical
[62] Han B, Bojalil R, Amezcua-Guerra LM, Springall R, Valderrama-Carvajal H, Wu J, Luo H. DcR3 as a diagnostic parameter and risk factor for systemic lupus erythema‐
[63] Otsuki T, Tomokuni A, Sakaguchi H, Aikoh T, Matsuki T, Isozaki Y, Hyodoh F, Ueki H, Kusaka M, Kita S, Ueki A. Over-expression of the decoy receptor 3 (DcR3) gene in peripheral blood mononuclear cells (PBMC) derived from silicosis patients. Clinical
[64] Knipping E, Krammer PH, Onel KB, Lehman TJ, Mysler E, Elkon KB. Levels of solu‐ ble Fas/APO-1/CD95 in systemic lupus erythematosus and juvenile rheumatoid ar‐
[65] Nozawa K, Kayagaki N, Tokano Y, Yagita H, Okumura K, Hasimoto H. Soluble Fas (APO-1, CD95) and soluble Fas ligand in rheumatic diseases. Arthritis and Rheuma‐
[66] Sahin M, Aydintug O, Tunc SE, Tutkak H, Naziroğlu M. Serum soluble Fas levels in patients with autoimmune rheumatic diseases. Clinical Biochemistry 2007;40(1-2)
[58] Nagata S, Golstein P. The Fas death factor. Science 1995;267(5203) 1449-56.
[54] Green DR. Apoptosis. Death deceiver. Nature 1998;396(6712) 629-30.
1994;263(5154) 1759-62.
57, 129-144, ISSN: 0065-2776
matism 2007;56(4) 1067-75.
tism 1997;40(6) 1126-9.
6-10.
and Experimental Immunology 2008;151(3) 383-90.
tosus. International Immunology 2008;20(8) 1067-75.
and Experimental Immunology 2000;119(2) 323-7.
thritis. Arthritis and Rheumatism 1995;38(12) 1735-7.
1998;10(8) 543-51.
2004;114(1) 195-7.
Cohen RL, Goddard AD, Botstein D, Ashkenazi A. Genomic amplification of a decoy receptor for Fas ligand in lung and colon cancer. Nature 1998;396(6712) 699-703.
[54] Green DR. Apoptosis. Death deceiver. Nature 1998;396(6712) 629-30.
[42] Wu P, Hyodoh F, Hatayama T, Sakaguchi H, Hatada S, Miura Y, Takata-Tomokuni A, Katsuyama H, Otsuki T. Induction of CD69 antigen expression in peripheral blood mononuclear cells on exposure to silica, but not by asbestos/chrysotile-A. Im‐
[43] Hayashi H, Maeda M, Murakami S, Kumagai N, Chen Y, Hatayama T, Katoh M, Miyahara N, Yamamoto S, Yoshida Y, Nishimura Y, Kusaka M, Fujimoto W, Otsuki T. Soluble interleukin-2 receptor as an indicator of immunological disturbance found in silicosis patients. International Journal of Immunopathology and Pharmacology
[44] Zerler B. The soluble interleukin-2 receptor as a marker for human neoplasia and im‐
[45] Murakami S. Soluble interleukin-2 receptor in cancer. Frontiers in Bioscience 2004;9,
[46] Witkowska, A.M. On the role of sIL-2R measurements in rheumatoid arthritis and cancers. Mediators of Inflammation Mediators of Inflammation 2005;2005(3) 121-30.
[47] Bien E, Balcerska A. Serum soluble interleukin 2 receptor alpha in human cancer of
[48] Bleesing J, Prada A, Siegel DM, Villanueva J, Olson J, Ilowite NT, Brunner HI, Griffin T, Graham TB, Sherry DD, Passo MH, Ramanan AV, Filipovich A, Grom AA. The di‐ agnostic significance of soluble CD163 and soluble interleukin-2 receptor alpha-chain in macrophage activation syndrome and untreated new-onset systemic juvenile idio‐
[49] Coaccioli S, Pinoca F, Giuliani M, Landucci P, Sabatini C, Puxeddu A. Definition of adult-onset rheumatoid arthritis from elderly-onset rheumatoid arthritis by studying T-lymphocyte subpopulations, their soluble receptors and soluble receptor of inter‐
[50] Suh CH, Kim HA. Cytokines and their receptors as biomarkers of systemic lupus er‐
[51] Maier LM, Anderson DE, Severson CA, Baecher-Allan C, Healy B, Liu DV, Wittrup KD, De Jager PL, Hafler DA. Soluble IL-2RA levels in multiple sclerosis subjects and the effect of soluble IL-2RA on immune responses. Journal of Immunology
[52] Lin WW, Hsieh SL. Decoy receptor 3: a pleiotropic immunomodulator and biomark‐ er for inflammatory diseases, autoimmune diseases and cancer. Biochemical Pharma‐
[53] Pitti RM, Marsters SA, Lawrence DA, Roy M, Kischkel FC, Dowd P, Huang A, Dona‐ hue CJ, Sherwood SW, Baldwin DT, Godowski PJ, Wood WI, Gurney AL, Hillan KJ,
ythematosus. Expert Review of Molecular Diagnostics 2008;8(2) 189-98.
http://dx.doi.org/10.1155/MI.2005.121 (accessed 22 November 2013).
adults and children: a review. Biomarkers 2008;13(1) 1-26.
pathic arthritis. Arthritis and Rheumatism 2007;56(3) 965-71.
leukin-2. Clinica Terapeutica 2007;158(4) 303-6.
2009;182(3) 1541-47.
cology 2011;81(7) 838-47.
munology Letters 2005;98(1) 145-52.
mune status. Cancer Cells 1991;3(12) 471-9.
2009;22(1) 53-62.
3085-90
262 Immune Response Activation
[67] Tomokuni A, Aikoh T, Matsuki T, Isozaki Y, Otsuki T, Kita S, Ueki H, Kusaka M, Kishimoto T, Ueki A. Elevated soluble Fas/APO-1 (CD95) levels in silicosis patients without clinical symptoms of autoimmune diseases or malignant tumours. Clinical and Experimental Immunology 1997;110(2) 303-9.
relationship with chronic inflammation and carcinogenesis. Journal of Biomedicine and Biotechnology. 2012;2012:492608. doi: 10.1155/2012/492608. Epub 2012 Feb 6. http://www.hindawi.com/journals/bmri/2012/492608/ (accessed 22November 2013)
Immunostimulation by Silica Particles and the Development of Autoimmune Dysregulation
http://dx.doi.org/10.5772/57544
265
[80] Miura Y, Nishimura Y, Katsuyama H, Maeda M, Hayashi H, Dong M, Hyodoh F, To‐ mita M, Matsuo Y, Uesaka A, Kuribayashi K, Nakano T, Kishimoto T, Otsuki T. In‐ volvement of IL-10 and Bcl-2 in resistance against an asbestos-induced apoptosis of T
[81] Maeda M, Nishimura Y, Hayashi H, Kumagai N, Chen Y, Murakami S, Miura Y, Hir‐ atsuka J, Kishimoto T, Otsuki, T. Reduction of CXC chemokine receptor 3 in an in vi‐ tro model of continuous exposure to asbestos in a human T-cell line, MT-2. American
[82] Maeda M, Nishimura Y, Hayashi H, Kumagai N, Chen Y, Murakami S, Miura Y, Hir‐ atsuka J, Kishimoto T, Otsuki T. Decreased CXCR3 expression in CD4+T cells ex‐ posed to asbestos or derived from asbestos-exposed patients. American Journal of
[83] Kumagai-Takei N, Nishimura Y, Maeda M, Hayashi H, Matsuzaki H, Lee S, Hiratsu‐ ka J, Otsuki T. Effect of asbestos exposure on differentiation of cytotoxic T lympho‐ cytes in mixed lymphocyte reaction of human peripheral blood mononuclear cells.
American Journal of Respiratory Cell and Molecular Biology 2013;49(1) 28-36.
Journal of Immunopathology and Pharmacology 2009;22(3) 579-90.
thology and Pharmacology 2009;22(4) 879-88.
[84] Nishimura Y, Miura Y, Maeda M, Kumagai N, Murakami S, Hayashi H, Fukuoka K, Nakano T, Otsuki T. Impairment in cytotoxicity and expression of NK cell-activating receptors on human NK cells following exposure to asbestos fibers. International
[85] Nishimura Y, Maeda M, Kumagai N, Hayashi H, Miura Y, Otsuki, T. Decrease in phosphorylation of ERK following decreased expression of NK cell-activating recep‐ tors in human NK cell line exposed to asbestos. International Journal of Immunopa‐
Journal of Respiratory Cell and Molecular Biology 2011;45(3) 470-9.
Respiratory Cell and Molecular Biology 2011;45(4) 795-803.
cells. Apoptosis 2006;11(10) 1825-35
relationship with chronic inflammation and carcinogenesis. Journal of Biomedicine and Biotechnology. 2012;2012:492608. doi: 10.1155/2012/492608. Epub 2012 Feb 6. http://www.hindawi.com/journals/bmri/2012/492608/ (accessed 22November 2013)
[67] Tomokuni A, Aikoh T, Matsuki T, Isozaki Y, Otsuki T, Kita S, Ueki H, Kusaka M, Kishimoto T, Ueki A. Elevated soluble Fas/APO-1 (CD95) levels in silicosis patients without clinical symptoms of autoimmune diseases or malignant tumours. Clinical
[68] Otsuki T, Sakaguchi H, Tomokuni A, Aikoh T, Matsuki T, Kawakami Y, Kusaka M, Ueki H, Kita S, Ueki A. Soluble Fas mRNA is dominantly expressed in cases with sili‐
[69] Fontenot JD, Rudensky AY. Molecular aspects of regulatory T cell development.
[70] Hori S, Sakaguchi S. Foxp3: a critical regulator of the development and function of
[71] Weaver CT, Harrington LE, Mangan PR, Gavrieli M, Murphy KM. Th17: an effector CD4 T cell lineage with regulatory T cell ties. Immunity 2006;24(6) 677-88.
[72] Korn T, Bettelli E, Oukka M, Kuchroo VK. IL-17 and Th17 Cells. Annual Review of
[73] Craighead JE, Gibbs AR, Pooley F. Mineralogy of asbestos. In: Craighead JE, Gibbs AR. (ed.) Asbestos and its diseases. New York, Oxford University Press. New York,
[74] Pfau JC, Sentissi JJ, Weller G, Putnam EA. Assessment of autoimmune responses as‐ sociated with asbestos exposure in Libby, Montana, USA. Environmental Health Per‐
[75] Noonan CW, Pfau JC, Larson TC, Spence MR. Nested case-control study of autoim‐ mune disease in an asbestos-exposed population. Environmental Health Perspectives
[76] Lemen RA. Epidemiology of asbestos-related diseases and the knowledge that led to what is known today. (2011) In Asbestos. Risk assessment, epidemiology, and health effects. 2nd ed. (eds,: Dodson, R.F. and Hammar, S.P.) CRC press. New York,
[77] Committee on Asbestos: selected health effects, board on population health and pub‐ lic health practices. Asbestos, selected cancers. Washington, D.C., The National Aca‐
[78] Kumagai-Takei N, Maeda M, Chen Y, Matsuzaki H, Lee S, Nishimura Y, Hiratsuka J, Otsuki T. Asbestos induces reduction of tumor immunity. Clinical and Developmen‐ tal Immunology 2011;2011:481439. doi: 10.1155/2011/481439. Epub 2011 Oct 4. http://
[79] Matsuzaki H, Maeda M, Lee S, Nishimura Y, Kumagai-Takei N, Hayashi H, Yama‐ moto S, Hatayama T, Kojima Y, Tabata R, Kishimoto T, Hiratsuka J, Otsuki T. Asbes‐ tos-induced cellular and molecular alteration of immunocompetent cells and their
www.hindawi.com/journals/cdi/2011/481439/ (accessed 22November 2013)
and Experimental Immunology 1997;110(2) 303-9.
cosis. Immunology 1998;94(2) 258-62.
immunology 2009;27 485-517.
spectives 2005;113(1) 25-30.
131-268, ISBN:978-1-4398-0968-6
demic Press. 2006. ISBN:0-309-10169-7.
2006;114(8) 1243-7.
2008. p23-38.
264 Immune Response Activation
Seminars in Immunology 2004;16(2) 73-80.
regulatory T cells. Microbes and Infection 2004;6(8) 745-51.
### *Edited by Guy Huynh Thien Duc*
The book Immune Response Activation is aiming to analyse the multifaceted aspects of the immune response, treating a number of representative cases in which the immune response is, on one hand, activated against pathogens, and, on the other hand, involved in pathologic settings, leading to allograft rejection, allergy and autoimmunity. The regulatory mechanisms in which the immune response can be modulated for rendering its effector components more efficient and/or not harmful to the organism is also dissected in translational purposes in cancer immunotherapy, local immunity against bacteria and viruses, as well as in allergy and autoimmunity.
Immune Response Activation
Immune Response Activation
*Edited by Guy Huynh Thien Duc*
Photo by selvanegra / iStock | doab | 2025-04-07T04:13:03.369732 | 20-4-2021 17:49 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffa90cf5-6ee0-48d5-8212-f65bf49c005f",
"url": "https://mts.intechopen.com/storage/books/3811/authors_book/authors_book.pdf",
"author": "",
"title": "Immune Response Activation",
"publisher": "IntechOpen",
"isbn": "9789535113744",
"section_idx": 6
} |
ffa9f88a-5af3-498d-8440-85875bd29e9d.0 | *Edited by Heidi Fuller and Monte Gates*
Restoration of motor function following spinal cord injury is a complex and challenging task. By reviewing emerging cellular, pharmacological, rehabilitative, as well as surgical approaches, this book seeks to highlight promising therapeutic strategies for the repair and regeneration of motor circuitry. The multidisciplinary nature of these approaches illustrates various routes to bridging the gap between the bench and the bedside and to identify the challenges that must be overcome in order to bring about a viable therapeutic strategy for spinal cord injury patients.
ISBN 978-953-51-2497-9
Recovery of Motor Function Following Spinal Cord Injury
Photo by toeytoey2530 / iStock
| doab | 2025-04-07T04:13:03.442437 | 20-4-2021 17:56 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffa9f88a-5af3-498d-8440-85875bd29e9d",
"url": "https://mts.intechopen.com/storage/books/5203/authors_book/authors_book.pdf",
"author": "",
"title": "Recovery of Motor Function Following Spinal Cord Injury",
"publisher": "IntechOpen",
"isbn": "9789535124986",
"section_idx": 0
} |
ffa9f88a-5af3-498d-8440-85875bd29e9d.1 | Recovery of Motor Function Following Spinal Cord Injury
*Edited by Heidi Fuller and Monte Gates*
| doab | 2025-04-07T04:13:03.442604 | 20-4-2021 17:56 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffa9f88a-5af3-498d-8440-85875bd29e9d",
"url": "https://mts.intechopen.com/storage/books/5203/authors_book/authors_book.pdf",
"author": "",
"title": "Recovery of Motor Function Following Spinal Cord Injury",
"publisher": "IntechOpen",
"isbn": "9789535124986",
"section_idx": 1
} |
ffa9f88a-5af3-498d-8440-85875bd29e9d.2 | **RECOVERY OF MOTOR FUNCTION FOLLOWING SPINAL CORD INJURY**
Edited by **Heidi Fuller** and **Monte Gates**
### **Recovery of Motor Function Following Spinal Cord Injury**
http://dx.doi.org/10.5772/61539 Edited by Heidi Fuller and Monte Gates
### **Contributors**
Mokhtar Arazpour, Sherif Amr, Juan C. Moreno, Diego Torricelli, Filipe Oliveira Barroso, Christina Francisca Vogelaar, Veronica Estrada, Lucia Slovinska, Juraj Blasko, Dasa Cizkova, Eva Szekiova, Miriam Nagyova, Antonio Ibarra, Salvatore Cuzzocrea, Emanuela Esposito, Irene Paterniti, Mengliang Zhang, Aimee Nelson, Aaron Bailey, Hunter Fassett, Tea Lulic, Jenin El-Sayes, John V. Priestley, Adina Michael-Titus, Juan Larrain, Victor S Tapia, Giles Plant, Laura Marquardt, Bhavaani Jayaram, Vanessa Doulames, Christine Plant
### **© The Editor(s) and the Author(s) 2016**
The moral rights of the and the author(s) have been asserted.
All rights to the book as a whole are reserved by INTECH. The book as a whole (compilation) cannot be reproduced, distributed or used for commercial or non-commercial purposes without INTECH's written permission. Enquiries concerning the use of the book should be directed to INTECH rights and permissions department (permissions@intechopen.com).
Violations are liable to prosecution under the governing Copyright Law.
Individual chapters of this publication are distributed under the terms of the Creative Commons Attribution 3.0 Unported License which permits commercial use, distribution and reproduction of the individual chapters, provided the original author(s) and source publication are appropriately acknowledged. If so indicated, certain images may not be included under the Creative Commons license. In such cases users will need to obtain permission from the license holder to reproduce the material. More details and guidelines concerning content reuse and adaptation can be foundat http://www.intechopen.com/copyright-policy.html.
### **Notice**
Statements and opinions expressed in the chapters are these of the individual contributors and not necessarily those of the editors or publisher. No responsibility is accepted for the accuracy of information contained in the published chapters. The publisher assumes no responsibility for any damage or injury to persons or property arising out of the use of any materials, instructions, methods or ideas contained in the book.
First published in Croatia, 2016 by INTECH d.o.o. eBook (PDF) Published by IN TECH d.o.o. Place and year of publication of eBook (PDF): Rijeka, 2019. IntechOpen is the global imprint of IN TECH d.o.o. Printed in Croatia
Legal deposit, Croatia: National and University Library in Zagreb
Additional hard and PDF copies can be obtained from orders@intechopen.com
Recovery of Motor Function Following Spinal Cord Injury Edited by Heidi Fuller and Monte Gates p. cm. Print ISBN 978-953-51-2497-9 Online ISBN 978-953-51-2498-6 eBook (PDF) ISBN 978-953-51-7290-1
| doab | 2025-04-07T04:13:03.442636 | 20-4-2021 17:56 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffa9f88a-5af3-498d-8440-85875bd29e9d",
"url": "https://mts.intechopen.com/storage/books/5203/authors_book/authors_book.pdf",
"author": "",
"title": "Recovery of Motor Function Following Spinal Cord Injury",
"publisher": "IntechOpen",
"isbn": "9789535124986",
"section_idx": 2
} |
ffa9f88a-5af3-498d-8440-85875bd29e9d.3 | We are IntechOpen, the world's leading publisher of Open Access books Built by scientists, for scientists
3,800+
Open access books available
116,000+
International authors and editors
120M+
Downloads
Our authors are among the
Top 1%
most cited scientists
12.2%
Contributors from top 500 universities
Selection of our books indexed in the Book Citation Index in Web of Science™ Core Collection (BKCI)
## Interested in publishing with us? Contact book.department@intechopen.com
Numbers displayed above are based on latest data collected. For more information visit www.intechopen.com
| doab | 2025-04-07T04:13:03.442853 | 20-4-2021 17:56 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffa9f88a-5af3-498d-8440-85875bd29e9d",
"url": "https://mts.intechopen.com/storage/books/5203/authors_book/authors_book.pdf",
"author": "",
"title": "Recovery of Motor Function Following Spinal Cord Injury",
"publisher": "IntechOpen",
"isbn": "9789535124986",
"section_idx": 3
} |
ffa9f88a-5af3-498d-8440-85875bd29e9d.4 | **Meet the editors**
Dr. Fuller received a BSc honours degree in Biomolecular Science at Bangor University, UK, and went on to undertake a PhD studentship, based at the Robert Jones and Agnes Hunt Orthopaedic Hospital in Oswestry, UK. Following several years of postdoctoral studies in which she developed an interest in applying proteomics technologies to study neurological injury and disease, Dr. Fuller
was recently appointed as a Lecturer in Medical Education at Keele University and Principle Investigator at the Keele Institute for Science and Technology in Medicine. Together with her research team, she aims to identify molecular and cellular targets for neurological regeneration and repair.
Dr. Gates graduated with a BS in Experimental Psychology and a BA in Philosophy from Austin Peay State University in the United States. From there, he went on to complete a PhD in Anatomy and Neurobiology at the University of Tennessee (UT), where he researched how glial cells affect the development and repair of neural circuits in the central nervous system. After postdoctor-
al studies in the laboratory of Dr. Anders Bjorklund at Lund University, Dr. Jeffrey Macklis at Harvard University, and Professor Stephen Dunnett at Cardiff University, he took up a lectureship position at Keele University. Here, his research focuses on the use of in vivo and in vitro models to understand how neuronal cells develop and find appropriate targets in the central nervous system, and how these processes may be manipulated as a strategy to repair motor circuitry defects.
## Contents
## Preface
Chapter 7 **Role of JAK-STAT Signalling on Motor Function Recovery after**
**Section 3 Surgical Approaches for Repair and Regeneration Following**
Chapter 8 **Bridging Defects in Chronic Spinal Cord Injury Using Peripheral**
**Section 4 Emerging Techniques for the Assessment and Rehabilitation of**
Chapter 11 **Non-invasive Brain Stimulation to Characterize and Alter Motor**
Chapter 9 **Stem Cell Therapies for Cervical Spinal Cord Injury 221**
Christine D. Plant and Giles W. Plant
Chapter 10 **Orthoses for Spinal Cord Injury Patients 259**
**Function after Spinal Cord Injury 277**
Chapter 12 **Emerging Techniques for Assessment of Sensorimotor Impairments after Spinal Cord Injury 305** Filipe Barroso, Diego Torricelli and Juan C. Moreno
**Nerve Grafts: From Basic Science to Clinical Experience 181**
Vanessa M. Doulames, Laura M. Marquardt, Bhavaani Jayaram,
Mokhtar Arazpour, Monireh Ahmadi Bani, Mohammad Ebrahim Mousavi, Mahmood Bahramizadeh and Mohammad Ali Mardani
Aaron Z. Bailey, Hunter J. Fassett, Tea Lulic, Jenin El Sayes and
**Spinal Cord Injury 153** Victor S. Tapia and Juan Larrain
**Spinal Cord Injury 179**
**Motor Recovery 257**
Aimee J. Nelson
Sherif M. Amr
**VI** Contents
Restoring motor function after a spinal cord injury (SCI) is one of the most challenging tasks medicine is facing today. To address this problem, and to translate research from the bench to the bedside, experimental models are required, which faithfully recapitulate the molecu‐ lar, cellular and physiological responses to the injury. From there, researchers can begin to characterize and manipulate the complex mechanisms to facilitate repair and regeneration of motor circuitry. In addition to modulating the body's own cellular responses to SCI, sur‐ gical approaches continue to explore how grafting and cell replacement strategies may en‐ hance endogenous repair mechanisms. Beyond these biological interventions, precision in assessing the degree of spinal cord injury and prognostic outlook, as well as tailoring reha‐ bilitative-based interventions, are evolving to enhance mobility following spinal cord injury.
This book seeks to give the reader insight into the current understanding of each of these approaches to restoring motor recovery following SCI. What should become obvious to the reader is the large diversity, in terms of the number of specialized areas of the science in‐ volved in attempting to address this issue, as well as the complexity of the biological factors that play a role in regeneration and repair of SCI.
We are grateful to the authors who have contributed a rich diversity of articles to this book and to our publishing process manager, Ms. Romina Rovan, for her indefatigable organiza‐ tion and support.
> **Dr Heidi Fuller** Postgraduate Medicine and Institute for Science and Technology in Medicine, Keele University, UK
**Dr Monte Gates** School of Medicine and Institute for Science and Technology in Medicine, Keele University, UK
**Models of Spinal Cord Injury**
## **Experimental Spinal Cord Injury Models in Rodents: Anatomical Correlations and Assessment of Motor Recovery**
Christina F. Vogelaar and Veronica Estrada
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/62947
### **Abstract**
Human traumatic spinal cord injury (SCI) causes disruption of descending motor and ascending sensory tracts, which leads to severe disturbances in motor functions. To date, no standard therapy for the regeneration of severed spinal cord axons in humans exists. Experimental SCI in rodents is essential for the development of new treatment strat‐ egiesandforunderstandingtheunderlyingmechanisms leadingtomotorrecovery.Here, we provide an overview of the main rodent models and techniques available for the investigation of neuronal regeneration and motor recovery after experimental SCI.
**Keywords:** spinal cord injury, regeneration, plasticity, rodent, motor recovery
### **1. Introduction**
The challenge of spinal cord injury (SCI) research is to find the right model for testing new treatment strategies. Although rodents differ from humans in many aspects, the research on primates is prohibited in many countries, and there are very strict regulations on experiment‐ ing with nonhuman primates [1]. Therefore, rodent models are the first choice for testing the effectiveness and mechanisms of new potential treatments for SCI. Rodents, especially mice, provide the additional advantage of transgenic technologies (knock out and knock in) that can be helpful in SCI research. In this chapter, an extensive description is provided on the current‐ ly available rodent SCI models, methods of treatment application, histological analysis of regenerating axons, and functional analysis of motor recovery.
© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
### **2. Anatomy of the longitudinal axon tracts in rodents and humans**
In order to understand the impact of SCI, it is important to have some basic knowledge about the long axon tracts that are interrupted by the lesion. Descending tracts control various motor functions. Sensory information from ascending tracts is also essential for posture, balance, and coordination of movements. Here, the main projections from the brain to the spinal cord and vice versa are summarized.
### **2.1. Descending motor tracts**
The descending tracts in the spinal cord (**Figure 1**, left-hand side) run from the brain and brainstem to the spinal cord and are all involved in motor control [2].
### *2.1.1. Corticospinal tract (CST)*
The corticospinal tract (CST) is variable between species. The motor cortex in rodents, generally referred to as the sensorimotor cortex (a rostrocaudal gradient of motor and sensory areas), is not as well defined as it is in humans, who have separate areas for sensory and motor cortex. The CST is responsible for the control of fine movements of distal musculature (e.g., fingers). Pyramidal neurons in layer V of the motor area give rise to the corticospinal axons that run via the internal capsule to the brainstem pyramids where they cross. It then depends on the species which path the majority of CST axons follow. In primates, almost all crossed fibers run in the lateral CST, located in the dorsolateral part of the lateral column. In rodents, most fibers are located in the dorsal CST (dCST), running in the ventral part of the dorsal columns. In some species, a ventral CST (vCST) is also observed. The CST axons terminate mainly in lamina 3– 6 of the grey matter. In humans, up to 20% of CST axons terminate directly on motoneurons in lamina 9. CST terminals are glutamatergic.
### *2.1.2. Rubrospinal tract (RST)*
The rubrospinal tract (RST) plays a role in general locomotion and in some species controls more skilled motor tasks together with the CST. It arises from the caudal magnocellular part of the red nucleus and crosses in the ventral tegmental decussation. The RST descends in the dorsal part of the spinal cord lateral column. The axons terminate in laminae 5 and 6 (some‐ times 7) in the cervical and lumbosacral enlargements corresponding to the limbs. In rats, direct termination on lamina 9 motoneurons has been reported. The RST is prominent in rodents, whereas in animals with a large lateral CST (e.g., primates and humans), the RST is smaller. RST axons use glutamate as neurotransmitter.
### *2.1.3. Reticulospinal tracts (ReST)*
The reticular formation in the brainstem plays a role in the preparation of movements and postural control. Reticulospinal tracts run medially and laterally in the ventral part of the spinal cord white matter. Whereas the medial reticulospinal tract (ReST) remains uncrossed, part of the lateral ReST fibers cross to the contralateral side. The ReST does not form a clear bundle but intermingles with fibers from other tracts, for example, the vestibulospinal and spinotha‐ lamic tracts. The axons terminate in laminae 5–9 and can be either glutamatergic or GABAergic [3].
### *2.1.4. Vestibulospinal tracts (VeST)*
**2. Anatomy of the longitudinal axon tracts in rodents and humans**
vice versa are summarized.
4 Recovery of Motor Function Following Spinal Cord Injury
**2.1. Descending motor tracts**
*2.1.1. Corticospinal tract (CST)*
*2.1.2. Rubrospinal tract (RST)*
*2.1.3. Reticulospinal tracts (ReST)*
in lamina 9. CST terminals are glutamatergic.
RST axons use glutamate as neurotransmitter.
In order to understand the impact of SCI, it is important to have some basic knowledge about the long axon tracts that are interrupted by the lesion. Descending tracts control various motor functions. Sensory information from ascending tracts is also essential for posture, balance, and coordination of movements. Here, the main projections from the brain to the spinal cord and
The descending tracts in the spinal cord (**Figure 1**, left-hand side) run from the brain and
The corticospinal tract (CST) is variable between species. The motor cortex in rodents, generally referred to as the sensorimotor cortex (a rostrocaudal gradient of motor and sensory areas), is not as well defined as it is in humans, who have separate areas for sensory and motor cortex. The CST is responsible for the control of fine movements of distal musculature (e.g., fingers). Pyramidal neurons in layer V of the motor area give rise to the corticospinal axons that run via the internal capsule to the brainstem pyramids where they cross. It then depends on the species which path the majority of CST axons follow. In primates, almost all crossed fibers run in the lateral CST, located in the dorsolateral part of the lateral column. In rodents, most fibers are located in the dorsal CST (dCST), running in the ventral part of the dorsal columns. In some species, a ventral CST (vCST) is also observed. The CST axons terminate mainly in lamina 3– 6 of the grey matter. In humans, up to 20% of CST axons terminate directly on motoneurons
The rubrospinal tract (RST) plays a role in general locomotion and in some species controls more skilled motor tasks together with the CST. It arises from the caudal magnocellular part of the red nucleus and crosses in the ventral tegmental decussation. The RST descends in the dorsal part of the spinal cord lateral column. The axons terminate in laminae 5 and 6 (some‐ times 7) in the cervical and lumbosacral enlargements corresponding to the limbs. In rats, direct termination on lamina 9 motoneurons has been reported. The RST is prominent in rodents, whereas in animals with a large lateral CST (e.g., primates and humans), the RST is smaller.
The reticular formation in the brainstem plays a role in the preparation of movements and postural control. Reticulospinal tracts run medially and laterally in the ventral part of the spinal cord white matter. Whereas the medial reticulospinal tract (ReST) remains uncrossed, part of the lateral ReST fibers cross to the contralateral side. The ReST does not form a clear bundle
brainstem to the spinal cord and are all involved in motor control [2].
The medial and lateral vestibulospinal tracts (VeSTs) are responsible for the initiation of limb and trunk extensor activity, which is important for posture. The lateral VeST arises from the lateral vestibular nucleus and does not cross, whereas the medial VeST originates from both the medial and the spinal vestibular nuclei and partially crosses to the contralateral side. Both run in the ventral white matter and terminate in laminae 7–8, providing glutamatergic input [3].
### *2.1.5. Raphespinal and coeruleospinal tracts*
The Raphe nuclei give rise to the raphespinal projections, which together with the coeruleo‐ spinal projections (from the locus coeruleus) modulate (among others) motor functions. The raphespinal projections include a non-serotonergic component that runs in the dorsolateral funiculus and is involved in gating pain, as well as a serotonergic component that runs in the ventrolateral white matter, terminating in the intermediate grey and on motoneurons in the ventral horn. The noradrenergic coeruleospinal fibers run without crossing in the ventral funiculus and project throughout the grey matter.
**Figure 1.** Spinal cord anatomy: schematic representation of the main ascending sensory tracts (right) and descending motor tracts (left) in a transverse section of the rodent spinal cord. Dotted areas represent locations where tracts are intermingled. Dashed lines indicate the location of the corticospinal tract in humans.
### **2.2. Ascending sensory tracts**
The ascending tracts in the spinal cord (**Figure 1**, right-hand side) convey sensory information from the periphery to central nervous system (CNS) areas involved in walking, posture, and information processing about noxious stimuli [4].
### *2.2.1. Gracile and cuneate tracts*
These two large ascending pathways contain axons from the dorsal root ganglia (DRGs) and provide sensory information from the limbs and trunk. In rodents, an additional dorsal column nucleus contains afferent axons from the tail. The tracts synapse in the gracile and cuneate nuclei located in the medulla oblongata. The second-order axons then cross the midline and run through the medial lemniscus to the thalamus. A subpopulation of DRG neurons synapses locally on dorsal horn neurons, whose axons also project to the gracile and cuneate nuclei. This is called the post-synaptic dorsal column pathway, whereas those DRG axons that do not synapse locally constitute the direct dorsal column pathway.
### *2.2.2. Spinothalamic, spinoreticular, and spinovestibular tracts*
Several sensory tracts run in the ventrolateral funiculus of the spinal cord. These include the spinothalamic tract that conveys nociceptive, thermal, crude touch, and pressure information from the DRGs to the thalamus. The spinoreticular tract provides pain information to brain‐ stem nuclei of the reticular formation. The spinovestibular tract is important for bringing proprioceptive signals to the vestibular nuclei. Several other tracts are present in the ventro‐ lateral funiculus, such as the spinomesencephalic, spinoparabrachial, spinohypothalamic, and spinocervical tracts, each providing information to specific brain regions, that is, mesence‐ phalon, parabrachial nuclei, hypothalamus, and lateral cervical nucleus in the upper cervical cord, respectively.
### *2.2.3. Spinocerebellar tracts*
Projection axons from the spinal cord to the cerebellum are located in the dorsolateral and ventrolateral funiculi (**Figure 1** right-hand side). They carry proprioceptive information from the muscles and tendons to the cerebellum, so that adjustments of posture and coordination of movements can take place.
### **2.3. The propriospinal system**
The spinal cord's "own" projection system refers to neurons that are located in the spinal cord, whose axons interconnect various spinal cord levels [5]. This so-called propriospinal system constitutes a large part of the white matter. It comprises interneurons that are connected to either other interneurons or directly to motoneurons. With respect to locomotor control, shortaxon propriospinal neurons are also called premotoneurons, because they modulate cortico‐ spinal and sensory input to motoneuron pools controlling fore- and hindlimb activity. The long-axon propriospinal neurons form connections between the cervical and lumbosacral enlargements and are responsible for coordination of fore- and hindlimbs. These axons run in the fasciculus proprius (**Figure 1**, right-hand side).
The propriospinal neurons also modulate input to the lumbar central pattern generator (CPG), a local system involved in reflexive stepping in total absence of supraspinal input [5]. Serotonin from brainstem neurons has been shown to play a major role in CPG activation [6, 7].
### **3. Rodent spinal cord injury models**
*2.2.1. Gracile and cuneate tracts*
6 Recovery of Motor Function Following Spinal Cord Injury
cord, respectively.
*2.2.3. Spinocerebellar tracts*
of movements can take place.
**2.3. The propriospinal system**
the fasciculus proprius (**Figure 1**, right-hand side).
These two large ascending pathways contain axons from the dorsal root ganglia (DRGs) and provide sensory information from the limbs and trunk. In rodents, an additional dorsal column nucleus contains afferent axons from the tail. The tracts synapse in the gracile and cuneate nuclei located in the medulla oblongata. The second-order axons then cross the midline and run through the medial lemniscus to the thalamus. A subpopulation of DRG neurons synapses locally on dorsal horn neurons, whose axons also project to the gracile and cuneate nuclei. This is called the post-synaptic dorsal column pathway, whereas those DRG axons that do not
Several sensory tracts run in the ventrolateral funiculus of the spinal cord. These include the spinothalamic tract that conveys nociceptive, thermal, crude touch, and pressure information from the DRGs to the thalamus. The spinoreticular tract provides pain information to brain‐ stem nuclei of the reticular formation. The spinovestibular tract is important for bringing proprioceptive signals to the vestibular nuclei. Several other tracts are present in the ventro‐ lateral funiculus, such as the spinomesencephalic, spinoparabrachial, spinohypothalamic, and spinocervical tracts, each providing information to specific brain regions, that is, mesence‐ phalon, parabrachial nuclei, hypothalamus, and lateral cervical nucleus in the upper cervical
Projection axons from the spinal cord to the cerebellum are located in the dorsolateral and ventrolateral funiculi (**Figure 1** right-hand side). They carry proprioceptive information from the muscles and tendons to the cerebellum, so that adjustments of posture and coordination
The spinal cord's "own" projection system refers to neurons that are located in the spinal cord, whose axons interconnect various spinal cord levels [5]. This so-called propriospinal system constitutes a large part of the white matter. It comprises interneurons that are connected to either other interneurons or directly to motoneurons. With respect to locomotor control, shortaxon propriospinal neurons are also called premotoneurons, because they modulate cortico‐ spinal and sensory input to motoneuron pools controlling fore- and hindlimb activity. The long-axon propriospinal neurons form connections between the cervical and lumbosacral enlargements and are responsible for coordination of fore- and hindlimbs. These axons run in
The propriospinal neurons also modulate input to the lumbar central pattern generator (CPG), a local system involved in reflexive stepping in total absence of supraspinal input [5]. Serotonin
from brainstem neurons has been shown to play a major role in CPG activation [6, 7].
synapse locally constitute the direct dorsal column pathway.
*2.2.2. Spinothalamic, spinoreticular, and spinovestibular tracts*
The choice of SCI models is important in view of comparability with human SCI, but practical issues should also be considered. Although human lesions are usually compressions (but some may be sharp wounds as well or a mixture of both), from the experimental point of view it might be important to have a more "clean" and reproducible cut. Treatment strategies that fail to cause regeneration through a spinal cord transection lesion will probably have equally small effects after contusion lesions. On the contrary, treatments that induce regeneration in a transection model should then be tested and optimized in a contusion model. Partial injury models are useful for the investigation of the locomotor recovery over time, since not only regeneration but also sprouting from spared axon tracts can occur (see Section 5). Models of complete transection are used to study regeneration without the possibility of plasticity processes bypassing the lesion. While the complete transection of the spinal cord is a very reproducible injury, disadvantages of this lesion model are the poor degree of regenerative
**Figure 2** Spinal cord lesions in rodent models of SCI: (A) schematic representation of a parasagittal section through the brain and spinal cord (modified from Paxinos & Watson, The Rat Brain in Stereotaxic Coordinates, 6th Edition). Tracts with clear localizations are indicated. It should be noted that these do not run in the same spinal cord section, with the RST (red) running more laterally than the CST (brown) and the cuneate and gracile tracts (blue). The dorsal hemisec‐ tion, complete transection, and the pyramidotomy lesions are represented as black bars. (B) Schematic drawings of transverse sections through the spinal cord with bilateral motor tracts depicted left and bilateral sensory tracts depict‐ ed right (for exact description of the tracts see **Figure 1**). Dashed areas represent the extent of tissue damage produced by the different injury paradigms. Note that motor and sensory tracts run in both spinal cord hemispheres but are de‐ picted separately for better understanding. (C) Histological parasagittal section through the spinal cord of YFP-H mice (The Jackson Laboratory) 7 days after a dorsal hemisection (compilation of 2 sagittal sections). Descending CST axons are intact rostrally and have degenerated caudally from the lesion site (dashed white line). Ventral CST fibers are not lesioned (plane of section causes apparent lack rostrally). Ascending dorsal column axons are intact caudally and have degenerated rostrally.
growth of the severed axons and the general inadequacy for most motor tests. In this section, the technical principles of each model in rats and mice are described.
### **3.1. Dorsal hemisection (Hx)**
Spinal cord transection lesions are generally applied using scissors, scalpel blades, or fine retractable wire knives. The advantage of wire knives (McHugh Millieux) is that a SCI can be performed with high precision, because they can be attached to a stereotactic frame. The dorsal Hx (Figure 2A–C) is the most used SCI paradigm for the investigation of the regeneration of CST and, depending on the extent of lateral lesion, it also includes the RST. It is mostly performed at thoracic level 8 (T8) and involves the laminectomy at T8-9, opening of the dura mater and subsequent lesioning of the spinal cord [8, 9]. For mice, microdissection spring scissors (Fine Science Tools) are used to hemisect the spinal cord. Since this procedure is inherently variable, the experimenter needs to test various depths to determine the desired extension of the lesion. A more controlled technique for dorsal Hx in mice was described by Hill et al. (2009) who used a so-called Vibraknife (LISA-Vibraknife; Louisville, KY) [10, 11]. Dorsal hemisection lesions are usually applied at thoracic spinal cord levels and result in the formation of a dense inhibitory scar [12, 13]. Depending on the severity of the lesion, the animals spontaneously recover a certain degree of walking that can be further ameliorated by regeneration promoting treatments.
### **3.2. Lateral Hx**
For lateral Hx, the lateral half of the spinal cord is transected in mostly the same technical procedure as the dorsal Hx, with the difference that the tracts on one side are left intact (Figure 2B). These lesions provide the advantage of an internal control situation [14], which is also reflected in the behavioral testing, where paw preferences are often scored (see Section 7.5.). Lateral Hx experiments are usually performed at cervical levels, allowing the analysis of both fore- and hindlimb recovery. Mostly, a lesion at cervical level C5 is produced, but some groups have specialized on the analysis of breathing musculature after a lesion at C2 [15].
### **3.3. Complete transection (Tx)**
For a complete transection (Figure 2B), small scissors are generally used to transect the spinal cord after having cut the meninges. Alternatively, the dura mater is opened just enough to allow the insertion of a spinal cord hook (Fine Science Tools) between dura and spinal cord. The hook is then used to lift the cord in order to completely cut the spinal cord. The dura mater can be closed with fine sutures (10-O) after the procedure. The complete Tx model is useful to investigate the effect of treatments on the axonal regeneration, and on (limited) recovery of locomotor function. After a complete SCI in rats, there is usually the development of fluidfilled cavities, whereas in mice this is generally not the case [16].
### **3.4. Contusion and compression injury**
growth of the severed axons and the general inadequacy for most motor tests. In this section,
Spinal cord transection lesions are generally applied using scissors, scalpel blades, or fine retractable wire knives. The advantage of wire knives (McHugh Millieux) is that a SCI can be performed with high precision, because they can be attached to a stereotactic frame. The dorsal Hx (Figure 2A–C) is the most used SCI paradigm for the investigation of the regeneration of CST and, depending on the extent of lateral lesion, it also includes the RST. It is mostly performed at thoracic level 8 (T8) and involves the laminectomy at T8-9, opening of the dura mater and subsequent lesioning of the spinal cord [8, 9]. For mice, microdissection spring scissors (Fine Science Tools) are used to hemisect the spinal cord. Since this procedure is inherently variable, the experimenter needs to test various depths to determine the desired extension of the lesion. A more controlled technique for dorsal Hx in mice was described by Hill et al. (2009) who used a so-called Vibraknife (LISA-Vibraknife; Louisville, KY) [10, 11]. Dorsal hemisection lesions are usually applied at thoracic spinal cord levels and result in the formation of a dense inhibitory scar [12, 13]. Depending on the severity of the lesion, the animals spontaneously recover a certain degree of walking that can be further ameliorated by
For lateral Hx, the lateral half of the spinal cord is transected in mostly the same technical procedure as the dorsal Hx, with the difference that the tracts on one side are left intact (Figure 2B). These lesions provide the advantage of an internal control situation [14], which is also reflected in the behavioral testing, where paw preferences are often scored (see Section 7.5.). Lateral Hx experiments are usually performed at cervical levels, allowing the analysis of both fore- and hindlimb recovery. Mostly, a lesion at cervical level C5 is produced, but some groups
For a complete transection (Figure 2B), small scissors are generally used to transect the spinal cord after having cut the meninges. Alternatively, the dura mater is opened just enough to allow the insertion of a spinal cord hook (Fine Science Tools) between dura and spinal cord. The hook is then used to lift the cord in order to completely cut the spinal cord. The dura mater can be closed with fine sutures (10-O) after the procedure. The complete Tx model is useful to investigate the effect of treatments on the axonal regeneration, and on (limited) recovery of locomotor function. After a complete SCI in rats, there is usually the development of fluid-
have specialized on the analysis of breathing musculature after a lesion at C2 [15].
filled cavities, whereas in mice this is generally not the case [16].
the technical principles of each model in rats and mice are described.
**3.1. Dorsal hemisection (Hx)**
8 Recovery of Motor Function Following Spinal Cord Injury
regeneration promoting treatments.
**3.3. Complete transection (Tx)**
**3.2. Lateral Hx**
Contusion injuries are the most widely used lesion type in SCI research, since the majority of human SCI involves a contusion or compression pathology. Several commercially available systems can be used to inflict standardized graded contusion injuries. These include the NYU MASCIS impactor (New York University Multicenter Animal SCI Study) [17], the OSU impactor (Ohio State University electromagnetic SCI device) [18, 19], the IH impactor (Infinite Horizon) [20], and the spinal cord compression device (Kopf Instruments). In general, a controlled pressure is exerted on the spinal cord after laminectomy by either dropping or placing a weight onto the cord, controlling the force and/or velocity [21]. Depending on the species (in rats more than in mice), contusion injury leads to cyst formation (Figure 2B), a feature also displayed by human SCI patients [19, 21]. Thoracic contusions are usually performed to induce dorsal bilateral lesions, whereas contusions at the cervical level are performed unilaterally [20].
Compression/decompression models are frequently used to investigate the occlusion of the central canal, another common symptom of SCI in human patients. To perform an experi‐ mental compression, injury clips, balloons, or forceps are widely used [21]. Vascular clips and calibrated forceps can be used to create graded and reproducible injuries. The clip compression model and the contusion injury model show some resemblances as they both inflict the injury via pressure to the outer surface of the spinal cord. These models can be fine-tuned so that injuries of varying degrees can be created. They lead to the formation of fluid-filled cysts which are surrounded by spared tissue. The remaining tissue continuity and axon sparing makes them also a suitable model for locomotor functional tests. For the same reason, however, SCI contusion and compression models are not as well suited as transection models to investigate the neuronal and axonal regeneration.
### **3.5. Pyramidotomy**
An exclusive CST-only lesion can be achieved by pyramidotomy, a transection at the height of the pyramids [22] (Figure 2A–B). The injury of the CST by pyramidotomy does not greatly affect locomotion in rodents. Rats and mice use the CST primarily for fine finger movements, which is greatly relevant for human patients. For the study of motor recovery, specific forepaw tests are used (see Section 7.5.). Since this lesion is usually performed unilaterally, the intact side serves as an internal control and is also used for studying plasticity-related regeneration mechanisms [23].
### **3.6. New SCI models**
Scientists are continuously looking for models that resemble the human injuries more closely. For example, two recent studies focused on lumbosacral SCI, a type of injury affecting an estimated one-third of patients [24, 25]. A model combining SCI and traumatic brain injury was recently introduced, because a proportion of SCI patients additionally suffer from head injuries, for example, due to traffic or diving accidents [26]. Finally, a recent publication on a closed-body SCI by applying a high-pressure air blast in mice provides a model resembling human traumatic SCI [27].
### **4. The application of treatments**
After the choice of the appropriate lesion model for therapy development, the next decision in experimental SCI is the technique to use to apply a treatment. The application method determines the timing, frequency, and duration of the treatment. This section provides technical details of current methods for applying treatments in the various SCI models.
**Figure 3** Strategies for the application of treatments and tracers exemplified in a schematic drawing of a dorsal Hx lesion: (A) treatments can be applied by (a) injection into the lesion site and/or adjacent tissue, (b) resection of the chronic lesion scar and subsequent matrix or cell implantation, (c) intrathecal lumbar injection, and (d) infusion over prolonged periods with minipumps and catheters either intraventricularly, intrathecally, or epidurally. The epidural catheter can be guided not only from rostral direction through the cisterna magna but also from the caudal side by performing an additional laminectomy [9, 28]. (B) Anterograde tracers are injected into the motor cortex and nucleus ruber in order to label CST and RST, respectively, and in the peripheral nerve to label the dorsal column axons. Injec‐ tion of a retrograde tracer caudally to the lesion site is applied in order to visualize the cell bodies corresponding to regenerated axons or local interneuron circuits.
### **4.1. Injection into the spinal cord parenchyma**
closed-body SCI by applying a high-pressure air blast in mice provides a model resembling
After the choice of the appropriate lesion model for therapy development, the next decision in experimental SCI is the technique to use to apply a treatment. The application method determines the timing, frequency, and duration of the treatment. This section provides technical details of current methods for applying treatments in the various SCI models.
**Figure 3** Strategies for the application of treatments and tracers exemplified in a schematic drawing of a dorsal Hx lesion: (A) treatments can be applied by (a) injection into the lesion site and/or adjacent tissue, (b) resection of the chronic lesion scar and subsequent matrix or cell implantation, (c) intrathecal lumbar injection, and (d) infusion over prolonged periods with minipumps and catheters either intraventricularly, intrathecally, or epidurally. The epidural catheter can be guided not only from rostral direction through the cisterna magna but also from the caudal side by performing an additional laminectomy [9, 28]. (B) Anterograde tracers are injected into the motor cortex and nucleus ruber in order to label CST and RST, respectively, and in the peripheral nerve to label the dorsal column axons. Injec‐ tion of a retrograde tracer caudally to the lesion site is applied in order to visualize the cell bodies corresponding to
human traumatic SCI [27].
**4. The application of treatments**
10 Recovery of Motor Function Following Spinal Cord Injury
regenerated axons or local interneuron circuits.
The simplest method for acutely applying a therapy is the (single) injection of a substance at the time of surgery. In some models, such as the dorsal Hx, the lesion site is open, so the treatment might potentially diffuse too quickly out of the area. Therefore, treatments are often injected in the intact tissue immediately adjacent to the lesion (Figure 3Aa). The injection volume should not be too high (<1 μl), and the injection should be performed slowly so that additional damage to the tissue is avoided. Controlled injection is achieved either by using a pump (e.g., the Pump 11 Elite Nanomite, Harvard Sachs Elektronik) or by introducing a delay of several minutes between injection and retraction of the needle. The injection method is most suitable for single acute treatments, because any additional doses will require additional surgery.
### **4.2. Lumbar injection into the CSF**
A therapeutic can be applied to the CSF by lumbar intrathecal injection (Figure 3Ac), described in detail for mice by Lu et al, 2013. Shortly, the animal is subjected to a brief inhalation narcosis and kept in half-sleep by keeping its head in a dark environment. The L5 vertebra can be localized between the iliac crests of the hip bones. A 30-gauge needle is used to puncture the skin and enter the spine between the L5 and L6 spinous process. When the dura mater is punctured, a reflective flick of the tail is induced and up to 5 μl liquid can be applied [29]. This method is useful for renewing treatments multiple times after the initial injection.
### **4.3. Intrathecal application via minipumps**
For continuous long-term application of liquid therapeutics, the use of minipumps is a standard delivery method (Figure 3Ad). Pumps can be implanted subcutaneously and attached to a catheter for intrathecal delivery. Minipumps either release the liquid via osmosis (Alzet®) or they use a programmable microprocessor (iPrecio®). They are commercially available in different sizes and with varying pumping rates and time periods. The subcutane‐ ously placed minipumps can be removed after the required delivery period. The minipumps are connected to a catheter which can be inserted in the brain for intraventricular infusion [30], or the catheter can be guided through the epidural space underneath the vertebrae toward the lesion site [19, 28]. It is important to consider that the catheter by itself can produce a com‐ pression of the spinal cord. This is especially problematic in mice, because of their size, although special mouse catheters are available commercially (Alzet®).
### **4.4. Cellular transplantation strategies**
Cell therapy is generally the focus in neurodegenerative diseases such as Alzheimer's or Parkinson's disease where the common goal is to replace degenerated neurons. In contrast, SCI is characterized by damage of the neuronal processes, whereas the corresponding cell bodies are located in various areas of the brain, brainstem, and DRGs, thus complicating cell replacement. Moreover, the projection neurons are thought to undergo atrophy in contrast to dying [31]. Therefore, cellular therapeutic approaches for experimental SCI concentrate on the spinal cord, where local cells are affected by the primary and secondary injury events. The therapies on the one hand aim to replace glial cells or local neurons. Peripheral nerve grafts or Schwann cell cables have been used to bridge the lesion [32, 33]. Transplanted oligodendrocyte precursor cells or Schwann cells have been shown to remyelinate axons, whereas olfactory ensheathing or mucosal cells may provide axon guidance and trophic support [34–36]. Cell therapies using embryonic stem cells, neural stem/progenitor cells, or induced pluripotent stem (iPS) cells mostly aim to provide local pools of neurons that might serve as relay stations, analogue to propriospinal neurons [37–40]. Stem cells might also differentiate into glial cells that can remyelinate axons. On the other hand, stem cells can bridge the lesion gap and promote regeneration by the secretion of trophic factors, the support of angiogenic events, or the inhibition of glutamate toxicity. These effects have been reported for mesenchymal stem cells, bone marrow mesenchymal stromal cells, or unrestricted somatic stem cells from umbilical cord blood [41–43].
The transplantation of (stem) cells is usually performed by injection of cell suspensions into the spinal cord parenchyma (Figure 3Aa). This can be performed acutely by injecting cells into the intact tissue adjacent to the lesion. Alternatively, the lesion is allowed to form over a certain time period (usually 7 days, also called subacute), and a new surgery is performed to inject the cells directly into the lesion site. Factors to consider are cell survival, migration, differen‐ tiation into neural/glial cell types, axon outgrowth, and synaptic contacts in the case of neuronal transplants and secretion of regeneration-supportive factors in the case of nonneural transplants.
### **4.5. Implantation of matrices**
Although many studies have proven the beneficial effects of autologous or heterologous cellular grafts in acute and chronic SCI models in animals [44, 45], the use of cell transplantation in human patients often remains a controversial issue [46, 47]. The search for artificial bioma‐ terials for the implantation into the injured spinal cord has been prompted due to the limited access to autologous donor material and immunological problems associated with allograft rejection.
Cavities or cysts that often form after SCI are a major obstacle impeding axonal regeneration. Therefore, the reconnection across the trauma cavity by means of scaffolds or matrices is a major focus in SCI research. In order to provide a favorable growth substrate for regenerating axons, a bridging material should provide and combine several structural, physicochemical, and molecular properties [48]. Materials should ideally be easily modifiable, serve as a scaffold for matrix molecules and/or cellular transplants, and further be immunologically inert and absorbable [49]. Positive results with acellular matrices have been obtained in numerous studies [45, 49–55]. Important advances have recently been reported in the development of biosynthetic conduits for spinal cord repair. Biosynthetic conduits equipped with ECM molecules and different cell lines, and supplemented with neurotrophic growth factors, have been shown to yield encouraging results in the treatment of experimental SCI [51].
In chronic SCI, cavity formation has occurred and a lesion scar has formed, which presents a stable physical and molecular barrier to axonal regeneration. Cavities and sites of scar resection can be treated with bridging or scaffolding materials. Interesting effects were achieved with a polyethylene glycol (PEG) treatment in a chronic SCI paradigm [56]. PEG was used to fill the cavity that was created by resection of the 5-week-old lesion scar in spinal cord-injured rats (Figure 3Ab). After 8 months, long-distance axonal regeneration through and beyond the graft was observed. The PEG matrix was repopulated by blood vessels, astrocytes, and Schwann cells, the latter remyelinating bundles of regenerating axons. These histological parameters were accompanied by long-lasting functional motor improvement. This study suggests that the chronically lesioned tracts are still able to regenerate when provided with the right extracellular environment [56].
### **4.6. Implantation of a mechanical microconnector system**
spinal cord, where local cells are affected by the primary and secondary injury events. The therapies on the one hand aim to replace glial cells or local neurons. Peripheral nerve grafts or Schwann cell cables have been used to bridge the lesion [32, 33]. Transplanted oligodendrocyte precursor cells or Schwann cells have been shown to remyelinate axons, whereas olfactory ensheathing or mucosal cells may provide axon guidance and trophic support [34–36]. Cell therapies using embryonic stem cells, neural stem/progenitor cells, or induced pluripotent stem (iPS) cells mostly aim to provide local pools of neurons that might serve as relay stations, analogue to propriospinal neurons [37–40]. Stem cells might also differentiate into glial cells that can remyelinate axons. On the other hand, stem cells can bridge the lesion gap and promote regeneration by the secretion of trophic factors, the support of angiogenic events, or the inhibition of glutamate toxicity. These effects have been reported for mesenchymal stem cells, bone marrow mesenchymal stromal cells, or unrestricted somatic stem cells from umbilical
The transplantation of (stem) cells is usually performed by injection of cell suspensions into the spinal cord parenchyma (Figure 3Aa). This can be performed acutely by injecting cells into the intact tissue adjacent to the lesion. Alternatively, the lesion is allowed to form over a certain time period (usually 7 days, also called subacute), and a new surgery is performed to inject the cells directly into the lesion site. Factors to consider are cell survival, migration, differen‐ tiation into neural/glial cell types, axon outgrowth, and synaptic contacts in the case of neuronal transplants and secretion of regeneration-supportive factors in the case of nonneural
Although many studies have proven the beneficial effects of autologous or heterologous cellular grafts in acute and chronic SCI models in animals [44, 45], the use of cell transplantation in human patients often remains a controversial issue [46, 47]. The search for artificial bioma‐ terials for the implantation into the injured spinal cord has been prompted due to the limited access to autologous donor material and immunological problems associated with allograft
Cavities or cysts that often form after SCI are a major obstacle impeding axonal regeneration. Therefore, the reconnection across the trauma cavity by means of scaffolds or matrices is a major focus in SCI research. In order to provide a favorable growth substrate for regenerating axons, a bridging material should provide and combine several structural, physicochemical, and molecular properties [48]. Materials should ideally be easily modifiable, serve as a scaffold for matrix molecules and/or cellular transplants, and further be immunologically inert and absorbable [49]. Positive results with acellular matrices have been obtained in numerous studies [45, 49–55]. Important advances have recently been reported in the development of biosynthetic conduits for spinal cord repair. Biosynthetic conduits equipped with ECM molecules and different cell lines, and supplemented with neurotrophic growth factors, have
been shown to yield encouraging results in the treatment of experimental SCI [51].
In chronic SCI, cavity formation has occurred and a lesion scar has formed, which presents a stable physical and molecular barrier to axonal regeneration. Cavities and sites of scar resection
cord blood [41–43].
12 Recovery of Motor Function Following Spinal Cord Injury
transplants.
rejection.
**4.5. Implantation of matrices**
Complete transections result in a gap between the two spinal cord stumps. Recently, a novel and unique connector device was described [57]. The purpose of this mechanical microcon‐ nector system is to reconnect severed spinal cord tissue stumps in the submillimeter range. The microconnector consists of two elliptical discs lined with numerous honeycombed holes. After implantation into the injured rat spinal cord, the device is connected to a vacuum pump, and the tissue stumps are brought into close apposition via the application of negative pressure. The connector discs have a rough surface, allowing the adherence of the spinal cord tissue. Additional features of the mechanical microconnector system are an internal canal system and an inlet tube, which can be connected either to a syringe or to an osmotic minipump to achieve application of therapeutics into the lesion area. Even the implantation of the device alone was sufficient for axon regeneration and led to a significant improvement of locomotor function following complete transection of the thoracic spinal cord [57].
### **4.7. Electrical stimulation and neuroprothesis**
Electric field stimulation has been shown to promote enhanced and/or oriented neurite outgrowth, thereby offering potential additional treatment strategies after PNS but also CNS injury [58–60]. For SCI treatment, epidural stimulation has been used to create electric fields to restore motor functions [61, 62]. Electrical current is applied at varying frequencies and intensities to the areas of the lumbosacral spinal cord, activating the CPG. The CPG can initiate stepping function even without any input from the brain. The lab of Grégoire Courtine developed a neuroprosthetic that achieves a high-fidelity control of leg kinematics. A closedloop system, using muscle activity and other kinematic parameters in real-time to feed back into the system, allowed neuromodulation during walking [63]. Another study used neuro‐ prosthetic intervention in the form of a Neurochip 2 recurrent brain-computer interface in a cervical hemisection model. The neurochip delivered electrical stimulation and measured in parallel the electromyographic (EMG) activity of the muscles, thus adjusting the stimulation according to the muscle activity. Animals that received this so-called targeted, activitydependent stimulation displayed increased skilled forepaw reaching as compared to animals receiving non-targeted stimulations or physical training [64].
Although it has no direct effect on the regeneration of axons after SCI, epidural stimulation is a very promising approach already used for the rehabilitation of SCI patients with promising results [65].
### **4.8. Exercise and training**
The first studies suggesting that exercise might stimulate motor recovery were performed using environment enrichment [66, 67]. During the last decades, several investigators devel‐ oped new experimental settings to perform motor training of animals. For example, forced walking on treadmills, training either bipedal or quadrupedal stepping, has been shown to improve locomotor recovery after SCI [68–70]. The combination of treadmill training with epidural stimulation and the administration of serotonergic and dopaminergic agonists seemed to be especially effective in restoring motor activity. Extensive plasticity of corticospi‐ nal, brainstem, and intraspinal connectivity was shown to underlie the observed functional recovery [61].
Recently, Starkey et al. (2015) developed a new type of cage with enriched environment over three floors with different types of training possibilities (e.g., grasping tasks, ladder walking, and climbing). This so-called "natural habitat cage" was combined with a new three-dimen‐ sional animal tracking system to allow high-impact, self-motivated training. Interestingly, differences were observed between the animals' overall activity and preference for certain tasks. Healthy as well as SCI animals trained in these cages performed better in experimental tests for fine motor control of fore- and hindlimb [71]. For forelimb training, a robotic rehabil‐ itation system was recently developed, in which the animal has to pull a bar to receive food. This setup could also be used to measure forelimb strength [72].
### **4.9. Other types of treatments**
Systemic treatments (intravenous, intraperitoneal, subcutaneous) are not discussed above, although they are clinically relevant. For treatments outside the spinal column, it should in general be known whether the applied therapeutic can cross the blood-brain barrier. A much higher concentration must be applied peripherally to achieve an effective concentration centrally. Since human SCI almost invariably involves surgery, the possibility of local treat‐ ment is given.
### **4.10. Combination treatments**
It has become more and more apparent that combination therapies will be necessary to successfully treat SCI. The above described matrices, cell transplantations, electrical stimula‐ tion, and training paradigms all offer possibilities of combination with trophic factors, pharmacological treatments, agonists or antagonists of neurotransmitters, anti-inhibitory treatments, and so forth. It seems likely that holistic treatments combining several regeneration mechanisms will be clinically more successful to target the multitude of SCI systems.
### **5. Possible treatment effects on neurons**
### **5.1. Regeneration versus sprouting**
Although it has no direct effect on the regeneration of axons after SCI, epidural stimulation is a very promising approach already used for the rehabilitation of SCI patients with promising
The first studies suggesting that exercise might stimulate motor recovery were performed using environment enrichment [66, 67]. During the last decades, several investigators devel‐ oped new experimental settings to perform motor training of animals. For example, forced walking on treadmills, training either bipedal or quadrupedal stepping, has been shown to improve locomotor recovery after SCI [68–70]. The combination of treadmill training with epidural stimulation and the administration of serotonergic and dopaminergic agonists seemed to be especially effective in restoring motor activity. Extensive plasticity of corticospi‐ nal, brainstem, and intraspinal connectivity was shown to underlie the observed functional
Recently, Starkey et al. (2015) developed a new type of cage with enriched environment over three floors with different types of training possibilities (e.g., grasping tasks, ladder walking, and climbing). This so-called "natural habitat cage" was combined with a new three-dimen‐ sional animal tracking system to allow high-impact, self-motivated training. Interestingly, differences were observed between the animals' overall activity and preference for certain tasks. Healthy as well as SCI animals trained in these cages performed better in experimental tests for fine motor control of fore- and hindlimb [71]. For forelimb training, a robotic rehabil‐ itation system was recently developed, in which the animal has to pull a bar to receive food.
Systemic treatments (intravenous, intraperitoneal, subcutaneous) are not discussed above, although they are clinically relevant. For treatments outside the spinal column, it should in general be known whether the applied therapeutic can cross the blood-brain barrier. A much higher concentration must be applied peripherally to achieve an effective concentration centrally. Since human SCI almost invariably involves surgery, the possibility of local treat‐
It has become more and more apparent that combination therapies will be necessary to successfully treat SCI. The above described matrices, cell transplantations, electrical stimula‐ tion, and training paradigms all offer possibilities of combination with trophic factors, pharmacological treatments, agonists or antagonists of neurotransmitters, anti-inhibitory treatments, and so forth. It seems likely that holistic treatments combining several regeneration
mechanisms will be clinically more successful to target the multitude of SCI systems.
This setup could also be used to measure forelimb strength [72].
results [65].
recovery [61].
ment is given.
**4.9. Other types of treatments**
**4.10. Combination treatments**
**4.8. Exercise and training**
14 Recovery of Motor Function Following Spinal Cord Injury
For researchers, the "holy grail" is the regeneration of the injured fibers through the lesion scar and the subsequent reinnervation of their targets. After an initial retraction phase, the axons of the above-described tracts usually start regrowing toward the lesion site. Treatments can increase the regenerative growth of various tracts through and beyond the lesion site [73]. Although this process could be called "sprouting," it is important to distinguish between regenerative sprouting of the severed tract with the goal of regrowth toward the original connections and plastic sprouting, with the goal to find alternatives routes (**Figure 4**). Func‐ tional recovery of locomotion can also be achieved through plasticity of intact fibers that may form contralateral sprouts or make new synapses with local propriospinal neurons (**Figure 4**). In the last decade, the propriospinal system became a major focus for SCI. It can serve as a detour for bypassing the scar. Injured descending axons have been shown to sprout and rewire to propriospinal neurons, whose axons are located in the spared tissue and project into the lower denervated spinal cord [69, 74]. Also, the propriospinal interneurons can sprout to innervate new targets below the lesion. In addition, these neurons might regenerate better than the projection neurons, because of the shorter distance of the axon stump to the cell body. They have been shown to upregulate growth-associated proteins and have a high intrinsic capacity for plasticity [75]. In partial injury models, such as dorsal or lateral transection or contusion,
**Figure 4** Axonal mechanisms leading to motor recovery exemplified for the dCST in two lesion models: (A) in the case of lateral Hx lesions, the contralateral intact dCST axons caudal to the lesion can sprout (s) and synapse with proprio‐ spinal interneurons (IN) connected to motoneurons (not necessarily at the same rostrocaudal level). The vCST on the intact side can also sprout to the ipsilateral side. Regeneration of the CST through the lesion is depicted by irregular lines since regenerating axons generally display a meandering and less straight course than the original tracts. (B) In the dorsal Hx model axons of the dCST can regenerate either through the lesion (irregular lines) or to form sprouts to make local connections with propriospinal interneurons whose axons run ventrally below the lesion and are connected to caudal motoneurons. The intact vCST can sprout and extend to the degenerated dCST tract or form new connections with interneurons that contact local motoneurons. Abbreviations: dCST: dorsal CST, IN: interneuron, MN: motoneur‐ on, reg: regenerating fiber, s: sprout, vCST: ventral CST.
both regeneration from injured tracts and sprouting from spared tracts can be studied. It may be of importance to note that sprouting can be undirected, so that aberrant neuronal circuits may be formed [76]. Treatments can enhance sprouting and direct the sprouts to establish functional circuits.
### **5.2. Neuroprotection**
After the primary insult, secondary damage due to, among others, inflammation, oxidative stress, and blood-brain barrier dysfunction causes the death of neurons (and glia) in the tissue surrounding the lesion [77]. The loss of local motor neurons leads to more extensive motor deficits in addition to the impairments caused by the injury to the descending motor tracts. The loss of spinal interneurons may disrupt intraspinal connections between motor centers. Therefore, a neuroprotective action of a treatment might, first, reduce functional impairments and, second, increase the possibility of local plasticity via interneurons (see Section 5.1.). For the analysis of neuroprotection, quantification of (moto-) neurons is performed at various distances rostrally and caudally from the lesion center [78]. A treatment could also lead to the protection of the brain and brainstem projection neurons from death or atrophy [8, 31, 79]. Quantification of the lesion size and spared white matter in standardized lesion models might also provide information about the protective effects of a treatment strategy.
### **6. Tracing and/or immunohistochemical (IHC) staining of motor and sensory tracts**
The next step in SCI research is the histological analysis of regenerated axons. Short-term studies (up to several weeks after the injury) give information about early injury events, whereas long-term studies (up to several months or even years after the injury) are useful to investigate long-term effects and behavioral outcomes with a treatment compared to a control. In order to visualize regenerating axons from the specific spinal cord tracts, these can be marked via axonal tracing (Figure 3B) or detected by immunohistochemistry (IHC). In this section, the main techniques used in experimental SCI in rodents are summarized.
### **6.1. Tracing methods**
### *6.1.1. Conventional tracing*
Axonal tracing is an important tool for the investigation of regeneration after SCI (Figure 3B), holding the advantage that specific axonal populations are precisely marked. Conventional tracers label axons and neurons via the axonal transport [80]. Neuronal tracers can label the axons anterogradely (toward the axon terminal) which is the preferred method for analyzing their sprouting and regeneration. Retrogradely transported tracers (towards the cell body) are injected at the distal side of a lesion, in order to quantify the number of neurons with regen‐ erated (distal) fibers and to visualize propriospinal neurons (see Section 6.4.).
Preferred application methods are pressure injection (liquid tracers), iontophoretic injection of electrically charged tracer molecules, or the insertion of dye crystals (carbocyanine dyes) [80]. The tracer can be detected via confocal microscopy using either its own fluorescence or IHC. The ideal survival time after the tracing depends on the tracer used, the distance between the site of tracer application and the area of interest, and the rate of its transport in the axons. A drawback of conventional tracing techniques is that in most cases not all axons of a neuronal population take up and transport the tracer substance. Many classical retrograde tracers are only, or more efficiently, taken up by injured axons and axon terminals, whereas the rate of the uptake by uninjured axons of passage is rather small. This can lead to nonspecific results [81].
Some examples for widely used monosynaptic neuronal tracers are the enzyme horseradish peroxidase, biotinylated dextran amine (BDA), and Fluoro Gold™. Examples of nonviral polysynaptic tracers are bacterial toxins, such as cholera toxin B. The drawback of nonviral polysynaptic tracers is, however, the dilution of signal after each synaptic step [82]. For the purpose of multisynaptic tracing, viral tracings are more suitable.
### *6.1.2. Viral tracings*
both regeneration from injured tracts and sprouting from spared tracts can be studied. It may be of importance to note that sprouting can be undirected, so that aberrant neuronal circuits may be formed [76]. Treatments can enhance sprouting and direct the sprouts to establish
After the primary insult, secondary damage due to, among others, inflammation, oxidative stress, and blood-brain barrier dysfunction causes the death of neurons (and glia) in the tissue surrounding the lesion [77]. The loss of local motor neurons leads to more extensive motor deficits in addition to the impairments caused by the injury to the descending motor tracts. The loss of spinal interneurons may disrupt intraspinal connections between motor centers. Therefore, a neuroprotective action of a treatment might, first, reduce functional impairments and, second, increase the possibility of local plasticity via interneurons (see Section 5.1.). For the analysis of neuroprotection, quantification of (moto-) neurons is performed at various distances rostrally and caudally from the lesion center [78]. A treatment could also lead to the protection of the brain and brainstem projection neurons from death or atrophy [8, 31, 79]. Quantification of the lesion size and spared white matter in standardized lesion models might
also provide information about the protective effects of a treatment strategy.
**6. Tracing and/or immunohistochemical (IHC) staining of motor and**
section, the main techniques used in experimental SCI in rodents are summarized.
erated (distal) fibers and to visualize propriospinal neurons (see Section 6.4.).
The next step in SCI research is the histological analysis of regenerated axons. Short-term studies (up to several weeks after the injury) give information about early injury events, whereas long-term studies (up to several months or even years after the injury) are useful to investigate long-term effects and behavioral outcomes with a treatment compared to a control. In order to visualize regenerating axons from the specific spinal cord tracts, these can be marked via axonal tracing (Figure 3B) or detected by immunohistochemistry (IHC). In this
Axonal tracing is an important tool for the investigation of regeneration after SCI (Figure 3B), holding the advantage that specific axonal populations are precisely marked. Conventional tracers label axons and neurons via the axonal transport [80]. Neuronal tracers can label the axons anterogradely (toward the axon terminal) which is the preferred method for analyzing their sprouting and regeneration. Retrogradely transported tracers (towards the cell body) are injected at the distal side of a lesion, in order to quantify the number of neurons with regen‐
functional circuits.
16 Recovery of Motor Function Following Spinal Cord Injury
**5.2. Neuroprotection**
**sensory tracts**
**6.1. Tracing methods**
*6.1.1. Conventional tracing*
When transneuronal tracing is desired, viral tracings are the method of choice. Transneuronal tracing is useful for the investigation of multisynaptic pathways and circuits [82]. The virus, which expresses a reporter gene in order to achieve the tracing, can replicate in the neurons and then infect other neurons which are connected via synapses. The virus replication further amplifies the signal, thereby avoiding the problem of signal dilution [8]. Very importantly, viral vector systems are very effective tools for gene therapeutic approaches. Frequently used viral systems used for axon tracing are adeno-associated viral vectors [83], lentiviral vectors [84], rabies virus [82], and herpes simplex virus [85]. The combination of viral tracings and gene therapy further offers the possibility to deliver a vector into specific areas.
A very elegant approach to investigate axonal pathways and their regeneration is the combi‐ nation of viral tracing with optical tissue clearing and light sheet laser scanning microscopy [86–88].
### **6.2. Anterograde tracing of defined tracts**
### *6.2.1. Motor cortex—CST tracing*
In SCI research, the CST is the most established model tract for the investigation of regeneration and the associated locomotor functional outcome. Its origin in the sensorimotor cortex and its course through the pyramidal decussations and, in rodents, the dorsal center part of the spinal cord allow a very precise labeling and localization of the tract. By using a stereotactic frame, precise injections of the tracer of choice are applied into the sensorimotor cortex [8, 89]. In general, tracing is performed 2 (mice) to 3 (rats) weeks before sacrifice of the animals for histological analysis. In the case of BDA, tissue sections need to be stained with streptavidin coupled to a fluorescent marker. Fluorescently labeled BDA is available, but the signal is usually still enhanced by post-staining. Analysis is performed by confocal microscopy, counting regenerating axon profiles in and beyond the lesion site.
### *6.2.2. Nucleus ruber—RST tracing*
The nucleus ruber can be traced in the same way as the CST; however, it is much smaller in size and therefore easier to miss [73].
### *6.2.3. Ascending sensory tracts: CTB tracing or CGRP staining*
Cholera toxin β (CTB) is a tracer that is transported anterogradely, retrogradely and, as a recent study suggested, even transneuronally [90]. This tracer is used frequently to label the ascend‐ ing sensory tracts in the dorsal column of the spinal cord. For this purpose, CTB is injected into the sciatic nerve that is crushed to achieve maximum uptake of the tracer [90]. This allows the specific analysis of the regeneration of ascending axons corresponding to the hindlimbs. In contrast, IHC staining for the marker calcitonin gene-related peptide CGRP allows the detection of axon profiles entering the spinal cord at all spinal segments. This, however, compromises the analysis of CGRP axons beyond the lesion, since axons from intact spinal levels above the lesion will also stain positively.
### **6.3. Raphespinal and coeruleospinal tracts**
Because of their neurotransmitters serotonin (5-HT) and noradrenaline (NA), whose key synthesizing enzyme is tyrosine hydroxylase (TH), the tracts descending from the Raphe nuclei and locus coeruleus can be investigated by IHC using 5-HT- and TH-specific antibodies [73]. Since their fibers run both ventrally and dorsally, care should be taken to analyze only areas that are relevant to the localization of the lesion, for example, only the dorsal funiculus in case of a dorsal hemisection, with respect to regeneration. The possibility of sprouting from ventral axons cannot be ruled out, and some types of interneurons also express 5-HT.
### **6.4. Retrograde tracing**
A very valuable tool for tracing regenerating neurons is retrograde tracing. When a tracer injected distally from the lesion site marks neurons proximally from the lesion site, these neurons have regenerated their fibers (provided the tracer is precisely located to the lesioned and not the spared region, and tracer diffusion can be ruled out). It can also answer the question whether axons from the intact side sprouted to the lesioned side. This technique was applied to show which brainstem nuclei were projecting into the distal cord [14] and to trace proprio‐ spinal interneuron networks [70]. If a retrograde tracer is applied to the spinal cord proximal to the lesion site, it can also be used to quantify the neurons "associated" with the lesion, for studying cell death or atrophy of neuronal populations [8].
### **7. Assessment of motor function**
usually still enhanced by post-staining. Analysis is performed by confocal microscopy,
The nucleus ruber can be traced in the same way as the CST; however, it is much smaller in
Cholera toxin β (CTB) is a tracer that is transported anterogradely, retrogradely and, as a recent study suggested, even transneuronally [90]. This tracer is used frequently to label the ascend‐ ing sensory tracts in the dorsal column of the spinal cord. For this purpose, CTB is injected into the sciatic nerve that is crushed to achieve maximum uptake of the tracer [90]. This allows the specific analysis of the regeneration of ascending axons corresponding to the hindlimbs. In contrast, IHC staining for the marker calcitonin gene-related peptide CGRP allows the detection of axon profiles entering the spinal cord at all spinal segments. This, however, compromises the analysis of CGRP axons beyond the lesion, since axons from intact spinal
Because of their neurotransmitters serotonin (5-HT) and noradrenaline (NA), whose key synthesizing enzyme is tyrosine hydroxylase (TH), the tracts descending from the Raphe nuclei and locus coeruleus can be investigated by IHC using 5-HT- and TH-specific antibodies [73]. Since their fibers run both ventrally and dorsally, care should be taken to analyze only areas that are relevant to the localization of the lesion, for example, only the dorsal funiculus in case of a dorsal hemisection, with respect to regeneration. The possibility of sprouting from ventral
A very valuable tool for tracing regenerating neurons is retrograde tracing. When a tracer injected distally from the lesion site marks neurons proximally from the lesion site, these neurons have regenerated their fibers (provided the tracer is precisely located to the lesioned and not the spared region, and tracer diffusion can be ruled out). It can also answer the question whether axons from the intact side sprouted to the lesioned side. This technique was applied to show which brainstem nuclei were projecting into the distal cord [14] and to trace proprio‐ spinal interneuron networks [70]. If a retrograde tracer is applied to the spinal cord proximal to the lesion site, it can also be used to quantify the neurons "associated" with the lesion, for
axons cannot be ruled out, and some types of interneurons also express 5-HT.
studying cell death or atrophy of neuronal populations [8].
counting regenerating axon profiles in and beyond the lesion site.
*6.2.3. Ascending sensory tracts: CTB tracing or CGRP staining*
levels above the lesion will also stain positively.
**6.3. Raphespinal and coeruleospinal tracts**
**6.4. Retrograde tracing**
*6.2.2. Nucleus ruber—RST tracing*
size and therefore easier to miss [73].
18 Recovery of Motor Function Following Spinal Cord Injury
In order to assess motor recovery after experimental SCI and putative regenerative treatments, several functional tests are available. The choice of the tests depends not only on the lesion model but also on the costs, because some tests require specific commercial systems.
### **7.1. Basso, Beattie, and Bresnahan (BBB) locomotor score and subscore (rat) and Basso Mouse Scale (BMS, mouse)**
The BBB open-field test, developed by Basso, Beattie, and Bresnahan [91], is an established test for the evaluation of hindlimb locomotor function of SCI rats. It is suitable for thoracic SCI models where it has become the first choice test to evaluate locomotor function [92]. The BBB score is based on the classification of hindlimb locomotor function using a scale which ranges from 0 (no spontaneous movement of the hindlimbs) to 21 (normal movement, coordinated walking pattern). For the evaluation procedure, the rats are placed in a defined open field where they are observed and evaluated by two trained observers. The animal's movements in the open field are scored over 4 minutes according to the criteria of the BBB locomotor rating scale [91]. The evaluation of coordination, an important parameter of the intermediate and late phases of the BBB, is not always clear without any doubt. This entails ratings in the mediumrange scale intervals often leading to an artificial plateau. Therefore, and because usually not all aspects of locomotion are influenced by a treatment, the determination of a BBB subscore can be helpful to improve the sensitivity of the test [93]. Furthermore, additional automatic gait analysis helps to avoid potential subjective evaluation of coordination [94]. An advantage of the BBB locomotor rating scale is that preoperative training—which is a general requirement for many locomotor behavioral tests—is not necessary. However, as is generally the case for behavioral tests, preoperative handling of the experimental animals and their familiarization with the test surroundings are useful. Additionally, adaptations of the original BBB locomotor scale have been described also for severe thoracic injuries such as complete spinal cord transection [56, 95]. Since such severe lesions result in maximum BBB scores of 8–10, the spreading of the low and intermediate BBB values (BBB 1–10) allows a distinct evaluation of less prominent locomotor behavioral improvements.
The small size and rapid speed of mice caused investigators to develop a mouse-specific scale, the BMS [96]. The procedure of the animal walking in an open field is basically the same as described above, but parameters like coordination, paw position, and trunk instability are evaluated in a slightly different way than for rats. Similarly, for unilateral cervical SCI new locomotor rating scales have been developed, such as the forelimb locomotor assessment scale (FLAS) [97] or the forelimb locomotor scale (FLS) [98].
### **7.2. Horizontal ladder rung test and Gridwalk**
The horizontal ladder walking test is used for the evaluation of fine motor control, coordina‐ tion, and foot placing accuracy, all of which require certain degrees of sensory feedback. Therefore, this test is particularly useful for the investigation of locomotion after a thoracic CST injury. Video analyses of the runs allow the assessment of multiple parameters [99]. The mistakes the animals make during walking are evaluated and classified into predefined categories. The test apparatus consists of metal rungs (3 mm in diameter) placed between Plexiglas walls in predefined intervals (1–5 cm for rats). The spacing patterns should be regularly alternated to make sure that the animals' locomotor function and not their cognitive functions are evaluated. Care should be taken to provide gaps between the rungs that are neither too narrow (mistakes being made by an animal might not be observable) nor too large (the animal cannot walk across without fear of falling, or without having to jump between rungs). During pretraining, the animals learn to cross the horizontal ladder without interrup‐ tion. Post-injury runs are recorded with a (high-speed) video camera from an angle slightly below the rod plane. This ensures the possibility to detect precise movements of all four paws and their digits. For evaluation, the predefined foot placing mistakes are counted. For mice, the procedure is similar, with smaller spacing (approximately 15 mm). For both species, several parameters of skilled walking can be observed, including correct placement, slight and deep slip, total miss, (partial) replacement, and correction [22, 99]. Alternative to the ladder test, the Gridwalk test makes use of grids to asses skilled walking.
### **7.3. Automated gait analysis methods**
### *7.3.1. CatWalk™*
The CatWalk™ system for automated quantitative gait analysis in SCI rats was developed by Hamers et al (2001). Classically, gait analysis in the form of footprint analyses has been (and still is by some groups) performed by painting the animal's paws with ink and letting it run on paper (or, an elegant variation, with developer and photographic paper) [100]. Static measures such as the distance between paws and toes could be measured, but no spatiotem‐ poral resolution was achieved. The Catwalk™ system consists of a glass plate through which fluorescent light is internally reflected. When a mouse or rat places its paw on the glass, the light is deflected from the glass and the paw print lights up. The intensity is related to the pressure or weight support, which provides additional information about the functionality of the paw. A high-speed camera placed below the glass plate records all the runs (originally, a mirror projected the light toward the camera, but the commercial version (Noldus) images directly). A narrow walkway corridor on top of the glass plate ensures that the animals walk in a straight line. After a few days of habituation training, the animals walk steadily through the corridor. Recording is performed in the dark, but the commercial setup has a lid with red light above the walkway, so that the outline of the animal is visualized. After analysis, main parameters of interest are the stride length (step size), the base of support (distance between left and right paws), the walking speed, the duration of the swing and stand phase, the regularity index as a measure of coordination, and the intensity of the prints. Many more parameters can be studied, the choice of which can be based on the animal model [101, 102]. The CatWalk™ system has been used in the following SCI models: thoracic CST Tx, RST Tx, and dorsal Hx [8, 103]; thoracic contusion [101, 102]; pyramidotomy models [22]; and lateral cervical spinal cord contusion [104, 105], and in recent studies assessing the effects of training and gene therapy [106, 107]. The CatWalk™ can furthermore be combined with the horizontal ladder test by placing the ladder above the glass plate, so that footslips light up because the animal touches the glass plate [108].
### *7.3.2. Automated gait analysis using treadmill*
mistakes the animals make during walking are evaluated and classified into predefined categories. The test apparatus consists of metal rungs (3 mm in diameter) placed between Plexiglas walls in predefined intervals (1–5 cm for rats). The spacing patterns should be regularly alternated to make sure that the animals' locomotor function and not their cognitive functions are evaluated. Care should be taken to provide gaps between the rungs that are neither too narrow (mistakes being made by an animal might not be observable) nor too large (the animal cannot walk across without fear of falling, or without having to jump between rungs). During pretraining, the animals learn to cross the horizontal ladder without interrup‐ tion. Post-injury runs are recorded with a (high-speed) video camera from an angle slightly below the rod plane. This ensures the possibility to detect precise movements of all four paws and their digits. For evaluation, the predefined foot placing mistakes are counted. For mice, the procedure is similar, with smaller spacing (approximately 15 mm). For both species, several parameters of skilled walking can be observed, including correct placement, slight and deep slip, total miss, (partial) replacement, and correction [22, 99]. Alternative to the ladder test, the
The CatWalk™ system for automated quantitative gait analysis in SCI rats was developed by Hamers et al (2001). Classically, gait analysis in the form of footprint analyses has been (and still is by some groups) performed by painting the animal's paws with ink and letting it run on paper (or, an elegant variation, with developer and photographic paper) [100]. Static measures such as the distance between paws and toes could be measured, but no spatiotem‐ poral resolution was achieved. The Catwalk™ system consists of a glass plate through which fluorescent light is internally reflected. When a mouse or rat places its paw on the glass, the light is deflected from the glass and the paw print lights up. The intensity is related to the pressure or weight support, which provides additional information about the functionality of the paw. A high-speed camera placed below the glass plate records all the runs (originally, a mirror projected the light toward the camera, but the commercial version (Noldus) images directly). A narrow walkway corridor on top of the glass plate ensures that the animals walk in a straight line. After a few days of habituation training, the animals walk steadily through the corridor. Recording is performed in the dark, but the commercial setup has a lid with red light above the walkway, so that the outline of the animal is visualized. After analysis, main parameters of interest are the stride length (step size), the base of support (distance between left and right paws), the walking speed, the duration of the swing and stand phase, the regularity index as a measure of coordination, and the intensity of the prints. Many more parameters can be studied, the choice of which can be based on the animal model [101, 102]. The CatWalk™ system has been used in the following SCI models: thoracic CST Tx, RST Tx, and dorsal Hx [8, 103]; thoracic contusion [101, 102]; pyramidotomy models [22]; and lateral cervical spinal cord contusion [104, 105], and in recent studies assessing the effects of training and gene therapy [106, 107]. The CatWalk™ can furthermore be combined with the horizontal
Gridwalk test makes use of grids to asses skilled walking.
**7.3. Automated gait analysis methods**
20 Recovery of Motor Function Following Spinal Cord Injury
*7.3.1. CatWalk™*
The CatWalk™ is semi-automated, because the animals must voluntarily walk across the walkway and need pretraining. Again, scientists are striving to improve the existing systems (Neckel, 2015). New fully automated gait analysis platforms have been developed, including the DigiGait™ (Mouse Specifics, Inc.) [109–111] and the TreadScan™ (Clever Sys Incorporated) systems [112]. These two systems use transparent treadmills allowing the animals' gait analysis at constant speed, including the possibility to measure at different speeds.
### *7.3.3. MotoRater and kinematic analysis*
The growing number of SCI models is accompanied by the need to modify the test systems. Recently, a new method for profiling locomotor recovery was developed in the lab of Martin Schwab [113]. This setup, now commercially available as the so-called MotoRater (TSE Systems), makes use of mirrors to image the mouse or rat that is walking in a Plexiglas basin from three sides (left, right, and below). The animals are tattooed on anatomical landmarks such as ilias crest, trochanter major of the hip, condylus lateralis of the knee, malleolus lateralis of the ankle, and the tip of the fifth toe. This way the walking is precisely monitored as stick diagrams and followed in time. As with the CatWalk™, the kinetics of even-ground walking patterns are analyzed. In contrast to the Catwalk, the researchers included new levels of difficulty in this system. A horizontal ladder is introduced to monitor precise paw placement and forelimb-hindlimb coordination. Alternatively, the basin is filled with water, either at levels where animals are wading (3 cm for rats, 1 cm for mice) or at levels where the animals have to swim. Wading brings the advantage that the water provides weight support. Further‐ more, the animal's strength can be measured, because of the desire of the animal to raise its body as much as possible out of the water. In the original article, three types of SCI were compared (dorsal Hx, ventral Hx, and lateral Hx). For each lesion model, various aspects of the test revealed to be suitable in different ways. For example, skilled walking and overground locomotion are most suitable for the evaluation of thoracic dorsal Hx. For thoracic ventral Hx, wading was described to be the better test and for cervical lateral Hx, the authors observed improvement of hindlimb movements during wading and swimming. Due to the forepaw impairment, cervical Hx animals can hardly perform the ladder test and are poor at normal even-ground locomotion. Further studies of the same group made use of the MotoRater to assess the contribution of the brain stem nuclei to locomotor recovery [14] and the effects of training on motor skills after SCI [71].
Another kinematic gait analysis system makes use of reflective markers at essentially the same hallmarks as the MotoRater system (iliac crest, hip, knee, ankle, metatarsophalangeal joint, and toe). A motion capture system (SIMI Reality Motion Systems) is used to analyze gait parameters combined with electromyogram recording (EMG) [69, 70].
### **7.4. Sensory testing**
Although less relevant than motor recovery, the recovery of sensory functions has a potential impact on locomotion. Furthermore, lesioned animals can develop neuropathic pain [114] which may be attenuated or, worse, aggravated by a treatment. Sensory tests performed after SCI include mechanical and nociceptive tests.
Sensorimotor reflexes can be tested by light touch to the paw, causing contact placing of the paw. Proprioceptive placing is elicited by stretching a tendon or joint [21]. Von Frey filaments are used to assess the animal's sensitivity to sub-threshold mechanical stimuli. For this purpose, filaments of increasing thickness are applied to the foot sole, exerting a defined force. This is normally not painful to the animal, so that only animals that suffer from mechanical allodynia (pain reaction from a normally non-painful stimulus) withdraw their paw from the filament. The minimum force eliciting a pain response is scored as paw withdrawal threshold [115]. Electronic versions of this test are available commercially (e.g., IITC Life Science, Ugo Basile). For the assessment of cutaneous hyperalgesia (increased pain from a pain-provoking stimulus), a hot plate or a commercial Plantar Test setup (e.g., Hargreaves Apparatus, Ugo Basil) [116] is used. The paw of interest is placed on a source of radiant heat or, in the case of the Planar Test, an infrared beam is precisely aimed at the central part of the animal's sole. The paw withdrawal time is recorded. Each paw is tested three times since the animal can also withdraw the paw spontaneously. Compared to the traditional hot plate test setup, the Plantar Test has the advantage of an automated, and therefore, accurate end-point detection [116].
For the majority of the sensory tests, the animal has to be able to move (withdraw) the paw. They can, therefore, generally not be performed with severely and completely spinal cordinjured animals that often lack the ability to perform limb movements below the level of the injury. For severely injured animals, the tail-flick test, a modification of the plantar hot plate test where the base of the tail is heated, can be applied [92].
### **7.5. Forelimb tests**
For cervical hemisection lesions and for pyramidotomy, specific tests to analyze forelimb motor recovery have been developed [21]. Since these lesions are usually one-sided, the healthy side serves as an internal control. First, new locomotor rating scales (alternatives for the BBB) have been developed, such as the forelimb locomotor assessment scale (FLAS) [97] or the forelimb locomotor scale (FLS) [98]. Second, broad tests for paw preference are applied, such as the cylinder test, where the choice of the weight-bearing forelimb is monitored [22], and the grooming test, where the preferred paw for grooming is scored. Popular tests assessing dexterity include pasta eating or the Irvine, Beatti, Bresneham (IBB) forelimb rating scale, where the forelimb function is assessed, while the rat is eating a round-shaped cereal [117]. Furthermore, tests for the assessment of fine finger movements include the single pelletgrasping test or the staircase test [21, 71]. In these skilled forepaw tests, mice or rats have to reach for and grasp sugar pellets through a slit in a Plexiglas wall or from wells in a staircase setup (Lafayette Instruments (rat), Campden Instruments (mouse)). Video analyses of the sessions allow the assessment of multiple parameters. Some groups use the horizontal ladder as well to score forepaw locomotion, but the animals are usually poor at performing this test. To further quantify grip a commercial grip strength meter is available (TSE Systems, Ugo Basile, Columbus Instruments), or the ability of the animal to keep its balance and hold on stably to an inclined plane (or cage grid) is measured.
### **7.6. Important considerations for functional testing**
**7.4. Sensory testing**
**7.5. Forelimb tests**
SCI include mechanical and nociceptive tests.
22 Recovery of Motor Function Following Spinal Cord Injury
test where the base of the tail is heated, can be applied [92].
Although less relevant than motor recovery, the recovery of sensory functions has a potential impact on locomotion. Furthermore, lesioned animals can develop neuropathic pain [114] which may be attenuated or, worse, aggravated by a treatment. Sensory tests performed after
Sensorimotor reflexes can be tested by light touch to the paw, causing contact placing of the paw. Proprioceptive placing is elicited by stretching a tendon or joint [21]. Von Frey filaments are used to assess the animal's sensitivity to sub-threshold mechanical stimuli. For this purpose, filaments of increasing thickness are applied to the foot sole, exerting a defined force. This is normally not painful to the animal, so that only animals that suffer from mechanical allodynia (pain reaction from a normally non-painful stimulus) withdraw their paw from the filament. The minimum force eliciting a pain response is scored as paw withdrawal threshold [115]. Electronic versions of this test are available commercially (e.g., IITC Life Science, Ugo Basile). For the assessment of cutaneous hyperalgesia (increased pain from a pain-provoking stimulus), a hot plate or a commercial Plantar Test setup (e.g., Hargreaves Apparatus, Ugo Basil) [116] is used. The paw of interest is placed on a source of radiant heat or, in the case of the Planar Test, an infrared beam is precisely aimed at the central part of the animal's sole. The paw withdrawal time is recorded. Each paw is tested three times since the animal can also withdraw the paw spontaneously. Compared to the traditional hot plate test setup, the Plantar Test has the advantage of an automated, and therefore, accurate end-point detection [116]. For the majority of the sensory tests, the animal has to be able to move (withdraw) the paw. They can, therefore, generally not be performed with severely and completely spinal cordinjured animals that often lack the ability to perform limb movements below the level of the injury. For severely injured animals, the tail-flick test, a modification of the plantar hot plate
For cervical hemisection lesions and for pyramidotomy, specific tests to analyze forelimb motor recovery have been developed [21]. Since these lesions are usually one-sided, the healthy side serves as an internal control. First, new locomotor rating scales (alternatives for the BBB) have been developed, such as the forelimb locomotor assessment scale (FLAS) [97] or the forelimb locomotor scale (FLS) [98]. Second, broad tests for paw preference are applied, such as the cylinder test, where the choice of the weight-bearing forelimb is monitored [22], and the grooming test, where the preferred paw for grooming is scored. Popular tests assessing dexterity include pasta eating or the Irvine, Beatti, Bresneham (IBB) forelimb rating scale, where the forelimb function is assessed, while the rat is eating a round-shaped cereal [117]. Furthermore, tests for the assessment of fine finger movements include the single pelletgrasping test or the staircase test [21, 71]. In these skilled forepaw tests, mice or rats have to reach for and grasp sugar pellets through a slit in a Plexiglas wall or from wells in a staircase setup (Lafayette Instruments (rat), Campden Instruments (mouse)). Video analyses of the sessions allow the assessment of multiple parameters. Some groups use the horizontal ladder as well to score forepaw locomotion, but the animals are usually poor at performing this test.
Several studies indicate that the choice of the motor tests should be based on the type of injury and the degree of impairment [113, 118]. For thoracic dorsal Hx, the horizontal ladder test and CatWalk™ gait analysis systems are suitable since even-ground walking and skilled walking are impaired, but display recovery over time. In the case of ventral Hx, the wading and swimming paradigms in the MotoRater provide more useful information on impairment and recovery. Cervical lateral Hx animals also perform better during wading and swimming. With regard to swimming, an assessment tool was developed in Sweden, where parameters like fore- and hindlimb usage, hindlimb alternation and position, trunk instability, body angle, and tail movements are precisely scored [119].
Care should, however, always be taken with the evaluation of the results. Animals can develop compensation strategies to perform a task in a different way than before the injury [118]. For example, animals primarily use their hindlimbs for swimming, but after a thoracic injury, they utilize their forelimbs. Therefore, distance or speed may recover, but the actual functional recovery of the hindlimbs might be still impaired. Another example is the grasping of food pellets that animals with forelimb impairment cannot do. Some animals tend to successfully develop a scooping strategy to retrieve pellets [118]. Investigators should be aware of this and monitor the strategies the animals use. The use of video equipment to accompany a test is therefore advisable.
The strain of the animals (and even the substrain produced by different suppliers) also plays an important role. Some animal strains perform better than others in tests which require the acquisition of certain skills [118, 120]. For example, in the staircase-skilled forepaw reaching test, Lister-hooded and Long-Evans rats perform much better than Lewis rats and Fischer rats [121, 122]. Housing is also of importance, since the amount of motor activity in the cage can provide training effects. This might mask a treatment effect, because the spontaneous recovery due to training may be too prominent. A popular cage enrichment in the form of sunflower seeds might compromise skilled grasping tests [118]. On the contrary, if the chosen test is too difficult for the animals in view of their impairment, recovery of function might be missed too. Other variables like circadian rhythms and stress can introduce variability. Therefore, it is vital to habituate the animals to the experimenters, to perform pre-injury recordings of the basal performance of the animals in the tests and to perform testing always at the same time of day under the same circumstances.
### **8. Discussion and conclusions**
This chapter provides an overview of the main rodent models, experimental treatment strategies, histological analysis methods, and motor tests that are available for the investigation of neuronal regeneration and locomotor function after experimental SCI. The choice of the appropriate model depends on the research question and on the type of human injury which the investigation is based on. There is ongoing controversy regarding the comparability of experimental blunt versus sharp lesions to the clinical situation of human patients. Contusion/ compression injuries are very suitable for studying human traumatic SCI. These types of injury maintain tissue continuity even in the most severe cases, which is also observed in the vast majority of human spinal cord traumata. However, spared tissue bridges might compromise the analysis of treatment effects in experimental SCI. Moreover, blunt force spinal cord traumata are often accompanied by sharp lesions like maceration, laceration, or transection, for example by bone splinters. Therefore, sharp transections are also valid models, not least because they are easier to control and reproduce.
SCI experiments in rodents are essential for the development of new treatment strategies. They aim to extensively test treatment effects on multiple nerve tracts, to elucidate their mechanisms of action and, using multiple motor and sensory tests, to shed light on their ability to restore function. It is highly important to know whether a treatment is effective via neuroprotection, spared axon sprouting, or axon regeneration, since this will influence the choice of treatment that suits the patient best. Patients with incomplete lesions may benefit from plasticityinducing treatments, whereas patients suffering from complete injuries require therapeutic strategies that induce regeneration. Patients with contusion lesions or complete injuries might further benefit from matrix or stem cell implantation to fill up cavities. When a treatment strategy displays promising effects in rodent SCI models, the next step will be to test it in a model system that is more close to human patients. In primates, the CST projects mainly dorsolaterally and originates from both left and right motor cortex, because a number of CST axons decussate along the spinal cord midline. These axons are capable of forming detour circuits reconnecting the motor cortex with denervated spinal cord areas in monkeys with lateral cervical Hx [123–125]. Due to the comparability with the anatomy of humans, the nonhuman primate cervical Hx model has been proposed to be a suitable model to test the recovery of forelimb skills after SCI [126].
Rodent research provided numerous important insights into the SCI field. To name a few, the regeneration and/or sprouting responses of tracts involved in locomotion, the involvement of the propriospinal system, the CPG circuits, and the ability to stimulate these without supra‐ spinal input all contributed to a better understanding of human spinal cord pathophysiology. Numerous treatments have been tested and have provided even more insights into how the various systems can be manipulated. However, to date, despite many years of extensive research, there are no clinical standard therapies for SCI which significantly increase the regenerative response to such a degree that they achieve strong (locomotor) improvements in human patients. This reflects the complexity of SCI. Although many treatments did not reach the clinic, they have been of enormous value to understanding the mechanisms of regeneration leading to functional motor recovery.
### **Author details**
of neuronal regeneration and locomotor function after experimental SCI. The choice of the appropriate model depends on the research question and on the type of human injury which the investigation is based on. There is ongoing controversy regarding the comparability of experimental blunt versus sharp lesions to the clinical situation of human patients. Contusion/ compression injuries are very suitable for studying human traumatic SCI. These types of injury maintain tissue continuity even in the most severe cases, which is also observed in the vast majority of human spinal cord traumata. However, spared tissue bridges might compromise the analysis of treatment effects in experimental SCI. Moreover, blunt force spinal cord traumata are often accompanied by sharp lesions like maceration, laceration, or transection, for example by bone splinters. Therefore, sharp transections are also valid models, not least
SCI experiments in rodents are essential for the development of new treatment strategies. They aim to extensively test treatment effects on multiple nerve tracts, to elucidate their mechanisms of action and, using multiple motor and sensory tests, to shed light on their ability to restore function. It is highly important to know whether a treatment is effective via neuroprotection, spared axon sprouting, or axon regeneration, since this will influence the choice of treatment that suits the patient best. Patients with incomplete lesions may benefit from plasticityinducing treatments, whereas patients suffering from complete injuries require therapeutic strategies that induce regeneration. Patients with contusion lesions or complete injuries might further benefit from matrix or stem cell implantation to fill up cavities. When a treatment strategy displays promising effects in rodent SCI models, the next step will be to test it in a model system that is more close to human patients. In primates, the CST projects mainly dorsolaterally and originates from both left and right motor cortex, because a number of CST axons decussate along the spinal cord midline. These axons are capable of forming detour circuits reconnecting the motor cortex with denervated spinal cord areas in monkeys with lateral cervical Hx [123–125]. Due to the comparability with the anatomy of humans, the nonhuman primate cervical Hx model has been proposed to be a suitable model to test the
Rodent research provided numerous important insights into the SCI field. To name a few, the regeneration and/or sprouting responses of tracts involved in locomotion, the involvement of the propriospinal system, the CPG circuits, and the ability to stimulate these without supra‐ spinal input all contributed to a better understanding of human spinal cord pathophysiology. Numerous treatments have been tested and have provided even more insights into how the various systems can be manipulated. However, to date, despite many years of extensive research, there are no clinical standard therapies for SCI which significantly increase the regenerative response to such a degree that they achieve strong (locomotor) improvements in human patients. This reflects the complexity of SCI. Although many treatments did not reach the clinic, they have been of enormous value to understanding the mechanisms of regeneration
because they are easier to control and reproduce.
24 Recovery of Motor Function Following Spinal Cord Injury
recovery of forelimb skills after SCI [126].
leading to functional motor recovery.
Christina F. Vogelaar1\* and Veronica Estrada2
\*Address all correspondence to: tineke.vogelaar@unimedizin-mainz.de
1 Institute of Microanatomy and Neurobiology, Johannes Gutenberg-University, Mainz, Germany
2 Molecular Neurobiology Laboratory, Department of Neurology, Heinrich-Heine-Universi‐ ty, Duesseldorf, Germany
### **References**
[22] Starkey ML, Barritt AW, Yip PK, Davies M, Hamers FP, McMahon SB, et al. Assessing behavioural function following a pyramidotomy lesion of the corticospinal tract in adult mice. Experimental Neurology. 2005;195(2):524–39.
[9] Vogelaar CF, Konig B, Krafft S, Estrada V, Brazda N, Ziegler B, et al. Pharmacological suppression of CNS scarring by deferoxamine reduces lesion volume and increases regeneration in an in vitro model for astroglial-fibrotic scarring and in rat spinal cord
[10] Hill RL, Zhang YP, Burke DA, Devries WH, Zhang Y, Magnuson DS, et al. Anatomical and functional outcomes following a precise, graded, dorsal laceration spinal cord
[11] Zhang YP, Walker MJ, Shields LB, Wang X, Walker CL, Xu XM, et al. Controlled cervical laceration injury in mice. Journal of Visualized Experiments: JoVE. 2013;75:e50030.
[12] Hermanns S, Klapka N, Muller HW. The collagenous lesion scar—an obstacle for axonal regeneration in brain and spinal cord injury. Restorative Neurology and
[13] Silver J, Miller JH. Regeneration beyond the glial scar. Nature Reviews Neuroscience.
[14] Zörner B, Bachmann LC, Filli L, Kapitza S, Gullo M, Bolliger M, et al. Chasing central nervous system plasticity: the brainstem's contribution to locomotor recovery in rats
[15] Mantilla CB, Greising SM, Stowe JM, Zhan WZ, Sieck GC. TrkB kinase activity is critical for recovery of respiratory function after cervical spinal cord hemisection. Experimen‐
[16] Surey S, Berry M, Logan A, Bicknell R, Ahmed Z. Differential cavitation, angiogenesis and wound-healing responses in injured mouse and rat spinal cords. Neuroscience.
[17] Young W. Spinal cord contusion models. Progress in Brain Research. 2002;137:231–55.
[18] Jakeman LB, Guan Z, Wei P, Ponnappan R, Dzwonczyk R, Popovich PG, et al. Trau‐ matic spinal cord injury produced by controlled contusion in mouse. Journal of
[19] Stokes BT, Jakeman LB. Experimental modelling of human spinal cord injury: a model that crosses the species barrier and mimics the spectrum of human cytopathology.
[20] Lee JH, Streijger F, Tigchelaar S, Maloon M, Liu J, Tetzlaff W, et al. A contusive model of unilateral cervical spinal cord injury using the infinite horizon impactor. Journal of
[21] Geissler SA, Schmidt CE, Schallert T. Rodent models and behavioral outcomes of cervical spinal cord injury. Journal of Spine. 2013;Suppl. 4. DOI:
injury in C57BL/6 mice. Journal of Neurotrauma. 2009;26(1):1–15.
injury in vivo. PLoS One. 2015;10(7):e0134371.
26 Recovery of Motor Function Following Spinal Cord Injury
Neuroscience. 2001;19(1–2):139–48.
tal Neurology. 2014;261:190–5.
Neurotrauma. 2000;17(4):299–319.
Spinal Cord. 2002;40(3):101–9.
10.4172/2165-7939.S4-001
Visualized Experiments: JoVE. 2012;65.
with spinal cord injury. Brain. 2014;137(Pt 6):1716–32.
2004;5(2):146–56.
2014;275:62–80.
[47] Ori S, Okada Y, Nishimura S, Sasaki T, Itakura G, Kobayashi Y, et al. Long-term safety issues of iPSC-based cell therapy in a spinal cord injury model: oncogenic transforma‐ tion with epithelial-mesenchymal transition. Stem Cell Reports. 2015;4(3):360–73.
[34] Tso D, McKinnon RD. Cell replacement therapy for central nervous system diseases.
[35] Papastefanaki F, Matsas R. From demyelination to remyelination: the road toward
[36] Jin Y, Bouyer J, Shumsky JS, Haas C, Fischer I. Transplantation of neural progenitor
[37] Slawinska U, Miazga K, Jordan LM. The role of serotonin in the control of locomotor movements and strategies for restoring locomotion after spinal cord injury. Acta
[38] Sharp KG, Yee KM, Steward O. A re-assessment of long distance growth and connec‐ tivity of neural stem cells after severe spinal cord injury. Experimental Neurology.
[39] Lu P, Wang Y, Graham L, McHale K, Gao M, Wu D, et al. Long-distance growth and connectivity of neural stem cells after severe spinal cord injury. Cell. 2012;150(6):1264–
[40] Lu Y, Wang MY. Neural stem cell grafts for complete spinal cord injury. Neurosurgery.
[41] Chen CT, Foo NH, Liu WS, Chen SH. Infusion of human umbilical cord blood cells ameliorates hind limb dysfunction in experimental spinal cord injury through antiinflammatory, vasculogenic and neurotrophic mechanisms. Pediatrics and Neonatol‐
[42] Schira J, Gasis M, Estrada V, Hendricks M, Schmitz C, Trapp T, et al. Significant clinical, neuropathological and behavioural recovery from acute spinal cord trauma by transplantation of a well-defined somatic stem cell from human umbilical cord blood.
[43] Schira J, Falkenberg H, Hendricks M, Waldera-Lupa DM, Kogler G, Meyer HE, et al. Characterization of regenerative phenotype of unrestricted somatic stem cells (USSC) from human umbilical cord blood (hUCB) by functional secretome analysis. Molecular
[44] Jones LL, Oudega M, Bunge MB, Tuszynski MH. Neurotrophic factors, cellular bridges and gene therapy for spinal cord injury. The Journal of Physiology. 2001;533(Pt 1):83–
[45] Samadikuchaksaraei A. An overview of tissue engineering approaches for manage‐ ment of spinal cord injuries. Journal of NeuroEngineering and Rehabilitation. 2007;4:15.
[46] Markakis EA, Redmond DE, Jr. Know thyself: autologous cell transplantation strategies
for brain repair. Experimental Neurology. 2005;196(1):6–8.
Neural Regeneration Research. 2015;10(9):1356–8.
Neurobiologiae Experimentalis. 2014;74(2):172–87.
2014;257:186–204.
28 Recovery of Motor Function Following Spinal Cord Injury
2012;71(6):N13–5.
ogy. 2008;49(3):77–83.
Brain. 2012;135(Pt 2):431–46.
and Cellular Proteomics. 2015;14(10):2630–43.
73.
9.
therapies for spinal cord injury. Glia. 2015;63(7):1101–25.
cells in chronic spinal cord injury. Neuroscience. 2016;320:69–82.
dose-dependently promotes motor skill acquisition and functional recovery. Neurore‐ habilitation and Neural Repair. 2014;28(6):594–605.
[72] Sharp KG, Duarte JE, Gebrekristos B, Perez S, Steward O, Reinkensmeyer DJ. Robotic rehabilitator of the rodent upper extremity: a system and method for assessing and training forelimb force production after neurological injury. Journal of Neurotrauma. 2016;33. DOI: 10.1089/neu.2015.3987
[59] Singh B, Xu QG, Franz CK, Zhang R, Dalton C, Gordon T, et al. Accelerated axon outgrowth, guidance, and target reinnervation across nerve transection gaps following a brief electrical stimulation paradigm. Journal of Neurosurgery. 2012;116(3):498–512.
[60] Young W. Electrical stimulation and motor recovery. Cell Transplant. 2015;24(3):429–
[61] van den Brand R, Heutschi J, Barraud Q, DiGiovanna J, Bartholdi K, Huerlimann M, et al. Restoring voluntary control of locomotion after paralyzing spinal cord injury.
[62] Lavrov I, Gerasimenko Y, Burdick J, Zhong H, Roy RR, Edgerton VR. Integrating multiple sensory systems to modulate neural networks controlling posture. Journal of
[63] Wenger N, Moraud EM, Gandar J, Musienko P, Capogrosso M, Baud L, et al. Spatio‐ temporal neuromodulation therapies engaging muscle synergies improve motor
[64] McPherson JG, Miller RR, Perlmutter SI. Targeted, activity-dependent spinal stimula‐ tion produces long-lasting motor recovery in chronic cervical spinal cord injury. Proceedings of the National Academy of Sciences U S A. 2015;112(39):12193–8.
[65] Angeli CA, Edgerton VR, Gerasimenko YP, Harkema SJ. Altering spinal cord excita‐ bility enables voluntary movements after chronic complete paralysis in humans. Brain.
[66] Koopmans GC, Brans M, Gomez-Pinilla F, Duis S, Gispen WH, Torres-Aleman I, et al. Circulating insulin-like growth factor I and functional recovery from spinal cord injury under enriched housing conditions. The European Journal of Neuroscience. 2006;23(4):
[67] Lankhorst AJ, ter Laak MP, van Laar TJ, van Meeteren NL, de Groot JC, Schrama LH, et al. Effects of enriched housing on functional recovery after spinal cord contusive
[68] Fouad K, Tetzlaff W. Rehabilitative training and plasticity following spinal cord injury.
[69] Courtine G, Song B, Roy RR, Zhong H, Herrmann JE, Ao Y, et al. Recovery of supra‐ spinal control of stepping via indirect propriospinal relay connections after spinal cord
[70] Shah PK, Garcia-Alias G, Choe J, Gad P, Gerasimenko Y, Tillakaratne N, et al. Use of quadrupedal step training to re-engage spinal interneuronal networks and improve
[71] Starkey ML, Bleul C, Kasper H, Mosberger AC, Zorner B, Giger S, et al. High-impact, self-motivated training within an enriched environment with single animal tracking
locomotor function after spinal cord injury. Brain. 2013;136(Pt 11):3362–77.
injury in the adult rat. Journal of Neurotrauma. 2001;18(2):203–15.
Experimental Neurology. 2012;235(1):91–9.
injury. Nature Medicine. 2008;14(1):69–74.
control after spinal cord injury. Nature Medicine. 2016;22(2):138–45.
46.
Science. 2012;336(6085):1182–5.
30 Recovery of Motor Function Following Spinal Cord Injury
2014;137(Pt 5):1394–409.
1035–46.
Neurophysiology. 2015;114(6):3306–14.
[96] Basso DM, Fisher LC, Anderson AJ, Jakeman LB, McTigue DM, Popovich PG. Basso Mouse Scale for locomotion detects differences in recovery after spinal cord injury in five common mouse strains. Journal of Neurotrauma. 2006;23(5):635–59.
[84] Parr-Brownlie LC, Bosch-Bouju C, Schoderboeck L, Sizemore RJ, Abraham WC, Hughes SM. Lentiviral vectors as tools to understand central nervous system biology in mammalian model organisms. Frontiers in Molecular Neuroscience. 2015;8:14. [85] McGovern AE, Davis-Poynter N, Rakoczy J, Phipps S, Simmons DG, Mazzone SB. Anterograde neuronal circuit tracing using a genetically modified herpes simplex virus
expressing EGFP. The Journal of Neuroscience Methods. 2012;209(1):158–67.
primates. eNeuro. 2015;2(2).
32 Recovery of Motor Function Following Spinal Cord Injury
One. 2015;10(5):e0124650.
95.
463–77.
2015;10(12):e0144030.
2002;16(3):467–76.
[86] Soderblom C, Lee DH, Dawood A, Carballosa M, Jimena Santamaria A, Benavides FD, et al. 3D imaging of axons in transparent spinal cords from rodents and nonhuman
[87] Schwarz MK, Scherbarth A, Sprengel R, Engelhardt J, Theer P, Giese G. Fluorescentprotein stabilization and high-resolution imaging of cleared, intact mouse brains. PLoS
[88] Ertürk A, Becker K, Jahrling N, Mauch CP, Hojer CD, Egen JG, et al. Three-dimensional imaging of solvent-cleared organs using 3DISCO. Nature Protocols. 2012;7(11):1983–
[89] Steward O, Zheng B, Ho C, Anderson K, Tessier-Lavigne M. The dorsolateral cortico‐ spinal tract in mice: an alternative route for corticospinal input to caudal segments following dorsal column lesions. The Journal of Comparative Neurology. 2004;472(4):
[90] Lai BQ, Qiu XC, Zhang K, Zhang RY, Jin H, Li G, et al. Cholera toxin B subunit shows transneuronal tracing after injection in an injured sciatic nerve. PLoS One.
[91] Basso DM, Beattie MS, Bresnahan JC. A sensitive and reliable locomotor rating scale
[92] Sedy J, Urdzikova L, Jendelova P, Sykova E. Methods for behavioral testing of spinal cord injured rats. Neuroscience and Biobehavioral Reviews. 2008;32(3):550–80. [93] Lankhorst AJ, Duis SE, ter Laak MP, Joosten EA, Hamers FP, Gispen WH. Functional recovery after central infusion of alpha-melanocyte-stimulating hormone in rats with
[94] Koopmans GC, Deumens R, Honig WM, Hamers FP, Steinbusch HW, Joosten EA. The assessment of locomotor function in spinal cord injured rats: the importance of objective
[95] Antri M, Orsal D, Barthe JY. Locomotor recovery in the chronic spinal rat: effects of long-term treatment with a 5-HT2 agonist. The European Journal of Neuroscience.
for open field testing in rats. Journal of Neurotrauma. 1995;12(1):1–21.
spinal cord contusion injury. Journal of Neurotrauma. 1999;16(4):323–31.
analysis of coordination. Journal of Neurotrauma. 2005;22(2):214–25.
hypersensitivity after experimental spinal cord injury. Journal of Neurotrauma. 2013;30(21):1805–11.
[121] Nikkhah G, Rosenthal C, Hedrich HJ, Samii M. Differences in acquisition and full performance in skilled forelimb use as measured by the 'staircase test' in five rat strains. Behavioural Brain Research. 1998;92(1):85–95.
[108] Smith JM, Lunga P, Story D, Harris N, Le Belle J, James MF, et al. Inosine promotes recovery of skilled motor function in a model of focal brain injury. Brain. 2007;130(Pt
[109] Neckel ND. Methods to quantify the velocity dependence of common gait measure‐ ments from automated rodent gait analysis devices. The Journal of Neuroscience
[110] Ek CJ, Habgood MD, Callaway JK, Dennis R, Dziegielewska KM, Johansson PA, et al. Spatio-temporal progression of grey and white matter damage following contusion
[111] Zhang B, Bailey WM, Braun KJ, Gensel JC. Age decreases macrophage IL-10 expression: implications for functional recovery and tissue repair in spinal cord injury. Experi‐
[112] Beare JE, Morehouse JR, DeVries WH, Enzmann GU, Burke DA, Magnuson DS, et al. Gait analysis in normal and spinal contused mice using the TreadScan system. Journal
[113] Zörner B, Filli L, Starkey ML, Gonzenbach R, Kasper H, Rothlisberger M, et al. Profiling locomotor recovery: comprehensive quantification of impairments after CNS damage
[114] Gwak YS, Kang J, Unabia GC, Hulsebosch CE. Spatial and temporal activation of spinal glial cells: role of gliopathy in central neuropathic pain following spinal cord injury in
[115] Detloff MR, Fisher LC, Deibert RJ, Basso DM. Acute and chronic tactile sensory testing after spinal cord injury in rats. Journal of Visualized Experiments: JoVE. 2012;62:e3247.
[116] Hargreaves K, Dubner R, Brown F, Flores C, Joris J. A new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia. Pain. 1988;32(1):77–88.
[117] Irvine KA, Ferguson AR, Mitchell KD, Beattie SB, Lin A, Stuck ED, et al. The Irvine, Beatties, and Bresnahan (IBB) Forelimb Recovery Scale: an assessment of reliability and
[118] Fouad K, Hurd C, Magnuson DS. Functional testing in animal models of spinal cord injury: not as straight forward as one would think. Frontiers in Integrative Neuro‐
[119] Xu N, Akesson E, Holmberg L, Sundstrom E. A sensitive and reliable test instrument to assess swimming in rats with spinal cord injury. Behavioural Brain Research.
[120] Kjell J, Sandor K, Josephson A, Svensson CI, Abrams MB. Rat substrains differ in the magnitude of spontaneous locomotor recovery and in the development of mechanical
injury in rat spinal cord. PLoS One. 2010;5(8):e12021.
4):915–25.
Methods. 2015;253:244–53.
34 Recovery of Motor Function Following Spinal Cord Injury
mental Neurology. 2015;273:83–91.
of Neurotrauma. 2009;26(11):2045–56.
in rodents. Nature Methods. 2010;7(9):701–8.
validity. Frontiers in Neurology. 2014;5:116.
science. 2013;7:85.
2015;291:172–83.
rats. Experimental Neurology. 2012;234(2):362–72.
## *In Vitro* **Models of Spinal Cord Injury**
Lucia Slovinska, Juraj Blasko, Miriam Nagyova, Eva Szekiova and Dasa Cizkova
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/63459
### **Abstract**
Living organisms are extremely complex functional systems. At present, there are many *in vivo* models of spinal cord injury (SCI) that allow the modeling of any type of central nervous system (CNS) injury, however, with some disadvantages. The production of injury models can be a highly invasive and time‐consuming process and requires high technical requirements, and costly financial issues should also be taken into account. Of course, a large number of animals have been used to obtain the relevant data of statistical significance. All of these aspects can be reduced by carrying out experiments in *in vitro* conditions. The primary advantage of *in vitro* method is that it simplifies the system under study. There are two major groups of *in vitro* model in use: cell culture and organotypic slice (OTS) culture. OTS is an intermediate system of the screening of *in vitro* cell culture and animal models and represents the *in vitro* system preserving the basic tissue architecture that able to closely mimic the cellular and physiological characteristics *in vivo*. *In vitro* models are the preferred methods for the study of acute or subacute pathophysiology after a trauma stimulus, enabling precise control on the extracellular environment, easy and repeatable access to the cells.
**Keywords:** *in vitro* model, spinal cord injury (SCI), cell culture, co‐culture, organotyp‐ ic slice (OTS) culture
### **1. Introduction**
Spinal cord injury (SCI) is a critical incurable difficulty affecting the quality of life of pa‐ tients and society equally. The spinal cord injury interferes predominantly with motorfunction defects, which substantially restrict normal daily activities, thus leading to serious psycho‐
© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
logical trauma. Today, the majority of patients are left paralyzed with unsuccessful treat‐ ment currently available. Many strategies, including surgical, transplantations, pharmacological, neurophysiological, and others approaches, have been used and examined in attempts to develop new and successful therapies that will allow patients to get the valuable life they lived before the injury [1].
First, before devising a new treatment for spinal cord injury, it is also important to know the processes of nerve tissue regeneration, neurogenesis, and gliogenesis, which can be influenced in many ways. The answer to these needs, a growing number of animal models have been introduced and utilized. The creation of *in vivo* injury models requires many invasive inter‐ ventions and is time consuming. The financial costs of such models as well as technical equipment are often high. However, the biggest problem of the *in vivo* models is the utilization of large amounts of experimental animals. All of these facts can be reduced by carrying out experiments in *in vitro* conditions. *In vitro* experimental models are needed to understand the cellular mechanisms and pathophysiology of the injured spinal cord (SC) and to provide evidence for further treatment. This is also the right place to mention the principles of the 3Rs (Replacement, Reduction, and Refinement). The principles of the 3Rs were developed as a framework for humane animal research and became embedded in national and international legislation regulating the use of animals in scientific procedures in most world countries. There are two major groups of *in vitro* model in use: cell culture and organotypic slice (OTS) culture.
### **2. Cell cultures**
After central nervous system (CNS) injury, the acute primary and the chronic secondary damage takes place. After initial mechanical insult, a rapid deformation of CNS tissue happens, causing an immediate cell death in the epicenter of injury site followed by a cascade of processes leading to the secondary damage. Traumatic injury disrupts spinal white matter tracts resulting in the loss of sensory and motor function. This loss of function is generally permanent due to the limited regenerative capacity of the CNS [2]. Cells used for *in vitro* modeling should have morphological, functional qualities, and relevant requirements depending on disease, injury, and ultimate goal of the proposed experiments. The commonly used *in vitro* cell culture models of SCI include:
**i.** Primary isolated cells, neurons, oligodendrocytes, astrocytes, or microglia cells [3, 4]. Primary cells are obtained from the dissociated neural tissues, spinal cord, or brain and are usually cultivated immediately after the isolation as "primo‐cultures." The origin of these cells could be embryonal, fetal, neonatal, or adult [5, 6]. It is well known that younger cells are more potent according to their strong proliferative activity, in contrary to terminally differentiated poor proliferating capacity owning adult cells [6]. By using the primary cultures, we can investigate the behavior and intercommu‐ nication of the individual or mixed cell populations. These populations preserve the main characteristics similar to the *in vivo* examined types of cells [7]. Very often used are also the immortalized cell lines, which are in fact the primary cells with genetical modification.
**ii.** Co‐culture of neuronal cells with different cell types, which are present in the glial scar [8, 9]. *In vitro* cultures offer simplified, high throughput systems to study disease, drug toxicity, and biological processes by controlling environmental factors and directly measuring cell responses [10].
### **2.1.** *In vitro* **model of glial scar**
logical trauma. Today, the majority of patients are left paralyzed with unsuccessful treat‐ ment currently available. Many strategies, including surgical, transplantations, pharmacological, neurophysiological, and others approaches, have been used and examined in attempts to develop new and successful therapies that will allow patients to get the valuable
First, before devising a new treatment for spinal cord injury, it is also important to know the processes of nerve tissue regeneration, neurogenesis, and gliogenesis, which can be influenced in many ways. The answer to these needs, a growing number of animal models have been introduced and utilized. The creation of *in vivo* injury models requires many invasive inter‐ ventions and is time consuming. The financial costs of such models as well as technical equipment are often high. However, the biggest problem of the *in vivo* models is the utilization of large amounts of experimental animals. All of these facts can be reduced by carrying out experiments in *in vitro* conditions. *In vitro* experimental models are needed to understand the cellular mechanisms and pathophysiology of the injured spinal cord (SC) and to provide evidence for further treatment. This is also the right place to mention the principles of the 3Rs (Replacement, Reduction, and Refinement). The principles of the 3Rs were developed as a framework for humane animal research and became embedded in national and international legislation regulating the use of animals in scientific procedures in most world countries. There are two major groups of *in vitro* model in use: cell culture and organotypic slice (OTS) culture.
After central nervous system (CNS) injury, the acute primary and the chronic secondary damage takes place. After initial mechanical insult, a rapid deformation of CNS tissue happens, causing an immediate cell death in the epicenter of injury site followed by a cascade of processes leading to the secondary damage. Traumatic injury disrupts spinal white matter tracts resulting in the loss of sensory and motor function. This loss of function is generally permanent due to the limited regenerative capacity of the CNS [2]. Cells used for *in vitro* modeling should have morphological, functional qualities, and relevant requirements depending on disease, injury, and ultimate goal of the proposed experiments. The commonly
**i.** Primary isolated cells, neurons, oligodendrocytes, astrocytes, or microglia cells [3, 4].
Primary cells are obtained from the dissociated neural tissues, spinal cord, or brain and are usually cultivated immediately after the isolation as "primo‐cultures." The origin of these cells could be embryonal, fetal, neonatal, or adult [5, 6]. It is well known that younger cells are more potent according to their strong proliferative activity, in contrary to terminally differentiated poor proliferating capacity owning adult cells [6]. By using the primary cultures, we can investigate the behavior and intercommu‐ nication of the individual or mixed cell populations. These populations preserve the main characteristics similar to the *in vivo* examined types of cells [7]. Very often used are also the immortalized cell lines, which are in fact the primary cells with genetical
life they lived before the injury [1].
38 Recovery of Motor Function Following Spinal Cord Injury
**2. Cell cultures**
used *in vitro* cell culture models of SCI include:
modification.
The characteristic manifestation of the secondary reactive phase is scar formation at the initial impact site, which represents the major obstruction to CNS regeneration. The glial scar is rich in different cells (microglia and reactive astrocytes), myelin‐associated inhibitors, and it physically surrounds the damaged tissue. To study the components of the glial scar, many models of co‐cultivation of neurons with others glial scar situated cells have been developed [8, 9, 11]. By combining two pivotal aspects of CNS damage, mechanical injury and co‐ cultivation with meningeal cells in *in vitro* scar formation model, it has been shown that astrocytes increase the expression of GFAP (glial fibrillary acidic protein), as well as the scar‐ associated markers, e.g., phosphacan, neurocan, and tenascin. Subsequent short‐time co‐ cultivation with different developmental neuronal populations led to significantly reduced neurite growth in this scar‐like model [9]. After spinal cord injury, demyelination of spared white matter significantly hinders spontaneous function recovery [12]. To understand the processes related to this phenomenon, a wide range of developed cultures in which the many stages of myelination can be followed over time are used [13–15]. These myelinating cultures involve plating dissociated rat spinal cord cells onto a confluent monolayer of neurosphere‐ derived astrocytes. Sorensen et al. [16] compared the different types of cells in the process of myelination promotion of CNS axons. The mixed cell population of dissociated embryonal spinal cord was cultivated on olfactory ensheathing cells (OECs), Schwann cells (SchCs), or neurosphere‐derived astrocytes to study the mutual interactions in context to affect the myelination. Myelin internodes and nodes of Ranvier were often found only in the cells cultivated onto neurosphere‐derived astrocytes. This model is very useful to investigate the CNS axonal myelination because it recapitulates the processes occurring in CNS *in vivo*. On the basis of these cultures, it has been shown that astrocytes have a direct role in promoting myelination either by releasing a soluble factor or by cell–cell contact [16]. Also, the phenotype of astrocytes has a crucial role in determining their effects on myelination. Myelination is poor when the myelinating cultures are plated on quiescent astrocytes but enhanced when they are plated on a more reactive/activated phenotype. Interestingly, quiescent astrocytes secreted CXCL10, which did not appear to directly affect the purified oligodendrocyte progenitor cell differentiation but had the effect on the ability of oligodendrocytes to ensheath axons [13]. Astrocytes exhibit dynamic cell process movement and changes in their membrane topogra‐ phy as they interact with axons and oligodendrocytes during the process of myelination [15]. These observations may have important implications with respect to the development of therapeutic strategies to promote CNS remyelination in demyelinating diseases. Also other models have been developed to investigate and to understand the molecular mechanisms of CNS axon regeneration. The process of axon regeneration can be monitored by neurite outgrowth assays, which often represent the phenotypic expression of regeneration progress. In neurite outgrowth assays, the neurons are cultivated on an adherent substrate, and the neurite growth is evaluated in specific medium conditions for extended periods of time, and then the overall length of the radial outgrowth is measured. Primary neurons (neurons derived from dorsal roots ganglion, cerebellum, cortex, or hippocampus) are considered more biologically relevant to neurons *in vivo* instead of induced pluripotent stem cells or any special cell lines [17].
### **2.2** *In vitro*
The majority of SCI occurring in human is due to the compression of spinal cord; therefore most *in vitro* models have been designed to mimic this type of injury. To simulate the spinal cord injury *in vitro*, the confluent layer of astrocytes and dissociated embryonic spinal cord cells plated onto them are cut with scalpel blade, the so‐called scratch model. Such treatment creates an empty area without any cells—lesion. With an increasing time of cultivation, the situation around the injury becomes comparable to the SCI *in vivo*. The microglial cells and astrocytes start to migrate to the injury site. Conversely, the decrease in neurite density and the process of demyelination start as the consequences of disrupting confluent cell layer. The main purpose of this model is to find out, to recognize any factors, molecules, biomaterials, even cells that can contribute, may be, to accelerate the potential regeneration of remaining cells destroyed by injury. The closure of this lesion could be modified using different Rho/ ROCK inhibitors [11] or by cAMP modulators, which promote neurite outgrowth and myelination [5]. *In vitro* models are the preferred methods for the study of acute or sub‐acute pathophysiology after a trauma stimulus enabling the precise control of the extracellular environment, easy and repeatable access to the cells.
To imitate the SCI in our experimental model, the P1 rat spinal cord cells were mechanically scratched across the cell monolayer with a pipette tip. The result of this handling was the linear cell‐free area/lesion which varied between 400 - 600 μm in width. After washing out the scraped cells with PBS (phosphate-buffered saline), cells were maintained for up to 10 days in culture. During the cultivation time, the cultures remained viable. Cell migration or wound closure of the scratched surface was observed and photographed at daily intervals. To investigate the potential stimulative effect of mesenchymal stem cells (MSCs) on the neural cells, we treated such injured cultures with conditioned medium obtained from rat bone marrow MSCs. In fact, MSCs cannot differentiate into fully functional neurons, or into others neural cells; the supporting effect of neuroregeneration and neuroprotection is mediated by specific neurotro‐ phic factors and cytokines produced by MSCs [18]. All experiments conformed to the Slovak Law for Animal Protection No. 377/2012 and 436/2012 transposed from the European Com‐ munities Council Directive (2010/63/EU) and were approved by the Institutional Ethical Committee for animal research. Our preliminary data showed that the cell‐free area in injured group treated with MSC conditioned medium became filled with reactive astrocytes, microglia cells, and oligodendrocyte in a shorter time when compared to untreated controls (**Figure 1**).
**Figure 1.** Fluorescent microscopy images of spinal cord culture injured by mechanical insult, control group (A) and group treated (B) with conditioned medium obtained from MSCs. The images were taken on day 2 after injury. Cells were stained with GFAP/astrocytes/green antibody, Iba/microglia cells/red antibody and DAPI/nuclear/blue marker, and (C) the detail of microglia cells infiltrating the lesion on day 2. Scale bar is 100 μm.
The cell culture approaches lack the complexity and physiological relevance of *in vivo* system. However, animal studies offer complexity, which cannot be modeled *in vitro* and which is also difficult to control and manipulate and therefore results in data that must be extrapolated to human systems [10].
### **3. Organotypic slices**
neurite growth is evaluated in specific medium conditions for extended periods of time, and then the overall length of the radial outgrowth is measured. Primary neurons (neurons derived from dorsal roots ganglion, cerebellum, cortex, or hippocampus) are considered more biologically relevant to neurons *in vivo* instead of induced pluripotent stem cells or any special
The majority of SCI occurring in human is due to the compression of spinal cord; therefore most *in vitro* models have been designed to mimic this type of injury. To simulate the spinal cord injury *in vitro*, the confluent layer of astrocytes and dissociated embryonic spinal cord cells plated onto them are cut with scalpel blade, the so‐called scratch model. Such treatment creates an empty area without any cells—lesion. With an increasing time of cultivation, the situation around the injury becomes comparable to the SCI *in vivo*. The microglial cells and astrocytes start to migrate to the injury site. Conversely, the decrease in neurite density and the process of demyelination start as the consequences of disrupting confluent cell layer. The main purpose of this model is to find out, to recognize any factors, molecules, biomaterials, even cells that can contribute, may be, to accelerate the potential regeneration of remaining cells destroyed by injury. The closure of this lesion could be modified using different Rho/ ROCK inhibitors [11] or by cAMP modulators, which promote neurite outgrowth and myelination [5]. *In vitro* models are the preferred methods for the study of acute or sub‐acute pathophysiology after a trauma stimulus enabling the precise control of the extracellular
To imitate the SCI in our experimental model, the P1 rat spinal cord cells were mechanically scratched across the cell monolayer with a pipette tip. The result of this handling was the linear cell‐free area/lesion which varied between 400 - 600 μm in width. After washing out the scraped cells with PBS (phosphate-buffered saline), cells were maintained for up to 10 days in culture. During the cultivation time, the cultures remained viable. Cell migration or wound closure of the scratched surface was observed and photographed at daily intervals. To investigate the potential stimulative effect of mesenchymal stem cells (MSCs) on the neural cells, we treated such injured cultures with conditioned medium obtained from rat bone marrow MSCs. In fact, MSCs cannot differentiate into fully functional neurons, or into others neural cells; the supporting effect of neuroregeneration and neuroprotection is mediated by specific neurotro‐ phic factors and cytokines produced by MSCs [18]. All experiments conformed to the Slovak Law for Animal Protection No. 377/2012 and 436/2012 transposed from the European Com‐ munities Council Directive (2010/63/EU) and were approved by the Institutional Ethical Committee for animal research. Our preliminary data showed that the cell‐free area in injured group treated with MSC conditioned medium became filled with reactive astrocytes, microglia cells, and oligodendrocyte in a shorter time when compared to untreated controls (**Figure 1**).
environment, easy and repeatable access to the cells.
cell lines [17].
40 Recovery of Motor Function Following Spinal Cord Injury
**2.2** *In vitro*
Organotypic cultures are whole slices of tissue cultured without dissociation of the individual cells. Organotypic slices (OTSs) preserve the basic structural and connective organization of their original tissue (organotypic) and represent an interim system sharing the properties of the cell culture *in vitro* and an animal *in vivo* model. The OTS culture allows long‐term maintenance of tissue architecture in "dish" [19].
OTS preparation is carried out according to the following steps. Briefly, after tissue removal under sterile conditions, the tissue is cut into 150–400 μm thick sections using a tissue chopper or vibratome. After washing, the slices are attached to a substrate and cultivated under appropriate conditions for several weeks. Today, we know many ways of OTS preparation depending on the final thickness of the slices and the survival time in culture. In *roller‐tube* *cultures*, the slices embedded in plasma clot or collagen matrix on glass coverslips are placed in plastic culture tube containing a small amount of medium and undergo continuous slow rotation which allows the oxygenation of these slices by regular changing of the liquid–gas interface [20]. Because the slices become flattened to a quasi monolayer during cultivation procedure, this technique is appropriate when individual living nerve cells have to be observed by normal optical conditions for several weeks [21]. For obtaining quantitatively more pieces of OTSs, the modified roller method for organotypic cultivation of free‐floating sections of postnatal rats can be used. This innovative procedure enables the simultaneous cultivation of multiple amount of OTS in one bottle (up to 50 and even more). For this reason, this method is advantageous for the investigation of cytotoxic injury of neural cells and appropriate immunocytochemical and molecular analysis [22]. Another, but not so frequently used, is the method, where the slices are placed either directly on collagen coated or embedded in collagen gels in Petri dishes and are totally covered by medium. These cultures survive only a few days and are mainly used for electrophysiology [21].
The most used and popular OTS‐cultivating methods are the membrane cultures, where the slices are directly placed at the air–medium interface on semiporous membranes and kept stationary during the entire cultivation. Slices can collect the nutrients from an appropriate growth medium from below via capillary process and oxygen from above (**Figure 2**) [21, 23].
**Figure 2.** Schematic illustration of the most used OTS cultivating methods called "interface method". Organotypic spi‐ nal cord slice (OTS‐SC) is placed on semiporous membrane through which OTS can obtain the nutrition from medium and oxygen from above.
At the beginning, experiments were carried out using hippocampal slices. However, cerebellar, cortical, and spinal cord explants have also been tested with positive results [23]. The advant‐ age of this technique is that these OTSs are never covered with embedding materials or media which allow the study of the effect of a therapeutic agent added directly onto the OTSs during all stages of culturing longer time while maintaining semi‐three‐dimensional structure of tissue. Positioning of slices on the membrane is extremely easy and may represent an advant‐ age also for co‐culture experiments. This technique also has some limitations. The problem may concern the accessibility of individual neurons, because the cultured slices remain a few cell layers thick and are covered by fibers and glial processes. During OTS preparation and cultivation, we should avoid some troubles. Slices should be placed in the middle of the membrane without contacting one another. In this position, every slice can be easily reached to perform the impact injury. If any impact injury is performed, the uninjured and damaged sections have to be situated on individual insert; otherwise, excytotoxins released from damaged sections may induce injury in the controls. Sections with defective or lacking layers should be rejected from the examination. It is important to start the concrete experiments about 12–14 days after the beginning of OTS cultivation [24]. In some studies, the time of OTS culture before starting the actual experiment is shorter (1 week) [25, 26]; so the mechanical damage caused by slicing disappears, and the cells can reach a more advanced developmental stage. In addition, this time period will allow the slices to acclimatize to the culturing conditions and mature. Organotypic cultures survive from weeks to months, providing options for long‐term studies, such as those studying different processes of neurodegeneration and recovery following excytotoxicity ischemia and traumatic injury.
Since the first description of organotypic hippocampal slice cultures based on the method of Gahwiler [20], the hippocampal regions are likely the most widely used organotypic slice culture model for the study of neural tissues [27]. Together with the accumulation of practical experiences and increased necessity to answer the questions dealing with nerve tissue regeneration after spinal cord injury (SCI), the development and entrenchment of organotypic spinal cord slice (OTS‐SC) models became more actual. Today, there is a wide choice of methods and OTS models of both spinal cord and brain derived from different animals mouse, rats, and rabbits [28], even the human too [29], which have been used to describe ways of monitoring spontaneous or induced neuronal degeneration in organotypic brain slice cultures [30], to study the different processes of neurodegeneration after tissue traumatic injury, and to investigate mechanisms and treatment strategies for the neurodegenerative disorders like stroke (ischemia) [31–33]. These models can simulate several pathological aspects of various neurological conditions depending on different factors, e.g., (i) the applied stimulus which causes damage, (ii) the age of OTS donor animal (iii) which parameter is the experiment focused on.
### **4. Insults**
*cultures*, the slices embedded in plasma clot or collagen matrix on glass coverslips are placed in plastic culture tube containing a small amount of medium and undergo continuous slow rotation which allows the oxygenation of these slices by regular changing of the liquid–gas interface [20]. Because the slices become flattened to a quasi monolayer during cultivation procedure, this technique is appropriate when individual living nerve cells have to be observed by normal optical conditions for several weeks [21]. For obtaining quantitatively more pieces of OTSs, the modified roller method for organotypic cultivation of free‐floating sections of postnatal rats can be used. This innovative procedure enables the simultaneous cultivation of multiple amount of OTS in one bottle (up to 50 and even more). For this reason, this method is advantageous for the investigation of cytotoxic injury of neural cells and appropriate immunocytochemical and molecular analysis [22]. Another, but not so frequently used, is the method, where the slices are placed either directly on collagen coated or embedded in collagen gels in Petri dishes and are totally covered by medium. These cultures survive only a few days
The most used and popular OTS‐cultivating methods are the membrane cultures, where the slices are directly placed at the air–medium interface on semiporous membranes and kept stationary during the entire cultivation. Slices can collect the nutrients from an appropriate growth medium from below via capillary process and oxygen from above (**Figure 2**) [21, 23].
**Figure 2.** Schematic illustration of the most used OTS cultivating methods called "interface method". Organotypic spi‐ nal cord slice (OTS‐SC) is placed on semiporous membrane through which OTS can obtain the nutrition from medium
At the beginning, experiments were carried out using hippocampal slices. However, cerebellar, cortical, and spinal cord explants have also been tested with positive results [23]. The advant‐ age of this technique is that these OTSs are never covered with embedding materials or media which allow the study of the effect of a therapeutic agent added directly onto the OTSs during all stages of culturing longer time while maintaining semi‐three‐dimensional structure of tissue. Positioning of slices on the membrane is extremely easy and may represent an advant‐ age also for co‐culture experiments. This technique also has some limitations. The problem may concern the accessibility of individual neurons, because the cultured slices remain a few cell layers thick and are covered by fibers and glial processes. During OTS preparation and cultivation, we should avoid some troubles. Slices should be placed in the middle of the membrane without contacting one another. In this position, every slice can be easily reached
and are mainly used for electrophysiology [21].
42 Recovery of Motor Function Following Spinal Cord Injury
and oxygen from above.
### **4.1. Ischemic insults**
Damage of nerve tissue may be caused by a number of insults. Previously, organotypic slice preparations have been used in several injury models including ischemia (experimental models with oxygen‐glucose deprivation) and cytotoxicity studies (exposure to glutamate receptor agonists) [24]. Ischemic insults can be generated by oxygen‐glucose deprivation of the cultures by placing them for 1h in glucose‐free medium saturated with 95% N2, 5% CO<sup>2</sup> using standard interface method [34]. To induce stroke‐like, energy failure conditions in slice cultures can be achieved in several other ways, like use of anoxic chambers or OTS submersion in glucose‐free medium saturated with nitrogen [35]. In a study of free radical‐induced damage in CNS using OTS exposed to H2O2 as a model of oxidative injury in the brain, the results provide evidence that glial cells in cultured slices are vulnerable to H2O2toxicity as pyramidal neurons, and that H2O2‐mediated cell death was significantly alleviated by antioxidants and independent of calcium influx in both glial and neuronal populations [36].
The hypoxic environment in SCI has been shown to inhibit recovery. One of the promising treatments for hypoxic ischemia sustained during SCI is gene therapy. Vascular endothelial growth factor (VEGF) increasingly has gained attention as a potential factor in gene therapy to replace or improve damaged neurons in SCI because of its role in neurogenesis and neuroprotection, astrocyte proliferation, and neurite outgrowth. The controlled release of VEGF in hypoxia‐inducible VEGF gene delivery system in the OTS‐SC may prove to be useful for providing neuroprotection and in stimulating the neural growth in the hypoxic environ‐ ment of the injured spinal cord [37]. The principal feature of acute SCI pathophysiology is excitotoxicity that is considered the main contributor to the clinical outcome. An established *in vitro* SCI model using excitotoxic damage evoked by transient kainate application to organotypic slice cultures (with predominant neuro‐ rather than glio‐toxicity) closely mimics the *in vivo* pathophysiology of SCI. The results from such model indicate that, after an excitotoxic stimulus, extracellular S100b, a useful biomarker in serum or cerebrospinal fluid for brain or spinal cord injury, was significantly elevated in association with ongoing neuronal damage and unscathed glia [38]. The application of glutamate is believed to mimic excitotox‐ icity that occurs as a consequence of ischemic stroke. Also, glutamate‐induced excitotoxicity is a major contributor to motor neuron (MN) degeneration in disorders such as amyotrophic lateral sclerosis, stroke, and spinal cord injury. On the basis of using OTS from different parts of SC (i.e., cervical, thoracic, and lumbar), it is found out that the MN susceptibility to glutamate‐induced excitotoxicity differs within diverse populations of MNs. This sensitivity is in correlation to the segmental variable expression of glutamate receptor sub‐types in MN populations. In the cervical and lumbar parts, the AMPA receptor is in dominance, whereas in thoracic segments, the NMDA receptors predominate. These results should be taken into account by devising a new drug screening experiment with cell culture system [39].
### **4.2. Mechanical insults**
Adamchik et al. [24] described two original models of moderate mechanical trauma containing the induced primary and secondary traumatic damage in organotypic slice cultures. The primary trauma injury was achieved by quickly rolling a stainless steel cylinder (weight, 0.9 g; diameter, 5 mm; length, 7 mm) twice on the organotypic slice. To model the secondary trauma injury, a weight of 0.137 g was dropped from a height of 2 mm on a localized area of the organotypic slice. The primary damage delivered by rotating the steel cylinder on sections induces injury following the head accidents, while the damage caused in this model for secondary injury, limited initially to a small area of the section, enables to monitor the expansion of dying cells.
Almost all SCI occurring in human are due to the compression of spinal cord; therefore, most *in vitro* models have been designed to mimic this type of injury. For that reason, an *in vitro* model of neurotrauma was developed and assessed by using organotypic slice culture of adult mice spinal cord to facilitate the investigation of primary and secondary mechanisms of cell death that occur after mechanical SCI [25]. The mechanical trauma was achieved by using a weight drop model of injury, where an impactor (pin) with a weight of 0.2g was dropped from 1.7 cm onto the center of the culture slice. They modified previously described methods for generating murine spinal trauma in OTS [40]. In contrary to other models, where the OTS‐SC from embryonic or neonatal mice was used, the modified model used the OTS from adult animals according to the different regenerative capabilities.
Organotypic slice cultures of spinal cord have been used in neuroscience research for a long time, but classically, these cultures were cut in the transverse plane. The transverse slices could be, however, obtained from only a single part of the spinal cord. In order to reflect the spinal cord microenvironment, the longitudinal cultures are required to maintain the basic cytoarch‐ itectonic organization of the intact spinal cord. In contrary to the previous model, Bonnici and Kapfhammer [41] used the spinal cord slices for their study that were cut not in the transverse but in the sagittal longitudinal plane such that several spinal cord segments were included in the slide culture. This OTS‐SC model is well suited for the study of axonal growth because the typical ventro‐dorsal polarity of the SC is maintained after a culture period of 2 weeks, and intrinsic spinal cord axons formed a strong fiber tract extending along the longitudinal axis of the slice. The axons became myelinated during the culture period, and synaptic contacts were present in these cultures. After mechanical lesions originated by completely transecting the SC tissue with a scalpel blade, the remaining spinal cord neurons had a considerable capability to regenerate the axons. The amount of renewed axons infiltrating the wound site decreased with increasing cultivation time and maturation of the culture. This indicates that the cellular differentiation and cytoarchitectonic organization of the spinal cord neurons are similar to that of slice cultures of the transverse plane and reflects the *in vivo* situation well [41]. Some scientists use the OTS for the study of radiation effect on cell damage. By determining the time‐ dependent course of choline uptake in mature organotypic slice cultures of rabbit and assessing the effects of continuous and single high‐dose irradiation on choline uptake, the results demonstrate that the delivery of continuous but relatively low‐dose rate gamma irradiation is more efficacious than single high‐dose external irradiation on high‐affinity choline uptake in nervous tissues [28].
### **5. Age of donor animal**
neurons, and that H2O2‐mediated cell death was significantly alleviated by antioxidants and
The hypoxic environment in SCI has been shown to inhibit recovery. One of the promising treatments for hypoxic ischemia sustained during SCI is gene therapy. Vascular endothelial growth factor (VEGF) increasingly has gained attention as a potential factor in gene therapy to replace or improve damaged neurons in SCI because of its role in neurogenesis and neuroprotection, astrocyte proliferation, and neurite outgrowth. The controlled release of VEGF in hypoxia‐inducible VEGF gene delivery system in the OTS‐SC may prove to be useful for providing neuroprotection and in stimulating the neural growth in the hypoxic environ‐ ment of the injured spinal cord [37]. The principal feature of acute SCI pathophysiology is excitotoxicity that is considered the main contributor to the clinical outcome. An established *in vitro* SCI model using excitotoxic damage evoked by transient kainate application to organotypic slice cultures (with predominant neuro‐ rather than glio‐toxicity) closely mimics the *in vivo* pathophysiology of SCI. The results from such model indicate that, after an excitotoxic stimulus, extracellular S100b, a useful biomarker in serum or cerebrospinal fluid for brain or spinal cord injury, was significantly elevated in association with ongoing neuronal damage and unscathed glia [38]. The application of glutamate is believed to mimic excitotox‐ icity that occurs as a consequence of ischemic stroke. Also, glutamate‐induced excitotoxicity is a major contributor to motor neuron (MN) degeneration in disorders such as amyotrophic lateral sclerosis, stroke, and spinal cord injury. On the basis of using OTS from different parts of SC (i.e., cervical, thoracic, and lumbar), it is found out that the MN susceptibility to glutamate‐induced excitotoxicity differs within diverse populations of MNs. This sensitivity is in correlation to the segmental variable expression of glutamate receptor sub‐types in MN populations. In the cervical and lumbar parts, the AMPA receptor is in dominance, whereas in thoracic segments, the NMDA receptors predominate. These results should be taken into
account by devising a new drug screening experiment with cell culture system [39].
Adamchik et al. [24] described two original models of moderate mechanical trauma containing the induced primary and secondary traumatic damage in organotypic slice cultures. The primary trauma injury was achieved by quickly rolling a stainless steel cylinder (weight, 0.9 g; diameter, 5 mm; length, 7 mm) twice on the organotypic slice. To model the secondary trauma injury, a weight of 0.137 g was dropped from a height of 2 mm on a localized area of the organotypic slice. The primary damage delivered by rotating the steel cylinder on sections induces injury following the head accidents, while the damage caused in this model for secondary injury, limited initially to a small area of the section, enables to monitor the
Almost all SCI occurring in human are due to the compression of spinal cord; therefore, most *in vitro* models have been designed to mimic this type of injury. For that reason, an *in vitro* model of neurotrauma was developed and assessed by using organotypic slice culture of adult mice spinal cord to facilitate the investigation of primary and secondary mechanisms of cell death that occur after mechanical SCI [25]. The mechanical trauma was achieved by using a
**4.2. Mechanical insults**
expansion of dying cells.
independent of calcium influx in both glial and neuronal populations [36].
44 Recovery of Motor Function Following Spinal Cord Injury
The age of the donor animal influences the viability of OTS, the degree of tissue organization, and the successfulness of experiment at all. Early postnatal periods (P0–P9) are ideally suited for culturing [26, 39, 42]. Since most diseases of the spinal cord concern the adults, some organotypic spinal cord slice culture models from grown rats were established [25, 26]. But the isolation of the cord from adult individuals is, however, much more problematic. The fully myelinated spinal cord is very sensitive to any damage, and cutting it with tissue chopper causes injury to about 60% of slices. The slice culture survives about 7 days maintaining proper parallel fiber tract architecture. Also, the constant high vitality and tissue organization are preserved up to 4–5 weeks in neonatal slices, while the vitality of adult slice cultures decreases significantly upon the first 5 days of cultivation [26]. On the contrary, spinal cord organotypic culture from adult mouse can be used to produce neurospheres that can be further differen‐ tiated into astrocytes and neuronal cells of different phenotypes. Although, it is not clear what exactly triggers the formation of neurospheres, one explanation could be, that neuronal loss and/or long‐term culturing of spinal cord slices resets the internal cell program and stimulates proliferation and further specification of neuronal stem cells (NSCs) [43]. Most researchers did not use cultures from animals older then P6, because neuronal survival in such cultures was more variable and a longitudinal fiber tract did not develop consistently [41]. The same goes for the transverse spinal cord slice cultures, in which neuronal survival also appears to be declining in cultures derived from older postnatal animals [44].
### **6. Utilization of organotypic slices**
Because the OTS system is accessible to experimental manipulations, OTSs represent a suitable model which may be used for different research projects not only for investigation of various parameters accompanying the CNS injury but also for the development of therapeutic application for such injury. Spinal cord injury leads to the death of various cell types, including neurons, oligodendrocytes, astrocytes, and precursor cells. A high percentage of astrocytes and oligodendrocytes in the white matter at the site of injury die within a short time after injury leading to axonal demyelination. Grafts of cells used as a treatment or as the source of supporting factors can be directly transplanted in the slices and in this way, the graft‐‐host interaction, behavior, and fate of grafted cells can be analyzed. The injured spinal cord microenvironment is very inhospitable for survival and differentiation of transplanted or endogenous NSCs. To mimic NSC grafting into the spinal cord, NSCs were seeded on top of cultured spinal cord slices with the Hamilton syringe between 7 and 10 days after the initial culture. In such culture system, possible effects of environmental manipulation on biological behavior of either exogenous grafted NSCs or endogenous neural progenitor cells in a complex tissue environment similar to that of *in vivo* system were studied. The majority of focally applied exogenous NSCs survived up to 4 weeks, and dispersing cells did not show any preference to either the white or the gray mater. Cultured spinal cord slices retained the non‐ neurogenic characteristics of *in vivo* spinal cord tissue since they did not support neuronal differentiation of either exogenous NSCs or endogenous neural progenitors [42]. Also, it is very much preferred to use the OTS derived from animals with some genetic alteration, which is manifested in special disease features. For example, to study the process of the myelination and myelin support, the OTSs obtained from the cerebellum of shiverer mice were used. These mice are genetically modified in myelin basic protein (MBP) gene resulting in wide CNS dysmyelination. After the transplantation of oligodendrocyte precursors into such modified sections, the direct observation of myelin produced by grafted cells is possible. This system is more advantageous in comparison with neuron‐oligodendrocyte cell co‐cultures for its propinquity with *in vivo* system [45].
The impact of individual cells on the regeneration of damaged CNS tissue can be studied not only by direct transplantation of cells to OTS but also by the so‐called co‐culture system, where the exogenous populations are cultivated underneath the slice co‐cultures without direct contact using a Transwell system. The cultured OTSs are then influenced by factors released from exogenous, co‐cultured cells. Indirect co‐cultivation of exogenous cells along with OTS‐ SC permits the analysis of the factors secreted into the medium by both cells and the spinal cord tissue and the determination of their mutual influence. For example, MSCs, the promising candidates for neuroregenerative cell therapies, promote neuronal regeneration by signifi‐ cantly enhanced fiber outgrowth, an effect that may be mediated by a higher expression of brain‐derived neurotrophic factor (BDNF) and basic fibroblast growth factor [46]. So, the role of the MSCs is in establishing the microenvironment to be more favorable for tissue repair.
tiated into astrocytes and neuronal cells of different phenotypes. Although, it is not clear what exactly triggers the formation of neurospheres, one explanation could be, that neuronal loss and/or long‐term culturing of spinal cord slices resets the internal cell program and stimulates proliferation and further specification of neuronal stem cells (NSCs) [43]. Most researchers did not use cultures from animals older then P6, because neuronal survival in such cultures was more variable and a longitudinal fiber tract did not develop consistently [41]. The same goes for the transverse spinal cord slice cultures, in which neuronal survival also appears to be
Because the OTS system is accessible to experimental manipulations, OTSs represent a suitable model which may be used for different research projects not only for investigation of various parameters accompanying the CNS injury but also for the development of therapeutic application for such injury. Spinal cord injury leads to the death of various cell types, including neurons, oligodendrocytes, astrocytes, and precursor cells. A high percentage of astrocytes and oligodendrocytes in the white matter at the site of injury die within a short time after injury leading to axonal demyelination. Grafts of cells used as a treatment or as the source of supporting factors can be directly transplanted in the slices and in this way, the graft‐‐host interaction, behavior, and fate of grafted cells can be analyzed. The injured spinal cord microenvironment is very inhospitable for survival and differentiation of transplanted or endogenous NSCs. To mimic NSC grafting into the spinal cord, NSCs were seeded on top of cultured spinal cord slices with the Hamilton syringe between 7 and 10 days after the initial culture. In such culture system, possible effects of environmental manipulation on biological behavior of either exogenous grafted NSCs or endogenous neural progenitor cells in a complex tissue environment similar to that of *in vivo* system were studied. The majority of focally applied exogenous NSCs survived up to 4 weeks, and dispersing cells did not show any preference to either the white or the gray mater. Cultured spinal cord slices retained the non‐ neurogenic characteristics of *in vivo* spinal cord tissue since they did not support neuronal differentiation of either exogenous NSCs or endogenous neural progenitors [42]. Also, it is very much preferred to use the OTS derived from animals with some genetic alteration, which is manifested in special disease features. For example, to study the process of the myelination and myelin support, the OTSs obtained from the cerebellum of shiverer mice were used. These mice are genetically modified in myelin basic protein (MBP) gene resulting in wide CNS dysmyelination. After the transplantation of oligodendrocyte precursors into such modified sections, the direct observation of myelin produced by grafted cells is possible. This system is more advantageous in comparison with neuron‐oligodendrocyte cell co‐cultures for its
The impact of individual cells on the regeneration of damaged CNS tissue can be studied not only by direct transplantation of cells to OTS but also by the so‐called co‐culture system, where the exogenous populations are cultivated underneath the slice co‐cultures without direct contact using a Transwell system. The cultured OTSs are then influenced by factors released
declining in cultures derived from older postnatal animals [44].
**6. Utilization of organotypic slices**
46 Recovery of Motor Function Following Spinal Cord Injury
propinquity with *in vivo* system [45].
Cho et al. [47] demonstrated that the treatment of lumbar slices of spinal cord with lysolecithin caused an important degree of cell death and demyelination of nerve fibers. However, after the co‐cultivation of lumbar sections with human MSCs (hMSCs), the earlier observed effects were elevated to a significant range. Since the spinal cord slice culture is totally isolated and does not have any circulating immune cells originated from the host, it excludes the possibility of any systemic immune response to transplanted human cells. The results demonstrated that direct hMSC transplantation into the demyelinated organotypic spinal cord slice culture alter the tissue microenvironment and clearly increased the survival of endogenous cells and promoted axonal outgrowth of nerves in this organotypic spinal cord slice culture. So, the OTS‐ SC cultures provide the opportunity to investigate the underlying mechanisms of nerve regeneration after injury in the adult nervous system and to identify the factors which enhance axonal outgrowth as well as cell survival. The ease by which organotypic slice cultures can be manipulated and inspected also make them interesting tools for studies of cell proliferation, chronic CNS astrogliosis [48] and migration [49].
The OTS could be applied also to study the neurons and glial damage during bacterial infection. In the course of pneumococcal meningitis, both spontaneous and antibiotic‐induced lysis of bacteria occurs. During this process, not only cell wall constituents but also nucleic acids of the bacteria are liberated. Therefore, organotypic hippocampal slices were utilized for the investigation of possible toxic effects of complete bacteria causing meningitis or their components on neuronal and glial tissue in the absence of leukocytes. The neuronal tissue of an infected host is damaged as a result of the intense inflammatory response against the attacking bacteria. Though OTS closely mimics the *in vivo* system, during the actual cultivation, it represents only the part of more complex unit derived from, which is closed without any circulating immune cells. According to this, by the measurement of neuron‐specific enolase in cultivation medium, as an additional indicator of neuronal damage, it was found out that not only heat‐inactivated whole bacteria but also diverse bacterial elements were capable of inducing nervous tissue damage, such as bacterial walls and nucleic acids [50]. Also, the organotypic brain slice cultures of bovine origin are a suitable model to study aspects of host– pathogen interaction in listeric encephalitis and potentially in other neuroinfectious diseases [51]. To evaluate the effect of HIV‐1 virus on neural cells, a method for culturing human fetal organotypic brain slices in the presence of live virus was developed. The brain‐‐derived OTSs were maintained on membranes, and co‐cultured with HIV‐1–infected T‐cells on the well bottom. Thus, OTSs were exposed to live virus during whole cultivation time, HIV‐1 proteins, and other molecules produced by the infected T‐cells. The results obtained from this model indicate the gliosis observed *in vivo.* This valuable method can be used to model the dynamics and the microenvironment of brain tissue exposed to HIV‐1 and can eventually be used to test therapies directed at preventing HIV‐1‐induced neural damage in CNS [52].
Within the wide range of available OTSs derived from different CNS sections, the OTS‐SC system represents an attractive model with some benefits, when compared to generally used *in vitro* and *in vivo* models. OTS‐SC can be kept under cultivation conditions for a long time. During the whole cultivation time, this system enables quick and direct evaluation and identification of huge diapason of different cellular, molecular relations involved in regener‐ ation processes. Moreover, OTS‐SCs can be utilized to improve the *ex vivo* neurological disorders and thus represent an intermediate step between *in vitro* experiments and animal models [53].
### **7. Experimental study**
The aim of our study was: (i) to imitate the spinal cord injury (SCI) in organotypic spinal cord slice (OTS‐SC) and (ii) to treat these injured OTS‐SCs (OTS‐SCIs) in the presence of rat mesenchymal stem cells (rMSCs). We imitate the SCI in the OTS‐SC not by direct mechanical impact, but the injury was induced by OTS‐SC treated with conditioned medium obtained from injured spinal cord (CM‐SCI). CM‐SCI represents a cocktail of cytokines and chemokines released by central lesion segments of injured spinal cord. We examined the impact of CM‐ SCI, as well as rMSCs on behavior of microglia cells, astrocytes, and neurons in the organotypic spinal cord slice cultures.
### **7.1. Methods**
The SCI was induced by modified balloon‐compression technique in adult rats [54], and after 3 days of animal survival, the central lesion segments were dissected and cultured for 24 h *in vitro* to obtain CM‐SCI used for OTS‐SC treatment. Similarly, we obtained the conditioned medium from control‐intact thoracic spinal cord (CM‐SC), which in our experiment served as a source of control‐conditioned medium. The OTS‐SCs were prepared from lumbal SCs of naive adult male rats. The isolated SCs were transversely cut with a Mcllwain tissue chopper into 400 μm OTS‐SC sections and cultivated for an adaptation for 5 days in 1ml/well of culture media containing DMEM, 10% fetal bovine serum, and 1% penicillin–streptomycin and incubated at 37°C in humidified atmosphere with 5% of CO2 using standard interface method. The study was performed with the approval and guidelines of the Institutional Animal Care and Use Committee of the Slovak Academy of Sciences and with the European Communities Council Directive (2010/63/EU) regarding the use of animals in Research, Slovak Law for Animal Protection No. 377/2012 and 436/2012.
### **7.2.** *In vitro* **model of spinal cord injury and treatment**
After adaptation, OTS‐SCs were divided into four groups according to the treatment and cultivated for the next 3 days. (i) **OTS‐SC** (OTS treated with CM‐SC), (ii) **OTS‐SCI** (OTS treated with CM‐SCI) (iii) **OTS‐SC + MSCs** (OTS co‐cultured with rMSCs in CM‐SC), and (iv) **OTS‐ SCI + MSCs** (OTS co‐cultured with rMSCs in CM‐SCI) (**Figure 3**).
and the microenvironment of brain tissue exposed to HIV‐1 and can eventually be used to test
Within the wide range of available OTSs derived from different CNS sections, the OTS‐SC system represents an attractive model with some benefits, when compared to generally used *in vitro* and *in vivo* models. OTS‐SC can be kept under cultivation conditions for a long time. During the whole cultivation time, this system enables quick and direct evaluation and identification of huge diapason of different cellular, molecular relations involved in regener‐ ation processes. Moreover, OTS‐SCs can be utilized to improve the *ex vivo* neurological disorders and thus represent an intermediate step between *in vitro* experiments and animal
The aim of our study was: (i) to imitate the spinal cord injury (SCI) in organotypic spinal cord slice (OTS‐SC) and (ii) to treat these injured OTS‐SCs (OTS‐SCIs) in the presence of rat mesenchymal stem cells (rMSCs). We imitate the SCI in the OTS‐SC not by direct mechanical impact, but the injury was induced by OTS‐SC treated with conditioned medium obtained from injured spinal cord (CM‐SCI). CM‐SCI represents a cocktail of cytokines and chemokines released by central lesion segments of injured spinal cord. We examined the impact of CM‐ SCI, as well as rMSCs on behavior of microglia cells, astrocytes, and neurons in the organotypic
The SCI was induced by modified balloon‐compression technique in adult rats [54], and after 3 days of animal survival, the central lesion segments were dissected and cultured for 24 h *in vitro* to obtain CM‐SCI used for OTS‐SC treatment. Similarly, we obtained the conditioned medium from control‐intact thoracic spinal cord (CM‐SC), which in our experiment served as a source of control‐conditioned medium. The OTS‐SCs were prepared from lumbal SCs of naive adult male rats. The isolated SCs were transversely cut with a Mcllwain tissue chopper into 400 μm OTS‐SC sections and cultivated for an adaptation for 5 days in 1ml/well of culture media containing DMEM, 10% fetal bovine serum, and 1% penicillin–streptomycin and incubated at 37°C in humidified atmosphere with 5% of CO2 using standard interface method. The study was performed with the approval and guidelines of the Institutional Animal Care and Use Committee of the Slovak Academy of Sciences and with the European Communities Council Directive (2010/63/EU) regarding the use of animals in Research, Slovak Law for
After adaptation, OTS‐SCs were divided into four groups according to the treatment and cultivated for the next 3 days. (i) **OTS‐SC** (OTS treated with CM‐SC), (ii) **OTS‐SCI** (OTS treated
therapies directed at preventing HIV‐1‐induced neural damage in CNS [52].
models [53].
**7. Experimental study**
48 Recovery of Motor Function Following Spinal Cord Injury
spinal cord slice cultures.
Animal Protection No. 377/2012 and 436/2012.
**7.2.** *In vitro* **model of spinal cord injury and treatment**
**7.1. Methods**
**Figure 3.** The design of the experiment. After adaptation, OTS‐SCs were divided into four groups according to the treatment and cultivated for the next three days. Control group is represented by OTS‐SCs treated with conditioned medium obtained from intact spinal cord. OTS‐SCs treated with conditioned medium from injured spinal cord repre‐ sent *in vitro* model of spinal cord injury. Both groups were co‐cultured with mesenchymal stem cells to investigate the potential influence of these cells on behavior of different cell population in OTSs.
Rat MSCs were isolated from bone marrow of the long bones (femur, tibia) of adult male Wistar rats (290–320 g). To confirm the phenotypic characteristics of rMSC of the third passage, the surface markers (CD29, CD90, and CD45) were analyzed by flow cytometry [55]. The co‐culture of rMSCs with OTS‐SCs started at the same time with CM‐SCI or CM‐SC influence. On the basis of immunohistochemistry analyses for cell markers (Iba‐microglial cells, GFAP‐astro‐ cytes, NeuN‐neurons), after 3 days of treatment, OTS‐SCs were analyzed using a Leica SP5X confocal microscope and an inverted Nikon ECLIPSE Ti fluorescence microscope. We exam‐ ined the percentage of NeuN and Iba positive cells calculated from the total number of DAPI (4′,6‐diamidino‐2‐phenylindole) - positive cells in 10 random visual fields (600 × 600 μm) in OTSs. The quantification of immunofluorescence intensity of GFAP positive cells was also analyzed by Image J software according to the previous protocol [56]. Data are presented as mean ± SEM. Statistical differences between groups were evaluated with ANOVA, and values of \**p* < 0.05 and \*\**p* < 0.01 were considered to be statistically significant.
### **7.3. Results**
During the cultivation time, slices preserved their morphological and structural integrity with clear differentiation of white and gray matter. In response to injury induced by conditioned medium from injured spinal cord, both astrocytes and microglia were activated, while astrocytes served as a stimulus for microglial‐mediated inflammation.
The highest positivity of the microglial populations was observed in the injured OTS‐SCI group (3.98 ± 0.64%) when compared to control OTS‐SC (1.36 ± 0.46%, \*\**p* < 0.01) and OTS‐SC + MSCs (1.22 ± 0.62%, \*\**p* < 0.01) (**Figure 4**). OTS‐SCI represents *in vitro* model of spinal cord injury. The treatment in *in vitro* model of spinal cord injury with MSCs, in OTS‐SCI + MSCs group, caused a significant decrease of microglial cell population (2.9 ± 0.85%, \**p* < 0.05) in comparison with the actual *in vitro* model of spinal cord injury. Astrocytes showed a similar pattern of behavior like microglia cells. The highest activation of astrocytes was observed in the OTS‐SCI (density/25.26 ± 3.29) in comparison with the OTS‐SC (density/14.31 ± 4.33, \**p* < 0.05) and OTS‐ SCI + MSCs (density/19.52 ± 5.22, \**p* < 0.05) groups (**Figure 5**). We observed an adverse impact of CM‐SCI on the neuron presence in slices. The smallest number of the neurons (4.38 ± 1.77%) was found in the injured OTS‐SCI slices. The co‐culture with MSCs elevated the amount of the neurons (13.78 ± 4.51%, \*\**p* < 0.01) in OTS‐SC + MSCs when compared to actual *in vitro* model (**Figure 6**).
**Figure 4.** The number of Iba positive microglia cells observed in OTS‐SCs treated with conditioned medium from in‐ tact/injured spinal cord OTS‐SC/OTS‐SCI and co‐cultured with mesenchymal stem cells. Cultivation of spinal cord sli‐ ces in CM‐SCI caused a significant increase in number of microglial positive cells when compared to controls. Besides, co‐cultivation of injured OTS with MSCs (OTS‐SCI + MSCs) inhibits the activation of microglial cells causing a signifi‐ cant decrease of Iba positive cells in comparison to untreated *in vitro* model of spinal cord injury (OTS‐SCI).
**7.3. Results**
50 Recovery of Motor Function Following Spinal Cord Injury
(**Figure 6**).
During the cultivation time, slices preserved their morphological and structural integrity with clear differentiation of white and gray matter. In response to injury induced by conditioned medium from injured spinal cord, both astrocytes and microglia were activated, while
The highest positivity of the microglial populations was observed in the injured OTS‐SCI group (3.98 ± 0.64%) when compared to control OTS‐SC (1.36 ± 0.46%, \*\**p* < 0.01) and OTS‐SC + MSCs (1.22 ± 0.62%, \*\**p* < 0.01) (**Figure 4**). OTS‐SCI represents *in vitro* model of spinal cord injury. The treatment in *in vitro* model of spinal cord injury with MSCs, in OTS‐SCI + MSCs group, caused a significant decrease of microglial cell population (2.9 ± 0.85%, \**p* < 0.05) in comparison with the actual *in vitro* model of spinal cord injury. Astrocytes showed a similar pattern of behavior like microglia cells. The highest activation of astrocytes was observed in the OTS‐SCI (density/25.26 ± 3.29) in comparison with the OTS‐SC (density/14.31 ± 4.33, \**p* < 0.05) and OTS‐ SCI + MSCs (density/19.52 ± 5.22, \**p* < 0.05) groups (**Figure 5**). We observed an adverse impact of CM‐SCI on the neuron presence in slices. The smallest number of the neurons (4.38 ± 1.77%) was found in the injured OTS‐SCI slices. The co‐culture with MSCs elevated the amount of the neurons (13.78 ± 4.51%, \*\**p* < 0.01) in OTS‐SC + MSCs when compared to actual *in vitro* model
**Figure 4.** The number of Iba positive microglia cells observed in OTS‐SCs treated with conditioned medium from in‐ tact/injured spinal cord OTS‐SC/OTS‐SCI and co‐cultured with mesenchymal stem cells. Cultivation of spinal cord sli‐ ces in CM‐SCI caused a significant increase in number of microglial positive cells when compared to controls. Besides, co‐cultivation of injured OTS with MSCs (OTS‐SCI + MSCs) inhibits the activation of microglial cells causing a signifi‐
cant decrease of Iba positive cells in comparison to untreated *in vitro* model of spinal cord injury (OTS‐SCI).
astrocytes served as a stimulus for microglial‐mediated inflammation.
**Figure 5.** The GFAP density detected in OTS‐SCs treated with conditioned medium from intact/injured spinal cord OTS‐SC/OTS‐SCI and co‐cultured with mesenchymal stem cells. The highest activation of astrocytes was observed in the OTS‐SCI and co‐cultivation of injured OTS with MSCs (OTS‐SCI + MSCs) caused a significant decrease of astro‐ cytes when compared to untreated *in vitro* model of spinal cord injury (OTS‐SCI).
**Figure 6.** The number of NeuN positive cells noted in spinal cord slices treated with conditioned medium from intact/ injured spinal cord OTS‐SC/OTS‐SCI and co‐cultured with mesenchymal stem cells. Cultivation of spinal cord slices in CM‐SCI caused a significant decrease in number of neuronal cells when compared to controls. MSCs have neurotro‐ phic effects resulting in an increase of neuronal cells in *in vitro* model of spinal cord injury treated with MSCs (OTS‐SCI + MSCs).
Our data revealed increased activity of microglia and astrocytes and elimination of neurons in experimental group treated with CM‐SCI. We observed the presence of hypertrophied astrocytes with increased proliferation activity within the white matter of injured experimental group in contrary to controls. Astrocytes are one of the major glial cell types that maintain homeostasis in the undamaged CNS. After injury, astrocytes become reactive and prevent regeneration; however, it has also been suggested that astrocytes can become activated and promote regeneration. This indicated that reactive astrogliosis may exert both beneficial and detrimental effects in a context‐specific manner determined by distinct molecular signaling cascades. Resident microglia in the control group OTS‐SC was in the resting state and cells were few in number and smaller in size (**Figure 7A**). After inducing damage effects in the OTS‐ SCI with CM‐SCI, quiescent microglial cells became activated, were bigger, and created ramified processes (**Figure 7B**).
**Figure 7.** Fluorescent microscopy images of the residual microglia (red) in organotypic spinal cord slices (400 μm) stained with specific Iba antibody. The immunoreactive cells in control OTS‐SC group (A) and injured OTS‐SCI group (B). Scale bar is 10 μm.
Our study showed that (i) OTS‐SC *in vitro* model closely mimics the post‐injury microenvir‐ onment affecting the microglia and astrocytes cell response following secondary damage and is suitable for various pharmacy‐therapeutic approaches, and that (ii) rMSCs are able to inhibit the activation of microglial cells and have neurotrophic effects on the neural cells. Thus, the effects shown in this study are directly mediated from various factors or proteins secreted from rMSCs. Our study represents the rMSC as a promising candidate for SCI treatment due to their potential to perform the immunomodulatory functions and to produce specific neurotrophic components and cytokines likewise to other results [57]. These techniques give us the possi‐ bility to achieve a moderate trauma in *in vitro* preparation, the organotypic slices that can be used to study short‐ and long‐term effects and mechanisms of traumatic injury. Our study supports the recent hypothesis regarding the role of MSCs in establishing microenvironment more favorable for tissue repair. MSCs have a great potential in therapy for a range of neural insults.
### **8. Summary**
Our data revealed increased activity of microglia and astrocytes and elimination of neurons in experimental group treated with CM‐SCI. We observed the presence of hypertrophied astrocytes with increased proliferation activity within the white matter of injured experimental group in contrary to controls. Astrocytes are one of the major glial cell types that maintain homeostasis in the undamaged CNS. After injury, astrocytes become reactive and prevent regeneration; however, it has also been suggested that astrocytes can become activated and promote regeneration. This indicated that reactive astrogliosis may exert both beneficial and detrimental effects in a context‐specific manner determined by distinct molecular signaling cascades. Resident microglia in the control group OTS‐SC was in the resting state and cells were few in number and smaller in size (**Figure 7A**). After inducing damage effects in the OTS‐ SCI with CM‐SCI, quiescent microglial cells became activated, were bigger, and created
**Figure 7.** Fluorescent microscopy images of the residual microglia (red) in organotypic spinal cord slices (400 μm) stained with specific Iba antibody. The immunoreactive cells in control OTS‐SC group (A) and injured OTS‐SCI group
Our study showed that (i) OTS‐SC *in vitro* model closely mimics the post‐injury microenvir‐ onment affecting the microglia and astrocytes cell response following secondary damage and is suitable for various pharmacy‐therapeutic approaches, and that (ii) rMSCs are able to inhibit the activation of microglial cells and have neurotrophic effects on the neural cells. Thus, the effects shown in this study are directly mediated from various factors or proteins secreted from rMSCs. Our study represents the rMSC as a promising candidate for SCI treatment due to their potential to perform the immunomodulatory functions and to produce specific neurotrophic components and cytokines likewise to other results [57]. These techniques give us the possi‐ bility to achieve a moderate trauma in *in vitro* preparation, the organotypic slices that can be used to study short‐ and long‐term effects and mechanisms of traumatic injury. Our study supports the recent hypothesis regarding the role of MSCs in establishing microenvironment more favorable for tissue repair. MSCs have a great potential in therapy for a range of neural
ramified processes (**Figure 7B**).
52 Recovery of Motor Function Following Spinal Cord Injury
(B). Scale bar is 10 μm.
insults.
Many *in vitro* models are invented, and others are still waiting for their further application, depending on the interest of researchers and key problems they are focused on. Many *in vitro* models are used in bright research field concerning the influence of any factor originated due to human activity we come into contact every day. For example, exposure of primary rat cortical neurons to extremely low‐frequency electromagnetic fields has only limited (devel‐ opmental) neurotoxic potential *in vitro* manifested by unchanged cell viability and neurite outgrowth [58]. Organotypic slice cultures have the potential to become powerful tools in the arsenal of drug discovery technology, lying at the interface between high‐throughput screen‐ ing and clinically relevant animal disease models. It has to be mentioned, however, that the cultured slices cannot represent all the *in vivo* properties. However, organotypic slices can model different histopathological aspects of neurological conditions and are a suitable *in vitro* system to address a wide range of questions concerning mechanisms of nerve regenera‐ tion after injury in the adult nervous system and to identify the factors supporting or prevent‐ ing axonal outgrowth as well as cell survival that will ultimately add to the development of therapeutic application for spinal cord injury. Organotypic culture technique is easy, efficient, and practical and allows reduction of animal number and preserves the three‐dimensional neuronal network. Although these *in vitro* organotypic slices have been able to model and reproduce many processes taking place after CNS injury with subsequent regeneration, some limitations exist, including the lack of any functioning vasculature, an incomplete immune system. Thus, OTSs have their own advantages and disadvantages with regard to stimulating *in vivo* conditions. Obviously, organotypic slices cannot replace *in vivo* models, which are still necessary and remain the only way to evaluate the functional outcome of a therapeutic strategy [59].
### **Acknowledgements**
We acknowledge the financial support given by the Grant Agency of the Slovak Academy of Sciences: VEGA 2/0169/13, VEGA 2/0125/15, and VEGA 2/0145/16.
### **Author details**
Lucia Slovinska\* , Juraj Blasko, Miriam Nagyova, Eva Szekiova and Dasa Cizkova
\*Address all correspondence to: slovinska@saske.sk
Institute of Neurobiology, Slovak Academy of Sciences, Kosice, Slovakia
### **References**
[14] Ioannidou K, Edgar JM, Barnett SC: Time‐lapse imaging of glial‐axonal interactions. Curr Protoc Neurosci 2012; 72: 2.23.21–22.23.14.
**References**
Focus 2008; 25: E2.
54 Recovery of Motor Function Following Spinal Cord Injury
1999; 49: 377–391.
1990; 10: 3594–3603.
75: 107–114.
[1] Rowland J, Hawryluk GW, Kwon B, Fehlings MG: Current status of acute spinal cord injury pathophysiology and emerging therapies: promise on the horizon. Neurosurg
[2] Fawcett JW, Asher RA: The glial scar and central nervous system repair. Brain Res Bull
[3] Rudge JS, Silver J: Inhibition of neurite outgrowth on astroglial scars in vitro. J Neurosci
[4] Tom VJ, Steinmetz MP, Miller JH, Doller CM, Silver J: Studies on the development and behavior of the dystrophic growth cone, the hallmark of regeneration failure, in an in vitro model of the glial scar and after spinal cord injury. J Neurosci 2004; 24: 6531–6539.
[5] Boomkamp SD, McGrath MA, Houslay MD, Barnett SC: Epac and the high affinity rolipram binding conformer of PDE4 modulate neurite outgrowth and myelination using an in vitro spinal cord injury model. Br J Pharmacol 2014; 171: 2385–2398.
[6] Slovinska L, Szekiova E, Blasko J, Devaux S, Salzet M, Cizkova D: Comparison of dynamic behavior and maturation of neural multipotent cells derived from different spinal cord developmental stages: an in vitro study. Acta Neurobiol Exp (Wars) 2015;
[7] Morrison B, 3rd, Saatman KE, Meaney DF, McIntosh TK: In vitro central nervous system models of mechanically induced trauma: a review. J Neurotrauma 1998; 15: 911–928.
[8] Shearer MC, Niclou SP, Brown D, Asher RA, Holtmaat AJ, Levine JM, Verhaagen J, Fawcett JW: The astrocyte/meningeal cell interface is a barrier to neurite outgrowth which can be overcome by manipulation of inhibitory molecules or axonal signalling
[9] Wanner IB, Deik A, Torres M, Rosendahl A, Neary JT, Lemmon VP, Bixby JL: A new in vitro model of the glial scar inhibits axon growth. Glia 2008; 56: 1691–1709.
[10] Hopkins AM, DeSimone E, Chwalek K, Kaplan DL: 3D in vitro modeling of the central
[11] Boomkamp SD, Riehle MO, Wood J, Olson MF, Barnett SC: The development of a rat in vitro model of spinal cord injury demonstrating the additive effects of Rho and ROCK inhibitors on neurite outgrowth and myelination. Glia 2012; 60: 441–456.
[12] Kim BG, Hwang DH, Lee SI, Kim EJ, Kim SU: Stem cell‐based cell therapy for spinal
[13] Nash B, Thomson CE, Linington C, Arthur AT, McClure JD, McBride MW, Barnett SC: Functional duality of astrocytes in myelination. J Neurosci 2011; 31: 13028–13038.
pathways. Mol Cell Neurosci 2003; 24: 913–925.
nervous system. Prog Neurobiol 2015; 125: 1–25.
cord injury. Cell Transplant 2007; 16: 355–364.
[41] Bonnici B, Kapfhammer JP: Spontaneous regeneration of intrinsic spinal cord axons in a novel spinal cord slice culture model. Eur J Neurosci 2008; 27: 2483–2492.
[28] Savas A, Warnke PC, Ginap T, Feuerstein TJ, Ostertag CB: The effects of continuous and single‐dose radiation on choline uptake in organotypic tissue slice cultures of rabbit
[29] Jeong D, Taghavi CE, Song KJ, Lee KB, Kang HW: Organotypic human spinal cord slice culture as an alternative to direct transplantation of human bone marrow precursor
[30] Noraberg J, Kristensen BW, Zimmer J: Markers for neuronal degeneration in organo‐
[31] Morrison B, 3rd, Eberwine JH, Meaney DF, McIntosh TK: Traumatic injury induces differential expression of cell death genes in organotypic brain slice cultures deter‐ mined by complementary DNA array hybridization. Neuroscience 2000; 96: 131–139.
[32] Ravikumar M, Jain S, Miller RH, Capadona JR, Selkirk SM: An organotypic spinal cord slice culture model to quantify neurodegeneration. J Neurosci Methods 2012; 211: 280–
[33] Noraberg J, Poulsen FR, Blaabjerg M, Kristensen BW, Bonde C, Montero M, Meyer M, Gramsbergen JB, Zimmer J: Organotypic hippocampal slice cultures for studies of brain damage, neuroprotection and neurorepair. Curr Drug Targets CNS Neurol Disord
[34] Sundstrom L, Morrison B, 3rd, Bradley M, Pringle A: Organotypic cultures as tools for
[35] Frantseva MV, Carlen PL, El‐Beheiry H: A submersion method to induce hypoxic damage in organotypic hippocampal cultures. J Neurosci Methods 1999; 89: 25–31.
[36] Feeney CJ, Frantseva MV, Carlen PL, Pennefather PS, Shulyakova N, Shniffer C, Mills LR: Vulnerability of glial cells to hydrogen peroxide in cultured hippocampal slices.
[37] An SS, Pennant WA, Ha Y, Oh JS, Kim HJ, Gwak SJ, Yoon DH, Kim KN: Hypoxia‐ induced expression of VEGF in the organotypic spinal cord slice culture. Neuroreport
[38] Mazzone GL, Nistri A: S100beta as an early biomarker of excitotoxic damage in spinal
[39] Nava J, Mayorenko II, Grehl T, Steinbuch HWM, Weis J, Brook GA: Differential pattern of neuroprotection in lumbar, cervical and thoracic spinal cord segments in an orga‐ notypic rat model of glutamate‐induced excitotoxicity. J Chem Neuroanat 2013; 53: 11–
[40] Balentine JD, Greene WB, Bornstein M: In vitro spinal cord trauma. Lab Invest 1988;
cord organotypic cultures. J Neurochem 2014; 130: 598–604.
functional screening in the CNS. Drug Discov Today 2005; 10: 993–1000.
cells for treating spinal cord injury. World Neurosurg 2011; 75: 533–539.
hippocampus. Neurol Res 2001; 23: 669–675.
56 Recovery of Motor Function Following Spinal Cord Injury
288.
2005; 4: 435–452.
Brain Res 2008; 1198: 1–15.
2011; 22: 55–60.
17.
58: 93–99.
typic slice cultures. Brain Res Protoc 1999; 3: 278–290.
**Modulating Cellular and Molecular Responses to Spinal Cord Injury**
[54] Vanicky I, Urdzikova L, Saganova K, Cizkova D, Galik J: A simple and reproducible model of spinal cord injury induced by epidural balloon inflation in the rat. J Neuro‐
[55] Nagyova M, Slovinska L, Blasko J, Grulova I, Kuricova M, Cigankova V, Harvanova D, Cizkova D: A comparative study of PKH67, DiI, and BrdU labeling techniques for tracing rat mesenchymal stem cells. In Vitro Cell Dev Biol Anim 2014; 50: 656–663. [56] Jones LL, Tuszynski MH: Spinal cord injury elicits expression of keratan sulfate proteoglycans by macrophages, reactive microglia, and oligodendrocyte progenitors.
[57] Bai L, Lennon DP, Caplan AI, DeChant A, Hecker J, Kranso J, Zaremba A, Miller R: Hepatocyte growth factor mediates MSCs stimulated functional recovery in animal
[58] de Groot MW, van Kleef RG, de Groot A, Westerink RH: In vitro developmental neurotoxicity following chronic exposure to 50 Hz extremely low‐frequency electro‐
[59] Daviaud N, Garbayo E, Schiller PC, Perez‐Pinzon M, Montero‐Menei CN: Organotypic cultures as tools for optimizing central nervous system cell therapies. Exp Neurol 2013;
magnetic fields in primary rat cortical cultures. Toxicol Sci 2016; 149: 433–440.
trauma 2001; 18: 1399–1407.
58 Recovery of Motor Function Following Spinal Cord Injury
J Neurosci 2002; 22: 4611–4624.
248: 429–440.
models of MS. Nat Neurosci 2012; 15: 862–870.
## **Protective Role of the Immune System in Spinal Cord Injury: Immunomodulation with Altered Peptide Ligands**
Paola Suárez‐Meade and Antonio Ibarra
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/63221
### **Abstract**
Spinal cord injury (SCI) is a phenomenon characterized by damage to the spinal cord and nerve roots, resulting in loss of physiological activity below the lesion. Injury to the spinal cord activates a cascade of cellular and molecular reactions in which the immune system plays an essentialrole, as there is an uncontrolled immune response that endows further damage to neural tissue. However, the activity of immune system at the site of injury can be modified in order to obtain a neuroprotective environment and promote SCI recovery. This strategy has been designed under the light of the innovative concept "protective autoimmunity" (PA) and can be stimulated with the use of altered peptide ligands (APL). Adequate immunomodulation with APL can be obtained with the peptide A91, which is a safe synthetic peptide derived from the myelin basic protein (MBP) that has proven to be effective in preclinical research. Immunization with A91 is carried out with the objective of preventing further damage and promoting neuropro‐ tection. This peptide has direct influence over SCI secondary mechanisms such as inflammation, lipid peroxidation, and apoptosis. Preclinical results suggest that immunization with A91 could be an effective treatment in the clinical field, providing a better quality of life to SCI patients.
**Keywords:** spinal cord injury, protective autoimmunity, A91, altered peptide ligand, immunization
### **1. Introduction**
Spinal cord injury (SCI) is perhaps one of the most devastating conditions as it results in a disruption of motor, sensory, and autonomic functions, leading to permanent neurological
© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
disability. According to the National Spinal Cord Injury Statistical Center (NSCISC), in 2014 over 276,000 people were suffering from SCI [1]. The incidence ranges over 12,500 new cases each year, with a prevalence of approximately 40 cases per million in the United States [2]. Epidemiological studies indicate that SCI has a higher incidence among male population, and people between 30 and 50 years old [3, 3]. Clinically, the most disabling outcomes of traumat‐ ic SCI are motor deficit and sensory loss. Nonetheless, due to autonomic dysfunction and depending on the level and severity of injury, SCI may also alter normal homeostasis and the respiratory, reproductive, urinary, and gastrointestinal systems [4–6]. Besides surgical intervention, the standard of care for SCI in several countries focuses on preventing shock and further damage with the use of methylprednisolone (MP), a synthetic glucocorticoid with anti‐ inflammatoryproperties.The NationalAcute Spinal CordInjury Study (NASCIS)trials suggest that high‐dose MP is effective for the management of acute SCI [5, 8]. However, further studies about NASCIS results indicate that the clinical benefits of the use of MP as a treatment for SCI are questionable [6, 10]. Acute MP therapy reduces cellular damage and secondary injury mechanisms butleads to risks of high‐dose steroids [7]. Forthis reason,MPadministration after SCI is still a controversial treatment, and it continues to be debated [8].
Another molecule that has been widely studied for the clinical treatment of SCI is Ganglioside, which is highly expressed in the cell membranes of the central nervous system (CNS). The synthetic form of this glycosphingolipid—monosialotetrahexosylganglioside (GM1)—was known for having anti‐apoptotic, anti‐excitotoxic, and neuroprotective properties [9, 14]. Therefore, several clinical trials were designed in order to test this drug on SCI [10]. In the most recent study, patients received NASCIS II doses of MP, and different doses of GM1 after the effect of MP was over. Although, GM1 treated patients presented a significant recovery, the beneficial effects related to the drug were inconsistent between different types of injuries and were lost during chronic SCI stages [11, 17]. These results suggest that GM1 has a limited effect and to the date, it is not an approved treatment for SCI [12].
Undoubtedly, the lack of an available treatment for SCI in the clinical field highlights the need to design new and safe therapies for injured patients. Nowadays, research is focused on targeting secondary SCI mechanisms with the objective of minimizing further damage, promoting regeneration and thus, improving functional recovery [13, 19]. One of the most important secondary injury mechanisms is the post‐traumatic inflammatory response, which is roughly characterized by the presence of injury‐dependent pro‐inflammatory cytokines and infiltration of peripheral immune cells to the damaged area [14–22]. It was previously thought that the presence of this uncontrolled immune response in the CNS was harmful and patho‐ logical. However, other findings state a controversial theory: immune cells could play an essential role in neuroprotection and regeneration of the spinal cord after injury [15]. A better understanding of how inflammation mediates secondary injury suggests that suppressing all immune responses with the use of glucocorticoids is no longer a rational treatment approach. This information has led scientists to further investigate the beneficial effects of the immune system in several neurological conditions, including SCI [16]. Focusing on the beneficial mechanisms of the immune system after trauma has opened the doors for the design of a new therapeutic strategy. Nevertheless, some questions should first be answered. For instance, how can the immune system be modulated to attain a protective microenvironment? What are the immune‐related elements capable of providing the required immune‐modulation? And how are they able to provide recovery after SCI?
### **2. Spinal cord injury pathophysiology**
disability. According to the National Spinal Cord Injury Statistical Center (NSCISC), in 2014 over 276,000 people were suffering from SCI [1]. The incidence ranges over 12,500 new cases each year, with a prevalence of approximately 40 cases per million in the United States [2]. Epidemiological studies indicate that SCI has a higher incidence among male population, and people between 30 and 50 years old [3, 3]. Clinically, the most disabling outcomes of traumat‐ ic SCI are motor deficit and sensory loss. Nonetheless, due to autonomic dysfunction and depending on the level and severity of injury, SCI may also alter normal homeostasis and the respiratory, reproductive, urinary, and gastrointestinal systems [4–6]. Besides surgical intervention, the standard of care for SCI in several countries focuses on preventing shock and further damage with the use of methylprednisolone (MP), a synthetic glucocorticoid with anti‐ inflammatoryproperties.The NationalAcute Spinal CordInjury Study (NASCIS)trials suggest that high‐dose MP is effective for the management of acute SCI [5, 8]. However, further studies about NASCIS results indicate that the clinical benefits of the use of MP as a treatment for SCI are questionable [6, 10]. Acute MP therapy reduces cellular damage and secondary injury mechanisms butleads to risks of high‐dose steroids [7]. Forthis reason,MPadministration after
Another molecule that has been widely studied for the clinical treatment of SCI is Ganglioside, which is highly expressed in the cell membranes of the central nervous system (CNS). The synthetic form of this glycosphingolipid—monosialotetrahexosylganglioside (GM1)—was known for having anti‐apoptotic, anti‐excitotoxic, and neuroprotective properties [9, 14]. Therefore, several clinical trials were designed in order to test this drug on SCI [10]. In the most recent study, patients received NASCIS II doses of MP, and different doses of GM1 after the effect of MP was over. Although, GM1 treated patients presented a significant recovery, the beneficial effects related to the drug were inconsistent between different types of injuries and were lost during chronic SCI stages [11, 17]. These results suggest that GM1 has a limited effect
Undoubtedly, the lack of an available treatment for SCI in the clinical field highlights the need to design new and safe therapies for injured patients. Nowadays, research is focused on targeting secondary SCI mechanisms with the objective of minimizing further damage, promoting regeneration and thus, improving functional recovery [13, 19]. One of the most important secondary injury mechanisms is the post‐traumatic inflammatory response, which is roughly characterized by the presence of injury‐dependent pro‐inflammatory cytokines and infiltration of peripheral immune cells to the damaged area [14–22]. It was previously thought that the presence of this uncontrolled immune response in the CNS was harmful and patho‐ logical. However, other findings state a controversial theory: immune cells could play an essential role in neuroprotection and regeneration of the spinal cord after injury [15]. A better understanding of how inflammation mediates secondary injury suggests that suppressing all immune responses with the use of glucocorticoids is no longer a rational treatment approach. This information has led scientists to further investigate the beneficial effects of the immune system in several neurological conditions, including SCI [16]. Focusing on the beneficial mechanisms of the immune system after trauma has opened the doors for the design of a new therapeutic strategy. Nevertheless, some questions should first be answered. For instance, how
SCI is still a controversial treatment, and it continues to be debated [8].
62 Recovery of Motor Function Following Spinal Cord Injury
and to the date, it is not an approved treatment for SCI [12].
Damage to the spinal cord (SC) causes anatomical and functional deficits to the CNS, that result in the appearance of several long‐term medical comorbidities. SCI is characterized by two different pathophysiological phases: primary and secondary injury [17]. Initial trauma to the SC—known as primary injury—is caused by a compressive or contusive mechanism that results in gross anatomical tissue disruption, and immediate hemostatic self‐defense events that produce further damage to CNS structures. Direct impact to the SC leads to vascular disturbances such as hemorrhage, ischemia, edema, and hypoperfusion, resulting in tissue necrosis [18]. Hemorrhage and edema formation can raise the risk of developing increased parenchymal pressure and produce more tissue damage [19]. Also, reactive gliosis, demyeli‐ nation, and axonal loss are often caused by immediate trauma and sustained compression to neural tissue [20]. Depending upon primary injury characteristics, there could be greater tissue damage and worse functional outcomes. Therefore, during this phase, treatment should focus on hemorrhage control to avoid necrosis and early decompression to stabilize intrathecal pressure.
As a consequence of primary injury, there is a cascade of biological reactions that occur minutes after injury and last for several weeks known as secondary injury [21]. This phase is quite complex, as it consists of the development of mechanisms like loss of ATP‐dependent cellular functions, ion homeostasis imbalance, excitotoxicity, oxidative stress, lipid peroxidation, inflammation, and apoptosis [22]. There are several secondary mechanisms that are strongly related to the ischemic event observed in SCI. Ischemia produces a depletion of the intracellular amount of ATP, leading to a reduction in the energy‐dependent cell function that preserve ion homeostasis [23]. Therefore, the sodium‐potassium pump cannot execute its physiological activity, resulting in an elevated potassium (K<sup>+</sup> ) efflux, and a high influx of sodium (Na<sup>+</sup> ), calcium (Ca2+), and chloride (Cl- ) into the cell. This homeostasis imbalance alters normal ion concentrations within the intracellular and extracellular spaces, producing a sustained membrane depolarization and a release of excitatory amino acid (EAA) neurotransmitters [24]. The pathological effect related to an increased concentration of EAA neurotransmitters, such as glutamate and aspartate, is known as excitotoxicity. This secondary injury mechanism is characterized by an overstimulation of the NMDA, kainate and AMPA glutamate receptors, which causes massive intracellular Ca2+ concentrations, resulting in a pathological neuronal excitation and cell death [25, 32]. Glial cells—especially oligodendrocytes—are very sensitive to excitotoxic damage because of their high expression of ionotropic glutamate receptors [26]. That is why an excessive glutamate accumulation related to SCI can produce oligodendrocyte death, and consequent white matter demyelination [27]. Also, because of glutamate excito‐ toxicity, there is an increased production of free radicals by reactive microglia, which contrib‐ ute to lipid peroxidation (LP), and mitochondrial dysfunction [28, 36]. Therefore, this sustained toxic microenvironment is postulated to be one of the most detrimental secondary injury mechanisms related to SCI.
Oxidative stress accompanies secondary injury damage, and is mainly characterized by an increased mitochondrial production of reactive oxygen species (ROS) and reactive nitrogen species (RNS) [29, 38]. The elevation in ROS and RNS concentration is closely related to the aforementioned high Ca2+ influx to the cells, as it stimulates free radical production [30]. Free radicals such as superoxide anion (O2 - ), nitric oxide (NO), and peroxynitrite (ONOO- ) create a toxic microenvironment by oxidizing nearby molecules producing neural energy failure, blood‐brain barrier dysfunction, vascular reactivity, and potentiating inflammation [31, 41]. At high concentrations, these molecules can become cytotoxic and worsen secondary injury mechanisms [32]. Oxidative stress also influences excitotoxicity by exacerbating Ca2+ deregu‐ lation and thus, glutamate concentrations. The pathological production of ROS that arises after trauma causes oxidative damage, especially on lipids, originating LP. LP is characterized by producing a disruption in the normal structure of the polyunsaturated fatty acids in the cell membrane, such as arachidonic acid, and linoleic acid. In LP, the high concentration of free radicals, results in functional compromise and cell death [33, 43]. Also, structural damage to the cell membrane produces a reduction in the generation and transmission of the electrical potentials leading to synapse dysfunction [34]. Altogether these mechanisms potentiate apoptosis, which is a form of programmed cell death characterized by cell shrinkage, nuclear pyknosis, and chromatin aggregation in a stressful environment [35]. This deleterious event, (apoptosis) occurs after SCI by stimulation of the apoptosis‐inducing factor (AIF) to the nucleus, or by direct mitochondrial disruption leading to a subsequent activation of caspase‐ 3 signaling pathways [36, 47]. Evidence supports that secondary injury mechanisms contribute to delayed tissue damage, exacerbating damage, and limiting recovery after traumatic SCI.
### **2.1. Pathophysiological involvement of the immune system**
The immune system has a pivotal and somehow controversial role within the pathophysiology of traumatic SCI. Immediately after trauma there is an activation of the inflammatory response that consists in the proliferation of resident microglia and astrocytes, a high concentration of pro‐inflammatory molecules, and infiltration of peripheral immune cells to the site of injury (**Figure 1**). SCI induces the activation of a series of inflammatory stimuli leading to an increased concentration of cytokines and inflammatory cells that will determine the extent of secondary damage [37]. Evidence suggests that in the presence of an excitotoxic and inflammatory microenvironment, microglial cells differentiate into a M1 pro‐inflammatory phenotype [38]. Under these conditions, activated microglia is capable of secreting interleukin 1β (IL‐1β), IL‐ 6, tumor necrosis factor‐alpha (TNFα), and macrophage colony‐stimulating factor (MCSF) which are pro‐inflammatory in nature [39].
A high free radical and cytotoxic substances secretion is more evident when there is microglial activation starting at day one, and increasing at 7 days post‐injury [40]. That is why the immune response is closely related to LP, as these cells are capable of boosting ROS and NO concen‐ trations, favoring oxidative stress, cell membrane dysfunction and thus, apoptosis [41, 50]. At the same time, TNFα stimulates astrocyte proliferation and growth, leading to the formation of the glial scar within the chronic stages of SCI impeding axonal regeneration through the site of injury [42, 53]. Glial cells act immediately after trauma; they secrete pro‐inflammatory molecules and promote inflammation, favoring the appearance of secondary injury mecha‐ nisms. In spite of the above mentioned deleterious effects, microglial—especially when differentiated into an M2 phenotype (IL‐10 and TGF‐beta)—and astrocyte activation could produce a beneficial effect through a high production of growth factors like brain derived neurotrophic factor (BDNF), and neurotrophin‐3 (NT‐3), essential for tissue repair [43]. Also, these cells are capable of expressing glutamate transporters that help reducing harmful
ute to lipid peroxidation (LP), and mitochondrial dysfunction [28, 36]. Therefore, this sustained toxic microenvironment is postulated to be one of the most detrimental secondary injury
Oxidative stress accompanies secondary injury damage, and is mainly characterized by an increased mitochondrial production of reactive oxygen species (ROS) and reactive nitrogen species (RNS) [29, 38]. The elevation in ROS and RNS concentration is closely related to the aforementioned high Ca2+ influx to the cells, as it stimulates free radical production [30]. Free
toxic microenvironment by oxidizing nearby molecules producing neural energy failure, blood‐brain barrier dysfunction, vascular reactivity, and potentiating inflammation [31, 41]. At high concentrations, these molecules can become cytotoxic and worsen secondary injury mechanisms [32]. Oxidative stress also influences excitotoxicity by exacerbating Ca2+ deregu‐ lation and thus, glutamate concentrations. The pathological production of ROS that arises after trauma causes oxidative damage, especially on lipids, originating LP. LP is characterized by producing a disruption in the normal structure of the polyunsaturated fatty acids in the cell membrane, such as arachidonic acid, and linoleic acid. In LP, the high concentration of free radicals, results in functional compromise and cell death [33, 43]. Also, structural damage to the cell membrane produces a reduction in the generation and transmission of the electrical potentials leading to synapse dysfunction [34]. Altogether these mechanisms potentiate apoptosis, which is a form of programmed cell death characterized by cell shrinkage, nuclear pyknosis, and chromatin aggregation in a stressful environment [35]. This deleterious event, (apoptosis) occurs after SCI by stimulation of the apoptosis‐inducing factor (AIF) to the nucleus, or by direct mitochondrial disruption leading to a subsequent activation of caspase‐ 3 signaling pathways [36, 47]. Evidence supports that secondary injury mechanisms contribute to delayed tissue damage, exacerbating damage, and limiting recovery after traumatic SCI.
The immune system has a pivotal and somehow controversial role within the pathophysiology of traumatic SCI. Immediately after trauma there is an activation of the inflammatory response that consists in the proliferation of resident microglia and astrocytes, a high concentration of pro‐inflammatory molecules, and infiltration of peripheral immune cells to the site of injury (**Figure 1**). SCI induces the activation of a series of inflammatory stimuli leading to an increased concentration of cytokines and inflammatory cells that will determine the extent of secondary damage [37]. Evidence suggests that in the presence of an excitotoxic and inflammatory microenvironment, microglial cells differentiate into a M1 pro‐inflammatory phenotype [38]. Under these conditions, activated microglia is capable of secreting interleukin 1β (IL‐1β), IL‐ 6, tumor necrosis factor‐alpha (TNFα), and macrophage colony‐stimulating factor (MCSF)
A high free radical and cytotoxic substances secretion is more evident when there is microglial activation starting at day one, and increasing at 7 days post‐injury [40]. That is why the immune response is closely related to LP, as these cells are capable of boosting ROS and NO concen‐ trations, favoring oxidative stress, cell membrane dysfunction and thus, apoptosis [41, 50]. At
), nitric oxide (NO), and peroxynitrite (ONOO-
) create a
**2.1. Pathophysiological involvement of the immune system**
which are pro‐inflammatory in nature [39].
mechanisms related to SCI.
radicals such as superoxide anion (O2
64 Recovery of Motor Function Following Spinal Cord Injury
**Figure 1. Cellular immune response in SCI**. **A.** M1 microglial differentiation, astrocyte activation, and IL‐1β, IL‐6, and TNFα secretion. **B.** Increase in the concentration of ROS, NO, and RNS. **C.** Peripheral infiltration of neutrophils and myeloperoxidase secretion. **D.** Macrophages arise from peripheral tissue and activate the adaptive immune response. **E.** Infiltration of T‐lymphocytes to the site of injury and INFγ secretion. Abbreviations: IL, interleukin; RNS, reactive nitrogen species; ROS, reactive oxygen species; NO, nitric oxide; TNFα, tumor necrosis factor‐alpha; INFγ, interferon‐ gamma; MYO, myeloperoxidase.
concentrations, leading to a reduction in excitotoxicity [44]. Evidence suggests that even though the immune system is considered to be completely pathological in nature, it can also provide beneficial effects for SCI repair.
In normal conditions, inflammatory reaction leads the response to a pathological outcome, promoting damage and spinal cord degeneration. The severity of SCI determines the intensity of the inflammatory response and the glial reaction to SCI. Glial cells are well distributed within the CNS, and have the ability to proliferate and migrate to the site of injury. In response to injury there is a high glial secretion of cytokines and chemokines—such as IL‐1, IL‐6, and TNFα—that allows migration of peripheral immune cells [45]. TNFα stimulates the expression of adhesion molecules like endothelial intracellular adhesion molecule ‐1 (ICAM‐1) and vascular cell adhesion molecule‐1 (VCAM‐1) altering the blood‐brain barrier permeability and inducing a peripheral infiltration of immune cells to the CNS (**Figure 1**). Cell infiltration to the CNS is considered the principal factor for tissue disruption and sustained neural damage. The first peripheral cells to arrive at the site of injury are neutrophils [46]. These cells have the ability to phagocyte and clear debris, but they also secrete ROS and RNS, as well as other pathological proteolytic and oxidative enzymes like myeloperoxidase, producing greater tissue damage [47, 58]. At the same time, macrophages arise from resident microglia or from peripheral monocytes with the objective of removing cell debris and stimulating angiogenesis [48]. These cells play an essential role in the immune response, as they help in the activation and reclusion of the adaptive immune system cells. Microglia, astrocytes and dendritic cells may act as antigen‐presenting cells (APC), leading to T lymphocyte activation, proliferation, and infiltration to the site of injury [49]. Lymphocytes recognize the signal and proliferate creating large numbers of clones, specific to the antigen being presented by the APC. At 3 days post‐injury, there is an evident infiltration of T‐cells to the CNS, these cells determine the intensity and continue on modulating the immune response to trauma [50]. In SCI, the presence of T‐lymphocytes is considered detrimental as they secrete a Th1 cytokine profile including interferon gamma (IFNγ) [51, 63]. IFNγ is a pro‐inflammatory cytokine capable of inducing free radical production, increasing IL‐6, IL‐12, IL‐1β, and TNFα concentrations, and activating apoptotic‐signaling pathways [52–65]. These events represent the immune activity within the acute phases of SCI and if it is not well controlled it could turn into a chronic and degenerative immune response.
When inflammation is modulated, i.e., we control the intensity of inflammatory response, the type, the action, and the arrival of immune cells, we could expect a change in the final outcome. With this regard, there is a strong evidence suggesting that lymphocytes could also have favorable activity, as they are capable of synthesizing several growth factors, like BDNF, NT3, and nerve growth factor (NGF) [53–68]. These molecules are known for being capable of promoting a protective and regenerative environment for CNS disorders [54].
To the date, it is clear that the immune response appearing after SCI could be pathological if it is not well controlled at the onset of SCI [55]. In a short period of time, this uncontrolled response leads to the extensive recognition of CNS peptides, proteins, lipids, or nucleic acids by the immune system [56]. This interaction between immune cells and CNS constituents like the MBP—promotes the activation of lymphocytes and thereby, the possible development of an autoimmune response [57, 72, 73]. As a result, higher levels of demyelination are noted, leading to loss of sensitive and motor synapses. Also, several studies have identified that B cells secrete self‐reactive antibodies and pro‐inflammatory cytokines, promoting autoimmune activity [58, 75]. Therefore, the immune response elicited after SCI is considered to be one of the most important secondary injury mechanisms, as it plays an essential role in stimulating the appearance of a neurotoxic microenvironment after injury. However, further studies about the immune system and its relationship with CNS damage suggest that it is not completely pathological, as it has protective and regenerative properties [59–78].
### **3. Protective autoimmunity**
concentrations, leading to a reduction in excitotoxicity [44]. Evidence suggests that even though the immune system is considered to be completely pathological in nature, it can also
In normal conditions, inflammatory reaction leads the response to a pathological outcome, promoting damage and spinal cord degeneration. The severity of SCI determines the intensity of the inflammatory response and the glial reaction to SCI. Glial cells are well distributed within the CNS, and have the ability to proliferate and migrate to the site of injury. In response to injury there is a high glial secretion of cytokines and chemokines—such as IL‐1, IL‐6, and TNFα—that allows migration of peripheral immune cells [45]. TNFα stimulates the expression of adhesion molecules like endothelial intracellular adhesion molecule ‐1 (ICAM‐1) and vascular cell adhesion molecule‐1 (VCAM‐1) altering the blood‐brain barrier permeability and inducing a peripheral infiltration of immune cells to the CNS (**Figure 1**). Cell infiltration to the CNS is considered the principal factor for tissue disruption and sustained neural damage. The first peripheral cells to arrive at the site of injury are neutrophils [46]. These cells have the ability to phagocyte and clear debris, but they also secrete ROS and RNS, as well as other pathological proteolytic and oxidative enzymes like myeloperoxidase, producing greater tissue damage [47, 58]. At the same time, macrophages arise from resident microglia or from peripheral monocytes with the objective of removing cell debris and stimulating angiogenesis [48]. These cells play an essential role in the immune response, as they help in the activation and reclusion of the adaptive immune system cells. Microglia, astrocytes and dendritic cells may act as antigen‐presenting cells (APC), leading to T lymphocyte activation, proliferation, and infiltration to the site of injury [49]. Lymphocytes recognize the signal and proliferate creating large numbers of clones, specific to the antigen being presented by the APC. At 3 days post‐injury, there is an evident infiltration of T‐cells to the CNS, these cells determine the intensity and continue on modulating the immune response to trauma [50]. In SCI, the presence of T‐lymphocytes is considered detrimental as they secrete a Th1 cytokine profile including interferon gamma (IFNγ) [51, 63]. IFNγ is a pro‐inflammatory cytokine capable of inducing free radical production, increasing IL‐6, IL‐12, IL‐1β, and TNFα concentrations, and activating apoptotic‐signaling pathways [52–65]. These events represent the immune activity within the acute phases of SCI and if it is not well controlled it could turn into a chronic and degenerative
When inflammation is modulated, i.e., we control the intensity of inflammatory response, the type, the action, and the arrival of immune cells, we could expect a change in the final outcome. With this regard, there is a strong evidence suggesting that lymphocytes could also have favorable activity, as they are capable of synthesizing several growth factors, like BDNF, NT3, and nerve growth factor (NGF) [53–68]. These molecules are known for being capable of
To the date, it is clear that the immune response appearing after SCI could be pathological if it is not well controlled at the onset of SCI [55]. In a short period of time, this uncontrolled response leads to the extensive recognition of CNS peptides, proteins, lipids, or nucleic acids by the immune system [56]. This interaction between immune cells and CNS constituents like the MBP—promotes the activation of lymphocytes and thereby, the possible development
promoting a protective and regenerative environment for CNS disorders [54].
provide beneficial effects for SCI repair.
66 Recovery of Motor Function Following Spinal Cord Injury
immune response.
The reactive immune response against self‐constituents appearing after injury has been widely studied, as it can be an excellent target in the design of new therapies for SCI treatment. It was previously thought that the presence of immune activity was detrimental and it had to be suppressed [60–81]. However, more recent findings suggest that such response and the infiltration of peripheral immune cells to the site of injury is a phenomenon destined to protect and restore the CNS after trauma. The phenomenon capable of inducing this beneficial actions has been termed protective autoimmunity (PA) and is a mechanism in which the adaptive immune cells—especially lymphocytes that recognize self‐constituents and potentiate an autoreactive response—help maintain tissue integrity in SCI. Researchers have demonstrated that PA is genetically encoded, and it is a physiological phenomenon linked to the inflamma‐ tory activity in the CNS, capable of providing protection in several neurodegenerative disorders [61, 83]. The immune system plays an essential role in tissue restoration, angiogen‐ esis, and is capable of increasing functional recovery in CNS trauma [62]. However, in normal conditions the intense inflammatory response appearing after injury overshadows the beneficial effects of PA. For that reason, it was thought that boosting PA after injury could promote the beneficial instead of deleterious effects of the immune response to injury. In an attempt to test this hypothesis, researchers performed a passive transfer of T‐cells specific to MBP, and demonstrated that it reduced tissue damage and improved motor recovery in rats with SCI [63, 86]. These studies suggested that the delayed adaptive immune response to injury and the low concentration of autoreactive T‐cells are the reasons why PA is not evident, but a higher and earlier presence of these cells could potentiate the beneficial effects of PA over SCI.
Shortly after researchers envisioned that active immunization could elicit a higher proliferation and migration of autoreactive (antigen specific) T‐cells to the CNS increasing the action of PA. Therefore, immunization with natural CNS components could help activate this response [64]. In line with this motion, several studies were performed indicating that, immunization with MBP can modulate the immune response, potentiate PA, and provide neuroprotection to the injured tissue [65, 88, 89]. In spite of the encouraging results, immune modulation with neural constituents increases the risk of developing an autoimmune disease such as experimental autoimmune encephalomyelitis (in rodents; EAE) or multiple sclerosis (in humans; MS) [66, 91]. For that reason, neural constituents were studied with the objective of creating a peptide capable of stimulating PA and reducing the risk of developing an autoimmune disease [67, 93]. These experiments led to the creation of altered peptide ligands (APL), which are synthetic peptides with changes in specific amino acid residues, critical for T‐cell receptor (TCR) binding [68]. In the normal immune response, the MBP has agonist properties as it interacts with the TCR, leading to lymphocyte differentiation toward a Th1 phenotype [69, 96, 97]. With the objective of altering TCR recognition, a specific amino acid substitution makes APL become partial agonists or antagonists capable of deviating the immune response [70, 98]. That is why APL are able to cause lymphocyte anergy or their differentiation to an anti‐inflammatory Th2 phenotype. This way, APL can alter the natural response of the immune system after SCI, by being able to change the whole Th1 (pro‐inflammatory) cytokine profile toward a Th2 (anti‐ inflammatory) profile. Therefore, by altering the immune response, APL represent a good therapeutic approach for the treatment of SCI and other neurodegenerative diseases [71, 99].
### **3.1. Modulation of protective autoimmunity with the use of altered peptide ligands**
A safe and effective way of increasing PA is through immunization with APL [72]. This strategy allows a change in the interaction with the TCR from an agonist to a partial agonist switching the response toward a Th2 cytokine pattern [73]. To create APL, identification of the essential residues of MBP for an acquired immune response to self‐determinants was investigated. Evidence indicates that the amino acid sequence including the 87 to 99 residues is the most encephalitogenic portion of the MBP and that it is essential for TCR recognition [74, 102]. This amino acid sequence was fundamental for the creation of several APL and the evaluation of their effect over the immune response [75]. Altering the amino acid sequence by substituting each residue of the encephalitogenic region of the MBP with alanine, led to the discovery of this group of APL [76, 104]. While trying to identify an ideal peptide to promote PA, several APL derived from MBP encephalitogenic epitopes like G91, A96, or A97, along with A91 were tested as therapeutic strategies for CNS trauma. These peptides were capable of controlling the MBP peptide induced autoimmune reaction by altering the MBP specific T‐cell responses [77]. Also, these APL demonstrated to be able of providing significant protection by reducing neuronal loss [78]. More importantly, they were limited the extent of the immune secondary injury mechanisms by inducing changes in the cytokine secretion profile of T‐cells and enhanced the recovery of motor activity [79, 99, 103]. Studied APL showed different levels of neuroprotection, however, the conclusion of these conducted studies was that the APL A91 provided the best therapeutic effects without the risk of an autoimmune response [80].
Evidence has demonstrated that lysine at the position 91 is an essential residue of the MBP p87–99 for the development of a Th1 immune response. With this respect, it has been shown that when the amino acid 91 (lysine) is replaced with glycine (G91), the peptide which is non‐ encephalitogenic, regulates the proliferative response and modifies the cytokine secretion profile (toward a Th2 profile) of encephalitogenic MBP 87–99 reactive T‐cells [81, 99, 103, 104]. The substitution of lysine at the position 91 for the amino acid alanine led to the creation of the APL: A91. This APL counteracts the production of pro‐inflammatory cytokines, generating a microenvironment with anti‐inflammatory features [82, 103, 104]. A91 (amino acid sequence: VHFFANIVTPRTP) is a safe synthetic non‐encephalitogenic peptide, capable of inhibiting the development of autoimmune disease while maintaining neuroprotection [83, 104]. This APL (A91) has proven to be an effective TCR partial agonist capable of modulating the immune response after CNS injury, and increasing the beneficial effects of PA. Immunization with A91peptide down regulates Th1 activity and increases the levels of a Th2 cytokine pattern (IL‐ 4 and IL‐10) creating an anti‐inflammatory microenvironment [84, 105, 106].
### **3.2. The altered peptide ligand A91 as a potential treatment for spinal cord injury**
capable of stimulating PA and reducing the risk of developing an autoimmune disease [67, 93]. These experiments led to the creation of altered peptide ligands (APL), which are synthetic peptides with changes in specific amino acid residues, critical for T‐cell receptor (TCR) binding [68]. In the normal immune response, the MBP has agonist properties as it interacts with the TCR, leading to lymphocyte differentiation toward a Th1 phenotype [69, 96, 97]. With the objective of altering TCR recognition, a specific amino acid substitution makes APL become partial agonists or antagonists capable of deviating the immune response [70, 98]. That is why APL are able to cause lymphocyte anergy or their differentiation to an anti‐inflammatory Th2 phenotype. This way, APL can alter the natural response of the immune system after SCI, by being able to change the whole Th1 (pro‐inflammatory) cytokine profile toward a Th2 (anti‐ inflammatory) profile. Therefore, by altering the immune response, APL represent a good therapeutic approach for the treatment of SCI and other neurodegenerative diseases [71, 99].
68 Recovery of Motor Function Following Spinal Cord Injury
**3.1. Modulation of protective autoimmunity with the use of altered peptide ligands**
A safe and effective way of increasing PA is through immunization with APL [72]. This strategy allows a change in the interaction with the TCR from an agonist to a partial agonist switching the response toward a Th2 cytokine pattern [73]. To create APL, identification of the essential residues of MBP for an acquired immune response to self‐determinants was investigated. Evidence indicates that the amino acid sequence including the 87 to 99 residues is the most encephalitogenic portion of the MBP and that it is essential for TCR recognition [74, 102]. This amino acid sequence was fundamental for the creation of several APL and the evaluation of their effect over the immune response [75]. Altering the amino acid sequence by substituting each residue of the encephalitogenic region of the MBP with alanine, led to the discovery of this group of APL [76, 104]. While trying to identify an ideal peptide to promote PA, several APL derived from MBP encephalitogenic epitopes like G91, A96, or A97, along with A91 were tested as therapeutic strategies for CNS trauma. These peptides were capable of controlling the MBP peptide induced autoimmune reaction by altering the MBP specific T‐cell responses [77]. Also, these APL demonstrated to be able of providing significant protection by reducing neuronal loss [78]. More importantly, they were limited the extent of the immune secondary injury mechanisms by inducing changes in the cytokine secretion profile of T‐cells and enhanced the recovery of motor activity [79, 99, 103]. Studied APL showed different levels of neuroprotection, however, the conclusion of these conducted studies was that the APL A91 provided the best therapeutic effects without the risk of an autoimmune response [80].
Evidence has demonstrated that lysine at the position 91 is an essential residue of the MBP p87–99 for the development of a Th1 immune response. With this respect, it has been shown that when the amino acid 91 (lysine) is replaced with glycine (G91), the peptide which is non‐ encephalitogenic, regulates the proliferative response and modifies the cytokine secretion profile (toward a Th2 profile) of encephalitogenic MBP 87–99 reactive T‐cells [81, 99, 103, 104]. The substitution of lysine at the position 91 for the amino acid alanine led to the creation of the APL: A91. This APL counteracts the production of pro‐inflammatory cytokines, generating a microenvironment with anti‐inflammatory features [82, 103, 104]. A91 (amino acid sequence: To increase neuroprotection, A91 was designed to boost PA and to act directly over secondary mechanisms in SCI. Immunization with A91 has been tested as a subcutaneous injection, which has resulted to be an effective and minimally invasive route of administration. The use of this strategy in preclinical studies indicates that active immunization with A91 with a single dose of 150–200 μg improves neurological recovery. It is important to note that in order to avoid the risk of an autoimmune disease while maintaining neuroprotection, immunization with myelin‐associated antigens should be fully controlled. A study evaluating the dose and therapeutic window of A91 indicates that beneficial effect of this peptide lies between 10 minutes up to 72 hours after SCI [85]. Further studies also indicate that A91 could be considered as a prophylactic therapeutic vaccine since its administration before SCI could provide high levels of neuroprotection and motor recovery [86]. These results could be of relevant benefit as an approach to provide with prophylactic measures to patients sustaining invasive spinal surgery procedures.
On the other hand, with the objective of evaluating the effect of A91 immunization in the presence of the gold standard treatment for SCI (MP) another study was carried out. The results of this investigation indicated that when these two treatments were administered at the same time, the beneficial properties of A91 were abolished [87]. However, the extensive therapeutic time window of A91 enables immediate MP administration and immunization with A91 up to 48 hours later [88, 109]. This approach has given the possibility of rescuing the beneficial effects elicited by A91, as animals subjected to this combined strategy, present neuroprotection and a higher motor recovery. These results also allow envisioning the possible clinical application of this therapy with no risk of avoiding the therapeutic effects theoretically provided by MP. In the search of increasing the neuroprotective effect of A91, double immu‐ nization has also been evaluated [89]. However, unexpectedly a higher concentration of this peptide eliminates the beneficial aftermath of the therapy [90]. Studies in our laboratory have also demonstrated that A91 can be applied at different SCI stages and still be effective. Our investigation suggests that adequate immunization must be performed immediately after injury or during the acute phase [91, 109]. Moreover, studies including vaccination of A91 in chronic SCI are now being conducted, and results have demonstrated to be profitable (Un‐ published data).
APL were designed to specifically target the immune‐related secondary injury mechanisms in order to attain neuroprotection, promote regeneration, and thus improving motor and sensory recovery in SCI. In line with this, previous studies have shown that immunization with A91 peptide produces an adequate T‐cell proliferation characterized by a Th2 phenotype, where the production of IL‐4 and IL‐10 is increased [92, 110]. Additionally, A91 specific T‐cells are capable of producing BDNF, which could be linked directly with the functional recovery appearing after immunization [93]. This anti‐inflammatory and permissive microenvironment controls the inflammatory response elicited by SCI, reducing some of the main harmful phenomena developed by inflammation. For instance, it has been demonstrated that A91‐ immunization is capable of inhibiting LP, which is closely related to the action of immune system and it is one of the most aggressive phenomena related to SCI [94]. LP is present immediately after injury reaching its maximum peak at 4–5 h and has a second increase between 24 hours and 5 days [95]. With this regard, a study conducted to evaluate the impact of A91‐immunization on LP showed that A91 is able to reduce the concentration of ROS at the site of injury having a strong impact over the second peak of this phenomenon [96]. A further study indicated that A91‐immunization counteracts the production of nitric oxide (NO) and down regulates the expression of the gene encoding for nitric oxide synthase (iNOS) [97]. These are some of the beneficial mechanisms that explain, at least in part, the effect of this strategy on LP.
Apoptotic cell death is another of the main destructive phenomena triggered after SCI. This phenomenon is activated by inflammatory cytokines, free radicals, excitotoxic agents, and increased levels of intracellular calcium [98]. After SCI, neurological recovery depends mainly on the extent of neuronal loss and the functionality of the residual neural tissue. Numerous studies have shown that many neurons die as a consequence of apoptosis. Therefore, regulat‐ ing apoptotic cell death might play an important role in the neurological recovery following SCI [99, 116]. Recent investigations on the field found that immunizing with A91 decreases caspase‐3 activity and TNFα concentrations, reducing the number of apoptotic cells, which is directly correlated with functional improvement after injury [100].
Altogether, the aforementioned observations provide clear evidence on the mechanisms by which A91‐immunization exerts its beneficial effects. Besides reducing secondary injury mechanisms, A91 peptide has also proven to prevent tissue damage, as immunized animals presented a higher number of myelinated axons and survival of rubrospinal neurons com‐ pared to controls [101, 109, 110]. These results were consistent throughout several SCI preclinical studies. Also it was noted that motor recovery had a direct correlation with neuronal survival, myelin preservation, and apoptosis reduction in treated groups [102, 107, 109, 110, 114, 117, 118]. As a consequence of these encouraging results, we have envisioned the possi‐ bility of combining this strategy with others that have also shown beneficial effects [103, 117, unpublished data]. For that reason, A91‐immunization was administered along with gluta‐ thione monoethyl ester (GSHE), which is an anti‐oxidant capable of accelerating the immune response and providing neuroprotection [104, 117]. The results of this study showed that after a contusive or a compressive SCI, this combination induced better motor recovery, higher number of myelinated axons, and better rubrospinal neuron survival than immunization alone. These results open an interesting scenery for clinical studies.
Finally, in order to consider A91‐immunization for being used at clinical settings, it is of relevance to contemplate vaccine safety. With this regard, immunizing with A91 shows no signs of autoimmune disease development, possibly due to its low affinity to major histocom‐ patibility complex (MHC) molecules [105]. Furthermore to evaluate vaccine safety, the clinical appearance of treated animals was assessed, and no weight variation or other clinical data of EAE in immunized animals was detected [106]. A91 has proven to be effective in several studies conducted at different time points showing the stability of the vaccine in promoting recovery. Preclinical results of studies evaluating vaccine efficacy indicate that this therapy could be possibly applied to SCI patients and improve their recovery and quality of life.
### **4. Concluding statement**
peptide produces an adequate T‐cell proliferation characterized by a Th2 phenotype, where the production of IL‐4 and IL‐10 is increased [92, 110]. Additionally, A91 specific T‐cells are capable of producing BDNF, which could be linked directly with the functional recovery appearing after immunization [93]. This anti‐inflammatory and permissive microenvironment controls the inflammatory response elicited by SCI, reducing some of the main harmful phenomena developed by inflammation. For instance, it has been demonstrated that A91‐ immunization is capable of inhibiting LP, which is closely related to the action of immune system and it is one of the most aggressive phenomena related to SCI [94]. LP is present immediately after injury reaching its maximum peak at 4–5 h and has a second increase between 24 hours and 5 days [95]. With this regard, a study conducted to evaluate the impact of A91‐immunization on LP showed that A91 is able to reduce the concentration of ROS at the site of injury having a strong impact over the second peak of this phenomenon [96]. A further study indicated that A91‐immunization counteracts the production of nitric oxide (NO) and down regulates the expression of the gene encoding for nitric oxide synthase (iNOS) [97]. These are some of the beneficial mechanisms that explain, at least in part, the effect of this strategy
Apoptotic cell death is another of the main destructive phenomena triggered after SCI. This phenomenon is activated by inflammatory cytokines, free radicals, excitotoxic agents, and increased levels of intracellular calcium [98]. After SCI, neurological recovery depends mainly on the extent of neuronal loss and the functionality of the residual neural tissue. Numerous studies have shown that many neurons die as a consequence of apoptosis. Therefore, regulat‐ ing apoptotic cell death might play an important role in the neurological recovery following SCI [99, 116]. Recent investigations on the field found that immunizing with A91 decreases caspase‐3 activity and TNFα concentrations, reducing the number of apoptotic cells, which is
Altogether, the aforementioned observations provide clear evidence on the mechanisms by which A91‐immunization exerts its beneficial effects. Besides reducing secondary injury mechanisms, A91 peptide has also proven to prevent tissue damage, as immunized animals presented a higher number of myelinated axons and survival of rubrospinal neurons com‐ pared to controls [101, 109, 110]. These results were consistent throughout several SCI preclinical studies. Also it was noted that motor recovery had a direct correlation with neuronal survival, myelin preservation, and apoptosis reduction in treated groups [102, 107, 109, 110, 114, 117, 118]. As a consequence of these encouraging results, we have envisioned the possi‐ bility of combining this strategy with others that have also shown beneficial effects [103, 117, unpublished data]. For that reason, A91‐immunization was administered along with gluta‐ thione monoethyl ester (GSHE), which is an anti‐oxidant capable of accelerating the immune response and providing neuroprotection [104, 117]. The results of this study showed that after a contusive or a compressive SCI, this combination induced better motor recovery, higher number of myelinated axons, and better rubrospinal neuron survival than immunization
Finally, in order to consider A91‐immunization for being used at clinical settings, it is of relevance to contemplate vaccine safety. With this regard, immunizing with A91 shows no
directly correlated with functional improvement after injury [100].
alone. These results open an interesting scenery for clinical studies.
on LP.
70 Recovery of Motor Function Following Spinal Cord Injury
Injury to the spinal cord stimulates the appearance of innate and adaptive immune responses, which could participate in either the pathogenesis or healing responses to trauma. The immune system should not be suppressed; instead, it must be modulated to attain its beneficial effect. That is why the use of immunosuppressant drugs like MP in the clinical field no longer seems a rational treatment. As a physiological hemostatic self‐defense mechanism, PA is an essential mechanism to be considered for the pathophysiology and treatment of SCI. Boosting PA with the use of APL is needed in order to increase the functional recovery in the immune related neurodegenerative diseases. In SCI, immunization with the APL A91 has proven to reduce part of the immune‐related secondary injury mechanisms without the risk of developing an autoimmune response. Preclinical results suggest that this therapy could be an effective treatment for SCI recovery, as it is closely related to a higher motor improvement, which is the most evident deficit in SCI patients. However, further studies related to the use of APL in SCI are needed to translate this therapy to the clinical field. For instance, we have to ensure that immunization with this peptide does not cause any side effect (i.e. hypersensitivity or autoimmunity). Additionally, further experiments should be performed in order to find out the best adjuvant to be used in humans, even the investigation should be directed to elucidate if the use of adjuvants is really necessary. Finally, the dosing of the peptide as well as the schedule of administration at clinical settings should also be investigated.
### **Author details**
Paola Suárez‐Meade<sup>1</sup> and Antonio Ibarra1,2\*
\*Address all correspondence to: iantonio65@yahoo.com; jose.ibarra@anahuac.mx
1 Faculty of Health Sciences, Anáhuac University, Universidad Anáhuac Av. Huixquilucan Edo. Mexico, Mexico
2 CAMINA Research Project A.C. Mexico City, Mexico
### **References**
[14] Geisler FH, Dorsey FC, Coleman WP. Past and current clinical studies with GM‐1 ganglioside in acute spinal cord injury. Ann Emerg Med. 1993; 22 (6): 104–1047.
**References**
23.
www.nscisc.uab.edu
72 Recovery of Motor Function Following Spinal Cord Injury
tions. Spinal Cord. 2012; 50: 365–372.
symptoms and signs. Prog Brain Res. 2006; 152: 1–8.
parameters. J Trauma Treat. 2014; 3 (4): 1–5.
Trial. J Neurosurg. 1998; 89 (5): 699–706.
of care. J Neurosurg. 2000; (Spine 1) 93: 1–7.
Perspect Dev Neurobiol. 1996; 3 (2): 93–100.
for old drugs. N Engl J Med. 2005; 353: 1711–1723.
processes. J Neurotrauma. 1992; 9 (Suppl 2): S507–516.
Neurosurg. 2014; 61 (1): 32–35.
Public Health. 2015; 22 (1): 9–18.
1990; 332 (20): 1405–1411.
[1] National Spinal Cord Injury Statistical Center, University of Alabama at Birmingham. 2015. Annual Statistical Report. Spinal cord injury facts and figures at glance. https://
[2] DeVivo MJ. Epidemiology of traumatic spinal cord injury: trends and future implica‐
[3] Chiu WT, Lin HC, Lam C, Chu SF, Chiang YH, et al. Epidemiology of traumatic spinal cord injury: comparisons between developed and developing countries. Asia Pac J
[4] Karlsson AK. Autonomic dysfunction in spinal cord injury: clinical presentation of
[5] Ibarra A, Rios‐Hoyo A, Suarez‐Meade P, Malagon E, Colin‐Rodriguez A. Influence of the level, severity and phase of spinal cord injury on hematological and biochemical
[6] Hagen EM. Acute complications of spinal cord injuries. World J Orthop. 2015; 6 (1): 17–
[7] Bracken MB, Shepard MJ, Collins WF, Holford T, Young W, et al. A randomized, controlled trial of methylprednisolone or naloxone in the treatment of acute spinal‐cord injury. Results of the Second National Acute Spinal Cord Injury Study. N Engl J Med.
[8] Bracken M, Shepard MJ, Holford T, Leo‐Summers L, Aldrich F, et al. Methylpredniso‐ lone or tirilazad mesylate administration after acute spinal cord injury: 1‐year follow up. Results of the Third National Acute Spinal Cord Injury Randomized controlled
[9] Hulbert J. Methylprednisolone for the treatment of acute spinal cord injury: point. Clin
[10] Hulbert J. Methylprednisolone for acute spinal cord injury: an inappropriate standard
[11] Rhen T, Cidlowski JA. Anti‐inflammatory action of glucocorticoids—new mechanisms
[12] Ferrari G, Greene LA. Prevention of neuronal apoptotic death by neurotrophic agents and Ganglioside GM1: insights and speculations regarding a common mechanism.
[13] Skaper SD, Leon A. Monosialogangliosides, neuroprotection, and neuronal repair
[44] Ibarra A, Diaz-Ruiz A. Inhibition of lipid peroxidation by cyclosporin after spinal cord injury in rats. Humana Press. 2003: 283–298.
[28] Rowland J, Hawryluk G, Kwon B, Fehlings M. Current status of acute spinal cord injury pathophysiology and emerging therapies: promise on the horizon. Neurosurg Focus.
[29] Aidemise Oyinbo C. Secondary injury mechanisms in traumatic spinal cord injury: a
[30] Xing C, Arai K, Lo EH, Hommel M. Pathophysiologic cascades in ischemic stroke. Int
[31] Mehta A, Prabhakar M, Kumar P, Deshmukh R, Sharma PL. Excitotoxicity: bridge to various triggers in neurodegenerative disorders. Eur J Pharmacol. 2013; 698: 6–18. [32] Sattler R, Tymianski M. Molecular mechanisms of glutamate receptor‐mediated
[33] Matute C, Domercq M, Sanchez‐Gomez MV. Glutamate‐mediated glial injury: mech‐
[34] Park E, Velumian A, Fehlings MG. The role of excitotoxicity in secondary mechanisms of spinal cord injury: a review with an emphasis on the implications for white matter
[35] David S, Kroner A. Repertoire of microglial and macrophage responses after spinal
[36] Azbill RD, Mu X, Bruce‐Keller A, Mattson MP, Springer JE. Impaired mitochondrial function, oxidative stress and altered antioxidant enzime activities following traumatic
[37] Cornelius C, Crupo R, Calabrese V, Graziano A, Milone P, et al. Traumatic brain injury: oxidative stress and neuroprotection. Antioxid Redox Signal. 2013; 19 (8): 836–853.
[38] Visavadiya PN, Patel SP, Venrooyen JL, Sullivan P, Rabchevsky A. Cellular and subcellular oxidative stress parameters following severe spinal cord injury. Redox Biol.
[39] Nathan C, Cunningham‐Brussel A. Beyong oxidative stress: an immunologist's guide
[40] Hall ED, Wang JA, Bosken JM, Singh IN. Lipid peroxidation in brain or spinal cord
[41] Gilgun‐Sherki Y, Rosenbaum Z, Melamed E, Offen D. Antioxidant therapy in acute central nervous system injury: current state. Pharmacol Rev. 2002; 54: 271–284.
[42] O'Connel KM, Littleton‐Kearney MT. The role of free radicals in traumatic brain injury.
[43] Farmer E, Mueller M. ROS‐mediated lipid peroxidation and RES‐activated signaling.
to reactive oxygen species. Net Rev Immunol. 2013; 13 (5): 349–361.
mitochondria after injury. J Bioenerg Biomembr. 2015. 48 (2): 169–174.
nugget of this multiply cascade. Acta Neurobiol Exp. 2011; 71: 281–299.
excitotoxic neuronal cell death. Mol Neurobiol. 2001; 24: 107–129.
anisms and clinical importance. Glia. 2006; 53: 212–224.
degeneration. J. Neurotrauma. 2004; 21(6): 754–774.
cord injury. Nat Rev Neurosci. 2011; 12: 388–399.
spinal cord injury. Brain Res. 1997; 765: 283–290.
Biol Res Nurs. 2012; 15 (3): 253–263.
Ann Rev Plant Biol. 2013; 64: 429–450.
2008; 25 (5): E2.
2016; 8: 59–67.
J Stroke. 2012; 7: 378–385.
74 Recovery of Motor Function Following Spinal Cord Injury
[72] Schwab JM, Zhang Y, Kopp M, Brommer B, Popovich PG. The paradox of chronic neuroinflammation, systemic immune suppression, autoimmunity after traumatic chronic spinal cord injury. Exp Neurol. 2014; 258: 121–129.
[58] Taoka Y, Okajima K, Uchiba M, Murakami K, Kushimoto S, Johno M. Role of neutro‐ phils in spinal cord injury in the rat. Neuroscience. 1997; 79 (4): 1177–1182.
[59] Greenhalgh A, David S. Differences in the phagocytic response of microglia and pripheral macrophages after spinal cord injury and its effects on cell death. J Neurosci.
[60] Popovich PG, Jones TB. Manipulating neuroinflammatory reactions in the injured
[61] Triveldi A, Olivas A, Noble‐Haeusslein L. Inflammation and spinal cord injury: infiltrating leukocytes as determinants of injury and repair processes. Clin Neurosci
[62] Hu JG, Shi LL, Chen YJ, Xie XM, Zhang N, et al. Differential effects of myelin basic protein‐activated Th1 and Th2 cells on the local immune microenvironment of injured
[63] Shaked I, Tchiresh D, Gersner R, Meiri G, Mordechai S, et al. Protective autoimmunity: interferon‐γ enables microglia to remove glutamate without evoking inflammatory
[64] Mir M, Tolosa L, Asensio V, Llado J, Olmos G. Complementary roles of tumor necrosis factor alpha and interferón gamma in inducible microglial nitric oxide generation. J
[65] Esposito E, Cuzzocrea S. TNF‐alpha as a therapeutic target in inflammatory diseases, ichemia‐reperfusion injury and trauma. Curr Med Chem. 2009; 16 (24): 3152–3167.
[66] Stadelmann C, Kerschensteriner M, Misgeld T, Bruck W, Hohlfeld R, et al. BDNF and gp145trkB in multiple sclerosis brain lesions: neuroprotective interactions between
[67] Chen Q, Smith G, Shine D. Immune activation is required for NT‐3‐induced azonal plasticity in chronic spinal cord injury. Exp Neurol. 1008; 209 (2): 497–509.
[68] Stantambrogio L, Benedetti M, Chao M, Muzaffar R, Kulig K, et al. Nerve growth factor
[69] Suarez‐Meade P, Carvajal HG, Yasuhara T, Taijiri N, Date I, et al. Regenerative medicine for central nervous system disorders: role of therapeutic molecules in stem
[70] Saltzman JW, Battaglino RA, Stott HL, Morse LR. Neurotoxic or neuroprotective? Current controversies in SCI‐induced autoimmunity. Curr Phys Med Rehabil Rep.
[71] Ankeny DP, Popovich PG. Mechanisms and implications of adaptive immune respons‐
es after traumatic spinal cord injury. Neuroscience. 2009; 158: 1112–1121.
spinal cord: back to basics. Trends Pharmacol Sci. 2003; 24 (1): 13–17.
2014; 34 (18): 6316–6322.
76 Recovery of Motor Function Following Spinal Cord Injury
Res. 2006; 6: 283–292.
spinal cord. Exp Neurol. 2016; 277: 190–201.
mediators. J Neurochem. 2005; 92: 997–1009.
Neuroimmunol. 2008; 204 (1–4): 101–109.
immune and neuronal cells. Brain. 2002; 125: 75–85.
cell therapy. Brain Circul. 2015; 1 (2): 125–132.
2013; 1: 174–177.
production by lymphocytes. J Immunol. 1994; 153: 4488–4495.
[99] Nel AE, Slaughter N. T‐Cell activaction through the antigen receptor. Part 2: role of signaling cascades in T‐cell differentiation, anergy, immune senescence, and develop‐ ment of immunotherapy. J Allergy Clin Immunol. 2002; 109 (6): 901–915.
[86] Kipnis J, Mizrahi T, Hauben E, Shaked I, Shevach E. Neuroprotective autoimmunity: naturally occurring CD4+CD25+ regulatory T cells suppress the ability to withstand
[87] Schwartz M, Raposo C. Protective autoimmunity: a unifying model for the immune
[88] Barouch R, Schwartz M. Autoreactive T cells induce neurotrophin production by immune and neural cells in injured rat optic nerve: implications for protective auto‐
[89] Aharoni R, Arnon R, Eilam R. Neurogenesis and neuroprotection induced by periph‐ eral immunomodulatory treatment of experimental autoimmune encephalomyelitis. J
[90] Sakai K, Sinha A, Mitchel D, Zamvil S, Rothbard J, et al. Involvement of distinct murine T‐cell receptors in the autoimmune encephalitogenic response to nested epitopes of
[91] Sakai K, Zamvil S, Mitchel DJ, Lim M, Rothbard J, et al. Characterization of a major encephalitogenic T cell epitope in SJL/J mice with synthetic oligopeptides of myelin
[92] Nicholson L, Greer J, Sobel R, Lees M, Kuchroo V. An altered peptide ligand mediates immune deviation and prevents autoimmune encephalomyelitis. Immunity. 1995; 3:
[93] Hauben E, Agranov E, Gothlif A, Nevo U, Cohen A, et al. Posttraumatic therapeutic vaccination with modified myelin self‐antigen prevents complete paralysis while
[94] Winghagen A, Scholz C, Hollsberg P, Fukaura H, Sette A, et al. Modulation of cytokine patters of human autoreactive T cell cones by a single amino acid substitution of their
[95] Jameson S, Bevan M. T cell receptor antagonists and partial agonists. Immunity. 1995;
[96] Nel A. T‐cell activation through the antigen receptor. Part 1: signaling components, signaling pathways, and signal integration at the T‐cell antigen receptor synapse. J
[97] Evavold BD, Allen PM. Sparation of IL‐4 production from Th cell proliferation by an
[98] Hauben E, Ibarra A, Mizrahi T, Barouch R, Agranov E, et al. Vaccination with a Nogo‐ A‐derived peptide after incomplete spinal cord injury promotes recovery via a T‐cell mediated neuroprotective response: comparison with other myelin antigens. PNAS.
myelin basic protein. Proc Natl Acad Sci. 1988; 85: 8608–8612.
avoiding autoimmune disease. J Clin Invest. 2001; 108: 591–599.
basic protein. J Neuroimmunol. 1988; 19: 21–32.
peptide ligand. Immunity. 1995; 2: 373–380.
Allergy Clin Immunol. 2002; 109 (5): 758–770.
2001; 98 (26): 15173–15178.
altered T cell receptor ligand. Science. 1991; 252: 1308–1310.
injury to the central nervous system. PNAS. 2002; 99 (24): 15620–15625.
network involved in CNS repair. Neuroscientist. 2014; 20 (4): 343–358.
immunity. FASEB J. 2002; 16: 1304–1306.
Neurosci. 2005; 25: 8217–8228.
78 Recovery of Motor Function Following Spinal Cord Injury
397–405.
2 (1): 1–11.
## **Role of the Neuroinflammation in the Degree of Spinal Cord Injury: New Therapeutic Strategies**
Irene Paterniti, Emanuela Esposito and
Salvatore Cuzzocrea
[111] Christie SD, Corneau B, Myers T, Sadi D, Purdy M, et al. Duration of lipid peroxidation after acute spinal cord injury in rats and the effect of methylprednisolone. J Neurosurg.
[112] García E, Silva‐García R, Mestre H, Flores N, Martiñón S. Immunization with A91 peptide or Copolymer‐1 reduces the production of nitric oxide and inducible nitric oxide synthase gene expression after spinal cord injury. J Neurosci Res. 2012; 90: 656–
[113] Lou J, Lenke LG, Ludwig FJ, O'Brien MF. Apoptosis as a mechanism of neuronal cell death following acute experimental spinal cord injury. Spinal Cord*.* 1998; 36 (10): 683–
[114] Lee YJ, Choi SY, Oh TH, Yune TY. Estradiol inhibits apoptotic cell death of oligoden‐ drocytes by inhibiting RhoA‐JNK3 activation after spinal cord injury. Endocrinology.
[115] Zhang N, Xu SJ, Chen WS, Yin Y, Wu YP. Inflammation and apoptosis in spinal cord
[116] Rodríguez‐Barrera R, Fernández‐Presas A, García E, Flores‐Romero A, Martiñón S, et al. Immunization with a neural‐derived peptide protects the spinal cord from apoptosis
[117] Hauben E, Ibarra A, Mizrahi T, Barouch R, Agranov E, et al. Vaccination with a Nogo‐ A‐derived peptide after incomplete spinal‐cord injury promotes recovery via a T‐cell mediated neuroprotective response: comparison with other myelin antigens. PNAS.
2008; 25 (5): E5.
80 Recovery of Motor Function Following Spinal Cord Injury
2012; 153 (8): 3815–3827.
2001; 98 (26): 15173–15178.
injury. Indian J Med Res. 2012; 135 (3): 287–296.
after traumatic injury. Bio Med Res. 2013: 2013: 1–8.
663.
690.
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/63222
### **Abstract**
A case of spinal cord injury (SCI) is defined as the occurrence of an acute traumatic lesion of neural elements in the spinal canal (spinal cord and cauda equina), resulting in temporary or permanent sensory and/or motor deficit. Most studies on traumatic SCI show a bimodal age distribution, with a first peak in young adulthood and a second peak in older adults. Spinal cord trauma activates a cascade of events that exacerbates the damage such as activation of inflammatory process that determinates cytokine and chemokine production and that generates reduction in functional recovery resulting in necrosis or apoptosis of neurons. However, the precise mechanism of SCI-induced inflammatory response remains not fully understood at present. Current strategy to treat damage to the spinal cord is limited, only the treatment with methylpredniso‐ lone (MP), if administered in excessive dose during the acute phase of the damage, could ameliorate patients with severe SCI. However, associated to the beneficial effects, there are growing evidence that high-dose of MP is correlated to increased risk of infec‐ tions, pneumonia and gastrointestinal bleeding. Therefore, there is a necessity to develop new therapies to treat SCI; one of these is to selectively reduce inflammation that possess unique role in the processes of injury and recovery.
**Keywords:** Immune response, oxidative stress, spinal cord injury, inflammation, neu‐ roprotection
### **1. Inflammation response after spinal cord injury**
Spinal cordinjury(SCI)isdamagecausedtothespinal cordthat compromisethemajorfunctions of the spinal cord and remains actually the most important cause of mortality in the society. In
© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
addition to its cost to the individual, physically as well as the health care system and society financially, SCI hasprofoundpsychosocial effects that aredevastating forpatients,families and friends.SCIusuallybeginswithasudden,traumaticblowtothespinethatfracturesordislocates vertebrae; long-term mechanical compression of the spinal cord gradually causes various pathologic changes in neural tissue. The pathophysiology of SCI comprises both primary and secondarymechanismsofinjury;the"primaryinjury"referstotheforcesthatimparttheprimary mechanical insult to the spinal cord, which in its mildest form causes a cord concussion with brief transient neurologic deficits and in its most severe form causes complete and permanent paralysis. The primary damage to tissue is followed by a second phase of tissue degeneration that might occur over weeks or even months and further generate progressive destruction of the tissue surrounding the necrotic core that expands from the injury "epicenter" and is known as secondary injury,thatis apersistence of some crucial events ofthe acutephase such as edema and apoptotic cell death as well as generation of oxidative stress, activation of immune system response and inflammation process. In particular we focus our attention on some mecha‐ nisms that occur after spinal cordinjury suchas onthe involvement ofthe inflammationprocess and immune system response.
Inflammation is a physiological process that removes damaged stimuli and initiates the healing process; however, if it persists and if it is over-activated, then the inflammation becomes devastating; inflammation is a key element in the pathophysiology of some disorders such as chronic pain, neurodegenerative pathology, trauma and spinal cord injury [1–5]. Principal proinflammatory markers that are both cellular components, such as neutrophils, macrophages and T cells, and non-cellular components, such as cytokines, prostaglandins and complement, have been found in spinal cord tissue that received a mechanical insult. Following spinal cord trauma, the site of injury is penetrated by neutrophils, that determinate release of cytokines that may progressive damage local tissue and recruit other inflammatory cells [6]; moreover, monocytes/macrophages are released and locally activated in microglia, that subsequently invade the injured tissue [7]. The pro-inflammatory cytokines that are produced at the site of injury, such as tumor necrosis factor (TNF-α), interleukins and interferons, mediate the inflammatory response and can generate further tissue damage [8,9]. Furthermore, cytokines can induce the expression of cyclo-oxygenase (COX) 2 and thus promote the breakdown of arachidonic acid into pro-inflammatory prostanoids (prostaglandins, prostacyclin and thromboxanes) that mediate vascular permeability/resistance and platelet aggregation/ adherance [10,11]. The involvement of the cyclo-oxygenases in the generation of these inflammatory mediators represents a potential target for intervention, because inhibitors of these enzymes are in widespread clinical use. Additionally, TNF-α contributes to the tissue injury induced by neutrophils by directly activating them [12,13] as well as by increasing the expression of such molecules as ICAM-1 and E-selectin, which cause the activated neutrophils to adhere to the surface of the endothelial cells; it has also been shown that the inhibition of neutrophil adhesion to the endothelial cell surface markedly reduces the severity of the SCI induced by compressive trauma [14]. These observations indicate that the interaction of activated neutrophils with the surface of the endothelial cells is another important mediator in the secondary tissue damage of the spinal cord.
### **1.1. Cytokine responses to inflammation**
addition to its cost to the individual, physically as well as the health care system and society financially, SCI hasprofoundpsychosocial effects that aredevastating forpatients,families and friends.SCIusuallybeginswithasudden,traumaticblowtothespinethatfracturesordislocates vertebrae; long-term mechanical compression of the spinal cord gradually causes various pathologic changes in neural tissue. The pathophysiology of SCI comprises both primary and secondarymechanismsofinjury;the"primaryinjury"referstotheforcesthatimparttheprimary mechanical insult to the spinal cord, which in its mildest form causes a cord concussion with brief transient neurologic deficits and in its most severe form causes complete and permanent paralysis. The primary damage to tissue is followed by a second phase of tissue degeneration that might occur over weeks or even months and further generate progressive destruction of the tissue surrounding the necrotic core that expands from the injury "epicenter" and is known as secondary injury,thatis apersistence of some crucial events ofthe acutephase such as edema and apoptotic cell death as well as generation of oxidative stress, activation of immune system response and inflammation process. In particular we focus our attention on some mecha‐ nisms that occur after spinal cordinjury suchas onthe involvement ofthe inflammationprocess
Inflammation is a physiological process that removes damaged stimuli and initiates the healing process; however, if it persists and if it is over-activated, then the inflammation becomes devastating; inflammation is a key element in the pathophysiology of some disorders such as chronic pain, neurodegenerative pathology, trauma and spinal cord injury [1–5]. Principal proinflammatory markers that are both cellular components, such as neutrophils, macrophages and T cells, and non-cellular components, such as cytokines, prostaglandins and complement, have been found in spinal cord tissue that received a mechanical insult. Following spinal cord trauma, the site of injury is penetrated by neutrophils, that determinate release of cytokines that may progressive damage local tissue and recruit other inflammatory cells [6]; moreover, monocytes/macrophages are released and locally activated in microglia, that subsequently invade the injured tissue [7]. The pro-inflammatory cytokines that are produced at the site of injury, such as tumor necrosis factor (TNF-α), interleukins and interferons, mediate the inflammatory response and can generate further tissue damage [8,9]. Furthermore, cytokines can induce the expression of cyclo-oxygenase (COX) 2 and thus promote the breakdown of arachidonic acid into pro-inflammatory prostanoids (prostaglandins, prostacyclin and thromboxanes) that mediate vascular permeability/resistance and platelet aggregation/ adherance [10,11]. The involvement of the cyclo-oxygenases in the generation of these inflammatory mediators represents a potential target for intervention, because inhibitors of these enzymes are in widespread clinical use. Additionally, TNF-α contributes to the tissue injury induced by neutrophils by directly activating them [12,13] as well as by increasing the expression of such molecules as ICAM-1 and E-selectin, which cause the activated neutrophils to adhere to the surface of the endothelial cells; it has also been shown that the inhibition of neutrophil adhesion to the endothelial cell surface markedly reduces the severity of the SCI induced by compressive trauma [14]. These observations indicate that the interaction of activated neutrophils with the surface of the endothelial cells is another important mediator
and immune system response.
82 Recovery of Motor Function Following Spinal Cord Injury
in the secondary tissue damage of the spinal cord.
Cytokines are small and non-structural proteins with no amino acid sequence motif, their biological activities allow us in turn to group them into different classes: exit 18 cytokines called interleukin (IL), some of these promote inflammation and are named pro-inflammatory cytokines such as IL1β and IL1α, IL6, IL8 and TNF-α, whereas other cytokines suppress the activity of pro-inflammatory cytokines and are called anti-inflammatory cytokines such as IL-4, IL-10, TGFβ. The hypothesis that some cytokine functions primarily induce inflammation while others suppress inflammation is essential to cytokine biology and also to clinical medicine.
Cytokines are secreted by a variety of immune cells such as T-lymphocytes and macrophages, as well as b non-immune cells such as fibroblasts; the physiological effects mediated by cytokines comprise the stimulation or inhibition of cell growth, cytotoxicity/apoptosis, antiviral activity and inflammatory responses. The main function of cytokines is the regulation of T-cell differentiation from undifferentiated cells to T-helper 1 and 2, regulatory T cells, and T-helper 17 cells [15]. These regulatory proteins include ILs, interferons (IFNs) and TNFs. Many of these cytokines have already been shown to be produced by neurons or glia in central nervous system (CNS) disorders in which they are notably increased.
The cytokine class of inflammatory mediators is secreted by microglia and astrocytes and their production is increased in inflammatory states; moreover they act by modulating the intensity and duration of the immune response. Pro-inflammatory cytokines and chemokines upregulate microbicidal activity of neutrophils, and they can be considered as additional immunomodulatory agents to treat serious or refractory infections in humans.
Through cytokines IL-1 the immune response is initiated, having a crucial role in the onset and expansion of a complex hormonal and cellular inflammatory cascade; the IL-1 family of cytokines includes IL-1α and IL-1β, which generate cell activation upon binding with specific membrane receptors and has been documented that IL-1 plays a role in neuronal degeneration. In astrocytes, IL-1 induces IL-6 production, stimulates iNOS activity [16], enhances neuronal acetylcholinesterase activity, microglial activation and additional IL-1 production, and astrocyte activation.
Another important pro-inflammatory cytokine is the IL-6, a multifunctional cytokine that plays an important role in host defence [17] and possesses main effects during the inflammatory response [18]. IL-6 is associated to the family of neuropoietin cytokine and it possesses direct and indirect neurotrophic effects on neurons [19]; moreover, IL-6 promotes astrogliosis [20], activates microglia [21], and stimulates the release of acute phase molecules.
### **1.2. Inflammatory mediator: crucial role of TNF-α**
Through all the cytokines involved in the secondary damage, TNF-α plays a crucial role; in fact it releases shortly after injury, it can accumulate rapidly at the site of injury, and it is produced by a number of different cell populations, such as neutrophils, macrophages and microglia, astrocytes and T cells [22]. Several cell types are able to produce TNF-α, including macrophages after its activation, dendritic cells, monocytes, NK cells, CD4+ T cells, CD8+ T cells, microglia and astrocytes. Macrophages/monocytes are able to produce TNF-α in the acute phase of inflammation and this cytokines drives several range of signalling events within cells, leading to necrosis or apoptosis.
Several biological functions are ascribed to the TNF-α and for this reason the mechanism of action is somewhat complex; although it inhibits the growth of tumor cells and it has an enhancing effect on the proliferation of normal cells [23]. TNF-α takes part in septic shock, autoimmunity, and inflammatory disorders. The major role of TNF-α is explicated as mediator in resistance against infections; moreover, it was postulated that TNF plays a pathological role in a number of autoimmune pathology such as graft vs host rejection or rheumatoid arthritis. Moreover, TNF-α possesses potent pro-inflammatory effects that are associated to its capacity to generate endothelial cell adhesion molecules and subsequently support neutrophil adher‐ ence to vascular endothelium. Neutrophils are exquisite targets of TNF-α that is under certain conditions it strengthens their expression of adhesion molecules induces their degranulation and successive release of lysosomal enzymes, causing the production of highly reactive oxygen species. TNF-α induces the migration of neutrophils mediating the production of chemotactic factors, including IL-8; this testifies cytokine networking involvement in inflammatory cell recruitment and an active role in inflammation.
TNF-α works by binding and clustering high-affinity receptors that are present in a great numbers on most cell membranes [24], the ligand/receptor complex is easily internalised via clathrin-coated pits and ends up in secondary lysozymes where it is degraded. Interestingly, the binding of TNF-α to the 75 kDa TNFR-2 is not sufficient to reach cytotoxicity, but rather binding to the 55 kDa TNFR-1 is sufficient to reach TNF-α mediated cell killing. TNF-α exerts its effects by activating several secondary proteins that provoke a variety of responses within the cell such as activation of gene transcription and/or production of reactive oxygen or nitrogen radicals (e.g., NO). Activated proteins include Gprotein, transcription factors such as NF-κB and AP-1 and serine and cysteine proteases, known as caspases. Many members of the TNF receptor superfamily have intracellular "death domains" which represent protein interaction domains each consisting of 65–80 amino acids; these proteins participate in TNFα mediated apoptosis process; many evidence demonstrated that TNF–TNFR interactions are implicated in the pathogenesis of CNS disorders such as EAE and MS. These interactions are able to monitor the disease outcome by modifying immune response and the interactions between CNS-resident cells and effector immune cells in the CNS.
However, recent studies showed a dual nature for TNF-α that it can be not only neurotoxic but also neuroprotective; a study conducted with transgenic mice for TNF-α receptors demonstrated that the mice lacking TNF-α showed more tissue loss and functional deficits compared to wild-type mice, implying that TNF-α mediated a neuroprotective effect [25]. The beneficial or deleterious effects of TNF-α dependent when it is being released and on cellular population that is acting on, the conflicting actions of TNF-α described above reflects a growing view of inflammation as a "dual-edged sword" having neurotoxic and neuroprotective properties [26].
Thus, comprehension of their profile, kinetics of expression and interactions between TNF-α ligands and their TNFRs on different CNS residents and infiltrating immune cells would aid to better design strategies to control neuroinflammation and CNS autoimmunity. Blockers of TNF-α have been acknowledged for human use in treating TNF-linked autoimmune and inflammatory disorders. Pathways downstream of receptor ligation supply critical points for interjection for planning new therapeutic strategies.
### **1.3. Microglia activation**
cells, microglia and astrocytes. Macrophages/monocytes are able to produce TNF-α in the acute phase of inflammation and this cytokines drives several range of signalling events within
Several biological functions are ascribed to the TNF-α and for this reason the mechanism of action is somewhat complex; although it inhibits the growth of tumor cells and it has an enhancing effect on the proliferation of normal cells [23]. TNF-α takes part in septic shock, autoimmunity, and inflammatory disorders. The major role of TNF-α is explicated as mediator in resistance against infections; moreover, it was postulated that TNF plays a pathological role in a number of autoimmune pathology such as graft vs host rejection or rheumatoid arthritis. Moreover, TNF-α possesses potent pro-inflammatory effects that are associated to its capacity to generate endothelial cell adhesion molecules and subsequently support neutrophil adher‐ ence to vascular endothelium. Neutrophils are exquisite targets of TNF-α that is under certain conditions it strengthens their expression of adhesion molecules induces their degranulation and successive release of lysosomal enzymes, causing the production of highly reactive oxygen species. TNF-α induces the migration of neutrophils mediating the production of chemotactic factors, including IL-8; this testifies cytokine networking involvement in inflammatory cell
TNF-α works by binding and clustering high-affinity receptors that are present in a great numbers on most cell membranes [24], the ligand/receptor complex is easily internalised via clathrin-coated pits and ends up in secondary lysozymes where it is degraded. Interestingly, the binding of TNF-α to the 75 kDa TNFR-2 is not sufficient to reach cytotoxicity, but rather binding to the 55 kDa TNFR-1 is sufficient to reach TNF-α mediated cell killing. TNF-α exerts its effects by activating several secondary proteins that provoke a variety of responses within the cell such as activation of gene transcription and/or production of reactive oxygen or nitrogen radicals (e.g., NO). Activated proteins include Gprotein, transcription factors such as NF-κB and AP-1 and serine and cysteine proteases, known as caspases. Many members of the TNF receptor superfamily have intracellular "death domains" which represent protein interaction domains each consisting of 65–80 amino acids; these proteins participate in TNFα mediated apoptosis process; many evidence demonstrated that TNF–TNFR interactions are implicated in the pathogenesis of CNS disorders such as EAE and MS. These interactions are able to monitor the disease outcome by modifying immune response and the interactions
However, recent studies showed a dual nature for TNF-α that it can be not only neurotoxic but also neuroprotective; a study conducted with transgenic mice for TNF-α receptors demonstrated that the mice lacking TNF-α showed more tissue loss and functional deficits compared to wild-type mice, implying that TNF-α mediated a neuroprotective effect [25]. The beneficial or deleterious effects of TNF-α dependent when it is being released and on cellular population that is acting on, the conflicting actions of TNF-α described above reflects a growing view of inflammation as a "dual-edged sword" having neurotoxic and neuroprotective
Thus, comprehension of their profile, kinetics of expression and interactions between TNF-α ligands and their TNFRs on different CNS residents and infiltrating immune cells would aid
cells, leading to necrosis or apoptosis.
84 Recovery of Motor Function Following Spinal Cord Injury
recruitment and an active role in inflammation.
between CNS-resident cells and effector immune cells in the CNS.
properties [26].
Moreover, other important mediators of inflammation that respond rapidly to disturbances within the microenvironment by change in morphology are the microglia, the expression "activated microglia" is used to define cells that change their immunophenotype and their morphology after a specific stimuli; the principal role of microglia at the lesion site is a rapid phagocytosis of fragments and induction of apoptosis [27]. The different response of microglia *in vitro* suggests that these cells may elicit unique functional properties and consequently control the inflammatory response at the injury site. Microglial activation has been well-known in the spinal cord tissue that has received a trauma and has shown to occur from caudal to lumbar enlargement, based on that there are papers supporting the role of microglia in pain after injury and showing activation of microglia post-SCI.
Thus, we postulated that activated spinal microglia have a role in chronic pain after SCI.
Microglia activate the innate immune system and are key regulators of inflammatory processes in CNS pathologies such as trauma and neurodegenerative diseases participation in both acute and chronic phase of the inflammatory responses. Activated microglia secrete cytotoxic substances including various cytokines such as TNF-α, IL-1, reactive free radicals, and nitric oxide. However, the principal effects of microglia at the levels of the lesion core are probably rapid phagocytosis of debris rather than induction of apoptosis. Microglia when activated can cause neuronal and glial toxicity through the release of cytokines, free radicals, eicosanoids, activated neutrophils, and macrophages [28]. On the other hand, microglia activation leads to beneficial effects producing growth factors that are fundamental for neuronal and tissue restoration. Moreover, it has been demonstrated that transplantation of peripherally activated macrophages has beneficial effects on spinal cord regeneration.
### **1.4. Apoptosis**
In the last decade the generation of apoptotic process after spinal cord trauma was also confirmed, apoptosis can be triggered by a variety of insults including cytokines, inflammatory injury, free radical damage, and excitotoxicity.
The apoptotic process after spinal cord trauma is activated in neurons, oligodendrocytes, microglia, and perhaps, astrocytes; apoptosis in microglia contributes to inflammatory secondary injury.
Two main pathways of apoptosis—extrinsic or receptor-dependent and intrinsic or receptorindependent—have been well characterized, and both appear to be active in SCI; the extrinsic or receptor-dependent pathway is mediated by Fas ligand and Fas receptor [29] and/or inducible nitric oxide synthase production by macrophages [30], while intrinsic or receptorindependent pathway is mediated via direct caspase-3 pro-enzyme activation [31] and/or mitochondrial damage, release of cytochrome c and activation of the inducer caspase-9, pathways of caspase-mediated apoptotic death [32].
Receptor-dependent apoptosis is evoked by extracellular signals, the most significant of which is TNF, so it is termed as "extrinsic" pathway; TNF is known to rapidly accumulate in the injured spinal cord, and activation of the Fas receptor of neurons, microglia, and oligoden‐ drocytes induces a programmed sequence of caspase activation. Moreover, additional control of cell death/survival is provided by the balance between major pro-apoptotic proteins such as Bax, Bad, and Bid and anti-apoptotic proteins such as Bcl-XL and Bcl-2.
Apoptotic cells were reached in the grey matter after injury, starting from 1h with a prolifer‐ ation during the other 8 h [33]. The number of apoptotic cells is increased at the site of injury and are related with axons degeneration [34].
Apoptotic process that is activated in the secondary injury in SCI has recently come under close study, and the precise contribution and potential therapeutic implications of apoptosis in SCI could be helpful to generate new therapeutic approach to treat the secondary events associated to spinal cord injury.
### **2. Inflammatory/immunologic response**
The inflammatory and immunological response to injury within the CNS is different than that which is occurring in other tissues [35]. The inflammatory and immunologic responses to injury involve activation of innate immune cells that provide immediate defence against inflammatory stimuli and in turn help to recruit cells of the adaptive immune system (i.e., Tand B-lymphocytes). The activation of immune system is driven by interactions involving presentation of antigen and release of various inflammatory mediators [36]. Also, cells present at the injury site may sequester debris and carry CNS antigens to secondary lymphoid organs [37], where trigger lymphocyte activation. Recent studies done on mice showed that the number of activated T and B cells increases in the spleen and bone marrow within 24 hours of trauma [38].
### **2.1. Lymphocytes infiltration**
T-lymphocytes are distributed in the intact spinal cord and gradually grow in number after trauma in parallel with the stimulation of microglia and influx of peripheral macrophages. Lymphocytes infiltrate the spinal cord tissue starting from 24 hours until 7 days after injury and return gradually back to normal levels in 4 to 5 weeks [14].
Under normal conditions, activated T cells can cross the Add Blood–brain barrier (BBB) and enter the CNS parenchyma. In contrast with other inflammatory cells enrolled after a trauma, the number of lymphocytes remains low [39]; however, T-lymphocytes play an important role in the CNS immune system, since on activation, T-lymphocytes may kill target cells and produce cytokines [40].
Once lymphocytes enter the lesion site, they persist indeterminately [38,41], whereas T and B cells increase at the lesion site at least 9 weeks post-injury [42,43], suggesting that cytokine/ chemokine gradients exist chronically and regulate integrin expression on endothelia and cells [44,45]. These chemokine gradients and adhesion molecules represent molecular targets for manipulating the effects of intraspinal lymphocytes after SCI [46–48]; the progressive increase in lymphocyte numbers may also be justified by lymphocyte activation and proliferation within the injured centre of spinal cord.
Moreover, induction of immune response could be generated as impaired nerve transmission; increasing the production of pro-inflammatory cytokines in chronic phase of SCI could worsen the damage increasing the axonal injury and demyelination. Furthermore, there are evidences that autoreactive lymphocytes promote neuronal survival *in vivo* through activation not only of autoreactive T cell but also through activation of other non-CNS reactive T cells or B cells such as resident microglia and infiltrating macrophages.
Thus, because lymphocytes remain for long term at the site of the lesion; new strategy of treatment could orientate on these cells that possess a fundamental role in regulating degen‐ erative and regenerative processes after injury.
### **3. Pharmacologic interventions for acute spinal cord injury**
The temporal profile of the secondary injury cascade provides a window within which it is theoretically possible to reduce the pathophysiological processes; many of the current pharmacological and surgical strategies for the treatment of SCI are based on minimising secondary injury and preserving neurological function following trauma to the brain and spinal cord. Over the past couple of decades, a myriad of agents have been nominated as putative neuroprotective therapies, several of these have been tested in the pre-clinical and clinical studies.
The following section briefly highlights some of the most promising neuroprotective ap‐ proaches that are being pursued.
### **3.1. Corticosteroids**
independent pathway is mediated via direct caspase-3 pro-enzyme activation [31] and/or mitochondrial damage, release of cytochrome c and activation of the inducer caspase-9,
Receptor-dependent apoptosis is evoked by extracellular signals, the most significant of which is TNF, so it is termed as "extrinsic" pathway; TNF is known to rapidly accumulate in the injured spinal cord, and activation of the Fas receptor of neurons, microglia, and oligoden‐ drocytes induces a programmed sequence of caspase activation. Moreover, additional control of cell death/survival is provided by the balance between major pro-apoptotic proteins such
Apoptotic cells were reached in the grey matter after injury, starting from 1h with a prolifer‐ ation during the other 8 h [33]. The number of apoptotic cells is increased at the site of injury
Apoptotic process that is activated in the secondary injury in SCI has recently come under close study, and the precise contribution and potential therapeutic implications of apoptosis in SCI could be helpful to generate new therapeutic approach to treat the secondary events
The inflammatory and immunological response to injury within the CNS is different than that which is occurring in other tissues [35]. The inflammatory and immunologic responses to injury involve activation of innate immune cells that provide immediate defence against inflammatory stimuli and in turn help to recruit cells of the adaptive immune system (i.e., Tand B-lymphocytes). The activation of immune system is driven by interactions involving presentation of antigen and release of various inflammatory mediators [36]. Also, cells present at the injury site may sequester debris and carry CNS antigens to secondary lymphoid organs [37], where trigger lymphocyte activation. Recent studies done on mice showed that the number of activated T and B cells increases in the spleen and bone marrow within 24 hours of
T-lymphocytes are distributed in the intact spinal cord and gradually grow in number after trauma in parallel with the stimulation of microglia and influx of peripheral macrophages. Lymphocytes infiltrate the spinal cord tissue starting from 24 hours until 7 days after injury
Under normal conditions, activated T cells can cross the Add Blood–brain barrier (BBB) and enter the CNS parenchyma. In contrast with other inflammatory cells enrolled after a trauma, the number of lymphocytes remains low [39]; however, T-lymphocytes play an important role in the CNS immune system, since on activation, T-lymphocytes may kill target cells and
and return gradually back to normal levels in 4 to 5 weeks [14].
as Bax, Bad, and Bid and anti-apoptotic proteins such as Bcl-XL and Bcl-2.
pathways of caspase-mediated apoptotic death [32].
86 Recovery of Motor Function Following Spinal Cord Injury
and are related with axons degeneration [34].
**2. Inflammatory/immunologic response**
associated to spinal cord injury.
trauma [38].
**2.1. Lymphocytes infiltration**
produce cytokines [40].
In the last decade one of the most used approaches to treat patients with a severe SCI is the use of corticosteroids that possess a well-recognized anti-inflammatory properties reducing spinal cord edema. However, the exact mechanisms by which corticosteroids mediated neuroprotection are not yet completely understood but seems that they induce a reduction of inflammatory cytokines production, modulation of the inflammatory/immune cells, inhibition of lipid peroxidation and reduction of oxidative stress. In this regard, methylprednisolone (MP) appears to be particularly efficacious compared with other glucocorticoids. According to the National Acute Spinal Cord Injury Study 2 (NASCIS 2) protocol, MP is usually admin‐ istered within 8 h after trauma in a high concentration of 30 mg/kg, followed by an infusion of 5.4 mg/kg/h for 23 hours [49,50]. MP is the only well-known pharmacological treatment of SCI injury patients; however, its administration has been shown to be controversial. Some toxic effects such as infection rates, pulmonary embolism, severe pneumonia and sepsis and even death secondary to respiratory complications appeared to be higher with steroid use.
The anti-inflammatory properties of MP are mediated by different mechanisms that are known as transrepression, this mechanism involves interference with pro-inflammatory transcription factor signalling (such as NF-κB) which is upstream of several inflammatory mediators including COX2, chemokines and cytokines.
Together with the minor functional improvement in humans and risk of adverse side effects, these studies highlighted the compelling need to develop better neuroprotective agents with more convincing efficacy.
### **3.2. Estrogen as a neuroprotective agent**
At present, several effects have been associated to estrogen that acts with different mechanism of action [51]. Recently, the neuroprotective and anti-inflammatory effect of estrogen are of great interest; leukocyte adhesion and microglia activation are also sensitive to estrogen and show significantly decreased superoxide dismutase production and phagocytic activity when treated with estrogen *in vitro* [52].
Moreover, cell death is associated to decreases blood flow in spinal cord and estrogen are involved in increasing post-traumatic blood flow induced by ischemia [53] and TBI [54]. Thus, estrogen treatment has been related with increased expression of pro-apoptotic Bcl-2 in the spinal cord injured tissue, this anti-apoptotic increase in Bcl-2 may be mediated by Akt activation with downstream phosphorylation of cAMP response element binding protein (CREB) [55].
However, estrogen may also act on Add N-methyl-D-aspartate receptor (NMDA) receptors, indicating a potential to limit secondary cell death due to excitotoxicity [56].
### **3.3. Nuclear hormone receptors family: PPARs receptor**
Other members of the nuclear hormone receptors family (NHRs) are now explored for their anti-inflammatory properties in experimental models, including SCI.
Peroxisome proliferator-activated receptors (PPARs) are part of the nuclear hormone receptor superfamily, upon ligand activation regulate gene expression and have been shown to be antiinflammatory in different model of inflammatory pathology, including SCI. PPAR exists as three isoforms (α, β/δ and γ) that control many cellular functions including lipid metabolism, glucose absorption, cell growth and differentiation, and inflammation.
One mechanism involves direct interaction of PPAR with pro-inflammatory transcription factors, most importantly NFκB and AP-1, and the subsequent reduction of gene transcription. Pioglitazone and rosiglitazone are PPARγ agonists that are in common clinical use for type II diabetes. Beyond their metabolic effects, interest in PPAR γ ligand has grown due to their antiinflammatory, neuroprotective, and even anti-neoplastic properties, suggesting its potential use after spinal cord trauma [57]. As such, PPAR agonists have already been clinically tested in other disorders with inflammatory pathology, such as Alzheimer's disease (AD), rheuma‐ toid arthritis (RA), and ischemia reperfusion injury, but not in SCI yet. Moreover, in the last few years a great interest has been focused on other PPAR receptors agonist such as for PPARα and β/δ receptors. One of the first reports indicating that PPARα is involved in attenuating inflammation demonstrated that the eicosanoid LTB4 binds and activates PPARα [58]. Several studies on inflammatory cytokine produced in aged mice demonstrated an active interaction of PPARα with NFκB; in fact the oxidative stress in different tissue leads to active NFκB. Treatment with PPARα agonists were found to restore the alteration of oxidative mediators, to inhibit the activation of NFκB and to remove IL-6 and IL-12 produced [59,60].
These features were accompanied by enhanced functional motor recovery and reduced hyperalgesia.
Since PPAR agonists are currently used in medical treatment of diabetes, clinical studies for stroke and different CNS pathologies are to be expected. The knowledge about anti-inflam‐ matory properties of PPAR ligands obtained from cell cultures and animal model of SCI demonstrate that PPARs signalling may be therapeutic targets after spinal cord injury.
### **4. Conclusion**
effects such as infection rates, pulmonary embolism, severe pneumonia and sepsis and even
The anti-inflammatory properties of MP are mediated by different mechanisms that are known as transrepression, this mechanism involves interference with pro-inflammatory transcription factor signalling (such as NF-κB) which is upstream of several inflammatory mediators
Together with the minor functional improvement in humans and risk of adverse side effects, these studies highlighted the compelling need to develop better neuroprotective agents with
At present, several effects have been associated to estrogen that acts with different mechanism of action [51]. Recently, the neuroprotective and anti-inflammatory effect of estrogen are of great interest; leukocyte adhesion and microglia activation are also sensitive to estrogen and show significantly decreased superoxide dismutase production and phagocytic activity when
Moreover, cell death is associated to decreases blood flow in spinal cord and estrogen are involved in increasing post-traumatic blood flow induced by ischemia [53] and TBI [54]. Thus, estrogen treatment has been related with increased expression of pro-apoptotic Bcl-2 in the spinal cord injured tissue, this anti-apoptotic increase in Bcl-2 may be mediated by Akt activation with downstream phosphorylation of cAMP response element binding protein
However, estrogen may also act on Add N-methyl-D-aspartate receptor (NMDA) receptors,
Other members of the nuclear hormone receptors family (NHRs) are now explored for their
Peroxisome proliferator-activated receptors (PPARs) are part of the nuclear hormone receptor superfamily, upon ligand activation regulate gene expression and have been shown to be antiinflammatory in different model of inflammatory pathology, including SCI. PPAR exists as three isoforms (α, β/δ and γ) that control many cellular functions including lipid metabolism,
One mechanism involves direct interaction of PPAR with pro-inflammatory transcription factors, most importantly NFκB and AP-1, and the subsequent reduction of gene transcription. Pioglitazone and rosiglitazone are PPARγ agonists that are in common clinical use for type II diabetes. Beyond their metabolic effects, interest in PPAR γ ligand has grown due to their antiinflammatory, neuroprotective, and even anti-neoplastic properties, suggesting its potential use after spinal cord trauma [57]. As such, PPAR agonists have already been clinically tested in other disorders with inflammatory pathology, such as Alzheimer's disease (AD), rheuma‐
indicating a potential to limit secondary cell death due to excitotoxicity [56].
anti-inflammatory properties in experimental models, including SCI.
glucose absorption, cell growth and differentiation, and inflammation.
**3.3. Nuclear hormone receptors family: PPARs receptor**
death secondary to respiratory complications appeared to be higher with steroid use.
including COX2, chemokines and cytokines.
88 Recovery of Motor Function Following Spinal Cord Injury
**3.2. Estrogen as a neuroprotective agent**
treated with estrogen *in vitro* [52].
(CREB) [55].
more convincing efficacy.
Thus, recent researches are moving to develop new pharmacological approaches that may offer an effective neuroprotection after spinal cord injury. After spinal cord injury, inflamma‐ tory reactions account for a large proportion of the secondary damage to neurons and oligodendrocytes.
Promising research is being carried out to better understand the aspects of inflammation, lipid peroxidation and apoptotic cell death that may be the target of pharmacologic intervention.
Few agents have been studied demonstrating efficacy in animal models of spinal cord injury and may become appropriate for testing in the human setting in the near future.
Clearly, much effort has to be done to bring experimental strategies to clinical fruition, but they do represent promising potential interventions.
Thus, the current medical and surgical interventions for the acutely cord-injured patient attempt to minimize the inflammatory process that possess a crucial role in generating and maintaining secondary damage associated to injury and defend the neural cells that initially survived the mechanical injury.
### **Author details**
Irene Paterniti, Emanuela Esposito\* and Salvatore Cuzzocrea
\*Address all correspondence to: eesposito@unime.it
Department of Chemical, Biological, Pharmaceutical and Environmental Sciences, University of Messina, Messina, Italy
### **References**
[13] Paterniti I, Genovese T, Crisafulli C, Mazzon E, Di Paola R, Galuppo M, et al. Treatment with green tea extract attenuates secondary inflammatory response in an experimental model of spinal cord trauma. Naunyn-Schmiedeberg's archives of pharmacology. 2009; 380(2): 179–92. doi:10.1007/s00210-009-0414-z.
**References**
39(6): 889–909.
90 Recovery of Motor Function Following Spinal Cord Injury
2012.09.008.
443–64.
712–24.
10.1186/1742-2094-10-43.
[1] Giovannini MG, Scali C, Prosperi C, Bellucci A, Vannucchi MG, Rosi S, et al. Betaamyloid-induced inflammation and cholinergic hypofunction in the rat brain in vivo: involvement of the p38MAPK pathway. Neurobiology of disease. 2002; 11(2): 257–74.
[2] Dauer W, Przedborski S. Parkinson's disease: mechanisms and models. Neuron. 2003;
[3] Tenorio G, Kulkarni A, Kerr BJ. Resident glial cell activation in response to perispinal inflammation leads to acute changes in nociceptive sensitivity: implications for the generation of neuropathic pain. Pain. 2013; 154(1): 71–81. doi:10.1016/j.pain.
[4] Najjar S, Pearlman DM, Alper K, Najjar A, Devinsky O. Neuroinflammation and psychiatric illness. Journal of neuroinflammation. 2013; 10: 43. doi:
[5] Iadecola C, Anrather J. The immunology of stroke: from mechanisms to translation.
[6] Popovich PG, Wei P, Stokes BT. Cellular inflammatory response after spinal cord injury in Sprague–Dawley and Lewis rats. The journal of comparative neurology. 1997; 377(3):
[7] Dusart I, Schwab ME. Secondary cell death and the inflammatory reaction after dorsal hemisection of the rat spinal cord. The European journal of neuroscience. 1994; 6(5):
[8] Bartholdi D, Schwab ME. Expression of pro-inflammatory cytokine and chemokine mRNA upon experimental spinal cord injury in mouse: an in situ hybridization study.
[9] Klusman I, Schwab ME. Effects of pro-inflammatory cytokines in experimental spinal
[10] Tonai T, Taketani Y, Ueda N, Nishisho T, Ohmoto Y, Sakata Y, et al. Possible involve‐ ment of interleukin-1 in cyclooxygenase-2 induction after spinal cord injury in rats.
[11] Dubois RN, Abramson SB, Crofford L, Gupta RA, Simon LS, Van De Putte LB, et al. Cyclooxygenase in biology and disease. FASEB journal : official publication of the Federation of American Societies for Experimental Biology. 1998; 12(12): 1063–73.
[12] Genovese T, Mazzon E, Di Paola R, Cannavo G, Muia C, Bramanti P, et al. Role of endogenous ligands for the peroxisome proliferators activated receptors alpha in the secondary damage in experimental spinal cord trauma. Experimental neurology. 2005;
Nature medicine. 2011; 17(7): 796–808. doi:10.1038/nm.2399.
The European journal of neuroscience. 1997; 9(7): 1422–38.
cord injury. Brain research. 1997; 762(1–2): 173–84.
194(1): 267–78. doi:10.1016/j.expneurol.2005.03.003.
Journal of neurochemistry. 1999; 72(1): 302–9.
[38] Ankeny DP, Lucin KM, Sanders VM, McGaughy VM, Popovich PG. Spinal cord injury triggers systemic autoimmunity: evidence for chronic B lymphocyte activation and lupus-like autoantibody synthesis. Journal of neurochemistry. 2006; 99(4): 1073–87. doi: 10.1111/j.1471-4159.2006.04147.x.
[25] Kim GM, Xu J, Xu J, Song SK, Yan P, Ku G, et al. Tumor necrosis factor receptor deletion reduces nuclear factor-kappaB activation, cellular inhibitor of apoptosis protein 2 expression, and functional recovery after traumatic spinal cord injury. The journal of neuroscience : the official journal of the society for neuroscience. 2001; 21(17): 6617–25.
[26] Bethea JR. Spinal cord injury-induced inflammation: a dual-edged sword. Progress in
[27] Shuman SL, Bresnahan JC, Beattie MS. Apoptosis of microglia and oligodendrocytes after spinal cord contusion in rats. Journal of neuroscience research. 1997; 50(5): 798–
[28] Schwartz M. Autoimmune involvement in CNS trauma is beneficial if well controlled. Progress in brain research. 2000; 128: 259–63. doi:10.1016/S0079-6123(00)28023-0.
[29] Leskovar A, Moriarty LJ, Turek JJ, Schoenlein IA, Borgens RB. The macrophage in acute neural injury: changes in cell numbers over time and levels of cytokine production in mammalian central and peripheral nervous systems. The journal of experimental
[30] Satake K, Matsuyama Y, Kamiya M, Kawakami H, Iwata H, Adachi K, et al. Nitric oxide via macrophage iNOS induces apoptosis following traumatic spinal cord injury. Brain
[31] Citron BA, Arnold PM, Sebastian C, Qin F, Malladi S, Ameenuddin S, et al. Rapid upregulation of caspase-3 in rat spinal cord after injury: mRNA, protein, and cellular localization correlates with apoptotic cell death. Experimental neurology. 2000; 166(2):
[32] Hartley CJ, Reddy AK, Madala S, Martin-McNulty B, Vergona R, Sullivan ME, et al. Hemodynamic changes in apolipoprotein E-knockout mice. American journal of
[33] Zurita M, Vaquero J, Zurita I. Presence and significance of CD-95 (Fas/APO1) expres‐ sion after spinal cord injury. Journal of neurosurgery. 2001; 94(2 Suppl): 257–64.
[34] Casha S, Yu WR, Fehlings MG. Oligodendroglial apoptosis occurs along degenerating axons and is associated with FAS and p75 expression following spinal cord injury in
[35] Schwartz M, Moalem G, Leibowitz-Amit R, Cohen IR. Innate and adaptive immune responses can be beneficial for CNS repair. Trends in neurosciences. 1999; 22(7): 295–
[36] Ling C, Sandor M, Fabry Z. In situ processing and distribution of intracerebrally injected OVA in the CNS. Journal of neuroimmunology. 2003; 141(1–2): 90–8.
[37] Karman J, Ling C, Sandor M, Fabry Z. Initiation of immune responses in brain is promoted by local dendritic cells. Journal of immunology. 2004; 173(4): 2353–61.
physiology-heart and circulatory physiology. 2000; 279(5): H2326–34.
research molecular brain research. 2000; 85(1–2): 114–22.
brain research. 2000; 128: 33–42. doi:10.1016/S0079-6123(00)28005-9.
808.
9.
biology. 2000; 203(Pt 12): 1783–95.
92 Recovery of Motor Function Following Spinal Cord Injury
213–26. doi:10.1006/exnr.2000.7523.
the rat. Neuroscience. 2001; 103(1): 203–18.
The New England journal of medicine. 1990; 322(20): 1405–11. doi:10.1056/ NEJM199005173222001.
## **Normal Distribution and Plasticity of Serotonin Receptors after Spinal Cord Injury and Their Impacts on Motor Outputs**
Mengliang Zhang
The New England journal of medicine. 1990; 322(20): 1405–11. doi:10.1056/
[51] Sribnick EA, Wingrave JM, Matzelle DD, Ray SK, Banik NL. Estrogen as a neuropro‐ tective agent in the treatment of spinal cord injury. Annals of the New York Academy
[52] Bruce-Keller AJ, Keeling JL, Keller JN, Huang FF, Camondola S, Mattson MP. Antiin‐ flammatory effects of estrogen on microglial activation. Endocrinology. 2000; 141(10):
[53] He Z, He YJ, Day AL, Simpkins JW. Proestrus levels of estradiol during transient global cerebral ischemia improves the histological outcome of the hippocampal CA1 region: perfusion-dependent and independent mechanisms. Journal of the neurological
[54] Roof RL, Hall ED. Estrogen-related gender difference in survival rate and cortical blood flow after impact-acceleration head injury in rats. Journal of neurotrauma. 2000; 17(12):
[55] Honda K, Shimohama S, Sawada H, Kihara T, Nakamizo T, Shibasaki H, et al. Nonge‐ nomic antiapoptotic signal transduction by estrogen in cultured cortical neurons.
[56] Nilsen J, Chen S, Brinton RD. Dual action of estrogen on glutamate-induced calcium signaling: mechanisms requiring interaction between estrogen receptors and src/ mitogen activated protein kinase pathway. Brain research. 2002; 930(1–2): 216–34. [57] Cuzzocrea S. Peroxisome proliferator-activated receptors gamma ligands and ischemia and reperfusion injury. Vascular pharmacology. 2004; 41(6): 187–95. doi:10.1016/j.vph.
[58] Devchand PR, Keller H, Peters JM, Vazquez M, Gonzalez FJ, Wahli W. The PPAR alphaleukotriene B4 pathway to inflammation control. Nature. 1996; 384(6604): 39–43. doi:
[59] Esposito E, Cuzzocrea S. Palmitoylethanolamide is a new possible pharmacological treatment for the inflammation associated with trauma. Mini reviews in medicinal
[60] Esposito E, Cuzzocrea S. Targeting the peroxisome proliferator-activated receptors (PPARs) in spinal cord injury. Expert opinion on therapeutic targets. 2011; 15(8): 943–
NEJM199005173222001.
94 Recovery of Motor Function Following Spinal Cord Injury
of Sciences. 2003; 993: 125–33; discussion 59–60.
3646–56. doi:10.1210/endo.141.10.7693.
1155–69. doi:10.1089/neu.2000.17.1155.
Journal of neuroscience research. 2001; 64(5): 466–75.
sciences. 2002; 193(2): 79–87.
2004.10.004.
10.1038/384039a0.
chemistry. 2013; 13(2): 237–55.
59. doi:10.1517/14728222.2011.581231.
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/63759
### **Abstract**
Following spinal cord injury (SCI) a series of anatomical and functional plastic changes occurinthespinalcord,includingreorganizationofthespinalneuronalnetwork,alteration of properties of interneurons and motoneurons as well as up- or down-regulation of different neurotransmitter receptors. In mammalian spinal cord, one of the important neurotransmitters, serotonin(5-HT),playsanessentialrole inmodulatingsensory,motor and autonomic functions. Following SCI, especially complete spinal cord lesion, the descending supply of 5-HT is lost. As a consequence different 5-HT receptors undergo variant degrees of plastic changes.
Inthis chapterIhave systematically reviewedthedistributionofdifferent 5-HTreceptors in the spinal cord and their plastic changes following SCI where applicable. In addition, the plastic changes of 5-HT supplying system in reaction to SCI have also been re‐ viewed. These results indicate that 5-HT receptors are important factors not only for modulation of normal motor function, their plastic changes are also critical for motor functional recovery and, quite often, for the development of certain pathological states after SCI. Pharmacological and/or genetic intervention of selected 5-HT receptors and/or intrinsic 5-HT producing system in the spinal cord may pave new ways for the restora‐ tion of motor functions after SCI.
**Keywords:** monoamine, monoamine receptor, spinal cord, motor control, intraspinal 5-HT cell
© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
### **1. Introduction**
Spinal cord injury (SCI) is a devastating condition with an incidence of 10–83 per million people per year worldwide according to statistical data from different countries [1]. It leads to an extensive and usually irreversible loss of sensory functions, voluntary motor control, and autonomic functions below injury level. A variety of primary and secondary complications occur depending on the severity of the injury and time course of its development. These symptoms may involve different systems, manifesting as, e.g., paralysis, spasticity, neuro‐ pathic pain, pulmonary and cardiovascular problems, osteoporosis, anemia, pressure ulcers, bladder and bowel problems, and sexual dysfunction [2–8]. Currently, there is no cure for SCI and thus improving quality of life, such as restoration of partial motor and sensory functions, has become a priority for setting up the treatment strategy. The primary cause of these problems is the loss of both descending and ascending projecting pathways in the spinal cord. The descending pathways include direct motor-initiating pathways, such as cortical spinal tracts, and modulatory pathways, such as serotonergic, dopaminergic, and noradrenergic pathways. These monoaminergic systems are so important that in either acute or chronic SCI direct stimulation of the receptors of these monoamines with drugs could regain locomotor activity in animals [9–17]. Thus, restoring the function of monoaminergic systems has become a key strategy to restore motor function and ameliorate secondary symptoms [18–23].
So far a great number of studies have been focusing on the serotonergic system in SCI. In the mammalian spinal cord, serotonin (5-HT) originates mainly in the raphe nuclei of the brain‐ stem and plays an important role in modulating sensory, motor, and autonomic function [24– 27]. Following SCI, especially complete spinal cord lesion, the descending supply of 5-HT is lost and as a consequence different 5-HT receptors undergo variant degrees of plastic changes [22, 28–36]. In addition, a potential intraspinal 5-HT-producing system in the spinal cord, i.e., aromatic L-amino acid decarboxylase (AADC) cells, also undergoes plastic changes to increase their potency to produce 5-HT from its precursor [37–40]. Although these plastic changes may induce pathological symptoms such as spasticity and chronic central pain [34, 41, 42], they are essential for spinal function recovery (for review see [43–45]). In this chapter, I will make a systematic review according to up-to-date literature related to the distribution and plastic changes of 5-HT receptors in the spinal cord in normal or SCI states with a note on the intraspinal 5-HT-producing cells.
### **2. 5-HT in the spinal cord in health**
In mammals, including humans, 5-HT axons in the spinal cord almost exclusively originate from the brainstem raphe nuclei [46–49]. Their terminals are distributed in all parts of the gray matter at all levels of the spinal cord [50, 51]. The cell bodies with descending 5-HT projec‐ tions are located in the caudal part of the raphe nuclei, which include the raphe magnus, raphe obscurus, raphe pallidus, ventral lateral medulla, and the area postrema [52]. In the spinal cord, 5-HT projecting fibers descend in the white matterthrough two differentroutes: one with fibers from the raphe magnus in the dorsal part of the lateral funiculus which terminates mainly in the dorsal horn, and the other with fibers from raphe obscurus and raphe pallidus in the ventralfuniculus which terminates mainly in the ventral horn and the intermediate zone.
In addition to the descending projecting system, there are indeed 5-HT neurons in the spinal cord although in normal states their contribution of 5-HT can be ignored. So far a small number of intraspinal 5-HT neurons have been reported in macaque monkey (ca. 150 cells per monkey [53]), rat (3–9 cells per rat [54, 55]), and mouse [51]. In the rat spinal cord, 5-HT cells were distributed in different parts of the spinal cord with the exception of cervical segments; in the monkey spinal cord most of the cells were observed in the cervical segments with a small number in other segments; whereas in the mouse spinal cord they were exclusively located in the sacral segments. In the rat spinal cord, the 5-HT cells were found to be located in lami‐ nae VII and X in the gray matter, whereas in monkey and mouse spinal cord they were exclusively found in lamina X.
Serotonin in the spinal cord plays an important role in sensory information processing, motor control, and autonomic function. Traditionally, it is hypothesized that 5-HT in the spinal cord exerts its effects by inhibiting sensory systems and facilitating motor systems [56]. However, now it is known that the effects of 5-HT are very complicated both in sensory and motor aspects. In sensory aspect, 5-HT not only has antinociceptive but also pronociceptive effects (e.g., [57, 58]). In motor aspect, except for facilitating motor output, 5-HT also inhibits motor behavior (e.g., [59, 60]). The different functions that 5-HT exert at different circumstances depend on many factors, such as the brainstem origins of the descending projections, the termination localizations of these fibers in the spinal cord, and the activation states of its different receptors.
### **3. 5-HT receptors in the spinal cord in health**
**1. Introduction**
96 Recovery of Motor Function Following Spinal Cord Injury
intraspinal 5-HT-producing cells.
**2. 5-HT in the spinal cord in health**
Spinal cord injury (SCI) is a devastating condition with an incidence of 10–83 per million people per year worldwide according to statistical data from different countries [1]. It leads to an extensive and usually irreversible loss of sensory functions, voluntary motor control, and autonomic functions below injury level. A variety of primary and secondary complications occur depending on the severity of the injury and time course of its development. These symptoms may involve different systems, manifesting as, e.g., paralysis, spasticity, neuro‐ pathic pain, pulmonary and cardiovascular problems, osteoporosis, anemia, pressure ulcers, bladder and bowel problems, and sexual dysfunction [2–8]. Currently, there is no cure for SCI and thus improving quality of life, such as restoration of partial motor and sensory functions, has become a priority for setting up the treatment strategy. The primary cause of these problems is the loss of both descending and ascending projecting pathways in the spinal cord. The descending pathways include direct motor-initiating pathways, such as cortical spinal tracts, and modulatory pathways, such as serotonergic, dopaminergic, and noradrenergic pathways. These monoaminergic systems are so important that in either acute or chronic SCI direct stimulation of the receptors of these monoamines with drugs could regain locomotor activity in animals [9–17]. Thus, restoring the function of monoaminergic systems has become
a key strategy to restore motor function and ameliorate secondary symptoms [18–23].
So far a great number of studies have been focusing on the serotonergic system in SCI. In the mammalian spinal cord, serotonin (5-HT) originates mainly in the raphe nuclei of the brain‐ stem and plays an important role in modulating sensory, motor, and autonomic function [24– 27]. Following SCI, especially complete spinal cord lesion, the descending supply of 5-HT is lost and as a consequence different 5-HT receptors undergo variant degrees of plastic changes [22, 28–36]. In addition, a potential intraspinal 5-HT-producing system in the spinal cord, i.e., aromatic L-amino acid decarboxylase (AADC) cells, also undergoes plastic changes to increase their potency to produce 5-HT from its precursor [37–40]. Although these plastic changes may induce pathological symptoms such as spasticity and chronic central pain [34, 41, 42], they are essential for spinal function recovery (for review see [43–45]). In this chapter, I will make a systematic review according to up-to-date literature related to the distribution and plastic changes of 5-HT receptors in the spinal cord in normal or SCI states with a note on the
In mammals, including humans, 5-HT axons in the spinal cord almost exclusively originate from the brainstem raphe nuclei [46–49]. Their terminals are distributed in all parts of the gray matter at all levels of the spinal cord [50, 51]. The cell bodies with descending 5-HT projec‐ tions are located in the caudal part of the raphe nuclei, which include the raphe magnus, raphe obscurus, raphe pallidus, ventral lateral medulla, and the area postrema [52]. In the spinal cord, 5-HT projecting fibers descend in the white matterthrough two differentroutes: one with
The diverse functions of 5-HT in the spinal cord are achieved through the activation of different 5-HT receptors. As seen in **Table 1**, seven families (or types) and at least 14 subfa‐ milies (or subtypes) of 5-HT receptors have been identified so far [61]. In these seven fami‐ lies, with exception of 5-HT3 receptors that are ligand-gated ion channels, all other families are G protein-coupled receptors [62]. To better understand the diversity of 5-HT functions, it is fundamentally important to have the knowledge of the anatomical localizations of differ‐ ent 5-HT receptors in the spinal cord. A majority, though not all, of 5-HT receptor subfami‐ lies have been found to be expressed in the spinal cord (for reviews see [63, 64]). Here, I will make a systematic review of these receptors in terms of their cellular as well as subcellular localizations in the spinal cord based on available data. One should keep in mind that due to the existence of a great number of splices and editing variants for several 5-HT receptors, a greater degree of operational diversity could be expected, though this issue is not the focus of the present review.
Normal Distribution and Plasticity of Serotonin Receptors after Spinal Cord Injury and Their Impacts on Motor Outputs http://dx.doi.org/10.5772/63759 99
**Receptor family (year of molecular gene clone)\***
**Receptor subfamily (year of molecular gene clone)\***
98 Recovery of Motor Function Following Spinal Cord Injury
5-HT1 1A (1987) Primary afferent
**Expression in the spinal cord**
fibers in dorsal horn [66]; neuronal somata in different laminae of gray matter [31,69]; axon initial segments [31,71];
1B (1992) Primary afferent fibers in dorsal horn [83]; neuronal somata at least in intermediate zone [38]
1D (1991) Primary afferent fibers in dorsal horn [92]; gamma motoneurons in ventral horn [94]
1F (1993) Substantia
5-HT2 2A (1988) Different laminae
gelatinosa of spinal dorsal horn [98]
of the spinal gray matter but largely in the superficial dorsal horn and lamina IX **Functions at normal states**
increases motoneuron excitability [60]; induces central fatigue [81]
Antinociception [73,76]; pronociception [77];
Antinociception [72]; autoreceptor [89]; inhibits activity of AADC cells
Antinociception, [93];
modulates proprioceptive circuits [94]
1E (1992) -- -- -- --
Antinociception [101,102]??
Pronociceptive [57,112]; antinociceptive [115]; excites motoneuron [117]; facilitates micturition reflex [119]; facilitates sexual behavior
[38]
**Expression changes following SCI**
Upregulation at least to 30 days after SCI [28,31]
Upregulation [32,33,35]
**Functions at SCI states**
Promotes motor functional recovery [13,14,172]
and polysynaptic
reflexes [99,174]
polysynaptic reflex [95]??
polysynaptic reflex [99]??
Locomotor functional recovery [12,14,43]; 5-HT supersensitivity and muscle spasm
[42]
\* Year of molecular gene clone is based on Wikipedia (https://en.wikipedia.org/wiki/5HT\_receptor) except 5-HT3 receptors; --: no data available; ??: data not conclusive.
**Table 1.** 5-HT receptors in the spinal cord and plastic changes following SCI.
### **3.1. 5-HT1 receptors**
As listed in **Table 1**, 5-HT1 receptors include five subfamilies: 1A, 1B, 1D, 1E, and 1F. Although all of these receptors have been detected in the brain by different techniques [65], there is no report to my knowledge demonstrating the expression of 5-HT1E receptors in the spinal cord. In addition, the evidence of the existence of 5-HT1F in the spinal cord is mainly from physio‐ logical experiments. Nonetheless, I will describe 5-HT1A, 1B, 1D, and 1F in the spinal cord according to the available data.
*5-HT1A receptors:* Data from both autoradiographic (radioligand binding) and immunohisto‐ chemical experiments have demonstrated the presence of 5-HT1A receptors in different regions of the spinal gray matter across different spinal segments in rats [31, 66–69]. 5-HT1A receptor binding sites were predominantly seen in the dorsal horn especially laminae I and II [67], where they were partly located in the primary afferent fibers [66]. In autoradiograph‐ ic images, it is difficult to identify the components located in cell bodies; however, with immunohistochemistry with a 5-HT1A antibody generated from an intracellular epitope, it was clear that 5-HT1A receptors were expressed in cell bodies in the spinal cord [31]. The immunolabeled cell bodies were located in different spinalregions in the gray matterincluding the dorsal horn, intermediate zone, and ventral horn [31, 69, 70]. In addition, using a differ‐ ent 5-HT1A antibody generated from an extracellular epitope, the receptors were also shown to be present in the axon hillock [31, 71]. There are no studies to investigate 5-HT1A receptor localization at an ultrastructural level and therefore no data available for their subcellular distribution.
It is commonly believed that 5-HT1A receptors exert an inhibitory effect in sensory informa‐ tion transmission including nociception in the spinal dorsal horn (e.g., [72–76]). For example, in mice intrathecal injection of 5-HT1A receptor agonist 8-hydroxy-2-(di-*n*-propylamino) tetralin (8-OH-DPAT) could inhibit the tail-flicks induced by noxious radiant heat [72]. In neonatal rat spinal cord in vitro experiments, Hochman et al. [73] showed that 8-OH-DPAT depressed evoked excitatory postsynaptic potentials (EPSPs) in the deep dorsal horn neu‐ rons, whereas 5-HT1A receptor antagonist WAY-100635 (*N*-[2-[4-(2-methoxyphenyl)-1-
piperazinyl]ethyl]-*N*-(2-pyridyl)cyclohexanecarboxamide) facilitated evoked EPSPs. Data from the same group also indicated that the activation of 5-HT1A receptors facilitated longterm depression in the dorsal horn neurons. However, the effects of 5-HT1A receptors in the dorsal horn are not always inhibitory. For example, the activation of 5-HT1A receptors has been reported to facilitate nociceptive [77, 78] and itch transmission [79], and to induce spontaneous tail-flicks [80].
**Receptor family (year of molecular gene clone)\***
**Receptor subfamily (year of molecular gene clone)\***
100 Recovery of Motor Function Following Spinal Cord Injury
**3.1. 5-HT1 receptors**
distribution.
according to the available data.
**Expression in the spinal cord**
with a dorsoventral density gradient [16,162]
**Table 1.** 5-HT receptors in the spinal cord and plastic changes following SCI.
receptors; --: no data available; ??: data not conclusive.
**Functions at normal states**
\* Year of molecular gene clone is based on Wikipedia (https://en.wikipedia.org/wiki/5HT\_receptor) except 5-HT3
As listed in **Table 1**, 5-HT1 receptors include five subfamilies: 1A, 1B, 1D, 1E, and 1F. Although all of these receptors have been detected in the brain by different techniques [65], there is no report to my knowledge demonstrating the expression of 5-HT1E receptors in the spinal cord. In addition, the evidence of the existence of 5-HT1F in the spinal cord is mainly from physio‐ logical experiments. Nonetheless, I will describe 5-HT1A, 1B, 1D, and 1F in the spinal cord
*5-HT1A receptors:* Data from both autoradiographic (radioligand binding) and immunohisto‐ chemical experiments have demonstrated the presence of 5-HT1A receptors in different regions of the spinal gray matter across different spinal segments in rats [31, 66–69]. 5-HT1A receptor binding sites were predominantly seen in the dorsal horn especially laminae I and II [67], where they were partly located in the primary afferent fibers [66]. In autoradiograph‐ ic images, it is difficult to identify the components located in cell bodies; however, with immunohistochemistry with a 5-HT1A antibody generated from an intracellular epitope, it was clear that 5-HT1A receptors were expressed in cell bodies in the spinal cord [31]. The immunolabeled cell bodies were located in different spinalregions in the gray matterincluding the dorsal horn, intermediate zone, and ventral horn [31, 69, 70]. In addition, using a differ‐ ent 5-HT1A antibody generated from an extracellular epitope, the receptors were also shown to be present in the axon hillock [31, 71]. There are no studies to investigate 5-HT1A receptor localization at an ultrastructural level and therefore no data available for their subcellular
It is commonly believed that 5-HT1A receptors exert an inhibitory effect in sensory informa‐ tion transmission including nociception in the spinal dorsal horn (e.g., [72–76]). For example, in mice intrathecal injection of 5-HT1A receptor agonist 8-hydroxy-2-(di-*n*-propylamino) tetralin (8-OH-DPAT) could inhibit the tail-flicks induced by noxious radiant heat [72]. In neonatal rat spinal cord in vitro experiments, Hochman et al. [73] showed that 8-OH-DPAT depressed evoked excitatory postsynaptic potentials (EPSPs) in the deep dorsal horn neu‐ rons, whereas 5-HT1A receptor antagonist WAY-100635 (*N*-[2-[4-(2-methoxyphenyl)-1-
**Expression changes following SCI**
[16,167,169] recovery [43,172]
**Functions at SCI states**
> As 5-HT1A receptors have been found to be present both at the neuronal somata and the axon initial segments in spinal ventral horn neurons, including motoneurons [31], it is speculated that 5-HT1A receptors may have heterogeneous functions on motor outputs, i.e., both excitatory and inhibitory [59, 60]. Indeed, using a turtle spinal cord, slice preparation, and intracellular recording technique, Perrier and Cotel [60] demonstrated that the activation of 5- HT1A receptors by 8-OH-DPAT induced an increased excitability in a large fraction (9/11) of sampled motoneurons; whereas in a small fraction of motoneurons (2/11) 8-OH-DPAT gave a hyperpolarizing effect. Subsequent data from the same research group have demonstrated that the inhibitory effect was due to the activation of extrasynaptic 5-HT1A receptors located at the axon initial segments, a mechanism supposed to underlie central fatigue [81, 82].
> *5-HT1B receptors:* To the best of my knowledge, there are no systematic immunohistochemi‐ cal studies on the distribution of 5-HT1B receptors in the spinal cord. Previous studies using autoradiography have demonstrated that 5-HT1B receptors did not have any dominant expression region in the spinal gray matter although the intermediate zone seemed to dwell slightly higher labeled profiles [67, 68]. Although it is difficult to differentiate labeled cell bodies from nerve fibers with autoradiography, results from studies using other methods including lesion, pharmacology, and reverse transcription polymerase chain reaction (RT-PCT) indicated the existence of this receptor subfamily both at presynaptic (primary afferent fibers) [83] and postsynaptic locations (motoneurons) [84, 85]. Recently by using immunohis‐ tochemistry, Wienecke et al. [38] have found that 5-HT1B receptors were indeed expressed in the cell bodies at least in the intermediate zone in the rat spinal cord.
> Similar to 5-HT1A receptors, a major role of 5-HT1B receptors is inhibition in sensory transmission, especially nociception [72, 86, 87]. For example, Eide et al. [72] showed that in mice intrathecal injection of 5-HT1B receptor antagonist RU-24969 (5-methoxy-3(1,2,3,6 tetrahydropyridin-4-yl)-1*H*-indole) induced a significant increase of tail-flick latencies. Another role of 5-HT1B receptors is to inhibit the release of 5-HT from its fibers via autore‐ ceptor mechanism. Although both 5-HT1A and 1B receptors have been demonstrated to be autoreceptors in the brain, in the spinal cord it is most likely that 5-HT1B receptors play a more important role than 5-HT1A receptors in controlling the release of 5-HT from 5-HT fiber terminals [88–90]. Thus, Brown et al. [89] using rat spinal cord synaptosomes preparation showed that the 5-HT1B receptor agonists 1-(*m*-trifluoromethylphenyl)piperazine and 1-(*m*chlorophenyl) piperazine concentration-dependently decreased [3*H*]5-HT release. In addi‐ tion, 5-HT1B receptors could also inhibit the production of 5-HT from AADC cells in the spinal cord through feed-forward mechanism so that the AADC cells do not produce 5-HT in normal physiological states [38]. Another function of 5-HT1B receptors, as noted in mice, is to delay
the maturation of γ-aminobutyric acid (GABA) phenotype in spinal cord during its develop‐ ment [91].
*5-HT1D receptors:* Data concerning the anatomical localization of 5-HT1D receptors in the spinal cord are relatively poor. The available data indicate that 5-HT1D receptors exist both in the dorsal and the ventral horn. Using immunohistochemistry, 5-HT1D receptors were found to be expressed in the primary afferent fibers in the superficial dorsal horn of the rat spinal cord [92, 93]. At the ultrastructural level, 5-HT1D receptors in the spinal cord dorsal horn were found to be localized exclusively within dense core vesicles of synaptic terminals and not the plasma membrane [92]. Using in situ hybridization and transgenic mouse model, Enjin et al. [94] found that 5-HT1D receptors were specifically expressed in γ motoneurons in the ventral horn as well as in some proprioceptive sensory neurons in different spinalregions which were coexpressed with parvalbumin (used as a proprioceptive neuronal marker in the study).
The available data indicate an effect of 5-HT1D receptors in the suppression of sensory, especially nociceptive, information transmission [58, 93, 95]. However, different results were also present. For example, in a formalin-induced hindpaw pain model, Jeong et al. [96] showed that the 5-HT1D receptor agonist GR-46611 (3-[3-(2-dimethylaminoethyl)-1*H*-indol-5-yl]-*N*-(4 methoxybenzyl)acrylamide) did not suppress the formalin-induced flinching responses and the 5-HT1D receptor antagonist BRL-15572 (3-(4-(4-chlorophenyl)piperazin-1-yl)-1,1-diphen‐ yl-2-propanol) failed to reverse the antinociceptive effects of 5-HT. Other functions of 5-HT1D receptors include, e.g., depressing the spinal monosynaptic reflex induced by endogenously released 5-HT [97], and shaping proprioceptive circuits to receive and relay accurate sensory information in the spinal cord motor network [94]. In addition, since 5-HT1D receptors were expressed in γ, but not α, motoneurons, they could be used as a marker to identify γ moto‐ neurons in the spinal cord [94].
*5-HT1F receptors:* To my knowledge, there is no systematic investigation concerning the expression of 5-HT1F receptors in the spinal cord of any species. However, there are indeed some pieces of evidence from studies using radioligand or physiological technique indicat‐ ing that this receptor subfamily is present in the spinal cord dorsal horn. Castro et al. [98] used [3*H*]sumatriptan as a radioligand in the presence of suitable concentrations of 5-carbox‐ amidotryptamine (5-CT) to define 5-HT1F receptors in the human spinal cord, and they found significant levels of binding sites in substantia gelatinosa in the cervical spinal cord. However, because 5-CT is not a specific 5-HT1F agonist, the data cannot be taken as granted. Using an in vitro sacrocaudal spinal cord preparation from spinal transected rats, Murray et al. [99] showed that the EPSPs and associated long polysynaptic reflexes were consistently inhibited by 5-HT1F-specific agonist LY-344864 (*N*-[(3*R*)-3-(dimethylamino)-2,3,4,9-tetrahydro-1*H*carbazol-6-yl]-4-fluorobenzamide), indicating the existence of 5-HT1F in the spinal cord. However, the exact locations of these receptors are yet to be determined.
Due to the lack of related anatomical data, it is difficult to unveil the functions of 5-HT1F in the spinal cord. Considering that 5-HT1F receptors are expressed in the trigeminal and spinal dorsal root ganglia [100] and their agonists have been used to treat migraine [101, 102], it is most likely that this receptor subfamily exerts an antinociceptive effect in spinal cord.
Normal Distribution and Plasticity of Serotonin Receptors after Spinal Cord Injury and Their Impacts on Motor Outputs http://dx.doi.org/10.5772/63759 103
### **3.2. 5-HT2 receptors**
the maturation of γ-aminobutyric acid (GABA) phenotype in spinal cord during its develop‐
*5-HT1D receptors:* Data concerning the anatomical localization of 5-HT1D receptors in the spinal cord are relatively poor. The available data indicate that 5-HT1D receptors exist both in the dorsal and the ventral horn. Using immunohistochemistry, 5-HT1D receptors were found to be expressed in the primary afferent fibers in the superficial dorsal horn of the rat spinal cord [92, 93]. At the ultrastructural level, 5-HT1D receptors in the spinal cord dorsal horn were found to be localized exclusively within dense core vesicles of synaptic terminals and not the plasma membrane [92]. Using in situ hybridization and transgenic mouse model, Enjin et al. [94] found that 5-HT1D receptors were specifically expressed in γ motoneurons in the ventral horn as well as in some proprioceptive sensory neurons in different spinalregions which were coexpressed with parvalbumin (used as a proprioceptive neuronal marker in the study).
The available data indicate an effect of 5-HT1D receptors in the suppression of sensory, especially nociceptive, information transmission [58, 93, 95]. However, different results were also present. For example, in a formalin-induced hindpaw pain model, Jeong et al. [96] showed that the 5-HT1D receptor agonist GR-46611 (3-[3-(2-dimethylaminoethyl)-1*H*-indol-5-yl]-*N*-(4 methoxybenzyl)acrylamide) did not suppress the formalin-induced flinching responses and the 5-HT1D receptor antagonist BRL-15572 (3-(4-(4-chlorophenyl)piperazin-1-yl)-1,1-diphen‐ yl-2-propanol) failed to reverse the antinociceptive effects of 5-HT. Other functions of 5-HT1D receptors include, e.g., depressing the spinal monosynaptic reflex induced by endogenously released 5-HT [97], and shaping proprioceptive circuits to receive and relay accurate sensory information in the spinal cord motor network [94]. In addition, since 5-HT1D receptors were expressed in γ, but not α, motoneurons, they could be used as a marker to identify γ moto‐
*5-HT1F receptors:* To my knowledge, there is no systematic investigation concerning the expression of 5-HT1F receptors in the spinal cord of any species. However, there are indeed some pieces of evidence from studies using radioligand or physiological technique indicat‐ ing that this receptor subfamily is present in the spinal cord dorsal horn. Castro et al. [98] used [3*H*]sumatriptan as a radioligand in the presence of suitable concentrations of 5-carbox‐ amidotryptamine (5-CT) to define 5-HT1F receptors in the human spinal cord, and they found significant levels of binding sites in substantia gelatinosa in the cervical spinal cord. However, because 5-CT is not a specific 5-HT1F agonist, the data cannot be taken as granted. Using an in vitro sacrocaudal spinal cord preparation from spinal transected rats, Murray et al. [99] showed that the EPSPs and associated long polysynaptic reflexes were consistently inhibited by 5-HT1F-specific agonist LY-344864 (*N*-[(3*R*)-3-(dimethylamino)-2,3,4,9-tetrahydro-1*H*carbazol-6-yl]-4-fluorobenzamide), indicating the existence of 5-HT1F in the spinal cord.
Due to the lack of related anatomical data, it is difficult to unveil the functions of 5-HT1F in the spinal cord. Considering that 5-HT1F receptors are expressed in the trigeminal and spinal dorsal root ganglia [100] and their agonists have been used to treat migraine [101, 102], it is
most likely that this receptor subfamily exerts an antinociceptive effect in spinal cord.
However, the exact locations of these receptors are yet to be determined.
ment [91].
102 Recovery of Motor Function Following Spinal Cord Injury
neurons in the spinal cord [94].
As listed in **Table 1**, 5-HT2 receptors include three subfamilies, i.e., 5-HT2A, 2B, and 2C. Before 1993, 5-HT2C receptors were named 5-HT1C receptors [103]. By using different experimental techniques including autoradiography, in situ hybridization, Western blot and immunohistochemistry all these three receptor subfamilies have been detected in the spinal cord in different species [104–109] although 5-HT2C subfamily is probably not expressed in cat spinal cord [105]. 5-HT2A and 2C receptors are likely among the most thoroughly investigated 5-HT receptors in the spinal cord due to theirimportant functions in both sensory information processing and motor control.
*5-HT2A receptors:* Using various techniques, the existence of 5-HT2A receptors has been confirmed at both mRNA and protein level in the spinal cord of different species [104–108, 110]. By using polymerase chain reaction (PCR) together with Southern hybridization technique, Helton et al. [105] showed that spinal cord tissues from rat, cat, monkey, and human contained 5-HT2A receptor mRNAs. By using in situ hybridization, Pompeiano et al. [104] showedthat 5-HT2Areceptor mRNAs were present atintermediatedensity in the ventral horn. This distribution pattern has been confirmed with immunohistochemistry [106, 108]. Doly et al. [108] systematically investigated the distribution of 5-HT2A receptor immunoreactivity in rat spinal cord and the results indicated that 5-HT2A receptors were distributed in allthe spinal segments with a similar immunolabeling pattern. The most prominent labeling was found in lamina IX, where many, but not all, motoneurons were labeled. At lumbar five to six level, 5- HT2A receptors were densely expressed in Onuf's nucleus that contains the motoneurons innervating pelvic striated muscles controlling sexual behavior and micturition [110]. Another region with dense immunolabeling was lamina II. The remaining laminae of the gray matter displayed a weak to moderate labeling. Doly et al. [108] also examined the subcellular localization of 5-HT2A receptors in laminae II and IX and they demonstrated that most immunolabeled profiles were postsynaptic, i.e., dendrites and cell somata, although a small number of axons and axon terminals were also labeled. The immunoreactive product was localized mainly on the plasma membrane where synaptic specifications were lacking.
The localization of 5-HT2A receptors in the neurons in both the spinal dorsal and ventral horn endows their functions in both sensory information transmission and motor control. For the sensory information modulation, evidence to date largely favors in their pronociceptive role [57, 111–114]. For example, Kjørsvik et al. [112] showed that in formalin-induced pain model, intrathecal injection of DOI ((±)-2,5-dimethoxy-4-iodoamphetamine hydrochloride), a 5- HT2A/2C receptor agonist, augmented nociceptive response in rats. This effect could be completely abolished when ketanserin (a 5-HT-2A receptor antagonist) was coadministrated. Rahman et al. [57] investigated the effects of DOI, ketanserin, and ritanserin (another 5-HT2A/ 2C antagonist) on the evoked responses of dorsal horn neurons to electrical, mechanical, and thermal stimulation, and found that the activation of 5-HT2A receptors facilitated the spinal nociceptive transmission under normal physiological states. However, there is also evidence indicating that 5-HT2A receptors exert an antinociceptive effect [115, 116]. For example, using a rat spinal cord slice preparation Xie et al. [115] found that bath application of 5-HT in‐ creased the frequency of spontaneous inhibitory postsynaptic currents in GABAergic and
glycinergic neurons in substantia gelatinosa. TCB-2 (4-bromo-3,6-dimethoxybenzocyclobut‐ en-1-yl)methylamine hydrobromide), a 5-HT2A receptor agonist, could mimic the 5-HT effect, and ketanserin could partially inhibit the effect of 5-HT and completely inhibit the effect of TCB-2.
It is not surprising that 5-HT2A receptors play a significant role in motor control consider‐ ing their intensive expression in ventral horn motoneurons. Indeed, direct stimulation of 5- HT2 receptors in motoneurons in vivo with selective 5-HT2 receptor agonist DOI or DOM (2,5 dimethoxy-alpha,4-dimethylbenzene ethamine hydrochloride) produced dose-related back muscle contractions and wet dog shakes which could be markedly attenuated by ritanserin, ketanserin, or mianserin, suggesting an effect from 5-HT2A receptors [117]. Their direct facilitating effect on spinal motoneurons was also evidenced by application of DOI in rat spinal cord in vitro experiments [118]. In addition, pharmacological experiments showed that 5- HT2A receptor activation facilitated micturition reflex and activated the external urethral sphincter in female rats [119], whereas in both male and female rats 5-HT2A receptors seemed to involve sexual behavior [120, 121].
*5-HT2B receptors:*In comparison with 5-HT2A and 5-HT2C, 5-HT2B receptors in the spinal cord have been less investigated and the available data are inconsistent. Using in situ hybridiza‐ tion expression of 5-HT2B receptors was not found in rat spinal cord and even brain [104]. However, using several other techniques, the receptors have been detected in the spinal cord. Thus, using reverse transcription polymerase chain reaction (RT-PCR) 5-HT2B receptor mRNAs have been detected in spinal cord tissue in many different species including rat, cat, monkey, and human [105]. Using microarray global gene expression technique [122] and immunohistochemistry [123], 5-HT2B receptors have been detected in the rat spinal cord motoneurons. In addition, using Western blot it was shown that the receptor proteins were also present in the dorsal part of the spinal cord [109]. Considering that with Western blot and immunohistochemistry 5-HT2B receptors have been demonstrated to be expressed in the dorsal root ganglia [109, 124], their expression in the spinal dorsal horn is likely originating from the primary afferent fibers from the dorsal root ganglia, although it cannot be excluded that they might be also expressed in neuronal somata (e.g., [125]).
It is presumable that 5-HT2B receptors are largely related with sensory information transmis‐ sion considering their localization in dorsal root ganglia and dorsal horn. Thus, evidence from a few studies related to 5-HT2B receptor functions in the spinal cord favors that the recep‐ tors are responsible for facilitating mechanical hyperalgesia, tactile allodynia, and nocicep‐ tion [109, 124, 126]. Studies concerning the receptors' motor function are rare. One study showed that 5-HT2B receptors likely increased urethral smooth muscle tone since the 5-HT2B receptor antagonist RS-127445 (3-(4-(4-chlorophenyl)piperazin-1-yl)-1,1-diphenyl-2-propa‐ nol) blocked increase in urethral pressure in female rats [127]. In frog, the activation of 5-HT2B receptors by α-methey-5-HT (a 5-HT2B receptor agonist) facilitates *N*-methyl-D-aspartate (NMDA)-induced depolarization of motoneurons [128]. In addition, 5-HT2B receptors seemed to modulate respiratory activity in rats [123].
*5-HT2C receptors:* 5-HT2C receptor mRNAs and proteins were widely distributed in differ‐ ent laminae of the spinal gray matter including motoneurons as demonstrated by different experimental methods, such as PCR [105], autoradiography [68, 129], in situ hybridization [104, 107], and immunohistochemistry [36, 130–132]. Using PCR combined with Southern hybridization 5-HT2C receptor mRNAs were detected in spinal cord tissue from rat, mon‐ key, and human, but not in cat spinal cord [105]. By using in situ hybridization, Fonseca et al. [107] showed that 5-HT2C receptor mRNAs were present at high levels in most parts of the spinal gray matter, except lamina II. Using immunohistochemistry, Ren et al. [36] have investigated 5-HT2C receptor expression in both normal and spinalized rats and the results showed that 5-HT2C receptors were widely distributed in different regions of the spinal gray matter (except lamina II) and were predominantly located in the neuronal somata and their dendrites although they also seemed to be present in axonal fibers in the superficial dorsal horn. Thus, the data from in situ hybridization and immunohistochemistry fit very well for this receptor subfamily.
glycinergic neurons in substantia gelatinosa. TCB-2 (4-bromo-3,6-dimethoxybenzocyclobut‐ en-1-yl)methylamine hydrobromide), a 5-HT2A receptor agonist, could mimic the 5-HT effect, and ketanserin could partially inhibit the effect of 5-HT and completely inhibit the effect of
It is not surprising that 5-HT2A receptors play a significant role in motor control consider‐ ing their intensive expression in ventral horn motoneurons. Indeed, direct stimulation of 5- HT2 receptors in motoneurons in vivo with selective 5-HT2 receptor agonist DOI or DOM (2,5 dimethoxy-alpha,4-dimethylbenzene ethamine hydrochloride) produced dose-related back muscle contractions and wet dog shakes which could be markedly attenuated by ritanserin, ketanserin, or mianserin, suggesting an effect from 5-HT2A receptors [117]. Their direct facilitating effect on spinal motoneurons was also evidenced by application of DOI in rat spinal cord in vitro experiments [118]. In addition, pharmacological experiments showed that 5- HT2A receptor activation facilitated micturition reflex and activated the external urethral sphincter in female rats [119], whereas in both male and female rats 5-HT2A receptors seemed
*5-HT2B receptors:*In comparison with 5-HT2A and 5-HT2C, 5-HT2B receptors in the spinal cord have been less investigated and the available data are inconsistent. Using in situ hybridiza‐ tion expression of 5-HT2B receptors was not found in rat spinal cord and even brain [104]. However, using several other techniques, the receptors have been detected in the spinal cord. Thus, using reverse transcription polymerase chain reaction (RT-PCR) 5-HT2B receptor mRNAs have been detected in spinal cord tissue in many different species including rat, cat, monkey, and human [105]. Using microarray global gene expression technique [122] and immunohistochemistry [123], 5-HT2B receptors have been detected in the rat spinal cord motoneurons. In addition, using Western blot it was shown that the receptor proteins were also present in the dorsal part of the spinal cord [109]. Considering that with Western blot and immunohistochemistry 5-HT2B receptors have been demonstrated to be expressed in the dorsal root ganglia [109, 124], their expression in the spinal dorsal horn is likely originating from the primary afferent fibers from the dorsal root ganglia, although it cannot be excluded
It is presumable that 5-HT2B receptors are largely related with sensory information transmis‐ sion considering their localization in dorsal root ganglia and dorsal horn. Thus, evidence from a few studies related to 5-HT2B receptor functions in the spinal cord favors that the recep‐ tors are responsible for facilitating mechanical hyperalgesia, tactile allodynia, and nocicep‐ tion [109, 124, 126]. Studies concerning the receptors' motor function are rare. One study showed that 5-HT2B receptors likely increased urethral smooth muscle tone since the 5-HT2B receptor antagonist RS-127445 (3-(4-(4-chlorophenyl)piperazin-1-yl)-1,1-diphenyl-2-propa‐ nol) blocked increase in urethral pressure in female rats [127]. In frog, the activation of 5-HT2B receptors by α-methey-5-HT (a 5-HT2B receptor agonist) facilitates *N*-methyl-D-aspartate (NMDA)-induced depolarization of motoneurons [128]. In addition, 5-HT2B receptors seemed
*5-HT2C receptors:* 5-HT2C receptor mRNAs and proteins were widely distributed in differ‐ ent laminae of the spinal gray matter including motoneurons as demonstrated by different
that they might be also expressed in neuronal somata (e.g., [125]).
to modulate respiratory activity in rats [123].
TCB-2.
to involve sexual behavior [120, 121].
104 Recovery of Motor Function Following Spinal Cord Injury
The wide distribution of 5-HT2C receptors in different laminae of the spinal cord endows the receptors' role both in sensory information transmission and motor control. For the sensory information transmission available data point to that 5-HT2C receptors play roles both in pronociception [57, 112] and antinociception [96, 115, 133]. For their role on pronociception, referto the part concerning 5-HT2A receptors since most studies used DOI (a common agonist for 5-HT2A and 2C receptors) to stimulate and ketanserin to inhibit the receptors in re‐ sponse to different nociceptive stimulations. However, because ketanserin is mainly a 5-HT2A receptor antagonist some researchers argued that the pronociception effect should prefera‐ bly attribute to 2-HT2A than 2C receptors. Their effects on antinociception seem to be adequately evidenced. For instance, in one study Jeong et al. [96] showed that in formalininduced pain model 5-HT2C receptor specific antagonist D-MC (*N*-ormethylclozapine/8 chloro-11-(1-piperazinyl)-5*H*-dibenzo[*b,e*][1,4]diazepine) could block the suppression effect of 5-HT, whereas 5-HT2C receptor specific agonist MK-212 (6-chloro-2-(1-piperazin‐ yl)pyrazine hydrochloride) could suppress the formalin response. In another study, Xie et al. [115] showed that a 5-HT2C receptor agonist WAY-161503 (8,9-dichloro-2,3,4,4a-tetrahy‐ dro-1*H*-pyrazino[1,2-*a*]quinoxalin-5(6*H*)) mimicked the 5-HT antinociceptive effect in dorsal horn neurons and this effect could be blocked by a 5-HT2C receptor antagonist, *N*-desmethyl‐ clozapine.
In motor control, one of the functions of 5-HT2C receptors is the induction of a long-lasting amplification of spinal reflex [73, 134–136]. Machacek et al. [134] showed that, in an in vitro neonatal rat spinal cord preparation, superfusion of 5-HT depressed reflex responses record‐ ed in the ventral roots which was induced by electrical stimulation of primary afferents. However, following 5-HT washout, a long-lasting reflex facilitation was observed. Further pharmacological analysis indicated that it was the activation of 5-HT2C but not 5-HT2A receptors that was required for this long-lasting reflex. Although 5-HT2C receptors play a major role in long-term motor reflexes, they also play an inhibitory role in some physiologi‐ cal states. For example, in mice, when activated by DOI, 5-HT2C receptors inhibited the locomotor activity which opposed the effects of 5-HT2A receptors [137]. In addition, 5-HT2C receptors were also demonstrated to inhibit the micturition reflex [119, 127, 138]. Thus, Conlon et al. [138] showed that, in guinea pigs, Ro-600175 ((α*S*)-6-chloro-5-fluoro-α-methyl-1*H*- indole-1-ethanamine fumarate), a 5-HT2C receptor agonist, increased peak urethral pressure in a dose-dependent manner. This effect was reversed by a selective 5-HT2C receptor antagonist SB-242084 (6-chloro-2,3-dihydro-5-methyl-*N*-[6-[(2-methyl-3-pyridinyl)oxy]-3 pyridinyl]-1*H*-indole-1-carboxyamide) but not a 5-HT2A or 2B receptor antagonist. Similar results were also observed in rats [119, 127]. Clinically, duloxetine, a serotonin-norepinephr‐ ine reuptake inhibitor, has been used to treat stress urinary incontinence. One possible mechanism for this effect may be the activation of 5-HT2C receptors in motoneurons in Onuf's nucleus, which leads to an increased activity of pudendal motor neurons and a subsequent increase in the strength of urethral sphincter contractions.
### **3.3. 5-HT3 receptors**
Although genes encoding the molecular structures of five different 5-HT3 receptor subfami‐ lies (A–E) have been cloned, it is demonstrated so far that only the 5-HT3A subfamily is functional in homomeric form, and all other subfamilies require coassembling with 5-HT3A as functional heteromers [62, 139]. Therefore, here 5-HT3 receptors will be described togeth‐ er as one receptor unity. In the spinal cord, 5-HT3 receptors in the beginning were detected mainly in the dorsal horn with autoradiography and immunohistochemistry in rats and humans [140, 141]. Later by using in situ hybridization and immunohistochemistry, they were found to be expressed in different laminae across the spinal gray matter [107, 142]. By using immunohistochemistry, Morales et al. [142] observed labeled cell bodies in the dorsal horn among the densely labeled fiber terminals. In the ventral horn, large neurons, likely moto‐ neurons, were densely labeled. By using in situ hybridization, Fonseca et al. [107] found that 5-HT3 receptor mRNAs were expressed in different laminae in the rat spinal cord with very low levels in the dorsal horn, slightly higher levels in laminae VI through X and the highest level in lamina IX. Maxwell et al. [143] have studied the ultrastructure of 5-HT3 immunola‐ beled profiles in the dorsal horn of the rat spinal cord. They found that 5-HT3-immunoposi‐ tive terminals invariably formed asymmetric synaptic junctions with dendritic profiles and often contained a mixture of granular and agranular vesicles. Immunoreactive cells were found to contain intense patches of reaction product within their cytoplasm. Although there are no data available regarding the subcellular morphology of 5-HT3 receptors in the ventral horn neurons, from the light microscopic data, it can be concluded that the receptor immunoreac‐ tive product was located at least in the cytoplasm [142].
It has long been known that 5-HT3 receptors modulate spinal nociceptive reflexes [144]. A large part of literature demonstrated a role of 5-HT3 receptors in antinociception in the spinal cord (e.g., [145, 146]). This is likely due to, for instance, that the receptors interact with inhibitory interneurons such as GABA interneurons in the dorsal horn [147]. However, there are also data indicated that 5-HT3 receptors have a pronociceptive effect. For example, Guo et al. [148] showed that when the spinal 5-HT3 receptors were activated by intrathecal injection of a selective 5-HT3 receptor agonist SR-57227 (1-(6-chloropyridin-2-yl)piperidin-4-amine) spinal glial hyperactivity, neuronal hyperexcitability, and pain hypersensitivity were in‐ duced in rats. These diverse effects could be explained by, e.g., the expression of 5-HT3 receptors in different neuronal components and/or the different effects of assemblies formed from different receptor subfamilies [149, 150].
Although 5-HT3 receptors are expressed in the ventral horn neurons data concerning their motor functions are scarce. A study by Guertin and Steuer [151] showed that in hindlimb paralyzed mice 5-HT3 receptor agonist SR-57227 could produce hindlimb movements although with a low score. This result indicates that 5-HT3 receptors might also modulate spinal motoneuron activity in normal states. In addition, the activation of 5-HT3 receptors by 2 methyl-5-HT, another 5-HT3 receptor agonist, could induce sideward tail-flick reflex in rats [117], which might likely involve both sensory and motor components in the spinal cord. Further, 5-HT3 receptors have been reported to inhibit micturition in cats in both normal and spinalized situation [152].
### **3.4. 5-HT4 receptors**
indole-1-ethanamine fumarate), a 5-HT2C receptor agonist, increased peak urethral pressure in a dose-dependent manner. This effect was reversed by a selective 5-HT2C receptor antagonist SB-242084 (6-chloro-2,3-dihydro-5-methyl-*N*-[6-[(2-methyl-3-pyridinyl)oxy]-3 pyridinyl]-1*H*-indole-1-carboxyamide) but not a 5-HT2A or 2B receptor antagonist. Similar results were also observed in rats [119, 127]. Clinically, duloxetine, a serotonin-norepinephr‐ ine reuptake inhibitor, has been used to treat stress urinary incontinence. One possible mechanism for this effect may be the activation of 5-HT2C receptors in motoneurons in Onuf's nucleus, which leads to an increased activity of pudendal motor neurons and a subsequent
Although genes encoding the molecular structures of five different 5-HT3 receptor subfami‐ lies (A–E) have been cloned, it is demonstrated so far that only the 5-HT3A subfamily is functional in homomeric form, and all other subfamilies require coassembling with 5-HT3A as functional heteromers [62, 139]. Therefore, here 5-HT3 receptors will be described togeth‐ er as one receptor unity. In the spinal cord, 5-HT3 receptors in the beginning were detected mainly in the dorsal horn with autoradiography and immunohistochemistry in rats and humans [140, 141]. Later by using in situ hybridization and immunohistochemistry, they were found to be expressed in different laminae across the spinal gray matter [107, 142]. By using immunohistochemistry, Morales et al. [142] observed labeled cell bodies in the dorsal horn among the densely labeled fiber terminals. In the ventral horn, large neurons, likely moto‐ neurons, were densely labeled. By using in situ hybridization, Fonseca et al. [107] found that 5-HT3 receptor mRNAs were expressed in different laminae in the rat spinal cord with very low levels in the dorsal horn, slightly higher levels in laminae VI through X and the highest level in lamina IX. Maxwell et al. [143] have studied the ultrastructure of 5-HT3 immunola‐ beled profiles in the dorsal horn of the rat spinal cord. They found that 5-HT3-immunoposi‐ tive terminals invariably formed asymmetric synaptic junctions with dendritic profiles and often contained a mixture of granular and agranular vesicles. Immunoreactive cells were found to contain intense patches of reaction product within their cytoplasm. Although there are no data available regarding the subcellular morphology of 5-HT3 receptors in the ventral horn neurons, from the light microscopic data, it can be concluded that the receptor immunoreac‐
It has long been known that 5-HT3 receptors modulate spinal nociceptive reflexes [144]. A large part of literature demonstrated a role of 5-HT3 receptors in antinociception in the spinal cord (e.g., [145, 146]). This is likely due to, for instance, that the receptors interact with inhibitory interneurons such as GABA interneurons in the dorsal horn [147]. However, there are also data indicated that 5-HT3 receptors have a pronociceptive effect. For example, Guo et al. [148] showed that when the spinal 5-HT3 receptors were activated by intrathecal injection of a selective 5-HT3 receptor agonist SR-57227 (1-(6-chloropyridin-2-yl)piperidin-4-amine) spinal glial hyperactivity, neuronal hyperexcitability, and pain hypersensitivity were in‐ duced in rats. These diverse effects could be explained by, e.g., the expression of 5-HT3
increase in the strength of urethral sphincter contractions.
106 Recovery of Motor Function Following Spinal Cord Injury
tive product was located at least in the cytoplasm [142].
**3.3. 5-HT3 receptors**
Data concerning 5-HT4 receptors in the spinal cord are scarce. As far as I know, so far only one article reported the expression of 5-HT4 receptors in the spinal cord [153]. Using immunohis‐ tochemistry, Suwa et al. [153] investigated 5-HT4 receptor cell distribution in the brain and spinal cord in juvenile rats and found a high density of immunostained neurons in the ventral horn of the spinal cord. In addition, several sympathetic neurons were also seen to be immunopositive.
Similarly, studies about the functions of 5-HT4 receptors in the spinal cord are also rare. Using pharmacological method, Godínez-Chaparro et al. [154] showed that 5-HT in the spinal cord promotes the development and maintenance of secondary allodynia and hyperalgesia caused by formalin stimulation via the activation of 5-HT4/6/7 receptors. Considering the anatomi‐ cal distribution of 5-HT4 receptors in the spinal cord, they should also play a role in some motor functions. Whether this is the case needs to be investigated further.
### **3.5. 5-HT5 receptors**
The 5-HT5 receptor family comprises two members: 5-HT5A and 5-HT5B. Although 5-HT5B receptors have been found in some structures in rat and mouse brain, they are not expressed in humans [155]. Therefore, I will only describe the 5-HT5A receptors in this chapter.
5-HT5A receptors were first identified in the rat spinal cord by Doly et al. [156]. They showed that in the rat spinal cord, 5-HT5A receptors were expressed with high density in the super‐ ficial dorsal horn (laminae I and II) especially in lamina II. In addition, they were also expressed in other regions including the intermediolateral nucleus in the thoracolumbar region and Onuf's nucleus in lumbosacral region. Ventral horn motoneurons in different regions were weakly labeled. Subcellularly, the receptors were found both in the cytoplasm and on the cell membrane of neuronal somata and dendrites. In the cytoplasm, they were associated with Golgi apparatus, rough endoplasmic reticulum, and vesicles. On the membrane, they were exclusively located on the postsynaptic density that was in clear contrast with the subcellu‐ lar localization of 5-HT2A receptors [108, 110].
According to theirlocations in the spinal cord, itis speculatedthat 5-HT5Areceptors are related to spinal modulation of pain, autonomic function, and control of micturition. Indeed, data from recent pharmacological studies have demonstrated that 5-HT-induced antinociceptive effect was mediated by spinal 5-HT5A receptors in several pain models, such as that induced by formalin, capsaicin, or acetic acid [58, 157]. For example, injection of 5-HT or 5-CT (an agonist of 5-HT5A receptors) could dose-dependently prevent nociception induced by formalin [157]. However, the speculated effects of the receptors on autonomic function and control of micturition yet need to be demonstrated.
### **3.6. 5-HT6 receptors**
Using RT-PCR, Gérard et al. [158] discovered the existence of 5-HT6 receptor mRNAs at a moderate level in the rat spinal cord. Later, using immunoautoradiogram, data from the same group showed that 5-HT6 receptor immunoreactive product seemed to be present across all laminae of the spinal gray matter with a denser labeling in the superficial dorsal horn and lamina IX [159]. No data are available about the subcellular distribution of 5-HT6 receptors in the spinal cord. However, according to the electron microscopic data from other brain regions such as the striatum and the hippocampus, 5-HT6 receptors were mainly associated with postsynaptic dendrites and no immunopositive axon terminals were found [159].
The exclusive roles of 5-HT6 receptors in the spinal cord yet remain to be elaborated. Using a formalin pain model in rats, Castañeda-Corral et al. [160] suggested that 5-HT6 receptors play a pronociceptive role in the spinal cord since intrathecal injection of EMD-386088 (5-chloro-2 methyl-3-(1,2,3,6-tetrahydro-4-pyridinyl)-1*H*-indole), a selective 5-HT6 receptor agonist, enhanced formalin-induced nociception. It is unknown whether 5-HT6 receptors play any role in motor control.
### **3.7. 5-HT7 receptors**
As there were no specific radioligands for 5-HT7 receptors available, data concerning their distribution in the spinal cord acquired using autoradiography had been inconclusive [161] until 2005 when Doly et al. [162] used immunohistochemistry to study their distribution in the rat spinal cord. Doly et al. [162] showed that in the rat lumbar spinal cord 5-HT7 receptors were mainly located in two superficial laminae of the dorsal horn. Except in the Onuf's nucleus where the labeling was relatively denser, immunolabeling in the ventral horn motoneuron region was generally weak. At subcellular level, the receptors were found in the postsynap‐ tic locations in neuronal cell bodies and dendrites as well as presynaptic locations in unmye‐ linated and thin myelinated axonal fibers. In addition, immunolabeling was also found in astrocytes. In cats, Noga et al. [16] found that in thoracolumbar spinal cord 5-HT7-immuno‐ labeled cells spread across different laminae in the gray matter with a dorsoventral density gradient.
Available data showed that 5-HT7 receptors exert multiple roles in modulating both sensory and motor behavior. In sensory aspect 5-HT7 receptors have been demonstrated to exert dual but diverse actions in nociception depending on the different situations. In healthy rats, 5-HT7
receptor agonists exerted a pronociceptive action but in neuropathic animals they exerted an antinociceptive action [163]. Yesilyurt et al. [164] showed that intrathecal application of SB-269970 ((2*R*)-1-[(3-hydroxyphenyl)sulfonyl]-2 -(2-(4-methyl-1-piperidinyl)ethyl)pyrroli‐ dine), a 5-HT7 receptor antagonist, blocked both opioid and nonopioid type stress-induced analgesia, and this effect was mediated by descending serotonergic pathways and the spinal 5-HT7 receptors. Several pieces of evidence have indicated that the nonopioid analgesic drug Nefopam also took its effects via activation of 5-HT7 receptors [165, 166].
5-HT7 receptors have been demonstrated to be critical for 5-HT-induced locomotor-like activity [16, 167–169]. Liu and Jordan [167] showed that in rats 5-HT7 receptor antagonists blocked locomotor-like activity induced by stimulating mid-medulla region and also de‐ creased step cycle duration. These results have been verified in 5-HT7 knockout mice [169]. Considering that 5-HT7 receptor antagonists blocked locomotor-like activity when applied only above the L3 segments [167], it is possible that they activated the central pattern genera‐ tor neurons above this level that expressed 5-HT7 receptors and responded to 5-HT stimula‐ tion. Indeed, in decerebrated cats in which locomotion was induced by electrical stimulation of the mesencephalic locomotor region abundant c-Fos immunoreactive cells were observed in laminae VII and VIII throughout the thoracolumbar segments [16].
### **4. Expression changes of different 5-HT receptors and their significance in motor outputs following SCI**
Following SCI the descending 5-HT projections are interrupted and as a consequence, a number of 5-HT receptors undergo different degrees of plastic changes below the lesion depending on the severity of injury, which on one hand will facilitate the re-establishment of neuronal circuits in the spinal cord below the lesion and thus promote functional recovery, and on the other hand will result in pathological symptoms. These changes can occur at both anatomical and functional levels. However, due to the fact that more functional than anatom‐ ical data are available, for some receptors only functional changes have been reported. So far 5-HT receptors that have been confirmed to undergo anatomical changes include, but are not limited to, 5-HT1 and 5-HT2 receptor families. In some other receptors, functional (activity) changes are indeed detected although there are no data revealing their anatomical changes. This group of receptors includes 5-HT3 and 5-HT7 receptors. So far there are no reports as to the anatomical or functional changes for 5-HT4, 5-HT5, and 5-HT6 receptors following SCI. Below I will mainly describe the 5-HT receptors that have been clearly demonstrated to undergo anatomical and/or functional changes in relation to their impact on motor outputs.
### **4.1. 5-HT1 receptors**
According to theirlocations in the spinal cord, itis speculatedthat 5-HT5Areceptors are related to spinal modulation of pain, autonomic function, and control of micturition. Indeed, data from recent pharmacological studies have demonstrated that 5-HT-induced antinociceptive effect was mediated by spinal 5-HT5A receptors in several pain models, such as that induced by formalin, capsaicin, or acetic acid [58, 157]. For example, injection of 5-HT or 5-CT (an agonist of 5-HT5A receptors) could dose-dependently prevent nociception induced by formalin [157]. However, the speculated effects of the receptors on autonomic function and control of
Using RT-PCR, Gérard et al. [158] discovered the existence of 5-HT6 receptor mRNAs at a moderate level in the rat spinal cord. Later, using immunoautoradiogram, data from the same group showed that 5-HT6 receptor immunoreactive product seemed to be present across all laminae of the spinal gray matter with a denser labeling in the superficial dorsal horn and lamina IX [159]. No data are available about the subcellular distribution of 5-HT6 receptors in the spinal cord. However, according to the electron microscopic data from other brain regions such as the striatum and the hippocampus, 5-HT6 receptors were mainly associated with
The exclusive roles of 5-HT6 receptors in the spinal cord yet remain to be elaborated. Using a formalin pain model in rats, Castañeda-Corral et al. [160] suggested that 5-HT6 receptors play a pronociceptive role in the spinal cord since intrathecal injection of EMD-386088 (5-chloro-2 methyl-3-(1,2,3,6-tetrahydro-4-pyridinyl)-1*H*-indole), a selective 5-HT6 receptor agonist, enhanced formalin-induced nociception. It is unknown whether 5-HT6 receptors play any role
As there were no specific radioligands for 5-HT7 receptors available, data concerning their distribution in the spinal cord acquired using autoradiography had been inconclusive [161] until 2005 when Doly et al. [162] used immunohistochemistry to study their distribution in the rat spinal cord. Doly et al. [162] showed that in the rat lumbar spinal cord 5-HT7 receptors were mainly located in two superficial laminae of the dorsal horn. Except in the Onuf's nucleus where the labeling was relatively denser, immunolabeling in the ventral horn motoneuron region was generally weak. At subcellular level, the receptors were found in the postsynap‐ tic locations in neuronal cell bodies and dendrites as well as presynaptic locations in unmye‐ linated and thin myelinated axonal fibers. In addition, immunolabeling was also found in astrocytes. In cats, Noga et al. [16] found that in thoracolumbar spinal cord 5-HT7-immuno‐ labeled cells spread across different laminae in the gray matter with a dorsoventral density
Available data showed that 5-HT7 receptors exert multiple roles in modulating both sensory and motor behavior. In sensory aspect 5-HT7 receptors have been demonstrated to exert dual but diverse actions in nociception depending on the different situations. In healthy rats, 5-HT7
postsynaptic dendrites and no immunopositive axon terminals were found [159].
micturition yet need to be demonstrated.
108 Recovery of Motor Function Following Spinal Cord Injury
**3.6. 5-HT6 receptors**
in motor control.
gradient.
**3.7. 5-HT7 receptors**
Among the different 5-HT1 subfamilies, 5-HT1A receptors were mostly investigated in terms of their anatomical and functional changes probably due to their direct involvement in functional recovery following SCI. For other subfamilies, such as 5-HT1B, 1D, and 1F, related studies were less abundant and results relating to their roles in motor functional outputs were also less conclusive.
*5-HT1A receptors:* Using autoradiography, the expression of 5-HT1A receptors have been shown to increase in the rat spinal cord 3 weeks following destruction of descending seroto‐ nergic fibers with 5,7-dihydroxytryptamine [83]. In cats, using the same technique, it was found that after spinal cord transection at T13 level binding density was significantly in‐ creased in laminae II, III, andXoflumbar segments at 15 and 30 days, but at 60 days the binding density recovered to the control level [28]. It should be addressed that using binding techni‐ que it is difficult to differentiate whether the labeled profiles were from neuronal cell bodies, their dendrites or axon fiber terminals. Using immunohistochemistry with two different 5- HT1A antibodies, one labeling neuronal somata and the other labeling the axon initial segments, Otoshi et al. [31] reported that 8 weeks following complete spinal transection at T7- T8 level the expression of 5-HT1A receptors in axon hillock was increased in laminae III, IV, VII, and IX, whereas the expression in neuronal somata and dendrites were increased in laminae VII and IX. This upregulation, both in the axon hillock and neuronal somata and dendrites, was dependent on the sensory input since the receptors were not upregulated when the spinal cord was isolated. Using in situ hybridization, Cornide-Petronio et al. [170] saw a similar upregulation time course for 5-HT1A receptors in lampreys as in cats. They reported an acute upregulation of 5-HT1A receptors in the spinal cord after SCI both in the rostral and caudal part of the lesion site, and the upregulation recovered to normal levels at 3 weeks. This quick recovery of 5-HT1A receptor expression may be due to that in lamprey serotonergic descending projecting neurons in the rhombencephalon could regenerate their axons across the lesion site after complete spinal cord transection [171]. Thus, the variation of results from these studies may partly reflect the different receptor responses to SCI in different species.
There is ample evidence that the activation of 5-HT1A receptors could induce motor func‐ tional recovery after SCI [13, 84, 172]. Antri et al. [13] found that, in thoracic spinal cord transected rats, after daily systemic application of 5-HT1A receptor agonist 8-OH-DPAT, locomotor function was significantly improved when compared with control spinalized animals. The agonist had both short- and long-term effects for motor functional improve‐ ment. A similarresult was also demonstrated by Jackson and White [84] that in acute C1 spinal cord transected rats, afterintravenous administration of 8-OH-DPAT, the excitability of spinal motoneurons was markedly enhanced. However, local application of 8-OH-DPAT directly into the ventral horn by microiontophoresis inhibited the glutamate-evoked firing of moto‐ neurons. These results indicate that when 8-OH-DPAT was directly applied into the vicinity of the motoneurons it may activate 5-HT1A receptors in the axon hillock which inhibits the motoneuron firing [81]. However, the marked increase in firing of motoneurons induced by systemic administration of 8-OH-DPAT suggests that 5-HT1A receptors in other locations, likely at sites presynaptic to the motoneurons, have also been activated. In addition to the functions described above, 5-HT1A receptors have been reported to increase bladder capaci‐ ty under saline or acid infused conditions in SCI cats [173].
*5-HT1B receptors:* In comparison with 5-HT1A receptors, studies concerning plastic changes of 5-HT1B receptors following SCI are sparse. Using autoradiography, Laporte et al. [83]
reported that when descending serotonergic projections were destroyed by 5,7-dihydroxy‐ tryptamine the labeling of 5-HT1B receptors was decreased (−12%) in the dorsal horn at the cervical but not at the lumbarlevel in rats. Interestingly, when noradrenergic systems had been lesioned by DSP-4 (*N*-(2-chloroethyl)-*N*-ethyl-2-bromobenzylamine) there was an increase of labeling of 5-HT1B receptors both at the cervical level (+31%) and the lumbar level (+17%). In a C2 hemisection rat model, using quantitative RT-PCR, Mantilla et al. [86] did not detect expression changes of 5-HT1B receptors in phrenic motoneurons 2 or 3 weeks following the injury.
Similar to their functions in normal physiological states, in SCI states 5-HT1B receptors also play an inhibitory role for motor outputs possibly through inhibitions on both mono- and polysynaptic reflexes. In acute C1 spinalized rats, Honda et al. [174] showed that serotoner‐ gic depression of monosynaptic reflex transmission induced by application of 5-HTP was mediated by 5-HT1B receptors. In S2 spinal cord transection model (also called tail spasticity model), Murray et al. [99] showed that polysynaptic EPSPs that trigger muscle spasms after SCI were inhibited by 5-HT1B receptors.
*5-HT1D and 1F receptors:* To my knowledge, no studies have investigated the expression changes of these two receptors following SCI. Functionally, the activation of 5-HT1D recep‐ tors seemed to reduce sensory transmission in both monosynaptic and polysynaptic reflexes in humans [95], and to inhibit bladder activity in cats after SCI [175]. The activation of 5-HT1F seemed to inhibit long polysynaptic reflexes in rats after SCI [99]. However because the agonists used in these studies were not specific enough for selected receptors the results were not conclusive.
### **4.2. 5-HT2 receptors**
studies were less abundant and results relating to their roles in motor functional outputs were
*5-HT1A receptors:* Using autoradiography, the expression of 5-HT1A receptors have been shown to increase in the rat spinal cord 3 weeks following destruction of descending seroto‐ nergic fibers with 5,7-dihydroxytryptamine [83]. In cats, using the same technique, it was found that after spinal cord transection at T13 level binding density was significantly in‐ creased in laminae II, III, andXoflumbar segments at 15 and 30 days, but at 60 days the binding density recovered to the control level [28]. It should be addressed that using binding techni‐ que it is difficult to differentiate whether the labeled profiles were from neuronal cell bodies, their dendrites or axon fiber terminals. Using immunohistochemistry with two different 5- HT1A antibodies, one labeling neuronal somata and the other labeling the axon initial segments, Otoshi et al. [31] reported that 8 weeks following complete spinal transection at T7- T8 level the expression of 5-HT1A receptors in axon hillock was increased in laminae III, IV, VII, and IX, whereas the expression in neuronal somata and dendrites were increased in laminae VII and IX. This upregulation, both in the axon hillock and neuronal somata and dendrites, was dependent on the sensory input since the receptors were not upregulated when the spinal cord was isolated. Using in situ hybridization, Cornide-Petronio et al. [170] saw a similar upregulation time course for 5-HT1A receptors in lampreys as in cats. They reported an acute upregulation of 5-HT1A receptors in the spinal cord after SCI both in the rostral and caudal part of the lesion site, and the upregulation recovered to normal levels at 3 weeks. This quick recovery of 5-HT1A receptor expression may be due to that in lamprey serotonergic descending projecting neurons in the rhombencephalon could regenerate their axons across the lesion site after complete spinal cord transection [171]. Thus, the variation of results from these studies may partly reflect the different receptor responses to SCI in different species.
There is ample evidence that the activation of 5-HT1A receptors could induce motor func‐ tional recovery after SCI [13, 84, 172]. Antri et al. [13] found that, in thoracic spinal cord transected rats, after daily systemic application of 5-HT1A receptor agonist 8-OH-DPAT, locomotor function was significantly improved when compared with control spinalized animals. The agonist had both short- and long-term effects for motor functional improve‐ ment. A similarresult was also demonstrated by Jackson and White [84] that in acute C1 spinal cord transected rats, afterintravenous administration of 8-OH-DPAT, the excitability of spinal motoneurons was markedly enhanced. However, local application of 8-OH-DPAT directly into the ventral horn by microiontophoresis inhibited the glutamate-evoked firing of moto‐ neurons. These results indicate that when 8-OH-DPAT was directly applied into the vicinity of the motoneurons it may activate 5-HT1A receptors in the axon hillock which inhibits the motoneuron firing [81]. However, the marked increase in firing of motoneurons induced by systemic administration of 8-OH-DPAT suggests that 5-HT1A receptors in other locations, likely at sites presynaptic to the motoneurons, have also been activated. In addition to the functions described above, 5-HT1A receptors have been reported to increase bladder capaci‐
*5-HT1B receptors:* In comparison with 5-HT1A receptors, studies concerning plastic changes of 5-HT1B receptors following SCI are sparse. Using autoradiography, Laporte et al. [83]
ty under saline or acid infused conditions in SCI cats [173].
also less conclusive.
110 Recovery of Motor Function Following Spinal Cord Injury
5-HT2, especially 5-HT2A and 2C, receptors are the most intensely studied receptors among all 5-HT receptor families mostly due to their considerable effects on the motor functional recovery and perhaps also on the pathological symptoms developed following SCI such as spasticity. Due to the close relationship and similar functions of these three receptor subfami‐ lies, their influences on the motor outputs after SCI will be described together.
*5-HT2A receptors:* Using different SCI animal models such as contusion, hemisection, and complete spinal transection, it is confirmed that 5-HT2A receptors underwent different degrees of plastic changes following SCI. In thoracic contusive rats using immunohistochem‐ istry, Lee et al. [30] found a moderate yet significantly increased 5-HT2A receptor expres‐ sion (~1.2-fold of control) in the motoneurons at L5-L6 level after 4 weeks of injury. With the same method in a C2 hemisection rat model Fuller et al. [29] demonstrated a significant upregulation (~1.7-fold of control) of 5-HT2A receptors in the ipsilateral phrenic motoneur‐ ons after 2 weeks of injury. In a S2 spinal transection model, Kong et al. [32, 33] reported a dramatic increase of 5-HT2A receptor immunoreactivity (~5.6-fold of control) in the moto‐ neurons below the lesion. The upregulation began as early as 1 day after injury, reached a maximal level by 28 days and lasted at least until 60 days, the longest time interval investi‐ gated. Similarly an upregulation of 5-HT2A receptors were also observed in chronic (6 weeks) thoracic spinal transected rats. Navarrett et al. [35] reported a ~1.3-fold increase of 5-HT2A receptor mRNAs in the spinal tissue (including both white and gray matter) below the lesion with RT-PCR. Upregulation of 5-HT2A receptor mRNAs was also reported in mice subject‐ ed to an acute thoracic spinal transection. Thus, Ung et al. [176] found that 3 h afterthe lesion 5- HT2A receptor mRNA expression levels were increased 2.5-fold in the lateral intermediate zone in L1-L2 segments and they remained to be elevated for at least 2 weeks. However, no significant change was found in the ventral horn motoneuron regions, which was strikingly different from that in the rat.
*5-HT2B receptors:* No expression changes have been reported for this receptor subfamily in the spinal cord following SCI. With microarray global gene analysis technique, Wienecke et al. [122] did not detect significant changes of 5-HT2B receptor mRNAs in S2 spinal transected rats.
*5-HT2C receptors:* Similarto 5-HT2A receptors, 5-HT2C receptor expression changes have been investigated in different SCI animal models with different techniques. However, discrepan‐ cies exist as to whether 5-HT2C receptors are upregulated or not following SCI. In chronic SCI rats (~15 weeks) severe contusion or transection at thoracic level (T9) could induce a signifi‐ cant increase of 5-HT2C receptor immunoreactivity in lumbar spinal cord [22]. The increase could reach 4–7-fold in the ventral horn and ~17-fold in the dorsal horn over control level. In thoracic (T8-T9) spinal transected neonatal rats, following 4 weeks injury 5-HT2C receptor immunoreactivity was seen to increase 4–5-fold both in the spinal dorsal and ventral horn [177]. In a S2 spinal transection rat model, data from our group [36] indicated that 5-HT2C receptor immunoreactivity increased in all parts of the spinal gray matter below the lesion from 2 weeks but did notreach a significant level until 3 weeks (~1.4-fold over control animals). The increase sustained thereafter and a maximal level was reached at 45 days (~1.7-fold) and maintained at 60 days, the longest investigated interval. It is somehow perplexing that although upregulation was observed at protein level no significant increase at mRNA level was detected in different SCI animal models [34, 35, 122, 178]. However, although no signifi‐ cant changes of the total amount of 5-HT2C receptor mRNAs were detected, Murray et al. [34] did detect certain constitutive isoforms, such as INI isoform, that were significantly in‐ creased in the motoneurons 6 weeks after S2 spinal transection.
*Functional significance for motor outputs of 5-HT2 receptors:* The activation of 5-HT2 receptors has different effects on motor outputs after SCI, which include promoting motor functional recovery and causing maladaptive pathological motor symptoms such as spasticity. Due to uncertainty of the expression of 5-HT2B receptors in the spinal cord most studies have been focused on 5-HT2A and 2C receptors. There is ample evidence for the roles of 5-HT2A and 2C receptors in motor functional recovery following SCI from different studies (e.g., [9, 12, 14, 34, 43, 84, 134, 176, 179, 180], for reviews see [44, 45]). For example, in thoracic spinalized cats, when a 5-HT2A/C receptor agonist (e.g., quipazine or DOI) was administrated α-motoneur‐ on excitability was increased and hindlimb motor activity was enhanced [9]. Similar results were also demonstrated in SCI rats [12, 14] and mice [176]. Although 5-HT2A and 2C recep‐ tors both play a common role for motor functional recovery each subfamily may also exert different functions in different aspects of the motor outputs. Using different agonists and antagonists specific for each ofthese two receptor subfamilies it is found that 5-HT2A receptors might contribute to locomotor network activation and locomotor-like movement generation [15, 43, 176]. Direct stimulation of 5-HT2C receptors with its agonist meta-chlorophenylpiper‐ azine improved weight-supported locomotion in adult rats spinalized as neonates [177, 181]. 5-HT2C, probably also 2B, but not 2A, receptors could become constitutively active which could contribute to motor functional recovery following SCI [34, 182].
receptor mRNAs in the spinal tissue (including both white and gray matter) below the lesion with RT-PCR. Upregulation of 5-HT2A receptor mRNAs was also reported in mice subject‐ ed to an acute thoracic spinal transection. Thus, Ung et al. [176] found that 3 h afterthe lesion 5- HT2A receptor mRNA expression levels were increased 2.5-fold in the lateral intermediate zone in L1-L2 segments and they remained to be elevated for at least 2 weeks. However, no significant change was found in the ventral horn motoneuron regions, which was strikingly
*5-HT2B receptors:* No expression changes have been reported for this receptor subfamily in the spinal cord following SCI. With microarray global gene analysis technique, Wienecke et al. [122] did not detect significant changes of 5-HT2B receptor mRNAs in S2 spinal transected rats.
*5-HT2C receptors:* Similarto 5-HT2A receptors, 5-HT2C receptor expression changes have been investigated in different SCI animal models with different techniques. However, discrepan‐ cies exist as to whether 5-HT2C receptors are upregulated or not following SCI. In chronic SCI rats (~15 weeks) severe contusion or transection at thoracic level (T9) could induce a signifi‐ cant increase of 5-HT2C receptor immunoreactivity in lumbar spinal cord [22]. The increase could reach 4–7-fold in the ventral horn and ~17-fold in the dorsal horn over control level. In thoracic (T8-T9) spinal transected neonatal rats, following 4 weeks injury 5-HT2C receptor immunoreactivity was seen to increase 4–5-fold both in the spinal dorsal and ventral horn [177]. In a S2 spinal transection rat model, data from our group [36] indicated that 5-HT2C receptor immunoreactivity increased in all parts of the spinal gray matter below the lesion from 2 weeks but did notreach a significant level until 3 weeks (~1.4-fold over control animals). The increase sustained thereafter and a maximal level was reached at 45 days (~1.7-fold) and maintained at 60 days, the longest investigated interval. It is somehow perplexing that although upregulation was observed at protein level no significant increase at mRNA level was detected in different SCI animal models [34, 35, 122, 178]. However, although no signifi‐ cant changes of the total amount of 5-HT2C receptor mRNAs were detected, Murray et al. [34] did detect certain constitutive isoforms, such as INI isoform, that were significantly in‐
*Functional significance for motor outputs of 5-HT2 receptors:* The activation of 5-HT2 receptors has different effects on motor outputs after SCI, which include promoting motor functional recovery and causing maladaptive pathological motor symptoms such as spasticity. Due to uncertainty of the expression of 5-HT2B receptors in the spinal cord most studies have been focused on 5-HT2A and 2C receptors. There is ample evidence for the roles of 5-HT2A and 2C receptors in motor functional recovery following SCI from different studies (e.g., [9, 12, 14, 34, 43, 84, 134, 176, 179, 180], for reviews see [44, 45]). For example, in thoracic spinalized cats, when a 5-HT2A/C receptor agonist (e.g., quipazine or DOI) was administrated α-motoneur‐ on excitability was increased and hindlimb motor activity was enhanced [9]. Similar results were also demonstrated in SCI rats [12, 14] and mice [176]. Although 5-HT2A and 2C recep‐ tors both play a common role for motor functional recovery each subfamily may also exert different functions in different aspects of the motor outputs. Using different agonists and antagonists specific for each ofthese two receptor subfamilies it is found that 5-HT2A receptors might contribute to locomotor network activation and locomotor-like movement generation
creased in the motoneurons 6 weeks after S2 spinal transection.
different from that in the rat.
112 Recovery of Motor Function Following Spinal Cord Injury
One maladaptive consequence for 5-HT2 receptor plastic changes is the induction of muscle spasm due to increased motoneuron excitability. The rapid and robust upregulation of 5-HT2A receptors in spinal motoneurons might be responsible for 5-HT supersensitivity after SCI [32, 33]. Activation of 5-HT2C receptors could induce long-lasting reflexes of motoneurons in the in vitro spinal cord preparations [134, 135]. These long-lasting reflexes could be enhanced when 5-HT2C receptors became constitutively active following SCI [34, 182]. One ofthe cellular mechanisms for the increased motoneuron excitability is that the activation of 5-HT2 recep‐ tors could enhance calcium and/or sodium persistent inward currents [42, 183].
In addition to promoting locomotion, 5-HT2 (especially 2A) receptors also play a significant role for respiratory functional recovery after cervical spinal hemisection [179] and in enhanc‐ ing bladder function in thoracic spinal transected rats [184].
It should be addressed that the upregulation of 5-HT2 receptors does not by all means lead the motoneurons toward a hyperexcitatory state. Recently one interesting finding is present‐ ed by Bos et al. [185] who showed that the activation of 5-HT2A receptors could increase K(+)- Cl(–) cotransporter (KCC2) expression on cell membrane in the spinal cord motoneurons after SCI. The upregulation of KCC2 receptors in turn increased the activity of GABA A and glycine receptors which would restore endogenous inhibition and reduce spasticity after SCI in rats [186]. These results indicate that the upregulation of 5-HT2A receptors would not simply produce an excitatory effect. Rather, their activation will trigger an inhibitory factor which attempts to balance the excitatory effect. This effect seems to be paradoxical to the 5-HT2A receptors' direct excitatory effect on motoneurons. However, this might be one of the com‐ mon functional mechanisms of a spinal network—different factors interact and compensate each other, and the final motor output is determined by a summed vector of these push-pull forces acted on the motoneurons. In fact, any motor action cannot be arisen solely from the activation of a single factor; instead it needs a concert activity of multiple factors including both excitatory and inhibitory. For instance, Hayashi et al. [22] have showed that in contu‐ sive SCIrats, application of 5-HT1A or 5-HT2C receptor agonists alone orin combination could not improve hindlimb motor function; rather, motor functional improvement could be achieved only when the 5-HT precursor 5-HTP was administrated, indicating that simultane‐ ously multiple 5-HT receptor activations are needed for motor functional recovery. This concept is reinforced by the results from Noga et al. [16] who showed that in paralyzed, decerebrated cats, when locomotion was induced by electrical stimulation of the mesence‐ phalic locomotor region, most locomotor-activated cells, labeled with c-Fos, colocalized with 5-HT1A, 5-HT2A, and 5-HT7 receptors in laminae VII and VIII in the thoracolumbar spinal region.
### **4.3. 5-HT3 receptors**
There are no data demonstrating the expression changes of 5-HT3 following SCI. Laporte et al. [83] did not detect 5-HT3 receptor expression changes with binding technique in rat spinal cords whose descending serotonergic projections had been destroyed by 5,7-dihydroxytrypt‐ amine. However, lesions in peripheral nerves could reduce 5-HT3 receptor expression in the spinal motoneurons in rats [187].
Most studies concerning 5-HT3 receptor functions have been focused on sensory aspects, especially nociception. Studies relating to theirfunctions in motor outputs after SCI are scarce. So far there is only one study having investigated the influence of the activation of 5-HT3 receptors on motor outputs in SCI mice. In this study, Guertin and Steuer [152] showed that in hindlimb paralyzed mice 5-HT3 receptor agonist SR-57227 could produce hindlimb movements although with a low score. This result, anyhow, provided the first evidence that 5-HT3 receptors could modulate spinal motoneuron activity after SCI.
### **4.4. 5-HT7 receptors**
So far there are no data available with respect to the expression changes of 5-HT7 receptors following SCI. However, the activation of 5-HT7 receptor has been shown to be related to locomotor functional recovery in SCI states [23, 43, 172]. Landry et al. [172] showed that in thoracic spinal transected mice systemic application of 8-OH-DPAT, an agonist for both 5- HT1A and 5-HT7 receptors, acutely induced hindlimb movements with characteristics similar to normal locomotion. When the animals were pretreated with SB-269970, a selective 5-HT7 receptor antagonist, 8-OH-DPAT-induced movements were reduced. Sławińska et al. [23] showed that, in thoracic spinal transected rats, grafting of neurons from the B1, B2, and B3 descending 5-HT system from the brainstem into the spinal cord below the lesion effectively restored coordinated plantar stepping, and the application of SB-269970 disrupted the interand intralimb coordination. Further evidence from the same group indicated that 8-OH-DPAT facilitated plantar stepping in chronic spinalized rats [43]. Since 5-HT7 receptor agonists mainly improved coordination movement the authors assumed that 5-HT7 receptors mainly facilitated the activity of the central pattern generator interneurons. This assumption is supported by anatomical data that 5-HT7 receptors were mainly expressed in the neurons in the dorsal horn and the intermediate zone of the spinal gray matter [16]. In addition to the effects on locomotion, it has been reported that the activation of 5-HT7 receptors by their agonist LP44 (4-[2-(methylthio)phenyl]-*N*-(1,2,3,4-tetrahydronaphthalen-1-yl)-1-piperazine‐ hexanamide) increased bladder voiding efficiency in chronic thoracic spinal transected rats [188].
### **5. Plasticity of 5-HT innervations after SCI**
It is evidenced that a small number of 5-HT receptors, or more precisely receptor isoforms, could become active following SCI without ligand activation—so-called constitutive activa‐ tion. However, for most of the 5-HT receptors that do not have constitutive isoforms the **4.3. 5-HT3 receptors**
**4.4. 5-HT7 receptors**
[188].
spinal motoneurons in rats [187].
114 Recovery of Motor Function Following Spinal Cord Injury
There are no data demonstrating the expression changes of 5-HT3 following SCI. Laporte et al. [83] did not detect 5-HT3 receptor expression changes with binding technique in rat spinal cords whose descending serotonergic projections had been destroyed by 5,7-dihydroxytrypt‐ amine. However, lesions in peripheral nerves could reduce 5-HT3 receptor expression in the
Most studies concerning 5-HT3 receptor functions have been focused on sensory aspects, especially nociception. Studies relating to theirfunctions in motor outputs after SCI are scarce. So far there is only one study having investigated the influence of the activation of 5-HT3 receptors on motor outputs in SCI mice. In this study, Guertin and Steuer [152] showed that in hindlimb paralyzed mice 5-HT3 receptor agonist SR-57227 could produce hindlimb movements although with a low score. This result, anyhow, provided the first evidence that
So far there are no data available with respect to the expression changes of 5-HT7 receptors following SCI. However, the activation of 5-HT7 receptor has been shown to be related to locomotor functional recovery in SCI states [23, 43, 172]. Landry et al. [172] showed that in thoracic spinal transected mice systemic application of 8-OH-DPAT, an agonist for both 5- HT1A and 5-HT7 receptors, acutely induced hindlimb movements with characteristics similar to normal locomotion. When the animals were pretreated with SB-269970, a selective 5-HT7 receptor antagonist, 8-OH-DPAT-induced movements were reduced. Sławińska et al. [23] showed that, in thoracic spinal transected rats, grafting of neurons from the B1, B2, and B3 descending 5-HT system from the brainstem into the spinal cord below the lesion effectively restored coordinated plantar stepping, and the application of SB-269970 disrupted the interand intralimb coordination. Further evidence from the same group indicated that 8-OH-DPAT facilitated plantar stepping in chronic spinalized rats [43]. Since 5-HT7 receptor agonists mainly improved coordination movement the authors assumed that 5-HT7 receptors mainly facilitated the activity of the central pattern generator interneurons. This assumption is supported by anatomical data that 5-HT7 receptors were mainly expressed in the neurons in the dorsal horn and the intermediate zone of the spinal gray matter [16]. In addition to the effects on locomotion, it has been reported that the activation of 5-HT7 receptors by their agonist LP44 (4-[2-(methylthio)phenyl]-*N*-(1,2,3,4-tetrahydronaphthalen-1-yl)-1-piperazine‐ hexanamide) increased bladder voiding efficiency in chronic thoracic spinal transected rats
It is evidenced that a small number of 5-HT receptors, or more precisely receptor isoforms, could become active following SCI without ligand activation—so-called constitutive activa‐ tion. However, for most of the 5-HT receptors that do not have constitutive isoforms the
5-HT3 receptors could modulate spinal motoneuron activity after SCI.
**5. Plasticity of 5-HT innervations after SCI**
presence of their ligand—5HT is a necessity for their activation. Then, where could 5-HT originate following SCI? After contusion spinal cord injury or partial spinal cord transec‐ tions there are still spared supraspinal serotonergic projections and these serotonergic fibers could undergo plastic changes to eventually supplement lost 5-HT supply in the injured spinal cord. For example, in contusion injury the density of spared 5-HT fibers below the lesion varies depending on the locations along the dorsoventral axis and the severity of the injury [22, 189, 190], and 5-HT innervation could be partially restored with time which is responsible at least partly for motor functional recovery [190]. In thoracic hemisection injury, the density of 5-HT fibers in the ipsilateral ventral horn varied from 8% to 30% of the control value according to studies with a postinjury interval 4–7 days [39, 191–194]. The data are inconsistent as to whether 5-HT fibers increase with time following hemisection. Some researchers reported a gradual increase of 5-HT fibers on the ipsilateral side below the lesion (e.g., [193, 195]), whereas some others did not see apparent changes (e.g., [39, 194]). For example, Saruhashi et al. [193] observed that 5-HT-immunoreactive fibers recovered from 20% to about 75% of the normal value in the ventral horn during first 4 weeks. Camand et al. [195] also reported a similar finding in the intermediate zone. However, Filli et al. [194] reported a decrease to about 10% by 4 weeks from about 30% at day 4 in ventral horn motoneuron region. Azam et al. [39] reported that by 60 days the density of 5-HT-immunoreacitve fibers in the intermediate zone was reduced to 11% from 23% at 5 days, whereas in the ventral horn it was not reduced (23% vs. 22%). Following complete spinal transection, some 2–15% of the normal content of 5-HT remained in the spinal cord below the lesion ([196]; for review see [25]).
The next question would be then where the residual 5-HT originates following complete spinal transection. One origin might be the intraspinal serotonergic neurons. However, intraspinal serotonergic neurons are very few in number and sparsely distributed (see Section 1). Then, what are the other possible sources? Recent findings from our [38] and Bennett's group [37] indicated that AADC cells in the spinal cord might be another origin. AADC is an essential enzyme for the synthesis of 5-HT, dopamine, and certain trace amines from their respective precursors. AADC cells are widely distributed in different regions of the spinal cord [38], not limiting to the area around the central canal as reported previously [197]. Following SCI the ability of AADC cells in the spinal cord to synthesize 5-HT/dopamine from 5-HTP/L-dopa was dramatically increased and 5-HT/dopamine produced in the AADC was responsible for the increased motoneuron ability recorded both in vivo and in vitro [37, 38, 40, 198]. Nonethe‐ less, we have to admit that without 5-HTP application 5-HT could not be detected in the spinal AADC cells. This might be due to that the detecting techniques used in related studies were not sensitive enough to disclose a small amount of 5-HT, orthe turnoverrate of 5-HT produced in the AADC cells was so high that once produced it was immediately metabolized. One piece of evidence to support the speculation that monoamine transmitters could be produced from AADC cells in the spinal cord after SCI comes from Hou et al. [199], who showed that the number of dopamine-producing cells, which contained both AADC and tyrosine hydroxy‐ lase, was increased in lumbosacral spinal cord in thoracic spinal transected rats; and with enzyme-linked immunosorbent assay (ELISA) dopamine with a content of about 10% of its normal value could be detected in this part of the spinal cord even without L-dopa applica‐ tion. At the same time they also showed that dopamine produced from the spinal cord below the lesion was implicated for micturition functional recovery following SCI.
### **6. Conclusions**
The serotonin system in the spinal cord is an important modulator for sensory, motor as well as autonomic function. This system normally exercises its function via the interaction between ligand 5-HT and a number of 5-HT receptors expressed in different structures in the spinal cord. Among the 14 5-HT receptor subfamilies at least 12 have been detected in the spinal cord, in which 5-HT1A, 2A, 2C, and 7 receptors have been demonstrated to be more important for motor control. In normal physiological states, these receptors coordinate with each other and also with other monoamine systems to enable smooth and controllable motor outputs for our physical activity [200]. Following SCI extensive plastic changes occur for a number of 5-HT receptor subfamilies, mostly with increased receptor numbers and/or constitutive activity. In addition, plasticity also occurs for descending serotonergic fibers (in the case of incomplete SCI) and intraspinal serotonin-producing cells. The coeffects of these plastic changes, togeth‐ er with the plastic changes of other monoamine systems, ultimately lead to increased moto‐ neuron excitability in the chronic phase [17, 21, 108]. These plastic changes have both beneficial and detrimental effects for spinal motor outputs after SCI. Although the plastic changes can result in an adaptive compensation of the lost transmitters and thus assist in motor function‐ al recovery, they can also result in a plethora of maladaptive problems, such as spasticity. In clinical practice, a strategy needs to be set so as to maximally take advantage of the positive side of the plasticity to promote motor functional recovery and meanwhile to reduce nega‐ tive effects to a minimal extent. To reach this endpoint, pharmacological and/or genetic interferences need to be utilized so thatthe activity ofdifferent 5-HTreceptors and 5-HTsupply in the spinal cord reach a new balance and consequently an appropriate motor behavior is generated. Such a strategy would certainly have great implications for future treatment of SCI patients.
### **Acknowledgements**
This work was supported by the Lundbeck Foundation and the Danish Medical Research Council.
### **Abbreviations**
5-CT: 5-carboxamidotryptamine
5-HT: 5-hydroxytryptomine, serotonin
8-OH-DPAT: 8-hydroxy-2-(di-*n*-propylamino) tetralin
AADC: aromatic L-amino acid decarboxylase
BRL-15572: 3-(4-(4-chlorophenyl)piperazin-1-yl)-1,1-diphenyl-2-propanol
C, T, L, and S: cervical, thoracic, lumbar, and sacral spinal cord, respectively
D-MC: *N*-ormethylclozapine/8-chloro-11-(1-piperazinyl)-5*H*-dibenzo[*b,e*][1,4]diazepine
DOI: (±)-2,5-dimethoxy-4-iodoamphetamine hydrochloride
DOM: 2,5-dimethoxy-alpha,4-dimethyl-benzene ethamine hydrochloride
DSP-4: *N*-(2-chloroethyl)-*N*-ethyl-2-bromobenzylamine
ELISA: enzyme-linked immunosorbent assay
EMD-386088: 5-chloro-2-methyl-3-(1,2,3,6-tetrahydro-4-pyridinyl)-1*H*-indole
EPSP: excitatory postsynaptic potential
GABA: γ-aminobutyric acid
tion. At the same time they also showed that dopamine produced from the spinal cord below
The serotonin system in the spinal cord is an important modulator for sensory, motor as well as autonomic function. This system normally exercises its function via the interaction between ligand 5-HT and a number of 5-HT receptors expressed in different structures in the spinal cord. Among the 14 5-HT receptor subfamilies at least 12 have been detected in the spinal cord, in which 5-HT1A, 2A, 2C, and 7 receptors have been demonstrated to be more important for motor control. In normal physiological states, these receptors coordinate with each other and also with other monoamine systems to enable smooth and controllable motor outputs for our physical activity [200]. Following SCI extensive plastic changes occur for a number of 5-HT receptor subfamilies, mostly with increased receptor numbers and/or constitutive activity. In addition, plasticity also occurs for descending serotonergic fibers (in the case of incomplete SCI) and intraspinal serotonin-producing cells. The coeffects of these plastic changes, togeth‐ er with the plastic changes of other monoamine systems, ultimately lead to increased moto‐ neuron excitability in the chronic phase [17, 21, 108]. These plastic changes have both beneficial and detrimental effects for spinal motor outputs after SCI. Although the plastic changes can result in an adaptive compensation of the lost transmitters and thus assist in motor function‐ al recovery, they can also result in a plethora of maladaptive problems, such as spasticity. In clinical practice, a strategy needs to be set so as to maximally take advantage of the positive side of the plasticity to promote motor functional recovery and meanwhile to reduce nega‐ tive effects to a minimal extent. To reach this endpoint, pharmacological and/or genetic interferences need to be utilized so thatthe activity ofdifferent 5-HTreceptors and 5-HTsupply in the spinal cord reach a new balance and consequently an appropriate motor behavior is generated. Such a strategy would certainly have great implications for future treatment of SCI
This work was supported by the Lundbeck Foundation and the Danish Medical Research
the lesion was implicated for micturition functional recovery following SCI.
**6. Conclusions**
116 Recovery of Motor Function Following Spinal Cord Injury
patients.
Council.
**Abbreviations**
**Acknowledgements**
5-CT: 5-carboxamidotryptamine
5-HT: 5-hydroxytryptomine, serotonin
8-OH-DPAT: 8-hydroxy-2-(di-*n*-propylamino) tetralin
AADC: aromatic L-amino acid decarboxylase
GR-46611: 3-[3-(2-dimethylaminoethyl)-1*H*-indol-5-yl]-*N*-(4-methoxybenzyl)acrylamide
KCC2: K(+)-Cl(–) cotransporter
L-dopa: L-3,4-dihydroxyphenylalanine
LP44: 4-[2-(methylthio)phenyl]-*N*-(1,2,3,4-tetrahydronaphthalen-1-yl)-1-piperazinehexana‐ mide
LY-344864: *N*-[(3*R*)-3-(dimethylamino)-2,3,4,9-tetrahydro-1*H*-carbazol-6-yl]-4-fluorobenza‐ mide
MK-212: 6-chloro-2-(1-piperazinyl)pyrazine hydrochloride
mRNA: messenger ribonucleic acid
NMDA: *N*-methyl-D-aspartate
PCR: polymerase chain reaction
Ro-600175: (α*S*)-6-chloro-5-fluoro-α-methyl-1*H*-indole-1-ethanamine fumarate
RS-127445: (3-(4-(4-chlorophenyl)piperazin-1-yl)-1,1-diphenyl-2-propanol)
RT-PCT: reverse transcription polymerase chain reaction
RU-24969: 5-methoxy-3(1,2,3,6-tetrahydropyridin-4-yl)-1*H*-indole
SB-242084: 6-chloro-2,3-dihydro-5-methyl-*N*-[6-[(2-methyl-3-pyridinyl)oxy]-3-pyridinyl]-1*H*indole-1-carboxyamide
SB-269970: (2*R*)-1-[(3-hydroxyphenyl)sulfonyl]-2 -(2-(4-methyl-1-piperidinyl)ethyl)pyrroli‐ dine
SCI: spinal cord injury
SR-57227: 1-(6-chloropyridin-2-yl)piperidin-4-amine
TCB-2: 4-bromo-3,6-dimethoxybenzocyclobuten-1-yl)methylamine hydrobromide
WAY-100635: *N*-[2-[4-(2-methoxyphenyl)-1-piperazinyl]ethyl]-*N*-(2-pyridyl)cyclohexanecar‐ boxamide
WAY-161503: 8,9-dichloro-2,3,4,4a-tetrahydro-1*H*-pyrazino[1,2-*a*]quinoxalin-5(6*H*)
### **Author details**
Mengliang Zhang1,2
Address all correspondence to: mzhang@sund.ku.dk
1 Department of Neuroscience and Pharmacology, University of Copenhagen, Copenhagen, Denmark
2 Neuronano Research Center, Department of Experimental Medical Science, Lund Univer‐ sity, Lund, Sweden
### **References**
[8] Guertin PA. New pharmacological approaches against chronic bowel and bladder problems in paralytics. World J Crit Care Med. 2016;5:1–6. DOI: 10.5492/wjccm.v5.i1.1.
TCB-2: 4-bromo-3,6-dimethoxybenzocyclobuten-1-yl)methylamine hydrobromide
WAY-161503: 8,9-dichloro-2,3,4,4a-tetrahydro-1*H*-pyrazino[1,2-*a*]quinoxalin-5(6*H*)
boxamide
Denmark
**Author details**
Mengliang Zhang1,2
sity, Lund, Sweden
2007.02.002.
2009.08.002.
2012.0040.
1748-1716.2006.01652.x.
**References**
Address all correspondence to: mzhang@sund.ku.dk
118 Recovery of Motor Function Following Spinal Cord Injury
210. DOI: 10.1097/BSD.0000000000000287.
WAY-100635: *N*-[2-[4-(2-methoxyphenyl)-1-piperazinyl]ethyl]-*N*-(2-pyridyl)cyclohexanecar‐
1 Department of Neuroscience and Pharmacology, University of Copenhagen, Copenhagen,
2 Neuronano Research Center, Department of Experimental Medical Science, Lund Univer‐
[1] Witiw CD, Fehlings MG. Acute spinal cord injury. J Spinal Disord Tech. 2015;28:202–
[2] Jensen MP, Kuehn CM, Amtmann D, Cardenas DD. Symptom burden in persons with spinal cord injury. Arch Phys Med Rehabil. 2007;88:638–645. DOI: 10.1016/j.apmr.
[3] Nielsen JB, Crone C, Hultborn H. The spinal pathophysiology of spasticity-from a basic science point of view. Acta Physiol (Oxf). 2007;189:171–180. DOI: 10.1111/j.
[4] Sheean G, McGuire JR. Spastic hypertonia and movement disorders: pathophysiolo‐ gy, clinical presentation, and quantification. PM R. 2009;1:827–833. DOI: 10.1016/j.pmrj.
[5] Gunduz H, Binak DF. Autonomic dysreflexia: an important cardiovascular complica‐ tion in spinal cord injury patients. Cardiol J. 2012;19:215–219. DOI: 10.5603/CJ.
[6] Felix ER. Chronic neuropathic pain in SCI: evaluation and treatment. Phys Med Rehabil
[7] Nas K, Yazmalar L, Şah V, Aydın A, Öneş K. Rehabilitation of spinal cord injuries.
Clin N Am. 2014;25:545–571. DOI: 10.1016/0006-8993(88)91548-X.
World J Orthop. 2015;6:8–16. DOI: 10.5312/wjo.v6.i1.8.
term pharmacological treatment or transplantation of embryonic neurons? Prog Brain Res. 2002;137:213–230.
[31] Otoshi CK, Walwyn WM, Tillakaratne NJ, Zhong H, Roy RR, Edgerton VR. Distribu‐ tion and localization of 5-HT(1A) receptors in the rat lumbar spinal cord after transec‐ tion and deafferentation. J Neurotrauma. 2009;26:575–584. DOI: 10.1089/neu.2008.0640.
term pharmacological treatment or transplantation of embryonic neurons? Prog Brain
[20] Kubasak MD, Jindrich DL, Zhong H, Takeoka A, McFarland KC, Muñoz-Quiles C, Roy RR, Edgerton VR, Ramón-Cueto A, Phelps PE. OEG implantation and step training enhance hindlimb-stepping ability in adult spinal transected rats. Brain. 2008;131:264–
[21] Courtine G, Gerasimenko Y, van den Brand R, Yew A, Musienko P, Zhong H, Song B, Ao Y, Ichiyama RM, Lavrov I, Roy RR, Sofroniew MV, Edgerton VR. Transformation of nonfunctional spinal circuits into functional states after the loss of brain input. Nat
[22] Hayashi Y, Jacob-Vadakot S, Dugan EA, McBride S, Olexa R, Simansky K, Murray M, Shumsky JS. 5-HT precursor loading, but not 5-HT receptor agonists, increases motor function after spinal cord contusion in adult rats. Exp Neurol. 2010;221:68–78. DOI:
[23] Sławińska U, Miazga K, Cabaj AM, Leszczyńska AN, Majczyński H, Nagy JI, Jordan LM. Grafting of fetal brainstem 5-HT neurons into the sublesional spinal cord of paraplegic rats restores coordinated hindlimb locomotion. Exp Neurol. 2013;247:572–
[24] Jacobs BL, Fornal CA. Serotonin and motor activity. Curr Opin Neurobiol. 1997;7:820–
[25] Schmidt BJ, Jordan LM. The role of serotonin in reflex modulation and locomotor rhythm production in the mammalian spinal cord. Brain Res Bull. 2000;53:689–710.
[26] Millan MJ. Descending control of pain. Prog Neurobiol. 2002;66:355–474. DOI: 10.1016/
[27] Pearlstein E, Ben Mabrouk F, Pflieger JF, Vinay L. Serotonin refines the locomotorrelated alternations in the in vitro neonatalrat spinal cord. EurJ Neurosci. 2005;21:1338–
[28] Giroux N, Rossignol S, Reader TA. Autoradiographic study of alpha1- and alpha2 noradrenergic and serotonin1A receptors in the spinal cord of normal and chronical‐ ly transected cats. J Comp Neurol. 1999;406:402–414. DOI: 10.1002/
[29] Fuller DD, Baker-Herman TL, Golder FJ, Doperalski NJ, Watters JJ, Mitchell GS. Cervical spinal cord injury upregulates ventral spinal 5-HT2A receptors. J Neurotrau‐
[30] Lee JK, Johnson CS, Wrathall JR. Up-regulation of 5-HT2 receptors is involved in the increased H-reflex amplitude after contusive spinal cord injury. Exp Neurol.
Res. 2002;137:213–230.
120 Recovery of Motor Function Following Spinal Cord Injury
276. DOI: 10.1093/brain/awm267.
10.1016/j.expneurol.2009.10.003.
581. DOI: 10.1016/j.expneurol.2013.02.008.
825. DOI: 10.1016/S0959-4388(97)80141-9.
1346. DOI: 10.1111/j.1460-9568.2005.03971.x.
(SICI)1096-9861(19990412)406:3<402::AID-CNE8>3.0.CO;2-F.
ma. 2005;22:203–213. DOI: 10.1089/neu.2005.22.203.
2007;203:502–511. DOI: 10.1016/j.expneurol.2006.09.003.
DOI: 10.1016/S0361-9230(00)00402-0.
S0301-0082(02)00009-6.
Neurosci. 2009;12:1333–1342. DOI: 10.1038/nn.2401.
chronic spinal cord injury. J Neurophysiol. 2006;96:1158–1170. DOI: 10.1152/jn. 01088.2005.
[56] Jacobs BL, Fornal CA. 5-HT and motor control: a hypothesis. Trends Neurosci. 1993;16:346–352. DOI: 10.1016/0166-2236(93)90090-9.
chronic spinal cord injury. J Neurophysiol. 2006;96:1158–1170. DOI: 10.1152/jn.
[43] Sławińska U, Miazga K, Jordan LM. 5-HT₂ and 5-HT₇ receptor agonists facilitate plantar stepping in chronic spinal rats through actions on different populations of spinal neurons. Front Neural Circuits. 2014;8:95. DOI: 10.3389/fncir.2014.00095. [44] Ghosh M, Pearse DD. The role of the serotonergic system in locomotor recovery after spinal cord injury. Front Neural Circuits. 2015;8:151. DOI: 10.3389/fncir.2014.00151.
[45] Nardone R, Höller Y, Thomschewski A, Höller P, Lochner P, Golaszewski S, Brigo F, Trinka E. Serotonergic transmission after spinal cord injury. J Neural Transm (Vienna).
[46] Dahlstroem A, Fuxe K. Evidence for the existence of monoamine containing neurons in the central nervous system. I. Demonstration of monamines in the cell bodies of brain
[47] Takeuchi Y, Kimura H, Sano Y. Immunohistochemical demonstration of serotonin neurons in the brainstem of the rat and cat. Cell Tissue Res. 1982;224:247–267. DOI:
[48] Skagerberg G, Björklund A. Topographic principles in the spinal projections of serotonergic and non-serotonergic brainstem neurons in the rat. Neuroscience.
[49] Hornung JP. The human raphe nuclei and the serotonergic system. J Chem Neuroa‐
[50] Azmitia EC, Gannon PJ. The primate serotonergic system: a review of human and animal studies and a report on macaca fasicularis. Adv Neurol. 1986;43:407–468. [51] Ballion B, Branchereau P, Chapron J, Viala D. Ontogeny of descending serotonergic innervation and evidence for intraspinal 5-HT neurons in the mouse spinal cord. Brain
[52] Jacobs BL, Azmitia EC. Structure and function of the brain serotonin system. Physiol
[53] Lamotte CC, Johns DR, de Lanerolle NC. Immunohistochemical evidence of indola‐ mine neurons in monkey spinal cord. J Comp Neurol.1982; 206:359–570. DOI: 10.1002/
[54] Newton BW, Maley BE, Hamill RW. Immunohistochemical demonstration of seroto‐ nin neurons in autonomic regions of the rat spinal cord. Brain Res. 1986;376:155–163.
[55] Newton BW, Hamill RW. The morphology and distribution of rat serotoninergic intraspinal neurons: an immunohistochemical study. Brain Res Bull. 1988;20:349–360.
Res Dev Brain Res. 2002;137:81–88. DOI: 10.1016/S0165-3806(02)00414–5.
2015;122:279–295. DOI: 10.1007/s00702-014-1241-z.
stem neurons. Acta Physiol Scand Suppl. 1964;62:1–55.
1985;15:445–480. DOI: 10.1016/0306-4522(85)90225-8.
nat. 2003;26:331–343. DOI: 10.1016/j.jchemneu.2003.10.002.
01088.2005.
122 Recovery of Motor Function Following Spinal Cord Injury
10.1007/BF00216872.
Rev. 1992;72:165–229.
DOI: 10.1016/0006-8993(86)90910-8.
DOI: 10.1016/0361-9230(88)90064-0.
cne.902060404.
nervous system. J Comp Neurol. 1996;365:289–305. DOI: 10.1002/(SI‐ CI)1096-9861(19960205)365:2<289::AID-CNE7>3.0.CO;2-1.
[81] Cotel F, Exley R, Cragg SJ, Perrier JF. Serotonin spillover onto the axon initial seg‐ ment of motoneurons induces central fatigue by inhibiting action potential initiation. Proc Natl Acad Sci USA. 2013;110:4774–4779. DOI: 10.1073/pnas.1216150110.
nervous system. J Comp Neurol. 1996;365:289–305. DOI: 10.1002/(SI‐
[70] Talley EM, Bayliss DA. Postnatal development of 5-HT(1A) receptor expression in rat somatic motoneurons. Brain Res Dev Brain Res. 2000;122:1–10. DOI: 10.1016/
[71] Kheck NM, Gannon PJ, Azmitia EC. 5-HT1A receptor localization on the axon hillock of cervical spinal motoneurons in primates. J Comp Neurol. 1995;355:211–220. DOI:
[72] Eide PK, Joly NM, Hole K. The role of spinal cord 5-HT1A and 5-HT1B receptors in the modulation of a spinal nociceptive reflex. Brain Res. 1990;536:195–200. DOI:
[73] Hochman S, Garraway SM, Machacek DW, Shay BL. 5-HT receptors and the neuro‐ modulatory control of spinal cord function. In: Cope TC, editor. Motor Neurobiology
[74] Nadeson R, Goodchild CS. Antinociceptive role of 5-HT1A receptors in rat spinal cord.
[75] Buritova J, Larrue S, Aliaga M, Besson JM, Colpaert F. Effects of the high-efficacy 5- HT1A receptor agonist, F 13640 in the formalin pain model: a c-Fos study. Eur J
[76] Kim JM, Jeong SW, Yang J, Lee SH, Kim WM, Jeong S, Bae HB, Yoon MH, Choi JI. Spinal 5-HT1A, not the 5-HT1B or 5-HT3 receptors, mediates descending serotoner‐ gic inhibition for late-phase mechanical allodynia of carrageenan-induced peripheral inflammation. Neurosci Lett. 2015;600:91–97. DOI: 10.1016/j.neulet.2015.05.058.
[77] Alhaider AA, Wilcox GL. Differential roles of 5-hydroxytryptamine1A and 5-hydrox‐ ytryptamine1B receptor subtypes in modulating spinal nociceptive transmission in
[78] Ali Z, Wu G, Kozlov A, Barasi S. The actions of 5-HT1 agonists and antagonists on nociceptive processing in the rat spinal cord: results from behavioural and electrophy‐ siological studies. Brain Res. 1994;661:83–90. DOI:10.1016/0006-8993(94)91184-3.
[79] Zhao ZQ, Liu XY, Jeffry J, Karunarathne WK, Li JL, Munanairi A, Zhou XY, Li H, Sun YG, Wan L, Wu ZY, Kim S, Huo FQ, Mo P, Barry DM, Zhang CK, Kim JY, Gautam N, Renner KJ, Li YQ, Chen ZF. Descending control of itch transmission by the serotoner‐ gic system via 5-HT1A-facilitated GRP–GRPR signaling. Neuron. 2014;84:821–834.
[80] Bervoets K, Rivet JM, Millan MJ. 5-HT1A receptors and the tail-flick response. IV. Spinally localized 5-HT1A receptors postsynaptic to serotoninergic neurones mediate
spontaneous tail-flicks in the rat. J Pharmacol Exp Ther. 1993;264:95–104.
CI)1096-9861(19960205)365:2<289::AID-CNE7>3.0.CO;2-1.
of the Spinal Cord. CRC Press: Boca Raton; 2001. p. 47–87.
Br J Anaesth. 2002;88:679–684. DOI: 10.1093/bja/88.5.679.
mice. J Pharmacol Exp Ther. 1993;265:378–385.
DOI: 10.1016/j.neuron.2014.10.003.
Pharmacol. 2005;514:121–130. DOI:10.1016/j.ejphar.2005.03.016.
S0165-3806(00)00036-5.
124 Recovery of Motor Function Following Spinal Cord Injury
10.1002/cne.903550205.
10.1016/0006-8993(90)90025-7.
[105] Helton LA, Thor KB, Baez M. 5-Hydroxytryptamine2A, 5-hydroxytryptamine2B, and 5-hydroxytryptamine2C receptor mRNA expression in the spinal cord of rat, cat, monkey and human. Neuroreport. 1994;5:2617–2620.
[93] Ahn AH, Basbaum AI. Tissue injury regulates serotonin 1D receptor expression: implications for the control of migraine and inflammatory pain. J Neurosci.
[94] Enjin A, Leão KE, Mikulovic S, Le Merre P, Tourtellotte WG, Kullander K. Sensorimo‐ tor function is modulated by the serotonin receptor 1d, a novel marker for gamma motor neurons. Mol Cell Neurosci. 2012;49:322–332. DOI: 10.1016/j.mcn.2012.01.003.
[95] D'Amico JM, Li Y, Bennett DJ, Gorassini MA. Reduction of spinal sensory transmis‐ sion by facilitation of 5-HT1B/D receptors in noninjured andspinal cord-injured
[96] Jeong CY, Choi JI, Yoon MH. Roles of serotonin receptor subtypes for the antinocicep‐ tion of 5-HT in the spinal cord ofrats. EurJ Pharmacol. 2004;502:205–211. DOI: 10.1016/
[97] Honda M, Imaida K, Tanabe M, Ono H. Endogenously released 5-hydroxytrypta‐ mine depresses the spinal monosynaptic reflex via 5-HT1D receptors. Eur J Pharma‐
[98] Castro ME, Pascual J, Romón T, del Arco C, del Olmo E, Pazos A. Differential distri‐
[99] Murray KC, Stephens MJ, Rank M, D'Amico J, Gorassini MA, Bennett DJ. Polysynap‐ tic excitatory postsynaptic potentials that trigger spasms after spinal cord injury in rats are inhibited by 5-HT1B and 5-HT1F receptors. J Neurophysiol. 2011;106:925–943.
[100] Classey JD, Bartsch T, Goadsby PJ. Distribution of 5-HT(1B), 5-HT(1D) and 5-HT(1F) receptor expression in rat trigeminal and dorsal root ganglia neurons: relevance to the selective anti-migraine effect of triptans. Brain Res. 2010;1361:76–85. DOI: 10.1016/
[101] Agosti RM. 5HT1F- and 5HT7-receptor agonists for the treatment of migraines. CNS Neurol Disord Drug Targets. 2007;6:235–237. DOI: 10.2174/187152707781387242. [102] Ferrari MD, Färkkilä M, Reuter U, Pilgrim A, Davis C, Krauss M, Diener HC; European COL-144 Investigators. Acute treatment of migraine with the selective 5-HT1F receptor agonist lasmiditan—a randomised proof-of-concept trial. Cephalalgia. 2010;30:1170–
[103] Humphrey PP, Hartig P, Hoyer D. A proposed new nomenclature for 5-HT receptors. Trends Pharmacol Sci. 1993;14:233–236. DOI: 10.1016/0165-6147(93)90016-D.
[104] Pompeiano M, Palacios JM, Mengod G. Distribution of the serotonin 5-HT2 receptor family mRNAs: comparison between 5-HT2A and 5-HT2C receptors. Brain Res Mol
Brain Res. 1994;23:163–178. DOI: 10.1016/0169-328X(94)90223-2.
human brain: focus on brainstem and spinal cord. Neuropharmacology. 1997;36:535–
H]sumatriptan binding sites (5-HT1B, 5-HT1D and 5-HT1F receptors) in
humans. J Neurophysiol. 2013;109:1485–1493. DOI: 10.1152/jn.00822.2012.
2006;26:8332–8338. DOI: 10.1523/JNEUROSCI.1989-06.2006.
col. 2004;503:55–61. DOI: 10.1016/j.ejphar.2004.09.045.
542. DOI: 10.1016/S0028-3908(97)00061-0.
1178. DOI: 10.1177/0333102410375512.
DOI: 10.1152/jn.01011.2010.
j.brainres.2010.09.004.
j.ejphar.2004.08.048.
126 Recovery of Motor Function Following Spinal Cord Injury
bution of [3
GABA/glycine and 5-HT release via 5-HT(2A) and 5-HT(3) receptors. Eur J Pharma‐ col. 2013;702:149–157. DOI: 10.1016/j.ejphar.2013.01.058.
[127] MbakiY, RamageAG. Investigation ofthe role of 5-HT2 receptor subtypes in the control of the bladder and the urethra in the anaesthetized female rat. Br J Pharmacol. 2008;155:343–356. DOI: 10.1038/bjp.2008.273.
GABA/glycine and 5-HT release via 5-HT(2A) and 5-HT(3) receptors. Eur J Pharma‐
[117] Fone KC, Robinson AJ, Marsden CA. Characterization of the 5-HT receptor subtypes involved in the motor behaviours produced by intrathecal administration of 5-HT agonists in rats. Br J Pharmacol. 1991;103:1547–1555. DOI: 10.1111/j.
[118] Yamazaki J, Fukuda H, Nagao T, Ono H. 5-HT2/5-HT1C receptor-mediated facilitato‐ ry action on unit activity of ventral horn cells in rat spinal cord slices. Eur J Pharma‐
[119] Mbaki Y, Gardiner J, McMurray G, Ramage AG. 5-HT 2A receptor activation of the external urethral sphincter and 5-HT 2C receptor inhibition of micturition: a study based on pharmacokinetics in the anaesthetized female rat. Eur J Pharmacol.
[120] Watson NV, Gorzalka BB. DOI-induced inhibition of copulatory behavior in male rats: reversal by 5-HT2 antagonists. Pharmacol Biochem Behav. 1991;39:605–612. DOI:
[121] Rössler AS, Bernabé J, Denys P, Alexandre L, Giuliano F. Effect of the 5-HT receptor agonist DOI on female rat sexual behavior. J Sex Med. 2006;3:432–441. DOI: 10.1111/j.
[122] Wienecke J, Westerdahl AC, Hultborn H, Kiehn O, Ryge J. Global gene expression analysis of rodent motor neurons following spinal cord injury associates molecular mechanisms with development of postinjury spasticity. J Neurophysiol. 2010;103:761–
[123] MacFarlane PM, Vinit S, Mitchell GS. Serotonin 2A and 2B receptor-induced phrenic motor facilitation: differential requirement for spinal NADPH oxidase activity.
[124] Lin SY, Chang WJ, Lin CS, Huang CY, Wang HF, Sun WH. Serotonin receptor 5-HT2B mediates serotonin-induced mechanical hyperalgesia. J Neurosci. 2011;31:1410–1418.
[125] Aira Z, Buesa I, García del Caño G, Salgueiro M, Mendiable N, Mingo J, Aguilera L, Bilbao J, Azkue JJ. Selective impairment of spinal mu-opioid receptor mechanism by plasticity of serotonergic facilitation mediated by 5-HT2A and 5-HT2B receptors. Pain.
[126] Cervantes-Durán C, Vidal-Cantú GC, Barragán-Iglesias P, Pineda-Farias JB, Bravo-Hernández M, Murbartián J, Granados-Soto V. Role of peripheral and spinal 5-HT2B receptors in formalin-induced nociception. Pharmacol Biochem Behav. 2012;102:30–35.
Neuroscience. 2011;178:45–55. DOI: 10.1016/j.neuroscience.2011.01.011.
col. 2013;702:149–157. DOI: 10.1016/j.ejphar.2013.01.058.
col. 1992;220:237–242. DOI: 10.1016/0014-2999(92)90753-Q.
2012;682:142–152. DOI: 10.1016/j.ejphar.2012.02.010.
1476-5381.1991.tb09825.x.
128 Recovery of Motor Function Following Spinal Cord Injury
10.1016/0091-3057(91)90135-O.
778. DOI: 10.1152/jn.00609.2009.
DOI: 10.1523/JNEUROSCI.4682-10.2011.
DOI: 10.1016/j.pbb.2012.03.015.
2012;153:1418–1425. DOI: 10.1016/j.pain.2012.03.017.
1743-6109.2006.00240.x.
(5-hydroxytryptamine) 3 receptor. Neuroscience. 2005;134:165–173. DOI: 10.1016/ j.neuroscience.2005.02.013.
[150] Huang J, Wang YY, Wang W, Li YQ, Tamamaki N, Wu SX. 5-HT(3A) receptor subunit is expressed in a subpopulation of GABAergic and enkephalinergic neurons in the mouse dorsal spinal cord. Neurosci Lett. 2008;441:1–6. DOI: 10.1016/j.neulet. 2008.04.105.
[138] Conlon K, Miner W, McCleary S, McMurray G. Identification of 5-HT(2C) mediated mechanisms involved in urethral sphincter reflexes in a guinea-pig model of urethral
[139] Thompson AJ, Lummis SC. Discriminating between 5-HT₃A and 5-HT₃AB receptors.
[140] Kia HK, Miquel MC, McKernan RM, Laporte AM, Lombard MC, Bourgoin S, Hamon M, Vergé D. Localization of 5-HT3 receptors in the rat spinal cord: immunohistochem‐
[141] Laporte AM, Doyen C, Nevo IT, Chauveau J, Hauw JJ, Hamon M. Autoradiographic mapping of serotonin 5-HT1A, 5-HT1D, 5-HT2A and 5-HT3 receptors in the aged human spinal cord. J Chem Neuroanat. 1996;11:67–75. DOI:
[142] Morales M, Battenberg E, Bloom FE. Distribution of neurons expressing immunoreac‐ tivity for the 5HT3 receptor subtype in the rat brain and spinal cord. J Comp Neurol. 1998;402:385–401. DOI: 10.1002/(SICI)1096-9861(19981221)402:3<385::AID-
[143] Maxwell DJ, Kerr R, Rashid S, Anderson E. Characterisation of axon terminals in the rat dorsal horn that are immunoreactive for serotonin 5-HT3A receptor subunits. Exp
[144] Glaum SR, Proudfit HK, Anderson EG. 5-HT3 receptors modulate spinal nociceptive
[145] Alhaider AA, Lei SZ, Wilcox GL. Spinal 5-HT3 receptor-mediated antinociception:
[146] Khasabov SG, Lopez-Garcia JA, Asghar AU, King AE. Modulation of afferent-evoked neurotransmission by 5-HT3 receptors in young rat dorsal horn neurones in vitro: a putative mechanism of 5-HT3 induced anti-nociception. Br J Pharmacol. 1999;127:843–
[147] Fukushima T, Ohtsubo T, Tsuda M, Yanagawa Y, Hori Y. Facilitatory actions of serotonin type 3 receptors on GABAergic inhibitory synaptic transmission in the spinal superficial dorsal horn. J Neurophysiol. 2009;102:1459–1471. DOI: 10.1152/jn.
[148] Guo W, Miyoshi K, Dubner R, Gu M, Li M, Liu J, Yang J, Zou S, Ren K, Noguchi K, Wei F. Spinal 5-HT3 receptors mediate descending facilitation and contribute to behavio‐ ral hypersensitivity via a reciprocal neuron–glial signaling cascade. Mol Pain.
[149] Conte D, Legg ED, McCourt AC, Silajdzic E, Nagy GG, Maxwell DJ. Transmitter content, origins and connections of axons in the spinal cord that possess theserotonin
reflexes. Brain Res. 1990;510:12–16. DOI: 10.1016/0006-8993(90)90721-M.
function. BJU Int. 2012;110:E113–117. DOI: 10.1111/j.1464-410X.2011.10756.x.
Br J Pharmacol. 2013;169:736–747. DOI: 10.1111/bph.12166.
istry and in situ hybridization. Neuroreport. 1995;6:257–261.
Brain Res. 2003;149:114–124. DOI: 10.1007/s00221-002-1339-7.
possible release of GABA. J Neurosci. 1991;11:1881–1888.
10.1016/0891-0618(96)00130-5.
130 Recovery of Motor Function Following Spinal Cord Injury
852. DOI: 10.1038/sj.bjp.0702592.
2014;10:35. DOI: 10.1186/1744-8069-10-35.
91160.2008.
CNE7>3.0.CO;2-Q.
[171] Cornide-Petronio ME, Ruiz MS, Barreiro-Iglesias A, Rodicio MC. Spontaneous regeneration of the serotonergic descending innervation in the sea lamprey after spinal cord injury. J Neurotrauma. 2011;28:2535–2540. DOI: 10.1089/neu.2011.1766.
[160] Castañeda-Corral G, Rocha-González HI, Araiza-Saldaña CI, Ambriz-Tututi M, Vidal-Cantú GC, Granados-Soto V. Role of peripheral and spinal 5-HT6 receptors accord‐ ing to the rat formalin test. Neuroscience. 2009;162:444–452. DOI: 10.1016/
[161] To ZP, Bonhaus DW, Eglen RM, Jakeman LB. Characterization and distribution of putative 5-ht7 receptors in guinea-pig brain. Br J Pharmacol. 1995;115:107–116. DOI:
[162] Doly S, Fischer J, Brisorgueil MJ, Vergé D, Conrath M. Pre- and postsynaptic localiza‐ tion of the 5-HT7 receptor in rat dorsal spinal cord: immunocytochemical evidence. J
[163] Viguier F, Michot B, Hamon M, Bourgoin S. Multiple roles of serotonin in pain control mechanisms—implications of 5-HT₇ and other 5-HT receptor types. Eur J Pharmacol.
[164] Yesilyurt O, Seyrek M, Tasdemir S, Kahraman S, Deveci MS, Karakus E, Halici Z, Dogrul A. The critical role of spinal 5-HT7 receptors in opioid and non-opioid type stress-induced analgesia. Eur J Pharmacol. 2015;762:402–410. DOI: 10.1016/j.ejphar.
[165] Dam LJ, Hai L, Ha YM. Role of the 5-HT(7) receptorin the effects of intrathecal nefopam in neuropathic pain in rats. Neurosci Lett. 2014;566:50–54. DOI: 10.1016/j.neulet.
[166] Lee HG, Kim WM, Kim JM, Bae HB, Choi JI. Intrathecal nefopam-induced antinoci‐ ception through activation of descending serotonergic projections involving spinal 5- HT7 but not 5-HT3 receptors. Neurosci Lett. 2015;587:120–125. DOI: 10.1016/j.neulet.
[167] Liu J, Jordan LM. Stimulation of the parapyramidal region of the neonatal rat brain stem produces locomotor-like activity involving spinal 5-HT7 and 5-HT2A receptors.
[168] Jordan LM, Liu J, Hedlund PB, Akay T, Pearson KG. Descending command systems for the initiation of locomotion in mammals. Brain Res Rev. 2008;57:183–191. DOI:
[169] Liu J, Akay T, Hedlund PB, Pearson KG, Jordan LM. Spinal 5-HT7 receptors are critical for alternating activity during locomotion: in vitro neonatal and in vivo adult studies using 5-HT7 receptor knockout mice. J Neurophysiol. 2009;102:337–348. DOI: 10.1152/
[170] Cornide-Petronio ME, Fernández-López B, Barreiro-Iglesias A, Rodicio MC. Traumat‐ ic injury induces changes in the expression of the serotonin 1A receptor in the spinal cord of lampreys. Neuropharmacology. 2014;77:369–378. DOI: 10.1016/j.neuropharm.
J Neurophysiol. 2005;94:1392–1404. DOI: 10.1152/jn.00136.2005.
Comp Neurol. 2005;490:256–269. DOI: 10.1002/cne.20667.
2013;716:8–16. DOI: 10.1016/j.ejphar.2013.01.074.
j.neuroscience.2009.04.072.
132 Recovery of Motor Function Following Spinal Cord Injury
2015.04.020.
2014.02.021.
2014.12.040.
jn.91239.2008.
2013.10.017.
10.1016/j.brainresrev.2007.07.019.
10.1111/j.1476-5381.1995.tb16327.x.
[194] Filli L, Zörner B, Weinmann O, Schwab ME. Motor deficits and recovery in rats with unilateral spinal cord hemisection mimic the Brown-Sequard syndrome. Brain. 2011;134:2261–2273. DOI: 10.1093/brain/awr167.
[182] Murray KC, Stephens MJ, Ballou EW, Heckman CJ, Bennett DJ. Motoneuron excitabil‐ ity and muscle spasms are regulated by 5-HT2B and 5-HT2C receptor activity. J
[183] Perrier JF, Hounsgaard J. 5-HT2 receptors promote plateau potentials in turtle spinal motoneurons by facilitating an L-type calcium current. J Neurophysiol. 2003;89:954–
[184] Chen J, Gu B, Wu G, Tu H, Si J, Xu Y, Andersson KE. The effect of the 5-HT2A/2C receptor agonist DOI on micturition in rats with chronic spinal cord injury. J Urol.
[185] Bos R, Sadlaoud K, Boulenguez P, Buttigieg D, Liabeuf S, Brocard C, Haase G, Bras H, Vinay L. Activation of 5-HT2A receptors upregulates the function of the neuronal K-Cl cotransporter KCC2. Proc Natl Acad Sci USA. 2013;110:348–353. DOI: 10.1073/
[186] Gackière F, Vinay L. Serotonergic modulation of post-synaptic inhibition and locomo‐ tor alternating pattern in the spinal cord. Front Neural Circuits. 2014;8:102. DOI:
[187] Rende M, Morales M, Brizi E, Bruno R, Bloom F, Sanna PP. Modulation of serotonin 5- HT3 receptor expression in injured adult rat spinal cord motoneurons. Brain Res.
[188] Gang W, Hongjian T, Jasheng C, Jiemin S, Zhong C, Yuemin X, Baojun G, Andersson KE. The effect of the 5-HT7 serotonin receptor agonist, LP44, on micturition in rats with chronic spinal cord injury. Neurourol Urodyn. 2014;33:1165–1170. DOI: 10.1002/nau.
[189] Faden AI, Gannon A, Basbaum AI. Use of serotonin immunocytochemistry as a marker of injury severity after experimental spinal trauma in rats. Brain Res. 1988;450:94–100.
[190] Holmes GM, Van Meter MJ, Beattie MS, Bresnahan JC. Serotonergic fiber sprouting to external anal sphincter motoneurons after spinal cord contusion. Exp Neurol.
[191] Hadjiconstantinou M, Panula P, Lackovic Z, Neff NH. Spinal cord serotonin: a biochemical and immunohistochemical study following transection. Brain Res.
[192] Bregman BS. Development of serotonin immunoreactivity in the rat spinal cord and its plasticity after neonatal spinal cord lesions. Brain Res. 1987;431:245–263. DOI:
[193] Saruhashi Y, Young W, Perkins R. The recovery of 5-HT immunoreactivity in lumbo‐ sacral spinal cord and locomotor function after thoracic hemisection. Exp Neurol.
Neurophysiol. 2011;105:731–748. DOI: 10.1152/jn.00774.2010.
2013;189:1982–1988. DOI: 10.1016/j.juro.2012.11.049.
1999;823:234–240. DOI: 10.1016/S0006-8993(99)01180-4.
2005;193:29–42. DOI: 10.1016/j.expneurol.2005.01.002.
1984;322:245–254. DOI: 10.1016/0006-8993(84)90114-8.
1996;39:203–213. DOI: 10.1006/exnr.1996.0094.
959. DOI: 10.1152/jn.00753.2002.
134 Recovery of Motor Function Following Spinal Cord Injury
pnas.1213680110.
22463.
10.3389/fncir.2014.00102.
DOI: 10.1016/0006-8993(88)91548-X.
10.1016/0165-3806(87)90213-6.
## **Docosahexaenoic Acid Promotes Recovery of Motor Function by Neuroprotection and Neuroplasticity Mechanisms**
John V. Priestley and Adina T. Michael-Titus
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/62709
### **Abstract**
The omega-3 polyunsaturated fatty acid, docosahexaenoic acid (DHA), has been shown to promote recovery of motor function after spinal cord injury. This is likely to be at least partly due to neuroprotective effects of DHA. However, recent studies have shown that DHA also supports neuroplasticity after injury, such as promoting sprouting of spared corticospinal tract (CST) axons. In this chapter, we review the published studies showing that DHA promotes recovery of motor function in rodent models of spinal cord injury (SCI),andconsidertheavailabledataontheunderlyingmechanisms.Thisincludeseffects on inflammation and on neuronal and oligodendrocyte survival at the injury site, and effects on spared CST axons and serotonergic axons. Current data support the hypothe‐ sis that DHA promotes recovery of motor function by both neuroprotection and neuroplasticity mechanisms. The significance of this, and the implications of combin‐ ing DHA with rehabilitation strategies, will be discussed.
**Keywords:** thoracic spinal cord injury, cervical spinal cord injury, central pattern gen‐ erator, V2a interneurons, DHA
### **1. Introduction**
### **1.1. Polyunsaturated fatty acids and their role in neurology**
Polyunsaturated fatty acids (PUFA) of the omega-3 and omega-6 series are lipids with major structural and signalling roles. The long-chain omega-3 PUFA docosahexaenoic acid (DHA) is present in significant concentrations in the central nervous system (CNS), where its synthesis
© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
from the dietary precursor alpha-linolenic acid (LNA) occurs through desaturation, elonga‐ tion andβ-oxidation reactions.Its intermediateprecursors are eicosapentaenoic acid(EPA) and docosapentaenoic acid (DPA). In humans, the conversion of LNA into DHA is below 5% [1, 2]. Therefore, an adequate dietary supply of DHA is required, especially because due to modern foodmanufacturing,therehasbeenasignificantdecreaseintheomega-3intakeandtheomega-3 to omega-6 PUFA dietary ratio in the Western diet [3]. These changes have been linked to an increase in the incidence and the prevalence of diseases such as cancer, cardiovascular dis‐ ease, rheumatoid arthritis, osteoporosis and asthma [4]. After absorption, the omega-3 PUFA can cross cellular membranes through specializedfatty acidtransporters andinteract with fatty acid-binding proteins (FABP). These proteins are a family of lipid chaperones involved in the extracellular andintracellulartransportoffattyacids [5].FABPbindlong-chainfattyacids,with different ligand selectivity and binding affinity profiles. Changes in FABP levels have been reported following injury. For example, B-FABP is found in the brain and has very high affinity for DHA; it is detected in plasma as a high specificity and sensitivity biomarker of brain injury [6]. More recently, Figueroa et al. [7] reported increases in tissue levels of FABP-5 after spinal cord injury.
The principal omega-3 PUFA in the brain is DHA, representing 10–20% of the total fatty acid composition. Retinal tissue is also highly enriched in DHA. This fatty acid is a component of phospholipids, in particular phosphatidylethanolamine and phosphatidylserine; it is concen‐ trated in cytoplasmic and synaptosomal membranes, growth cones, microsomal and mito‐ chondrial membranes, and is also a component of myelin [8]. DHA-enriched phospholipids are present in the innerleaflet of the cytoplasmic membrane, and DHA chains are very flexible and transition rapidly between conformational states. This creates a unique microenviron‐ ment for the functioning of many proteins embedded in membranes, such as neurotransmit‐ ter receptors [9]. After release from membrane phospholipids by phospholipases, long-chain omega-3 PUFA such as DHA can be metabolized and produce a wide range of metabolites, designated with the generic term of "docosanoids", which include protectins, D-series resolvins and maresins. These DHA metabolites have intrinsic biological effects, which are mediated through specific receptors [10, 11]. Therefore, some of the effects induced by DHA may be due not to a direct action of the fatty acid but to specific metabolites which are derived from it, and which could attain significant levels, especially under conditions of chronic DHA supplementation [12].
DHA has an essential role in normal neurodevelopment, in particular for vision and cogni‐ tion [13]. Supplementation with DHA has been shown to improve various aspects of learn‐ ing and memory in children [14]. More than two decades ago, Martinez and collaborators [15] carried out seminal studies in children with peroxysomal disorders, which lead to a major DHA deficit, and showed that supplementation with DHA can lead to significant improve‐ ment in neurological function, and this is accompanied by improved myelination of the immature brain. At around the same time, studies initiated by Lazdunski and collaborators were providing evidence that omega-3 PUFA had significant neuroprotective potential under conditions which lead to CNS injury, such as seizures and ischaemia [16]. DHA has been increasingly linked to a variety of conditions, such as Alzheimer's disease and Parkinson's
disease, and also schizophrenia, depression and attention deficit-hyperactivity disorder [17– 19], and this has consolidated the idea that dietary supplementation with DHA could have therapeutic value across a spectrum of major disorders in neurology and psychiatry.
### **1.2. Pathways controlling movement in the rat and possible strategies for promoting recovery of movement following spinal cord injury (SCI)**
from the dietary precursor alpha-linolenic acid (LNA) occurs through desaturation, elonga‐ tion andβ-oxidation reactions.Its intermediateprecursors are eicosapentaenoic acid(EPA) and docosapentaenoic acid (DPA). In humans, the conversion of LNA into DHA is below 5% [1, 2]. Therefore, an adequate dietary supply of DHA is required, especially because due to modern foodmanufacturing,therehasbeenasignificantdecreaseintheomega-3intakeandtheomega-3 to omega-6 PUFA dietary ratio in the Western diet [3]. These changes have been linked to an increase in the incidence and the prevalence of diseases such as cancer, cardiovascular dis‐ ease, rheumatoid arthritis, osteoporosis and asthma [4]. After absorption, the omega-3 PUFA can cross cellular membranes through specializedfatty acidtransporters andinteract with fatty acid-binding proteins (FABP). These proteins are a family of lipid chaperones involved in the extracellular andintracellulartransportoffattyacids [5].FABPbindlong-chainfattyacids,with different ligand selectivity and binding affinity profiles. Changes in FABP levels have been reported following injury. For example, B-FABP is found in the brain and has very high affinity for DHA; it is detected in plasma as a high specificity and sensitivity biomarker of brain injury [6]. More recently, Figueroa et al. [7] reported increases in tissue levels of FABP-5 after spinal
The principal omega-3 PUFA in the brain is DHA, representing 10–20% of the total fatty acid composition. Retinal tissue is also highly enriched in DHA. This fatty acid is a component of phospholipids, in particular phosphatidylethanolamine and phosphatidylserine; it is concen‐ trated in cytoplasmic and synaptosomal membranes, growth cones, microsomal and mito‐ chondrial membranes, and is also a component of myelin [8]. DHA-enriched phospholipids are present in the innerleaflet of the cytoplasmic membrane, and DHA chains are very flexible and transition rapidly between conformational states. This creates a unique microenviron‐ ment for the functioning of many proteins embedded in membranes, such as neurotransmit‐ ter receptors [9]. After release from membrane phospholipids by phospholipases, long-chain omega-3 PUFA such as DHA can be metabolized and produce a wide range of metabolites, designated with the generic term of "docosanoids", which include protectins, D-series resolvins and maresins. These DHA metabolites have intrinsic biological effects, which are mediated through specific receptors [10, 11]. Therefore, some of the effects induced by DHA may be due not to a direct action of the fatty acid but to specific metabolites which are derived from it, and which could attain significant levels, especially under conditions of chronic DHA
DHA has an essential role in normal neurodevelopment, in particular for vision and cogni‐ tion [13]. Supplementation with DHA has been shown to improve various aspects of learn‐ ing and memory in children [14]. More than two decades ago, Martinez and collaborators [15] carried out seminal studies in children with peroxysomal disorders, which lead to a major DHA deficit, and showed that supplementation with DHA can lead to significant improve‐ ment in neurological function, and this is accompanied by improved myelination of the immature brain. At around the same time, studies initiated by Lazdunski and collaborators were providing evidence that omega-3 PUFA had significant neuroprotective potential under conditions which lead to CNS injury, such as seizures and ischaemia [16]. DHA has been increasingly linked to a variety of conditions, such as Alzheimer's disease and Parkinson's
cord injury.
138 Recovery of Motor Function Following Spinal Cord Injury
supplementation [12].
Locomotion in rodents is controlled by a central pattern generator (CPG) which generates the basic rhythm of limb flexion and extension and which, forthe hindlimbs, is located in the lower thoracic and upper lumbar spinal cord. In recent years, great progress has been made in dissecting the organization of the CPG, by identifying subtypes of neurons based on their developmental expression of specific transcription factors (summarized in [20, 21]). Genetic ablation of specific transcription factors can then be used to probe the functional roles of different neuronal subtypes (see e.g. [22, 23]). **Figure 1** illustrates one possible model of the
**Figure 1.** Schematic model of the rodent CPG controlling hind limb movement. Excitatory neurons are shown in red, and inhibitory neurons are shown in green. Rhythm generator (RG), pattern formation (PF) and motoneurons (Mn) are present for both flexion and extension on both sides of the spinal cord. V0d, V0v, V0c, V1, V2a, V2b and V3 are differ‐ ent types of interneurons which in some cases correspond to existing well-characterized neuronal types (e.g. V1 neu‐ rons include Renshaw cells). V2a neurons are shown in blue and have direct projections to motoneurons, plus projections to V2b/V1 PF inhibitory interneurons and to V0v excitatory commissural interneurons. Afferent feedback and descending control occur bilaterally, but for simplicity have only been shown ipsilaterally. The diagram is based on diagrams contained in Harris-Warrick [20] and Rybak and colleagues [21].
CPG based on studies reported in [22, 23]. Flexor and extensor motoneurons are thought to have separate rhythm generators (RG) which provide input to a pattern formation (PF) network that controls flexor/extensor alternation. The RG network also connects to commis‐ sural interneurons which coordinate left/right hindlimb movement. As an example of the complexity of the system, we have highlighted the V2a class of excitatory interneuron in **Figure 1**. V2a neurons can be identified by their expression of the homeodomain protein Chx10 and receive input from the RG centres. They project to three different targets, namely (1) motoneurons, (2) V2b inhibitory interneurons that are part of the PF centre, and (3) V0v commissural interneurons [21]. They therefore can influence locomotion by direct monosy‐ naptic inputs to motoneurons and by regulating the flexor/extensor and left/right coordinat‐ ing centres. However, V2a interneurons are not only present in the lumbar CPG, they also occurin the cervical spinal cordandplay a key role in skilledforelimb movement, by conveying an internal copy of the premotor signals. This is possible because of their dual innervation of cervical motoneurons and precerebellar interneurons [22].
The CPG can generate alternating hindlimb movements independently, but in the intact animal, the CPG is modulated by afferent feedback and by descending control. The effects of SCI depend on the spinal level of the injury. Rats subject to complete transection above the level of the lumbar locomotor centres show some spontaneous recovery of hindlimb move‐ ment [24], and this can be increased by a variety of treatments, including weight supported treadmill training [24], delivery of growth factors [25, 26] and implantation of various exogenous cells [27–29]. At least some of this recovery of function appears to be related to regeneration or sprouting of descending bulbospinal [30, 31] and corticospinal [25] path‐ ways. In contrast, the lasting paraplegia that results from a SCI at lumbar level is thought to result from loss of spinal interneurons, possibly including components of the CPG [32]. An efficacious treatment for SCI might need to preserve and/or restore descending control pathways and repair damage to intrinsic networks such as the CPG.
### **2. The effects of PUFA treatment on recovery of movement following experimental SCI**
### **2.1. Thoracic hemisection SCI**
DHA promotes recovery of quadrupedal locomotion in rats and mice following thoracic hemisection [33], compression [34–36] and contusion [37] SCI, when acutely administered post-injury. Recovery is further improved if an acute injection of DHA is followed by a DHAsupplemented diet (**Figure 2**). The DHA treatment results in increased neuronal and oligo‐ dendrocyte survival [34] and decreased macrophage and microglial activation [36, 38] at the injury site. However, since the motoneuron pools that contribute to locomotion are not at thoracic level, it is likely that the effects on locomotion are due to preservation of descend‐ ing axons passing through the injury site rather than preservation of thoracic neurons themselves. Analysis of axon numbers at the SCI site using generic axonal markers such as
neurofilament and myelin basic protein (MBP) reveals that DHA preserves axons in white matter tracts [39] but, with the exception of serotonin (5-hydroxytryptamine, 5-HT), the origin and role of those axons have not been much explored. As outlined above, corticospinal and bulbospinal projections play an important role in motor control but mainly innervate moto‐ neurons indirectly, via interneurons, and so are difficult to analyse (however, see Section 4 on pyramidotomy below). In contrast, 5-HT axons are derived from the medullary raphe nuclei, innervate many motoneurons monosynaptically and play key roles in modulating motoneur‐ on excitability [40]. DHA treatment preserves the number of 5-HT axons at the level of a thoracic compression injury [39], including axons that appear to contact motoneurons (see **Figure 3**). It is therefore likely that at least some of the motor recovery is due to increased preservation of 5-HT axons that pass through the injury site and innervate motoneurons at lumbar level. 5-HT axons have a great capacity to sprout and make new connections [41]. Another possibility is therefore that some 5-HT axons are spared by the SCI but that DHA promotes sprouting and restoration of synaptic circuits by these axons. These possibilities have not been examined directly following thoracic SCI, but the effects of DHA on 5-HT axons have been studied in much more detail following cervical injury.
CPG based on studies reported in [22, 23]. Flexor and extensor motoneurons are thought to have separate rhythm generators (RG) which provide input to a pattern formation (PF) network that controls flexor/extensor alternation. The RG network also connects to commis‐ sural interneurons which coordinate left/right hindlimb movement. As an example of the complexity of the system, we have highlighted the V2a class of excitatory interneuron in **Figure 1**. V2a neurons can be identified by their expression of the homeodomain protein Chx10 and receive input from the RG centres. They project to three different targets, namely (1) motoneurons, (2) V2b inhibitory interneurons that are part of the PF centre, and (3) V0v commissural interneurons [21]. They therefore can influence locomotion by direct monosy‐ naptic inputs to motoneurons and by regulating the flexor/extensor and left/right coordinat‐ ing centres. However, V2a interneurons are not only present in the lumbar CPG, they also occurin the cervical spinal cordandplay a key role in skilledforelimb movement, by conveying an internal copy of the premotor signals. This is possible because of their dual innervation of
The CPG can generate alternating hindlimb movements independently, but in the intact animal, the CPG is modulated by afferent feedback and by descending control. The effects of SCI depend on the spinal level of the injury. Rats subject to complete transection above the level of the lumbar locomotor centres show some spontaneous recovery of hindlimb move‐ ment [24], and this can be increased by a variety of treatments, including weight supported treadmill training [24], delivery of growth factors [25, 26] and implantation of various exogenous cells [27–29]. At least some of this recovery of function appears to be related to regeneration or sprouting of descending bulbospinal [30, 31] and corticospinal [25] path‐ ways. In contrast, the lasting paraplegia that results from a SCI at lumbar level is thought to result from loss of spinal interneurons, possibly including components of the CPG [32]. An efficacious treatment for SCI might need to preserve and/or restore descending control
cervical motoneurons and precerebellar interneurons [22].
140 Recovery of Motor Function Following Spinal Cord Injury
pathways and repair damage to intrinsic networks such as the CPG.
**experimental SCI**
**2.1. Thoracic hemisection SCI**
**2. The effects of PUFA treatment on recovery of movement following**
DHA promotes recovery of quadrupedal locomotion in rats and mice following thoracic hemisection [33], compression [34–36] and contusion [37] SCI, when acutely administered post-injury. Recovery is further improved if an acute injection of DHA is followed by a DHAsupplemented diet (**Figure 2**). The DHA treatment results in increased neuronal and oligo‐ dendrocyte survival [34] and decreased macrophage and microglial activation [36, 38] at the injury site. However, since the motoneuron pools that contribute to locomotion are not at thoracic level, it is likely that the effects on locomotion are due to preservation of descend‐ ing axons passing through the injury site rather than preservation of thoracic neurons themselves. Analysis of axon numbers at the SCI site using generic axonal markers such as
**Figure 2.** Locomotor performance of thoracic compression SCI rats injected with saline or DHA and of animals injected with DHA and then fed on a DHA-enriched diet. Results for days 1–7 are shown in the bottom half of the graph, plot‐ ted against the bottom X-axis. Results for 2–6 weeks are shown in the top half of the graph, plotted against the top Xaxis. Both DHA- and saline-treated animals showed similar low levels of locomotor function 1 day after surgery, as assessed on the BBB open field task. During the first week, DHA-treated animals had improved motor function recov‐ ery compared with saline-treated animals (days 5, 6 and 7; \**p* < 0.05). Between week 2 and week 6, DHA-treated ani‐ mals continued to perform better than saline-treated animals (\**p* < 0.05). In addition, from the 4th week onwards, animals injected with DHA followed by maintenance on the DHA-enriched diet, performed better than animals only injected with DHA (weeks 4, 5 and 6; <sup>Δ</sup>*p* < 0.05). Error bars represent SEMs. Reproduced from [34] with permission.
**Figure 3.** Immunofluorescence micrographs showing thin, 5-HT immunoreactive (IR) axons (red) that innervate NeuN labelled neurons (green) in the ventral horn (VH). In naïve animals (A), a dense network of beaded 5-HT-positive fi‐ bres can be seen surrounding large NeuN neurons. At 6 weeks post-thoracic compression injury, a marked loss of 5- HT-positive fibres was seen in saline-treated animals (B) compared with the groups receiving DHA injection alone (C), or in combination with a DHA-enriched diet (D). Quantitative analysis (E) at 6 weeks post-injury revealed a significant loss of 5-HT IR in the VH of saline-treated animals following injury. A significant amelioration in the loss of 5-HT IR was observed in the VH in both DHA-treatment groups (significantly different from \*naïve or #saline-treated animals at *p* < 0.05; scale bar = 100 μm). Reproduced from [39] with permission.
### **2.2. Experimental cervical SCI**
The majority of experimental studies have focused on thoracic injury, but the most common clinical injury is at cervical level (57% of SCI). Recently, we have therefore examined the effects of DHA in a cervical (C5 level) hemisection SCI model [42]. The cervical level of injury also has the advantage of allowing us to examine the effects of DHA treatment on a skilled motor task, namely retrieval offood using the forepaws. Rats are trained to retrieve food by extending a forelimb through a gap in a perspex box and grasping food pellets which are on a staircase [43]. Following SCI at C5 level, rats lose the ability to recover the food pellets, but recover partial function if treated acutely with DHA (**Figure 4**). Following cervical hemisection, 5-HT axons show a remarkable degree of plasticity. Rostral to the hemisection, there is an increase in 5-HT immunoreactivity which is maintained up to 3 weeks. Caudal to the hemisection, there is a transient loss of 5-HT immunoreactivity at 1 week, but levels are restored by 3 weeks (**Figure 5B**). DHA treatment results in an even greater number of 5-HT immunoreactive axons caudal to the injury (**Figure 5B**), and increased density of 5-HT axons contacting motoneur‐ ons (**Figure 5C**–**E**). However, 5-HT axons are not the only ones to sprout in response to DHA treatment. Corticospinal tract (CST) axons on the intact side of the spinal cord were identi‐ fied by anterograde labelling with the tracer biotinylated dextran amine (BDA). Following hemisection, a small number of these axons sprout across the midline to innervate the denervated side of the spinal cord. However, after DHA treatment, this number doubles [42]. It is therefore possible that the recovery of skilled movement is due to sprouting of 5-HT and CST axons and formation of novel circuits. However, since the increase in axons occurs at the same level as the hemisection, it is possible that the increase in 5-HT and CST axons is due to neuroprotection rather than neuroplasticity. We have therefore examined the effect of DHA treatment in a third model of SCI, namely unilateral pyramidotomy in the mouse [42].
**Figure 3.** Immunofluorescence micrographs showing thin, 5-HT immunoreactive (IR) axons (red) that innervate NeuN labelled neurons (green) in the ventral horn (VH). In naïve animals (A), a dense network of beaded 5-HT-positive fi‐ bres can be seen surrounding large NeuN neurons. At 6 weeks post-thoracic compression injury, a marked loss of 5- HT-positive fibres was seen in saline-treated animals (B) compared with the groups receiving DHA injection alone (C), or in combination with a DHA-enriched diet (D). Quantitative analysis (E) at 6 weeks post-injury revealed a significant loss of 5-HT IR in the VH of saline-treated animals following injury. A significant amelioration in the loss of 5-HT IR was observed in the VH in both DHA-treatment groups (significantly different from \*naïve or #saline-treated animals
The majority of experimental studies have focused on thoracic injury, but the most common clinical injury is at cervical level (57% of SCI). Recently, we have therefore examined the effects
at *p* < 0.05; scale bar = 100 μm). Reproduced from [39] with permission.
**2.2. Experimental cervical SCI**
142 Recovery of Motor Function Following Spinal Cord Injury
**Figure 4.** DHA enhances functional recovery following cervical lateral hemisection in the rat. A, B, In the Montoya staircase test, all injured animals lost the ability to displace (gross motor function) or eat (fine motor function) the food pellet after cervical lateral hemisection 2 d after injury. The animals treated with vehicle did not recover any food re‐ trieval ability (blue squares), but DHA-treated animals gradually recovered food retrieval ability (red circles) from around 2 weeks onwards, compared with uninjured sham operated animals (green triangles). \**p* < 0.05, \*\**p* < 0.01, \*\*\**p* < 0.001, compared to vehicle-treated animals. Results represent mean ± SEM. Reproduced from [42] with permission.
**Figure 5.** DHA enhances 5-HT (serotonin) fibre regrowth following cervical lateral hemisection in the rat. (*A*) Serotonin fibres in the rostral part of the lesion site were significantly increased in both treatment groups at 3 weeks post-injury, but only the DHA-treated group (red circles), not the vehicle group (blue squares), had significantly increased seroto‐ nin in the caudal region compared with the sham operated animals (green triangles). (*B*) Serotonin fibres in the rostral part of the lesion site were significantly increased at 1 week (red square) and 3 weeks (blue circle) compared with the sham operated group after lateral cervical hemisection. However, in the caudal region, 1 and 3 weeks after injury were similar to the sham operated group. 1 week after hemisection, there was a loss of serotonin but by 3 weeks levels had returned to values similar to sham operated animals. *N* = 5 or 6 per group. #*p* < 0.05, 1 week hemisection versus sham group. \**p* < 0.05, 3 week hemisection versus sham group. (*C*, *D*) The images were captured 1750 μm caudal to the lesion site. Double labelling shows serotonin fibres (green) in contact with ChAT immunoreactive motor neurons (red) in the ventral horn of vehicle- and DHA-treated rats. Insets, dashed boxes at higher magnification. (*E*) Quantitative analysis reveals that there is a significant increase in the density of serotonin fibres contacting motor neurons in the DHA-treated group compared with the control group. N = 5 or 6 per group. \**p* < 0.05, DHA versus vehicle group. †*p* < 0.05, sham versus vehicle group. #*p* < 0.05, DHA versus sham group. Scale bar, 100 μm. Reproduced from [42] with permission.
#### **2.3. Pyramidotomy**
The CST axons in rodents run in the medullary pyramids, before decussating and descend‐ ing at the base of the dorsal columns. Unilateral pyramidotomy transects these CST axons and
**Figure 6.** DHA enhances functional recovery via connection with Chx10 interneurons in pyramidotomy mice. (*A–D*), Confocal images of mouse cervical spinal cord transverse sections at the intermediate laminae of the CST-denervated side, showing examples of BDA-labelled CST collaterals (green) in the vicinity of blue fluorescent Nissl-stained cells (Neurotrace 435/455), some of which are V2a interneurons identified by immunostaining for Chx10 (red). (*A*, *B*) Dash‐ ed boxes represent high magnification in *A*', and in *C*, *D*, which reveal contacts (arrows) between BDA-labelled collat‐ erals and Chx10 interneurons (asterisks). (*A'*) *A Z*-stack comprising 10 × 0.66 μm optical images. (*C*), A stack comprising 25 × 0.72 μm optical images. (*D*, A) Single 0.72 μm optical image. Examination of the single optical image confirms that the BDA-labelled CST fibre contacts (arrow) one of the Chx10 interneurons. (*E*) Quantitative analysis re‐ vealed a significant increase in the number of Chx10 interneurons contacted by BDA-labelled CST collaterals following DHA treatment (red bar) compared with vehicle treatment (blue bar). (*F*) A strong negative correlation was observed between the numbers of Chx10 interneurons with BDA-labelled CST contacts and the numbers of forelimb misplace‐ ments. Data were taken from DHA-treated (red circle) and vehicle-treated (blue square) animals. \*\*\**p*<0.001, DHA ver‐ sus vehicle group. Scale bar, 20 μm. Reproduced from [42] with permission.
**Figure 5.** DHA enhances 5-HT (serotonin) fibre regrowth following cervical lateral hemisection in the rat. (*A*) Serotonin fibres in the rostral part of the lesion site were significantly increased in both treatment groups at 3 weeks post-injury, but only the DHA-treated group (red circles), not the vehicle group (blue squares), had significantly increased seroto‐ nin in the caudal region compared with the sham operated animals (green triangles). (*B*) Serotonin fibres in the rostral part of the lesion site were significantly increased at 1 week (red square) and 3 weeks (blue circle) compared with the sham operated group after lateral cervical hemisection. However, in the caudal region, 1 and 3 weeks after injury were similar to the sham operated group. 1 week after hemisection, there was a loss of serotonin but by 3 weeks levels had returned to values similar to sham operated animals. *N* = 5 or 6 per group. #*p* < 0.05, 1 week hemisection versus sham group. \**p* < 0.05, 3 week hemisection versus sham group. (*C*, *D*) The images were captured 1750 μm caudal to the lesion site. Double labelling shows serotonin fibres (green) in contact with ChAT immunoreactive motor neurons (red) in the ventral horn of vehicle- and DHA-treated rats. Insets, dashed boxes at higher magnification. (*E*) Quantitative analysis reveals that there is a significant increase in the density of serotonin fibres contacting motor neurons in the DHA-treated group compared with the control group. N = 5 or 6 per group. \**p* < 0.05, DHA versus vehicle group. †*p* < 0.05, sham versus vehicle group. #*p* < 0.05, DHA versus sham group. Scale bar, 100 μm. Reproduced from [42] with
The CST axons in rodents run in the medullary pyramids, before decussating and descend‐ ing at the base of the dorsal columns. Unilateral pyramidotomy transects these CST axons and
permission.
**2.3. Pyramidotomy**
144 Recovery of Motor Function Following Spinal Cord Injury
results in the loss of CST axons in the spinal cord on the contralateral side. Using BDA to label the CSTaxons on the non-lesionedside, we observedthat a small number of CSTaxons respond to the pyramidotomy by sprouting across the midline to innervate the lesioned side (**Figure 6A**). The number of sprouting CST axons is doubled following DHA treatment [42]. Furthermore, we have shown that the sprouting axons contact a particular class of interneur‐ on, namely the V2a propriospinal neurons described above. V2a neurons were identified by their expression of the Chx10 transcription factor (**Figure 6A**–**D**). V2a interneurons have been shown to play an important role in skilled reaching [22]. Following DHA treatment and pyramidotomy, an increased number of V2a interneurons receive contacts from CST axons (**Figure 6E**), and the recovery of forelimb movement shows a tight correlation with the number of V2a neurons thatreceive contacts (**Figure 6F**). It therefore appears that DHA treatment after pyramidotomy results in the formation of a novel circuit (CST to contralateral V2a interneur‐ ons) that promotes recovery of skilled reaching. This is in addition to the effects of DHA on sprouting 5-HT axons, and neuroprotective effects on neuronal cell bodies and axons.
### **3. Discussion**
Motor pathways are so complex that it is difficult to establish a direct causal relationship between the effects of a therapeutic agent on spinal circuitry and an improvement in a particular motor task. However, the data reviewed above indicates that there is a strong correlation between the neuroprotective and neuroplasticity effects of DHA and the recov‐ ery of motor function. This dual effect of DHA, together with its established safety profile and the recent demonstration of the efficacy in SCI of a multinutrient combination containing DHA and EPA [44], makes DHA a particularly promising candidate for development as a thera‐ peutic agent in SCI. There is a substantial literature on the neuroprotective effects of DHA (reviewed above) but the effects on neuroplasticity are more novel and less studied. We have shown that DHA treatment following SCI upregulates the microRNA miR-21 and suppress‐ es phosphatase and tensin homolog (PTEN), a central negative regulator of the phosphatidy‐ linositol 3-kinase (PI3K) signalling pathway [42]. However, these effects are quite widespread and it is not known how they result in beneficial axon sprouting. Bareyre and colleagues [45] have shown that new connections form after SCI, but inappropriate connections are lost because they are not used. A similar mechanism may mould the sprouting and connections promoted by DHA, in which case, there may be great benefit in combining DHA treatment with rehabilitation and task specific training.
### **3.1. Therapeutic implications**
The magnitude of the effects induced by the acute administration of DHA and reported in various models of SCI is comparable with that described with various other therapeutic approaches which are already being explored in clinical trials in neurotrauma, such as progesterone, erythropoietin, riluzole and minocycline [46]. DHA has clear potential for clinical translation in SCI; therefore, it is appropriate to consider the issues which remain to be clarified in order to improve the chances of translational success.
results in the loss of CST axons in the spinal cord on the contralateral side. Using BDA to label the CSTaxons on the non-lesionedside, we observedthat a small number of CSTaxons respond to the pyramidotomy by sprouting across the midline to innervate the lesioned side (**Figure 6A**). The number of sprouting CST axons is doubled following DHA treatment [42]. Furthermore, we have shown that the sprouting axons contact a particular class of interneur‐ on, namely the V2a propriospinal neurons described above. V2a neurons were identified by their expression of the Chx10 transcription factor (**Figure 6A**–**D**). V2a interneurons have been shown to play an important role in skilled reaching [22]. Following DHA treatment and pyramidotomy, an increased number of V2a interneurons receive contacts from CST axons (**Figure 6E**), and the recovery of forelimb movement shows a tight correlation with the number of V2a neurons thatreceive contacts (**Figure 6F**). It therefore appears that DHA treatment after pyramidotomy results in the formation of a novel circuit (CST to contralateral V2a interneur‐ ons) that promotes recovery of skilled reaching. This is in addition to the effects of DHA on
sprouting 5-HT axons, and neuroprotective effects on neuronal cell bodies and axons.
Motor pathways are so complex that it is difficult to establish a direct causal relationship between the effects of a therapeutic agent on spinal circuitry and an improvement in a particular motor task. However, the data reviewed above indicates that there is a strong correlation between the neuroprotective and neuroplasticity effects of DHA and the recov‐ ery of motor function. This dual effect of DHA, together with its established safety profile and the recent demonstration of the efficacy in SCI of a multinutrient combination containing DHA and EPA [44], makes DHA a particularly promising candidate for development as a thera‐ peutic agent in SCI. There is a substantial literature on the neuroprotective effects of DHA (reviewed above) but the effects on neuroplasticity are more novel and less studied. We have shown that DHA treatment following SCI upregulates the microRNA miR-21 and suppress‐ es phosphatase and tensin homolog (PTEN), a central negative regulator of the phosphatidy‐ linositol 3-kinase (PI3K) signalling pathway [42]. However, these effects are quite widespread and it is not known how they result in beneficial axon sprouting. Bareyre and colleagues [45] have shown that new connections form after SCI, but inappropriate connections are lost because they are not used. A similar mechanism may mould the sprouting and connections promoted by DHA, in which case, there may be great benefit in combining DHA treatment
The magnitude of the effects induced by the acute administration of DHA and reported in various models of SCI is comparable with that described with various other therapeutic approaches which are already being explored in clinical trials in neurotrauma, such as progesterone, erythropoietin, riluzole and minocycline [46]. DHA has clear potential for
**3. Discussion**
146 Recovery of Motor Function Following Spinal Cord Injury
with rehabilitation and task specific training.
**3.1. Therapeutic implications**
It is important to note that in the studies published so far with DHA and reporting beneficial effects of this fatty acid in a variety of conditions, there are two types of administration: DHA administered using a chronic oral supplementation route and DHA administered as an acute bolus, using an injectable route. Are the cellular targets and mechanisms activated by DHA the same in the two types of treatment? When DHA is administered chronically, one of the consequences of this regime is the structural enrichment in DHA in membranes, and be‐ cause DHA has unique molecular structural characteristics, this influences the dynamics of membrane components, and may change the activity of ion channels and G-protein-coupled receptors (GPCR). DHA incorporation in membranes changes the properties of specialized domains such as the lipid rafts and caveolae, and this can modify signalling [47]. When DHA is administered acutely as a bolus, its half-life in plasma is very short (approximately 2 min), as shown in PET imaging studies carried out in healthy volunteers [48]; therefore, efficacy may be due to the activation of different mechanisms. Long-chain omega-3 PUFA such as DHA and EPA have various specific molecular targets, including ion channels, GPCR and nuclear receptors [17, 47]. They include voltage-sensitive Na+ and Ca2+ channels and two-pore domain background K+ channels, such as TREK-1 (which is also a target forriluzole), the GPCRs GPR40 and GPR120, and transcription factors such as the retinoid X receptors (RXR) and peroxi‐ some proliferator-activated receptors (PPAR). RXR can heterodimerize with PPAR or with retinoic acid receptors (RAR), and changes in the expression of these receptors have been reported after SCI [49]. Several studies indicate that the activation of retinoid signalling supports axonal regeneration [50].
The spontaneous partial recovery of function which occurs after SCI may be due to neuro‐ plasticity, compensation and repair [51]. Rehabilitation through training is at present the most successful treatment for patients with SCI, to enhance the recovery of some neurological function. Rehabilitation enhances the spontaneous plasticity changes that occur after injury and enhances the activity of sensorimotor pathways, as documented in experimental SCI [52]. Exercise leads to a reconfiguration of cortical representation maps, it modifies the biophysi‐ cal properties of motoneuron membranes, changes the activity of spinal inhibitory circuits and also increases the levels of neurotrophic factors, such as brain-derived neurotrophic factor (BDNF) and neurotrophin-3 (NT-3) [53]. Liu and collaborators [54] have shown that exercise increases the expression of miR21 and decreases PTEN mRNA levels, a profile of effects similar to our observations with DHA. Considering this similarity of action, a goal of future studies should be the exploration of a combination of DHA and exercise, to establish whether what could be achieved is true synergism or only additivity. The concept of combining treatment with exercise is supported by examples such as the successful combination of exercise with strategies such as the reduction in the glial scar components using chondroitinase ABC [55]. However, not all combinations achieve an optimum effect, as shown by the negative results obtained by combining exercise with antibodies against the myelin-derived inhibitor Nogo A [56]. Therefore, it will be essential to continue to characterize the targets and mode of action of DHA, so that it becomes clear what mechanisms could be harnessed and amplified by combining DHA (acute administration and chronic exposure) and rehabilitation training, to optimize outcome in SCI patients.
### **Acknowledgements**
We acknowledge the generous support of Spinal Research, Corporate Action Trust, Barts Charity and Chang Gung Memorial Hospital (Taiwan) for the DHA studies in the various models of SCI discussed here.
### **Author details**
John V. Priestley\* and Adina T. Michael-Titus
\*Address all correspondence to: j.v.priestley@qmul.ac.uk
Centre for Neuroscience and Trauma, Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK
### **References**
[8] HorrocksLA, FarooquiAA. Docosahexaenoic acid in the diet: its importance in maintenance and restoration of neural membrane function. Prostaglandins Leukot Essent Fatty Acids. 2004; 70: 361–372.
combining DHA (acute administration and chronic exposure) and rehabilitation training, to
We acknowledge the generous support of Spinal Research, Corporate Action Trust, Barts Charity and Chang Gung Memorial Hospital (Taiwan) for the DHA studies in the various
Centre for Neuroscience and Trauma, Blizard Institute, Barts and The London School of
[1] BrennaJT, SalemNJr., SinclairAJ, CunnaneSC. alpha-Linolenic acid supplementation and conversion to n-3 long-chain polyunsaturated fatty acids in humans. Prostaglan‐
[2] BrennaJT. Efficiency of conversion of alpha-linolenic acid to long chain n-3 fatty acids
[3] SimopoulosAP. Evolutionary aspects of omega-3 fatty acids in the food supply.
[4] SimopoulosAP. The importance of the omega-6/omega-3 fatty acid ratio in cardiovas‐ cular disease and other chronic diseases. Exp Biol Med (Maywood ). 2008; 233: 674–688.
[5] FuruhashiM, HotamisligilGS. Fatty acid-binding proteins: role in metabolic diseases
[6] PelsersMM, GlatzJF. Detection of brain injury by fatty acid-binding proteins. Clin Chem
[7] FigueroaJD, IllanMS, LiceroJ, CorderoK, MirandaJD, DeLM. Fatty acid binding protein 5 (FABP5) modulates docosahexaenoic acid (DHA)-induced recovery in rats
undergoing spinal cord injury. J Neurotrauma. March 2016, ahead of print.
and potential as drug targets. Nat Rev Drug Discov. 2008; 7: 489–503.
Medicine and Dentistry, Queen Mary University of London, London, UK
in man. Curr Opin Clin Nutr Metab Care. 2002; 5: 127–132.
Prostaglandins Leukot Essent Fatty Acids. 1999; 60: 421–429.
and Adina T. Michael-Titus
dins Leukot Essent Fatty Acids. 2009; 80: 85–91.
Lab Med. 2005; 43: 802–809.
\*Address all correspondence to: j.v.priestley@qmul.ac.uk
optimize outcome in SCI patients.
148 Recovery of Motor Function Following Spinal Cord Injury
**Acknowledgements**
models of SCI discussed here.
**Author details**
John V. Priestley\*
**References**
[33] KingVR, HuangWL, DyallSC, CurranOE, PriestleyJV, Michael-TitusAT. Omega-3 fatty acids improve recovery, whereas omega-6 fatty acids worsen outcome, after spinal cord injury in the adult rat. J Neurosci. 2006; 26: 4672–4680.
[22] AzimE, JiangJ, AlstermarkB, JessellTM. Skilled reaching relies on a V2a propriospinal
[23] CroneSA, QuinlanKA, ZagoraiouL, DrohoS, RestrepoCE, LundfaldL, EndoT, SetlakJ, JessellTM, KiehnO, SharmaK. Genetic ablation of V2a ipsilateral interneurons dis‐ rupts left-right locomotor coordination in mammalian spinal cord. Neuron. 2008; 60:
[24] ZhangY, JiSR, WuCY, FanXH, ZhouHJ, LiuGL. Observation of locomotor functional recovery in adult complete spinal rats with BWSTT using semiquantitative and
[25] FernandezE, PalliniR, LaurettiL, MercantiD, SerraA, CalissanoP, PiepmeierJM, TatorCH. Spinal cord transection in adultrats—effects of local infusion of nerve growth
[26] BoyceVS, ParkJ, GageFH, MendellLM. Differential effects of brain-derived neurotro‐ phic factor and neurotrophin-3 on hindlimb function in paraplegic rats. EurJ Neurosci.
[27] KubasakMD, JindrichDL, ZhongH, TakeokaA, McFarlandKC, Munoz-QuilesC, RoyRR, EdgertonVR, Ramon-CuetoA, PhelpsPE. OEG implantation and step training enhance hindlimb-stepping ability in adult spinal transected rats. Brain. 2008; 131: 264–
[28] RapalinoO, LazarovSpieglerO, AgranovE, VelanGJ, YolesE, FraidakisM, SolomonA, GepsteinR, KatzA, BelkinM, HadaniM, SchwartzM. Implantation of stimulated homologous macrophages results in partialrecovery of paraplegic rats. Nat Med. 1998;
[29] SlawinskaU, MajczynskiH, DjavadianR. Recovery of hindlimb motor functions after spinal cord transection is enhanced by grafts of the embryonic raphe nuclei. Exp Brain
[30] FouadK, SchnellL, BungeMB, SchwabME, LiebscherT, PearseDD. Combining Schwann cell bridges and olfactory-ensheathing glia grafts with chondroitinase promotes locomotor recovery after complete transection of the spinal cord. J Neurosci. 2005; 25:
[31] Lopez-ValesR, ForesJ, NavarroX, VerduE. Chronic transplantation of olfactory ensheathing cells promotes partial recovery after complete spinal cord transection in
[32] HadiB, ZhangYP, BurkeDA, ShieldsCB, MagnusonDS. Lasting paraplegia caused by loss of lumbar spinal cord interneurons in rats: no direct correlation with motor neuron
factor on the corticospinal tract axons. Neurosurgery. 1993; 33: 889–893.
internal copy circuit. Nature. 2014; 508: 357–363.
150 Recovery of Motor Function Following Spinal Cord Injury
qualitative methods. Spinal Cord. 2007; 45: 496–501.
70–83.
2012; 35: 221–232.
276.
4: 814–821.
1169–1178.
Res. 2000; 132: 27–38.
the rat. Glia. 2007; 55: 303–311.
loss. J Neurosurg. 2000; 93: 266–275.
tion improves outcome in experimental spinal cord injury. Neurobiol Dis. 2015; 82: 504– 515.
## **Role ofJAK-STAT Signalling on Motor Function Recovery after Spinal Cord Injury**
Victor S. Tapia and Juan Larrain
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/63418
### **Abstract**
tion improves outcome in experimental spinal cord injury. Neurobiol Dis. 2015; 82: 504–
[45] BareyreFM, KerschensteinerM, RaineteauO, MettenleiterTC, WeinmannO, SchwabME. The injured spinal cord spontaneously forms a new intraspinal circuit in
[46] PriestleyJV, Michael-TitusAT, TetzlaffW. Limiting spinal cord injury by pharmacolog‐
[48] UmhauJC, ZhouW, CarsonRE, RapoportSI, PolozovaA, DemarJ, HusseinN, Bhatta‐ charjeeAK, MaK, EspositoG, MajchrzakS, HerscovitchP, EckelmanWC, KurdzielKA, SalemNJr. Imaging incorporation of circulating docosahexaenoic acid into the human
[49] Van NeervenS, MeyJ. RAR/RXR and PPAR/RXR signaling in spinal cord injury. PPAR
[50] PuttaguntaR & DiGS. Retinoic acid signaling in axonal regeneration. Front Mol
[51] CurtA, van HedelHJ, KlausD, DietzV. Recovery from a spinal cord injury: signifi‐ cance of compensation, neural plasticity, and repair. J Neurotrauma. 2008; 25: 677–685.
[52] IchiyamaRM, CourtineG, GerasimenkoYP, YangGJ, van den BrandR, LavrovIA, ZhongH, RoyRR, EdgertonVR. Step training reinforces specific spinal locomotor
[53] FouadK & TetzlaffW. Rehabilitative training and plasticity following spinal cord injury.
[54] LiuG, DetloffMR, MillerKN, SantiL, HouleJD. Exercise modulates microRNAs that affect the PTEN/mTOR pathway in rats after spinal cord injury. Exp Neurol. 2012; 233:
[55] Garcia-AliasG, FawcettJW. Training and anti-CSPG combination therapy for spinal
[56] MaierIC, IchiyamaRM, CourtineG, SchnellL, LavrovI, EdgertonVR, SchwabME. Differential effects of anti-Nogo-A antibody treatment and treadmill training in rats
circuitry in adult spinal rats. J Neurosci. 2008; 28: 7370–7375.
with incomplete spinal cord injury. Brain. 2009; 132: 1426–1440.
[47] CalderPC. Mechanisms of action of (n-3) fatty acids. J Nutr. 2012; 142: 592S–599S.
brain using positron emission tomography. J Lipid Res. 2009; 50: 1259–1268.
adult rats. Nat Neurosci. 2004; 7: 269–277.
Res. 2007; 2007: 29275.
152 Recovery of Motor Function Following Spinal Cord Injury
Neurosci. 2011; 4: 59.
Exp Neurol. 2012; 235: 91–99.
cord injury. Exp Neurol. 2012; 235: 26–32.
447–456.
ical intervention. Handb Clin Neurol. 2012; 109: 463–484.
515.
JAK-STAT signalling is a main transduction pathway of cytokines and growth factors, which is involved in several biological processes including cell proliferation, cell differentiation, axon regeneration, apoptosis and inflammation. After spinal cord injury several cytokines activate the JAK-STAT pathway, thereby modulating several cell responses. In this chapter we discuss how regulation of this signalling pathway could improve motor recovery after injury by modulation of axon regeneration, neuropro‐ tection, glial scar formation, demyelination and inflammatory response. Studies with gene over-expression, gene deletion and *in vitro* approaches will be discussed for understanding the cell-specific response to JAK-STAT signalling, with a focus on preclinical treatment with IL6-family cytokines, hematopoietic cytokines and IL10.
**Keywords:** cytokine, JAK-STAT, STAT3, axon regeneration, glial scar, inflammation
### **1. Introduction**
Worldwide, an estimate of 180,000 cases of spinal cordinjuries (SCI) occur yearly [1]. SCIresults in the complete or partial loss of motor and sensory functions below the lesion site. This type of injury causes irreversible paralysis, chronic pain, loss of bladder, bowel and sexual function, amongst others dysfunctions, impairing quality and increasing the cost of life [2, 3].
The pathology of SCI in mammals starts with an acute phase during the first days of injury, which includes massive cell death and inflammatory response. The acute response is fol‐ lowed by a second phase during the first week after injury consisting of tissue replacement, where the loss of cells is replaced by a glial scar. After the second week it finalizes with a third
© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
phase which continues for months involving chronic tissue remodelling, remyelination and circuit remodelling [4]. Although some spontaneous repair after SCI has been described in mammals including humans, it contributes poorly to motor and functional recovery. Progenitor cells proliferate and differentiate but mostly to glial cells while no neurogenesis occurs [5, 6]. Besides there is only a partial axon regeneration and circuitremodelling response that contribute to a limited compensatory recovery [2].
Several reviews have been published describing growth factors and cytokines that regulate cell response in SCI [2, 7, 8]. Advancements using these factors to improve spinal cord repair are being made, and preclinical treatments have been achieved by local delivery of growth factors and by physiological delivery with i.v. or i.p. injections (intravenous or intraperito‐ neal, respectively) [2]. Between the plethora of signalling pathways activated in response to tissue injury, the JAK-STAT signalling pathway is one of the main pathways that has been extensively studied because of its broad effect on the response to injury. In this chapter, we discuss the activation and role of JAK-STAT signalling in response to SCI and studies focused on modulating JAK-STAT to improve motorrecovery. First, we briefly describe the JAK-STAT pathway components, followed by a discussion of its role in different cellular processes including: axonal regeneration, neuroprotection, glial response and its effects on the inflam‐ matory response.
### **2. JAK-STAT signalling pathway**
The JAK-STAT signalling pathway is involved in transmitting information from the extracel‐ lular milieu to gene promoters in the nucleus. The basic components of the JAK-STAT pathway are depicted in **Figure 1**. Several cytokines, growth factors and even hormones signal through the JAK-STAT pathway. Currently, there are 38 protein ligands and 36 cell surface receptor combinations have been described [9]. Besides the several combinations of protein ligands and receptor complexes, in mammals there are four JAK (Janus kinase) tyrosine kinases: JAK1, JAK2, JAK3 and Tyk2 (Non-receptor tyrosine-protein kinase TYK2); and seven STAT (Signal transducer and activator oftranscription) transcription factors: STAT1, STAT2, STAT3, STAT4, STAT5a, STAT5b and STAT6 [9].
Activation of the JAK-STAT pathway begins after ligand binding to the receptor subunits, which forms homodimers, heterodimers or heteromultimers depending on the family receptor (**Table 1** presents ligands and receptors related to SCI). After multimerization, intracellular transduction is initiated through the recruitment of JAK kinases. JAKs phosphor‐ ylate receptor subunits and STAT transcription factors. After tyrosine phosphorylation, cytosolic STATs dimerize and are translocated to the nucleus to bind specific DNA regulato‐ ry sequences and regulate gene expression [10].
The JAK-STAT signalling pathway has different negative inhibitors. The most important inhibitorfor preclinical studies is the classical negative feedback loop of suppressor of cytokine signalling (SOCS) proteins, which are target genes for STATs proteins and switch off JAK proteins. Over-expression or deletion of SOCS3, one of the eight mammalian SOCS proteins, has been extensively used to modulate endogenous pathway activation [11–13].
phase which continues for months involving chronic tissue remodelling, remyelination and circuit remodelling [4]. Although some spontaneous repair after SCI has been described in mammals including humans, it contributes poorly to motor and functional recovery. Progenitor cells proliferate and differentiate but mostly to glial cells while no neurogenesis occurs [5, 6]. Besides there is only a partial axon regeneration and circuitremodelling response
Several reviews have been published describing growth factors and cytokines that regulate cell response in SCI [2, 7, 8]. Advancements using these factors to improve spinal cord repair are being made, and preclinical treatments have been achieved by local delivery of growth factors and by physiological delivery with i.v. or i.p. injections (intravenous or intraperito‐ neal, respectively) [2]. Between the plethora of signalling pathways activated in response to tissue injury, the JAK-STAT signalling pathway is one of the main pathways that has been extensively studied because of its broad effect on the response to injury. In this chapter, we discuss the activation and role of JAK-STAT signalling in response to SCI and studies focused on modulating JAK-STAT to improve motorrecovery. First, we briefly describe the JAK-STAT pathway components, followed by a discussion of its role in different cellular processes including: axonal regeneration, neuroprotection, glial response and its effects on the inflam‐
The JAK-STAT signalling pathway is involved in transmitting information from the extracel‐ lular milieu to gene promoters in the nucleus. The basic components of the JAK-STAT pathway are depicted in **Figure 1**. Several cytokines, growth factors and even hormones signal through the JAK-STAT pathway. Currently, there are 38 protein ligands and 36 cell surface receptor combinations have been described [9]. Besides the several combinations of protein ligands and receptor complexes, in mammals there are four JAK (Janus kinase) tyrosine kinases: JAK1, JAK2, JAK3 and Tyk2 (Non-receptor tyrosine-protein kinase TYK2); and seven STAT (Signal transducer and activator oftranscription) transcription factors: STAT1, STAT2, STAT3, STAT4,
Activation of the JAK-STAT pathway begins after ligand binding to the receptor subunits, which forms homodimers, heterodimers or heteromultimers depending on the family receptor (**Table 1** presents ligands and receptors related to SCI). After multimerization, intracellular transduction is initiated through the recruitment of JAK kinases. JAKs phosphor‐ ylate receptor subunits and STAT transcription factors. After tyrosine phosphorylation, cytosolic STATs dimerize and are translocated to the nucleus to bind specific DNA regulato‐
The JAK-STAT signalling pathway has different negative inhibitors. The most important inhibitorfor preclinical studies is the classical negative feedback loop of suppressor of cytokine signalling (SOCS) proteins, which are target genes for STATs proteins and switch off JAK
that contribute to a limited compensatory recovery [2].
154 Recovery of Motor Function Following Spinal Cord Injury
matory response.
**2. JAK-STAT signalling pathway**
STAT5a, STAT5b and STAT6 [9].
ry sequences and regulate gene expression [10].
Although it will not be discussed in this chapter, it has to be considered that ligands binding to receptor complexes also activate other intracellular signalling cascades besides the JAK-STAT pathway [10]. Phosphorylation of receptors induces activation of ERK1/2 and AKT pathways. Moreover, STATs can also be activated independently to the canonical JAK-STAT signalling pathway. Growth factor activation of RTKs (receptor tyrosine kinases) and NRTKs (non-receptor tyrosine kinases) can activate STATs. Hormone and chemokine binding to G protein–coupled receptors can also activate JAKs proteins to phosphorylate STATs [10].
**Figure 1.** Basic components of the JAK-STAT signaling pathway. After binding of cytokines, receptors multimerize (violet boxes) recruiting to the membrane and activating JAK kinases (orange boxes) that initiate substrate phosphory‐ lation (letter P). STAT transcription factors (pink boxes) are phosphorylated, dimerized and transported to the nucleus. STAT dimers regulate gene transcription. Among others, a classic target is SOCS that forms a negative feedback loop inhibiting JAK function. Alternative signaling pathways (grey boxes) for JAK-STAT are JAK-receptor complex activa‐ tion of ERK1/2 and AKT pathways; STATs phosphorylation by RTKs, NRTKs or JAKs associated to G-coupled protein receptors.
### **3. JAK-STAT pathway activation in response to SCI**
### **3.1. Cytokine expression in response to SCI**
Before discussing the role of the JAK-STAT pathway in motor recovery, we will discuss the endogenous expression of cytokines and activation of STAT proteins in response to SCI. The upregulation of the IL6 family cytokines, IL6 (Interleukin 6), LIF (Leukemia inhibitory factor), OSM (Oncostatin M), IL11 (Interleukin 11) and CNTF (Ciliary neurotrophic factor) have been well characterized in response to SCI. Early IL6 and LIF upregulation has been detected in the acute phase after SCI, with a peak of expression around 6–12 hpi (hours post-injury) and basal levels at 4 dpi (days post-injury) [14, 15]. Different neural cell types contribute to IL6 and LIF expression. It has been shown that IL6 has a broad cell type expression, being detected in neurons, astrocytes and macrophages, while LIF is mainly expressed in meningeal astro‐ cytes [14]. IL6 mRNA expression correlates with IL6 protein levels analyzed in another study, which detected an increase in IL6 concentration from 3 hours to 3 dpi [16]. In these studies the upregulation of cytokines occurs in the surrounding area of the lesion site, around 1 or 2 mm from the epicentre.
Early upregulation of OSM is also detected after SCI, with a strong peak at 6 hpi, but on the contrary to IL6 and LIF, upregulation of OSM is still detected until 1 mpi (months postinjury) [15]. IL11 and CNTF protein upregulation has also been detected after the acute phase. A steady increase in protein levels has been detected during the first week of injury for IL11 and during the first month for CNTF. In both cases no later times were analyzed to deter‐ mine when the cytokines reached basal levels, therefore there is a possibility that they had an extended upregulation [17, 18]. IL11 cell expression has not been analyzed; on the contrary, it has been shown that CNTF has a chronic expression in glial cell types after spinal cord injury. CNTF has been detected in oligodendrocytes during the first month and in astrocytes up to 4 mpi [18, 19]. This chronic expression of CNTF is more spatially restricted than the early expressed cytokines IL6 and LIF, being detected only in the lesion borders [19].
**Table 1.** JAK-STAT ligands, receptors and transducers involved in spinal cord injury. Cytokines are shown with associated receptors and JAK-STAT components. It is shown if cytokines are up-regulated in response to SCI (U) or if they have been used in preclinical (P) or clinical (C) studies for spinal cord recovery.
### **3.2. JAK-STAT signalling in response to SCI**
**3. JAK-STAT pathway activation in response to SCI**
Before discussing the role of the JAK-STAT pathway in motor recovery, we will discuss the endogenous expression of cytokines and activation of STAT proteins in response to SCI. The upregulation of the IL6 family cytokines, IL6 (Interleukin 6), LIF (Leukemia inhibitory factor), OSM (Oncostatin M), IL11 (Interleukin 11) and CNTF (Ciliary neurotrophic factor) have been well characterized in response to SCI. Early IL6 and LIF upregulation has been detected in the acute phase after SCI, with a peak of expression around 6–12 hpi (hours post-injury) and basal levels at 4 dpi (days post-injury) [14, 15]. Different neural cell types contribute to IL6 and LIF expression. It has been shown that IL6 has a broad cell type expression, being detected in neurons, astrocytes and macrophages, while LIF is mainly expressed in meningeal astro‐ cytes [14]. IL6 mRNA expression correlates with IL6 protein levels analyzed in another study, which detected an increase in IL6 concentration from 3 hours to 3 dpi [16]. In these studies the upregulation of cytokines occurs in the surrounding area of the lesion site, around 1 or 2 mm
Early upregulation of OSM is also detected after SCI, with a strong peak at 6 hpi, but on the contrary to IL6 and LIF, upregulation of OSM is still detected until 1 mpi (months postinjury) [15]. IL11 and CNTF protein upregulation has also been detected after the acute phase. A steady increase in protein levels has been detected during the first week of injury for IL11 and during the first month for CNTF. In both cases no later times were analyzed to deter‐ mine when the cytokines reached basal levels, therefore there is a possibility that they had an extended upregulation [17, 18]. IL11 cell expression has not been analyzed; on the contrary, it has been shown that CNTF has a chronic expression in glial cell types after spinal cord injury. CNTF has been detected in oligodendrocytes during the first month and in astrocytes up to 4 mpi [18, 19]. This chronic expression of CNTF is more spatially restricted than the early
**Cytokine Receptors JAKs STATs Endogenous levels,**
Jak1 STAT1-3 U/P
Jak2 STAT3 U/P
Tyk2 STAT3 U/P
STAT3 U
**preclinical or clinical trials**
expressed cytokines IL6 and LIF, being detected only in the lesion borders [19].
**3.1. Cytokine expression in response to SCI**
156 Recovery of Motor Function Following Spinal Cord Injury
from the epicentre.
**Cytokine family**
IL6 IL6 Gp130 –
IL6R
LIFR
IL11R
LIFR – CNTFR
OSM Gp130 – STAT3 U/P
LIF Gp130 –
CNTF Gp130 –
IL11 Gp130 –
In concordance with the upregulation of cytokines in response to SCI, several studies have characterized JAK-STAT pathway activation in spinal cord cells. Consistent with the transi‐ ent increase of IL6 concentration, in the same study it was detected an increase in gp130 dimerization and JAK1 phosphorylation [16]. As expected forthe activation of the gp130/JAK1 axis, STAT1 and STAT3 are phosphorylated in response to SCI. pSTAT1 has an acute in‐ crease which reach basal levels at 2 dpi [20]. On the contrary, pSTAT3 has an extended increase which differs between studies. An increase in pSTAT3 has been detected up to 7 dpi in some studies [12, 16], while in another no basal levels were reached even after 2 weeks post-injury (wpi) [21]. Nevertheless, all studies agree that STAT3 has a longer activation than STAT1. Together with the temporal difference, protein localization suggests that STAT3 has a more prominent role in regulating gene expression than STAT1 in cell response. While pSTAT3 has been detected in neuron nuclei after spinal cord injury, pSTAT1 has been only detected in the cytoplasm [16, 20].
STAT3 activation has been characterized in several spinal cord cell types. In the acute injury, nuclear pSTAT3 is observed in neurons of the anterior horns [16] and transcriptional activity is also supported by the detection of SOCS3+ neurons [21]. Nuclear pSTAT3 and SOCS3 has been also detected in microglia/macrophage during the acute phase of injury [16, 21] and nuclear pSTAT3 in oligodendrocytes and astrocytes during the first week near the lesion site [12, 22]. Due to a prominent role in glial scar formation, spatial STAT3 activation on astro‐ cytes has also been highly defined. It has been determined that during the first week pSTAT3+ astrocytes appear at the border of the lesion with elevated predominance on the first 500 μm near the injury [12, 22].
After the first week of injury, contusion models show chronic pSTAT3 signalling in glial cell types located in the lesion borders. pSTAT3 has been detected in oligodendrocyte precursor cells (OPCs) and oligodendrocytes at 2 wpi [18, 19]. This activation decreases from 1 to 5 wpi, but at that time is still higher than on uninjured spinal cords [19]. pSTAT3+ astrocytes have been detected even further, at 12 wpi in the lesion borders [19].
### **4. Role of JAK-STAT in axon regeneration and collateral sprouting**
To recover motorfunction after SCI, new connections have to be established after neuron death and axon degeneration [2]. These circuits can be established by two different cellular respons‐ es that should not be confused: Axon regeneration is the process where axons from severed neurons regrow from the injured tip or from a lateral growth distant to the injury (**Figure 2ab**). On the contrary, collateral sprouting is a compensatory growth from undamaged axons initiated in response to injury (**Figure 2c**) [23].
While there is limited axon regeneration through the lesion area in response to complete SCI (**Figure 2a**), there is a compensatory remodelling on brain and spinal circuits when the spinal
**Figure 2.** Axonal regeneration and collateral sprouting in spinal cord injury. Models used to evaluate axon regenera‐ tion and circuit remodeling, showing corticospinal tract (CST, in blue) as an example. Remodeling can be evaluated by axonal regeneration or collateral sprouting (both in red arrows) and new innervations can be connected to propiospi‐ nal neurons (PSNs, interneurons in orange) or motor neurons (in green). CNS injury is indicated as a darkened triangle in every model. **A**, In SCI models (usually contusion and hemisection) axonal regeneration of injured CSTs can be eval‐ uated through and beyond the lesion area. **B**, in a hemisection model, innervation of new targets by injured CSTs can be evaluated. Regenerative axons from CSTs can innervate PSNs, which then connect to denervated motoneurons. **C**, in a unilateral pyramidotomy injury paradigm, collateral sprouting of uninjured CSTs to denervated PSNs and motor neurons can be measured. Anterograde, retrograde or retrograde trans-synaptic labelling are used depending of the injury model.
cord lesion is incomplete [2]. The corticospinal tract (CST) has been used as a model to study this recovery process. CST starts in the motor cortex and connects to spinal motor neurons, controlling voluntary motor function [23]. After spinal cord injury, CST circuit remodelling can be achieved by axon regeneration innervating long descendent interneurons that in‐ crease connection to motoneurons (**Figure 2b**) [24] or by collateral sprouting of uninjuredaxons to denervated motoneurons (**Figure 2c**) [25]. In this section we will discuss the role of JAK-STAT pathway on the promotion of axon regeneration and collateral sprouting after SCI. Briefly, we will discuss the results involving dorsalroot ganglion (DRG) and optic nerve injury, to follow with studies of axon growth and collateral sprouting of the CSTs after SCI.
### **4.1. Role of JAK-STAT pathway in axon regeneration in SCI and other nerve injury models**
### *4.1.1. Axon regeneration after dorsal root ganglion or optic nerve axotomy*
but at that time is still higher than on uninjured spinal cords [19]. pSTAT3+ astrocytes have
To recover motorfunction after SCI, new connections have to be established after neuron death and axon degeneration [2]. These circuits can be established by two different cellular respons‐ es that should not be confused: Axon regeneration is the process where axons from severed neurons regrow from the injured tip or from a lateral growth distant to the injury (**Figure 2ab**). On the contrary, collateral sprouting is a compensatory growth from undamaged axons
While there is limited axon regeneration through the lesion area in response to complete SCI (**Figure 2a**), there is a compensatory remodelling on brain and spinal circuits when the spinal
**Figure 2.** Axonal regeneration and collateral sprouting in spinal cord injury. Models used to evaluate axon regenera‐ tion and circuit remodeling, showing corticospinal tract (CST, in blue) as an example. Remodeling can be evaluated by axonal regeneration or collateral sprouting (both in red arrows) and new innervations can be connected to propiospi‐ nal neurons (PSNs, interneurons in orange) or motor neurons (in green). CNS injury is indicated as a darkened triangle in every model. **A**, In SCI models (usually contusion and hemisection) axonal regeneration of injured CSTs can be eval‐ uated through and beyond the lesion area. **B**, in a hemisection model, innervation of new targets by injured CSTs can be evaluated. Regenerative axons from CSTs can innervate PSNs, which then connect to denervated motoneurons. **C**, in a unilateral pyramidotomy injury paradigm, collateral sprouting of uninjured CSTs to denervated PSNs and motor neurons can be measured. Anterograde, retrograde or retrograde trans-synaptic labelling are used depending of the
**4. Role of JAK-STAT in axon regeneration and collateral sprouting**
been detected even further, at 12 wpi in the lesion borders [19].
initiated in response to injury (**Figure 2c**) [23].
158 Recovery of Motor Function Following Spinal Cord Injury
injury model.
The DRG model has been useful to study axon regeneration due to regenerative and nonregenerative branches. Sensory neurons from DRG target peripheral tissue with branches (i.e. sciatic nerve) that regenerate after axotomy. On the contrary, DRG neurons also target the dorsal column of the spinal cord, which do not regenerate after injury. This difference in the regenerative capacity has been partially relatedto the JAK-STATpathway activation by studies on cytokine deficient mice, but conclusive results have been obtained in a study assessing the role of STAT3 in axon regeneration. While STAT3 deletion reduced axon growth of the severed sciatic nerve, STAT3 viral over-expression in the DRG improves axon regeneration of the nonregenerative nerve [26].
Similar to the activation of the JAK-STAT pathway in the DRG model, axon growth has been also promoted in the non-regenerative optic nerve. JAK-STAT activation by CNTF treatment or deletion of the SOCS3 negative feedback improved axon regeneration in optic nerve injury [11]. The studies in both injury models demonstrate that the JAK-STAT pathway, specifically the axis of STAT3/SOCS3, promotes axon regeneration. Therefore, a possible mechanism to improve motor recovery after SCI is through modulation of the JAK-STAT pathway to induce axon regeneration.
### *4.1.2. Axon regeneration after the spinal cord injury*
A hemisection study showed that only a small subset of cortical neurons presented an increment in STAT3 and pSTAT3 levels after SCI, suggesting that on the contrary of periph‐ eral nerves, in the spinal cord there is a lack ofretrograde JAK-STAT activation [27]. As in DRG and optic nerve injury models, insufficient JAK-STAT activation could be related to the lack of axon regeneration. Several studies have shown that improvement of local extrinsic factors, such as cytokines administration, or improvement of intrinsic factors, such as the genetic modulation of the JAK-STAT pathway in neurons, improves axon regeneration after SCI, suggesting that modulation of JAK-STAT can be a bona fide target for designing novel therapies for spinal cord repair.
Administration of CNTF, G-CSF (granulocyte-colony stimulating factor), IL6 and LIF by intrathecal or i.p. injections results in improved axon regeneration through the lesion site and motorrecovery after SCI. CNTF intrathecal injections during the first 10 days or administration in sodium-hyaluronate particles placed in the transection site have improved functional recovery after SCI [28, 29]. Although CST axon regeneration has not been assessed in CNTF treatment, a study showed increased retrograde labelling of the rubrospinal tracts [28]. Another study showed improved axonal sprouting through the lesion site, although it was not evaluated if these axons were from local neurons or descended from the brain [29].
Axon regeneration has also been assessed after spinal cord administration of G-CSF during 2 weeks after a hemisection [30]. CST axon regeneration was improved in the lesion site although no axon growth caudal to the injury was detected. The local G-CSF delivery also results in improvement of motor function and a similar result was observed in a transgenic mice line expressing G-CSF under the control of the MapKII promoter that is specifically activated in cortical and spinal cord neurons [30].
On the contrary to CNTF and G-CSF treatments, LIF has been administered daily by intraper‐ itoneal injection since it is actively transported through the BBB with a mechanism involving the LIFRα receptor [31, 32]. LIF treatment for 14 days in a hemisection injury model im‐ proved the number of retrograde labelled CST and rubrospinal neurons. Improvement in motor recovery was assessed by RotaRod and Platform hang tests [32].
Although CNTF, G-CSF and LIF treatments showed an improvement in axon regeneration and motor function in rodents, in these studies the activation of the JAK-STAT pathway in cortical neurons was not assessed. Besides, these studies cannot distinguish if the improve‐ ment in axon regeneration and functional recovery is due to the modulation of the intrinsic program of cortical neurons or due to the modulation of other spinal cord cell types. Contrary to these studies, local IL6 administration has been shown to activate the JAK-STAT pathway in spinal and cortical neurons [33, 34]. IL6 intrathecal administration during the first week after a hemisection injury increased pSTAT3 levels in the spinal cord and in cortical neurons, while up-regulated the regeneration associated gene GAP-43 mRNA [34]. Although no motor recovery analyses were assessed, anterograde labelling showed an increase in the number of synapsin1+ CST axons near the lesion site [33].
The IL6 treatment suggests that cytokine delivery can promote the regenerative response of motoneurons by JAK-STAT pathway activation. In addition, specific STAT3 over-expression in the cortex has shown that this pathway can promote axon regeneration. After overexpression mediated by adenovirus injection in the cortex, pSTAT3 levels increased in cortical neurons and CST axonal regeneration improved through and beyond the lesion site of a hemisection [27].
### **4.2. Role of JAK-STAT in circuit remodelling by collateral sprouting**
Although improvement in axon regeneration (**Figure 2a**) and motor recovery has been achieved by cytokine delivery, no relation has been established between the observed axon regeneration in the lesion site and the motorrecovery. On the contrary, studies involving JAK-
STAT pathway and circuitremodelling by collateral sprouting (**Figure 2c**) have suggested that motor recovery is accomplished by innervation of uninjured axons to denervated motoneur‐ ons and interneurons.
Administration of CNTF, G-CSF (granulocyte-colony stimulating factor), IL6 and LIF by intrathecal or i.p. injections results in improved axon regeneration through the lesion site and motorrecovery after SCI. CNTF intrathecal injections during the first 10 days or administration in sodium-hyaluronate particles placed in the transection site have improved functional recovery after SCI [28, 29]. Although CST axon regeneration has not been assessed in CNTF treatment, a study showed increased retrograde labelling of the rubrospinal tracts [28]. Another study showed improved axonal sprouting through the lesion site, although it was not
Axon regeneration has also been assessed after spinal cord administration of G-CSF during 2 weeks after a hemisection [30]. CST axon regeneration was improved in the lesion site although no axon growth caudal to the injury was detected. The local G-CSF delivery also results in improvement of motor function and a similar result was observed in a transgenic mice line expressing G-CSF under the control of the MapKII promoter that is specifically activated in
On the contrary to CNTF and G-CSF treatments, LIF has been administered daily by intraper‐ itoneal injection since it is actively transported through the BBB with a mechanism involving the LIFRα receptor [31, 32]. LIF treatment for 14 days in a hemisection injury model im‐ proved the number of retrograde labelled CST and rubrospinal neurons. Improvement in
Although CNTF, G-CSF and LIF treatments showed an improvement in axon regeneration and motor function in rodents, in these studies the activation of the JAK-STAT pathway in cortical neurons was not assessed. Besides, these studies cannot distinguish if the improve‐ ment in axon regeneration and functional recovery is due to the modulation of the intrinsic program of cortical neurons or due to the modulation of other spinal cord cell types. Contrary to these studies, local IL6 administration has been shown to activate the JAK-STAT pathway in spinal and cortical neurons [33, 34]. IL6 intrathecal administration during the first week after a hemisection injury increased pSTAT3 levels in the spinal cord and in cortical neurons, while up-regulated the regeneration associated gene GAP-43 mRNA [34]. Although no motor recovery analyses were assessed, anterograde labelling showed an increase in the number of
The IL6 treatment suggests that cytokine delivery can promote the regenerative response of motoneurons by JAK-STAT pathway activation. In addition, specific STAT3 over-expression in the cortex has shown that this pathway can promote axon regeneration. After overexpression mediated by adenovirus injection in the cortex, pSTAT3 levels increased in cortical neurons and CST axonal regeneration improved through and beyond the lesion site of a
Although improvement in axon regeneration (**Figure 2a**) and motor recovery has been achieved by cytokine delivery, no relation has been established between the observed axon regeneration in the lesion site and the motorrecovery. On the contrary, studies involving JAK-
motor recovery was assessed by RotaRod and Platform hang tests [32].
**4.2. Role of JAK-STAT in circuit remodelling by collateral sprouting**
evaluated if these axons were from local neurons or descended from the brain [29].
cortical and spinal cord neurons [30].
160 Recovery of Motor Function Following Spinal Cord Injury
synapsin1+ CST axons near the lesion site [33].
hemisection [27].
**Table 2.** Pre-clinical studies with cytokine for motor recovery of SCI. Summary of JAK-STAT cytokine preclinical studies, for any cytokine it is shown administration timing and method. Time indicated as acute means a single dose at the moment of injury or daily doses from 1 to 5 dpi. Methods of administration are indicated as intrathecal (i.t.), intraperitoneal (i.p.), intravenous (i.v.) or subcutaneous injection; embedded on hyaluronate or gel foam placed in the lesion site; and by SC (spinal cord) viral expression. Contrary to the other studies, MR16-1 indicates a treatment with a neutralizing antibody against IL6 receptor.
Promotion of collateral sprouting by the hematopoietic cytokine EPO (Erythropoietin) has been assessed in a model of traumatic brain injury thatresults in denervation of the CST circuit. EPO i.p. injections during the acute phase resulted in increased collateral sprouting of uninjured CST fibers in the cervical and lumbar area and improved motor function [35]. Although it has not been shown that EPO activate the JAK-STAT signalling pathway in neurons, cortical viral infection has been useful to determine the role of the pathway in collateral sprouting. STAT3 viral over-expression in uninjured mice induced collateral
sprouting and innervations to propiospinal neuron (PSN, a type of interneuron) [27]. This viral expression was also assessed in a unilateral pyramidotomy model, which is an injury at the level of the medulla oblongata that severs half of the CST (**Figure 2c**) and it is useful to evaluate how collateral sprouting connects to spinal neurons. In this approach it was shown that STAT3 induced collateral sprouting, innervation of PSNs and motoneurons, and improvement in motorrecovery which was assessed by behavioural and electrophysiologicaltests [27]. Cortical SOCS3 deletion also induced collateral sprouting in the pyramidotomy paradigm [13], suggesting that an endogenous JAK-STAT pathway activation promotes circuit remodelling. After finding a transient CNTF expression in denervated neurons caudal to injury, the same study showed that combination of cortical SOCS3 deletion and spinal cord CNTF overexpression improved collateral sprouting [13].
### **5. Role of JAK-STAT in local neuron response**
In response to SCI the number of local motor and interneurons decrease with no neurogene‐ sis to generate new neurons [2]. To improve motor recovery, promotion of axon regenera‐ tion and collateral sprouting should be accompanied by neuroprotective strategies, since number and dendrite distribution of local spinal neurons would be beneficial for circuit remodelling. We will start this section discussing *in vitro* studies related to the neuroprotec‐ tive role of the JAK-STAT pathway, following *in vivo* studies with cytokine delivery and knockdown treatments after SCI.
### **5.1. Role of JAK-STAT in neuron survival**
At a molecular level, studies in non-neural cells have shown that STAT transcription factors have differentroles in apoptosis regulation. A comparative *in vitro* study of STAT1 and STAT3 showed that while STAT1 inhibits the expression of the anti-apoptotic genes Bcl-2 and Bcl-X, STAT3 promoted their expression [36]. STAT5 proteins, which are activated by a different set of cytokines (Table 1), also promote cell survival by Bcl-X upregulation [37].
*In vitro* studies have shown that the neuroprotective role of JAK-STAT cytokines is partly associated to STAT3 activation and expression of anti-apoptotic genes. IL10 activated STAT3 and promoted Bcl-2 and Bcl-xl expression in embryonic spinal cord neurons [38]. IL6 and OSM cytokines also activated STAT3 and promoted anti-apoptotic gene expression in cortical neurons and neuroblastoma cells [39]. In cultured cerebellar granule neurons, G-CSF also promoted STAT3 activation while increased Bcl-2 levels [40]. It should be considered that theses cytokines also activate the AKT pathway [39, 40]. Both inhibition of the JAK and AKT pathway decreased the neuroprotective effect, suggesting that probably both pathways contribute to the anti-apoptotic response [38, 40].
Related to the previous *in vitro* studies, *in vivo* studies in CNS injury suggest a protective role of STAT3 while an opposite role for STAT1. On a focal cerebral ischemia model, it was shown that STAT1 deficient mice presented a reduced infarct size and apoptotic cell number [41]. On the contrary, on a permanent cerebral ischemia model the IL6/STAT3 axis has been associat‐
ed to positive outcomes. While the neutralization with an anti-IL6R antibody resulted in decreased STAT3 activation and increased lesion area, IL6 administration reduced the lesion size [42, 43]. Although these studies indicate a neuroprotective role of STAT3, specific cell type modulation should be assessed to differentiate the direct modulation on neuron response from indirect neuroprotection by modulation of glial and inflammatory cell types. At this mo‐ ment, there is only one study where specific STAT3 deletion in neurons is analyzed in an *in vivo* CNS injury model. Mice with STAT3 deletion in neurons (using the Neurofilament L promoter) showed a decrease in motoneuron number after facial nerve injury and diminish‐ ed expression of anti-apoptotic genes [44].
### **5.2. JAK-STAT modulation of neuronal protection in SCI**
sprouting and innervations to propiospinal neuron (PSN, a type of interneuron) [27]. This viral expression was also assessed in a unilateral pyramidotomy model, which is an injury at the level of the medulla oblongata that severs half of the CST (**Figure 2c**) and it is useful to evaluate how collateral sprouting connects to spinal neurons. In this approach it was shown that STAT3 induced collateral sprouting, innervation of PSNs and motoneurons, and improvement in motorrecovery which was assessed by behavioural and electrophysiologicaltests [27]. Cortical SOCS3 deletion also induced collateral sprouting in the pyramidotomy paradigm [13], suggesting that an endogenous JAK-STAT pathway activation promotes circuit remodelling. After finding a transient CNTF expression in denervated neurons caudal to injury, the same study showed that combination of cortical SOCS3 deletion and spinal cord CNTF over-
In response to SCI the number of local motor and interneurons decrease with no neurogene‐ sis to generate new neurons [2]. To improve motor recovery, promotion of axon regenera‐ tion and collateral sprouting should be accompanied by neuroprotective strategies, since number and dendrite distribution of local spinal neurons would be beneficial for circuit remodelling. We will start this section discussing *in vitro* studies related to the neuroprotec‐ tive role of the JAK-STAT pathway, following *in vivo* studies with cytokine delivery and
At a molecular level, studies in non-neural cells have shown that STAT transcription factors have differentroles in apoptosis regulation. A comparative *in vitro* study of STAT1 and STAT3 showed that while STAT1 inhibits the expression of the anti-apoptotic genes Bcl-2 and Bcl-X, STAT3 promoted their expression [36]. STAT5 proteins, which are activated by a different set
*In vitro* studies have shown that the neuroprotective role of JAK-STAT cytokines is partly associated to STAT3 activation and expression of anti-apoptotic genes. IL10 activated STAT3 and promoted Bcl-2 and Bcl-xl expression in embryonic spinal cord neurons [38]. IL6 and OSM cytokines also activated STAT3 and promoted anti-apoptotic gene expression in cortical neurons and neuroblastoma cells [39]. In cultured cerebellar granule neurons, G-CSF also promoted STAT3 activation while increased Bcl-2 levels [40]. It should be considered that theses cytokines also activate the AKT pathway [39, 40]. Both inhibition of the JAK and AKT pathway decreased the neuroprotective effect, suggesting that probably both pathways
Related to the previous *in vitro* studies, *in vivo* studies in CNS injury suggest a protective role of STAT3 while an opposite role for STAT1. On a focal cerebral ischemia model, it was shown that STAT1 deficient mice presented a reduced infarct size and apoptotic cell number [41]. On the contrary, on a permanent cerebral ischemia model the IL6/STAT3 axis has been associat‐
of cytokines (Table 1), also promote cell survival by Bcl-X upregulation [37].
expression improved collateral sprouting [13].
162 Recovery of Motor Function Following Spinal Cord Injury
knockdown treatments after SCI.
**5.1. Role of JAK-STAT in neuron survival**
contribute to the anti-apoptotic response [38, 40].
**5. Role of JAK-STAT in local neuron response**
G-CSF and IL10 are cytokines with several studies showing beneficial outcomes in neuropro‐ tection, inflammatory response and motor recovery after SCI [45, 46]. It has been shown that a pro-inflammatory response is negative for neuron survival [47]; therefore, in these cyto‐ kine treatments it is difficult to differentiate a direct modulation on neuronal survival and the effect on neuro-inflammation. Since it has been shown that spinal cord neurons express G-CSF and IL10 receptors [40, 48], we will discuss a small set of studies that focused in neuroprotec‐ tion and in a following section we will discuss the modulation of the immune response in SCI.
G-CSF has been locally administrated to evaluate neuroprotection in SCI. Subcutaneous G-CSF treatment during the first five days after compression improved motor recovery, decreased apoptotic neurons in the acute phase and improved neuron number in the remod‐ elling phase (6 wpi) [40]. Other studies with intrathecal administration of G-CSF for 2 weeks, commented before for assessing axonal regeneration, also showed neuroprotective results. This long-term treatment reduced apoptotic cell number and increased Bcl-xL expression in the spinal cord [30]. Two independent IL10 viral over-expression studies also improved motor function [48, 49] and one of these studies showed decreased pro-apoptotic protein levels and increased Bcl-2 levels, resulting in a higher neuron number [48]. Although it was shown by *in vitro* studies that G-CSF and IL10 treatments activate STAT3 in spinal neurons [40, 48], it has not been demonstrated that this activation occurs on *in vivo* treatments. *In vivo* analyses of IL10 over-expression only focused in neuronal NF-κβ activation [48]. Therefore, it is an open question if these cytokines promotes neuron survival by JAK-STAT activation or by other signalling pathway, as NF-κβ gene regulation.
Although by results of cytokine treatments the role of STAT3 in neuroprotection remains inconclusive, knockdown of the inhibitory protein SOCS3 has contributed to this proposi‐ tion [21]. Local over-expression of shSOCS3 diminished SOCS3 mRNA levels. This knock‐ down increased pSTAT3 levels in a transection model and it decreased the Bax/Bcl-2 protein ratio, while improving neuron number near the lesion site [21]. A following study of the same group showed that SOCS3 knockdown increased dendrite growth in dorsal and ventral horns near the injury site [50].
Finally, consistent with the pro-apoptotic role of STAT1 on *in vitro* studies [36], STAT1 knockdown has shown positive outcomes after SCI. Administration of STAT1 siRNA reduced STAT1 protein levels in the spinal cord. Although no analysis of neuronal survival was assessed, 1 day post-contusion the STAT1 knockdown mice had increased its Bcl-2 levels. siRNA treated mice also improved motor function assessed by BMS open-field score.
### **6. Role of JAK-STAT in glial response**
In response to SCI, apoptosis of astrocytes and oligodendrocytes occurs in the acute phase of injury and then glial cells dynamically respond during weeks and months after injury. A glial hallmark of CNS injury is the glial scar, formed by reactive astrocytes, OPCs and meningeal cells [51]. This compact scar gathers around damaged tissue, inflammatory cells and fibro‐ blasts. The glial scar is necessary to contain secondary injury, as disruption of the astrocyte scar with different transgenic models leads to increased cell death, demyelination and reduced functional motor recovery [12, 52, 53]. Although the glial scar is necessary for the contain‐ ment of the lesion, reactive astrocytes express and secrete inhibitory molecules for axonal growth and sprouting, as chondroitin sulfate proteoglycans (CSPGs) [51]. Altogether with the glial scar formation, a demyelination process occurs during weeks after injury by oligoden‐ drocyte apoptosis. Although mature oligodendrocytes do not proliferate to recover cell number, OPCs start to proliferate and differentiate and remyelination proceeds nearthe lesion area [6]. Accordingly to the glial response commented above, in this section we will discuss the following topics: Cytokine modulation of the glial response to reduce secondary damage, the regulation ofreactive astrocyte and the neural stem progenitor cells (NSPCs) by the STAT3/ SOCS3 axis.
### **6.1. Cytokines modulate glial response in SCI**
Several studies with the hematopoietic and IL6-family cytokines have assessed astrocyte and oligodendrocyte responses with positive outcomes on tissue preservation and motor recov‐ ery. Acute administration of GM-CSF or G-CSF by i.p. or i.v. injections during the acute phase reduced lesion area, while increasing the spared myelin area and improving motor recovery after SCI [54–56]. Both cytokine treatments reduced NG2 levels (OPC marker) during the first week of injury [54, 55]. In one of these studies it was also shown that G-CSF up-regulated BclxL and reduced apoptosis in oligodendrocytes [56]. Therefore, these results suggest that these cytokines maintain spared myelin by a protective mechanism and not by promotion of OPC proliferation.
GM-CSF and G-CSF also modulate the reactive astrocyte response during the first month after SCI, as seen by the reduction of GFAP levels and the CSPG neurocan [54, 55]. Another study with improvement in motorrecovery by G-CSF intrathecal administration during the first day of injury also showed a reduction in the CSPG proteins neurocan and phosphocan [57]. Altogether with the previous studies, this last study suggest that G-CSF decrease the reac‐ tive gliosis because the treatmentreduced TGF-β levels, a growth factorthat promotes reactive gliosis; reduced vimentin levels, an astrocyte filament induced in glial scar; and presented less astrocyte proliferation [57]. Acute EPO i.p. injection also has similar results in astrocyte regulation after SCI. EPO treatment reduced lesion area and GFAP levels, while improving preservation of myelin and motor recovery at 2 wpi [58]. A following study showed possible differences with GM-CSF and G-CSF on the OPC response. Although these cytokines reduce NG2 levels, it was shown that EPO treatment increased it at 4 wpi [59].
The IL6-family of cytokines also regulates glial response. LIF i.p. administration has been shown to reduce oligodendrocyte apoptosis [60, 61], although *in vivo* LIF signalling on oligodendrocyte is not clear. On one hand, one study found that SCI induced LIFRβ expres‐ sion on oligodendrocytes and that LIF treatment induced pSTAT3 and pAKT in these cells [60]. On the contrary, an alternative study did not find LIFRβ expression on oligodendrocytes and suggested a LIF positive modulation of microglia [61]. Although studies with LIF i.p. treat‐ ment have not assessed the astrocyte response in SCI, it has been shown that this treatment incremented Nestin+ cell number near the lesion site [32]. CNTF treatment, previously commented for improved motor recovery and axon regeneration, also increased the density of astrocytes [28]. Finally, gel foam administration of OSM reduced the lesion area and preserved MBP (myelin basic protein), but on the contrary to the other IL6-family cytokines, it reduced GFAP levels near the lesion site. This glial modulation was accompanied by improved serotonergic fiber outgrowth and motor recovery [15].
### **6.2. STAT modulation of reactive astrocytes**
assessed, 1 day post-contusion the STAT1 knockdown mice had increased its Bcl-2 levels.
In response to SCI, apoptosis of astrocytes and oligodendrocytes occurs in the acute phase of injury and then glial cells dynamically respond during weeks and months after injury. A glial hallmark of CNS injury is the glial scar, formed by reactive astrocytes, OPCs and meningeal cells [51]. This compact scar gathers around damaged tissue, inflammatory cells and fibro‐ blasts. The glial scar is necessary to contain secondary injury, as disruption of the astrocyte scar with different transgenic models leads to increased cell death, demyelination and reduced functional motor recovery [12, 52, 53]. Although the glial scar is necessary for the contain‐ ment of the lesion, reactive astrocytes express and secrete inhibitory molecules for axonal growth and sprouting, as chondroitin sulfate proteoglycans (CSPGs) [51]. Altogether with the glial scar formation, a demyelination process occurs during weeks after injury by oligoden‐ drocyte apoptosis. Although mature oligodendrocytes do not proliferate to recover cell number, OPCs start to proliferate and differentiate and remyelination proceeds nearthe lesion area [6]. Accordingly to the glial response commented above, in this section we will discuss the following topics: Cytokine modulation of the glial response to reduce secondary damage, the regulation ofreactive astrocyte and the neural stem progenitor cells (NSPCs) by the STAT3/
Several studies with the hematopoietic and IL6-family cytokines have assessed astrocyte and oligodendrocyte responses with positive outcomes on tissue preservation and motor recov‐ ery. Acute administration of GM-CSF or G-CSF by i.p. or i.v. injections during the acute phase reduced lesion area, while increasing the spared myelin area and improving motor recovery after SCI [54–56]. Both cytokine treatments reduced NG2 levels (OPC marker) during the first week of injury [54, 55]. In one of these studies it was also shown that G-CSF up-regulated BclxL and reduced apoptosis in oligodendrocytes [56]. Therefore, these results suggest that these cytokines maintain spared myelin by a protective mechanism and not by promotion of OPC
GM-CSF and G-CSF also modulate the reactive astrocyte response during the first month after SCI, as seen by the reduction of GFAP levels and the CSPG neurocan [54, 55]. Another study with improvement in motorrecovery by G-CSF intrathecal administration during the first day of injury also showed a reduction in the CSPG proteins neurocan and phosphocan [57]. Altogether with the previous studies, this last study suggest that G-CSF decrease the reac‐ tive gliosis because the treatmentreduced TGF-β levels, a growth factorthat promotes reactive gliosis; reduced vimentin levels, an astrocyte filament induced in glial scar; and presented less astrocyte proliferation [57]. Acute EPO i.p. injection also has similar results in astrocyte regulation after SCI. EPO treatment reduced lesion area and GFAP levels, while improving
siRNA treated mice also improved motor function assessed by BMS open-field score.
**6. Role of JAK-STAT in glial response**
164 Recovery of Motor Function Following Spinal Cord Injury
**6.1. Cytokines modulate glial response in SCI**
SOCS3 axis.
proliferation.
Contrary to motoneuron response in SCI, astrocyte response has been studied with cell-specific deletions of STAT3 and SOCS3. Deletions on Nestin or GFAP expressing cells, two genes upregulated in reactive astrocytes, have elucidated the role of the JAK-STAT signalling path‐ way on reactive gliosis. Mice with Nestin:STAT3KO or GFAP:STAT3KO had an increased lesion area and decreased glial scar after 2 weeks of a contusive SCI [12, 22]. Consistent with the protective role of the glial scar, this was accompanied by increased demyelination and inflammatory response, altogether with a lack of motor recovery after injury. On the contra‐ ry, the deletion of the negative feedback in a Nestin:SOCS3KO mice had prolonged and increased pSTAT3 levels in response to SCI, resulting in reduced lesion area, early and increased glial scar surrounding this area, and improved motor recovery [12].
These *in vivo* studies demonstrated that the glial JAK-STAT signalling is necessary for secondary damage contention and further studies have elucidated the cellular mechanisms modulated by this pathway. First, it seems that JAK-STAT pathway is necessary for astro‐ cyte survival in response to injury, as an *in vitro* study showed that GFAP:STAT3KO astro‐ cytes had increased necrosis and protein release after a mechanical injury [62]. Besides cell survival, JAK-STAT controls gene expression associated to reactive gliosis. STAT3 controls the expression of glial filaments, as it is known that GFAP is a target gene for STAT3 [63] and in GFAP:STAT3KO mice, reduced levels of GFAP and vimentin were detected [22]. Cell morphology is also modulated by JAK-STAT pathway. In response to SCI, astrocytes near the lesion site change their cell morphology and orientation to form the glial scar. In GFAP:STAT3KO mice, astrocytes failed to change orientation and did not form the closed boundaries of the glial scar [53].
Another STAT transcription factor that controls astrocyte response is STAT5, which is activated by GM-CSF. As commented before, this cytokine reduced GFAP and CSPG levels with an increase in motor recovery [55]. An *in vitro* study with astrocytes activated with TGFβ3, which up-regulates CSPG proteins, found that GM-CSF increased pSTAT5, pAKT and pRaf levels [64]. The cytokine also reduced TGF-β signalling and CSPGs expression. GM-CSF effects were blocked by JAK and PI3K inhibitors, suggesting that STAT5 and PI3K signalling could reduce the levels of axon inhibitory proteins secreted by reactive astrocytes [64]. Following studies should clarify if STAT5 has a similar role *in vivo*.
### **6.3. JAK-STAT modulation of neural stem progenitor cells**
Adult neural stem progenitor cells (NSPCs) can generate new astrocytes and oligodendro‐ cytes in homeostasis and in response to injury [5, 6]. NSPCs can also generate new neurons but only in some CNS specific areas, the dentate gyrus and the subventricular zone (SVZ) [65]. The role on the JAK-STAT pathway to modulate neurogenesis and gliogenesis is not clear and probably depends on the specific context, the cytokine types and cellular phenotype. At one hand, *in vitro* studies show that the JAK-STAT pathway induces gliogenesis in opposition to neurogenic differentiation, as seen in the blocking of IL6/STAT3 axis in NSPCs [66, 67]. On the other hand, there are some *in vivo* studies that have shown a JAK-STAT role for specific cytokines in adult NSPC proliferation and neurogenesis. In the dentate gyrus, CNTF and STAT3 deficient mice had reduced NSPC proliferation and neurogenesis [68]. Using cyto‐ kine injections and deficient mice it has been also shown that IL15 and IL10 regulate NSPC proliferation in the SVZ [69, 70].
Contrary to the neurogenesis of the dentate gyrus and the SVZ, in the spinal cord there is only gliogenesis. This cell differentiation occurs in three main cell types that proliferate in re‐ sponse to injury: ependymal cells, astrocytes and OPCs [6]. The role of the JAK-STAT pathway on proliferation and differentiation of ependymal cells and astrocytes has not been assessed, as transgenic mice with specific cell type deletion has been done only for Nestin and GFAP promoters, which are markers for both NSPCs and mature astrocytes. On the contrary, understanding on OPCs modulation has been done *in vivo* by NG2 knockout cells [71]. In response to injury, NG2+ cells proliferate and increase in number near the lesion site. NG2 proliferation is incremented in a NG2:SOCS3KO mice. On the contrary, in the NG2:STAT3KO mice no differences in NG2 cell number and proliferation were detected [71]. This could be explained by the modulation of OPC proliferation by other STAT transcription factors or by other pathways activated by JAK proteins (**Figure 1**), which are inhibited by SOCS3. JAK-STAT modulation of NG2 cell differentiation was also assessed in this study. Although STAT3 did not regulate OPC proliferation, in the same study it was shown that NG2:STAT3KO mice had reduced differentiation to mature oligodendrocytes during the first week of injury, but not at chronic phases (1mpi) [71].
### **7. Role of JAK-STAT in the inflammatory response**
In response to injury neutrophils, macrophages and lymphocytes infiltrate to the spinal cord [72]. There are several studies where reduction of cell infiltration improves motor recovery
and tissue sparing [73, 74]. Even so, the reduction of functional recovery after SCI in mice with total ablation of infiltrated macrophages indicates that the inflammatory response is also necessary for proper recovery [75]. These differences could be explained taking into account that immune cells have different and opposite phenotypes. Microglia and macrophages, the main effectors during spinal cord inflammation, are capable of polarization that leads to either pro-inflammatory (M1 type) or anti-inflammatory cells (M2 type). While M1 macrophages participate in secondary damage as well as in axonal retraction observed after SCI, M2 macrophages are proposed to be protective and promote axonal growth [47]. Considering the role of cytokines in the regulation of the immune response, in this section we discuss studies where JAK-STAT signalling has been shown to reduce the inflammatory response or change macrophage phenotype for improvement in spinal cord functional recovery.
### **7.1. Anti-inflammatory cytokines**
with an increase in motor recovery [55]. An *in vitro* study with astrocytes activated with TGFβ3, which up-regulates CSPG proteins, found that GM-CSF increased pSTAT5, pAKT and pRaf levels [64]. The cytokine also reduced TGF-β signalling and CSPGs expression. GM-CSF effects were blocked by JAK and PI3K inhibitors, suggesting that STAT5 and PI3K signalling could reduce the levels of axon inhibitory proteins secreted by reactive astrocytes [64]. Following
Adult neural stem progenitor cells (NSPCs) can generate new astrocytes and oligodendro‐ cytes in homeostasis and in response to injury [5, 6]. NSPCs can also generate new neurons but only in some CNS specific areas, the dentate gyrus and the subventricular zone (SVZ) [65]. The role on the JAK-STAT pathway to modulate neurogenesis and gliogenesis is not clear and probably depends on the specific context, the cytokine types and cellular phenotype. At one hand, *in vitro* studies show that the JAK-STAT pathway induces gliogenesis in opposition to neurogenic differentiation, as seen in the blocking of IL6/STAT3 axis in NSPCs [66, 67]. On the other hand, there are some *in vivo* studies that have shown a JAK-STAT role for specific cytokines in adult NSPC proliferation and neurogenesis. In the dentate gyrus, CNTF and STAT3 deficient mice had reduced NSPC proliferation and neurogenesis [68]. Using cyto‐ kine injections and deficient mice it has been also shown that IL15 and IL10 regulate NSPC
Contrary to the neurogenesis of the dentate gyrus and the SVZ, in the spinal cord there is only gliogenesis. This cell differentiation occurs in three main cell types that proliferate in re‐ sponse to injury: ependymal cells, astrocytes and OPCs [6]. The role of the JAK-STAT pathway on proliferation and differentiation of ependymal cells and astrocytes has not been assessed, as transgenic mice with specific cell type deletion has been done only for Nestin and GFAP promoters, which are markers for both NSPCs and mature astrocytes. On the contrary, understanding on OPCs modulation has been done *in vivo* by NG2 knockout cells [71]. In response to injury, NG2+ cells proliferate and increase in number near the lesion site. NG2 proliferation is incremented in a NG2:SOCS3KO mice. On the contrary, in the NG2:STAT3KO mice no differences in NG2 cell number and proliferation were detected [71]. This could be explained by the modulation of OPC proliferation by other STAT transcription factors or by other pathways activated by JAK proteins (**Figure 1**), which are inhibited by SOCS3. JAK-STAT modulation of NG2 cell differentiation was also assessed in this study. Although STAT3 did not regulate OPC proliferation, in the same study it was shown that NG2:STAT3KO mice had reduced differentiation to mature oligodendrocytes during the first week of injury, but not at
In response to injury neutrophils, macrophages and lymphocytes infiltrate to the spinal cord [72]. There are several studies where reduction of cell infiltration improves motor recovery
studies should clarify if STAT5 has a similar role *in vivo*.
166 Recovery of Motor Function Following Spinal Cord Injury
proliferation in the SVZ [69, 70].
chronic phases (1mpi) [71].
**7. Role of JAK-STAT in the inflammatory response**
**6.3. JAK-STAT modulation of neural stem progenitor cells**
IL10 is a cytokine that has been studied in SCI for anti-inflammatory modulation and neuroprotection [46]. Endogenous IL10 expression has not been assessed in detail after SCI and it is only known that M2 infiltrated macrophages express Il10 surrounding the glial scar [75]. Using IL10 deficient mice it has been found that this cytokine is necessary for control‐ ling the inflammatory response and apoptosis. After SCI, IL10 deficient mice have an in‐ creased expression of pro-inflammatory genes, increased levels of the pro-apoptotic proteins and decreased motor recovery [76]. Another study which replaced normal monocytes with IL10 deficient monocytes also resulted in reduced motor recovery, suggesting that this cytokine is necessary for the positive macrophage anti-inflammatory response [75]. IL10 treatments in SCI has resulted in motor recovery improvement by inflammatory and neuro‐ protective regulation. Acute i.p. administration of IL10 reduced lesion volume in two independent studies [77, 78]. But while one study showed that IL10 improved motor recov‐ ery and reduced TNFα expression in the injured spinal cord and in infiltrated macrophages [77], the other study did not find any functional recovery [78]. Although we previously commented other studies with motor recovery improvement by local viral over-expression of IL10 [48, 49], on those studies the inflammatory response was not assessed.
Another important anti-inflammatory cytokine is IL4, which signals by the STAT6 transcrip‐ tion factor. Although the neutralization of this cytokine by anti-IL4 antibody administration increased the inflammatory response, neitherthe IL4-neutralization treatment nor a study with STAT6 deficient mice found differences in motor recovery after SCI [79, 80]. These results indicate that Il4 signalling is not necessary for motor recovery. Further studies should test this cytokine in the promotion of the anti-inflammatory response in SCI.
In addition, there are some G-CSF studies which have shown immune modulation upon SCI. Previously commented studies for G-CSF modulation on glial cells also assessed immune response. Daily intravenous administration of G-CSF during the first 4 dpi showed that treatment not only improved motor recovery and oligodendrocyte protection, but also suppressed upregulation of the pro-inflammatory cytokines TNFα and IL1β and reduced IL1β + neutrophils infiltration [56]. Another treatment with G-CSF intrathecal injection during the first day reduced macrophage/microglia cell number and TNFα levels during the first 2 wpi [57]. Although G-CSF is well known for neuroprotective modulation [45], only one study has elucidated a biological mechanism for G-CSF inflammatory modulation on SCI. While in this study was found an increase in microglial number at 7dpi in the G-CSF treated mice, *in vitro* analyses indicated that this cytokine induced a M2 microglial phenotype, reducing proinflammatory genes while inducing IL10 expression [81]. A further characterization should be done to demonstrate that G-CSF induces a M2 phenotype *in vivo*.
### **7.2. Pro-inflammatory cytokines**
IL12 is a pro-inflammatory cytokine produced by macrophages and dendritic cells which signals via STAT4 [82]. Although it was shown that gel foam administration of IL12 regulat‐ ed immune and OPC response, this treatment slightly improved motor recovery in C57BL/6 mice and did not improve function in BALB/c strain [83]. A similar result was found after SCI in STAT4 deficient mice, no improvement was found in comparison WT mice [79]. These studies suggest that IL12 signalling is not functionally important for motor recovery.
The IL6-cytokine family has shown different effects in inflammatory response after spinal cord injury. IL11 deficient mice have been assessed in SCI, with no significant differences in motor recovery or macrophage infiltration [17]. LIF treatment incremented Mac1 levels, a marker for macrophage/microglial, and IGF-1 expression on these cells [61], while gel foam administra‐ tion of OSM reduced T-cell infiltration [15]. Although we previously discussed IL6 intrathe‐ cal administration for axon regeneration modulation [33, 34], IL6 is a pro-inflammatory cytokine which can be blocked to decrease inflammatory response. To avoid Il6 mediated immune response, a single i.p. administration of MR16-1, an anti-mouse IL6 receptor anti‐ body, has been tested with positive outcomes. MR16-1 improved motor recovery after a contusive SCI [66, 84]. This treatment also reduced the lesion area, astrocyte proliferation and increased spared myelin area and neurofilament and serotonergic fibers near the lesion area [66, 85]. The mechanism underlying the MR16-1 improvement in tissue and functional recovery has been associated to microglial and infiltrated macrophage function. MR16-1 accelerated inflammatory resolution, as number of total macrophage/microglial began to decrease earlierin the treated group [85]. Further phenotypic characterization of immune cells lead to the discovery that in MR16-1 treated mice, the microglial cells had an increased phagocytic phenotype. It was also shown that macrophages had a predominant M2 pheno‐ type and that anti-inflammatory cytokines were up-regulated while pro-inflammatory cytokines were down-regulated [84, 85].
### **8. Conclusion and clinical implications**
Cytokine upregulation and JAK-STAT signalling activation are endogenous mechanism that are activated in response to SCI and can be used to improve motor recovery. Cell type specific and *in vitro* studies have identified the role of STATs modulation in spinal cord cells. STAT3 has been the most extensively studied STAT transcription factor in SCI, but promising results have been found in other STATs and further studies should continue to determine the roles of STAT1 and STAT5 in SCI. Moreover, further studies with transgenic models should focus in other specific cell responses not studied up to now, like motoneuron and NSPC responses.
Several preclinical studies have shown positive outcomes for motor recovery and tissue sparing in JAK-STAT cytokine treatments (**Table 2**). Thus, modulation of the JAK-STAT signalling presents an opportunity to modulate neuron and glial response after an SCI in clinical settings. The hematopoietic cytokines have been already used in clinical studies for several pathologies, therefore are advanced in comparison with other JAK-STAT cytokines. Discussion of G-CSF and EPO treatments for SCI can be found in previous chapters [45, 86], while for GM-CSF there is one SCI clinical study finished which consisted in cytokine administration with transplant of bone marrow cells [87]. Although clinical studies have not been assessed for other cytokines in CNS trauma, there are clinical studies involving other pathologies that could be translated to SCI. The anti-inflammatory and neuroprotective mechanism of IL10 could also be assessed with a recombinant human IL10 that has been used for HIV infection and other several pathologies [46]. Emfilermin is a recombinant human LIF that has been tested, although with a lack of effectiveness, in clinical trials for embryo implantation and peripheral neuropathy [88, 89] that could also be used in SCI or other CNS diseases. Finally, tocilizumab, an anti-human IL6R, has shown positive outcomes in clinical studies for rheumatoid arthritis [90] and could be used for anti-inflammatory treatment in the spinal cord acute response.
### **Acknowledgements**
[57]. Although G-CSF is well known for neuroprotective modulation [45], only one study has elucidated a biological mechanism for G-CSF inflammatory modulation on SCI. While in this study was found an increase in microglial number at 7dpi in the G-CSF treated mice, *in vitro* analyses indicated that this cytokine induced a M2 microglial phenotype, reducing proinflammatory genes while inducing IL10 expression [81]. A further characterization should be
IL12 is a pro-inflammatory cytokine produced by macrophages and dendritic cells which signals via STAT4 [82]. Although it was shown that gel foam administration of IL12 regulat‐ ed immune and OPC response, this treatment slightly improved motor recovery in C57BL/6 mice and did not improve function in BALB/c strain [83]. A similar result was found after SCI in STAT4 deficient mice, no improvement was found in comparison WT mice [79]. These
The IL6-cytokine family has shown different effects in inflammatory response after spinal cord injury. IL11 deficient mice have been assessed in SCI, with no significant differences in motor recovery or macrophage infiltration [17]. LIF treatment incremented Mac1 levels, a marker for macrophage/microglial, and IGF-1 expression on these cells [61], while gel foam administra‐ tion of OSM reduced T-cell infiltration [15]. Although we previously discussed IL6 intrathe‐ cal administration for axon regeneration modulation [33, 34], IL6 is a pro-inflammatory cytokine which can be blocked to decrease inflammatory response. To avoid Il6 mediated immune response, a single i.p. administration of MR16-1, an anti-mouse IL6 receptor anti‐ body, has been tested with positive outcomes. MR16-1 improved motor recovery after a contusive SCI [66, 84]. This treatment also reduced the lesion area, astrocyte proliferation and increased spared myelin area and neurofilament and serotonergic fibers near the lesion area [66, 85]. The mechanism underlying the MR16-1 improvement in tissue and functional recovery has been associated to microglial and infiltrated macrophage function. MR16-1 accelerated inflammatory resolution, as number of total macrophage/microglial began to decrease earlierin the treated group [85]. Further phenotypic characterization of immune cells lead to the discovery that in MR16-1 treated mice, the microglial cells had an increased phagocytic phenotype. It was also shown that macrophages had a predominant M2 pheno‐ type and that anti-inflammatory cytokines were up-regulated while pro-inflammatory
Cytokine upregulation and JAK-STAT signalling activation are endogenous mechanism that are activated in response to SCI and can be used to improve motor recovery. Cell type specific and *in vitro* studies have identified the role of STATs modulation in spinal cord cells. STAT3 has been the most extensively studied STAT transcription factor in SCI, but promising results have been found in other STATs and further studies should continue to determine the roles of
studies suggest that IL12 signalling is not functionally important for motor recovery.
done to demonstrate that G-CSF induces a M2 phenotype *in vivo*.
**7.2. Pro-inflammatory cytokines**
168 Recovery of Motor Function Following Spinal Cord Injury
cytokines were down-regulated [84, 85].
**8. Conclusion and clinical implications**
Special thanks to Daniel Guzmán for critical reading and valuable suggestions. This work was supported by ICM (P07/011-F, P09/016-F) and FONDECYT (1141162).
### **Author details**
Victor S. Tapia and Juan Larrain\*
\*Address all correspondence to: jlarrain@bio.puc.cl
Center for Aging and Regeneration, Millennium Nucleus in Regenerative Biology, Faculty of Biological Sciences, Pontifical Catholic University of Chile, Santiago, Chile
### **References**
[1] Lee B.B., Cripps R.A., Fitzharris M., Wing P.C. The global map fortraumatic spinal cord injury epidemiology: update 2011, global incidence rate. Spinal Cord. 2014;52:110–116. DOI: 10.1038/sc.2012.158
[16] Yamauchi K., Osuka K., Takayasu M., Usuda N., Nakazawa A., Nakahara N., et al. Activation of JAK/STAT signalling in neurons following spinal cord injury in mice. Journal of Neurochemistry. 2006;96(4):1060–1070. DOI: 10.1111/j. 1471-4159.2005.03559.x
[2] Thuret S., Moon L.D.F., Gage F.H. Therapeutic interventions after spinal cord injury.
[4] Burda J.E., Sofroniew M.V. Reactive gliosis and the multicellular response to CNS damage and disease. Neuron. 2014;81(2):229–248. DOI: 10.1016/j.neuron.2013.12.034
[5] Horky L., Galimi F. Fate of endogenous stem/progenitor cells following spinal cord injury. Journal of Comparative Neurology. 2006;498(4):525–538. DOI: 10.1002/cne.
[6] Barnabé-Heider F., Göritz C., Sabelström H., Takebayashi H., Pfrieger F.W., Meletis K., et al. Origin of new glial cells in intact and injured adult spinal cord. Cell Stem Cell.
[7] Bastien D., Lacroix S. Cytokine pathways regulating glial and leukocyte function after spinal cord and peripheral nerve injury. Experimental Neurology. 2014;258:62–77.
[8] Kundi S., Bicknell R., Ahmed Z. The role of angiogenic and wound-healing factors after spinal cord injury in mammals. Neuroscience Research. 2013;76:1–9. DOI: 10.1016/
[9] Murray P. The JAK-STAT signaling pathway: input and outputintegration. The Journal
[10] Rawlings J.S., Rosler K.M., Harrison D. The JAK/STAT signaling pathway. Journal of
[11] Smith P., Sun F., Park K., Cai B., Wang C. SOCS3 deletion promotes optic nerve regeneration in vivo. Neuron. 2009;64(5):617–623. DOI: 10.1016/j.neuron.
[12] Okada S., Nakamura M., Katoh H., Miyao T., Shimazaki T., Ishii K., et al. Conditional ablation of Stat3 or Socs3 discloses a dual role for reactive astrocytes after spinal cord
[13] Jin D., Liu Y., Sun F., Wang X., Liu X., He Z. Restoration of skilled locomotion by sprouting corticospinal axons induced by co-deletion of PTEN and SOCS3. Nature
[14] Pineau I., Lacroix S. Proinflammatory cytokine synthesis in the injured mouse spinal cord: multiphasic expression pattern and identification of the cell types involved.
[15] Slaets H., Nelissen S., Janssens K., Vidal P.M., Lemmens E., Stinissen P., et al. Oncostatin M reduces lesion size and promotes functional recovery and neurite outgrowth after spinal cord injury. Molecular Neurobiology. 2014;50(3):1142–1151. DOI: 10.1007/
Journal of Comparative Neurology. 2007;500(2):267–285. DOI: 10.1002/cne
injury. Nature Medicine. 2006;12(7):829–834. DOI: 10.1038/nm1425
Communications. 2015;6:8074. DOI: 10.1038/ncomms9074
of Immunology. 2007;178:2623–2629. DOI: 10.4049/jimmunol.178.5.2623
Cell Science. 2004;117:1281–1283. DOI: 10.1242/jcs.00963
2010;7(4):470–482. DOI: 10.1016/j.stem.2010.07.014
DOI: 10.1016/j.expneurol.2014.04.006
j.neures.2013.03.013
2009.11.021.SOCS3
s12035-014-8795-5
Nature Reviews Neuroscience. 2006;7(8):628–643. DOI: 10.1038/nrn1955
21065.Fate
170 Recovery of Motor Function Following Spinal Cord Injury
[3] Cannon B. Sensation and loss. Nature. 2013;503(7475):S2–3. DOI: 10.1038/503S2a
[40] Nishio Y., Koda M., Kamada T. Granulocyte colony-stimulating factor attenuates neuronal death and promotes functional recovery after spinal cord injury in mice. Journal of Neuropathology and Experimental Neurology. 2997;66(8):724–731.
[28] Ye J., Cao L., Cui R., Huang A., Yan Z., Lu C., et al. The effects of ciliary neurotrophic factor on neurological function and glial activity following contusive spinal cord injury in the rats. Brain Research. 2004;997(1):30–39. DOI: 10.1016/j.brainres.2003.10.036 [29] Wang N., Zhang S., Zhang F., Yang Z.Y., Li X.G. Sodium hyaluronate-CNTF gelati‐ nous particles promote axonal growth, neurogenesis and functional recovery after
spinal cord injury. Spinal Cord. 2014;52(7):517–523. DOI: 10.1038/sc.2014.54
1471-4159.2010.06659.x
172 Recovery of Motor Function Following Spinal Cord Injury
DOI: 10.1016/j.expneurol.2012.03.019
DOI: 10.1016/j.brainres.2010.07.046
Chemistry. 1999;274(32):22165–22169.
1471-4159.2009.06263.x.
[30] Pitzer C., Klussmann S., Krüger C., Letellier E., Plaas C., Dittgen T., et al. The hemato‐ poietic factor granulocyte-colony stimulating factor improves outcome in experimen‐ tal spinal cord injury. Journal of Neurochemistry. 210;113(4):930–942. DOI: 10.1111/j.
[31] Pan W., Cain C., Yu Y., Kastin A.J. Receptor-mediated transport of LIF across bloodspinal cord barrier is upregulated after spinal cord injury. Journal of Neuroimmunol‐
[32] Li Y., Zang D. The neuron regrowth is associated with the proliferation of neural precursor cells after leukemia inhibitory factor administration following spinal cord injury in mice. PloS One. 2014;9(12):e116031. DOI: 10.1371/journal.pone.0116031 [33] Yang P., Wen H., Ou S., Cui J., Fan D. IL-6 promotes regeneration and functional recovery after cortical spinal tract injury by reactivating intrinsic growth program of neurons and enhancing synapse formation. Experimental Neurology. 2012;236:19–27.
[34] Yang P., Qin Y., Bian C., Zhao Y., Zhang W. Intrathecal delivery of IL-6 reactivates the intrinsic growth capacity of pyramidal cells in the sensorimotor cortex after spinal cord
[35] ZhangY., XiongY., Mahmood A., MengY., LiuZ. Sprouting of corticospinaltract axons from the contralateral hemisphere into the denervated side of the spinal cord is associated with functional recovery in adult rat. Brain Research. 2010;1353:249–257.
[36] Stephanou A., Brar B. Opposing actions of STAT-1 and STAT-3 on the Bcl-2 and Bcl-x
[37] Silva M., Benito A., Sanz C., Prosper F. Erythropoietin can induce the expression of bclxLthrough stat5 in erythropoietin-dependent progenitor cell lines. Journal of Biological
[38] Zhou Z., Peng X., Insolera R. Interleukin-10 provides direct trophic support to neurons. Journal of Neurochemistry. 2009;110(5):1617–1627. DOI: 10.1111/j.
[39] Park K.W., Nozell S.E., Benveniste E.N. Protective role of STAT3 in NMDA and glutamate-induced neuronal death: negative regulatory effect of SOCS3. PloS One.
promoters. Cell Death and Differentiation. 2000;7(3):329–330.
1012;7(11):e50874. DOI: 10.1371/journal.pone.0050874
injury. PloS One. 2015;10(5):e0127772. DOI: 10.1371/journal.pone.0127772
ogy. 2006;174:119–125. DOI: 10.1016/j.jneuroim.2006.02.006
[63] Takizawa T., Nakashima K., Namihira M. DNA methylation is a critical cell-intrinsic determinant of astrocyte differentiation in the fetal brain. Developmental Cell. 2001;1(6):749–758. DOI: 10.1016/S1534-5807(01)00101-0
[52] Faulkner J.R., Herrmann J.E., Woo M.J., Tansey K.E., Doan N.B., Sofroniew M.V. Reactive astrocytes protect tissue and preserve function after spinal cord injury. Journal of Neuroscience. 2004;24(9):2143–2155. DOI: 10.1523/JNEUROSCI.3547-03.2004
[53] Wanner I.B., Anderson M., Song B., Levine J., Fernandez A., Gray-Thompson Z., et al. Glial scar borders are formed by newly proliferated, elongated astrocytes that inter‐ act to corral inflammatory and fibrotic cells via STAT3-dependent mechanisms after spinal cord injury. Journal of Neuroscience. 2013;33(31):12870–12886. DOI: 10.1523/
[54] Huang X., Kim J.M., Kong T.H., Park S.R., Ha Y., Kim M.H., et al. GM-CSF inhibits glial scar formation and shows long-term protective effect after spinal cord injury. Journal
[55] Chung J., Kim M., Yoon Y., Kim K. Effects of granulocyte colony–stimulating factor and granulocyte-macrophage colony–stimulating factor on glial scar formation after spinal cord injury in rats. Journal of Neurosurgery: Spine. 2014;21(6):966–973. DOI:
[56] Kadota R., Koda M., Kawabe J., Hashimoto M., Nishio Y., Mannoji C., et al. Granulo‐ cyte colony-stimulating factor (G–CSF) protects oligodendrocyte and promotes hindlimb functional recovery after spinal cord injury in rats. PloS One.
[57] ChenW.F., Chen C.H., Chen N.F., Sung C.S., WenZ.H. Neuroprotective effects of direct intrathecal administration of granulocyte colony-stimulating factor in rats with spinal cord injury. CNS Neuroscience & Therapeutics. 2015;21(9):698–707. DOI: 10.1111/cns.
[58] Gorio A., Madaschi L. Methylprednisolone neutralizes the beneficial effects of erythropoietin in experimental spinal cord injury. Proceedings of the National Academy of Sciences of the United States of America. 2005;102(45):16379–16384. DOI:
[59] Vitellaro-Zuccarello L., Mazzetti S., Madaschi L., Bosisio P., Gorio A., De Biasi S. Erythropoietin-mediated preservation of the white matter in rat spinal cord injury.
[60] Azari M., Profyris C. Leukemia inhibitory factor arrests oligodendrocyte death and demyelination in spinal cord injury. Journal of Neuropathology and Experimental
[61] Kerr B.J., Patterson P.H. Leukemia inhibitory factor promotes oligodendrocyte survival
[62] Levine J., Kwon E., Paez P., Yan W., Czerwieniec G., Loo J., et al. Traumatically injured astrocytes release a proteomic signature modulated by STAT3-dependent cell surviv‐
Neuroscience. 2007;144(3):865–877. DOI: 10.1016/j.neuroscience.2006.10.023
Neurology. 2006;65(6):914–929. DOI: 10.1097/01.jnen.0000235855.77716.25
after spinal cord injury. Glia. 2005;51:73–79. DOI: 10.1002/glia.20177
al. Glia. 2015; 64(5):668-694. DOI: 10.1002/glia.22953
of the Neurological Science. 2009;277:87–97. DOI: 10.1016/j.jns.2008.10.02
JNEUROSCI.2121-13.2013
174 Recovery of Motor Function Following Spinal Cord Injury
10.3171/2014.8.SPINE131090
10.1073/pnas.0508479102
12429
2012;7(11):e50391. DOI: 10.1371/journal.pone.0050391
spinal cord: A potential neuroprotective treatment. Journal of Neuroimmunology. 2004;156:42–57. DOI: 10.1016/j.jneuroim.2004.07.002
alternative pathway of macrophage activation after spinal cord injury in mice. Journal of Neuroinflammation. 2012;9:40. DOI: 10.1186/1742-2094-9-40
[85] Mukaino M., Nakamura M., Yamada O., Okada S., Morikawa S., Renault-Mihara F., et al. Anti-IL-6-receptor antibody promotes repair of spinal cord injury by inducing microglia-dominant inflammation. Experimental Neurology. 2010;224:403–414. DOI: 10.1016/j.expneurol.2010.04.020
spinal cord: A potential neuroprotective treatment. Journal of Neuroimmunology.
[75] Shechter R., London A., Varol C., Raposo C., Cusimano M., Yovel G., et al. Infiltrat‐ ing blood-derived macrophages are vital cells playing an anti-inflammatory role in recovery from spinal cord injury in mice. PLoS Medicine. 2009;6(7):e1000113. DOI:
[76] GenoveseT., EspositoE., MazzonE., DiPaolaR., CaminitiR., BramantiP., et al. Absence of endogenous interleukin-10 enhances secondary inflammatory process after spinal cord compression injury in mice. Journal of Neurochemistry. 2009;108(6):1360–1372.
[77] Bethea J.R., Nagashima H., Acosta M.C., Briceno C., Gomez F., Marcillo A.E., et al. Systemically administered interleukin-10 reduces tumor necrosis factor-alpha produc‐ tion and significantly improves functional recovery following traumatic spinal cord.
[78] Takami T., Oudega M., Bethea J.R., Wood P.M., Kleitman N., Bunge M.B. Methylpred‐ nisolone and interleukin-10 reduce gray matter damage in the contused Fischer rat thoracic spinal cord but do not improve functional outcome. Journal of Neurotrauma.
[79] Fraidakis M., Kiyotani T., Pernold K. Recovery from spinal cord injury in tumor necrosis factor-alpha, signal transducers and activators of transcription 4 and signal transducers and activators of transcription 6 null mice. Neuroimmunology. 2007;18(2):
[80] Lee S.I., Jeong S.R., Kang Y.M., Han D.H., Jin B.K., Namgung U., et al. Endogenous expression of interleukin-4 regulates macrophage activation and confines cavity formation after traumatic spinal cord injury. Journal of Neuroscience Research.
[81] Guo Y., Zhang H., Yang J., Liu S., Bing L., Gao J., et al. Granulocyte colony-stimulat‐ ing factor improves alternative activation of microglia under microenvironment of spinal cord injury. Neuroscience. 2013;238:1–10. DOI: 10.1016/j.neuroscience.
[82] Vignali D., Kuchroo V.K. IL-12 family cytokines: immunological playmakers. Nature
[83] Yaguchi M., Ohta S., Toyama Y., Kawakami Y., Toda M. Functional recovery after spinal cord injury in mice through activation of microglia and dendritic cells afterIL-12 administration. Journal of Neuroscience Research. 2008;86(9):1972–1980. DOI: 10.1002/
[84] Guerrero A.R., Uchida K., Nakajima H., Watanabe S., Nakamura M., Johnson W.E., et al. Blockade of interleukin-6 signaling inhibits the classic pathway and promotes an
Journal of Neurotrauma. 1999;16(10):851–863. DOI: 10.1089/neu.1999.16.851
2004;156:42–57. DOI: 10.1016/j.jneuroim.2004.07.002
2002;19(5):653–666. DOI: 10.1089/089771502753754118
Immunology. 2012;13(8):722–728. DOI: 10.1038/ni.2366
2010;88:2409–2419. DOI: 10.1002/jnr.22411
10.1371/journal.pmed.1000113
176 Recovery of Motor Function Following Spinal Cord Injury
185–189.
2013.01.047
jnr.21658
DOI: 10.1111/j.1471-4159.2009.05899.x
## **Surgical Approaches for Repair and Regeneration Following Spinal Cord Injury**
### **Bridging Defects in Chronic Spinal Cord Injury Using Peripheral Nerve Grafts: From Basic Science to Clinical Experience Bridging Defects in Chronic Spinal Cord Injury Using Peripheral Nerve Grafts: From Basic Science to Clinical Experience**
Sherif M. Amr Sherif M. Amr
Additional information is available at the end of the chapter Additional information is available at the end of the chapter
#### **Abstract Abstract**
http://dx.doi.org/10.5772/64211
http://dx.doi.org/10.5772/64211
Nerve grafting of the injured spinal cord should pursue a sixfold attack: lysing the fibrosis/gliosis to an extent that allows settling of the basal lamina preventing mean‐ while collapse of the neural tissue matrix; supplying the tissue matrix with a suitable scaffold, on which the basal lamina can settle; basal lamina synthesis; seeding the basal lamina with cell adhesion molecules; providing the axonal growth cone with neurite outgrowth promoting factors that allow its distal progression; supplying the axonal growth cone with neurotrophic factors that power its continued growth. In addition to this, the intrinsic properties of the neurons should be stimulated, possibly through modulating the function of astrocytes by heparin, aspirin and other factors. Nerve side grafting of the cord increases the incidence of nerve regeneration by applying additional grafts extending from the side of the donor end of the cord to the side of the recipient end. Also, it allows the surgeon to enhance regeneration through a partially regenerat‐ ed cord. During surgery, after establishment of CSF circulation, a long-lasting indwelling catheter has to be inserted for postoperative drug and cell delivery. This allows for continual lysis of the gliosis by chondroitinase ABC, sialidase and other factors. spinal cord injury, nerve grafting, indwelling intrathecal catheter **Chapter 8** Nerve grafting of the injured spinal cord should pursue a sixfold attack: lysing the fibrosis/gliosis to an extent that allows settling of the basal lamina preventing mean‐ while collapse of the neural tissue matrix; supplying the tissue matrix with a suitable scaffold, on which the basal lamina can settle; basal lamina synthesis; seeding the basallamina with cell adhesion molecules; providing the axonal growth cone with neurite outgrowth promoting factors that allow its distal progression; supplying the axonal growth cone with neurotrophic factors that power its continued growth. In addition to this, the intrinsic properties of the neurons should be stimulated, possibly through modulating the function of astrocytes by heparin, aspirin and other factors. Nerve side graftingofthecordincreasestheincidenceofnerveregenerationbyapplyingadditionalgrafts extending from the side of the donor end of the cord to the side of the recipient end. Also, it allows the surgeon to enhance regeneration through a partially regenerat‐ ed cord. During surgery, after establishment of CSF circulation, a long-lasting indwelling catheter has to be inserted for postoperative drug and cell delivery. This allows for continual lysis of the gliosis by chondroitinase ABC, sialidase and other factors.
**Keywords:** spinal cord injury, nerve grafting, indwelling intrathecal catheter implan‐ tation, lysis of the gliosis, chondroitinase ABC **Keywords:** spinal cord injury, nerve grafting, indwelling intrathecal catheter implan‐ tation, lysis of the gliosis, chondroitinase ABC
#### **1. Introduction 1. Introduction**
Since 1903, when Tello and Cajal demonstrated that the central nervous system (CNS) could regenerate [1, 2], experimentalneurosciencehas advancedour knowledge repairing the injured Since 1903, when Tello and Cajal demonstrated that the central nervous system (CNS) could regenerate [1, 2], experimentalneurosciencehas advancedour knowledge repairing the injured
© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative CommonsAttribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
spinal cord. Several cellular transplantation strategies have been recommended with some clinical success [3–6]. Clinically, however, the injured spinal cord is usually extensively gliotic, cystic, evendisrupted,necessitating bridging the injury zone first before contemplating cellular transplantation. Placing peripheral nerve grafts to bridge the injury zone has been successful experimentally[7–13],yetonlyanecdotal clinical evidence supports it[14, 15].Ina reviewarticle on bridging spinal cord injuries, Fawcett[16] commented on this disparity between experimen‐ tal and clinical neuroscience: 'Sadly, we have yet to achieve a treatment that is licensed for this purpose in human patients'. The aim of this review is to enable clinicians to put the findings made by neuroscientists into clinical practice and to provide neuroscientists with upcoming ideas investigating the clinical issues physicians face. avoiding any tension at the repair for even minimal can end up with 182Following Spinal Cord Injury
### **2. Changing concepts of nerve grafting**
### **2.1. Basic concepts of nerve grafting**
Autogenous nerve grafting is the standard for repair of irreducible nerve gaps [17]. The basic principles of nerve grafting include the following (**Figure 1**):
**Figure 1.** The basic principles of conventional end-to-end grafting include trimming both proximal and distal nerve ends up to healthy nerve fascicles; avoiding any tension at the repair site, avoiding any shearing stress at the repair site, fascicular grafting, end-to-end grafting suturing fascicles at proximal nerve ends to their counterparts at distal nerve ends after grouping them topographically, using small caliber sutures (9/0 or 10/0 sutures), and healthy vascular bed. The basic principles of conventional end-to-end grafting include trimming both proximal and distal ends up to healthy nerve avoiding any tension at the repair avoiding any shearing stress at the fascicular end-to-end grafting suturing fascicles at proximal nerve ends to their counterparts at using (9/0 or healthy
In the absence of a proximal nerve end, such as in brachial plexus avulsions, nerve transfer (neurotisation) refers to using an expendable nearby donor nerve as a substitute, grafting it to the original recipient. The principles of nerve transfer include: In the absence of a proximal nerve such as in brachial plexus nerve
spinal cord. Several cellular transplantation strategies have been recommended with some clinical success [3–6]. Clinically, however, the injured spinal cord is usually extensively gliotic, cystic, evendisrupted,necessitating bridging the injury zone first before contemplating cellular transplantation. Placing peripheral nerve grafts to bridge the injury zone has been successful experimentally[7–13],yetonlyanecdotal clinical evidence supports it[14, 15].Ina reviewarticle on bridging spinal cord injuries, Fawcett[16] commented on this disparity between experimen‐ tal and clinical neuroscience: 'Sadly, we have yet to achieve a treatment that is licensed for this purpose in human patients'. The aim of this review is to enable clinicians to put the findings made by neuroscientists into clinical practice and to provide neuroscientists with upcoming
Autogenous nerve grafting is the standard for repair of irreducible nerve gaps [17]. The basic
regeneration;avoiding stress site because inflammatory 182Following Spinal Cord Injury
irreducible The basic principles of conventional end-to-end grafting include trimming both proximal and distal
**Figure 1.** The basic principles of conventional end-to-end grafting include trimming both proximal and distal nerve ends up to healthy nerve fascicles; avoiding any tension at the repair site, avoiding any shearing stress at the repair site, fascicular grafting, end-to-end grafting suturing fascicles at proximal nerve ends to their counterparts at distal nerve ends after grouping them topographically, using small caliber sutures (9/0 or 10/0 sutures), and healthy vascular
ends up to healthy nerve avoiding any tension at the repair avoiding any shearing stress at the fascicular end-to-end grafting suturing fascicles at proximal nerve ends to their counterparts at
small
avoiding any tension at the repair for even minimal can end up with
ending up with fibrosis and hampering progression of regeneration;
ideas investigating the clinical issues physicians face.
principles of nerve grafting include the following (**Figure 1**):
**2. Changing concepts of nerve grafting**
**of nerve concepts of**
**2.1. Basic concepts of nerve grafting**
2 Recovery of Motor Function in Spinal Cord Injury
hampering progression of regeneration;
bed.
Autogenous grafts act as immunogenically inert scaffolds, providing appropriate neurotro‐ phic factors and viable Schwann cells for axonal regeneration [17]. Autogenous grafts act as immunogenically inert scaffolds, providing appropriate
### **2.2. Molecular aspects of peripheral nerve regeneration**
Advances in the understanding of molecular pathways and their physiological role have provided us with new insights as to the mechanism of axonal (peripheral nerve) regeneration [18, 19]. [17].Advances in the understanding of molecular pathways and their physiological role provided us with new insights as to the mechanism of axonal (peripheral nerve) regenera‐ tion [18, 19].
Fibrous tissue and chondroitin sulphate proteoglycans secreted by astrocytes provide the necessary scaffold for settling of the basal lamina and subsequent basement membrane synthesis. *Neurite outgrowth promoting factors* are basement membrane-related extracellular matrix proteins (such as laminin (LN), fibronectin (FN), heparin sulphate proteoglycans (HSP) and tenascin), which pave the proper path by supplying orientation and adhesiveness for axons. *Neurotrophic factors* are specific trophic agents that power peripheral nerve regeneration. They include (a) the neurotrophins (nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3) and neurotrophin-4/5(NT-4/5)); (b) the neurokines (ciliary neurotrophic factor (CNTF) and leukaemia inhibitory factor (LIF)); (c) the transforming growth factor (TGF)-b family (TGF-b1, TGF-b2, TGF-b3, glial cell line¦derived neurotrophic factor (GDNF)). *Schwann cells* secrete (a) cell adhesion molecules (CAMs), such as N-CAM, Ng-CAM/L1, N-cadherin and L2/HNK-1; (b) produce basement membrane that contains many extracellular matrix proteins, such as laminin (LN), fibronectin (FN), heparin sulphate proteoglycans (HSP) and tenascin; (c) secrete neurotrophic factors and their receptors. In addition to this, axonal regeneration is determined by trophic factors from activated perineural glial cells (astrocytes), trophic factors from efferent axons by anterograde transport, geneinduced trophic factors by intracrine or autocrine transport, trophic factors from retrograde target cell support, trophic factors from retroaxonal transport and trophic factors from Fibrous tissue and chondroitin sulphate proteoglycans secreted by astrocytes provide necessary scaffold for settling of the basal lamina and subsequent basement *Neurite outgrowth promoting* are basement membrane-related matrix (such as (LN), heparin sulphate and which pave the proper path by supplying orientation and adhesiveness *Neurotrophic factors* are specific trophic agents that power peripheral nerve regenera‐tion. They include (a) the neurotrophins (nerve growth factor (NGF), brain-derived neurotro‐ phic factor (BDNF), neurotrophin-3 (NT-3) and neurotrophin-4/5(NT-4/5)); (b) the neurokines (ciliary neurotrophic factor (CNTF) and leukaemia inhibitory factor (LIF)); (c) the transforming growth factor (TGF)-b family (TGF-b1, TGF-b2, TGF-b3, glial cell line¦derived neurotrophic factor (GDNF)). *Schwann cells* secrete (a) cell adhesion molecules (CAMs), such as N-CAM, Ng-CAM/L1, N-cadherin and L2/HNK-1; (b) produce basement membrane that contains many extracellular matrix proteins, such as laminin (LN), fibronectin (FN), heparin sulphate proteoglycans (HSP) and tenascin; (c) secrete neurotrophic factors and their recep‐ tors. In addition to this, axonal regeneration is determined by trophic factors from activated perineural glial cells (astrocytes), trophic factors from efferent axons by anterograde trans‐ port, gene-induced trophic factors by intracrine or autocrine transport, trophic factors from retrograde target cell support, trophic factors from retroaxonal transport and trophic factors recruited macrophages (secretory products, cytokines). Activated mesenchymal cells contrib‐ ute to repair and vascularisation (**Figure 2**). from recruited macrophages (secretory products, cytokines). Activated mesenchymal cellscontribute to repair and vascularisation (**Figure 2**).
**Figure 2.** Molecular aspects of axonal regeneration. (I) *Neurite outgrowth promoting factors* [such as laminin (LN), fibro‐ nectin (FN), heparin sulphate proteoglycans (HSP) and tenascin] pave the proper path by supplying orientation and adhesiveness for axons. (II) *Neurotrophic factors* are the specific trophic agents that power peripheral nerve regenera‐ tion. They include (a) neurotrophins (NGF, BDNF, NT-4/5); (b) neurokines (CNTF, LIF); (c) (TGF)-β family (TGF-β1, TGF-β2, TGF-β3, GDNF). (III) *Schwann cells* secrete (a) cell adhesion molecules (CAMs), such as N-CAM, Ng-CAM/L1, N-cadherin, andL2/HNK-1; (b) basement membrane that contains many extracellular matrix proteins, such as laminin (LN), fibronectin (FN), heparin sulphate proteoglycans (HSP) and tenascin; (c) neurotrophic factors and their receptors. In addition to this, axonal regeneration is determined by trophic factors from activated perineural glial cells (astro‐ cytes) (IV), trophic factors from efferents by anterograde transport (V), gene-induced trophic factors by intracrine (VI) or autocrine (VII) transport, trophic factors from retrograde target cell support (VIII), trophic factors from retroaxonal transport (IX), trophic factors from recruited macrophages (secretory products X, cytokines XI). Activated mesenchy‐ mal cells contribute to repair and vascularisation (XII). Among other molecules, heparin and aspirin modulate astro‐ cytic function stimulating them to secrete axonal trophic factors (IV). **2.** Molecular aspects of axonal *Neurite outgrowth promoting* [such as laminin (LN), heparin sulphate and pave the proper path by supplying orientation adhesiveness for *Neurotrophic* are the specific trophic agents that power peripheral nerve They BDNF, (TGF)-β (TGF-β1,TGF-β3, GDNF). (III) N-cadherin, andL2/HNK-1; (b) basement membrane that contains many extracellular matrix proteins, such as lamininIn addition to this, axonal regeneration is determined by trophic factors from activated perineural glial cells (astro‐trophic factors from efferents by anterograde gene-induced trophic factors by or (VII) trophic factors from retrograde target cell trophic factors from transport (IX), trophic factors from recruited macrophages (secretory products X, cytokines XI). Activated mal cells contribute to repair and vascularisation (XII). Among other molecules, heparin and aspirin modulate
Based on the previous considerations, axonal sprouting and nerve grafting are based on a sixfold attack (**Figure 3**): Based on the previous axonal sprouting and nerve grafting are based on
– lysing the fibrosis/gliosis in the injury zone to an extent that allows settling of the basal lamina preventing meanwhile collapse of the neural tissue matrix; or excision of the fibrotic segment and replacing it with nerve grafts; lysing the fibrosis/gliosis in the injury zone to an extent that allows settling ofthe basal preventing meanwhile collapse of the neural tissue matrix; or excision of the fibrotic seg‐ment and replacing it with nerve grafts;
– supplying the tissue matrix with a suitable scaffold, on which the basal lamina can settle,
**Figure 3.** The sixfold attack: lysing the fibrosis/gliosis to an extent that allows settling of the basal lamina preventing meanwhile collapse of the neural tissue matrix; supplying the tissue matrix with a suitable scaffold, on which the basal lamina can settle; basal lamina synthesis; seeding the basal lamina with cell adhesion molecules; providing the axonal growth cone with neurite outgrowth promoting factors (FGF) that allow its distal progression; supplying the axonal growth cone with neurotrophic factors that allow its continued growth. The sixfold lysing the fibrosis/gliosis to an extent that allows settling of the basal lamina lamina can basal lamina seeding the basal lamina with cell adhesion providing the growth cone with neurite outgrowth promoting that allow its distal supplying the
– basal lamina synthesis;
recruited macrophages (secretory products, cytokines). Activated mesenchymal cells contrib‐
from recruited macrophages (secretory products, cytokines). Activated mesenchymal cells
**Figure 2.** Molecular aspects of axonal regeneration. (I) *Neurite outgrowth promoting factors* [such as laminin (LN), fibro‐ nectin (FN), heparin sulphate proteoglycans (HSP) and tenascin] pave the proper path by supplying orientation and adhesiveness for axons. (II) *Neurotrophic factors* are the specific trophic agents that power peripheral nerve regenera‐ tion. They include (a) neurotrophins (NGF, BDNF, NT-4/5); (b) neurokines (CNTF, LIF); (c) (TGF)-β family (TGF-β1, TGF-β2, TGF-β3, GDNF). (III) *Schwann cells* secrete (a) cell adhesion molecules (CAMs), such as N-CAM, Ng-CAM/L1, N-cadherin, andL2/HNK-1; (b) basement membrane that contains many extracellular matrix proteins, such as laminin (LN), fibronectin (FN), heparin sulphate proteoglycans (HSP) and tenascin; (c) neurotrophic factors and their receptors. In addition to this, axonal regeneration is determined by trophic factors from activated perineural glial cells (astro‐ cytes) (IV), trophic factors from efferents by anterograde transport (V), gene-induced trophic factors by intracrine (VI) or autocrine (VII) transport, trophic factors from retrograde target cell support (VIII), trophic factors from retroaxonal transport (IX), trophic factors from recruited macrophages (secretory products X, cytokines XI). Activated mesenchy‐ mal cells contribute to repair and vascularisation (XII). Among other molecules, heparin and aspirin modulate astro‐
N-cadherin, (b) basement membrane that contains many extracellular matrix proteins, such as lamininIn addition to this, axonal regeneration is determined by trophic factors from activated perineural glial cells (astro‐
trophic factors from efferents by anterograde gene-induced trophic factors by or trophic factors from retrograde target cell trophic factors from transport (IX), trophic factors from recruited macrophages (secretory products X, cytokines XI). Activated mal cells contribute to repair and vascularisation (XII). Among other molecules, heparin and aspirin modulate
adhesiveness for *Neurotrophic* are the specific trophic agents that power peripheral nerve
Molecular aspects of axonal *Neurite outgrowth promoting* [such as laminin (LN), heparin sulphate and pave the proper path by supplying orientation
They (TGF)-β (TGF-β1,TGF-β3, GDNF). (III) *cells* N-CAM,
Based on the previous considerations, axonal sprouting and nerve grafting are based on a
Based on the previous axonal sprouting and nerve grafting are based on
):lysing the fibrosis/gliosis in the injury zone to an extent that allows settling ofthe basal
– lysing the fibrosis/gliosis in the injury zone to an extent that allows settling of the basal lamina preventing meanwhile collapse of the neural tissue matrix; or excision of the fibrotic segment
preventing meanwhile collapse of the neural tissue matrix; or excision of the fibrotic seg‐
– supplying the tissue matrix with a suitable scaffold, on which the basal lamina can settle,
cytic function stimulating them to secrete axonal trophic factors (IV).
sixfold attack (**Figure 3**):
and replacing it with nerve grafts;
ment and replacing it with nerve grafts;
tissue matrix 184Following Spinal Cord Injury
ute to repair and vascularisation (**Figure 2**).
contribute to repair and vascularisation (**Figure 2**).
4 Recovery of Motor Function in Spinal Cord Injury
– seeding the basal lamina with cell adhesion molecules;
– providing the axonal growth cone with neurite outgrowth promoting factors that allow its distal progression;
– supplying the axonal growth cone with neurotrophic factors that power its continued growth.
### **2.3. Changing concepts of nerve grafting: side grafting**
The previous conditions prevailing, if the side of a motor nerve is injured, the axonal growth cone may be enticed to grow off motor nerve side to the injured end of another motor nerve, the so-called recipient end to donor side coaptation. Described independently by Balance and Harris over a century ago (in 1903), interest in end-to-side coaptation has been rekindled by Viterbo et al. [17]. In its essence, it involves grafting donor side to recipient end after stimulating donor side collateral sprouting by mechanical trauma or axotomy (**Figure 4(a)**). An indirect application of it is increasing the incidence of nerve regeneration after conventional end-toend grafting by applying additional grafts extending from the side of the donor end to the side of the recipient end [20] (**Figure 4(b)**). In nerve transfer, the latter technique allows the surgeon to use a single high axonal load donor for multiple recipients without producing cocontractions (e.g., major brachial plexus root to several peripheral nerves and caudal cord to cauda equina) [20] (**Figure 4(c)**). Also, partially regenerated nerves cannot be surgically cut and nerve grafted leading to loss of already regained function; the latter technique allows the surgeon to enhance regeneration through them (**Figure 4(d)**). supplying the axonal growth cone with neurotrophic factors that powerits continuedgrowth.coaptation. Harris over <sup>a</sup> century ago interest in end-to-side coaptation has been rekindled Viterbo et al. [17]. In its essence, it involves grafting donor side to recipient end after stimu‐ lating donor side collateral sprouting by mechanical trauma or axotomy (**Figure 4(a)**). An indirect application of it is increasing the incidence of nerve regeneration after conventional end-to-end grafting by applying additional grafts extending from the side of the donor end tothe side of the recipient end [20] (**Figure 4(b)**). In nerve transfer, the latter technique allows the surgeon to use a single high axonal load donor for multiple recipients without producing cocontractions (e.g., major brachial plexus root to several peripheral nerves and caudal cord to cauda equina) [20] (**Figure 4(c)**). Also, partially regenerated nerves cannot be surgically cut and nerve grafted leading to loss of already regained function; the latter technique allows thesurgeon to enhance regeneration through them (**Figure 4(d)**).
**Figure 4.** (a) End-to-side grafting involves grafting donor side to recipient end after stimulating donor side collateral sprouting by mechanical trauma or axotomy. (b) An indirect application of side grafting is increasing the incidence of nerve regeneration after conventional end-to-end grafting by applying additional grafts extending from the side of the donor end to the side of the recipient end. (c) Side grafting allows the surgeon to use a single high axonal load donor for multiple recipients without producing cocontractions. (d) Partially regenerated nerves cannot be surgically cut and nerve grafted sacrifying already regained function; side grafting allows the surgeon to enhance regeneration through them.
#### **2.4. Mathematical modelling and channel-carrying capacity applied to nerve grafting: necessary concepts for subsequent computer-assisted fabrication of artificial nerve grafts grafting:**
Can we manipulate the molecular mechanisms of the sixfold attack to increase neurite outgrowth into the side grafts? Manipulating molecular mechanisms is based on the sensitivity of the axonal growth cone to spatial molecular concentration gradients [21, 22]. In a study by Rosoff et al. [21], axonal growth has been shown to be enhanced by a steep nerve growth factor (NGF) spatial concentration gradient. There is a narrow range, however, between the lowest NGF concentration necessary for axonal growth stimulation and the highest NCF concentra‐ tion beyond which axonal growth is competitively blocked. As the lower and upper limits of the concentration gradient should fall within this narrow range, the maximal distance for axonal growth cone progression guided by that gradient would be far less than the length of the neural defect. Utilising synthetic nerve graft scaffolds and observing the sixfold attack, axonal growth can be hypothetically made to bridge the whole length of the neural gap by seeding the scaffolds with multiple NGF spatial concentration gradients [22]. Neurite out‐ growth has been modelled as a non-linear partial differential equation, that is solved by an iterative mathematical process suitable for numerical analysis and for subsequent computerassisted fabrication of artificial nerve grafts [22]. By diffusion, these NGF spatial concentration gradients might also enhance axonal growth within the adjacent natural nerve side grafts. Can we manipulate the molecular mechanisms of the sixfold attack to increase outgrowth into the side grafts? Manipulating molecular mechanisms is based on the sensitiv‐ ity of the axonal growth cone to spatial molecular concentration gradients [21, 22]. In a study by Rosoff et al. [21], axonal growth has been shown to be enhanced by a steep nerve growth factor (NGF) spatial concentration gradient. There is a narrow range, however, between the lowest NGF concentration necessary for axonal growth stimulation and the highest NCF concentration beyond which axonal growth is competitively blocked. As the lower and upperlimits of the concentration gradient should fall within this narrow range, the maximal distance for axonal growth cone progression guided by that gradient would be far less than the length of the neural defect. Utilising synthetic nerve graft scaffolds and observing the sixfold axonal growth can be hypothetically made to bridge the whole length of the neural gap seeding the scaffolds with multiple NGF spatial concentration gradients [22]. Neurite outgrowth has been modelled as a non-linear partial differential equation, that is solved by suitable assisted fabrication of artificial nerve grafts [22]. By diffusion, these NGF spatial concentra‐ tion gradients might also enhance axonal growth within the adjacent natural nerve side grafts.
Alternatively, and also to increase neurite outgrowth into the side grafts and decrease aberrant neural sprouting, multiple microspheres embedding chemical attractive and repulsive cues and placed along nerve side grafts may be used to guide axonal growth [23]. Preliminary experiments conducted with embryonic rat hippocampal neurons and calcium alginate microspheres have been encouraging [24]. A mathematical model has been developed based on the diffusion gradient of the implanted microspheres; a genetic algorithm has been used to study its proper spatial implementation [23]. and also to increase neurite outgrowth into the side grafts and decrease neural multiple microspheres embedding chemical attractive and repulsive and placed along nerve side grafts may be used to guide axonal experiments conducted with embryonic rat hippocampal neurons and calcium encouraging A has the gradient microspheres; algorithm study its A more accurate mathematical model for axonal growth cone progression has been provid‐
A more accurate mathematical model for axonal growth cone progression has been provided based on sensory pinch test data [25]. Disadvantages of this model, however, include the assumptions that the initial delay is the major cause of variability, and that delay to scarring of the neural bed lies within the initial delay and that the regeneration rate is linear (constant). ed based on sensory pinch test data [25]. Disadvantages of this model, however, include of the neural bed lies within the initial delay and that the regeneration rate is
Can we quantify the molecular mechanisms of the sixfold attack so that nearly 100% of all axons sprouting from the proximal spinal cord reach the distal spinal cord through the side grafts and simultaneously minimise the probability of aberrant neural sprouting to nearly 0%? Unless incorporated in information theory, which is a theory based on mathematical proba‐ bility [26, 27], the mathematical models mentioned above do not provide a numerical solution for this problem. This is imperative, however, for subsequent computer-assisted fabrication of artificial nerve grafts. The Shannon-Hartley channel-carrying capacity principle, a central concept in information theory, refers to the intrinsic property of any information channel to accept all information from the donor and transmit it noiseless to the recipient. Applied to nerve grafting, the channel-carrying capacity of a nerve graft scaffold is its ability to transmit all axons sprouting from the proximal cord to the distal cord and simultaneously minimise the probability of aberrant neural sprouting. Can we quantify the molecular mechanisms of the sixfold attack so that of axons sprouting from the proximal spinal cord reach the distal spinal cord through the grafts and simultaneously minimise the probability of aberrant neural sprouting to nearly 0%?Unless incorporated in information theory, which is <sup>a</sup> theory based on mathematical the mathematical models mentioned above do not provide a numerical for this This is for subsequent computer-assisted fabrication artificial nerve The Shannon-Hartley channel-carrying capacity a concept in information theory, refers to the intrinsic property of any information channel toaccept all information from the donor and transmit it noiseless to the Applied nerve grafting, the channel-carrying capacity of a nerve graft scaffold is its ability to trans‐mit all axons sprouting from the proximal cord to the distal cord and simultaneously mini‐ mise the probability of aberrant neural sprouting.
**Figure 4.** (a) End-to-side grafting involves grafting donor side to recipient end after stimulating donor side collateral sprouting by mechanical trauma or axotomy. (b) An indirect application of side grafting is increasing the incidence of nerve regeneration after conventional end-to-end grafting by applying additional grafts extending from the side of the donor end to the side of the recipient end. (c) Side grafting allows the surgeon to use a single high axonal load donor for multiple recipients without producing cocontractions. (d) Partially regenerated nerves cannot be surgically cut and nerve grafted sacrifying already regained function; side grafting allows the surgeon to enhance regeneration through
them.
6 Recovery of Motor Function in Spinal Cord Injury 186Following Spinal Cord Injury
### **3. Peripheral nerve grafting of the injured spinal cord as an application of the concept of side grafting and the sixfold attack grafting of injured spinal cord as an**
### **3.1. Technical aspects of peripheral nerve grafting and repair of the injured spinal cord**
An indirect application of side grafting, as mentioned previously, is, first, increasing the incidence of nerve regeneration after conventional end-to-end grafting by applying additional grafts extending from the side of the donor end to the side of the recipient end; and, second, preserving partially regenerated nerves which cannot be surgically cut and nerve grafted leading to loss of already regained function [20]. Both of these apply to the spinal cord; compared to its high axonal load, the cross-sectional area of the spinal cord is too small for efficient end-to-end grafting. In addition, some kind of regeneration may have occurred in a contused cord in a completely paraplegic patient; this may take the form of less than Grade 3 motor power improvement, which, according to ASIA standards, is not enough to be consid‐ ered motor or neurological level progression [28]. Third, side grafting produces less trauma to the spinal cord. In fact, the glial tissue secreted by astrocytes provides the necessary supporting tissue for axons and neurons [2, 29]. In side grafting, fine pial incisions are used. This process is far less traumatic than freshening of the cord ends during end-to-end grafting, a procedure which would lead to excessive glial tissue secretion and subsequent blocking of regeneration. An indirect application of side as mentioned increasing incidence of nerve regeneration after conventional end-to-end grafting by applying addition‐ al grafts extending from the side of the donor end to the side of the recipient end; and, preserving partially regenerated nerves which cannot be surgically cut and nerve leading to loss of already regained Both of these apply to the spinal compared to its high axonal the cross-sectional area of the spinal cord is too small Grade ered motor or neurological level side grafting produces less to the spinal In the glial tissue secreted by astrocytes provides the and This process is far less traumatic than freshening of the cord ends during end-to-end graft‐ing, a procedure which would lead to excessive glial tissue secretion and subsequent block‐ ing of regeneration. outgrowth promoting and neurotrophic the use of 188Following Spinal Cord Injury
### **3.2. Donor nerves**
Clinically, the sural nerve is the most commonly used donor nerve, other suitable donor nerves include the medial and lateral cutaneous nerves of the forearm, dorsal cutaneous branch of the ulnar nerve, superficial and deep peroneal nerves, intercostal nerves, and the posterior and lateral cutaneous nerves of the thigh [17]. the sural nerve is the most commonly used donor other suitable donor nervesinclude the medial and lateral cutaneous nerves of the dorsal cutaneous branch the ulnar superficial and deep peroneal intercostal and the posterior
Pre-degenerated (segments of nerve cut and left in situ for 7–10 days prior to harvesting) peripheral nerves are infiltrated by regenerating axons to a greater extent (both in number and distance) than freshly cut and harvested nerves [30]. The use of pre-degenerated nerves has been contested by other authors, however [31]. nerves of the (segments of nerve cut and left in situ 7–10 days prior to harvesting) peripheral nerves are infiltrated by regenerating axons to a greater (both in number than freshly cut and harvested The use of pre-degenerated nerves
### **3.3. Experimental applications of side grafting**
Based on the work of David and Aguayo [5, 7], numerous studies have confirmed axonal outgrowth after nerve grafting of the injured spinal cord [30, 32–38]; axons growing within peripheral nerve grafts have not only retained their physiological properties [39] but have synapsed with neurons near the point of central nervous system re-entry as well [40]. Based on the work of David and numerous studies have confirmed outgrowth after nerve grafting of the injured spinal axons growing peripheral nerve grafts have not only retained their physiological but
### **3.4. Re-evaluation of the use of peripheral nerve grafts to bridge spinal cord defects**
Nerve grafts supply the injured spinal cord with five factors of the sixfold attack: a suitable scaffold, on which the basal lamina can settle; basal laminae; cell adhesion molecules; neurite outgrowth promoting factors; and neurotrophic factors. Nevertheless, the use of peripheral Nerve grafts supply the injured spinal cord with five factors of the sixfold a
nerve grafts has been challenged [5, 16, 35, 41, 42]. According to experimental observation, damaged spinal cord axons might grow from the cranial cord into peripheral nerve grafts but would not leave them to enter the caudal cord [5]. Schwann cells might even promote gliosis [16, 41]. nerve grafts has been According to experimental
**3. Peripheral nerve grafting of the injured spinal cord as an application of**
**3.1. Technical aspects of peripheral nerve grafting and repair of the injured spinal cord**
An indirect application of side as mentioned increasing incidence of nerve regeneration after conventional end-to-end grafting by applying addition‐
al grafts extending from the side of the donor end to the side of the recipient end; and, preserving partially regenerated nerves which cannot be surgically cut and nerve leading to loss of already regained Both of these apply to the spinal compared to its high axonal the cross-sectional area of the spinal cord is too small
ered motor or neurological level side grafting produces less to the spinal In the glial tissue secreted by astrocytes provides the
lamina outgrowth promoting and neurotrophic the use of 188Following Spinal Cord Injury
An indirect application of side grafting, as mentioned previously, is, first, increasing the incidence of nerve regeneration after conventional end-to-end grafting by applying additional grafts extending from the side of the donor end to the side of the recipient end; and, second, preserving partially regenerated nerves which cannot be surgically cut and nerve grafted leading to loss of already regained function [20]. Both of these apply to the spinal cord; compared to its high axonal load, the cross-sectional area of the spinal cord is too small for efficient end-to-end grafting. In addition, some kind of regeneration may have occurred in a contused cord in a completely paraplegic patient; this may take the form of less than Grade 3 motor power improvement, which, according to ASIA standards, is not enough to be consid‐ ered motor or neurological level progression [28]. Third, side grafting produces less trauma to the spinal cord. In fact, the glial tissue secreted by astrocytes provides the necessary supporting tissue for axons and neurons [2, 29]. In side grafting, fine pial incisions are used. This process is far less traumatic than freshening of the cord ends during end-to-end grafting, a procedure which would lead to excessive glial tissue secretion and subsequent blocking of
enough
Clinically, the sural nerve is the most commonly used donor nerve, other suitable donor nerves include the medial and lateral cutaneous nerves of the forearm, dorsal cutaneous branch of the ulnar nerve, superficial and deep peroneal nerves, intercostal nerves, and the posterior and
include the medial and lateral cutaneous nerves of the dorsal cutaneous branch the ulnar superficial and deep peroneal nerves, intercostal nerves, and the posterior
the sural nerve is the most commonly used donor other suitable donor
29]. grafting, pial This process is far less traumatic than freshening of the cord ends during end-to-end graft‐ ing, a procedure which would lead to excessive glial tissue secretion and subsequent block‐
Pre-degenerated (segments of nerve cut and left in situ for 7–10 days prior to harvesting) peripheral nerves are infiltrated by regenerating axons to a greater extent (both in number and distance) than freshly cut and harvested nerves [30]. The use of pre-degenerated nerves has
peripheral nerves are infiltrated by regenerating axons to a greater (both in number
(segments of nerve cut and left in situ for days prior to
than freshly cut and harvested The use of pre-degenerated nerves
Based on the work of David and Aguayo [5, 7], numerous studies have confirmed axonal outgrowth after nerve grafting of the injured spinal cord [30, 32–38]; axons growing within peripheral nerve grafts have not only retained their physiological properties [39] but have
Nerve grafts supply the injured spinal cord with five factors of the sixfold attack: a suitable scaffold, on which the basal lamina can settle; basal laminae; cell adhesion molecules; neurite outgrowth promoting factors; and neurotrophic factors. Nevertheless, the use of peripheral
synapsed with neurons near the point of central nervous system re-entry as well [40].
Based on the work of David and numerous studies have confirmed outgrowth after nerve grafting of the injured spinal axons growing peripheral nerve grafts have not only retained their physiological but
**3.4. Re-evaluation of the use of peripheral nerve grafts to bridge spinal cord defects**
Nerve grafts supply the injured spinal cord with five factors of the sixfold a
**the concept of side grafting and the sixfold attack**
8 Recovery of Motor Function in Spinal Cord Injury
regeneration.
**3.2. Donor**
**3.2. Donor nerves**
ing of regeneration.
lateral cutaneous nerves of the thigh [17].
nerves of thigh [17].
been contested by other authors, however [31].
**3.3. Experimental applications of side grafting**
**Figure 5.** (a) Excessive fibrosis in the presence of neurotophic factors and adequate neuron growth potential leads to hypertrophy of both donor and recipient ends of the damaged neural tissue (excessive neuroma formation; dystrophic growth cones). (b) Excessive fibrosis in the absence of neurotophic factors and adequate neuron growth potential leads to atrophy of both donor and recipient ends of the damaged neural tissue (atrophic neuroma formation; atrophic growth cones). Excessive fibrosis in the presence of neurotophic factors and adequate neuron growth potential leads (excessive to atrophy of both donor and recipient ends of the damaged neural (atrophic neuroma
It might be assumed that failure of growth cone progression is due to missing of the sixth factor in the sixfold attack (lysing the fibrosis/gliosis to an extent that allows settling of the basal lamina preventing meanwhile collapse of the neural tissue matrix). However, failure of axons to regenerate following peripheral nerve grafting may also result from intrinsic properties of the neurons and the absence of neurotrophic factors [43–46]. These considerations help us assess neural tissue during surgery. Excessive fibrosis in the presence of neurotophic factors and adequate neuron growth potential leads to hypertrophy of both donor and recipient ends of the damaged neural tissue (excessive neuroma formation and dystrophic growth cones) [47, 48] (**Figure 5(a)**), whereas excessive fibrosis in the absence of neurotophic factors and adequate neuron growth potential leads to atrophy of both donor and recipient ends of the damaged neural tissue (atrophic neuroma formation; atrophic growth cones) (**Figure 5(b)**). In conclu‐ It might be assumed that failure of growth cone progression is due to missing of the sixth in the sixfold (lysing the fibrosis/gliosis to an extent that allows settling of the lamina failure nerve grafting may the neurons and the absence of neurotrophic [43–46]. These considerations help in the and adequate neuron growth potential leads to hypertrophy of both donor and recipient of the damaged neural (excessive neuroma formation and dystrophic growth whereas excessive fibrosis in the absence of neurotophic factors and adequateneuron growth potential leads to atrophy of both donor and recipient ends of the sion, in addition to the sixfold attack, the intrinsic properties of the neurons should be stimulated to produce neurites. in addition to the sixfold the intrinsic properties of the neurons should
### **4. Lysing the gliosis in conjunction with nerve grafting**
### **4.1. The gliosis: blocking of regeneration mechanically and through the activation of RhoA**
Injured axons encounter a series of inhibitory factors that are non-permissive for growth [4, 29, 41, 49]. These include myelin inhibitors (Nogo-A ,MAG108 (myelin-associated glycopro‐ tein) and OMgp109 (oligodendrocyte myelin glycoprotein)); chondroitin sulphate proteogly‐ cans (neurocan, versican, aggrecan, brevican, phosphacan and NG2); and semaphorins and ephrins. Upregulated in response to spinal cord injury [50–52], they act not only by mechanical blocking of neural regeneration but by inhibiting axon outgrowth neuronal receptors and subsequent activation of signalling pathways known to be involved in the activation of RhoA and the rise in intracellular calcium. Injured axons encounter a series of inhibitory factors that are non-permissive for 49]. These include myelin inhibitors (Nogo-A ,MAG108 (myelin-associated glycoprotein)); chondroitin sulphate versican, phosphacan and and semaphorins ephrins. Upregulated in response to spinal cord injury [50–52], they act not only by mechani‐cal blocking of neural regeneration but by inhibiting axon outgrowth neuronal receptors signalling in the Chondroitinase ABC cleaves the inhibitory chondroitin sulphate glycosaminoglycan
### **4.2. Lysing the gliosis by chondroitinase ABC**
Chondroitinase ABC cleaves the inhibitory chondroitin sulphate glycosaminoglycan chains from the core protein, reducing the inhibition by chondroitin sulphate proteoglycans [53, 54]. Houle et al. [55] have demonstrated CNS axons regenerating through a peripheral nerve graft and entering the caudal spinal cord following chondroitinase ABC treatment. the core sulphate demonstrated regenerating a peripheral nerve
#### *4.2.1. Chondroitinase ABC in combination with growth factors and scaffolds ABC in with growth scaffolds*
Combined with glial-derived neurotrophic factor (GDNF) delivery, chondroitinase ABC promotes axon extension through peripheral nerve bridges [50, 56]. Combined with growth factors, chondroitinase ABC enhances the activation and oligodendrocyte differentiation of endogenous precursor cells after spinal cord injury and attenuates astrogliosis. When added to polycaprolactone or poly (acrylonitrile)/poly(vinyl chloride) (PAN/PVC) scaffolds, chon‐ droitinase ABC allows regenerating axons to exit the distal end of the scaffold and continue on to distal targets [57–60]. Combined with glial-derived neurotrophic (GDNF) delivery, chondroitinase promotes axon extension through peripheral nerve Combined with chondroitinase ABC enhances the activation and oligodendrocyte differentiation spinal astrogliosis. addedto polycaprolactone or poly (acrylonitrile)/poly(vinyl chloride) (PAN/PVC) scaffolds, chondroitinase ABC allows regenerating axons to exit the distal end of the scaffold and continue on to distal targets [57–60].
#### *4.2.2. Chondroitinase ABC in combination with cell therapies ABC in with cell*
Chondroitinase ABC has improved recovery of function in synergy with mesenchymal stromal cells [D43], or a Schwann cell bridge and olfactory ensheathing cells or in combination with transplanted neural precursor cells and a growth factor (GF) cocktail containing EGF, FGF2 and platelet-derived growth factor (PDGF)-AA [61, 62]. Chondroitinase ABC has improved recovery offunction in synergy with mesenchymal or a Schwann cell bridge and olfactory ensheathing cells or in combination transplanted neural precursor cells and a growth cocktail containing (PDGF)-AA [61,
#### *4.2.3. Thermostabilisation, delivery, dosage and complications of chondroitinase ABC: chondroitinase ABC might be a weak enzyme delivery, dosage ABC might*
Thermostabilisation of chondroitinase ABC with the sugar trehalose can reduce its tempera‐ ture-dependent loss of activity [50, 63]. Delivery of chondroitinase ABC is predominantly ABC ture-dependent loss of Delivery of chondroitinase ABC is
intrathecal using osmotic minipumps [30, 55, 64]. Nevertheless, the enzyme can also be loaded into lipid microtubes or possibly poly (lactic-co-glycolic acid) (PLGA), providing a means for its gradual release over 1–2 weeks [65]. It can also be loaded into fibrin gel scaffolds before injecting it intrathecally; this ensures its continuous release for at least 3 weeks [66]. intrathecal using osmotic the enzyme can also be its gradual release It can also be loaded into fibrin gel scaffolds
The effect of chondroitinase ABC is dose-dependent [67]. Injected intrathecally in acute injuries at a low dose (1 or 5 IUs), chondroitinase ABC may enhance axonal progression; injected at a high dose (50 IU), it may produce subarachnoid haemorrhage. In subacute or chronic injuries, low-dose injection produces no or limited functional recovery, whilst high-dose injection (50 or 100 IUs) produce lysis of the gliosis [68, 69]. The effect of chondroitinase ABC is Injected intrathecally in acute high it may produce subarachnoid In subacute or chronic
Single-dose chondroitinase ABC is not considered enough. Therefore, multiple single injec‐ tions at 0, 1, 2 and 4 weeks have been recommended [70]. Daily injections for 2 weeks at 0.06 Units per dose have also been recommended [71]. Four weeks of treatment have promoted recovery more than 2 weeks [72]. One way to ensure the continued release of chondroitinase ABC is neuron transfection with a vector containing the gene encoding chondroitinase ABC [73]. Another way is loading scaffolds with chondroitinase ABC. Single-dose chondroitinase ABC is not considered enough. Therefore, multiple single injections at 0, 1, 2 and 4 weeks have been recommended [70]. Daily injections for 2 weeks at 0.06 Units per dose have also been recommended [71]. Four weeks of treatment have promot‐ ed recovery more than 2 weeks [72]. One way to ensure the continued release of chondroiti‐ nase ABC is neuron transfection with a vector containing the gene encoding chondroitinase ABC [73]. Another way is loading scaffolds with chondroitinase ABC.
Although chondroitinase ABC allows substantial structural plasticity in the spinal cord, it is not sufficient to enhance locomotor recovery unless combined with neural precursor cell transplantation and in vivo infusion of growth factors [74]. In a rat spinal cord injury model, Wilems et al. [62] have used fibrin scaffolds loaded with neurotrophic factors (item I), antiinhibitory molecules (item II) and encapsulated embryonic stem-cell-derived progenitor motor neurons (pMNs) (item III). Fibrin scaffolds containing items I and II but not item III have had lower chondroitin sulphate proteoglycan levels compared to scaffolds containing items II and III. This shows the importance of combining neurotrophic factors with chondroitinase ABC and cellular transplants. Scaffolds containing item III, but not item II, have shown differentia‐ tion into neuronal cell types, axonal extension and the ability to integrate into host tissue. However, the combination of items II and III have led to reduced cell survival and increased macrophage infiltration. This shows that cellular transplants not only claim priority over chondroitinase ABC, but that chondroitinase ABC may be a weak enzyme as well. Because of this fact, a combination treatment of zymosan and chondroitinase ABC has been recommended [75]. Lastly, adipose-derived stem cell transplantation has been found to produce the same effect as chondroitinase ABC administration [76]. Thus, chondroitinase ABC might be a weak enzyme. not sufficient to enhance locomotor recovery unless combined with neural precursor Wilems et have used fibrin scaffolds loaded with neurotrophic factors (item inhibitory (item and encapsulated embryonic stem-cell-derived progenitor (item lower chondroitin sulphate proteoglycan levels compared to scaffolds containing items II This shows the importance of combining neurotrophic factors with chondroitinase II, tion into neuronal cell axonal extension and the ability to integrate into host tissue.items and reduced cell and macrophage This shows that cellular transplants not only claim priority this fact, a combination treatment of zymosan and chondroitinase ABC has been recommend‐ ed [75]. Lastly, adipose-derived stem cell transplantation has been found to produce the ABC [76]. ABC In <sup>a</sup> rat model of spinal cord contusion injury the effects of sialidase (Vibrio
### **4.3. Sialidase as an alternative**
sion, in addition to the sixfold attack, the intrinsic properties of the neurons should be
in addition to the sixfold the intrinsic properties of the neurons should
**4.1. The gliosis: blocking of regeneration mechanically and through the activation of RhoA** Injured axons encounter a series of inhibitory factors that are non-permissive for growth [4, 29, 41, 49]. These include myelin inhibitors (Nogo-A ,MAG108 (myelin-associated glycopro‐ tein) and OMgp109 (oligodendrocyte myelin glycoprotein)); chondroitin sulphate proteogly‐ cans (neurocan, versican, aggrecan, brevican, phosphacan and NG2); and semaphorins and ephrins. Upregulated in response to spinal cord injury [50–52], they act not only by mechanical blocking of neural regeneration but by inhibiting axon outgrowth neuronal receptors and subsequent activation of signalling pathways known to be involved in the activation of RhoA
ephrins. Upregulated in response to spinal cord injury [50–52], they act not only by mechani‐
cal blocking of neural regeneration but by inhibiting axon outgrowth neuronal receptors of involved RhoArise in
phosphacan and and semaphorins
Injured axons encounter a series of inhibitory factors that are non-permissive for 41, 49]. These include myelin inhibitors (Nogo-A ,MAG108 (myelin-associated
Chondroitinase ABC cleaves the inhibitory chondroitin sulphate glycosaminoglycan chains from the core protein, reducing the inhibition by chondroitin sulphate proteoglycans [53, 54]. Houle et al. [55] have demonstrated CNS axons regenerating through a peripheral nerve graft
Chondroitinase ABC cleaves the inhibitory chondroitin sulphate glycosaminoglycan
Combined with glial-derived neurotrophic factor (GDNF) delivery, chondroitinase ABC promotes axon extension through peripheral nerve bridges [50, 56]. Combined with growth factors, chondroitinase ABC enhances the activation and oligodendrocyte differentiation of endogenous precursor cells after spinal cord injury and attenuates astrogliosis. When added to polycaprolactone or poly (acrylonitrile)/poly(vinyl chloride) (PAN/PVC) scaffolds, chon‐ droitinase ABC allows regenerating axons to exit the distal end of the scaffold and continue
cord attenuates to polycaprolactone or poly (acrylonitrile)/poly(vinyl chloride) (PAN/PVC) scaffolds, chondroitinase ABC allows regenerating axons to exit the distal end of the scaffold and
chondroitinase ABC enhances the activation and oligodendrocyte differentiation
reducing the by chondroitin [53, Houle have axons through the caudal following chondroitinase Combined with glial-derived neurotrophic chondroitinase
promotes axon extension through peripheral nerve bridges Combined with
Chondroitinase ABC has improved recovery of function in synergy with mesenchymal stromal cells [D43], or a Schwann cell bridge and olfactory ensheathing cells or in combination with transplanted neural precursor cells and a growth factor (GF) cocktail containing EGF, FGF2
Chondroitinase ABC has improved recovery offunction in synergy with mesenchymal or a Schwann cell bridge and olfactory ensheathing cells or in combination
transplanted neural precursor cells and a growth cocktail containing
*cell therapies*
*4.2.3. Thermostabilisation, delivery, dosage and complications of chondroitinase ABC: chondroitinase*
Thermostabilisation of chondroitinase ABC with the sugar trehalose can reduce its tempera‐ ture-dependent loss of activity [50, 63]. Delivery of chondroitinase ABC is predominantly
and entering the caudal spinal cord following chondroitinase ABC treatment.
*4.2.1. Chondroitinase ABC in combination with growth factors and scaffolds*
ture-dependent loss of Delivery of chondroitinase ABC is 190Following Spinal Cord Injury
*4.2.2. Chondroitinase ABC in combination with cell therapies*
and platelet-derived growth factor (PDGF)-AA [61, 62].
Thermostabilisation of the
**4. Lysing the gliosis in conjunction with nerve grafting**
stimulated to produce neurites.
10 Recovery of Motor Function in Spinal Cord Injury
and the rise in intracellular calcium.
**4.2. Lysing the**
on to distal targets [57–60].
continue on to distal targets [57–60].
*ABC might be a weak enzyme*
*a weak*
**4.2. Lysing the gliosis by chondroitinase ABC**
In a rat model of spinal cord contusion injury the effects of sialidase (Vibrio cholerae) and chondroitinase ABC (ChABC, Proteus vulgaris) have been tested [77]. Immunohistochemistry has revealed that infused sialidase has acted robustly throughout the spinal cord grey and white matter, whereas ChABC activity has been more intense superficially. Sialidase treatment alone has resulted in improved behavioural and anatomical outcomes. chondroitinase ABC (ChABC, Proteus vulgaris) have been tested [77]. Immunohistochemis‐ try has revealed that infused sialidase has acted robustly throughout the spinal cord grey white matter, whereas ChABC activity has been more intense superficially. Sialidase treat‐ment alone has resulted in improved behavioural and anatomical outcomes.
### **4.4. Anti-Nogo**
Blocking myelin-associated inhibitors with Nogo-A monoclonal antibodies or with Nogoreceptor competitive agonist peptide, NEP1-40 has been shown to increase axonal regeneration [50]. Combination therapies, such as cross-linking the Nogo-66 receptor antibody into a hyaluronic acid hydrogel [50], a combination of methylprednisolone and NEP1-40 and Nogoreceptor vaccination combined with neural stem cell transplantation have improved neural fibre regeneration. Blocking myelin-associated inhibitors with Nogo-A monoclonal antibodies or with Nogoreceptor competitive agonist peptide, NEP1-40 has been shown to increase axonal regenera‐ tion [50]. Combination therapies, such as cross-linking the Nogo-66 receptor antibody into a hyaluronic acid hydrogel [50], a combination of methylprednisolone and NEP1-40 and Nogoreceptor vaccination combined with neural stem cell transplantation have improved neural
#### **4.5. Rho inhibition** fibre regeneration.
Many of the inhibitory signals described above (ephrins, Nogo) converge on the intracellular molecule Rho-A, which is a key mediator of actin depolymerisation and hence inhibition of axonal elongation. Blocking Rho-A with Rho inhibitor 'cethrin' might overcome its effect. A synthetic membrane-permeable peptide mimetic of the protein tyrosine phosphatase σ (PTPσ), wedge domain can bind to PTPσ and relieve chondroitin sulphate proteoglycan-mediated inhibition [78]. **4.5. Rho inhibition** Many of the inhibitory signals described above (ephrins, Nogo) converge on the intracellu‐ lar molecule Rho-A, which is a key mediator of actin depolymerisation and hence inhibition of axonal elongation. Blocking Rho-A with Rho inhibitor 'cethrin' might overcome its effect. A synthetic membrane-permeable peptide mimetic of the protein tyrosine phosphatase σ
#### **4.6. Reversing the inhibition of phosphoinositide 3-kinase (PI3K) by cell permeable phosphopeptide (PI3Kpep)** (PTPσ), wedge domain can bind to PTPσ and relieve chondroitin sulphate proteoglycanmediated inhibition [78].
Phosphoinositide 3-kinase (PI3K) is a lipid kinase activated by axon growth promoting signals. Chondroitin sulphate proteoglycans inhibit phosphoinositide 3-kinase signalling in axons and growth cones, an effect that can be reversed by cell permeable phosphopeptide (PI3Kpep). The latter acts by R-Ras-PI3K signalling [79]. **4.6. Reversing the inhibition of phosphoinositide 3-kinase (PI3K) by cell permeable phosphopeptide (PI3Kpep)** Phosphoinositide 3-kinase (PI3K) is a lipid kinase activated by axon growth promoting signals.
#### **4.7. Rolipram** Chondroitin sulphate proteoglycans inhibit phosphoinositide 3-kinase signalling in axons and
Increased intracellular levels of cyclic adenosine monophosphate (cAMP) and protein kinase A have been associated with CNS ability to overcome the gliosis [50]. Rolipram, a phospho‐ diesterase4 inhibitor, can increase intracellular cAMP levels [50]. growth cones, an effect that can be reversed by cell permeable phosphopeptide (PI3Kpep). The latter acts by R-Ras-PI3K signalling [79]. **4.7. Rolipram**
#### **4.8. Improving blood vessel formation** Increased intracellular levels of cyclic adenosine monophosphate (cAMP) and protein kinase
Improving blood vessel formation might reduce cell death and promote angiogenesis within the injury zone. Neural stem cells modified to express vascular endothelial growth factor have improved white matter sparing following thoracic contusion spinal cord injury [50]. Bioma‐ terial poly-lactic-co-glycolic acid (PLGA) scaffolds loaded with neural stem cells and endo‐ thelial cells have shown increased vessel and neurofilament density at the injury centre [50]. A have been associated with CNS ability to overcome the gliosis [50]. Rolipram, a phospho‐ diesterase4 inhibitor, can increase intracellular cAMP levels [50]. **4.8. Improving blood vessel formation**
Improving blood vessel formation might reduce cell death and promote angiogenesis within
#### **5. Modulating astrocyte function enhances the intrinsic properties of neurons to stimulate neurite outgrowth into peripheral nerve grafts** the injury zone. Neural stem cells modified to express vascular endothelial growth factor have improved white matter sparing following thoracic contusion spinal cord injury [50]. Biomaterial poly-lactic-co-glycolic acid (PLGA) scaffolds loaded with neural stem cells and
In addition to the sixfold attack, the intrinsic properties of the neurons have to be stimulated to produce neurites. This can take place by modulating astrocyte function (**Figure 2**). endothelial cells have shown increased vessel and neurofilament density at the injury centre [50].
#### **5.1. Endogenous inhibitors of axonal regeneration (intrinsic properties of neurons) 5. Modulating astrocyte function enhances the intrinsic properties of**
**4.4. Anti-Nogo**
12 Recovery of Motor Function in Spinal Cord Injury
192Following Spinal Cord Injury
fibre regeneration.
fibre regeneration.
**4.5. Rho inhibition**
**4.5. Rho inhibition**
inhibition [78].
**4.7. Rolipram**
**4.7. Rolipram**
[50].
**phosphopeptide (PI3Kpep)**
**phosphopeptide (PI3Kpep)**
mediated inhibition [78].
latter acts by R-Ras-PI3K signalling [79].
latter acts by R-Ras-PI3K signalling [79].
**4.8. Improving blood vessel formation**
**4.8. Improving blood vessel formation**
Blocking myelin-associated inhibitors with Nogo-A monoclonal antibodies or with Nogoreceptor competitive agonist peptide, NEP1-40 has been shown to increase axonal regeneration [50]. Combination therapies, such as cross-linking the Nogo-66 receptor antibody into a hyaluronic acid hydrogel [50], a combination of methylprednisolone and NEP1-40 and Nogoreceptor vaccination combined with neural stem cell transplantation have improved neural
Blocking myelin-associated inhibitors with Nogo-A monoclonal antibodies or with Nogoreceptor competitive agonist peptide, NEP1-40 has been shown to increase axonal regenera‐ tion [50]. Combination therapies, such as cross-linking the Nogo-66 receptor antibody into a hyaluronic acid hydrogel [50], a combination of methylprednisolone and NEP1-40 and Nogoreceptor vaccination combined with neural stem cell transplantation have improved neural
Many of the inhibitory signals described above (ephrins, Nogo) converge on the intracellular molecule Rho-A, which is a key mediator of actin depolymerisation and hence inhibition of axonal elongation. Blocking Rho-A with Rho inhibitor 'cethrin' might overcome its effect. A synthetic membrane-permeable peptide mimetic of the protein tyrosine phosphatase σ (PTPσ), wedge domain can bind to PTPσ and relieve chondroitin sulphate proteoglycan-mediated
Many of the inhibitory signals described above (ephrins, Nogo) converge on the intracellu‐ lar molecule Rho-A, which is a key mediator of actin depolymerisation and hence inhibition of axonal elongation. Blocking Rho-A with Rho inhibitor 'cethrin' might overcome its effect. A synthetic membrane-permeable peptide mimetic of the protein tyrosine phosphatase σ (PTPσ), wedge domain can bind to PTPσ and relieve chondroitin sulphate proteoglycan-
**4.6. Reversing the inhibition of phosphoinositide 3-kinase (PI3K) by cell permeable**
**4.6. Reversing the inhibition of phosphoinositide 3-kinase (PI3K) by cell permeable**
Phosphoinositide 3-kinase (PI3K) is a lipid kinase activated by axon growth promoting signals. Chondroitin sulphate proteoglycans inhibit phosphoinositide 3-kinase signalling in axons and growth cones, an effect that can be reversed by cell permeable phosphopeptide (PI3Kpep). The
Increased intracellular levels of cyclic adenosine monophosphate (cAMP) and protein kinase A have been associated with CNS ability to overcome the gliosis [50]. Rolipram, a phospho‐
Phosphoinositide 3-kinase (PI3K) is a lipid kinase activated by axon growth promoting signals. Chondroitin sulphate proteoglycans inhibit phosphoinositide 3-kinase signalling in axons and growth cones, an effect that can be reversed by cell permeable phosphopeptide (PI3Kpep). The
Improving blood vessel formation might reduce cell death and promote angiogenesis within the injury zone. Neural stem cells modified to express vascular endothelial growth factor have improved white matter sparing following thoracic contusion spinal cord injury [50]. Bioma‐ terial poly-lactic-co-glycolic acid (PLGA) scaffolds loaded with neural stem cells and endo‐ thelial cells have shown increased vessel and neurofilament density at the injury centre [50].
Increased intracellular levels of cyclic adenosine monophosphate (cAMP) and protein kinase A have been associated with CNS ability to overcome the gliosis [50]. Rolipram, a phospho‐
**5. Modulating astrocyte function enhances the intrinsic properties of neurons to stimulate neurite outgrowth into peripheral nerve grafts**
to produce neurites. This can take place by modulating astrocyte function (**Figure 2**).
In addition to the sixfold attack, the intrinsic properties of the neurons have to be stimulated
Improving blood vessel formation might reduce cell death and promote angiogenesis within the injury zone. Neural stem cells modified to express vascular endothelial growth factor have improved white matter sparing following thoracic contusion spinal cord injury [50]. Biomaterial poly-lactic-co-glycolic acid (PLGA) scaffolds loaded with neural stem cells and endothelial cells have shown increased vessel and neurofilament density at the injury centre
diesterase4 inhibitor, can increase intracellular cAMP levels [50].
diesterase4 inhibitor, can increase intracellular cAMP levels [50].
Endogenous inhibitors of axonal regeneration include the molecule phosphatase and tensin homologue (PTEN), loss of neuronal cAMP and deactivation/activation of certain transcription factors [46]. **neurons to stimulate neurite outgrowth into peripheral nerve grafts** In addition to the sixfold attack, the intrinsic properties of the neurons have to be stimulated
The molecule phosphatase and tensin homologue (PTEN) on chromosome 10 is a tumour suppressor. Its deletion has been shown to increase post-embryonic neural regeneration after injury. Its inhibition–mediated regeneration is partly mediated by the inhibitor of the mecha‐ nistic target of rapamycin (mTOR) pathway; it is also mediated by glycogen synthesis kinase GSK-3β. to produce neurites. This can take place by modulating astrocyte function (**Figure 2**). **5.1. Endogenous inhibitors of axonal regeneration (intrinsic properties of neurons)** Endogenous inhibitors of axonal regeneration include the molecule phosphatase and tensin homologue (PTEN), loss of neuronal cAMP and deactivation/activation of certain transcrip‐
When levels of cAMP at the growth cone are high, the effect on the growth cone is chemoat‐ traction, whereas when they are low, the effect is chemorepulsion. tion factors [46]. The molecule phosphatase and tensin homologue (PTEN) on chromosome 10 is a tumour
Certain transcription factors are positive regulators of axonal growth (e.g., members of the Krüppel-like factors (KLFs) present in retinal ganglion cells (RGCs); STAT3 a transcription factor, part of the JAK-STAT signaling pathway; members of the Jun and Fos families, components of the transcription factor AP-1 and ATF3). Other transcription factors are negative regulators of axonal growth. Nuclear factor IL-3 (NFIL3) represses CREB-mediated transcription and expression of regeneration-associated genes such as arginase and GAP-43. suppressor. Its deletion has been shown to increase post-embryonic neural regeneration after injury. Its inhibition–mediated regeneration is partly mediated by the inhibitor of the mechanistic target of rapamycin (mTOR) pathway; it is also mediated by glycogen synthesis kinase GSK-3β. When levels of cAMP at the growth cone are high, the effect on the growth cone is chemoat‐ traction, whereas when they are low, the effect is chemorepulsion.
It follows that combatting endogenous inhibitors of axonal regeneration include the following: Certain transcription factors are positive regulators of axonal growth (e.g., members of the
– Inactivation of GSK-3β by neurotrophins. This increases collapsin response mediator protein-2 (CRMP-2) stabilisation of microtubules and increases axon elongation in developing neurites. Krüppel-like factors (KLFs) present in retinal ganglion cells (RGCs); STAT3 a transcription factor, part of the JAK-STAT signaling pathway; members of the Jun and Fos families, components of the transcription factor AP-1 and ATF3). Other transcription factors are
– Local administration of taxol. Microtubules and actin microfilaments are critical for regen‐ eration [80]. They potentiate the effect of GAP-43. Thus, local administration of taxol, a microtubule-stabilising agent, increases neurite outgrowth [81]. negative regulators of axonal growth. Nuclear factor IL-3 (NFIL3) represses CREB-mediated transcription and expression of regeneration-associated genes such as arginase and GAP-43. It follows that combatting endogenous inhibitors of axonalregeneration include the following:
– Elevating cAMP levels by local injection of a phosphodiesterase inhibitor. This improves axonal regeneration. Inactivation of GSK-3β by neurotrophins. This increases collapsin response mediatorprotein-2 (CRMP-2) stabilisation of microtubules and increases axon elongation in develop‐
– Conditioning lesions. Conditioning lesions [46] are based on the observation that double level nerve lesions regenerate better than single level nerve lesions. Thus sciatic nerve transection prior to a spinal cord lesion improves regeneration within the injured spinal cord. ing neurites. – Local administration of taxol. Microtubules and actin microfilaments are critical for regen‐ eration [80]. They potentiate the effect of GAP-43. Thus, local administration of taxol, a
– Cell adhesion molecules. Their synthesis is increased after peripheral nerve injury but not after CNS injury [82, 83]. Expression of cell adhesion molecules by neurons can be induced by virally mediated vectors or by injecting them into the CNS injury site [84, 85]. microtubule-stabilising agent, increases neurite outgrowth [81]. – Elevating cAMP levels by local injection of a phosphodiesterase inhibitor. This improves axonal regeneration.
Many of these functions can be activated by modulating astrocyte function. – Conditioning lesions. Conditioning lesions [46] are based on the observation that double
#### **5.2. Astrocyte trophic effects on neurite outgrowth and neurogenesis** transection prior to a spinal cord lesion improves regeneration within the injured spinal cord.
Astrocytes release a variety of trophic factors [86–88]. These trophic factors include nerve growth factor, basic fibroblast growth factor, transforming growth factor-β, platelet-derived growth factor, brain-derived neurotrophic factor, ciliary neurotrophic factor and others. – Cell adhesion molecules. Their synthesis is increased after peripheral nerve injury but not after CNS injury [82, 83]. Expression of cell adhesion molecules by neurons can be induced by virally mediated vectors or by injecting them into the CNS injury site [84, 85].
level nerve lesions regenerate better than single level nerve lesions. Thus sciatic nerve
Reactive astrocytes increase the expression of several of these, notably nerve growth factor, basic fibroblast growth factor, brain-derived neurotrophic factor and neuregulins, which can stimulate neurite outgrowth. Reactive astrocytes also overexpress neuropilin-1 and vascular endothelial growth factor, which act in concert to promote angiogenesis after cerebral ische‐ mia. Hevin (SPARC-like protein 1), a synaptogenic protein released by astrocytes, forms a relay between a neurexin (pre-synaptic) and a neuroligin (post-synaptic); this relay is crucial for the synaptogenesis [89]. Many of these functions can be activated by modulating astrocyte function. **5.2. Astrocyte trophic effects on neurite outgrowth and neurogenesis** Astrocytes release a variety of trophic factors [86–88]. These trophic factors include nerve growth factor, basic fibroblast growth factor, transforming growth factor-β, platelet-derived growth factor, brain-derived neurotrophic factor, ciliary neurotrophic factor and others. Reactive astrocytes increase the expression of several of these, notably nerve growth factor,
#### **5.3. Modulating the function of astrocytes and heparin** basic fibroblast growth factor, brain-derived neurotrophic factor and neuregulins, which can stimulate neurite outgrowth. Reactive astrocytes also overexpress neuropilin-1 and vascular
#### *5.3.1. Role of heparin in lysing the gliosis* endothelial growth factor, which act in concert to promote angiogenesis after cerebral ischemia. Hevin (SPARC-like protein 1), a synaptogenic protein released by astrocytes, forms
Both unfractionated and low molecular weight heparins inhibit thrombin activation [90]. In addition, they have a fibrolytic (gliolytic) effect and can modulate astrocyte function. a relay between a neurexin (pre-synaptic) and a neuroligin (post-synaptic); this relay is crucial for the synaptogenesis [89].
Clinically, perineural application of condensed polytetrafluoroethylene-extractum cepaeheparin-allantoin gel during peripheral nerve surgery improves functional recovery [91]. The anti-fibrotic effects of heparin are well documented after flexor tendon surgery of the hand [92], in the resolution of intraperitoneal fibrosis [93, 94] and in improving various scar types [95]. **5.3. Modulating the function of astrocytes and heparin** *5.3.1. Role of heparin in lysing the gliosis* Both unfractionated and low molecular weight heparins inhibit thrombin activation [90]. In
Among other actions, heparan sulphate/heparin influences fibroblast growth factor responsi‐ ble for cell proliferation, differentiation, signal transduction and angiogenesis [96, 97]. Heparin is known to inhibit fibroblast growth factor (FGF)-2-stimulated DNA synthesis as well as gene expression of FGF-2 and its receptor in AT2 pneumocytes [98]. Probably via syndecan-1, the presence of heparin at high concentrations reduces the activity of FGF-7, which is responsible for enhancement of keratinocytes migration and proliferation. Heparin enhances the action of FGF-1, which regulates the proliferation of fibroblasts, endothelial and epithelial cells, and influences angiogenesis via effect on the activity of endothelial cells. Heparin can enhance the stability of FGF-1 and might determine the formation of FGF1-FGFR (fibroblast growth factor receptor) active complex. addition, they have a fibrolytic (gliolytic) effect and can modulate astrocyte function. Clinically, perineural application of condensed polytetrafluoroethylene-extractum cepaeheparin-allantoin gel during peripheral nerve surgery improves functional recovery [91]. The anti-fibrotic effects of heparin are well documented after flexor tendon surgery of the hand [92], in the resolution of intraperitoneal fibrosis [93, 94] and in improving various scar types [95]. Among other actions, heparan sulphate/heparin influences fibroblast growth factor responsi‐ ble for cell proliferation, differentiation, signal transduction and angiogenesis [96, 97]. Heparin is known to inhibit fibroblast growth factor (FGF)-2-stimulated DNA synthesis as well as gene expression of FGF-2 and its receptor in AT2 pneumocytes [98]. Probably via syndecan-1, the
#### *5.3.2. Possible role of heparin in modulating astrocyte function* presence of heparin at high concentrations reduces the activity of FGF-7, which is responsi‐ ble for enhancement of keratinocytes migration and proliferation. Heparin enhances the action
Astrocyte stress response and trophic effects are mediated by the FGF family member, on which heparin exerts a profound influence [96, 99]. Fibroblast growth factor 1 (FGF1) has been shown to maintain the survival of neurons and induce neurite outgrowth [100]. Basic fibroblast growth factor (FGF-2) has been found to increase neuronal survival and neurite extension in foetal rat hippocampal neurons when bound to heparin substrates [101]. The length and sulphated position of heparin regulate FGF-2-dependent astrocytic transformation (stellation), native heparin significantly promoting FGF-2-dependent astrocytic transformation, whereas heparin hexasaccharide and 2-O-, 6-O- and N-desulphated heparins inhibit it [102]. Heparin affin regulatory peptide (HARP, pleiotrophin, heparin-binding growth-associated molecule) promotes neurite outgrowth and synaptic development. High levels of heparin affin regula‐ tory peptide HARP mRNA and protein are induced in transformed astrocytes [103, 104]. of FGF-1, which regulates the proliferation of fibroblasts, endothelial and epithelial cells, and influences angiogenesis via effect on the activity of endothelial cells. Heparin can enhance the stability of FGF-1 and might determine the formation of FGF1-FGFR (fibroblast growth factor receptor) active complex. *5.3.2. Possible role of heparin in modulating astrocyte function* Astrocyte stress response and trophic effects are mediated by the FGF family member, on which heparin exerts a profound influence [96, 99]. Fibroblast growth factor 1 (FGF1) has been shown to maintain the survival of neurons and induce neurite outgrowth [100]. Basic fibroblastgrowth factor (FGF-2) has been found to increase neuronal survival and neurite extension in foetal rat hippocampal neurons when bound to heparin substrates [101]. The length and Glypican-1 is a major high-affinity ligand of the Slit proteins, both of which are strongly upregulated in reactive astrocytes, suggesting their possible role in the inhibitory environment preventing axonal regeneration after injury. Heparins inhibit glypican-Slit interactions [105, 106]. sulphated position of heparin regulate FGF-2-dependent astrocytic transformation (stella‐ tion), native heparin significantly promoting FGF-2-dependent astrocytic transformation, whereas heparin hexasaccharide and 2-O-, 6-O- and N-desulphated heparins inhibit it [102].Heparin affin regulatory peptide (HARP, pleiotrophin, heparin-binding growth-associated
molecule) promotes neurite outgrowth and synaptic development. High levels of heparin affin
#### **5.4. Possible role of aspirin in modulating astrocyte function** regulatory peptide HARP mRNA and protein are induced in transformed astrocytes [103, 104].
Reactive astrocytes increase the expression of several of these, notably nerve growth factor, basic fibroblast growth factor, brain-derived neurotrophic factor and neuregulins, which can stimulate neurite outgrowth. Reactive astrocytes also overexpress neuropilin-1 and vascular endothelial growth factor, which act in concert to promote angiogenesis after cerebral ische‐ mia. Hevin (SPARC-like protein 1), a synaptogenic protein released by astrocytes, forms a relay between a neurexin (pre-synaptic) and a neuroligin (post-synaptic); this relay is crucial for the
Astrocytes release a variety of trophic factors [86–88]. These trophic factors include nerve growth factor, basic fibroblast growth factor, transforming growth factor-β, platelet-derived growth factor, brain-derived neurotrophic factor, ciliary neurotrophic factor and others. Reactive astrocytes increase the expression of several of these, notably nerve growth factor, basic fibroblast growth factor, brain-derived neurotrophic factor and neuregulins, which can stimulate neurite outgrowth. Reactive astrocytes also overexpress neuropilin-1 and vascular endothelial growth factor, which act in concert to promote angiogenesis after cerebral ischemia. Hevin (SPARC-like protein 1), a synaptogenic protein released by astrocytes, forms a relay between a neurexin (pre-synaptic) and a neuroligin (post-synaptic); this relay is crucial
Many of these functions can be activated by modulating astrocyte function.
**5.2. Astrocyte trophic effects on neurite outgrowth and neurogenesis**
Both unfractionated and low molecular weight heparins inhibit thrombin activation [90]. In
Clinically, perineural application of condensed polytetrafluoroethylene-extractum cepaeheparin-allantoin gel during peripheral nerve surgery improves functional recovery [91]. The anti-fibrotic effects of heparin are well documented after flexor tendon surgery of the hand [92], in the resolution of intraperitoneal fibrosis [93, 94] and in improving various scar types
Among other actions, heparan sulphate/heparin influences fibroblast growth factor responsi‐ ble for cell proliferation, differentiation, signal transduction and angiogenesis [96, 97]. Heparin is known to inhibit fibroblast growth factor (FGF)-2-stimulated DNA synthesis as well as gene expression of FGF-2 and its receptor in AT2 pneumocytes [98]. Probably via syndecan-1, the presence of heparin at high concentrations reduces the activity of FGF-7, which is responsible for enhancement of keratinocytes migration and proliferation. Heparin enhances the action of FGF-1, which regulates the proliferation of fibroblasts, endothelial and epithelial cells, and influences angiogenesis via effect on the activity of endothelial cells. Heparin can enhance the stability of FGF-1 and might determine the formation of FGF1-FGFR (fibroblast growth factor
Clinically, perineural application of condensed polytetrafluoroethylene-extractum cepaeheparin-allantoin gel during peripheral nerve surgery improves functional recovery [91]. The anti-fibrotic effects of heparin are well documented after flexor tendon surgery of the hand [92], in the resolution of intraperitoneal fibrosis [93, 94] and in improving various scar types
Among other actions, heparan sulphate/heparin influences fibroblast growth factor responsi‐ ble for cell proliferation, differentiation, signal transduction and angiogenesis [96, 97]. Heparin is known to inhibit fibroblast growth factor (FGF)-2-stimulated DNA synthesis as well as gene expression of FGF-2 and its receptor in AT2 pneumocytes [98]. Probably via syndecan-1, the presence of heparin at high concentrations reduces the activity of FGF-7, which is responsi‐ ble for enhancement of keratinocytes migration and proliferation. Heparin enhances the action of FGF-1, which regulates the proliferation of fibroblasts, endothelial and epithelial cells, and influences angiogenesis via effect on the activity of endothelial cells. Heparin can enhance the stability of FGF-1 and might determine the formation of FGF1-FGFR (fibroblast growth factor
Both unfractionated and low molecular weight heparins inhibit thrombin activation [90]. In
Astrocyte stress response and trophic effects are mediated by the FGF family member, on which heparin exerts a profound influence [96, 99]. Fibroblast growth factor 1 (FGF1) has been shown to maintain the survival of neurons and induce neurite outgrowth [100]. Basic fibroblast growth factor (FGF-2) has been found to increase neuronal survival and neurite extension in foetal rat hippocampal neurons when bound to heparin substrates [101]. The length and sulphated position of heparin regulate FGF-2-dependent astrocytic transformation (stellation), native heparin significantly promoting FGF-2-dependent astrocytic transformation, whereas heparin hexasaccharide and 2-O-, 6-O- and N-desulphated heparins inhibit it [102]. Heparin affin regulatory peptide (HARP, pleiotrophin, heparin-binding growth-associated molecule) promotes neurite outgrowth and synaptic development. High levels of heparin affin regula‐ tory peptide HARP mRNA and protein are induced in transformed astrocytes [103, 104].
Astrocyte stress response and trophic effects are mediated by the FGF family member, on which heparin exerts a profound influence [96, 99]. Fibroblast growth factor 1 (FGF1) has been shown to maintain the survival of neurons and induce neurite outgrowth [100]. Basic fibroblastgrowth factor (FGF-2) has been found to increase neuronal survival and neurite extension in foetal rat hippocampal neurons when bound to heparin substrates [101]. The length and
addition, they have a fibrolytic (gliolytic) effect and can modulate astrocyte function.
addition, they have a fibrolytic (gliolytic) effect and can modulate astrocyte function.
synaptogenesis [89].
receptor) active complex.
receptor) active complex.
[95].
[95].
**5.3. Modulating the function of astrocytes and heparin**
**5.3. Modulating the function of astrocytes and heparin**
*5.3.2. Possible role of heparin in modulating astrocyte function*
*5.3.2. Possible role of heparin in modulating astrocyte function*
*5.3.1. Role of heparin in lysing the gliosis*
*5.3.1. Role of heparin in lysing the gliosis*
for the synaptogenesis [89].
14 Recovery of Motor Function in Spinal Cord Injury
194Following Spinal Cord Injury
Ciliary neurotrophic factor (CNTF) is a promyelinating trophic factor. Acetylsalicylic acid (aspirin) increases mRNA and protein expression of CNTF in primary mouse and human astrocytes in a dose- and time-dependent manner. Aspirin-induced astroglial CNTF is also functionally active; supernatants of aspirin-treated astrocytes of wild type, but not Cntf null, mice increase myelin-associated proteins in oligodendrocytes and protected oligodendrocytes from TNF-α insult [107]. Glypican-1 is a major high-affinity ligand of the Slit proteins, both of which are strongly upregulated in reactive astrocytes, suggesting their possible role in the inhibitory environ‐ ment preventing axonal regeneration after injury. Heparins inhibit glypican-Slit interactions [105, 106]. **5.4. Possible role of aspirin in modulating astrocyte function**
Ciliary neurotrophic factor (CNTF) is a promyelinating trophic factor. Acetylsalicylic acid
#### **5.5. Possible role of hyaluronic acid salts in modulating astrocyte function** (aspirin) increases mRNA and protein expression of CNTF in primary mouse and human
scarring as part of future repair strategies after spinal cord injury [108].
The presence of high molecular weight hyaluronic acid (hyaluronic acid with limited degra‐ dation) after spinal cord injury decreases glial scarring. High molecular weight hyaluronic acid stabilised against degradation mitigates astrocyte activation in vitro and in vivo. Therefore, hyaluronic-acid-based hydrogel systems hold great potential for minimising undesired scarring as part of future repair strategies after spinal cord injury [108]. astrocytes in a dose- and time-dependent manner. Aspirin-induced astroglial CNTF is also functionally active; supernatants of aspirin-treated astrocytes of wild type, but not Cntf null, mice increase myelin-associated proteins in oligodendrocytes and protected oligodendro‐ cytes from TNF-α insult [107]. **5.5. Possible role of hyaluronic acid salts in modulating astrocyte function**
#### **6. Combining peripheral nerve grafts with scaffolds: the scaffold as a drug release system** dation) after spinal cord injury decreases glial scarring. High molecular weight hyaluronic acid stabilised against degradation mitigates astrocyte activation in vitro and in vivo. Therefore, hyaluronic-acid-based hydrogel systems hold great potential for minimising undesired
The presence of high molecular weight hyaluronic acid (hyaluronic acid with limited degra‐
The defect within the spinal cord is too large to be bridged by nerve grafts alone; besides, the myelin sheath within them is inhibitory to axonal growth. The rationale for polymer implants is twofold, to replace a damaged area of the cord with just such a structural matrix [109] and to provide it with a synthetic scaffold, in which myelin is absent. Combining peripheral nerve grafts with scaffolds has gained more acceptance because of the importance of seeding the scaffolds with multiple nerve growth factor (NGF) spatial concentration gradients in order to promote axonal growth both within the scaffolds and the nerve grafts [22]. **6. Combining peripheral nerve grafts with scaffolds: the scaffold as a drug release system** The defect within the spinal cord is too large to be bridged by nerve grafts alone; besides, the
Biomaterial scaffolds in spinal cord injury have been reviewed by Madigan et al. [109] and Straley et al. [110]. Commonly used scaffolds include natural polymers (in vivo extracellular matrix polymers, polymers derived from blood, and polymers from marine life) and synthetic polymers (poly-hydroxy acid polymers and synthetic hydrogels). Examples for in vivo extracellular matrix polymers are collagen solutions and the glycosaminoglycan hyaluronic acid. Examples for polymers derived from blood are plasma-derived polymers, fibronectin and fibrin. Examples for polymers from marine life are agarose, alginate and chitosan. Synthetic polymers include poly-hydroxy acid polymers and synthetic hydrogels. Compared to natural polymers, they offer wider scope to design and control the characteristics of the material. Poly-hydroxy acid polymers are biodegradable materials based on polyesters of lactic myelin sheath within them is inhibitory to axonal growth. The rationale for polymer im‐ plants is twofold, to replace a damaged area of the cord with just such a structural matrix [109] and to provide it with a synthetic scaffold, in which myelin is absent. Combining peripheral nerve grafts with scaffolds has gained more acceptance because of the importance of seeding the scaffolds with multiple nerve growth factor (NGF) spatial concentration gradients in order to promote axonal growth both within the scaffolds and the nerve grafts [22]. Biomaterial scaffolds in spinal cord injury have been reviewed by Madigan et al. [109] and Straley et al. [110]. Commonly used scaffolds include natural polymers (in vivo extracellular matrix polymers, polymers derived from blood, and polymers from marine life) and synthet‐ ic polymers (poly-hydroxy acid polymers and synthetic hydrogels). Examples for in vivo and glycolic acid (PLA and PGA) and their co-polymer PLGA. Synthetic hydrogels are based on polyethylene glycol, a biodegradable synthetic polymer of ethylene oxide units. Poly(2 hydroxyethyl methacrylate) (pHEMA) compounds and pHEMA-co-methyl methacrylate (pHEMA-MMA) are used as spinal cord scaffolds. extracellular matrix polymers are collagen solutions and the glycosaminoglycan hyaluronic acid. Examples for polymers derived from blood are plasma-derived polymers, fibronectin and fibrin. Examples for polymers from marine life are agarose, alginate and chitosan. Synthetic polymers include poly-hydroxy acid polymers and synthetic hydrogels. Com‐
Whatever macroengineering and microengineering procedures scaffolds are subjected to allow for axonal growth, this will not occur unless the scaffold is seeded with basal lamina and supplied with neurite outgrowth promoting factors and neurotrophic factors (**Figure 3**). These factors can be released from the scaffold material itself, from integrated micro- or nano-spheres or tubules of a different material, or by means of a scaffold's capacity to support a genetically modified cell line in vivo [58, 109, 110]. pared to natural polymers, they offer wider scope to design and control the characteristics of the material. Poly-hydroxy acid polymers are biodegradable materials based on polyesters of lactic and glycolic acid (PLA and PGA) and their co-polymer PLGA. Synthetic hydrogels are based on polyethylene glycol, a biodegradable synthetic polymer of ethylene oxide units. Poly(2-hydroxyethyl methacrylate) (pHEMA) compounds and pHEMA-co-methyl methacry‐ late (pHEMA-MMA) are used as spinal cord scaffolds.
Whatever macroengineering and microengineering procedures scaffolds are subjected to allow for axonal growth, this will not occur unless the scaffold is seeded with basal lamina and
or tubules of a different material, or by means of a scaffold's capacity to support a genetical‐
#### **7. Augmentation of peripheral nerve grafts with cellular transplants** supplied with neurite outgrowth promoting factors and neurotrophic factors (**Figure 3**). These factors can be released from the scaffold material itself, from integrated micro- or nano-spheres
Nearly half of the spinal cord injuries occur at the thoracolumbar junction (D12-L1), the site of the conus medullaris. We could potentially nerve graft the spinal cord (part of the CNS) directly to the cauda equina (part of the peripheral nervous system) without resorting to cellular transplants to restitute the neuronal and astrocytic components of the CNS. There is mounting evidence, however, that cellular transplants potentiate the effect of other factors and might even recruit endogenous neural precursor cells [62, 76]. ly modified cell line in vivo [58, 109, 110]. **7. Augmentation of peripheral nerve grafts with cellular transplants** Nearly half of the spinal cord injuries occur at the thoracolumbar junction (D12-L1), the site of
#### **7.1. Types of cellular transplants used to augment peripheral nerve grafts** the conus medullaris. We couldpotentially nerve graftthe spinal cord (part ofthe CNS) directly to the cauda equina (part of the peripheral nervous system) without resorting to cellular
In a meta-analysis reviewing cellular transplantation strategies in spinal cord injury, Tetzlaff et al. [111], have come to the following conclusions. Schwann cells are the most extensively studied cell type. They are reported both to remyelinate-injured spinal cord axons and to form a permissive substrate for their regeneration [112]. However, compared to neural precursors such as oligodendrocyte precursors or neural precursor/stem cells, they provoke a more robust astrocytic reaction, resulting in less effective integration into the host spinal cord [113]. Olfactory ensheathing cells demonstrate good integration into host spinal cord [114]. However, there is no robust evidence of improvement after their transplantation [115]. They also appear to require adjuvant treatment to increase efficacy (e.g., Schwann cells, Matrigel, rolipram, cAMP and neurotrophic factors) [116,117]. Neural stem/progenitor cells appear to integrate well into the host spinal cord with improved outcomes [118]. They differentiate mostly into astroglial cells, less so oligodendrocytes seen but rarely into neurons [119]. Suspicion has been raised as to their role in axonal regeneration [120, 121]. Fate-restricted neural and glial precursor have the potential to remyelinate injured axons [122]. More evidence is needed, however, to confirm this observation [123]. Bone marrow stromal cells have some bridging capacity in sharp transaction models [124]. However, their integration in the injured spinal cord is very limited. There is no convincing differentiation into neural cells despite claims to the contrary [125]. They are reported to stimulate neurite outgrowth over neural proteogly‐ cans, myelin-associated glycoprotein and Nogo-A [126, 127]. transplants to restitute the neuronal and astrocytic components of the CNS. There is mount‐ ing evidence, however, that cellulartransplants potentiate the effect of otherfactors and might even recruit endogenous neural precursor cells [62, 76]. **7.1. Types of cellular transplants used to augment peripheral nerve grafts** In a meta-analysis reviewing cellular transplantation strategies in spinal cord injury, Tetzlaff et al. [111], have come to the following conclusions. Schwann cells are the most extensively studied cell type. They are reported both to remyelinate-injured spinal cord axons and to form a permissive substrate for their regeneration [112]. However, compared to neural precursors such as oligodendrocyte precursors or neural precursor/stem cells, they provoke a more robust astrocytic reaction, resulting in less effective integration into the host spinal cord [113]. Olfactory ensheathing cellsdemonstrate goodintegration into host spinal cord [114]. However, there is no robust evidence of improvement after their transplantation [115]. They also appear to require adjuvant treatment to increase efficacy (e.g., Schwann cells, Matrigel, rolipram, cAMP and neurotrophic factors) [116,117]. Neural stem/progenitor cells appear to integrate well into the host spinal cord with improved outcomes [118]. They differentiate mostly into astroglial cells, less so oligodendrocytes seen but rarely into neurons [119]. Suspicion has been raised as to their role in axonal regeneration [120, 121]. Fate-restricted neural and glial precursor have the potential to remyelinate injured axons [122]. More evidence is needed,
#### **7.2. Cellular transplants in combination strategy** however, to confirm this observation [123]. Bone marrow stromal cells have some bridging
and glycolic acid (PLA and PGA) and their co-polymer PLGA. Synthetic hydrogels are based on polyethylene glycol, a biodegradable synthetic polymer of ethylene oxide units. Poly(2 hydroxyethyl methacrylate) (pHEMA) compounds and pHEMA-co-methyl methacrylate
extracellular matrix polymers are collagen solutions and the glycosaminoglycan hyaluronic acid. Examples for polymers derived from blood are plasma-derived polymers, fibronectin and fibrin. Examples for polymers from marine life are agarose, alginate and chitosan. Synthetic polymers include poly-hydroxy acid polymers and synthetic hydrogels. Com‐ pared to natural polymers, they offer wider scope to design and control the characteristics of the material. Poly-hydroxy acid polymers are biodegradable materials based on polyesters of lactic and glycolic acid (PLA and PGA) and their co-polymer PLGA. Synthetic hydrogels are based on polyethylene glycol, a biodegradable synthetic polymer of ethylene oxide units. Poly(2-hydroxyethyl methacrylate) (pHEMA) compounds and pHEMA-co-methyl methacry‐
Whatever macroengineering and microengineering procedures scaffolds are subjected to allow for axonal growth, this will not occur unless the scaffold is seeded with basal lamina and supplied with neurite outgrowth promoting factors and neurotrophic factors (**Figure 3**). These factors can be released from the scaffold material itself, from integrated micro- or nano-spheres or tubules of a different material, or by means of a scaffold's capacity to support a genetically
**7. Augmentation of peripheral nerve grafts with cellular transplants**
**7. Augmentation of peripheral nerve grafts with cellular transplants**
**7.1. Types of cellular transplants used to augment peripheral nerve grafts**
**7.1. Types of cellular transplants used to augment peripheral nerve grafts**
Nearly half of the spinal cord injuries occur at the thoracolumbar junction (D12-L1), the site of the conus medullaris. We could potentially nerve graft the spinal cord (part of the CNS) directly to the cauda equina (part of the peripheral nervous system) without resorting to cellular transplants to restitute the neuronal and astrocytic components of the CNS. There is mounting evidence, however, that cellular transplants potentiate the effect of other factors and might
Whatever macroengineering and microengineering procedures scaffolds are subjected to allow for axonal growth, this will not occur unless the scaffold is seeded with basal lamina and supplied with neurite outgrowth promoting factors and neurotrophic factors (**Figure 3**). These factors can be released from the scaffold material itself, from integrated micro- or nano-spheres or tubules of a different material, or by means of a scaffold's capacity to support a genetical‐
In a meta-analysis reviewing cellular transplantation strategies in spinal cord injury, Tetzlaff et al. [111], have come to the following conclusions. Schwann cells are the most extensively studied cell type. They are reported both to remyelinate-injured spinal cord axons and to form a permissive substrate for their regeneration [112]. However, compared to neural precursors such as oligodendrocyte precursors or neural precursor/stem cells, they provoke a more robust astrocytic reaction, resulting in less effective integration into the host spinal cord [113]. Olfactory ensheathing cells demonstrate good integration into host spinal cord [114]. However, there is no robust evidence of improvement after their transplantation [115]. They also appear to require adjuvant treatment to increase efficacy (e.g., Schwann cells, Matrigel, rolipram, cAMP and neurotrophic factors) [116,117]. Neural stem/progenitor cells appear to integrate well into the host spinal cord with improved outcomes [118]. They differentiate mostly into astroglial cells, less so oligodendrocytes seen but rarely into neurons [119]. Suspicion has been raised as to their role in axonal regeneration [120, 121]. Fate-restricted neural and glial precursor have the potential to remyelinate injured axons [122]. More evidence is needed, however, to confirm this observation [123]. Bone marrow stromal cells have some bridging capacity in sharp transaction models [124]. However, their integration in the injured spinal cord is very limited. There is no convincing differentiation into neural cells despite claims to the contrary [125]. They are reported to stimulate neurite outgrowth over neural proteogly‐
In a meta-analysis reviewing cellular transplantation strategies in spinal cord injury, Tetzlaff et al. [111], have come to the following conclusions. Schwann cells are the most extensively studied cell type. They are reported both to remyelinate-injured spinal cord axons and to form a permissive substrate for their regeneration [112]. However, compared to neural precursors such as oligodendrocyte precursors or neural precursor/stem cells, they provoke a more robust astrocytic reaction, resulting in less effective integration into the host spinal cord [113]. Olfactory ensheathing cellsdemonstrate goodintegration into host spinal cord [114]. However, there is no robust evidence of improvement after their transplantation [115]. They also appear to require adjuvant treatment to increase efficacy (e.g., Schwann cells, Matrigel, rolipram, cAMP and neurotrophic factors) [116,117]. Neural stem/progenitor cells appear to integrate well into the host spinal cord with improved outcomes [118]. They differentiate mostly into astroglial cells, less so oligodendrocytes seen but rarely into neurons [119]. Suspicion has been raised as to their role in axonal regeneration [120, 121]. Fate-restricted neural and glial precursor have the potential to remyelinate injured axons [122]. More evidence is needed,
Nearly half of the spinal cord injuries occur at the thoracolumbar junction (D12-L1), the site of the conus medullaris. We couldpotentially nerve graftthe spinal cord (part ofthe CNS) directly to the cauda equina (part of the peripheral nervous system) without resorting to cellular transplants to restitute the neuronal and astrocytic components of the CNS. There is mount‐ ing evidence, however, that cellulartransplants potentiate the effect of otherfactors and might
(pHEMA-MMA) are used as spinal cord scaffolds.
late (pHEMA-MMA) are used as spinal cord scaffolds.
even recruit endogenous neural precursor cells [62, 76].
even recruit endogenous neural precursor cells [62, 76].
cans, myelin-associated glycoprotein and Nogo-A [126, 127].
modified cell line in vivo [58, 109, 110].
ly modified cell line in vivo [58, 109, 110].
16 Recovery of Motor Function in Spinal Cord Injury
196Following Spinal Cord Injury
Because of the previous controversies, the use of a combination strategy including Schwann cells has been recommended by Bunge [128]. The following combination strategy has been suggested [128]: Schwann cells, neuroprotective agents and growth factors administered in various ways, such as, olfactory ensheathing cell (OEC) implantation, chondroitinase addition or elevation of cyclic AMP. A targeted approach has been proposed by Kadoya et al. [43]. It includes the following: a peripheral conditioning lesion (bilateral sciatic nerve crush), mesen‐ chymal stem cell transplantation mixed with neurotrophin-3 (NT-3) and creating a neurotro‐ phic factor gradient by injection of lentivirus expressing neurotrophin-3 (NT-3) just proximal to the site of the lesion. capacity in sharp transaction models [124]. However, their integration in the injured spinal cord is very limited. There is no convincing differentiation into neural cells despite claims to the contrary [125]. They are reported to stimulate neurite outgrowth over neural proteogly‐ cans, myelin-associated glycoprotein and Nogo-A [126, 127]. **7.2. Cellular transplants in combination strategy** Because of the previous controversies, the use of a combination strategy including Schwann cells has been recommended by Bunge [128]. The following combination strategy has been suggested [128]: Schwann cells, neuroprotective agents and growth factors administered in
#### **7.3. Number of cellular transplant injections** various ways, such as, olfactory ensheathing cell (OEC) implantation, chondroitinase addi‐ tion or elevation of cyclic AMP. A targeted approach has been proposed by Kadoya et al. [43].
A third unresolved issue is the number of injections that the patient has to receive. Mackay-Sim et al. [129] have used a single intraoperative injection. Multiple injections have been recorded by other authors [130, 131]. It includes the following: a peripheral conditioning lesion (bilateral sciatic nerve crush), mesenchymal stem cell transplantation mixed with neurotrophin-3 (NT-3) and creating a neurotrophic factor gradient by injection of lentivirus expressing neurotrophin-3 (NT-3) just proximal to the site of the lesion.
### **7.4. Inducing mobilisation of neural precursor cells 7.3. Number of cellular transplant injections**
Stem cell transplantation has the potential to recruit endogenous neural stem cells [132]. Neural stem cells exist in the mammalian developing and adult nervous system (mainly in the hippocamous and subventricular area). Multiple cell-intrinsic regulators coordinate neural stem cell maintenance, self-renewal and migration into injured areas. Essential intracellular regulators include the orphan nuclear receptor TLX, the high-mobility-group DNA binding protein Sox2, the basic helix-loop-helix transcription factor Hes, the tumour suppressor gene Pten, the membrane-associated protein Numb and its cytoplasmic homolog Numblike. Manipulating these factors among others [133–135] by injecting them through indwelling catheters might induce mobilisation of neural stem cells to the injured spinal cord area. A third unresolved issue is the number of injections that the patient has to receive. Mackay-Sim et al. [129] have used a single intraoperative injection. Multiple injections have been recorded by other authors [130, 131]. **7.4. Inducing mobilisation of neural precursor cells** Stem celltransplantation has the potentialto recruit endogenous neural stem cells [132]. Neural stem cells exist in the mammalian developing and adult nervous system (mainly in the hippocamous and subventricular area). Multiple cell-intrinsic regulators coordinate neural
stem cell maintenance, self-renewal and migration into injured areas. Essential intracellular
Pten, the membrane-associated protein Numb and its cytoplasmic homolog Numblike.
#### **8. Clinical application** regulators include the orphan nuclear receptor TLX, the high-mobility-group DNA binding protein Sox2, the basic helix-loop-helix transcription factor Hes, the tumour suppressor gene
#### **8.1. Pre-operative assessment (neurological and radiographic evaluation)** Manipulating these factors among others [133–135] by injecting them through indwelling
Patients should be evaluated pre-operatively and at monthly intervals. Motor power and sensation should be evaluated using ASIA standards [28]. Confounding factors during motor power evaluation include fake muscle contractions produced by movements of the trunk and cocontractions between abdominal muscles and different muscle groups. Optional elements of ASIA neurologic impairment assessment should be included because the abdominal muscles and medial hamstrings are the first muscles to regain power. catheters might induce mobilisation of neural stem cells to the injured spinal cord area. **8. Clinical application 8.1. Pre-operative assessment (neurological and radiographic evaluation)**
Radiographic evaluation should include plain anteroposterior and lateral radiographs and pre-operative magnetic resonance imaging (MRI). The injury zone on the MRI is determined by the superior and inferior extents of the gliosis; nerve grafts have to be extended beyond the injury zone [136] (**Figure 6**). Patients should be evaluated pre-operatively and at monthly intervals. Motor power sensation should be evaluated using ASIA standards [28]. Confounding factors during motor power evaluation include fake muscle contractions produced by movements of the trunk and cocontractions between abdominal muscles and different muscle groups. Optional elements
**Figure 6.** Cervical vertebral C5,6 fracture dislocation in a quadriplegic patient; the gliosis extends from the inferior bor‐ der of cervical vertebra C4 to the superior border of cervical vertebra C6. The injury zone on the MRI is determined by the superior and inferior extents of the gliosis; nerve grafts have to be extended beyond the injury zone.
### **8.2. Timing of grafting**
Spinal cord injury is considered chronic months to years after injury [6]. At this stage, the primary and secondary injuries have ceased. As the zone of gliosis may extend superiorly or inferiorly during the secondary injury phase, the patient should be operated upon at least 2 months after the injury, i.e., when it has become chronic. **Figure 6.** Cervical vertebral C5,6 fracture dislocation in a quadriplegic patient; the gliosis extends from the inferior border of cervical vertebra C4 to the superior border of cervical vertebra C6. The injury zone on the MRI is determined
Improvement is also independent of the time delay between the date of injury and the date of definitive surgery. This is supported by the observations made by Li and Raisman [137], who have noted that sprouts from cut corticospinal axons persist despite the presence of astrocytic scarring in long-term lesions of the adult rat spinal cord. by the superior and inferior extents of the gliosis; nerve grafts have to be extended beyond the injury zone. **8.2. Timing of grafting** Spinal cord injury is considered chronic months to years after injury [6]. At this stage, the
#### **8.3. Operative technique, establishment of CSF circulation and establishment of a continuous drug delivery system (indwelling catheter implantation for post-operative drug and cell administration)** inferiorly during the secondary injury phase, the patient should be operated upon at least 2 months after the injury, i.e., when it has become chronic. Improvement is also independent of the time delay between the date of injury and the date of
primary and secondary injuries have ceased. As the zone of gliosis may extend superiorly or
After exploring the injured cord through a posterior spinal laminectomy incision, sural nerves are side-grafted to the cord, especially on its ventral aspect [20, 138] (**Figures 7**(**a** and **b**) and **8**). The cord graft construct invariably adheres to the dura preventing CSF circulation. The definitive surgery. This is supported by the observations made by Li and Raisman [137], who have noted that sprouts from cut corticospinal axons persist despite the presence of astrocyt‐ ic scarring in long-term lesions of the adult rat spinal cord.
latter is important for nutrient, cell and growth factor transport [139, 140]. To prevent the cord adhering to the dura, it has become the author's practice to wrap the cord graft construct with a silicone membrane (**Figure 9**). In fact, silicone chambers or tubes have been used as scaffolds for peripheral nerve regeneration [141–143]. Interest to use them as scaffolds for cellular growth has been rekindled [144, 145]. They have been modified physically to increase porosity or coated to allow for growth of mesenchymal stem cells [146–148] or even to allow neuronlike differentiation of mesenchymal stem cells [149]. They have been modified physically to increase porosity or chemically (with hyaluronic acid and hyaluronic acid--collagen conjugate) to allow for growth of neural cells [150, 151] or to inhibit glial tissue formation [152, 153]. To establish a continuous drug and cell delivery system, an indwelling percutaneous catheter is placed in the interstitium between the membrane and the dura; the dura is finally closed (**Figure 9**). **Figure 10** is a schematic drawing of the hypothetical spinal cord-graft-scaffoldcatheter construct. **8.3. Operative technique, establishment of CSF circulation and establishment of a continuous drug delivery system (indwelling catheterimplantation for post-operative drug and cell administration)** After exploring the injured cord through a posterior spinal laminectomy incision, sural nerves are side-grafted to the cord, especially on its ventral aspect [20, 138] (**Figures 7**(**a** and **b**) and **8**). The cord graft construct invariably adheres to the dura preventing CSF circulation. The latter is important for nutrient, cell and growth factor transport [139, 140]. To prevent the cord adhering to the dura, it has become the author's practice to wrap the cord graft construct with a silicone membrane (**Figure 9**). In fact, silicone chambers or tubes have been used as scaf‐
**Figure 6.** Cervical vertebral C5,6 fracture dislocation in a quadriplegic patient; the gliosis extends from the inferior bor‐ der of cervical vertebra C4 to the superior border of cervical vertebra C6. The injury zone on the MRI is determined by
of ASIA neurologic impairment assessment should be included because the abdominal
Radiographic evaluation should include plain anteroposterior and lateral radiographs and pre-operative magnetic resonance imaging (MRI). The injury zone on the MRI is determined by the superior and inferior extents of the gliosis; nerve grafts have to be extended beyond the
muscles and medial hamstrings are the first muscles to regain power.
Spinal cord injury is considered chronic months to years after injury [6]. At this stage, the primary and secondary injuries have ceased. As the zone of gliosis may extend superiorly or inferiorly during the secondary injury phase, the patient should be operated upon at least 2
Improvement is also independent of the time delay between the date of injury and the date of definitive surgery. This is supported by the observations made by Li and Raisman [137], who have noted that sprouts from cut corticospinal axons persist despite the presence of astrocytic
**Figure 6.** Cervical vertebral C5,6 fracture dislocation in a quadriplegic patient; the gliosis extends from the inferior border of cervical vertebra C4 to the superior border of cervical vertebra C6. The injury zone on the MRI is determined
**continuous drug delivery system (indwelling catheter implantation for post-operative drug**
Spinal cord injury is considered chronic months to years after injury [6]. At this stage, the primary and secondary injuries have ceased. As the zone of gliosis may extend superiorly or inferiorly during the secondary injury phase, the patient should be operated upon at least 2
After exploring the injured cord through a posterior spinal laminectomy incision, sural nerves are side-grafted to the cord, especially on its ventral aspect [20, 138] (**Figures 7**(**a** and **b**) and **8**). The cord graft construct invariably adheres to the dura preventing CSF circulation. The
Improvement is also independent of the time delay between the date of injury and the date of definitive surgery. This is supported by the observations made by Li and Raisman [137], who have noted that sprouts from cut corticospinal axons persist despite the presence of astrocyt‐
**8.3. Operative technique, establishment of CSF circulation and establishment of a**
by the superior and inferior extents of the gliosis; nerve grafts have to be extended beyond the injury zone.
the superior and inferior extents of the gliosis; nerve grafts have to be extended beyond the injury zone.
months after the injury, i.e., when it has become chronic.
scarring in long-term lesions of the adult rat spinal cord.
months after the injury, i.e., when it has become chronic.
ic scarring in long-term lesions of the adult rat spinal cord.
**8.2. Timing of grafting**
18 Recovery of Motor Function in Spinal Cord Injury
198Following Spinal Cord Injury
injury zone [136] (**Figure 6**).
**and cell administration)**
**8.2. Timing of grafting**
**Figure 7.** (a) Through a midline dorsal incision, a formal laminectomy is performed preserving the facet joints and pedicles. The dura is exposed. The dura is incised longitudinally and held with stay sutures exposing the cord lesion. The yellow arrows points to the defect in the cord. (b) Spinal cord lesion without defect but with a completely gliotic segment in a paraplegic patient suffering from a dorsolumbar fracture dislocation. The gliotic segment extends from the cord (white arrow) up to the cauda equine (yellow arrow).
**Figure 8.** Sural nerve grafts having been side-grafted to the cord.
**Figure 9.** To prevent the cord adhering to the dura, it has become the author's practice to wrap the cord graft construct with a silicone membrane. To establish a continuous drug and cell delivery system, an indwelling catheter is placed in the interstitium between the membrane and the dura; the dura is finally closed.
Bridging Defects in Chronic Spinal Cord Injury Using Peripheral Nerve Grafts: From Basic Science to Clinical Experience 21 in Chronic http://dx.doi.org/10.5772/64211 201
**Figure 10.** A schematic drawing of the hypothetical spinal cord-graft-scaffold-catheter construct: (I) end-to-end grafts; (II) side grafts; (III) synthetic scaffolds; (IV) cellular transplants; (V) modulated astrocytes; (VI) silicone membrane; (VII) dura mater; (VIII) indwelling catheter for post-operative delivery of neurolyzing agents (chondroitinase ABC, heparin), neurotrophic factors, neurite outgrowth promoting factors, injectable scaffolds and cellular transplants; (IX) skin. **Figure 9.** To prevent the cord adhering to the dura, it has become the author's practice to wrap the cord graft construct with a silicone membrane. To establish a continuous drug and cell delivery system, an indwelling catheter is placed in the interstitium between the membrane and the dura; the dura is finally closed.
### **8.4. Post-operative drug and cellular transplant administration through the catheter**
**Figure 8.** Sural nerve grafts having been side-grafted to the cord.
the cord (white arrow) up to the cauda equine (yellow arrow).
20 Recovery of Motor Function in Spinal Cord Injury
200Following Spinal Cord Injury
**Figure 9.** To prevent the cord adhering to the dura, it has become the author's practice to wrap the cord graft construct with a silicone membrane. To establish a continuous drug and cell delivery system, an indwelling catheter is placed in
**Figure 7.** (a) Through a midline dorsal incision, a formal laminectomy is performed preserving the facet joints and pedicles. The dura is exposed. The dura is incised longitudinally and held with stay sutures exposing the cord lesion. The yellow arrows points to the defect in the cord. (b) Spinal cord lesion without defect but with a completely gliotic segment in a paraplegic patient suffering from a dorsolumbar fracture dislocation. The gliotic segment extends from
the interstitium between the membrane and the dura; the dura is finally closed.
**Figure 8.** Sural nerve grafts having been side-grafted to the cord.
The catheter can be used for post-operative administration of growth factors, neurolyzing agents, cellular transplants or even scaffolds. The author's practice has been as follows. Starting from the fifth post-operative day calcium heparin (5000 IU) is injected every second day through the catheter. Chondroitinase ABC (5 IU, Sigma) is dissolved in 2 cc normal saline and injected on a weekly basis.
### **8.5. Catheter-related complications: keeping the indwelling catheter for 18 months or more around the spinal cord as a means of establishing a continuous spinal drug delivery system**
Catheter-related complications include tension headache, meningitis, fibrous track formation, catheter slippage, difficult catheter insertion and catheter blockage. Fibrous track formation is noted by increased pressure on injection through the catheter, associated with increased serosanguinous discharge from the catheter skin exit site due to drug extrusion. Tension headache can be avoided by decreasing the volume of injection; meningitis and early catheter blockage and slippage are avoided by proper catheter care. Complications associated with fibrous track formation, such as difficult catheter insertion, late catheter blockage and slippage (during months 9–18) can only be avoided by inserting the catheter in the interstitium between the silicone membrane and the dura. In this way, the catheter need not be exchanged over 18 months. **Figure 10.** A schematic drawing of the hypothetical spinal cord-graft-scaffold-catheter construct: (I) end-to-end grafts; (II) side grafts; (III) synthetic scaffolds; (IV) cellular transplants; (V) modulated astrocytes; (VI) silicone membrane; (VII) dura mater; (VIII) indwelling catheter for post-operative delivery of neurolyzing agents (chondroitinase ABC,
#### **8.6. Author's clinical experience with heparin and chondroitinase ABC: its clinical safety and limitations** heparin), neurotrophic factors, neurite outgrowth promoting factors, injectable scaffolds and cellular transplants; (IX) skin.
Delayed wound healing and sinus formation is related to repeated calheparin injection. Its incidence decreases when calheparin is administered every other day. **8.4. Post-operative drug and cellular transplant administration through the catheter** The catheter can be used for post-operative administration of growth factors, neurolyzing
A vasovagal reaction occurs, when chondroitinase ABC is rapidly injected intrathecally. Its manifestations are cough, hypotension, general irritability and spinal cord irritability mani‐ fested by lower limb twitches. A vasovagal reaction does not occur, when the enzyme is injected extradurally or slowly intrathecally. agents, cellulartransplants or even scaffolds. The author's practice has been as follows. Starting from the fifth post-operative day calcium heparin (5000 IU) is injected every second day through the catheter. Chondroitinase ABC (5 IU, Sigma) is dissolved in 2 cc normal saline and injected on a weekly basis.
#### **8.7. Results of surgery 8.5. Catheter-related complications: keeping the indwelling catheter for 18 months or more around the spinal cord as a means of establishing a continuous spinal drug delivery system**
In a clinical study [14], the right and left antero-lateral quadrant of the cord at T7-8 levels have been nerve grafted to homolateral L2-4 lumbar ventral roots. Eight months after surgery, voluntary contractions of bilateral adductors and of the left quadriceps have been observed. Catheter-related complications include tension headache, meningitis, fibrous track forma‐ tion, catheter slippage, difficult catheter insertion and catheter blockage. Fibrous track formation is noted by increased pressure on injection through the catheter, associated with
Similar improvements have been observed in another study [15] after nerve side-grafting and augmentation by single-stage mesenchymal cell transplantation. Improvement has been hampered by cocontractions between abdominal muscles and different muscle groups. It has also been hampered by spasticity increased serosanguinous discharge from the catheter skin exit site due to drug extrusion. Tension headache can be avoided by decreasing the volume of injection; meningitis and early catheter blockage and slippage are avoided by proper catheter care. Complications associat‐ ed with fibrous track formation, such as difficult catheter insertion, late catheter blockage and slippage (during months 9–18) can only be avoided by inserting the catheter in the interstiti‐
In a not yet published study, the author has observed that repeated heparin, chondroitinase ABC and cellular transplant injection through an indwelling catheter placed in the interstitium between the membrane and the dura has led to the disappearance of cocontractions between abdominal muscles and different muscle groups. All patients have had pre-operative bouts of a moderate dull aching pain in the abdomen, back and both legs caused by adherence of the cord to the dura and the bony spinal canal. It has been completely resolved by inserting the silicone barrier membrane in the interstitium between the spinal cord and the dura. um between the silicone membrane and the dura. In this way, the catheter need not be exchanged over 18 months. **8.6. Author's clinical experience with heparin and chondroitinase ABC: its clinical safety and limitations** Delayed wound healing and sinus formation is related to repeated calheparin injection. Its incidence decreases when calheparin is administered every other day.
Studies using cellular transplantation alone in spinal cord injuries have reported similar motor and sensory score improvement [154–157]. In all these studies, spontaneous or treatmentinduced anatomical neural plasticity as well as the adaptive reorganisation of the neural pathways occurring after injury and acting to restore some of the lost function have to be taken into consideration [4]. A vasovagal reaction occurs, when chondroitinase ABC is rapidly injected intrathecally. Its manifestations are cough, hypotension, general irritability and spinal cord irritability mani‐ fested by lower limb twitches. A vasovagal reaction does not occur, when the enzyme is injected extradurally or slowly intrathecally. **8.7. Results of surgery**
#### **8.8. False positive results** In a clinical study [14], the right and left antero-lateral quadrant of the cord at T7-8 levels have
On evaluating results of surgery after grafting the cord to the cauda equina in thoracolumbar lesions, it should be noted that false positive results could be obtained from intercostal nerves (peripheral nerves) regenerating into the cauda equina (peripheral nerves) via nerve grafts (**Figure 11**). been nerve grafted to homolateral L2-4 lumbar ventral roots. Eight months after surgery, voluntary contractions of bilateral adductors and of the left quadriceps have been observed. Similar improvements have been observed in another study [15] after nerve side-grafting and augmentation by single-stage mesenchymal cell transplantation. Improvement has been Bridging Defects in Chronic Spinal Cord Injury Using Peripheral Nerve Grafts: From Basic Science to Clinical Experience 23 in Chronic http://dx.doi.org/10.5772/64211 203
**Figure 11.** In thoracolumbar lesions, false positive results could be obtained from intercostal nerves (peripheral nerves) at the cranial cord (white arrow) regenerating into to the caua equina (peripheral nerves) (yellow arrow) via nerve grafts. Studies using cellulartransplantation alone in spinal cord injuries have reported similar motor and sensory score improvement [154–157]. In all these studies, spontaneous or treatment-
induced anatomical neural plasticity as well as the adaptive reorganisation of the neural
pathways occurring afterinjury and acting to restore some of the lost function have to be taken
### **9. Post-operative target organ derived trophic support**
the silicone membrane and the dura. In this way, the catheter need not be exchanged over 18
**Figure 10.** A schematic drawing of the hypothetical spinal cord-graft-scaffold-catheter construct: (I) end-to-end grafts; (II) side grafts; (III) synthetic scaffolds; (IV) cellular transplants; (V) modulated astrocytes; (VI) silicone membrane; (VII) dura mater; (VIII) indwelling catheter for post-operative delivery of neurolyzing agents (chondroitinase ABC, heparin), neurotrophic factors, neurite outgrowth promoting factors, injectable scaffolds and cellular transplants; (IX)
**8.6. Author's clinical experience with heparin and chondroitinase ABC: its clinical safety**
Delayed wound healing and sinus formation is related to repeated calheparin injection. Its
The catheter can be used for post-operative administration of growth factors, neurolyzing agents, cellulartransplants or even scaffolds. The author's practice has been as follows. Starting from the fifth post-operative day calcium heparin (5000 IU) is injected every second day through the catheter. Chondroitinase ABC (5 IU, Sigma) is dissolved in 2 cc normal saline and
**8.4. Post-operative drug and cellular transplant administration through the catheter**
A vasovagal reaction occurs, when chondroitinase ABC is rapidly injected intrathecally. Its manifestations are cough, hypotension, general irritability and spinal cord irritability mani‐ fested by lower limb twitches. A vasovagal reaction does not occur, when the enzyme is
In a clinical study [14], the right and left antero-lateral quadrant of the cord at T7-8 levels have been nerve grafted to homolateral L2-4 lumbar ventral roots. Eight months after surgery, voluntary contractions of bilateral adductors and of the left quadriceps have been observed.
Catheter-related complications include tension headache, meningitis, fibrous track forma‐ tion, catheter slippage, difficult catheter insertion and catheter blockage. Fibrous track formation is noted by increased pressure on injection through the catheter, associated with increased serosanguinous discharge from the catheter skin exit site due to drug extrusion. Tension headache can be avoided by decreasing the volume of injection; meningitis and early catheter blockage and slippage are avoided by proper catheter care. Complications associat‐ ed with fibrous track formation, such as difficult catheter insertion, late catheter blockage and slippage (during months 9–18) can only be avoided by inserting the catheter in the interstiti‐ um between the silicone membrane and the dura. In this way, the catheter need not be
**8.5. Catheter-related complications: keeping the indwelling catheter for 18 months or more around the spinal cord as a means of establishing a continuous spinal drug delivery system**
Similar improvements have been observed in another study [15] after nerve side-grafting and augmentation by single-stage mesenchymal cell transplantation. Improvement has been hampered by cocontractions between abdominal muscles and different muscle groups. It has
In a not yet published study, the author has observed that repeated heparin, chondroitinase ABC and cellular transplant injection through an indwelling catheter placed in the interstitium between the membrane and the dura has led to the disappearance of cocontractions between abdominal muscles and different muscle groups. All patients have had pre-operative bouts of a moderate dull aching pain in the abdomen, back and both legs caused by adherence of the cord to the dura and the bony spinal canal. It has been completely resolved by inserting the
**8.6. Author's clinical experience with heparin and chondroitinase ABC: its clinical safety**
Studies using cellular transplantation alone in spinal cord injuries have reported similar motor and sensory score improvement [154–157]. In all these studies, spontaneous or treatmentinduced anatomical neural plasticity as well as the adaptive reorganisation of the neural pathways occurring after injury and acting to restore some of the lost function have to be taken
A vasovagal reaction occurs, when chondroitinase ABC is rapidly injected intrathecally. Its manifestations are cough, hypotension, general irritability and spinal cord irritability mani‐ fested by lower limb twitches. A vasovagal reaction does not occur, when the enzyme is
Delayed wound healing and sinus formation is related to repeated calheparin injection. Itsincidence decreases when calheparin is administered every other day.
On evaluating results of surgery after grafting the cord to the cauda equina in thoracolumbar lesions, it should be noted that false positive results could be obtained from intercostal nerves (peripheral nerves) regenerating into the cauda equina (peripheral nerves) via nerve grafts
Similar improvements have been observed in another study [15] after nerve side-grafting and augmentation by single-stage mesenchymal cell transplantation. Improvement has been
In a clinical study [14], the right and left antero-lateral quadrant of the cord at T7-8 levels have been nerve grafted to homolateral L2-4 lumbar ventral roots. Eight months after surgery, voluntary contractions of bilateral adductors and of the left quadriceps have been observed.
silicone barrier membrane in the interstitium between the spinal cord and the dura.
incidence decreases when calheparin is administered every other day.
injected extradurally or slowly intrathecally.
months.
skin.
**and limitations**
22 Recovery of Motor Function in Spinal Cord Injury
202Following Spinal Cord Injury
**8.7. Results of surgery**
injected on a weekly basis.
also been hampered by spasticity
exchanged over 18 months.
**and limitations**
into consideration [4].
**8.7. Results of surgery**
(**Figure 11**).
**8.8. False positive results**
injected extradurally or slowly intrathecally.
Target organ derived neurotrophic factors, the so-called neurotrophins (nerve growth factor(NGF), brain-derived neurotrophic factor (BDNF), neurotrophin-3(NT-3) and neurotro‐ phin-4/5(NT-4/5)), are transported by retrograde axonal via an endosomal mechanism involving dyneins [158, 159]. Neurotrophins contribute a lot to axonal progression; in the injured spinal cord, this stimulus is lost. After grafting the spinal cord, an important question is whether this stimulus can be restituted by injecting neurotrophins into target organs postoperatively, or whether axonal progression into specific nerves can be restituted by injecting neurotrophins into the specific muscles supplied by them. Experimental evidence points to this [158, 159]. By the same token, it can be questioned whether other neurotrophic factors and neurolyzing agents can be similarly injected into target muscles. Both in vitro and in vivo local infusion of fibroblastic growth factors (FGFs) have been found to rescue motoneuron death induced by spinal cord injury [160]. However, evidence for retrograde axonal transport of heparin-binding growth factors is lacking [161]. into consideration [4]. **8.8. False positive results** On evaluating results of surgery after grafting the cord to the cauda equina in thoracolum‐ bar lesions, it should be noted that false positive results could be obtained from intercostal nerves (peripheral nerves) regenerating into the cauda equina (peripheral nerves) via nerve grafts (**Figure 11**).
### **10. Conclusions**
We have outlined current experimental and clinical experience applying nerve side grafts to the injured spinal cord. Nerve side grafting increases the incidence of nerve regeneration by applying additional grafts extending from the side of the donor end of the cord to the side of the recipient end. A partially regenerated cord cannot be surgically cut and end grafted; nerve side grafting can enhance regeneration through it without incriminating already regained function. Nevertheless, side grafting will fail, unless the gliosis is counteracted or lysed by chondroitinase ABC, sialidase, anti-Nogo, Rho inhibitors and other factors. Side grafting will also fail unless neurons are stimulated to produce neurites. Modulating the function of astrocytes by heparin, aspirin and other factors is one method to stimulate the intrinsic properties of the neurons to produce neurites. Side grafting should be augmented by artificial scaffolds and cellular transplants. Clinically, to prevent the cord adhering to the dura and reestablish CSF circulation, it has become the author's practice to wrap the cord graft construct with a silicone membrane. To establish a continuous drug and cell delivery system, an indwelling catheter is placed in the interstitium between the membrane and the dura; the dura is finally closed. Post-operative injection of paralysed muscles with neurotrophic factors stimulates neurite outgrowth by target-organ-derived neurotrophic support.
#### **Author details Figure 11.** In thoracolumbar lesions, false positive results could be obtained from intercostal nerves (peripheral nerves)
Sherif M. Amr\* grafts.
Address all correspondence to: sherifamrh@hotmail.com
The Department of Orthopaedics and Traumatology, Cairo University, Cairo, Egypt **9. Post-operative target organ derived trophic support**
at the cranial cord (white arrow) regenerating into to the caua equina (peripheral nerves) (yellow arrow) via nerve
Target organ derived neurotrophic factors, the so-called neurotrophins (nerve growth
involving dyneins [158, 159]. Neurotrophins contribute a lot to axonal progression; in the
#### **References** factor(NGF), brain-derived neurotrophic factor (BDNF), neurotrophin-3(NT-3) and neurotro‐ phin-4/5(NT-4/5)), are transported by retrograde axonal via an endosomal mechanism
[5] Reier PJ: Cellular transplantation strategies for spinal cord injury and translational neurobiology. Neurorx 2004 October;1(4):424–451. **10. Conclusions**
**10. Conclusions**
24 Recovery of Motor Function in Spinal Cord Injury
204Following Spinal Cord Injury
**Author details**
Sherif M. Amr\*
grafts.
**References**
2003;96(6):259–261.
Neurosci 2006;7:644–653.
We have outlined current experimental and clinical experience applying nerve side grafts to the injured spinal cord. Nerve side grafting increases the incidence of nerve regeneration by applying additional grafts extending from the side of the donor end of the cord to the side of the recipient end. A partially regenerated cord cannot be surgically cut and end grafted; nerve side grafting can enhance regeneration through it without incriminating already regained function. Nevertheless, side grafting will fail, unless the gliosis is counteracted or lysed by chondroitinase ABC, sialidase, anti-Nogo, Rho inhibitors and other factors. Side grafting will also fail unless neurons are stimulated to produce neurites. Modulating the function of astrocytes by heparin, aspirin and other factors is one method to stimulate the intrinsic properties of the neurons to produce neurites. Side grafting should be augmented by artificial scaffolds and cellular transplants. Clinically, to prevent the cord adhering to the dura and reestablish CSF circulation, it has become the author's practice to wrap the cord graft construct with a silicone membrane. To establish a continuous drug and cell delivery system, an indwelling catheter is placed in the interstitium between the membrane and the dura; the dura is finally closed. Post-operative injection of paralysed muscles with neurotrophic factors
stimulates neurite outgrowth by target-organ-derived neurotrophic support.
The Department of Orthopaedics and Traumatology, Cairo University, Cairo, Egypt
**Figure 11.** In thoracolumbar lesions, false positive results could be obtained from intercostal nerves (peripheral nerves) at the cranial cord (white arrow) regenerating into to the caua equina (peripheral nerves) (yellow arrow) via nerve
RM, translator). New York: Oxford University Press, 1991.
ment of spinal cord injuries. J Neuroeng Rehabil 2007;4:15.
transport of heparin-binding growth factors is lacking [161].
[1] DeFelipe J, Jones EG: Cajal's degeneration and regeneration of the nervous system (May
Target organ derived neurotrophic factors, the so-called neurotrophins (nerve growth factor(NGF), brain-derived neurotrophic factor (BDNF), neurotrophin-3(NT-3) and neurotro‐ phin-4/5(NT-4/5)), are transported by retrograde axonal via an endosomal mechanism involving dyneins [158, 159]. Neurotrophins contribute a lot to axonal progression; in the injured spinal cord, this stimulus is lost. After grafting the spinal cord, an important ques‐ tion is whether this stimulus can be restituted by injecting neurotrophins into target organs post-operatively, or whether axonal progression into specific nerves can be restituted by injecting neurotrophins into the specific muscles supplied by them. Experimental evidence points to this [158, 159]. By the same token, it can be questioned whether other neurotrophic factors and neurolyzing agents can be similarly injected into target muscles. Both in vitro and in vivo local infusion of fibroblastic growth factors (FGFs) have been found to rescue moto‐ neuron death induced by spinal cord injury [160]. However, evidence for retrograde axonal
[2] Raisman G: A promising therapeutic approach to spinal cord repair. J R Soc Med
[3] Samadikuchaksaraei A: An overview of tissue engineering approaches for manage‐
[4] Bradbury EJ, McMahon SB. Spinal cord repair strategies: why do they work? Nat Rev
Address all correspondence to: sherifamrh@hotmail.com
**9. Post-operative target organ derived trophic support**
- [15] Amr SM, Gouda A, Koptan WT, Galal AA, Abdel-Fattah DS, Rashed LA, Atta HM, Abdel-Aziz MT: Bridging defects in chronic spinal cord injury using peripheral nerve grafts combined with a chitosan-laminin scaffold and enhancing regeneration through them by cotransplantation with bone-marrow derived mesenchymal stem cells; a series of 14 patients. J Spinal Cord Medicine 2014;37(1):54–71 [2] Raisman G: A promising therapeutic approach to spinal cord repair. J R Soc Med 2003;96(6):259–261. [3] Samadikuchaksaraei A: An overview of tissue engineering approaches for manage‐ ment of spinal cord injuries. J Neuroeng Rehabil 2007;4:15.
- [16] Fawcett JW: Bridging spinal cord injuries. J Biol 2008;7:25. [4] Bradbury EJ, McMahon SB. Spinal cord repair strategies: why do they work? Nat Rev Neurosci 2006;7:644–653.
[31] Hill CE, Brodak DM, Bartlett Bunge M: Dissociated predegenerated peripheralnerve transplants for spinal cord injury repair: a comprehensive assessment of their effects on regeneration and functional recovery compared to Schwann cell transplants. J Neurotrauma 2012;29(12):2226–2243. doi: 10.1089/neu.2012.2377. PubMed PMID: 22655857; PubMed Central PMCID: PMC3472680. [17] Ray WZ: Mackinnon SE. Management of nerve gaps: autografts, allografts, nervetransfers, and end-to-side neurorrhaphy. Exp Neurol 2010 May ;223(1):77–85. doi: 10.1016/j.expneurol.2009.03.031. [18] Frostick SP, Yin Q, Kemp G: Schwann cells, neurotrophic factors, and peripheral nerve regeneration. Microsurgery 1998;18:397–405.
[17] Ray WZ: Mackinnon SE. Management of nerve gaps: autografts, allografts, nerve transfers, and end-to-side neurorrhaphy. Exp Neurol 2010 May ;223(1):77–85. doi:
[5] Reier PJ: Cellular transplantation strategies for spinal cord injury and translational
[18] Frostick SP, Yin Q, Kemp G: Schwann cells, neurotrophic factors, and peripheral nerve
[6] Hyun JK, Kim HW. Clinical and experimental advances in regeneration of spinal cord injury. J Tissue Eng. 2010 Nov 2; Vol. 2010 : 1–20: Article ID 650857. doi: 10.4061/2010/650857. PubMed PMID: 21350645; PubMed Central PMCID:
[19] Blottner D, Baumgarten HG: Neurotrophy and regeneration in vivo. Acta Anat
[7] David S, Aguayo AJ: Axonal elongation into peripheral nervous system ´bridges´ after
[20] Amr SM, Moharram AN: Repair of brachial plexus lesions by end-to-side side-to-side grafting neurorrhaphy: experience based on 11 cases. Microsurgery 2005, 25(2):126–
[8] Tuszynski MH, Petersen DA, Ray J, Baird A, Nakahara Y, Gage FH: Fibroblasts genetically modified to produce nerve growth factor induce robust neuritic ingrowth
central nervous system injury in adult rats. Science 1981;214:931–933.
regeneration is reversed by Schwann cells. Exp Neurol 1994;126:44–60.
after grafting to the spinal cord. Exp Neurol 1994;126:1–14.
[21] Rosoff WJ, Urbach JS, Esrick MA, McAllister RG, Richards LJ, Goodhill GJ. A new chemotaxis assay shows the extreme sensitivity of axons to molecular gradients. Nat Neurosci 2004;7(6):678–682. Epub 2004 May 25. Erratum in: Nat Neurosci. 2004;7(7):
[9] Guenard V, Aebischer P, Bunge RP: The astrocyte inhibition of peripheral nerve
[22] Tse TH, Chan BP, Chan CM, Lam J: Mathematical modeling of guided neurite extension in an engineered conduit with multiple concentration gradients of nerve growth factor
[10] Oudega M, Varon S, Hagg T: Regeneration of adult rat sensory axons into intraspinal nerve grafts: promoting effects of conditioning lesion and graft predegeneration. Exp
[23] Ciofani G, Sergi PN, Carpaneto J, Micera S. A hybrid approach for the control of axonal outgrowth: preliminary simulation results. Med Biol Eng Comput 2011;49(2):163–170.
[11] Liu S, Kadi K, Boisset N, Lacroix C, Said G, Tadie M: Reinnervation of denervated lumbar ventral roots and their target muscle by thoracic spinal motoneurons via an implanted nerve autograft in adult rats after spinal cord injury. J Neurosci Res 1999
[24] Ciofani G, Raffa V, Menciassi A, Micera S, Dario P: A drug delivery system based on alginate microspheres: mass-transport test and in vitro validation. Biomed Microdevi‐
[12] Liu S, Aghakhani N, Boisset N, Said G, Tadie M: Innervation of the caudal denervat‐ ed ventral roots and their target muscles by the rostral spinal motoneurons after implanting a nerve autograft in spinal cord-injured adult marmosets. J Neurosurg 2001
[25] Holmquist B, Kanje M, Kerns JM, Danielsen N. A mathematical model for regeneration rate and initial delay following surgical repair of peripheral nerves. J Neurosci
[13] Dam-Hieu P, Liu S, Choudhri T, Said G, Tadie M: Regeneration of primary sensory axons into the adult rat spinal cord via a peripheral nerve graft bridging the lumbar
[26] Cover TM, Thomas JA: Elements of Information Theory, 2nd ed. Hoboken, New Jersey:
dorsal roots to the dorsal column. J Neurosci Res 2002 May 1;68(3):293–304.
[27] Effenberger F. A Primer on Information Theory, with Applications to Neuroscience. In: Rakocevic G., Djukic T., Filipovic N., Milutinović V. (editors) Computational Medicine
[14] Tadie M, Liu S, Robert R, Guiheneuc P, Pereon Y, Perrouin-Verbe B, Mathe JF: Partial return of motor function in paralyzed legs after surgical bypass of the lesion site by nerve autografts three years after spinal cord injury. J Neurotrauma 2002 Aug;19(8):
[28] Standards for Neurological Classification of Spinal Injury. American Spinal Injury
[15] Amr SM, Gouda A, Koptan WT, Galal AA, Abdel-Fattah DS, Rashed LA, Atta HM, Abdel-Aziz MT: Bridging defects in chronic spinal cord injury using peripheral nerve grafts combined with a chitosan-laminin scaffold and enhancing regeneration through them by cotransplantation with bone-marrow derived mesenchymal stem cells; a series
[29] Harel NY, Strittmatter SM: Can regenerating axons recapitulate developmental guidance during recovery from spinal cord injury? Nat Rev Neurosci 2006;7:603–616.
[30] Côté MP, Amin AA, Tom VJ, Houle JD: Peripheral nerve grafts support regeneration
doi: 10.1007/s11517-010-0687-x. Epub 2010 Oct 6. PubMed PMID: 20924708.
(NGF). Ann Biomed Eng 2007;35(9):1561–1572. Epub 2007 May 23.
Methods. 1993 Jun;Vol.48(1–2):27–33. PubMed PMID: 8377520.
in Data Mining and Modeling. New York: Springer 2013.
after spinal cord injury. Neurotherapeutics 2011;8(2):294–303.
of 14 patients. J Spinal Cord Medicine 2014;37(1):54–71
[16] Fawcett JW: Bridging spinal cord injuries. J Biol 2008;7:25.
Association (ASIA) Illinois: Chicago, 1996.
10.1016/j.expneurol.2009.03.031.
1994;150:235–245.
PMC3042682.
26 Recovery of Motor Function in Spinal Cord Injury
206Following Spinal Cord Injury
ces 2007;9:395–403.
909–916.
Jan;94(1 Suppl):82–90.
Jun 1;56(5):506–517.
Neurol 1994;129:194–206.
John Wiley & Sons, Inc., 2006.
146.
785.
regeneration. Microsurgery 1998;18:397–405.
neurobiology. Neurorx 2004 October;1(4):424–451.
[56] Tom VJ, Sandrow-Feinberg HR, Miller K, Santi L, Connors T, Lemay MA, Houle JD: Combining peripheral nerve grafts and chondroitinase promotes functional axonal regeneration in the chronically injured spinal cord. J Neurosci 2009;29:14881–14890 [PMC free article] [PubMed]. doi: 10.1016/j.neuron.2009.09.016. PubMed PMID: 19874785; PubMedCentral PMCID:PMC2773653. [44] Neumann S, Bradke F, Tessier-Lavigne M, Basbaum AI: Regeneration of sensory axons within the injured spinal cord induced by intraganglionic cAMP elevation. Neuron.
[44] Neumann S, Bradke F, Tessier-Lavigne M, Basbaum AI: Regeneration of sensory axons within the injured spinal cord induced by intraganglionic cAMP elevation. Neuron.
[31] Hill CE, Brodak DM, Bartlett Bunge M: Dissociated predegenerated peripheralnervetransplants for spinal cord injury repair: a comprehensive assessment of their effects
on regeneration and functional recovery compared to Schwann cell transplants. J Neurotrauma 2012;29(12):2226–2243. doi: 10.1089/neu.2012.2377. PubMed PMID:
[45] Qiu J, Cai D, Dai H, McAtee M, Hoffman PN, Bregman BS, Filbin MT: Spinal axon regeneration induced by elevation of cyclic AMP. Neuron 2002;34:895–903 [PubMed].
[46] Ferguson TA, Son YJ: Extrinsic and intrinsic determinants of nerve regeneration. J Tissue Eng 2011;2(1):2041731411418392. doi: 10.1177/2041731411418392. Epub 2011 Sep
[32] Bray GM, Villegas-Perez MP, Vidal-Sanz M, Aguayo AJ: The use of peripheral nerve grafts to enhance neuronal survival, promote growth and permit terminal reconnec‐
[47] Tom VJ, Steinmetz MP, Miller JH, Doller CM, Silver J: Studies on the development and behavior of the dystrophic growth cone, the hallmark of regeneration failure, in an in vitro model of the glial scar and after spinal cord injury. J Neurosci 2004;24(29):6531–
[34] Guzen FP, de Almeida Leme RJ, de Andrade MS, de Luca BA, Chadi G: Glial cell linederived neurotrophic factor added to a sciatic nerve fragment grafted in a spinal cord gap ameliorates motor impairments in rats and increases local axonal growth. Restor
[33] Bunge MB: Bridging areas of injury in the spinal cord. Neuroscientist 2001;7:325–339.
[48] Blits B, Carlstedt TP, Ruitenberg MJ, de Winter F, Hermens WT, Dijkhuizen PA, Claasens JW, Eggers R, Sluis R van der, Tenenbaum L, Boer GJ, Verhaagen J: Rescue and sprouting of motoneurons following ventral root avulsion and reimplantation combined with intraspinal adeno-associated viral vector-mediated expression of glial cell line-derived neurotrophic factor or brain-derived neurotrophic factor. Exp Neurol
[35] Cheng H, Cao Y, Olson L: Spinal cord repair in adult paraplegic rats partial restora‐
[36] Lee YS, Hsiao I, Lin VW: Peripheral nerve grafts and aFGF restore partial hindlimb
function in adult paraplegic rats. J Neurotrauma 2002;19:1203–1216.
[49] Yiu G, He Z: Glial inhibition of CNS axon regeneration. Nat Rev Neurosci 2006;7:617–
[37] Campos L, Meng Z, Hu G, Chiu DT, Ambron RT, Martin JH: Engineering novel spinal circuits to promote recovery after spinal injury. J Neurosci 2004;24:2090–2101.
[50] McCreedy DA, Sakiyama-Elbert SE: Combination therapies in the CNS: engineering the environment. Neurosci Lett 2012;519(2):115–121. doi: 10.1016/j.neulet.2012.02.025.
[38] Gauthier P, Rega P, Lammari-Barreault N, Polentes J: Functional reconnections established by central respiratory neurons regenerating axons into a nerve graft bridging the respiratory centers to the cervical spinal cord. J Neurosci Res 2002;70:65–
[51] Bradbury EJ, Carter LM: Manipulating the glial scar: chondroitinase ABC as a therapy for spinal cord injury. Brain Res Bull 2011;84:306316 [PubMed: 20620201].
[39] Decherchi P, Lammari-Barreault N, Gauthier P: Regeneration of respiratory path‐
[52] Galtrey CM, Fawcett JW: The role of chondroitin sulfate proteoglycans in regeneration and plasticity in the central nervous system. Brain Res Rev 2007;54:118 [PubMed:
[40] Aguayo AJ, Bray GM, Rasminsky MM, Zwimpfer T, Carter D, Vidal-Sanz M: Synaptic connections made by axons regenerating in the central nervous system of adult
ways within spinal peripheral nerve grafts. Exp Neurol 1996;137:1–14.
[53] Yamagata T, Saito H, Habuchi O, Suzuki S: Purification and properties of bacterial chondroitinases and chondrosulfatases. J Biol Chem 1968;243:15231535 [PubMed:
[41] Plant GW, Bates ML, Bunge MB: Inhibitory proteoglycan immunoreactivity is higher at the caudal than the rostral Schwann cell graft-transected spinal cord interface. Mol
[54] Bradbury EJ, Moon LD, Popat RJ, King VR, Bennett GS, Patel PN, Fawcett JW, McMa‐ hon SB: Chondroitinase ABC promotes functional recovery after spinal cord injury.
[42] Thuret S, Moon LDF, Gage FH: Therapeutic interventions after spinal cord injury. Nat
[55] Houle JD, Tom VJ, Mayes D, Wagoner G, Phillips N, Silver J: Combining an autologous peripheral nervous system "bridge" and matrix modification by chondroitinase allows robust, functional regeneration beyond a hemisection lesion of the adult rat spinal cord.
[43] Kadoya K, Tsukada S, Lu P, Coppola G, Geschwind D, Filbin MT, Blesch A,Tuszyn‐ ski MH: Combined intrinsic and extrinsic neuronal mechanisms facilitate bridging axonal regeneration one year after spinal cord injury. Neuron. 2009;64(2):165–172.
13. PubMed PMID: 22292105; PubMed Central PMCID: PMC3251917.
tions in the central nervous system of adult rats. J Exp Biol 1987;132:5–19.
22655857; PubMed Central PMCID: PMC3472680.
tion of hind limb function. Science 1996;273:510–513.
2002;34:885–893 [PubMed].
28 Recovery of Motor Function in Spinal Cord Injury
208Following Spinal Cord Injury
6539. PubMed PMID: 15269264.
Neurol Neurosci 2009;27(1):1–16.
Epub 2012 Feb 17. Review. PMID: 22343313.
mammals. J Exp Biol 1990;153:199–224.
Nature 2002;416:636–640 [PubMed: 11948352].
J Neurosci 2006;26(28):7405–7415.
Cell Neurosci 2001;17:471–487.
Rev Neurosci 2006;7:628–643.
2004;189(2):303–316.
627.
81.
17222456].
5647268].
gous peripheral nervous system "bridge" and matrix modification by chondroitinase
lesioned spinal cord. J Control Release 2010;147(1):24–29. doi: 10.1016/j.jconrel. 2010.06.026. allows robust, functional regeneration beyond a hemisection lesion of the adult rat spinal cord. J Neurosci 2006;26(28):7405–7415.
the contusion model of spinal cord injury in rats. Iran J Basic Med Sci 2014;17(9):685– 693. PubMed PMID: 25691946; PubMed Central PMCID:PMC4322153. lesioned spinal cord. J Control Release 2010;147(1):24–29. doi: 10.1016/j.jconrel. 2010.06.026.
[77] Mountney A, Zahner MR, Sturgill ER, Riley CJ, Aston JW, Oudega M, Schramm LP, Hurtado A, Schnaar RL: Sialidase, chondroitinase ABC, and combination therapy after spinal cord contusion injury. J Neurotrauma 2013;30(3):181–190. doi: 10.1089/neu. 2012.2353 [67] Hunanyan AS, García-Alías G, Alessi V, Levine JM, Fawcett JW, Mendell LM, Arvanian VL: Role of chondroitin sulfate proteoglycans in axonal conduction in Mammalian spinal cord. J Neurosci 2010;30(23):7761–7769. doi: 10.1523/JNEUROSCI.4659-09.2010.
lesioned spinal cord. J Control Release 2010;147(1):24–29. doi: 10.1016/j.jconrel.
allows robust, functional regeneration beyond a hemisection lesion of the adult rat
[67] Hunanyan AS, García-Alías G, Alessi V, Levine JM, Fawcett JW, Mendell LM, Arvanian VL: Role of chondroitin sulfate proteoglycans in axonal conduction in Mammalian spinal cord. J Neurosci 2010;30(23):7761–7769. doi: 10.1523/JNEUROSCI.4659-09.2010.
[56] Tom VJ, Sandrow-Feinberg HR, Miller K, Santi L, Connors T, Lemay MA, Houle JD: Combining peripheral nerve grafts and chondroitinase promotes functional axonal regeneration in the chronically injured spinal cord. J Neurosci 2009;29:14881–14890
[68] Cheng CH, Lin CT, Lee MJ, Tsai MJ, Huang WH, Huang MC, Lin YL, Chen CJ, Huang WC, Cheng H: Local delivery of high-dose chondroitinase ABC in the sub-acute stage promotes axonal outgrowth and functional recovery after complete spinal cord transection. PLoS One. 2015;10(9):e0138705. doi:10.1371/journal.pone.0138705. eCol‐
[57] Hwang DH, Kim HM, Kang YM, Joo IS, Cho CS, Yoon BW, Kim SU, Kim BG. Combination of multifaceted strategies to maximize the therapeutic benefits of neural stem cell transplantation for spinal cord repair. Cell Transplant. 2011;20(9):1361–1379.
[69] Shields LB, Zhang YP, Burke DA, Gray R, Shields CB: Benefit of chondroitinase ABC on sensory axon regeneration in a laceration model of spinal cord injury in the rat. Surg
[58] Kumar P, Choonara YE, Modi G, Naidoo D, Pillay V: Multifunctional therapeutic delivery strategies for effective neuro-regeneration following traumatic spinal cord
[70] Iseda T, Okuda T, Kane-Goldsmith N, Mathew M, Ahmed S, Chang YW, Young W, Grumet M: Single, high-dose intraspinal injection of chondroitinase reduces glycosa‐ minoglycans in injured spinal cord and promotes corticospinal axonal regrowth after hemisection but not contusion. J Neurotrauma 2008;25(4):334–349. doi: 10.1089/neu.
[59] Jain A, Kim YT, McKeon RJ, Bellamkonda RV: In situ gelling hydrogels for conformal repair of spinal cord defects, and local delivery of BDNF after spinal cord injury.
[71] Caggiano AO, Zimber MP, Ganguly A, Blight AR, Gruskin EA: Chondroitinase ABCI improves locomotion and bladder function following contusion injury of the rat spinal
[61] Fouad K, Schnell L, Bunge MB, Schwab ME, Liebscher T, Pearse DD. Combining Schwann cell bridges and olfactory-ensheathing glia grafts with chondroitinase promotes locomotor recovery after complete transection of the spinal cord. J
[60] Kreider BQ, Messing A, Doan H, Kim SU, Lisak RP, Pleasure DE. Enrichment of Schwann cell cultures from neonatal rat sciatic nerve by differential adhesion. Brain
[72] Mondello SE, Jefferson SC, Tester NJ, Howland DR: Impact of treatment duration and lesion size on effectiveness of chondroitinase treatment post-SCI. Exp Neurol
[62] Wilems TS, Pardieck J, Iyer N, Sakiyama-Elbert SE: Combination therapy of stem cell derived neural progenitors and drug delivery of anti-inhibitory molecules for spinal cord injury. Acta Biomater 2015. pii: S1742–7061(15)30112–4. doi: 10.1016/j.actbio.2015.09.018. [Epub ahead of print]
[73] Ma Y, Liu M, Li Y: Secretion of bacterial chondroitinase ABC from bone marrow stromal cells by glycosylation site mutation: a promising approach for axon regeneration. Med Hypotheses 2011;77(5):914–916. doi: 10.1016/j.mehy.2011.08.010. Epub 2011 Aug 31.
[74] Kumar P, Choonara YE, Modi G, Naidoo D, Pillay V(1): Multifunctional therapeutic delivery strategies for effective neuro-regeneration following traumatic spinal cord
[63] Lee H, McKeon RJ, Bellamkonda RV. Sustained delivery of thermostabilized chABC enhances axonal sprouting and functional recovery after spinal cord injury. Proc Natl Acad Sci U S A. 2010 Feb 23;107(8):3340–3345. doi: 10.1073/pnas.0905437106. Epub 2009
[75] Steinmetz MP, Horn KP, Tom VJ, Miller JH, Busch SA, Nair D, Silver DJ, Silver J: Chronic enhancement of the intrinsic growth capacity of sensory neurons combined with the degradation of inhibitory proteoglycans allows functional regeneration of sensory axons through the dorsal root entry zone in the mammalian spinal cord. J
[64] Tom VJ, Houlé JD: Intraspinal microinjection of chondroitinase ABC following injury promotes axonal regeneration out of a peripheral nerve graft bridge. Exp Neurol
Nov 2. PubMed PMID: 19884507; PubMed Central PMCID: PMC2840440.
[76] Sarveazad A, Bakhtiari M, Babahajian A, Janzade A, Fallah A, Moradi F, Soleimani M, Younesi M, Goudarzi F, Mohammad Taghi Joghataei: Comparison of human adiposederived stem cells and chondroitinase ABC transplantation on locomotor recovery in
10.1016/j.jconrel.2015.06.031 [Epub ahead of print] [PubMed PMID: 26122130]. [66] Hyatt AJ, Wang D, Kwok JC, Fawcett JW, Martin KR: Controlled release of chondroi‐ tinase ABC from fibrin gel reduces the level of inhibitory glycosaminoglycan chains
[65] Wilems TS, Sakiyama-Elbert SE: Sustained dual drug delivery of anti-inhibitory molecules for treatment of spinal cord injury. J Control Release 2015;213:103–111. doi:
2015;267:64–77. doi: 10.1016/j.expneurol.2015.02.028. Epub 2015 Feb 26.
Neurol 2008;69(6):568–577; discussion 577. doi: 10.1016/j.surneu.2008.02.009.
injury. Curr Pharm Des 2015;21(12):1517–1528 [PubMed PMID: 25594407].
doi: 10.3727/096368910X557155. Epub 2011 Mar 7. PubMed PMID: 21396156.
2010.06.026.
30 Recovery of Motor Function in Spinal Cord Injury
210Following Spinal Cord Injury
[PMC free article] [PubMed].
spinal cord. J Neurosci 2006;26(28):7405–7415.
lection 2015.
2007.0289.
cord. J Neurotrauma 2005;22(2):226–239.
Neurosci 2005;25:1169–1178 [PubMed].
Res 1981;207:433444 [PubMed: 7008901].
Biomaterials 2006;27:497504 [PubMed: 16099038].
injury. Curr Pharm Des 2015;21(12):1517–1528.
Neurosci 2005;25(35):8066–8076.
2008;211(1):315–319.
Younesi M, Goudarzi F, Mohammad Taghi Joghataei: Comparison of human adiposederived stem cells and chondroitinase ABC transplantation on locomotor recovery in
Lung Cell Mol Physiol 2004;287(1):L191–L200. Epub 2004 Feb 13. PubMed PMID: 14966081. [88] Hering TM, Beller JA, Calulot CM, Centers A, Snow DM: Proteoglycans of reactive rat cortical astrocyte cultures: abundance of N-unsubstituted glucosamine-enriched
[99] Santos-Silva A, Fairless R, Frame MC, Montague P, Smith GM, Toft A, Riddell JS, Barnett SC: FGF/heparin differentially regulates Schwann cell and olfactory ensheath‐ ing cell interactions with astrocytes: a role in astrocytosis. J Neurosci 2007;27(27):7154– 7167. heparan sulfate. Matrix Biol 2015;41:8–18. doi: 10.1016/j.matbio.2014.11.006. [89] Yates D: Synaptogenesis. Asynaptic bridge. Nat Rev Neurosci 2016;17,135. doi:10.1038/ nrn.2016.12 Published online 11 February 2016. [90] Harter K, Levine M, Henderson SO: Anticoagulation drug therapy: a review. West J
[88] Hering TM, Beller JA, Calulot CM, Centers A, Snow DM: Proteoglycans of reactive rat cortical astrocyte cultures: abundance of N-unsubstituted glucosamine-enriched
the contusion model of spinal cord injury in rats. Iran J Basic Med Sci 2014;17(9):685–
Hurtado A, Schnaar RL: Sialidase, chondroitinase ABC, and combination therapy afterspinal cord contusion injury. J Neurotrauma 2013;30(3):181–190. doi: 10.1089/neu.
[89] Yates D: Synaptogenesis. Asynaptic bridge. Nat Rev Neurosci 2016;17,135. doi:10.1038/
[77] Mountney A, Zahner MR, Sturgill ER, Riley CJ, Aston JW, Oudega M, Schramm LP,
[90] Harter K, Levine M, Henderson SO: Anticoagulation drug therapy: a review. West J Emerg Med 2015;16(1):11–17. doi: 10.5811/westjem.2014.12.22933. Epub 2015 Jan 12.
[78] Lang BT, Cregg JM, DePaul MA, Tran AP, Xu K, Dyck SM, Madalena KM, Brown BP,Weng YL, Li S, Karimi-Abdolrezaee S, Busch SA, Shen Y, Silver J: Modulation of the
[91] Kahraman A, Kahveci R: Evaluating the effect of polytetrafluoroethylene and extrac‐ tum cepae-heparin-allantoin gel in peripheral nerve injuries in a rat model. Plast Surg (Oakv) 2015 Spring;23(1):9–14. PubMed PMID: 25821766; PubMed Central PMCID:
proteoglycan receptor PTPσ promotes recovery after spinal cord injury. Nature 2015;518(7539):404–408. doi: 10.1038/nature13974. Epub 2014 Dec 3. PubMed PMID:
[92] Akbari H, Rahimi AA, Ghavami Y, Mousavi SJ, Fatemi MJ. Effect of heparin on postoperative adhesion in flexor tendon surgery of the hand. J Hand Microsurg 2015;7(2): 244–249. Epub 2015 Aug 26. PubMed PMID: 26578825; PubMed Central PMCID:
[80] Tetzlaff W, Alexander SW, Miller FD, Bisby MA: Response of facial and rubrospinal neurons to axotomy: changes in mRNA expression for cytoskeletal proteins and
[79] Silver L, Michael JV, Goldfinger LE, Gallo G: Activation of PI3K and R-Ras signaling promotes the extension of sensory axons on inhibitory chondroitin sulfate proteogly‐ cans. Dev Neurobiol 2014;74(9):918–933. doi: 10.1002/dneu.22174. Epub 2014 Mar 27.
[93] Aydoseli A, Tahta A, Aras Y, Sabancı A, Keskin M, Balik E, Onder S, Sencer A, Izgi N: Use of antifibrotics to prevent ventriculoperitoneal shunt complications due to intraabdominal fibrosis: experimental study in a rat model. J Neurol Surg A Cent Eur Neurosurg 2015;76(3):219–223. doi: 10.1055/s-0034-1389369. Epub 2015 Mar 26.
[81] Sengottuvel V, Leibinger M, Pfreimer M, Andreadaki A, Fischer D: Taxol facilitates axon regeneration in the mature CNS. J Neurosci 2011;31:2688–2699 [PubMed]. [82] Bates CA, Becker CG, Miotke JA, Meyer RL: Expression of polysialylated NCAM but not L1 or N-cadherin by regenerating adult mouse optic fibers in vitro. Exp Neurol
[94] Alonso Jde M, Rodrigues KA, Yamada AL, Watanabe MJ, Alves AL, Rodrigues CA, Hussni CA: Peritoneal reactivity evaluation in horses subjected to experimental small colon enterotomy and treated with subcutaneous heparin. Vet Med Int 2014;2014:385392. doi: 10.1155/2014/385392. Epub 2014 Nov 11. PubMed PMID:
[83] Chaisuksunt V, Zhang Y, Anderson PN, et al: Axonal regeneration from CNS neu‐ rons in the cerebellum and brainstem of adult rats: correlation with the patterns of expression and distribution of messenger RNAs for L1, CHL1, c-jun and growth-
[95] Ho WS, Ying SY, Chan PC, Chan HH: Use of onion extract, heparin, allantoin gel in prevention of scarring in chinese patients having laser removal of tattoos: a prospective randomized controlled trial. Dermatol Surg 2006;32(7):891–896. PubMed PMID:
[84] Chen J, Wu J, Apostolova I, et al: Adeno-associated virus-mediated L1 expression promotes functional recovery after spinal cord injury. Brain 2007;130:954–969
[96] Billings PC, Pacifici M: Interactions of signaling proteins, growth factors and other proteins with heparan sulfate: mechanisms and mysteries. Connect Tissue Res
[85] Roonprapunt C, Huang W, Grill R, et al: Soluble cell adhesion molecule L1-Fc pro‐ motes locomotor recovery in rats after spinal cord injury. J Neurotrauma 2003;20:871–
[97] Olczyk P, Mencner Ł, Komosinska-Vassev K: Diverse roles of heparan sulfate and heparin in wound repair. Biomed Res Int 2015;2015:549417. doi: 10.1155/2015/549417. Epub 2015 Jul 7. Review. PubMed PMID: 26236728; PubMed Central PMCID:
[86] Chen Y and Swanson RA: Astrocytes and brain injury. Rev Article J Cereb Blood Flow
[98] Newman DR, Li CM, Simmons R, Khosla J, Sannes PL: Heparin affects signaling pathways stimulated by fibroblast growth factor-1 and -2 in type II cells. Am J Physiol
[87] Mohn TC, Koob AO: Adult astrogenesis and the etiology of cortical neurodegenera‐
heparan sulfate. Matrix Biol 2015;41:8–18. doi: 10.1016/j.matbio.2014.11.006.
693. PubMed PMID: 25691946; PubMed Central PMCID:PMC4322153.
Review. PubMed PMID: 25671002; PubMed Central PMCID: PMC4307693.
nrn.2016.12 Published online 11 February 2016.
25470046; PubMed Central PMCID: PMC4336236.
GAP-43. J Neurosci 1991;11:2528–2544 [PubMed].
PMC4364148.
2012.2353
32 Recovery of Motor Function in Spinal Cord Injury
212Following Spinal Cord Injury
PMC4642480.
16875470.
[PubMed].
PMC4508384.
882 [PubMed].
Metab 2003;23(2):137–149.
PubMed PMID: 25811104.
1999;155:128–139 [PubMed].
25436172; PubMed Central PMCID: PMC4243600.
associated protein-43. Neuroscience 2000;100:87–108 [PubMed].
2015;56(4):272–280. doi: 10.3109/03008207.2015.1045066.
tion. J Exp Neurosci 2015:9(s) 25–34. doi:10.4137/JEn.s25520.
PubMed Central PMCID: PMC2981799.
[121] Jin Y, Bouyer J, Shumsky JS, Haas C, Fischer I: Transplantation of neural progenitor cells in chronic spinal cord injury. Neuroscience 2016;320:69–82. doi: 10.1016/j.neuro‐ science.2016.01.066. Epub 2016 Feb 4. PubMedPMID: 26852702. [110] Straley KS, Foo CW, Heilshorn SC: Biomaterial design strategies for the treatment ofspinal cord injuries. J Neurotrauma 2010;27(1):1–19. doi: 10.1089/neu.2009.0948. Review. PubMed PMID: 19698073; PubMed Central PMCID: PMC2924783.
[110] Straley KS, Foo CW, Heilshorn SC: Biomaterial design strategies for the treatment of spinal cord injuries. J Neurotrauma 2010;27(1):1–19. doi: 10.1089/neu.2009.0948.
Lung Cell Mol Physiol 2004;287(1):L191–L200. Epub 2004 Feb 13. PubMed PMID:
[111] Tetzlaff W, Okon EB, Karimi-Abdolrezaee S, Hill CE, Sparling JS, Plemel JR, Plunet WT, Tsai EC, Baptiste D, Smithson LJ, Kawaja MD, Fehlings MG, Kwon BK: A systematic review of cellular transplantation therapies for spinal cord injury. J Neurotrauma
[99] Santos-Silva A, Fairless R, Frame MC, Montague P, Smith GM, Toft A, Riddell JS, Barnett SC: FGF/heparin differentially regulates Schwann cell and olfactory ensheath‐ ing cell interactions with astrocytes: a role in astrocytosis. J Neurosci 2007;27(27):7154–
[112] Duncan ID, Milward EA: Glial cell transplants: experimental therapies of myelin
[100] Yi-Chao Hsu, Su-Liang Chen, Dan-Yen Wang: Ing-Ming Chiu stem cell-based thera‐
[113] Santos-Silva A, Fairless R, Frame MC, Montague P, Smith GM, Toft A, Riddell JS, Barnett SC: FGF/heparin differentially regulates Schwann cell and olfactory ensheath‐ ing cell interactions with astrocytes: a role in astrocytosis. J Neurosci 2007;27(27):7154–
[101] Walicke P, Cowan WM, Ueno N, Baird A, Guillemin R: Fibroblast growth factor promotes survival of dissociated hippocampal neurons and enhances neurite exten‐
[114] Ramon-Cueto A, Plant GW, Avila J, Bunge MB: Long-distance axonal regeneration in the transected adult rat spinal cord is promoted by olfactory ensheathing glia trans‐
[102] Nagayasu T, Miyata S, Hayashi N, Takano R, Kariya Y, Kamei K: Heparin structures in FGF-2-dependent morphological transformation of astrocytes. J Biomed Mater Res
[115] Lu P, Yang H, Culbertson M, Graham L, Roskams AJ, Tuszynski MH: Olfactory ensheathing cells do not exhibit unique migratory or axonal growth-promoting
[103] Poulsen FR, Lagord C, Courty J, Pedersen EB, Barritault D, Finsen B: Increased synthesis of heparin affin regulatory peptide in the perforant path lesioned mouse hippocampal formation. Exp Brain Res 2000;135(3):319–330. PubMed PMID: 11146810.
properties after spinal cord injury. J Neurosci 2006;26:11120–11130 [PubMed]. [116] Pearse DD, Sanchez AR, Pereira FC, Andrade CM, Puzis R, Pressman Y, Golden K, Kitay BM, Blits B, Wood PM, Bunge MB: Transplantation of Schwann cells and/or olfactory ensheathing glia into the contused spinal cord: survival, migration, axon
[104] Wanaka A, Carroll SL, Milbrandt J: Developmentally regulated expression of pleiotro‐ phin, a novel heparin binding growth factor, in the nervous system of the rat. Brain Res
[117] Ruitenberg MJ, Plant GW, Hamers FP, Wortel J, Blits B, Dijkhuizen PA, Gispen WH, Boer GJ, Verhaagen J: Ex vivo adenoviral vector-mediated neurotrophin gene transfer to olfactory ensheathing glia: effects on rubrospinal tract regeneration, lesion size, and functional recovery after implantation in the injured rat spinal cord. J Neurosci
[106] Lau E, Margolis RU: Inhibitors of slit protein interactions with the heparin sulphate proteoglycan glypican-1: potential agents for the treatment of spinal cord injury. Clin Exp Pharmacol Physiol 2010;37(4):417–421. doi: 10.1111/j.1440-1681.2009.05318.x.
[105] Hagino S, Iseki K, Mori T, Zhang Y, Sakai N, Yokoya S, Hikake T, Kikuchi S, Wanaka A: Expression pattern of glypican-1 mRNA after brain injury in mice. Neurosci Lett
[118] Alexanian AR, Crowe MJ, Kurpad SN: Efficient differentiation and integration of lineage-restricted neural precursors in the traumatically injured adult cat spinal cord.
[107] Modi KK, Sendtner M, Pahan K: Up-regulation of ciliary neurotrophic factor in astrocytes by aspirin: implications for remyelination in multiple sclerosis. J Biol
[119] Cao QL, Zhang YP, Howard RM, Walters WM, Tsoulfas P, Whittemore SR: Pluripotent stem cells engrafted into the normal or lesioned adult rat spinal cord are restricted to
cord injury. J Neural Eng 2011;8(4):046033. doi: 10.1088/1741-2560/8/4/046033.
[108] Khaing ZZ, Milman BD, Vanscoy JE, Seidlits SK, Grill RJ, Schmidt CE: High molecu‐ lar weight hyaluronic acid limits astrocyte activation and scar formation after spinal
[120] Macias MY, Syring MB, Pizzi MA, Crowe MJ, Alexanian AR, Kurpad SN: Pain with no gain: allodynia following neural stem cell transplantation in spinal cord injury. Exp
[109] Madigan NN, McMahon S, O'Brien T, Yaszemski MJ, Windebank AJ: Current tissue engineering and novel therapeutic approaches to axonal regeneration following spinal cord injury using polymer scaffolds. Respir Physiol Neurobiol 2009;169(2):183–199. doi: 10.1016/j.resp.2009.08.015. Epub 2009 Sep 6. Review. PubMed PMID: 19737633;
association, and functional recovery. Glia 2007;55:976–1000 [PubMed].
Dev Brain Res 1993;72(1):133–144. PubMed PMID: 8453763.
Review. PubMed PMID: 19698073; PubMed Central PMCID: PMC2924783.
2011;28(8):1611–1682. Epub 2010 Apr 20. Review.
diseases. Brain Pathol 1995;5:301–310 [PubMed].
sion. Proc Natl Acad Sci U S A 1986;83(9):3012–3016.
py in neural repair. Biomed J 2013;36:98–105.
plants. J Neurosci 1998;18:3803–3815 [PubMed].
A 2005;74(3):374–380. PubMed PMID: 15973728.
2003;23:7045–7058 [PubMed].
J Neurosci Methods 2006;150:41–46 [PubMed].
2003;349(1):29–32. PubMed PMID: 12946579.
Neurol 2006;201:335–348 [PubMed].
PubMed Central PMCID: PMC2981799.
a glial lineage. Exp Neurol 2001;167:48–58 [PubMed].
Chem 2013;288(25):18533–18545. doi: 10.1074/jbc.M112.447268.
7167.
7167.
34 Recovery of Motor Function in Spinal Cord Injury
214Following Spinal Cord Injury
14966081.
[144] Peláez RJ, Afonso CN, Vega F, Recio-Sánchez G, Torres-Costa V, Manso-Silván M, García-Ruiz JP, Martín-Palma RJ: Laser fabrication of porous silicon-based platforms for cell culturing. J Biomed Mater Res B Appl Biomater 2013;101(8):1463–1468. doi: 10.1002/jbm.b.32966. [132] Sullivan R, Duncan K, Dailey T, Kaneko Y, Tajiri N, Borlongan CV: A possible new focus for stroke treatment—migrating stem cells. Expert Opin Biol Ther 2015;15(7):949– 958. doi: 10.1517/14712598.2015.1043264. [133] Klingener M, Chavali M, Aguirre A: N-cadherin promotes recruitment and migra‐
[132] Sullivan R, Duncan K, Dailey T, Kaneko Y, Tajiri N, Borlongan CV: A possible new focus for stroke treatment—migrating stem cells. Expert Opin Biol Ther 2015;15(7):949–
[121] Jin Y, Bouyer J, Shumsky JS, Haas C, Fischer I: Transplantation of neural progenitor cells in chronic spinal cord injury. Neuroscience 2016;320:69–82. doi: 10.1016/j.neuro‐
[133] Klingener M, Chavali M, Aguirre A: N-cadherin promotes recruitment and migration of neural progenitor cells from the SVZ stem cell niche into demyelinated lesions. Int J Dev Neurosci 2015;47(Pt A):107. doi: 10.1016/j.ijdevneu.2015.04.292. PubMed PMID:
hedgehog results in improved function and white matter sparing in the spinal cords ofadult rats after contusion. Spine J 2004;4:16–26 [PubMed].
[122] Bambakidis NC, Miller RH: Transplantation of oligodendrocyte precursors and sonic
[134] Kim HJ, Shaker MR, Cho B, Cho HM, Kim H, Kim JY, Sun W: Dynamin-related protein 1 controls the migration and neuronal differentiation of subventricular zone-derived
[123] Cao QL, Howard RM, Dennison JB, Whittemore SR. Differentiation of engrafted neuronal-restricted precursor cells is inhibited in the traumatically injured spinal cord.
[135] Chen Q, Zhang M, Li Y, Xu D, Wang Y, Song A, Zhu B, Huang Y, Zheng JC: CXCR7 mediates neural progenitor cells migration to CXCL12 independent of CXCR4. Stem Cells 2015;33(8):2574–2585. doi: 10.1002/stem.2022. Epub 2015 May 13. PubMed PMID:
[125] Yano S, Kuroda S, Lee JB, Shichinohe H, Seki T, Ikeda J, Nishimura G, Hida K, Tamura M, Iwasaki Y: In vivo fluorescence tracking of bone marrow stromal cells transplant‐ ed into a pneumatic injury model of rat spinal cord. J Neurotrauma 2005;22:907–918
[124] Ankeny DP, McTigue DM, Jakeman LB: Bone marrow transplants provide tissue protection and directional guidance for axons after contusive spinal cord injury in rats.
[136] Potter K, Saifuddin A: MRI of chronic spinal cord injury. Br J Radiol 2003;76:347–352. [137] Li Y, Raisman G: Sprouts from cut corticospinal axons persist in the presence of astrocytic scarring in long-term lesions of the adult rat spinal cord. Exp Neurol
[138] Amr SM, Essam AM, Abdel-Meguid AM, Kholeif AM, Moharram AN, El-Sadek RE: Direct cord implantation in brachial plexus avulsions: revised technique using a single stage combined anterior (first) posterior (second) approach and end-to-side side-to-
[127] Wright KT, El Masri W, Osman A, Roberts S, Chamberlain G, Ashton BA, Johnson WE: Bone marrow stromal cells stimulate neurite outgrowth over neural proteoglycans (CSPG), myelin associated glycoprotein and Nogo-A. Biochem Biophys Res Com‐
[126] Wright KT, Masri WE, Osman A, Roberts S, Trivedi J, Ashton BA, Johnson WE: The cell culture expansion of bone marrow stromal cells from humans with spinal cord injury: implications for future cell transplantation therapy. Spinal Cord 2008;46(12):811–817.
[139] Nakagomi T, Molnár Z, Nakano-Doi A, Taguchi A, Saino O, Kubo S, Clausen M, Yoshikawa H, Nakagomi N, Matsuyama T: Ischemia-induced neural stem/progenitor cells in the pia mater following cortical infarction. Stem Cells Dev 2011 Dec;20(12):2037–
[128] Bunge MB: Novel combination strategies to repair the injured mammalian spinal cord.
[140] Johanson C, Stopa E, Baird A, Sharma H: Traumatic brain injury and recovery mecha‐ nisms: peptide modulation of periventricular neurogenic regions by the choroid
[129] Mackay-Sim A, Feron F, Cochrane J, Bassingthwaighte L, Bayliss C, Davies W, Fronek P, Gray C, Kerr G, Licina P, Nowitzke A, Perry C, Silburn PAS, Urquhart S, Geraghty T: Autologous olfactory ensheathing cell transplantation in human paraplegia: a 3-year
[141] Francel PC, Francel TJ, Mackinnon SE, Hertl C: Enhancing nerve regeneration across a silicone tube conduit by using interposed short-segment nerve grafts. J Neurosurg
[130] Li H, Wen Y, Luo Y, Lan X, Wang D, Sun Z, Hu L: [Transplantation of bone marrow mesenchymal stem cells into spinal cord injury: a comparison of delivery different times] [Article in Chinese] Zhongguo Xiu Fu Chong Jian Wai Ke Za Zhi 2010;24(2):180–
[142] Spector JG, Lee P, Derby A, Roufa DG: Comparison of rabbit facial nerve regeneration in nerve growth factor-containing silicone tubes to that in autologous neural grafts.
[131] Ichim TE, Solano F, Lara F, Paris E, Ugalde F, Rodriguez JP, Minev B, Bogin V, Ramos F, Woods EJ, Murphy MP, Patel AN, Harman RJ, Riordan NH: Feasibility of combina‐ tion allogeneic stem celltherapy for spinal cord injury: a case report. Int Archives
[143] Le Beau JM, Powell HC, Ellisman MH: Node of Ranvier formation along fibres regenerating through silicone tube implants: a freeze-fracture and thin-section electron
microscopic study. J Neurocytol 1987;16(3):347–358. PubMed PMID: 3612184.
Ann Otol Rhinol Laryngol 1995;104(11):875–885.PubMed PMID: 8534028.
plexus-CSF nexus. J Neural Transm 2011;118(1):115–133. Epub 2010 Oct 10.
1997;87(6):887–892. PubMed PMID: 9384400.
side grafting neurorrhaphy. J Brachial Plex Peripher Nerve Inj 2009;4:8.
neural progenitor cells. Sci Rep 2015;5:15962. doi: 10.1038/srep15962.
science.2016.01.066. Epub 2016 Feb 4. PubMedPMID: 26852702.
958. doi: 10.1517/14712598.2015.1043264.
Exp Neurol 2002;177:349–359 [PubMed].
Exp Neurol 2004;190:17–31 [PubMed].
26531615.
36 Recovery of Motor Function in Spinal Cord Injury
216Following Spinal Cord Injury
25833331.
[PubMed].
2051.
184.
Med 2010;3:30.
1995;134:102–111.
mun 2007;354(2):559–566.
J Spinal Cord Med 2008;31:262–269.
clinical trial. Brain 2008;131:2376–2386.
[151] Yue Z, Liu X, Molino PJ, Wallace GG: Bio-functionalisation of polydimethylsiloxane with hyaluronic acid and hyaluronic acid--collagen conjugate for neural interfacing.
[152] Minev IR, Moshayedi P, Fawcett JW, Lacour SP: Interaction of glia with a compliant, microstructured silicone surface. Acta Biomater 2013;9(6):6936–6942. doi: 10.1016/
[153] Patel KR, Tang H, Grever WE, Simon Ng KY, Xiang J, Keep RF, Cao T, McAllister JP II: Evaluation of polymer and self-assembled monolayer-coated silicone surfaces to reduce neural cell growth. Biomaterials 2006;27(8):1519–1526. Epub 2005 Sep 19.
Biomaterials 2011;32(21):4714–4724. doi: 10.1016/j.biomaterials.2011.03.032.
j.actbio.2013.02.048. Epub 2013 Mar 14. PubMed PMID: 23499849.
## **Stem Cell Therapies for Cervical Spinal Cord Injury**
Vanessa M. Doulames, Laura M. Marquardt, Bhavaani Jayaram, Christine D. Plant and Giles W. Plant
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/63580
### **Abstract**
[154] Huang H, Chen L, Wang H, Xiu B, Li B, Wang R, Zhang J, Zhang F, Gu Z, Li Y, Song Y, Hao W, Pang S, Sun J: Influence of patients' age on functional recovery after transplantation of olfactory ensheathing cells into injured spinal cord injury. Chin Med
[144] Peláez RJ, Afonso CN, Vega F, Recio-Sánchez G, Torres-Costa V, Manso-Silván M, García-Ruiz JP, Martín-Palma RJ: Laser fabrication of porous silicon-based platforms for cell culturing. J Biomed Mater Res B Appl Biomater 2013;101(8):1463–1468. doi:
[155] Lima C, Pratas-Vital J, Escada P, Hasse-Ferreira A, Capucho C, Peduzzi JD: Olfactory mucosa autografts in human spinal cord injury: a pilot clinical study. J Spinal Cord
[145] Punzón-Quijorna E, Sánchez-Vaquero V, Muñoz-Noval A, Pérez-Roldán MJ, Martín-Palma RJ, Rossi F, Climent-Font A, Manso-Silván M, Ruiz JP, Torres-Costa V: Nanostructured porous silicon micropatterns as a tool forsubstrate-conditioned cell research. Nanoscale Res Lett 2012;7(1):396. PubMed PMID: 22799489; PubMed Central
[156] Park HC, Shim YS, Ha Y, Yoon SH, Park SR, Choi BH, Park HS: Treatment of complete spinal cord injury patients by autologous bone marrow cell transplantation and administration of granulocyte-macrophage colony stimulating factor. Tissue Eng
[146] Collart-Dutilleul PY, Panayotov I, Secret E, Cunin F, Gergely C, Cuisinier F, Martin M: Initial stem cell adhesion on porous silicon surface: molecular architecture of actin cytoskeleton and filopodial growth. Nanoscale Res Lett 2014;9(1):564. doi: 10.1186/1556-276X-9-564. eCollection 2014. PubMed PMID: 25386101; PubMed Central
[157] Yoon SH, Shim YS, Park YH, Chung JK, Nam JH, Kim MO, Park HC, Park SR, Min BH, Kim EY, Choi BH, Park H, Ha Y: Complete spinal cord injury treatment using autolo‐ gous bone marrow cell transplantation and bone marrow stimulation with granulocyte macrophage-colony stimulating factor: phase I/II clinical trial. Stem Cells 2007;25(8):
[158] Chowdary PD, Che DL, Cui B: Neurotrophin signaling via long-distance axonal transport. Annu Rev Phys Chem 2012;63:571–594. doi: 10.1146/annurev-phys‐
[147] Pandis C, Trujillo S, Matos J, Madeira S, Ródenas-Rochina J, Kripotou S, Kyritsis A, Mano JF, Gómez Ribelles JL: Porous polylactic acid-silica hybrids: preparation, characterization, and study of mesenchymal stem cell osteogenic differentiation. Macromol Biosci 2015;15(2):262–274. doi: 10.1002/mabi.201400339. Epub 2014 Oct 10.
[159] Chowdary PD, Che DL, Kaplan L, Chen O, Pu K, Bawendi M, Cui B: Nanoparticleassisted optical tethering of endosomes reveals the cooperative function of dyneins in retrograde axonal transport. Sci Rep 2015;5:18059. doi: 10.1038/srep18059. PubMed
[148] Zhou P, Cheng X, Xia Y, Wang P, Zou K, Xu S, Du J: Organic/inorganic composite membranes based on poly(L-lactic-co-glycolic acid) and mesoporous silica for effective bone tissue engineering. ACS Appl Mater Interfaces 2014;6(23):20895–20903. doi:
[160] Dono R: Fibroblast growth factors as regulators of central nervous system development and function. Am J Physiol Regul Integr Comp Physiol 2003;284(4):R867–R881. Review.
[149] Kim H, Kim I, Choi HJ, Kim SY, Yang EG: Neuron-like differentiation of mesenchy‐ mal stem cells on silicon nanowires. Nanoscale 2015;7(40):17131–17138. doi: 10.1039/
[161] Hendry IA, Belford DA: Lack of retrograde axonal transport of the heparin-binding growth factors by chick ciliary neurones. Int J Dev Neurosci 1991;9(3):243–250. PubMed
[150] Ai H, Lvov YM, Mills DK, Jennings M, Alexander JS, Jones SA: Coating and selective deposition of nanofilm on silicone rubber for cell adhesion and growth. Cell Biochem
[151] Yue Z, Liu X, Molino PJ, Wallace GG: Bio-functionalisation of polydimethylsiloxane with hyaluronic acid and hyaluronic acid--collagen conjugate for neural interfacing.
[152] Minev IR, Moshayedi P, Fawcett JW, Lacour SP: Interaction of glia with a compliant, microstructured silicone surface. Acta Biomater 2013;9(6):6936–6942. doi: 10.1016/
[153] Patel KR, Tang H, Grever WE, Simon Ng KY, Xiang J, Keep RF, Cao T, McAllister JP II: Evaluation of polymer and self-assembled monolayer-coated silicone surfaces to reduce neural cell growth. Biomaterials 2006;27(8):1519–1526. Epub 2005 Sep 19.
Biomaterials 2011;32(21):4714–4724. doi: 10.1016/j.biomaterials.2011.03.032.
j.actbio.2013.02.048. Epub 2013 Mar 14. PubMed PMID: 23499849.
chem-032511-143704. Epub 2012 Jan 30. Review. PubMed PMID: 22404590.
PMID: 26656461; PubMed Central PMCID: PMC4674899.
Biophys 2003;38(2):103–114. PubMed PMID: 12777710.
J (Engl) 2003,116:1488–1491.
10.1002/jbm.b.32966.
PMCID: PMC3458952.
PMCID: PMC4217708.
38 Recovery of Motor Function in Spinal Cord Injury
218Following Spinal Cord Injury
2005;11:913–922.
2066–2073.
PubMed PMID: 12626354.
c5nr05787f. PubMed PMID: 26422757.
10.1021/am505493j.
PubMed PMID: 25303745.
PMID: 1718148.
Med 2006;29:191–203; discussion 204–206.
Cervical-level injuries account for the majority of presented spinal cord injuries (SCIs), yet there are few therapies that successfully improve the overall quality of life for patients. Regenerative therapies aimed at ameliorating deficits in respiratory and motor function are urgently needed. Cellular transplantation strategies are a promising therapeutic avenue. These strategies seek to overcome the inhibitory environment of the injury site, increase native regenerative capacities, provide scaffolding to bridge the lesion, or replace injury-lost neurons and glia.
Numerous considerations must be taken into account, however, when designing effective cellular transplantation therapies, most notably of which is cell source. Each cell source offers its own unique attributes—both positive and negative—that directly correspond with functional outcomes and clinical translation. Here we discuss three different cell types currently used in cellular transplantation strategies to treat cervical SCIs: mesenchymal stem cells (MSCs), embryonic stem cells (ESCs), and induced pluripotent stem cells (iPSCs). By illustrating the characteristics of each cell type and outlining the studies and clinical trials in which they have been featured, we hope to provide the reader with a detailed understanding of both their capabilities and also their potential drawbacks in experimental and clinical settings.
**Keywords:** cervical spinal cord injury, stem cell therapies, cellular transplantation, functional outcomes, regenerative strategies
### **1. Introduction**
### **1.1. The impact of cervical SCI**
Spinal cord injuries (SCIs) create a formidable encumbrance on the US healthcare service with over half of all injuries occurring atthe cervical level. While most causes of SCI can be attributed
© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
to accidents or violence, the incidence of cervical-specific injuries continues to rise from particularly distinctive causes [1,2]. This is in part due to the ever-increasing spectrum ofinjury types, such as those sustained in direct military environments or as a result of changes in tactical armor design [3–6]. Others include improvements in emergency medicine leading to better survival rates [7], the growth of the aging population as a result of advances in preventative care [8–10], and lifestyle choices leading to structural degradation of the cervical spine [11,12].
Survivors of cervical SCI are faced with dramatic life changes owing to lengthy and repeat‐ ed hospitalizations and the need for full- or part-time caretakers, overall resulting in a loss of personal independence. Combined with a frequent inability to maintain employment or contribute to the workforce, patients incur substantial financial expenses—over the course of their lifetime, a 25-year-old SCI patient can expect to accrue up to \$4.5 million in direct costs alone. Overall, SCI costs the nation upward of \$40.5 billion annually, as per a 2009 report by the Christopher and Dana Reeve Foundation [1,2]. Although recent advances have resulted in increased survival rates, quality of life still remains poor; patients encounter a gradation of sensory deficits, respiratory deficits, motor dysfunction, and paralysis based on their specif‐ ic injury location. Therapies designed to ameliorate some of these complications, even partially, are drastically needed and would make a radical impact in easing the financial, emotional, and physical burden experienced by cervical SCI patients.
### **1.2. The pathophysiology of cervical SCI**
The cervical spinal cord contains the long tracts connecting the rostral and caudal portions of the central nervous system (CNS), as well as sensory and motor neurons. Cervical SCI in mammals initiates large zones of necrosis at the site of injury, creating gaps in the circuitry and preventing communication within the CNS. Axons within the spinal cord fail to regener‐ ate after injury and retract toward the soma from the lesion border. Overall this culminates in crucial changes to normal upper limb function in mammals and disrupts motor function in humans resulting in paralysis and diaphragm-mediated respiration [13].
SCI is characterized by two distinct phases: primary and secondary injuries. During primary injury, the delicate spinal cord tissue is mechanically compromised due to shearing and compression forces, either by direct contact or inadvertently through manipulation of the vertebrae. This leads to mechanical injury, disruptions in vasculature and respiration, neurogenic shock, inflammation, membrane compromise, and alterations in ion and neuro‐ transmitter levels [14–16]. While the primary injury phase leads to an immediate and often serious impairment of neurological function, it is the secondary injury phase that typically dictates the full magnitude of injury. There are approximately 25 established mechanisms to date by which this occurs, but still much ambiguity as to how these pathways converge upon each other to determine the full manifestation of injury [17,18]. Overall, this biochemical cascade activates the ischemic pathway, inflammation and immune responses, swelling, and neuronal apoptosis and leads to neurotransmitter imbalances that underlie excitotoxicity [19– 25].
### **1.3. Regeneration and plasticity of the CNS**
to accidents or violence, the incidence of cervical-specific injuries continues to rise from particularly distinctive causes [1,2]. This is in part due to the ever-increasing spectrum ofinjury types, such as those sustained in direct military environments or as a result of changes in tactical armor design [3–6]. Others include improvements in emergency medicine leading to better survival rates [7], the growth of the aging population as a result of advances in preventative care [8–10], and lifestyle choices leading to structural degradation of the cervical
Survivors of cervical SCI are faced with dramatic life changes owing to lengthy and repeat‐ ed hospitalizations and the need for full- or part-time caretakers, overall resulting in a loss of personal independence. Combined with a frequent inability to maintain employment or contribute to the workforce, patients incur substantial financial expenses—over the course of their lifetime, a 25-year-old SCI patient can expect to accrue up to \$4.5 million in direct costs alone. Overall, SCI costs the nation upward of \$40.5 billion annually, as per a 2009 report by the Christopher and Dana Reeve Foundation [1,2]. Although recent advances have resulted in increased survival rates, quality of life still remains poor; patients encounter a gradation of sensory deficits, respiratory deficits, motor dysfunction, and paralysis based on their specif‐ ic injury location. Therapies designed to ameliorate some of these complications, even partially, are drastically needed and would make a radical impact in easing the financial,
The cervical spinal cord contains the long tracts connecting the rostral and caudal portions of the central nervous system (CNS), as well as sensory and motor neurons. Cervical SCI in mammals initiates large zones of necrosis at the site of injury, creating gaps in the circuitry and preventing communication within the CNS. Axons within the spinal cord fail to regener‐ ate after injury and retract toward the soma from the lesion border. Overall this culminates in crucial changes to normal upper limb function in mammals and disrupts motor function in
SCI is characterized by two distinct phases: primary and secondary injuries. During primary injury, the delicate spinal cord tissue is mechanically compromised due to shearing and compression forces, either by direct contact or inadvertently through manipulation of the vertebrae. This leads to mechanical injury, disruptions in vasculature and respiration, neurogenic shock, inflammation, membrane compromise, and alterations in ion and neuro‐ transmitter levels [14–16]. While the primary injury phase leads to an immediate and often serious impairment of neurological function, it is the secondary injury phase that typically dictates the full magnitude of injury. There are approximately 25 established mechanisms to date by which this occurs, but still much ambiguity as to how these pathways converge upon each other to determine the full manifestation of injury [17,18]. Overall, this biochemical cascade activates the ischemic pathway, inflammation and immune responses, swelling, and neuronal apoptosis and leads to neurotransmitter imbalances that underlie excitotoxicity [19–
emotional, and physical burden experienced by cervical SCI patients.
humans resulting in paralysis and diaphragm-mediated respiration [13].
**1.2. The pathophysiology of cervical SCI**
220 222Recovery of Motor Function Following Spinal Cord Injury
spine [11,12].
25].
Prior evidence suggested that the adult mammalian CNS did not regenerate, predominantly due to the unlikely event of axonal regeneration through the inhibitory milieu of the spinal lesion [26]. However, some degree of functional recovery is often seen, possibly as a result of reorganization of spared circuitry from innate axonal sprouting of spared and intact fibers [26– 28]. Experimental evidence has shown that this process can be influenced and axonal regen‐ eration encouraged via the use of other synergistic therapies. These include the addition of neurotrophic growth factors [29–32], the deletion of inhibitory factors typically associated with the lesion [33–35], and rehabilitation regimens and physical activity [36–38]. Despite this, the innate regenerative capacities of the CNS are often overwhelmed by the extent of injury and functional recovery is limited at best.
Given SCI's multifactorial pathophysiology and the inherent complexity of the CNS, any potentially successful treatment must be effective in positively addressing multiple deficits. Cellular transplantation therapies offer an attractive means of accomplishing this by repopu‐ lating SCI-lost neurons and glia, increasing native regenerative capacities through trophic and immunomodulatory factors secreted by transplanted cells, and providing scaffolding to bridge the inhibitory milieu of the lesion site [20,29,30,32,39–43]. Furthermore, the potency of stem cells makes them an ideal candidate by circumventing the impediments of harvesting and transplanting adult neurons. By promoting neurite regeneration and replenishing appropri‐ ate cell populations, it may be possible to reconnect rostral and caudal neural circuitry and restore function.
In recent decades, the therapeutic promise of cellular intraspinal transplantation has gained significant interest and has eventuated in preliminary clinical trials. Numerous preclinical experiments have been developed to target SCI using peripheral nerve bridges, Schwann cells,
**Figure 1.** A summary of the different stem cell types. Panel (A) shows the origin and isolation of the cells, (B) shows their differentiation potential, and (C) describes transplantation methods used and their intended goals for treatment of cervical SCI.
olfactory glia, mesenchymal stem cells (MSCs), embryonic-derived stem cells, and induced pluripotent stem cells (iPSCs). However, the vast majority of these do not address cervicallevel injury [44–52].
Here, we highlight the current literature on embryonic, mesenchymal, and induced pluripo‐ tent stem cell-based cellulartransplantation with an emphasis on cervical-level SCI. We discuss the benefits and disadvantages of each cellular source, and consider possible future therapeu‐ tic avenues. **Figure 1** provides a summary of these cellular types and their origins, including transplantation methods used and their intended goals for treatment of cervical SCI.
### **2. Mesenchymal stem cells (MSCs)**
### **2.1. Isolation and purification of MSCs**
Transplantation of mesenchymal stem cells (MSCs), also known as bone marrow stromal cells (BMSCs) or mesenchymal progenitor cells, is a strategy currently being investigated to ameliorate the array of deleterious effects following SCI. The terms mesenchymal stromal cells and mesenchymal stem cells have been interchangeably used in the published literature; however there are demonstrable differences between the cells. Mesenchymal stem cells are a subset of stromal cells that maintain the same fibroblast morphology and specific marker expression; however they also have the potential for self-renewal and to differentiate into adipocytes, chondrocytes, and osteoblasts *in vitro* [53–55]. For cultured cells to be defined as MSCs, they should demonstrate the following: (1) adherence to plastic under culture condi‐ tions, (2) expression of CD105, CD73, and CD90, (3) lack of expression of CD45, CD34, CD14/ CD11b, CD79/CD19, and HLA-DR surface markers, and (4) possession of the transdifferen‐ tiation potential to mesodermal lineages *in vitro* [56].
MSCs reside in a wide variety of tissues but are typically extracted from bone marrow and adipose tissue and to a lesser extent the umbilical cord. Their wide distribution and perivas‐ cular origin [57] account fortheir capability to sense and respond to injury by secreting trophic and anti-inflammatory factors [58,59]. The first report of MSCs isolation from bone marrow was by Friedenstein and colleagues. [60]. The spindle-shaped cells isolated were defined as colony-forming units (CFU) with the potential for *in vitro* culture for further transplantation *in vivo* [61]. It was only in 1999 that Pittenger and colleagues established the multi-lineage differentiation potential of MSCs into distinct mesodermal lineages [62]. Originally, Frieden‐ stein and colleagues cultured MSCs by plastic adherence. Since then, many groups have modified this technique by expanding MSCs as a suspension culture [63–65].
Since its origin, bone marrow-derived MSC culture techniques have been continuously validated and improved. While isolation via plastic adherence is effective, the isolation of such cells does not yield a purified population of MSCs leading to varied growth kinetics and differentiation capabilities [66]. The complications of a heterogeneous population can be overcome by the purification of MSCs by using single specific surface markers such as Stro-1, CD271, Stro-3, CD73, and CD2000 [65,67]. Transplantation of MSCs, selected as above, following SCI in a rat model translated to marked improvement in functional recovery and increased tissue sparing [68].
Compared to iPSCs or embryonic stem cells (ESCs), MSCs overcome the ethical concerns of isolation, as MSCs can be extracted from one's own bone marrow or adipose tissue [69]. Recent data has indicated the principal therapeutic advantages of MSCs are their neuroprotective [70,71] and immunomodulatory [71,72] properties.
### **2.2. Regenerative potential of MSCs**
olfactory glia, mesenchymal stem cells (MSCs), embryonic-derived stem cells, and induced pluripotent stem cells (iPSCs). However, the vast majority of these do not address cervical-
Here, we highlight the current literature on embryonic, mesenchymal, and induced pluripo‐ tent stem cell-based cellulartransplantation with an emphasis on cervical-level SCI. We discuss the benefits and disadvantages of each cellular source, and consider possible future therapeu‐ tic avenues. **Figure 1** provides a summary of these cellular types and their origins, including
Transplantation of mesenchymal stem cells (MSCs), also known as bone marrow stromal cells (BMSCs) or mesenchymal progenitor cells, is a strategy currently being investigated to ameliorate the array of deleterious effects following SCI. The terms mesenchymal stromal cells and mesenchymal stem cells have been interchangeably used in the published literature; however there are demonstrable differences between the cells. Mesenchymal stem cells are a subset of stromal cells that maintain the same fibroblast morphology and specific marker expression; however they also have the potential for self-renewal and to differentiate into adipocytes, chondrocytes, and osteoblasts *in vitro* [53–55]. For cultured cells to be defined as MSCs, they should demonstrate the following: (1) adherence to plastic under culture condi‐ tions, (2) expression of CD105, CD73, and CD90, (3) lack of expression of CD45, CD34, CD14/ CD11b, CD79/CD19, and HLA-DR surface markers, and (4) possession of the transdifferen‐
MSCs reside in a wide variety of tissues but are typically extracted from bone marrow and adipose tissue and to a lesser extent the umbilical cord. Their wide distribution and perivas‐ cular origin [57] account fortheir capability to sense and respond to injury by secreting trophic and anti-inflammatory factors [58,59]. The first report of MSCs isolation from bone marrow was by Friedenstein and colleagues. [60]. The spindle-shaped cells isolated were defined as colony-forming units (CFU) with the potential for *in vitro* culture for further transplantation *in vivo* [61]. It was only in 1999 that Pittenger and colleagues established the multi-lineage differentiation potential of MSCs into distinct mesodermal lineages [62]. Originally, Frieden‐ stein and colleagues cultured MSCs by plastic adherence. Since then, many groups have
Since its origin, bone marrow-derived MSC culture techniques have been continuously validated and improved. While isolation via plastic adherence is effective, the isolation of such cells does not yield a purified population of MSCs leading to varied growth kinetics and differentiation capabilities [66]. The complications of a heterogeneous population can be overcome by the purification of MSCs by using single specific surface markers such as Stro-1, CD271, Stro-3, CD73, and CD2000 [65,67]. Transplantation of MSCs, selected as above,
modified this technique by expanding MSCs as a suspension culture [63–65].
transplantation methods used and their intended goals for treatment of cervical SCI.
level injury [44–52].
**2. Mesenchymal stem cells (MSCs)**
tiation potential to mesodermal lineages *in vitro* [56].
**2.1. Isolation and purification of MSCs**
222 224Recovery of Motor Function Following Spinal Cord Injury
The ubiquitous presence ofMSCs around blood vessels makes them more amenable to respond to cues from tissue damage. Recentreports attribute the immunomodulatory function of MSCs as suitable for regenerative therapies and thereby tissue repair. Experimental evidence has shown that the immunomodulatory effect of MSCs is possibly due to their ability to sup‐ press T-cell proliferation by secreting soluble factors such as TGF-beta and hepatocyte growth factor and not via apoptosis [73]. MSCs can also exert their immunomodulatory effect by shifting the balance in favor of regulatory T-cells, known suppressors of the immune system that are triggered by anti-inflammatory cytokines such as IL10. Aggarwal and Pittenger [74] co-cultured populations of immune cells with human MSCS and demonstrated that MSCs altered the secretory cytokine profile, restored balance between helper T-cells and macro‐ phages, reduced pro-inflammatory cytokines, and increased anti-inflammatory molecules, thus favoring the induction of a tolerant anti-inflammatory phenotype [74,75]. Shifting and quenching the inflammatory response then redirects the body's resources toward tissue repair and growth [76,77]. Thus far, preclinical studies have demonstrated the potential for MSCs and their secreted factors to repair damaged tissue through their immunomodulatory and neuroprotective properties.
### **2.3. Treating cervical SCI with MSCs: toward clinical application**
Clinically, the majority of interventions in treating SCI are pharmaceutically based and designed primarily to manage pain and control inflammation. With recent advances in stem cell therapy, there has been an increased interest in clinical studies evaluating the safety and efficacy of stem cells. MSCs are considered a favorable option for transplantation due to a number of factors: ease of isolation, rapid clinical expansion of cultures [78], ability to be cryopreserved and regenerated without loss of potency [79,80], minimal risk of tumorigenic‐ ity [81,82], multipotent capabilities, and the clinical possibility for autologous transplanta‐ tion [83,84]. Furthermore, MSC transplantation has been tested widely in clinical trials and considered safe in a variety of neurological, cardiovascular, and immunological diseases [85]. As such, there is great potential for MSCs as a treatment for SCI, which has been well documented within the literature [53,68,86,87]. There are a number of clinical trials (both ongoing and completed) to test the potential of using MSCs to treat SCI (for the most current information, please refer to www.clinicaltrials.gov). Despite promising results of MSC therapies in animal SCI models and potential for clinical translation, there is yet to be an FDA approved treatment available for SCI patients.
Further investigation is needed to fully understand the basic delivery of MSCs and the mechanistic role in cervical SCI. It has not been established that engraftment and differentia‐ tion of MSCs are even needed for a therapeutic effect, and less than 1% of MSCs survive for longer than a week when systemically administered [88–90]. This survival rate would be similar in intraspinal injection. A transplantation study by Paul and colleagues compared the efficacy of hMSC engraftment when delivered either via lumbar puncture (LP), intravenous transfer (IV), or direct injection into the injury site in a rat C5 subtotal hemisection model [91]. Based on their results assessing engraftment volume (direct injection > LP > IV), glial scar‐ ring (no difference seen after 21 days of MSC transfer), and host immune response (direct injection had the highest host immune response), it can be concluded that MSC delivery via LP is a viable alternative. LP can overcome some of the difficulties of delivering MSCs in patients at the clinical setting. As a validation, a clinical trial performed by a Japanese group evaluated the effect of MSCs treatment in a single patient with a C5 fracture dislocation [92]. On day 13, he received an autologous MSC transplant via lumbar puncture and showed gradual improvement over the 6-month period in both motor and sensory scores graded according to the American Spinal Injury Association (ASIA) Scoring for International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI). Though this is a single case reported, it appears to be promising for the future of MSCs as an intervention in cervical SCI.
### **2.4. Sources of MSCs**
### *2.4.1. From bone marrow stromal cells (BMSCs)*
Bone marrow stem/stromal cells (BMSCs) are the most commonly used stem cell source for transplantation in experimental SCI models. These multipotent cells are derived from the heterogeneous stroma of bone marrow, which is also comprised of hematopoietic stem cells (HSC). MSCs are separated from other cells (like HSCs) by the expression of distinctive cell surface markers [53,67]. Though there has been much speculation about the transdifferentia‐ tion capacity of MSCs into neuronal and glial lineages [71,93], Hofstetter and colleagues reported that transplanted MSCs could not be induced to differentiate toward a neuronal fate, either *in vitro* or *in vivo*, in spite of the fact the MSCs displayed weak expression of NeuN (a neuronal marker) [94].
As one of the first identified sources of MSCs, BMSCs have been well studied for their antiinflammatory, neurotrophic, and neuroprotective functions in SCI [95,96]. BMSCs transplant‐ ed into the rat spinal cord after a contusion injury demonstrated sensorimotor enhancements, partly due to their anti-inflammatory effects by attenuating activation of microglia and astrocytes [97]. To further support theirtherapeutic role in SCI, BMSC transplantation has been shown to reduce cavity formation, enhance axonal growth, and also prevent neuronal apoptosis [86]. While the majority of the studies have explored MSCs potential in a thoracic injury model, very few studies demonstrate their effect in a cervical SCI model; here we summarize studies relevant to cervical SCI.
One of the earliest studies to investigate transplantation within the cervical region investigat‐ ed the use of rat-derived BMSCs in a combinatorial approach with cyclic adenosine mono‐ phosphate (cAMP; a neuronal stimulator) and neurotrophin-3 (NT-3; neurotrophic factor) [98]. Using a dorsal column injury as a model system, cAMP and NT-3 were injected 5 days prior to a C4 transection at L4 to precondition the DRG soma. BMSCs were then transplanted 7 days post injury. This combinatorial approach showed successful axon regeneration throughout and beyond the injury site after 12 weeks that was augmented by preconditioning with cAMP and NT-3. Despite this, functional recovery was not supported. The authors attribute the failure to the regenerating axons notreaching theirtarget in the gracile nucleus, a majorregion that intercepts sensory information.
Further investigation is needed to fully understand the basic delivery of MSCs and the mechanistic role in cervical SCI. It has not been established that engraftment and differentia‐ tion of MSCs are even needed for a therapeutic effect, and less than 1% of MSCs survive for longer than a week when systemically administered [88–90]. This survival rate would be similar in intraspinal injection. A transplantation study by Paul and colleagues compared the efficacy of hMSC engraftment when delivered either via lumbar puncture (LP), intravenous transfer (IV), or direct injection into the injury site in a rat C5 subtotal hemisection model [91]. Based on their results assessing engraftment volume (direct injection > LP > IV), glial scar‐ ring (no difference seen after 21 days of MSC transfer), and host immune response (direct injection had the highest host immune response), it can be concluded that MSC delivery via LP is a viable alternative. LP can overcome some of the difficulties of delivering MSCs in patients at the clinical setting. As a validation, a clinical trial performed by a Japanese group evaluated the effect of MSCs treatment in a single patient with a C5 fracture dislocation [92]. On day 13, he received an autologous MSC transplant via lumbar puncture and showed gradual improvement over the 6-month period in both motor and sensory scores graded according to the American Spinal Injury Association (ASIA) Scoring for International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI). Though this is a single case reported, it appears to be promising for the future of MSCs as an intervention in
Bone marrow stem/stromal cells (BMSCs) are the most commonly used stem cell source for transplantation in experimental SCI models. These multipotent cells are derived from the heterogeneous stroma of bone marrow, which is also comprised of hematopoietic stem cells (HSC). MSCs are separated from other cells (like HSCs) by the expression of distinctive cell surface markers [53,67]. Though there has been much speculation about the transdifferentia‐ tion capacity of MSCs into neuronal and glial lineages [71,93], Hofstetter and colleagues reported that transplanted MSCs could not be induced to differentiate toward a neuronal fate, either *in vitro* or *in vivo*, in spite of the fact the MSCs displayed weak expression of NeuN (a
As one of the first identified sources of MSCs, BMSCs have been well studied for their antiinflammatory, neurotrophic, and neuroprotective functions in SCI [95,96]. BMSCs transplant‐ ed into the rat spinal cord after a contusion injury demonstrated sensorimotor enhancements, partly due to their anti-inflammatory effects by attenuating activation of microglia and astrocytes [97]. To further support theirtherapeutic role in SCI, BMSC transplantation has been shown to reduce cavity formation, enhance axonal growth, and also prevent neuronal apoptosis [86]. While the majority of the studies have explored MSCs potential in a thoracic injury model, very few studies demonstrate their effect in a cervical SCI model; here we
cervical SCI.
**2.4. Sources of MSCs**
neuronal marker) [94].
*2.4.1. From bone marrow stromal cells (BMSCs)*
224 226Recovery of Motor Function Following Spinal Cord Injury
summarize studies relevant to cervical SCI.
MSCs can secrete anti-inflammatory molecules and neurotrophic factors, which can lead to immunomodulation and tissue regeneration. In addition, they can also be engineered to secrete factors specific to CNS regeneration. In 2005, Lu and colleagues examined the ability of BMSCs to promote repair in the injured cord by secreting growth factors that overcome the inhibito‐ ry environment of the lesion [99]. Native or neurally induced rat-derived BMSCs were modified to either express human brain-derived neurotrophic factor (BDNF) in an acute injury or NT-3 in a chronic injury model [100]. By 1 month post transplantation within a C3 dorsal column lesion, BMSC grafts (both native and neurally induced) supported the growth of host and sensory motor axons, a finding that was augmented by either BDNF or NT-3 transduc‐ tion. Transduction with neurotrophic factors substantially increased the number of coerulo‐ spinal, raphespinal, sensory, and motor axons penetrating the lesion site. Supporting these results, Novikova and colleagues transplanted BMSCs that were pretreated with Schwann-cell differentiating factors into a rat C4 hemisection model and demonstrated that BMSCs prestimulated to secrete neurotrophic factors can also contribute to inhibition of astrogliosis and the post-injury inflammatory response [101].
Another group investigated the combination of BMSCs with rat-derived neural progenitor cells (NPCs) co-transplanted in a rat C3 dorsal hemisection transection model [102]. As demyelination is a significant obstacle following SCI, the strategy was employed with the hypothesis that BMSCs would drive NPCs toward a mature oligodendrocyte lineage. In contrast with their *in vitro* data where BMSCs sufficiently redirected NPC differentiation toward an oligodendrocyte lineage, the group's *in vivo* data failed to demonstrate the same 6 weeks post transplant.
Human-derived BMSCs are a clinically attractive transplantation strategy because they can be reintroduced into patients as autografts or allografts. One preclinical study examined the transplantation of human-derived BMSCs from the iliac crest of four different healthy donors into a rat C3–C4 hemisection model [66]. Regardless of donor source, BMSCs survived and filled the lesion site with minimal migration and substantial axonal growth by 2 weeks post transplantation. By 11 weeks post transplant, no BMSCs were present within the lesion site, having been replaced by host oligodendroglial cells. Axonal infiltration into the lesion site and functional recovery varied by donor; however there was no direct correlation between the amount of axonal growth and forelimb function. One possible explanation for this is that individual donor-based distinctions in the secretory profile of BMSCs contributed to the varied outcome.
As challenging as the pitfalls within previous results may appear, there are currently two active clinical trials (Phase I NCT02574572 and Phase II/III NCT0167644) approved by the FDA to study the safety and efficacy of BMSCs transplanted in patients with chronic SCI. In conjunc‐ tion with human studies, a clinical study tested the dose-dependent efficacy of autologous human-derived BMSCs in 13 patients with a mix of cervical (five patients) and thoracic (eight patients) chronic SCI. They found no deleterious effect on the patients; however only one of the 13 patients showed improvement [103]. It is possible that the absence of statistical significance can be attributed to the difference in observation of positive outcomes. For example, while one patient showed an improvement in motor power, two other patients had a patchy improvement in sensory outcome below the level of injury. At the chronic stage, the formation of glial scar tissue around the injury lesion is possibly too dense for growing axons to penetrate. While the transplantation of allogenic BMSCs was deemed safe in patients with chronic SCI, there is a need for preclinical studies to establish the mechanism of MSC's contributory role in chronic SCI.
While spinal cord injury models are in the majority represented by a contusion or hemisec‐ tion injury, damage via herniated discs is also quite a common cause of disability. A herniat‐ ed disc under traumatic events can lead to spinal cord injury. A discectomy is the surgical removal of the herniated nucleus pulposus of a vertebral disc to reduce pressure on the spinal cord or radiating nerves. Clinically, a discectomy is treated by fusing artificial prostheses to replace the intervertebral disc. In a preliminary proof-of-concept study utilizing an ovine model, one group sought to replicate the intervertebral disc by formulating allogeneic BMSCs with a chondrogenic agent, pentosan polysulfate, to form a cartilaginous matrix when implanted into the animal at the C3–C4 and C4–C5 levels [104]. The implant was devoid of any adverse events, and histological evidence showed predominantly cartilaginous tissue within the interbody cages. This was further confirmed by CT scans at 3 months post trans‐ plantation that showed significant bone formation in the cohort receiving BMSCs with pentosan polysulfate when compared to the cohort that received BMSCs alone. Although this particular study had its limitations, it further illustrates the potential for MSCs in preserving spinal function and advancing regenerative medicine.
### *2.4.2. From adipose tissue (AMSCs)*
Adipose tissue is equally attractive as a cell therapy source, due to its minimally invasive harvesting procedure. From a clinical standpoint, adipose-derived stem cells (AMSCs) can easily be obtained from the large quantities of fat tissue that are removed by routine and safe procedures such as liposuction and abdominoplasties. Adipose tissue also contains suppor‐ tive stroma that can be isolated and differentiated toward mesodermal lineages [105].
While a vast majority of the literature indicates that AMSCs support axonal growth, a study utilizing transplanted human-derived AMSCs in a rat C3–C4 hemisection was found to significantly reduce sprouting of the descending serotonergic fibers at the injured site [106]. The authors attributed this to several cumulative factors including enhanced survival of neurons and axons, attenuating the need for excessive sprouting, and reduced astroglial and microglial reactivity favoring the growth of the fibers into and across the transection site. Consistent with other MSC transplantation studies, there was no improvement in functional recovery despite promising microanatomical changes.
Use of AMSCs in humans has been validated for safety and toxicity in both a preclinical testing and a Phase I clinical trial [82]. It is important that every batch of stem cells prepared for transplantation into humans is processed under strict GMP conditions and the cultured cells are verified for absence of toxins and tumorigenic potential in preclinical testing. The pur‐ pose of this study was to evaluate any tumorigenic potential for hAMSCs. Twelve weeks post transplantation, the safety of transplanted AMSCs showed no significant difference in adverse events, ECG monitoring, and physical examinations. Though there was no statistical signifi‐ cance in ASIA score, individual patient scores showed improvement at different levels for motor and/or sensory assessment.
### *2.4.3. From human umbilical cord blood (UMSCs)*
individual donor-based distinctions in the secretory profile of BMSCs contributed to the varied
As challenging as the pitfalls within previous results may appear, there are currently two active clinical trials (Phase I NCT02574572 and Phase II/III NCT0167644) approved by the FDA to study the safety and efficacy of BMSCs transplanted in patients with chronic SCI. In conjunc‐ tion with human studies, a clinical study tested the dose-dependent efficacy of autologous human-derived BMSCs in 13 patients with a mix of cervical (five patients) and thoracic (eight patients) chronic SCI. They found no deleterious effect on the patients; however only one of the 13 patients showed improvement [103]. It is possible that the absence of statistical significance can be attributed to the difference in observation of positive outcomes. For example, while one patient showed an improvement in motor power, two other patients had a patchy improvement in sensory outcome below the level of injury. At the chronic stage, the formation of glial scar tissue around the injury lesion is possibly too dense for growing axons to penetrate. While the transplantation of allogenic BMSCs was deemed safe in patients with chronic SCI, there is a need for preclinical studies to establish the mechanism of MSC's
While spinal cord injury models are in the majority represented by a contusion or hemisec‐ tion injury, damage via herniated discs is also quite a common cause of disability. A herniat‐ ed disc under traumatic events can lead to spinal cord injury. A discectomy is the surgical removal of the herniated nucleus pulposus of a vertebral disc to reduce pressure on the spinal cord or radiating nerves. Clinically, a discectomy is treated by fusing artificial prostheses to replace the intervertebral disc. In a preliminary proof-of-concept study utilizing an ovine model, one group sought to replicate the intervertebral disc by formulating allogeneic BMSCs with a chondrogenic agent, pentosan polysulfate, to form a cartilaginous matrix when implanted into the animal at the C3–C4 and C4–C5 levels [104]. The implant was devoid of any adverse events, and histological evidence showed predominantly cartilaginous tissue within the interbody cages. This was further confirmed by CT scans at 3 months post trans‐ plantation that showed significant bone formation in the cohort receiving BMSCs with pentosan polysulfate when compared to the cohort that received BMSCs alone. Although this particular study had its limitations, it further illustrates the potential for MSCs in preserving
Adipose tissue is equally attractive as a cell therapy source, due to its minimally invasive harvesting procedure. From a clinical standpoint, adipose-derived stem cells (AMSCs) can easily be obtained from the large quantities of fat tissue that are removed by routine and safe procedures such as liposuction and abdominoplasties. Adipose tissue also contains suppor‐
While a vast majority of the literature indicates that AMSCs support axonal growth, a study utilizing transplanted human-derived AMSCs in a rat C3–C4 hemisection was found to
tive stroma that can be isolated and differentiated toward mesodermal lineages [105].
outcome.
contributory role in chronic SCI.
226 228Recovery of Motor Function Following Spinal Cord Injury
*2.4.2. From adipose tissue (AMSCs)*
spinal function and advancing regenerative medicine.
Human umbilical cord blood-derived MSCs (UMSCs) offer various therapeutic advantages in SCI treatment with reversal of SCI pathophysiology (downregulation of apoptotic genes and secretion of neurotrophic factors) in as little as 5 days post injury [107]. In an example using a thoracic SCI, transplantation of UMSCs was reported to have been transdifferentiated toward neuronal and oligodendroglial phenotypes. This was viewed as being a successful strategy as evidenced by improved functional motor outcome [86]. These transdifferentiated oligoden‐ drocytes supported the injured spinal cord in remyelination by secreting neurotrophic factors [108]. UMSCs demonstrate a potential application in the treatment of SCI by their reported ability to transdifferentiate into neuronal lineages. Forfurtherinformation on UMSCs in thoracic injury, readers are referred to the review by Park and colleagues [109].
The first clinically based translational study to use UMSCs was in a rat model of radiation myelopathy, in which significant improvement of the microenvironment previously affected by radiation therapy was observed [110]. Radiation myelopathy is a rare, yet serious compli‐ cation of cancer radiotherapy. Even though radiation myelopathy is not classified as a traditional SCI, there are many similarities in pathophysiology such as vascular damage and demyelination. This particular case study was relevant, as the group studied the efficacy of UMSCs in a radiation myelopathy modelrestricted to the cervical spine. Administration of the UMSCs clearly improved both microvessel and endothelial cell density, along with function‐ al improvement in blood flow. Concurrent with other studies utilizing MSCs, UMSCs were able to reverse injury induced inflammation by reducing pro-inflammatory and increasing anti-inflammatory cytokines within the spinal cord. **Table 1** shows eightpreclinical andclinical trials involving the transplantation of MSCs, describing study design, injury model, ob‐ served outcomes, and noted adverse effects.
**Study and reference (PMID)**
Intrathecal transplantation of autologous adiposederived mesenchymal stem cells for treating spinal cord injury: a
human trial (26208177)
Long-term results of spinal cord injury therapy using mesenchymal stem cells derived from bone marrow in humans (22127044)
**Study designs Preclinical**
)
)
Cells: autologous MSCs harvested from iliac bone
228 230Recovery of Motor Function Following Spinal Cord Injury
Dosing: MSCs (8 × 106
directly injected into spinal cord and (4 × 107
were injected into intradural space Sample size: *N*=10 patients with ASIA class A or B injury caused by traumatic cervical SCI After 4 and 8 weeks of first injection, 5 × 107 injected via lumbar
tapping
recordings
Cells: autologous ADMSCs isolated from lipoaspirates of patient's subcutaneous fat tissue
Dosing: 9 × 107
2 months
patient intrathecally through lumbar tapping at day 1, 1 month, and
Sample size: *N*=14 patients with 12 for ASIA A, 1 for B, 1 for D Six patients were injured
at cervical, 1 at cervicothoracic, 6 at thoracic, and 1 at lumbar
Measurements: MRI, hematological
levels
cells per
Measurements: grading of motor power, MRI, electrophysiological
**or clinical trials**
**Functional Observations**
Clinical At 6-month follow-up: - Six out of 10 patients showed improvement in motor skills - Three out of 10 showed gradual improvement in daily activities
Clinical At 8-month follow-up: - ASIA motor score was improved in five patients - ASIA sensory score was improved in 10 patients - Voluntary and contraction improvement was seen in two patients - One patient showed median
nerve
improvement in somatosensory
**Histological/ imaging datas**
MRI showed a decrease in cavity size and the appearance of fiberlike low-signal intensity streaks
MRI: no significant interval change for presence of tumorous growth EEG: no significant change after transplantation
**Adverse events**
None observed
Urinary tract infection, headache, nausea, and vomiting were observed in three patients
**Table 1.** Tabulated summary of preclinical and clinical trials involving the transplantation of MSCs; for each trial the study design, injury model, observed outcomes and adverse effects are described.
### **3. Embryonic-derived stem cells (ESCs)**
### **3.1. Derivation of embryonic stem cells**
**Study and reference (PMID)**
Localized delivery of brain-derived neurotrophic factor-
mesenchymal stem cells enhances functional recovery following cervical spinal cord
Bone morphogenetic proteins prevent bone marrow stromal cell-
mediated oligodendroglial differentiation of transplanted adult neural progenitor cells in the injured spinal
cord (23770801)
expressing
injury (25093762) **Study designs Preclinical**
Cells: WT-MSCs or BDNF-MSCs
MSCs derived from BM of adult transgenic rats expressing GFP and transduced with murine leukemia virus encoding
MSCs characterized by cell surface expression of CD105 and lack of CD45
injected intraspinally at C2 at time of injury Measurements: immunohistochemistry, electromyogram (EMG)
Cells: NPC isolated from sub-ventricular zone and BMSC isolated from bone
Dosing: immediately following the transection, cell suspensions (2 μl) containing either only BMSCs (0.6 × 105 BMSCs/μl; *N*=5), only NPCs (1.8 × 105 NPCs/μl; *N*=5), or a mixture of NPCs and BMSCs (1.2 × 105
NPCs/μl mixed
*N* =6) were administered by a single injection into
BMSCs/μl;
with 0.3 × 105
marrow
cells
Dosing: 2 × 105
BDNF
230 232Recovery of Motor Function Following Spinal Cord Injury
**or clinical trials**
Preclinical: unilateral spinal cord hemisection at C2 in Sprague-Dawley rats; male
Preclinical: C3 complete transection in adult female Fischer 344 rats
**Functional Observations**
showed significant improvement in postoperative EMG
**Histological/ imaging datas**
N/A Retrograde labeling
None In vitro assays
demonstrate blocking of BMP signaling enables BMSC-induced differentiation of NPCs to oligodendrocytes
with CTB showed localization of MSCs near injection site and primarily in the white matter At day 14 after transplantation, all rats treated with BDNF-MSCs showed functional recovery of diaphragm
**Adverse events**
N/A
Embryonic stem cells (ESCs) represent an intriguing avenue to pursue in the race to under‐ stand and treat neurodegenerative pathophysiology. Their pluripotency makes them an extremely versatile option when compared to the limitations of MSCs, and the basic cell culture techniques governing them have been established for decades. ESCs are pluripotent cells derived from the inner cell mass harvested at the embryonic blastocyst stage. Unlike multi‐ potent stem cells, such as MSCs, that are limited to mesoderm lineages, ESCs have the ability to give rise to all three germ layer lineages: the ectoderm, endoderm, and mesoderm. The capacity to generate neuronal lineages under directed conditioning makes them an ideal target for cell-based therapies of the nervous system [111–113]. Furthermore, ESCs possess the ability to self-renew, a key characteristic of stem cells that allows for a potentially indefinite source of cells.
ESCs were first harvested and cultured in the early 1980s from murine sources and eventual‐ ly followed by human sources in the late 1990s [113–116]. Key differences in the *in vitro* culturing of these cells were responsible for the almost 20-year gap between the successful harvesting of murine versus human-derived ESCs. Murine ESCs are capable of surviving without the support of fibroblast feeder layers but require the addition of leukemia inhibito‐ ry factor (LIF). This led to earlier xeno-free expansion and characterization than human sources, which were later found to rely on the addition of fibroblast growth factor (FGF) to retain their pluripotency and stem cell characteristics [117].
Despite the positive attributes of ESCs, there are certain shortcomings that cannot be over‐ looked. Current research utilizing human-derived ESCs is still limited by the lack of chemi‐ cally defined, *in vitro* culture conditions and is often dependent on the extracellular matrix and growth factor support of Matrigel substrates [118–120]. Additionally, the harvesting and culture of human-derivedESCs raise ethical concerns andextendedculturing has beendebated to lead to karyotypic stability. *In vivo* use of ESCs often leads to teratoma formation and the need for immunosuppressant drugs, all of which pose a problem for usage in clinical studies [111,121–123]. Current research strategies have centered around developing high-efficiency, high-purity differentiation protocols in order to generate committed or progenitor cell populations that limit their teratogenicity and enhance their therapeutic potential.
### **3.2. The regenerative potential of ESCs**
In treating SCI, groups have looked to generate neural and glial-specific lineages such as oligodendrocytes, astrocytes, and neurons from ESCs. The majority of differentiation proto‐ cols lead to high astrocyte and oligodendrocyte populations and relatively few neurons, likely due to the proliferative capacity of the supporting glial cell types compared to neurons [40,111,124–126]. Differentiation methods developed by the McDonald and Keirstead groups have extensively researched the high-efficiency generation of oligodendrocyte progenitor cells (OPCs), which have proven useful in improving myelination and functional outcomes of increased weight support and partial hindlimb gait coordination after thoracic SCI [40,127– 131].
Elegantly designed directed differentiation protocols from the Jessell research group using retinoic acid and sonic hedgehog have led to generation of spinal motor neurons from ESCs, which expressed progenitor motor neuron (pMN) marker Olig2 followed by classic motor neuron markers, Isl1, Hb9, and choline acetyltransferase (ChAT) [132]. When grafted into embryonic day 27 chick spinal cords, these cells have shown integration into the ventral spinal cord and acquire appropriate native motor pool identities [132,133], indicating a potential cell source for lost motor neuron pools after SCI. The Keirstead group has also investigated the
spontaneous differentiation of human ESC-derivedneural progenitor cells (NPCs) into various neuronal phenotypes such as cholinergic, serotonergic, dopaminergic (DA) and/or noradre‐ nergic, and medium spiny striatal neurons [130,134]. Unfortunately, when ES-derived NPCs are transplanted as a heterogenic population, over-proliferation of undifferentiated stem cells can occur and potentially lead to tumor formation [135]. Further research has investigated the ways to mitigate this over-proliferation through antibiotic selection in order to generate highpurity progenitor motor neurons and committed motor neuron populations [136–138].
The first study to demonstrate that stem cells could be induced toward a cortical projection lineage utilized murine ES cells [139]. Appropriate culture conditions and *in vitro* patterning led to the generation of neural precursor cells that initially express forebrain progenitorspecific genes and later features of cortical pyramidal neurons. Most interestingly, transplan‐ tation of these cultures into various locations in the mouse brain led to area-specific axonal and dendritic growth, connectivity and integration within the host CNS, and axon extension to developmentally appropriate targets. These results suggest that it is possible to drive stem cells toward committed neural subtypes that can integrate into anatomically relevant cir‐ cuits *in vivo*.
With all of these various cell populations, numerous cell-based treatment strategies for SCI have been investigated; however the overwhelming majority of these studies have been completed in thoracic injury models. Despite cervical injuries comprising more than 50% of the SCI population, very little preclinical research has been completed to date in this repre‐ sentative model [111].
### **3.3. Treating cervical SCI with ESCs**
potent stem cells, such as MSCs, that are limited to mesoderm lineages, ESCs have the ability to give rise to all three germ layer lineages: the ectoderm, endoderm, and mesoderm. The capacity to generate neuronal lineages under directed conditioning makes them an ideal target for cell-based therapies of the nervous system [111–113]. Furthermore, ESCs possess the ability to self-renew, a key characteristic of stem cells that allows for a potentially indefinite source
ESCs were first harvested and cultured in the early 1980s from murine sources and eventual‐ ly followed by human sources in the late 1990s [113–116]. Key differences in the *in vitro* culturing of these cells were responsible for the almost 20-year gap between the successful harvesting of murine versus human-derived ESCs. Murine ESCs are capable of surviving without the support of fibroblast feeder layers but require the addition of leukemia inhibito‐ ry factor (LIF). This led to earlier xeno-free expansion and characterization than human sources, which were later found to rely on the addition of fibroblast growth factor (FGF) to
Despite the positive attributes of ESCs, there are certain shortcomings that cannot be over‐ looked. Current research utilizing human-derived ESCs is still limited by the lack of chemi‐ cally defined, *in vitro* culture conditions and is often dependent on the extracellular matrix and growth factor support of Matrigel substrates [118–120]. Additionally, the harvesting and culture of human-derivedESCs raise ethical concerns andextendedculturing has beendebated to lead to karyotypic stability. *In vivo* use of ESCs often leads to teratoma formation and the need for immunosuppressant drugs, all of which pose a problem for usage in clinical studies [111,121–123]. Current research strategies have centered around developing high-efficiency, high-purity differentiation protocols in order to generate committed or progenitor cell
populations that limit their teratogenicity and enhance their therapeutic potential.
In treating SCI, groups have looked to generate neural and glial-specific lineages such as oligodendrocytes, astrocytes, and neurons from ESCs. The majority of differentiation proto‐ cols lead to high astrocyte and oligodendrocyte populations and relatively few neurons, likely due to the proliferative capacity of the supporting glial cell types compared to neurons [40,111,124–126]. Differentiation methods developed by the McDonald and Keirstead groups have extensively researched the high-efficiency generation of oligodendrocyte progenitor cells (OPCs), which have proven useful in improving myelination and functional outcomes of increased weight support and partial hindlimb gait coordination after thoracic SCI [40,127–
Elegantly designed directed differentiation protocols from the Jessell research group using retinoic acid and sonic hedgehog have led to generation of spinal motor neurons from ESCs, which expressed progenitor motor neuron (pMN) marker Olig2 followed by classic motor neuron markers, Isl1, Hb9, and choline acetyltransferase (ChAT) [132]. When grafted into embryonic day 27 chick spinal cords, these cells have shown integration into the ventral spinal cord and acquire appropriate native motor pool identities [132,133], indicating a potential cell source for lost motor neuron pools after SCI. The Keirstead group has also investigated the
retain their pluripotency and stem cell characteristics [117].
**3.2. The regenerative potential of ESCs**
232 234Recovery of Motor Function Following Spinal Cord Injury
of cells.
131].
The first study to use embryonic-derived stem cells for therapy in a cervical SCI model was from Sharp and colleagues, looking at the use of human ESC-derived OPCs in a severe midline contusion cervical SCI rat model [140]. The hESC-derived OPCs used in this study were derived using the same protocol [130,131,141,142] as those used in earlier thoracic injury models from the Keirstead group [39,142] and in the Geron™-sponsored, clinical trial (first to use human ESCs as a therapeutic) that was halted in 2011 due to financial and safety con‐ cerns [143,144]. This study found that transplanting hESC-derived OPCs 7 days post injury led to a decrease in lesion area accompanied by robust white and gray matter sparing 8 weeks post transplant. Transplanted cells remained localized within the lesion epicenter with minimal rostral or caudal migration. Notably, the preservation of endogenous motor neu‐ rons was correlated with increased forelimb function. Finally, the use of transplanted hESCderived OPCs led to differential changes in spinal cord gene expression for neurons, growth factors, apoptosis, and inflammation [140]. As such, injured animals with transplanted cells saw gene expression levels more closely matching their uninjured counterparts when compared to transplant-naïve animals. Since the discontinuation of the Geron™-sponsored thoracic injury clinical trial, Asterias Biotherapeutics, Inc., has undertaken the use of hESC AST-OPC1 cells for clinical use in a new Phase I/IIa efficacy and safety trial for cervical-level injuries (NCT02302157).
While most of the literature utilizes acute injury models to study the intervention of stem cells, one group investigated the use of hESC-derived OPCs in a rat cervical chronic injury model. In the 2013 study by Sun and colleagues, Olig2+ GFP+ OPCs derived from mouse ESCs were transplanted 4 months post rat cervical irradiation injury [145]. The irradiation injury model is a chronic injury commonly caused by radiotherapy for cancer treatment that results in the death of oligodendrocytes, which leads to severe demyelination and increased axon death. By 8 weeks post transplant, there was significantly less demyelination in addition to im‐ proved forelimb locomotor function using a clinical degree of weakness scale. Transplanted hESC-OPCs differentiated primarily into mature oligodendrocytes that expressed myelin basic protein. These cells were produced using a retinoic acid/sonic hedgehog differentiation protocol similar to the ones developed by the Jessell and Zhang groups [132,146].
The Keirstead group has also exploited the potential for ESCs to differentiate into neural lineages for use in cervical SCI treatments. They demonstrated successful use of hESC-derived progenitor motor neurons (pMNs) in a cervical SCI contusion model. The pMNs were generated using a retinoic acid differentiation protocol over 28 days and were shown to be Olig1/2+, as well as Tuj1 and Hb9 positive. Electrophysiology of these cells indicated gluta‐ mate receptors after 8 weeks in culture and could innervate both human and rodent muscle tissues. *In vivo*, reduced SCI pathology, orlesion size, and greater endogenous neuron survival and growth were observed in hESC-pMN transplanted spinal cords. Furthermore, these outcomes correlated with increased functional recovery on the balance beam task [147]. The authorsdidnote, however, thatdifferentiation oftransplantedcells was dependent on location. Cells that were found in the distal ventral horn led to increased differentiation, whereas cells confined to the site of injury reverted to a progenitor state.
Other embryonic or fetal-derived tissues have been used in spinal cord injury therapies, including whole fetal spinal cord transplants (pioneered by Reier and Anderson) and fetal neural progenitor/stem cells taken from brain and spinal tissue [51,111,123,148–157]. In the study by Diener and Bregman, transplantation of fetal spinal cord tissues into cervical hemisections led to supraspinal growth, axon projections, improved skilled forelimb func‐ tion indicating transplanted cell survival, and potential local circuit formation. Further investigation of fetal spinal cord lineage-restricted NPCs by various groups has indicated differentiation of transplanted cells into all three neural lineages, in both cervical and thoracic level injuries, with long distance cell migration and axon growth leading to functional improvement [123,148–151,158]. Two clinical trials have begun since 2013 investigating the use of fetal-derived tissues for spinal cord injury [144]. StemCells, Inc.® have begun a safety and efficacy study using human-derived CNS-stem cells (proprietary source) in cervical SCI after completing a thoracic injury trial [159]. Neuralstem, Inc., is also investigating the use of fetal stem cells in SCI [160,161] after promising results using such cells in ALS treatment. This trial, however, is determining the efficacy of these cells in a thoracic injury model. While these cells have proven useful in improving functional outcomes after SCI in preclinical trials, they are subject to the same ethical concerns as those for ESCs and have limited differentiation capacity in comparison [162].
In treating SCI, ESCs demonstrate great potential in promoting axon regeneration and functional recovery, but the lack of full characterization of cell safety, efficacy, and pheno‐ type has significantly limited their clinical applications. Furthermore, the use of ESCs in cervical SCI models is limited to less than a handful of peer-reviewed studies. As cervical injuries are becoming more prevalent, future studies must be designed that better replicate clinical injuries in order to more accurately test cell therapeutic strategies.
### **4. Induced pluripotent stem cells (iPSCs)**
### **4.1. Derivation of iPSCs**
While most of the literature utilizes acute injury models to study the intervention of stem cells, one group investigated the use of hESC-derived OPCs in a rat cervical chronic injury model. In the 2013 study by Sun and colleagues, Olig2+ GFP+ OPCs derived from mouse ESCs were transplanted 4 months post rat cervical irradiation injury [145]. The irradiation injury model is a chronic injury commonly caused by radiotherapy for cancer treatment that results in the death of oligodendrocytes, which leads to severe demyelination and increased axon death. By 8 weeks post transplant, there was significantly less demyelination in addition to im‐ proved forelimb locomotor function using a clinical degree of weakness scale. Transplanted hESC-OPCs differentiated primarily into mature oligodendrocytes that expressed myelin basic protein. These cells were produced using a retinoic acid/sonic hedgehog differentiation
protocol similar to the ones developed by the Jessell and Zhang groups [132,146].
confined to the site of injury reverted to a progenitor state.
234 236Recovery of Motor Function Following Spinal Cord Injury
capacity in comparison [162].
The Keirstead group has also exploited the potential for ESCs to differentiate into neural lineages for use in cervical SCI treatments. They demonstrated successful use of hESC-derived progenitor motor neurons (pMNs) in a cervical SCI contusion model. The pMNs were generated using a retinoic acid differentiation protocol over 28 days and were shown to be Olig1/2+, as well as Tuj1 and Hb9 positive. Electrophysiology of these cells indicated gluta‐ mate receptors after 8 weeks in culture and could innervate both human and rodent muscle tissues. *In vivo*, reduced SCI pathology, orlesion size, and greater endogenous neuron survival and growth were observed in hESC-pMN transplanted spinal cords. Furthermore, these outcomes correlated with increased functional recovery on the balance beam task [147]. The authorsdidnote, however, thatdifferentiation oftransplantedcells was dependent on location. Cells that were found in the distal ventral horn led to increased differentiation, whereas cells
Other embryonic or fetal-derived tissues have been used in spinal cord injury therapies, including whole fetal spinal cord transplants (pioneered by Reier and Anderson) and fetal neural progenitor/stem cells taken from brain and spinal tissue [51,111,123,148–157]. In the study by Diener and Bregman, transplantation of fetal spinal cord tissues into cervical hemisections led to supraspinal growth, axon projections, improved skilled forelimb func‐ tion indicating transplanted cell survival, and potential local circuit formation. Further investigation of fetal spinal cord lineage-restricted NPCs by various groups has indicated differentiation of transplanted cells into all three neural lineages, in both cervical and thoracic level injuries, with long distance cell migration and axon growth leading to functional improvement [123,148–151,158]. Two clinical trials have begun since 2013 investigating the use of fetal-derived tissues for spinal cord injury [144]. StemCells, Inc.® have begun a safety and efficacy study using human-derived CNS-stem cells (proprietary source) in cervical SCI after completing a thoracic injury trial [159]. Neuralstem, Inc., is also investigating the use of fetal stem cells in SCI [160,161] after promising results using such cells in ALS treatment. This trial, however, is determining the efficacy of these cells in a thoracic injury model. While these cells have proven useful in improving functional outcomes after SCI in preclinical trials, they are subject to the same ethical concerns as those for ESCs and have limited differentiation
Induced pluripotent stem cells (iPSCs) are created by reprogramming adult somatic fibro‐ blasts to revert to a pluripotent stem cell state initially via retroviral delivery of Oct3/4, Sox2, c-Myc, and Klf4 [163–165]. Now iPSCs can be generated via multiple processes, each with its own merits and limitations. Viral transduction is easy to use and reproducible, yields iPSCs efficiently, and is controlled. However there is an increased risk of insertional mutagenesis and transgene reactivation, incomplete splicing, and clone-to-clone variation [166–168]. Reprogramming factors can also be fused to cell-penetrating peptides or introduced through plasmids, which requires no genomic modification but is also a time consuming and poten‐ tially inefficient process [169]. Finally iPSCs can be induced via mRNA introduction, which is a highly effective and rapid method; it also requires no genomic modification and is deemed safe due to the transient nature of mRNA. However, repeated transfections are typically required [170,171].
While MSCs are hindered by their limited potency and the harvesting of ESCs is subject to ethical constraints, the pluripotency and source of iPSCs circumvent some of these issues, therefore making them a promising alternative in cellular transplantation therapies. iPSCs share many similarities with ESCs and provide a comparative alternative in that they have the same morphology, gene markers, and potential to form teratomas (ability to differentiate into all three germ layers) [163,165]. Furthermore, the use of iPSCs opens up new possibilities for clinical consideration—ethical concerns are diminished and, in the case of potential trans‐ plants, cells can be harvested directly from the patient, therefore avoiding the need for immunosuppression.
### **4.2. Cell reprogramming technologies for controlled differentiation**
There has been a strong motivation to create iPSC differentiation protocols that drive stem cells toward the three main neural lineages *in vitro*. Methods to generate functional neurons have been of particular interest so as to study the differences in neuronal networks in both healthy and impaired states [172–177]. In one report, mature human fibroblasts were direct‐ ly programmed into synaptically active functional neurons via a cocktail of miR-124, BRN2, and MYT1L [178]. An additional group found that Ascl1 (which has pioneer factor proper‐ ties) in conjunction with BRN2, and MYT1L, successfully drove murine fibroblasts into neuronal cells with appropriate morphology, expression, and formation of functional synapses [179]. One report demonstrated that the overexpression of neurogenin 2 efficiently transformed iPSCs into functional neurons that were able to spontaneously form excitatory synaptic networks. Furthermore, these networks both synaptically integrate once transplant‐ ed into the mouse brain and exhibit plasticity [180]. The majority of research utilizes iPSCs driven toward a neural progenitor state [181,182]; however iPSC differentiation has also been useful in disease modeling. As an example, midbrain dopaminergic (DA) neuron pheno‐ types have been generated, which has been particularly useful in studying Parkinson's disease (PD), typically characterized by the loss of these DA neurons [183–186]. In one study by the Pera group, a stable iPSC line was derived from a PD patient that carried the most common PD-associated genetic mutation and differentiated into midbrain DA neurons. These iPSC-derived DA neurons exhibited classic hallmarks of PD-related damage including accumulation of α-synuclein and oxidative stress, susceptibility to H2O2-induced CASP3 activation, and sensitivity to 6-OHDA and proteasome inhibition. Additionally, other groups have shown that iPSCs can be successfully driven toward glutamatergic, GABAergic, motor, and retinal neuron phenotypes [187–199]. While not specific to SCI, these results demon‐ strate that developing differentiation protocols capable of generating specific neural sub‐ types can open up new research avenues in understanding and creating therapies for neuropathologies.
There is also interest in reliably driving iPSCs toward functional glial subtypes, as glial cells are heavily affected in the process of neurodegeneration. In a study by Krencik and Zhang, exogenous patterning molecules were used to transform iPSCs into a neuroepithelial pheno‐ type. From there, administration of mitogens allowed for the generation of astroglial progen‐ itors, which could then be further differentiated into functional astrocytes via ciliary neurotrophic factor [200]. Another protocol utilized the forced expression of Sox10, Olig2, and Zfb536 to directly reprogram rodent fibroblasts into oligodendrocyte precursor cells. The resulting population of precursors exhibited typical morphology and gene expression and gave rise to mature oligodendrocytes that could ensheath dorsal root ganglion cells *in vitro* and form myelin *in vivo* [201].
The intended goal behind SCI therapies is to ameliorate damage and restore the circuitry within the CNS. Cellulartransplantation offers an innovative means in accomplishing this, but is obviously extremely dependent upon the characteristics and capabilities of the transplant‐ ed cell type. Driving human iPSCs toward neuronal lineages via reproducible and robust differentiation protocols represents a practical interface between developmental neurobiolo‐ gy and SCI research; it may be possible to tailor iPSCs toward a more developmentally appropriate, specific neuronal cell type capable of restoring CNS connectivity rather than the uncharacterized progenitor populations previously used with limited functional recovery.
### **4.3. Treating cervical SCI with iPSCs**
Similar to MSC and ESC-focused SCI therapies, there is a scarcity of targeted preclinical therapies for SCI using iPSC transplantation. Therapies do show positive outcomes yet they are limited in number; to date there are only five published studies using either rodent or simian models of thoracic SCI [202]. In these studies, iPSCs were driven toward neural stem spheres [203], neural stem cells (NSCs) [202,204], and neurospheres [205,206] with all except one study experiencing amelioration of the inhibitory nature of the lesion site, synaptic integration of transplanted cells, and significant functional improvement in transplanted animals.
neuronal cells with appropriate morphology, expression, and formation of functional synapses [179]. One report demonstrated that the overexpression of neurogenin 2 efficiently transformed iPSCs into functional neurons that were able to spontaneously form excitatory synaptic networks. Furthermore, these networks both synaptically integrate once transplant‐ ed into the mouse brain and exhibit plasticity [180]. The majority of research utilizes iPSCs driven toward a neural progenitor state [181,182]; however iPSC differentiation has also been useful in disease modeling. As an example, midbrain dopaminergic (DA) neuron pheno‐ types have been generated, which has been particularly useful in studying Parkinson's disease (PD), typically characterized by the loss of these DA neurons [183–186]. In one study by the Pera group, a stable iPSC line was derived from a PD patient that carried the most common PD-associated genetic mutation and differentiated into midbrain DA neurons. These iPSC-derived DA neurons exhibited classic hallmarks of PD-related damage including accumulation of α-synuclein and oxidative stress, susceptibility to H2O2-induced CASP3 activation, and sensitivity to 6-OHDA and proteasome inhibition. Additionally, other groups have shown that iPSCs can be successfully driven toward glutamatergic, GABAergic, motor, and retinal neuron phenotypes [187–199]. While not specific to SCI, these results demon‐ strate that developing differentiation protocols capable of generating specific neural sub‐ types can open up new research avenues in understanding and creating therapies for
There is also interest in reliably driving iPSCs toward functional glial subtypes, as glial cells are heavily affected in the process of neurodegeneration. In a study by Krencik and Zhang, exogenous patterning molecules were used to transform iPSCs into a neuroepithelial pheno‐ type. From there, administration of mitogens allowed for the generation of astroglial progen‐ itors, which could then be further differentiated into functional astrocytes via ciliary neurotrophic factor [200]. Another protocol utilized the forced expression of Sox10, Olig2, and Zfb536 to directly reprogram rodent fibroblasts into oligodendrocyte precursor cells. The resulting population of precursors exhibited typical morphology and gene expression and gave rise to mature oligodendrocytes that could ensheath dorsal root ganglion cells *in vitro*
The intended goal behind SCI therapies is to ameliorate damage and restore the circuitry within the CNS. Cellulartransplantation offers an innovative means in accomplishing this, but is obviously extremely dependent upon the characteristics and capabilities of the transplant‐ ed cell type. Driving human iPSCs toward neuronal lineages via reproducible and robust differentiation protocols represents a practical interface between developmental neurobiolo‐ gy and SCI research; it may be possible to tailor iPSCs toward a more developmentally appropriate, specific neuronal cell type capable of restoring CNS connectivity rather than the uncharacterized progenitor populations previously used with limited functional recovery.
Similar to MSC and ESC-focused SCI therapies, there is a scarcity of targeted preclinical therapies for SCI using iPSC transplantation. Therapies do show positive outcomes yet they are limited in number; to date there are only five published studies using either rodent or
neuropathologies.
236 238Recovery of Motor Function Following Spinal Cord Injury
and form myelin *in vivo* [201].
**4.3. Treating cervical SCI with iPSCs**
Therapies targeting cervical SCI are equally as limited. There are four published studies to date that have examined acute, subacute, and chronic iPSC transplantation following cervical SCI. Li and colleagues evaluated respiratory function following transplantation of iPSCderived astrocytes engineered to overexpress GLT1, an astroglial glutamate transporter [207]. Both rats and mice underwent a C4 contusion injury resulting in chronic diaphragm dysfunc‐ tion and phrenic motor neuron deterioration. Immediately post injury, subjects received two separate intraspinal injections rostral and caudal to the lesion within the ventral horn. At time points ranging from 2 days to 4 weeks post transplant, it was demonstrated that transplant‐ ed grafts survived and differentiated into GFAP-positive astrocytes, were not tumorigenic, and had less than 10% proliferation (evidenced by Ki67 staining). Furthermore, lesion area and volume were reduced within 1 mm rostral and caudal to the lesion epicenter and innerva‐ tion of the diaphragm neuromuscular junction was preserved in animals that had received iPSC-derived astrocyte transplants that overexpressed GLT1. Through analysis of spontane‐ ous electromyogram (EMG) activity, GLT1-overexpressing astrocyte transplants significant‐ ly magnified EMG amplitude in the dorsal region of the hemidiaphragm, further demonstrating preservation of diaphragmatic respiratory function.
Another study by Lu and colleagues examined the effect of human iPSC-derived NSCs harvested from an elderly donor in a C5 lateral hemisection rat model [208]. While the chosen cell population was minimally characterized, *in vitro* analysis demonstrated reduced expres‐ sion of Tra1-81 and SSEA4 (pluripotency markers) and maintained expression of nestin and Sox2 (NSC-associated markers). Two weeks post injury, NSCs were intraspinally co-trans‐ planted with a fibrin matrix and a raft of growth factors. By 3 months post transplantation, there was evidence that the grafted cells had survived, distributed throughout the lesion, and integrated with host axons. The majority of grafted cells expressed NeuN and mature neuronal markers MAP2 and Tuj1 alongside mature astrocytic marker GFAP, suggesting preferential differentiation into neuronal and astrocytic lineages. There was also evidence of prolifera‐ tion and spinal motor neuron identity within a small percentage of transplanted cells via the expression of Ki67 and ChAT, respectively. Most notably, a remarkable amount of axonal growth was present extending from the lesion site toward the olfactory bulbs and lumbar spine sections. Despite robust axonal growth, no behavioral recovery was observed.
In consideration of the substantial lack of existing chronic cervical SCI data, Nutt and colleagues investigated an early chronic injury model mimicking the deficits seen in human injury [209]. Four weeks following a cervical contusion injury at C4, iPSC-derived neural progenitor cells and fibroblasts were co-transplanted rostral and caudal to the lesion in a rat model. Immunohistochemical analysis suggested the differentiation of transplanted cells into mature neurons as well as the intermingling of the host CNS with transplanted cells, as evidenced by NeuN/FOX-3 labeling. Despite interactions between host and donor cells, transplanted cells did not express glutamate receptors. Furthermore, transplanted cells were not positive for serotonin but did express GABA and were shown co-localized with host positive choline acetyltransferase. Behavioralrecovery was weak; grasping and weight bearing were only slightly improved by transplants.
**Table 2.** A comparative scheme of the characteristics of MSCs, ESCs and iPSCs described in this chapter.
In contrast, Kobayashi and colleagues examined the safety and efficacy of iPSC-derived NSC transplants in a simian model of cervical SCI [210]. Human iPSCs were cultured and in‐ duced to form neurospheres and passaged a secondary and tertiary time prior to transplan‐ tation. Adult female marmosets were given a moderate contusion at the C5 level and received an intraspinal injection of cultured iPSC-neurospheres 9 days post injury at the lesion site. By 12 weeks post transplant, hematoxylin-eosin staining revealed that the grafted cells survived and were positive for NeuN, GFAP, and Olig 1 indicating differentiation into all three neural subtypes. Additionally, animals that received transplants had reduced cystic cavity size, no evidence of tumorigenicity, increased angiogenesis, and a higher amount of neurofi‐ laments and descending motor axons at the lesion center. Severe demyelination was evident surrounding the lesion site in both transplanted and control groups; however, the amount was significantly exacerbated in animals that did not receive cellular transplantation. These findings were further supported by MRI and myelin mapping in which myelin sparing was more evident in the transplanted group and an intramedullary high-signal intensity area in the lesion site of the control group. Calcitonin-generated peptide fibers, which are involved in spinal pain mechanisms, did not differ between transplanted and control groups. In nonhu‐ man primates, contusion at the C5 level in a severed central cord injury model leads to tetraplegia with an expected gradual improvement in motor function. By 8 weeks post injury, there were significant differences in the open field test, bar grip strength test, and cage climbing tests between transplanted and control groups, which stayed consistent throughout the study indicating some level of functional recovery due to transplantation. Promising results from the various cell-based therapies have demonstrated varying degrees of axonal regeneration and functional recovery. **Table 2** provides a summary of the characteristics of MSC, ESC, and iPSC types and also notes their functional outcomes, tumorigenicity, and clinical trial stage to date.
### **5. Summary and conclusions**
transplanted cells did not express glutamate receptors. Furthermore, transplanted cells were not positive for serotonin but did express GABA and were shown co-localized with host positive choline acetyltransferase. Behavioralrecovery was weak; grasping and weight bearing
**MSCs ESCs iPSCs**
ectoderm
blastocyst)
High; often requires immunosuppression
tissue
endoderm, mesoderm, and
Embryonic (inner cell mass of
Fetal tissue Somatic (adult) cells
All three germ layers endoderm, mesoderm, and
Induced or reprogrammed to
Easily accessible sources (e.g.,
induced to be pluripotent
Low; safe for autologous
transplantation
ectoderm
"stemness"
skin)
Strong concerns Minimal; even skin cells can be
Yes; by transdifferentiation Yes; by transdifferentiation
were only slightly improved by transplants.
238 240Recovery of Motor Function Following Spinal Cord Injury
umbilical cord
from adult tissue
Lineage differentiation Mesodermal lineage All three germ layers—
Ease of isolation Easily accessible sources Difficult; isolated from fetal
isolated from the patient
Low; safe for autologous
neuroprotection
transplantation
Differentiation potential Multipotent Pluripotent Pluripotent
Functional outcome Mild to moderate Moderate Mild to moderate
Tumorigenicity No tumor formation Teratoma formation Teratoma formation Clinical trial Advanced to Phase III Advanced to Phase II Preclinical only
**Table 2.** A comparative scheme of the characteristics of MSCs, ESCs and iPSCs described in this chapter.
In contrast, Kobayashi and colleagues examined the safety and efficacy of iPSC-derived NSC transplants in a simian model of cervical SCI [210]. Human iPSCs were cultured and in‐ duced to form neurospheres and passaged a secondary and tertiary time prior to transplan‐ tation. Adult female marmosets were given a moderate contusion at the C5 level and received an intraspinal injection of cultured iPSC-neurospheres 9 days post injury at the lesion site. By 12 weeks post transplant, hematoxylin-eosin staining revealed that the grafted cells survived and were positive for NeuN, GFAP, and Olig 1 indicating differentiation into all three neural subtypes. Additionally, animals that received transplants had reduced cystic cavity size, no evidence of tumorigenicity, increased angiogenesis, and a higher amount of neurofi‐ laments and descending motor axons at the lesion center. Severe demyelination was evident surrounding the lesion site in both transplanted and control groups; however, the amount was
Immunomodulatory Yes Low Low
Source Bone marrow, adipose,
Derivation Purified by surface markers
Ethical issues Minimal; cells can be
Immunogenicity/ autologous
Axonal regrowth Yes; by tissue sparing and
Great care and consideration must be taken when choosing an optimal stem cell type as a potential cellular transplantation treatment for cervical SCIs. Of the three main stem cell types discussed here, there are distinct advantages and disadvantages to each. The use of MSCs in treating nervous system injuries remains a hotly debated topic due to their limited survival, differentiation potential, and functional recovery outcomes. Nevertheless, their immunomo‐ dulatory properties and growth factor secretion make them potentially beneficial for use in combinatorial strategies especially if delivered noninvasively. ESCs offer significantly more differentiation potential for neural applications than adult stem cells and have the added benefit of promoting functional recovery in cervical SCI models. However, the lack of detailed cell characterization, need for immunosuppression, and overall ethical concerns have led to only a single cervical SCI clinical trial. Moreover, the use of ESCs in preclinical cervical SCI studies is limited to only two ESC-derived phenotypes (OPCs and pMNs). Significantresearch must still be performed to fully explore alternative appropriate cell types that can potential‐ ly promote functional regeneration. Finally iPSCs, the newest technology in stem cell sour‐ ces, propose an interesting alternative to fate-limited MSCs and ethically restricted ESCs in treating cervical SCI. Promising preclinical data has indicated iPSC-based therapies can improve functional outcomes after injury; however, their recent discovery highlights the need for careful characterization and exploration of secure differentiation protocols. Further studies must still be completed before iPSCs can be approved for clinical applications.
While stem cell transplantation therapies have shown promise in promoting post-injury regeneration, both anatomical and functional recovery still remain imperfect; no preclinical or clinical study to date has dramatically restored significant recovery in patients. Experimen‐ tal evidence has shown that native regeneration and plasticity occur in limited amounts following injury. These innate processes can be enhanced via the addition of neurotrophic and immunomodulatory factors, the removal of lesion-associated inhibitory factors, and injuryappropriate rehabilitation regimens and physical activity. It would be of great interest to determine whether combinatorial approaches utilizing stem cell transplantation in conjunc‐ tion with the strategies described above provides a synergistic effect within the living system. Furthermore, the vast majority of cell transplantation studies utilize cell populations driven toward immature final phenotypes. The pluripotent capabilities of ESCs and iPSCs provide the freedom to drive these cell types toward numerous definitive lineages or ages. This, however, will be defined by developing appropriate differentiation protocols that can be used in both preclinical and clinical settings. It is possible that transplanting more mature cells results in the establishment and integration of meaningful circuitry within the host nervous system to restore and promote functional recovery.
The broad scope of stem cell therapies offers a myriad of therapeutic potential. However, due to the limited number of preclinical and clinical studies, extensive logistical questions remain regarding how to optimize their usage. Nonetheless, the great strides made in designing and improving effective stem cell therapies for enhancing function promises an exciting future for the field of spinal cord injury repair.
### **Author details**
Vanessa M. Doulames, Laura M. Marquardt, Bhavaani Jayaram, Christine D. Plant and Giles W. Plant\*
\*Address all correspondence to: gplant@stanford.edu
Department of Neurosurgery, Stanford University School of Medicine, Stanford, CA, USA
### **References**
[5] Inoue, T., et al., *Combined SCI and TBI: recovery of forelimb function after unilateral cervical spinal cord injury (SCI) isretarded by contralateraltraumatic brain injury (TBI), and ipsilateral TBI balances the effects of SCI on paw placement*. Exp Neurol, 2013. 248: p. 136–47.
appropriate rehabilitation regimens and physical activity. It would be of great interest to determine whether combinatorial approaches utilizing stem cell transplantation in conjunc‐ tion with the strategies described above provides a synergistic effect within the living system. Furthermore, the vast majority of cell transplantation studies utilize cell populations driven toward immature final phenotypes. The pluripotent capabilities of ESCs and iPSCs provide the freedom to drive these cell types toward numerous definitive lineages or ages. This, however, will be defined by developing appropriate differentiation protocols that can be used in both preclinical and clinical settings. It is possible that transplanting more mature cells results in the establishment and integration of meaningful circuitry within the host nervous
The broad scope of stem cell therapies offers a myriad of therapeutic potential. However, due to the limited number of preclinical and clinical studies, extensive logistical questions remain regarding how to optimize their usage. Nonetheless, the great strides made in designing and improving effective stem cell therapies for enhancing function promises an exciting future for
Vanessa M. Doulames, Laura M. Marquardt, Bhavaani Jayaram, Christine D. Plant and
Department of Neurosurgery, Stanford University School of Medicine, Stanford, CA, USA
[1] Cahill, A., et al. *One Degree of Separation: Paralysis and Spinal Cord Injury in the United States,* 2009, Christopher and Dana Reeves Paralysis Resource Center, p. 1–28.
[2] National Spinal Cord Injury Statistical Center. *Facts and Figures at a Glance*, 2015,
[3] Wagner, C.S. and A.R. Lehman, *Cervicalspine and neck injuries*, in *Musculoskeletal Injuries in the Military*, L.K. Cameron and D.B. Owens, Editors, 2016, Springer New York: New
[4] Breeze, J., et al., *Defining combat helmet coverage for protection against explosively propel‐*
University of Alabama at Birmingham: Birmingham, AL.
*led fragments*. J R Army Med Corps, 2015. 161(1): p. 9–13.
system to restore and promote functional recovery.
\*Address all correspondence to: gplant@stanford.edu
the field of spinal cord injury repair.
240 242Recovery of Motor Function Following Spinal Cord Injury
York, NY, p. 229–45.
**Author details**
Giles W. Plant\*
**References**
[35] Simonen, M., et al., *Systemic deletion of the myelin-associated outgrowth inhibitor Nogo-A improves regenerative and plastic responses after spinal cord injury.* Neuron, 2003. 38(2): p. 201–11.
[21] Dumont, R.J., et al., *Acute spinal cord injury, part I: pathophysiologic mechanisms.* Clin
[22] Hausmann, O.N., *Post-traumatic inflammation following spinal cord injury.* Spinal Cord,
[23] Lu, J., K.W. Ashwell, and P. Waite, *Advances in secondary spinal cord injury: role of*
[24] Mautes, A.E., et al., *Vascular events after spinal cord injury: contribution to secondary*
[25] Park, E., A.A. Velumian, and M.G. Fehlings, *The role of excitotoxicity in secondary mechanisms ofspinal cord injury: a review with an emphasis on theimplicationsfor whitematter*
[26] Horner, P.J. and F.H. Gage, *Regenerating the damaged central nervoussystem.* Nature, 2000.
[27] Bernstein, J.J. and M.E. Bernstein, *Axonal regeneration and formation of synapses proxi‐ mal to the site of lesion following hemisection of the rat spinal cord.* Exp Neurol, 1971. 30(2):
[28] Ramer, M.S., J.V. Priestley, and S.B. McMahon, *Functional regeneration of sensory axons*
[29] Bregman, B.S., et al., *Neurotrophic factors increase axonal growth after spinal cord injury and*
[30] Grill, R., et al., *Cellular delivery of neurotrophin-3 promotes corticospinal axonal growth and partial functional recovery after spinal cord injury.* J Neurosci, 1997. 17(14): p. 5560–72.
[31] McTigue, D.M., et al., *Neurotrophin-3 and brain-derived neurotrophic factor induce oligodendrocyte proliferation and myelination of regenerating axons in the contused adult rat*
[32] Ye, J.H. and J.D. Houle, *Treatment of the chronically injured spinal cord with neurotrophic factors can promote axonal regeneration from supraspinal neurons.* Exp Neurol, 1997. 143(1):
[33] Kim, G.M., et al., *Tumor necrosis factor receptor deletion reduces nuclear factor-kappaB activation, cellular inhibitor of apoptosis protein 2 expression, and functional recovery after*
[34] Nishio, Y., et al., *Deletion of macrophage migration inhibitory factor attenuates neuronal death and promotes functional recovery after compression-induced spinal cord injury in mice.* Acta
Neuropharmacol, 2001. 24(5): p. 254–64.
*apoptosis.* Spine (Phila Pa 1976), 2000. 25(14): p. 1859–66.
*pathogenesis.* Phys Ther, 2000. 80(7): p. 673–87.
*degeneration.* J Neurotrauma, 2004. 21(6): p. 754–74.
*into the adult spinal cord.* Nature, 2000. 403(6767): p. 312–6.
*spinal cord.* J Neurosci, 1998. 18(14): p. 5354–65.
Neuropathol, 2009. 117(3): p. 321–8.
*transplantation in the adult rat.* Exp Neurol, 1997. 148(2): p. 475–94.
*traumatic spinal cord injury.* J Neurosci, 2001. 21(17): p. 6617–25.
2003. 41(7): p. 369–78.
242 244Recovery of Motor Function Following Spinal Cord Injury
407(6807): p. 963–70.
p. 336–51.
p. 70–81.
[66] Neuhuber, B., et al., *Axon growth and recovery of function supported by human bone marrow stromal cells in the injured spinal cord exhibit donor variations.* Brain Res, 2005. 1035(1): p. 73–85.
[51] Reier, P.J., B.S. Bregman, and J.R. Wujek, *Intraspinal transplantation of embryonic spinal cord tissue in neonatal and adult rats.* J Comp Neurol, 1986. 247(3): p. 275–96.
[52] Reubinoff, B.E., et al., *Neural progenitors from human embryonic stem cells.* Nat Biotech‐
[53] Hodgetts, S.I., P.J. Simmons, and G.W. Plant, *Human mesenchymal precursor cells (Stro-1(+)) from spinal cord injury patients improve functional recovery and tissue sparing in*
[54] Horwitz, E.M., et al., *Clarification of the nomenclature for MSC: The International Society*
[55] Bianco, P., et al., *The meaning, the sense and the significance: translating the science of*
[56] Dominici, M., et al., *Minimal criteria for defining multipotent mesenchymalstromal cells. The International Society for Cellular Therapy position statement.* Cytotherapy, 2006. 8(4): p.
[57] Crisan, M., et al., *A perivascular origin for mesenchymal stem cells in multiple human organs.*
[58] Bai, L., et al., *Hepatocyte growth factor mediates mesenchymal stem cell-induced recovery in*
[59] Honczarenko, M., et al., *Human bone marrow stromal cells express a distinct set of biologi‐*
[60] Friedenstein, A.J., R.K. Chailakhjan, and K.S. Lalykina, *The development of fibroblast colonies in monolayer cultures of guinea-pig bone marrow and spleen cells.* Cell Tissue Kinet,
[61] Friedenstein, A.J., et al., *Stromal cells responsible for transferring the microenvironment of the hemopoietic tissues. Cloning in vitro and retransplantation in vivo.* Transplantation, 1974.
[62] Pittenger, M.F., et al., *Multilineage potential of adult human mesenchymal stem cells.*
[63] Castro-Malaspina, H., et al., *Characterization of human bone marrow fibroblast colony-*
[64] Goshima, J., V.M. Goldberg, and A.I. Caplan, *The osteogenic potential of culture-expand‐ ed rat marrow mesenchymal cells assayed in vivo in calcium phosphate ceramic blocks.* Clin
[65] Simmons, P.J. and B. Torok-Storb, *Identification of stromal cell precursors in human bone marrow by a novel monoclonal antibody, STRO-1.* Blood, 1991. 78(1): p. 55–62.
*forming cells (CFU-F) and their progeny.* Blood, 1980. 56(2): p. 289–301.
*an acute spinal cord injury rat model.* Cell Transplant, 2013. 22(3): p. 393–412.
*for Cellular Therapy position statement.* Cytotherapy, 2005. 7(5): p. 393–5.
*mesenchymal stem cells into medicine.* Nat Med, 2013. 19(1): p. 35–42.
*multiple sclerosis models.* Nat Neurosci, 2012. 15(6): p. 862–70.
*cally functional chemokine receptors.* Stem Cells, 2006. 24(4): p. 1030–41.
nol, 2001. 19(12): p. 1134–40.
244 246Recovery of Motor Function Following Spinal Cord Injury
Cell Stem Cell, 2008. 3(3): p. 301–13.
1970. 3(4): p. 393–403.
Science, 1999. 284(5411): p. 143–7.
Orthop Relat Res, 1991(262): p. 298–311.
17(4): p. 331–40.
315–7.
[96] Uccelli, A., et al., *Neuroprotective features of mesenchymal stem cells.* Best Pract Res Clin Haematol, 2011. 24(1): p. 59–64.
[80] Lee, M.W., et al., *Isolation of mesenchymal stem cells from cryopreserved human umbilical*
[81] Lalu, M.M., et al., *Safety of celltherapy with mesenchymalstromal cells (SafeCell): a systematic*
[82] Ra, J.C., et al., *Safety of intravenous infusion of human adipose tissue-derived mesenchymal*
[83] Fouillard, L., et al., *Infusion of allogeneic-related HLA mismatched mesenchymal stem cells for the treatment of incomplete engraftment following autologous haematopoietic stem cell*
[84] Marmont, A.M., et al., *Allogeneic bone marrow transplantation (BMT) forrefractory Behcet's disease with severe CNS involvement.* Bone Marrow Transplant, 2006. 37(11): p. 1061–3.
[85] Parekkadan, B. and J.M. Milwid, *Mesenchymal stem cells as therapeutics.* Annu Rev
[86] Dasari, V.R., K.K. Veeravalli, and D.H. Dinh, *Mesenchymal stem cells in the treatment of*
[87] Vawda, R. and M.G. Fehlings, *Mesenchymal cells in the treatment of spinal cord injury:*
[88] White, S.V., et al., *Intravenous transplantation of mesenchymal progenitors distribute solely to the lungs and improve outcomes in cervical spinal cord injury.* Stem Cells, 2016. DOI:
[89] Lee, R.H., et al., *Intravenous hMSCs improve myocardial infarction in mice because cells embolized in lung are activated to secrete the anti-inflammatory protein TSG-6.* Cell Stem
[90] Zangi, L., et al., *Direct imaging of immune rejection and memory induction by allogeneic*
[91] Paul, C., et al., *Grafting of human bone marrow stromal cells into spinal cord injury: a comparison of delivery methods.* Spine (Phila Pa 1976), 2009. 34(4): p. 328–34.
[92] Saito, F., et al., *Spinal cord injury treatment with intrathecal autologous bone marrow stromal cell transplantation: the first clinical trial case report.* J Trauma, 2008. 64(1): p. 53–9. [93] Kopen, G.C., D.J. Prockop, and D.G. Phinney, *Marrow stromal cells migrate throughout forebrain and cerebellum, and they differentiate into astrocytes after injection into neonatal*
[94] Hofstetter, C.P., et al., *Marrow stromal cells form guiding strands in the injured spinal cord*
[95] Martinez, A.M., et al., *Neurotrauma and mesenchymal stem cells treatment: from experimen‐*
*and promote recovery.* Proc Natl Acad Sci U S A, 2002. 99(4): p. 2199–204.
*tal studies to clinical trials.* World J Stem Cells, 2014. 6(2): p. 179–94.
*mesenchymal stromal cells.* Stem Cells, 2009. 27(11): p. 2865–74.
*mouse brains.* Proc Natl Acad Sci U S A, 1999. 96(19): p. 10711–6.
*spinal cord injuries: A review.* World J Stem Cells, 2014. 6(2): p. 120–33.
*current & future perspectives.* Curr Stem Cell Res Ther, 2013. 8(1): p. 25–38.
*review and meta-analysis of clinical trials.* PLoS One, 2012. 7(10): p. e47559.
*stem cells in animals and humans.* Stem Cells Dev, 2011. 20(8): p. 1297–308.
*cord blood.* Int J Hematol, 2005. 81(2): p. 126–30.
246 248Recovery of Motor Function Following Spinal Cord Injury
*transplantation.* Leukemia, 2007. 21(3): p. 568–70.
Biomed Eng, 2010. 12: p. 87–117.
10.1002/STEM.2364
Cell, 2009. 5(1): p. 54–63.
[127] Cloutier, F., et al., *Transplantation of human embryonic stem cell-derived oligodendrocyte progenitors into rat spinal cord injuries does not cause harm.* Regen Med, 2006. 1(4): p. 469– 79.
[111] Tetzlaff, W., et al., *A systematic review of cellular transplantation therapies for spinal cord*
[112] Tsuji, O., et al., *Cell therapy forspinal cord injury by neuralstem/progenitor cells derived from*
[113] Willerth, S.M., *Neural tissueengineering using embryonic and induced pluripotentstem cells.*
[114] Evans, M.J. and M.H. Kaufman, *Establishment in culture of pluripotential cells from mouse*
[115] Martin, G.R., *Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma stem cells.* Proc Natl Acad Sci U S A, 1981. 78(12): p. 7634–
[116] Thomson, J.A., et al., *Embryonic stem cell lines derived from human blastocysts.* Science,
[117] Pera, M.F. and A.O. Trounson, *Human embryonic stem cells: prospects for development.*
[118] Braam, S.R., et al., *Feeder-free culture of human embryonic stem cells in conditioned medium*
[119] McElroy, S.L. and R.A. Reijo Pera, *Culturing human embryonic stem cells in feeder-free*
[120] Xu, C., et al., *Feeder-free growth of undifferentiated human embryonic stem cells.* Nat Biotech,
[121] Ko, J.Y., et al., *Human embryonic stem cell-derived neural precursors as a continuous, stable, and on-demand source for human dopamine neurons.* J Neurochem, 2007. 103(4): p. 1417–
[122] Richards, M., et al., *Human feederssupport prolonged undifferentiated growth of human inner*
[123] Ruff, C.A., J.T. Wilcox, and M.G. Fehlings, *Cell-based transplantation strategies to promote*
[124] Bain, G., et al., *Embryonic stem cells express neuronal properties in vitro.* Dev Biol, 1995.
[125] Kumagai, G., et al., *Roles of ES cell-derived gliogenic neuralstem/progenitor cellsin functional*
[126] Salewski, R.P., et al., *Transplantation of induced pluripotent stem cell-derived neural stem cells mediate functional recovery following thoracic spinal cord injury through remyelination*
*cell masses and embryonic stem cells.* Nat Biotechnol, 2002. 20(9): p. 933–6.
*plasticity following spinal cord injury.* Exp Neurol, 2012. 235(1): p. 78–90.
*recovery after spinal cord injury.* PLoS One, 2009. 4(11): p. e7706.
*of axons.* Stem Cells Transl Med, 2015. 4(7): p. 743–54.
*for efficient genetic modification.* Nat Protoc, 2008. 3(9): p. 1435–43.
*conditions.* CSH Protoc, 2008. 2008: doi: 10.1101/pdb.prot5044.
*injury.* J Neurotrauma, 2011. 28(8): p. 1611–82.
Stem Cell Res Ther, 2011. 2(2): p. 17.
248 250Recovery of Motor Function Following Spinal Cord Injury
1998. 282(5391): p. 1145–7.
2001. 19(10): p. 971–974.
168(2): p. 342–57.
8.
29.
*embryos.* Nature, 1981. 292(5819): p. 154–6.
Development, 2004. 131(22): p. 5515–25.
*iPS/ES cells.* Neurotherapeutics, 2011. 8(4): p. 668–76.
[159] StemCells, Inc., *Study of human central nervous system (CNS) stem cell transplantation in cervical spinal cord injury. In: ClinicalTrials.gov [Internet]*, 2000-2015, National Library of Medicine (US): Bethesda. MD, *Jan 16. Available from: https://clinicaltrials.gov/ct2/show/ NCT02163876 NLM Indentifier: NCT02163876.*
[143] Baker, M., *Stem-cell pioneer bows out.* Nature, 2011. 479(7374): p. 459.
p. 18.
p. 371–80.
1998. 18(2): p. 779–93.
250 252Recovery of Motor Function Following Spinal Cord Injury
194(1): p. 230–42.
*cord injury.* Cell, 2012. 150(6): p. 1264–73.
Neurotrauma, 1994. 11(4): p. 369–77.
Neurobiol, 1985. 11(2): p. 81–104.
Neurosci, 1998. 18(2): p. 763–78.
*and cat spinal cord.* Exp Neurol, 1992. 115(1): p. 177–88.
*spinal cord.* Brain Pathol, 1995. 5(4): p. 451–57.
[144] Hayden, E.C., *Funding windfall rescues abandoned stem-cell trial.* Nature, 2014. 510(7503):
[145] Sun, Y., et al., *Transplantation of oligodendrocyte precursor cells improves locomotion deficits*
[146] Du, Z.W., et al., *Induced expression of Olig2 is sufficient for oligodendrocyte specification but not for motoneuron specification and astrocyte repression.* Mol Cell Neurosci, 2006. 33(4):
[147] Rossi, S.L., et al., *Histological and functional benefitfollowing transplantation of motor neuron*
[148] Diener, P.S. and B.S. Bregman, *Fetal spinal cord transplants support growth of supraspinal and segmental projections after cervicalspinal cord hemisection in the neonatalrat.* J Neurosci,
[149] Lepore, A.C., et al., *Neural precursor cells can be delivered into the injured cervical spinal cord by intrathecal injection at the lumbar cord.* Brain Res, 2005. 1045(1–2): p. 206–16.
[150] Lepore, A.C. and I. Fischer, *Lineage-restricted neural precursors survive, migrate, and differentiate following transplantation into the injured adult spinal cord.* Exp Neurol, 2005.
[151] Lu, P., et al., *Long-distance growth and connectivity of neural stem cells after severe spinal*
[152] Giovanini, M.A., et al., *Characteristics of human fetalspinal cord graftsin the adultratspinal cord: influences of lesion and grafting conditions.* Exp Neurol, 1997. 148(2): p. 523–43.
[153] Reier, P.J., et al., *Workshop on intraspinal transplantation and clinical application.* J
[154] Reier, P.J., et al., *Fetal cell grafts into resection and contusion/compression injuries of the rat*
[155] Anderson, D.K., D.R. Howland, and P.J. Reier, *Fetal neural grafts and repair of the injured*
[156] Anderson, D.K., et al., *Delayed grafting of fetal CNS tissue into chronic compression lesions*
[157] Reier, P.J., *Neural tissue grafts and repair of the injured spinal cord.* Neuropathol Appl
[158] Diener, P.S. and B.S. Bregman, *Fetalspinal cord transplantssupportthe development oftarget reaching and coordinated postural adjustments after neonatal cervical spinal cord injury.* J
*of the adult cat spinal cord.* Restor Neurol Neurosci, 1991. 2(4): p. 309–25.
*in rats with spinal cord irradiation injury.* PLoS One, 2013. 8(2): p. e57534.
*progenitors to the injured rat spinal cord.* PLoS One, 2010. 5(7): p. e11852.
[190] Faravelli, I., et al., *iPSC‐based models to unravel key pathogenetic processes underlying motor neuron disease development.* J Clin Med, 2014. 3(4): p. 1124–45.
[173] Han, S.S., L.A. Williams, and K.C. Eggan, *Constructing and deconstructing stem cell models*
[174] Marchetto, M.C. and F.H. Gage, *Modeling brain disease in a dish: really?* Cell Stem Cell,
[175] Guo, J., et al., *Contribution of mouse embryonic stem cells and induced pluripotent stem cells to chimerasthrough injection and coculture ofembryos.* Stem Cells Int, 2014. 2014: p. 409021.
[176] Parmar, M. and A. Bjorklund, *Generation of transplantable striatal projection neurons from*
[177] Barker, R.A., *Developing stem cell therapies for Parkinson's disease: waiting until the time is*
[178] Ambasudhan, R., et al., *Direct reprogramming of adult human fibroblasts to functional*
[179] Vierbuchen, T., et al., *Direct conversion of fibroblaststo functional neurons by defined factors.*
[180] Zhang, Y., et al., *Rapid single‐step induction of functional neurons from human pluripotent*
[181] Molyneaux, B.J., et al., *Neuronal subtype specification in the cerebral cortex.* Nat Rev
[182] Arlotta, P., et al., *Neuronal subtype‐specific genes that control corticospinal motor neuron*
[183] Byers, B., et al., *SNCA triplication Parkinson's patient's iPSC‐derived DA neurons accumulate alpha‐synuclein and are susceptible to oxidative stress.* PLoS One, 2011. 6(11): p. e26159.
[184] Nguyen, H.N., et al., *LRRK2 mutant iPSC‐derived DA neurons demonstrate increased*
[185] Dolmetsch, R. and D.H. Geschwind, *The human brain in a dish: the promise of iPSC‐derived*
[186] Pera, M.F., *Stem cells: the dark side of induced pluripotency.* Nature, 2011. 471(7336): p. 46–
[187] Boulting, G.L., et al., *A functionally characterized test set of human induced pluripotent stem*
[188] Brennand, K.J. and F.H. Gage, *Concise review: the promise of human induced pluripotent stem cell‐based studies of schizophrenia.* Stem Cells, 2011. 29(12): p. 1915–22.
[189] Brennand, K.J., et al., *Modelling schizophrenia using human induced pluripotent stem cells.*
*susceptibility to oxidative stress.* Cell Stem Cell, 2011. 8(3): p. 267–80.
*neurons under defined conditions.* Cell Stem Cell, 2011. 9(2): p. 113–8.
*of neurological disease.* Neuron, 2011. 70(4): p. 626–44.
*human ESCs.* Cell Stem Cell, 2012. 10(4): p. 349–50.
*right.* Cell Stem Cell, 2014. 15(5): p. 539–42.
Nature, 2010. 463(7284): p. 1035–41.
Neurosci, 2007. 8(6): p. 427–37.
*neurons.* Cell, 2011. 145(6): p. 831–4.
Nature, 2011. 473(7346): p. 221–5.
*cells.* Nat Biotechnol, 2011. 29(3): p. 279–86.
7.
*stem cells.* Neuron, 2013. 78(5): p. 785–98.
*development in vivo.* Neuron, 2005. 45(2): p. 207–21.
2012. 10(6): p. 642–5.
252 254Recovery of Motor Function Following Spinal Cord Injury
**Emerging Techniques for the Assessment and Rehabilitation of Motor Recovery**
[205] Nori, S., et al., *Grafted human‐induced pluripotent stem‐cell‐derived neurospheres promote motor functional recovery after spinal cord injury in mice.* Proc Natl Acad Sci U S A, 2011.
[206] Tsuji, O., et al., *Therapeutic potential of appropriately evaluated safe‐induced pluripotent stem cells for spinal cord injury.* Proc Natl Acad Sci U S A, 2010. 107(28): p. 12704–9.
[207] Li, K., et al., *Human iPS cell‐derived astrocyte transplants preserve respiratory function after*
[208] Lu, P., et al., *Long‐distance axonal growth from human induced pluripotent stem cells after*
[209] Nutt, S.E., et al., *Caudalized human iPSC‐derived neural progenitor cells produce neurons and glia but fail to restore function in an early chronic spinal cord injury model.* Exp Neurol, 2013.
[210] Kobayashi, Y., et al., *Pre‐evaluated safe human iPSC‐derived neural stem cells promote functionalrecovery afterspinal cord injury in common marmoset withouttumorigenicity.* PLoS
*spinal cord injury.* Exp Neurol, 2015. 271: p. 479–92.
*spinal cord injury.* Neuron, 2014. 83(4): p. 789–96.
108(40): p. 16825–30.
254 256Recovery of Motor Function Following Spinal Cord Injury
248: p. 491–503.
One, 2012. 7(12): p. e52787.
## **Orthoses for Spinal Cord Injury Patients**
Mokhtar Arazpour, Monireh Ahmadi Bani, Mohammad Ebrahim Mousavi, Mahmood Bahramizadeh and Mohammad Ali Mardani
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/64092
### **Abstract**
There are some limitations for patients with spinal cord injury (SCI) when walking with assistive devices. Heavy energy expenditure and walking high loads on the upper limb joints are two main reasons of high rejection rate of orthosis by these patients . Many devices have been designed to enable people with paraplegia to ambulate in an upright position as a solution of these limitations such as mechanical orthoses, hybrid orthoses and powered orthoses. All these devices are designed to solve the problem of standing and walking, but there are some other important notes, which should be considered. For example, the size and weight of external orthoses, donning and doffing, cumbersome‐ ness and independency for using are very important.
**Keywords:** spinal cord injury, orthoses, walking, assistive devices
### **1. Introduction**
Lower limb paralysis resulting from spinal cord injury (SCI) causes inability to walk. Trauma is one of the main causes of SCI that occurs mostly in young people (aged between 16 and 30 years).In 2006, more than 2000patients with SCI were in theUnitedStates andthere are more than 10,000 new cases each year [1]. The ability to walk is the ultimate goal of rehabilitation in SCIpatients.Ambulation is always influencedinpatients with SCI basedon the lesion level and the resulting different levels of muscle paralysis, sensory impairment, spasticity and the lack of trunk control. Walking anomalies in patients with SCI included the absence of active sagittal
© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
plane motions in the hip and knee joints, increased ankle plantarflexion during swing phase and an inability to positioning the lower extremity for initial foot contact and the existence of the so-called "foot slap" based on paralysis of ankle dorsiflexor muscles. Different types of orthoses are designed to reduce complication of inability to walk.
In this chapter, we introduce different types of orthoses for patients with spinal cord injury and explain important parameters (walking, stability, energy expenditure and independen‐ cy) for all existent orthoses by current documents.
### **2. Different types of orthoses**
In overall view, there were three mechanisms in orthoses to walking in spinal cord injury patients: mechanical orthoses, hybrid orthoses and power orthoses (exoskeleton).
### **2.1. Mechanical orthoses**
In this category, three types of the mechanical orthoses included were hip-knee-ankle-foot orthoses, reciprocating gait orthoses and medial linkage orthoses.
### *2.1.1. Hip-knee-ankle-foot orthoses*
A simple hip joint with one degree of freedom generally was used in the Hip Knee Ankle Foot Orthosis (HKAFO). Paraplegic patients use a swing through walking pattern during ambula‐ tion with this type of orthoses. Walking with this pattern produces a high rate of energy consumption and high rate of loads on the upper limb joints (shoulder joints and wrist) in paraplegic patients [2–5].
### *2.1.2. Reciprocating gait orthoses*
Reciprocating gait orthoses (RGO) were introduced in the late 1960s. The ultimate goal was mobilizing lower limb by trunk extension. Therefore, hip extension in one side created hip flexion in other side. In using this kind of orthosis, patients were able to walk reciprocally, doff and don the orthosis independently, but cannot stand up without assistance [6, 7]. In addi‐ tion, walking on a ramp and incline is difficult for paraplegic patients when using the original form of RGO [6]. To resolve this problem, the hip joints in this orthosis must have a two condition locking systems: firstly, in a full extension locking position and secondly in 20° flexion from the first position to walking on ramp and incline [6].
A modified version of RGOs is referred to as the Advanced Reciprocating Gait Orthosis (ARGO) with a one pull-push cable within the pelvic section developed to assist walking performance in paraplegic subjects. In comparison, the Hip Guidance Orthosis (HGO), RGO and ARGO demonstrated the same motion in pattern and magnitude, but in using the ARGO the pelvic displayed a jerky movement pattern. A more developed RGO is defined as the Isocentric Reciprocating Gait Orthosis (IRGO) and was introduced by Motlock in 1992 [8].
### *2.1.3. Medial linkage orthosis (MLO)*
plane motions in the hip and knee joints, increased ankle plantarflexion during swing phase and an inability to positioning the lower extremity for initial foot contact and the existence of the so-called "foot slap" based on paralysis of ankle dorsiflexor muscles. Different types of
In this chapter, we introduce different types of orthoses for patients with spinal cord injury and explain important parameters (walking, stability, energy expenditure and independen‐
In overall view, there were three mechanisms in orthoses to walking in spinal cord injury
In this category, three types of the mechanical orthoses included were hip-knee-ankle-foot
A simple hip joint with one degree of freedom generally was used in the Hip Knee Ankle Foot Orthosis (HKAFO). Paraplegic patients use a swing through walking pattern during ambula‐ tion with this type of orthoses. Walking with this pattern produces a high rate of energy consumption and high rate of loads on the upper limb joints (shoulder joints and wrist) in
Reciprocating gait orthoses (RGO) were introduced in the late 1960s. The ultimate goal was mobilizing lower limb by trunk extension. Therefore, hip extension in one side created hip flexion in other side. In using this kind of orthosis, patients were able to walk reciprocally, doff and don the orthosis independently, but cannot stand up without assistance [6, 7]. In addi‐ tion, walking on a ramp and incline is difficult for paraplegic patients when using the original form of RGO [6]. To resolve this problem, the hip joints in this orthosis must have a two condition locking systems: firstly, in a full extension locking position and secondly in 20°
A modified version of RGOs is referred to as the Advanced Reciprocating Gait Orthosis (ARGO) with a one pull-push cable within the pelvic section developed to assist walking performance in paraplegic subjects. In comparison, the Hip Guidance Orthosis (HGO), RGO and ARGO demonstrated the same motion in pattern and magnitude, but in using the ARGO the pelvic displayed a jerky movement pattern. A more developed RGO is defined as the Isocentric Reciprocating Gait Orthosis (IRGO) and was introduced by Motlock in 1992 [8].
patients: mechanical orthoses, hybrid orthoses and power orthoses (exoskeleton).
orthoses are designed to reduce complication of inability to walk.
orthoses, reciprocating gait orthoses and medial linkage orthoses.
flexion from the first position to walking on ramp and incline [6].
cy) for all existent orthoses by current documents.
**2. Different types of orthoses**
258 260Recovery of Motor Function Following Spinal Cord Injury
**2.1. Mechanical orthoses**
paraplegic patients [2–5].
*2.1.2. Reciprocating gait orthoses*
*2.1.1. Hip-knee-ankle-foot orthoses*
Another type of mechanical orthosis is medial linkage orthosis (MLO). There are variations in this type of orthosis, which include the WalkAbout (WA) [9], Moorong [10] and PrimeWalk (PW) [11] and the Hip and Ankle Linked Orthosis (HALO) [12]. These orthosis are based on a medial single hip joint, which provides artificial hip joint movements. Although it is noted that there is no congruency between the anatomical and mechanical hip joints in the Walkabout MLO, this limitation has been resolved in the Primewalk and Moorong MLO [9–11].
Evaluation of the hip joint mechanism when using the IRGO has been reported to demon‐ strate improved gait parameters, energy consumption compared to other RGOs, HKAFO and MLOs [13]. Therefore, the prescribed option in improvement of walking parameters be‐ tween mechanical orthoses is the IRGO system. The difference between walking with optimal mechanical orthoses (IRGO) and healthy subjects walking is high. Donning and doffing with this type of orthosis, however, is difficult due to its bulky structure and increased weight. MLOs have less donning and doffing time and light structure compared to IRGOs, but these types of orthoses do not have a reciprocating gait mechanism and pelvic rigid structure. Therefore, the users are forced to use high energy consumption during walking, which in turn can create a poor posture compared to walking with IRGO conditions [13–15].
Walking with a mechanical orthoses is not ideal for SCI patients and is a view based on the associated problems during walking with them that includes high loads on upper limb joints and high rate of energy consumption. Some authors have also stated that walking with mechanical orthoses is boring and exhausted [2, 4, 7, 16]. Recent efforts to improve orthoses for SCI patients have led to systems of orthoses that combine the mechanical orthoses with functional electrical stimulation of selected lower extremity muscles and powered hip orthoses [3, 17–19]. Generally, walking parameters and energy consumption improved with new generation of orthoses [19, 20].
### **2.2. Hybrid orthoses**
Hybrid orthoses are a kind of orthoses, which are activated by functional electrical stimula‐ tion (FES). FES is the application of external electrical stimulation to paralyzed muscles to restore their function [6]. The first reported use of FES to facilitate walking for SCI individu‐ als was in 1980, involving the stimulation of quadriceps muscle to enhance stability during stance phase. Two kinds of hybrid orthoses are available: hybrid orthosis based on available mechanical designs (HGO, LSU RGO, ARGO and MLO) and hybrid orthoses based on the new designs (modular hybrid, wrapped spring clutch and spring brake orthoses).
Although some benefits are noted about hybrid orthosis, there are some limitation associat‐ ed with the use of orthoses, FES or hybrid orthoses, which include premature muscle fatigue, false triggering of nearby muscles and being heavy and orthotics cumbersomeness.
### **2.3. Power orthoses (exoskeleton)**
Powered orthoses (exoskeleton) are kinds of orthoses, which activate with external power. The mechanical orthoses have a simple structure and user-friendly design. This type of orthoses, however, has not progressed in development in recent years with the progression, since technology of the powered orthoses appears to be the main focus of research in rehabilita‐ tion and assisted walking and ambulation in SCI patients. There is currently only a limited range of powered orthoses, but there is some evidence of an increase in temporal spatial parameters when walking with powered orthoses [21, 22].
One successful approach used in patient rehabilitation is the use of partial weight bearing when walking on a treadmill using suspension via an overhead harness. This technique, known as body weight support treadmill training (BWSTT), has been used to encourage the regeneration of stepping and walking in incomplete and complete SCI subjects [23, 24]. One of the limitations of BWSTT is that it can only be used in a clinical environment. Provision of hip extension at the end of stance phase, adequate weight bearing through the lower limb during stance phase and the ability to shift weight to the lateral side in the double support phase are essential for BWSTT [25]. However, agreement on the ideal parameters of such gait training does not currently exist [26]. The speed of the treadmill [27], cadence used, and the amount of body unloading are parameters, which can affect gait training in SCI subjects using this system [28]. Therefore, due to these limitations, powered gait orthosis (PGOs) may offer potential in providing an alternative form of gait training for SCI patients.
PGOs can be used as a gait training system to facilitate ambulation in both the clinical situation and in the home via an external power supply using electric motors, pneumatic and/or hydraulic actuators [9]. The first design and construction of PGOs were done in mid-1970s [29]. Belforte [30, 31], and Kang et al. [32], both developed powered orthoses with pneumatic actuators, which demonstrated a positive effect on walking by paraplegic patients, and further developments were subsequently reported by Ruthenberg et al. [33], Ohta et al. [34] and Arazpour et al. [35, 36].
Robert Bogue analyzed some recently developed exoskeletons for military, civil and medical applications and described brain-computer and systems in this review paper [37]. Ferris et al. in another review paper evaluated the recent powered lower limb orthosis for assisting treadmill stepping in disable persons. Practice starting, turning, stopping and avoiding obstacles during overground walking reported main advantage of using powered orthoses as rehabilitation aids in gait rehabilitation after neurological injury. In this study, the powered orthosis that used with bodyweight supported treadmill training for gait rehabilitation in clinical environment analyzed [38]. In the review paper, Dollar and Herr reported the history of the lower limb active orthoses. They reported a design overview of hardware, actuation, sensory and control systems for most of the powered orthosis that have been fabricated until 2007 [39]. Powered orthosis in this literature included full lower limb exoskeletons, modular active orthoses, single joint active orthoses and other orthotic devices. They conducted that the research directions in the future will focus more on the development of light weight exoskeletons and active orthotic devices. All powered orthoses that fabricated for disable persons discussed in the Dollar and Herr review paper, but the effect of advanced powered orthosis on walking in paraplegia patients was not analyzed in their study [39]. Duerinck et al. reported an overview ofthe influence ofthe ankle-footrobot-assisted rehabilitation orthoses to the attributes of normal gait in patients with spinal cord injury. They conducted that pneumatic artificial muscles in combination with proportional myoelectric control can restore the attributes of normal gait. In this review, only powered ankle foot orthosis evaluated on walking in SCI patients [40]. Del-Amae et al. reviewed powered gait orthosis that used FES on the muscles as natural actuators to generate gait in persons with spinal cord injury. Their paper explained an overview of hybrid lower limb exoskeletons, related technologies and advan‐ ces in actuation and control systems, but the efficiency of this type of assistive devices was not evaluated on walking in SCI patients [41].
### **3. Daily living of spinal cord injury patients**
**2.3. Power orthoses (exoskeleton)**
260 262Recovery of Motor Function Following Spinal Cord Injury
Arazpour et al. [35, 36].
parameters when walking with powered orthoses [21, 22].
potential in providing an alternative form of gait training for SCI patients.
Powered orthoses (exoskeleton) are kinds of orthoses, which activate with external power. The mechanical orthoses have a simple structure and user-friendly design. This type of orthoses, however, has not progressed in development in recent years with the progression, since technology of the powered orthoses appears to be the main focus of research in rehabilita‐ tion and assisted walking and ambulation in SCI patients. There is currently only a limited range of powered orthoses, but there is some evidence of an increase in temporal spatial
One successful approach used in patient rehabilitation is the use of partial weight bearing when walking on a treadmill using suspension via an overhead harness. This technique, known as body weight support treadmill training (BWSTT), has been used to encourage the regeneration of stepping and walking in incomplete and complete SCI subjects [23, 24]. One of the limitations of BWSTT is that it can only be used in a clinical environment. Provision of hip extension at the end of stance phase, adequate weight bearing through the lower limb during stance phase and the ability to shift weight to the lateral side in the double support phase are essential for BWSTT [25]. However, agreement on the ideal parameters of such gait training does not currently exist [26]. The speed of the treadmill [27], cadence used, and the amount of body unloading are parameters, which can affect gait training in SCI subjects using this system [28]. Therefore, due to these limitations, powered gait orthosis (PGOs) may offer
PGOs can be used as a gait training system to facilitate ambulation in both the clinical situation and in the home via an external power supply using electric motors, pneumatic and/or hydraulic actuators [9]. The first design and construction of PGOs were done in mid-1970s [29]. Belforte [30, 31], and Kang et al. [32], both developed powered orthoses with pneumatic actuators, which demonstrated a positive effect on walking by paraplegic patients, and further developments were subsequently reported by Ruthenberg et al. [33], Ohta et al. [34] and
Robert Bogue analyzed some recently developed exoskeletons for military, civil and medical applications and described brain-computer and systems in this review paper [37]. Ferris et al. in another review paper evaluated the recent powered lower limb orthosis for assisting treadmill stepping in disable persons. Practice starting, turning, stopping and avoiding obstacles during overground walking reported main advantage of using powered orthoses as rehabilitation aids in gait rehabilitation after neurological injury. In this study, the powered orthosis that used with bodyweight supported treadmill training for gait rehabilitation in clinical environment analyzed [38]. In the review paper, Dollar and Herr reported the history of the lower limb active orthoses. They reported a design overview of hardware, actuation, sensory and control systems for most of the powered orthosis that have been fabricated until 2007 [39]. Powered orthosis in this literature included full lower limb exoskeletons, modular active orthoses, single joint active orthoses and other orthotic devices. They conducted that the research directions in the future will focus more on the development of light weight exoskeletons and active orthotic devices. All powered orthoses that fabricated for disable Daily living including independence, orthosis donning and doffing, sitting and standing, walking on slope surfaces and curbs and toileting have been reported in few documents and literatures. Insufficient evidence exists as to the effect of powered and hybrid orthoses on daily living parameters in comparison with mechanical orthosis because these kinds of orthoses have evaluated in laboratory.
About mechanical orthoses, researchers have noted that in comparison to the RGOs, wear‐ ing the WO and HGOs has been reported to be easier to don and doff [42, 43]. In comparison between WO and IRGO, the WO has also been demonstrated to provide easier sitting and standing [15]. In comparison between HGO and RGO, it was demonstrated that the HGO has been reported to be easier to don and doff [42]. There are only three documents that ana‐ lyzed mechanical orthosis on sloped surfaces and have demonstrated that the RGO was slightly more effective than the HGO when SCI subjects were negotiating curbs or slopes [10, 15, 42]. In a comparison between a MLOs and RGOs, the IRGO was significantly better than the WO (p value = 0.03) in 1:12 and 1:26 gradients [15]. It has demonstrated that using the Moorong orthoses in negotiating slopes surfaces was more functional than the WO due to the ability to align the orthotic and anatomic joints [44]. There is no evidence of any significant difference between any types of orthosis with regards to toileting.
Mechanical orthoses have different effects in improvement of parameters, which affect the daily living activities of SCI patients. No study was found, which considered these factors comprehensively in a selection of patients. From an independence view, the light and uncomplicated structure of the MLOs with removable hip joints and without a rigid spinal component made them easier for donning and doffing than RGOs [21, 22, 45]. The lumbar portion of IRGOs compared to MLOs limited the pelvis and lumbar joints. Unlocked joints in the IRGO caused more crutch support. Therefore, wearing IRGOs makes it more difficult than the MLO to provide sitting and standing [15], and additional document is required to demonstrate how assistive devices can help to provide activity of daily living more easily in the SCI patients. Therefore, from independence view, the comparison between MLO and RGO will be beneficial in this field.
### **4. Energy expenditure of spinal cord injury patients**
Excessive energy expenditure and increased applied force on upper limb joints are two most important factors that increase rejection rates of orthoses in paraplegia patients. There are different methods of evaluation of energy expenditure including O2 cost (ml kg−1 m−1), O2 consumption (ml kg−1 min−1), the PCI (beat per m), HR (beat per min), O2 uptake (l min−1) and the respiratory exchange ratio in SCI patients when using orthoses in all of the documents. However, PCI was introduced as a most sensitive indicator for evaluation of energy expendi‐ ture in SCI patients [46]. In other hand, there is little evidence regarding the efficacy of powered gait orthoses (PGOs) and hybrid orthoses when directly compared to mechanical orthoses in reducing energy expenditure in SCI subjects. However, documents have demonstrated that PGOs and hybrid orthoses have less energy expenditure in comparison with mechanical orthoses [5, 47, 48].
SCI patients do not use their mechanical orthoses, with abandonment rates of 61–90% for children with myelomeningocele [49, 50] and 46–54% in adults with spinal cord injury [10, 51– 53] due to the high level of energy expenditure needed to ambulate.
One of main reasons for the development of PGOs was to potentially reduce energy con‐ sumption when walking with an orthosis. A healthy subject walk with 0.176 mL/kg/m energy expenditure [2] and a SCI subject walk with WBCO 5.41 J/kg/s energy expenditure [54]. The difference between healthy normal subject and walking with a powered orthosis is therefore substantial, but is still improved when compared to that noted when paraplegic subjects walk with mechanical orthoses. There is not enough literature on energy cost of walking with powered orthosis. A further understanding of the energy cost of powered orthosis ambula‐ tion is therefore required for patients with SCI.
Kawashima et al., in evaluation of the weight bearing control orthoses on energy consump‐ tion stated that the energy consumption of walking has been quoted as being 5.41 J/kg/s [54] for SCI persons, while the mean of this parameter was reported 0.176 mL/kg/m in healthy participants [2]. Arazpour et al. in evaluation of the wearing PGOs compared to the mechan‐ ical orthoses such as HKAFO and IRGO reported that wearing PGO provided improved speed of walking and distance walked. The value of the PCI reduced in using the PGO compared to other mechanical orthoses. Active hip and knee joints in providing activated sagittal plane motions were announced as responsible for these results [55]. In a comparison between the WPAL and Primewalk orthoses on energy consumption in four people with paraplegia, Tanabe et al. reported that the PCI exhibited was reduced when using the WPAL [56]. The effort of walking has been shown to be reduced when using a powered motorized ARGO compared to mechanical ARGO in SCI patients [34]. Based on limited studies in evaluation of PGOs on energy consumption in paraplegic patients, a further understanding of this param‐ eter is therefore required for patients with SCI when using PGOs.
About hybrid orthoses, Nene and Patrick [57] in evaluation of using hybrid orthoses (combination of Parawalker and electrical stimulation of the gluteal muscles) in three subjects reported that the rate of the energy cost was 11.78 J/kg/m and 10.95 J/kg/m (a 7.1% reduction)
in without functional electrical stimulation (FES) and with FES condition, respectively. Using FES announced a considerable responsibility for reduction in the vertical crutch impulse values (mean 21%) [57]. The lowest energy costs (kcal/kg min) were associated with the RGO and FES, followed by the RGO, HGO, LLB and FES for walking speeds below 28 m/s.
**4. Energy expenditure of spinal cord injury patients**
262 264Recovery of Motor Function Following Spinal Cord Injury
53] due to the high level of energy expenditure needed to ambulate.
eter is therefore required for patients with SCI when using PGOs.
tion is therefore required for patients with SCI.
orthoses [5, 47, 48].
Excessive energy expenditure and increased applied force on upper limb joints are two most important factors that increase rejection rates of orthoses in paraplegia patients. There are different methods of evaluation of energy expenditure including O2 cost (ml kg−1 m−1), O2 consumption (ml kg−1 min−1), the PCI (beat per m), HR (beat per min), O2 uptake (l min−1) and the respiratory exchange ratio in SCI patients when using orthoses in all of the documents. However, PCI was introduced as a most sensitive indicator for evaluation of energy expendi‐ ture in SCI patients [46]. In other hand, there is little evidence regarding the efficacy of powered gait orthoses (PGOs) and hybrid orthoses when directly compared to mechanical orthoses in reducing energy expenditure in SCI subjects. However, documents have demonstrated that PGOs and hybrid orthoses have less energy expenditure in comparison with mechanical
SCI patients do not use their mechanical orthoses, with abandonment rates of 61–90% for children with myelomeningocele [49, 50] and 46–54% in adults with spinal cord injury [10, 51–
One of main reasons for the development of PGOs was to potentially reduce energy con‐ sumption when walking with an orthosis. A healthy subject walk with 0.176 mL/kg/m energy expenditure [2] and a SCI subject walk with WBCO 5.41 J/kg/s energy expenditure [54]. The difference between healthy normal subject and walking with a powered orthosis is therefore substantial, but is still improved when compared to that noted when paraplegic subjects walk with mechanical orthoses. There is not enough literature on energy cost of walking with powered orthosis. A further understanding of the energy cost of powered orthosis ambula‐
Kawashima et al., in evaluation of the weight bearing control orthoses on energy consump‐ tion stated that the energy consumption of walking has been quoted as being 5.41 J/kg/s [54] for SCI persons, while the mean of this parameter was reported 0.176 mL/kg/m in healthy participants [2]. Arazpour et al. in evaluation of the wearing PGOs compared to the mechan‐ ical orthoses such as HKAFO and IRGO reported that wearing PGO provided improved speed of walking and distance walked. The value of the PCI reduced in using the PGO compared to other mechanical orthoses. Active hip and knee joints in providing activated sagittal plane motions were announced as responsible for these results [55]. In a comparison between the WPAL and Primewalk orthoses on energy consumption in four people with paraplegia, Tanabe et al. reported that the PCI exhibited was reduced when using the WPAL [56]. The effort of walking has been shown to be reduced when using a powered motorized ARGO compared to mechanical ARGO in SCI patients [34]. Based on limited studies in evaluation of PGOs on energy consumption in paraplegic patients, a further understanding of this param‐
About hybrid orthoses, Nene and Patrick [57] in evaluation of using hybrid orthoses (combination of Parawalker and electrical stimulation of the gluteal muscles) in three subjects reported that the rate of the energy cost was 11.78 J/kg/m and 10.95 J/kg/m (a 7.1% reduction) Merati et al. [5] compared the energy cost of locomotion demonstrated by 14 SCI patients (lesion level C7 ± T11) during ambulation with different orthoses (the HGO, Parawalker, RGO and RGO + FES). They observed that during locomotion at maximal speed, HR peak values were 160 ± 16, 155 ± 31 and 154 ± 31 bts/min and VO2 1/kg peak values were 18.0 ± 6.1, 18.5 ± 5.4 and 19.1 ± 7.2 for PW, RGO and RGO + FNS, respectively. During orthosis-assisted locomo‐ tion at maximal speed, HR peak values were 150 ± 13, 131 ± 21 and 155 ± 23 bts/min, and VO2 1/kg peak values were 13.4 ± 3.0, 13.8 ± 3.5 and 17.2 ± 4.8 for PW, RGO and RGO + FNS, respectively. They also reported that maximal ventilations at VO2 peak were 63.8 ± 24.0, 68.9 ± 27.1 and 67.6 ± 23.91 1/min during wheelchair ambulation, and 71.8 ± 7.3, 76.5 ± 21.3 and 72.3 ± 12.2 m/kg/min during orthosis locomotion for PW, RGO and RGO + FNS, respec‐ tively [5]. The PEDro score for this study was a 4/10, which equals the highest score assigned in this group.
In mechanical orthoses, connector cable ARGO maintains the posture and reduce PCI and IRGO (2.6 beat/m) has been shown to be more effective than an RGO (3.6 beat/m). Partici‐ pants in this study reported less fatigue when they used the IRGO [46]. In comparison between RGOs and MLOs, RGOs have less energy expenditure, because MLOs do not have flexionassist system forthe SCI patients and have limited trunk stability, which made patients expend additional effort to maintain upright stance [58]. Also in comparison between MLOs, there is no significant difference in PCI.
RGOs produce less energy expenditure than MLOs and KAFOs [58, 59]. This is because RGOs offer more support for pelvis and trunk and possess a reciprocal system [58]. Energy expen‐ diture and endurance have indirect correlation [60]. Muscle dysfunction, locked position of the ankle and knee joints, flexion of the trunk for providing stability and ambulation with assistive devices such as walker and crutch increase energy consumption in paraplegic subjects in wearing and walking with mechanical orthoses. Ankle and knee joints fixed position can increase energy consumption by up to 33% [60, 61]. Development of the ankle and knee joints of mechanical orthoses via powered or movable structures can be announced as the feasible approach to reduce energy expenditure. The energy consumption of using RGOs reported low rate compared to other mechanical orthoses such as MLOs and HKAFOs, based on the reciprocal section of RGOs announced as the responsible for this result. But the effect of the reciprocal link in providing of the reciprocal motion was demonstrated less [62–64]. Using RGOs for SCI patients ratherthan MLOs is cumbersome due to their bulky structure and using an MLO has high energy consumption compared to RGOs, but MLOs are more user-friend‐ ly for people with SCI [15, 65]. As a result, development of the MLOs with additional structure to provide reciprocal motion announced as the better approach in reduction of the energy expenditure in the SCI subjects. Therefore, the comparison of MLOs and IRGOs on the energy cost, energy expenditure and endurance of the ambulation to provide more effective mechan‐
ical orthoses in orthotic rehabilitation of ambulation in the SCI subjects will be important and beneficial.
In comparison between IRGO andRGO on energy consumption in four paraplegic SCI subjects with T3–T12 level of injury, the PCI was reduced when wearing the IRGO (2.6 beat/m), compared to condition that the RGO was used (3.6 beat/m). SCI patients announced less rate of the fatigue while walking with the IRGO [46]. Walking with standard ARGOs demonstrat‐ ed the reduction of the energy consumption compared to an ARGO without a reciprocal connector cable (5.4 beat/m vs. 5.8 beat/m), although this difference was not statistically significant [66]. The standard ARGO provided best trunk posture in the SCI patients. This positioning is critical and important in the subjects with high level of injury. Evaluation of using WO on five SCI patients demonstrated that energy expenditure was 9.61 (ml/kg/min) [67]. The rate of the O2 consumption and cost were announced as 13.79 (ml/kg/min) and 1.28 (ml/kg/m), respectively, in walking with standard ARGO in six SCI subjects [68].
### **5. Temporal spatial parameters**
The findings of literature show that there is no high level of document to demonstrate that PGOs are better than mechanical orthoses such as RGOs and HGO in improving temporal spatial gait parameters in SCI subjects. From financial view, the PGOs are so expensive, and also the more orthotic gait training time and effort required providing them functionally. Walking with this type of orthosis does not propose any acceptable improvement in tempo‐ ral spatial parameters compared to mechanical orthoses. Consequently, more additional attempt on structure of the powered orthoses is required to provide more acceptable pow‐ ered device for SCI patient wearing and using [69]. There shows to be no significant differ‐ ence in reported speed of walking subjects in walking with different RGOs in the SCI patients. The following text demonstrates the findings resulted by the literature, which analyzed the effect of common mechanical orthoses on temporal spatial parameters in the SCI patients.
### **5.1. Stride length**
In comparison between the RGO and HGO, there was no significant difference in stride length [67]. In comparison of the standard ARGO with RGO without connector cable, 7% increase in stride length (0.89 vs. 0.83) was reported. Improvement in providing vertical positioning ofthe trunk by the connector cable announced the main responsible in the standard ARGO [62]. The mean of stride length was reported to be 0.56 m in SCI patients when using theWO [70]. In newly developed orthoses, Genda et al. demonstrated that the hip-ankle-linked orthoses (HALO), increased stride length by 3% compared to the WO (1.03 vs. 1.00 m) [12]. Six percent increased stride length was demonstrated when ankle foot orthosis with dorsiflex‐ ion-assist ankle joints were used in the ARGO (0.94 vs. 1 m) compared to the ARGO associat‐ ed with solid ankle-foot orthosis. In this study, it was claimed that the moveable ankle joint in the ankle-foot orthosis section could improve stride length [71]. When comparing two medial linked orthoses (the Primewalk and Walkabout orthosis). Onogi et al. reported that there was significant difference between them in this parameter and the mean of stride length was increased by 19% [72].
### **5.2. Cadence**
ical orthoses in orthotic rehabilitation of ambulation in the SCI subjects will be important and
In comparison between IRGO andRGO on energy consumption in four paraplegic SCI subjects with T3–T12 level of injury, the PCI was reduced when wearing the IRGO (2.6 beat/m), compared to condition that the RGO was used (3.6 beat/m). SCI patients announced less rate of the fatigue while walking with the IRGO [46]. Walking with standard ARGOs demonstrat‐ ed the reduction of the energy consumption compared to an ARGO without a reciprocal connector cable (5.4 beat/m vs. 5.8 beat/m), although this difference was not statistically significant [66]. The standard ARGO provided best trunk posture in the SCI patients. This positioning is critical and important in the subjects with high level of injury. Evaluation of using WO on five SCI patients demonstrated that energy expenditure was 9.61 (ml/kg/min) [67]. The rate of the O2 consumption and cost were announced as 13.79 (ml/kg/min) and 1.28
(ml/kg/m), respectively, in walking with standard ARGO in six SCI subjects [68].
The findings of literature show that there is no high level of document to demonstrate that PGOs are better than mechanical orthoses such as RGOs and HGO in improving temporal spatial gait parameters in SCI subjects. From financial view, the PGOs are so expensive, and also the more orthotic gait training time and effort required providing them functionally. Walking with this type of orthosis does not propose any acceptable improvement in tempo‐ ral spatial parameters compared to mechanical orthoses. Consequently, more additional attempt on structure of the powered orthoses is required to provide more acceptable pow‐ ered device for SCI patient wearing and using [69]. There shows to be no significant differ‐ ence in reported speed of walking subjects in walking with different RGOs in the SCI patients. The following text demonstrates the findings resulted by the literature, which analyzed the effect of common mechanical orthoses on temporal spatial parameters in the SCI patients.
In comparison between the RGO and HGO, there was no significant difference in stride length [67]. In comparison of the standard ARGO with RGO without connector cable, 7% increase in stride length (0.89 vs. 0.83) was reported. Improvement in providing vertical positioning ofthe trunk by the connector cable announced the main responsible in the standard ARGO [62]. The mean of stride length was reported to be 0.56 m in SCI patients when using theWO [70]. In newly developed orthoses, Genda et al. demonstrated that the hip-ankle-linked orthoses (HALO), increased stride length by 3% compared to the WO (1.03 vs. 1.00 m) [12]. Six percent increased stride length was demonstrated when ankle foot orthosis with dorsiflex‐ ion-assist ankle joints were used in the ARGO (0.94 vs. 1 m) compared to the ARGO associat‐ ed with solid ankle-foot orthosis. In this study, it was claimed that the moveable ankle joint in the ankle-foot orthosis section could improve stride length [71]. When comparing two medial linked orthoses (the Primewalk and Walkabout orthosis). Onogi et al. reported that there was
**5. Temporal spatial parameters**
264 266Recovery of Motor Function Following Spinal Cord Injury
**5.1. Stride length**
beneficial.
In a comparison between the RGO and the HGO, Whittle et al. reported no significant difference between them in this parameter [70]. However, Winchester et al. announced that although cadence was better in the IRGO than the RGO, there was no significant difference between them [46]. Ijzerman et al. in an evaluation of the effect of a connector cable reported that using an ARGO with a connector cable increased cadence compared to without one, but not significantly (31.3 vs. 30.3 steps/min) [66].
Four studies assessed the cadence in walking with orthoses associated with medial single hip joints in SCI patients. The mean of this parameter was reported as being 70.02 steps/min when using the WO [67] and 50.9 steps/min when walking with any type of medial single hip joints orthosis [73]. Significant difference in cadence between the HALO and PW devices (74.1 vs. 58.9) was reported by Genda et al. [12]. In a recently published study, the mean of cadence demonstrated a significant difference in value (40.8 vs. 48 steps/min, respectively) when using the WO and the PW [72].
### **5.3. Speed of walking**
The mean of speed of walking has been reported to be 0.214 m/s when using the Parawalker orthosis during ambulation by SCI patients [45]. When using the ARGO, the mean of this parameter was 0.16 m/s [68]. The mean of speed of walking was reported to be similar between the HGO and the RGO (0.24 m/s) by Whittle et al. [70]. In comparison between RGOs, the IRGO was shown to produce a higher speed of walking (0.22) compared to a cable-type RGO (0.21), but there was no significant difference between them [46]. When comparing the KAFO and the IRGO, there were significant differences noted between them in this parameter (p = 0.009) [59]. The effect of a connector cable when using the ARGO has no significant effect on improving walking speed when compared to an ARGO without one (0.24 vs. 0.23 m/s) [62].
The effect of PGOs on walking speed, cadence and step length exhibited by SCI subjects is dependent on which joints are actuated in an orthosis (e.g. the hip or knee actuated separate‐ ly or by being synchronized). Few studies have directly compared these parameters directly between separately or synchronized movement of the hip and knee joint conditions during walking with PGOs.
Kang et al. [74] in evaluation of the powered IRGO (via using pneumatic actuators in the hip joints) compared to a mechanical IRGO after three months orthotic gait training in three SCI subjects reported that evaluated parameters such as walking speed, pelvic tilt, flexion and extension angles of the knee and hip joints, stance and swing phase times improved. Walking speed was increased by 26% and the percentage swing phase during walking was increased by 25% when walking with the PGO.
Powered lower limb orthoses such as the ReWalk powered orthosis (Argo Medical Technol‐ ogies), the wearable power-assist locomotor (WPAL) and the eLEGS powered orthosis
(Berkeley Bionics) are all examples of commercially developed powered orthoses designed for walking by paraplegic subjects. The Hybrid Assistive Limb (HAL)—6LB which has six electric motors—bilaterally at the hip, knee and ankle joints, the HAL-5 Type-C, which is previous HAL for paraplegia patients, with only four power units are another examples of powerlower limb orthosis. Commercially developed HAL has four power units and uses EMG signals of leg to synchronize the motion support with the wearer's movement, but complete paraple‐ gia patients cannot use this version of HAL.
Speed of walking reported difference between mechanical, hybrid and powered gait ortho‐ ses. There was statistically significant difference between orthotic walking and walking in healthy subjects. Since walking with powered orthoses (mechanical orthoses that powered with external actuators in the hip or knee joints) needs to keep balance and other assistive devices such as walker or crutch need to activate the powered joints, therefore, the speed of walking is thought to be adversely affected by these conditions. Commercially developed exoskeletons have shown the potential to significantly improve speed of walking.
Since there were some studies about increased step length and cadence in wearing powered gait orthoses during walking, more research in this field is required. The mechanism of the additional external actuator to mechanical orthosis in the powered gait orthoses in improve‐ ment of speed of walking needs more research.
### *5.3.1. Quiet standing*
Mechanical orthosis has been analyzed in a few researches in stability by quite standing. Anterior–posterior (AP) and mediolateral (ML) COP displacement in amplitude and veloci‐ ty of were measured in SCI patients. Baardman evaluated the effect of connectorlink in ARGO and reported no difference between ARGO with (35.22 mm in AP and 41.72 mm in ML COP displacement) and without (37.94 mm in AP and 34.53 mm in ML direction) cable in quiet standing in stability [62].
In another study, Middleton et al. compared KAFO with and without a single medial linkage on sway amplitude in SCI patients announced that AP sway amplitude when using a single medial linkage with KAFO was half that when wearing a KAFO without it, but in ML direction amplitude there was no significant difference [75]. In a research Abe et al. compared stabili‐ ty with KAFO, RGO and WO and reported significant difference between KAFO with two other orthoses [75].
### *5.3.2. Performance*
Such as other parameters in PGOs there is no document in relation to performance and powered orthoses. About mechanical orthoses, there are three studies in relation to ampli‐ tude and velocity of COP displacement. The first study was reported by Baardman et al., which done on cable connector in the ARGO on performance. These researchers reported no effect of connector cable on performance (COP displacement in quite standing) but demonstrated this cable in ARGO can reduce upper limb load [62].
Middleton et al. compared KAFO with and without single medial linkage and announced no significant difference in sway path and amplitude in the A-P direction between them in SCI patients, but use of KAFO with a medial linkage decreased sway amplitude (p = 0.008) and increased M-L direction of sway path (p = 0.021). Therefore, this research proved using a single medial linkage can increase stability and balance in the mediolateral direction [75]. In the third study, Abe et al. demonstrated that sway amplitude was not significantly different between the WO and the RGO but the KAFO had significantly less stability [76].
### **6. Future research topics in this field**
The following topics can be performed in orthotics rehabilitation of the SCI subjects:
### **7. Summary remarks**
(Berkeley Bionics) are all examples of commercially developed powered orthoses designed for walking by paraplegic subjects. The Hybrid Assistive Limb (HAL)—6LB which has six electric motors—bilaterally at the hip, knee and ankle joints, the HAL-5 Type-C, which is previous HAL for paraplegia patients, with only four power units are another examples of powerlower limb orthosis. Commercially developed HAL has four power units and uses EMG signals of leg to synchronize the motion support with the wearer's movement, but complete paraple‐
Speed of walking reported difference between mechanical, hybrid and powered gait ortho‐ ses. There was statistically significant difference between orthotic walking and walking in healthy subjects. Since walking with powered orthoses (mechanical orthoses that powered with external actuators in the hip or knee joints) needs to keep balance and other assistive devices such as walker or crutch need to activate the powered joints, therefore, the speed of walking is thought to be adversely affected by these conditions. Commercially developed
Since there were some studies about increased step length and cadence in wearing powered gait orthoses during walking, more research in this field is required. The mechanism of the additional external actuator to mechanical orthosis in the powered gait orthoses in improve‐
Mechanical orthosis has been analyzed in a few researches in stability by quite standing. Anterior–posterior (AP) and mediolateral (ML) COP displacement in amplitude and veloci‐ ty of were measured in SCI patients. Baardman evaluated the effect of connectorlink in ARGO and reported no difference between ARGO with (35.22 mm in AP and 41.72 mm in ML COP displacement) and without (37.94 mm in AP and 34.53 mm in ML direction) cable in quiet
In another study, Middleton et al. compared KAFO with and without a single medial linkage on sway amplitude in SCI patients announced that AP sway amplitude when using a single medial linkage with KAFO was half that when wearing a KAFO without it, but in ML direction amplitude there was no significant difference [75]. In a research Abe et al. compared stabili‐ ty with KAFO, RGO and WO and reported significant difference between KAFO with two
Such as other parameters in PGOs there is no document in relation to performance and powered orthoses. About mechanical orthoses, there are three studies in relation to ampli‐ tude and velocity of COP displacement. The first study was reported by Baardman et al., which done on cable connector in the ARGO on performance. These researchers reported no effect of connector cable on performance (COP displacement in quite standing) but demonstrated
exoskeletons have shown the potential to significantly improve speed of walking.
gia patients cannot use this version of HAL.
266 268Recovery of Motor Function Following Spinal Cord Injury
ment of speed of walking needs more research.
this cable in ARGO can reduce upper limb load [62].
*5.3.1. Quiet standing*
standing in stability [62].
other orthoses [75].
*5.3.2. Performance*
thought to be adversely affected by these conditions. Commercially developed exoskele‐ tons have shown the potential to significantly improve speed of walking.
### **Author details**
Mokhtar Arazpour\* , Monireh Ahmadi Bani, Mohammad Ebrahim Mousavi, Mahmood Bahramizadeh and Mohammad Ali Mardani
\*Address all correspondence to: M.Arazpour@yahoo.com
Department of Orthotics and Prosthetics, University of Social Welfare and Rehabilitation Sciences, Tehran, Iran
### **References**
[8] Motlock WM. principles of orthotic management for child and adult paraplegia and clinical experience with the isocentric RGO. Proceeding of 7th world congress of the international society in prosthetic and orthotics 1992; Chicago, p 28.
thought to be adversely affected by these conditions. Commercially developed exoskele‐
, Monireh Ahmadi Bani, Mohammad Ebrahim Mousavi,
[1] Varma AK, Das A, Wallace IV G, Barry J, Vertegel AA, Ray SK, et al. Spinal cord injury: a review of current therapy, future treatments, and basic science frontiers. Neuro‐
[2] Bernardi M, Macaluso A, Sproviero E, Castellano V, Coratella D, Felici F, et al. Cost of walking and locomotor impairment. Journal of Electromyography and Kinesiology
[3] Hirokawa S, Grimm M, Solomonow M, Baratta R, Shoji H, D'ambrosia R. Energy consumption in paraplegic ambulation using the reciprocating gait orthosis and electric stimulation of the thigh muscles. Archives of Physical Medicine and Rehabilitation
[4] Johnson W, Fatone S, Gard S. Walking mechanics of persons who use reciprocating gait orthoses. Journal of Rehabilitation Research and Development 2009;46(3):435.
[5] Merati G, Sarchi P, Ferrarin M, Pedotti A, Veicsteinas A. Paraplegic adaptation to assisted-walking: energy expenditure during wheelchair versus orthosis use. Spinal
[6] Nene A, Hermens H, Zilvold G. Paraplegic locomotion: a review. Spinal Cord
[7] Solomonow M, Baratta R, Shoji H, et al., eds. FES powered locomotion of paraplegics fitted with the LSU reciprocating gait orthoses (RGO). Proceeding of Annual Interna‐
tional Conferance on IEEE Engineering. Med Biol Soc 1988;10:1672.
Department of Orthotics and Prosthetics, University of Social Welfare and Rehabilitation
tons have shown the potential to significantly improve speed of walking.
Mahmood Bahramizadeh and Mohammad Ali Mardani
\*Address all correspondence to: M.Arazpour@yahoo.com
chemical Research 2013;38(5):895–905.
**Author details**
268 270Recovery of Motor Function Following Spinal Cord Injury
Mokhtar Arazpour\*
Sciences, Tehran, Iran
1999;9(2):149–57.
1990;71(9):687–94.
Cord 2000;38(1):37–44.
1996;34(9):507–24.
**References**
[33] Ruthenberg B, Wasylewski N, Beard J. An experimental device for investigating the force and powerrequirements of a powered gait orthosis. Development 1997;34(2):203– 13.
[20] Petrofsky J, Smith JB. Physiologic costs of computer-controlled walking in persons with paraplegia using a reciprocating-gait orthosis. Archives of Physical Medicine and
[21] Arazpour M, Chitsazan A, Hutchins SW, Ghomshe FT, Mousavi ME, Takamjani EE, et al. Design and simulation of a new powered gait orthosis for paraplegic patients.
[22] Audu ML, To CS, Kobetic R, Triolo RJ. Gait evaluation of a novel hip constraint orthosis with implication for walking in paraplegia. Neural Systems and Rehabilitation
[23] Hornby TG, Zemon DH, Campbell D. Robotic-assisted, body-weight-supported treadmill training in individuals following motor incomplete spinal cord injury.
[24] Colombo G, Joerg M, Schreier R, Dietz V. Treadmill training of paraplegic patients using a robotic orthosis. Journal of Rehabilitation Research and Development
[25] Van Der Salm A, Nene AV, Maxwell DJ, Veltink PH, Hermens HJ, IJzerman MJ. Gait impairments in a group of patients with incomplete spinal cord injury and their relevance regarding therapeutic approaches using functional electrical stimulation.
[26] Hidler J. What is next for locomotor-based studies? Journal of Rehabilitation Research
[27] Harkema SJ, Hurley SL, Patel UK, Requejo PS, Dobkin BH, Edgerton VR. Human lumbosacral spinal cord interprets loading during stepping. Journal of Neurophysiol‐
[28] Beres-Jones JA, Harkema SJ. The human spinal cord interprets velocity-dependent
[29] Vukobratovic M, Hristic D, Stojiljkovic Z. Development of active anthropomorphic exoskeletons. Medical and Biological Engineering and Computing 1974;12(1):66–80.
[30] Belforte G, Gastaldi L, Sorli M. Pneumatic active gait orthosis. Mechatronics 2001;11(3):
[31] Belforte G, Eula G, Appendino S, Sirolli S. Pneumatic interactive gait rehabilitation orthosis: design and preliminary testing. Proceedings of the Institution of Mechanical
[32] Kang S, Ryu J, Moon I, Kim K, Mun M. Walker gait analysis of powered gait orthosis for paraplegic. In: Proceedings of world congress on medical physics and biomedical engineering IFMBE, COEX Seoul, Korea, 27 August –1 September 2006, vol. 14, pp.
Engineers, Part H: Journal of Engineering in Medicine 2011;225(2):158–69.
afferent input during stepping. Brain 2004;127(10):2232–46.
Prosthetics and Orthotics International 2012;36(1):125–30.
Engineering, IEEE Transactions on 2010;18(6):610–8.
Rehabilitation 1991;72(11):890–6.
270 272Recovery of Motor Function Following Spinal Cord Injury
Physical Therapy 2005;85(1):52–66.
Artificial Organs 2005;29(1):8–14.
and Development 2005;42(1):xi.
ogy 1997;77(2):797–811.
2889–2891. Springer, 2007.
301–23.
2000;37(6):693–700.
[59] Leung AK, Wong AF, Wong EC, Hutchins SW. The physiological cost index of walking with an isocentric reciprocating gait orthosis among patients with T12–L1 spinal cord injury. Prosthetics and Orthotics International 2009;33(1):61–8.
[46] Winchester P, Carollo J, Parekh R, Lutz L, Aston J. A comparison of paraplegic gait performance using two types of reciprocating gait orthoses. Prosthetics and Orthotics
[47] Nene A, Patrick J. Energy cost of paraplegic locomotion using the ParaWalker electrical stimulation ``hybrid" orthosis. Archives of Physical Medicine and Rehabili‐
[48] Hirokawa S, Grimm M. Energy consumption in paraplegic ambulation using the reciprocating gait orthosis and electric stimulation of the thigh muscles. Archives of
[49] Katz-Leurer M, Weber C, Smerling-Kerem J, Rottem H, Meyer S. Prescribing the reciprocal gait orthosis for myelomeningocele children: a different approach and
[50] Sykes L, Edwards J, Powell ES, Ross ERS. The reciprocating gait orthosis: long-term usage patterns. Archives of Physical Medicine and Rehabilitation 1995;76(8):779–83.
[51] Franceschini M, Baratta S, Zampolini M, Loria D, Lotta S. Reciprocating gait orthoses: a multicenter study of their use by spinal cord injured patients. Archives of Physical
[52] Scivoletto G, Petrelli A, Di Lucente L, Giannantoni A, Fuoco U, D'Ambrosio F, et al. One year follow up of spinal cord injury patients using a reciprocating gait orthosis:
[53] Jaspers P, Peeraer L, Van Petegem W, Van der Perre G. The use of an advanced reciprocating gait orthosis by paraplegic individuals: a follow-up study. Spinal Cord
[54] Kawashima N, Sone Y, Nakazawa K, Akai M, Yano H. Energy expenditure during walking with weight-bearing control (WBC) orthosis in thoracic level of paraplegic
[55] Arazpour M, Bani M, Hutchins S, Jones R. The physiological cost index of walking with mechanical and powered gait orthosis in patients with spinal cord injury. Spinal
[56] Tanabe S, Saitoh E, Hirano S, Katoh M, Takemitsu T, Uno A, et al. Design of the Wearable Power-Assist Locomotor (WPAL) for paraplegic gait reconstruction.
[57] Nene A, Patrick J. Energy cost of paraplegic locomotion with the ORLAU ParaWalker.
[58] Harvey LA, Davis GM, Smith MB, Engel S. Energy expenditure during gait using the walkabout and isocentric reciprocal gait orthoses in persons with paraplegia. Archives
Disability and Rehabilitation: Assistive Technology 2013;8(1):84–91.
of Physical Medicine and Rehabilitation 1998;79(8):945–9.
Physical Medicine and Rehabilitation 1990;71(9):687.
Medicine and Rehabilitation 1997;78(6):582–6.
preliminary report. Spinal Cord 2000;38(9):555–8.
patients. Spinal Cord 2003;41(9):506–10.
clinical outcome. Pediatric Rehabilitation 2004;7(2):105–9.
International 1993;17(2):101–6.
272 274Recovery of Motor Function Following Spinal Cord Injury
tation 1990;71(2):116.
1997;35(9):585–9.
Cord 2013;51(5):356–9.
Spinal Cord 1989;27(1):5–18.
orthoses in paraplegic patients. Prosthetics and Orthotics International 2013;37(2):161– 7.
## **Non-invasive Brain Stimulation to Characterize and Alter Motor Function after Spinal Cord Injury**
Aaron Z. Bailey, Hunter J. Fassett, Tea Lulic, Jenin El Sayes and Aimee J. Nelson
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/63351
### **Abstract**
orthoses in paraplegic patients. Prosthetics and Orthotics International 2013;37(2):161–
[72] Onogi K, Kondo I, Saitoh E, Kato M, Oyobe T. Comparison of the effects of sliding-type and hinge-type joints of knee-ankle-foot orthoses on temporal gait parameters in patients with paraplegia. Japanese Journal of Comprehensive Rehabilitation Science
[73] Suzuki T, Sonoda S, Saitoh E, Onogi K, Fujino H, Teranishi T, et al. Prediction of gait outcome with the knee–ankle–foot orthosis with medial hip joint in patients with spinal cord injuries: a study using recursive partitioning analysis. Spinal Cord 2007;45(1):57–
[74] Middleton JW, Sinclair PJ, Smith RM, Davis GM. Postural control during stance in paraplegia: effects of medially linked versus unlinked knee-ankle-foot orthoses.
[75] Abe K. Comparison of static balance, walking velocity, and energy consumption with knee-ankle-foot orthosis, walkabout orthosis, and reciprocating gait orthosis in thoracic-level paraplegic patients. JPO: Journal of Prosthetics and Orthotics 2006;18(3):
Archives of Physical Medicine and Rehabilitation 1999;80(12):1558–65.
7.
63.
87–91.
2010;1:1–6.
274 276Recovery of Motor Function Following Spinal Cord Injury
Advances in transcranial magnetic stimulation (TMS) now permit the precise assess‐ ment of circuitry in human motor cortices that contribute to movement. Further, TMS approaches are used to promote neural plasticity within cortical and spinal circuitry in an attempt to create short-term changes in motor control. This review is focused on the application of TMS techniques in the study of characterizing and promoting neural plasticity within individuals presenting with chronic spinal cord injury. We review TMS research performed in individuals with SCI and consider new opportunities for the use of TMS approaches to promote neural plasticity for improving motor recovery.
**Keywords:** spinal cord injury, motor recovery, noninvasive brain stimulation, trans‐ cranial magnetic stimulation, neural plasticity
### **1. Introduction**
Noninvasive brain stimulation (NIBS) approaches are used to both characterize and modify neural activity within the targeted cortices and the spinal cord. These approaches are, there‐ fore, well-suited for understanding and attempting to improve functional recovery following spinal cord injury. There is, however, a gap in our present understanding of how the cortical physiology is altered in individuals with spinal cord injury (SCI). In this review, we focus on the characterization of motor cortical circuitry in individuals with SCI using transcranial magnetic stimulation (TMS). This basic knowledge is essentialto developing therapies that aim to induce long-term changes in neural circuits that participate in motor control. NIBS ap‐
© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
proaches are also capable of inducing short-term plasticity within motor cortical neural circuits and may be combined with other techniques to promote greater functional changes in individ‐ ualswithSCI.Therefore,the secondgoalofthis reviewis tosurveythe literature thathasutilized NIBS approaches for promoting neural plasticity within the motor system in individuals with SCI.Wesubsequentlyexplorenewopportunities fortheuseofNIBStocharacterizeandpromote neural plasticity in SCI for improving motor recovery.
### **2. Noninvasive brain stimulation to characterize motor function**
NIBS provides an opportunity to assess the neural circuitry within motor cortical areas as well as the excitatory and inhibitory influences on neural output to specific muscles. These circuits may originate within the motor cortex (intracortical) and/or operate between cortical areas (intercortical) or hemispheres (interhemispheric). Following SCI, the excitability of the motor cortex and the corresponding cortical circuits may be altered, and these changes may be, in part, related to the functional reorganization of the central nervous system. TMS techniques offer the opportunity to characterize alterations in cortical circuits following SCI and are described following. Importantly, these techniques are ideal for monitoring changes that accompany motor recovery in SCI populations.
### **2.1. Motor-evoked potentials (MEPs)**
TMS delivered over the motor cortex trans-synaptically activates corticospinal neurons that ultimately give rise to a motor-evoked potential (MEP) recorded in the target muscle. The amplitude of the MEP represents the overall corticospinal excitability including contribu‐ tions from the upper and lower motorneurons and the functional integrity of the corticospi‐ nal path [1]. The MEP amplitude may be modulated by factors such as action planning [2, 3], pharmaceuticals [4], or neuromuscular diseases [5–7]. The evaluation of MEP amplitude is a commonly used method to quantify changes in excitability within the descending motor pathways following experimental manipulation [8] or during the process of recovery [9]. The MEP latency provides an additional source of neurophysiological information and repre‐ sents the conduction time for descending signals originating within the motor cortex to travel to the muscle effector, a measure that serves a diagnostic purpose for a number of disease states [10].
Following SCI, MEP amplitude and latency are often altered. When using TMS to assess muscles caudal to the spinal lesion level, smaller amplitude and longer latency MEPs are reported [11, 12]. This is thought to be due to the structural damage to descending corticospi‐ nal fibers including fewer intact axons and significant demyelination surrounding the lesion [13]. In individuals with SCI, such decrements in MEP amplitude may be lessened by introducing low level active contraction in the muscle of interest [14]. However, voluntary control of a muscle is not necessarily indicative of corticospinal connectivity to that particu‐ lar muscle since MEPs may be elicited in muscles not under volitional control in individuals with SCI [15].
The origin of MEPs is complex, as pyramidal tract neurons are influenced directly and indirectly (trans-synaptically) in response to cortical stimulation. Additionally, MEPs depend on spinal motorneuron excitability and are influenced by cortical areas outside of motor cortex such as the premotor and supplementary motor areas [8]. As such, the ambiguity in the origin of the MEP creates challenges in its interpretation. Nonetheless, MEP characteristics have a value for assessing motor recovery following SCI such that both MEP amplitude and latency, in the acute stage of injury, predict the recovery of ambulation and hand function [10].
### **2.2. Motor threshold (MT)**
proaches are also capable of inducing short-term plasticity within motor cortical neural circuits and may be combined with other techniques to promote greater functional changes in individ‐ ualswithSCI.Therefore,the secondgoalofthis reviewis tosurveythe literature thathasutilized NIBS approaches for promoting neural plasticity within the motor system in individuals with SCI.Wesubsequentlyexplorenewopportunities fortheuseofNIBStocharacterizeandpromote
NIBS provides an opportunity to assess the neural circuitry within motor cortical areas as well as the excitatory and inhibitory influences on neural output to specific muscles. These circuits may originate within the motor cortex (intracortical) and/or operate between cortical areas (intercortical) or hemispheres (interhemispheric). Following SCI, the excitability of the motor cortex and the corresponding cortical circuits may be altered, and these changes may be, in part, related to the functional reorganization of the central nervous system. TMS techniques offer the opportunity to characterize alterations in cortical circuits following SCI and are described following. Importantly, these techniques are ideal for monitoring changes that
TMS delivered over the motor cortex trans-synaptically activates corticospinal neurons that ultimately give rise to a motor-evoked potential (MEP) recorded in the target muscle. The amplitude of the MEP represents the overall corticospinal excitability including contribu‐ tions from the upper and lower motorneurons and the functional integrity of the corticospi‐ nal path [1]. The MEP amplitude may be modulated by factors such as action planning [2, 3], pharmaceuticals [4], or neuromuscular diseases [5–7]. The evaluation of MEP amplitude is a commonly used method to quantify changes in excitability within the descending motor pathways following experimental manipulation [8] or during the process of recovery [9]. The MEP latency provides an additional source of neurophysiological information and repre‐ sents the conduction time for descending signals originating within the motor cortex to travel to the muscle effector, a measure that serves a diagnostic purpose for a number of disease
Following SCI, MEP amplitude and latency are often altered. When using TMS to assess muscles caudal to the spinal lesion level, smaller amplitude and longer latency MEPs are reported [11, 12]. This is thought to be due to the structural damage to descending corticospi‐ nal fibers including fewer intact axons and significant demyelination surrounding the lesion [13]. In individuals with SCI, such decrements in MEP amplitude may be lessened by introducing low level active contraction in the muscle of interest [14]. However, voluntary control of a muscle is not necessarily indicative of corticospinal connectivity to that particu‐ lar muscle since MEPs may be elicited in muscles not under volitional control in individuals
**2. Noninvasive brain stimulation to characterize motor function**
neural plasticity in SCI for improving motor recovery.
276 278Recovery of Motor Function Following Spinal Cord Injury
accompany motor recovery in SCI populations.
**2.1. Motor-evoked potentials (MEPs)**
states [10].
with SCI [15].
MT is defined as the minimal stimulation intensity that elicits a MEP in a target muscle. MT can be assessed in two ways. Resting motor threshold (RMT) is often obtained by finding the lowest stimulation intensity that produces a MEP of at least 50 μV in 5 out of 10 consecutive trials, while the target muscle is relaxed [16, 17]. Furthermore, motor threshold can be determined while the participant maintains low levels of muscle contraction (i.e., active motor threshold or AMT). AMT is commonly defined as the lowest stimulation intensity to elicit MEPs of at least 200 μV in 5 out of 10 consecutive trials, while the participant maintains a contraction corresponding to 10–15% of their maximum [16]. RMT represents the resting membrane excitability throughout the efferent system at both cortical and spinal levels [18]. During this resting state, spinal motorneurons require a number of descending volleys from the cortex to summate in order to produce a MEP. This requirement for greater corticospinal summation at rest has been speculated to be due to glutamatergic synaptic activity [18]. Pharmacological studies have shown that AMT and RMT depend on the axon membrane excitability, specifically on voltage-gated sodium channel activity [19]. Application of various drugs that block the action of these channels such as carbamazepine and lamotrigine in‐ crease the required stimulation intensity to elicit RMT and AMT (see Paulus et al. 2008 for review) [18]. Despite the factthat pharmacological studies have not been able to isolate changes in RMT without a simultaneous change in AMT [18], differences in levels of synaptic excita‐ bility have been inferred from studies of MT latency, where RMT is seen to have a longer latency than AMT [20]. By maintaining a small contraction, low levels of synaptic firing are sustained and are thought to be more reliant on levels of excitability at the axonal mem‐ brane.[18].
Individuals with SCI may demonstrate increases in RMT and AMT. Davey et al. [11] found that RMT and AMT were higher for muscles below the level of injury whereas muscles innervated by fibers rostral to the injury had thresholds similarto the uninjured control group. Other research has shown that alterations following SCI are more prevalent when examin‐ ing AMT [21, 22]. Bailey et al. [22] demonstrated increased AMT elicited from the flexor carpi radialis muscle of the forearm in participants with chronic cervical SCI. They suggested that this increase in AMT without a corresponding increase in RMT may be due to abnormalities in regulation of membrane excitability via voltage-gated ion channels. Collectively, the published data suggest that RMT may or may not be altered while AMT is increased in SCI participants. However, it has been established that changes in both RMT and AMT are dependent on the location of the target muscle relative to the level of the lesion. Further, changes in MT in chronic SCI are likely to differ depending on the severity and location of injury.
### **2.3. Motor cortical maps**
The motor cortex is somatotopically organized in that cortical representations of the muscles are found in relatively predictable locations in the medial–lateral direction [23]. The extent of cortical territory occupied by a given muscle representation is thought to correspond to the amount of dexterity and fine motor control needed by the corresponding muscle such that larger representations exist for small distal muscles of the hand while smaller representa‐ tions are presentforlarger proximal muscles ofthe arm and shoulder [12]. Most cortical muscle representations correspond to effectors of the contralateral limb although some degrees of ipsilateral projections have been found in proximal upper limb muscles in uninjured individ‐ uals [24]. Motor cortical maps can be obtained by delivering TMS pulses at suprathreshold intensities using a spatial grid aligned with the precentral gyrus and recording MEPs from target muscles [25]. Motor cortical maps are typically evaluated by quantifying the area, volume, and center of gravity for the representation of each muscle of interest. Map area and volume scale linearly with increasing TMS intensity and are considered a reliable method to evaluate cortical organization [26].
Studies of motor maps following SCI revealed enlarged representations of the most proximal‐ ly spared muscles thattended to shiftinto corticalterritory corresponding to the muscles below the level of injury [27, 28]. Takeover of de-innervated territory was supported by findings in which the representation of the extensor digitorum communis muscle was shifted anterolaterally (into the hand region) following incomplete cervical SCI [21, 29]. Brouwer and Hopkins-Fosseel [12] quantified motor maps in SCI participants for a variety of contracted upper arm muscles and found that proximal muscles were represented to a greater extent than distal hand muscles, and there was an extensive overlap of different muscle representations when compared to uninjured participants. This overlap may be beneficial for organization as the neurons projecting to single muscles are not focally distributed and may act to minimize complete functional losses following neurological injury. Cortical reorganization following SCI occurs rapidly as evident by the changes in the shape and location of biceps brachii motor cortical maps only 6–17 days following SCI [29]. More recently, a study using neuroimaging techniques demonstrated that representations of the tongue were significantly shifted in their location within motor cortex in participants with cervical SCI compared to uninjured partici‐ pants [30]. Collectively, the motor cortical maps indicate that the reorganization of motor cortex somatotopy takes place following SCI, suggesting that plasticity may compensate for the loss of neuronal communication between the cortex and the muscles of the body. Although the mechanisms for such reorganization are unknown in humans, it may be speculated that changes in cortical organization are shaped through GABA mediated lateral inhibitory circuits [31].
A number of practical considerations are important for cortical mapping studies using TMS. Mapping protocols require many grid points with approximately 3–10 pulses delivered to each point to capture cortical organization with appropriate resolution. This may require a number of hours to generate a motor map if a large number of points are used [26]. Furthermore, heightened motor thresholds commonly accompany SCI, requiring higher stimulation intensities to elicit MEPs, and this may impact the comfort of the participant and increase the likelihood of coil heating.
### **2.4. Short-interval intracortical inhibition (SICI)**
changes in MT in chronic SCI are likely to differ depending on the severity and location of
The motor cortex is somatotopically organized in that cortical representations of the muscles are found in relatively predictable locations in the medial–lateral direction [23]. The extent of cortical territory occupied by a given muscle representation is thought to correspond to the amount of dexterity and fine motor control needed by the corresponding muscle such that larger representations exist for small distal muscles of the hand while smaller representa‐ tions are presentforlarger proximal muscles ofthe arm and shoulder [12]. Most cortical muscle representations correspond to effectors of the contralateral limb although some degrees of ipsilateral projections have been found in proximal upper limb muscles in uninjured individ‐ uals [24]. Motor cortical maps can be obtained by delivering TMS pulses at suprathreshold intensities using a spatial grid aligned with the precentral gyrus and recording MEPs from target muscles [25]. Motor cortical maps are typically evaluated by quantifying the area, volume, and center of gravity for the representation of each muscle of interest. Map area and volume scale linearly with increasing TMS intensity and are considered a reliable method to
Studies of motor maps following SCI revealed enlarged representations of the most proximal‐ ly spared muscles thattended to shiftinto corticalterritory corresponding to the muscles below the level of injury [27, 28]. Takeover of de-innervated territory was supported by findings in which the representation of the extensor digitorum communis muscle was shifted anterolaterally (into the hand region) following incomplete cervical SCI [21, 29]. Brouwer and Hopkins-Fosseel [12] quantified motor maps in SCI participants for a variety of contracted upper arm muscles and found that proximal muscles were represented to a greater extent than distal hand muscles, and there was an extensive overlap of different muscle representations when compared to uninjured participants. This overlap may be beneficial for organization as the neurons projecting to single muscles are not focally distributed and may act to minimize complete functional losses following neurological injury. Cortical reorganization following SCI occurs rapidly as evident by the changes in the shape and location of biceps brachii motor cortical maps only 6–17 days following SCI [29]. More recently, a study using neuroimaging techniques demonstrated that representations of the tongue were significantly shifted in their location within motor cortex in participants with cervical SCI compared to uninjured partici‐ pants [30]. Collectively, the motor cortical maps indicate that the reorganization of motor cortex somatotopy takes place following SCI, suggesting that plasticity may compensate for the loss of neuronal communication between the cortex and the muscles of the body. Although the mechanisms for such reorganization are unknown in humans, it may be speculated that changes in cortical organization are shaped through GABA mediated lateral inhibitory
A number of practical considerations are important for cortical mapping studies using TMS. Mapping protocols require many grid points with approximately 3–10 pulses delivered to each point to capture cortical organization with appropriate resolution. This may require a number
injury.
**2.3. Motor cortical maps**
278 280Recovery of Motor Function Following Spinal Cord Injury
evaluate cortical organization [26].
circuits [31].
SICI is an inhibitory circuit within the motor cortex that is comprised of low-threshold GABAergic interneurons [18] and can be evaluated using paired-pulse TMS. By delivering an initial subthreshold conditioning stimulus (CS) approximately 1–5 ms prior to a suprathres‐ hold test stimulus (TS) to the motor cortex, the resulting MEP is reduced when compared to MEPs produced when the TS is applied alone [32]. The subthreshold CS is typically set at an intensity of 70–90% of AMT, showing that the MEP modulation is of cortical origin as no descending volleys to the muscles are produced at these intensities [32]. The mechanism of inhibition involves presynaptic inhibition as inhibitory postsynaptic potentials are generat‐ ed on neurons upstream of the corticospinal output neurons [18]. GABAA receptors are implicated in SICI since The delivery of Lorazepam or Diazepam (GABAA Receptor agonist) reduces SICI [33, 34].
It is well documented that SICI is reduced in individuals with SCI. One case report exam‐ ined SICI in the relaxed extensor digitorum communis muscle using various interstimulus intervals and a single CS intensity [35]. This method was also used elsewhere in the relaxed first dorsal interosseous muscle [36]. Both cases demonstrated a reduction in SICI when compared to controls [35, 36]. However, the aforementioned studies used a single CS and TS intensity, making interpretation difficult, as the required stimulator output to produce MEPs in SCI participants is typically greater relative to able-bodied individuals. Roy et al. [37] explored SICI in SCI participants by measuring recruitment curves whereby CS intensities were altered from 60 to 110% AMT. They recorded responses from both tibialis anterior in the lower leg and the first dorsal interosseous muscle of the hand during contraction of each muscle. The results indicated a reduction of SICI responses in SCI participants compared to controls, although both retained a U-shaped recruitment curve. Individuals with SCI were seen to have a smaller range of CS intensities eliciting inhibition of the MEP when compared to the recruitment curves of uninjured individuals [37]. Therefore, recruitment characteris‐ tics of SICI appear to be unchanged following SCI while the overall amount of inhibition is reduced. The authors speculate that this may be due to spinal inhibitory mechanisms, or to reductions in GABAergic activity following cortical reorganization that occurs with SCI. Mi et al. [38] examined SICI in chronic cervical SCI. Similar to the results found by Roy et al. [37], the authors identified a smaller range of CS intensities to elicit SICI in the flexor carpi radialis muscle ofthe forearm compared to controls. In this case, significant inhibition ofthe TS-evoked response was only seen at 90% AMT in SCI compared to the wider range of 70, 80, and 90% AMT in controls [38]. This may reflect the reduced excitability of corticospinal neurons following SCI as the pattern of recruitment is maintained, while the range of inhibition is reduced. Therefore, impairment in GABAergic circuitry mediating SICI in the motor cortex has been established following lesion to the spinal cord.
### **2.5. Long-interval intracortical inhibition (LICI)**
LICI represents intracortical circuitry that acts to modulate MEPs and can be observed via paired-pulse TMS to the motor cortex. The inhibitory pathways involved in LICI can be probed by delivering a suprathreshold CS before a subsequent suprathreshold TS with an interval between 50 and 200 ms to the motor cortex [39, 40]. This produces two MEPs; however, the amplitude of the second MEP is reduced (i.e., inhibited) compared to the MEP evoked by delivery of the single TS pulse. Similar to SICI, the interneuron interactions leading to MEP inhibition are mediated by GABA as seen in pharmacological studies [41, 42]. The longer interstimulus intervals leading to the inhibition are due to GABAB receptor activity which has been confirmed in studies using GABAB receptor agonists such as baclofen to increase the inhibitory response [43].
Limited research has examined alterations in LICI after SCI. Barry et al. [43] compared LICI in SCI versus uninjured controls in the resting and active muscle states to examine the effects of GABAB agonist Baclofen. They examined the first dorsal interosseous muscle of the hand, which showed a decrease in LICI in the active state in the control group as well as in SCI individuals that were taking Baclofen. However, SCI participants not medicated with Baclofen did not show a reduction in LICI during active contraction. This finding suggests abnormal GABAB receptor activity following SCI as the MEP responses continue to be further inhibit‐ ed with increasing tonic efferent activity introduced by sustained muscle contraction. Interestingly, this decrease in LICI associated with muscle contraction appears to be rectified by the application of Baclofen, normalizing the inhibitory response in SCI to that of the control group. Another study examined the recruitment of the LICI circuitry in the active muscle state by applying CS intensities at 10% increments from 90 to 130% AMT to capture the pattern of GABAB mediated response in SCI [38]. An increase in LICI was observed in SCI participants at higher CS intensities (120 and 130% AMT), while the uninjured group did not show any LICI with increasing CS intensity. The authors attributed the increase in inhibition to Baclofen, which as a GABAB agonist may return LICI circuitry to its normative state.
### **2.6. Intracortical facilitation (ICF)**
ICF can be observed within motor cortex, demonstrating the presence of excitatory circuits impacting the descending corticospinal output. ICF is elicited by delivering a subthreshold TMS pulse over the motor representation of a muscle followed by a suprathreshold TS after an interval of 7–20 ms [32]. This paired-pulse TMS method elicits a net facilitation of the MEP as seen by increased response amplitudes when compared to MEPs evoked from the single TS [33]. Pharmacological approaches reveal that ICF is the net effect of both inhibitory and excitatory contributions. Application of GABAA receptor agonists such as Diazepam result in decreased ICF, showing an influence of GABAergic inhibition onto corticospinal output neurons [44]. This demonstrates that although ICF is affected by GABAergic activity [44], there is a large excitatory component resulting in net facilitation. Thus, it is speculated that that NMDA receptors play a role in the facilitation observed in ICF [45]. These receptors respond to glutamate when nearby AMPA receptors are active to excite postsynaptic neurons. The role of NMDA receptor activity in ICF is supported by the finding that the application of NMDA receptor antagonists also causes a decrease in facilitation, implicating their role in the physiology of ICF [45].
A single case study examined ICF in a participant with cervical myelopathy using a CS intensity of 80% RMT and TS intensity of 120% RMT at a number of different interstimulus intervals. ICF was only observed when the interstimulus interval was 10 ms [35]. Further research with larger sample sizes and control groups should examine alterations in ICF in individuals with SCI. Collection of recruitment curves for ICF may provide a more compre‐ hensive view of excitatory mechanisms contributing to corticospinal output following SCI.
### **2.7. Short and long latency afferent inhibition (SAI/LAI)**
**2.5. Long-interval intracortical inhibition (LICI)**
280 282Recovery of Motor Function Following Spinal Cord Injury
inhibitory response [43].
**2.6. Intracortical facilitation (ICF)**
LICI represents intracortical circuitry that acts to modulate MEPs and can be observed via paired-pulse TMS to the motor cortex. The inhibitory pathways involved in LICI can be probed by delivering a suprathreshold CS before a subsequent suprathreshold TS with an interval between 50 and 200 ms to the motor cortex [39, 40]. This produces two MEPs; however, the amplitude of the second MEP is reduced (i.e., inhibited) compared to the MEP evoked by delivery of the single TS pulse. Similar to SICI, the interneuron interactions leading to MEP inhibition are mediated by GABA as seen in pharmacological studies [41, 42]. The longer interstimulus intervals leading to the inhibition are due to GABAB receptor activity which has been confirmed in studies using GABAB receptor agonists such as baclofen to increase the
Limited research has examined alterations in LICI after SCI. Barry et al. [43] compared LICI in SCI versus uninjured controls in the resting and active muscle states to examine the effects of GABAB agonist Baclofen. They examined the first dorsal interosseous muscle of the hand, which showed a decrease in LICI in the active state in the control group as well as in SCI individuals that were taking Baclofen. However, SCI participants not medicated with Baclofen did not show a reduction in LICI during active contraction. This finding suggests abnormal GABAB receptor activity following SCI as the MEP responses continue to be further inhibit‐ ed with increasing tonic efferent activity introduced by sustained muscle contraction. Interestingly, this decrease in LICI associated with muscle contraction appears to be rectified by the application of Baclofen, normalizing the inhibitory response in SCI to that of the control group. Another study examined the recruitment of the LICI circuitry in the active muscle state by applying CS intensities at 10% increments from 90 to 130% AMT to capture the pattern of GABAB mediated response in SCI [38]. An increase in LICI was observed in SCI participants at higher CS intensities (120 and 130% AMT), while the uninjured group did not show any LICI with increasing CS intensity. The authors attributed the increase in inhibition to Baclofen,
which as a GABAB agonist may return LICI circuitry to its normative state.
ICF can be observed within motor cortex, demonstrating the presence of excitatory circuits impacting the descending corticospinal output. ICF is elicited by delivering a subthreshold TMS pulse over the motor representation of a muscle followed by a suprathreshold TS after an interval of 7–20 ms [32]. This paired-pulse TMS method elicits a net facilitation of the MEP as seen by increased response amplitudes when compared to MEPs evoked from the single TS [33]. Pharmacological approaches reveal that ICF is the net effect of both inhibitory and excitatory contributions. Application of GABAA receptor agonists such as Diazepam result in decreased ICF, showing an influence of GABAergic inhibition onto corticospinal output neurons [44]. This demonstrates that although ICF is affected by GABAergic activity [44], there is a large excitatory component resulting in net facilitation. Thus, it is speculated that that NMDA receptors play a role in the facilitation observed in ICF [45]. These receptors respond to glutamate when nearby AMPA receptors are active to excite postsynaptic neurons. The role of NMDA receptor activity in ICF is supported by the finding that the application of NMDA
SAI demonstrates the influence of a peripherally evoked afferent volley on the MEP and is observed when an electrical stimulus applied to a peripheral nerve is followed by a single suprathreshold TMS pulse to the motor cortex. The temporal interval between the peripher‐ al and cortical stimulation corresponds to the conduction time for the afferent volley to reach the somatosensory cortex. This timing causes the afferent information to arrive shortly before the suprathreshold TMS pulse, which is subsequently inhibited compared to the MEP evoked by the TS pulse delivered alone. This circuit is commonly studied in intrinsic muscles of the hand where the interstimulus interval is roughly 18–21 ms [46, 47]. The exact mechanisms underpinning this pathway remains unclear, although pharmacological intervention using drugs such as Scopolamine (cholinergic antagonist) or Diazepam indicate that SAI is mediat‐ ed by both cholinergic and GABAergic circuits [48, 49]. LAI is another example of sensory modulation to cortical pathways involved in MEP generation. The method of eliciting LAI is similar to that of SAI with the exception of a longer interstimulus interval of ~100–200 ms [46, 50]. LAI is speculated to modulate corticospinal output by exclusively GABAergic systems, particularly those mediated by GABAB receptors [49]. The modulation of MEPs induced by sensory input suggests that SAI and LAI provide insight into sensorimotor integration [51].
The literature surrounding afferent regulation of the motor cortex following SCI is minimal. Roy et al. [14] stimulated the common peroneal nerve at the ankle to modify MEPs elicited from tibialis anterior using interstimulus intervals ranging from 30 to 80 ms. Results indicat‐ ed an inhibition of MEPs from tibialis anterior when the internal between the TS pulse and the peripheral electrical stimulation corresponded to the latency of the afferent volley from the common peroneal nerve. However, this effect was reduced in the SCI group. The decrease in SAI relative to controls is expected in this population, as the transmission of afferent informa‐ tion from the periphery to the cortex is largely dependent on spinal mechanisms that are likely to be affected by damage to the spinal cord. Further, a recent study examined SAI within the flexor carpi radialis muscle in cervical SCI participants and found that, compared to unin‐ jured controls SAI was reduced when the muscle was in the active and resting state [22]. The authors speculated that this impairment in SAI circuitry is due to plasticity effects in the processing and/or transmission of the information within the cortex as well as reductions in afferent transmission to the cortex. Thus, literature regarding sensory modulation of motor output suggests that afferent integration is impaired following SCI.
### **2.8. Cortical silent period (CSP)**
The CSP represents descending inhibitory signals during active contraction of the muscle. When maintaining a constant level of muscle contraction, stimulation of the motor cortex at suprathreshold intensities produces a CSP in the active contralateral muscle. The duration of the CSP may provide some indication offunctional ability. For example, the ability to modulate muscle activity following large descending efferent volleys may impact motor tasks involv‐ ing fine motor control where small, precise movements are required. The CSP is commonly ~200 ms in duration and is impacted by both cortical and spinal factors. Initial inhibition (~50– 75 ms) is attributed to the refractory period of efferent fibers within the spinal cord follow‐ ing suprathreshold TMS, while the remaining duration is thought to originate from the cortex [52–54]. This effect is physiologically mediated by multiple GABAergic systems as treatment with pharmaceuticals that reduce synaptic clearance of GABA increase CSP durations [42, 55]. At the receptor level, GABAA appears to modulate the CSP at lower TMS intensities as benzodiazepines increase the CSP duration [52]. However, at high TMS stimu‐ lation intensities, treatment with benzodiazepines shortens the CSP [56]. The latter finding suggests that GABAA activity suppresses the effects of GABAB receptors when such inputs are presented to the motor cortex [56].
The CSP duration appears to be altered in SCI. Shimizu et al. [36] investigated the CSP in three individuals with cervical SCI. In all participants, there was no observable CSP in the foot muscle, and in two individuals, the CSP was additionally absent in the hand muscles. The loss of this silent period suggests that there may have been reorganization at the level of the cortex or hyper excitability of the cortex as GABA systems are suppressed [36]. Barry et al. [43] examined the GABA system involved in producing the CSP in intrinsic hand muscles of chronic, incomplete SCI groups that either were or were not taking Baclofen. Baclofen did not alterthe CSP in SCI, and all SCI participants displayed a longer CSP duration than age-matched uninjured controls [43]. Another study examined the CSP, while generating motor maps of forearm extensor muscles and reported an inverse relationship between the duration of the CSP and the amount of spinal cord atrophy via inspection of MRI images [21]. Therefore, when increases in the CSP are observed in SCI, it is likely due to intact inhibitory corticospinal projections to the muscle although spinal modulation of inhibitory circuits is impaired.
### **2.9. Inter-hemispheric inhibition (IHI)**
IHI is a circuit that results from the extensive interconnectivity between the two cerebral hemispheres via the corpus callosum. The integrity of this cross-communication can be assessed by paired-pulse TMS over the motor cortices. By providing a suprathreshold CS over the motor representation of a muscle in one hemisphere followed by a suprathreshold TS over the homologous motor representation in the opposite hemisphere, inhibition of the MEP in response to the CS is observed [57, 58]. This suppression occurs at interstimulus intervals of 10 and 40 ms [57]. Modulation of MEP amplitude with IHI is mediated via GABAergic systems as application of GABAB agonist Baclofen induces an increase in inhibition at both paired-pulse intervals [59].
IHI in SCI has not been studied extensively. One report used an interstimulus interval of 10 ms for probing IHI in cervical SCI and found that there was neither enhancement nor suppression of the MEP when the participants were at rest or maintaining a contraction at a level that corresponded to either 30 or 70% MVC [60]. The authors speculate that the lack of change in IHI may be due to interactions with other circuitry that has been affected by injury such as SICI. One challenge in measuring IHI in the SCI population is that a MEP must be obtained with amplitude that is large enough to observe suppression following modulation from the opposite hemisphere. Decreased muscle responses are well established in SCI, making it difficult to elicit a MEP that is large enough to be modulated by the CS of the opposite hemisphere. Therefore, the IHI protocol may need to be adapted for the SCI population.
### **2.10. Summary**
**2.8. Cortical silent period (CSP)**
282 284Recovery of Motor Function Following Spinal Cord Injury
presented to the motor cortex [56].
**2.9. Inter-hemispheric inhibition (IHI)**
intervals [59].
The CSP represents descending inhibitory signals during active contraction of the muscle. When maintaining a constant level of muscle contraction, stimulation of the motor cortex at suprathreshold intensities produces a CSP in the active contralateral muscle. The duration of the CSP may provide some indication offunctional ability. For example, the ability to modulate muscle activity following large descending efferent volleys may impact motor tasks involv‐ ing fine motor control where small, precise movements are required. The CSP is commonly ~200 ms in duration and is impacted by both cortical and spinal factors. Initial inhibition (~50– 75 ms) is attributed to the refractory period of efferent fibers within the spinal cord follow‐ ing suprathreshold TMS, while the remaining duration is thought to originate from the cortex [52–54]. This effect is physiologically mediated by multiple GABAergic systems as treatment with pharmaceuticals that reduce synaptic clearance of GABA increase CSP durations [42, 55]. At the receptor level, GABAA appears to modulate the CSP at lower TMS intensities as benzodiazepines increase the CSP duration [52]. However, at high TMS stimu‐ lation intensities, treatment with benzodiazepines shortens the CSP [56]. The latter finding suggests that GABAA activity suppresses the effects of GABAB receptors when such inputs are
The CSP duration appears to be altered in SCI. Shimizu et al. [36] investigated the CSP in three individuals with cervical SCI. In all participants, there was no observable CSP in the foot muscle, and in two individuals, the CSP was additionally absent in the hand muscles. The loss of this silent period suggests that there may have been reorganization at the level of the cortex or hyper excitability of the cortex as GABA systems are suppressed [36]. Barry et al. [43] examined the GABA system involved in producing the CSP in intrinsic hand muscles of chronic, incomplete SCI groups that either were or were not taking Baclofen. Baclofen did not alterthe CSP in SCI, and all SCI participants displayed a longer CSP duration than age-matched uninjured controls [43]. Another study examined the CSP, while generating motor maps of forearm extensor muscles and reported an inverse relationship between the duration of the CSP and the amount of spinal cord atrophy via inspection of MRI images [21]. Therefore, when increases in the CSP are observed in SCI, it is likely due to intact inhibitory corticospinal projections to the muscle although spinal modulation of inhibitory circuits is impaired.
IHI is a circuit that results from the extensive interconnectivity between the two cerebral hemispheres via the corpus callosum. The integrity of this cross-communication can be assessed by paired-pulse TMS over the motor cortices. By providing a suprathreshold CS over the motor representation of a muscle in one hemisphere followed by a suprathreshold TS over the homologous motor representation in the opposite hemisphere, inhibition of the MEP in response to the CS is observed [57, 58]. This suppression occurs at interstimulus intervals of 10 and 40 ms [57]. Modulation of MEP amplitude with IHI is mediated via GABAergic systems as application of GABAB agonist Baclofen induces an increase in inhibition at both paired-pulse
Please refer to **Table 1** for a summary of the previous findings. Changes in motor cortical excitability and circuitry follow SCI and yield alterations in corticospinal output. The availa‐ ble information regarding changes in cortical circuitry for motor output in SCI is limited, making a comprehensive view of neurophysiological rehabilitation difficult in this popula‐ tion. Thus, further studies should identify and quantify aberrant motor cortical circuits in SCI.
**Table 1.** Cortical circuitry in spinal cord injury.
### **3. NIBS to induce plasticity in individuals with SCI**
NIBS provide an opportunity to modulate the neural circuits that are altered following SCI and have the potential to improve motor function. There are several NIBS protocols that have been used to promote plasticity with the two main forms including repetitive TMS (rTMS), and transcranial direct current stimulation (tDCS). RTMS and tDCS protocols are each founded in the principle of homosynaptic plasticity, while TMS and tDCS protocols paired with peripheral nerve stimulation have effects based on spike-timing dependent plasticity (STDP). Homosynaptic plasticity refers to plasticity occurring at a single synapse that is undergoing stimulation [61, 62]. STDP refers to plasticity induced by timing two stimuli, typically a cortical and a peripheral stimulus, to activate both the presynaptic and postsynaptic neurons coinci‐
dentally. These protocols are able to produce effects that resemble long-term potentiation (LTP) or long-term depression (LTD) of synaptic connectivity. Typically, activation of the presynaptic neuron priorto the postsynaptic neuron lends to the generation of LTP while LTD occurs when the postsynaptic neuron is activated first [61]. STDP can be timed such that the two stimuli coincide at different levels of the central nervous system to create either spinal or cortical plasticity. We provide an overview of the literature that has used NIBS approaches to promote plasticity in the motor system in individuals with SCI.
### **3.1. Repetitive TMS (rTMS)**
**Measure Source Classification Muscle tested Response**
Ischemic Myelopathy Tetraplegic Para/Tetraplegic Tetraplegic
Tetraplegic Tetraplegic
Myelopathy
Tetraplegia Tetraplegia
Tetraplegic Tetraplegic Tetraplegic
NIBS provide an opportunity to modulate the neural circuits that are altered following SCI and have the potential to improve motor function. There are several NIBS protocols that have been used to promote plasticity with the two main forms including repetitive TMS (rTMS), and transcranial direct current stimulation (tDCS). RTMS and tDCS protocols are each founded in the principle of homosynaptic plasticity, while TMS and tDCS protocols paired with peripheral nerve stimulation have effects based on spike-timing dependent plasticity (STDP). Homosynaptic plasticity refers to plasticity occurring at a single synapse that is undergoing stimulation [61, 62]. STDP refers to plasticity induced by timing two stimuli, typically a cortical and a peripheral stimulus, to activate both the presynaptic and postsynaptic neurons coinci‐
Saturno et al. [35] Shimizu et al. [36] Roy et al. [37] Mi et al. [38]
284 286Recovery of Motor Function Following Spinal Cord Injury
Barry et al. [43] Mi et al. [38]
Roy et al. [14] Bailey et al. [22]
Bunday and Perez [60]
**Table 1.** Cortical circuitry in spinal cord injury.
Shimizu et al. [36] Barry et al. [43] Freund et al. [21]
Saturno et al. [35] Ischemic
**3. NIBS to induce plasticity in individuals with SCI**
**Short-interval intracortical inhibition**
**Long-interval intracortical inhibition**
**Intracortical facilitation**
**Short-latency afferent inhibition**
**Cortical silent period**
**Interhemispheric inhibition**
Wrist extensors, Tongue
EDC FDI, FHB TA, FDI FDI
FDI FDI
TA FCR
FDI, FHB FDI EDC
Tetraplegic FDI No change
↑ Map area of Biceps, Bicep map
↓ Map volumes, No change in area No change in activation volume, peak response site shifted medially
↓ magnitude of active SICI, similar recruitment with increasing CS
No change in magnitude, reduced range of intensities showing SICI
no change with baclofen, ↑ active
LICI without baclofen ↑ active LICI
↓ active and resting SAI
↑ CSP duration, correlated with
↓ SAI (absent)
↓ CSP duration
spinal atrophy ↑ CSP duration
shifted laterally
for tongue\*
↓ SICI ↓SICI in FDI
intensity
EDC No change
RTMS is thought to induce homosynaptic plasticity within the target cortex. RTMS may increase or decrease cortical activity via alterations in the activity of glutamatergic NMDA receptors [63]. The frequency, intensity, and duration ofrTMS determine whetherthe plasticity effect is LTD or LTP-like [63, 64]. Typically, rTMS delivered over motor cortex at a frequen‐ cy of <1 Hz results in LTD-like effects while frequencies >5 Hz yield LTP-like effects [63]. In SCI, the primary focus ofrTMS is to modulate descending projections and strengthen the intact corticospinal connections.
In a study by Belci et al. [65], rTMS was applied overthe motor cortex forfive consecutive days using a protocol that included doublets of TMS delivered at a frequency of 10 Hz with each doublet being separated by 10 s for a total 720 stimuli. Following real rTMS, measures of CSP were reduced. Further, motor and sensory function, determined by the assessment of motor and sensory function by the American Spinal Injury Association (ASIA), was improved, as was performance in the timed peg-board task. Further, the sensory perceptual threshold to an electrical stimulus was reduced suggesting alterations within the somatosensory cortex. These measures remained improved for 2 weeks post-intervention while sham rTMS revealed no benefit [65]. Ellaway et al. [66] delivered real versus sham rTMS via 2 s trains of 5 Hz stimu‐ lation separated by 8 s for a total of 15 min over the motor cortex in individuals with SCI. The real and sham interventions were performed over 5 consecutive days. Measurements of the ASIA score, active research arm test (ARAT) which includes testing grasp, grip, pinch and gross movement, electrical stimulation perceptual threshold, MEPs, AMT, and CSP were assessed before and following intervention. Results revealed that AMT was increased and ARAT improved following rTMS compared to the sham intervention, without changes in other measures. Another study delivered real versus sham high-frequency (20 Hz) rTMS over 15 consecutive days to assess improvement in lower limb function [67]. Measurements includ‐ ed lower extremities motor score (LEMS), modified Ashworth scale (MAS), walking index for SCI, ten-meter walking test, step length, cadence assessment, and a timed up and go (TUG) test. At the cessation of the rTMS intervention, LEMS improved, MAS score was reduced, gait was altered (i.e., increased velocity, cadence, step length) and an improvement was ob‐ served in TUG without changes in the walking index. Effects returned to baseline 2 weeks following the last intervention [67]. Finally, an rTMS protocol consisting of 4 pulse trains (quadropulse) delivered at a frequency of 250–500 Hz with an inter-train interval of 5–6 s (250– 360/day) for either 1 or 5 consecutive days (i.e., ~1000–1440 pulses per day) was delivered to individuals with SCI [68]. Hand dexterity, MEPs, and spinal excitability were measured. After a single session of rTMS, there was no change in hand dexterity although MEPs increased and spinal excitability decreased. Further, after 5 consecutive days of rTMS there was a 10% increase in hand dexterity. These data indicate that rTMS delivered over consecutive days is more effective than single session at promoting motor improvements [68].
### **3.2. Transcranial direct current stimulation (TDCS)**
TDCS is a neuromodulating technique that utilizes homosynaptic plasticity, similar to rTMS. However, unlike rTMS, tDCS does not produce neuronal action potentials since the static field produced by tDCS does not produce the rapid depolarization of neurons [69]. TDCS modi‐ fies spontaneous neuronal excitability by either creating tonic depolarization or hyperpolari‐ zation [69]. The direction of current flow appears to determine whether the protocol results in LTD or LTP-like effects; anodal tDCS results in neuron depolarization (LTP) while cathodal tDCS results in neuron hyperpolarization (LTD) [69, 70]. There exists one study that has shown evidence for promoting plasticity in individuals with SCI using tDCS. Murray et al. [71] delivered anodal tDCS for 20 min once per week over 3 weeks at varying intensities; 1, 2 mA and sham. Measurements of MEP amplitude, sensory threshold, and muscle strength were performed before and after intervention. MEP amplitude was significantly increased after delivery of 2 mA anodal tDCS but not 1 mA or sham conditions. Sensory threshold was significantly reducedafter both 2 and 1 mAbut not sham stimulation. No protocol was effective at changing muscle strength [71]. It is evident that the extent of research in tDCS aimed at altering motor function in individuals with SCI is limited. However, early results show that tDCS may be an effective tool to alter corticospinal excitability and improve motor function.
### **3.3. Paired associative stimulation (PAS) and spinal associative stimulation (SAS)**
PAS is a plasticity inducing protocol that utilizes STDP to alter cortical or spinal function. Traditionally, PAS involves timing peripheral electrical nerve stimulation with a TMS pulse over the motor cortex. When these two stimuli are timed to arrive in the cortex at approxi‐ mately the same time, the protocol is known as PAS and when they are timed to coincide at the level of the spinal cord, the technique is known as spinal associative stimulation (SAS). PAS delivered at an ISI of 25 ms and targeting motor cortex has been effective at increasing MEP amplitude in uninjured individuals [72], while PAS delivered at an ISI of 20 ms has been effective at altering spinal excitability by increasing the amplitude of the spinal H-reflex [73]. There have been three studies performed in SCI where PAS or variations of PAS have been used to induce cortical and/or spinal plasticity promoting functional recovery.
Roy et al. [14] assessed corticospinal excitability in the lower limb before and following PAS delivered at an interval to promote near coincident activation within the motor cortex (i.e., the N20 latency of somatosensory-evoked potential +6 ms). The protocol included 120 pairs of peripheral afferent stimuli and TMS pulses at a frequency of 0.2 Hz [14]. MEPs were meas‐ ured during tonic 20% contraction (active MEPs) and at rest (resting MEPs). Results indicat‐ ed significant increases to resting but no change to active MEPs [14]. Bunday and Perez [74] tested the SAS protocol in individuals with SCI and delivered 100 pairs of stimuli at a frequency of 0.1 Hz and timed the ISI such that the TMS efferent volley and the PNS antidromic efferent
volley arrived at the corticospinal-motorneuronal synapse at the C7 spinal level nearly simultaneously. Measurements of MEPs, spinal F-waves, voluntary motor output, and manual dexterity were performed post-intervention. Results showed increases in MEPs, no change in F-waves, and increases to both maximum voluntary EMG/force and a reduction in the time to complete the 9-hole pegboard task [74]. Finally, Yamaguchi et al. [75] paired peripheral nerve stimulation with anodal tDCS. During a 20 min, 1 mA anodal tDCS protocol peripheral electrical stimulation trains consisting of 10 pulses delivered at a frequency of 100 Hz were delivered every 2 s. Measurements of reciprocal inhibition from the tibialis anterior and ankle movement were assessed after stimulation. Results indicate reduced reciprocal inhibition, and increased ankle movements following the combination of peripheral electrical stimulation and anodal tDCS [75].
### **3.4. Summary**
a single session of rTMS, there was no change in hand dexterity although MEPs increased and spinal excitability decreased. Further, after 5 consecutive days of rTMS there was a 10% increase in hand dexterity. These data indicate that rTMS delivered over consecutive days is
TDCS is a neuromodulating technique that utilizes homosynaptic plasticity, similar to rTMS. However, unlike rTMS, tDCS does not produce neuronal action potentials since the static field produced by tDCS does not produce the rapid depolarization of neurons [69]. TDCS modi‐ fies spontaneous neuronal excitability by either creating tonic depolarization or hyperpolari‐ zation [69]. The direction of current flow appears to determine whether the protocol results in LTD or LTP-like effects; anodal tDCS results in neuron depolarization (LTP) while cathodal tDCS results in neuron hyperpolarization (LTD) [69, 70]. There exists one study that has shown evidence for promoting plasticity in individuals with SCI using tDCS. Murray et al. [71] delivered anodal tDCS for 20 min once per week over 3 weeks at varying intensities; 1, 2 mA and sham. Measurements of MEP amplitude, sensory threshold, and muscle strength were performed before and after intervention. MEP amplitude was significantly increased after delivery of 2 mA anodal tDCS but not 1 mA or sham conditions. Sensory threshold was significantly reducedafter both 2 and 1 mAbut not sham stimulation. No protocol was effective at changing muscle strength [71]. It is evident that the extent of research in tDCS aimed at altering motor function in individuals with SCI is limited. However, early results show that tDCS may be an effective tool to alter corticospinal excitability and improve motor function.
**3.3. Paired associative stimulation (PAS) and spinal associative stimulation (SAS)**
used to induce cortical and/or spinal plasticity promoting functional recovery.
PAS is a plasticity inducing protocol that utilizes STDP to alter cortical or spinal function. Traditionally, PAS involves timing peripheral electrical nerve stimulation with a TMS pulse over the motor cortex. When these two stimuli are timed to arrive in the cortex at approxi‐ mately the same time, the protocol is known as PAS and when they are timed to coincide at the level of the spinal cord, the technique is known as spinal associative stimulation (SAS). PAS delivered at an ISI of 25 ms and targeting motor cortex has been effective at increasing MEP amplitude in uninjured individuals [72], while PAS delivered at an ISI of 20 ms has been effective at altering spinal excitability by increasing the amplitude of the spinal H-reflex [73]. There have been three studies performed in SCI where PAS or variations of PAS have been
Roy et al. [14] assessed corticospinal excitability in the lower limb before and following PAS delivered at an interval to promote near coincident activation within the motor cortex (i.e., the N20 latency of somatosensory-evoked potential +6 ms). The protocol included 120 pairs of peripheral afferent stimuli and TMS pulses at a frequency of 0.2 Hz [14]. MEPs were meas‐ ured during tonic 20% contraction (active MEPs) and at rest (resting MEPs). Results indicat‐ ed significant increases to resting but no change to active MEPs [14]. Bunday and Perez [74] tested the SAS protocol in individuals with SCI and delivered 100 pairs of stimuli at a frequency of 0.1 Hz and timed the ISI such that the TMS efferent volley and the PNS antidromic efferent
more effective than single session at promoting motor improvements [68].
**3.2. Transcranial direct current stimulation (TDCS)**
286 288Recovery of Motor Function Following Spinal Cord Injury
Please referto **Table 2** for a summary of the previous findings. Collectively these studies reveal promising indications that rTMS, TCDS, and PAS approaches can modulate cortical function leading to short-term improvements in the motor system in individuals with SCI. Although, these results are not always unanimous, (i.e., differential effects on MEPs and CSP), they may relate to the specific protocol parameters as existing studies in SCI have utilized variable stimulation parameters. Further research should determine the most effective protocol at yielding changes to neural physiology and improvements in motor function in individuals with SCI. NIBS combined with othertechniques might be a promising new avenue forresearch for the ultimate goal of creating long-term functional improvements.
**Table 2.** Summary of NIBS to promote plasticity in SCI.
**Intervention Source Parameters Classification Response**
extremities motor score - improved modified ashworth scale - improved gait mechanics
*Physiological*
*Physiological*
*Physiological* - increased motor-evoked potentials
Incomplete tetraplegic
Incomplete para /tetraplegic Incomplete tetraplegic Incomplete para /tetraplegic
ency of 20 Hz
288 290Recovery of Motor Function Following Spinal Cord Injury
in trains at a freq uency of 250–500 Hz - 5–6 s between trains
over 20 min
of 0.2 Hz
0.1 Hz
**tDCS** Murray et al. [71] - Anodal tDCS at 1 or 2 mA
**PAS/SAS** Roy et al. [37]
Bunday and Perez [74] Yamaguchi et al. [75]
### **4. Coupling NIBS with movement protocols**
A primary goal of motor Training is to improve functional ability by repeated exposure to a particular task, such as treadmill training to improve walking ability. In clinical populations, motor training can promote plastic changes in unaffected motor networks by increasing the efficacy of synaptic transmission [76]. Therefore, there is an opportunity to promote neural plasticity via motortraining in intact cortical andspinal motor circuitry in individuals with SCI.
### **4.1. Pairing NIBS with motor training**
Previous studies have shown that motor training can influence corticospinal excitability. In participants with SCI, locomotor resistance training using Lokomat facilitates spinal reflexes at 20 and 80 ms in the soleus muscle and improves gait quality as assessed by LEMS, walk‐ ing index for SCI and velocity [77], as well as MEP amplitudes in tibialis anterior [78]. Treadmill training in SCI participants for ~2 months (5 sessions per week for 1 h) increases MEP amplitudes in tibialis anterior, increases manual muscle strength in ankle dorsiflexors as measured by 11-point manual muscle strength score, and increases the duration of the CSP [79]. Although motor training alone promotes motor recovery, the functional outcomes are often limited and patients still exhibit substantial motor impairments.
Motor training and NIBS are each, independently effective at promoting plasticity in SCI participants. Therefore, the combination of the two may lead to plasticity effects that exceed their individual components. Few studies have tested the effects of pairing NIBS with motor training to facilitate functional motorrecovery. Gomes-Osman et al. [80] evaluated upperlimb function in SCI participants. They delivered 10 Hz rTMS (800 pulses at 80% RMT) with repetitive task practice involving 30 s of practice with the 9-hole pegboard task. The results revealed a decrease in the time to complete the Jebsen Taylor test following real and sham rTMS paired with repetitive task practice. However, the effects were larger following real versus sham rTMS paired with repetitive task practice. Measures of RMT and AMT were unchanged [80]. Alexeeva et al. [68] combined rTMS with motor training in SCI participants. Participants experienced 5 consecutive days of each intervention: rTMS, motor training, and rTMS + motor training. A washout period of at least 4 weeks elapsed between each interven‐
tion. RTMS consisted of 4-pulse trains with a ~0.2–1.5 Hz train delivery rate at an intensity set to 80–90% RMT (i.e., quadropulse rTMS). Motor training consisted of hand tasks in partici‐ pants 1 and 2 (10 tasks performed 10 times each: grasp and release, hand pronation and supination, isometric and concentric contractions of the wrist, thumb, and interphalangeal joints) and locomotor training in participant 3 (walking on a treadmill at a self-selected pace for 30 min with belt speed ≥0.05 m/s). In participant 1 and 2, "rTMS" and "motor training" improved 9-hole pegboard task performance and increased MEP amplitudes without changing SICI, ICF, or CSP. The combined intervention led to the greatest improvements in 9 hole pegboard task. In participant 3, the largest improvements in treadmill walking speeds were seen following the combined interventions as well [68]. Collectively, studies pairing NIBS with motortraining reveal that largerfunctional gains may be induced compared to the effects of NIBS or motor training delivered in isolation.
### **4.2. Pairing NIBS with aerobic exercise**
Using aerobic exercise to prime the brain prior to NIBS may also lead to larger changes in corticospinal excitability. Regular physical activity and aerobic exercise promote plasticity by increasing levels of growth factors including brain-derived neurotrophic factor (BDNF) and insulin-like growth factor (IGF-1) [81, 82]. Further, aerobic exercise can modify plasticity [83] via increases in cerebral blood flow [84] and angiogenesis [85]. Rojas Vega et al. [86] investi‐ gated the effects of aerobic exercise on BDNF and IGF-1 serum concentrations in SCI partici‐ pants. The SCI participants completed an aerobic exercise session on a hand-bike that included a 10-min warm-up followed by a timed trial over a distance of 42 km. Blood samples were collected before, after the warm-up, and immediately following aerobic exercise. The warmup resulted in a 1.5-fold increase in BDNF concentration, although no significant differences were seen between pre- and post-aerobic exercise measures. Additionally, IGF-1 concentra‐ tions were increased following both the warm-up and aerobic exercise [86] suggesting that aerobic exercise has the ability to prime the central nervous system for neuroplastic changes. This has been supported by recent evidence that aerobic exercise, such as cycling, is able to prime the brain priorto NIBS in healthy participants [87, 88]. In uninjured individuals, priming the cortex with aerobic exercise priorto PAS enhances the plasticity effectrelative to PAS alone as measured by increases in the MEP recruitment curve slope [87]. Additionally, recent evidence suggests that individuals who exercise regularly are more prone to motor cortex plasticity following PAS relative to sedentary/low physically active individuals. Those who exercise regularly had increased MEP amplitudes and steeperinput/outputrecruitment curves after intervention, while no significant changes were seen in sedentary/low physically active individuals [89].
### **4.3. Summary**
There are challenges for future research focused on combining NIBS with aerobic exercise or motor training. For example, one consideration involves timing the delivery of NIBS with respect to aerobic exercise and/or motor training. Thus far, NIBS has been delivered simulta‐ neously [80] or in advance of [68] motor training yet their combined effect may be influ‐
enced by their order of delivery [90]. In addition, the outcome of pairing protocols may be dependent on parameters of NIBS, aerobic exercise, or motor training, such as intensity and number of sessions; multiple sessions may be needed to induce any significant changes in motor function [91]. Therefore, pairing NIBS with aerobic exercise and/or motor training has the potential to drive neuroplastic changes in SCI participants that may exceed the function‐ al gains achieved by a singular intervention, but further investigation is required.
### **5. New opportunities for NIBS to promote motor recovery in SCI**
tion. RTMS consisted of 4-pulse trains with a ~0.2–1.5 Hz train delivery rate at an intensity set to 80–90% RMT (i.e., quadropulse rTMS). Motor training consisted of hand tasks in partici‐ pants 1 and 2 (10 tasks performed 10 times each: grasp and release, hand pronation and supination, isometric and concentric contractions of the wrist, thumb, and interphalangeal joints) and locomotor training in participant 3 (walking on a treadmill at a self-selected pace for 30 min with belt speed ≥0.05 m/s). In participant 1 and 2, "rTMS" and "motor training" improved 9-hole pegboard task performance and increased MEP amplitudes without changing SICI, ICF, or CSP. The combined intervention led to the greatest improvements in 9 hole pegboard task. In participant 3, the largest improvements in treadmill walking speeds were seen following the combined interventions as well [68]. Collectively, studies pairing NIBS with motortraining reveal that largerfunctional gains may be induced compared to the effects
Using aerobic exercise to prime the brain prior to NIBS may also lead to larger changes in corticospinal excitability. Regular physical activity and aerobic exercise promote plasticity by increasing levels of growth factors including brain-derived neurotrophic factor (BDNF) and insulin-like growth factor (IGF-1) [81, 82]. Further, aerobic exercise can modify plasticity [83] via increases in cerebral blood flow [84] and angiogenesis [85]. Rojas Vega et al. [86] investi‐ gated the effects of aerobic exercise on BDNF and IGF-1 serum concentrations in SCI partici‐ pants. The SCI participants completed an aerobic exercise session on a hand-bike that included a 10-min warm-up followed by a timed trial over a distance of 42 km. Blood samples were collected before, after the warm-up, and immediately following aerobic exercise. The warmup resulted in a 1.5-fold increase in BDNF concentration, although no significant differences were seen between pre- and post-aerobic exercise measures. Additionally, IGF-1 concentra‐ tions were increased following both the warm-up and aerobic exercise [86] suggesting that aerobic exercise has the ability to prime the central nervous system for neuroplastic changes. This has been supported by recent evidence that aerobic exercise, such as cycling, is able to prime the brain priorto NIBS in healthy participants [87, 88]. In uninjured individuals, priming the cortex with aerobic exercise priorto PAS enhances the plasticity effectrelative to PAS alone as measured by increases in the MEP recruitment curve slope [87]. Additionally, recent evidence suggests that individuals who exercise regularly are more prone to motor cortex plasticity following PAS relative to sedentary/low physically active individuals. Those who exercise regularly had increased MEP amplitudes and steeperinput/outputrecruitment curves after intervention, while no significant changes were seen in sedentary/low physically active
There are challenges for future research focused on combining NIBS with aerobic exercise or motor training. For example, one consideration involves timing the delivery of NIBS with respect to aerobic exercise and/or motor training. Thus far, NIBS has been delivered simulta‐ neously [80] or in advance of [68] motor training yet their combined effect may be influ‐
of NIBS or motor training delivered in isolation.
**4.2. Pairing NIBS with aerobic exercise**
290 292Recovery of Motor Function Following Spinal Cord Injury
individuals [89].
**4.3. Summary**
Motor recovery in SCI participants via NIBS and paired protocols are promising. However, other forms of NIBS including theta burst stimulation (TBS), rapid rate paired associative stimulation (rPAS), or transspinal direct current stimulation (ts-CCS) have the potential to induce plasticity and promote motor recovery and have yet to be explored in SCI.
TBS is a form of rTMS at low intensity that delivers continuous (cTBS) or intermittent (iTBS) high-frequency pulses inducing homosynaptic plasticity in the stimulated area. TBS effects depend on corticospinal output depend on the nature of stimulation. iTBS over motor cortex increases the amplitude of MEPs [92, 93], while cTBS over motor cortex decreases the amplitude of MEPs [92, 94], although this pattern is not always observed [95–97]. In partici‐ pants with stroke, iTBS improves hand function [98] and increases MEP amplitudes [98, 99]. In addition, applying multiple sessions of cTBS in participants with amyotrophic lateral sclerosis decrease MEP amplitudes and increase RMT [100]. Although it has yet to be tested, TBS over motor cortex representation of the affected muscle in SCI may have the potential to modulate intracortical (i.e., SICI, ICF) and corticospinal circuitry (i.e., MEP amplitude). Hence, TBS, like rTMS, may be a suitable tool to influence synaptic interactions by strengthening the residual connections [92] and therefore increase motor output from the affected muscle. Further, by modulating plasticity within the cortex, indirect changes in the spinal circuitry may occur. Hence, TBS may provide an alternate method to induce plasticity in the cortex that may lead to motor recovery in SCI participants.
RPAS is based on the principles of STDP and involves pairing 5 Hz rTMS with peripheral nerve stimulation at a specific interstimulus interval [101]. Unlike PAS, which requires ~30 min to deliver, rPAS provides a particularly fast method (i.e., ~3–4 min for 600 pulses) to induce increases in corticospinal excitability. RPAS over the motor cortex increases MEP ampli‐ tudes and reduces SAI [101–104] in uninjured individuals. However, rPAS has yet to be investigated in clinical populations presenting with motor impairments. In rPAS, the pairing of TMS with nerve stimuli activates both, afferent and efferent pathways. Recently, it has been speculated that reorganization in the cortex following afferent stimulation may be crucial in neurorehabilitation of the hand [105]. Since rPAS is highly efficient in increasing corticospi‐ nal excitability in healthy adults for prolonged periods of time, it may provide a useful tool to promote sensory-motor coupling [102] in SCI participants.
Another promising NIBS technique involves the delivery of 40 min of constant current stimulation to the spinal cord [106]. Long-lasting transspinal constant current stimulation (ts-
CCS) alters cortical, corticospinal, and spinal plasticity in uninjured participants. Knikou et al. [106] found that both cathodal tsCCS and anodal tsCCS decreased afferent-mediated MEP facilitation, increased MEP amplitudes, and decreased transspinal-evoked potentials (TEPs) of knee flexors. Further, cathodal tsCCS increased TMS-mediated tibialis anterior flexor reflex facilitation, while anodal tsCCS decreased TMS-mediated TA flexor reflex facilitation and decreased post-activation depression of TEPs for the soleus H-reflex. This technique pro‐ vides a way to directly stimulate the spinal cord, changing the synaptic efficacy between descending motor axons and spinal motorneurons, cortical interneurons and descending motor axons and Ia afferents and motorneurons [106]. While more research is required to determine the efficacy of this tool at modulating changes in the cortex and spinal cord to promote motor recovery in clinical populations, it may have the potential to improve volun‐ tary motor function in SCI participants.
### **Author details**
Aaron Z. Bailey, Hunter J. Fassett, Tea Lulic, Jenin El Sayes and Aimee J. Nelson\*
\*Address all correspondence to: nelsonaj@mcmaster.ca
Department of Kinesiology, McMaster University, Hamilton, Ontario, Canada
### **References**
[6] Neva JL, Lakhani B, Brown KE, Wadden KP, Mang CS, Ledwell NHM, et al. Multiple measures of corticospinal excitability are associated with clinical features of multiple sclerosis. Behav Brain Res 2016;297:187–95. doi:10.1016/j.bbr.2015.10.015
CCS) alters cortical, corticospinal, and spinal plasticity in uninjured participants. Knikou et al. [106] found that both cathodal tsCCS and anodal tsCCS decreased afferent-mediated MEP facilitation, increased MEP amplitudes, and decreased transspinal-evoked potentials (TEPs) of knee flexors. Further, cathodal tsCCS increased TMS-mediated tibialis anterior flexor reflex facilitation, while anodal tsCCS decreased TMS-mediated TA flexor reflex facilitation and decreased post-activation depression of TEPs for the soleus H-reflex. This technique pro‐ vides a way to directly stimulate the spinal cord, changing the synaptic efficacy between descending motor axons and spinal motorneurons, cortical interneurons and descending motor axons and Ia afferents and motorneurons [106]. While more research is required to determine the efficacy of this tool at modulating changes in the cortex and spinal cord to promote motor recovery in clinical populations, it may have the potential to improve volun‐
Aaron Z. Bailey, Hunter J. Fassett, Tea Lulic, Jenin El Sayes and Aimee J. Nelson\*
Department of Kinesiology, McMaster University, Hamilton, Ontario, Canada
Neurosci 2012;32:806–16. doi:10.1523/JNEUROSCI.4299-12.2012
excitability after stroke. Brain 2003;126:470–81.
[1] Siebner HR, Rothwell J. Transcranial magnetic stimulation: new insights into repre‐ sentational cortical plasticity. Exp Brain Res 2003;148:1–16. doi:10.1007/
[2] Duque J, Labruna L, Verset S, Olivier E, Ivry RB. Dissociating the role of prefrontal and premotor cortices in controlling inhibitory mechanisms during motor preparation. J
[3] Sinclair C, Hammond GR. Excitatory and inhibitory processes in primary motor cortex during the foreperiod of a warned reaction time task are unrelated to response
[4] Ziemann U. TMS and drugs. Clin Neurophysiol 2004;115:1717–29. doi:10.1016/j.clinph.
[5] Bütefisch CM, Netz J, Wessling M, Seitz RJ, Hömberg V. Remote changes in cortical
expectancy. Exp Brain Res 2009;194:103–13. doi:10.1007/s00221-008-1684-2
tary motor function in SCI participants.
292 294Recovery of Motor Function Following Spinal Cord Injury
\*Address all correspondence to: nelsonaj@mcmaster.ca
**Author details**
**References**
s00221-002-1234-2
2004.03.006.
[31] DeFelipe J, ConleyM, Jones EG. Long-range focal collateralization of axons arising from corticocortical cells in monkey sensory-motor cortex. J Neurosci 1986;6:3749–66.
[18] Paulus W, Classen J, Cohen LG, Large CH, Di Lazzaro V, Nitsche M, et al. State of the art: pharmacologic effects on cortical excitability measures tested by transcranial
magnetic stimulation. Brain Stimul 2008;1:151–63. doi:10.1016/j.brs.2008.06.002
doi:10.1111/(ISSN)1469-7793
294 296Recovery of Motor Function Following Spinal Cord Injury
(ISSN)1469-7793
brain/60.4.389
2010.11.002
1990;510:130–4
1995;7:245–50.
2002;58:794–801.
[19] Hodgkin AL, Huxley AF. A quantitative description of membrane current and its application to conduction and excitation in nerve. J Physiol (Lond) 1952;117:500–44.
[20] Day BL, Dressler D, Maertens de Noordhout A, Marsden CD, Nakashima K, Rothwell JC, et al. Electric and magnetic stimulation of human motor cortex: surface EMG and single motor unit responses. J Physiol (Lond) 1989;412:449–73. doi:10.1111/
[21] Freund P, Thompson AJ, Rothwell J, Craggs M, Bestmann S. Corticomotor represen‐ tation to a human forearm muscle changes following cervical spinal cord injury. Eur J
[22] Bailey AZ, Mi YP, Nelson AJ. Short-latency afferent inhibition in chronic spinal cord
[23] Penfield W, Boldrey E. Somatic motor and sensory representation in the cerebral cortex of man as studied by electrical stimulation. Brain J Neurol 1937;60:389–443. doi:10.1093/
[24] Raineteau O, Schwab ME. Plasticity of motor systems after incomplete spinal cord
[25] Romero JR, Ramirez DM, Aglio LS, Gugino LD. Brain mapping using transcranial magnetic stimulation. Neurosurg Clin N Am 2011;22:141–52–vii. doi:10.1016/j.nec.
[26] van de Ruit M, Grey MJ. The TMS Map Scales with Increased Stimulation Intensity and Muscle Activation. Brain Topogr 2016;29:56–66. doi:10.1007/s10548-015-0447-1
[27] Levy WJ, Amassian VE, Traad M, Cadwell J. Focal magnetic coil stimulation reveals motor cortical system reorganized in humans after traumatic quadriplegia. Brain Res
[28] Cohen LG, Roth BJ, Wassermann EM, Topka H, Fuhr P, Schultz J, et al. Magnetic stimulation of the human cerebral cortex, an indicator of reorganization in motor pathways in certain pathological conditions. J Clin Neurophysiol 1991;8:56–65.
[29] Streletz LJ, Belevich JK, Jones SM, Bhushan A, Shah SH, Herbison GJ. Transcranial magnetic stimulation: cortical motor maps in acute spinal cord injury. Brain Topogr
[30] Mikulis DJ, Jurkiewicz MT, McIlroy WE, Staines WR, Rickards L, Kalsi-Ryan S, et al. Adaptation in the motor cortex following cervical spinal cord injury. Neurology
Neurosci 2011;34:1839–46. doi:10.1111/j.1460-9568.2011.07895.x
injury. Transl Neurosci 2015;6:1–9. doi:10.1515/tnsci-2015-0025
injury. Nat Rev Neurosci 2001;2:263–73. doi:10.1038/35067570
[56] Inghilleri M, Berardelli A, Marchetti P, Manfredi M. Effects of diazepam, baclofen and thiopental on the silent period evoked by transcranial magnetic stimulation in humans. Exp Brain Res 1996;109:467–72.
[44] Mohammadi B, Krampfl K, Petri S, Bogdanova D, Kossev A, Bufler J, et al. Selective and nonselective benzodiazepine agonists have different effects on motor cortex
[45] Schwenkreis P, Witscher K, Janssen F, Addo A, Dertwinkel R, Zenz M, et al. Influ‐ ence of the N-methyl-D-aspartate antagonist memantine on human motor cortex
[46] Chen R, Corwell B, Hallett M. Modulation of motor cortex excitability by median nerve
[47] Hirashima F, Yokota T. Influence of peripheral nerve stimulation on human motor cortical excitability in patients with ventrolateral thalamic lesion. Arch Neurol
[48] Di Lazzaro V, Pilato F, Dileone M, Tonali PA, Ziemann U. Dissociated effects of diazepam and lorazepam on short-latency afferent inhibition. J Physiol (Lond)
[49] Di Lazzaro V, Oliviero A, Profice P, Pennisi MA, Di Giovanni S, Zito G, et al. Muscarinic receptor blockade has differential effects on the excitability of intracortical circuits in
[50] Sailer A, Molnar GF, Cunic DI, Chen R. Effects of peripheral sensory input on cortical
[51] Tokimura H, Di Lazzaro V, Tokimura Y, Oliviero A, Profice P, Insola A, Mazzone P, Tonali P, Rothwell JC. Short latency inhibition of human hand motor cortex by
[52] Kimiskidis VK, Papagiannopoulos S, Sotirakoglou K, Kazis DA, Kazis A, Mills KR. Silent period to transcranial magnetic stimulation: construction and properties of stimulus-response curves in healthy volunteers. Exp Brain Res 2005;163:21–31. doi:
[53] Fuhr P, Agostino R, Hallett M. Spinal motor neuron excitability during the silent period after cortical stimulation. Electroencephalogr Clin Neurophysiol 1991;81:257–62. [54] Ziemann U, Netz J, Szelényi A, Hömberg V. Spinal and supraspinal mechanisms contribute to the silent period in the contracting soleus muscle after transcranial
magnetic stimulation of human motor cortex. Neurosci Lett 1993;156:167–71.
[55] Pierantozzi M, Marciani MG, Palmieri MG, Brusa L, Galati S, Caramia MD, et al. Effect of Vigabatrin on motorresponses to transcranial magnetic stimulation: an effective tool to investigate in vivo GABAergic cortical inhibition in humans. Brain Res 2004;1028:1–
excitability. Muscle Nerve 2006;33:778–84. doi:10.1002/mus.20531
excitability. Neurosci Lett 1999;270:137–40.
1997;54:619–24.
296 298Recovery of Motor Function Following Spinal Cord Injury
10.1007/s00221-004-2134-4
8. doi:10.1016/j.brainres.2004.06.009
and digit stimulation. Exp Brain Res 1999;129:77–86.
2005;569:315–23. doi:10.1113/jphysiol.2005.092155
the human motor cortex. Exp Brain Res 2000;135:455–61.
inhibition in humans. J Physiol (Lond) 2002;544:617–29.
somatosensory input from the hand. J Physiol 2000;523:503–13.
[79] Thomas SL, Gorassini MA. Increases in corticospinal tract function by treadmill training after incomplete spinal cord injury. J Neurophysiol 2005;94:2844–55. doi: 10.1152/jn.00532.2005
[68] Alexeeva N, Calancie B. Efficacy of QuadroPulse rTMS for improving motor function after spinal cord injury: three case studies. J Spinal Cord Med 2016;39:50–7. doi:
[69] Nitsche MA, Cohen LG, Wassermann EM, Priori A, Lang N, Antal A, et al. Transcra‐ nial direct current stimulation: state of the art 2008. Brain Stimul 2008;1:206–23. doi:
[70] Kuo H-I, Paulus W, Batsikadze G, Jamil A, Kuo M-F, Nitsche MA. Chronic enhance‐ ment of serotonin facilitates excitatory transcranial direct current stimulation-in‐ duced neuroplasticity. Neuropsychopharmacology 2015. doi:10.1038/npp.2015.270
[71] Murray LM, Edwards DJ, Ruffini G, Labar D, Stampas A, Pascual-leone A, et al. Intensity dependent effects of transcranial direct current stimulation on corticospinal excitability in chronic spinal cord injury. Arch Phys Med Rehabil 2015;96:S114–21.
[72] Player MJ, TaylorJL, Alonzo A, Loo CK. Paired associative stimulation increases motor cortex excitability more effectively than theta-burst stimulation. Clin Neurophysiol
[73] Lamy J-C, Russmann H, Shamim EA, Meunier S, Hallett M. Paired associative stimulation induces change in presynaptic inhibition of Ia terminals in wrist flexors in
[74] Bunday KL, Perez MA. Motor recovery after spinal cord injury enhanced by strength‐ ening corticospinal synaptic transmission. Curr Biol 2012;22:2355–61. doi:10.1016/j.cub.
[75] Yamaguchi T, Fujiwara T, TsaiY-A, Tang S-C, Kawakami M, Mizuno K, et al. The effects of anodal transcranial direct current stimulation and patterned electrical stimulation on spinal inhibitory interneurons and motor function in patients with spinal cord
[76] Rioult-Pedotti MS, Friedman D, Hess G, Donoghue JP. Strengthening of horizontal cortical connections following skill learning. Nat Neurosci 1998;1:230–4. doi:
[77] Benito Penalva J, Opisso E, Medina J, Corrons M, Kumru H, Vidal J, et al. H reflex modulation by transcranial magnetic stimulation in spinal cord injury subjects after gait training with electromechanical systems. Spinal Cord 2010;48:400–6. doi:
[78] Chisholm AE, Peters S, Borich MR, Boyd LA, Lam T. Short-term cortical plasticity associated with feedback-error learning after locomotor training in a patient with incomplete spinal cord injury. Phys Ther 2015;95:257–66. doi:10.2522/ptj.20130522
humans. J Neurophysiol 2010;104:755–64. doi:10.1152/jn.00761.2009
injury. Exp Brain Res 2016:1–10. doi:10.1007/s00221-016-4561-4
10.1179/2045772314Y.0000000279
10.1016/j.brs.2008.06.004
298 300Recovery of Motor Function Following Spinal Cord Injury
doi:10.1016/j.apmr.2014.11.004
2012.10.046
10.1038/678
10.1038/sc.2009.151
2012;123:2220–6. doi:10.1016/j.clinph.2012.03.081
[103] Tsang P, Bailey AZ, Nelson AJ. Rapid-rate paired associative stimulation over the primary somatosensory cortex. Plos One 2015;10:e0120731. doi:10.1371/journal.pone. 0120731
[91] Froc DJ, Chapman CA, Trepel C, Racine RJ. Long-term depression and depotentia‐ tion in the sensorimotor cortex of the freely moving rat. J Neurosci 2000;20:438–45.
[92] Huang Y-Z, Edwards MJ, Rounis E, Bhatia KP, Rothwell JC. Theta burst stimulation of the human motor cortex. Neuron 2005;45:201–6. doi:10.1016/j.neuron.2004.12.033
[93] Zafar N, Paulus W, Sommer M. Comparative assessment of best conventional with best theta burst repetitive transcranial magnetic stimulation protocols on human motor cortex excitability. Clin Neurophysiol 2008;119:1393–9. doi:10.1016/j.clinph.2008.02.006
[94] Wu SW, Shahana N, Huddleston DA, Gilbert DL. Effects of 30Hz θ burst transcranial magnetic stimulation on the primary motor cortex. J Neurosci Methods 2012;208:161–
[95] Gentner R, Wankerl K, Reinsberger C, Zeller D, Classen J. Depression of human corticospinal excitability induced by magnetic theta-burst stimulation: evidence of rapid polarity-reversing metaplasticity. Cereb Cortex 2008;18:2046–53. doi:10.1093/
[96] Hamada M, Murase N, Hasan A, Balaratnam M, Rothwell JC. The role of interneuron networks in driving human motor cortical plasticity. Cereb Cortex 2013;23:1593–605.
[97] López-Alonso V, Cheeran B, Río-Rodríguez D, Fernández-Del-Olmo M. Inter-individ‐ ual variability in response to non-invasive brain stimulation paradigms. Brain Stimul
[98] Talelli P, Greenwood RJ, Rothwell JC. Exploring Theta Burst Stimulation as an intervention to improve motor recovery in chronic stroke. Clin Neurophysiol
[99] Di Lazzaro V, Pilato F, Dileone M, Profice P, Capone F, Ranieri F, et al. Modulating cortical excitability in acute stroke: a repetitive TMS study. Clin Neurophysiol
[100] Munneke MAM, Rongen JJ, Overeem S, Schelhaas HJ, Zwarts MJ, Stegeman DF. Cumulative effect of 5 daily sessions of θ burst stimulation on corticospinal excitabil‐ ity in amyotrophic lateral sclerosis. Muscle Nerve 2013;48:733–8. doi:10.1002/mus.
[101] Quartarone A, Rizzo V, Bagnato S, Morgante F, Sant'Angelo A, Girlanda P, et al. Rapidrate paired associative stimulation of the median nerve and motor cortex can pro‐ duce long-lasting changes in motor cortical excitability in humans. J Physiol (Lond)
[102] Naro A, Russo M, AbdelKader M, Manganotti P, Genovesi V, Marino M, et al. A local signature of LTP-like plasticity induced by repetitive paired associative stimulation.
4. doi:10.1016/j.jneumeth.2012.05.014
2014;7:372–80. doi:10.1016/j.brs.2014.02.004
2007;118:333–42. doi:10.1016/j.clinph.2006.10.014
2008;119:715–23. doi:10.1016/j.clinph.2007.11.049
2006;575:657–70. doi:10.1113/jphysiol.2006.114025
Brain Topogr 2015;28:238–49. doi:10.1007/s10548-014-0396-0
cercor/bhm239
23818
doi:10.1093/cercor/bhs147
300 302Recovery of Motor Function Following Spinal Cord Injury
## **Emerging Techniques for Assessment of Sensorimotor Impairments after Spinal Cord Injury**
Filipe Barroso, Diego Torricelli and Juan C. Moreno
Additional information is available at the end of the chapter
### **Abstract**
http://dx.doi.org/10.5772/64182
Gait function can be altered after incomplete spinal cord (iSCI) lesions. Muscular weakness, co‐activation of antagonist muscles, and altered muscle mechanics are likely to provoke abnormal gait and postural movements. Functional scales are available for assessment of functional walking in SCI patients, such as walking index for spinal cord injury (WISCI II), timed up and go (TUG) test, 10‐meter walk test (10MWT), and 6‐ minute walk test (6MWT). Novel metrics for a more detailed comprehension of neuromuscular control in terms of degree of voluntary motor control have been recently proposed. This section describes novel techniques based on muscle synergy and frequency domain analysis of electromyographic signals. Such techniques are illustrated as potential tools for assessment of motor function after SCI with experi‐ mental data and a case study describing a diagnostic scenario. This chapter presents a discussion of the current status of the emerging metrics for assessment of sensorimo‐ tor impairments. Conclusions are given with respect to the availability of enriched information about neuromuscular behavior between functional tasks (walking and pedalling) and the potential relevance of these new techniques to improve the efficacy of treatment to improve locomotion after iSCI.
**Keywords:** Rehabilitation, spinal cord injury, walking, functional scales
### **1. Introduction**
Gait function can be altered after incomplete spinal cord (iSCI) lesions. Muscular weakness, co‐ activation of antagonist muscles, and altered muscle mechanics are likely to provoke abnor‐ malgaitandposturalmovements.Humanwalkinginvolves the coordinationof severalmuscles and its correct activation. One of the main goals of treatment after SCI is to recover the ability
© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
to walk again. The assessment of the neurorehabilitation process has traditionally been done based on the qualitative methods (classic clinical scales) or subjective assessment from physio‐ therapists (based on clinical gaze) [1]. Traditional techniques are prone to low reliability and, as a consequence, may result in inadequate or costly interventions. More importantly, clinical tests focused on behavioral outcomes provide little information about the underlying differen‐ ces between healthy and impaired nervous system [1]. It is becoming more and more clear that a more profound understanding of impairments may be crucial not only to prescribe effective treatments to individual patients but also to gather comparative results and evidence that are needed to develop novel therapies. Thus, motor neurorehabilitation should be informed by more reliable and repeatable metrics that allow a quantitative assessment of motor control performance and recovery.
Gait analysis is broadly known as means to adequately assess and follow‐up patients and supports a clinical decision on the best treatment [2]. Clinical gait analysis involves a variety of techniques including kinematic or joint motion measurements; kinetic or joint torque assessment, electromyographic (EMG) measurements, and video analysis. Measures derived from gait analysis provide a detailed and quantitative description. This might further be used to extract important information to select a task‐oriented approach that might enhance therapeutic response, which cannot be provided by clinical evaluation alone [2]. Instrument‐ ed clinical gait analysis, despite its objectivity, is not straightforward in practice. Thus, more concise indexes of gait function are to be developed to assess the changes in gait function over time and evaluate interventions. Gait analysis is broadly known as means to adequately assess and follow‐up patients andclinical of techniques including kinematic or joint motion kinetic or joint torqueprovide to extract important information to select a task‐oriented approach that might enhancetherapeutic response, which cannot be provided by clinical evaluation alone [2]. Instrumented clinical gait analysis, despite its objectivity, is not straightforward in practice. Thus, concise indexes of gait function are to be developed to assess the changes in gait function Most widely used functional scales for gait rehabilitation in SCI patients are walking index
Most widely used functional scales for gait rehabilitation in SCI patients are walking index for spinal cord injury (WISCI II), timed up and go (TUG) test, 10‐meter walk test (10MWT), and 6‐minute walk test (6MWT). As mentioned above, to achieve adequate treatment, it is crucial to investigate not only the functional effect but also the mechanisms underlying the im‐ paired function. Recently, the use of quantitative metrics based on electromyography and biomechanical features is bringing a new insight into the motor recovery mechanisms and performance outcomes after neural damage. EMG features provide useful information concerning brain motor control strategies [3]. Muscle activation patterns and muscle syner‐ gies have been proposed as a potential technique to measure motor recovery following therapeutic interventions [4]. Muscle synergies, understood as groups of co‐activated muscles that are responsible for task execution in different conditions, can explain the way that central nervous system (CNS) solves control of multiple muscles and degrees of freedom by means of a smaller number of neural parameters [5]. This brings a more comprehensive understand‐ ing of the underlying motor strategies responsible for impaired locomotion. to achieve to investigate not only the functional effect but also the mechanisms underlying the impaired function. Recently, the use of quantitative metrics based on electromyography and biome‐ chanical features is bringing a new insight into the motor recovery mechanisms and perform‐ ance outcomes after neural damage. EMG features provide useful information concerningbrain motor control strategies [3]. Muscle activation patterns and muscle synergies have been proposed as a potential technique to measure motor recovery following therapeutic interven‐ tions [4]. Muscle synergies, understood as groups of co‐activated muscles that are responsible for task execution in different conditions, can explain the way that central nervous system (CNS) solves control of multiple muscles and degrees of freedom by means of a smaller numberof neural parameters [5]. This brings a more comprehensive understanding of the underlying motor strategies responsible for impaired locomotion. of chapter to present emerging data to support therapeutic interventions in SCI patients who are commonly affected spastic paresis and require targeted relearning and activation of a residual motor Novel metrics based on computational such as muscle coherence and muscle for in terms of degree of voluntary motor Conclusions are given with respect to of information 3042in Spinal Cord Injury
The aim of this chapter is to present the emerging indexes based on EMG and biomechanical data to support therapeutic interventions in SCI patients who are commonly affected with spastic paresis and require targeted relearning and activation of a residual motor function. Novel metrics based on computational methods, such as muscle coherence and muscle synergy analysis, are presented as tools for a more detailed comprehension of neuromuscular control in terms of degree of voluntary motor control. Conclusions are given with respect to the availability of enriched information about neuromuscular behavior between functional tasks
(walking and pedalling) and the potential relevance of these new techniques to improve the efficacy of treatment to improve locomotion after iSCI.
### **2. Sensorimotor impairments after SCI**
to walk again. The assessment of the neurorehabilitation process has traditionally been done based on the qualitative methods (classic clinical scales) or subjective assessment from physio‐ therapists (based on clinical gaze) [1]. Traditional techniques are prone to low reliability and, as a consequence, may result in inadequate or costly interventions. More importantly, clinical tests focused on behavioral outcomes provide little information about the underlying differen‐ ces between healthy and impaired nervous system [1]. It is becoming more and more clear that a more profound understanding of impairments may be crucial not only to prescribe effective treatments to individual patients but also to gather comparative results and evidence that are needed to develop novel therapies. Thus, motor neurorehabilitation should be informed by more reliable and repeatable metrics that allow a quantitative assessment of motor control
Gait analysis is broadly known as means to adequately assess and follow‐up patients and supports a clinical decision on the best treatment [2]. Clinical gait analysis involves a variety of techniques including kinematic or joint motion measurements; kinetic or joint torque assessment, electromyographic (EMG) measurements, and video analysis. Measures derived from gait analysis provide a detailed and quantitative description. This might further be used to extract important information to select a task‐oriented approach that might enhance therapeutic response, which cannot be provided by clinical evaluation alone [2]. Instrument‐ ed clinical gait analysis, despite its objectivity, is not straightforward in practice. Thus, more concise indexes of gait function are to be developed to assess the changes in gait function over
to extract important information to select a task‐oriented approach that might enhancetherapeutic response, which cannot be provided by clinical evaluation alone [2]. Instrumented
of techniques including kinematic or joint motion kinetic or joint description. This
Gait analysis is broadly known as means to adequately assess and follow‐up patients and
Most widely used functional scales for gait rehabilitation in SCI patients are walking index for spinal cord injury (WISCI II), timed up and go (TUG) test, 10‐meter walk test (10MWT), and 6‐minute walk test (6MWT). As mentioned above, to achieve adequate treatment, it is crucial to investigate not only the functional effect but also the mechanisms underlying the im‐ paired function. Recently, the use of quantitative metrics based on electromyography and biomechanical features is bringing a new insight into the motor recovery mechanisms and performance outcomes after neural damage. EMG features provide useful information concerning brain motor control strategies [3]. Muscle activation patterns and muscle syner‐ gies have been proposed as a potential technique to measure motor recovery following therapeutic interventions [4]. Muscle synergies, understood as groups of co‐activated muscles that are responsible for task execution in different conditions, can explain the way that central nervous system (CNS) solves control of multiple muscles and degrees of freedom by means of a smaller number of neural parameters [5]. This brings a more comprehensive understand‐
to investigate not only the functional effect but also the mechanisms underlying the impaired function. Recently, the use of quantitative metrics based on electromyography and biome‐ chanical features is bringing a new insight into the motor recovery mechanisms and perform‐ ance outcomes after neural damage. EMG features provide useful information concerningbrain motor control strategies [3]. Muscle activation patterns and muscle synergies have been proposed as a potential technique to measure motor recovery following therapeutic interven‐ tions [4]. Muscle synergies, understood as groups of co‐activated muscles that are responsiblefor task execution in different conditions, can explain the way that central nervous system (CNS) solves control of multiple muscles and degrees of freedom by means of a smaller number of neural parameters [5]. This brings a more comprehensive understanding of the underlying
(WISCI (10MWT), it is
clinical gait analysis, despite its objectivity, is not straightforward in practice. Thus, concise indexes of gait function are to be developed to assess the changes in gait function Most widely used functional scales for gait rehabilitation in SCI patients are walking index
in terms of degree of voluntary motor Conclusions are given with respect to of enriched 3042in Spinal Cord Injury
The aim of this chapter is to present the emerging indexes based on EMG and biomechanical data to support therapeutic interventions in SCI patients who are commonly affected with spastic paresis and require targeted relearning and activation of a residual motor function. Novel metrics based on computational methods, such as muscle coherence and muscle synergy analysis, are presented as tools for a more detailed comprehension of neuromuscular control in terms of degree of voluntary motor control. Conclusions are given with respect to the availability of enriched information about neuromuscular behavior between functional tasks
of is EMG data to support therapeutic interventions in SCI patients who are commonly affected spastic paresis and require targeted relearning and activation of a residual motor Novel metrics based on computational methods, such as muscle coherence and muscle
ing of the underlying motor strategies responsible for impaired locomotion.
motor strategies responsible for impaired locomotion.
performance and recovery.
306 Recovery of Motor Function Following Spinal Cord Injury
time and evaluate interventions.
The spinal lesion leads to sensorimotor impairments in both upper and lower extremities. Ambulation results limited due to the sensory and proprioceptive impairments and muscle spasticity, commonly found even in subjects that reach a sufficient level of ambulation. The mechanisms involved in muscle spasticity are related to the lesion in the CNS that leads to changes in the excitability of spinal reflexes and loss of supraspinal drive. This results in abnormal muscle function and leads to altered mechanical muscle properties. In addition to this, proprioceptive and sensory impairment lead to altered or loss afferent feedback to the CNS, which in turn progressively affects motor control and leads to unstable, non‐physiolog‐ ical gait. In particular, symptoms such as muscular weakness, co‐activation of antagonist muscles, and altered muscle mechanics provoke abnormal gait and postural movements. (walking and and the potential relevance of these new techniques to improve iSCI.The spinal lesion leads to sensorimotor impairments in both upper and lower Ambulation results limited due to the sensory and proprioceptive impairments and musclecommonly found even in subjects that reach a sufficient level of mechanisms involved in muscle spasticity are related to the lesion in the CNS that leads changes in the excitability of spinal reflexes and loss of supraspinal This results abnormal muscle function and leads to altered mechanical muscle In addition proprioceptive and sensory impairment lead to altered or loss afferent feedback to ical gait. In particular, symptoms such as muscular weakness, co‐activation of antagonist
The relationship between hypertonia and gait function is still not clear controversial [6]. The difficulty to classify iSCI subjects as spastic or not is a well‐known problem [7]. Spasticity in SCI patients is mostly associated (clinically) with the presence of flexor and extensor spasms triggered by cutaneous stimulation. In practice, it is important to determine whether an impaired gait is mainly caused by disabling paresis but also altered afferent feedback to the CNS. Novel metrics to assess motor control based on detailed EMG analysis are required to complement and optimize interventions that focus on the clinical signs of spasticity, such as exaggerated reflexes and muscle tone (e.g., medication). mechanics provoke relationship classify well‐known [7]. Spasticity inpatients triggered by cutaneous In it is important to determine whether impaired gait is mainly caused by disabling paresis but also altered afferent feedback to Novel metrics to assess motor control based on detailed EMG analysis are required The main goal of clinical assessments is to quantify the motor recovery by observing measuring the functional changes that occur in the patient after the This is done by scales, which quantify the patient's residual functions under a wide therefore support scientifically the clinical practice in making effective choice on the treatment,
### **3. Clinical assessment of walking ability after SCI**
The main goal of clinical assessments is to quantify the motor recovery by observing and measuring the functional changes that occur in the patient after the injury. This is normally done by scales, which quantify the patient's residual functions under a wide spectrum, and therefore support scientifically the clinical practice in making effective choice on the treat‐ ment, studying cost‐benefit ofthe rehabilitation process and quantifying objectively the degree of incapacity or handicap. Nevertheless, the sensitivity and reliability of current scales are still limited. For instance, the score assignment still relies on a strong subjective component, which causes high intra and inter‐rater variability, even when the scale is carried out by experi‐ enced examiners [8]. In addition, scales often have similar content and purpose, which makes difficult to decide which of the available tests are superior and should be used as outcome measures. Therefore, there is a need to incorporate new and more objective methods to assess of motorrecovery into the existing panorama. In the following, we present and briefly describe the main measures used in clinical settings to evaluate the functional recovery after SCI. studying cost‐benefit of the rehabilitation process and quantifying objectively the degree of incapacity or handicap. Nevertheless, the sensitivity and reliability of current scales are For the score assignment still relies on a strong subjective causes high intra and inter‐rater variability, even when the scale is carried out by experienced examiners [8]. In addition, scales often have similar content and purpose, which makes difficultto decide which of the available tests are superior and should be used as outcome measures. Therefore, there is a need to incorporate new and more objective methods to assess of motorrecovery into the existing panorama. In the following, we present and briefly describe the main measures used in clinical settings to evaluate the functional recovery after SCI.
### **3.1. ASIA impairment scale (AIS)**
This scale [9] provides a correct assessment of the severity of the SCI and can assess the developmental stage of the lesion. The evaluation is mainly based on motor exploration and includes complementary tests for sensory and muscle assessment. Ten muscles are evaluat‐ ed (covering upper and lower limbs) to get muscle balance score between 0 and 5. The sensory examination measures superficial and deep sensitivity on a scale of 0–2 in 28 sensitive key points (for example, the big toe to C6 or armpit for L2). The sum of the motor and sensory scores reflects the global degree of impairment. According to international standards estab‐ lished by the American Association of SCI AIS, this can be classified into five levels, ranging between complete lesion (A, sensory but not motor function is preserved below the neurolog‐ ical level and includes the sacral segments S4–S5) and normality (E, the sensorimotor func‐ tions are completely preserved). This provides a correct assessment of the severity of the SCI and can assess includes complementary tests for sensory and muscle assessment. Ten muscles are evaluated (covering upper and lower limbs) to get muscle balance score between 0 and 5. The examination measures superficial and deep sensitivity on a scale sensitive (for the big toe to or armpit for L2). The sum of the motor and scores reflects the global degree of impairment. According to international standards of SCI this can into (A, but not is ical level and includes the sacral segments S4–S5) and normality (E, the sensorimotor functionsare completely preserved).
#### **3.2. Walking index for spinal cord injury (WISCI II) Walking index injury (WISCI**
This index [10] specifically measures the functionality of walking and aims to quantify the secondary physical limitations in SCI. A recent study on functional measures of gait in people with SCI [11] showed how previous scales (e.g., 10 m and timed up and go, TUG) were unable to discriminate the improvement in function of patients who needed assistive devices during walking, such as walker or crutches. These data suggest that the WISCII scale serves as a valid tool for measuring changes in locomotion and effectiveness of treatment after SCI when technical help or the assistance of an external person is required to realize gait function. The scoring method is based on visual observation of the patient's ability during walking over at least 10 m. Walking ability is classified in this index, from the level with the most severe disabilities (0), to less severe disability (20). Within this range, the patient is classified according to the use of devices, braces, and physical assistance of one or two people. This [10] specifically measures the functionality of walking and aims to quantify secondary limitations in A of gait peoplewith [11] showed how previous (e.g., m and timed up and were unablediscriminate the patients who needed assistive devices such as walker or These data suggest that the WISCII scale serves as a validtool for measuring changes in locomotion and effectiveness of treatment after SCI whenthe external is required to realize gait is during walking least 10 m. Walking ability is classified in this from the level with the most to less severe Within this range, the patient is classified accordingto the assistance of stroke and Parkinson's and then transferred as an alternative to evaluate patients 3064in Spinal Cord Injury
#### **3.3. Timed up and go (TUG) test 3.3. Timed up and go**
Originally developed to study the balance in senior age [12], this test demonstrated its validity as a measure of walking performance in neurologically injured subjects [13]. This tool measures the time (in seconds) that patients take to get up from a chair, walk 3 m, and sit back. Walk 3 m is the traditional test TUG, but there are adaptations in which the patient has to walk 10 or 7 m. The score ranges from one (normal) to five points (very abnormal). Times above 20 s are predictors of falls. This test captures the complex interaction between balance and move‐ ment, including planning, initiation, and implementation, and completes a series of linked movements that are common in activities of daily living. The validity of the TUG has been studied in different works and found to have a sensitivity and specificity of 87%. Originally developed to study the balance in senior this test demonstrated its as a measure of walking performance in neurologically injured This tool m is the traditional test TUG, but there are adaptations in which the patient has to walk 10 or 7 m. The score ranges from one (normal) to five points (very abnormal). Times above 20 s arepredictors of falls. This test captures the complex interaction between balance and movement,including planning, initiation, and implementation, and completes a series of linked move‐ ments that are common in activities of daily living. The validity of the TUG has been studiedin different works and found to have a sensitivity and specificity of 87%.
#### **3.4. 10‐meter walk test (10MWT) test (10MWT)**
This test quantifies the walking speed by measuring the time (in seconds) that the patient takes to travel a distance of 10 m [13]. This scale was firstly proposed to assess gait in patients with stroke and Parkinson's disease, and then transferred as an alternative to evaluate patients with This test quantifies the walking speed by measuring the (in seconds) that the patient takestravel a with SCI. A study [14] comparing several clinical scales of motion concluded that 10MWT is a more reliable and easier handling tool than the test of 6‐minute walk test (see next section). As a drawback, this test requires additional information on technical aids needed during walking. A comparing several clinical scales of motion concluded 10MWT is a reliable and easier handling tool than the test of 6‐minute walk test (see next section). As adrawback, this test requires additional information on technical aids needed during walking.
#### **3.5. 6‐minute walk test (6MWT) (6MWT)**
**3.1. ASIA impairment scale (AIS)**
308 Recovery of Motor Function Following Spinal Cord Injury
tions are completely preserved).
are completely preserved).
**3.3. Timed up and go (TUG) test**
**3.4. 10‐meter walk test (10MWT)**
**3.4. 10‐meter walk test (10MWT)**
**3.2. Walking index for spinal cord injury (WISCI II)**
**(TUG) test**
This scale [9] provides a correct assessment of the severity of the SCI and can assess the developmental stage of the lesion. The evaluation is mainly based on motor exploration and includes complementary tests for sensory and muscle assessment. Ten muscles are evaluat‐ ed (covering upper and lower limbs) to get muscle balance score between 0 and 5. The sensory examination measures superficial and deep sensitivity on a scale of 0–2 in 28 sensitive key points (for example, the big toe to C6 or armpit for L2). The sum of the motor and sensory scores reflects the global degree of impairment. According to international standards estab‐ lished by the American Association of SCI AIS, this can be classified into five levels, ranging between complete lesion (A, sensory but not motor function is preserved below the neurolog‐ ical level and includes the sacral segments S4–S5) and normality (E, the sensorimotor func‐
This provides a correct assessment of the severity of the SCI and can assess includes complementary tests for sensory and muscle assessment. Ten muscles are evaluated
(covering upper and lower limbs) to get muscle balance score between 0 and 5. The sensoryexamination measures superficial and deep sensitivity on a scale sensitive points (for the big toe to or armpit for The sum of the motor and
(A, function below
scores reflects the global degree of impairment. According to international standards the American SCI five
**spinal cord (WISCI II)**
This index [10] specifically measures the functionality of walking and aims to quantify the secondary physical limitations in SCI. A recent study on functional measures of gait in people with SCI [11] showed how previous scales (e.g., 10 m and timed up and go, TUG) were unable to discriminate the improvement in function of patients who needed assistive devices during walking, such as walker or crutches. These data suggest that the WISCII scale serves as a valid tool for measuring changes in locomotion and effectiveness of treatment after SCI when technical help or the assistance of an external person is required to realize gait function. The scoring method is based on visual observation of the patient's ability during walking over at least 10 m. Walking ability is classified in this index, from the level with the most severe disabilities (0), to less severe disability (20). Within this range, the patient is classified according
This specifically measures the functionality of walking and aims to quantify physical in functional measures of with SCI showed how previous 10 m and timed up and were improvement in of needed assistive such as walker or These data suggest that the WISCII scale serves as a tool for measuring changes in locomotion and effectiveness of treatment after SCI technical or assistance person required to realize function. is based on the during walking over
m. Walking ability is classified in this index, from the level with the most disabilities (0), to less severe (20). Within this range, the patient is classified of devices, braces, and physical
stroke and Parkinson's and then transferred as an alternative to evaluate patients 3064in Spinal Cord Injury
ical level and includes the sacral segments S4–S5) and normality (E, the sensorimotor functions
Originally developed to study the balance in senior age [12], this test demonstrated its validity as a measure of walking performance in neurologically injured subjects [13]. This tool measures the time (in seconds) that patients take to get up from a chair, walk 3 m, and sit back. Walk 3 m is the traditional test TUG, but there are adaptations in which the patient has to walk 10 or 7 m. The score ranges from one (normal) to five points (very abnormal). Times above 20 s are predictors of falls. This test captures the complex interaction between balance and move‐ ment, including planning, initiation, and implementation, and completes a series of linked movements that are common in activities of daily living. The validity of the TUG has been
Originally developed to study the balance in senior this test demonstrated its as a measure of walking performance in neurologically injured This tool m is the traditional test TUG, but there are adaptations in which the patient has to walk 10 or
7 m. The score ranges from one (normal) to five points (very abnormal). Times above 20 s arepredictors of falls. This test captures the complex interaction between balance and movement, including planning, initiation, and implementation, and completes a series of linked move‐ ments that are common in activities of daily living. The validity of the TUG has been studiedin different works and found to have a sensitivity and specificity of 87%.
This test quantifies the walking speed by measuring the time (in seconds) that the patient takes to travel a distance of 10 m [13]. This scale was firstly proposed to assess gait in patients with stroke and Parkinson's disease, and then transferred as an alternative to evaluate patients with
This test quantifies the walking speed by measuring the time (in that the patient
to the use of devices, braces, and physical assistance of one or two people.
studied in different works and found to have a sensitivity and specificity of 87%.
10 m [13]. scale
This test calculates the distance (in meters) that patients are able to walk in six minutes [15]. Initially, it was used to measure cardiovascular exercise capacity in elderly patients with heart or lung disease, as a submaximal test of the aerobic capacity of the individual. It is one of the most widely used tests to evaluate gait motor recovery in neurologically injured subjects [16]. This test calculates the distance (in meters) that patients are able to walk in six minutes [15].Initially, it was used to measure cardiovascular exercise capacity in elderly patients with heartor lung disease, as a submaximal test of the aerobic capacity of the individual. It is one of themost widely used tests to evaluate gait motor recovery in neurologically injured subjects [16].
#### 3.6. Spinal cord (functional) independence measures (FIM and SCIM) 3.6. Spinal cord (functional) independence measures (FIM and SCIM)
The functional independence measure (FIM) scale was originally developed to assess disability of individuals with stroke [17]. This tool has been then widely applied for the evaluation of the functionality in the SCI. It is a standard and validated scale studying the state of function‐ alrecovery of patients from hospital admission to discharge. It is also used globally to monitor the progress of the functionality of patients throughout the rehabilitation treatment. It measures 18 activities, of which 13 items referto the motor area and five to the cognitive status. These 18 items are grouped into six blocks: self‐care, sphincter control, transfers, locomotion, communication, and awareness of the external world. It includes seven levels ranging from total dependence to full independence [18,19]. The functional independence measure (FIM) scale was originally developed to assess disabilityof individuals with stroke [17]. This tool has been then widely applied for the evaluation ofthe functionality in the SCI. It is a standard and validated scale studying the state of functional recovery of patients from hospital admission to discharge. It is also used globally to the progress of the functionality of patients throughout the rehabilitation measures 18 activities, of which 13 items refer to the motor area and five to the cognitive status.self‐care, sphincter communication, and awareness of the external world. It includes seven levels ranging fromtotal dependence to full independence [18,19].The SCIM scale was developed specifically to improve some aspects of the FIM scale [13].
The SCIM scale was developed specifically to improve some aspects of the FIM scale [13]. This scale consists of 16 items (score range 0–100) and includes three levels of activity: (1) self‐ care (feeding, grooming, bathing, and dressing), (2) respiration and sphincter management, and (3) mobility (bed and transfers and indoor/outdoor). Scores are obtained by adding up the individual items. The item scores are weighted related to the assumed clinical relevance. scale consists of 16 items (score range 0–100) and includes three levels of activity: (1) self‐care (feeding, grooming, bathing, and dressing), (2) respiration and sphincter management, and (3)
### **4. Scales to assess clinical conditions of spasticity**
The correct measurement of spasticity is of vital importance in SCI for three main reasons: (i) evaluating the effectiveness of anti‐spasticity treatments, (ii) optimizing individually the amount of antispastic medication to patients, and (iii) understanding the pathophysiological mechanisms underlying this disorder [20]. However, although spasticity is usually easy to recognize, its quantification is a much more complex process [21]. A unique scale that can provide a general and objective assessment of spasticity is still not available in the literature; therefore, the various tests and clinical measures are usually combined in order to evaluate different aspects of spasticity. The existing clinical measurements include scales that meas‐ ure muscle hypertonia (modified Ashworth scale (MAS), EAM) [22], the spasms often suffered by patients (frequency spasms scale Penn) [23], and the reflected hyperexcitability after SCI (scale of severity of spasms (SCATS) [24]. In the following sections, these three methods will be briefly described. mobility (bed and transfers and indoor/outdoor). Scores are obtained by adding up theindividual items. The item scores are weighted related to the assumed clinical relevance.The correct measurement of spasticity is of vital importance in SCI for three main reasons: (i)evaluating the effectiveness of anti‐spasticity treatments, (ii) optimizing individually theamount of antispastic medication to patients, and (iii) understanding the pathophysiologicalmechanisms underlying this [20]. However, although spasticity is usually easy torecognize, its quantification is a much more complex process [21]. A unique scale that canprovide a general and objective assessment of spasticity is still not available in the literature;therefore, the various tests and clinical measures are usually combined in order to evaluatedifferent aspects of spasticity. The existing clinical measurements include scales that measure muscle hypertonia (modified Ashworth scale (MAS), EAM) [22], the spasms often suffered bypatients (frequency spasms scale Penn) [23], and the reflected hyperexcitability after SCI (scale of severity of spasms (SCATS) [24]. In the following sections, these three methods will be briefly described.
### **4.1. Modified Ashworth scale (MAS)**
This scale measures the resistance of muscle to passive stretching [22]. The resistance value is scored between 0 and 4, being "0" the absence of any increase in tone during movement; "1" slight increase in tone and muscle response at the start of movement or increased resistance at the end of the movement; "1+" slight increase in muscle resistance movement followed by minimal resistance throughout the remainder of the range of motion; "2" significant increase in muscle endurance during most arc joint movement, but the joint moves easily; "3" marked increase in muscle strength or passive movement is difficult in flexion or extension; and "4" the affected joints are rigid in flexion or extension. Two physiotherapists should perform this test independently. slight increase in tone and muscle response at the start of movement or increased significant or prior to the The scores on this scale are is the total caused
### **4.2. Penn spasm frequency scale (PSFS)**
The PENN scale is a tool to quantify spasm frequency suffered by the patient during the 24 h priorto the test [25]. The scores on this scale are between 0 and 4, where "0" is the total absence of spasms, "1" spasms caused only by stimulation, "2" spasms that occur less than once every hour, "3" spasms occur more than once every hour, and "4" spasms that occur more than 10 times an hour. As for the MAS scale, two physiotherapists should perform this test independ‐ ently. occur than 10perform
### **4.3. Spinal cord assessment tool for spastic reflexes (SCATS)**
This scale is a physiologically based measure for spastic reflexes for use in individuals with SCI. It measures the reflected hyperexcitability, which includes clonus, flexor spasms, and cramps in the extensors. This scale is developed to provide a measure of primary spastic reaction, addressing categories of spasms as follows: (1) clonus, (2) flexor spasms, and (3) extensor spasms. The spasm is triggered and rated with a score ranging from 0 to 3. The SCATS does not provide information on patient perspective, which is an important aspect since spasms are sometimes perceived beneficial to the patient. This tool is simple and quick (<5 s) to administer but despite its simplicity it has not been widely accepted yet. The measure could be conducted during a home visit or at a clinic/hospital. A study [24] evaluated the validity of this scale, demonstrating that it correlated significantly with kinematic and electromyograph‐ ic measures, and with Ashworth scores. This scale is a physiologically based measure for spastic reflexes for use in individuals It measures the reflected which includes flexor cramps in the This scale is developed to provide a measure of primary spasticaddressing categories of spasms as follows: flexor extensor The spasm is triggered and rated with a score ranging The does not provide information on patient which is an important aspect spasms are sometimes perceived beneficial to the patient. This tool is simple and quick (<5 s)to administer but despite its simplicity it has not been widely accepted yet. The measure at a clinic/hospital. this scale, demonstrating that it correlated significantly with kinematic and electromyographic measures, and with Ashworth scores. EMG 3086in Spinal Cord Injury
### **5. Emerging metrics to assess sensorimotor impairments after spinal cord injury**
Despite its ability to assess and provide information about underlying mechanisms and changes in motor control, electromyography (EMG) is still rarely employed in the clinical setting to assess SCI patients [26]. Thus, novel approaches using EMG should be explored, as EMG can be very valuable to understand compensatory strategies and for further rehabilita‐ tion processes. Despite its ability to assess and provide information about underlying mechanisms changes in motor is still rarely employed in the to assess using
This section describes novel techniques based on muscle synergies and frequency‐domain analysis of EMG signals (muscle coherence). Such techniques are illustrated as potential tools for assessment of motor function after SCI with experimental data and a case study describ‐ ing a diagnostic scenario. This section describes novel techniques based on muscle synergies and for assessment of motor function after SCI with experimental data and a case study describing a diagnostic scenario.
### **5.1. EMG coherence**
**4.1. Modified Ashworth scale (MAS)**
310 Recovery of Motor Function Following Spinal Cord Injury
**4.2. Penn spasm frequency scale (PSFS)**
ic measures, and with Ashworth scores.
measures, and with Ashworth scores.
**4.3. Spinal cord assessment tool for spastic reflexes (SCATS)**
test independently.
ently.
**injury**
tion processes.
This scale measures the resistance of muscle to passive stretching [22]. The resistance value is scored between 0 and 4, being "0" the absence of any increase in tone during movement; "1" slight increase in tone and muscle response at the start of movement or increased resistance at the end of the movement; "1+" slight increase in muscle resistance movement followed by minimal resistance throughout the remainder of the range of motion; "2" significant increase in muscle endurance during most arc joint movement, but the joint moves easily; "3" marked increase in muscle strength or passive movement is difficult in flexion or extension; and "4" the affected joints are rigid in flexion or extension. Two physiotherapists should perform this
slight increase in tone and muscle response at the start of movement or increased of motion; "2" muscle extension; "4"
The PENN scale is a tool to quantify spasm frequency suffered by the patient during the 24 h priorto the test [25]. The scores on this scale are between 0 and 4, where "0" is the total absence of spasms, "1" spasms caused only by stimulation, "2" spasms that occur less than once every hour, "3" spasms occur more than once every hour, and "4" spasms that occur more than 10 times an hour. As for the MAS scale, two physiotherapists should perform this test independ‐
prior to the The scores on this scale are is the total only "2" occur spasms 10
quantify spasm during the
independ‐
This scale is a physiologically based measure for spastic reflexes for use in individuals with SCI. It measures the reflected hyperexcitability, which includes clonus, flexor spasms, and cramps in the extensors. This scale is developed to provide a measure of primary spastic reaction, addressing categories of spasms as follows: (1) clonus, (2) flexor spasms, and (3) extensor spasms. The spasm is triggered and rated with a score ranging from 0 to 3. The SCATS does not provide information on patient perspective, which is an important aspect since spasms are sometimes perceived beneficial to the patient. This tool is simple and quick (<5 s) to administer but despite its simplicity it has not been widely accepted yet. The measure could be conducted during a home visit or at a clinic/hospital. A study [24] evaluated the validity of this scale, demonstrating that it correlated significantly with kinematic and electromyograph‐
Despite its ability to assess and provide information about underlying mechanisms changes in motor is still rarely employed in the SCI patients [26]. EMG 3086in Spinal Cord Injury
This scale is a physiologically based measure for spastic reflexes for use in individuals SCI. It measures the reflected which includes flexor cramps in the This scale is developed to provide a measure of primary reaction, addressing categories of spasms as flexor extensor The spasm is triggered and rated with a score ranging The does not provide information on patient perspective, which is an important aspect spasms are sometimes perceived beneficial to the patient. This tool is simple and quick (<5 s)
to administer but despite its simplicity it has not been widely accepted The measure or at this demonstrating that it correlated significantly with kinematic and electromyographic
**5. Emerging metrics to assess sensorimotor impairments after spinal cord**
Despite its ability to assess and provide information about underlying mechanisms and changes in motor control, electromyography (EMG) is still rarely employed in the clinical setting to assess SCI patients [26]. Thus, novel approaches using EMG should be explored, as EMG can be very valuable to understand compensatory strategies and for further rehabilita‐ Neurophysiological studies demonstrate only limited spontaneous recovery of voluntary motor function after incomplete SCI diagnosed with the American Spinal Injury Association (ASIA) Impairment Scale (AIS) [9,27]. Thus, there is a clinical need to identify new compre‐ hensive outcome measures that may be used to assess sensorimotor impairments after incomplete SCI (iSCI). Effective neurophysiological measures should have clinical relevance, reflecting the recovery of volunteer motor force and gait function, as well as the develop‐ ment of maladaptive motor plasticity, such as the spasticity syndrome, which is known to limit the recovery of voluntary motor strength, gait, and daily life activities following SCI [28–34]. These measures may facilitate clinical diagnosis and guidance of individualized neuroreha‐ bilitation programs. Neurophysiological studies demonstrate only limited spontaneous recovery of Spinal Impairment [9,27]. Thus, there is a clinical need to identify new hensive outcome measures that may be used to assess sensorimotor impairments (iSCI). neurophysiological measures should have clinical relevance,reflecting the recovery of volunteer motor force and gait function, as well as the developmentof maladaptive motor plasticity, such as the spasticity syndrome, which is known to limit the recovery of voluntary motor strength, gait, and daily life activities following SCI [28–34]. These measures may facilitate clinical diagnosis and guidance of individualized neurorehabilitation programs.
### *5.1.1. Muscle coherence*
Lower limb EMG coherence analysis has been used as an indirect measure of voluntary motor control, gait function [35], and spasticity [27] after iSCI. This is a frequency‐domain measure of the similarity between two independent EMG signals, having the potential to assess the descending motor drive [36]. The fact that it just needs EMG recording to be calculated makes it suitable for clinical applications. On the other hand, muscle coherence activity estimation can be obtained from the same muscle (intramuscular coherence) [37,38], as well as between two different muscles (intermuscular coherence) [38–40]. Lower limb EMG coherence analysis has been used as an indirect measure of voluntary control, gait function [35], and spasticity [27] iSCI. a of the similarity between two independent EMG having the potential to assess descending [36]. fact EMG to be it suitable for clinical applications. On the other hand, muscle coherence activity obtained well betweentwo different muscles (intermuscular coherence)
Smith et al. [41] have shown that the synchronization of motoneuron discharges (coherence) was reduced after spinal cord injury and also that it could be better recorded during isomet‐ ric muscle contractions. The analysis of muscle coherence has been used to diagnose lesions in voluntary motor control mechanisms and functionality of SCI patients, for example, gait control [37,38]. Several studies support the use of muscle coherence analysis as an indirect measure of the common descending tracts during the execution of specific motor tasks. For instance, it has been suggested the existence of an association between the force of a maxi‐ mum isometric contraction and corticospinal activity, based on the evidence of a reduction of intracortical inhibition modulated by training muscle strength [42–45]. Despite the potential of this technique, few studies have investigated which type of movement would be optimal for the use of muscle coherence calculation as a greater diagnostic measure. et [41] shown that synchronization (coherence)was reduced after spinal cord injury and also that it could be better recorded during isometric muscle contractions. The analysis of muscle coherence has been used to diagnose lesions in voluntary motor control mechanisms and functionality of SCI patients, for example, [37,38]. Several studies support the use of muscle coherence analysis as an indirectmeasure of the common descending tracts during the execution of specific motor it has been suggested the existence of an association between the force of a maximum isometric contraction and corticospinal activity, based on the evidence of a reduction intracortical inhibition modulated by training muscle Despite the of this few studies have investigated which type of movement would be indirect measure of adaptive neuroplasticity during the rehabilitation phase [46], while 309 7
Measuring the residual activity of tibialis anterior may represent an interesting diagnostic measure of functionality after SCI, mainly because this muscle receives increased innerva‐ tion from the corticospinal system [45]. In fact, ankle dorsiflexion movement has been used as an indirect measure of adaptive neuroplasticity during the rehabilitation phase [46], while the for the use of muscle coherence calculation as a greater diagnostic measure.Measuring the residual activity of tibialis anterior may represent an interesting measure of functionality after SCI, mainly because this muscle receives increased innervation from the corticospinal system [45]. In fact, ankle dorsiflexion movement has been used as an recording of tibialis anterior co‐activation during plantarflexion may be also used as indica‐ tor of maladaptive mechanisms after SCI, such as spasticity or its associated symptoms [47]. recording of tibialis anterior co‐activation during plantarflexion may be also used as indicatorof maladaptive mechanisms after SCI, such as spasticity or its associated symptoms [47].
### *5.1.2. An emerging methodology to assess muscle coherence*
This section sums up the methodology and main results presented by [27], who analyzed intramuscular TA coherence within specific frequency bands between 10–60 Hz from sub‐ jects who suffered an iSCI, while they performed different movement tasks. This section sums up the methodology and main results presented by who analyzedintramuscular TA coherence within specific frequency bands between 10–60 Hz from subjectswho suffered an iSCI, while they performed different movement tasks.
Muscle coherence between two TA EMG signals (intramuscular coherence) was calculated during the periods of activation of this muscle during controlled ankle dorsiflexion move‐ ments. Specifically, EMG was recorded during the following ankle dorsiflexion tasks: (I) isometric activation at 50, 75, and 100% of maximal voluntary torque (MVT), (II) isokinetic activation at 60 and 120°/s and III) isotonic dorsiflexion at 50% MVT. Periods of activation were visually determined after processing EMG signals. This analysis of EMGs started with demeaning of raw signals, followed by band‐pass filtering (3–700 Hz) and rectification. Muscle coherence between two TA EMG signals (intramuscular coherence) was calculatedduring the periods of activation of this muscle during controlled ankle dorsiflexion ments. EMG was recorded during the following ankle dorsiflexion tasks: (I)isometric activation at and 100% of maximal voluntary torque (MVT), (II) activation at 60 and 120°/s and III) isotonic dorsiflexion at 50% MVT. Periods of activation werevisually determined after processing EMG signals. This analysis of EMGs started Muscle coherence was computed using the function "mscohere" of Matlab (Version output of this function is the magnitude‐squared coherence estimate of the input signals (both
Muscle coherence was computed using the function "mscohere" of Matlab (Version 7.11). The output of this function is the magnitude‐squared coherence estimate of the input signals (both TA signals), using Welch's averaged modified periodogram method [48]. Finally, four bandwidths were analyzed: 10–16, 15–30, 24–40, and 40–60 Hz. For each band, it was comput‐ ed the mean magnitude squared coherence estimate. A non‐significant higher level of coherence activity was identified within the 10–16 Hz band in the iSCI spasticity group, when comparing the isometric activation at 100% of MVT for subjects diagnosed without or with spasticity. For the other bands of frequency, no differences were identified for TA intramus‐ cular coherence [27]. TA signals), using Welch's averaged modified periodogram method [48]. Finally, fourwere analyzed: 10–16, 15–30, 24–40, and 40–60 Hz. For each band, it was computed the mean magnitude squared coherence estimate. A non‐significant higher level of coherence activity was identified within the 10–16 Hz band in the iSCI spasticity group, when comparingthe isometric activation at 100% of MVT for subjects diagnosed without or with spasticity. For the other bands of frequency, no differences were identified for TA intramuscular coherence [27].
There was also a negative correlation between the TA muscle coherence measured during isometric activation at 100% of MVT and specific symptoms of SCI spasticity muscle hyper‐ tonia, passive resistive torque, and involuntary muscle contractions. The modified Ashworth scores correlated negatively with TA coherence within the 24–60 Hz frequency band. The severity spasms measured with the SCATS scale also presented a negative correlation the 40– 60 Hz band. On the other hand, a positive correlation was found between the Penn score and the TA coherence for the 15–30 Hz band. A positive correlation was also identified between the degree of clonus activity and TA coherence within the 10–16 Hz bandwidth [27]. There was also a negative correlation between the TA muscle coherence measured during isometric activation at 100% of MVT and specific symptoms of SCI spasticity muscle scores correlated negatively with TA coherence within 24–60 Hz frequency band. the 60 Hz band. On the other hand, a positive correlation was found between the Penn score andthe TA coherence for 15–30 Hz band. A positive correlation was also identified between
TA coherence calculated as the ratio of 120/60°/s isokinetic activation was higher in the iSCI spastic subgroup, when compared with the non‐spastic group [27]. In summary, the results presented by Bravo‐Esteban et al. [27] suggest that TA muscle coherence activity estimation during the execution of controlled ankle dorsiflexion movements may be used to assess the level of spasticity in iSCI patients. TA coherence calculated as the ratio of 120/60°/s isokinetic activation was higher in the spastic subgroup, when compared with the non‐spastic group In summary, the resultspresented by Bravo‐Esteban et al. [27] suggest that TA muscle coherence activity estimationduring the execution of controlled ankle dorsiflexion movements may be used to assess level of spasticity in patients.
#### **5.2. Analysis of muscle synergies during cycling synergies during cycling**
Given the redundancy of the musculoskeletal system, a long‐stand idea is that the central nervous system (CNS) controls muscle activation through the use of a synergistic organiza‐ Given the redundancy of the musculoskeletal system, a long‐stand idea is that the centralnervous system (CNS) controls muscle activation through the use of a synergistic organization tion constituted by basic control elements called synergies (or motor modules) [49–51]. In 1967, Nikolai Bernstein [52] proposed the existence of muscle synergies as a simplified strategy of motor control. Muscle synergies are functional sets of muscles, co‐activated by varying their timing and/or neural drive. This synergistic control is thought to underlie the execution of different biomechanical tasks [53]. constituted by basic control elements called synergies (or motor modules) [49–51]. In [52] the of synergies timing and/or neural drive. This synergistic control is thought to underlie the execution of
Muscle synergies may offer new insight into the underlying motor strategies responsible for impaired locomotion [1]. Thus, muscle synergies have been proposed as a potential techni‐ que to measure motor recovery [4]. Nevertheless, it is still difficult to assess or predict motor performance in those patients who lack the required muscle strength to walk during the early stage of the rehabilitation, even with some body weight supports. Given the similarities in kinematics and muscle control with walking [54], cycling may be explored as a novel frame‐ work to assess motor performance in iSCI patients. into impaired locomotion [1]. Thus, muscle synergies have been proposed as a potential technique to measure motor recovery [4]. Nevertheless, it is still difficult to assess or predict performance in those patients who lack the required muscle strength to walk during the earlystage of the rehabilitation, even with some body weight supports. Given the similarities inkinematics and muscle control with walking [54], cycling be frame‐
### *5.2.1. Comparison of two iSCI patients*
Based on the findings presented in our previous research [3], confirming the hypothesis that similar synergistic features are shared between walking and cycling, we further performed a research testing the hypothesis that muscle synergies outcomes extracted during cycling can be used as indicators of gait performance in iSCI patients. Preliminary results of two pa‐ tients are presented in this section. Based the in our [3], confirming the hypothesis that shared walking athe during be used as indicators of gait performance in iSCI patients. Preliminary results of two patients are presented in this section.
### *5.2.1.1. Subjects*
recording of tibialis anterior co‐activation during plantarflexion may be also used as indica‐ tor of maladaptive mechanisms after SCI, such as spasticity or its associated symptoms [47].
recording of tibialis anterior co‐activation during plantarflexion may be also used as indicatorof maladaptive mechanisms after SCI, such as spasticity or its associated symptoms [47].
This section sums up the methodology and main results presented by [27], who analyzed intramuscular TA coherence within specific frequency bands between 10–60 Hz from sub‐
This section sums up the methodology and main results presented by who analyzedintramuscular TA coherence within specific frequency bands between 10–60 Hz from subjects
*coherence*
Muscle coherence between two TA EMG signals (intramuscular coherence) was calculated during the periods of activation of this muscle during controlled ankle dorsiflexion move‐ ments. Specifically, EMG was recorded during the following ankle dorsiflexion tasks: (I) isometric activation at 50, 75, and 100% of maximal voluntary torque (MVT), (II) isokinetic activation at 60 and 120°/s and III) isotonic dorsiflexion at 50% MVT. Periods of activation were visually determined after processing EMG signals. This analysis of EMGs started with demeaning of raw signals, followed by band‐pass filtering (3–700 Hz) and rectification.
Muscle coherence was computed using the function "mscohere" of Matlab (Version 7.11). The output of this function is the magnitude‐squared coherence estimate of the input signals (both TA signals), using Welch's averaged modified periodogram method [48]. Finally, four bandwidths were analyzed: 10–16, 15–30, 24–40, and 40–60 Hz. For each band, it was comput‐ ed the mean magnitude squared coherence estimate. A non‐significant higher level of coherence activity was identified within the 10–16 Hz band in the iSCI spasticity group, when comparing the isometric activation at 100% of MVT for subjects diagnosed without or with spasticity. For the other bands of frequency, no differences were identified for TA intramus‐
TA signals), using Welch's averaged modified periodogram method [48]. Finally, fourwere analyzed: 10–16, 15–30, 24–40, and 40–60 Hz. For each band, it was computed the mean magnitude squared coherence estimate. A non‐significant higher level of coherence activity was identified within the 10–16 Hz band in the iSCI spasticity group, when comparingthe isometric activation at 100% of MVT for subjects diagnosed without or with spasticity. For the other bands of frequency, no differences were identified for TA intramuscular coherence
Muscle coherence between two TA EMG signals (intramuscular coherence) was calculatedduring the periods of activation of this muscle during controlled ankle dorsiflexion ments. EMG was recorded during the following ankle dorsiflexion tasks: (I)isometric activation at and 100% of maximal voluntary torque (MVT), (II) activation at 60 and 120°/s and III) isotonic dorsiflexion at 50% MVT. Periods of activation werevisually determined after processing EMG signals. This analysis of EMGs started Muscle coherence was computed using the function "mscohere" of Matlab (Version output of this function is the magnitude‐squared coherence estimate of the input signals (both
There was also a negative correlation between the TA muscle coherence measured during isometric activation at 100% of MVT and specific symptoms of SCI spasticity muscle hyper‐ tonia, passive resistive torque, and involuntary muscle contractions. The modified Ashworth scores correlated negatively with TA coherence within the 24–60 Hz frequency band. The severity spasms measured with the SCATS scale also presented a negative correlation the 40– 60 Hz band. On the other hand, a positive correlation was found between the Penn score and the TA coherence for the 15–30 Hz band. A positive correlation was also identified between
There was also a negative correlation between the TA muscle coherence measured duringisometric activation at 100% of MVT and specific symptoms of SCI spasticity muscle modified scores correlated negatively with TA coherence within 24–60 Hz frequency band. spasms also 60 Hz band. On the other hand, a positive correlation was found between the Penn score andthe TA coherence for 15–30 Hz band. A positive correlation was also identified betweenTA coherence calculated as the ratio of 120/60°/s isokinetic activation was higher in the
TA coherence calculated as the ratio of 120/60°/s isokinetic activation was higher in the iSCI spastic subgroup, when compared with the non‐spastic group [27]. In summary, the results presented by Bravo‐Esteban et al. [27] suggest that TA muscle coherence activity estimation during the execution of controlled ankle dorsiflexion movements may be used to assess the
spastic subgroup, when compared with the non‐spastic group In summary, the resultspresented by Bravo‐Esteban et al. suggest that TA muscle coherence activity estimationduring the execution of controlled ankle dorsiflexion movements may be used to assess
Given the redundancy of the musculoskeletal system, a long‐stand idea is that the central nervous system (CNS) controls muscle activation through the use of a synergistic organiza‐
Given the redundancy of the musculoskeletal system, a long‐stand idea is that the central nervous system (CNS) controls muscle activation through the use of a synergistic organization
the degree of clonus activity and TA coherence within the 10–16 Hz bandwidth [27].
jects who suffered an iSCI, while they performed different movement tasks.
who suffered an iSCI, while they performed different movement tasks.
*5.1.2. An emerging methodology to assess muscle coherence*
312 Recovery of Motor Function Following Spinal Cord Injury
3108in Spinal Cord Injury
cular coherence [27].
[27].
level of spasticity in iSCI patients.
**5.2. Analysis of muscle synergies during cycling**
**5.2. Analysis of muscle synergies cycling**
Two iSCI patients gave their written consent to participate in this study and for data publica‐ tion, after being informed about the procedures and possible discomfort associated with the experiments, in accordance with the Declaration of Helsinki. The local Toledo Paraplegics Hospital (Spain) Clinical Ethical Committee approved this study (07/05/2013 N°47). Two iSCI patients consent in tion,
**Table 1.** Individual iSCI patients' description, as well as the amount of physical assistance needed and gait performance.
Detailedinformation ofthe patients is presentedin**Table 1**. Both patients receivedthe standard rehabilitation program of the hospital. Inclusion criteria were as follows: aged between 18 and 80 years; motor incomplete spinal lesion (AIS C‐D) of traumatic and non‐traumatic etiology, with a prognosis ofrecovery of the walking function; evolution of at least 1.5 months. Exclusion criteria were as follows: supraspinal or peripheral neurological involvement; history Detailed information of the patients is presented in . Both patients received the standardrehabilitation program of the hospital. Inclusion criteria were as follows: aged between 18 and 80 years; motor incomplete spinal lesion (AIS C‐D) of traumatic and non‐traumatic etiology, with a prognosis of recovery of the walking function; evolution of at least 1.5 months. Exclusion criteria were as follows: supraspinal or peripheral neurological involvement; history of of epilepsy; musculoskeletal involvement of lower limbs or spasticity higher than 3 (meas‐ ured with the Modified Ashworth Scale) for each joint, either for extension or flexion. epilepsy; musculoskeletal involvement of lower limbs or spasticity higher than 3 (measured with the Modified Ashworth Scale) for each joint, either for extension or flexion.
### *5.2.1.2. Experimental protocol*
Prior to the experiment, a trained physiotherapist performed a set of clinical evaluations in order to inform about the functional status of the patients. The gait performance was evalu‐ ated using the timed up and go (TUG) test [55] and the 10‐meter test [56]. In order to quanti‐ fy the amount of assistance required by the subjects (10 m walk), the walking index for spinal cord injury (WISCI II) was applied. This is a 21‐point scale that ranges from 0 (patient unable to stand and/or participate in assisted walking) to 20 (patient ambulates 10 m with no devices, no braces and no physical assistance) [10]. Prior to the a trained physiotherapist performed a set of clinical evaluations order to inform about the functional status of the patients. The gait performance was evaluated using the timed up and go (TUG) test [55] and the 10‐meter test [56]. In order to quantify the amount of assistance required by the subjects (10 m walk), the walking index for spinal cord injury (WISCI II) was applied. This is a 21‐point scale that ranges from 0 (patient unable to
On the day of the experiment, patients received their standard rehabilitation therapy in the morning. In the afternoon, each iSCI patient performed four cycling trials (at 30, 42, 50, and 60 rpm, revolutions per minute) of 30 s duration each, with 60 s resting between trials. These trials were performed on an electronically braked cycle ergometer (MOTOmed viva2, Reck, Betzenweiler, Germany) in the passive mode. For each patient, the order of the trials was randomized to avoid biased results. Patients were asked to perform the experiment while sat on a regular chair. stand and/or participate in assisted walking) to 20 (patient ambulates 10 m with no On the day of the experiment, patients received their standard rehabilitation therapy in In the each iSCI patient performed four cycling (at 30, 42, 50, <sup>60</sup>rpm, revolutions per minute) of 30 s duration each, with 60 s resting between trials. Thesetrials were performed on an electronically braked cycle (MOTOmed viva2, in the passive For each the order of the trials the An auditory metronome was used in order to synchronize cycling frequency
An auditory metronome was used in order to synchronize patients' cycling frequency with the desired cadence. An EMG amplifier (EMG‐USB, OT Bioelettronica, Torino, Italy) with recording bandwidth of 10–750 Hz, overall gain of 1000 V/V, and acquisition frequency of 2048 Hz was used to record surface electromyography (sEMG) of 13 muscles of the most affected leg of patients. The recorded muscles were as follows: Gluteus Medius, Adductor Longus, Sartorius, Tibialis Anterior, Rectus Femoris, Tensor Fascia Latae, Vastus Lateralis, Vastus Medialis, Biceps Femoris, Semitendinosus, Soleus, Gastrocnemius Lateralis, and Gastrocnemius Medialis. The most affected side was determined based on the muscle score [57] of quadriceps, hamstrings, TA, and gastrocnemius for both limbs. the desired cadence. An EMG amplifier (EMG‐USB, OT Bioelettronica, Torino, Italy) withrecording bandwidth of Hz, overall gain V/V, and acquisition frequency 2048 Hz was used to record surface muscles of the affected leg of The recorded muscles were as Gluteus Tibialis Rectus Tensor Fascia Vastus Vastus Biceps Gastrocnemius Gastrocnemius The most affected side was determined based on the muscle [57] of quadriceps, hamstrings, TA, and gastrocnemius for both limbs.
Bipolar sEMG electrodes (Ag‐AgCl, Ambur Neuroline 720, Ambu, Ballerup, Denmark) were fastened with a 2‐cm inter‐electrode distance on each recorded muscle, following SENIAM recommendations for sEMG recording procedures [58]. One angular sensor (Vishay, Malvern, PA) was applied to estimate the crank angle at a sampling frequency of 100 Hz. Segmenta‐ tion of pedaling cycles was then performed based on the bottom dead center (BDC) position of the pedal. Ambu, Ballerup, Denmark) werefastened with a 2‐cm inter‐electrode distance on each recorded muscle, following recommendations for sEMG recording [58]. One angular (Vishay, was applied to estimate the crank angle at a sampling frequency of 100 Hz. Segmentation of pedaling cycles was then performed based on the bottom dead center (BDC) position of the pedal.
### *5.2.1.3. Muscle synergies analysis*
For each patient and trial, 10 continuous pedaling cycles were selected for analysis. The selected raw EMG signals were pre‐processed using high‐pass filtering at 20 Hz, demeaning, rectification, and low‐pass filtering at 5 Hz, resulting in the EMG envelopes [49,59]. Muscle synergies were extracted using a Non‐Negative Matrix Factorization (NNMF) algorithm [60]. A detailed explanation of the procedure to extract muscles synergies is presented in [49] and [3]. For each trial, the NNMF algorithm was run four times, considering as input two to five For each patient and continuous pedaling cycles were selected for EMG were pre‐processed 20 Hz, demeaning, rectification, and low‐pass filtering at 5 Hz, resulting in the EMG envelopes [49,59]. Musclewere extracted Non‐Negative Factorization (NNMF) [60].A detailed explanation of the procedure to extract muscles synergies is presented in [49] and[3]. times, to synergies. In order to avoid local minima, for each run, the NNMF was repeated 40 times and the repetition with the lowest reconstruction error was selected. avoid was the repetition with the lowest reconstruction error was selected.
To assess whether the recorded EMGs were well described as a combination of the identi‐ fied synergies, two indicators of the quality of reconstruction of the EMG data were used: the *variability accounted for* (VAFtotal) [49] and the *coefficient of determination* (*r*<sup>2</sup> ) [61]. Both VAFtotal and *r*<sup>2</sup> have been adopted in most studies on muscle synergies [49,61]. VAFtotal has been suggested to be more stringent than *r*<sup>2</sup> , since it is sensitive to both shape and amplitude of the To assess whether the recorded EMGs were well described as a combination of the identified synergies, two indicators of the quality of reconstruction of the EMG data were used: the *variability accounted for* (VAFtotal) [49] and the *coefficient of determination* (*r*<sup>2</sup> ) [61]. Both VAFtotal and *r*<sup>2</sup> have been adopted in most studies on muscle synergies [49,61]. VAFtotal has been suggested to be more stringent than *r*<sup>2</sup> , since it is sensitive to both shape and amplitude of the signals, whereas *r*<sup>2</sup> only addresses similarity in shape.
only addresses similarity in shape.
### *5.2.1.4. Results*
signals, whereas *r*<sup>2</sup>
of epilepsy; musculoskeletal involvement of lower limbs or spasticity higher than 3 (meas‐
epilepsy; musculoskeletal involvement of lower limbs or spasticity higher than 3 (measured
Prior to the experiment, a trained physiotherapist performed a set of clinical evaluations in order to inform about the functional status of the patients. The gait performance was evalu‐ ated using the timed up and go (TUG) test [55] and the 10‐meter test [56]. In order to quanti‐ fy the amount of assistance required by the subjects (10 m walk), the walking index for spinal cord injury (WISCI II) was applied. This is a 21‐point scale that ranges from 0 (patient unable to stand and/or participate in assisted walking) to 20 (patient ambulates 10 m with no devices,
Prior to the a trained physiotherapist performed a set of clinical evaluations order to inform about the functional status of the patients. The gait performance was evaluated using the timed up and go (TUG) test [55] and the 10‐meter test [56]. In order to quantify the amount of assistance required by the subjects (10 m walk), the walking index for spinal cord injury (WISCI II) was applied. This is a 21‐point scale that ranges from 0 (patient unable to
On the day of the experiment, patients received their standard rehabilitation therapy in the morning. In the afternoon, each iSCI patient performed four cycling trials (at 30, 42, 50, and 60 rpm, revolutions per minute) of 30 s duration each, with 60 s resting between trials. These trials were performed on an electronically braked cycle ergometer (MOTOmed viva2, Reck, Betzenweiler, Germany) in the passive mode. For each patient, the order of the trials was randomized to avoid biased results. Patients were asked to perform the experiment while sat
stand and/or participate in assisted walking) to 20 (patient ambulates 10 m with no assistance) [10].On the day of the patients received their standard rehabilitation therapy in In the each iSCI patient performed four cycling (at 30, 42, 50,
An auditory metronome was used in order to synchronize patients' cycling frequency with the desired cadence. An EMG amplifier (EMG‐USB, OT Bioelettronica, Torino, Italy) with recording bandwidth of 10–750 Hz, overall gain of 1000 V/V, and acquisition frequency of 2048 Hz was used to record surface electromyography (sEMG) of 13 muscles of the most affected leg of patients. The recorded muscles were as follows: Gluteus Medius, Adductor Longus, Sartorius, Tibialis Anterior, Rectus Femoris, Tensor Fascia Latae, Vastus Lateralis, Vastus Medialis, Biceps Femoris, Semitendinosus, Soleus, Gastrocnemius Lateralis, and Gastrocnemius Medialis. The most affected side was determined based on the muscle score
the desired cadence. An EMG amplifier (EMG‐USB, OT Bioelettronica, Torino, Italy) withrecording bandwidth of 10–750 Hz, overall gain V/V, and acquisition frequency
<sup>60</sup>rpm, revolutions per minute) of 30 s duration each, with 60 s resting between trials. Thesetrials were performed on an electronically braked cycle (MOTOmed viva2, in the passive For each the order of the trials to avoid An auditory metronome was used in order to synchronize cycling frequency
Bipolar sEMG electrodes (Ag‐AgCl, Ambur Neuroline 720, Ambu, Ballerup, Denmark) were fastened with a 2‐cm inter‐electrode distance on each recorded muscle, following SENIAM recommendations for sEMG recording procedures [58]. One angular sensor (Vishay, Malvern, PA) was applied to estimate the crank angle at a sampling frequency of 100 Hz. Segmenta‐ tion of pedaling cycles was then performed based on the bottom dead center (BDC) position
Ambu, Ballerup, Denmark) werefastened with a 2‐cm inter‐electrode distance on each recorded muscle, following
recommendations for sEMG recording [58]. One angular (Vishay, was applied to estimate the crank angle at a sampling frequency of 100 Hz. Segmentation of pedaling cycles was then performed based on the bottom dead center (BDC) position of the
2048 Hz was used to record surface (sEMG) muscles of the affected leg of The recorded muscles were as Gluteus Tibialis Rectus Tensor Fascia Vastus Vastus Biceps Gastrocnemius Gastrocnemius The most affected side was determined based on the muscle [57] of quadriceps, hamstrings, TA, and gastrocnemius for both limbs.
For each patient and trial, 10 continuous pedaling cycles were selected for analysis. The selected raw EMG signals were pre‐processed using high‐pass filtering at 20 Hz, demeaning, rectification, and low‐pass filtering at 5 Hz, resulting in the EMG envelopes [49,59]. Muscle synergies were extracted using a Non‐Negative Matrix Factorization (NNMF) algorithm [60]. A detailed explanation of the procedure to extract muscles synergies is presented in [49] and [3]. For each trial, the NNMF algorithm was run four times, considering as input two to five
rectification, and low‐pass filtering at 5 Hz, resulting in the EMG envelopes [49,59]. Muscle
[60].A detailed explanation of the procedure to extract muscles synergies is presented in [49] and[3]. trial, algorithm run
For each patient and continuous pedaling cycles were selected for
[57] of quadriceps, hamstrings, TA, and gastrocnemius for both limbs.
ured with the Modified Ashworth Scale) for each joint, either for extension or flexion.
with the Modified Ashworth Scale) for each joint, either for extension or flexion.
*5.2.1.2. Experimental protocol*
314 Recovery of Motor Function Following Spinal Cord Injury
31210in Spinal Cord Injury
on a regular chair.
of the pedal.
pedal.
*5.2.1.3. Muscle synergies analysis*
no braces and no physical assistance) [10].
*5.2.1.4. Results* Patient 02 presented better motor performance than patient 01. In the case of WISCI II, patient 01 scored 15 points, whereas patient 02 score 19 points. Patient 01 needed 19.3 s to perform TUG test, whereas patient 02 needed 25.1 s to perform the same test. Patient 01 Patient 02 presented better motor performance than patient 01. In the case of WISCI II, patient 01 scored 15 points, whereas patient 02 score 19 points. Patient 01 needed 19.3 s to perform TUG test, whereas patient 02 needed 25.1 s to perform the same test. Patient 01 needed 27.7 to perform the 10‐meter test, whereas patient 02 just needed 10.2 s to perform the same test.
needed 27.7 to perform the 10‐meter test, whereas patient 02 just needed 10.2 s to perform the same test. As a general trend, patient 02 presented higher values of VAFtotal and *r*<sup>2</sup> than patient 01, for all speeds and number of synergies. Also, the higher the number of synergies used to recon‐ struct EMG data, the better the data were reconstructed (higher VAFtotal and *r*<sup>2</sup> values). Both patients reached their minimum values of VAFtotal at 50 rpm, using two synergies to recon‐ struct EMG data. VAFtotal values in such condition were 79% (**Figure 1A**) for patient 01 and As a general trend, patient 02 presented higher values of VAFtotal and *r*<sup>2</sup> than patient 01, for all speeds and number of synergies. Also, the higher the number of synergies used to reconstruct EMG data, the better the data were reconstructed (higher VAFtotal and *r*<sup>2</sup> values). Both patients reached their minimum values of VAFtotal at 50 rpm, using two synergies to reconstruct EMG data. VAFtotal values in such condition were 79% (**Figure 1A**) for patient 01 and 76% (**Fig‐ ure 1B**) for patient 02. Patient 01 reached 92% as maximum VAFtotal value. This value was observed when using five synergies at 60 rpm. For patient 02, a maximum VAFtotal of 95% was reached when using five synergies at 42 rpm.
76% (**Figure 1B**) for patient 02. Patient 01 reached 92% as maximum VAFtotal value. This value
**Figure 1.** Variability accounted for (VAFtotal) for patient 01 (A) and patient 02 (B), as well as coefficient of determination (*r*2 ) for patient 01 (C) and patient 02 (D), according to the number of synergies, for each of the four speeds (30, 42, 50, and 60 rpm).
**Figure 1.** Variability accounted for (VAFtotal) for patient 01 (A) and patient 02 (B), as well as coefficient of determination (*r*2 ) for patient 01 (C) and patient 02 (D), according to the number of synergies, for each of the four speeds (30, 42, 50, and 60 rpm). In the case of *r*<sup>2</sup> values, a minimum of 0.49 (**Figure 1C**) and 0.52 (**Figure 1D**) were obtained for patient 01 (at 50 rpm) and patient 02 (at 60 rpm), respectively. Both values were obtained using two synergies. On the other hand, maximum *r*<sup>2</sup> values of 0.82 and 0.89 were obtained for patient 01 (at 60 rpm) and patient 02 (at 42 rpm), respectively. Both values were obtained using five synergies. The quality of reconstruction indicators seems to correlate positively with WISCI II scores, that is, patient 02 presented higher values of VAFtotal and *r*<sup>2</sup> and also higher In the case of *r*<sup>2</sup> values, a minimum of 0.49 (**Figure 1C**) and 0.52 (**Figure 1D**) were obtained for patient 01 (at 50 rpm) and patient 02 (at 60 rpm), respectively. Both values were obtained using two synergies. On the other hand, maximum *r*<sup>2</sup> values of 0.82 and 0.89 were obtained for patient 01 (at 60 rpm) and patient 02 (at 42 rpm), respectively. Both values were obtained using five synergies. The quality of reconstruction indicators seems to correlate positively with WISCI II scores, that is, patient 02 presented higher values of VAFtotal and *r*<sup>2</sup> and also higher values of WISCI II. On the other hand, quality of reconstruction indicators seem to correlate negatively with TUG and 10‐meter tests, that is, patient 02 presented higher values of VAFtotal and *r*<sup>2</sup> and also needed less time to perform these two gait performance tests.
#### values of WISCI II. On the other hand, quality of reconstruction indicators seem to correlate negatively with TUG and 10‐meter tests, that is, patient 02 presented higher values of VAFtotal *5.2.1.5. Discussion*
and *r*<sup>2</sup> and also needed less time to perform these two gait performance tests. *5.2.1.5. Discussion* Based on the observed common muscle synergies between cycling and walking [3], this Based on the observed common muscle synergies between cycling and walking [3], this preliminary study tested the hypothesis that the analysis of muscle synergies during cycling correlates with gait performance scales in iSCI patients. These preliminary results corroborate this hypothesis.
preliminary study tested the hypothesis that the analysis of muscle synergies during cycling correlates with gait performance scales in iSCI patients. These preliminary results corrobo‐ rate this hypothesis. Results showed positive correlations between WISCI II and EMG reconstruction goodness scores (VAFtotal and *r*<sup>2</sup> ) when recording a set of 13 muscles of the most affected leg. In the case of gait speed tests, reconstruction goodness scores correlated negatively with TUG and 10‐ meter tests. Our results are in agreement with the fact that patients with lower amount of required assistance and good walking performance present higher signal‐to‐noise ratio in Results showed positive correlations between WISCI II and EMG reconstruction goodness scores (VAFtotal and *r*<sup>2</sup> ) when recording a set of 13 muscles of the most affected leg. In the case of gait speed tests, reconstruction goodness scores correlated negatively with TUG and 10‐ meter tests. Our results are in agreement with the fact that patients with lower amount of required assistance and good walking performance present higher signal‐to‐noise ratio in EMG signals than patients with poor walking performance, as severely impaired subjects usually present reduced signal‐to‐noise ratio in the EMG signals due to reduced signal strength [60]. As a consequence of lower signal‐to‐noise, lower VAFtotal, and *r*<sup>2</sup> values are expected [5].
EMG signals than patients with poor walking performance, as severely impaired subjects usually present reduced signal‐to‐noise ratio in the EMG signals due to reduced signal
#### strength [60]. As a consequence of lower signal‐to‐noise, lower VAFtotal, and *r*<sup>2</sup> values are **6. Conclusions**
expected [5].
**6. Conclusions** Recovery after SCI is greatly dependent on the severity of the injury as well as the treatment Recovery after SCI is greatly dependent on the severity of the injury as well as the treatment provided in each phase of the lesion, being the subacute phase (up to approximately 6 months) the best phase to promote plastic changes in the CNS. Functional Improvements such as the increase of gait speed increased recovered distance or better WISCI II scores have been observed in acute SCI patients if compared to chronic SCI patients [62].
provided in each phase of the lesion, being the subacute phase (up to approximately 6 months) the best phase to promote plastic changes in the CNS. Functional Improvements such as the increase of gait speed increased recovered distance or better WISCI II scores have been observed in acute SCI patients if compared to chronic SCI patients [62]. The application of simple diagnostic measures that might provide comprehensive informa‐ tion regarding the state of adaptive and maladaptive motor control mechanisms after SCI could play a crucial role in guiding rehabilitation strategies in the clinic. Neurophysiological The application of simple diagnostic measures that might provide comprehensive information regarding the state of adaptive and maladaptive motor control mechanisms after SCI could play a crucial role in guiding rehabilitation strategies in the clinic. Neurophysiological measures should be preferred over qualitative clinical measures, as they are more independ‐ ent, provide objective data and can be performed in less cooperative patients [63]. However, standardization and clinical validation should be achieved to allow for wide use in the clinical setting.
measures should be preferred over qualitative clinical measures, as they are more independ‐ ent, provide objective data and can be performed in less cooperative patients [63]. However, The analysis of muscle synergies during cycling can be explored as a novel approach for the quantitative assessment of gait performance. This analysis can complement current assessment standardization and clinical validation should be achieved to allow for wide use in the clinical setting. The analysis of muscle synergies during cycling can be explored as a novel approach for the quantitative assessment of gait performance. This analysis can complement current assess‐ procedures. On the other hand, the analysis of intramuscular coherence of tibialis anterior can also be explored as a measure of spasticity during the subacute phase of recovery in SCI patients, as it provides information on the mechanisms of maladaptive plasticity (specifically spasticity) [27].
ment procedures. On the other hand, the analysis ofintramuscular coherence oftibialis anterior can also be explored as a measure of spasticity during the subacute phase of recovery in SCI patients, as it provides information on the mechanisms of maladaptive plasticity (specifical‐ ly spasticity) [27]. In the future, additional researches are needed in order to validate TA muscle coherence analysis and use it in the clinical setting for the assessment of sensorimotor impairments in SCI patients. Specifically, longitudinal studies have to be performed since the initial stage of In the future, additional researches are needed in order to validate TA muscle coherence analysis and use it in the clinical setting for the assessment of sensorimotor impairments in SCI patients. Specifically, longitudinal studies have to be performed since the initial stage of recovery, comparing muscle coherence with clinical and functional scales. This will be also useful to provide new information on the neurophysiological mechanisms present during the recovery stage in SCI patients.This work has been partially funded by grant from the European Commission, within the Seventh Framework Programme (IFP7-ICT-2013-10-611695: BioMot - Smart Wearable Robots with Bioinspired Sensory-Motor Skills).
recovery, comparing muscle coherence with clinical and functional scales. This will be also useful to provide new information on the neurophysiological mechanisms present during the recovery stage in SCI patients.This work has been partially funded by grant from the European
#### Commission, within the Seventh Framework Programme (IFP7-ICT-2013-10-611695: BioMot - **Author details**
**Figure 1.** Variability accounted for (VAFtotal) for patient 01 (A) and patient 02 (B), as well as coefficient of determination
In the case of *r*<sup>2</sup> values, a minimum of 0.49 (**Figure 1C**) and 0.52 (**Figure 1D**) were obtained for patient 01 (at 50 rpm) and patient 02 (at 60 rpm), respectively. Both values were obtained using
patient 01 (at 50 rpm) and patient 02 (at 60 rpm), respectively. Both values were obtained using two synergies. On the other hand, maximum *r*<sup>2</sup> values of 0.82 and 0.89 were obtained for patient 01 (at 60 rpm) and patient 02 (at 42 rpm), respectively. Both values were obtained using five synergies. The quality of reconstruction indicators seems to correlate positively with WISCI II scores, that is, patient 02 presented higher values of VAFtotal and *r*<sup>2</sup> and also higher values of WISCI II. On the other hand, quality of reconstruction indicators seem to correlate negatively with TUG and 10‐meter tests, that is, patient 02 presented higher values of VAFtotal
01 (at 60 rpm) and patient 02 (at 42 rpm), respectively. Both values were obtained using five synergies. The quality of reconstruction indicators seems to correlate positively with WISCI II scores, that is, patient 02 presented higher values of VAFtotal and *r*<sup>2</sup> and also higher values of WISCI II. On the other hand, quality of reconstruction indicators seem to correlate negatively with TUG and 10‐meter tests, that is, patient 02 presented higher values of VAFtotal and *r*<sup>2</sup>
Based on the observed common muscle synergies between cycling and walking [3], this preliminary study tested the hypothesis that the analysis of muscle synergies during cycling correlates with gait performance scales in iSCI patients. These preliminary results corrobo‐
Results showed positive correlations between WISCI II and EMG reconstruction goodness
Based on the observed common muscle synergies between cycling and walking [3], this preliminary study tested the hypothesis that the analysis of muscle synergies during cycling correlates with gait performance scales in iSCI patients. These preliminary results corroborate
Results showed positive correlations between WISCI II and EMG reconstruction goodness
of gait speed tests, reconstruction goodness scores correlated negatively with TUG and 10‐ meter tests. Our results are in agreement with the fact that patients with lower amount of required assistance and good walking performance present higher signal‐to‐noise ratio in EMG signals than patients with poor walking performance, as severely impaired subjects usually present reduced signal‐to‐noise ratio in the EMG signals due to reduced signal strength
of gait speed tests, reconstruction goodness scores correlated negatively with TUG and 10‐ meter tests. Our results are in agreement with the fact that patients with lower amount of required assistance and good walking performance present higher signal‐to‐noise ratio in EMG signals than patients with poor walking performance, as severely impaired subjects usually present reduced signal‐to‐noise ratio in the EMG signals due to reduced signal strength [60]. As a consequence of lower signal‐to‐noise, lower VAFtotal, and *r*<sup>2</sup> values are
Recovery after SCI is greatly dependent on the severity of the injury as well as the treatment provided in each phase of the lesion, being the subacute phase (up to approximately 6 months) the best phase to promote plastic changes in the CNS. Functional Improvements such as the increase of gait speed increased recovered distance or better WISCI II scores have been
The application of simple diagnostic measures that might provide comprehensive information regarding the state of adaptive and maladaptive motor control mechanisms after SCI could play a crucial role in guiding rehabilitation strategies in the clinic. Neurophysiological measures should be preferred over qualitative clinical measures, as they are more independ‐ ent, provide objective data and can be performed in less cooperative patients [63]. However, standardization and clinical validation should be achieved to allow for wide use in the clinical
Recovery after SCI is greatly dependent on the severity of the injury as well as the treatment provided in each phase of the lesion, being the subacute phase (up to approximately 6 months) the best phase to promote plastic changes in the CNS. Functional Improvements such as the increase of gait speed increased recovered distance or better WISCI II scores have been
The application of simple diagnostic measures that might provide comprehensive informa‐ tion regarding the state of adaptive and maladaptive motor control mechanisms after SCI could play a crucial role in guiding rehabilitation strategies in the clinic. Neurophysiological measures should be preferred over qualitative clinical measures, as they are more independ‐ ent, provide objective data and can be performed in less cooperative patients [63]. However,
The analysis of muscle synergies during cycling can be explored as a novel approach for the quantitative assessment of gait performance. This analysis can complement current assessment
observed in acute SCI patients if compared to chronic SCI patients [62].
observed in acute SCI patients if compared to chronic SCI patients [62].
[60]. As a consequence of lower signal‐to‐noise, lower VAFtotal, and *r*<sup>2</sup>
) when recording a set of 13 muscles of the most affected leg. In the case
) when recording a set of 13 muscles of the most affected leg. In the case
and also needed less time to perform these two gait performance tests.
also needed less time to perform these two gait performance tests.
) for patient 01 (C) and patient 02 (D), according to the number of synergies, for each of the four speeds (30, 42, 50,
values, a minimum of 0.49 (**Figure 1C**) and 0.52 (**Figure 1D**) were obtained for
values of 0.82 and 0.89 were obtained for patient
and
values are expected [5].
(*r*2
and 60 rpm).
and *r*<sup>2</sup>
*5.2.1.5. Discussion*
this hypothesis.
*5.2.1.5. Discussion*
rate this hypothesis.
scores (VAFtotal and *r*<sup>2</sup>
scores (VAFtotal and *r*<sup>2</sup>
expected [5].
setting.
**6. Conclusions**
**6. Conclusions**
In the case of *r*<sup>2</sup>
316 Recovery of Motor Function Following Spinal Cord Injury
31412in Spinal Cord Injury
two synergies. On the other hand, maximum *r*<sup>2</sup>
Filipe Barroso1,2, Diego Torricelli2 and Juan C. Moreno2\*
Smart Wearable Robots with Bioinspired Sensory-Motor Skills).
**Author details** \*Address all correspondence to: jc.moreno@csic.es
Filipe Barroso1,2, Diego Torricelli2 and Juan C. Moreno2\* 1 Department of Physiology, Feinberg School of Medicine – Northwestern University, Chicago, IL, USA
1 Department of Physiology, Feinberg School of Medicine – Northwestern University,
\*Address all correspondence to: jc.moreno@csic.es 2 Neural Rehabilitation Group, Cajal Institute, Spanish National Research Council (CSIC), Madrid, Spain
#### 2 Neural Rehabilitation Group, Cajal Institute, Spanish National Research Council (CSIC), **References**
Chicago, IL, USA
- Rehabil, v. 17, n. 1, p. 16‐24, 2011. [2] NADEAU, S. et al. Guiding task‐oriented gait training after stroke or spinal cord injury [3] BARROSO, F. O. et al. Shared muscle synergies in human walking and cycling. Journal of Neurophysiology, v. 112, n. 8, p. 1984‐1998, 2014.
- by means of a biomechanical gait analysis. Progress in Brain Research, v. 192, p. 161‐ 180, 2011. [4] ROUTSON, R. L. et al. The influence of locomotor rehabilitation on module quality and post‐stroke hemiparetic walking performance. Gait Posture, v. 38, n. 3, p. 511‐517, 2013.
[17] HAMILTON, B. B.; GRANGER, C. V. Disability outcomes following inpatient rehabil‐ itation for stroke. Physical Therapy, v. 74, n. 5, p. 494‐503, 1994. [18] KIRSHBLUM, S. et al. Late neurologic recovery aftertraumatic spinal cord injury. Arch [19] CATZ, A. et al. A multicenter international study on the spinal cord independence measure, version III: Rasch psychometric validation. Spinal Cord, v. 4, n. 275‐291, p. 45, 2007.
[3] BARROSO, F. O. et al. Shared muscle synergies in human walking and cycling. Journal
[5] STEELE, K. M.; TRESCH, M.; PERREAULT, E. The number and choice of muscles impact the results of muscle synergy analyses. Front Comput Neurosci, v. 7, 2013.
[4] ROUTSON, R. L. et al. The influence of locomotorrehabilitation on module quality and post‐stroke hemiparetic walking performance. Gait Posture, v. 38, n. 3, p. 511‐517, 2013.
[6] DUFFELL, L. D.; BROWN, G. L.; MIRBAGHERI, M. M. Facilitatory effects of anti‐ spastic medication on robotic locomotor training in people with chronic incomplete
[5] STEELE, K. M.; TRESCH, M.; PERREAULT, E. The number and choice of muscles impact the results of muscle synergy analyses. Front Comput Neurosci, v. 7, 2013.
[7] REICHENFELSER, W. et al. Monitoring of spasticity and functional ability in individ‐ uals with incomplete spinal cord injury with a functional electrical stimulation cycling
[6] DUFFELL, L. D.; BROWN, G. L.; MIRBAGHERI, M. M. Facilitatory effects of anti‐ spastic medication on robotic locomotor training in people with chronic incomplete
[7] REICHENFELSER, W. et al. Monitoring of spasticity and functional ability in individ‐ uals with incomplete spinal cord injury with a functional electrical stimulation cycling
[8] SAVIC, G. et al. Inter‐rater reliability of motor and sensory examinations performed according to American Spinal Injury Association standards. Spinal Cord, v. 45, p. 444‐
[8] SAVIC, G. et al. Inter‐rater reliability of motor and sensory examinations performed according to American Spinal Injury Association standards. Spinal Cord, v. 45, p. 444‐
[9] MAYNARD, F. M. J. et al. International standards for neurological and functional classification of spinal cord injury. American Spinal Injury Association. Spinal Cord, v.
[9] MAYNARD, F. M. J. et al. International standards for neurological and functional classification of spinal cord injury. American Spinal Injury Association. Spinal Cord,
[11] SAENSOOK, W. et al. Discriminative ability of the three functional tests in independent ambulatory patients with spinal cord injury who walked with and without ambulatory
[10] DITTUNO, P. L.; DITUNNO, J. F. Walking index for spinal cord injury (WISCI II): scale
[10] DITTUNO, P. L.; DITUNNO, J. F. Walking index for spinal cord injury (WISCI II): scale
[12] PODSIADLO, D.; RICHARDSON, S. The timed "Up & Go": a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc, v. 39, n. 2, p. 142‐148, 1991.
[11] SAENSOOK, W. et al. Discriminative ability of the three functional tests in independ‐ ent ambulatory patients with spinal cord injury who walked with and without
[13] FURLAN, J. C. et al. Assessment of disability in patients with acute traumatic spinalcord injury: a systematic review of the literature. J Neurotrauma, v. 28, p. 1413‐1430,
[12] PODSIADLO, D.; RICHARDSON, S. The timed "Up & Go": a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc, v. 39, n. 2, p. 142‐148, 1991.
[14] VAN HEDEL, H. J.; WIRZ, M.; DIETZ, V. Assessing walking ability in subjects with spinal cord injury: validity and reliability of 3 walking tests. Arch Phys Med Rehabil.,
[13] FURLAN, J. C. et al. Assessment of disability in patients with acute traumatic spinal cord injury: a systematic review of the literature. J Neurotrauma, v. 28, p. 1413‐1430,
[14] VAN HEDEL, H. J.; WIRZ, M.; DIETZ, V. Assessing walking ability in subjects with spinal cord injury: validity and reliability of 3 walking tests. Arch Phys Med Rehabil.,
[16] MAANUM, G. et al. Walking ability and predictors of performance on the 6‐minute walk test in adults with spastic cerebral palsy. Dev Med Child Neurol, v. 6, n. e126‐
[15] GUYATT, G. H. et al. The 6‐miunte walk: a new measure of exercise capacity in patients
[15] GUYATT, G. H. et al. The 6‐miunte walk: a new measure of exercise capacity in patients
[17] HAMILTON, B. B.; GRANGER, C. V. Disability outcomes following inpatient rehabil‐
[16] MAANUM, G. et al. Walking ability and predictors of performance on the 6‐minute walk test in adults with spastic cerebral palsy. Dev Med Child Neurol, v. 6, n. e126‐
[18] KIRSHBLUM, S. et al. Late neurologic recovery after traumatic spinal cord injury. Arch
with chronic heart failure. Can Med Assoc J, v. 8, n. 919‐923, p. 132, 1985.
itation for stroke. Physical Therapy, v. 74, n. 5, p. 494‐503, 1994.
Phys Med Rehabil, v. 85, p. 1811‐1817, 2004.
with chronic heart failure. Can Med Assoc J, v. 8, n. 919‐923, p. 132, 1985.
ambulatory assistive devices. J Spinal Cord Med, v. 37, n. 2, p. 212‐217, 2014.
of Neurophysiology, v. 112, n. 8, p. 1984‐1998, 2014.
spinal cord injury. J Neuroeng Rehabil, v. 12, n. 29.
spinal cord injury. J Neuroeng Rehabil, v. 12, n. 29.
system. J Rehabil Med, v. 44, n. 5, p. 444‐449, 2012.
system. J Rehabil Med, v. 44, n. 5, p. 444‐449, 2012.
revision. Spinal Cord, v. 39, n. 12, p. 654‐656, 2001.
revision. Spinal Cord, v. 39, n. 12, p. 654‐656, 2001.
assistive devices. J Spinal Cord Med, v. 37, n. 2, p. 212‐217, 2014.
451, 2007.
451, 2007.
2011.
2011.
v. 35, n. 5, p. 266‐274, 1997.
35, n. 5, p. 266‐274, 1997.
318 Recovery of Motor Function Following Spinal Cord Injury
31614in Spinal Cord Injury
v. 86, n. 2, p. 190‐196, 2005.
v. 86, n. 2, p. 190‐196, 2005.
e132, p. 52, 2010.
e132, p. 52, 2010.
[31] BOAKYE, M. et al. Quantitative testing in spinal cord injury: overview of reliability
and predictive validity. J Neurosurg Spine., v. 17, n. 1, p. 141‐150, 2012.
[46] WIRTH, B.; VAN HEDEL, H. J.; CURT, A. Changes in corticospinal function and ankle motor control during recovery from incomplete spinal cord injury. J Neurotrauma, v. 25, p. 467–478, 2008. [47] GOMEZ‐SORIANO, J. et al. Voluntary ankle flexor activity and adaptive coactivation gain is decreased by spasticity during subacute spinal cord injury. Exp Neurol, v. 224,p. 507–516, 2010.
[32] PETERSEN, J. A. et al. Spinal cord injury: one‐year evolution of motor‐evoked potentials and recovery of leg motor function in 255 patients. Neurorehabil Neural
[33] CURT, A.; KECK, M. E.; DIETZ, V. Functional outcome following spinal cord injury:
significance of motor‐evoked potentials and ASIA scores. Arch Phys Med Rehabil, v.79, p. 81‐86, 1998.
[33] CURT, A.; KECK, M. E.; DIETZ, V. Functional outcome following spinal cord injury: significance of motor‐evoked potentials and ASIA scores. Arch Phys Med Rehabil, v.
[34] YANG, J. F. et al. Volitional muscle strength in the legs predicts changes in walking speed following locomotor training in people with chronic Spinal Cord Injury. Phys
[34] YANG, J. F. et al. Volitional muscle strength in the legs predicts changes in walking speed following locomotor training in people with chronic Spinal Cord Injury. Phys
[35] HANSEN, N. L. et al. Reduction of common synaptic drive to ankle dorsiflexormoto‐ neurons during walking in patients with spinal cord lesion. J Neurophysiol, v. 94, p.
[35] HANSEN, N. L. et al. Reduction of common synaptic drive to ankle dorsiflexormoto‐ neurons during walking in patients with spinal cord lesion. J Neurophysiol, v. 94, p.
[36] Boonstra, T. W. (2013). The potential of corticomuscular and intermuscular coherence for research on human motor control. Frontiers in Human Neuroscience, 7, 855. http://
[36] Boonstra, T. W. (2013). The potential of corticomuscular and intermuscular coherence for research on human motor control. Frontiers in Human Neuroscience, 7, 855. http://
[37] HALLIDAY, D. M. et al. Functional coupling of motor units is modulated during
[37] HALLIDAY, D. M. et al. Functional coupling of motor units is modulated during
[38] BARTHELEMY, D. et al. Impaired transmission in the corticospinal tract and gait disability in spinal cord injured persons. J Neurophysiol, v. 104, p. 1167–1176, 2010.
[38] BARTHELEMY, D. et al. Impaired transmission in the corticospinal tract and gait disability in spinal cord injured persons. J Neurophysiol, v. 104, p. 1167–1176, 2010.
[39] HANSEN, N. L. et al. Synchronization of lower limb motor unit activity during walking
[39] HANSEN, N. L. et al. Synchronization oflowerlimb motor unit activity during walking
coherence accompanying improvements in locomotor skills in incomplete spinal cordinjury. J Neurophysiol, v. 95, p. 2580–2589, 2006.
[40] NORTON, J. A.; GORASSINI, M. A. Changes in cortically related intermuscular
[40] NORTON, J. A.; GORASSINI, M. A. Changes in cortically related intermuscular coherence accompanying improvements in locomotor skills in incomplete spinal cord
[41] SMITH, H. C. et al. Motor unit discharge characteristics during voluntary contraction in patients with incomplete spinal cord injury. Exp Physiol, v. 84, p. 1151–1160, 1999.
[41] SMITH, H. C. et al. Motor unit discharge characteristics during voluntary contraction in patients with incomplete spinal cord injury. Exp Physiol, v. 84, p. 1151–1160, 1999.
[42] CROS, D.; SOTO, O.; CHIAPPA, K. H. Transcranial magnetic stimulation during voluntary action: directional facilitation of outputs and relationships to force genera‐
[42] CROS, D.; SOTO, O.; CHIAPPA, K. H. Transcranial magnetic stimulation during voluntary action: directional facilitation of outputs and relationships to force genera‐
[43] ZOGHI, M.; NORDSTROM, M. A. Progressive suppression of intracortical inhibitionduring graded isometric contraction of a hand muscle is not influenced by hand
[43] ZOGHI, M.; NORDSTROM, M. A. Progressive suppression of intracortical inhibition during graded isometric contraction of a hand muscle is not influenced by hand
[44] WEIER, A. T.; PEARCE, A. J.; KIDGELL, D. J. Strength training reduces intracortical
[44] WEIER, A. T.; PEARCE, A. J.; KIDGELL, D. J. Strength training reduces intracortical
[45] BROUWER, B.; ASHBY, P. Corticospinal projections to lower limb motoneurons in
[45] BROUWER, B.; ASHBY, P. Corticospinal projections to lower limb motoneurons in
[46] WIRTH, B.; VAN HEDEL, H. J.; CURT, A. Changes in corticospinal function and ankle motor control during recovery from incomplete spinal cord injury. J Neurotrauma, v.
walking in human subjects. J Neurophysiol, v. 89, p. 960–968, 2003.
walking in human subjects. J Neurophysiol, v. 89, p. 960–968, 2003.
in human subjects. J Neurophysiol, v. 86, p. 1266–1276, 2001.
in human subjects. J Neurophysiol, v. 86, p. 1266–1276, 2001.
injury. J Neurophysiol, v. 95, p. 2580–2589, 2006.
preference. Exp Brain Res, v. 177, n. 2, p. 266–274, 2007.
preference. Exp Brain Res, v. 177, n. 2, p. 266–274, 2007.
inhibition. Acta Physiol (Oxf), v. 206, n. 2, p. 109–119, 2012.
inhibition. Acta Physiol (Oxf), v. 206, n. 2, p. 109–119, 2012.
tion. Brain Res, v. 1185, p. 103–116, 2007.
tion. Brain Res, v. 1185, p. 103–116, 2007.
man. Exp Brain Res, v. 89, p. 649–654, 1992.
25, p. 467–478, 2008.
man. Exp Brain Res, v. 89, p. 649–654, 1992.
Repair, v. 26, n. 8, p. 939‐948, 2012.
320 Recovery of Motor Function Following Spinal Cord Injury
31816in Spinal Cord Injury
Ther, v. 91, n. 6, p. 931‐943, 2011.
Ther, v. 91, n. 6, p. 931‐943, 2011.
doi.org/10.3389/fnhum.2013.00855
doi.org/10.3389/fnhum.2013.00855
79, p. 81‐86, 1998.
934‐942, 2005.
934‐942, 2005.
Roessingh Research and Development, 1999.
[58] HUG, F. et al. Is interindividual variability of EMG patterns in trained cyclists related to different muscle synergies? J Appl Physiol, v. 108, n. 6, p. 1727‐1736, 2010. [59] LEE, D. D.; SEUNG, S. Learning the parts of objects by non‐negative matrix factoriza‐ tion. Nature, v. 401, p. 788‐791, 1999.
[60] TORRES‐OVIEDO, G.; MACPHERSON, J. M.; TING, L. H. Muscle synergy organiza‐ tion is robust across a variety of postural perturbations. J Neurophysiol, v. 96, n. 3, p.
[61] DOBKIN, B. et al. Weight‐supported treadmill vs over‐ground training for walking
[62] XIE, J.; BOAKYE, M. Electrophysiological outcomes after spinal cord injury. Neurosurg
[63] YANG, J. F.; MUSSELMAN, K. E. Training to achieve over ground walking after spinal cord injury: a review of who, what, when, and how. J Spinal Cord Med, v. 35, n. 5, p.
after acute incomplete SCI. Neurology, v. 66, n. 4, p. 484–493, 2006.
1530‐1546, 2006.
32018 Recovery of Motor Function in Spinal Cord Injury
293‐304, 2012.
Focus, v. 25, n. 5, p. E11, 2008.
## *Edited by Heidi Fuller and Monte Gates*
Restoration of motor function following spinal cord injury is a complex and challenging task. By reviewing emerging cellular, pharmacological, rehabilitative, as well as surgical approaches, this book seeks to highlight promising therapeutic strategies for the repair and regeneration of motor circuitry. The multidisciplinary nature of these approaches illustrates various routes to bridging the gap between the bench and the bedside and to identify the challenges that must be overcome in order to bring about a viable therapeutic strategy for spinal cord injury patients.
Recovery of Motor Function Following Spinal Cord Injury
Recovery of Motor Function
Following Spinal Cord Injury
*Edited by Heidi Fuller and Monte Gates*
Photo by toeytoey2530 / iStock | doab | 2025-04-07T04:13:03.442930 | 20-4-2021 17:56 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffa9f88a-5af3-498d-8440-85875bd29e9d",
"url": "https://mts.intechopen.com/storage/books/5203/authors_book/authors_book.pdf",
"author": "",
"title": "Recovery of Motor Function Following Spinal Cord Injury",
"publisher": "IntechOpen",
"isbn": "9789535124986",
"section_idx": 4
} |
ffbd80dd-28a6-46cd-af21-a432f1dad4e5.0 | *Edited by Abdellah El Maghraoui*
The World Health Organization (WHO) has established dual-energy x-ray absorptiometry (DXA) as the best densitometric technique for assessing bone mineral density (BMD) in postmenopausal women and has based the definitions of osteopenia and osteoporosis on its results. DXA enables accurate diagnosis of osteoporosis, estimation of fracture risk and monitoring of patients undergoing treatment. Additional features of DXA include measurement of BMD at multiple skeletal sites, vertebral fracture assessment and body composition assessment, including fat mass and lean soft tissue mass of the whole body and the segments. This book contains reviews and original studies about DXA and its different uses in clinical practice (diagnosis of osteoporosis, monitoring of BMD measurement) and in medical research in several situations (e.g. assessment of morphological asymmetry in athletes, estimation of resting energy expenditure, assessment of vertebral strength and vertebral fracture risk, or study of dry bones such as the ulna).
ISBN 978-953-307-877-9
Dual Energy X-Ray Absorptiometry
Photo by Rost-9D / iStock
## Dual Energy X-Ray Absorptiometry
*Edited by Abdellah El Maghraoui*
**DUAL ENERGY X-RAY**
**ABSORPTIOMETRY**
Edited by **Abdellah El Maghraoui**
**INTECHOPEN.COM**
## **DUAL ENERGY X-RAY ABSORPTIOMETRY**
Edited by **Abdellah El Maghraoui**
**INTECHOPEN.COM**
http://dx.doi.org/10.5772/1114 Edited by Abdellah El Maghraoui
#### **Contributors**
Soledad Aguado-Henche, Rosa Rodriguez\_Torres, Asunción Bosch-Martín, Spottorno-Rubio Pia, Yannis Dionyssiotis, Abdellah El Maghraoui, Joonas Sirola, Toni Rikkonen, Risto Honkanen, Jukka Jurvelin, Marjo Tuppurainen, Heikki Kröger, Kazuhiro Imai, Anne Blais, Magdalena Krzykała, Chiyoko Usui, Motoko Taguchi, Kazuko Ishikawa-Takata, Mitsuru Higuchi
#### **© The Editor(s) and the Author(s) 2012**
The moral rights of the and the author(s) have been asserted.
All rights to the book as a whole are reserved by INTECH. The book as a whole (compilation) cannot be reproduced, distributed or used for commercial or non-commercial purposes without INTECH's written permission. Enquiries concerning the use of the book should be directed to INTECH rights and permissions department (permissions@intechopen.com).
Violations are liable to prosecution under the governing Copyright Law.
Individual chapters of this publication are distributed under the terms of the Creative Commons Attribution 3.0 Unported License which permits commercial use, distribution and reproduction of the individual chapters, provided the original author(s) and source publication are appropriately acknowledged. If so indicated, certain images may not be included under the Creative Commons license. In such cases users will need to obtain permission from the license holder to reproduce the material. More details and guidelines concerning content reuse and adaptation can be foundat http://www.intechopen.com/copyright-policy.html.
#### **Notice**
Statements and opinions expressed in the chapters are these of the individual contributors and not necessarily those of the editors or publisher. No responsibility is accepted for the accuracy of information contained in the published chapters. The publisher assumes no responsibility for any damage or injury to persons or property arising out of the use of any materials, instructions, methods or ideas contained in the book.
First published in Croatia, 2012 by INTECH d.o.o. eBook (PDF) Published by IN TECH d.o.o. Place and year of publication of eBook (PDF): Rijeka, 2019. IntechOpen is the global imprint of IN TECH d.o.o. Printed in Croatia
Legal deposit, Croatia: National and University Library in Zagreb
Additional hard and PDF copies can be obtained from orders@intechopen.com
Dual Energy X-Ray Absorptiometry Edited by Abdellah El Maghraoui p. cm. ISBN 978-953-307-877-9 eBook (PDF) ISBN 978-953-51-6742-6
| doab | 2025-04-07T04:13:03.511092 | 20-4-2021 17:18 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffbd80dd-28a6-46cd-af21-a432f1dad4e5",
"url": "https://mts.intechopen.com/storage/books/602/authors_book/authors_book.pdf",
"author": "",
"title": "Dual Energy X-Ray Absorptiometry",
"publisher": "IntechOpen",
"isbn": "9789533078779",
"section_idx": 0
} |
ffbd80dd-28a6-46cd-af21-a432f1dad4e5.1 | We are IntechOpen, the world's leading publisher of Open Access books Built by scientists, for scientists
4,100+
Open access books available
116,000+
International authors and editors
120M+
Downloads
Our authors are among the
Top 1%
most cited scientists
12.2%
Contributors from top 500 universities
Selection of our books indexed in the Book Citation Index in Web of Science™ Core Collection (BKCI)
### Interested in publishing with us? Contact book.department@intechopen.com
Numbers displayed above are based on latest data collected. For more information visit www.intechopen.com
## **Meet the editor**
Prof. El Maghraoui was born in 1969, and is currently a Professor of Rheumatology at the Rabat Medicine and Pharmacy University. He is also the head of the Rheumatology Department, Military Hospital Mohammed V, Rabat, Morocco, and the vice president of the Moroccan Society of Rheumatology. He graduated from the Rabat Medicine and Pharmacy University, and received di-
plomas in Podology from the Paris V University, in Muscular Ultrasound from the Paris VI University, and in Methodology of Clinical Research from the Bordeaux University. Prof. El Maghraoui has authored more than 90 PubMed indexed articles, and has published a book titled L'ostéoporose in 2010. He is also the founder, designer and Webmaster of www.rhumato. info, a website for public health professionals, dedicated to rheumatology.
Contents
**Preface IX**
**Part 1 Bone Mineral Density 1**
Abdellah El Maghraoui
Abdellah El Maghraoui
**Part 2 Body Composition 43**
Chapter 1 **Interpreting a DXA Scan in Clinical Practice 3**
Chapter 3 **Bone Loss and Seasonal Variation in Serial DXA**
Chapter 4 **The Validity of Body Composition Measurement Using Dual Energy X-Ray Absorptiometry for Estimating Resting Energy Expenditure 45**
Kazuko Ishikawa-Takata and Mitsuru Higuchi
Chapter 7 **Internal Design of the Dry Human Ulna by DXA 97** S. Aguado-Henche, A. Bosch-Martín, P. Spottorno-Rubio and R. Rodríguez-Torres
**of Morphological Asymmetry in Athletes 59**
Chiyoko Usui, Motoko Taguchi,
**of Central Nervous System 75**
Chapter 5 **Dxa as a Tool for the Assessment**
Magdalena Krzykała
Chapter 6 **Body Composition in Disabilities**
Yannis Dionyssiotis
**Part 3 Miscellaneous 95**
Chapter 2 **Monitoring DXA Measurement in Clinical Practice 19**
**Densitometry – A Population-Based Study 29** Joonas Sirola, Toni Rikkonen, Risto Honkanen,
Jukka S. Jurvelin, Marjo Tuppurainen and Heikki Kröger
### Contents
#### **Preface XI**
#### **Part 2 Body Composition 43**
#### **Part 3 Miscellaneous 95**
- P. Spottorno-Rubio and R. Rodríguez-Torres
Chapter 9 **Dietary Protein and Bone Health 121** Anne Blais, Emilien Rouy and Daniel Tomé
### Preface
Dual-energy x-ray absorptiometry (DXA) is now recognized as the reference method to measure bone mineral density (BMD) with acceptable accuracy errors, good precision and reproducibility. The World Health Organization (WHO) has established DXA as the best densitometric technique for assessing BMD in postmenopausal women and has based the definitions of osteopenia and osteoporosis on its results. DXA allows accurate diagnosis of osteoporosis, estimation of fracture risk and monitoring of patients undergoing treatment. Correct performance of BMD measurements using DXA requires rigorous attention to detail in positioning and analysis. Performing DXA studies incorrectly can lead to major mistakes in diagnosis and therapy. Measurement error must be considered when evaluating serial assessments. A clear understanding of the interpretation of serial measurements and the statistical principles impacting upon their interpretation is necessary to determine whether a change is real and not simply random fluctuation. Moreover, it is important to keep in mind that fracture-protection benefit may be realized before BMD gains are detected. Physicians interested in osteoporosis management, even if not directly involved in the performance and interpretation of DXA, should be familiar with the principles of DXA to minimize serious errors and allow proper use of bone densitometry.
Additional features of DXA include measurement of BMD at multiple skeletal sites, vertebral fracture assessment (also called vertebral morphometry) using evaluation of vertebral heights with a special software to determine vertebral body dimensions, body composition assessment including bone mineral content, fat mass and lean soft tissue mass of the whole body and the segments.
This book contains reviews and original studies about DXA and many different uses in clinical practice (diagnosis of osteoporosis, monitoring of BMD measurement) and in medical research in several situations (e.g. assessment of morphological asymmetry in athletes, estimation of resting energy expenditure, consequences of dietary protein or disabilities of the central nervous system on BMD and body composition, assessment of vertebral strength and vertebral fracture risk, or study of dry bones such as the ulna). It will certainly help the researchers and/or physicians
#### X Preface
using DXA to improve their knowledge of the wide range of potential uses that this technique allows.
> **Prof. A. El Maghraoui** Rheumatology Department, Military Hospital Mohammed V, Rabat, Morocco
X Preface
technique allows.
using DXA to improve their knowledge of the wide range of potential uses that this
**Prof. A. El Maghraoui** Rheumatology Department, Military Hospital Mohammed V,
> Rabat, Morocco
**Part 1**
**Bone Mineral Density**
**Part 1**
**Bone Mineral Density**
**1**
*Morocco*
Abdellah El Maghraoui
**Interpreting a DXA Scan in Clinical Practice**
Osteoporosis is a metabolic bone disorder characterized by low bone mass and microarchitectural deterioration, with a subsequent increase in bone fragility and susceptibility to fracture. Dual-energy x-ray absorptiometry (DXA) is recognized as the reference method to measure bone mineral density (BMD) with acceptable accuracy errors and good precision and reproducibility(Blake and Fogelman 2007). The World Health Organization (WHO) has established DXA as the best densitometric technique for assessing BMD in postmenopausal women and based the definitions of osteopenia and osteoporosis on its results (table 1)(Kanis 1994; Kanis, Borgstrom et al. 2005). DXA allows accurate diagnosis of osteoporosis, estimation of fracture risk, and monitoring of patients undergoing treatment. Additional features of DXA include measurement of BMD at multiple skeletal sites, safety of performance, short investigation time, and ease of use(Hans, Downs et al. 2006; Lewiecki, Binkley et al. 2006). A DXA measurement can be completed in about 5 minutes with minimal radiation exposure (about one tenth that of a standard chest x-ray for
> *Diagnosis T-score* Normal >–1.0 Osteopenia <–1.0, >–2.5 Osteoporosis <–2.5
Severe osteoporosis <–2.5 plus fragility fractures
As with many other diagnostic examinations, DXA scans should be critically assessed by the interpreting physician and densitometrist for abnormalities that may affect BMD measurements. In clinical practice, recognition of diverse artifacts and disease processes that may influence BMD results can be of major importance in the optimal interpretation of DXA scans(Roux 1998). Physicians not directly involved in the performance and interpretation of DXA should be familiar enough to detect common positioning and scanning problems, to know what should appear on a report, what questions to ask if the necessary information is not on the report, how to apply the results in patient management, and when to do and how
to interpret a second measurement to monitor treatment(Watts 2004).
**1. Introduction**
a quick hips and spine exam).
Table 1. WHO Osteoporosis Classification
**2. Principle of DXA scanning**
*Rheumatology Department, Military Hospital Mohammed V, Rabat,*
### **Interpreting a DXA Scan in Clinical Practice**
Abdellah El Maghraoui
*Rheumatology Department, Military Hospital Mohammed V, Rabat, Morocco*
#### **1. Introduction**
Osteoporosis is a metabolic bone disorder characterized by low bone mass and microarchitectural deterioration, with a subsequent increase in bone fragility and susceptibility to fracture. Dual-energy x-ray absorptiometry (DXA) is recognized as the reference method to measure bone mineral density (BMD) with acceptable accuracy errors and good precision and reproducibility(Blake and Fogelman 2007). The World Health Organization (WHO) has established DXA as the best densitometric technique for assessing BMD in postmenopausal women and based the definitions of osteopenia and osteoporosis on its results (table 1)(Kanis 1994; Kanis, Borgstrom et al. 2005). DXA allows accurate diagnosis of osteoporosis, estimation of fracture risk, and monitoring of patients undergoing treatment. Additional features of DXA include measurement of BMD at multiple skeletal sites, safety of performance, short investigation time, and ease of use(Hans, Downs et al. 2006; Lewiecki, Binkley et al. 2006). A DXA measurement can be completed in about 5 minutes with minimal radiation exposure (about one tenth that of a standard chest x-ray for a quick hips and spine exam).
Table 1. WHO Osteoporosis Classification
#### **2. Principle of DXA scanning**
As with many other diagnostic examinations, DXA scans should be critically assessed by the interpreting physician and densitometrist for abnormalities that may affect BMD measurements. In clinical practice, recognition of diverse artifacts and disease processes that may influence BMD results can be of major importance in the optimal interpretation of DXA scans(Roux 1998). Physicians not directly involved in the performance and interpretation of DXA should be familiar enough to detect common positioning and scanning problems, to know what should appear on a report, what questions to ask if the necessary information is not on the report, how to apply the results in patient management, and when to do and how to interpret a second measurement to monitor treatment(Watts 2004).
Interpreting a DXA Scan in Clinical Practice 5
worldwide. Thus, the WHO diagnostic criteria for osteoporosis define osteoporosis in terms of
The T-score is calculated using the formula: (patient's BMD - young normal mean)/SD of young normal. For example, if a patient has a BMD of 0.700 g/cm2, the young normal mean is 1.000 g/cm2, and the young normal standard deviation is 0.100 g/cm2, then this patient's T-score would be (0.700 - 1.000)/0.100, or –0.300/0.100, or –3.0(Watts 2004). A T-score of 0 is equal to the young normal mean value, -1.0 is 1 SD low, -2.0 is 2 SD low, etc. Although the WHO classification was not intended to be applied to individual patients, it works well to define ''normal'' (T-score –1.0 and above) and ''osteoporosis'' (T-score –2.5 and below). Several large studies have shown an unacceptably high risk of fracture in post-menopausal women who have T-scores of –2.5 and below. Thus, this threshold is the cornerstone of the patient's assessment. For the therapeutic decisions, however, other risk factors are
In addition to the T-scores, DXA reports also provide Z-scores, which are calculated similarly to the T-score, except that the patient's BMD is compared with an age-matched (and race- and gender-matched) mean, and the result expressed as a standard deviation score(Watts 2004). In premenopausal women, a low Z-score (below -2.0) indicates that bone
Most official groups recommend screening healthy women for osteoporosis at age 65, and testing higher-risk women earlier(Baddoura, Awada et al. 2006). In Europe the recommendations are to screen for risk factors of osteoporosis and to perform BMD measurement in women with such risks. The International Society for Clinical Densitometry (ISCD) recommends screening men without risk factors for osteoporosis at age 70, and screening higher-risk men earlier. Risk factors include dementia, poor health, recent falls, prolonged immobilization, smoking, alcohol abuse, low body weight, history of fragility fracture in a first-degree relative, estrogen deficiency at an early age (<45 years), and steroid use for more than 3 months. Of course, BMD testing is an appropriate tool in the evaluation of patients who have diseases (e.g. hyperthyroidism, hyperparathyroidism, celiac disease, etc.) or use medications (e.g. glucocorticoids, GnRH agonists, aromatase inhibitors etc.) that might cause bone loss. Another indication is radiographic evidence of ''osteopenia'' or a
Recently, many epidemiological studies have validated risk assessment indices for osteoporosis in women. The purpose of the risk assessment indices is not to diagnose osteoporosis or low BMD, but to identify women who are more likely to have low BMD (Hillier, Stone et al. 2007). Such indices, while not identifying all cases of osteoporosis, increase the efficiency of BMD measurement by focusing on subjects who are at increased risk (Cadarette, Jaglal et al. 2000; Gnudi and Sitta 2005; Salaffi, Silveri et al. 2005). The easiest to use in clinical practice is certainly the Osteoporosis Self-assessment Tool (OST). The calculated risk index is based on self-reported age and weight: [(weight in kilograms – age in years) × 0.2, truncated to an integer]. It was developed and validated in several studies in Asian and White women(Richy, Ethgen et al. 2004; El Maghraoui, Guerboub et al. 2007; El Maghraoui, Habbassi
et al. 2007) and men (Adler, Tran et al. 2003; Ghazi, Mounach et al. 2007).
density is lower than expected and should trigger a search for an underlying cause.
a T-score below −2.5 and osteopenia when T-score is between -2.5 and -1.
considered such as prevalent fractures, age and low body mass index.
**3. Who should have a DXA measurement?**
vertebral fracture).
Several different types of DXA systems are available, but they all operate on similar principles. A radiation source is aimed at a radiation detector placed directly opposite the site to be measured. The patient is placed on a table in the path of the radiation beam. The source/detector assembly is then scanned across the measurement region. The attenuation of the radiation beam is determined and is related to the BMD (Blake and Fogelman 2002; Blake and Fogelman 2003).
Because DXA scanners use two X-ray energies in the presence of three types of tissue (bone mineral, lean tissue and adipose tissue), there are considerable errors arising from the inhomogeneous distribution of adipose tissue in the human body(Tothill and Avenell 1994) (which can be studied either through cadaver studies(Svendsen, Hassager et al. 1995), CT imaging to delineate the distribution of adipose tissue external to bone(Kuiper, van Kuijk et al. 1996; Lee, Wren et al. 2007) or MRI to measure the percentage of marrow fat inside bone(Griffith, Yeung et al. 2006)). These studies suggest BMD measurement errors of around 5 to 8%.
DXA technology can measure virtually any skeletal site, but clinical use has been concentrated on the lumbar spine, proximal femur, forearm, and total body (Hans, Downs et al. 2006). DXA systems are available as either full table systems (capable of multiple skeletal measurements, including the spine and hip) or as peripheral systems (limited to measuring the peripheral skeleton). Because of their versatility, and the ability to measure the skeletal sites of greatest clinical interest, full table DXA systems are the current clinical choice for osteoporosis assessment. Peripheral DXA systems, portable and less expensive than full table systems, are more frequently used as screening and early risk assessment tools; they cannot be used for treatments follow-up. Spine and proximal femur scans represent the majority of the clinical measurements performed using DXA. Most full table DXA systems are able to perform additional scans, including lateral spine BMD measurements, body composition study, assessment of vertebral fractures, measurements of children and infants, assessment of bone around prosthetic implants, small-animal studies and measurements of excised bone specimens. However, for children measurement, the exam should be undertaken by clinicians skilled in interpretation of scans in children in centers that have an adapted paediatric software.
Early DXA systems used a pencil beam geometry and a single detector, which was scanned across the measurement region. Modern full table DXA scanners use a fan-beam source and multiple detectors, which are swept across the measurement region. Fan beam provides the advantage of decreased scan times compared to single-beam systems, but these machines typically cost more because of the need for multiple X-ray detectors. Fan-beam systems use either a single-view or multiview mode to image the skeleton (Lewiecki and Borges 2006).
In clinical practice, BMD measurements are widely used to diagnose osteoporosis and measurement in bone mass are commonly used as a surrogate for fracture risk (Price, Walters et al. 2003). BMD is the measured parameter, and allows the calculation of the bone mineral content (BMC) in grams and the two-dimensional projected area in cm2 of the bone(s) being measured; thus the units of BMD are g/cm2. The BMD values (in g/cm2) are not used for diagnosing osteoporosis. Instead, a working group of the WHO proposed to define osteoporosis on the basis of the T-score (which is the difference between the measured BMD and the mean value of young adults, expressed in standard deviations (SD) for a normal population of the same gender and ethnicity)(Watts 2004). Despite its limitations; this definition, which concerns only postmenopausal women and men over 50, is currently applied
Several different types of DXA systems are available, but they all operate on similar principles. A radiation source is aimed at a radiation detector placed directly opposite the site to be measured. The patient is placed on a table in the path of the radiation beam. The source/detector assembly is then scanned across the measurement region. The attenuation of the radiation beam is determined and is related to the BMD (Blake and Fogelman 2002;
Because DXA scanners use two X-ray energies in the presence of three types of tissue (bone mineral, lean tissue and adipose tissue), there are considerable errors arising from the inhomogeneous distribution of adipose tissue in the human body(Tothill and Avenell 1994) (which can be studied either through cadaver studies(Svendsen, Hassager et al. 1995), CT imaging to delineate the distribution of adipose tissue external to bone(Kuiper, van Kuijk et al. 1996; Lee, Wren et al. 2007) or MRI to measure the percentage of marrow fat inside bone(Griffith, Yeung et al. 2006)). These studies suggest BMD measurement errors of around
DXA technology can measure virtually any skeletal site, but clinical use has been concentrated on the lumbar spine, proximal femur, forearm, and total body (Hans, Downs et al. 2006). DXA systems are available as either full table systems (capable of multiple skeletal measurements, including the spine and hip) or as peripheral systems (limited to measuring the peripheral skeleton). Because of their versatility, and the ability to measure the skeletal sites of greatest clinical interest, full table DXA systems are the current clinical choice for osteoporosis assessment. Peripheral DXA systems, portable and less expensive than full table systems, are more frequently used as screening and early risk assessment tools; they cannot be used for treatments follow-up. Spine and proximal femur scans represent the majority of the clinical measurements performed using DXA. Most full table DXA systems are able to perform additional scans, including lateral spine BMD measurements, body composition study, assessment of vertebral fractures, measurements of children and infants, assessment of bone around prosthetic implants, small-animal studies and measurements of excised bone specimens. However, for children measurement, the exam should be undertaken by clinicians skilled in interpretation of scans in children in centers that have an
Early DXA systems used a pencil beam geometry and a single detector, which was scanned across the measurement region. Modern full table DXA scanners use a fan-beam source and multiple detectors, which are swept across the measurement region. Fan beam provides the advantage of decreased scan times compared to single-beam systems, but these machines typically cost more because of the need for multiple X-ray detectors. Fan-beam systems use either a single-view or multiview mode to image the skeleton (Lewiecki and Borges 2006). In clinical practice, BMD measurements are widely used to diagnose osteoporosis and measurement in bone mass are commonly used as a surrogate for fracture risk (Price, Walters et al. 2003). BMD is the measured parameter, and allows the calculation of the bone mineral content (BMC) in grams and the two-dimensional projected area in cm2 of the bone(s) being measured; thus the units of BMD are g/cm2. The BMD values (in g/cm2) are not used for diagnosing osteoporosis. Instead, a working group of the WHO proposed to define osteoporosis on the basis of the T-score (which is the difference between the measured BMD and the mean value of young adults, expressed in standard deviations (SD) for a normal population of the same gender and ethnicity)(Watts 2004). Despite its limitations; this definition, which concerns only postmenopausal women and men over 50, is currently applied
Blake and Fogelman 2003).
adapted paediatric software.
5 to 8%.
worldwide. Thus, the WHO diagnostic criteria for osteoporosis define osteoporosis in terms of a T-score below −2.5 and osteopenia when T-score is between -2.5 and -1.
The T-score is calculated using the formula: (patient's BMD - young normal mean)/SD of young normal. For example, if a patient has a BMD of 0.700 g/cm2, the young normal mean is 1.000 g/cm2, and the young normal standard deviation is 0.100 g/cm2, then this patient's T-score would be (0.700 - 1.000)/0.100, or –0.300/0.100, or –3.0(Watts 2004). A T-score of 0 is equal to the young normal mean value, -1.0 is 1 SD low, -2.0 is 2 SD low, etc. Although the WHO classification was not intended to be applied to individual patients, it works well to define ''normal'' (T-score –1.0 and above) and ''osteoporosis'' (T-score –2.5 and below). Several large studies have shown an unacceptably high risk of fracture in post-menopausal women who have T-scores of –2.5 and below. Thus, this threshold is the cornerstone of the patient's assessment. For the therapeutic decisions, however, other risk factors are considered such as prevalent fractures, age and low body mass index.
In addition to the T-scores, DXA reports also provide Z-scores, which are calculated similarly to the T-score, except that the patient's BMD is compared with an age-matched (and race- and gender-matched) mean, and the result expressed as a standard deviation score(Watts 2004). In premenopausal women, a low Z-score (below -2.0) indicates that bone density is lower than expected and should trigger a search for an underlying cause.
#### **3. Who should have a DXA measurement?**
Most official groups recommend screening healthy women for osteoporosis at age 65, and testing higher-risk women earlier(Baddoura, Awada et al. 2006). In Europe the recommendations are to screen for risk factors of osteoporosis and to perform BMD measurement in women with such risks. The International Society for Clinical Densitometry (ISCD) recommends screening men without risk factors for osteoporosis at age 70, and screening higher-risk men earlier. Risk factors include dementia, poor health, recent falls, prolonged immobilization, smoking, alcohol abuse, low body weight, history of fragility fracture in a first-degree relative, estrogen deficiency at an early age (<45 years), and steroid use for more than 3 months. Of course, BMD testing is an appropriate tool in the evaluation of patients who have diseases (e.g. hyperthyroidism, hyperparathyroidism, celiac disease, etc.) or use medications (e.g. glucocorticoids, GnRH agonists, aromatase inhibitors etc.) that might cause bone loss. Another indication is radiographic evidence of ''osteopenia'' or a vertebral fracture).
Recently, many epidemiological studies have validated risk assessment indices for osteoporosis in women. The purpose of the risk assessment indices is not to diagnose osteoporosis or low BMD, but to identify women who are more likely to have low BMD (Hillier, Stone et al. 2007). Such indices, while not identifying all cases of osteoporosis, increase the efficiency of BMD measurement by focusing on subjects who are at increased risk (Cadarette, Jaglal et al. 2000; Gnudi and Sitta 2005; Salaffi, Silveri et al. 2005). The easiest to use in clinical practice is certainly the Osteoporosis Self-assessment Tool (OST). The calculated risk index is based on self-reported age and weight: [(weight in kilograms – age in years) × 0.2, truncated to an integer]. It was developed and validated in several studies in Asian and White women(Richy, Ethgen et al. 2004; El Maghraoui, Guerboub et al. 2007; El Maghraoui, Habbassi et al. 2007) and men (Adler, Tran et al. 2003; Ghazi, Mounach et al. 2007).
Interpreting a DXA Scan in Clinical Practice 7
by the use of positioning devices. Internal rotation may be improved by having the patient flex the foot before doing the internal rotation, and then relaxing the foot after the strap is in place. This amount of internal rotation presents the long axis of the femoral neck perpendicular to the X-ray beam, providing the greatest area and the lowest bone mineral content (and the lowest BMD), and is confirmed on the scan by seeing little or none of the lesser trochanter (Fig. 1b)(Lekamwasam and Lenora 2003; 2004). If the desired amount of internal rotation cannot be achieved, as is often the case in patients with hip arthritis or short femoral necks, the technologist should place the patient comfortably in a position that is likely to be reproducible
(a)
The software marks regions of interest in the spine and hip, but the technologist can and should make adjustments if needed. The spine region of interest consists of the L1 through L4 vertebrae (Fig. 1a). Correct placement of the top and bottom of the spine ''box'' is critical.
Fig. 1. Correct positioning and analysis of the L1–L4 spine (a) and the proximal femur
(b) (c)
(Lunar (b) and Hologic (c)).
**5.2 DXA scan analysis**
in a subsequent scan (Hamdy, Kiebzak et al. 2006; Lewiecki, Binkley et al. 2006).
#### **4. Site of measurement of BMD**
The ISCD recommends obtaining BMD measurements of the posteroanterior spine and hip(Leib, Binkley et al. 2006). The lateral spine and Ward's triangle region of the hip should not be used for diagnosis, because these sites overestimate osteoporosis and results can be false-positive. Evidence suggests that the femur (neck or total hip) is the optimum site for predicting the risk of hip fracture and the spine is the optimum site for monitoring response to treatment. Thus, many authors recommend hip measure alone for the fracture risk assessment(Kanis, Johnell et al. 2000; Kanis, Oden et al. 2001; Kanis 2002; Johnell, Kanis et al. 2005; Kanis, Seeman et al. 2005; Arabi, Baddoura et al. 2007). In very obese patients, those with primary hyperparathyroidism, or those in whom the hip or the spine, or both, cannot be measured or interpreted, BMD may be measured in the forearm, using a 33% radius on the nondominant forearm.
#### **5. Interpreting a DXA scan**
The most important informations to check are the correct identification of the patient, his date of birth and also the sex and ethnicity which are mandatory to calculate T-scores. Sex is used by all manufacturers to calculate T-scores (i.e. T-scores for women are calculated using a female normative database, while T-scores for men are calculated using a male normative database). Although all manufacturers use race in calculating Z-scores, there is inconsistency in the way race is handled when calculating T-scores. Norland and Hologic are using race in calculating T-scores (i.e. T-scores for Caucasians are calculated using a Caucasian normative database, T-scores for Blacks are calculated using a normative database for Blacks); however, GE Lunar and recent Hologic machines use the database for young-normal Caucasians to calculate T-scores, regardless of the race of the subject. The ISCD recommends the latter approach for use in North America (Baim, Wilson et al. 2005) because using race-adjusted T-scores results in a similar prevalence of ''osteoporosis'' in every racial group, despite the fact that age-specific fracture rates can be very different.
#### **5.1 Positioning**
The main purpose of the DXA scan image is to check if the patient is positioned correctly, something that the technologist must determine before the patient leaves the testing centre. Positioning should also be doublechecked by the clinician who interprets the test(Roux 1998). There is many available resources for BMD technologists and physicians training, such as ISCD or International Osteoporosis Foundation (IOF) courses.
A scan with correct positioning of the spine is shown in Fig. 1a: the patient is straight on the table (spine is straight on the image), not rotated (spinous processes are centered), and centered in the field (roughly equal soft tissue fields on either side of the spine). Patients with scoliosis cannot be positioned with the spine straight on the table; moreover with severe scoliosis degenerative changes can occur that invalidate the spine measurement. The scan should extend up sufficiently far to include part of the lowest vertebra with ribs (which is usually T12) and low enough to show the pelvic brim (which is usually the level of the L4–L5 interspace). Most testing centers will elevate the patient's knees with a foam block (hip at a 90° angle to the spine) to try to partially flatten the normal lumbar lordosis. For proper positioning of the hip, the patient should have the femur straight on the table (shaft parallel to the edge of the picture), with 15–25° of internal rotation, which can be achieved
The ISCD recommends obtaining BMD measurements of the posteroanterior spine and hip(Leib, Binkley et al. 2006). The lateral spine and Ward's triangle region of the hip should not be used for diagnosis, because these sites overestimate osteoporosis and results can be false-positive. Evidence suggests that the femur (neck or total hip) is the optimum site for predicting the risk of hip fracture and the spine is the optimum site for monitoring response to treatment. Thus, many authors recommend hip measure alone for the fracture risk assessment(Kanis, Johnell et al. 2000; Kanis, Oden et al. 2001; Kanis 2002; Johnell, Kanis et al. 2005; Kanis, Seeman et al. 2005; Arabi, Baddoura et al. 2007). In very obese patients, those with primary hyperparathyroidism, or those in whom the hip or the spine, or both, cannot be measured or interpreted, BMD may be measured in the forearm, using a 33% radius on
The most important informations to check are the correct identification of the patient, his date of birth and also the sex and ethnicity which are mandatory to calculate T-scores. Sex is used by all manufacturers to calculate T-scores (i.e. T-scores for women are calculated using a female normative database, while T-scores for men are calculated using a male normative database). Although all manufacturers use race in calculating Z-scores, there is inconsistency in the way race is handled when calculating T-scores. Norland and Hologic are using race in calculating T-scores (i.e. T-scores for Caucasians are calculated using a Caucasian normative database, T-scores for Blacks are calculated using a normative database for Blacks); however, GE Lunar and recent Hologic machines use the database for young-normal Caucasians to calculate T-scores, regardless of the race of the subject. The ISCD recommends the latter approach for use in North America (Baim, Wilson et al. 2005) because using race-adjusted T-scores results in a similar prevalence of ''osteoporosis'' in every racial group, despite the fact that age-specific fracture rates can be very different.
The main purpose of the DXA scan image is to check if the patient is positioned correctly, something that the technologist must determine before the patient leaves the testing centre. Positioning should also be doublechecked by the clinician who interprets the test(Roux 1998). There is many available resources for BMD technologists and physicians training,
A scan with correct positioning of the spine is shown in Fig. 1a: the patient is straight on the table (spine is straight on the image), not rotated (spinous processes are centered), and centered in the field (roughly equal soft tissue fields on either side of the spine). Patients with scoliosis cannot be positioned with the spine straight on the table; moreover with severe scoliosis degenerative changes can occur that invalidate the spine measurement. The scan should extend up sufficiently far to include part of the lowest vertebra with ribs (which is usually T12) and low enough to show the pelvic brim (which is usually the level of the L4–L5 interspace). Most testing centers will elevate the patient's knees with a foam block (hip at a 90° angle to the spine) to try to partially flatten the normal lumbar lordosis. For proper positioning of the hip, the patient should have the femur straight on the table (shaft parallel to the edge of the picture), with 15–25° of internal rotation, which can be achieved
such as ISCD or International Osteoporosis Foundation (IOF) courses.
**4. Site of measurement of BMD**
the nondominant forearm.
**5.1 Positioning**
**5. Interpreting a DXA scan**
by the use of positioning devices. Internal rotation may be improved by having the patient flex the foot before doing the internal rotation, and then relaxing the foot after the strap is in place. This amount of internal rotation presents the long axis of the femoral neck perpendicular to the X-ray beam, providing the greatest area and the lowest bone mineral content (and the lowest BMD), and is confirmed on the scan by seeing little or none of the lesser trochanter (Fig. 1b)(Lekamwasam and Lenora 2003; 2004). If the desired amount of internal rotation cannot be achieved, as is often the case in patients with hip arthritis or short femoral necks, the technologist should place the patient comfortably in a position that is likely to be reproducible in a subsequent scan (Hamdy, Kiebzak et al. 2006; Lewiecki, Binkley et al. 2006).
(a)
Fig. 1. Correct positioning and analysis of the L1–L4 spine (a) and the proximal femur (Lunar (b) and Hologic (c)).
#### **5.2 DXA scan analysis**
The software marks regions of interest in the spine and hip, but the technologist can and should make adjustments if needed. The spine region of interest consists of the L1 through L4 vertebrae (Fig. 1a). Correct placement of the top and bottom of the spine ''box'' is critical.
Interpreting a DXA Scan in Clinical Practice 9
(e) Fig. 2. Examples among some common spine scanning problems: (a) The spine is too close to the right side of the image (b) Vertebral levels are mis-identified (c) Metal button over L4 (d)
(a) (b)
(c) (d)
Scoliosis, and osteophyte at L3–L4 (e) Laminectomy
The intervertebral lines can be moved or angled, if necessary. There must be sufficient soft tissue on both sides of the spine; otherwise BMD will be under estimated. The hip regions of interest include the femoral neck, trochanter, and total hip (Fig. 1b). Ward's region and the intertrochanteric region are not relevant (and can be deleted from the results reports. The default hip analysis includes a midline that must be placed correctly for the other sites to be identified correctly. The preferred position for the rectangular femoral neck box differs for the different manufacturers. For GE Lunar, the femoral neck box is located by the analysis program at the narrowest and lowest density section of the neck; typically this will be about half way between the femoral head and the trochanter (Fig. 1b). For Hologic the box is on the distal part of the femoral neck (Fig. 1c). This induces a large difference among these 2 measurements, because of a gradient of BMD all along the femoral neck (the proximal being the highest, the distal being the lowest). Thus careful checking of the femoral neck box is mandatory.
The image should be evaluated for artifacts (e.g. surgical clips, navel rings, barium sulphate, metal from zipper, coin, clip, or other metallic object) or local structural change (e.g. osteophytes, syndesmophytes, compression fractures, aortic calcification). Almost all artifacts and local structural change will spuriously elevate BMD(El Maghraoui 2004). This is especially true for spinal degenerative change, which can elevate spine BMD by 2, 3, or more T-score. In the spine, absent bone (laminectomy or spina bifida) or vertebral rotation (idiopathic scoliosis) will spuriously lower BMD. All evaluable vertebrae should be used, but vertebrae that are affected by local structural change should be deleted from the analysis. Most agree that decisions can be based on two vertebrae; the use of a single vertebra is not recommended. If all vertebrae are affected, the spine should be reported as ''invalid,'' with no BMD or T-score results given. Figure 2 and 3 show examples from common spine and hips scanning problems.
Finally, physicians must keep in mind to actively look for secondary osteoporosis in front of low BMD value, either by thorough history taking or with biochemical studies before stating about post menopausal osteoporosis.
#### **6. Vertebral fracture assessment (VFA)**
For assessing vertebral heights (also called vertebral morphometry), a special software is used to determine vertebral body dimensions. The computer (with the help of the technologist) places points on the superior and inferior endplates of each vertebra. The vertebral heights are calculated and compared to each other as well as to the expected normal dimensions. With the advent of higher-resolution DXA systems, visual assessment of fractures is also possible from DXA-based lateral spine images (Figure 4). In this situation, the DXA system essentially functions as a digital X-ray imaging device. Visual assessment is performed from a computer monitor or high-resolution printout. To optimize the assessment, the use of high-definition dual-energy images has been recommended (Rea, Li et al. 2000; Chapurlat, Duboeuf et al. 2006; Olenginski, Newman et al. 2006). Using a DXA system for assessing vertebral fracture status has several advantages. The evaluation of spine fractures can be performed without a conventional lateral spine X-ray. This can be done at the same time and at the same place as the BMD measurement, with much less radiation than a conventional spine X-ray. Moreover, VFA is a technology for diagnosing vertebral fractures that may alter diagnostic classification, improve fracture risk stratification,
The intervertebral lines can be moved or angled, if necessary. There must be sufficient soft tissue on both sides of the spine; otherwise BMD will be under estimated. The hip regions of interest include the femoral neck, trochanter, and total hip (Fig. 1b). Ward's region and the intertrochanteric region are not relevant (and can be deleted from the results reports. The default hip analysis includes a midline that must be placed correctly for the other sites to be identified correctly. The preferred position for the rectangular femoral neck box differs for the different manufacturers. For GE Lunar, the femoral neck box is located by the analysis program at the narrowest and lowest density section of the neck; typically this will be about half way between the femoral head and the trochanter (Fig. 1b). For Hologic the box is on the distal part of the femoral neck (Fig. 1c). This induces a large difference among these 2 measurements, because of a gradient of BMD all along the femoral neck (the proximal being the highest, the distal being the lowest). Thus careful checking of the femoral neck box is
The image should be evaluated for artifacts (e.g. surgical clips, navel rings, barium sulphate, metal from zipper, coin, clip, or other metallic object) or local structural change (e.g. osteophytes, syndesmophytes, compression fractures, aortic calcification). Almost all artifacts and local structural change will spuriously elevate BMD(El Maghraoui 2004). This is especially true for spinal degenerative change, which can elevate spine BMD by 2, 3, or more T-score. In the spine, absent bone (laminectomy or spina bifida) or vertebral rotation (idiopathic scoliosis) will spuriously lower BMD. All evaluable vertebrae should be used, but vertebrae that are affected by local structural change should be deleted from the analysis. Most agree that decisions can be based on two vertebrae; the use of a single vertebra is not recommended. If all vertebrae are affected, the spine should be reported as ''invalid,'' with no BMD or T-score results given. Figure 2 and 3 show examples from
Finally, physicians must keep in mind to actively look for secondary osteoporosis in front of low BMD value, either by thorough history taking or with biochemical studies before stating
For assessing vertebral heights (also called vertebral morphometry), a special software is used to determine vertebral body dimensions. The computer (with the help of the technologist) places points on the superior and inferior endplates of each vertebra. The vertebral heights are calculated and compared to each other as well as to the expected normal dimensions. With the advent of higher-resolution DXA systems, visual assessment of fractures is also possible from DXA-based lateral spine images (Figure 4). In this situation, the DXA system essentially functions as a digital X-ray imaging device. Visual assessment is performed from a computer monitor or high-resolution printout. To optimize the assessment, the use of high-definition dual-energy images has been recommended (Rea, Li et al. 2000; Chapurlat, Duboeuf et al. 2006; Olenginski, Newman et al. 2006). Using a DXA system for assessing vertebral fracture status has several advantages. The evaluation of spine fractures can be performed without a conventional lateral spine X-ray. This can be done at the same time and at the same place as the BMD measurement, with much less radiation than a conventional spine X-ray. Moreover, VFA is a technology for diagnosing vertebral fractures that may alter diagnostic classification, improve fracture risk stratification,
mandatory.
common spine and hips scanning problems.
**6. Vertebral fracture assessment (VFA)**
about post menopausal osteoporosis.
(a) (b)
(c) (d)
(e)
Fig. 2. Examples among some common spine scanning problems: (a) The spine is too close to the right side of the image (b) Vertebral levels are mis-identified (c) Metal button over L4 (d) Scoliosis, and osteophyte at L3–L4 (e) Laminectomy
Interpreting a DXA Scan in Clinical Practice 11
and identify patients likely to benefit from pharmacological therapy who otherwise might not be treated(Olenginski, Newman et al. 2006; Roux, Fechtenbaum et al. 2007). Despite the apparent advantages, the future of VFA using DXA remains unclear. Skeletal radiologists have criticized the technique for being insensitive and inaccurate for detecting vertebral fractures in particular at the upper thoracic spine. A DXA image is of lower resolution than a conventional X-ray and might fail to identify other potential problems or diseases that would be apparent on a spine film. However, VFA allows ruling out vertebral fracture at levels where vertebral fracture is most common, i.e. the lumbar and the mid and lower thoracic levels, and the pencil beam mode of assessment eliminates parallax errors in viewing the vertebral body, which can sometimes make a normal vertebral body appear to have been compressed in a routine spine x-ray(Duboeuf, Bauer et al. 2005; Jacobs-Kosmin,
Sandorfi et al. 2005; Chapurlat, Duboeuf et al. 2006; Damiano, Kolta et al. 2006).
Fig. 4. Vertebral fracture assessment from a dual x-ray absorptiometry image of the spine.
fractures or spinal radiographs could benefit from a VFA.
At this time, DXA devices are not generally accepted as a surrogate for spinal X-rays, though they may provide a useful screening tool in higher-risk patients when spinal X-rays are unavailable. For example, individuals over 65, subjects reporting significant height loss or patients on long term glucocorticoid therapy who have not had previous vertebral
(e)
Fig. 3. Examples among some common hip scanning problems: (a) The scan did not go far enough laterally and part of the femoral head is missing. (b) The femur is adducted (c) The femur is abducted (d) Suboptimal internal rotation (too much of the lesser trochanter is showing) (e) Abnormal bone (history of hip fracture and osteosynthesis)
(e) Fig. 3. Examples among some common hip scanning problems: (a) The scan did not go far enough laterally and part of the femoral head is missing. (b) The femur is adducted (c) The femur is abducted (d) Suboptimal internal rotation (too much of the lesser trochanter is
showing) (e) Abnormal bone (history of hip fracture and osteosynthesis)
(a) (b)
(c) (d)
and identify patients likely to benefit from pharmacological therapy who otherwise might not be treated(Olenginski, Newman et al. 2006; Roux, Fechtenbaum et al. 2007). Despite the apparent advantages, the future of VFA using DXA remains unclear. Skeletal radiologists have criticized the technique for being insensitive and inaccurate for detecting vertebral fractures in particular at the upper thoracic spine. A DXA image is of lower resolution than a conventional X-ray and might fail to identify other potential problems or diseases that would be apparent on a spine film. However, VFA allows ruling out vertebral fracture at levels where vertebral fracture is most common, i.e. the lumbar and the mid and lower thoracic levels, and the pencil beam mode of assessment eliminates parallax errors in viewing the vertebral body, which can sometimes make a normal vertebral body appear to have been compressed in a routine spine x-ray(Duboeuf, Bauer et al. 2005; Jacobs-Kosmin, Sandorfi et al. 2005; Chapurlat, Duboeuf et al. 2006; Damiano, Kolta et al. 2006).
Fig. 4. Vertebral fracture assessment from a dual x-ray absorptiometry image of the spine.
At this time, DXA devices are not generally accepted as a surrogate for spinal X-rays, though they may provide a useful screening tool in higher-risk patients when spinal X-rays are unavailable. For example, individuals over 65, subjects reporting significant height loss or patients on long term glucocorticoid therapy who have not had previous vertebral fractures or spinal radiographs could benefit from a VFA.
Interpreting a DXA Scan in Clinical Practice 13
3. Anatomic discordance is owing to differences in the composition of bone envelopes tested. An example is the difference in T-scores found for the posteroanterior lumbar
4. Artifactual discordance occurs when dense synthetic manmade substances are within the field of ROI of the test: e.g. barium sulphate, metal from zipper, coin, clip, or other
5. And finally, technical discordance occurs because of device errors, technician variability, patients' movements, and variation due to other unpredictable sources. With respect to positioning error, some studies showed that either excessive internal or external rotation of the femur during test acquisition resulted in a BMD difference of as much as 10% compared with correct positioning. We demonstrated in a previous study that DXA in vivo reproducibility is two-fold better in the hips than the spine especially when measuring both hips(El Maghraoui, Do Santos Zounon et al. 2005). Finally, technical discordance can occur due to the normative reference data used by the device software to analyze the test(Liao, Wu et al. 2003; McMahon, Nightingale et al. 2004; Lewiecki, Binkley et al. 2006). This type of discordance occurs when the average BMD of the normative group used to calculate the T-score is significantly different from the
spine and the supine lateral lumbar spine in the same patient.
average value found for the whole population.
**7.2 Consequences of T-score discordance on osteoporosis management**
The high prevalence of T-score discordance could induce some problems for the physicians in decision-making regarding these patients. In general, high prevalence of discordance between lumbar spine and hip T-scores suggests some defects in the cut-off values for definition of osteoporosis and osteopenia proposed with the WHO. The inconsistencies in the diagnostic classification of osteoporosis between skeletal sites lend credence to the notion that BMD should be used as only one of the factors in making therapeutic decisions when evaluating patients with osteoporosis. An international team convened by the WHO is trying to develop a globally applicable measure of absolute fracture risk based upon multiple risk factors including BMD. This could silence much of the controversy regarding the choice of reference data for T-score calculation and usefulness of relatively arbitrary densitometric categorizations. However, one can speculate that discordance in individual fracture risk estimation with this new absolute fracture risk will still be observed as it will be
Correct performance of BMD measurements using DXA requires rigorous attention to detail in positioning and analysis. When DXA studies are performed incorrectly, it can lead to major mistakes in diagnosis and therapy. Measurement error must be considered when evaluating serial assessments. A clear understanding of the interpretation of serial measurements and the statistical principles impacting upon their interpretation is necessary
lumbar osteoporosis.
metallic object.
based on different sites BMD.
**8. Conclusion**
Maghraoui 2004; Khan, Hanley et al. 2006). This will lead to higher prevalence of
#### **7. Concordance between measurement sites**
It is recommended to measure the PA lumbar spine and proximal femur and classifying the patient based on the lowest T-score from three sites (lumbar spine, femoral neck, and total hip). Although the BMDs at different anatomic regions are correlated, the agreement between sites is low when it comes to classifying individual subjects as osteoporotic or not. Thus, T-score discordance between the lumbar spine and hip testing sites is a commonly observed phenomenon in densitometery. T-score discordance is the observation that the T-score of an individual patient varies from one key measurement site to another.
#### **7.1 Prevalence and risk factors of T-score discordance**
Various studies have analyzed the prevalence and impact of T-score discordance on the management of osteoporosis(Faulkner, von Stetten et al. 1999; Woodson 2000; O'Gradaigh, Debiram et al. 2003; Moayyeri, Soltani et al. 2005). Only two studies focused on risk factors of this commonly observed discordance (Moayyeri, Soltani et al. 2005; El Maghraoui, Mouinga Abayi et al. 2007; El Maghraoui, Mouinga Abayi et al. 2007). Five different causes for occurrence of discordance between the spine and the hip sites have been described(Woodson 2000).
It is recommended to measure the PA lumbar spine and proximal femur and classifying the patient based on the lowest T-score from three sites (lumbar spine, femoral neck, and total hip). Although the BMDs at different anatomic regions are correlated, the agreement between sites is low when it comes to classifying individual subjects as osteoporotic or not. Thus, T-score discordance between the lumbar spine and hip testing sites is a commonly observed phenomenon in densitometery. T-score discordance is the observation that the T-score of an individual patient varies from one key measurement
Various studies have analyzed the prevalence and impact of T-score discordance on the management of osteoporosis(Faulkner, von Stetten et al. 1999; Woodson 2000; O'Gradaigh, Debiram et al. 2003; Moayyeri, Soltani et al. 2005). Only two studies focused on risk factors of this commonly observed discordance (Moayyeri, Soltani et al. 2005; El Maghraoui, Mouinga Abayi et al. 2007; El Maghraoui, Mouinga Abayi et al. 2007). Five different causes for occurrence of discordance between the spine and the hip sites have been
1. Physiologic discordance is related to the skeleton's natural adaptive reaction to normal external and internal factors and forces. Mechanical strain especially related to weight bearing plays a key role in this kind of discordance. An example of this type of discordance is the difference observed between the dominant and nondominant total hip(Hamdy, Kiebzak et al. 2006). The explanation is that weight bearing can cause rise in bone density especially in the hip and femur regions. Moreover, the spine and hips usually start out with different T-scores (the spine is said to reach peak at least 5 yrs before the hip)(Blank, Malone et al. 2006). And finally, bone loss observed with age in an individual may be more rapid and important in trabecular than cortical bone is another explanation(Agarwal and Camacho 2006). Trabecular bones (typical of lumbar area) are known to have a more rapid rate of deprivation in early post-menopausal state in comparison to cortical bone (typical of
2. The second type of discordance described as pathophysiologic discordance is seen secondary to a disease. Common examples observed in the elderly include vertebral osteophytosis, vertebral end plate and facet sclerosis, osteochondrosis, and aortic calcification(Bolotin 2001; Theodorou and Theodorou 2002). Another important cause in younger patients is ankylosing spondylitis syndesmophytes(El Maghraoui, Borderie et al. 1999; Maillefert, Aho et al. 2001; El Maghraoui 2004; El Maghraoui 2004; El Maghraoui, Do Santos Zounon et al. 2005). The abnormal calcium deposition within the field of the DXA region of interest (ROI) leads to the falsely elevated spine T-score. A second subtype is a true discordance resulting from a more decreased BMD in the lumbar spine than the hips. Indeed, most of the aetiologies of the secondary osteoporosis (such as glucocorticoid excess, hyperthyroidism, malabsorption, liver disease, rheumatoid arthritis) first affect spinal column(El
**7. Concordance between measurement sites**
**7.1 Prevalence and risk factors of T-score discordance**
site to another.
described(Woodson 2000).
proximal femur).
Maghraoui 2004; Khan, Hanley et al. 2006). This will lead to higher prevalence of lumbar osteoporosis.
#### **7.2 Consequences of T-score discordance on osteoporosis management**
The high prevalence of T-score discordance could induce some problems for the physicians in decision-making regarding these patients. In general, high prevalence of discordance between lumbar spine and hip T-scores suggests some defects in the cut-off values for definition of osteoporosis and osteopenia proposed with the WHO. The inconsistencies in the diagnostic classification of osteoporosis between skeletal sites lend credence to the notion that BMD should be used as only one of the factors in making therapeutic decisions when evaluating patients with osteoporosis. An international team convened by the WHO is trying to develop a globally applicable measure of absolute fracture risk based upon multiple risk factors including BMD. This could silence much of the controversy regarding the choice of reference data for T-score calculation and usefulness of relatively arbitrary densitometric categorizations. However, one can speculate that discordance in individual fracture risk estimation with this new absolute fracture risk will still be observed as it will be based on different sites BMD.
#### **8. Conclusion**
Correct performance of BMD measurements using DXA requires rigorous attention to detail in positioning and analysis. When DXA studies are performed incorrectly, it can lead to major mistakes in diagnosis and therapy. Measurement error must be considered when evaluating serial assessments. A clear understanding of the interpretation of serial measurements and the statistical principles impacting upon their interpretation is necessary
Interpreting a DXA Scan in Clinical Practice 15
Blake, G. M. and I. Fogelman (2002). "Dual energy x-ray absorptiometry and its clinical
Blake, G. M. and I. Fogelman (2003). "DXA scanning and its interpretation in osteoporosis."
Blake, G. M. and I. Fogelman (2007). "The role of DXA bone density scans in the diagnosis
Blank, R. D., D. G. Malone, et al. (2006). "Patient variables impact lumbar spine dual energy
Bolotin, H. H. (2001). "Inaccuracies inherent in dual-energy X-ray absorptiometry in vivo
Cadarette, S. M., S. B. Jaglal, et al. (2000). "Development and validation of the Osteoporosis
Chapurlat, R. D., F. Duboeuf, et al. (2006). "Effectiveness of instant vertebral assessment to
Damiano, J., S. Kolta, et al. (2006). "Diagnosis of vertebral fractures by vertebral fracture
Duboeuf, F., D. C. Bauer, et al. (2005). "Assessment of vertebral fracture using densitometric
El Maghraoui, A., D. Borderie, et al. (1999). "Osteoporosis, body composition, and bone
El Maghraoui, A., A. A. Do Santos Zounon, et al. (2005). "Reproducibility of bone mineral
El Maghraoui, A., A. A. Guerboub, et al. (2007). "Body mass index and gynecological
El Maghraoui, A., A. Habbassi, et al. (2007). "Validation and comparative evaluation of
El Maghraoui, A., D. A. Mouinga Abayi, et al. (2007). "Prevalence and risk factors of
El Maghraoui, A., D. A. Mouinga Abayi, et al. (2007). "Discordance in diagnosis of osteoporosis using spine and hip bone densitometry." *J Clin Densitom* (in press). Faulkner, K. G., E. von Stetten, et al. (1999). "Discordance in patient classification using T-
density measurements using dual X-ray absorptiometry in daily clinical practice."
factors as determinants of bone mass in healthy Moroccan women." *Maturitas*
four osteoporosis risk indexes in Moroccan menopausal women." *Arch Osteop* (in
discordance in diagnosis of osteoporosis using spine and hip bone densitometry."
detect prevalent vertebral fracture." *Osteoporos Int* 17(8): 1189-95.
El Maghraoui, A. (2004). "L'ostéoprose cortisonique." *Presse Med* 33(17): 1213-7. El Maghraoui, A. (2004). "La spondylarthrite ankylosante." *Presse Med* 33(20): 1459-64. El Maghraoui, A. (2004). "Osteoporosis and ankylosing spondylitis." *Joint Bone Spine* 71(4):
turnover in ankylosing spondylitis." *J Rheumatol* 26(10): 2205-9.
bone mineral densitometry may flaw osteopenic/osteoporotic interpretations and mislead assessment of antiresorptive therapy effectiveness." *Bone* 28(5): 548-
Risk Assessment Instrument to facilitate selection of women for bone
and treatment of osteoporosis." *Postgrad Med J* 83(982): 509-517.
X-ray absorptiometry precision." *Osteoporos Int* 17(5): 768-74.
applications." *Semin Musculoskelet Radiol* 6(3): 207-18.
*Hosp Med* 64(9): 521-5.
densitometry." *Cmaj* 162(9): 1289-94.
assessment." *J Clin Densitom* 9(1): 66-71.
*Osteoporos Int* 16(12): 1742-8.
*Ann Rheum Dis* 66: 271-2.
scores." *J Clin Densitom* 2(3): 343-50.
morphometry." *J Clin Densitom* 8(3): 362-8.
55.
291-5.
56(4): 375-82.
press).
to determine whether a change is real and not simply random fluctuation. Moreover, it is important to keep in mind that fracture-protection benefit may be realized before BMD gains are detected. Physicians interested in osteoporosis management, even if not directly involved in the performance and interpretation of DXA, should be familiar with the principles outlined here to minimize serious errors and allow proper use of bone densitometry.
#### **9. Abbreviations**
BMC: bone mineral content BMD: bone mineral density CV: coefficient of variation DXA: dual-energy x-ray absorptiometry IOF: international osteoporosis foundation ISCD: international society for clinical densitometry LSC: least significant change OST: osteoporosis self-assessment tool PE: precision error ROI: region of interest SD: standard deviation SDD: smallest detectable difference VFA: vertebral fracture assessment
WHO: world health organization
#### **10. References**
to determine whether a change is real and not simply random fluctuation. Moreover, it is important to keep in mind that fracture-protection benefit may be realized before BMD gains are detected. Physicians interested in osteoporosis management, even if not directly involved in the performance and interpretation of DXA, should be familiar with the principles outlined here to minimize serious errors and allow proper use of bone
Adler, R. A., M. T. Tran, et al. (2003). "Performance of the Osteoporosis Self-assessment Screening Tool for osteoporosis in American men." *Mayo Clin Proc* 78(6): 723-7. Agarwal, M. and P. Camacho (2006). "Bone densitometry. Interpretation and pitfalls."
Arabi, A., R. Baddoura, et al. (2007). "Discriminative ability of dual-energy X-ray
Baddoura, R., H. Awada, et al. (2006). "An audit of bone densitometry practice with reference to ISCD, IOF and NOF guidelines." *Osteoporos Int* 17(7): 1111-5. Baim, S., C. R. Wilson, et al. (2005). "Precision assessment and radiation safety for dual-
Clinical Densitometry." *J Clin Densitom* 8(4): 371-8.
absorptiometry site selection in identifying patients with osteoporotic fractures."
energy X-ray absorptiometry: position paper of the International Society for
densitometry.
**9. Abbreviations**
BMC: bone mineral content BMD: bone mineral density CV: coefficient of variation
LSC: least significant change
PE: precision error
**10. References**
ROI: region of interest SD: standard deviation
DXA: dual-energy x-ray absorptiometry IOF: international osteoporosis foundation
OST: osteoporosis self-assessment tool
SDD: smallest detectable difference VFA: vertebral fracture assessment WHO: world health organization
*Postgrad Med* 119(1): 17-23.
*Bone* 40(4): 1060-5.
ISCD: international society for clinical densitometry
Interpreting a DXA Scan in Clinical Practice 17
Lekamwasam, S. and R. S. Lenora (2003). "Effect of leg rotation on hip bone mineral density
Lewiecki, E. M., N. Binkley, et al. (2006). "DXA quality matters." *J Clin Densitom* 9(4): 388-
Lewiecki, E. M. and J. L. Borges (2006). "Bone density testing in clinical practice." *Arq Bras*
Liao, E. Y., X. P. Wu, et al. (2003). "Establishment and evaluation of bone mineral density
Maillefert, J. F., L. S. Aho, et al. (2001). "Changes in bone density in patients with ankylosing spondylitis: a two-year follow-up study." *Osteoporos Int* 12(7): 605-9. McMahon, K., J. Nightingale, et al. (2004). "Discordance in DXA male reference ranges." *J*
Moayyeri, A., A. Soltani, et al. (2005). "Discordance in diagnosis of osteoporosis using spine
O'Gradaigh, D., I. Debiram, et al. (2003). "A prospective study of discordance in diagnosis of
Olenginski, T. P., E. D. Newman, et al. (2006). "Development and evaluation of a vertebral
Price, R. I., M. J. Walters, et al. (2003). "Impact of the analysis of a bone density reference
Rea, J. A., J. Li, et al. (2000). "Visual assessment of vertebral deformity by X-ray
Richy, F., O. Ethgen, et al. (2004). "Primary prevention of osteoporosis: mass screening
Roux, C., J. Fechtenbaum, et al. (2007). "Mild prevalent and incident vertebral fractures are
Salaffi, F., F. Silveri, et al. (2005). "Development and validation of the osteoporosis
Svendsen, O. L., C. Hassager, et al. (1995). "Impact of soft tissue on in vivo accuracy of bone
Theodorou, D. J. and S. J. Theodorou (2002). "Dual-energy X-ray absorptiometry in clinical
Tothill, P. and A. Avenell (1994). "Errors in dual-energy X-ray absorptiometry of the lumbar
range on determination of the T-score." *J Clin Densitom* 6(1): 51-62.
Roux, C. (1998). "Densitométrie osseuse et ostéoporose." *J Radiol* 79(9): 821-3.
likely to have low bone density." *Clin Rheumatol* 24(3): 203-11.
risk factors for new fractures." *Osteoporos Int*.
osteoporosis using spine and proximal femur bone densitometry." *Osteoporos Int*
fracture assessment program using IVA and its integration with mobile DXA." *J*
absorptiometry: a highly predictive method to exclude vertebral deformity."
scenario or prescreening with questionnaires? An economic perspective." *J Bone*
prescreening risk assessment (OPERA) tool to facilitate identification of women
mineral measurements in the spine, hip, and forearm: a human cadaver study." *J*
practice: application and interpretation of scans beyond the numbers." *Clin Imaging*
spine owing to fat distribution and soft tissue thickness during weight change." *Br J*
reference databases appropriate for diagnosis and evaluation of osteoporosis in
measurements." *J Clin Densitom* 6(4): 331-6.
Chinese women." *J Bone Miner Metab* 21(3): 184-92.
and hip bone densitometry." *BMC Endocr Disord* 5(1): 3.
*Endocrinol Metabol* 50(4): 586-95.
*Clin Densitom* 7(2): 121-6.
*Clin Densitom* 9(1): 72-7.
*Osteoporos Int* 11(8): 660-8.
*Miner Res* 19(12): 1955-60.
*Bone Miner Res* 10(6): 868-73.
26(1): 43-9.
*Radiol* 67(793): 71-5.
14(1): 13-8.
92.
Ghazi, M., A. Mounach, et al. (2007). "Performance of the osteoporosis risk assessment tool
Gnudi, S. and E. Sitta (2005). "Clinical risk factor evaluation to defer postmenopausal
Griffith, J. F., D. K. Yeung, et al. (2006). "Vertebral marrow fat content and diffusion and
Hamdy, R., G. M. Kiebzak, et al. (2006). "The prevalence of significant left-right differences
Hans, D., R. W. Downs, Jr., et al. (2006). "Skeletal sites for osteoporosis diagnosis: the 2005
Hillier, T. A., K. L. Stone, et al. (2007). "Evaluating the value of repeat bone mineral density
Jacobs-Kosmin, D., N. Sandorfi, et al. (2005). "Vertebral deformities identified by vertebral
Johnell, O., J. A. Kanis, et al. (2005). "Predictive value of BMD for hip and other fractures." *J*
Kanis, J. A. (1994). "Assessment of fracture risk and its application to screening for
Kanis, J. A. (2002). "Diagnosis of osteoporosis and assessment of fracture risk." *Lancet*
Kanis, J. A., F. Borgstrom, et al. (2005). "Assessment of fracture risk." *Osteoporos Int* 16(6):
Kanis, J. A., O. Johnell, et al. (2000). "Risk of hip fracture according to the World Health Organization criteria for osteopenia and osteoporosis." *Bone* 27(5): 585-90. Kanis, J. A., A. Oden, et al. (2001). "The burden of osteoporotic fractures: a method for
Kanis, J. A., E. Seeman, et al. (2005). "The perspective of the International Osteoporosis
Khan, A. A., D. A. Hanley, et al. (2006). "Standards for performing DXA in individuals with
Kuiper, J. W., C. van Kuijk, et al. (1996). "Accuracy and the influence of marrow fat on
Lee, D. C., T. A. L. Wren, et al. (2007). "Correcting DXA pediatric bone mineral density measurments to account for fat inhomogeneity." *ASBMR* W514: (Abstract). Leib, E. S., N. Binkley, et al. (2006). "Position Development Conference of the International
Foundation on the official positions of the International Society for Clinical
quantitative CT and dual-energy X-ray absorptiometry measurements of the
Society for Clinical Densitometry. Vancouver, BC, July 15-17, 2005." *J Rheumatol*
setting intervention thresholds." *Osteoporos Int* 12(5): 417-27.
Densitometry." *Osteoporos Int* 16(5): 456-9, discussion 579-80.
secondary causes of osteoporosis." *J Clin Densitom* 9(1): 47-57.
femoral neck in vitro." *Osteoporos Int* 6(1): 25-30.
in hip bone mineral density." *Osteoporos Int* 17(12): 1772-80.
ISCD Official Positions." *J Clin Densitom* 9(1): 15-21.
fractures." *Arch Intern Med* 167(2): 155-60.
density." *J Clin Densitom* 8(3): 267-72.
*Bone Miner Res* 20(7): 1185-94.
*Osteoporos Int* 4(6): 368-81.
359(9321): 1929-36.
33(11): 2319-21.
581-9.
women from bone mineral density measurement: an Italian study." *J Clin Densitom*
perfusion indexes in women with varying bone density: MR evaluation." *Radiology*
measurement and prediction of fractures in older women: the study of osteoporotic
fracture assessment: associations with clinical characteristics and bone mineral
postmenopausal osteoporosis: synopsis of a WHO report. WHO Study Group."
in Moroccan men." *Clin Rheumatol*.
8(2): 199-205.
241(3): 831-8.
**2**
*Morocco*
**Monitoring DXA Measurement**
*Rheumatology Department, Military Hospital Mohammed V, Rabat,*
Osteoporosis is a worldwide major public health problem. Bone densitometry has become the "gold standard" in its diagnosis and treatment evaluation(El Maghraoui and Roux 2008). With its advantages of high precision, short scan times, low radiation dose, and stable calibration, dual-energy x-ray absorptiometry (DXA) has been established by the World Health Organization (WHO) as the technique of reference for assessing bone mineral density (BMD) in postmenopausal women and based the definitions of osteopenia and osteoporosis on its results. Recently, efficient therapeutic options for treatment of osteoporosis have been developed which create possibilities of effective intervention. Therefore, screening for and treatment of osteoporosis are widely practised in postmenopausal women and in people with an increased risk of osteoporosis because of underlying diseases (e.g. chronic rheumatic diseases especially when treated by corticosteroids)(Phillipov, Seaborn et al. 2001). Moreover, BMD measurement is needed to select patients for osteoporosis treatment, as there is no proof that drugs for osteoporosis (other than hormone replacement therapy [HRT]) are beneficial in women with clinical risk
It has also become more and more common to perform a second DXA measurement to monitor BMD status or the effect of therapeutic intervention. When a second measurement is performed on a patient, the clinician needs to distinguish between a true change in BMD and a random fluctuation related to variability in the measurement procedure. The reproducibility of DXA measurements is claimed to be good. Such variability is due to multiple causes, such as device errors, technician variability, patients' movements, and variation due to other unpredictable sources (Nguyen, Sambrook et al. 1997; Lodder, Lems
The precision error is usually expressed as the coefficient of variation (CV), which is the ratio of the standard deviation (SD) to the mean of the measurements, although several other statistics to express reproducibility exist such as the smallest detectable difference (SDD) or the least significant change (LSC). The SDD represents a cut-off that can be measured in an individual and is usually considered more useful than the CV in clinical practice (Fuleihan,
**1. Introduction**
et al. 2004).
factors for fractures but normal BMD values.
Testa et al. 1995; Ravaud, Reny et al. 1999).
**in Clinical Practice**
Abdellah El Maghraoui
### **Monitoring DXA Measurement in Clinical Practice**
Abdellah El Maghraoui *Rheumatology Department, Military Hospital Mohammed V, Rabat, Morocco*
#### **1. Introduction**
18 Dual Energy X-Ray Absorptiometry
Watts, N. B. (2004). "Fundamentals and pitfalls of bone densitometry using dual-energy X-
Woodson, G. (2000). "Dual X-ray absorptiometry T-score concordance and discordance between the hip and spine measurement sites." *J Clin Densitom* 3(4): 319-24.
ray absorptiometry (DXA)." *Osteoporos Int* 15(11): 847-54.
Osteoporosis is a worldwide major public health problem. Bone densitometry has become the "gold standard" in its diagnosis and treatment evaluation(El Maghraoui and Roux 2008). With its advantages of high precision, short scan times, low radiation dose, and stable calibration, dual-energy x-ray absorptiometry (DXA) has been established by the World Health Organization (WHO) as the technique of reference for assessing bone mineral density (BMD) in postmenopausal women and based the definitions of osteopenia and osteoporosis on its results. Recently, efficient therapeutic options for treatment of osteoporosis have been developed which create possibilities of effective intervention. Therefore, screening for and treatment of osteoporosis are widely practised in postmenopausal women and in people with an increased risk of osteoporosis because of underlying diseases (e.g. chronic rheumatic diseases especially when treated by corticosteroids)(Phillipov, Seaborn et al. 2001). Moreover, BMD measurement is needed to select patients for osteoporosis treatment, as there is no proof that drugs for osteoporosis (other than hormone replacement therapy [HRT]) are beneficial in women with clinical risk factors for fractures but normal BMD values.
It has also become more and more common to perform a second DXA measurement to monitor BMD status or the effect of therapeutic intervention. When a second measurement is performed on a patient, the clinician needs to distinguish between a true change in BMD and a random fluctuation related to variability in the measurement procedure. The reproducibility of DXA measurements is claimed to be good. Such variability is due to multiple causes, such as device errors, technician variability, patients' movements, and variation due to other unpredictable sources (Nguyen, Sambrook et al. 1997; Lodder, Lems et al. 2004).
The precision error is usually expressed as the coefficient of variation (CV), which is the ratio of the standard deviation (SD) to the mean of the measurements, although several other statistics to express reproducibility exist such as the smallest detectable difference (SDD) or the least significant change (LSC). The SDD represents a cut-off that can be measured in an individual and is usually considered more useful than the CV in clinical practice (Fuleihan, Testa et al. 1995; Ravaud, Reny et al. 1999).
Monitoring DXA Measurement in Clinical Practice 21
two observations for each subject, the standard deviation of the differences (SDdiff) estimates the within variability of the measurements. Most disagreements between measurements are expected to be between limits called ''limits of agreement'' defined as dz(1-a/2) SDdiff where d is the mean difference between the pairs of measurements and z(1-a/2) is the 100(1-a/2)th centile of the normal distribution. The value d is an estimate of the mean systematic bias of measurement 1 to measurement 2. d is expected to be 0 because a true change in BMD is not assumed to occur during the interval between the two BMD measurements. Defining a to be 5%, the limits of agreement are +1.96SDdiff and -1.96SDdiff. Thus, about twice the standard deviation (SD) of the difference scores gives the 95% limits of agreement for the two measurements by the machine. A test is considered to be capable of detecting a difference, in
**3. Clinical implications of bone mineral density reproducibility measurement** In clinical practice, two absolute values (g/cm2) have to be compared, rather than two percentages (T-scores). When serial measurements are obtained in a patient, only changes greater than the LSC (in %) or the SDD (in g/cm2) can be ascribed to treatment effects.
We studied recently the in vivo short term variability of BMD measurement by DXA in three groups of subjects with a wide range of BMD values: healthy young volunteers, postmenopausal women and patients with chronic rheumatic diseases (most of them taking corticosteroids). In all studied subjects, reproducibility expressed by different means was good and independent from clinical and BMD status. Thus, the clinician interpreting a repeated DXA scan of a subject should be aware that a BMD change exceeding the LSC is significant, in our centre arising from a BMD change of at least 3.56% at the total hip and 5.60% at the spine. Expressed as SDD, a BMD change should exceed 0.02 g/cm2 at the total hip and 0.04 g/cm2 at the spine before it can be considered a significant change(El Maghraoui, Do Santos Zounon et al. 2005). Indeed, it has become usual to perform repeated DXA measurement: in postmenopausal women to monitor efficacy of treatment and in patients with chronic rheumatic diseases where high prevalence of bone loss has been demonstrated (Maillefert, Aho et al. 2001; Johnson, Petkov et al. 2005) especially when long term corticosteroid therapy is used. In the reports published, variability is usually expressed as CV and the figures for short term variability are lower than the ones we found [7-9]. However, two studies showed variability data more in line with our results. In Ravaud et al.(Ravaud, Reny et al. 1999) study, two samples of healthy (n=70) and elderly (n=57) postmenopausal women showed a CV (%) of 0.9 and 1.8, respectively, at the spine, and of 0.9 and 2.3, respectively, at the total hip. Eastell showed an LSC (%) of 5.4 at the lumbar spine and 8 at the total hip, respectively, in osteoporotic postmenopausal women (Eastell 1996). It has been suggested that the varying results of reproducibility studies might be explained by the ''population'' investigated; a phantom and healthy young subjects are likely to show more favourable variability than postmenopausal women, possibly in part because of easier positioning for measurement(Gluer, Blake et al. 1995). However, our study failed to show better variability, expressed as CV (%), in young healthy volunteers (El Maghraoui, Do Santos Zounon et al. 2005). Another reason advocated was that osteoarthritis in postmenopausal women may contribute to poorer variability than found in healthy young subjects. The SDD values found in our study were comparable to the figures presented by Ravaud et al. (Ravaud, Reny et al. 1999). In the first group of postmenopausal
absolute units, of at least the magnitude of the limits of agreement.
Smaller changes may be related to measurement error.
#### **2. Methods of bone mineral density reproducibility measurement**
Precision errors are evaluated by performing repeated scans on a representative set of individuals to characterize the reproducibility of the technique. Most published studies examine the short-term precision error, based on repeated measurements of each subject performed over a time period of no more than 2 weeks. Over such a short period, no true change in BMD is expected.
#### **2.1 The coefficient of variation (CV)**
The CV, the most commonly presented measure for BMD variability, is the SD corrected for the mean of paired measurements. CV, expressed as a percentage, is calculated as CV (%) = (√((∑(a-b)2)/2n))/((Ma+Mb)/2)x100 where a and b are the first and the second measurement, Ma and Mb are the mean values for the two groups, and n is the number of paired observations.
Reproducibility is far better for BMD measurement than for most laboratory tests. Reproducibility expressed by the CV is usually 1–2% at the spine on anteroposterior images and 2–3% at the proximal femur in individuals with normal BMD values; the difference between the two sites is ascribable to greater difficulties with repositioning and examining the femur, as compared to the spine. However, these data obtained under nearly experimental conditions may not apply to everyday clinical practice. Reproducibility depends heavily on quality assurance factors, including tests to control the quality and performance of the machine, as well as the experience of the operator. Assessment of machine performance requires daily scanning of a phantom (which may be anthropomorphic or not), followed by calculation of the in vitro coefficient of variation (CV), which serves to evaluate short-term and longterm performance and to detect drift in measurement accuracy. These in vitro data, however, do not necessarily reflect in vivo reproducibility, which should be evaluated at each measurement centre. Measurements are obtained either three times in each of 15 patients or twice in each of 30 patients, and the CV (m/*r*) is calculated from the mean (m) and standard deviation *(r)* of these repeated measurements. The CV is expressed as a percentage and depends on mean BMD values(Phillipov, Seaborn et al. 2001). The standard deviation reflects measurement error, which is a characteristic of machine performance and is independent from the value measured.
#### **2.2 The least significant change (LSC)**
For two point measurements in time, a BMD change exceeding 2√2 times the precision error (PE) of a technique is considered a significant change (with 95% confidence): the corresponding change criterion has been termed "least significant change" or LSC. LSC = 2.8 x PE; where PE is the largest precision error of the technique used (or more easily the CV expressed in percentage). This smallest change that is considered statistically significant is also expressed in percentage(Gluer 1999).
#### **2.3 The smallest detectable difference (SDD)**
The measurement error can be calculated using Bland and Altman's 95% limits of agreement method(Bland and Altman 1986). Precision expressed by this method gives an absolute and metric estimate of random measurement error, also called SDD. In this case, where there are
Precision errors are evaluated by performing repeated scans on a representative set of individuals to characterize the reproducibility of the technique. Most published studies examine the short-term precision error, based on repeated measurements of each subject performed over a time period of no more than 2 weeks. Over such a short period, no true
The CV, the most commonly presented measure for BMD variability, is the SD corrected for the mean of paired measurements. CV, expressed as a percentage, is calculated as CV (%) = (√((∑(a-b)2)/2n))/((Ma+Mb)/2)x100 where a and b are the first and the second measurement, Ma and Mb are the mean values for the two groups, and n is the number of paired
Reproducibility is far better for BMD measurement than for most laboratory tests. Reproducibility expressed by the CV is usually 1–2% at the spine on anteroposterior images and 2–3% at the proximal femur in individuals with normal BMD values; the difference between the two sites is ascribable to greater difficulties with repositioning and examining the femur, as compared to the spine. However, these data obtained under nearly experimental conditions may not apply to everyday clinical practice. Reproducibility depends heavily on quality assurance factors, including tests to control the quality and performance of the machine, as well as the experience of the operator. Assessment of machine performance requires daily scanning of a phantom (which may be anthropomorphic or not), followed by calculation of the in vitro coefficient of variation (CV), which serves to evaluate short-term and longterm performance and to detect drift in measurement accuracy. These in vitro data, however, do not necessarily reflect in vivo reproducibility, which should be evaluated at each measurement centre. Measurements are obtained either three times in each of 15 patients or twice in each of 30 patients, and the CV (m/*r*) is calculated from the mean (m) and standard deviation *(r)* of these repeated measurements. The CV is expressed as a percentage and depends on mean BMD values(Phillipov, Seaborn et al. 2001). The standard deviation reflects measurement error, which is a characteristic of machine performance and is independent from
For two point measurements in time, a BMD change exceeding 2√2 times the precision error (PE) of a technique is considered a significant change (with 95% confidence): the corresponding change criterion has been termed "least significant change" or LSC. LSC = 2.8 x PE; where PE is the largest precision error of the technique used (or more easily the CV expressed in percentage). This smallest change that is considered statistically significant is
The measurement error can be calculated using Bland and Altman's 95% limits of agreement method(Bland and Altman 1986). Precision expressed by this method gives an absolute and metric estimate of random measurement error, also called SDD. In this case, where there are
**2. Methods of bone mineral density reproducibility measurement**
change in BMD is expected.
observations.
the value measured.
**2.2 The least significant change (LSC)**
also expressed in percentage(Gluer 1999).
**2.3 The smallest detectable difference (SDD)**
**2.1 The coefficient of variation (CV)**
two observations for each subject, the standard deviation of the differences (SDdiff) estimates the within variability of the measurements. Most disagreements between measurements are expected to be between limits called ''limits of agreement'' defined as dz(1-a/2) SDdiff where d is the mean difference between the pairs of measurements and z(1-a/2) is the 100(1-a/2)th centile of the normal distribution. The value d is an estimate of the mean systematic bias of measurement 1 to measurement 2. d is expected to be 0 because a true change in BMD is not assumed to occur during the interval between the two BMD measurements. Defining a to be 5%, the limits of agreement are +1.96SDdiff and -1.96SDdiff. Thus, about twice the standard deviation (SD) of the difference scores gives the 95% limits of agreement for the two measurements by the machine. A test is considered to be capable of detecting a difference, in absolute units, of at least the magnitude of the limits of agreement.
#### **3. Clinical implications of bone mineral density reproducibility measurement**
In clinical practice, two absolute values (g/cm2) have to be compared, rather than two percentages (T-scores). When serial measurements are obtained in a patient, only changes greater than the LSC (in %) or the SDD (in g/cm2) can be ascribed to treatment effects. Smaller changes may be related to measurement error.
We studied recently the in vivo short term variability of BMD measurement by DXA in three groups of subjects with a wide range of BMD values: healthy young volunteers, postmenopausal women and patients with chronic rheumatic diseases (most of them taking corticosteroids). In all studied subjects, reproducibility expressed by different means was good and independent from clinical and BMD status. Thus, the clinician interpreting a repeated DXA scan of a subject should be aware that a BMD change exceeding the LSC is significant, in our centre arising from a BMD change of at least 3.56% at the total hip and 5.60% at the spine. Expressed as SDD, a BMD change should exceed 0.02 g/cm2 at the total hip and 0.04 g/cm2 at the spine before it can be considered a significant change(El Maghraoui, Do Santos Zounon et al. 2005). Indeed, it has become usual to perform repeated DXA measurement: in postmenopausal women to monitor efficacy of treatment and in patients with chronic rheumatic diseases where high prevalence of bone loss has been demonstrated (Maillefert, Aho et al. 2001; Johnson, Petkov et al. 2005) especially when long term corticosteroid therapy is used. In the reports published, variability is usually expressed as CV and the figures for short term variability are lower than the ones we found [7-9]. However, two studies showed variability data more in line with our results. In Ravaud et al.(Ravaud, Reny et al. 1999) study, two samples of healthy (n=70) and elderly (n=57) postmenopausal women showed a CV (%) of 0.9 and 1.8, respectively, at the spine, and of 0.9 and 2.3, respectively, at the total hip. Eastell showed an LSC (%) of 5.4 at the lumbar spine and 8 at the total hip, respectively, in osteoporotic postmenopausal women (Eastell 1996). It has been suggested that the varying results of reproducibility studies might be explained by the ''population'' investigated; a phantom and healthy young subjects are likely to show more favourable variability than postmenopausal women, possibly in part because of easier positioning for measurement(Gluer, Blake et al. 1995). However, our study failed to show better variability, expressed as CV (%), in young healthy volunteers (El Maghraoui, Do Santos Zounon et al. 2005). Another reason advocated was that osteoarthritis in postmenopausal women may contribute to poorer variability than found in healthy young subjects. The SDD values found in our study were comparable to the figures presented by Ravaud et al. (Ravaud, Reny et al. 1999). In the first group of postmenopausal
Monitoring DXA Measurement in Clinical Practice 23
measured. Thus, these results enhance to encourage the use of the measurement of both hips
(a)
0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 **Mean score**
(b)
0 0,511,5 2 **Mean score**
Fig. 1. Graph of the difference score against the mean score of the total hip (a) and the lumbar spine (b) BMD measurements (g/cm2) in our centre [17] using Bland and Altman
method. The outermost (solid) lines represent the SDD.
to improve the reproducibility of DXA at this site.
**Difference score**
0,02
0,04
0,06
0
**Diffrence score**
0,02
0,04
0,06
0
women (mean age 53 years) they describe, the SDD was 0.02 (g/cm2) at the total hip and 0.02 at the lumbar spine. In the second group described, women with a mean age of 80 years, these figures were 0.04 and 0.04, respectively. In Lodder et al.(Lodder, Lems et al. 2004) study (Ninety five women, mean age 59.9 years), the SDD was 0.04 (g/cm2) at the total hip and 0.05 at the lumbar spine. The SDD values of the children studied in this study tended to be lower than the values in the postmenopausal women (table I). Using the SDD, one can state that a (BMD) change larger than the figure found is a true (BMD) change in 95% of the cases. The characteristics of the Bland and Altman method thus allow direct insight into the variability of the measurement under study (figure 1).
It has been shown that reproducibility expressed using the SDD is independent of the BMD value whereas reproducibility expressed using the CV or the derived LSC depend on the BMD value. Ravaud et al.(Ravaud, Reny et al. 1999) reported that using SD, the values of the cut-offs are 0.024, 0.030, 0.020, and 0.021 g/cm2 for postmenopausal women aged 70 years and 0.040, 0.033, 0.033, and 0.038 g/cm2 for postmenopausal osteoporotic women aged >70 years at the spine, femoral neck, greater trochanter, and total hip, respectively. Using CV, cut-offs vary depending on the BMD level. In postmenopausal women aged >70 years, for a BMD level between 0.600 g/cm2 and 1.000 g/cm2, the cut-offs derived from CV vary between 0.015 g/cm2 and 0.024 g/cm2, 0.024 g/cm2 and 0.041 g/cm2, 0.018 g/cm2 and 0.030 g/cm2, 0.015 g/cm2 and 0.025 g/cm2 for the spine femoral neck, greater trochanter, and total hip, respectively. In postmenopausal osteoporotic women aged >70 years, for the same range of BMD level, cut-offs vary between 0.031 g/cm2 and 0.051 g/cm2, 0.038 g/cm2 and 0.063 g/cm2, 0.043 g/cm2 and 0.071 g/cm2, 0.038 g/cm2 and 0.063 g/cm2 for the same bone sites. Consequently, to express variability on a percentage basis using CV leads to underestimate variability in patients with low BMD and to overestimate variability in patients with high BMD. Previous reports in the literature, as well as Ravaud, Lodder's data and our data (table 1) demonstrate that absolute precision errors derived from SD are constant across a wide range of BMD values and independent of the level of BMD. Because of therapeutic consequences, the clinician should be especially careful in judging an apparent BMD change in patients with osteoporosis. Influence of age on BMD reproducibility is controversial. Previous studies have suggested that BMD measurement errors were independent of age even some studies suggested that SDD may vary in extreme ages (children and elderly) probably because of age-related factors other than BMD. However, a few data exist for reproducibility of DXA in women over 70. Ravaud et al. data, as well as those of Fuleihan (Fuleihan, Testa et al. 1995), and Maggio et al. (Maggio, McCloskey et al. 1998) show that the measurement error is greater in older osteoporotic subjects. Several factors such as difficulties in repositioning could explain the increase of measurement error in this kind of patients. Therefore, the use of the SDD in the evaluation of an apparent BMD change gives a more conservative approach than the use of the CV at low BMD. Because of its independence from the BMD level and its expression in absolute units, the SDD is a preferable measure for use in daily clinical practice as compared with the CV and the derived LSC.
In contrast with all previous publications about DXA reproducibility, we found in our centre better results for the hip BMD variability than the lumbar spine. This is due to the fact that our study was the first to use the mean measure of the two femurs (dual femur). In this study, we showed in a group of young healthy volunteers that the SDD was ±0.0218 g/cm2 when both femurs were measured whereas it was ±0.0339 g/cm2 when only one femur was
women (mean age 53 years) they describe, the SDD was 0.02 (g/cm2) at the total hip and 0.02 at the lumbar spine. In the second group described, women with a mean age of 80 years, these figures were 0.04 and 0.04, respectively. In Lodder et al.(Lodder, Lems et al. 2004) study (Ninety five women, mean age 59.9 years), the SDD was 0.04 (g/cm2) at the total hip and 0.05 at the lumbar spine. The SDD values of the children studied in this study tended to be lower than the values in the postmenopausal women (table I). Using the SDD, one can state that a (BMD) change larger than the figure found is a true (BMD) change in 95% of the cases. The characteristics of the Bland and Altman method thus allow direct
It has been shown that reproducibility expressed using the SDD is independent of the BMD value whereas reproducibility expressed using the CV or the derived LSC depend on the BMD value. Ravaud et al.(Ravaud, Reny et al. 1999) reported that using SD, the values of the cut-offs are 0.024, 0.030, 0.020, and 0.021 g/cm2 for postmenopausal women aged 70 years and 0.040, 0.033, 0.033, and 0.038 g/cm2 for postmenopausal osteoporotic women aged >70 years at the spine, femoral neck, greater trochanter, and total hip, respectively. Using CV, cut-offs vary depending on the BMD level. In postmenopausal women aged >70 years, for a BMD level between 0.600 g/cm2 and 1.000 g/cm2, the cut-offs derived from CV vary between 0.015 g/cm2 and 0.024 g/cm2, 0.024 g/cm2 and 0.041 g/cm2, 0.018 g/cm2 and 0.030 g/cm2, 0.015 g/cm2 and 0.025 g/cm2 for the spine femoral neck, greater trochanter, and total hip, respectively. In postmenopausal osteoporotic women aged >70 years, for the same range of BMD level, cut-offs vary between 0.031 g/cm2 and 0.051 g/cm2, 0.038 g/cm2 and 0.063 g/cm2, 0.043 g/cm2 and 0.071 g/cm2, 0.038 g/cm2 and 0.063 g/cm2 for the same bone sites. Consequently, to express variability on a percentage basis using CV leads to underestimate variability in patients with low BMD and to overestimate variability in patients with high BMD. Previous reports in the literature, as well as Ravaud, Lodder's data and our data (table 1) demonstrate that absolute precision errors derived from SD are constant across a wide range of BMD values and independent of the level of BMD. Because of therapeutic consequences, the clinician should be especially careful in judging an apparent BMD change in patients with osteoporosis. Influence of age on BMD reproducibility is controversial. Previous studies have suggested that BMD measurement errors were independent of age even some studies suggested that SDD may vary in extreme ages (children and elderly) probably because of age-related factors other than BMD. However, a few data exist for reproducibility of DXA in women over 70. Ravaud et al. data, as well as those of Fuleihan (Fuleihan, Testa et al. 1995), and Maggio et al. (Maggio, McCloskey et al. 1998) show that the measurement error is greater in older osteoporotic subjects. Several factors such as difficulties in repositioning could explain the increase of measurement error in this kind of patients. Therefore, the use of the SDD in the evaluation of an apparent BMD change gives a more conservative approach than the use of the CV at low BMD. Because of its independence from the BMD level and its expression in absolute units, the SDD is a preferable measure for use in daily clinical practice as compared with the
In contrast with all previous publications about DXA reproducibility, we found in our centre better results for the hip BMD variability than the lumbar spine. This is due to the fact that our study was the first to use the mean measure of the two femurs (dual femur). In this study, we showed in a group of young healthy volunteers that the SDD was ±0.0218 g/cm2 when both femurs were measured whereas it was ±0.0339 g/cm2 when only one femur was
insight into the variability of the measurement under study (figure 1).
CV and the derived LSC.
measured. Thus, these results enhance to encourage the use of the measurement of both hips to improve the reproducibility of DXA at this site.
(b) 0 0,511,5 2 **Mean score**
Fig. 1. Graph of the difference score against the mean score of the total hip (a) and the lumbar spine (b) BMD measurements (g/cm2) in our centre [17] using Bland and Altman method. The outermost (solid) lines represent the SDD.
Mean difference, mean of the difference between the first and the second BMD measurement; SD difference, SD of the difference between the first and the second BMD measurement; SDD, smallest detectable difference (g/cm2); CV, coefficient of variation (%); LSC, least significant change (%); ICC, intraclass correlation coefficient. Table 1. Comparison of BMD measurement reproducibility evaluation in two studies including various groups of patients.
Monitoring DXA Measurement in Clinical Practice 25
Although the variability as expressed by the CV, and especially the SDD, is reassuring, showing good short term variability at group level, the wide range of the differences in BMD and the derived T-scores indicates considerable individual differences between two
indicates that in some patients the diagnosis, based on the diagnostic thresholds of the
In summary, reproducibility of BMD measurement by DXA in different kinds of patients (postmenopausal women, patients with chronic rheumatic diseases, elderly…) expressed by different means is good at a group level. However, the clinician must remain aware that an apparent BMD change in an individual patient may represent a precision error. At each measurement centre, the SDD should be calculated from in vivo reproducibility data. In clinical practice, the SDD should be used to estimate the significance of observed changes, in
The first factor is the time interval between two measurements in the same patient which must be long enough to allow occurrence of a change greater than the SDD or the LSC. Therefore, it depends on the expected rate of change in BMD measurement (which varies according to whether the measurement site is composed predominantly of trabecular or of cortical bone) and the reproducibility of BMD measurement at that site. Thus, in clinical practice, a treatment-induced BMD increase can only be detected in general after 2 years. However, in patients receiving long term steroid therapy, the changes in BMD may be so important that they can be detected at 1 year. Thus, although the spine may not be the best site for the diagnosis of osteoporosis given the high prevalence of spinal degenerative disease, it is the most sensitive site for detecting changes over time. However, our study showed that measurement of both femurs (called "dual femur" in Lunar machines)
In another side, the changes in BMD measurements are influenced by the ability of osteoporosis treatments to increase the BMD at the different skeletal sites. For some treatments such as teriparatide and the more potent bisphosphonates, statistically significant changes in spine BMD occur on time scales of 1 to 2 years in the majority of patients, although for other treatments, such as raloxifene, the changes are often not large enough to be statistically significant. Recently, the strontium ranelate trials show BMD increases at the spine and hip 2-5 times larger than those for BPs and SERMs, and comparable or greater than with teriparatide. However, it is important to appreciate that much of this effect is due to the higher atomic number of strontium compared with calcium. Thus, strontium in bone attenuates X-rays much more strongly than calcium, and BMD is overestimated compared with the true mass of bone mineral present. This effect may persist for many years after the patient stops treatment and may affect the relationship between
Thus, treatment dosages cannot be adjusted on the basis of BMD changes. Moreover, there is no proof that repeating BMD measurements improves compliance, as most patients discontinue antiresorptive medications after a few months because of administration
constraints, side effects, cost of medications or lack of interest.
ΔT scores, for example,
consecutive BMD measurements in some patients. The range in
WHO, would change owing to the measurement variability.
**5. Other factors influencing DXA monitoring**
increases the reproducibility at this site.
BMD and fracture risk.
absolute values.
Mean difference, mean of the difference between the first and the second BMD measurement; SD difference, SD of the difference between the first and the second BMD measurement; SDD, smallest detectable difference (g/cm2); CV, coefficient of variation (%); LSC, least significant change (%); ICC, intraclass correlation coefficient. Table 1. Comparison of BMD measurement reproducibility evaluation in two studies including various groups of patients.
Although the variability as expressed by the CV, and especially the SDD, is reassuring, showing good short term variability at group level, the wide range of the differences in BMD and the derived T-scores indicates considerable individual differences between two consecutive BMD measurements in some patients. The range in ΔT scores, for example, indicates that in some patients the diagnosis, based on the diagnostic thresholds of the WHO, would change owing to the measurement variability.
In summary, reproducibility of BMD measurement by DXA in different kinds of patients (postmenopausal women, patients with chronic rheumatic diseases, elderly…) expressed by different means is good at a group level. However, the clinician must remain aware that an apparent BMD change in an individual patient may represent a precision error. At each measurement centre, the SDD should be calculated from in vivo reproducibility data. In clinical practice, the SDD should be used to estimate the significance of observed changes, in absolute values.
#### **5. Other factors influencing DXA monitoring**
The first factor is the time interval between two measurements in the same patient which must be long enough to allow occurrence of a change greater than the SDD or the LSC. Therefore, it depends on the expected rate of change in BMD measurement (which varies according to whether the measurement site is composed predominantly of trabecular or of cortical bone) and the reproducibility of BMD measurement at that site. Thus, in clinical practice, a treatment-induced BMD increase can only be detected in general after 2 years. However, in patients receiving long term steroid therapy, the changes in BMD may be so important that they can be detected at 1 year. Thus, although the spine may not be the best site for the diagnosis of osteoporosis given the high prevalence of spinal degenerative disease, it is the most sensitive site for detecting changes over time. However, our study showed that measurement of both femurs (called "dual femur" in Lunar machines) increases the reproducibility at this site.
In another side, the changes in BMD measurements are influenced by the ability of osteoporosis treatments to increase the BMD at the different skeletal sites. For some treatments such as teriparatide and the more potent bisphosphonates, statistically significant changes in spine BMD occur on time scales of 1 to 2 years in the majority of patients, although for other treatments, such as raloxifene, the changes are often not large enough to be statistically significant. Recently, the strontium ranelate trials show BMD increases at the spine and hip 2-5 times larger than those for BPs and SERMs, and comparable or greater than with teriparatide. However, it is important to appreciate that much of this effect is due to the higher atomic number of strontium compared with calcium. Thus, strontium in bone attenuates X-rays much more strongly than calcium, and BMD is overestimated compared with the true mass of bone mineral present. This effect may persist for many years after the patient stops treatment and may affect the relationship between BMD and fracture risk.
Thus, treatment dosages cannot be adjusted on the basis of BMD changes. Moreover, there is no proof that repeating BMD measurements improves compliance, as most patients discontinue antiresorptive medications after a few months because of administration constraints, side effects, cost of medications or lack of interest.
Monitoring DXA Measurement in Clinical Practice 27
Bland, J. M. and D. G. Altman (1986). "Statistical methods for assessing agreement between
Cummings, S. R., D. B. Karpf, et al. (2002). "Improvement in spine bone density and
Eastell, R. (1996). "Assessment of bone density and bone loss." *Osteoporos Int* 6 Suppl 2: 3-5. El Maghraoui, A. and C. Roux (2008). "DXA scanning in clinical practice." *QJM* 101(8): 605-
El Maghraoui, A., A. A. Do Santos Zounon, et al. (2005). "Reproducibility of bone mineral
El Maghraoui, A., L. Achemlal, et al. (2006). "Monitoring of dual-energy X-ray absorptiometry measurement in clinical practice." *J Clin Densitom* 9(3): 281-6. Fuleihan, G. E., M. A. Testa, et al. (1995). "Reproducibility of DXA absorptiometry: a model
Gluer, C. C. (1999). "Monitoring skeletal changes by radiological techniques." *J Bone Miner*
Gluer, C. C., G. Blake, et al. (1995). "Accurate assessment of precision errors: how to measure the reproducibility of bone densitometry techniques." *Osteoporos Int* 5(4): 262-70. Hochberg, M. C., P. D. Ross, et al. (1999). "Larger increases in bone mineral density during
Hochberg, M. C., S. Greenspan, et al. (2002). "Changes in bone density and turnover explain
Johnson, S. L., V. I. Petkov, et al. (2005). "Improving osteoporosis management in patients
Li, Z., M. P. Meredith, et al. (2001). "A method to assess the proportion of treatment effect
Lodder, M. C., W. F. Lems, et al. (2004). "Reproducibility of bone mineral density
Maggio, D., E. V. McCloskey, et al. (1998). "Short-term reproducibility of proximal femur
Maillefert, J. F., L. S. Aho, et al. (2001). "Changes in bone density in patients with ankylosing spondylitis: a two-year follow-up study." *Osteoporos Int* 12(7): 605-9. Nguyen, T. V., P. N. Sambrook, et al. (1997). "Sources of variability in bone mineral density
Phillipov, G., C. J. Seaborn, et al. (2001). "Reproducibility of DXA: potential impact on serial measurements and misclassification of osteoporosis." *Osteoporos Int* 12(1): 49-54. Ravaud, P., J. L. Reny, et al. (1999). "Individual smallest detectable difference in bone
Roux, C., P. Garnero, et al. (2005). "Recommendations for monitoring antiresorptive therapies in postmenopausal osteoporosis." *Joint Bone Spine* 72(1): 26-31.
measurements: implications for study design and analysis of bone loss." *J Bone*
with antiresorptive agents." *J Clin Endocrinol Metab* 87(4): 1586-92.
explained by a surrogate endpoint." *Stat Med* 20(21): 3175-88.
measurement in daily practice." *Ann Rheum Dis* 63(3): 285-9.
bone mineral density in the elderly." *Calcif Tissue Int* 63(4): 296-9.
mineral density measurements." *J Bone Miner Res* 14(8): 1449-56.
alendronate therapy are associated with a lower risk of new vertebral fractures in women with postmenopausal osteoporosis. Fracture Intervention Trial Research
the reductions in incidence of nonvertebral fractures that occur during treatment
reduction in risk of vertebral fractures during treatment with antiresorptive drugs."
density measurements using dual X-ray absorptiometry in daily clinical practice."
two methods of clinical measurement." *Lancet* 1(8476): 307-10.
for bone loss estimates." *J Bone Miner Res* 10(7): 1004-14.
**7. References**
17.
*Am J Med* 112(4): 281-9.
*Osteoporos Int* 16(12): 1742-8.
Group." *Arthritis Rheum* 42(6): 1246-54.
with fractures." *Osteoporos Int* 16(9): 1079-85.
*Res* 14(11): 1952-62.
*Miner Res* 12(1): 124-35.
Above all, BMD is used as a surrogate marker for the fracture risk, yet BMD increases do not reliably reflect a reduction in the fracture risk. Although bisphosphonates, raloxifene, and strontium have not been compared in the same study, they seem to produce comparable reductions in the risk of vertebral fractures, of about 30–50%, whereas BMD changes differ markedly across medications. Studies have shown that BMD gains explain only a small proportion of the vertebral fracture risk reduction: 28% with risedronate(Li, Meredith et al. 2001), 16% with alendronate(Cummings, Karpf et al. 2002), and 4% with raloxifene(Sarkar, Mitlak et al. 2002). It has been suggested that the percentage of BMD change may be related to the change in the relative risk of fracture(Wasnich and Miller 2000). In one study, a linear relationship was found between these two parameters, but a 1% increase in spinal BMD was associated with an only 3% decrease in the relative risk of vertebral fracture(Cummings, Karpf et al. 2002). For peripheral fractures, in contrast, the risk reduction is clearly related to the BMD gain(Hochberg, Greenspan et al. 2002). Common sense indicates that a BMD increase during treatment should be preferable over a BMD decrease. However, data showing that the fracture risk may decrease despite a reduction in BMD have been reported (Watts, Geusens et al. 2005). It has also been shown that the fracture risk was more heavily dependent on BMD at baseline than on BMD changes during treatment (Hochberg, Ross et al. 1999).
#### **6. Conclusion**
Serial BMD measurements can be used to monitor current antiresorptive treatments (raloxifene, bisphosphonates or strontium ranelate). However, adequate quality-control procedures must be used(Roux, Garnero et al. 2005). Measurement error must be considered when evaluating serial assessments. A clear understanding of the interpretation of serial measurements and the statistical principles impacting upon their interpretation is necessary to determine whether a change is real and not simply random fluctuation. It is inadequate to simply use the manufacturer's default precision error, which may underestimate the precision error in the clinical setting. Thus, every centre should calculate its own precision error from in vivo reproducibility data. International societies interested in osteoporosis diagnosis and management such as the International Society for Clinical Densitometry or the International Osteoporosis Foundation should add to their guidelines at least two recommendations about DXA monitoring highlighted in this paper: the measurement of both hips improves the reproducibility at this site and DXA measurement centres should determine and use the individual SDD. Indeed, the use of the SDD is preferable to the use of the CV and LSC because of its independence from BMD level and its expression in absolute units. The exact definition and advices for the measure and use of these parameters in clinical practice should be clearly explained. It is clear that the choice of the optimum site for performing follow-up scans depends on the ratio of the BMD treatment effect to the precision of the measurements. The larger this ratio, the more statistically significant the observed changes are likely to be. Actually, all data agree in showing that the spine is the optimum site. In clinical practice, BMD measurements have to be spaced at least 2 years apart. The main goal of serial BMD measurement is to check that no further bone loss has occurred; estimation of BMD gains is the secondary objective. This should be explained to the patients, many of whom expect to recover normal BMD values.
#### **7. References**
26 Dual Energy X-Ray Absorptiometry
Above all, BMD is used as a surrogate marker for the fracture risk, yet BMD increases do not reliably reflect a reduction in the fracture risk. Although bisphosphonates, raloxifene, and strontium have not been compared in the same study, they seem to produce comparable reductions in the risk of vertebral fractures, of about 30–50%, whereas BMD changes differ markedly across medications. Studies have shown that BMD gains explain only a small proportion of the vertebral fracture risk reduction: 28% with risedronate(Li, Meredith et al. 2001), 16% with alendronate(Cummings, Karpf et al. 2002), and 4% with raloxifene(Sarkar, Mitlak et al. 2002). It has been suggested that the percentage of BMD change may be related to the change in the relative risk of fracture(Wasnich and Miller 2000). In one study, a linear relationship was found between these two parameters, but a 1% increase in spinal BMD was associated with an only 3% decrease in the relative risk of vertebral fracture(Cummings, Karpf et al. 2002). For peripheral fractures, in contrast, the risk reduction is clearly related to the BMD gain(Hochberg, Greenspan et al. 2002). Common sense indicates that a BMD increase during treatment should be preferable over a BMD decrease. However, data showing that the fracture risk may decrease despite a reduction in BMD have been reported (Watts, Geusens et al. 2005). It has also been shown that the fracture risk was more heavily dependent on BMD at baseline than on BMD changes during treatment (Hochberg, Ross et
Serial BMD measurements can be used to monitor current antiresorptive treatments (raloxifene, bisphosphonates or strontium ranelate). However, adequate quality-control procedures must be used(Roux, Garnero et al. 2005). Measurement error must be considered when evaluating serial assessments. A clear understanding of the interpretation of serial measurements and the statistical principles impacting upon their interpretation is necessary to determine whether a change is real and not simply random fluctuation. It is inadequate to simply use the manufacturer's default precision error, which may underestimate the precision error in the clinical setting. Thus, every centre should calculate its own precision error from in vivo reproducibility data. International societies interested in osteoporosis diagnosis and management such as the International Society for Clinical Densitometry or the International Osteoporosis Foundation should add to their guidelines at least two recommendations about DXA monitoring highlighted in this paper: the measurement of both hips improves the reproducibility at this site and DXA measurement centres should determine and use the individual SDD. Indeed, the use of the SDD is preferable to the use of the CV and LSC because of its independence from BMD level and its expression in absolute units. The exact definition and advices for the measure and use of these parameters in clinical practice should be clearly explained. It is clear that the choice of the optimum site for performing follow-up scans depends on the ratio of the BMD treatment effect to the precision of the measurements. The larger this ratio, the more statistically significant the observed changes are likely to be. Actually, all data agree in showing that the spine is the optimum site. In clinical practice, BMD measurements have to be spaced at least 2 years apart. The main goal of serial BMD measurement is to check that no further bone loss has occurred; estimation of BMD gains is the secondary objective. This should be explained to
the patients, many of whom expect to recover normal BMD values.
al. 1999).
**6. Conclusion**
**3**
*Finland*
**Bone Loss and Seasonal Variation in Serial DXA**
Osteoporosis is a disease of increased skeleton fragility accompanied by low BMD and microarchitectural deterioration. Osteoporosis and bone fragility result in significant morbidity and medical and social costs (Dennison et al., 2005; Cummings et al., 2002). The risk of fractures is greater among women with low BMD although it explains only part of the increased fracture tendency among the elderly (National Osteoporosis Foundation, 1998).The diagnosis of osteoporosis is currently based on axial dual X-ray absorptiometry (DXA) measurements (National Osteoporosis Foundation 1998). In addition to being applicable for fracture prediction, axial DXA has a role in treatment monitoring protocols (Miller et al., 1996; Sowers et al., 1997). Furthermore, serial central DXA measurements have been used for research purposes to evaluate the risk- and preventive factors for postmenopausal bone loss (Burger et al., 1998; Hannan et al., 2000; Sirola et al., 2003) Perimenopausal bone loss rates of over -2 percent /year in spinal and over -1 percent /year in the femoral region have generally been reported (Harris and Dawson-Hughes, 1992; Pouilles et al., 1993; Pouilles et al., 1995; Prior et al., 1998; Ito et al., 1999). In postmenopausal women, age related bone loss continues at age specific rate after the initial fastening during
There are two forms of vitamin D, ergocalciferol (vitamin D2) and cholecalciferol (vitamin D3). Cholecalciferol is the metabolically active form of vitamin D. Vitamin D is produced with either the effect of ultraviolet B radiation or ingested with nutrition and the metabolically active form is produced in the kidneys. It has been suggested that there might be a seasonal variation in bone turnover as assessed with both BMD and biochemical markers (Rosen et al., 1994.; Storm et al., 1998; Rapuri et al., 2002). The sun-light related vitamin D production which varies according to season seems to contribute to this phenomenon (Rapuri et al., 2002; Dawson-Hughes et al., 1997). However, other studies have not demonstrated any such effect as measured either by BMD or by levels of bone marker
**1. Introduction**
the menopausal transition (Hansen et al., 1995).
**Densitometry – A Population-Based Study**
Jukka S. Jurvelin3, Marjo Tuppurainen1,4 and Heikki Kröger1,2,
Joonas Sirola1,2, Toni Rikkonen1, Risto Honkanen1,
*2Department of Orthopedics, Traumatology and Hand Surgery 3Department of Clinical Physiology & Nuclear Medicine*
*1University of Eastern Finland, Campus of Kuopio,*
*4Department of Obstetrics and Gynaecology*
*Bone and Cartilage Research Unit*
*2,3,4Kuopio University Hospital,*
### **Bone Loss and Seasonal Variation in Serial DXA Densitometry – A Population-Based Study**
Joonas Sirola1,2, Toni Rikkonen1, Risto Honkanen1, Jukka S. Jurvelin3, Marjo Tuppurainen1,4 and Heikki Kröger1,2, *1University of Eastern Finland, Campus of Kuopio, Bone and Cartilage Research Unit 2Department of Orthopedics, Traumatology and Hand Surgery 3Department of Clinical Physiology & Nuclear Medicine 4Department of Obstetrics and Gynaecology 2,3,4Kuopio University Hospital, Finland*
#### **1. Introduction**
28 Dual Energy X-Ray Absorptiometry
Sarkar, S., B. H. Mitlak, et al. (2002). "Relationships between bone mineral density and
Wasnich, R. D. and P. D. Miller (2000). "Antifracture efficacy of antiresorptive agents are related to changes in bone density." *J Clin Endocrinol Metab* 85(1): 231-6. Watts, N. B., P. Geusens, et al. (2005). "Relationship between changes in BMD and
10.
2097-104.
incident vertebral fracture risk with raloxifene therapy." *J Bone Miner Res* 17(1): 1-
nonvertebral fracture incidence associated with risedronate: reduction in risk of nonvertebral fracture is not related to change in BMD." *J Bone Miner Res* 20(12):
> Osteoporosis is a disease of increased skeleton fragility accompanied by low BMD and microarchitectural deterioration. Osteoporosis and bone fragility result in significant morbidity and medical and social costs (Dennison et al., 2005; Cummings et al., 2002). The risk of fractures is greater among women with low BMD although it explains only part of the increased fracture tendency among the elderly (National Osteoporosis Foundation, 1998).The diagnosis of osteoporosis is currently based on axial dual X-ray absorptiometry (DXA) measurements (National Osteoporosis Foundation 1998). In addition to being applicable for fracture prediction, axial DXA has a role in treatment monitoring protocols (Miller et al., 1996; Sowers et al., 1997). Furthermore, serial central DXA measurements have been used for research purposes to evaluate the risk- and preventive factors for postmenopausal bone loss (Burger et al., 1998; Hannan et al., 2000; Sirola et al., 2003) Perimenopausal bone loss rates of over -2 percent /year in spinal and over -1 percent /year in the femoral region have generally been reported (Harris and Dawson-Hughes, 1992; Pouilles et al., 1993; Pouilles et al., 1995; Prior et al., 1998; Ito et al., 1999). In postmenopausal women, age related bone loss continues at age specific rate after the initial fastening during the menopausal transition (Hansen et al., 1995).
> There are two forms of vitamin D, ergocalciferol (vitamin D2) and cholecalciferol (vitamin D3). Cholecalciferol is the metabolically active form of vitamin D. Vitamin D is produced with either the effect of ultraviolet B radiation or ingested with nutrition and the metabolically active form is produced in the kidneys. It has been suggested that there might be a seasonal variation in bone turnover as assessed with both BMD and biochemical markers (Rosen et al., 1994.; Storm et al., 1998; Rapuri et al., 2002). The sun-light related vitamin D production which varies according to season seems to contribute to this phenomenon (Rapuri et al., 2002; Dawson-Hughes et al., 1997). However, other studies have not demonstrated any such effect as measured either by BMD or by levels of bone marker
Bone Loss and Seasonal Variation in Serial DXA Densitometry – A Population-Based Study 31
offs was based on the assumption that highest BMDs (reflecting serum vitamin D concentration) would be present at late summer and early fall season (within group 2) and the lowest at late winter (within group 3) whereas group 1 would present an intermediate. Also, in order to reveal the greatest differences in seasonal BMD variation between two successive measurements a numeric value of a Seasonal Difference Index (SDI) was calculated as follows:
SDI = (Group number at baseline - Group number at follow-up) Accordingly, the numeric value of SDI varied from -2 to +2 (Table 1). In all, 521 (54.6 %) women were measured within the same season at both measurements (SDI group 0). In 212 (22.2 %) women the follow-up measurement was carried out in a later season (SDI groups -1 and -2) and in 221 (23.2 %) women in earlier season (SDI groups +1 and +2) than the baseline measurement. There was a significant variation in distribution of women into measurement seasons between and within baseline and follow-up measurements (Figure 1 ). Accordingly, any conclusions over differences between specific seasons were considered to
**Follow-up season (Group number) Baseline season (Group number)** January-April (1) May-August (2) September-December (3)
Table 1. Numeric values for the "Seasonal Difference Index" (SDI)a according to baseline
Women were divided into two groups according to their use of hormone therapy (HRT) (tablets and plasters) which was defined as the use of hormonal products for menopausal symptoms. *HRT users* (n=393) had used HRT continuously or occasionally during the follow-up regardless of whether or not they had used hormonal therapy before the baseline (14 women used HRT only before baseline, and were excluded in analysis on HRT effect). *HRT non-users* (n=547) had never used estrogen containing products aimed at postmenopausal therapy. In the OSTPRE cohort, the majority of HRT users were taking estrogen-progesterone combination products (56.2% of all HRT). No data was available on whether HRT was continuous or sequential. Forty-five percent of HRT users (occasional or continuous) had been treated with HRT also prior to baseline. The duration of HRT varied from one month to 7.5 years. The information about the use of hormonal products was obtained from the questionnaires. Comparison between self-reported use of HRT and the national prescription records of The Social Insurance Institution, Finland (KELA), for the whole OSTPRE cohort in 1996–2001 revealed that 97.8% of those who had received an oestrogen drug prescription reported HRT use in inquiries. On the other hand, in 25.5% of the self-reported non-users of HRT some oestrogen use (short-term, median 6.0 months) was
January-April (1) 0 -1 -2 May-August (2) 1 0 -1 September-December (3) 2 1 0
be precluded due to this uneven distribution.
*a) SDI=[Season Group at baseline]-[Season Group at follow-up]*
and follow-up DXA measurement season
**2.3 Other variables 2.3.1 Hormone therapy**
recorded (Sirola et al., 2003a).
compounds (Patel et al, 2001; Blumsohn et al., 2003)**.** Consequently, the effect of the season when densitometry was performed on bone density is still unresolved. Furthermore, the role of a seasonal difference between two distant follow-up bone density measurements in postmenopausal bone loss has not been studied and thus, nothing is known about the effect of this phenomenon on treatment monitoring or other longitudinal data collection. If present, such seasonal variation could have significant effect on the evaluation of prospective data.
The purpose of the present study was to investigate the effect of densitometry season on early postmenopausal BMD and bone loss in a subset of 954 Finnish women selected from the population-based Osteoporosis Risk Factor and Prevention (OSTPRE)-study.
### **2. Subjects and methods**
#### **2.1 Study population**
The study population was a randomly selected part of the prospective Kuopio Osteoporosis Risk Factor and Prevention (OSTPRE) study cohort. The OSTPRE cohort was established in 1989 and included all women born in 1932-1941 and who were resident in Kuopio Province, Finland (n=14 220). A postal inquiry including questions about health disorders, medications including HRT, gynaecologic history, nutritional habits, physical activity, lifestyle habits, and anthropometric information was sent to these women at baseline in 1989 (Honkanen et al, 1991). The 5-year (in 1994 follow-up questionnaires were sent to the 13 100 women who responded at baseline and responses were received from 11 954 at 5-year follow-up. The study protocol has been approved by the ethics committee of University of Kuopio and Kuopio University Hospital. Informed written consent from the participants was collected with the postal inquiries.
Of the 13 100 respondents in 1989, 11 055 (84.4 %) were willing to undergo DXA densitometry. A random sample of 2 362 women was selected for densitometry out of which 2025 women actually underwent the procedure during 1989-91. The questionnaire information was updated at the time of bone densitometry. In all, 1 873 women underwent both baseline (1989-91) and follow-up (1994-97) measurements and 1 551 of these had valid serial measurements for both lumbar spine and femoral neck (excluding severe bone deformities, see section *Bone mass measurements*).
For the present study, the following women were additionally excluded: 1) hysterectomized women (for whom it was impossible to define the menopausal status) and bilaterally ovariectomized women (n=445), 2) premenopausal women (n=152). Accordingly, the final study population consisted of 954 women (beginning of menopause either before or during follow-up) aged 48 to 59 years at baseline densitometry. The beginning of menopause was defined as 12 months' amennorhea (WHO Scientific Group, 1996) and the duration of menopause varied from 1 to 26 years at follow-up densitometry. The duration of follow-up varied from 3.8 to 7.9 years (mean 5.8 years).
#### **2.2 Seasonal Difference Index (SDI)**
The study population was divided into three equal groups within the year according to month of measurement at baseline and follow-up: Group 1 (from January to April), Group 2 (from May to August) and Group 3 (from September to December). The basis for selecting these cutoffs was based on the assumption that highest BMDs (reflecting serum vitamin D concentration) would be present at late summer and early fall season (within group 2) and the lowest at late winter (within group 3) whereas group 1 would present an intermediate. Also, in order to reveal the greatest differences in seasonal BMD variation between two successive measurements a numeric value of a Seasonal Difference Index (SDI) was calculated as follows:
#### SDI = (Group number at baseline - Group number at follow-up)
Accordingly, the numeric value of SDI varied from -2 to +2 (Table 1). In all, 521 (54.6 %) women were measured within the same season at both measurements (SDI group 0). In 212 (22.2 %) women the follow-up measurement was carried out in a later season (SDI groups -1 and -2) and in 221 (23.2 %) women in earlier season (SDI groups +1 and +2) than the baseline measurement. There was a significant variation in distribution of women into measurement seasons between and within baseline and follow-up measurements (Figure 1 ). Accordingly, any conclusions over differences between specific seasons were considered to be precluded due to this uneven distribution.
**Follow-up season (Group number)**
*a) SDI=[Season Group at baseline]-[Season Group at follow-up]*
Table 1. Numeric values for the "Seasonal Difference Index" (SDI)a according to baseline and follow-up DXA measurement season
#### **2.3 Other variables**
30 Dual Energy X-Ray Absorptiometry
compounds (Patel et al, 2001; Blumsohn et al., 2003)**.** Consequently, the effect of the season when densitometry was performed on bone density is still unresolved. Furthermore, the role of a seasonal difference between two distant follow-up bone density measurements in postmenopausal bone loss has not been studied and thus, nothing is known about the effect of this phenomenon on treatment monitoring or other longitudinal data collection. If present, such seasonal variation could have significant effect on the evaluation of prospective data.
The purpose of the present study was to investigate the effect of densitometry season on early postmenopausal BMD and bone loss in a subset of 954 Finnish women selected from
The study population was a randomly selected part of the prospective Kuopio Osteoporosis Risk Factor and Prevention (OSTPRE) study cohort. The OSTPRE cohort was established in 1989 and included all women born in 1932-1941 and who were resident in Kuopio Province, Finland (n=14 220). A postal inquiry including questions about health disorders, medications including HRT, gynaecologic history, nutritional habits, physical activity, lifestyle habits, and anthropometric information was sent to these women at baseline in 1989 (Honkanen et al, 1991). The 5-year (in 1994 follow-up questionnaires were sent to the 13 100 women who responded at baseline and responses were received from 11 954 at 5-year follow-up. The study protocol has been approved by the ethics committee of University of Kuopio and Kuopio University Hospital. Informed written consent from the participants
Of the 13 100 respondents in 1989, 11 055 (84.4 %) were willing to undergo DXA densitometry. A random sample of 2 362 women was selected for densitometry out of which 2025 women actually underwent the procedure during 1989-91. The questionnaire information was updated at the time of bone densitometry. In all, 1 873 women underwent both baseline (1989-91) and follow-up (1994-97) measurements and 1 551 of these had valid serial measurements for both lumbar spine and femoral neck (excluding severe bone
For the present study, the following women were additionally excluded: 1) hysterectomized women (for whom it was impossible to define the menopausal status) and bilaterally ovariectomized women (n=445), 2) premenopausal women (n=152). Accordingly, the final study population consisted of 954 women (beginning of menopause either before or during follow-up) aged 48 to 59 years at baseline densitometry. The beginning of menopause was defined as 12 months' amennorhea (WHO Scientific Group, 1996) and the duration of menopause varied from 1 to 26 years at follow-up densitometry. The duration of follow-up
The study population was divided into three equal groups within the year according to month of measurement at baseline and follow-up: Group 1 (from January to April), Group 2 (from May to August) and Group 3 (from September to December). The basis for selecting these cut-
the population-based Osteoporosis Risk Factor and Prevention (OSTPRE)-study.
**2. Subjects and methods**
was collected with the postal inquiries.
deformities, see section *Bone mass measurements*).
varied from 3.8 to 7.9 years (mean 5.8 years).
**2.2 Seasonal Difference Index (SDI)**
**2.1 Study population**
#### **2.3.1 Hormone therapy**
Women were divided into two groups according to their use of hormone therapy (HRT) (tablets and plasters) which was defined as the use of hormonal products for menopausal symptoms. *HRT users* (n=393) had used HRT continuously or occasionally during the follow-up regardless of whether or not they had used hormonal therapy before the baseline (14 women used HRT only before baseline, and were excluded in analysis on HRT effect). *HRT non-users* (n=547) had never used estrogen containing products aimed at postmenopausal therapy. In the OSTPRE cohort, the majority of HRT users were taking estrogen-progesterone combination products (56.2% of all HRT). No data was available on whether HRT was continuous or sequential. Forty-five percent of HRT users (occasional or continuous) had been treated with HRT also prior to baseline. The duration of HRT varied from one month to 7.5 years. The information about the use of hormonal products was obtained from the questionnaires. Comparison between self-reported use of HRT and the national prescription records of The Social Insurance Institution, Finland (KELA), for the whole OSTPRE cohort in 1996–2001 revealed that 97.8% of those who had received an oestrogen drug prescription reported HRT use in inquiries. On the other hand, in 25.5% of the self-reported non-users of HRT some oestrogen use (short-term, median 6.0 months) was recorded (Sirola et al., 2003a).
Bone Loss and Seasonal Variation in Serial DXA Densitometry – A Population-Based Study 33
The *height* and *weight* were measured with a stadiometer and calibrated scale by study
*Nutritional calcium* intake of each participant was estimated according to self-reported ingestion of milk products in postal inquiries. The following questions were asked:" How many deciliters of fluid milk products (milk, sour milk, yoghurt, etc.) do you consume daily? " and "How many slices of cheese do you eat daily?". The amount of calcium was approximated
Women were divided into two categories (yes, no) according to the presence/absence of *bone affecting diseases or medications* at baseline. Bone affecting diseases/medications have been described previously by Kröger et al. (Kröger et al. 1994). Diseases were: renal disease, liver disease, insulin-dependent diabetes, malignancies, rheumatoid arthritis, endocrine abnormalities (parathyroid/thyroid glands, adrenals), malabsorption (including lactose malabsorption), total/partial gastrectomy, postovariectomy status, premenopausal amenorrhea, alcoholism and long-term immobilisation. Medications were: corticosteroids, diuretics, cytotoxic drugs, anticonvulsive drugs, anabolic steroids, calcitonin,
*Physical activity level* was calculated based on combined physical activity in work and leisure based on self-reports in the postal inquiries. The physical activity was categorised into low,
The bone mineral density of lumbar spine (L2-L4) and left femoral neck was determined using the same dual X-ray absorptiometry (DXA) (Lunar DPX, Madison, Wisconsin, USA) equipment at both the baseline and five year measurements. The measurements were carried out in Kuopio University Hospital by specially trained personnel. The short term reproducibility of this method has been shown to be 0.9 % for lumbar spine and 1.5 % for femoral neck BMD measurements (Kröger et al., 1992).The long-term reproducibility (coefficient of variation) of the DXA instrument for BMD during the study period, as determined by regular phantom measurements, was 0.4 % (Komulainen et al., 1998). Each DXA measurement print was reviewed and women with bone deformities (osteoarthritis, osteophytes, scoliosis and compression fractures) in either area were excluded from the analyses. At the time of densitometry, also the weight and height of each participant was
Statistical analyses were carried out with the Statistical Package for Social Sciences (SPSS) for Windows, version 17. The annual BMD changes for both measurement sites were calculated as follows: [(BMD at the 5-year follow-up - BMD at baseline) / duration of follow-up in years] and reported as percentage of baseline BMD. In categorical analyses, uni- and multivariate analysis of variance was used and Tukey (crude models) and Least Significant Difference (adjusted models) -post hoc tests were utilized to study differences between multiple groups when applicable. Adjustment for age, height, weight, months since menopause, mean calcium intake, use of HRT (no, occasional, continuous), physical activity level (low, moderate, high),
to be 120 mg/dl for fluid milk products and 87 mg/slice for cheese (Sirola et al., 2003b).
**2.3.2 Adjusting variables**
bisphosphonates, vitamin D.
moderate and high (Kröger et al., 1994).
**2.4 Bone mass measurements**
measured in a controlled situation.
**2.5 Statistical methods**
group nurses at each bone densitometry.
Follow-up measurement month Fig. 1. Distribution of the study population accroding to DXA measurement month.
#### **2.3.2 Adjusting variables**
32 Dual Energy X-Ray Absorptiometry
Baseline measurement month
Follow-up measurement month
Fig. 1. Distribution of the study population accroding to DXA measurement month.
Number of women
Number of women
200
180
160
140
120
100
80
60
40
20 0
120
100
80
60
40
20
0
987654321 121110
987654321 121110
The *height* and *weight* were measured with a stadiometer and calibrated scale by study group nurses at each bone densitometry.
*Nutritional calcium* intake of each participant was estimated according to self-reported ingestion of milk products in postal inquiries. The following questions were asked:" How many deciliters of fluid milk products (milk, sour milk, yoghurt, etc.) do you consume daily? " and "How many slices of cheese do you eat daily?". The amount of calcium was approximated to be 120 mg/dl for fluid milk products and 87 mg/slice for cheese (Sirola et al., 2003b).
Women were divided into two categories (yes, no) according to the presence/absence of *bone affecting diseases or medications* at baseline. Bone affecting diseases/medications have been described previously by Kröger et al. (Kröger et al. 1994). Diseases were: renal disease, liver disease, insulin-dependent diabetes, malignancies, rheumatoid arthritis, endocrine abnormalities (parathyroid/thyroid glands, adrenals), malabsorption (including lactose malabsorption), total/partial gastrectomy, postovariectomy status, premenopausal amenorrhea, alcoholism and long-term immobilisation. Medications were: corticosteroids, diuretics, cytotoxic drugs, anticonvulsive drugs, anabolic steroids, calcitonin, bisphosphonates, vitamin D.
*Physical activity level* was calculated based on combined physical activity in work and leisure based on self-reports in the postal inquiries. The physical activity was categorised into low, moderate and high (Kröger et al., 1994).
#### **2.4 Bone mass measurements**
The bone mineral density of lumbar spine (L2-L4) and left femoral neck was determined using the same dual X-ray absorptiometry (DXA) (Lunar DPX, Madison, Wisconsin, USA) equipment at both the baseline and five year measurements. The measurements were carried out in Kuopio University Hospital by specially trained personnel. The short term reproducibility of this method has been shown to be 0.9 % for lumbar spine and 1.5 % for femoral neck BMD measurements (Kröger et al., 1992).The long-term reproducibility (coefficient of variation) of the DXA instrument for BMD during the study period, as determined by regular phantom measurements, was 0.4 % (Komulainen et al., 1998). Each DXA measurement print was reviewed and women with bone deformities (osteoarthritis, osteophytes, scoliosis and compression fractures) in either area were excluded from the analyses. At the time of densitometry, also the weight and height of each participant was measured in a controlled situation.
#### **2.5 Statistical methods**
Statistical analyses were carried out with the Statistical Package for Social Sciences (SPSS) for Windows, version 17. The annual BMD changes for both measurement sites were calculated as follows: [(BMD at the 5-year follow-up - BMD at baseline) / duration of follow-up in years] and reported as percentage of baseline BMD. In categorical analyses, uni- and multivariate analysis of variance was used and Tukey (crude models) and Least Significant Difference (adjusted models) -post hoc tests were utilized to study differences between multiple groups when applicable. Adjustment for age, height, weight, months since menopause, mean calcium intake, use of HRT (no, occasional, continuous), physical activity level (low, moderate, high),
Bone Loss and Seasonal Variation in Serial DXA Densitometry – A Population-Based Study 35
**Variable Group 1 Group 2 Group 3 Total (n=424) (n=192) (n=338) (n=954)**
Duration of follow-up yrs 5,9(0,5) 5,8(0,4) 5,8(0,5) 5,8 (0,5)
Duration of menopause, months 89,4(54,1) 86,2(48,1) 106,6(50,4) 95,0(52,3)d Baseline age, yrs 53,5(3,0) 53,5(2,9) 54,2(2,7) 53,7(2,9)d Baseline height, cm 160,9(5,1) 161,6(5,5) 160,8(5,3) 161,0(5,3)
Baseline weight, kg 69,1(12,4) 68,6(11,4) 69,3(11,9) 69,1(12,0) Weight change (%) 2,9(5,2) 2,9(5,6) 3,0(5,3) 2,9(5,3) Grip strength, kPa 62,0(16,5) 62,2(15,6) 62,4(16,0) 62,2(16,1) Mean calcium intake, mg/day 789(343) 813(311) 799(319) 797(328) Baseline lumbar BMD, g/cm2 1,13(0,17) 1,13(0,16) 1,11(0,16) 1,12(0,16) Baseline femoral neck BMD, g/cm2 0,93(0,13) 0,93(0,13) 0,92(0,12) 0,92(0,13) FU lumbar BMD, g/cm2 1,07(0,17) 1,09(0,15) 1,07(0,16) 1,08(0,16) FU femoral neck BMD, g/cm2 0,88(0,12) 0,90(0,13) 0,89(0,12) 0,89(0,12)
No use 55,8 56,0 64,7 59,0 c Occasional (<90 % of FU) 32,2 34,6 28,2 31,2 Continuous (>90 % of FU) 12,1 9,4 7,2 9,8
No bone affecting disease/medication 60,4 64,1 64,7 62,7 Any previous fracture at baseline 20,9 23,3 20,6 21,3
Previous wrist fracture at baseline 7,2 7,7 5,5 6,7 Alcohol >1 drink/week 35,1 34,4 32,3 33,9 Smoking 9,4 11,1 8,8 9,5 High overall physical activity levelb 29,9 37,8 32,8 32,5
*to December). b) Three categorical variable: low, moderate, high. c) p<0.05 / d) p<0.001*
*a) Season Groups: Group 1 (from January to April), Group 2 (from May to August) and Group 3 (from September*
Table 2. Baseline characteristics of the study population according to season groupa (n=954 ).
**Means (SD) of continuous variables**
**B. Distribution of category variables (%)**
Use of HRT during follow-up
**Season groupa**
duration of follow-up (years) and the use of bone affecting medications or diseases (yes/no) (including vitamin D supplements) as covariates was performed.
#### **3. Results**
The baseline data revealed that there were no significant differences between the three season groups with respect to age, duration of menopause or HRT use (Table 2). There were no differences in the cross-sectional BMD between the season groups at baseline or at the five year follow-up (Table 2).
In order to evaluate a possible contribution of seasonal differences to the follow-up BMD values, the association of SDI with mean annual bone loss was investigated (Figure 2). The bone loss rate in SDI categories -2 and 0 was greater than in SDI categories +1 and +2 (p<0.01) in both lumbar and femoral regions (Figure 2). In lumbar spine, the difference in bone loss rate between SDI categories -1 and +1 was also significant (p=0.015). In femoral neck there was no significant difference between SDI category -1 and the other categories. These effects were independent of any adjustments.
*p-values refer to the differences of the respective SDI group in comparison to SDI 1 and 2 for lumbar spine (LS=full line) and femoral neck (FN=dotted line). a) adjusted for age, height, weight, months since menopause, calcium intake, use of HT (no, occasional, continuous), overall physical activity level (low, moderate, high), duration of follow-up (years) and use of bone affecting medications or diseases (yes/no)*
Fig. 2. Effect of Seasonal Difference Index (SDI) on mean annual bone loss rate (%) in early postmenopausal women (n=954). Analysis of covariancea
duration of follow-up (years) and the use of bone affecting medications or diseases (yes/no)
The baseline data revealed that there were no significant differences between the three season groups with respect to age, duration of menopause or HRT use (Table 2). There were no differences in the cross-sectional BMD between the season groups at baseline or at the
In order to evaluate a possible contribution of seasonal differences to the follow-up BMD values, the association of SDI with mean annual bone loss was investigated (Figure 2). The bone loss rate in SDI categories -2 and 0 was greater than in SDI categories +1 and +2 (p<0.01) in both lumbar and femoral regions (Figure 2). In lumbar spine, the difference in bone loss rate between SDI categories -1 and +1 was also significant (p=0.015). In femoral neck there was no significant difference between SDI category -1 and the other categories.
*p-values refer to the differences of the respective SDI group in comparison to SDI 1 and 2 for lumbar spine (LS=full line) and femoral neck (FN=dotted line). a) adjusted for age, height, weight, months since menopause, calcium intake, use of HT (no, occasional, continuous), overall physical activity level (low, moderate, high),*
p < 0.01 (LS)
Fig. 2. Effect of Seasonal Difference Index (SDI) on mean annual bone loss rate (%) in early
*duration of follow-up (years) and use of bone affecting medications or diseases (yes/no)*
p < 0.001 (FN) p < 0.001 (FN)
p < 0.01 (FN)
postmenopausal women (n=954). Analysis of covariancea
p < 0.01 (LS)
(including vitamin D supplements) as covariates was performed.
These effects were independent of any adjustments.
**3. Results**
five year follow-up (Table 2).
*a) Season Groups: Group 1 (from January to April), Group 2 (from May to August) and Group 3 (from September to December). b) Three categorical variable: low, moderate, high. c) p<0.05 / d) p<0.001*
Table 2. Baseline characteristics of the study population according to season groupa (n=954 ).
Bone Loss and Seasonal Variation in Serial DXA Densitometry – A Population-Based Study 37
p < 0.01
B: SDI category 1 and 2
Fig. 3. Effect of hormone therapy on lumbar spine (full line) and femoral neck (dotted line) bone loss according to seaosal difference index (SDI) category. Analysis of variance
(ANOVA, n=954).
A: SDI category -2 to 0
p > 0.05
A: SDI category -2 to 0
B: SDI category 1 and 2
Fig. 3. Effect of hormone therapy on lumbar spine (full line) and femoral neck (dotted line) bone loss according to seaosal difference index (SDI) category. Analysis of variance (ANOVA, n=954).
Bone Loss and Seasonal Variation in Serial DXA Densitometry – A Population-Based Study 39
Secondly, in prospective studies, possible distortion in results due to seasonal differences in risk-factor analyses and treatment effects on bone loss rate should be closely considered. Accordingly, it could be worthwhile to create a "seasonal difference index" for each population based on the respective DXA measurement data. The effect of seasonal differences on densitometry-based risk-factors for postmenopausal bone loss remains to be resolved. It might be that seasonal dependency could interact with certain factors lessening their true impacts. Also, the impact of these differences on fracture prediction remains to be
The interpretation of seasonal indices needs to be undertaken with caution. In the present study, the differences between the SDI categories do not necessarily provide information of the exact season that each participant was measured (e.g. SDI +1 could represent a difference between baseline season category 3 and a 5-year season category 2 or a baseline category 2 and a 5-year category 1). Also, the baseline and follow-up season groups (groups 1, 2, 3) itself included quite heterogeneous population. For example, the women measured during the first months of group 1 (January-February) were likely to have significantly lower BMDs in comparison to women presenting the other end (March-April) during which sun light exposure would be higher. We used this season categorisation order to categorize the measurement months into three equal groups within a year (i.e. four months per group: Spring (January-April), Summer (May-August) and Fall/Winter (September-December). This was based on the rapid changes in sun light exposure in the northern latitudes: sun light hours rapidly increase during may and decrease rapidly during september. Furthermore, the distribution of women into measurement months within each group was found to differ between baseline and follow-up densitometry. However, the goal of the present study was not to identify specific "risk seasons", but only to assess variability in the estimated bone loss rate attributable to the seasons when the two successive measurements had been done. Thus, some arbitrary cut-offs, in terms of season groups, were forced to be decided. Accordingly, the hypothesis of high or low bone density according to sun exposure (and vitamin D levels) in baseline and follow-up was precluded by skew distribution of women into different measurement months. Before adaptation for wider use, these indices would need further testing and refinement in order to optimize the categorisation for local
This is the first long-term population-based study investigating the contribution of seasonal difference between two successive DXA measurements on postmenopausal bone loss. Previous studies have shown significant alterations in vitamin D and PTH levels attributable to season (Rapuri et al., 2002; Dawson-Hughes et al., 1997).which could also provide a pathophysiological mechanism for the seasonal bone effects. Some studies have also found seasonal variation in cross-sectional BMD data, bone markers and bone loss rate (Rosen et al., 1994.; Storm et al., 1998; Rapuri et al., 2002; Dawson-Hughes et al., 1997). However, other studies have failed to found any evidence on altered bone metabolism related to seasons as measured with either BMD or bone markers (Patel et al, 2001; Blumsohn et al., 2003). The ability to observe seasonal effects is likely to depend on geographical location. Another study conducted in northern latitudes (Gerdhem et al, 2004) failed to demonstrate any seasonal variation in cross-sectional study design in Sweden. The present large study population, living at northern latitudes (latitude 63 degrees), might facilitate the detection of seasonal differences. However, the present study also showed no significant cross-sectional
resolved in future studies.
purposes.
variation in BMD.
In ANOVA, SDI explained 2.3 percent (R2=0.023) and 1.3 percent (R2=0.013) of the bone mass changes in lumbar spine and femoral neck, respectively. Furthermore, in linear regression models, SDI was positively associated with both lumbar spine and femoral neck bone loss (p<0.001) independent of all adjustments.
In order to mimic the possible effect of SDI on a treatment monitoring protocol, we investigated the effects of HRT on bone loss according to SDI (Figure 3). HRT users had significantly lower bone loss rate in SDI categories -2 to 0 in comparison to HRT non-users (lumbar spine and femoral neck) (p<0.01). In SDI categories 1 to 2, there was no statistically significant difference between HRT users and non-users. These results were not affected by adjustments.
#### **4. Discussion**
The present study evaluated the effect of season on BMD and bone loss with a randomly selected population-based sample of 954 Finnish women. The seasonal difference between two successive axial DXA measurements, estimated with the "Seasonal Difference Index" (SDI), influenced the evaluation of postmenopausal bone loss rate. In addition, this factor interfered with the evaluation of protective bone effects of HRT. The exact direction of these relationships, in terms of specific seasons, was found to complex partly due to the study setting.
The present study sample was large and randomly selected. There were few differences in the baseline variables between the three season groups. The DXA measurements were carried out with same equipment and measurement staff and all bone deformities were excluded. In addition, phantom calibration was performed regularly which should exclude any significant seasonal changes attributable to equipment performance. Furthermore, comprehensive adjustment for any potential confounders was used in the analyses. Hence, it is most unlikely that any major confounding could have occurred in the present study.
Some weaknesses of the present study should also be considered. The follow-up time was relatively long with considerable inter-individual variation. Although the results were adjusted for duration of follow-up and reported in annual percent changes the follow-up period in treatment monitoring is usually only one or two years. However, the long followup and large sample probably facilitated the detection of bone mass changes in the present study. Furthermore, the present study assumed that the pattern of bone loss between the approximately five-year follow-up was linear, masking any possible short-term non-linear patterns. However, adjusting for these changes would have required DXA measurement at the very least at 1 year intervals. Lastly, the lack of information of serum vitamin D levels precluded causal conclusions. However, adjustment for bone affecting medication (including vitamin D containing products, medications and supplements) and calcium intake was performed eliminating bias due to these factors.
The results of the present study have two major applications. Firstly, in treatment monitoring, an attempt should be made to measure bone density within the same season. Naturally, the seasonal limits depend on amount of seasonal variation in sunlight exposure of the study population and should be closely studied. The sub-division used in the present study provides one example in DXA measurements suitable for Scandinavian latitudes.
In ANOVA, SDI explained 2.3 percent (R2=0.023) and 1.3 percent (R2=0.013) of the bone mass changes in lumbar spine and femoral neck, respectively. Furthermore, in linear regression models, SDI was positively associated with both lumbar spine and femoral neck
In order to mimic the possible effect of SDI on a treatment monitoring protocol, we investigated the effects of HRT on bone loss according to SDI (Figure 3). HRT users had significantly lower bone loss rate in SDI categories -2 to 0 in comparison to HRT non-users (lumbar spine and femoral neck) (p<0.01). In SDI categories 1 to 2, there was no statistically significant difference between HRT users and non-users. These results were not affected by
The present study evaluated the effect of season on BMD and bone loss with a randomly selected population-based sample of 954 Finnish women. The seasonal difference between two successive axial DXA measurements, estimated with the "Seasonal Difference Index" (SDI), influenced the evaluation of postmenopausal bone loss rate. In addition, this factor interfered with the evaluation of protective bone effects of HRT. The exact direction of these relationships, in terms of specific seasons, was found to complex partly due to the study
The present study sample was large and randomly selected. There were few differences in the baseline variables between the three season groups. The DXA measurements were carried out with same equipment and measurement staff and all bone deformities were excluded. In addition, phantom calibration was performed regularly which should exclude any significant seasonal changes attributable to equipment performance. Furthermore, comprehensive adjustment for any potential confounders was used in the analyses. Hence, it is most unlikely that any major confounding could have occurred in the present study.
Some weaknesses of the present study should also be considered. The follow-up time was relatively long with considerable inter-individual variation. Although the results were adjusted for duration of follow-up and reported in annual percent changes the follow-up period in treatment monitoring is usually only one or two years. However, the long followup and large sample probably facilitated the detection of bone mass changes in the present study. Furthermore, the present study assumed that the pattern of bone loss between the approximately five-year follow-up was linear, masking any possible short-term non-linear patterns. However, adjusting for these changes would have required DXA measurement at the very least at 1 year intervals. Lastly, the lack of information of serum vitamin D levels precluded causal conclusions. However, adjustment for bone affecting medication (including vitamin D containing products, medications and supplements) and calcium
The results of the present study have two major applications. Firstly, in treatment monitoring, an attempt should be made to measure bone density within the same season. Naturally, the seasonal limits depend on amount of seasonal variation in sunlight exposure of the study population and should be closely studied. The sub-division used in the present study provides one example in DXA measurements suitable for Scandinavian latitudes.
intake was performed eliminating bias due to these factors.
bone loss (p<0.001) independent of all adjustments.
adjustments.
setting.
**4. Discussion**
Secondly, in prospective studies, possible distortion in results due to seasonal differences in risk-factor analyses and treatment effects on bone loss rate should be closely considered. Accordingly, it could be worthwhile to create a "seasonal difference index" for each population based on the respective DXA measurement data. The effect of seasonal differences on densitometry-based risk-factors for postmenopausal bone loss remains to be resolved. It might be that seasonal dependency could interact with certain factors lessening their true impacts. Also, the impact of these differences on fracture prediction remains to be resolved in future studies.
The interpretation of seasonal indices needs to be undertaken with caution. In the present study, the differences between the SDI categories do not necessarily provide information of the exact season that each participant was measured (e.g. SDI +1 could represent a difference between baseline season category 3 and a 5-year season category 2 or a baseline category 2 and a 5-year category 1). Also, the baseline and follow-up season groups (groups 1, 2, 3) itself included quite heterogeneous population. For example, the women measured during the first months of group 1 (January-February) were likely to have significantly lower BMDs in comparison to women presenting the other end (March-April) during which sun light exposure would be higher. We used this season categorisation order to categorize the measurement months into three equal groups within a year (i.e. four months per group: Spring (January-April), Summer (May-August) and Fall/Winter (September-December). This was based on the rapid changes in sun light exposure in the northern latitudes: sun light hours rapidly increase during may and decrease rapidly during september. Furthermore, the distribution of women into measurement months within each group was found to differ between baseline and follow-up densitometry. However, the goal of the present study was not to identify specific "risk seasons", but only to assess variability in the estimated bone loss rate attributable to the seasons when the two successive measurements had been done. Thus, some arbitrary cut-offs, in terms of season groups, were forced to be decided. Accordingly, the hypothesis of high or low bone density according to sun exposure (and vitamin D levels) in baseline and follow-up was precluded by skew distribution of women into different measurement months. Before adaptation for wider use, these indices would need further testing and refinement in order to optimize the categorisation for local purposes.
This is the first long-term population-based study investigating the contribution of seasonal difference between two successive DXA measurements on postmenopausal bone loss. Previous studies have shown significant alterations in vitamin D and PTH levels attributable to season (Rapuri et al., 2002; Dawson-Hughes et al., 1997).which could also provide a pathophysiological mechanism for the seasonal bone effects. Some studies have also found seasonal variation in cross-sectional BMD data, bone markers and bone loss rate (Rosen et al., 1994.; Storm et al., 1998; Rapuri et al., 2002; Dawson-Hughes et al., 1997). However, other studies have failed to found any evidence on altered bone metabolism related to seasons as measured with either BMD or bone markers (Patel et al, 2001; Blumsohn et al., 2003). The ability to observe seasonal effects is likely to depend on geographical location. Another study conducted in northern latitudes (Gerdhem et al, 2004) failed to demonstrate any seasonal variation in cross-sectional study design in Sweden. The present large study population, living at northern latitudes (latitude 63 degrees), might facilitate the detection of seasonal differences. However, the present study also showed no significant cross-sectional variation in BMD.
Bone Loss and Seasonal Variation in Serial DXA Densitometry – A Population-Based Study 41
Gerdhem P, Mallmin H, Akesson K, Obrant KJ. Seasonal variation in bone density in
Hannan MT, Felson DT, Dawson-Hughes B, Tucker KL, Cupples LA, Wilson PW, Kiel DP
Hansen MA, Overgaard K, Christiansen C (1995) Spontaneous postmenopausal bone loss in different skeletal areas followed up for 15 years. *J Bone Miner Res* 10:205-210 Harris S, Dawson-Hughes B (1992) Rates of change in bone mineral density of spine, heel, femoral neck and radius in healthy postmenopausal women. *Bone Miner* 17:87-95 Honkanen R, Tuppurainen M, Alhava E, Saarikoski S (1991) Kuopio Osteoporosis Risk
*University of Kuopio, Community Health, Statistics and Reviews 3/1991*; Kuopio. Ito M, Nakamura T, Tsurusaki K, Uetani M, Hayashi K (1999) Effects of menopause on age-
Komulainen MH, Kröger H, Tuppurainen MT, Heikkinen A-M, Alhava E, Honkanen R,
postmenopausal women; a 5 year randomised trial. *Maturitas* 31:45-54 Komulainen MH, Kröger H, Tuppurainen MT, Heikkinen A-M, Alhava E, Honkanen R,
population-based 5-year randomised trial. *J Clin Endoc Metab* 84:546-552 Kröger H, Heikkinen J, Laitinen K, Kotaniemi A (1992) Dual-energy X-ray absorptiometry in
Kröger H, Tuppurainen M, Honkanen R, Alhava E, Saarikoski S (1994) Bone mineral density
Miller PD, Bonnick SL, Rosen CJ (1996) Consensus of an international panel on the clinical
National Osteoporosis Foundation (1998) Osteoporosis: review of the evidence for
Patel R, Collins D, Bullock S, Swaminathan R, Blake GM, Fogelman I (2001) The effect of
Pouilles JM, Tremollieres F, Ribot C (1993) The effects of menopause on longitudinal bone
Pouilles JM, Tremollieres F, Ribot C (1995) Effect of menopause on femoral and vertebral
Prior JC (1998) Perimenopause: the complex endocrinology of the menopausal transition.
Rapuri PB, Kinyamu HK, Gallagher JC, Haynatzka V (2002) Seasonal changes in calcitropic
hormones, bone markers and bone mineral density in elderly women. *J Clin*
a double-masked crossover study. *Osteoporos Int* 12:319-25
loss from the spine. *Calcif Tissue Int* 52:340-343
bone loss. *J Bone Miner Res* 10:1531-1536
*Endoc Res* 19(4):397-428
*Endocrinol Metab* 87:2024-32
(2000) Risk factors for longitudinal bone loss in elderly men and women: the
Factor and Prevention Study. Baseline postal inquiry in 1989. *Publications of the*
dependent bone loss in the axial and appendicular skeletons in healthy Japanese
Saarikoski S (1998) HRT and vit D in prevention of non-vertebral fractures in
Jurvelin J, Saarikoski S (1999) Prevention of femoral and lumbar bone loss with hormone replacement therapy and vitamin D3 in early postmenopausal women: a
normal women: a cross-sectional study of 717 Finnish volunteers. *Osteoporos Int*
and risk factors for osteoporosis-a population based study of 1600 perimenopausal
utility of bone mass measurements in the detection of low bone mass in the adult
prevention, diagnosis, and treatment and cost-effectiveness analysis-status report.
season and vitamin D supplementation on bone mineral density in healthy women:
postmenopausal women. *J Clin Densitom* 2004;7(1):93-100
women. *Osteoporos Int* 19:377-83.
women. *Calcif Tissue Int* 55:1-7
population. *Calcif Tissue Int* 58:207-214
*Osteoporosis Int* 18 (suppl 4):S1-S88
2:135-140.
Framingham osteoporosis study. *J Bone Miner Res* 15:710-720.
The present study also attempted to evaluate the possible bias in treatment monitoring resulting from seasonal difference via an investigation of how the SDI could affect the bone effect of HRT. In fact, variability in the protective effect of HRT between the SDI groups was detected. This serves as preliminary example of the extensiveness of the distortion in estimation of bone loss rate caused by seasonal difference. The inclusion of occasional use of HRT and less-than-perfect validity of HRT use in "non-users" probably mdified the results (Sirola et al., 2003a) but the trend was clear. Also, the effects may have been affected by lack of power due to small group sizes. However, seasonal densitometry difference may help in the identification of "non-responders" to HRT and be included in the list of other contributing factors (Sirola et al., 2003c; Komulainen et al., 1999). Previously, it has been suggested that calcium may flatten the seasonal differences in bone loss rate among elderly women8 and that there might be seasonal variation in the bone response to vitamin D (Dawson-Hughes et al, 1991). The present study also showed abolition in the difference in the bone loss rate between SDI categories in HRT users.
In summary, seasonal differences should receive closer attention in treatment monitoring protocols and longitudinal risk factor studies. In future studies, the seasonal densitometry difference should be considered as a potential confounder and its effect on risk factor and treatment monitoring data should be assessed. In addition, factors that might lessen the seasonal changes in the bone loss rate, such as calcium, vitamin D and other bone drugs, should be identified. Our study raises, for the first time, the question of whether the results of longitudinal DXA measurements might be significantly distorted by seasonal differences especially in northern latitudes.
#### **5. Acknowledgements**
This work has been financially supported by Kuopio University Hospital, EVO-grant
#### **6. References**
The present study also attempted to evaluate the possible bias in treatment monitoring resulting from seasonal difference via an investigation of how the SDI could affect the bone effect of HRT. In fact, variability in the protective effect of HRT between the SDI groups was detected. This serves as preliminary example of the extensiveness of the distortion in estimation of bone loss rate caused by seasonal difference. The inclusion of occasional use of HRT and less-than-perfect validity of HRT use in "non-users" probably mdified the results (Sirola et al., 2003a) but the trend was clear. Also, the effects may have been affected by lack of power due to small group sizes. However, seasonal densitometry difference may help in the identification of "non-responders" to HRT and be included in the list of other contributing factors (Sirola et al., 2003c; Komulainen et al., 1999). Previously, it has been suggested that calcium may flatten the seasonal differences in bone loss rate among elderly women8 and that there might be seasonal variation in the bone response to vitamin D (Dawson-Hughes et al, 1991). The present study also showed abolition in the difference in
In summary, seasonal differences should receive closer attention in treatment monitoring protocols and longitudinal risk factor studies. In future studies, the seasonal densitometry difference should be considered as a potential confounder and its effect on risk factor and treatment monitoring data should be assessed. In addition, factors that might lessen the seasonal changes in the bone loss rate, such as calcium, vitamin D and other bone drugs, should be identified. Our study raises, for the first time, the question of whether the results of longitudinal DXA measurements might be significantly distorted by seasonal differences
This work has been financially supported by Kuopio University Hospital, EVO-grant
sectional study. *J Bone Miner Res* 18:1274-1281
postmenopausal women. *Ann Intern Med* 115:505-12
Blumsohn A, Naylor KE, Timm W, Eagleton AC, Hannon RA, Eastell R (2003) Absence of
Burger H, de Laet CE, van Daele PL, Weel AE, Witteman JC, Hofman A, Pols HA (1998) Risk
Cummings SR, Melton LJ (2002) Epidemiology and outcomes of osteoporotic fractures.
Dawson-Hughes B, Dallal GE, Krall EA, Harris S, Sokoll LJ, Falconer G (1991) Effects of
Dawson-Hughes B, Harris SS, Dallal GE (1997) Plasma calcidiol, season and serum
Dennison E, Cole Z, Cooper C (2005) Diagnosis and epidemiology of osteoporosis. *Curr*
marked seasonal change in bone turnover: a longitudinal and multicenter cross-
factors for increased bone loss in an elderly population: the Rotterdam Study. *Am J*
vitamin D supplementation on wintertime and overall bone loss in healthy
parathyroid hormone concentration in healthy elderly men and women. *Am J Clin*
the bone loss rate between SDI categories in HRT users.
especially in northern latitudes.
*Epidemiol* 147:871-9.
*Lancet* 359:1761–1767
*Opin Rheumatol* 17(4):456–461
*Nutr* 65:67-71
**5. Acknowledgements**
**6. References**
**Part 2**
**Body Composition**
## **Part 2**
**Body Composition**
42 Dual Energy X-Ray Absorptiometry
Rosen CJ, Morrison A, Zhou H, Storm D, Hunter SJ, Musgrave K, Chen T, Wei W, Holick
Sirola J, Kröger H, Honkanen R, Jurvelin JS, Sandini L, Tuppurainen M, Saarikoski S (2003b)
Sirola J, Kröger H, Jurvelin JS, Sandini L, Tuppurainen M, Saarikoski S, Honkanen R (2003c)
Sowers MF (1997) Clinical epidemiology and osteoporosis. Measures and their
Storm D, Eslin R, Porter ES, Musgrave K, Vereault D, Patton C, Kessenich C, Mohan S, Chen
bone mineral density and calciotropic hormones. *Bone Miner* 25:83-92 Sirola J, Kröger H, Honkanen R, Sandini L, Tuppurainen M, Jurvelin JS, Saarikoski S (2003a)
prevent weight loss- related bone loss*? Osteoporosis Int* 14: 27-33
population based approach. *J Bone Miner Res* 18: 1036-1043
loss: a prospective study. *Calcif Tissue Int* 72:659-665
interpretation. *Endocrin Metab North Amer* 26:219-231
MF (1994) Elderly women in northern New England exhibit seasonal changes in
Risk factors associated with peri- and postmenopausal bone loss - does HRT
Smoking may impair the bone protective effects of nutritional calcium- a
Interaction of nutritional calcium and HRT in prevention of postmenopausal bone
T, Holick MF, Rosen CJ (1998) Calcium supplementation prevents seasonal bone loss and changes in biochemical markers of bone turnover in elderly New England women: a randomized placebo-controlled trial*. J Clin Endocrinol Metab* 83:3817-25 WHO Scientific Group (1996) Research on the menopause in the 1990's. *A report of the WHO Scientific Group*. World Health Organisation, Geneva, Switzerland, vol 866:1-79.
**4**
*Japan*
**The Validity of Body Composition Measurement**
Tissue energy production varies over time, and for practical purposes, can be organized into three main components, resting energy expenditure (REE), the thermic effect of food, and the thermic effect of physical activity (Ravussin & Bogardus, 1989). To maintain body weight, energy from food intake must equal energy expenditure. The energy requirement is defined as the average dietary energy intake that is predicted to maintain energy balance in healthy adults of a given age, gender, weight, height, and level of physical activity consistent with good health, and can be estimated from REE. REE reflects underlying tissue composition, mass, and metabolic activity (Elia, 1992) and accounts for 60–80% of total daily energy expenditure compared with the thermic effects of feeding and physical activity (~10% and ~15-30%, respectively). The REE has a large impact on the regulation of body mass and energy balance. In the field of obesity research, the presence and genesis of between-individual REE differences is a topic of great interest. To date, some earlier studies demonstrated that body mass, especially fat-free mass (FFM), has been a useful candidate in estimating REE (Ravussin & Bogardus, 1989; Fukagawa et al., 1990; Tataranni & Ravussin, 1995). Therefore, it is important to accurately evaluate REE and to measure body
A strategy for exploring between-individual differences in REE is to apply a tissue organ prediction model (Elia, 1992; Wang et al., 2000; Gallagher et al., 1998 & 2000). Each tissue and organ mass is quantified using either computed tomography (CT) or magnetic resonance imaging (MRI), and assigned assumed specific resting metabolic rates (Elia, 1992). The product of tissue organ mass and specific resting metabolic rate is then taken as the
composition by methods that are practical, precise, and accurate.
tissue organ REE, with the sum representing total body REE.
**1. Introduction**
**Using Dual Energy X-Ray Absorptiometry for**
**Estimating Resting Energy Expenditure**
Chiyoko Usui1,2, Motoko Taguchi3,
*1Department of Health Promotion and Exercise, National Institute of Health and Nutrition,*
*3Japan Women's College of Physical Education,*
*4Department of Nutritional Education, National Institute of Health and Nutrition, 5Faculty of Sport Sciences, Waseda University,*
Kazuko Ishikawa-Takata4 and Mitsuru Higuchi5
*2Research Fellow of the Japan Society for the Promotion of Science,*
### **The Validity of Body Composition Measurement Using Dual Energy X-Ray Absorptiometry for Estimating Resting Energy Expenditure**
Chiyoko Usui1,2, Motoko Taguchi3, Kazuko Ishikawa-Takata4 and Mitsuru Higuchi5 *1Department of Health Promotion and Exercise, National Institute of Health and Nutrition, 2Research Fellow of the Japan Society for the Promotion of Science, 3Japan Women's College of Physical Education, 4Department of Nutritional Education, National Institute of Health and Nutrition, 5Faculty of Sport Sciences, Waseda University, Japan*
#### **1. Introduction**
Tissue energy production varies over time, and for practical purposes, can be organized into three main components, resting energy expenditure (REE), the thermic effect of food, and the thermic effect of physical activity (Ravussin & Bogardus, 1989). To maintain body weight, energy from food intake must equal energy expenditure. The energy requirement is defined as the average dietary energy intake that is predicted to maintain energy balance in healthy adults of a given age, gender, weight, height, and level of physical activity consistent with good health, and can be estimated from REE. REE reflects underlying tissue composition, mass, and metabolic activity (Elia, 1992) and accounts for 60–80% of total daily energy expenditure compared with the thermic effects of feeding and physical activity (~10% and ~15-30%, respectively). The REE has a large impact on the regulation of body mass and energy balance. In the field of obesity research, the presence and genesis of between-individual REE differences is a topic of great interest. To date, some earlier studies demonstrated that body mass, especially fat-free mass (FFM), has been a useful candidate in estimating REE (Ravussin & Bogardus, 1989; Fukagawa et al., 1990; Tataranni & Ravussin, 1995). Therefore, it is important to accurately evaluate REE and to measure body composition by methods that are practical, precise, and accurate.
A strategy for exploring between-individual differences in REE is to apply a tissue organ prediction model (Elia, 1992; Wang et al., 2000; Gallagher et al., 1998 & 2000). Each tissue and organ mass is quantified using either computed tomography (CT) or magnetic resonance imaging (MRI), and assigned assumed specific resting metabolic rates (Elia, 1992). The product of tissue organ mass and specific resting metabolic rate is then taken as the tissue organ REE, with the sum representing total body REE.
The Validity of Body Composition Measurement
**2.2 Dual energy X-ray absorptiometry (DXA)**
(FFM) and fat mass (FM) were calculated by BW and % body fat.
**2.3 Method for the calculation of tissue organ mass**
**2. Body composition analysis 2.1 Anthropometric measurements**
meters (kg/m2).
gastrointestinal tract.
**3. Resting energy expenditure (REE)**
**3.1 Method for measuring REE**
Using Dual Energy X-Ray Absorptiometry for Estimating Resting Energy Expenditure 47
Body weight (BW) was measured to the nearest 0.1 kg by using an electronic scale (Inner Scan; Tanita Co., Japan and UC-321; A&D Co., Ltd., Tokyo, Japan), and height (Ht) was measured to thenearest 0.1cm by using a stadiometer (YL-65 and ST-2M; Yagami Inc., Japan). BW and Ht were measured with subjects wearing light clothing without shoes. Body mass index (BMI) was calculated by dividing BW in kilograms by the square of height in
The percentage of fat (% body fat) and bone mineral content (BMC) of the whole body and appendicular lean soft tissue (LST) were measured by DXA (Hologic QDR-4500 DXA Scanner and Hologic QDT DXA Scanner; Hologic Inc., Whaltham, MA, USA). Fat-free mass
Tissue organ mass was calculated using the previously reported prediction model as follows. Bone mass (BM) was calculated by multiplying BMC times 1.85 (Snyder et al., 1975; Heymsfield et al., 1990). Adipose tissue mass (AT) was assumed to be 85% of total body fat (Heymsfield et al., 2002), leading to the model based on FM. Thus, AT was calculated by multiplying FM times 1.18. Skeletal muscle mass (SM) was calculated using the prediction model of Kim et al. (2002). Finally, residual mass (RM) was calculated as the difference between BW and the sum of the calculated BM, AT and SM. Residual mass includes all of the high-metabolic-rate tissues and organs such as heart, brain, liver, kidneys, spleen, and
SM (kg) = 1.13×LST (kg) - 0.02×age (years) + 0.97
Participants came to the laboratory either on the previous night and stayed overnight, or came in the morning. In the latter case, subjects were asked to minimize any walking while en route from their home to the laboratory before REE determination. The measured REE (REEm) was directly measured by open-circuit indirect calorimetry. Measurements were performed between 0700 and 0900 h after 10–12 h of fasting, except for water, in a room at constant room temperature (23–25℃). After entering the laboratory, subjects rested in the supine position for at least 30 min, and a face mask was put on. In the case of overnight stay, the subjects were quietly awakened at 0630 and were attached a face mask while remaining in bed for 30 minutes. Two samples of expired air were collected in Douglas bags for a duration of 10 min each, and the mean value was used for the analysis. For young subjects,
BM (kg) = BMC (g)×1.85/1000 AT (kg) = FM (kg)×1.18
RM (kg) = BW - (BM + AT + SM)
REE = k1×tissue organ mass1 + k2×tissue organ mass2 + k3×tissue organ mass3 + ···
Dual-energy X-ray absorptiometry (DXA) can easily and accurately assess the body composition, including bone mineral content (BMC), fat mass (FM), and lean soft tissue mass (LST) of the whole body and the segments (Mazes et al., 1990; Svendsen et al., 1991 & 1993). In addition, recent studies support the use of DXA as a means of providing a "metabolic map" for exploring between individual or group differences in observed REE values (Hunter et al., 2001; Hayes et al., 2002; Usui et al., 2009). Specifically, the LST of the arms and legs, as provided by regional DXA estimates, represents primarily lowmetabolic rate skeletal muscle tissue. In contrast, LST of the head and trunk includes all of the tissue organs with high metabolic rates, such as brain, heart, liver, kidneys, spleen, and gastrointestinal tract. Total body-fat estimates reflect low metabolic rate adipose tissue. The various regional and whole-body estimates thus provide a qualitative metabolic body composition map of high and low metabolic rate components. Hayes et al. (2002) demonstrated that REE can be estimated from five measured DXA values: body weight, total body fat mass, bone mineral content, appendicular lean mass, and head area. Their study showed that no bias was detected between measured and predicted REEs (Hayes et al., 2002). These results were quite good and support the overall concept that predictable relations exist between major tissue organ level components and heat production at rest. In view of the finding of Hayes et al. (2002), we improved the estimation model from the five component model to four component model (adipose tissue, skeletal muscle, bone, and residual tissue organs). It was also evaluated the possibility that measurement of the magnitude and distribution of fundamental somatic heat-producing units using DXA can be used to estimate REE in both young and elderly women with different aerobic fitness levels (Usui et al., 2009). We suggested that REE in adult women can be estimated from four tissue organ components by using DXA regardless of age and aerobic fitness levels.
On the other hand, the traditional REE estimation approach, now widely applied, is to link REE with body composition determinants using a two-compartment model consisting of FM and FFM (Wang et al., 2000). Empirical regression models are developed with REE set as the dependent variable, and age, race, sex, height, and weight (FFM and FM) set as potential predictor variables. The Harris–Benedict equation (Harris & Benedict, 1919), Schofield equation (Schofield, 1985), and the Food and Agriculture Organization of the United Nations/World Health Organization/United Nations University (FAO/WHO/UNU) equation (FAO/WHO/UNU, 1985) are internationally used. In Japan, Dietary Reference Intakes for Japanese (DRI (Japan)) provides basal metabolic rate (BMR ≒ REE) standards according to sex and age categories (Ministry of Heaith, Labour and Welfare of Japan, 2009). REE can be calculated as BMR standards multiplied by body weight. In addition, Ganpule et al. (2007) recently developed new predictive equations (NIHN (Japan)) for REE in Japanese.
In the present chapter, we evaluated the validity of body composition measurement using DXA for estimating REE. The REEs which are predicted by the equations of Harris– Benedict, Schofield, FAO/WHO/UNU, DRI (Japan), and NIHN (Japan) will then be crossvalidated against our measured and predicted values by using DXA.
#### **2. Body composition analysis**
46 Dual Energy X-Ray Absorptiometry
REE = k1×tissue organ mass1 + k2×tissue organ mass2 + k3×tissue organ mass3 + ···
Dual-energy X-ray absorptiometry (DXA) can easily and accurately assess the body composition, including bone mineral content (BMC), fat mass (FM), and lean soft tissue mass (LST) of the whole body and the segments (Mazes et al., 1990; Svendsen et al., 1991 & 1993). In addition, recent studies support the use of DXA as a means of providing a "metabolic map" for exploring between individual or group differences in observed REE values (Hunter et al., 2001; Hayes et al., 2002; Usui et al., 2009). Specifically, the LST of the arms and legs, as provided by regional DXA estimates, represents primarily lowmetabolic rate skeletal muscle tissue. In contrast, LST of the head and trunk includes all of the tissue organs with high metabolic rates, such as brain, heart, liver, kidneys, spleen, and gastrointestinal tract. Total body-fat estimates reflect low metabolic rate adipose tissue. The various regional and whole-body estimates thus provide a qualitative metabolic body composition map of high and low metabolic rate components. Hayes et al. (2002) demonstrated that REE can be estimated from five measured DXA values: body weight, total body fat mass, bone mineral content, appendicular lean mass, and head area. Their study showed that no bias was detected between measured and predicted REEs (Hayes et al., 2002). These results were quite good and support the overall concept that predictable relations exist between major tissue organ level components and heat production at rest. In view of the finding of Hayes et al. (2002), we improved the estimation model from the five component model to four component model (adipose tissue, skeletal muscle, bone, and residual tissue organs). It was also evaluated the possibility that measurement of the magnitude and distribution of fundamental somatic heat-producing units using DXA can be used to estimate REE in both young and elderly women with different aerobic fitness levels (Usui et al., 2009). We suggested that REE in adult women can be estimated from four tissue organ components by using DXA
On the other hand, the traditional REE estimation approach, now widely applied, is to link REE with body composition determinants using a two-compartment model consisting of FM and FFM (Wang et al., 2000). Empirical regression models are developed with REE set as the dependent variable, and age, race, sex, height, and weight (FFM and FM) set as potential predictor variables. The Harris–Benedict equation (Harris & Benedict, 1919), Schofield equation (Schofield, 1985), and the Food and Agriculture Organization of the United Nations/World Health Organization/United Nations University (FAO/WHO/UNU) equation (FAO/WHO/UNU, 1985) are internationally used. In Japan, Dietary Reference Intakes for Japanese (DRI (Japan)) provides basal metabolic rate (BMR ≒ REE) standards according to sex and age categories (Ministry of Heaith, Labour and Welfare of Japan, 2009). REE can be calculated as BMR standards multiplied by body weight. In addition, Ganpule et al. (2007) recently developed new predictive equations
In the present chapter, we evaluated the validity of body composition measurement using DXA for estimating REE. The REEs which are predicted by the equations of Harris– Benedict, Schofield, FAO/WHO/UNU, DRI (Japan), and NIHN (Japan) will then be cross-
validated against our measured and predicted values by using DXA.
regardless of age and aerobic fitness levels.
(NIHN (Japan)) for REE in Japanese.
#### **2.1 Anthropometric measurements**
Body weight (BW) was measured to the nearest 0.1 kg by using an electronic scale (Inner Scan; Tanita Co., Japan and UC-321; A&D Co., Ltd., Tokyo, Japan), and height (Ht) was measured to thenearest 0.1cm by using a stadiometer (YL-65 and ST-2M; Yagami Inc., Japan). BW and Ht were measured with subjects wearing light clothing without shoes. Body mass index (BMI) was calculated by dividing BW in kilograms by the square of height in meters (kg/m2).
#### **2.2 Dual energy X-ray absorptiometry (DXA)**
The percentage of fat (% body fat) and bone mineral content (BMC) of the whole body and appendicular lean soft tissue (LST) were measured by DXA (Hologic QDR-4500 DXA Scanner and Hologic QDT DXA Scanner; Hologic Inc., Whaltham, MA, USA). Fat-free mass (FFM) and fat mass (FM) were calculated by BW and % body fat.
#### **2.3 Method for the calculation of tissue organ mass**
Tissue organ mass was calculated using the previously reported prediction model as follows. Bone mass (BM) was calculated by multiplying BMC times 1.85 (Snyder et al., 1975; Heymsfield et al., 1990). Adipose tissue mass (AT) was assumed to be 85% of total body fat (Heymsfield et al., 2002), leading to the model based on FM. Thus, AT was calculated by multiplying FM times 1.18. Skeletal muscle mass (SM) was calculated using the prediction model of Kim et al. (2002). Finally, residual mass (RM) was calculated as the difference between BW and the sum of the calculated BM, AT and SM. Residual mass includes all of the high-metabolic-rate tissues and organs such as heart, brain, liver, kidneys, spleen, and gastrointestinal tract.
> BM (kg) = BMC (g)×1.85/1000 AT (kg) = FM (kg)×1.18 SM (kg) = 1.13×LST (kg) - 0.02×age (years) + 0.97 RM (kg) = BW - (BM + AT + SM)
#### **3. Resting energy expenditure (REE)**
#### **3.1 Method for measuring REE**
Participants came to the laboratory either on the previous night and stayed overnight, or came in the morning. In the latter case, subjects were asked to minimize any walking while en route from their home to the laboratory before REE determination. The measured REE (REEm) was directly measured by open-circuit indirect calorimetry. Measurements were performed between 0700 and 0900 h after 10–12 h of fasting, except for water, in a room at constant room temperature (23–25℃). After entering the laboratory, subjects rested in the supine position for at least 30 min, and a face mask was put on. In the case of overnight stay, the subjects were quietly awakened at 0630 and were attached a face mask while remaining in bed for 30 minutes. Two samples of expired air were collected in Douglas bags for a duration of 10 min each, and the mean value was used for the analysis. For young subjects,
The Validity of Body Composition Measurement
2007) equations (Table 1).
**4.2 Statistical analysis**
**REE**
**4.1 Subjects**
**3.3 Predictive equations of REE for women**
Faculty of Sport Sciences of Waseda University, in Saitama.
Statistical significance was set at p < 0.05 for all predictors.
Using Dual Energy X-Ray Absorptiometry for Estimating Resting Energy Expenditure 49
Other predictive REEs were calculated using the Harris–Benedict (Harris & Benedict, 1919), Schofield (Schofield, 1985), FAO/WHO/UNU (FAO/WHO/UNU, 1985), DRI (Japan) (Ministry of Heaith, Labour and Welfare of Japan, 2009), and NIHN (Japan) (Ganpule et al.,
**4. The validity of body composition measurement using DXA for estimating**
The data used for the current analysis were collected from different 3 experimental studies that followed a similar methodology (Usui et al., 2009; Taguchi et al., 2011; Hasegawa et al., 2011). Total of 288 healthy women (216 young women; age: 21.8 ± 2.0 (18-29) years and 72 elderly women; age: 63.3 ± 6.4 (50-77) years) were recruited for the study. The elderly subjects had passed three years or more (13.5 ± 7.4 years) after menopause. None had used medications that affect bone and estrogen replacement was eliminated from the analysis. All subjects were informed about the purpose and possible risks of the study and were then provided written informed consent. These studies were conducted according to the guidelines laid down in the Declaration of Helsinki. All procedures involving human subjects were approved by the Ethical Committee of the National Institute of Health and Nutrition in Tokyo, the Institutional Ethical Committee Review Board of Japan Women's College of Physical Education in Tokyo, and the Human Research Ethical Committee of the
Results are presented as the mean ± standard deviation (SD). Statistical analyses were carried out with the Sigma Stat 3.5 (Systat Software Inc., CA, USA). Statistical analysis was performed using the Student's t-test for parametric variables and the Mann-Whitney rank sum test for nonparametric variables to determine differences between young and elderly women. In addition to the mean ± SD of the difference, total error was used to determine how accurately predicted REE matched measured REE. This statistic includes two sources of variation, one attributable to the lack of association between the two sets of measurement (standard error of estimate) and one attributable to the difference between the means (van der Ploeg et al., 2001). Statistical significance of differences between measured and predicted values in all subjects was analyzed by one-way repeatedmeasures analysis of variance (ANOVA) and Dunnett's post hoc test. Evaluation of bias between measured and predicted REE was conducted with a Bland–Altman analysis (Bland and Altman, 1986). Fixed bias was indicated if the 95% coefficient interval (CI) for the difference between measured and predicted REE did not include zero. Predictive equations that are accurate display a tight prediction interval around zero. The 95%CI below zero signify an underestimation, while the 95%CI above zero signify an overestimation. The correlation of the difference between measured and predicted REE values and the mean from the both REE values was utilized to assess proportional bias.
all measurements were made during the follicular phase of the menstrual cycle. An oxygen and carbon dioxide analyzer (Arco-1000A; Arco system, Japan and AE-300; Minato Medical Science, Tokyo) was used to analyze the rate of oxygen consumption and carbon dioxide production. The volume of expired air was determined using a dry gas volume meter (DC-5; Shinagawa, Japan) and converted to standard temperature, standard pressure and dry gas. Gas exchange results were converted to REE (kcal/day) using Weir's equation (Weir, 1949). To examine whether overnight stay before the REE measurement caused a significant difference in the observed REE, analysis of covariance with REE as the dependent variable and age, height, FFM, and FM as covariates was employed. No significant effect of the measurement conditions was observed (overnight stay: 1169 ± 12 kcal/day (mean ± standard error (SE)), came in the early morning on the day: 1170 ± 7 kcal/day (mean ± SE), F = 0.001, p = 0.980).
#### **3.2 Method for estimating REE**
Estimation of REE using DXA was obtained based on the sum of four body compartments (BM, AT, SM and RM) times the corresponding tissue respiration rates as follows. The specific resting metabolic rate of the four compartments was assumed from previously reported data, bone (2.3 kcal/kg), AT (4.5 kcal/kg), skeletal muscle (13 kcal/kg) and residual (54 kcal/kg) (Holliday et al., 1967; Grande, 1989; Elia, 1992; Hayes et al., 2002; Heymsfield et al., 2002).
REE (kcal/day) = 2.3×BM + 4.5×AT + 13×SM + 54×RM
BM: bone mass (kg), AT: adipose tissue mass (kg), SM: skeletal muscle mass (kg), RM: residual mass (kg), BW: body weight (kg), Ht: height (cm), A: age (years).
Table 1. Predictive equations of resting energy expenditure for women used in the present study.
#### **3.3 Predictive equations of REE for women**
Other predictive REEs were calculated using the Harris–Benedict (Harris & Benedict, 1919), Schofield (Schofield, 1985), FAO/WHO/UNU (FAO/WHO/UNU, 1985), DRI (Japan) (Ministry of Heaith, Labour and Welfare of Japan, 2009), and NIHN (Japan) (Ganpule et al., 2007) equations (Table 1).
#### **4. The validity of body composition measurement using DXA for estimating REE**
#### **4.1 Subjects**
48 Dual Energy X-Ray Absorptiometry
all measurements were made during the follicular phase of the menstrual cycle. An oxygen and carbon dioxide analyzer (Arco-1000A; Arco system, Japan and AE-300; Minato Medical Science, Tokyo) was used to analyze the rate of oxygen consumption and carbon dioxide production. The volume of expired air was determined using a dry gas volume meter (DC-5; Shinagawa, Japan) and converted to standard temperature, standard pressure and dry gas. Gas exchange results were converted to REE (kcal/day) using Weir's equation (Weir, 1949). To examine whether overnight stay before the REE measurement caused a significant difference in the observed REE, analysis of covariance with REE as the dependent variable and age, height, FFM, and FM as covariates was employed. No significant effect of the measurement conditions was observed (overnight stay: 1169 ± 12 kcal/day (mean ± standard error (SE)), came in the early morning on the day: 1170 ± 7 kcal/day (mean ± SE), F
Estimation of REE using DXA was obtained based on the sum of four body compartments (BM, AT, SM and RM) times the corresponding tissue respiration rates as follows. The specific resting metabolic rate of the four compartments was assumed from previously reported data, bone (2.3 kcal/kg), AT (4.5 kcal/kg), skeletal muscle (13 kcal/kg) and residual (54 kcal/kg) (Holliday et al., 1967; Grande, 1989; Elia, 1992; Hayes et al., 2002;
REE (kcal/day) = 2.3×BM + 4.5×AT + 13×SM + 54×RM
**30-59 (0.034×BW + 3.538)×1000/4.186 60 over (0.038×BW + 2.755)×1000/4.186**
**NIHN (Japan) - (0.0481×BW + 0.0234×Ht - 0.0138×A - 0.9708)×1000/4.186**
BM: bone mass (kg), AT: adipose tissue mass (kg), SM: skeletal muscle mass (kg), RM: residual mass
Table 1. Predictive equations of resting energy expenditure for women used in the present
**30-59 (36.4×BW - 104.6×Ht/100 + 3619)/4.186 60 over (38.5×BW + 2665.2×Ht/100 - 1264)/4.186**
**DXA - 2.3×BM + 4.5×AT + 13×SM + 54×RM**
**FAO/WHO/UNU 18-29 (55.6×BW + 1397.4×Ht/100 + 146)/4.186**
**30-49 21.7 ×BW 50 over 20.7 ×BW**
**Schofield 18-29 (0.062×BW + 2.036)×1000/4.186**
**Harris-Benedict - 655.0955 + 9.5634×BW + 1.8496×Ht - 4.6756×A**
= 0.001, p = 0.980).
Heymsfield et al., 2002).
**Predictive equations**
study.
**(kcal/day) Age range**
**DRI (Japan) 18-29 22.1 ×BW**
(kg), BW: body weight (kg), Ht: height (cm), A: age (years).
**3.2 Method for estimating REE**
The data used for the current analysis were collected from different 3 experimental studies that followed a similar methodology (Usui et al., 2009; Taguchi et al., 2011; Hasegawa et al., 2011). Total of 288 healthy women (216 young women; age: 21.8 ± 2.0 (18-29) years and 72 elderly women; age: 63.3 ± 6.4 (50-77) years) were recruited for the study. The elderly subjects had passed three years or more (13.5 ± 7.4 years) after menopause. None had used medications that affect bone and estrogen replacement was eliminated from the analysis. All subjects were informed about the purpose and possible risks of the study and were then provided written informed consent. These studies were conducted according to the guidelines laid down in the Declaration of Helsinki. All procedures involving human subjects were approved by the Ethical Committee of the National Institute of Health and Nutrition in Tokyo, the Institutional Ethical Committee Review Board of Japan Women's College of Physical Education in Tokyo, and the Human Research Ethical Committee of the Faculty of Sport Sciences of Waseda University, in Saitama.
#### **4.2 Statistical analysis**
Results are presented as the mean ± standard deviation (SD). Statistical analyses were carried out with the Sigma Stat 3.5 (Systat Software Inc., CA, USA). Statistical analysis was performed using the Student's t-test for parametric variables and the Mann-Whitney rank sum test for nonparametric variables to determine differences between young and elderly women. In addition to the mean ± SD of the difference, total error was used to determine how accurately predicted REE matched measured REE. This statistic includes two sources of variation, one attributable to the lack of association between the two sets of measurement (standard error of estimate) and one attributable to the difference between the means (van der Ploeg et al., 2001). Statistical significance of differences between measured and predicted values in all subjects was analyzed by one-way repeatedmeasures analysis of variance (ANOVA) and Dunnett's post hoc test. Evaluation of bias between measured and predicted REE was conducted with a Bland–Altman analysis (Bland and Altman, 1986). Fixed bias was indicated if the 95% coefficient interval (CI) for the difference between measured and predicted REE did not include zero. Predictive equations that are accurate display a tight prediction interval around zero. The 95%CI below zero signify an underestimation, while the 95%CI above zero signify an overestimation. The correlation of the difference between measured and predicted REE values and the mean from the both REE values was utilized to assess proportional bias. Statistical significance was set at p < 0.05 for all predictors.
The Validity of Body Composition Measurement
**predicted REE**
Dunnett's post hoc test).
Table 4. Bland–Altman analysis.
**4.4 Discussion**
against the measured REE of healthy women.
**REEm (kcal/day) 1170 ± 167**
Using Dual Energy X-Ray Absorptiometry for Estimating Resting Energy Expenditure 51
**DXA 1181 ± 152 11 ± 104 105 Harris-Benedict 1324 ± 123 154 ± 121 <sup>c</sup> 196 Schofield 1273 ± 119 104 ± 111 <sup>c</sup> 151 FAO/WHO/UNU 1281 ± 118 111 ± 109 <sup>c</sup> 156 DRI (Japan) 1191 ± 166 22 ± 114 115 NIHN (Japan) 1184 ± 136 14 ± 110 111**
Means ± SD.; a Mean error = predicted REE – measured REE. ; b Total error (kcal/day) = √(∑(predicted REE – measured REE)2/n). ; c p<0.05 vs. measured REE (one-way repeated-measures ANOVA and
Table 3. Measured and predicted REE and cross-validation of REE prediction equations
**predicted REE** r *p* values
95%CI
**DXA** -1.1 ~ 23.0 N.S. -0.152 p = 0.010 **Harris-Benedict** 140.2 ~ 168.3 -0.395 p < 0.001 **Schofield** 90.8 ~ 116.5 -0.471 p < 0.001 **FAO/WHO/UNU** 98.5 ~ 123.8 -0.481 p < 0.001 **DRI (Japan)** 8.4 ~ 34.8 -0.010 p = 0.868 N.S. **NIHN (Japan)** 1.6 ~ 27.2 -0.307 p < 0.001
With the increasing availability of imaging methods, CT, MRI, and DXA, estimating methods of REE are developing and providing a "new" viewpoint of REE from the modeling perspective of organs and tissues. The importance of the developed approach is that it not only provides a qualitative estimate of REE, but of the actual distribution of heatproducing tissue components. As whole-body CT and MRI cannot assess body composition easily and quickly, previous investigators have explored the use of DXA as an available and practical alternative for developing qualitative tissue organ REE predictions. Hayes et al. (2002) demonstrated that REE can be estimated from five measured DXA values: body weight, total body fat mass, bone mineral content, appendicular lean mass, and head area. Their findings suggested that when REE is reviewed in the context of major heat-producing
**All subjects Total error<sup>b</sup> (n = 288) (kcal/day) (kcal/day) (kcal/day)**
**REE Mean error<sup>a</sup>**
Bland-Altman analysis
fixed bias proportional bias
#### **4.3 Results**
Table 2 presents the comparisons of subjects characteristics and body composition. Ht was significantly lower in the elderly women (50 over age group) than in young women (18-29 age group). However, no significant difference in BW was noted between two groups. The elderly women had significantly higher levels of % body fat and FM, and lower levels of FFM than the young group. When the FFM was separated into BM, SM, and RM, young women had significantly higher BM, SM, and RM than elderly women.
Means ± SD, Ht: Height, BW: body weight, BMI: body mass index, FFM: fat-free mass, FM: fat mass, BM: bone mass, AT: adipose tissue mass, SM: skeletal muscle mass, RM: residual mass, \* p<0.05 vs. 18- 29 age group (Student's t-test), † p<0.05 vs. 18-29 age group (Mann-Whitney rank sum test).
Table 2. Physical characteristics in the healthy women.
Table 3 show measured and predicted REEs, and the mean and total errors of predicted REEs. No significant differences were observed from the REEm in predicted REE values by using DXA, DRI (Japan) equation, and NIHN (Japan) equation in all subjects. On the other hand, the mean errors of REE predicted by the Harris–Benedict, Schofield, and FAO/WHO/UNU equations were significantly higher than the REEm in all subjects. Total error of estimated REE by using DXA was lowest in the group. Total error using NIHN (Japan) equation was the second to the lowest. In particular, total error of the Harris– Benedict equation was largest in healthy women.
The systemic bias between the measured and predicted REEs by the Bland-Altman analysis are shown in Table 4 and Fig. 1. Fixed bias was not present in only estimated REE by using DXA. However, the REE estimated by using DXA had a proportional bias (Fig. 1 (a)). Systematic bias (fixed and proportional bias) was presented in predicted REE by the Harris– Benedict, Schofield, FAO/WHO/UNU, and NIHN (Japan) equations.
Means ± SD.; a Mean error = predicted REE – measured REE. ; b Total error (kcal/day) = √(∑(predicted REE – measured REE)2/n). ; c p<0.05 vs. measured REE (one-way repeated-measures ANOVA and Dunnett's post hoc test).
Table 3. Measured and predicted REE and cross-validation of REE prediction equations against the measured REE of healthy women.
Table 4. Bland–Altman analysis.
#### **4.4 Discussion**
50 Dual Energy X-Ray Absorptiometry
Table 2 presents the comparisons of subjects characteristics and body composition. Ht was significantly lower in the elderly women (50 over age group) than in young women (18-29 age group). However, no significant difference in BW was noted between two groups. The elderly women had significantly higher levels of % body fat and FM, and lower levels of FFM than the young group. When the FFM was separated into BM, SM, and RM, young
**Age (years) 32.1 ± 18.4 21.8 ± 2.0 63.3 ± 6.4** † **Ht (cm) 159.7 ± 6.8 161.6 ± 6.2 153.9 ± 5.2 \* BW (kg) 54.8 ± 7.4 54.9 ± 7.8 54.3 ± 6.0**
**FFM (kg) 41.4 ± 5.8 42.6 ± 5.9 37.7 ± 3.5** † **FM (kg) 13.4 ± 4.0 12.3 ± 3.4 16.6 ± 3.9** † **% body fat 24.3 ± 5.8 22.3 ± 4.6 30.2 ± 4.8 \* BM (kg) 3.8 ± 0.7 4.0 ± 0.6 3.1 ± 0.5** † **AT (kg) 15.8 ± 4.7 14.5 ± 4.0 19.6 ± 4.6** † **SM (kg) 19.5 ± 3.5 20.5 ± 3.3 16.5 ± 1.9** † **RM (kg) 15.7 ± 2.1 15.9 ± 2.2 15.2 ± 1.5** †
Means ± SD, Ht: Height, BW: body weight, BMI: body mass index, FFM: fat-free mass, FM: fat mass, BM: bone mass, AT: adipose tissue mass, SM: skeletal muscle mass, RM: residual mass, \* p<0.05 vs. 18-
Table 3 show measured and predicted REEs, and the mean and total errors of predicted REEs. No significant differences were observed from the REEm in predicted REE values by using DXA, DRI (Japan) equation, and NIHN (Japan) equation in all subjects. On the other hand, the mean errors of REE predicted by the Harris–Benedict, Schofield, and FAO/WHO/UNU equations were significantly higher than the REEm in all subjects. Total error of estimated REE by using DXA was lowest in the group. Total error using NIHN (Japan) equation was the second to the lowest. In particular, total error of the Harris–
The systemic bias between the measured and predicted REEs by the Bland-Altman analysis are shown in Table 4 and Fig. 1. Fixed bias was not present in only estimated REE by using DXA. However, the REE estimated by using DXA had a proportional bias (Fig. 1 (a)). Systematic bias (fixed and proportional bias) was presented in predicted REE by the Harris–
29 age group (Student's t-test), † p<0.05 vs. 18-29 age group (Mann-Whitney rank sum test).
Table 2. Physical characteristics in the healthy women.
Benedict equation was largest in healthy women.
Benedict, Schofield, FAO/WHO/UNU, and NIHN (Japan) equations.
**All 18-29 50 over n = 288 n = 216 n = 72**
**Age range**
**) 21.5 ± 2.6 21.0 ± 2.5 22.9 ± 2.3 \***
women had significantly higher BM, SM, and RM than elderly women.
**4.3 Results**
**BMI (kg/m2**
With the increasing availability of imaging methods, CT, MRI, and DXA, estimating methods of REE are developing and providing a "new" viewpoint of REE from the modeling perspective of organs and tissues. The importance of the developed approach is that it not only provides a qualitative estimate of REE, but of the actual distribution of heatproducing tissue components. As whole-body CT and MRI cannot assess body composition easily and quickly, previous investigators have explored the use of DXA as an available and practical alternative for developing qualitative tissue organ REE predictions. Hayes et al. (2002) demonstrated that REE can be estimated from five measured DXA values: body weight, total body fat mass, bone mineral content, appendicular lean mass, and head area. Their findings suggested that when REE is reviewed in the context of major heat-producing
The Validity of Body Composition Measurement
ratio of organ–tissue mass to body size.
women.
Using Dual Energy X-Ray Absorptiometry for Estimating Resting Energy Expenditure 53
units differing in metabolic activity, normal-weight men and women are actually quite similar. In addition, this five component model allowed REE prediction over a wide range of FFM, because the mean REE prediction errors in intermediate weight, underweight and obese subjects were not significantly different between groups (Bosy-Westphal et al., 2004). These findings provided indirect evidence for a view that, for practical purposes within humans, the specific metabolic rate is constant with increasing organ mass, and that interindividual REE differences in under- or overweight are more a reflection of an altered
On the other hand, there is evidence that REE is lower in the elderly, even after adjustment for tissue and organ mass (Gallagher et al., 2000) or age-related differences in body composition (Van Pelt et al., 1997, 2001). This raises the important question of the validity of assumed tissue- and organ-specific metabolic rates. The specific metabolic rates applied in the present study represent literature averages. The extent to which these heat production rates are valid remains unknown. Moreover, the specific metabolic rates for some components reported in the literature vary widely. However, we demonstrated that estimation based on the four tissue organs by using DXA allows successful calculation of REE in female adults regardless of age and aerobic fitness levels. (Usui et al., 2009). Our findings suggested the possibility that REE is regulated mainly by the mass of the tissue organs with lower and higher metabolic rates, including skeletal muscle and intestinal organs, rather than a decline in the specific metabolic rate of different tissue organs associated with advancing age and decreasing aerobic fitness levels in young and elderly
In addition, as for the four component model of REE prediction, Taguchi et al. (2011) evaluated that the relationship between REE and body composition in Japanese female athletes with a wide range of body sizes. Their results indicated, in good agreement with previous study (Bosy-Westphal et al., 2004), that REE for female athletes with a wide range of body sizes can be attributed to changes in organ tissue mass, and not changes in organ tissue metabolic rate. Moreover, Hasegawa et al. (2011) investigated the differences in body composition and REE between young women with low (BMI < 18.5 kg/m2) and normal BMI (18.5≦BMI<25 kg/m2), and suggested the possibility that low BMI subjects with normal menstrual cycle do not have any differences in specific metabolic rates of different tissue organs compared to those with normal BMI. According to Bland-Altman analysis in the present chapter, although there is a proportional bias in estimation of REE by using DXA, this proportional bias is relatively small compared with other estimation methods for REE (Table 4 and Fig. 1). Furthermore, fixed bias was not present in only estimated REE by using DXA. These findings suggested that estimation of the four tissue organs by using DXA can
easily, practically, precisely, and accurately predict REE in healthy men and women.
The present chapter also focused on the validity of body composition measurement using DXA for estimating REE. Our results showed that predicted REE values by using DXA, DRI (Japan) equations, and NIHN (Japan) equation were not significantly different from the REEm, while Harris–Benedict, Schofield, and FAO/WHO/UNU equations significantly higher than the REEm in healthy women (Table 3). DRI (Japan) and NIHN (Japan) equations were developed based on the data for Japanese subjects with standard body size (Ministry of Heaith, Labour and Welfare of Japan, 2009; Ganpule et al., 2007). Schofield and FAO/WHO/UNU equations were developed based on data from a population of many
Plots of the differences between measured REE and predicted REE. Open circle (○): young subjects (n = 216), black circle (●): elderly subjects (n = 72), the solid line ( ): regression line in all subjects (n = 288), REEm: measured by expiratory gas exchange, REEDXA: estimated by using DXA, REEH-B: predicted by Harris– Benedict equation, REEscho: predicted by Schofield equations, REEF/W/U: predicted by FAO/WHO/UNU equations, REEDRI: predicted by DRI (Japan) equations, REENIHN: predicted by NIHN (Japan) equation.
Fig. 1. Bland–Altman plot.
**REENIHN**
Plots of the differences between measured REE and predicted REE. Open circle (○): young subjects (n = 216), black circle (●): elderly subjects (n = 72), the solid line ( ): regression line in all subjects (n = 288), REEm: measured by expiratory gas exchange, REEDXA: estimated by using DXA, REEH-B: predicted by Harris– Benedict equation, REEscho: predicted by Schofield equations, REEF/W/U: predicted by FAO/WHO/UNU equations, REEDRI: predicted by DRI (Japan) equations, REENIHN: predicted by NIHN (Japan) equation.
**minus REEm (kcal/day)**
**REEF/W/U**
**minus REEm (kcal/day)**
**REEH-B**
**minus REEm (kcal/day)**
600 800 1000 1200 1400 1600 1800 2000
**all; r = -0.395, p < 0.001**
**all; r = -0.481, p < 0.001**
**all; r = -0.307, p < 0.001**
**mean REEm and REEH-B (kcal/day)**
600 800 1000 1200 1400 1600 1800 2000
**mean REEm and REEF/W/U (kcal/day)**
600 800 1000 1200 1400 1600 1800 2000
**mean REEm and REENIHN (kcal/day)**
Fig. 1. Bland–Altman plot.
**REEDRI**
**minus REEm (kcal/day)**
**REEScho**
**minus REEm (kcal/day)**
**REEDXA**
**minus REEm (kcal/day)**
600 800 1000 1200 1400 1600 1800 2000
**all; r = -0.152, p < 0.01**
(a) (b)
**all; r = -0.471, p < 0.001**
(c) (d)
(e) (f)
**mean REEm and REEDXA (kcal/day)**
600 800 1000 1200 1400 1600 1800 2000
**mean REEm and REEScho (kcal/day)**
600 800 1000 1200 1400 1600 1800 2000
**mean REEm and REEDRI (kcal/day)**
units differing in metabolic activity, normal-weight men and women are actually quite similar. In addition, this five component model allowed REE prediction over a wide range of FFM, because the mean REE prediction errors in intermediate weight, underweight and obese subjects were not significantly different between groups (Bosy-Westphal et al., 2004). These findings provided indirect evidence for a view that, for practical purposes within humans, the specific metabolic rate is constant with increasing organ mass, and that interindividual REE differences in under- or overweight are more a reflection of an altered ratio of organ–tissue mass to body size.
On the other hand, there is evidence that REE is lower in the elderly, even after adjustment for tissue and organ mass (Gallagher et al., 2000) or age-related differences in body composition (Van Pelt et al., 1997, 2001). This raises the important question of the validity of assumed tissue- and organ-specific metabolic rates. The specific metabolic rates applied in the present study represent literature averages. The extent to which these heat production rates are valid remains unknown. Moreover, the specific metabolic rates for some components reported in the literature vary widely. However, we demonstrated that estimation based on the four tissue organs by using DXA allows successful calculation of REE in female adults regardless of age and aerobic fitness levels. (Usui et al., 2009). Our findings suggested the possibility that REE is regulated mainly by the mass of the tissue organs with lower and higher metabolic rates, including skeletal muscle and intestinal organs, rather than a decline in the specific metabolic rate of different tissue organs associated with advancing age and decreasing aerobic fitness levels in young and elderly women.
In addition, as for the four component model of REE prediction, Taguchi et al. (2011) evaluated that the relationship between REE and body composition in Japanese female athletes with a wide range of body sizes. Their results indicated, in good agreement with previous study (Bosy-Westphal et al., 2004), that REE for female athletes with a wide range of body sizes can be attributed to changes in organ tissue mass, and not changes in organ tissue metabolic rate. Moreover, Hasegawa et al. (2011) investigated the differences in body composition and REE between young women with low (BMI < 18.5 kg/m2) and normal BMI (18.5≦BMI<25 kg/m2), and suggested the possibility that low BMI subjects with normal menstrual cycle do not have any differences in specific metabolic rates of different tissue organs compared to those with normal BMI. According to Bland-Altman analysis in the present chapter, although there is a proportional bias in estimation of REE by using DXA, this proportional bias is relatively small compared with other estimation methods for REE (Table 4 and Fig. 1). Furthermore, fixed bias was not present in only estimated REE by using DXA. These findings suggested that estimation of the four tissue organs by using DXA can easily, practically, precisely, and accurately predict REE in healthy men and women.
The present chapter also focused on the validity of body composition measurement using DXA for estimating REE. Our results showed that predicted REE values by using DXA, DRI (Japan) equations, and NIHN (Japan) equation were not significantly different from the REEm, while Harris–Benedict, Schofield, and FAO/WHO/UNU equations significantly higher than the REEm in healthy women (Table 3). DRI (Japan) and NIHN (Japan) equations were developed based on the data for Japanese subjects with standard body size (Ministry of Heaith, Labour and Welfare of Japan, 2009; Ganpule et al., 2007). Schofield and FAO/WHO/UNU equations were developed based on data from a population of many
The Validity of Body Composition Measurement
**6. Acknowledgments**
Japanese Olympic Committee.
1522-1555
Online ISSN 1522-1555
8923, Online ISSN 1749-6632
0954-3007, Online ISSN 1476-5640
pp. 307–310, Print ISSN 0140-6736
0565, Online ISSN 1476-5497
Raven Press, ISBN 978-088-1678-71-0, New York
**7. References**
Using Dual Energy X-Ray Absorptiometry for Estimating Resting Energy Expenditure 55
We express our appreciation to the subjects for their cooperation in this study. We thank the members of the National Institute of Health and Nutrition and Waseda University for their help in this experiment. This study was supported by a Research Grant for Academic Frontier Projects from the Ministry of Education, Culture, Sports, Science and Technology (05F-02), a Waseda University Grant for Special Research Projects (2005A-932, 2006B-242), Health Sciences Research Grants from the Ministry of Health, Labor and Welfare, Medical Health Care Research Grants from the Consolidated Research Institute for Advanced Science and Medical Care at Waseda University, Research grant for Waseda University Global COE program: Sports Sciences for Promotion of Active Life (projectsIII-2), Research Grant from Japan Women's College of Physical Education, and Research Grants from the
Bland, J.M. & Altman, D.G. (1986). Statistical methods for assessing agreement between two
Bosy-Westphal, A.; Reinecke, U.; Schlörke, T.; Illner, K.; Kutzner, D.; Heller M. & Müller,
Elia, M. (1992). Organ and tissue contribution to metabolic rate. In: *Energy Metabolism: Tissue*
FAO/WHO/UNU. (1985). *Energy and protein requirements, report of a joint FAO/WHO/UNU expert consultation. Technical Report Series 724*, WHO, ISBN 92-4-120724-8, Geneva. Fukagawa, N.K.; Bndini, L.G. & Young, J.B. (1990). Effect of age on body composition and
Gallagher, D.; Belmonte, D.; Deurenberg, P.; Wang, Z.; Krasnow, N.; Pi-Sunyer, F.X. &
Gallagher, D.; Allen, A.; Wang, Z.; Heymsfield, S.B. & Krasnow, N. (2000). Smaller organ
Ganpule, A.A.; Tanaka, S.; Ishikawa–Takata, K. & Tabata, I. (2007). Interindividual
methods of clinical measurement. *The Lancet*, Vol.327, No.8476, (08 February 1986),
M.J. (2004). Effect of organ and tissue masses on resting energy expenditure in underweight, normal weight and obese adults. *International Journal of Obesity and Related Metabolic Disorders,* Vol.28, No.1, (January 2004), pp. 72–79, Print ISSN 0307-
*Determinants and Cellular Corollaries*, Kinney, J. M. & Tucker, H. N. (eds), pp. 61–80,
resting metabolic rate. *American Journal of Physiology Endocrinology and Metabolism*, Vol.259, No.2, (August 1990), pp. E233–E238, Print ISSN 0193-1849, Online ISSN
Heymsfield S.B. (1998). Organ-tissue mass measurement allows modeling of REE and metabolically active tissue mass. *American Journal of Physiology Endocrinology and Metabolism*, Vol.275, No.2, (August 1998), pp. E249–E258, Print ISSN 0193-1849,
tissue mass in the elderly fails to explain lower resting metabolic rate. *Annals of the New York Academy of Sciences,* Vol.904, (May 2000), pp. 449–455, Print ISSN 0077-
variability in sleeping metabolic rate in Japanese subjects. *European Journal of Clinical Nutrition*, Vol.61, No.11, (Epub February 2007), pp. 1256–1261, Print ISSN
races (Schofield, 1985; FAO/WHO/UNU, 1985). However, the data used to develop the Schofield equations were mostly from young European military and police recruits with 45% being of Italian descent. Harris–Benedict equation was developed using data obtained in healthy normal weight Caucasian men and women aged 15–74 years. Asians are reported to have lower REE than Europeans by 10–12% (Hayter et al., 1994), even after adjustment for body composition. According to a recent study by Miyake et al. (2011), mean difference and total error values were smaller using the DRI (Japan) equations and NIHN (Japan) equation than the internationally used equations (Harris–Benedict, Schofield, FAO/WHO/UNU) in both sexes. In particular, the total error was lower for the NIHN (Japan) equation than the other equations in most age groups (Miyake et al., 2011). In the present chapter, the mean errors of the predicted REEs by using DXA, DRI (Japan) equations, and NIHN (Japan) equation were smaller than those of internationally used equations in all subjects (Table 3). In addition, total error of estimated REE by using DXA, DRI (Japan) equations, and NIHN (Japan) equation were also lower than other total errors (105 kcal/day, 115 kcal/day, 111 kcal/day, respectively). Moreover, while these three prediction methods had some kind of bias, the bias was relatively small compared with other internationally used equations (Table 4 & Fig. 1 (a), (e), (f)). The results of this chapter supported the notions that the DRI (Japan) and NIHN (Japan) equations of REE are more accurate in healthy Japanese subjects, and that estimating REE from four tissue organ components by using DXA will be one of the useful methods for the dietary management (energetic assessment) in healthy Japanese subjects.
Our investigation has a few limitations. First, we did not test middle-aged (30–49 years) and very old (> 80 years) women and the adult men of all ages. It is unclear whether age-related functional decline in cells or tissue organs make a major impact on the heat production. Second, it was performed only in a Japanese population. Hayter et al. (1994) reported that Asians have lower REE than Europeans by 10–12%, even after adjustment for body composition. Third, because all the participants were healthy adult women, little is known as to whether this four component model for estimating REE by using DXA can accurately estimate REE in patients with a chronic disease, such as diabetes mellitus or cardiac disorder. Recently, DXA and regression modeling of REE showed that skeletal muscle is hypermetabolic in patients with HIV lipoatrophy (Kosmiski et al., 2009). Accordingly, the specific metabolic rate of each tissue organ may differ from those in healthy adults over 80 years of age, other ethnic subjects, or patients with a chronic disease. Future studies are needed to extend these observations and to analyze gender-related, genetics, hormonal, ethnic, and other determinant factors of REE.
#### **5. Conclusion**
The results of the present chapter suggest that REE in healthy adult women can be accurately estimated from four tissue organ components by using DXA, because the mean error, total error, and systematic bias between the predicted and measured REE was relatively small compared with other estimation methods for REE. Therefore, we expect that the DXA will be one of the useful methods not only for examining body composition and bone mineral density but also for estimating REE in a broad range of fields. In the future, DXA method may offer a bridge to effective prevention, treatment and care with the dietary management and exercise regimen for life style-related disease, such as obesity, osteoporosis, and type 2 diabetes.
#### **6. Acknowledgments**
54 Dual Energy X-Ray Absorptiometry
races (Schofield, 1985; FAO/WHO/UNU, 1985). However, the data used to develop the Schofield equations were mostly from young European military and police recruits with 45% being of Italian descent. Harris–Benedict equation was developed using data obtained in healthy normal weight Caucasian men and women aged 15–74 years. Asians are reported to have lower REE than Europeans by 10–12% (Hayter et al., 1994), even after adjustment for body composition. According to a recent study by Miyake et al. (2011), mean difference and total error values were smaller using the DRI (Japan) equations and NIHN (Japan) equation than the internationally used equations (Harris–Benedict, Schofield, FAO/WHO/UNU) in both sexes. In particular, the total error was lower for the NIHN (Japan) equation than the other equations in most age groups (Miyake et al., 2011). In the present chapter, the mean errors of the predicted REEs by using DXA, DRI (Japan) equations, and NIHN (Japan) equation were smaller than those of internationally used equations in all subjects (Table 3). In addition, total error of estimated REE by using DXA, DRI (Japan) equations, and NIHN (Japan) equation were also lower than other total errors (105 kcal/day, 115 kcal/day, 111 kcal/day, respectively). Moreover, while these three prediction methods had some kind of bias, the bias was relatively small compared with other internationally used equations (Table 4 & Fig. 1 (a), (e), (f)). The results of this chapter supported the notions that the DRI (Japan) and NIHN (Japan) equations of REE are more accurate in healthy Japanese subjects, and that estimating REE from four tissue organ components by using DXA will be one of the useful methods for the dietary management (energetic assessment) in healthy Japanese subjects.
Our investigation has a few limitations. First, we did not test middle-aged (30–49 years) and very old (> 80 years) women and the adult men of all ages. It is unclear whether age-related functional decline in cells or tissue organs make a major impact on the heat production. Second, it was performed only in a Japanese population. Hayter et al. (1994) reported that Asians have lower REE than Europeans by 10–12%, even after adjustment for body composition. Third, because all the participants were healthy adult women, little is known as to whether this four component model for estimating REE by using DXA can accurately estimate REE in patients with a chronic disease, such as diabetes mellitus or cardiac disorder. Recently, DXA and regression modeling of REE showed that skeletal muscle is hypermetabolic in patients with HIV lipoatrophy (Kosmiski et al., 2009). Accordingly, the specific metabolic rate of each tissue organ may differ from those in healthy adults over 80 years of age, other ethnic subjects, or patients with a chronic disease. Future studies are needed to extend these observations and to analyze gender-related, genetics, hormonal,
The results of the present chapter suggest that REE in healthy adult women can be accurately estimated from four tissue organ components by using DXA, because the mean error, total error, and systematic bias between the predicted and measured REE was relatively small compared with other estimation methods for REE. Therefore, we expect that the DXA will be one of the useful methods not only for examining body composition and bone mineral density but also for estimating REE in a broad range of fields. In the future, DXA method may offer a bridge to effective prevention, treatment and care with the dietary management and exercise regimen for life style-related disease, such as obesity,
ethnic, and other determinant factors of REE.
osteoporosis, and type 2 diabetes.
**5. Conclusion**
We express our appreciation to the subjects for their cooperation in this study. We thank the members of the National Institute of Health and Nutrition and Waseda University for their help in this experiment. This study was supported by a Research Grant for Academic Frontier Projects from the Ministry of Education, Culture, Sports, Science and Technology (05F-02), a Waseda University Grant for Special Research Projects (2005A-932, 2006B-242), Health Sciences Research Grants from the Ministry of Health, Labor and Welfare, Medical Health Care Research Grants from the Consolidated Research Institute for Advanced Science and Medical Care at Waseda University, Research grant for Waseda University Global COE program: Sports Sciences for Promotion of Active Life (projectsIII-2), Research Grant from Japan Women's College of Physical Education, and Research Grants from the Japanese Olympic Committee.
#### **7. References**
The Validity of Body Composition Measurement
ISSN 1938-3207
Oxford.
3207
ISSN 1881-7742
ISSN 1476-5497
Online ISSN 1476-5640
0954-3007, Online ISSN 1476-5640
5–41, ISSN 0263-8290
ISSN 0002-9165, Online ISSN 1938-3207
0002-9165, Online ISSN 1938-3207
Using Dual Energy X-Ray Absorptiometry for Estimating Resting Energy Expenditure 57
Ministry of Heaith, Labour and Welfare of Japan. (2009). *Dietary reference intakes for Japanese, 2010,* Daiichi Shuppan, ISBN 978-4-8041-1219-0, Tokyo. (in Japanese) Miyake, R.; Tanaka, S.; Ohkawara, K.; Ishikawa-Takata, K., Hikihara Y.; Taguri, E.;
(June 2011), pp. 224-232, Print ISSN 0301-4800, Online ISSN 1881-7742 Ravussin, E. & Bogardus, C. (1989). Relationship of genetics, age, and physical fitness to
Schofield, W.N. (1985). Predicting basal metabolic rate, new standards and review of
Snyder, W.S.; Cook, M.J.; Nasset, E.S.; Karhausen, L.R.; Howells, G.P. & Tipton, I.H. (1975).
Svendsen, O.L.; Haarbo, J.; Heitman, B.L.; Gotfredsen, A. & Christiansen, C. (1991).
Svendsen, O.L.; Haarbo, J.; Hassager, C. & Christiansen, C. (1993). Accuracy of
Taguchi, M.; Ishikawa-Takata, K.; Tatsuta, W.; Katsuragi, C.; Usui, C.; Sakamoto, S. &
Tataranni, P.A. & Ravussin, E. (1995). Variability in metabolic rate: biological sites of
Usui, C.; Takahashi, E.; Gando, Y.; Sanada, K.; Oka, J.; Miyachi, M.; Tabata, I. & Higuchi, M.
van der Ploeg, G.E.; Gunn, S.M.; Withers, R.T.; Modra, A.C.; Keeves, J.P. & Chatterton, B.E.
*Journal of Clinical Nutrition*, Vol.51, No.6, (June 1990), pp. 1106–1112, Print ISSN
Kayashita J. & Tabata, I. (2011). Validity of predictive equations for basal metabolic rate in Japanese adults. *Journal of Nutritional Science and Vitaminology*, Vol.57, No.3 ,
daily energy expenditure and fuel utilization. *American Journal of Clinical Nutrition*, Vol.49, No.5 supplements, (May 1989), pp. S968–S975, Print ISSN 0002-9165, Online
previous work. *Human Nutrition. Clinical Nutrition*, Vol.39, supplements, (1985), pp.
*Report of the Task Group of Reference Man,* Pergamon Press, ISBN 00-8-017024-2,
Measurement of body fat in elderly subjects by dual energy x-ray absorptiometry, bioelectrical impedance, and anthropometry. *American Journal of Clinical Nutrition*, Vol.53, No.5, (May 1991), pp. 1117–1123, Print ISSN 0002-9165, Online ISSN 1938-
measurements of body composition by dual energy x-ray absorptiometry in vivo. *American Journal of Clinical Nutrition*, Vol.57, No.5, (May 1993), pp. 605–608, Print
Higuchi M. (2011). Resting energy expenditure can be assessed by fat-free mass in female athletes regardless of body size. *Journal of Nutritional Science and Vitaminology*, Vol.57, No.1, (January 2011), pp. 22-29, Print ISSN 0301-4800, Online
regulation. *International Journal of Obesity and Related Metabolic Disorders*, Vol.19, No.4 supplements, (October 1995), pp. S102–S106, Print ISSN 0307-0565, Online
(2009). Resting energy expenditure can be assessed by dual-energy X-ray absorptiometry in women regardless of age and fitness. *European Journal of Clinical Nutrition*, Vol.63, No.4, (Epub February 2008), pp. 529-535, Print ISSN 0954-3007,
(2001). Predicting the resting metabolic rate of young Australian males. *European Journal of Clinical Nutrition*, Vol.55, No. 3, (March 2001), pp. 145–152, Print ISSN
Grande, F. (1980). Energy expenditure of organs and tissues. In: *Assessment of Energy*
*Research*, Kinney, J.M. (ed.), pp. 88–92, Ross Laboratories, Columbus, Ohio Harris, J.A. & Benedict, F.G. (1919). *A Biometric Study of Basal Metabolism in Man,*
Hasegawa, A.; Usui, C.; Kawano, H.; Sakamoto, S. & Higuchi, M. (2011). Characteristics of
Hayes, M.; Chustek, M.; Wang, Z.; Gallagher, D.; Heshka, S.; Spungen, A.; Bauman, W. &
Hayter, J.E. & Henry, C.J.K. (1994). A re–examination of basal metabolic rate predictive
Heymsfield, S.B.; Smith, R.; Aulet, M.; Bensen, B.; Lichtman, S.; Wang, J. & Pierson, RN.
Heymsfield, S.B.; Gallagher, D.; Kotler, D.P.; Wang, Z.; Allison, D.B. & Heshka, S. (2002).
Holliday, M.A.; Potter, D.; Jarrah, A. & Bearg, S. (1967). The relation of metabolic rate to
Hunter, G.R.; Weinsier, R.L.; Gower, B.A. & Wetzstein, C. (2001). Age-related decrease in
(February 2001), pp. 333–337, Print ISSN 0002-9165, Online ISSN 1938-3207 Kim, J.; Wang, Z.; Heymsfield, S.B.; Baumgartner, R.N. & Gallagher, D. (2002). Total-body
Kosmiski, L.A.; Ringham, B.M.; Grunwald, G.K. & Bessesen, D.H. (2009). Dual-energy X-ray
Mazes, R.B.; Barden, H.S.; Bisek, J.P. & Hanson J. (1990). Dual energy x-ray absorptiometry
2002), pp. 969–977, ISSN 1930-7381, Online ISSN 1930-739x
pp. 214–218, Print ISSN 0002-9165, Online ISSN 1938-3207
195, Print ISSN 0031-3998, Online ISSN 1530-0447
383, Print ISSN 0002-9165, Online ISSN 1938-3207
4176-76-9, Washington D.C.
Online ISSN 1522-1555
Online ISSN 1938-3207
ISSN 0301-4800, Online ISSN 1881-7742
Print ISSN 0954-3007, Online ISSN 1476-5640
*Metabolism in Health and Disease: Report of the First Ross Conference on Medical*
Washington, Publication, No. 279, Carnegie Institute of Washington, ISBN 978-114-
body composition and resting energy expenditure in lean young women. *Journal of Nutritional Science and Vitaminology*, Vol.57, No.1, (January 2011), pp. 74-79, Print
Heymsfield S.B. (2002). DXA: potential for creating a metabolic map of organ-tissue resting energy expenditure components. *Obesity Research*, Vol.10, No.10, (October
equations: the importance of geographic origin of subjects in sample selection. *European Journal of Clinical Nutrition*, Vol.48, No.10, (October 1994), pp. 702-707,
(1990). Appendicular skeletal muscle mass: measurement by dualphoton absorptiometry. *American Journal of Clinical Nutrition*, Vol.52, No.2, (August 1990),
Body-size dependence of resting energy expenditure can be attributed to nonenergetic homogeneity of fat-free mass. *American Journal of Physiology Endocrinology and Metabolism*, Vol.282, No.1, (January 2002), pp. E132–E138, Print ISSN 0193-1849,
body weight and organ size. *Pediatric Research,* Vol.1, No.3, (May 1967), pp. 185–
resting energy expenditure in sedentary white women: effects of regional differences in lean and fat mass. *American Journal of Clinical Nutrition*, Vol.73, No.2,
skeletal muscle mass: estimation by a new dual-energy X-ray absorptiometry method. *American Journal of Clinical Nutrition*, Vol.76, No.2, (August 2002), pp. 378–
absorptiometry modeling to explain the increased resting energy expenditure associated with the HIV lipoatrophy syndrome. *American Journal of Clinical Nutrition*, Vol.90, No.6, (December 2009), pp. 1525–1531, Print ISSN 0002-9165,
for total-body and regional bone-mineral and soft-tissue composition. *American*
*Journal of Clinical Nutrition*, Vol.51, No.6, (June 1990), pp. 1106–1112, Print ISSN 0002-9165, Online ISSN 1938-3207
**5**
*Poland*
Magdalena Krzykała
**Dxa as a Tool for the Assessment of**
*University School of Physical Education in Poznań*
*Anthropology and Biometry Department,*
**Morphological Asymmetry in Athletes**
Symmetry and asymmetry – two opposite phenomenon, does coexist in nature and both are very essential for science. There are many definition of symmetry, depending on research area. In biology, a dominant view is the left-right bilateral symmetry describes health and high genetic quality (Gould & Gould, 1989). In physics, symmetry includes all features of a physical system that exhibit the property of symmetry—that is, under certain transformations, aspects of these systems are "unchanged according to a particular observation" (en.wikipedia.org). In mathematics, the intellectual pursuit of the Universal formulation of symmetry (Group Theory) has led to major discoveries in physics, and to Einstein's general relativity theory (Engler, 2005). In chemistry, left-right balance is a critical component in the notion of symmetry and refers to regular arrangements of molecules and the more symmetrical, the more aesthetic (Muller, 2003). Additionally symmetry and chemistry have been in interplay in spectroscopy, crystallography, reactivity and conformational analysis. Besides, symmetry considerations continua to assist chemistry in systematizing and interpreting observation and in discovering new reactions, molecules, and other material (Hargittai & Hargittai, 2005). In art, symmetry refers to left-right, topbottom balance (of forms, colors, lines etc) in the composition as a whole being essential
On the one hand symmetry means proper proportions, harmony, and balance between two elements of some totality and is connected with beauty. On the other hand there is another definition of this term- bilateral symmetry (right-left symmetry). Both of them concern body build (morphological symmetry), but can also refer to some human movements (Starosta,
According to Frey (1949) "symmetry signifies rest and binding, asymmetry motion and loosening, the one order and law, the other arbitrariness and accident, the one formal
According to Webster dictionary (Webster, 1991) symmetry is quality of possessing exactly corresponding parts on either side of an axis and in biology – regularity in form or similarity of structure, whereas asymmetry it is lack of symmetry, uneven disposition on each side of
component of art's aesthetic quality (Jacobsen et al, 2006).
rigidity and constrains, the other life, play and freedom".
**1. Introduction**
1990).
an central line or point.
### **Dxa as a Tool for the Assessment of Morphological Asymmetry in Athletes**
Magdalena Krzykała
*University School of Physical Education in Poznań Anthropology and Biometry Department, Poland*
#### **1. Introduction**
58 Dual Energy X-Ray Absorptiometry
Van Pelt, R.E.; Dinneno, F.A.; Seals, D.R. & Jones, P.P. (2001). Age-related decline in RMR in
Van Pelt, R.E.; Jones, P.P.; Davy, K.P.; Desouza, C.A.; Tanaka, H.; Davy, B.M. & Seals D.R.
Wang, Z.; Heshka, S.; Gallagher, D.; Boozer, C.N.; Kotler, D.P. & Heymsfield, S.B. (2000).
Weir, J.B. (1949). New methods for calculating metabolic rate with special reference to
pp. E633–E639, Print ISSN 0193-1849, Online ISSN 1522-1555
Online ISSN 1522-1555
Print ISSN 0022-3751, Online ISSN 1469-7793
1997), pp. 3208–3212, Print ISSN 0021-972x, Online ISSN 1945-7197
physically active men: relation to exercise volume and energy intake. *American Journal of Physiology Endocrinology and Metabolism*, Vol.281, No.3, (September 2001),
(1997). Regular exercise and the age-related decline in resting metabolic rate in women. *The Journal of Clinical Endocrinology and Metabolism*, Vol.82, No.10, (October
Resting energy expenditure fat-free mass relationship: new insights provided by body composition modeling. *American Journal of Physiology Endocrinology and Metabolism*, Vol.279, No.3, (September 2000), pp. E539–E545, Print ISSN 0193-1849,
protein metabolism. *Journal of Physiology*, Vol.109, No.1-2, (August 1949), pp. 1-9,
Symmetry and asymmetry – two opposite phenomenon, does coexist in nature and both are very essential for science. There are many definition of symmetry, depending on research area. In biology, a dominant view is the left-right bilateral symmetry describes health and high genetic quality (Gould & Gould, 1989). In physics, symmetry includes all features of a physical system that exhibit the property of symmetry—that is, under certain transformations, aspects of these systems are "unchanged according to a particular observation" (en.wikipedia.org). In mathematics, the intellectual pursuit of the Universal formulation of symmetry (Group Theory) has led to major discoveries in physics, and to Einstein's general relativity theory (Engler, 2005). In chemistry, left-right balance is a critical component in the notion of symmetry and refers to regular arrangements of molecules and the more symmetrical, the more aesthetic (Muller, 2003). Additionally symmetry and chemistry have been in interplay in spectroscopy, crystallography, reactivity and conformational analysis. Besides, symmetry considerations continua to assist chemistry in systematizing and interpreting observation and in discovering new reactions, molecules, and other material (Hargittai & Hargittai, 2005). In art, symmetry refers to left-right, topbottom balance (of forms, colors, lines etc) in the composition as a whole being essential component of art's aesthetic quality (Jacobsen et al, 2006).
On the one hand symmetry means proper proportions, harmony, and balance between two elements of some totality and is connected with beauty. On the other hand there is another definition of this term- bilateral symmetry (right-left symmetry). Both of them concern body build (morphological symmetry), but can also refer to some human movements (Starosta, 1990).
According to Frey (1949) "symmetry signifies rest and binding, asymmetry motion and loosening, the one order and law, the other arbitrariness and accident, the one formal rigidity and constrains, the other life, play and freedom".
According to Webster dictionary (Webster, 1991) symmetry is quality of possessing exactly corresponding parts on either side of an axis and in biology – regularity in form or similarity of structure, whereas asymmetry it is lack of symmetry, uneven disposition on each side of an central line or point.
Dxa as a Tool for the Assessment of Morphological Asymmetry in Athletes 61
whole body composition, but research show that regional BMD, FFM and FM distribution is equally important, in relation to training and performance (Bell et al., 2005). Tend to this
Many researches proved that morphological asymmetry - the difference between the right and the left sides of the body exist in sport (Dorado et al., 2002; Auerbach & Ruff 2006; Starosta, 1990) and it is very important to observe the scale of this phenomenon in order to its elimination if it will be necessary. Morphological asymmetry can concerns both side-toside differences between extremities, pelvis, trunk and total body with upper and lower body diversification. Analysis concerns mostly body dimension (length of limbs), level of body fat, lean and body density. In humans, some level of asymmetry in body dimensions is rather norm than the exception (Al-Eisa et al., 2004). It is stated that the lifelong preference for one extremity -e.g. the left arm or the left leg -as well as a predilection for a certain direction when turning around or rotating about one's longitudinal axis could lead to asymmetry which occurred in morphological characteristics and which can even be osseous in the case of competitive athletes. The body response on training stimulus will be vary
It is obvious that morphological side–to–side diversification depends on sport specificity. Participation in asymmetric sport disciplines is connected with asymmetric changes in soft tissues (Ducher et al., 2005; Haapasalo et al., 1998). As was noticed, soft tissues indicate greater lateral differences than skeletal measurements (Van Dusen, 1939; Chhibber & Singh, 1970; Kimura & Asaeda, 1974;). Observation from literature suggests that in most cases the upper limb is laterally dominant in size on the right side while the lower limb is larger on the left side (Singh, 1970; McGrew & Marchant, 1997). Some studies stated that asymmetries are more pronounced in the upper extremities in comparison to lower extremities when the right side tends to be larger than the left (Munter, 1936; Tomkinson et al., 2003; Malina & Buschang, 2004; Ulijaszek & Mascie-Taylor, 2005). Malinowski (2004) stated that average, the right arm and forearm are longer, and larger are their circumferences. Left hand is longer and narrowest. The right upper limb is longer about 1 cm in comparison to left one, whereas left lower limb is longer about 10-13 mm. Also left foot, left thigh and calf
As was stated by some authors (Manning et al., 2002) small percentage changes in left or right trait size may result in large percentage changes in asymmetries, so precise and the most suitable method is need to evaluated regional morphological diversification among
There are different methods which could be use to do such analysis, which differ in the time, expense and accuracy of the results. These procedures are subject to some error which can result from measurement procedures, from the equations selected to calculate body fat
The most traditional method to assess the size of particularly body segments is anthropometry. It uses circumferences, SKF thicknesses, skeletal breadth and segment lengths for total and regional body composition evaluation (Heyward & Wagner, 2004). Some standardized procedures should be taken into consideration to increase the accuracy
percent, or from both. A standard error for most procedure is about 3 to 4 %.
direction more and more scientists conduct research in this area.
according to its timing, duration and intensity (Malina, 1979).
circumferences are longer.
very specific part of population – athletes.
**2.1 Methods for regional body composition assessment**
The opposite word to symmetry is asymmetry. Three kinds of asymmetry were distinguished by Wolański (1955): 1/ morphological – differences in size and shape of organs or body parts situated on left or right side of the body; 2/ functional – connected with one of hemispheres domination (usually left); 3/ dynamic – differences between left and right extremities in strength, muscles hardness and elasticity.
Bilateral asymmetry in humans, as was stated by same authors (Zeidel & Hessamian, 2010), was fashioned by millions of years of adaptive evolution and it implies perfection.
There are three types of bilateral asymmetry by Van Valen (1962): 1/directional asymmetry (DA): when some traits develop more on one side than the other, e.g., the human brain; 2/ antisymmetry: asymmetric development is typical, but unpredictable, e.g., larger signaling claw of the male fiddler crabs or handedness in human; 3/ fluctuating asymmetry (FA): "randomly produced deviations from perfect symmetry of two sides of quantitative traits in an individual for which the population mean of R-L differences is zero and their variability is near-normally distributed".
Normal human body asymmetry appear very soon and the manifestation of morphological asymmetry intensify with aging what is connected with functional asymmetry (Malinowski, 2004). Level of asymmetry among population reflects its developmental stability, hence differences between right and left bilateral trait are very good predictor of developmental stability both on the individually and population level. Fluctuating asymmetry has been the most widely used measure of developmental stability (Palmer & Strobeck, 1992).
Potentially human body is divided into two symmetrical parts but in fact there are some deviations (aberrations): 1) internal aberrations: asymmetry of even (kidney) and odd (pancreas, heart, spleen, liver, stomach etc) of internal organs. It refers to size, shape, location, constitution or function, 2) external aberration: refer to extremities asymmetry and handedness. There is strict connection of those aberrations with hemisphere domination (Czachowska-Sieszycka, 1983). The more increasing of the left hemisphere in domination, the more differences in left and right size of the brain. The fact that human brain is asymmetrically organized is known for about 140 years. Platon stated that symmetry is ideal (beauty is symmetrical and proportional) and that there are perfect harmony between one side to the other. That is why anthropologists in XIX w thought that human brain must be symmetrical. Paul Broca debunked this theory (Broca, 1865) and stated that right hemisphere damaged cripple speech rarely but left hemisphere - in most cases. On the basis of this information he found out that speech centers are localized only on the one side of the brain. Nowadays, the research including asymmetry of human brain are more and more advanced due to a new technology. It is valid, because the information about right and left hemisphere asymmetry could also help in better understanding of human body morphological diversification.
#### **2. Morphological asymmetry in sport**
In sport there is a need to seek some acts to achievements the highest results, especially in highest level athletes. Among other factors like training methods modification or biological regeneration also certain level of morphological parameters is very important. Body compartments, among other factors, play an important role in physical performance (Petersen et al., 2006). Generally, the body composition of athletes are consider in terms of
The opposite word to symmetry is asymmetry. Three kinds of asymmetry were distinguished by Wolański (1955): 1/ morphological – differences in size and shape of organs or body parts situated on left or right side of the body; 2/ functional – connected with one of hemispheres domination (usually left); 3/ dynamic – differences between left
Bilateral asymmetry in humans, as was stated by same authors (Zeidel & Hessamian, 2010),
There are three types of bilateral asymmetry by Van Valen (1962): 1/directional asymmetry (DA): when some traits develop more on one side than the other, e.g., the human brain; 2/ antisymmetry: asymmetric development is typical, but unpredictable, e.g., larger signaling claw of the male fiddler crabs or handedness in human; 3/ fluctuating asymmetry (FA): "randomly produced deviations from perfect symmetry of two sides of quantitative traits in an individual for which the population mean of R-L differences is zero and their variability
Normal human body asymmetry appear very soon and the manifestation of morphological asymmetry intensify with aging what is connected with functional asymmetry (Malinowski, 2004). Level of asymmetry among population reflects its developmental stability, hence differences between right and left bilateral trait are very good predictor of developmental stability both on the individually and population level. Fluctuating asymmetry has been the
Potentially human body is divided into two symmetrical parts but in fact there are some deviations (aberrations): 1) internal aberrations: asymmetry of even (kidney) and odd (pancreas, heart, spleen, liver, stomach etc) of internal organs. It refers to size, shape, location, constitution or function, 2) external aberration: refer to extremities asymmetry and handedness. There is strict connection of those aberrations with hemisphere domination (Czachowska-Sieszycka, 1983). The more increasing of the left hemisphere in domination, the more differences in left and right size of the brain. The fact that human brain is asymmetrically organized is known for about 140 years. Platon stated that symmetry is ideal (beauty is symmetrical and proportional) and that there are perfect harmony between one side to the other. That is why anthropologists in XIX w thought that human brain must be symmetrical. Paul Broca debunked this theory (Broca, 1865) and stated that right hemisphere damaged cripple speech rarely but left hemisphere - in most cases. On the basis of this information he found out that speech centers are localized only on the one side of the brain. Nowadays, the research including asymmetry of human brain are more and more advanced due to a new technology. It is valid, because the information about right and left hemisphere asymmetry could also help in better understanding of human body
In sport there is a need to seek some acts to achievements the highest results, especially in highest level athletes. Among other factors like training methods modification or biological regeneration also certain level of morphological parameters is very important. Body compartments, among other factors, play an important role in physical performance (Petersen et al., 2006). Generally, the body composition of athletes are consider in terms of
was fashioned by millions of years of adaptive evolution and it implies perfection.
most widely used measure of developmental stability (Palmer & Strobeck, 1992).
and right extremities in strength, muscles hardness and elasticity.
is near-normally distributed".
morphological diversification.
**2. Morphological asymmetry in sport**
whole body composition, but research show that regional BMD, FFM and FM distribution is equally important, in relation to training and performance (Bell et al., 2005). Tend to this direction more and more scientists conduct research in this area.
Many researches proved that morphological asymmetry - the difference between the right and the left sides of the body exist in sport (Dorado et al., 2002; Auerbach & Ruff 2006; Starosta, 1990) and it is very important to observe the scale of this phenomenon in order to its elimination if it will be necessary. Morphological asymmetry can concerns both side-toside differences between extremities, pelvis, trunk and total body with upper and lower body diversification. Analysis concerns mostly body dimension (length of limbs), level of body fat, lean and body density. In humans, some level of asymmetry in body dimensions is rather norm than the exception (Al-Eisa et al., 2004). It is stated that the lifelong preference for one extremity -e.g. the left arm or the left leg -as well as a predilection for a certain direction when turning around or rotating about one's longitudinal axis could lead to asymmetry which occurred in morphological characteristics and which can even be osseous in the case of competitive athletes. The body response on training stimulus will be vary according to its timing, duration and intensity (Malina, 1979).
It is obvious that morphological side–to–side diversification depends on sport specificity. Participation in asymmetric sport disciplines is connected with asymmetric changes in soft tissues (Ducher et al., 2005; Haapasalo et al., 1998). As was noticed, soft tissues indicate greater lateral differences than skeletal measurements (Van Dusen, 1939; Chhibber & Singh, 1970; Kimura & Asaeda, 1974;). Observation from literature suggests that in most cases the upper limb is laterally dominant in size on the right side while the lower limb is larger on the left side (Singh, 1970; McGrew & Marchant, 1997). Some studies stated that asymmetries are more pronounced in the upper extremities in comparison to lower extremities when the right side tends to be larger than the left (Munter, 1936; Tomkinson et al., 2003; Malina & Buschang, 2004; Ulijaszek & Mascie-Taylor, 2005). Malinowski (2004) stated that average, the right arm and forearm are longer, and larger are their circumferences. Left hand is longer and narrowest. The right upper limb is longer about 1 cm in comparison to left one, whereas left lower limb is longer about 10-13 mm. Also left foot, left thigh and calf circumferences are longer.
As was stated by some authors (Manning et al., 2002) small percentage changes in left or right trait size may result in large percentage changes in asymmetries, so precise and the most suitable method is need to evaluated regional morphological diversification among very specific part of population – athletes.
#### **2.1 Methods for regional body composition assessment**
There are different methods which could be use to do such analysis, which differ in the time, expense and accuracy of the results. These procedures are subject to some error which can result from measurement procedures, from the equations selected to calculate body fat percent, or from both. A standard error for most procedure is about 3 to 4 %.
The most traditional method to assess the size of particularly body segments is anthropometry. It uses circumferences, SKF thicknesses, skeletal breadth and segment lengths for total and regional body composition evaluation (Heyward & Wagner, 2004). Some standardized procedures should be taken into consideration to increase the accuracy
Dxa as a Tool for the Assessment of Morphological Asymmetry in Athletes 63
chemical composition of the underlying tissue (Pietrobelli et al., 1996). Two kinds of absorptiometry were distinguished: single-photon absorptiometry and dual-photon absorptiometry which were precisely described in Lukaski comprehensive study (1987). It was stated that both technologies provide highly reproducible estimates of lean tissue mainly in adults, because the coefficient of variation for them for repeated measurements is less than 2 %. The advantage of DXA technology is reduced of radiation exposure and more readily available data about changes in bone mineral content in whole and regional skeleton. As was stated by Webber (1995) the x-ray exposure is low and corresponding to natural radiation and radioactivity received during 5 days of normal living. Additionally, this method is more relevant in longitudinal study to track changes in body composition (Malina et al., 2004) because of the good precision of particular DXA device (about 1% for BMC and 2-3% for total body fat) for assessing whole body composition (De Lorenzo et al.,
Kistrop with co-authors (Kistrop et al., 2000) found that measuring total and regional body composition by DXA can improve the prediction of energy expenditure what could be a
Each method has its own advantages and disadvantages, also DXA. Some authors observed (Calbet et al., 1998) that this method giving a general idea about the specific sport loading influence on BMC. On the other hand it is not possible to assess the effect of sport participation to particular bones during DXA analysis because it includes all bone structures. There are also some concerns about measurements bias related to the impact of a significant change in the lean tissue hydration. Some study however show not significantly alters the estimates for the bone, lean and fat mass (Pietrobelli et al., 1998). Another problem is that thicknesses are different depends on body regions and individual body shapes, so there could be large variation in percent fat in DXA study (Steward & Hannan, 2000b; Lohman & Chen, 2005). When comparing results of regional body composition from DXA scanners of different manufacturers, caution must be done because of some factors which are different among both of them like: pixel size, X-ray voltages or algorithms for shape and edge detection (Tothill et al., 1994; Steward & Hannan, 2000a). As was stated by some authors (Lohman & Chen, 2005) differences between DXA instruments from leading manufactures like Hologic, Lunar and Norland were 7% for body fat and about 15% for bone mineral content, some times ago. Nowadays, DXA software are more and more accurate with new calibration modes increasing precision and accuracy (O'Connor, 2006). iDXA (Lunar) for example have a little bit precision with total-body assessment and almost identical values in regional BMD measurements in comparison to Lunar Prodigy (Faulkner, 2006). Despite all limitations DXA is a widely used method, owing to its ease of use, availability, low-radiation exposure, good accuracy and reproducibility for the assessment of regional body composition. It is also seems to be less dependent on biological consistency than other methods (Haarbo et al., 1991; Kohrt, 1995) and is most useful for research purposes (Speiser et al., 2005). It was documented that regional adiposity by DXA is potentially more accurate than anthropometry measures, and more practical than computed tomography or magnetic resonance imaging scans (Henche et al., 2008). Taking into account all of those data, DXA seems to be one of the most suitable method for evaluation side-to-
valid information for athletes who try to lose fat mass or gain lean muscle.
side morphological diversification among athletes.
1998).
and reliability of measurements like taking three measurements within ± 0.2 for body segments with relatively small girths (calf, arm, forearm) and three measurements within ± 1.0 cm for longer body segments like waist, abdomen and buttock; or using a small sliding caliper with greater precision to measure breadth of smaller segments (elbow, wrist) and so on (Wilmore et al., 1988). Schell and co-authors (Schell et al., 1985) emphasized that asymmetry of paired dimensions is a big methodological problem in anthropometry. Additionally, analysis of asymmetry should be done with advanced statistical tests. Moreover, interpretation of given results must be done with cautions especially, if observed differences are so small that could be lower than technical errors of measurement (Moreno et al., 2002).
There are also some newer methods to assess regional diversification of some tissues. Traditional BIA method, allows to assess only whole body composition (wrist-to-ankle), however segmental BIA (SBIA) system based on eight symmetrically re-positioned electrodes can give information about resistance and reactance values from right and left arms and legs and torso (www.rjlsystem.com/pdf-files/segmental\_bia.pdf). Because of human body segments are not uniform in length or cross-sectional area, there will be different resistance to the flow of current through them (Heyward & Wagner., 2004). Fuller and Elia (1989) stated that those body segments which have small cross-sectional areas have simultaneously the greatest effect on impedance. The resistance ratio of upper limb-trunklower limb theoretically equal 13.8 : 1 : 11.8. SBIA measuring particularly body segments like upper and lower limb or trunk may be a good method for person with altered fluid distribution, so also for athletes (Organ et al., 1994). Segmental muscle changes for example during physical activity are visible by studying the individual arms and legs as a comparative percentage or a percent change over time. Segmental measurements can be successful made by traditional electrodes Akern Sre (Florence, Italy) or by segmental body composition analyzers (Omron, Tanita, BioSpace) which incorporate foot and hand contact points (where standing person holding two rods with fingers and thumb of each hand) what has disadvantage: high resistance of the body ankle and wrist are included in the measurements. Additionally the ankle, wrist, lower leg and forearm contribute more than 50 percent of the measurement (Scheltinge et al., 1991).
MRI (Magnetic Resonans Imaging) – another method to assess regional diversification, can be used to assess whole mass and muscle content of any body area, giving high-quality images of different tissues (Malina et al., 2004). Also computerized tomography CT is often used in regional determination of body segments with body fat, muscle and bone area assessment (Buckley et al., 1987; Forbes et al., 1988; Jordao et al., 2004). Magnetic resonance imaging and computed tomography are regard as the reference methods for regional body composition assessment but routine use of them is impeded by access, high cost, and in case of CT, significant exposure to ionizing radiation. Some authors propose then DXA (Dual Energy X-ray absorptiometry) as a more valid and precise method for measuring total and regional body composition (Chen et al., 2007; Andreoli et al., 2009). Generally, DXA, CT and MRI methods are consider to be a standard for precision and accuracy for body composition measurements (Chettle & Fremlin, 1984; Ellis, 2000). DXA is primary method to estimate body composition of the total body and specific regions (bone mineral, fat-free soft tissue and fat). The main assumption of absorptiometry method relies on measuring attenuation of X-rays with high-and-low-photon energies what is dependent on the thickness, density and
and reliability of measurements like taking three measurements within ± 0.2 for body segments with relatively small girths (calf, arm, forearm) and three measurements within ± 1.0 cm for longer body segments like waist, abdomen and buttock; or using a small sliding caliper with greater precision to measure breadth of smaller segments (elbow, wrist) and so on (Wilmore et al., 1988). Schell and co-authors (Schell et al., 1985) emphasized that asymmetry of paired dimensions is a big methodological problem in anthropometry. Additionally, analysis of asymmetry should be done with advanced statistical tests. Moreover, interpretation of given results must be done with cautions especially, if observed differences are so small that could be lower than technical errors of measurement (Moreno
There are also some newer methods to assess regional diversification of some tissues. Traditional BIA method, allows to assess only whole body composition (wrist-to-ankle), however segmental BIA (SBIA) system based on eight symmetrically re-positioned electrodes can give information about resistance and reactance values from right and left arms and legs and torso (www.rjlsystem.com/pdf-files/segmental\_bia.pdf). Because of human body segments are not uniform in length or cross-sectional area, there will be different resistance to the flow of current through them (Heyward & Wagner., 2004). Fuller and Elia (1989) stated that those body segments which have small cross-sectional areas have simultaneously the greatest effect on impedance. The resistance ratio of upper limb-trunklower limb theoretically equal 13.8 : 1 : 11.8. SBIA measuring particularly body segments like upper and lower limb or trunk may be a good method for person with altered fluid distribution, so also for athletes (Organ et al., 1994). Segmental muscle changes for example during physical activity are visible by studying the individual arms and legs as a comparative percentage or a percent change over time. Segmental measurements can be successful made by traditional electrodes Akern Sre (Florence, Italy) or by segmental body composition analyzers (Omron, Tanita, BioSpace) which incorporate foot and hand contact points (where standing person holding two rods with fingers and thumb of each hand) what has disadvantage: high resistance of the body ankle and wrist are included in the measurements. Additionally the ankle, wrist, lower leg and forearm contribute more than 50
MRI (Magnetic Resonans Imaging) – another method to assess regional diversification, can be used to assess whole mass and muscle content of any body area, giving high-quality images of different tissues (Malina et al., 2004). Also computerized tomography CT is often used in regional determination of body segments with body fat, muscle and bone area assessment (Buckley et al., 1987; Forbes et al., 1988; Jordao et al., 2004). Magnetic resonance imaging and computed tomography are regard as the reference methods for regional body composition assessment but routine use of them is impeded by access, high cost, and in case of CT, significant exposure to ionizing radiation. Some authors propose then DXA (Dual Energy X-ray absorptiometry) as a more valid and precise method for measuring total and regional body composition (Chen et al., 2007; Andreoli et al., 2009). Generally, DXA, CT and MRI methods are consider to be a standard for precision and accuracy for body composition measurements (Chettle & Fremlin, 1984; Ellis, 2000). DXA is primary method to estimate body composition of the total body and specific regions (bone mineral, fat-free soft tissue and fat). The main assumption of absorptiometry method relies on measuring attenuation of X-rays with high-and-low-photon energies what is dependent on the thickness, density and
percent of the measurement (Scheltinge et al., 1991).
et al., 2002).
chemical composition of the underlying tissue (Pietrobelli et al., 1996). Two kinds of absorptiometry were distinguished: single-photon absorptiometry and dual-photon absorptiometry which were precisely described in Lukaski comprehensive study (1987). It was stated that both technologies provide highly reproducible estimates of lean tissue mainly in adults, because the coefficient of variation for them for repeated measurements is less than 2 %. The advantage of DXA technology is reduced of radiation exposure and more readily available data about changes in bone mineral content in whole and regional skeleton. As was stated by Webber (1995) the x-ray exposure is low and corresponding to natural radiation and radioactivity received during 5 days of normal living. Additionally, this method is more relevant in longitudinal study to track changes in body composition (Malina et al., 2004) because of the good precision of particular DXA device (about 1% for BMC and 2-3% for total body fat) for assessing whole body composition (De Lorenzo et al., 1998).
Kistrop with co-authors (Kistrop et al., 2000) found that measuring total and regional body composition by DXA can improve the prediction of energy expenditure what could be a valid information for athletes who try to lose fat mass or gain lean muscle.
Each method has its own advantages and disadvantages, also DXA. Some authors observed (Calbet et al., 1998) that this method giving a general idea about the specific sport loading influence on BMC. On the other hand it is not possible to assess the effect of sport participation to particular bones during DXA analysis because it includes all bone structures. There are also some concerns about measurements bias related to the impact of a significant change in the lean tissue hydration. Some study however show not significantly alters the estimates for the bone, lean and fat mass (Pietrobelli et al., 1998). Another problem is that thicknesses are different depends on body regions and individual body shapes, so there could be large variation in percent fat in DXA study (Steward & Hannan, 2000b; Lohman & Chen, 2005). When comparing results of regional body composition from DXA scanners of different manufacturers, caution must be done because of some factors which are different among both of them like: pixel size, X-ray voltages or algorithms for shape and edge detection (Tothill et al., 1994; Steward & Hannan, 2000a). As was stated by some authors (Lohman & Chen, 2005) differences between DXA instruments from leading manufactures like Hologic, Lunar and Norland were 7% for body fat and about 15% for bone mineral content, some times ago. Nowadays, DXA software are more and more accurate with new calibration modes increasing precision and accuracy (O'Connor, 2006). iDXA (Lunar) for example have a little bit precision with total-body assessment and almost identical values in regional BMD measurements in comparison to Lunar Prodigy (Faulkner, 2006). Despite all limitations DXA is a widely used method, owing to its ease of use, availability, low-radiation exposure, good accuracy and reproducibility for the assessment of regional body composition. It is also seems to be less dependent on biological consistency than other methods (Haarbo et al., 1991; Kohrt, 1995) and is most useful for research purposes (Speiser et al., 2005). It was documented that regional adiposity by DXA is potentially more accurate than anthropometry measures, and more practical than computed tomography or magnetic resonance imaging scans (Henche et al., 2008). Taking into account all of those data, DXA seems to be one of the most suitable method for evaluation side-toside morphological diversification among athletes.
Dxa as a Tool for the Assessment of Morphological Asymmetry in Athletes 65
asymmetry in the upper body of fast bowlers in cricket was observed in Grobbelaar study (Grobbelaar et al., 2000). It concerns the relaxed arm (3.8 %), tensed arm (4.7%), forearm
The morphological characteristics of fencers show a typical asymmetry of the limbs as a result of the asymmetrical sport activity practicing what is advantage in gaining success. Fencing produces typical functional asymmetries that emphasize the very high level of specific function, strength and control required in this sport discipline (Roi et al., 2008). Muscle mass of dominant lower limb of high-class fencer is bigger than contrlateral limb (Nystrom et al., 1990). Tsolakis et al (2006) stated, on the example of fencers, that there was difference in morphological asymmetry of arm and leg depended on age, where arm asymmetries were specific for the age of 10-13 years, whereas leg asymmetries were observed among 14-17 year-old athletes. Because of asymmetrical nature of this sport, some specific injuries are characteristics in the shoulders, the back and the pelvic girdle, so there is
necessity of including prevention in daily fencers training (Kucera & Henn, 2003).
(professional national league and semi-professional state league).
years. Similar findings were done by Prives (1969).
Manning and Pickup (Manning, & Pickup, 1998) stated that the national league athletes would be more symmetric, because symmetry is positively related with physical performance in adult males. Besides, symmetric males run faster than asymmetric males (Manning & Pickup, 1998). According to Chinn study "Morphological asymmetry is more pointed out in high level athlete in case of asymmetrical sport disciplines" (Chinn et al., 1974). Tomkinson et al stated (2003) that there are no differences in variance in fluctuating asymmetry between adult basketball and soccer, competing at two different standards
There is also some assumption that symmetry positive correlated with body size what mean that the larger men is (higher and heavier), the more symmetric he is. But in women body size correlated negatively with symmetry (Manning, 1995). Additionally this author observed positive correlation between body mass index (BMI) and asymmetry among women (the bigger BMI= the bigger asymmetry). On the other hand some research indicated that the bigger person (both men and women)– the greater asymmetry (Graham et al., 1998). Studies of scientists from many disciplines suggest that age, sex and environmental stress like extreme unilateral work are the most influential factors for morphological asymmetry (Wolański, 1962; Singh, 1970; Mascie-Taylor et al., 1981; Malina, 1983). Some authors also found (Hetland et al., 1998) that there are different factors significantly contributed to the regional body composition. On the basic of 108 (86 recreational and 22 elite runners) male long-distance runners they stated that training was the strongest determinant in the legs and the arms, whereas androgenic activity was important in the abdominal region. Some data suggest that persistent unilateral training may also influence specific bone lengths and width. As was stated by Buskirk et al (1956) on the example of seven tennis players (nationally level), the athletes had greater length and width of bones of the dominant hands and forearms than the non-dominant extremity. Authors suggested that those laterality differences were results of vigorous exercise on the bone growth during the adolescent
In Jones et al study (1977) there was a significant hypertrophy of cortical bone of the humerus in the dominant arm compared to the non-playing arm of 84 male professional tennis players. There were sex differences in cortical thickness of the dominant arm
(2.8%) and ½ chest circumferences (6.4%).
#### **2.2 Morphological asymmetry of athletes - research overview**
Many studies with athletes body asymmetry exploited anthropometric method. Malina and Buschang (2004) observed greater hypertrophy in the musculature of the dominant side in athletes. Also another authors (Calbet et al., 2001; Dorado et al., 2002) found, on the example of golfers, that they had muscle hypertrophy in the dominant compared with non-dominant arm under training influence. The right-left differences in morphological parameters (mainly in forearm girth, arm girth, elbow width) were observed also in 134 athletes aged 21-32 years, engaged in many different asymmetric movement sports like tennis, canoeing, kayaking and boxing (Krawczyk et al., 1998). According to those research, most significant asymmetry was found among tennis players, then in kayakers, canoeists, rowers and skaters and concerned mainly forearm girth. The extreme directional asymmetry in the use of the limbs among tennis players is connected with physiological and anatomical changes in those body segments (Lucki, 2006). On the basis of 25 male collegiate tennis players age 19- 24 from four NCAA Division I she stated that there was significant asymmetry between the limbs, related to cross-sectional measurements (circumferences, widths). Some authors (Maughan et al., 1986) observed greater proportion of muscle and smaller proportion of fat in dominant arm than the opposite limb in tennis players. Numerous studies indicated increased bone density in the dominant limb among tennis players (Ruff et al., 1994; Ducher et al., 2005; Lucki & Nicolay, 2007). Analysis of the bilateral asymmetry among collegiate (19-24 years old) tennis players showed that forearm circumferences of the dominant limb was greater than in the opposite limb (3-10% in female and 2-13% in male tennis players) (Lucki & Nikolay, 2007). Jone's (Jones et al., 1977) study examining site-specific accretion of bone of professional tennis players, with differences being up to 30%. Tarociński (1977) stated that difference in circumferences of the left and the right upper extremities was 1-2 cm in male athlete (age 12-15) playing tennis for four years, whereas in Marchwicki study (1927) those differences was smaller and equal 0,3 – 1,1 cm. also among young tennis players. In lower extremities, in turn, there was smaller diversification with left leg advantage (Tarociński, 1977). Similar finding was done later on the example of baseball players (Komi 1996; Bubanj & Obradovic, 2002).
Abraho and Mello (2008) made comparison of the young athletes playing tennis for at most two years (age 6-10 years) with the male instructors between 22 y and 37 years of age engaged in this sport discipline at least eight years. They noticed the increase of the incidence of right somatic measures superior to the left, because the excessive time of training of asymmetric sport. This situation has influence on postural deviation according to those research.
When one considers that carrying loads with the preferred hand means a stress on the arm muscles of the same side and a simultaneous activation of the contra-lateral muscles for the stabilization of one's balance, these functional asymmetries become plausible. It is also remarkable that hurdlers, high jumpers and pole vaulters exhibit higher muscle contractility in their swing leg than in their take-off leg (Absauomov, 1976) because of the higher mechanical load. Kruger et al (2005) determined the degree of upper body morphological asymmetry in 19 elite international male javelin throwers, age of 26,4 ± 4.4 years. They found larger variables on the dominant side for thirteen of the fourteen variables (especially for triceps skinfold 5.9%, half-chest girth 4.9%,, forearm girth 3.9%, biceps skinfold 2.5%) what could have health consequences and performance limiting effects. The morphological
Many studies with athletes body asymmetry exploited anthropometric method. Malina and Buschang (2004) observed greater hypertrophy in the musculature of the dominant side in athletes. Also another authors (Calbet et al., 2001; Dorado et al., 2002) found, on the example of golfers, that they had muscle hypertrophy in the dominant compared with non-dominant arm under training influence. The right-left differences in morphological parameters (mainly in forearm girth, arm girth, elbow width) were observed also in 134 athletes aged 21-32 years, engaged in many different asymmetric movement sports like tennis, canoeing, kayaking and boxing (Krawczyk et al., 1998). According to those research, most significant asymmetry was found among tennis players, then in kayakers, canoeists, rowers and skaters and concerned mainly forearm girth. The extreme directional asymmetry in the use of the limbs among tennis players is connected with physiological and anatomical changes in those body segments (Lucki, 2006). On the basis of 25 male collegiate tennis players age 19- 24 from four NCAA Division I she stated that there was significant asymmetry between the limbs, related to cross-sectional measurements (circumferences, widths). Some authors (Maughan et al., 1986) observed greater proportion of muscle and smaller proportion of fat in dominant arm than the opposite limb in tennis players. Numerous studies indicated increased bone density in the dominant limb among tennis players (Ruff et al., 1994; Ducher et al., 2005; Lucki & Nicolay, 2007). Analysis of the bilateral asymmetry among collegiate (19-24 years old) tennis players showed that forearm circumferences of the dominant limb was greater than in the opposite limb (3-10% in female and 2-13% in male tennis players) (Lucki & Nikolay, 2007). Jone's (Jones et al., 1977) study examining site-specific accretion of bone of professional tennis players, with differences being up to 30%. Tarociński (1977) stated that difference in circumferences of the left and the right upper extremities was 1-2 cm in male athlete (age 12-15) playing tennis for four years, whereas in Marchwicki study (1927) those differences was smaller and equal 0,3 – 1,1 cm. also among young tennis players. In lower extremities, in turn, there was smaller diversification with left leg advantage (Tarociński, 1977). Similar finding was done later on the example of baseball
Abraho and Mello (2008) made comparison of the young athletes playing tennis for at most two years (age 6-10 years) with the male instructors between 22 y and 37 years of age engaged in this sport discipline at least eight years. They noticed the increase of the incidence of right somatic measures superior to the left, because the excessive time of training of asymmetric sport. This situation has influence on postural deviation according to
When one considers that carrying loads with the preferred hand means a stress on the arm muscles of the same side and a simultaneous activation of the contra-lateral muscles for the stabilization of one's balance, these functional asymmetries become plausible. It is also remarkable that hurdlers, high jumpers and pole vaulters exhibit higher muscle contractility in their swing leg than in their take-off leg (Absauomov, 1976) because of the higher mechanical load. Kruger et al (2005) determined the degree of upper body morphological asymmetry in 19 elite international male javelin throwers, age of 26,4 ± 4.4 years. They found larger variables on the dominant side for thirteen of the fourteen variables (especially for triceps skinfold 5.9%, half-chest girth 4.9%,, forearm girth 3.9%, biceps skinfold 2.5%) what could have health consequences and performance limiting effects. The morphological
**2.2 Morphological asymmetry of athletes - research overview**
players (Komi 1996; Bubanj & Obradovic, 2002).
those research.
asymmetry in the upper body of fast bowlers in cricket was observed in Grobbelaar study (Grobbelaar et al., 2000). It concerns the relaxed arm (3.8 %), tensed arm (4.7%), forearm (2.8%) and ½ chest circumferences (6.4%).
The morphological characteristics of fencers show a typical asymmetry of the limbs as a result of the asymmetrical sport activity practicing what is advantage in gaining success. Fencing produces typical functional asymmetries that emphasize the very high level of specific function, strength and control required in this sport discipline (Roi et al., 2008). Muscle mass of dominant lower limb of high-class fencer is bigger than contrlateral limb (Nystrom et al., 1990). Tsolakis et al (2006) stated, on the example of fencers, that there was difference in morphological asymmetry of arm and leg depended on age, where arm asymmetries were specific for the age of 10-13 years, whereas leg asymmetries were observed among 14-17 year-old athletes. Because of asymmetrical nature of this sport, some specific injuries are characteristics in the shoulders, the back and the pelvic girdle, so there is necessity of including prevention in daily fencers training (Kucera & Henn, 2003).
Manning and Pickup (Manning, & Pickup, 1998) stated that the national league athletes would be more symmetric, because symmetry is positively related with physical performance in adult males. Besides, symmetric males run faster than asymmetric males (Manning & Pickup, 1998). According to Chinn study "Morphological asymmetry is more pointed out in high level athlete in case of asymmetrical sport disciplines" (Chinn et al., 1974). Tomkinson et al stated (2003) that there are no differences in variance in fluctuating asymmetry between adult basketball and soccer, competing at two different standards (professional national league and semi-professional state league).
There is also some assumption that symmetry positive correlated with body size what mean that the larger men is (higher and heavier), the more symmetric he is. But in women body size correlated negatively with symmetry (Manning, 1995). Additionally this author observed positive correlation between body mass index (BMI) and asymmetry among women (the bigger BMI= the bigger asymmetry). On the other hand some research indicated that the bigger person (both men and women)– the greater asymmetry (Graham et al., 1998).
Studies of scientists from many disciplines suggest that age, sex and environmental stress like extreme unilateral work are the most influential factors for morphological asymmetry (Wolański, 1962; Singh, 1970; Mascie-Taylor et al., 1981; Malina, 1983). Some authors also found (Hetland et al., 1998) that there are different factors significantly contributed to the regional body composition. On the basic of 108 (86 recreational and 22 elite runners) male long-distance runners they stated that training was the strongest determinant in the legs and the arms, whereas androgenic activity was important in the abdominal region. Some data suggest that persistent unilateral training may also influence specific bone lengths and width. As was stated by Buskirk et al (1956) on the example of seven tennis players (nationally level), the athletes had greater length and width of bones of the dominant hands and forearms than the non-dominant extremity. Authors suggested that those laterality differences were results of vigorous exercise on the bone growth during the adolescent years. Similar findings were done by Prives (1969).
In Jones et al study (1977) there was a significant hypertrophy of cortical bone of the humerus in the dominant arm compared to the non-playing arm of 84 male professional tennis players. There were sex differences in cortical thickness of the dominant arm
Dxa as a Tool for the Assessment of Morphological Asymmetry in Athletes 67
respectively: p= 0.0086) and of the trunk (1.111 ± 0.0609 g/cm2 vs. 1.137 ± 0.0729 g/cm2, respectively: p=0.0008). In turn there were no differences between right arm 1.017 g/cm2 and for left arm 1.014 g/cm2 as well as between total bone mineral density (the right and the left sides of the body) (1.326 ± 0.0741 g/cm2 vs. 1.338 ± 0.0754 g/cm2, respectively: p=0.1310). The larger amount of body density was in the left legs, left trunk and total body of athletes. The individual differences between the left and right sides of a particularly body segment show that the biggest diversity among them was in the case of legs. Significant differences in side-to-side lean morphology were observed in every measured parameter. The lean of the arms, legs, trunk and total was higher in the left side of the body. The lean of right arm was 3738 ± 454.9 g vs. 4046 ± 420.7 g for left arm respectively (p=0.0000), lean of right leg was 10578 ± 1050.9 g vs. 10904 ± 1054.5 g, respectively (p=0.0000). Also lean amount was larger in the left side of trunk in comparison to the right (difference: 14371 ± 1580.6 g to 13996 ± 1436.7 g). Significance difference was observed in total lean (30444 ± 2759.5 g for right vs. 31208 ± 2814.5 g for left total, respectively (p=0.0073). There was clear similarity in arms and legs among field hockey players, as opposed to the trunk and total LEAN. Similar findings were observed in a case of body fat analyzing. Across all body segments, the amount of
morphological asymmetry was significantly greater in the left side of the body.
athletes combined nor within individual sport disciplines.
In a study of thirty competitive young adult Rugby Union players, Bell et al (2005) observed the arm-lower body contrast in lean soft tissue mass suggesting, that upper and lower limbs contribute equally to playing performance. Also bone mineral mass (BMM) was different according to playing position. The legs were dominant in forwards (-0.76) and the trunk in backs (0.67). The measurements were done using DXA (Hologic QDR/1000W. software version V5.73). Using DXA (QDR-1500, Hologic Corporation, Waltham, MA) in ten healthy postmenopausal tennis players Sanchis-Moysi et al (2004) observed linear relationship between the extent of years of training and the magnitude of the inter-arm difference in bone size- participation in tennis after menarche is connecting with greater bone size and mass in the loaded arm. Some data suggested also that muscle mass is independent predictor of regional and total BMD on the example of gymnasts and elite swimmers with DXA method (Taaffe et al., 1995). Also Calbet et al (2001) study showed a high correlation between the BMD and the muscle mass in the left leg among 33 right-leg dominant male soccer players. Nevil et al (2003) considered the effects of specific training on BMD asymmetry between arms and legs among Caucasians athletes from ten sport disciplines (runners, cyclists, triathletes, racket players, rock climbers, swimmers, rugby players, rowers, kayakers and bodybuilders), at a university or higher level, trained for a minimum of four hour per week, for a minimum of three years. Using DXA on a Hologic QDR 1000W (Bedford, MA) scanner they stated that the differences between the dominant and non-dominant arms were significant but different depends on sport. For example it was greater in the racket players and lesser in the rowers. In the case of right and left legs, there was no asymmetric difference in BMD between them for all
The McClanahan et al (2002) study investigated the effects of participation in different sports (NCAA Division I-A baseball, basketball, football, golf, soccer, tennis, cross-country, indoor/outdoor track and volleyball) on side-o-side diversification in bone mineral density (BMD) of the upper and lower limbs. On the basic of 184 collegiate athletes both sexes, and using dual energy X-ray absorptiometry, they revealed greater BMD of the right arms compared with the left arms for all teams. The largest differences were found in the case of
compared to non-dominant arm among athletes (man had 35% greater cortical thickness of the dominant arm, whereas women - 28%). Hypertrophy of the humeral cortex has been reported also in the throwing arms of professional baseball pitchers (King et al., 1969).
Asymmetry could have different value and it is stated that when it access some level, it may hinder some special activity practice, could negatively affect the health (being connected with functional asymmetry and some changes for example in the area of backbone). According to Wilk et al study (2002) therapist should focus on restoring the functional shoulder asymmetry during rehabilitation of shoulder injuries in the case of throwing athlete. Asymmetries between lower limbs during athletic movements are thought to increase the risk of injury and compromise performance (Cronin, 2010). In the case of athletes who are engaged with sport discipline including one-sided hips rotation (tennis, golf, squash) and suffer from LBP, more lower limbs asymmetry and pelvic asymmetry (LLD-leg length discrepancy) were observed (Egan et al., 1995; Van Dillen et al., 2008). Some authors asked the question if the asymmetry lead to the back pain or did the back pain lead to the asymmetry (Bussey, 2010). This is not yet determined evolutionary dilemma. She stated that asymmetries may be functional adaptations, meaning that the body has successfully adapted to the asymmetrical loading demands of the sport in order to decrease the excessive strain in some tissues. Similar finding did previously Koszczyc (1991) who observed that shoulder and whole upper limb asymmetry (which can appear in progressive stadium of ontogeny) could be effect of body adaptation to increased mechanical loads. Bussey (2010) conducted the study based on 60 women divided into three groups consisted of elite athletes at the national and international level (bilateral group: triathlon, crosscountry running, single scull rowing ; unilateral group: field hockey, ice hockey, speedskating) and control group. She speculated that some level of pelvis asymmetry may be a natural effect of lateral dominance which decrease with bilateral activities like running or cycling or increase with unilateral nature of sport discipline like hockey.
Field hockey is one-side dominant sport. It means that athletes have preference of one side of the body over the other. There are many different specific movements in this discipline. Many of them involve a rapid rotation of the hips, shoulders and arms for example push-in movement (Kerr & Ness, 2006). This movement stated with the body counter-rotated right side of the body is behind the left side (with respect to the direction of ball trajectory). Similar findings was described by Mc Laughlin (1997). Kerr found that pelvic and shoulder girdle maximum angular velocities occurred concurrently from left arm to the right arm. Many published studies focused on many components of field hockey like the hit (Burges-Limerick et al., 1991), the push (Alexander, 1983), the slap shot (Cresswell & Elliott, 1987), the push-in and trap phases of the penalty corner (Kerr & Ness, 2006), describing specificity of characteristic movements, which could appeared in athletes morphology.
Krzykała (2010) studied the effects of specific one-side dominant training on morphological asymmetry, on the basic of twenty competitive male field hockey players 18-34 years of age. All athletes had played representative hockey at a senior (national) polish level of the game. Average training years of all competitors was 17 y and hesitated between 13 y and 24 y. The anthropometric characteristics in both sides of the body was done by dual energy x-ray absorptiometry, using LUNAR PRODIGY ADVANCE (GE, Madison, WI, USA) densitometer with enCORE software (GE Helthcare v.10.50.086). There were significant differences in body mass density of the legs (1.576 ± 0.0909 g/cm2 vs. 1.611 ± 0.1062 g/cm2,
compared to non-dominant arm among athletes (man had 35% greater cortical thickness of the dominant arm, whereas women - 28%). Hypertrophy of the humeral cortex has been reported also in the throwing arms of professional baseball pitchers (King et al., 1969).
Asymmetry could have different value and it is stated that when it access some level, it may hinder some special activity practice, could negatively affect the health (being connected with functional asymmetry and some changes for example in the area of backbone). According to Wilk et al study (2002) therapist should focus on restoring the functional shoulder asymmetry during rehabilitation of shoulder injuries in the case of throwing athlete. Asymmetries between lower limbs during athletic movements are thought to increase the risk of injury and compromise performance (Cronin, 2010). In the case of athletes who are engaged with sport discipline including one-sided hips rotation (tennis, golf, squash) and suffer from LBP, more lower limbs asymmetry and pelvic asymmetry (LLD-leg length discrepancy) were observed (Egan et al., 1995; Van Dillen et al., 2008). Some authors asked the question if the asymmetry lead to the back pain or did the back pain lead to the asymmetry (Bussey, 2010). This is not yet determined evolutionary dilemma. She stated that asymmetries may be functional adaptations, meaning that the body has successfully adapted to the asymmetrical loading demands of the sport in order to decrease the excessive strain in some tissues. Similar finding did previously Koszczyc (1991) who observed that shoulder and whole upper limb asymmetry (which can appear in progressive stadium of ontogeny) could be effect of body adaptation to increased mechanical loads. Bussey (2010) conducted the study based on 60 women divided into three groups consisted of elite athletes at the national and international level (bilateral group: triathlon, crosscountry running, single scull rowing ; unilateral group: field hockey, ice hockey, speedskating) and control group. She speculated that some level of pelvis asymmetry may be a natural effect of lateral dominance which decrease with bilateral activities like running or
cycling or increase with unilateral nature of sport discipline like hockey.
of characteristic movements, which could appeared in athletes morphology.
Field hockey is one-side dominant sport. It means that athletes have preference of one side of the body over the other. There are many different specific movements in this discipline. Many of them involve a rapid rotation of the hips, shoulders and arms for example push-in movement (Kerr & Ness, 2006). This movement stated with the body counter-rotated right side of the body is behind the left side (with respect to the direction of ball trajectory). Similar findings was described by Mc Laughlin (1997). Kerr found that pelvic and shoulder girdle maximum angular velocities occurred concurrently from left arm to the right arm. Many published studies focused on many components of field hockey like the hit (Burges-Limerick et al., 1991), the push (Alexander, 1983), the slap shot (Cresswell & Elliott, 1987), the push-in and trap phases of the penalty corner (Kerr & Ness, 2006), describing specificity
Krzykała (2010) studied the effects of specific one-side dominant training on morphological asymmetry, on the basic of twenty competitive male field hockey players 18-34 years of age. All athletes had played representative hockey at a senior (national) polish level of the game. Average training years of all competitors was 17 y and hesitated between 13 y and 24 y. The anthropometric characteristics in both sides of the body was done by dual energy x-ray absorptiometry, using LUNAR PRODIGY ADVANCE (GE, Madison, WI, USA) densitometer with enCORE software (GE Helthcare v.10.50.086). There were significant differences in body mass density of the legs (1.576 ± 0.0909 g/cm2 vs. 1.611 ± 0.1062 g/cm2, respectively: p= 0.0086) and of the trunk (1.111 ± 0.0609 g/cm2 vs. 1.137 ± 0.0729 g/cm2, respectively: p=0.0008). In turn there were no differences between right arm 1.017 g/cm2 and for left arm 1.014 g/cm2 as well as between total bone mineral density (the right and the left sides of the body) (1.326 ± 0.0741 g/cm2 vs. 1.338 ± 0.0754 g/cm2, respectively: p=0.1310). The larger amount of body density was in the left legs, left trunk and total body of athletes. The individual differences between the left and right sides of a particularly body segment show that the biggest diversity among them was in the case of legs. Significant differences in side-to-side lean morphology were observed in every measured parameter. The lean of the arms, legs, trunk and total was higher in the left side of the body. The lean of right arm was 3738 ± 454.9 g vs. 4046 ± 420.7 g for left arm respectively (p=0.0000), lean of right leg was 10578 ± 1050.9 g vs. 10904 ± 1054.5 g, respectively (p=0.0000). Also lean amount was larger in the left side of trunk in comparison to the right (difference: 14371 ± 1580.6 g to 13996 ± 1436.7 g). Significance difference was observed in total lean (30444 ± 2759.5 g for right vs. 31208 ± 2814.5 g for left total, respectively (p=0.0073). There was clear similarity in arms and legs among field hockey players, as opposed to the trunk and total LEAN. Similar findings were observed in a case of body fat analyzing. Across all body segments, the amount of morphological asymmetry was significantly greater in the left side of the body.
In a study of thirty competitive young adult Rugby Union players, Bell et al (2005) observed the arm-lower body contrast in lean soft tissue mass suggesting, that upper and lower limbs contribute equally to playing performance. Also bone mineral mass (BMM) was different according to playing position. The legs were dominant in forwards (-0.76) and the trunk in backs (0.67). The measurements were done using DXA (Hologic QDR/1000W. software version V5.73). Using DXA (QDR-1500, Hologic Corporation, Waltham, MA) in ten healthy postmenopausal tennis players Sanchis-Moysi et al (2004) observed linear relationship between the extent of years of training and the magnitude of the inter-arm difference in bone size- participation in tennis after menarche is connecting with greater bone size and mass in the loaded arm. Some data suggested also that muscle mass is independent predictor of regional and total BMD on the example of gymnasts and elite swimmers with DXA method (Taaffe et al., 1995). Also Calbet et al (2001) study showed a high correlation between the BMD and the muscle mass in the left leg among 33 right-leg dominant male soccer players. Nevil et al (2003) considered the effects of specific training on BMD asymmetry between arms and legs among Caucasians athletes from ten sport disciplines (runners, cyclists, triathletes, racket players, rock climbers, swimmers, rugby players, rowers, kayakers and bodybuilders), at a university or higher level, trained for a minimum of four hour per week, for a minimum of three years. Using DXA on a Hologic QDR 1000W (Bedford, MA) scanner they stated that the differences between the dominant and non-dominant arms were significant but different depends on sport. For example it was greater in the racket players and lesser in the rowers. In the case of right and left legs, there was no asymmetric difference in BMD between them for all athletes combined nor within individual sport disciplines.
The McClanahan et al (2002) study investigated the effects of participation in different sports (NCAA Division I-A baseball, basketball, football, golf, soccer, tennis, cross-country, indoor/outdoor track and volleyball) on side-o-side diversification in bone mineral density (BMD) of the upper and lower limbs. On the basic of 184 collegiate athletes both sexes, and using dual energy X-ray absorptiometry, they revealed greater BMD of the right arms compared with the left arms for all teams. The largest differences were found in the case of
Dxa as a Tool for the Assessment of Morphological Asymmetry in Athletes 69
Al-Eisa, E.; Egan, D. & Wassersug, R. (2004). Fluctuating asymmetry and low back pain.
Alexander, M. (1983). The footwork pattern in the push stroke. *Counterattack*, Vol.3, No. 1,
Andreoli, A.; Scalzo, G.; Masala, S.; Tarantino, U. & Guglielmi, G. (2009). Body composition
Auerbach, BM. & Ruff, CB. (2006). Limb bone bilateral asymmetry: variability and
Bell, W. (2005). The regional placement of bone mineral mass, fat mass, and lean soft tissue
Broca, P. (1865). Sur le siege de la Faculte du langage articule. *Bulletin de la Societe de*
Bubanj, S. & Obradovic, B. (2002). Mechanical Force and Bones Density, Facta Universitatis.
Buckley, DC.; Kudsk, KA.; Rose, BA.; Fatzinger, P.; Koetting, CA. & Schlatter, M. (1987).
body mass. *Journal of the American Dietetic Association*, Vol.87, pp. 196-199. Burgess-Limerick, R.; Abernethy, B. & Neal, RJ. (1991). Experience and back swing
Buskirk, ER.; Andersen, KL. & Brozek, J. (1956). Unilateral activity and bone and muscle development in the forearm. *Research Quarterly*, Vol.27, pp. 127-131. Bussey, MD. (2010). Does the demand for asymmetric functional lower body postures in
Calbet, JA.; Moysi, JS.; Dorado, C.; Rodriguez, LP. (1998). Bone mineral content and density in professionals tennis players. *Calcified Tissue International*, Vol.62, pp. 491-496. Calbet, JAL.; Dorado, C.; Diaz-Herrera, P. & Rodriguez-Rodriguez, LP. (2001). High femoral
Chen, Z.; Wang, Z.; Lohman, T. et al. (2007). Dual-energy X-ray absorptiometry is a valid
Chettle, DR. & Fremlin, JH. (1984). Techniques of in vivo neutron activation analysis. *Physics*
Chhibber, SR. & Singh, I. (1970). Asymmetry in muscle weight and one sided dominance in
Chilibeck, PD.; Davison, KS.; Sale, DG.; Webber, CE. & Faulkner, RA. (2000). Effect of
lifespan. *American Journal of Human Biology*, Vol.12, No. 5, pp. 633-637. Chinn, CJ.; Priest, JD. & Kent, BE. (1974). Upper extremity range of motion, grip strength, and girth in highly skilled tennis players. *Physical Therapy*, Vol.54, pp. 474-482. Cresswell, A. & Elliott, B. (1987). The slap shot or drive in field hockey: a dilemma. *Sport*
the human lower limbs. *Journal of Anatomy*, Vol.106, pp. 553-556.
assessment by dual-energy X-ray absprptiometry (DXA). *Radiologia Medica*, Vol.114,
commonality among modern humans. *Journal of Human Evolution*, Vol.50, pp. 203-218.
mass in young adult Rugby Union players. *Ergonomics*, Vol.48, No.11-14, pp. 1462-
Anthropometric and computed tomography measurements of lower extremity lean
movement time variability: a short note concerning a serendipitous observation.
lateral sports related to structural asymmetry of the pelvis*. Journal of Science &*
bone mineral content and density in male football (soccer) players. *Medicine and*
tool for assessing skeletal muscle mass in older women. *Journal of Nutrition*,
physical activity on bone mineral density assessed by limb dominance across the
*Evolution and Human Behavior*, Vol.25, pp. 31-37.
1472. ISSN 0014-0139 print / ISSN 1366-5847 online.
Series:Physical Education and Sport, Vol.1, No.9, pp. 37-50.
*Science in Sports and Exercise*, Vol.33, No.10, pp. 1682-1687.
*Anthropologie* (Paris), Vol.6, pp. 377- 393.
*Human Movement Science,* Vol.10, pp. 621-627.
*in Medicine and Biology*, Vol.29, pp. 1011-1043.
*Medicine in Sport*, Vol.13, pp. 360-364.
Vol.137, pp. 2775-2780.
*Coach*, Vol.9, pp. 21-23.
pp. 14-18.
pp. 286-300.
men's and women's tennis and men's baseball. The differences in the lower limbs were less common in men and no significant in women.
One limb performance during some certain activities causes its higher fitness level than another limb. Thus training limb become dominate limb, in spite of athletes predisposition to right handedness, left handedness or ambidextratity (Starosta, 1990). Chilibeck et al (2000) asked the question if more bone deposition is a result of greater use of dominant limbs during physical activity or it is connected with fact that the dominant limb is genetically larger. They noticed, using dual energy x-ray absorptiometry, that a greater lifetime of preferential loading of the dominant arm is connected with a greater differences between arms in older group. Also earlier research reported greater mineralization for the dominant humerus but not for dominant radius and ulna, compared to the non-dominant ones, among baseball players 8-19 years of age (Watson, 1973). Additionally, they observed differences in the mineral content between the dominant and non-dominant increasing with age. Tomkinson study show that the increased metabolic rate that accompanies training will tend to increase fluctuating asymmetry in males athletes, at least if the training were conducted during the athlete's growth phase, when disturbances will have the greatest effect (Tomkinson et al., 2003).
#### **3. Conclusion**
All above information indicate that DXA is suitable method in athlete study. It allows to gain many valid data about not only overall but also regional body composition. Additionally it is precious method in longitudinal study of body composition assessment, tracking regional changes of some tissues during training program.
That kind of information could be important in a better understanding morphological characteristics of athletes, talent development, sports results improvement and with some injuries prevention. Further longitudinal research in this area could provide deeper insight into characteristic responses on specific training that require asymmetrical limb (side of body) use. Information about side-to-side differences in bone density may be valid for example for strength and conditioning professionals who want to include bilateral training programs minimizing stress-related injuries and maximizing sport results (McClanahan et al., 2002). It could be important to individualized symmetrization process among the athletes, because they do not have the same level of morphological differences. It is obvious that this process should be done methodologically (Starosta, 1990). After the few first stages, movements symmetrization could be used during training and the competitions, what could give quite big advantage over the opponents during the game. It relays on development through the training less fit side of the body and preservation harmony between symmetrical and asymmetrical movement (motor) preparation (Starosta, 1990).
#### **4. References**
men's and women's tennis and men's baseball. The differences in the lower limbs were less
One limb performance during some certain activities causes its higher fitness level than another limb. Thus training limb become dominate limb, in spite of athletes predisposition to right handedness, left handedness or ambidextratity (Starosta, 1990). Chilibeck et al (2000) asked the question if more bone deposition is a result of greater use of dominant limbs during physical activity or it is connected with fact that the dominant limb is genetically larger. They noticed, using dual energy x-ray absorptiometry, that a greater lifetime of preferential loading of the dominant arm is connected with a greater differences between arms in older group. Also earlier research reported greater mineralization for the dominant humerus but not for dominant radius and ulna, compared to the non-dominant ones, among baseball players 8-19 years of age (Watson, 1973). Additionally, they observed differences in the mineral content between the dominant and non-dominant increasing with age. Tomkinson study show that the increased metabolic rate that accompanies training will tend to increase fluctuating asymmetry in males athletes, at least if the training were conducted during the athlete's growth phase,
All above information indicate that DXA is suitable method in athlete study. It allows to gain many valid data about not only overall but also regional body composition. Additionally it is precious method in longitudinal study of body composition assessment,
That kind of information could be important in a better understanding morphological characteristics of athletes, talent development, sports results improvement and with some injuries prevention. Further longitudinal research in this area could provide deeper insight into characteristic responses on specific training that require asymmetrical limb (side of body) use. Information about side-to-side differences in bone density may be valid for example for strength and conditioning professionals who want to include bilateral training programs minimizing stress-related injuries and maximizing sport results (McClanahan et al., 2002). It could be important to individualized symmetrization process among the athletes, because they do not have the same level of morphological differences. It is obvious that this process should be done methodologically (Starosta, 1990). After the few first stages, movements symmetrization could be used during training and the competitions, what could give quite big advantage over the opponents during the game. It relays on development through the training less fit side of the body and preservation harmony between
when disturbances will have the greatest effect (Tomkinson et al., 2003).
tracking regional changes of some tissues during training program.
symmetrical and asymmetrical movement (motor) preparation (Starosta, 1990).
Abrahao, MRA. & Mello, D. (2008). Anthropometric differences between the right and the
Absauomov, TM. et al. (1976). Kontraktionsgeschwin-digkeit von Muskeln und Ihre Veranderung im sportlichen Training. *Leistungssport*, Vol.1, pp. 58-61.
left hemi-body of tennis instructor adults and children beginners in the sport and incidence of standard postural deviation*. Fitness Performance Journal,* Vol.7, No.4,
common in men and no significant in women.
**3. Conclusion**
**4. References**
pp. 264-270.
Dxa as a Tool for the Assessment of Morphological Asymmetry in Athletes 71
Henche, SA.; Torres, RR. & Pellico, LG. (2008). An evaluation of patterns of change in total
Heyward, VH. & Wagner, DR. (2004). *Applied body composition assessment*. Sd edition. Human
Jacobsen, T; Schubotz, RI. & Hofer, L. (2006). Brain correlates of aesthetic judgment of
Jones, H.; Priest, J.; Hayes, W.; Tichenor, C. et al. (1977). Humeral hypertrophy in response
Jordao, AA.; Bellucci, AD.; Dutra de Oliveira, JE. & Marchini, JS. (2004). Midarm
Kimura, K. & Asaeda, S. (1974). On the morphological and functional laterality in human
King, JW.; Grelsford, HJ. & Tullos, HS. (1969). Analysis of the pitching arm of the professional baseball pitcher. *Clinical Orthopedics*, Vol.67, pp. 116-123. Kistorp, CN.; Toubro, S.; Astrup, A. & Svendsen, OL. (2000). Measurements of body
Komi, PV. (1996).Strength and Power in Sport, in: volume III of The Encyclopedia of Sports
Koszczyc, T. (1991). *Asymetria morfologiczna I dynamiczna oraz możliwości jej kształtowania u dzieci w młodszym wieku szkolnym.* Studia i Monografie. AWF Wrocław. Poland. Krawczyk, B.; Skład, M.; Majle, B. & Jackiewicz, A. (1998). Lateran asymmetry in upper and
Kruger, A.; de Ridder, H.; Underhay, C. & Grobbelaae, H. (2005). Die voorkoms van
Krzykała, M. (2010). Location of body fat by Dual-Energy X-Ray Absorptiometry in
Lohman, TG. & Chen, Z. (2005). Dual-Energy X-Ray Absorptiometry. In *Human Body*
*Composition* (2nd ed: 63-77). Champaign, IL:Human Kinetics.
assessment data. *International Journal of Obesity*, Vol.28, pp. 1451-1455. Kerr, R. & Ness, K. (2006). Kinematics of the Field Hockey Penalty Corner Push-in. *Sports*
computerized tomography fat, muscle and total areas correlation with nutritional
extremities, especially in the lower limb. *Journal of the Anthropological Society of*
composition by dual-energy x-ray absorptiometry improve prediction of energy expenditure. *Annals of the New York Academy of Science*, Vol.904, No.1, pp. 79-84. Kohrt, WM. (1995). Body composition by DXA: tried and true? *Medicine and Science in Sports*
lower limb measurements in selected groups of male athletes. *Biology of Sport*,
morfologiese asimmetrie by aliteinternasionale manlike spiesgooiers. *South African Journal for Research in Sport, Physical Education and Recreation,* Vol.27, No.2, pp. 47-55. Kucera, K. & Henn, S. (2003). Pravention und physiotherapie von verletzungen und
uberlastungsschaden im fechten. *Sport Orthopedy and Traumatoly*, Vol.19, pp. 273-280.
professional male field hockey players in relation to their field position. *Studies in*
to exercise. *Journal of Bone and Joint Surgery*, 59a, pp. 204-208.
*Science in Sports*, Vol.8, pp. 102-108.
*Biomechanics,* Vol.5, No.1, pp. 47-61.
*Nippon*, Vol.1982, pp. 189-207.
*and Exercise*, Vol.27, pp.1349-53.
Vol.15, No.1, pp. 33-38.
Medicine (284-88). Blackwell science.
*Physical Culture and Tourism,* XVI, 2: 179-184.
beauty. *Neurolmage*, Vol.29, pp. 276-285.
Kinetics.
and regional body fat mass in healthy Spanish subjects using dual-energy X-ray absorptiometry (DXA). *European Journal of Clinical Nutrition*, Vol.62, pp. 1440-1448. Hetland, ML.; Haarbo, J. & Christianen, C. (1998). Regional body composition determined
by dual-energy x-ray absorptiometry. Relation to training, sex hormones, and serum lipids in male long-distance runners. *Scandinavian Journal of Medicine and*
Cronin, J. (2010). Leg Asymmetries During Running in Australia Rulet Football Players With
Czachowska-Sieszycka, B. (1983). Funkcjonalna asymetria mózgu. W: *Biologiczne podstawy zaburzeń psychoruchowego rozwoju dziecka*. Red. Budohoska, W. Warszawa. De Lorenzo, A.; Bertini, I.; Candeloro, N.; Iacopino, L.; Andreoli, A. & Van Loan, MD. (1998).
active adolescents. *British Journal of Sports Medicine*, Vol.32, pp. 215-219. Dorado, C.; Sanches Moysi, J.; Vicente, G.; Serrano, JA.; Rodriguez, LP. & Calbet, JAL.
Ducher, G.; Courteix, D.; Meme, S. et al. (2005). Bone geometry in response to long-term
resonance imaging study in tennis players. *Bone*, Vol.37, No.4, pp. 457-466. Egan, DA.; Cole, J. & Twomey, L. (1995). An alternative method for the measurement of
Ellis, KJ. (2000). Human Body composition: in vivo methods. *Physiological Review*, Vol.80, pp.
Engler, G. (2005). Einstain, his theories, and his aesthetic considerations. *International Studies*
Forbes, GB.; Brown, M. & Griffiths, HJB. (1988). Arm muscle plus bone area: anthropometric and CAT scan compared. *American Journal of Clinical Nutrition*, Vol.47, pp. 929-931. Faulkner, KG. (2006). Accuracy and precision of the Lunar iDXA, a new fan-beam
Frey, D. (1949). Zum Problem der Symmetrie in der bildenden Kunst. *Studium Generale*,
Fuller, NJ. & Elia, M. (1989). Potential use of bioelectrical impedance of "whole body" and of
Grobbelaar, HW. & de Ridder, Hans. (2000). Asymmetry in the upper body of high school fast
Hargittai, M. & Hargittai, I. (2005). Symmetry in chemistry. *European Review*, Vol.13, No.2,
*International Congress on Sport Science,* Brisbane, Australia, September 2000. Haapasalo, H.; Kannus, P.; Sivanen, H.; Pasanen, M.; Uusirasi, K.; Heinonen, A. et al. (1998).
tennis players. *Journal of Bone and Mineral Research,* Vol.13, pp. 310-9. Haarbo, J.; Gotfredsen, A.; Hassager, C. & Christiansen, C. (1991). Validation of body
Gould, JL. &Gould, CG. (1989). *Sexual selection*. Scientific American Library. New York. Graham, JH.; Freeman, DC. & Emlen, JM. (1993). Antisymmetry, directional asymmetry, and
dynamic morphogenesis. *Genetica*, Vol.89, pp. 121-137.
pp. 51-75. Academia Europeans, United Kingdom.
body segments for the assessment of body composition: A comparison with densitometry and anthropometry. *European Journal of Clinical Nutrition,* Vol.43, pp.
bowlers in cricket, *2000 Pre-Olympic Congress Sports Medicine and Physical Education.*
Effect of long-term unilateral activity on bone mineral density of female junior
composition by dual energy X-ray absorptiometry (DEXA). *Clin Physiol*, Vol.11, pp.
densitometer. *Journal of Clinical Densitometry*, Vol.9, No.2, pp. 237.
Suplement 1,1.
649-680.
779-791.
331-41.
Vol.2, pp. 268-278.
en.wikipedia.org
*Journal of Sports Science,* 20, pp. 591-597.
*in the Philosophy Science*, Vol.19, pp. 21-30.
*deformities,* IOS Press, Washington, DC, pp. 171-177.
Previous Hamstring Injuries. *Journal of Strength & Conditioning Research,* Vol.24,
Comparison of different techniques to measure body composition in moderately
(2002). Bone mass, bone mineral density and muscle mass in professional golfers.
tennis playing and its relationship with muscle volume: a quantitative magnetic
pelvic skeletal asymmetry and the association between asymmetry and back pain. In: *M. D'Amico, A. Merolli, G. Santabrogio, Editors, Three dimensional analysis of spinal*
Dxa as a Tool for the Assessment of Morphological Asymmetry in Athletes 73
Nystrom, J.; Lindwall, O.; Ceci, R.; Harmenberg, J.; Svadenhag, J. & Ekblom, B. (1990).
O'Connor, MK. (2006). Evaluation of the new Lunar iDXA bone densitometer. *Journal of*
Organ, LW.; Bradham, GB.; Gore, DT. & Lozier, SL. (1994). Segmental bioelectrical
Palmer, AR. & Strobeck, C. (1992). Fluctuating asymmetry as a measure of developmental
Petersen, HL.; Peterson, CT.; Reddy, MB.; Hanson, KB.; Swain, JH.; Sharp, RL. & Alekel, DL.
Pietrobelli, A.; Wang, Z.; Formica, C. et al. (1998). Dual-energy X-ray absorptiometry: fat
Pietrobelli, A.; Formica, C.; Wang, Z. & Heymsfield, SB. (1996). Dual-energy X-ray
Prives, MG. (1969). Influence of labor and sport upon skeletal structure in man. *Anatomical*
Roi, GS. & Bianchedi, D. (2008). The Science of Fencing. Implications for Performance and
Ruff, CB.; Walker, A. & Trinkaus, E. (1994). Postcranial robusticity in Homo. III Ontogeny.
Sanchis-Moysi, J.; Dorado, C.; Vicente-Rodriguez, G.; Milutinovic, L.; Garces, GL. & Calbet,
postmenopausal recreational tennis players. *Maturitas*, Vol.48, pp. 289-298. Schell, LM.; Johnson, FE.; Smith, DR. & Paolone, AM. (1985). Directional asymmetry of body
Scheltinge, MR.; Jacobs, D.; Kimbrough, TD. & Wilmore, DW. (1991).Alternations in Body
Singh, I. (1970). Functional asymmetry in the lower limbs. *Acta Anatomica*, Vol.77, pp. 131- 138. Speiser, PW.; Rudolf, MCJ.; Anhalt, H.; Camacho-Hubner, C.; Chiarelli, F.; Eliakim, A. et al.
Starosta, W. (1990). *Symmetry and asymmetry of movements in sport.* Institutute of Sport,
Steward, AD. & Hannan, WJ. (2000a). Sub-Regional Tissue Morphometry in Male Athletes
Steward, AD. & Hannan WJ. (200b). Prediction of fat and fat-free mass in male athletes
*Sport Nutrition and Exercise Metabolism*, Vol.10, pp. 157-169.
JAL. (2004). Inter-arm asymmetry in bone mineral content and bone area in
dimension among white adolescents? *American Journal of Physical Anthropology*,
Fluid Content Can Be Detected by Bioelectrical Impedance Analysis. *Journal of*
(2005). Consensus statement:n childhood obesity. *Journal of Clinical Endocrinology*
and Controls Using Dual X- Ray Absorptiometry (DXA). *International Journal of*
using X-ray absorptiometry as the reference method. *Journal of Sports Sciences*,
Injury Prevention. *Sports Medicine*, Vol.38, No.6, pp. 465-481.
*American Journal of Physical Anthropology*, 93, pp. 35-54.
*Journal of Sports Medicine*, Vol.11, No. 2, pp. 136-139.
*Clinical Densitometry*, Vol.9, No.2, pp. 237.
*Acta Zoologica Fennica*, Vol.191, pp. 57-72.
*Journal of Physiology,* Vol.271, pp. E941-E951.
*Physiology*, Vol.77, pp. 98-112.
*Physiology*, 1988, 274: 808-816.
*Record*, Vol.136, No.261.
Vol.67, pp. 317-322.
Warsaw, Poland.
Vol.18, No.1, pp. 263-274.
*Surgical Research*, Vol.50, pp. 461-468.
*and Metabolism*, Vol.90, pp. 1871-1887.
Vol.16, pp. 281-295.
Physiological and morphological characteristics of world class fencers. *International*
impedance analysis: Theory and application of a new technique. *Journal of Applied*
stability: implications of non-normal distributions and power of statistical tests.
(2006). Body composition, dietary intake, and iron status of female collegiate swimmers and divers. *International Journal of Sport Nutrition and Exercise Metabolism*,
estimation errors due to variation in soft tissue hydration. *American Journal of*
absorptiometry body composition model: Review of physical concepts. *American*
Lucki, N. (2006). Physiological Adaptations of the Upper Limb to the Biomechanical
Lukaski, HC. (1987). Methods for the assessment of human body composition: traditional
Malina, RM. (1979). The effects of exercise on specific tissues, dimensions, and functions
Malina, RM. (1983). Human growth, maturation, and regular physical activity. *Acta Medica*
Malina, RM.; Bouchard, C. & Bar-Or, O. (2004). Growth, Maturation and Physical Activity.
Malinowski A. (2004). *Auksologia. Rozwój osobniczy człowieka w ujęciu biomedycznym*. Zielona
Manning, JT. (1995). Fluctuating asymmetry and body weight in men and women: implications for sexual selection. *Ethology and Sociobiology*, Vol.16, pp. 145-153. Manning, JT. & Pickup LJ. (1998). Symmetry and performance in middle distance runners.
Manning , JT,: Gage, AR.; Diver, MJ.; Scutt, D. & Fraser, WD. (2002). Short-Term Changes in
Marchwicki, I. (1927). Obwody ramienia I przedramienia oraz ich asymetria u chłopców polskich w wieku od 9 do 19 lat. *Przegląd Antropologiczny*, tom II. Poland. Mascie-Taylor, CGN.; MacLarnon, AM.; Lanigan, PM. & McManus, IC. (1981). Foot-length
Maughan, RJ.; Abel, RW.; Watson, JS. & Weir, J. (1986). Forearm Composition and Muscle Function in Trained and Untrained Limbs. *Clinical Physiology*, Vol.6, pp. 389- 396. McClanahan, BS.; Harmon –Clayton, K.; Ward, KD.; Klesges, RC.; Vukadinovich, CM. &
McGrew, WC. & Marchant, LF. (1997). On the other hand: current issue in and meta-analysis
McLaughlin, P. (1997). Three-dimensional biomechanical analysis of the hockey drag flick. Full
Moreno, LA.; Rodriquez, G.; Guillen, J.; Rabanaque, MJ.; Leon, JF. & Arino, A. (2002).
Munter, AH. (1936). A study of the lengths of the long bones of the arms and legs in man, with special reference to Anglo-Saxon skeletons. *Biometrika,* Vol.28, pp. 258-294. Nevill, MA.; Holder, RL. & Steward, AD. (2003). Modeling elite male athlete's peripheral bone mass, assessed using regional dual x-ray absorptiometry. *Bone,* Vol.32, pp. 62-68.
Muller, A. (2003). Chemistry: The beauty of symmetry. *Science*, 300, pp. 749-750.
asymmetry, sex, and handedness. *Science,* Vol.212, pp. 1416-1417.
Asymmetry and Hormones in Men. *Evolution and Human Behaviour,* Vol. 23, pp. 95-42.
Cantler, ED. (2002). Side-to-side comparisons of bone mineral density in upper and lower limbs of collegiate athletes. *Journal of Strength and Conditioning Research*,
the behavioral laterality of hand function in nonhuman primates. *Yearbook of*
report to the Australian Sports Commission. Belconnen: Australian Sports Commision.
Anthropometric measurements in both sides of the body in the assessment of nutritional status in prepubertal children. *European Journal of Clinical Nutrition*,
*Undergraduate Research.* The University of North Carolina at Asheville. Lucki, N. & Nicolay, CW. (2007). Phenotypic Plasticity and Fuctional Asymmetry in
and new. *American Journal of Clinical Nutrition,* Vol.46, pp. 537-56.
during growth. *Studies in Physical Anthropology*, Vol.5, pp. 21-52.
*International Journal of Sports Medicine*, Vol.19, pp. 205-209.
*Journal of Human Biology*, Vol.19, pp. 566-577.
*Auxologica*, Vol.15, pp. 5-27.
Vol.16, No.4, pp. 586-590.
Vol.56, pp. 1208-1215.
*Physical Anthropology*, Vol.40, pp. 201-232.
Human Kinetics.
Góra. Poland.
Environment of Playing Tennis. *Proceedings of The National Conference On*
Response to Grip Forces Exterted by Intercollegiate Tennis Players. *American*
**6**
*Greece*
**Body Composition in Disabilities**
*2University of Athens, Laboratory for Research of the Musculoskeletal System*
Disability leads to immobilisation associated with profound changes in body composition. The potential risks involved with these changes i.e. loss of lean tissue mass (LM) and bone mineral density (BMD) vs. gain in fat mass (FM) in body composition have implications for the health of the disabled individuals (Jones et al., 1998). Body fat has been identified as a significant predictor of mortality in humans making body composition measurement to quantify nutritional and health status an important issue for human health. (Seidell et al., 1996; Bender et al., 1998; Van Der Ploeg et al., 2003). Moreover, some disorders such as carbohydrate intolerance, insulin resistance, lipid abnormalities, and heart disease occur prematurely and at a higher prevalence in disabled populations may be related to adverse changes in body composition that result from immobilization and skeletal muscle
In traumatic and pathological lesions of the central nervous system (CNS) there are differences according to the evolution or not of the lesion (i.e. progressive multiple sclerosis vs. complete paraplegia), the type of injury (i.e. lesion with a level of injury vs. upper motor neuron pyramidal lesion), life expectancy, the residual mobility and functionality, the ability to walk and stand (i.e. incomplete paraplegia vs. quadriplegia vs. high-low paraplegia) and drug treatment (i.e. frequent corticosteroid therapy in multiple sclerosis vs. long-term therapy with anticoagulants in paraplegia). In addition there are differences in the degree of spasticity which is likely to play a regulatory role in maintaining bone density (Dionyssiotis et al., 2011a). We need to take into account the element of fatigue and muscle weakness in disabilities, especially in diseases like multiple sclerosis, which significantly reduces the
The relative difference in energy expenditure between individuals with multiple sclerosis (MS) and able-bodied subjects is probably lower than the relative difference in physical activity, because individuals with MS have a higher energy expenditure of physical activity (Olgiati et al., 1988). Reduced physical activity (and probably reduced energy expenditure) in MS need to be accompanied by a reduction in energy intake otherwise body fat will increase (Lambert et al., 2002). Subjects with those motor disorders often face problems of depression and limit mobility (Dionyssiotis, 2011b). Moreover, in children with cerebral palsy (CP) studies suggest that increased stretch reflexes and muscle tone, weakness of
**1. Introduction**
denervation (Spungen et al., 2003).
mobility of these patients (Krupp et al., 2010).
**of Central Nervous System**
*1Physical and Social Rehabilitation Center Amyntæo*
Yannis Dionyssiotis1,2
### **Body Composition in Disabilities of Central Nervous System**
#### Yannis Dionyssiotis1,2
*1Physical and Social Rehabilitation Center Amyntæo 2University of Athens, Laboratory for Research of the Musculoskeletal System Greece*
#### **1. Introduction**
74 Dual Energy X-Ray Absorptiometry
The New Lexicon Webster's Encyclopedic Dictionary of the English Language. Deluxe
Taafe, DR.; Snow-Harter, C.; Connolly, DA.; Robinson, TL.; Brown, MD. & Marcus, R.
Tarociński, A. (1977). Kształtowanie się asymetrii morfologicznej i funkcjonalnej u
Tomkinson, GR.; Popovic, N. & Martin, M. (2003). Bilateral symmetry and the competitive
Tothill, P.; Arenell, A. & Reid, DM. (1994). Precision and accuracy of measurements of
Ulijaszek, SJ. & Mascie-Taylor, CGN. (2005). Anthropometry: the individual and the population, Chapter 2, 2nd, New York. Cambridge University Press. Watson, RC. (1973). Bone growth and physical activity in young males, Unpublished
Webber, CE. (1995). Dual photon transitions measurements of bone mass and body
Wilk, KJ.; Meister, K. & Andrews, JR. (2002). Current concepts in the rehabilitation of the
Wilmore, JH.; Frisancho, RA.; Gordon, CC.; Himes, JH.; Martin, AD.; Martorell, R. &
Wolański, N. (1955). Z badań nad tak zwaną maksymalną siłą mięśniową dłoni człowieka i
Wolański, N. (1962). Wpływ funkcji kończyn (boczności) na kształtowanie asymetrycznej
Van Dillen, LR.; Bloom, NJ.; Gombatto, SP. & Susco, TM. (2008). Hip rotation range of
Van Dusen, CR. (1939). An anthropometric study of the upper extremities of children.
Van Valen, A. (1962). A study of fluctuating asymmetry. *Evolution*, Vol. 16, pp. 125-142. Zeidel, DW. & Hessamian, M. (2010). Asymmetry and Symmetry in the Beauty of Human
doctoral dissertation, Madison: University of Wisconsin.
Martorell, pp. 27- 38. Champaign: Human Kinetics.
related sports. *Physical Therapy in Sport,* Vol. 9, pp. 72-81.
*Pediatric Exercise Science*. Champaign, IL: Human Kinetics: 57-76.
(1995). Differential effects of swimming versus weight-bearing activity on bone mineral status of amenorrhea athletes. *Journal of Bone and Mineral Research*, Vol.10,
młodzieży uprawiającej tenis ziemny. *Wychowanie Fizyczne i Sport*, Vol.2, pp. 41-46.
standard attained in elite and sub-elite sport. *Journal of Sports Science*, Vol.21, pp.
whole-body bone mineral: comparison between Hologic, Lunar and Norland dual energy X-ray absorptiometers. *British Journal of Radiology*, Vol.67, pp. 1210- 1217. Tsolakis, CH.; Bogdanis, GC. & Vagenas, G. (2006). Anthropometric profile and limb
asymmetries in young male and female fencers. *Journal of Human Movement Studies*,
composition during growth. In: C Blimke, Jr, O Bar-Or (eds), *New Horizons in*
overhead throwing athlete. *American Journal of Sports Medicine*, Vol.31, No.1, pp.
Seefeldt, RD. (1988). Body breadth equipment and measurement techniques. In *Anthropometric standardization reference manual,* ed. TG. Lohman, AF Roche, and R.
wartość tego pomiaru dla praktyki wychowania fizycznego. *Kultura Fizyczna*, Vol.2.
budowy ciała w aspekcie onto- i filogenezy. *Przegląd Antropologiczny*, XXVIII, 1, pp.
motion in people with and without low back pain who participate in rotation-
Edition. 1991. Lexicon Publications, Inc. New York.
No.4, pp. 586-593.
Vol. 5, pp. 201-216.
201-211.
136-51.
Poland.
27-59. Poland.
www.rjlsystem.com/pdf-files/segmental\_bia.pdf
*Human Biology*, Vol.11, pp. 277-284.
Faces. *Symmetry*, Vol.2, pp. 136-149.
Disability leads to immobilisation associated with profound changes in body composition. The potential risks involved with these changes i.e. loss of lean tissue mass (LM) and bone mineral density (BMD) vs. gain in fat mass (FM) in body composition have implications for the health of the disabled individuals (Jones et al., 1998). Body fat has been identified as a significant predictor of mortality in humans making body composition measurement to quantify nutritional and health status an important issue for human health. (Seidell et al., 1996; Bender et al., 1998; Van Der Ploeg et al., 2003). Moreover, some disorders such as carbohydrate intolerance, insulin resistance, lipid abnormalities, and heart disease occur prematurely and at a higher prevalence in disabled populations may be related to adverse changes in body composition that result from immobilization and skeletal muscle denervation (Spungen et al., 2003).
In traumatic and pathological lesions of the central nervous system (CNS) there are differences according to the evolution or not of the lesion (i.e. progressive multiple sclerosis vs. complete paraplegia), the type of injury (i.e. lesion with a level of injury vs. upper motor neuron pyramidal lesion), life expectancy, the residual mobility and functionality, the ability to walk and stand (i.e. incomplete paraplegia vs. quadriplegia vs. high-low paraplegia) and drug treatment (i.e. frequent corticosteroid therapy in multiple sclerosis vs. long-term therapy with anticoagulants in paraplegia). In addition there are differences in the degree of spasticity which is likely to play a regulatory role in maintaining bone density (Dionyssiotis et al., 2011a). We need to take into account the element of fatigue and muscle weakness in disabilities, especially in diseases like multiple sclerosis, which significantly reduces the mobility of these patients (Krupp et al., 2010).
The relative difference in energy expenditure between individuals with multiple sclerosis (MS) and able-bodied subjects is probably lower than the relative difference in physical activity, because individuals with MS have a higher energy expenditure of physical activity (Olgiati et al., 1988). Reduced physical activity (and probably reduced energy expenditure) in MS need to be accompanied by a reduction in energy intake otherwise body fat will increase (Lambert et al., 2002). Subjects with those motor disorders often face problems of depression and limit mobility (Dionyssiotis, 2011b). Moreover, in children with cerebral palsy (CP) studies suggest that increased stretch reflexes and muscle tone, weakness of
Body Composition in Disabilities of Central Nervous System 77
No differences were found in BMI between paraplegics in the acute phase of injury and controls, which is a finding in accordance with other studies in which, despite the same BMI, the body composition and the distribution of fat and fat free mass were altered in patients with spinal cord injury, with the fat free mass being statistically significantly lower in paraplegic patients in total body composition and in the lower, but not the upper limbs. As far as the fat mass is concerned, it was statistically significantly higher (kilograms and %)
These findings show that using the BMI does not contribute substantially in determining the body composition of paraplegics and lowers the percentage of fat in this population, finding that agrees with other studies and shows that the anthropometric measurement with BMI in paraplegics, underestimates fat in body composition when measurements are compared
Body mass index is a very simple measurement of fat; however it does not distinguish the individual components of weight. The applicability of conventional BMI cut off values is into question (Buchholz, 2005; McDonald et al., 2007). BMI is an insensitive marker of obesity in subjects with SCI and measuring fat with BMI in chronic paraplegic patients is not
To standardize or index physiological variables, such as resting metabolic rate and power fat free mass (FFM) is usually used (Van Der Ploeg et al., 2003). Skeletal muscle represents 50% of the non fat component in the total body (Clarys et al., 1984; Modlesky et al., 2004) and exact quantification of the amount of skeletal muscle is important to assess nutritional status, disease risk, danger of illnesses, physical function, atrophic effects of aging, and
Because muscle wasting is a common sign of cerebral palsy (CP), even in well nourished children, the validity of using muscle wasting as evidence or measurement of malnutrition in CP is in doubt. Studies found that the triceps, midthigh, and calf skinfold thicknesses of the affected side were greater than those of the no affected side among children with hemiplegic CP (Stevenson et al., 1995). Useful information regarding fat provide triceps, subscapular skinfolds and arm-fat area (Patrick & Gisel, 1990). Other studies support the concept that the validity of skinfold thickness as an assessment of limb fat storage is dependent on the preservation of limb muscles (Ingemann-Hansen T et al., 1977) and suggested good sensitivity and specificity of triceps skinfold thickness for predicting midupper arm fat area probably were attributable to good preservation of mid-upper arm
In disabled children techniques for measuring skinfolds are well established and standardised (Lohman et al., 1988) and equations are available for calculation of body fat from skin fold thickness (Slaughter et al., 1988) although unvalidated in this population, as are normative values for skinfold thickness (Frisancho, 1981; Kuperminc & Stevenson, 2008). Consequently, use of skinfold thickness as a measurement, especially for the affected limb, should be used with discretion in the assessment of children with CP, who tend to have
In cerebral palsy neither bioelectrical impedance analysis nor predictive equations for skinfold thickness generated from normal, able-bodied adults accurately determined
in the total body composition in the upper and lower limbs (Maimoun et al., 2006).
enough to determine subject's percentage of fat in the body (Olle et al., 1993).
muscle-wasting diseases (Forbes, 1987; Mojtahedi et al., 2008).
muscles among children with CP (Samson-Fang et al., 2000).
muscle wasting.
with healthy subjects (Jones et al., 1998).
involved musculature, and severe limitation of movement reduce the capacity to perform normal movements creating ambulation barriers limiting physical activity. The dependency on mobility devices, common in all disabilities, and the frequent periods of immobilization after multiple operative procedures contribute to the hypoactivity status of such children. It could be assumed that, under these conditions, body composition may be significantly compromised (Chad et al., 2000).
Studies found that lean mass of the contralateral limb was lower compared to the ipsilateral limb in upper motor neuron injury, as occurs in stroke (Ryan et al., 2000; 2002). Similar findings of reduced muscle mass and increased intramuscular fat have been also published in individuals with incomplete spinal cord injury (SCI) (Gorgey et al., 2007) suggesting that reduced muscle mass is fundamentally related to poor fitness and physical performance capacity after stroke (Hafer-Macko et al., 2008).
On the other side the clinical equivalence of diseases with different physiopathology, location, evolution, etc. could be similar; i.e. a severe form of MS can result in a wheelchair bound patient a clinical figure equivalent to paraplegia or a MS patient may have a more appropriate walking gait pattern vs. a patient with incomplete paraplegia but may also be unable to walk at all, is bedridden and vice versa (Dionyssiotis, 2011b; 2011c; 2011d). In addition to these differences and according to osteoporosis the role of factors which do not change, such as race or gender of patients has not been yet clarified, although there are few studies in women debating that bone mass in women with disabilities is more affected than men (Smeltzer et al., 2005; Coupaud et al., 2009).
Therefore, the purpose of this chapter was to present the bone-mineral density, bonemineral content, and bone-mineral-free lean and fat tissue mass alterations of ambulatory and non-ambulatory subjects with disabilities of the central nervous system.
#### **2. Body composition measurements**
#### **2.1 Anthropometric and various techniques of body composition measurements**
In a study which investigated a chronic spinal cord injury (SCI) population with paraplegia (Dionyssiotis, 2008a, Dionyssiotis et al., 2008b) values of body mass index (BMI, kg/m2) did not present statistical significance in relation to the controls, which is a finding in line with the literature (Maggioni et al., 2003; Mamoun et al., 2004).. Nevertheless, there are studies which demonstrate the usefulness of BMI as an indicator of obesity, in body composition in people with spinal cord injury (Gupta et al., 2006). These studies, however, included both tetraplegics and middle-aged people unlike the Greek one which included relatively young individuals (Dionyssiotis et al., 2008a). Whether the criteria of BMI may assess obesity in people with spinal cord injury the latest studies show the opposite (McDonald et al., 2007).
BMI of a male paraplegic group was slightly greater compared to a tetraplegic one and distribution of BMI by level of injury was similar with 37.5% and 40.5% of the male tetraplegic and male paraplegic groups, respectively, falling into the recommended BMI range. Approximately 50% in each male group were overweight by BMI, and 12.5% and 10.8%, respectively, were classified as obese. Overall, when compared with the general population-observed distribution by BMI, a greater proportion of men with SCI fell into the desirable BMI range and fewer fell into the obese category (Groah et al., 2009).
involved musculature, and severe limitation of movement reduce the capacity to perform normal movements creating ambulation barriers limiting physical activity. The dependency on mobility devices, common in all disabilities, and the frequent periods of immobilization after multiple operative procedures contribute to the hypoactivity status of such children. It could be assumed that, under these conditions, body composition may be significantly
Studies found that lean mass of the contralateral limb was lower compared to the ipsilateral limb in upper motor neuron injury, as occurs in stroke (Ryan et al., 2000; 2002). Similar findings of reduced muscle mass and increased intramuscular fat have been also published in individuals with incomplete spinal cord injury (SCI) (Gorgey et al., 2007) suggesting that reduced muscle mass is fundamentally related to poor fitness and physical performance
On the other side the clinical equivalence of diseases with different physiopathology, location, evolution, etc. could be similar; i.e. a severe form of MS can result in a wheelchair bound patient a clinical figure equivalent to paraplegia or a MS patient may have a more appropriate walking gait pattern vs. a patient with incomplete paraplegia but may also be unable to walk at all, is bedridden and vice versa (Dionyssiotis, 2011b; 2011c; 2011d). In addition to these differences and according to osteoporosis the role of factors which do not change, such as race or gender of patients has not been yet clarified, although there are few studies in women debating that bone mass in women with disabilities is more affected than
Therefore, the purpose of this chapter was to present the bone-mineral density, bonemineral content, and bone-mineral-free lean and fat tissue mass alterations of ambulatory
**2.1 Anthropometric and various techniques of body composition measurements**
In a study which investigated a chronic spinal cord injury (SCI) population with paraplegia (Dionyssiotis, 2008a, Dionyssiotis et al., 2008b) values of body mass index (BMI, kg/m2) did not present statistical significance in relation to the controls, which is a finding in line with the literature (Maggioni et al., 2003; Mamoun et al., 2004).. Nevertheless, there are studies which demonstrate the usefulness of BMI as an indicator of obesity, in body composition in people with spinal cord injury (Gupta et al., 2006). These studies, however, included both tetraplegics and middle-aged people unlike the Greek one which included relatively young individuals (Dionyssiotis et al., 2008a). Whether the criteria of BMI may assess obesity in people with spinal cord injury the latest studies show the opposite (McDonald et al., 2007). BMI of a male paraplegic group was slightly greater compared to a tetraplegic one and distribution of BMI by level of injury was similar with 37.5% and 40.5% of the male tetraplegic and male paraplegic groups, respectively, falling into the recommended BMI range. Approximately 50% in each male group were overweight by BMI, and 12.5% and 10.8%, respectively, were classified as obese. Overall, when compared with the general population-observed distribution by BMI, a greater proportion of men with SCI fell into the
and non-ambulatory subjects with disabilities of the central nervous system.
desirable BMI range and fewer fell into the obese category (Groah et al., 2009).
compromised (Chad et al., 2000).
capacity after stroke (Hafer-Macko et al., 2008).
men (Smeltzer et al., 2005; Coupaud et al., 2009).
**2. Body composition measurements**
No differences were found in BMI between paraplegics in the acute phase of injury and controls, which is a finding in accordance with other studies in which, despite the same BMI, the body composition and the distribution of fat and fat free mass were altered in patients with spinal cord injury, with the fat free mass being statistically significantly lower in paraplegic patients in total body composition and in the lower, but not the upper limbs. As far as the fat mass is concerned, it was statistically significantly higher (kilograms and %) in the total body composition in the upper and lower limbs (Maimoun et al., 2006).
These findings show that using the BMI does not contribute substantially in determining the body composition of paraplegics and lowers the percentage of fat in this population, finding that agrees with other studies and shows that the anthropometric measurement with BMI in paraplegics, underestimates fat in body composition when measurements are compared with healthy subjects (Jones et al., 1998).
Body mass index is a very simple measurement of fat; however it does not distinguish the individual components of weight. The applicability of conventional BMI cut off values is into question (Buchholz, 2005; McDonald et al., 2007). BMI is an insensitive marker of obesity in subjects with SCI and measuring fat with BMI in chronic paraplegic patients is not enough to determine subject's percentage of fat in the body (Olle et al., 1993).
To standardize or index physiological variables, such as resting metabolic rate and power fat free mass (FFM) is usually used (Van Der Ploeg et al., 2003). Skeletal muscle represents 50% of the non fat component in the total body (Clarys et al., 1984; Modlesky et al., 2004) and exact quantification of the amount of skeletal muscle is important to assess nutritional status, disease risk, danger of illnesses, physical function, atrophic effects of aging, and muscle-wasting diseases (Forbes, 1987; Mojtahedi et al., 2008).
Because muscle wasting is a common sign of cerebral palsy (CP), even in well nourished children, the validity of using muscle wasting as evidence or measurement of malnutrition in CP is in doubt. Studies found that the triceps, midthigh, and calf skinfold thicknesses of the affected side were greater than those of the no affected side among children with hemiplegic CP (Stevenson et al., 1995). Useful information regarding fat provide triceps, subscapular skinfolds and arm-fat area (Patrick & Gisel, 1990). Other studies support the concept that the validity of skinfold thickness as an assessment of limb fat storage is dependent on the preservation of limb muscles (Ingemann-Hansen T et al., 1977) and suggested good sensitivity and specificity of triceps skinfold thickness for predicting midupper arm fat area probably were attributable to good preservation of mid-upper arm muscles among children with CP (Samson-Fang et al., 2000).
In disabled children techniques for measuring skinfolds are well established and standardised (Lohman et al., 1988) and equations are available for calculation of body fat from skin fold thickness (Slaughter et al., 1988) although unvalidated in this population, as are normative values for skinfold thickness (Frisancho, 1981; Kuperminc & Stevenson, 2008). Consequently, use of skinfold thickness as a measurement, especially for the affected limb, should be used with discretion in the assessment of children with CP, who tend to have muscle wasting.
In cerebral palsy neither bioelectrical impedance analysis nor predictive equations for skinfold thickness generated from normal, able-bodied adults accurately determined
Body Composition in Disabilities of Central Nervous System 79
To determine whether bioelectrical impedance analysis and anthropometry can be used to determine body composition for clinical and research purposes in children with cerebral palsy 8 individuals (two female, mean age=10 years, mean gross motor function classification=4.6 [severe motor impairment]) recruited from an outpatient tertiary care setting underwent measurement of fat mass, fat-free mass, and percentage body fat using BIA, anthropometry (two and four skinfold equations), and dual-energy x-ray absorptiometry. Correlation were excellent for determination of fat-free mass for all methods (i.e., all were above 0.9) and moderate for determination of fat mass and percent body fat (range=0.4 to 0.8). Moreover, skinfolds were better predictors of percent body fat, while bioelectrical impedance was a better predictor for fat mass (Liu et al., 2005). On the contrary another study investigated the pattern of body composition in 136 subjects with spastic quadriplegic cerebral palsy, 2 to 12 years of age, by anthropometric measures, or by anthropometric and total body water (TBW) measures (n = 28), compared with 39 control subjects. Body composition and nutritional status indicators were significantly reduced. Calculation of body fat from two skinfolds correlated best with measures of fat mass from
Magnetic resonance imaging (MRI) provides remarkably accurate estimates of skeletal muscle in vivo (Modlesky et al., 2004). MRI and also quantitative computed tomography (QCT) have been validated in studies of humancadavers in the assessment of regional skeletal muscle (Mitsiopoulos et al., 1998). Although, these devices have disadvantages of
Recently, dual-energy X-ray absorptiometry (DXA) has gained acceptance as a reference method for body composition analysis (Mahon et al., 2007; LaForgia et al., 2009). Originally designed to determine bone density, DXA technology has subsequently been adopted for the assessment of whole body composition and offers estimation rapidly, non-invasively and with minimal radiation exposure (Van Der Ploeg et al., 2003; Dionyssiotis et al., 2008a). Moreover, is well tolerated in subjects who would be unable to tolerate other body composition techniques, such as underwater weighing (hydro-densitometry) (Laskey, 1996). DXA software determines the bone mineral and soft tissue composition in different regions of the body being a three-compartment model that quantifies: (i) bone mineral density and content (BMD, BMC), (ii) fat mass (FM); and (iii) lean mass (LM), half of which is closely correlated with muscle mass and also yields regional as well as total body values (Rittweger
DXA analyzes differently the dense pixels in body composition. Soft tissue pixels are analyzed for two materials: fat and fat-free tissue mass. Variations in the fat mass/fat free tissue mass composition of the soft tissue produce differences in the respective attenuation coefficients at both energy levels. The ratio at the two main energy peaks is automatically calculated of the X-ray attenuation providing separation of the soft tissue compartment into fat mass and fat-free tissue mass (lean mass) (Peppler & Mazess, 1981; Pietrobelli et al., 1996). A bone-containing pixel is analyzed for "bone mass" (bone mineral content, BMC) and soft tissue as the two materials. Thus, the fat mass/fat free tissue mass of the soft tissue component of the bone pixels cannot be measured, but only estimated
TBW (Stallings et al., 1995; Kuperminc & Stevenson, 2008).
et al., 2000) for example in the arms, legs, and trunk (figure 1).
high radiation exposure and are expensive.
(Ferretti et al., 2001).
**2.2 Dual-energy X-ray absorptiometry (DXA)**
percentage body fat (Hildreth et al., 1997). Body mass index (BMI), triceps skinfold thickness, subscapular skinfold thickness, suprailiac skinfold thickness, and circumferences of the biceps, waist, forearm, and knee were all significantly correlated with percentage body fat (Bandini et al., 1991).
BMI in patients with MS was statistically less compared to age comparable controls (Formica et al., 1997). In a recent study both total body fat and mass percent showed consistent significant dependence from BMI, as among normal subjects. Multiple linear regression analysis of bone mineral percent at all studied sites showed consistent dependence from BMI (increased with higher BMI) for both patient and control groups (Sioka et al., 2011).
Changes in body composition in spinal cord injured subjects can be assessed with various techniques including isotope-labelled water (Jones et al., 1998) total body potassium counting (Lussier et al., 1983; Spungen et al., 1992) anthropometric measures (Bulbulian et al., 1987) hydrodensitometry (Lussier et al., 1983; Sedlock, 1990) dual photon absorptiometry (DPA) (Spungen et al., 1992; Changlai, 1996) and dual energy X-ray absorptiometry (DXA) (Jones et al., 1998). However, some of these methods are not particularly suitable for use in the SCI population.
The hydrodensitometric model was regarded as the "gold standard" for body composition assessment. This model partitions the body into two compartments of constant densities [fat mass: 0.9007 g/cm3 and FFM: 1.100 g/cm3] and assumes that the relative amounts of the FFM components [water, protein, protein, bone mineral (BM), and non-BM] are fixed (Brozek et al., 1963; Van Der Ploeg et al., 2003). Hydrodensitometry is clearly inappropriate for individuals who deviate from these fixed and/or assumed values (e.g., children, elderly, blacks, obese), and its application is, therefore, somewhat limited (Womersley et al., 1976; Schutte, 1984; Lohman, 1986; Fuller et al., 1996).
Bioelectrical impedance analysis (BIA) has been used to measure cerebral palsy subjects. However, the inclusion of weight in the BIA predictive equation may reduce its accuracy in determining change in lean body mass (Forbes et al., 1992). The inability of BIA to accurately predict percentage body fat in the sample may be related to several factors. In the BIA method where the impedance of a geometrical system (i.e., the human body) is dependent on the length of the conductor (height) and its configuration, it is almost impossible to measure accurately height in subjects with CP because of their muscle contractures. An over- or underestimation of height by 2.5 cm can result in a l.0-L error in the estimation of TBW, producing a small error in the estimation of percentage body fat (< 5%). The second major problem is body asymmetry which renders the assumption of a symmetrical configuration of the human body invalid in this case. (National Institutes of Health Technology Assessment Conference Statement, 1994; Hildreth et al., 1997).
Isotope dilution measures the water compartment of the whole body rather than a single area assumed to mimic the composition of the whole body. Thus, the use of a stable isotope to measure body composition is ideal for people with CP because it is non-invasive, does not require the subject to remain still for the measurement, and is independent of height and body symmetry. However, the prohibitive cost of the isotopes and the need for a mass spectrometry facility and highly trained technicians make this method impractical for routine clinical use (Hildreth et al., 1997).
percentage body fat (Hildreth et al., 1997). Body mass index (BMI), triceps skinfold thickness, subscapular skinfold thickness, suprailiac skinfold thickness, and circumferences of the biceps, waist, forearm, and knee were all significantly correlated with percentage
BMI in patients with MS was statistically less compared to age comparable controls (Formica et al., 1997). In a recent study both total body fat and mass percent showed consistent significant dependence from BMI, as among normal subjects. Multiple linear regression analysis of bone mineral percent at all studied sites showed consistent dependence from BMI (increased with higher BMI) for both patient and control groups
Changes in body composition in spinal cord injured subjects can be assessed with various techniques including isotope-labelled water (Jones et al., 1998) total body potassium counting (Lussier et al., 1983; Spungen et al., 1992) anthropometric measures (Bulbulian et al., 1987) hydrodensitometry (Lussier et al., 1983; Sedlock, 1990) dual photon absorptiometry (DPA) (Spungen et al., 1992; Changlai, 1996) and dual energy X-ray absorptiometry (DXA) (Jones et al., 1998). However, some of these methods are not particularly suitable for use in
The hydrodensitometric model was regarded as the "gold standard" for body composition assessment. This model partitions the body into two compartments of constant densities [fat mass: 0.9007 g/cm3 and FFM: 1.100 g/cm3] and assumes that the relative amounts of the FFM components [water, protein, protein, bone mineral (BM), and non-BM] are fixed (Brozek et al., 1963; Van Der Ploeg et al., 2003). Hydrodensitometry is clearly inappropriate for individuals who deviate from these fixed and/or assumed values (e.g., children, elderly, blacks, obese), and its application is, therefore, somewhat limited (Womersley et al., 1976;
Bioelectrical impedance analysis (BIA) has been used to measure cerebral palsy subjects. However, the inclusion of weight in the BIA predictive equation may reduce its accuracy in determining change in lean body mass (Forbes et al., 1992). The inability of BIA to accurately predict percentage body fat in the sample may be related to several factors. In the BIA method where the impedance of a geometrical system (i.e., the human body) is dependent on the length of the conductor (height) and its configuration, it is almost impossible to measure accurately height in subjects with CP because of their muscle contractures. An over- or underestimation of height by 2.5 cm can result in a l.0-L error in the estimation of TBW, producing a small error in the estimation of percentage body fat (< 5%). The second major problem is body asymmetry which renders the assumption of a symmetrical configuration of the human body invalid in this case. (National Institutes of Health
Isotope dilution measures the water compartment of the whole body rather than a single area assumed to mimic the composition of the whole body. Thus, the use of a stable isotope to measure body composition is ideal for people with CP because it is non-invasive, does not require the subject to remain still for the measurement, and is independent of height and body symmetry. However, the prohibitive cost of the isotopes and the need for a mass spectrometry facility and highly trained technicians make this method impractical for
Technology Assessment Conference Statement, 1994; Hildreth et al., 1997).
body fat (Bandini et al., 1991).
(Sioka et al., 2011).
the SCI population.
Schutte, 1984; Lohman, 1986; Fuller et al., 1996).
routine clinical use (Hildreth et al., 1997).
To determine whether bioelectrical impedance analysis and anthropometry can be used to determine body composition for clinical and research purposes in children with cerebral palsy 8 individuals (two female, mean age=10 years, mean gross motor function classification=4.6 [severe motor impairment]) recruited from an outpatient tertiary care setting underwent measurement of fat mass, fat-free mass, and percentage body fat using BIA, anthropometry (two and four skinfold equations), and dual-energy x-ray absorptiometry. Correlation were excellent for determination of fat-free mass for all methods (i.e., all were above 0.9) and moderate for determination of fat mass and percent body fat (range=0.4 to 0.8). Moreover, skinfolds were better predictors of percent body fat, while bioelectrical impedance was a better predictor for fat mass (Liu et al., 2005). On the contrary another study investigated the pattern of body composition in 136 subjects with spastic quadriplegic cerebral palsy, 2 to 12 years of age, by anthropometric measures, or by anthropometric and total body water (TBW) measures (n = 28), compared with 39 control subjects. Body composition and nutritional status indicators were significantly reduced. Calculation of body fat from two skinfolds correlated best with measures of fat mass from TBW (Stallings et al., 1995; Kuperminc & Stevenson, 2008).
Magnetic resonance imaging (MRI) provides remarkably accurate estimates of skeletal muscle in vivo (Modlesky et al., 2004). MRI and also quantitative computed tomography (QCT) have been validated in studies of humancadavers in the assessment of regional skeletal muscle (Mitsiopoulos et al., 1998). Although, these devices have disadvantages of high radiation exposure and are expensive.
#### **2.2 Dual-energy X-ray absorptiometry (DXA)**
Recently, dual-energy X-ray absorptiometry (DXA) has gained acceptance as a reference method for body composition analysis (Mahon et al., 2007; LaForgia et al., 2009). Originally designed to determine bone density, DXA technology has subsequently been adopted for the assessment of whole body composition and offers estimation rapidly, non-invasively and with minimal radiation exposure (Van Der Ploeg et al., 2003; Dionyssiotis et al., 2008a). Moreover, is well tolerated in subjects who would be unable to tolerate other body composition techniques, such as underwater weighing (hydro-densitometry) (Laskey, 1996). DXA software determines the bone mineral and soft tissue composition in different regions of the body being a three-compartment model that quantifies: (i) bone mineral density and content (BMD, BMC), (ii) fat mass (FM); and (iii) lean mass (LM), half of which is closely correlated with muscle mass and also yields regional as well as total body values (Rittweger et al., 2000) for example in the arms, legs, and trunk (figure 1).
DXA analyzes differently the dense pixels in body composition. Soft tissue pixels are analyzed for two materials: fat and fat-free tissue mass. Variations in the fat mass/fat free tissue mass composition of the soft tissue produce differences in the respective attenuation coefficients at both energy levels. The ratio at the two main energy peaks is automatically calculated of the X-ray attenuation providing separation of the soft tissue compartment into fat mass and fat-free tissue mass (lean mass) (Peppler & Mazess, 1981; Pietrobelli et al., 1996). A bone-containing pixel is analyzed for "bone mass" (bone mineral content, BMC) and soft tissue as the two materials. Thus, the fat mass/fat free tissue mass of the soft tissue component of the bone pixels cannot be measured, but only estimated (Ferretti et al., 2001).
Body Composition in Disabilities of Central Nervous System 81
in spinal cord injured subjects (Modlesky et al., 2004). DXA technique has been used in assessment of SCI and appears to be tolerated well by this population (Szollar et al., 1997;
Fig. 2. Whole body and regional distribution of fat mass, lean mass, bone mineral content (BMC) and bone mineral density (BMD) from control male subject using whole body DEXA Norland X-36 and values of measured parameters. Modified and translated with permission
Spinal cord injury (SCI) always results in substantial and rapid bone loss predominately in areas below the neurological level of injury. The predominant finding of SCI on bone is a large loss of bone during the first year of injury (Spungen et al., 2003) and an ongoing demineralisation 3 years after trauma in tibia (Biering-Sörensen et al., 1988) with a progressive bone loss over 12 to 16 months prior to stabilizing (Lazo et al., 2001) was
Cancellous bone is more affected than cortical bone after SCI. In a prospective study, six acute tetraplegics were followed up for 12 months, and the trabecular and cortical BMD's of the tibia were found to be decreased by 15 and 7% (Frey-Rindova et al., 2000), while in
Uebelhart et al., 1995; Chow et al., 1996).
from Dionyssiotis, 2008a.
**3.1 Spinal cord injury**
demonstrated.
**3. Physiopathological context**
Fig. 1. Whole body and regional distribution of fat mass, lean mass, bone mineral content (BMC) and bone mineral density (BMD) from paraplegic subject thoracic 6 using whole body DXA (Norland X-36, Fort Atkinson, Wisconsin, USA) and values of measured parameters. Modified and translated with permission from Dionyssiotis, 2008a.
The important issue on this is the investigation of distribution of bone mineral, fat and mass throughout the body. These changes induce the risk for diseases such as diabetes, coronary heart disease, dyslipidaimias and osteoporosis (Bauman et al., 1992; Bauman & Spungen, 1994; Kocina, 1997; Garland et al., 1992). There is a need to quantify the alterations in body composition to prevent these diseases and their complications. Studies also reported that bone density measurements at one site cannot usefully predict the bone density elsewhere (Heymsfield et al., 1989) because different skeletal regions, even with similar quantities of trabecular or cortical bone, may respond variably in different physiopathological conditions (Laskey, 1996).
In disabled conditions the accuracy of skeletal muscle measured by DXA may be compromised when muscle atrophy is present. A lower ratio of muscle to adipose-tissuefree mass indicates a lower proportion of muscle in the fat-free soft tissue mass. Crosssectional area of skeletal muscle in the thighs after SCI is extensively reduced (Castro et al., 1999). If this is the case muscle mass would be overestimated by prediction models that assume that muscle represents all or a certain proportion of the fat-free soft tissue mass, i.e.
Fig. 1. Whole body and regional distribution of fat mass, lean mass, bone mineral content (BMC) and bone mineral density (BMD) from paraplegic subject thoracic 6 using whole body DXA (Norland X-36, Fort Atkinson, Wisconsin, USA) and values of measured parameters. Modified and translated with permission from Dionyssiotis, 2008a.
The important issue on this is the investigation of distribution of bone mineral, fat and mass throughout the body. These changes induce the risk for diseases such as diabetes, coronary heart disease, dyslipidaimias and osteoporosis (Bauman et al., 1992; Bauman & Spungen, 1994; Kocina, 1997; Garland et al., 1992). There is a need to quantify the alterations in body composition to prevent these diseases and their complications. Studies also reported that bone density measurements at one site cannot usefully predict the bone density elsewhere (Heymsfield et al., 1989) because different skeletal regions, even with similar quantities of trabecular or cortical bone, may respond variably in different physiopathological conditions
In disabled conditions the accuracy of skeletal muscle measured by DXA may be compromised when muscle atrophy is present. A lower ratio of muscle to adipose-tissuefree mass indicates a lower proportion of muscle in the fat-free soft tissue mass. Crosssectional area of skeletal muscle in the thighs after SCI is extensively reduced (Castro et al., 1999). If this is the case muscle mass would be overestimated by prediction models that assume that muscle represents all or a certain proportion of the fat-free soft tissue mass, i.e.
(Laskey, 1996).
in spinal cord injured subjects (Modlesky et al., 2004). DXA technique has been used in assessment of SCI and appears to be tolerated well by this population (Szollar et al., 1997; Uebelhart et al., 1995; Chow et al., 1996).
Fig. 2. Whole body and regional distribution of fat mass, lean mass, bone mineral content (BMC) and bone mineral density (BMD) from control male subject using whole body DEXA Norland X-36 and values of measured parameters. Modified and translated with permission from Dionyssiotis, 2008a.
#### **3. Physiopathological context**
#### **3.1 Spinal cord injury**
Spinal cord injury (SCI) always results in substantial and rapid bone loss predominately in areas below the neurological level of injury. The predominant finding of SCI on bone is a large loss of bone during the first year of injury (Spungen et al., 2003) and an ongoing demineralisation 3 years after trauma in tibia (Biering-Sörensen et al., 1988) with a progressive bone loss over 12 to 16 months prior to stabilizing (Lazo et al., 2001) was demonstrated.
Cancellous bone is more affected than cortical bone after SCI. In a prospective study, six acute tetraplegics were followed up for 12 months, and the trabecular and cortical BMD's of the tibia were found to be decreased by 15 and 7% (Frey-Rindova et al., 2000), while in
Body Composition in Disabilities of Central Nervous System 83
The neurological level of the lesion i.e. the extent of impairment of motor and sensory function is important, because tetraplegics are more likely to lose more bone mass throughout the skeleton than paraplegics (Tsuzuku et al., 1999). In paraplegics legs' BMC was reduced vs. controls, independently of the neurological level of injury and negatively correlated with the duration of paralysis in total paraplegic group, but after investigation according to the neurological level of injury this correlation was due to the strong correlation of high paraplegics' legs BMC with the duration of paralysis, meaning that the neurological level of injury determines the extent of bone loss (Dionyssiotis et al., 2009). The similar severity of demineralization in the sublesional area was shown between paraplegics and tetraplegics, and the extent of the bone loss may be variable (Demirel et al., 1998;
The duration of paralysis has an inverse relationship with leg percentage-matched BMD and trunk percentage-matched BMD (Clasey et al., 2004). In addition in complete paraplegics, with high (thoracic 4-7) and low (thoracic 8-12) neurological level of injury, upper limbs FM and lower limbs BMD were correlated with the duration of paralysis in total paraplegic group but after investigation according the neurological level of injury this correlation was due to the strong correlation of high paraplegics' lower limbs BMD with the duration of paralysis. The explanation of this strong correlation could possibly lie on higher incidence of standing in the group of low paraplegics and direct effect of loading lower limbs while standing and walking with orthotic equipment. Moreover, the association of the duration of paralysis with parameters below and above the neurological level of injury (upper limbs FM) raises the question of the existence of a hormonal mechanism as an influential regulator in paraplegics' body composition (Dionyssiotis, 2008a; Dionyssiotis et al., 2008b; 2009).
Actually, little is known regarding the nature and time frame of the influence of complete SCI on human skeletal muscle because published data are coming from cross-sectional studies, where different groups with few subjects have been examined at different times, usually in the chronic phase of paralysis. Disuse was thought to be the mechanism responsible for the skeletal muscle atrophy in paraplegics, but muscle fibres following SCI begin to change their functional properties early post injury. Muscle fiber cross-sectional area (CSA) has been suggested to decline from 1 to 17 months after injury and thereafter to reach its nadir. Conversion to type II fibers has been suggested to occur between 4 months and 2 years after injury, resulting in even slow-twitch muscle becoming predominantly fast twitch thereafter (Castro et al., 1999). Metabolic enzymes levels in skeletal muscle might be expected to be reduced after SCI because of inactivation. In support of this contention, succinic dehydrogenase (SDH) activity, a marker of aerobic-oxidative capacity, has been reported to be 47–68% below control values in fibers of tibialis anterior muscle years after
The muscle atrophy in SCI is of central type and depends on the disuse and loss of upper connections of the lower motor neuron, sometimes associated to the loss of anterior horn cells and transinaptic degeneration. The last alteration may be responsible for the denervation changes seen in early stages post SCI. In the later stages (10-17 months post SCI) diffuse muscle atrophy with reduction of the muscle fascicle dimension is associated to fat infiltration and endomysial fibrosis. In all stages post SCI, almost all patients showed myopathic changes, as internal nuclei, fibre degeneration and cytoplasmic vacuolation due
Tsuzuku et al., 1999; Dauty et al., 2000).
injury in support of this contention (Scelsi, 2001).
to lipid accumulation (Scelsi, 2001)
paraplegics trabecular metaphysical-epiphyseal areas of the distal femur and the proximal tibia are the most affected sites (Jiang et al., 2006). A cross-sectional study (Dauty et al., 2000) in SCI subjects demonstrated a significant demineralization at the distal femur (-52%) and the proximal tibia (-70%), respectively.
There is no demineralization of the upper limbs in paraplegics. Studies reported a minor increase of BMD while at the lumbar spine trabecular bone demineralization remains relatively low compared to long bones cortical bone demineralization of (Dauty et al., 2000). Normal (Chantraine et al., 1986; Biering-Sorensen et al., 1988; Kunkel et al., 1993) or even higher than normal values were found (Ogilvie et al., 1993), a phenomenon known as "dissociated hip and spine demineralization" (Leslie, 1993) One reason for preservation of bone mass in the vertebral column is because of its continued weight-bearing function in paraplegics but also lumbar spine arthrosis, bone callus, vertebral fracture, aortic calcification, osteosynthesis material, etc. Degenerative changes in the spine may be the most possible reason to give falsely higher values of BMD (Dauty et al., 2000).
Fig. 3. The picture depicts the analysis of bone mineral density (BMD) in high and low level paraplegics and controls. A statistically significant reduction in total BMD (p<0.001) and lower limbs BMD in body composition compared to able-bodied males was observed. On the contrary, upper limbs BMD was higher in low paraplegics and controls, an unexpected finding explained in the paper of Dionyssiotis et al., 2008b. Diagram modified and translated from Dionyssiotis, 2008a.
paraplegics trabecular metaphysical-epiphyseal areas of the distal femur and the proximal tibia are the most affected sites (Jiang et al., 2006). A cross-sectional study (Dauty et al., 2000) in SCI subjects demonstrated a significant demineralization at the distal femur (-52%) and
There is no demineralization of the upper limbs in paraplegics. Studies reported a minor increase of BMD while at the lumbar spine trabecular bone demineralization remains relatively low compared to long bones cortical bone demineralization of (Dauty et al., 2000). Normal (Chantraine et al., 1986; Biering-Sorensen et al., 1988; Kunkel et al., 1993) or even higher than normal values were found (Ogilvie et al., 1993), a phenomenon known as "dissociated hip and spine demineralization" (Leslie, 1993) One reason for preservation of bone mass in the vertebral column is because of its continued weight-bearing function in paraplegics but also lumbar spine arthrosis, bone callus, vertebral fracture, aortic calcification, osteosynthesis material, etc. Degenerative changes in the spine may be the
**BMD ANALYSIS**
**High paraplegics Low paraplegics**
| doab | 2025-04-07T04:13:03.511752 | 20-4-2021 17:18 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffbd80dd-28a6-46cd-af21-a432f1dad4e5",
"url": "https://mts.intechopen.com/storage/books/602/authors_book/authors_book.pdf",
"author": "",
"title": "Dual Energy X-Ray Absorptiometry",
"publisher": "IntechOpen",
"isbn": "9789533078779",
"section_idx": 1
} |
ffbd80dd-28a6-46cd-af21-a432f1dad4e5.2 | #
**Controls**
**\* p < 0,05 vs control # p < 0,05 vs high**
**\* \* \* \***
Fig. 3. The picture depicts the analysis of bone mineral density (BMD) in high and low level paraplegics and controls. A statistically significant reduction in total BMD (p<0.001) and lower limbs BMD in body composition compared to able-bodied males was observed. On the contrary, upper limbs BMD was higher in low paraplegics and controls, an unexpected
**Lower Limbs BMD Upper Limbs-BMD Total BMD**
finding explained in the paper of Dionyssiotis et al., 2008b. Diagram modified and
translated from Dionyssiotis, 2008a.
**0**
**0.2**
**0.4**
**0.6**
**mean value in grcm-2**
**0.8**
**1**
**1.2**
**1.4**
most possible reason to give falsely higher values of BMD (Dauty et al., 2000).
the proximal tibia (-70%), respectively.
The neurological level of the lesion i.e. the extent of impairment of motor and sensory function is important, because tetraplegics are more likely to lose more bone mass throughout the skeleton than paraplegics (Tsuzuku et al., 1999). In paraplegics legs' BMC was reduced vs. controls, independently of the neurological level of injury and negatively correlated with the duration of paralysis in total paraplegic group, but after investigation according to the neurological level of injury this correlation was due to the strong correlation of high paraplegics' legs BMC with the duration of paralysis, meaning that the neurological level of injury determines the extent of bone loss (Dionyssiotis et al., 2009). The similar severity of demineralization in the sublesional area was shown between paraplegics and tetraplegics, and the extent of the bone loss may be variable (Demirel et al., 1998; Tsuzuku et al., 1999; Dauty et al., 2000).
The duration of paralysis has an inverse relationship with leg percentage-matched BMD and trunk percentage-matched BMD (Clasey et al., 2004). In addition in complete paraplegics, with high (thoracic 4-7) and low (thoracic 8-12) neurological level of injury, upper limbs FM and lower limbs BMD were correlated with the duration of paralysis in total paraplegic group but after investigation according the neurological level of injury this correlation was due to the strong correlation of high paraplegics' lower limbs BMD with the duration of paralysis. The explanation of this strong correlation could possibly lie on higher incidence of standing in the group of low paraplegics and direct effect of loading lower limbs while standing and walking with orthotic equipment. Moreover, the association of the duration of paralysis with parameters below and above the neurological level of injury (upper limbs FM) raises the question of the existence of a hormonal mechanism as an influential regulator in paraplegics' body composition (Dionyssiotis, 2008a; Dionyssiotis et al., 2008b; 2009).
Actually, little is known regarding the nature and time frame of the influence of complete SCI on human skeletal muscle because published data are coming from cross-sectional studies, where different groups with few subjects have been examined at different times, usually in the chronic phase of paralysis. Disuse was thought to be the mechanism responsible for the skeletal muscle atrophy in paraplegics, but muscle fibres following SCI begin to change their functional properties early post injury. Muscle fiber cross-sectional area (CSA) has been suggested to decline from 1 to 17 months after injury and thereafter to reach its nadir. Conversion to type II fibers has been suggested to occur between 4 months and 2 years after injury, resulting in even slow-twitch muscle becoming predominantly fast twitch thereafter (Castro et al., 1999). Metabolic enzymes levels in skeletal muscle might be expected to be reduced after SCI because of inactivation. In support of this contention, succinic dehydrogenase (SDH) activity, a marker of aerobic-oxidative capacity, has been reported to be 47–68% below control values in fibers of tibialis anterior muscle years after injury in support of this contention (Scelsi, 2001).
The muscle atrophy in SCI is of central type and depends on the disuse and loss of upper connections of the lower motor neuron, sometimes associated to the loss of anterior horn cells and transinaptic degeneration. The last alteration may be responsible for the denervation changes seen in early stages post SCI. In the later stages (10-17 months post SCI) diffuse muscle atrophy with reduction of the muscle fascicle dimension is associated to fat infiltration and endomysial fibrosis. In all stages post SCI, almost all patients showed myopathic changes, as internal nuclei, fibre degeneration and cytoplasmic vacuolation due to lipid accumulation (Scelsi, 2001)
Body Composition in Disabilities of Central Nervous System 85
injury in most patients with SCI to develop autonomic dysreflexia. With SCIs above the level of T6, there is reduced SNS outflow and supraspinal control to the splanchnic outflow and the lower-extremity blood vessels while serum leptin levels in men with SCI correlated not only with BMI but also with the neurologic deficit. This finding supports the notion that decentralization of sympathetic nervous activity relieves its inhibitory tone on leptin secretion, because subjects with tetraplegia have a more severe deficit of sympathetic
No significant difference between ambulatory multiple sclerosis (MS) patients and non MS controls in body composition was found despite lower physical activity in ambulatory MS patients (Lambert et al., 2002). In MS subjects there was no significant relation between any of the body composition measures and the level of disability as measured by the Expanded Disability Status Scale (EDSS). Others found no difference in body fat percent between ambulatory MS patients (Formica et al., 1997) and lower physical activity in ambulatory MS patients vs. controls (Ng & Kent-Braun, 1997). A possible explanation for the similar body composition may be lower energy intake in MS individuals who are ambulatory and greater energy cost of physical activity (walking) in MS than it is with non MS controls (Lambert et
A significant inverse relation between free fat mass (FFM) and EDSS score when ambulatory and non ambulatory MS subjects were combined was found (Formica et al., 1997). On the contrary others without including non ambulatory subjects did not find a significant inverse relation between FFM percent and EDSS score (Lambert et al., 2002). It would seem apparent that ambulatory patients with MS and controls would strengthen the inverse
The finding of no relation between EDSS score and body fat percent (Lambert et al., 2002) fits well with studies which found no significant relation between the level of physical activity, and the level of disability in individuals with MS (Ng & Kent-Braun, 1997) because MS would likely have a much greater effect on physical activity than on energy intake. According to these findings it appears that the level of disability of ambulatory individuals with MS does not predict body composition. This suggests that a significant level of disability does not force these individuals to be physically inactive and does not result in a greater body fat content. There are many detrimental manifestations of excess body fat, such as hyperlipidemia, insulin resistance, and type II diabetes (Lambert et al., 2002). The largest component of FFM is muscle mass (Lohman, 1986). If muscle mass is lower in individuals with MS than in controls, it may also contribute to the impaired ability to ambulate and perform other activities of daily living. Muscle fiber size from biopsy specimens of the tibialis anterior were 26% smaller than specimens from control subjects (Kent-Braun et al., 1997). Thus, at least for this small muscle, muscle mass was lower in MS. This relationship may not hold for other muscle groups or for whole-body muscle mass (Lambert et al., 2002). Another reason for skeletal muscle alterations is glucocorticoid usage. The prolonged duration of glucocorticoid causes catabolism of skeletal muscle. Decreased amino acid transport into muscle and increased glutamine synthesis activity with resultant muscle atrophy are some of the concomitant effects of glucocorticoid use on skeletal muscle.
nervous activity (Wang et al., 2005).
relation between FFM and EDSS score.
**3.2 Multiple sclerosis**
al., 2002).
It is evident that other co-factors as spasticity and microvascular damage, contribute to the induction of the marked morphological and enzyme histochemical changes seen in the paralyzed skeletal muscle (Scelsi, 2001). Small fibers, predominantly fast-twitch muscle, and low mitochondrial content have been reported years after injury in cross-sectional studies. These data have been interpreted to suggest that human skeletal muscle shows plasticity (Castro et al., 1999).
On the contrary, force loss during repetitive contractions evoked by surface electrical stimulation (ES) of skeletal muscle in humans does not appear to be altered within a few months of injury (Shields, 1995) but it is greater a year or more after SCI (Hillegass & Dudley, unpublished observations). The greater fatigue, when evident, was partially attributed to lower metabolic enzyme levels (Scelsi, 2001).
Muscular loading of the bones has been thought to play a role in the maintenance of bone density (de Bruin et al., 1999; Dionyssiotis et al., 2011d). However, the ability to stand or ambulate itself does not improve BMD or prevent osteoporosis after SCI.
Controversial results have also been reported regarding the effect of spasticity on BMD in SCI paraplegics. A cross-sectional study of 41 SCI paraplegics reported less reduction of BMD in the spastic paraplegics SCI patients compared to the flaccid paraplegic SCI patients (Demirel et al., 1998). Others reported that spasticity may be protective against bone loss in SCI patients, however, without any preserving effect in the tibia (Dionyssiotis et al., 2011a; Eser et al., 2005). A possible explanation for that could lie in the fact paraplegics to be above thoracic (T)12 level with various degrees of spasticity according to the Ashworth scale. In addition, muscle spasms affecting the lower leg would mainly be extension spasms resulting in plantar flexion thus creating little resistance to the contracting muscles. Furthermore, the measuring sites of the tibia did not include any muscle insertions of either the knee or the ankle extensor muscles (Dionyssiotis et al., 2011a, 2011d). Other investigators also have not been able to establish a correlation between BMD and muscle spasticity (Lofvenmark et al., 2009).
The hormone leptin is secreted by fat cells and helps regulate body weight and energy consumption (Fruhbeck et al., 1998). The percentage of fat in people is positively correlated with the amount of leptin in the circulation (Maffei et al., 1995). In SCI, when compared with healthy subjects, higher levels of leptin have been found, possibly due to greater fat tissue storage (Bauman et al., 1996). Leptin activates the sympathetic nervous system (SNS) through a central administration. The disruption of the sympathetic nervous system i.e. in tetraplegia and high level paraplegia may modify the secretion and activity of the leptin, because the sympathetic preganglionic neurons become atrophic in these subgroups (Elias et al., 1998; Correia et al., 2001) leading to disturbed irritation from leptin below the neurological level of injury. In addition, extensive obesity is known to reduce lipolytic sensitivity (Haque et al., 1999; Horowitz et al., 1999, 2000).
In high level spinal cord injuries there is a disorder of the autonomic nervous system and combined to the fact that the hormone leptin activates the sympathetic nervous system through central control it could be suggested that "the closure of paths" of the central nervous system disrupts the effect of leptin and possibly increases the risk of obesity in SCI subjects with high-level injury (Krassioukov et al., 1999; Jeon et al., 2003). However, after separation of SCI subjects into those with an injury above or below Thoracic (T) 6, leptin levels were significantly higher in the former group. T6 appears to be the lowest level of injury in most patients with SCI to develop autonomic dysreflexia. With SCIs above the level of T6, there is reduced SNS outflow and supraspinal control to the splanchnic outflow and the lower-extremity blood vessels while serum leptin levels in men with SCI correlated not only with BMI but also with the neurologic deficit. This finding supports the notion that decentralization of sympathetic nervous activity relieves its inhibitory tone on leptin secretion, because subjects with tetraplegia have a more severe deficit of sympathetic nervous activity (Wang et al., 2005).
#### **3.2 Multiple sclerosis**
84 Dual Energy X-Ray Absorptiometry
It is evident that other co-factors as spasticity and microvascular damage, contribute to the induction of the marked morphological and enzyme histochemical changes seen in the paralyzed skeletal muscle (Scelsi, 2001). Small fibers, predominantly fast-twitch muscle, and low mitochondrial content have been reported years after injury in cross-sectional studies. These data have been interpreted to suggest that human skeletal muscle shows plasticity
On the contrary, force loss during repetitive contractions evoked by surface electrical stimulation (ES) of skeletal muscle in humans does not appear to be altered within a few months of injury (Shields, 1995) but it is greater a year or more after SCI (Hillegass & Dudley, unpublished observations). The greater fatigue, when evident, was partially
Muscular loading of the bones has been thought to play a role in the maintenance of bone density (de Bruin et al., 1999; Dionyssiotis et al., 2011d). However, the ability to stand or
Controversial results have also been reported regarding the effect of spasticity on BMD in SCI paraplegics. A cross-sectional study of 41 SCI paraplegics reported less reduction of BMD in the spastic paraplegics SCI patients compared to the flaccid paraplegic SCI patients (Demirel et al., 1998). Others reported that spasticity may be protective against bone loss in SCI patients, however, without any preserving effect in the tibia (Dionyssiotis et al., 2011a; Eser et al., 2005). A possible explanation for that could lie in the fact paraplegics to be above thoracic (T)12 level with various degrees of spasticity according to the Ashworth scale. In addition, muscle spasms affecting the lower leg would mainly be extension spasms resulting in plantar flexion thus creating little resistance to the contracting muscles. Furthermore, the measuring sites of the tibia did not include any muscle insertions of either the knee or the ankle extensor muscles (Dionyssiotis et al., 2011a, 2011d). Other investigators also have not been able to establish a
The hormone leptin is secreted by fat cells and helps regulate body weight and energy consumption (Fruhbeck et al., 1998). The percentage of fat in people is positively correlated with the amount of leptin in the circulation (Maffei et al., 1995). In SCI, when compared with healthy subjects, higher levels of leptin have been found, possibly due to greater fat tissue storage (Bauman et al., 1996). Leptin activates the sympathetic nervous system (SNS) through a central administration. The disruption of the sympathetic nervous system i.e. in tetraplegia and high level paraplegia may modify the secretion and activity of the leptin, because the sympathetic preganglionic neurons become atrophic in these subgroups (Elias et al., 1998; Correia et al., 2001) leading to disturbed irritation from leptin below the neurological level of injury. In addition, extensive obesity is known to reduce lipolytic
In high level spinal cord injuries there is a disorder of the autonomic nervous system and combined to the fact that the hormone leptin activates the sympathetic nervous system through central control it could be suggested that "the closure of paths" of the central nervous system disrupts the effect of leptin and possibly increases the risk of obesity in SCI subjects with high-level injury (Krassioukov et al., 1999; Jeon et al., 2003). However, after separation of SCI subjects into those with an injury above or below Thoracic (T) 6, leptin levels were significantly higher in the former group. T6 appears to be the lowest level of
attributed to lower metabolic enzyme levels (Scelsi, 2001).
ambulate itself does not improve BMD or prevent osteoporosis after SCI.
correlation between BMD and muscle spasticity (Lofvenmark et al., 2009).
sensitivity (Haque et al., 1999; Horowitz et al., 1999, 2000).
(Castro et al., 1999).
No significant difference between ambulatory multiple sclerosis (MS) patients and non MS controls in body composition was found despite lower physical activity in ambulatory MS patients (Lambert et al., 2002). In MS subjects there was no significant relation between any of the body composition measures and the level of disability as measured by the Expanded Disability Status Scale (EDSS). Others found no difference in body fat percent between ambulatory MS patients (Formica et al., 1997) and lower physical activity in ambulatory MS patients vs. controls (Ng & Kent-Braun, 1997). A possible explanation for the similar body composition may be lower energy intake in MS individuals who are ambulatory and greater energy cost of physical activity (walking) in MS than it is with non MS controls (Lambert et al., 2002).
A significant inverse relation between free fat mass (FFM) and EDSS score when ambulatory and non ambulatory MS subjects were combined was found (Formica et al., 1997). On the contrary others without including non ambulatory subjects did not find a significant inverse relation between FFM percent and EDSS score (Lambert et al., 2002). It would seem apparent that ambulatory patients with MS and controls would strengthen the inverse relation between FFM and EDSS score.
The finding of no relation between EDSS score and body fat percent (Lambert et al., 2002) fits well with studies which found no significant relation between the level of physical activity, and the level of disability in individuals with MS (Ng & Kent-Braun, 1997) because MS would likely have a much greater effect on physical activity than on energy intake. According to these findings it appears that the level of disability of ambulatory individuals with MS does not predict body composition. This suggests that a significant level of disability does not force these individuals to be physically inactive and does not result in a greater body fat content. There are many detrimental manifestations of excess body fat, such as hyperlipidemia, insulin resistance, and type II diabetes (Lambert et al., 2002). The largest component of FFM is muscle mass (Lohman, 1986). If muscle mass is lower in individuals with MS than in controls, it may also contribute to the impaired ability to ambulate and perform other activities of daily living. Muscle fiber size from biopsy specimens of the tibialis anterior were 26% smaller than specimens from control subjects (Kent-Braun et al., 1997). Thus, at least for this small muscle, muscle mass was lower in MS. This relationship may not hold for other muscle groups or for whole-body muscle mass (Lambert et al., 2002).
Another reason for skeletal muscle alterations is glucocorticoid usage. The prolonged duration of glucocorticoid causes catabolism of skeletal muscle. Decreased amino acid transport into muscle and increased glutamine synthesis activity with resultant muscle atrophy are some of the concomitant effects of glucocorticoid use on skeletal muscle.
Body Composition in Disabilities of Central Nervous System 87
percentile for age, even if most of the CP children had a low height and weight for age. In female subjects anthropometric measurements were highly correlated with measures of body fatness. Measuring fat by 18O dilution a hydration factor of 0.73 was assumed for FFM. A possible increase in the hydration factor would diminish measured FFM meaning that body fat appears increased. Moreover muscle spasms and spasticity in CP subjects deplete body glycogen. If glycogen is reduced the intracellular water would be reduced and the ratio extracellular water/total body water would increase. The same could result with a loss
Other important issues according alterations of body composition are the completeness of lesions (an absence of sensory or motor function below the neurological level, including the lowest sacral segment), because body composition seems to be worst than subjects with incomplete lesions (partial preservation of motor and/or sensory function below the neurological level, including the lowest sacral segment) (Sabo et al., 1991; Demirel et al., 1998; Garland et al., 1992) and aging which contributes to major alterations of body
In disabled subjects the most important issue according to body composition is how to promote optimal body weight to reduce risk of diseases such as coronary heart disease, noninsulin dependent diabetes mellitus, lipid abnormalities and fractures because of bone loss. Dietary changes, individualized physical activity programs and medication should be taken in mind in therapy when we deal with this subgroup of subjects. However, selfmanagement of dietary changes to improve weight control and disease should be the case, which means they need to follow diets with lower energy intake and at the same time to eat
We need to take in mind that healthy BMI values often underestimate body fat and may mask the adiposity and spasticity did not defend skeletal muscle mass and bone, supporting the concept that in neurologic disabilities the myopathic muscle could not recognize correctly the stimulation because of the neurogenic injury. Moreover, disabled subjects mostly transfer much of the weight-bearing demands of daily activities to their upper extremities reducing the weight-bearing of the affected paralyzed muscles triggering a cycle of added muscle atrophy which interacts with the continuous catabolic action caused by the neurogenic factor. Finally, an irreversible (once established) decline in bone mineral density, bone mineral content as well as geometric characteristics of bone is expected and the duration of lesion-injury is positively correlated with the degree of
Further research about body composition is needed in all physical disabilities and more longitudinal studies to quantitate and monitor body composition changes and to modify our therapeutic interventions. However, prevention rather than treatment may have the greatest potential to alleviate these major complications. Therapies should focus on how to perform weight bearing, standing or therapeutically walking activities early in the rehabilitation
of body cell mass or an increase in the hydration factor (Bandini et al., 1991).
regularly foods rich in nutrients (Groah et al., 2009).
program to gain benefits according to muscles and bones.
**4. Conclusions**
composition.
bone loss.
Endogenous glucocorticoid excess also produces generalized osteoporosis, most prevalent in trabecular-rich skeletal regions (Formica et al., 1997).
Beside corticosteroids, immunomodulatory, antiepileptic and antidepressant drugs usually used in individuals with MS, high incidence of vitamin D deficiency, molecular mechanisms and disuse-loss of mechanical stimuli in bone have an effect on bone integrity (most believe that immobilization of these patients is a minor factor in the etiology of osteoporosis) (Dionyssiotis, 2011).
#### **3.3 Stroke**
Longitudinal studies of body composition in the elderly have shown that body cell mass decreases with age and is lower in women than in men (Steen et al., 1985). A decline in body fat in both the dependent and independent groups nine weeks after admission was found, indicating consumption of energy stores. In contrast, the change of body cell mass between admission and after 9 weeks was significantly greater in the dependent patients compared with the independent (Unosson et al., 1994). Immobilized individuals lose muscle mass irrespective of nutritional intake because of reduced synthesis of proteins, while the rate of breakdown of proteins is unchanged (Schonheyder et al., 1954). During the recovery period the stroke patients seemed to break down body fat to compensate for energy needs, independent of their functional condition. However, change of body cell mass appeared to relate to the patients' functional condition after stroke (Unosson et al., 1994).
A study in 35 stroke patients compared the body composition, including lean tissue mass, fat tissue mass, and bone mineral content, of the paretic leg with that of the non affected leg in patients with stroke and evaluated the effects of time since stroke, spasticity, and motor recovery on the body composition specifically within the first year after stroke found lean tissue mass and bone mineral content of the paretic side to be significantly lower than those of the non affected side; a significant correlation was found between the lean tissue mass and bone mineral content of both the paretic and non affected legs after adjusting for age and weight. On the contrary bone mineral content and lean tissue mass of both the paretic and non affected sides were negatively correlated with time since stroke in patients with stroke for less than 1 year and a higher lean tissue mass and bone mineral content were found in patients with moderate to high spasticity in comparison with patients with low or no spasticity (Celik et al., 2008).
#### **3.4 Cerebral palsy**
Bone mineralization in children with CP has been found lower (bone-mineral values for the total body and total proximal femur) than sex- and age-matched able bodied children. This is illustrated by the BMC Z – scores determined at each skeletal site. The factors that contribute to low bone mineralization include genetic, hormonal, and nutritional problems (especially calcium and vitamin D) and weight-bearing physical activity, oral-motor dysfunction and anticonvulsant medication (Henderson et al., 1995).
Free fat mass (FFM) in cerebral palsy subjects was found significantly lower than that in a normal adolescent population. In 60% of the studied population body fat exceeded the 90th
percentile for age, even if most of the CP children had a low height and weight for age. In female subjects anthropometric measurements were highly correlated with measures of body fatness. Measuring fat by 18O dilution a hydration factor of 0.73 was assumed for FFM. A possible increase in the hydration factor would diminish measured FFM meaning that body fat appears increased. Moreover muscle spasms and spasticity in CP subjects deplete body glycogen. If glycogen is reduced the intracellular water would be reduced and the ratio extracellular water/total body water would increase. The same could result with a loss of body cell mass or an increase in the hydration factor (Bandini et al., 1991).
#### **4. Conclusions**
86 Dual Energy X-Ray Absorptiometry
Endogenous glucocorticoid excess also produces generalized osteoporosis, most prevalent
Beside corticosteroids, immunomodulatory, antiepileptic and antidepressant drugs usually used in individuals with MS, high incidence of vitamin D deficiency, molecular mechanisms and disuse-loss of mechanical stimuli in bone have an effect on bone integrity (most believe that immobilization of these patients is a minor factor in the etiology of osteoporosis)
Longitudinal studies of body composition in the elderly have shown that body cell mass decreases with age and is lower in women than in men (Steen et al., 1985). A decline in body fat in both the dependent and independent groups nine weeks after admission was found, indicating consumption of energy stores. In contrast, the change of body cell mass between admission and after 9 weeks was significantly greater in the dependent patients compared with the independent (Unosson et al., 1994). Immobilized individuals lose muscle mass irrespective of nutritional intake because of reduced synthesis of proteins, while the rate of breakdown of proteins is unchanged (Schonheyder et al., 1954). During the recovery period the stroke patients seemed to break down body fat to compensate for energy needs, independent of their functional condition. However, change of body cell mass appeared to
A study in 35 stroke patients compared the body composition, including lean tissue mass, fat tissue mass, and bone mineral content, of the paretic leg with that of the non affected leg in patients with stroke and evaluated the effects of time since stroke, spasticity, and motor recovery on the body composition specifically within the first year after stroke found lean tissue mass and bone mineral content of the paretic side to be significantly lower than those of the non affected side; a significant correlation was found between the lean tissue mass and bone mineral content of both the paretic and non affected legs after adjusting for age and weight. On the contrary bone mineral content and lean tissue mass of both the paretic and non affected sides were negatively correlated with time since stroke in patients with stroke for less than 1 year and a higher lean tissue mass and bone mineral content were found in patients with moderate to high spasticity in comparison with patients with low or
Bone mineralization in children with CP has been found lower (bone-mineral values for the total body and total proximal femur) than sex- and age-matched able bodied children. This is illustrated by the BMC Z – scores determined at each skeletal site. The factors that contribute to low bone mineralization include genetic, hormonal, and nutritional problems (especially calcium and vitamin D) and weight-bearing physical activity, oral-motor
Free fat mass (FFM) in cerebral palsy subjects was found significantly lower than that in a normal adolescent population. In 60% of the studied population body fat exceeded the 90th
dysfunction and anticonvulsant medication (Henderson et al., 1995).
relate to the patients' functional condition after stroke (Unosson et al., 1994).
in trabecular-rich skeletal regions (Formica et al., 1997).
(Dionyssiotis, 2011).
no spasticity (Celik et al., 2008).
**3.4 Cerebral palsy**
**3.3 Stroke**
Other important issues according alterations of body composition are the completeness of lesions (an absence of sensory or motor function below the neurological level, including the lowest sacral segment), because body composition seems to be worst than subjects with incomplete lesions (partial preservation of motor and/or sensory function below the neurological level, including the lowest sacral segment) (Sabo et al., 1991; Demirel et al., 1998; Garland et al., 1992) and aging which contributes to major alterations of body composition.
In disabled subjects the most important issue according to body composition is how to promote optimal body weight to reduce risk of diseases such as coronary heart disease, noninsulin dependent diabetes mellitus, lipid abnormalities and fractures because of bone loss. Dietary changes, individualized physical activity programs and medication should be taken in mind in therapy when we deal with this subgroup of subjects. However, selfmanagement of dietary changes to improve weight control and disease should be the case, which means they need to follow diets with lower energy intake and at the same time to eat regularly foods rich in nutrients (Groah et al., 2009).
We need to take in mind that healthy BMI values often underestimate body fat and may mask the adiposity and spasticity did not defend skeletal muscle mass and bone, supporting the concept that in neurologic disabilities the myopathic muscle could not recognize correctly the stimulation because of the neurogenic injury. Moreover, disabled subjects mostly transfer much of the weight-bearing demands of daily activities to their upper extremities reducing the weight-bearing of the affected paralyzed muscles triggering a cycle of added muscle atrophy which interacts with the continuous catabolic action caused by the neurogenic factor. Finally, an irreversible (once established) decline in bone mineral density, bone mineral content as well as geometric characteristics of bone is expected and the duration of lesion-injury is positively correlated with the degree of bone loss.
Further research about body composition is needed in all physical disabilities and more longitudinal studies to quantitate and monitor body composition changes and to modify our therapeutic interventions. However, prevention rather than treatment may have the greatest potential to alleviate these major complications. Therapies should focus on how to perform weight bearing, standing or therapeutically walking activities early in the rehabilitation program to gain benefits according to muscles and bones.
Body Composition in Disabilities of Central Nervous System 89
Coupaud S, McLean AN, Allan DB. Role of peripheral quantitative computed tomography
Dauty M, Perrouin Verbe B, Maugars Y, Dubois C, Mathe JF. Supralesional and sublesional bone mineral density in spinal cord-injured patients. Bone. 2000;27:305-9. Demirel G, Yilmaz H, Paker N, Onel S. Osteoporosis after spinal cord injury. Spinal Cord.
Dionyssiotis Y, Lyritis GP, Papaioannou N, Papagelopoulos P, Thomaides T. Influence of
Dionyssiotis Y, Trovas G, Galanos A, Raptou P, Papaioannou N, Papagelopoulos P,
Dionyssiotis Y. Changes in bone density and strength of the tibia and alterations of lean and
Dionyssiotis, Y. Bone loss and fractures in multiple sclerosis: focus on epidemiologic and
Dionyssiotis, Y. Spinal cord injury-related bone impairment and fractures: an update on
Dionyssiotis, Y, Lyritis GP, Mavrogenis AF, Papagelopoulos PJ. Factors influencing bone
Elias CF, Lee C, Kelly J, Aschkenasi C, Ahima RS, Couceyro PR, Kuhar MJ, Saper CB,
Ferretti J.L., Cointry G.R., Capozza R.F., Zanchetta J.R. Dual energy X-ray absorptiometry.
T.J.Peters (eds),Greenwich Medical Media, Ltd., London, 2001; p.451-458. Forbes GB, Simon W, Amatruda JM. Is bioimpedance a good predictor of body-composition
Forbes GB. Human body composition: growth, aging, nutrition, and activity. New York:
Formica CA, Cosman F, Nieves J, Herbert J, Lindsay R. Reduced bone mass and fat-free
Frey-Rindova P, de Bruin ED, Stussi E, Dambacher MA, Dietz V. Bone mineral density in
glucocorticoid Use. Calcif Tissue Int. 1997 Aug;61(2):129-33.
of the Musculoskeletal System, University of Athens, Athens, 2008a. Dionyssiotis Υ, Petropoulou Κ, Rapidi CA, Papagelopoulos PJ, Papaioannou N, Galanos A,
http://cirrie.buffalo.edu/encyclopedia/en/article/340/
physiopathological features. Int J Gen Med. 2011b; 4: 505-9.
loss in paraplegia. Hippokratia. 2011a ; 15(1):54-9.
Clinical Densitometry. 2008b;11: 437-43.
Interact. 2011c; 11(3):257-65.
cord. Neuron. 1998;21:1375-85.
change? Am J Clin Nutr 1992;56:4-6.
Springer-Verlag; 1987.
Oct;38(10):989-95.
Dev. 2009;46(8):1037-44.
1998;36:8
in identifying disuse osteoporosis in paraplegia. Skeletal Radiol. 2009
neurological level of injury in bones, muscles, and fat in paraplegia. J Rehabil Res
Petropoulou K, Lyritis GP. Bone loss and mechanical properties of tibia in spinal cord injured men. J Musculoskelet Neuronal Interact. 2007 Jan-Mar;7(1):62-8. Dionyssiotis Y. (2011d). Bone Loss in Spinal Cord Injury and Multiple Sclerosis. In: JH Stone,
M Blouin, editors. International Encyclopedia of Rehabilitation, av. online:
fat mass in chronic paraplegic men. Doctoral Dissertation Laboratory for Research
Papadaki P, and Lyritis GP. Body Composition in Paraplegic Men. Journal of
epidemiology and physiopathological mechanisms. J Musculoskelet Neuronal
Elmquist JK. Leptin activates hypothalamic CART neurons projecting to the spinal
Skeletal Muscle: Pathology, Diagnosis and Management of Disease. V.R.Preedy,
mass in women with multiple sclerosis: effects of ambulatory status and
upper and lower extremities during 12 months after spinal cord injury measured by peripheral quantitative computed tomography. Spinal Cord. 2000;38:26–32.
#### **5. References**
Bandini LG, Schoeller DA, Fukagawa NK, Wykes LJ, Dietz WH. Body composition and
Bauman WA, Spungen AM, Raza M, Rothstein J, Zhang RL, Zhong YG, Tsuruta M, Shahidi
Bauman WA, Spungen AM, Zhong YG, Mobbs CV. Plasma leptin is directly related to body adiposity in subjects with spinal cord injury. Horm Metab Res. 1996;28:732-6. Bauman WA, Spungen AM. 1994 Disorders of carbohydrate and lipid metabolism in
Bender R, Trautner C, Spraul M, Berger M. Assessment of excess mortality in obesity. Am J
Biering-Sorensen F, Bohr HH, Schaadt OP. Longitudinal study of bone mineral content in
Buchholz AC, Bugaresti JM. A review of body mass index and waist circumference as
Castro MJ, Apple DF Jr, Hillegass EA, and Dudley GA. Influence of complete spinal cord
Castro MJ, Apple DF Jr, Staron RS, Campos GE, Dudley GA. Influence of complete spinal
Celik B, Ones K, Ince N. Body composition after stroke. Int J Rehabil Res. 2008 Mar;31(1):93-
Chad KE, McKay HA, Zello GA, Bailey DA, Faulkner RA, Snyder RE. Body composition in
Chantraine A, Nusgens B, Lapiere CM. Bone remodelling during the development of
Chow YW, Inman C, Pollintine P, Sharp CA, Haddaway MJ, el Masry W, Davie
Clarys JP, Martin AD, Drinkwater DT. Gross tissue weights in the human body by cadaver
Clasey JL, Janowiak AL, Gater DR Relationship between regional bone density
Correia ML, Morgan DA, Mitchell JL, Sivitz WI, Mark AL, Haynes WG. Role of
osteoporosis in paraplegia. Calcif Tissue Int. 1986;38:323-7.
energy expenditure in adolescents with cerebral palsy or myelodysplasia.Pediatr
R, Pierson RN Jr, Wang J, et al. Coronary artery disease: metabolic risk factors and latent disease in individuals with paraplegia. Mt Sinai J Med. 1992 Mar;59(2):163-8.
veterans with paraplegia or quadriplegia: A model of premature aging. Metabolism
the lumbar spine, the forearm and the lower extremities after spinal cord injury.
markers of obesity and coronary heart disease risk in persons with chronic spinal
injury on skeletal muscle cross-sectional area within the first 6 months of injury.
cord injury on skeletal muscle within 6 mo of injury. J Appl Physiol. 1999b;86:350-8.
nutritionally adequate ambulatory and non-ambulatory children with cerebral palsy and a healthy reference group. Dev Med Child Neurol. 2000 May;42(5):334-9.
MW.Ultrasound bone densitometry and dual energy X-ray absorptiometry in patients with spinal cord injury: a cross-sectional study. Spinal Cord. 1996
measurements and the time since injury in adults with spinal cord injuries. Arch
corticotrophin-releasing factor in effects of leptin on sympathetic nerve activity and
**5. References**
Res. 1991 Jan;29(1):70-7.
Epidemiol. 1998;147:42-8.
Eur J Clin Invest. 1990; 20:330-5.
cord injury. Spinal Cord. 2005;43:513-8.
Eur J Appl Physiol 80: 373–378, 1999a.
dissection. Hum Biol. 1984;56:459-73.
arterial pressure. Hypertension. 2001;38:384-8.
Phys Med Rehabil. 2004;85:59–64
.43: 749.756.
6.
Dec;34(12):736-41.
Body Composition in Disabilities of Central Nervous System 91
Kent-Braun JA, Ng AV, Castro M, et al. Strength, skeletal muscle composition and enzyme
Kent-Braun JA, Sharma KR, Weiner MW, Miller RG. Effects of exercise on muscle activation
Krupp, LB, Serafin DJ, Christodoulou C. Multiple sclerosis-associated fatigue. Expert Rev
Kunkel CF, Scremin AM, Eisenberg B, Garcia JF, Roberts S, Martinez S. Effect of ''standing''
Kuperminc MN, Stevenson RD. Growth and nutrition disorders in children with cerebral
LaForgia J, Dollman J, Dale MJ, Withers RT, Hill AM. Validation of DXA body composition estimates in obese men and women. Obesity (Silver Spring). 2009;17:821-6. Lambert CP, Archer RL, Evans WJ. Body composition in ambulatory women with multiple
Laskey MA. Dual-energy X-ray absorptiometry and body composition. Nutrition.1996
Lazo MG, Shirazi P, Sam M, Giobbie-Hurder A, Blacconiere MJ, Muppidi M. Osteoporosis and risk of fracture in men with spinal cord injury. Spinal Cord. 2001;39:208-14. Leslie WD, Nance PW. Dissociated hip and spine demineralization: a specific finding in
Liu LF, Roberts R, Moyer-Mileur L, Samson-Fang L. Determination of body composition in
Lohman, TG.; Roche, AF.; Martorell, R. Anthropometric standardization reference manual.
Maffei M, Halaas J, Ravussin E, Pratley RE, Lee GH, Zhang Y, Fei H, Kim S, Lallone R,
Mahon AK, Flynn MG, Iglay HB, Stewart LK, Johnson CA, McFarlin BK, Campbell WW.
women: comparison of methods. J Nutr Health Aging. 2007;11:203-13. Maimoun L, Fattal C, Micallef JP, Peruchon E, Rabischong P. Bone loss in spinal cordinjured patients: from physiopathology to therapy. Spinal Cord. 2006;44:203-10.
children with cerebral palsy: bioelectrical impedance analysis and anthropometry vs dual-energy x-ray absorptiometry. J Am Diet Assoc. 2005 May;105(5):794-7. Lohman TG. Applicability of body composition techniques and constants for children and
youth. In: Pandolf KB, editor. Exercise and sport sciences reviews. Vol 14. New
Ranganathan S, et al. Leptin levels in human and rodent: measurement of plasma leptin and ob RNA in obese and weight-reduced subjects. Nat Med. 1995;1:1155-
Measurement of body composition changes with weight loss in postmenopausal
sympathetic preganglionic neurons after spinal cord injury. Spinal Cord. 1999;37:6-
on spasticity, contracture, and osteoporosis in paralyzed males. Arch Phys Med
activity in multiple sclerosis. J Appl Physiol 1997;83:1998-2004.
13.
Neurother. 2010;10:1437-47.
palsy. Dev Disabil Res Rev. 2008;14(2):137-46.
sclerosis. Arch Phys Med Rehabil 2002;83:1559-61.
spinal cord injury. Arch Phys Med Rehabil. 1993; 74:960-4.
Rehabil. 1993;74:73–8.
Jan;12(1):45-51.
61.
York: Macmillan; 1986. p 325-57.
Human Kinetics Books; Champaign: 1988
and metabolism in multiple sclerosis.Muscle Nerve 1994;17:1162-9. Kocina P. Body composition of spinal cord injured adults. Sports Medicine. 1997; 23:48-60. Krassioukov AV, Bunge RP, Pucket WR, Bygrave MA. The changes in human spinal
Frisancho RA. New norms of upper limb fat and muscle areas for assessment of nutritional
Fruhbeck G, Jebb SA, Prentice AM. Leptin: physiology and pathophysiology. Clin Physiol.
Garland DE, Stewart CA, Adkins RH, Hu SS, Rosen C, Liotta FJ, Weinstein DA. 1992.
Gorgey AS, Dudley GA. Skeletal muscle atrophy and increased intramuscular fat after
Groah SL, Nash MS, Ljungberg IH, Libin A, Hamm LF, Ward E, Burns PA, Enfield G.
Hafer-Macko CE, Ryan AS, Ivey FM, Macko RF. Skeletal muscle changes after hemiparetic
Haque MS, Minokoshi Y, Hamai M, Iwai M, Horiuchi M, Shimazu T. Role of the
Henderson RC, Lin PP, Greene WB. (1995). Bone-mineral density in children and
Heymsfield SB, Wang J, Heshka S, Kehayias JJ, Pierson RN. Dual-photon absorptiometry:
Hildreth HG, Johnson RK, Goran MI, Contompasis SH. Body composition in adults with
Horowitz JF, Coppack SW, Paramore D, Cryer PE, Zhao G, Klein S. Effect of short-term
Horowitz JF, Klein S. Whole body and abdominal lipolytic sensitivity to epinephrine is
Ingemann-Hansen T, Halkjaer-Kristensen J. Lean and fat component of the human thigh: the
Jeon JY, Steadward RD, Wheeler GD, Bell G, McCargar L, Harber V. Intact sympathetic
Jiang SD, Dai LY, Jiang LS. Osteoporosis after spinal cord injury. Osteoporos Int.
Jones LM, Goulding A, Gerrard DF. DEXA: a practical and accurate tool to demonstrate total
with spinal cord injury. J Clin Endocrinol Metab. 2003;88:402-7.
established methods. Am J Clin Nutr. 1989 Jun;49(6):1283-9.
technique. Am J Clin Nutr. 1997 Dec;66(6):1436-42.
Nutrient intake and body habitus after spinal cord injury: an analysis by sex and
stroke and potential beneficial effects of exercise intervention strategies. J Rehabil
sympathetic nervous system and insulin in enhancing glucose uptake in peripheral tissues after intrahypothalamic injection of leptin in rats. Diabetes. 1999;48:1706-12.
adolescents who have spastic cerebral palsy.*Journal of Bone and Joint Surgery* 77A:
comparison of bone mineral and soft tissue mass measurements in vivo with
cerebral palsy by dual-energy X-ray absorptiometry, bioelectrical impedance analysis, and skinfold anthropometry compared with the 18O isotope-dilution
fasting on lipid kinetics in lean and obese women. Am J Physiol. 1999;276:E278-
suppressed in upper body obese women. Am J Physiol Endocrinol Metab.
effects of immobilization in plaster and subsequent physical training. *Scand J*
nervous system is required for leptin effects on resting metabolic rate in people
and regional bone loss, lean tissue loss and fat mass gain in paraplegia. Spinal
Osteoporosis after spinal cord injury. J Orthop Res .10 :371.378.
incomplete spinal cord injury. Spinal Cord 2007;45(4):304–309.
level of injury. J Spinal Cord Med. 2009;32:25-33.
status. Am J Clin Nutr 1981;34:2540–2545.
1998;18:399-419.
Res Dev. 2008;45(2):261-72.
1671–81.
84.
2000;278:E1144-52.
2006;17:180-92.
Cord. 1998;36:637-40
*Rehabil Med.* 1977;9:67–72
Body Composition in Disabilities of Central Nervous System 93
Samson-Fang LJ, Stevenson RD. Identification of malnutrition in children with cerebral
Scelsi R. Skeletal muscle pathology after spinal cord injury. Basic Appl Myol. 2001;11:75-
Schonheyder F, Heilskov NCS, Olesen K. Isotopic studies on the mechanism of negative
Seidell JC, Verschuren WM, van Leer EM, Kromhout D. Overweight, underweight, and
Shields RK, Dudley-Javoroski S. Musculoskeletal adaptations in chronic spinal cord injury:
Shields RK. Muscular, skeletal, and neural adaptations following spinal cord injury. J
Sioka C, Fotopoulos A, Georgiou A, Papakonstantinou S, Pelidou SH, Kyritsis AP, Kalef-
Slaughter MH, Lohman TG, Boileau RA, et al. Skinfold equations for estimation of body
Smeltzer SC, Zimmerman V, Capriotti T. Osteoporosis risk and low bone mineral density in women with physical disabilities. Arch Phys Med Rehabil. 2005;86:582-6. Spungen AM, Adkins RH, Stewart CA, Wang J, Pierson RN Jr, Waters RL, Bauman WA.
Stallings VA, Cronk CE, Zemel BS, Charney EB. Body composition in children with spastic
Steen B, Lundgren BK, Isaksson B. Body composition at age 70, 75, 79, and 81 years: a
Szollar SM, Martin EM, Parthemore JG, Sartoris DJ, Deftos LJ. Densitometric patterns of spinal cord injury associated bone loss. Spinal Cord. 1997 Jun;35(6):374-82. Tsuzuku S, Ikegami Y, Yabe K. Bone mineral density differences between paraplegic and quadriplegic patients: a cross-sectional study. Spinal Cord. 1999; 37:358-61. Uebelhart D, Demiaux-Domenech B, Roth M, Chantraine A. Bone metabolism in spinal cord
Unosson M, Ek AC, Bjurulf P, von Schenck H, Larsson J. Feeding dependence and
Van Der Ploeg GE, Withers RT, Laforgia J. Percent body fat via DEXA: comparison with a
nutritional status after acute stroke. Stroke 1994, 25(2):366-371.
four-compartment model. J Appl Physiol. 2003;94:499-506.
Illness in the Elderly. New York, NY: Pergamon Press, Inc; 1985:49-52. Stevenson RD, Roberts CD, Vogtle L. The effects of non-nutritional factors on growth in
quadriplegic cerebral palsy. J Pediatr. 1995 May;126(5 Pt 1):833-9.
fatness in children and youth. Hum Biol 1988;60:709–723.
sectional study. J Appl Physiol. 2003;95: 2398–2407.
cerebral palsy. Dev Med Child Neurol. 1995;37: 124–130
2000;42:162–168
1996;156:958-63.
Repair. 2007;21:169-79.
Densitom. 2011 Aug 9.
Paraplegia 1995; 33: 669-673.
Orthop Sports Phys Ther. 2002;32:65-74.
85.
188.
palsy: poor performance of weight-for-height centiles. Dev Med Child Neurol.
nitrogen balance produced by immobilization. Scand Clin Lab Invest.1954;6:178-
mortality. A prospective study of 48.287 men and women. Arch Intern Med.
effects of long-term soleus electrical stimulation training. Neurorehabil Neural
Ezra JA. Body composition in ambulatory patients with multiple sclerosis. J Clin
Factors influencing body composition in persons with spinal cord injury: a cross-
longitudinal population study. In: Chandra RK, ed. Nutrition, Immunity and
injured individuals and in others who have prolonged immobilisation. A review.
McDonald CM, Abresch-Meyer AL, Nelson MD, Widman LM. Body mass index and body
Mitsiopoulos N, Baumgartner RN, Heymsfield SB, Lyons W, Gallagher D, and Ross R.
imaging and computerized tomography. J Appl Physiol. 85: 115–122, 1998. Modlesky CM, Bickel CS, Slade JM, Meyer RA, Cureton KJ, Dudley GA. Assessment of
National Institutes of Health Technology Assessment Conference Statement. Bioelectrical
Ng AV, Kent-Braun JA. Quantitation of lower physical activity in persons with multiple
Ogilvie C, Bowker P, Rowley DI. The physiological benefits of paraplegic orthotically aided
Olgiati R, Burgunder JM, Mumenthaler M. Increased energy cost of walking in multiple
Olle MM, Pivarnik JM, Klish WJ, Morrow JR Jr. Body composition of sedentary and
Patrick J, Gisel E. Nutrition for the feeding impaired child. J Neuro Rehab 1990;4:115–119. Peppler WW, Mazess RB. 1981. Total body bone mineral and lean body mass by dual-
Pietrobelli A, Formica C, Wang AM, Heymsfield SB. 1996. Dual-energy X-ray
Rittweger J, Beller G, Ehrig J, Jung C, Koch U, Ramolla J, Schmidt F, Newitt D, Majumdar S,
Ryan AS, Dobrovolny CL, Silver KH, Smith GV, Macko RF. Cardiovascular fitness after
Ryan AS, Dobrovolny CL, Smith GV, Silver KH, Macko RF. Hemiparetic muscle atrophy
Sabo D, Blaich S, Wenz W, Hohmann M, Loew M, Gerner HJ. Osteoporosis in patients
electrical conductivity. Arch Phys Med Rehabil. 1993;74:706-10.
photon absorptiometry. Calcif Tissue Int 33:353-359
Physiol 271 (Endocrinol Metab 34): E941-E951
with spinal cord injury. J Spinal Cord Med. 2007;30:S97-104.
cord injury. Spinal Cord. 2008 Mar;46:192-7.
sclerosis. Med Sci Sports Exerc 1997;29: 517-23.
Institutes of Health, 1994:12-4
1988;69:846-9.
Bone. 2000;27:319-26.
2002;83(12):1703–1707.
2000;9:185–191.
8.
walking. Paraplegia. 1993;31:111-5.
composition measures by dual x-ray absorptiometry in patients aged 10 to 21 years
Cadaver validation of skeletal muscle measurement by magnetic resonance
skeletal muscle mass in men with spinal cord injury using dual-energy X-ray absorptiometry and magnetic resonance imaging. J Appl Physiol. 2004;96:561-5. Mojtahedi MC, Valentine RJ, Arngrímsson SA, Wilund KR, Evans EM. The association
between regional body composition and metabolic outcomes in athletes with spinal
impedance analysis in body composition measurement. Bethesda, MD: National
sclerosis: effect of spasticity, ataxia, and weakness. Arch Phys Med Rehabil
physically active spinal cord injured individuals estimated from total body
absorptiometry body composition model: review of physical concepts. Am J
Schiessl H, Felsenberg D. Bone-muscle strength indices for the human lower leg.
stroke: Role of muscle mass and gait deficit severity. J Stroke Cerebro Dis
and increased intramuscular fat in stroke patients. Arch Phys Med Rehabil
with paralysis after spinal cord injury: a cross sectional study in 46 male patients with dual-energy X-ray absorptiometry. Arch Orthop Trauma Surg. 2001;121:75–
**Part 3**
**Miscellaneous**
Wang YH, Huang TS, Liang HW, Su TC, Chen SY, Wang TD. Fasting serum levels of adiponectin, ghrelin, and leptin in men with spinal cord injury. Arch Phys Med Rehabil. 2005;86:1964-8.
## **Part 3**
### **Miscellaneous**
94 Dual Energy X-Ray Absorptiometry
Wang YH, Huang TS, Liang HW, Su TC, Chen SY, Wang TD. Fasting serum levels of
Rehabil. 2005;86:1964-8.
adiponectin, ghrelin, and leptin in men with spinal cord injury. Arch Phys Med
**7**
*Spain*
**Internal Design of the Dry Human Ulna by DXA**
Dual energy X-ray absorptiometry (DXA) has been used to study dry bones such as spine, femur, jaw... to detect the first onsets of the ossification centers; In clinical practice, DXA is widely used for diagnosis and evaluation of osteoporosis, and a new generation of DXA scanners offers software for performing vertebral morphometry analysis (Blake & Fogelman, 1997). Also, many bone analyses have been performed on experimental animals using DXA
Studies on the spatial distribution of bone mineral density (BMD) in the whole bone, reflecting its morphological pattern are scarce (Gómez-Pellico et al., 1993 & Fernández-Camacho et al., 1996). In addition, there are just few studies regarding the anthropometric
In order to improve the treatment of the elbow's injury, knowledge related to the resistance of the bone is important to understand the origin of the fractures as well as to improve elbow fracture recovery (Heep, 2007). Most studies investigate the humeral component,
In order to develop an implant that carries out the mechanical characteristics of a native bone, we must study the trabecular architecture of the human ulna proximal extremity.
The ulna is a long bone, placed at the medial side of the forearm, parallel to the radius. It is divisible into a body and two extremities. Its upper extremity, of great thickness and strength, forms a large part of the elbow-joint; the bone diminishes in size from above downward, its lower extremity being very small, and excluded from the wrist-joint by the
The upper extremity presents two curved processes, the olecranon and the coronoid process; and two concave, articular cavities, the trochlear and radial notches. The olecranon is a large, thick, curved eminence, situated at the upper and back part of the ulna. The coronoid process is a triangular eminence projecting forward from the upper and front part of the ulna. Its base is continuous with the body of the bone, and of
interposition of an articular disk (the ulna articulates with the humerus and radius).
**1. Introduction**
(Tsujio et al., 2009).
characteristics of the human ulna (Weber et al., 2009).
**1.1 Brief anatomy of the human ulna**
considerable strength.
while the ulna component is not being studied as much (Goto, 2009).
S. Aguado-Henche, A. Bosch-Martín,
*University of Alcalá*
P. Spottorno-Rubio and R. Rodríguez-Torres
### **Internal Design of the Dry Human Ulna by DXA**
S. Aguado-Henche, A. Bosch-Martín, P. Spottorno-Rubio and R. Rodríguez-Torres *University of Alcalá Spain*
#### **1. Introduction**
Dual energy X-ray absorptiometry (DXA) has been used to study dry bones such as spine, femur, jaw... to detect the first onsets of the ossification centers; In clinical practice, DXA is widely used for diagnosis and evaluation of osteoporosis, and a new generation of DXA scanners offers software for performing vertebral morphometry analysis (Blake & Fogelman, 1997). Also, many bone analyses have been performed on experimental animals using DXA (Tsujio et al., 2009).
Studies on the spatial distribution of bone mineral density (BMD) in the whole bone, reflecting its morphological pattern are scarce (Gómez-Pellico et al., 1993 & Fernández-Camacho et al., 1996). In addition, there are just few studies regarding the anthropometric characteristics of the human ulna (Weber et al., 2009).
In order to improve the treatment of the elbow's injury, knowledge related to the resistance of the bone is important to understand the origin of the fractures as well as to improve elbow fracture recovery (Heep, 2007). Most studies investigate the humeral component, while the ulna component is not being studied as much (Goto, 2009).
In order to develop an implant that carries out the mechanical characteristics of a native bone, we must study the trabecular architecture of the human ulna proximal extremity.
#### **1.1 Brief anatomy of the human ulna**
The ulna is a long bone, placed at the medial side of the forearm, parallel to the radius. It is divisible into a body and two extremities. Its upper extremity, of great thickness and strength, forms a large part of the elbow-joint; the bone diminishes in size from above downward, its lower extremity being very small, and excluded from the wrist-joint by the interposition of an articular disk (the ulna articulates with the humerus and radius).
The upper extremity presents two curved processes, the olecranon and the coronoid process; and two concave, articular cavities, the trochlear and radial notches. The olecranon is a large, thick, curved eminence, situated at the upper and back part of the ulna. The coronoid process is a triangular eminence projecting forward from the upper and front part of the ulna. Its base is continuous with the body of the bone, and of considerable strength.
Internal Design of the Dry Human Ulna by DXA 99
To begin the scan, (figure 1) the starting point was placed 0, 5 cm directly above the upper extremity. A baseline point was marked under the lower extremity. A third point (goal line)
This technique has high accuracy and precision, approaching 1%. The speed of scanning was 60 mms, with an interlinear space of 1 mm and point by point resolution of 1 mm horizontal x 1 mm vertical. The defined exploration was completed as outlined in an average time of 10-15
min. Scan acquisition and scan analyses were performed by one investigator (figure 2).
was marked 1 cm from the more lateral part of the bone.
Fig. 1. Definition of the exploration area
Fig. 2. Densitometric image of the ulna.
Dry ulna calculations were performed for the following magnitudes:
**BMD**: Bone mineral density, in grams / cm2.
**BMC**: Bone mineral content, in grams. BMC is defined as the mass of
dependent parameter (Schoenau, 2004).
mineral contained in an entire bone or as the mass of mineral per unit bone length. Bone mineral content is obviously a size-
Its antero-inferior surface is concave, and marked by a rough impression for the insertion of the brachialis muscle. The trochlear notch is a large depression, formed by the olecranon and the coronoid process, and serving for articulation with the trochlea of the humerus. The notch is concave from above downward, and divided into a medial and a lateral portion by a smooth ridge running from the summit of the olecranon to the tip of the coronoid process. The radial notch is a narrow, oblong, articular depression on the lateral side of the coronoid process; it receives the circumferential articular surface of the head of the radius. The lower extremity of the ulna is small, and presents two eminences; the lateral and larger is a rounded, articular eminence, termed the head of the ulna; the medial, narrower and more projecting is a non-articular eminence named the styloid process.
The ulna is ossified from three centers: one for the body, the inferior extremity, and the top of the olecranon. Ossification begins about the eighth week of fetal life. About the fourth year, a center appears in the middle of the ulnar head, and soon extends into the styloid process. About the tenth year, a center appears in the olecranon near its extremity. The upper epiphysis joins the body about the sixteenth year and the lower about the twentieth.
#### **2. Objective**
In this chapter, we set out to show, by means of densitometric analysis with dual energy Xray absorptiometry (DXA) the internal design of the human ulna, to verify that the bone tissue distribution is not homogeneous and that this corresponds to the trabecular architecture of the bone.
#### **3. Material and method**
A random sample of 41 dry right ulnas from the skeletal collection of the Anatomy and Embriology Department of the University of Alcala was studied excluding those bones which presented any alterations or damage. A Norland XR-26 densitometer, software 2.5 (Norland Co., Fort Atkinson, WI, USA; Emsor SA, Madrid) was used for all studies. Each scan session was preceded by a calibration routine using a standard calibration block supplied by the manufacturer.
The bone is placed well centred on the examining board. It is important to check for stability so as not to vary its position during the study. Cotton gauze may be needed for an optimal stabilization. The bones are exposed directly, without any water or other materials that may resemble soft tissue.
For the densitometric analysis of the human ulna structure two projections were performed: lateral and antero-posterior.
For the study in two positions, the reference will be the ridge of the trochlear notch of the epiphyseal ulna (incisura trochlearis) which corresponds to the throat of the trochlea – humerus- (Gómez-Oliveros, 1962).
Its antero-inferior surface is concave, and marked by a rough impression for the insertion of the brachialis muscle. The trochlear notch is a large depression, formed by the olecranon and the coronoid process, and serving for articulation with the trochlea of the humerus. The notch is concave from above downward, and divided into a medial and a lateral portion by a smooth ridge running from the summit of the olecranon to the tip of the coronoid process. The radial notch is a narrow, oblong, articular depression on the lateral side of the coronoid process; it receives the circumferential articular surface of the head of the radius. The lower extremity of the ulna is small, and presents two eminences; the lateral and larger is a rounded, articular eminence, termed the head of the ulna; the medial, narrower and more
The ulna is ossified from three centers: one for the body, the inferior extremity, and the top of the olecranon. Ossification begins about the eighth week of fetal life. About the fourth year, a center appears in the middle of the ulnar head, and soon extends into the styloid process. About the tenth year, a center appears in the olecranon near its extremity. The upper epiphysis joins the body about the sixteenth year and the lower about the twentieth.
In this chapter, we set out to show, by means of densitometric analysis with dual energy Xray absorptiometry (DXA) the internal design of the human ulna, to verify that the bone tissue distribution is not homogeneous and that this corresponds to the trabecular
A random sample of 41 dry right ulnas from the skeletal collection of the Anatomy and Embriology Department of the University of Alcala was studied excluding those bones which presented any alterations or damage. A Norland XR-26 densitometer, software 2.5 (Norland Co., Fort Atkinson, WI, USA; Emsor SA, Madrid) was used for all studies. Each scan session was preceded by a calibration routine using a standard calibration block
The bone is placed well centred on the examining board. It is important to check for stability so as not to vary its position during the study. Cotton gauze may be needed for an optimal stabilization. The bones are exposed directly, without any water or other materials that may
For the densitometric analysis of the human ulna structure two projections were performed:
For the study in two positions, the reference will be the ridge of the trochlear notch of the epiphyseal ulna (incisura trochlearis) which corresponds to the throat of the trochlea –
Anteroposterior Position: The axis of the ridge of the trochlear notch is perpendicular to
Lateral Position: The axis of the ridge of the trochlear notch is parallel to the axis of the
projecting is a non-articular eminence named the styloid process.
**2. Objective**
architecture of the bone.
**3. Material and method**
supplied by the manufacturer.
lateral and antero-posterior.
examining board.
humerus- (Gómez-Oliveros, 1962).
the axis of the examining board.
resemble soft tissue.
To begin the scan, (figure 1) the starting point was placed 0, 5 cm directly above the upper extremity. A baseline point was marked under the lower extremity. A third point (goal line) was marked 1 cm from the more lateral part of the bone.
Fig. 1. Definition of the exploration area
This technique has high accuracy and precision, approaching 1%. The speed of scanning was 60 mms, with an interlinear space of 1 mm and point by point resolution of 1 mm horizontal x 1 mm vertical. The defined exploration was completed as outlined in an average time of 10-15 min. Scan acquisition and scan analyses were performed by one investigator (figure 2).
Fig. 2. Densitometric image of the ulna.
Dry ulna calculations were performed for the following magnitudes:
Internal Design of the Dry Human Ulna by DXA 101
Projection LAT Mínimum Máximum Mean SD Total BMD 0,38 0,94 0,67 0,14 Total BMC 13,15 46,74 28,72 8,50 Total Area 32,81 52,51 41,84 5,25 Total Lenght 20,40 28,05 24,50 1,66 Total Width 3,15 6,60 4,08 0,59
Table 2. Descriptive statistics of the total ulna in projection lateral (n=41). SD: Standard
AP-Area 10,58 1,57 9,08 1,33 8,12 1,42 7,02 1,01 6,22 0,94 AP-Lenght 4,88 0,33 4,88 0,33 5,03 0,41 4,88 0,33 4,88 0,33 LAT-Area 11,68 1,52 8,94 1,24 8,09 1,07 6,98 0,82 6,24 1,00 LAT-Lenght 4,88 0,33 4,88 0,33 5,03 0,43 4,88 0,33 4,88 0,33 Table 3. Descriptive statistics of the regions of interest in projections antero-posterior (AP)
Table 4. BMD (in grams/cm2) and BMC (in grams) of the regions of interest.
Dual energy X-ray absorptiometry (DXA) allows us to gather quantitative information on bone mineral content (BMC) and bone mineral density (BMD) of the bone (Wahner et al., 1985). As previously reported (Hvid et al., 1985), there is a close relationship between bone
In literature, there are various studies on long bones, such as the femur, the tibia, the humerus and the radius (Wahner et al., 1985; Kawashima & Uhthoff, 1991; Gómez-Pellico et
ROI 1 ROI 2 ROI 3 ROI 4 ROI 5 Media SD Media SD Media SD Media SD Media SD
deviation.
**5. Discussion**
mass and bone strength.
and lateral (LAT). n=41. SD: Standard deviation.
For the purpose of this survey, in both projections five equal regions of interest (ROI) were selected: proximal (ROI-1), proximal-intermediate (ROI-2), intermediate (ROI-3), distalintermediate (ROI-4) and distal (ROI-5). The total region corresponded to the area of the full length and height of the bone (figure 3).
All statistical calculations were performed using Statgraphics Plus (version 5.1) and SPSS (Statistical Package for Social Sciences), version 15.0. The means and standard deviation (SD) for bone mineral density (BMD) and bone mineral content (BMC) were calculated. The bone densities and the bone contents of the various regions of the ulna in the 2 projections were compared by Student's *t* test for paired samples.
Fig. 3. Regions of interest.
#### **4. Results**
DXA indicates that the higher BMD is in the proximal-intermediate region (R2), which is the part of the ulna that bears the higher force of traction. The higher BMC is found in the proximal region (R1) which corresponds to the coronoid process. Lower BMD and BMC are found in the distal region (R5). The total BMD shows significant statistical differences (p ≤ 0.001), which indicates the heterogeneous nature of the distribution of bone mass in the studied bone.
In tables 1 - 4 we present the statistic descriptions of the densitometry variables studied in both.
Table 1. Descriptive statistics of the total ulna in projection antero-posterior (n=41). SD: Standard deviation.
For the purpose of this survey, in both projections five equal regions of interest (ROI) were selected: proximal (ROI-1), proximal-intermediate (ROI-2), intermediate (ROI-3), distalintermediate (ROI-4) and distal (ROI-5). The total region corresponded to the area of the full
All statistical calculations were performed using Statgraphics Plus (version 5.1) and SPSS (Statistical Package for Social Sciences), version 15.0. The means and standard deviation (SD) for bone mineral density (BMD) and bone mineral content (BMC) were calculated. The bone densities and the bone contents of the various regions of the ulna in the 2 projections
DXA indicates that the higher BMD is in the proximal-intermediate region (R2), which is the part of the ulna that bears the higher force of traction. The higher BMC is found in the proximal region (R1) which corresponds to the coronoid process. Lower BMD and BMC are found in the distal region (R5). The total BMD shows significant statistical differences (p ≤ 0.001), which indicates the heterogeneous nature of the distribution of bone mass in the
In tables 1 - 4 we present the statistic descriptions of the densitometry variables studied in
Projection A-P Mínimum Máximum Mean SD Total BMD 0,40 0,97 0,69 0,14 Total BMC 13,10 46,40 28,75 8,63 Total Area 30,73 51,54 41,05 5,70 Total Lenght 20,40 27,90 24,51 1,66 Total Width 2,55 4,35 3,35 0,40
Table 1. Descriptive statistics of the total ulna in projection antero-posterior (n=41). SD:
**AREA**: Measured area, in square centimetres (cm2). **LENGTH**: Total length of the bone, in centimetres (cm). **WIDTH**: Total width of the bone, in centimetres (cm).
length and height of the bone (figure 3).
Fig. 3. Regions of interest.
**4. Results**
studied bone.
Standard deviation.
both.
were compared by Student's *t* test for paired samples.
Table 2. Descriptive statistics of the total ulna in projection lateral (n=41). SD: Standard deviation.
Table 3. Descriptive statistics of the regions of interest in projections antero-posterior (AP) and lateral (LAT). n=41. SD: Standard deviation.
Table 4. BMD (in grams/cm2) and BMC (in grams) of the regions of interest.
#### **5. Discussion**
Dual energy X-ray absorptiometry (DXA) allows us to gather quantitative information on bone mineral content (BMC) and bone mineral density (BMD) of the bone (Wahner et al., 1985). As previously reported (Hvid et al., 1985), there is a close relationship between bone mass and bone strength.
In literature, there are various studies on long bones, such as the femur, the tibia, the humerus and the radius (Wahner et al., 1985; Kawashima & Uhthoff, 1991; Gómez-Pellico et
Internal Design of the Dry Human Ulna by DXA 103
The human ulna presents a heterogeneous distribution of the BMD. This study confirms that the higher mechanical requirements in the ulna are in the proximal extremity. The differences found in the ulna BMD allows us a better understanding of the construction systematics and their functional activity. We conclude that bone densitometry, measured by the DXA technique, is useful for assessing trabecular architecture of the human skeleton. This study may provide some useful information on plate application for the treatment of
Blake, GM. & Fogelman I. (1997). Technical principles, *Seminars in Nuclear Medicine*
D'Amelio, P., Panattoni, GL. & Isaia GC. (2002). Densitometric study of human developing
Chen, WC., Hsu, WY & Wu, JJ (1991). Stress fracture of the diaphysis of the ulna,
Fernández Camacho, FJ., Morante Martínez, P., Rodríguez Torres, R., Cortés García, A. &
Fowler, K. & Chung, C. (2006). Normal MR imaging anatomy of the elbow, *Radiologic Clinics*
Gómez Pellico, L., Morante Martínez, P., & Dankloff Mora, C. (1993). Definición
Gómez-Oliveros, L. (1962). *Lecciones de Anatomía Humana. Osteología. Tercera parte. Miembros.*
Goto, A., Murase, T., Hashimoto, J., Oka, K., Yoshikawa, H. & Sugamoto, K. (2009).
Griffith, JF., Genant, HK. (2008). Bone mass and architecture determination: state of the art, *Best practice & Research Clinical Endocrinology & Metabolism.*22(5):737-764. Hepp, P., Josten, C. (2007). Biology and Biomechanics in Osteosynthesis of Proximal
Hvid, I., Jensen, NC., Bünger, C., Solund, K., Djurhuus, JC. (1985). Bone mineral assay: its
Kalkwarf, HJ., Laor, T & Bean JA. (2011). Fracture risk in children with forearm injury is
Kawashima, T. & Uhthoff HK. (1991). Pattern of bone loss of the proximal femur: a
Kim, JM., Mudgal, CS., Konopka, JF & Júpiter JB. (2011). Complications of total elbow arthroplasty, *Journal of American Academy of Orthopaedic Surgeons.* 19(6):328-339.
areal bone density (by DXA), *Osteoporosis International* 22:607-616.
Gómez Pellico, L. (1996). Densitometric analysis of the human calcaneus, *Journal of*
densitométrica de la morfología estructural del huesos del esqueleto humano, *Jano*
Morphologic análisis of the medulary canal in rheumatoid elbows, *Journal of*
Humerus Fractures, *European Journal of Trauma and Emergency Surgery* 33(4):337-344.
relation to the mechanical strength of cancellous bone, *Engineering in Medicine*
associated with volumetric bone density and cortical area (by peripheral QCT) and
radiologic, densitometric, and histomorphometric study, *Journal of Orthopaedic*
dry bones: a review, *Journal of Clinical Densitometric* 5(1):73-78.
*International Orthopaedics* 15: 197-198.
*Shoulder and Elbow Surgery.*18(1):33-37.
**6. Conclusions**
the elbow injuries.
**7. References**
27(3):210-228.
*Anatomy* 189:205-209.
XLV:637-640.
14:79-83.
*Research* 9(5):630-640.
Madrid. Marban.
*of North America* 44(4):553-567.
al., 1993; D'Amelio et al., 2002) but we haven't found any references that study the distribution of the BMD in the ulna that describes it´s construction systematics .
According to Wolf's law (Viladot , 2001), the bone adapts its size, shape and structure to the mechanical requirements it receives. Furthermore, Pauwels (Pauwels, 1980; Miralles, 1998) suggests that the mass of the cortical bone is distributed along its axis proportionally to the amount of tensions it receives. In our analysis, we find a wider BMD in the intermediate-proximal region (ROI-2), which corresponds to the region of the bone exposed to the mechanical flexions and to the transmission of weight charges while the superior member is in the extended position. Three different soft tissue structures insert in or attach to the coronoid process, the articular capsule, the tendon of the brachialis muscle, and the anterior band of the ulnar collateral ligament (Fowler & Chung, 2006). Furthermore, the transmission of weight charges travels through the coronoid apophysis, situated in the proximal region, which present the wider BMC in both projections, however, DXA is unable to distinguish between cortical and trabecular bone (Griffith & Genant, 2008).
This and other similar studies will contribute to a better understanding of stress related fractures which are quite scarce in the ulna and cannot easily be found in literature (Chen, WC et al., 1991). Most fractures occur in the middle third of the diaphysis and surrounding areas as a result of mechanical stressing forces of the forearm in a specific position, especially in athletes (Rettig, 1983).
Some authors have agreed on the homogeneous nature of the different diaphysarial regions of the long bones that they study (femur, humerus and tibia) (Gómez Pellico et al., 1993; Fernández-Camacho et al., 1996). As far as the ulna is concerned, BMD displays a more heterogeneous distribution, since we find statistical differences in all studied regions and on the entire bone.
In studies of the dry femur with DXA, epiphysiary regions are those with less BMD which, added to the mechanical requirements of the physiology of the articulation, would explain how hip ostheoporotic fractures occur (Gómez-Pellico et al., 1993). Furthermore, most studies with DXA are based on BMD variations related to loss of bone mass of pathologic nature (McCarthy et al., 1991). This also happens with "in vivo" studies of the radius. Due to the high rate fractures of the distal radius in children, the use this bone to measure BMD, is increasing, essentially as thus to predict the risk of fracture (Kalkwarf et al., 2011). The study of the dry radius with DXA would define its construction systematics.
The results obtained with the DXA technique showed that BMD agrees with the arrangement of the trabecular system in the human ulna, previously described by some authors (Testut & Latarjet, 1949; Gómez-Oliveros, 1962).
In addition, fractures of the coronoid process are rarely seen as an isolated injury. They are encountered more frecuently in association with radial head fractures (Weber et al., 2009).
Due to frequent complications associated with reconstructive surgery for the elbow, implant loosening, periprosthetic fracture, implant failure… (Kim, 2011), that remains higher than arthroplasty of other joints (Sanchez-Sotelo, 2011), the findings that result from this study could contribute to the improvement of elbow prosthesis.
#### **6. Conclusions**
102 Dual Energy X-Ray Absorptiometry
al., 1993; D'Amelio et al., 2002) but we haven't found any references that study the
According to Wolf's law (Viladot , 2001), the bone adapts its size, shape and structure to the mechanical requirements it receives. Furthermore, Pauwels (Pauwels, 1980; Miralles, 1998) suggests that the mass of the cortical bone is distributed along its axis proportionally to the amount of tensions it receives. In our analysis, we find a wider BMD in the intermediate-proximal region (ROI-2), which corresponds to the region of the bone exposed to the mechanical flexions and to the transmission of weight charges while the superior member is in the extended position. Three different soft tissue structures insert in or attach to the coronoid process, the articular capsule, the tendon of the brachialis muscle, and the anterior band of the ulnar collateral ligament (Fowler & Chung, 2006). Furthermore, the transmission of weight charges travels through the coronoid apophysis, situated in the proximal region, which present the wider BMC in both projections, however, DXA is unable to distinguish between cortical and trabecular bone (Griffith &
This and other similar studies will contribute to a better understanding of stress related fractures which are quite scarce in the ulna and cannot easily be found in literature (Chen, WC et al., 1991). Most fractures occur in the middle third of the diaphysis and surrounding areas as a result of mechanical stressing forces of the forearm in a specific position,
Some authors have agreed on the homogeneous nature of the different diaphysarial regions of the long bones that they study (femur, humerus and tibia) (Gómez Pellico et al., 1993; Fernández-Camacho et al., 1996). As far as the ulna is concerned, BMD displays a more heterogeneous distribution, since we find statistical differences in all studied regions and on
In studies of the dry femur with DXA, epiphysiary regions are those with less BMD which, added to the mechanical requirements of the physiology of the articulation, would explain how hip ostheoporotic fractures occur (Gómez-Pellico et al., 1993). Furthermore, most studies with DXA are based on BMD variations related to loss of bone mass of pathologic nature (McCarthy et al., 1991). This also happens with "in vivo" studies of the radius. Due to the high rate fractures of the distal radius in children, the use this bone to measure BMD, is increasing, essentially as thus to predict the risk of fracture (Kalkwarf et al., 2011). The study
The results obtained with the DXA technique showed that BMD agrees with the arrangement of the trabecular system in the human ulna, previously described by some
In addition, fractures of the coronoid process are rarely seen as an isolated injury. They are encountered more frecuently in association with radial head fractures (Weber et al., 2009).
Due to frequent complications associated with reconstructive surgery for the elbow, implant loosening, periprosthetic fracture, implant failure… (Kim, 2011), that remains higher than arthroplasty of other joints (Sanchez-Sotelo, 2011), the findings that result from this study
of the dry radius with DXA would define its construction systematics.
authors (Testut & Latarjet, 1949; Gómez-Oliveros, 1962).
could contribute to the improvement of elbow prosthesis.
distribution of the BMD in the ulna that describes it´s construction systematics .
Genant, 2008).
the entire bone.
especially in athletes (Rettig, 1983).
The human ulna presents a heterogeneous distribution of the BMD. This study confirms that the higher mechanical requirements in the ulna are in the proximal extremity. The differences found in the ulna BMD allows us a better understanding of the construction systematics and their functional activity. We conclude that bone densitometry, measured by the DXA technique, is useful for assessing trabecular architecture of the human skeleton. This study may provide some useful information on plate application for the treatment of the elbow injuries.
#### **7. References**
**8**
*Japan*
Kazuhiro Imai1,2,3
*Ex Vivo* **and** *In Vivo* **Assessment of Vertebral**
**by Dual Energy X-Ray Absorptiometry**
*2Department of Orthopaedic Surgery, School of Medicine, Tokyo University,*
*3Department of Orthopaedic Surgery, Tokyo Metropolitan Geriatric Medical Center,*
Osteoporosis is defined as a skeletal disorder characterized by loss of bone mass, decreased bone strength and resulting in increased risk of bone fracture. The disease is progressive with age, especially in postmenopausal women [1]. Osteoporotic hip fractures and vertebral fractures have become a major social problem because the elderly population continues to increase. Hip fractures account for about 10% of all osteoporosis-related fractures [2]. Hip fractures are particularly devastating and have a particularly negative impact on morbidity. Survivors often suffer severe and prolonged physical and social limitations, and fail to recover normal activity [3]. Vertebral fractures affect approximately 25% of postmenopausal women [4]. Vertebral fractures can be associated with chronic disabling pain and incur loss
In addition to this increased awareness of osteoporosis as a significant health problem, there has been the emergence of several novel drugs that appear to be effective at reducing the risk of fracture, such as bisphosphonates. Consequently, clinicians and researchers are emphasizing the importance of early detection of osteoporosis, aggressive fracture prevention, and monitoring of patients who have high risk of fractures. Fracture risk associated with osteoporosis consisted of bone strength reduction and tendency to fall, therefore it is essential to measure bone strength to assess the risk of fracture. Bone strength reflects the integration of bone density and bone quality, which are influenced by bone
Traditionally, measurement of areal bone mineral density (aBMD) by dual energy X-ray absorptiometry (DXA) has served as the means by which to best diagnose osteoporosis and evaluate fracture risk [6]. In 1994, the World Health Organization (WHO) published a set of diagnostic criteria to define osteoporosis in postmenopausal Caucasian women, using aBMD values measured by DXA [7]. Measurement of aBMD by DXA has been the standard method for diagnosing osteoporosis, in addition to assessing fracture risk and therapeutic effects. However, a variety of problems exist with DXA, which include its relatively high cost, the absence of DXA in many communities, especially in less-developed countries.
architecture, bone turnover, accumulation of damage, and mineralization [5].
**1. Introduction**
of normal activity.
*1Department of Orthopaedic Surgery, Mishuku Hospital, Tokyo,*
**Strength and Vertebral Fracture Risk Assessed**
### *Ex Vivo* **and** *In Vivo* **Assessment of Vertebral Strength and Vertebral Fracture Risk Assessed by Dual Energy X-Ray Absorptiometry**
Kazuhiro Imai1,2,3
*1Department of Orthopaedic Surgery, Mishuku Hospital, Tokyo, 2Department of Orthopaedic Surgery, School of Medicine, Tokyo University, 3Department of Orthopaedic Surgery, Tokyo Metropolitan Geriatric Medical Center, Japan*
#### **1. Introduction**
104 Dual Energy X-Ray Absorptiometry
McCarthy, CK., Steinberg, GG., Agren, M., Leahey, D., Wyman, E., & Baran DT (1991).
Rettig, AC. (1983). Stress fracture of the ulna in an adolescent tournament tennis player.
Sanchez-Sotelo, J. (2011). Total elbow arthroplasty, *The Open Orthopaedics Journal* 16(5):115-
Schoenau, E., Land, C., Stabrey, A., Remer, T. & Kroke, A. (2004). The bone mass concept: problems in short stature, *European Journal of Endocrinology* 151:S87-S91. Testut, L. & Latarjet, A. (1949). *Tratado de Anatomía Humana. Volume 1* Barcelona: Salvat. Tsujio, M., Mizorogi, T., Kitamura, I., Maeda, Y., Nishijima, K., Kuwahara, S., Ohno, T.,
Viladot Voegeli, A. (2001) *Lecciones Básicas de Biomecánica del Aparato Locomotor.* Barcelona:
Wahner, HW., Eastell, R. & Riggs, BL. (1985). Bone mineral density of the radius: Where do
Weber, MF., Barbosa, DM., Belentani, C., Ramos, PM., Trudell, D. & Resnick, D. (2009).
we stand?, *The Journal of Nuclear Medicine* 26(11):1339-1341.
Niida, S., Nagoya, M., Saito, R. & Tanaka, S. (2009). Bone mineral analisis through dual energy X-ray absorptiometry in laboratory animals, *Journal of Veterinary*
Coronoid process of the ulna: paleopathologic and anatomic study with imaging correlation. Emphasis on the anteromedial "facet", *Skeletal Radiology* 38(1):61-67.
Miralles Marrero, R. (1998). *Biomecánica Clínica del Aparato Locomotor.* Barcelona: Masson. Pauwels, F. (1980) *Biomechanics of the Locomotor Apparatus. Contribution on the functional of the*
*of Bone and Joint Surgery* 73(5):774-778.
*Medical Science* 71(11):1493-1497.
123.
Springer.
*Locomotor Apparatus.* Nueva York: Srpinger.
*American Journal of Sports Medicine* 11:103-106.
Quantifying bone loss from the proximal femur after total hip arthroplasty, *Journal*
Osteoporosis is defined as a skeletal disorder characterized by loss of bone mass, decreased bone strength and resulting in increased risk of bone fracture. The disease is progressive with age, especially in postmenopausal women [1]. Osteoporotic hip fractures and vertebral fractures have become a major social problem because the elderly population continues to increase. Hip fractures account for about 10% of all osteoporosis-related fractures [2]. Hip fractures are particularly devastating and have a particularly negative impact on morbidity. Survivors often suffer severe and prolonged physical and social limitations, and fail to recover normal activity [3]. Vertebral fractures affect approximately 25% of postmenopausal women [4]. Vertebral fractures can be associated with chronic disabling pain and incur loss of normal activity.
In addition to this increased awareness of osteoporosis as a significant health problem, there has been the emergence of several novel drugs that appear to be effective at reducing the risk of fracture, such as bisphosphonates. Consequently, clinicians and researchers are emphasizing the importance of early detection of osteoporosis, aggressive fracture prevention, and monitoring of patients who have high risk of fractures. Fracture risk associated with osteoporosis consisted of bone strength reduction and tendency to fall, therefore it is essential to measure bone strength to assess the risk of fracture. Bone strength reflects the integration of bone density and bone quality, which are influenced by bone architecture, bone turnover, accumulation of damage, and mineralization [5].
Traditionally, measurement of areal bone mineral density (aBMD) by dual energy X-ray absorptiometry (DXA) has served as the means by which to best diagnose osteoporosis and evaluate fracture risk [6]. In 1994, the World Health Organization (WHO) published a set of diagnostic criteria to define osteoporosis in postmenopausal Caucasian women, using aBMD values measured by DXA [7]. Measurement of aBMD by DXA has been the standard method for diagnosing osteoporosis, in addition to assessing fracture risk and therapeutic effects. However, a variety of problems exist with DXA, which include its relatively high cost, the absence of DXA in many communities, especially in less-developed countries.
*Ex Vivo* and *In Vivo* Assessment of Vertebral Strength
introduced by fat within the vertebral bone marrow.
widespread use in studies of osteoporosis.
have sustained a prior fracture [7].
part-body examinations being only a few microsieverts (µSv).
and Vertebral Fracture Risk Assessed by Dual Energy X-Ray Absorptiometry 107
the appendicular skeleton. The development of dual-photon absorptiometry (DPA) and, more recently, dual-energy X-ray absorptiometry (DXA) have resolved at least some of these problems. The different thickness of soft tissue can be accommodated by simultaneous measurement of the transmission of gamma-rays of two different energies, which makes the techniques applicable to any part of the body, but particularly the lumbar spine and hip.
The theory underlying DPA and DXA requires that there are only two components present – bone and soft tissue of uniform composition. In practice, fat forms a further component with attenuation characteristics that differ from those of water, muscle and most organs. A uniform layer of fat is unimportant, but fat is distributed non-uniformly in the region of the lumbar spine and may cause errors of up to 10% in spinal bone mineral. Errors can also be
Total body bone mineral can be measured by DPA, but instrumental problems are greater because of the wide range of count rates and the non-uniform distribution of fat, which introduces errors. However, total body bone mineral measured by neutron activation analysis. As with SPA, the radiation dose for DPA is low, the effective dose equivalent for
Recently, sources of gamma radiation have been replaced by X-ray generators. The necessary pairs of effective energies can be obtained either by K-edge filtering, using cerium or samarium, or by rapidly switching the generator potential. The advantages of these approaches are a higher beam intensity and therefore faster scan, improved spatial resolution with easier identification of vertebrae, and better precision. The absence of source decay also eliminates problems associated with decreasing count rates over the lifetime of the source.
Like DPA, DXA determines bone mineral density from an anterior-posterior image. The sites most commonly measured are the lumbar spine, generally L2-L4, including the intervertebral discs. Other sites include the hip, forearm, whole body and skeletal segments. The error in reproducibility *in vitro* is 1-2%. DXA has been reported to have a high shortterm and long-term precision *in vivo*, which is about twice that of DPA. This has led to its
A recent development has been scanning of the lumbar spine in the lateral position, which has the advantage of eliminating the posterior arch and the spines of the vertebrae as well as aortic calcification from the measurement. Its limitations are the increased soft tissue mass and overlap of the projected image by the ribs and pelvis, so that only one or two vertebrae are measured. Lateral scanning provides a measurement of vertebral depth which, together with the antero-posterior area, can provide a volumetric measurement for calculating bone mineral mass per unit volume. Whether this volumetric density measure is a better predictor of fracture is unknown. The technique may be useful in assessment of bone density in children, allowing accurate assessment of vertebral size. The precision error of measurement of the vertebral body and mid-slice *in vivo* is of the order of 2% [15]. DXA has now largely replaced DPA for screening because of its greater precision, ease of use and freedom from several technical artifacts. The WHO defines osteoporosis as a value for aBMD by DXA 2.5 standard deviation (SD) or more below the mean for young Caucasian adult women (T-score diagnostic criteria of -2.5), based on data that this criterion identified 30% of all postmenopausal women as having osteoporosis, more than half of whom would
Therefore, aBMD by DXA is not a suitable screening method for fracture risk in terms of accessibleness and cost. In addition, the correlations between bone strength and aBMD by DXA are reported to be 0.51-0.80 [8-11], which indicates aBMD only accounts for 50 to 80% of bone strength. And the application of aBMD measurements in isolation cannot identify individuals who eventually experience bone fracture because of the low sensitivity of the test [12].
Recently, quantitative ultrasound (QUS) is emerging as a relatively low-cost and readily accessible alternative means to identify osteoporosis, evaluate fracture risk, and initiate osteoporosis treatment. More recently, finite element (FE) method based on data from computed tomography (CT) has been used to assess bone strength, fracture risk, and therapeutic effects on osteoporosis.
#### **2. Dual energy X-ray absorptiometry (DXA)**
In the 1960s, a new method of measuring aBMD, called single-photon absorptiometry (SPA), was developed. In this method, a single-energy photon beam is passed through bone and soft tissue to a detector. The amount of mineral in the path is then quantified. This method most commonly uses a gamma-ray source coupled with a scintillation detector, which together scan across the area of interest [13]. The amount of the bone mineral in the tissue traversed by a well collimated gamma-ray beam is derived from its attenuation through bone plus soft tissue relative to that through soft tissue alone. The overall thickness of the soft tissue is standardized, usually by immersing the limb in water or cuffing with a fluidfilled bag. The value obtained is proportional to the bone mineral content of the segment scanned. The value may be divided by the bone width (yielding a result in g/cm) or by an estimate of the cross-sectional area to give a value for bone mineral density in g/cm2. The technique has been applied to the femur, humerus, metacarpal, os calcis, hand and foot, but the most commonly used site is the forearm. The most frequently used source is 125I (27keV), but has the major drawback of a relatively short half-life (60 days).
Accuracy may be compromised by a non-uniform thickness of fat, which has attenuation characteristics different from those of water or lean soft tissue. In some equipment, the program assumes the fat to be a uniform shell around the bone and makes a correction, but the correction requires a number of assumptions that influence the accuracy of the method. The heterogeneity of surrounding tissues is nevertheless considerably less than that of tissue surrounding axial sites such as the spine. Although true *in vivo* estimates of accuracy have not been made, errors in cadaveric studies of excised bone have sufficiently low to make the technique attractive for screening [14].
The radiation dose of SPA is very low and applied to a small volume of tissue, giving an effective dose equivalent of < 1µSv. Typical scanning times are 10-15 minutes. Single-energy X-ray absorptiometry (SXA) is a newly developed technique suitable for scanning appendicular sites. It avoids the need for isotopes and is likely to replace SPA.
The proximal femur and the vertebral bodies, with their associated processes, are very irregular bones that are difficult to delineate. Furthermore, they are surrounded by a widely varying amount of fat and muscle mass. The ratio of bone mass to soft tissue is thus lower in the spine or hip than in the forearm, and standardization of soft tissue by immersion in water is not feasible for these sites. These and other factors limit the use of SPA or SXA to
Therefore, aBMD by DXA is not a suitable screening method for fracture risk in terms of accessibleness and cost. In addition, the correlations between bone strength and aBMD by DXA are reported to be 0.51-0.80 [8-11], which indicates aBMD only accounts for 50 to 80% of bone strength. And the application of aBMD measurements in isolation cannot identify individuals who eventually experience bone fracture because of the low sensitivity of the
Recently, quantitative ultrasound (QUS) is emerging as a relatively low-cost and readily accessible alternative means to identify osteoporosis, evaluate fracture risk, and initiate osteoporosis treatment. More recently, finite element (FE) method based on data from computed tomography (CT) has been used to assess bone strength, fracture risk, and
In the 1960s, a new method of measuring aBMD, called single-photon absorptiometry (SPA), was developed. In this method, a single-energy photon beam is passed through bone and soft tissue to a detector. The amount of mineral in the path is then quantified. This method most commonly uses a gamma-ray source coupled with a scintillation detector, which together scan across the area of interest [13]. The amount of the bone mineral in the tissue traversed by a well collimated gamma-ray beam is derived from its attenuation through bone plus soft tissue relative to that through soft tissue alone. The overall thickness of the soft tissue is standardized, usually by immersing the limb in water or cuffing with a fluidfilled bag. The value obtained is proportional to the bone mineral content of the segment scanned. The value may be divided by the bone width (yielding a result in g/cm) or by an estimate of the cross-sectional area to give a value for bone mineral density in g/cm2. The technique has been applied to the femur, humerus, metacarpal, os calcis, hand and foot, but the most commonly used site is the forearm. The most frequently used source is 125I (27keV),
Accuracy may be compromised by a non-uniform thickness of fat, which has attenuation characteristics different from those of water or lean soft tissue. In some equipment, the program assumes the fat to be a uniform shell around the bone and makes a correction, but the correction requires a number of assumptions that influence the accuracy of the method. The heterogeneity of surrounding tissues is nevertheless considerably less than that of tissue surrounding axial sites such as the spine. Although true *in vivo* estimates of accuracy have not been made, errors in cadaveric studies of excised bone have sufficiently low to make the
The radiation dose of SPA is very low and applied to a small volume of tissue, giving an effective dose equivalent of < 1µSv. Typical scanning times are 10-15 minutes. Single-energy X-ray absorptiometry (SXA) is a newly developed technique suitable for scanning
The proximal femur and the vertebral bodies, with their associated processes, are very irregular bones that are difficult to delineate. Furthermore, they are surrounded by a widely varying amount of fat and muscle mass. The ratio of bone mass to soft tissue is thus lower in the spine or hip than in the forearm, and standardization of soft tissue by immersion in water is not feasible for these sites. These and other factors limit the use of SPA or SXA to
appendicular sites. It avoids the need for isotopes and is likely to replace SPA.
test [12].
therapeutic effects on osteoporosis.
technique attractive for screening [14].
**2. Dual energy X-ray absorptiometry (DXA)**
but has the major drawback of a relatively short half-life (60 days).
the appendicular skeleton. The development of dual-photon absorptiometry (DPA) and, more recently, dual-energy X-ray absorptiometry (DXA) have resolved at least some of these problems. The different thickness of soft tissue can be accommodated by simultaneous measurement of the transmission of gamma-rays of two different energies, which makes the techniques applicable to any part of the body, but particularly the lumbar spine and hip.
The theory underlying DPA and DXA requires that there are only two components present – bone and soft tissue of uniform composition. In practice, fat forms a further component with attenuation characteristics that differ from those of water, muscle and most organs. A uniform layer of fat is unimportant, but fat is distributed non-uniformly in the region of the lumbar spine and may cause errors of up to 10% in spinal bone mineral. Errors can also be introduced by fat within the vertebral bone marrow.
Total body bone mineral can be measured by DPA, but instrumental problems are greater because of the wide range of count rates and the non-uniform distribution of fat, which introduces errors. However, total body bone mineral measured by neutron activation analysis. As with SPA, the radiation dose for DPA is low, the effective dose equivalent for part-body examinations being only a few microsieverts (µSv).
Recently, sources of gamma radiation have been replaced by X-ray generators. The necessary pairs of effective energies can be obtained either by K-edge filtering, using cerium or samarium, or by rapidly switching the generator potential. The advantages of these approaches are a higher beam intensity and therefore faster scan, improved spatial resolution with easier identification of vertebrae, and better precision. The absence of source decay also eliminates problems associated with decreasing count rates over the lifetime of the source.
Like DPA, DXA determines bone mineral density from an anterior-posterior image. The sites most commonly measured are the lumbar spine, generally L2-L4, including the intervertebral discs. Other sites include the hip, forearm, whole body and skeletal segments. The error in reproducibility *in vitro* is 1-2%. DXA has been reported to have a high shortterm and long-term precision *in vivo*, which is about twice that of DPA. This has led to its widespread use in studies of osteoporosis.
A recent development has been scanning of the lumbar spine in the lateral position, which has the advantage of eliminating the posterior arch and the spines of the vertebrae as well as aortic calcification from the measurement. Its limitations are the increased soft tissue mass and overlap of the projected image by the ribs and pelvis, so that only one or two vertebrae are measured. Lateral scanning provides a measurement of vertebral depth which, together with the antero-posterior area, can provide a volumetric measurement for calculating bone mineral mass per unit volume. Whether this volumetric density measure is a better predictor of fracture is unknown. The technique may be useful in assessment of bone density in children, allowing accurate assessment of vertebral size. The precision error of measurement of the vertebral body and mid-slice *in vivo* is of the order of 2% [15]. DXA has now largely replaced DPA for screening because of its greater precision, ease of use and freedom from several technical artifacts. The WHO defines osteoporosis as a value for aBMD by DXA 2.5 standard deviation (SD) or more below the mean for young Caucasian adult women (T-score diagnostic criteria of -2.5), based on data that this criterion identified 30% of all postmenopausal women as having osteoporosis, more than half of whom would have sustained a prior fracture [7].
*Ex Vivo* and *In Vivo* Assessment of Vertebral Strength
evaluating bone fracture risk.
and Vertebral Fracture Risk Assessed by Dual Energy X-Ray Absorptiometry 109
QUS bone assessment method has been recently introduced as an alternative for peripheral bone mass assessment, reflecting bone strength, bone density, and bone elasticity or fragility, and may be superior to aBMD by DXA [26]. The advantages of this method over Xray-based techniques, which include low cost, portability, and no radiation exposure, have encouraged the use of this method for defining a stage of development of osteoporosis and
There are several reports for assessing bone conditions *in vivo* using QUS method and apparatus. QUS devices can be classified mostly into 3 groups, related to the type of ultrasound transmission. Trabecular sound transmission is best for measuring the heel [27]. Cortical transverse transmission currently only is used in phalanx contact devices [28]. And cortical axial transmission presently is being investigated for use in phalanges, the radius, and the tibia [28]. Heel devices currently appear to have the most clinical applications, where QUS are being used and evaluated for the prediction of fracture risk, the diagnosis of osteoporosis, the initiation of osteoporosis treatment, the monitoring of osteoporosis treatment, and osteoporosis case finding. For these purposes, the recommended parameter of interest in clinical routine is a composite score, e.g., heel stiffness index or Quantitative Ultrasound Index (QUI) combining the results of broad-band ultrasound attenuation (BUA)
At the present time, there is good evidence that QUS can discriminate those with osteoporotic fractures from age-matched controls without osteoporotic fracture [29,30]. The power of heel QUS to predict fracture observed in cross-sectional studies has been confirmed prospectively in some populations as defined by sex, age, and ethnic background. This is particularly true of heel QUS and for hip and spinal fractures. However, because of methodological issues, it is difficult to compare studies. Nonetheless, it is possible to make the following generations. Using QUS of the heel, the increase in relative risk for each standard deviation decrease in stiffness index (SI) is approximately 2.0 for the hip and spine
The evidence from studies is good that the heel QUS SI using QUS devices is predictive of hip fracture risk in Caucasian and Asian women over age 55 and of any fracture risk in Asian women over age 55. Cortical axial transmission devices have no prospectively proven clinical utility, although clinical use in adults of phalanx QUS devices using cortical transverse transmission is also limited. These results for heel QUS are roughly the same as for DXA by BMD in terms of hip and spine fracture risk per SD decrease [12,42]. Discordant results between heel QUS and DXA, which are not infrequent, are not necessarily an indication of methodological error but rather due to the independence between the 2
Diagnosing osteoporosis using QUS is less supported by evidence and more complicated and problematic than assessing fracture risk. To start with, the T-score diagnostic criteria of - 2.5, classically used for DXA aBMD, cannot be applied to QUS without discrepancies in the numbers of women diagnosed with osteoporosis because of tremendous variations in QUS measurements by skeletal site, because different QUS devices yield different results, and because of the relatively poor correlation between heel QUS and hip/spine DXA measurements. If the prevalence of osteoporosis is defined as -2.5 SD from the mean
**4. Quantitative ultrasound (QUS) bone assessment method**
and speed of sound (SOS), as measured in meters per second.
and roughly 1.5 for all fractures combined [31-41].
techniques.
#### **3. Quantitative computed tomography (QCT)**
In quantitative computed tomography (QCT), a thin transverse slice through the body is imaged. Under appropriate conditions, the image can be quantified to give a measure of volumetric bone mineral density (vBMD) (mg/cm3), and cancellous bone can be measured independently of surrounding cortical bone and aortic calcification. Developments have been concentrated in two directions: the construction of special equipment using a radionuclide source for measurements of the forearm, and the adaptation of X-ray CT machines installed for general radiology to measure vBMD. The attraction of the technique is that cancellous bone can be examined separately from cortical bone. It also gives a true value for mineral density (mg/cm3) unlike other techniques.
A dedicated forearm scanner was first described in the mid 1970s [16,17]. The photon source is 125I and is mounted in a gantry with a sodium iodide scintillation detector. A linear scan is performed at each of 48 angular positions. Computer reconstruction generates an image in which a region of interest in the cancellous bone of the distal ulna is selected. Since 1980s, QCT has been used as a means for non-invasive quantitative determination of bone mineral of the spine [18,19].
A lateral plane projection scan is necessary for precise slice positioning through the centers of the vertebrae. Comparison between the CT Hounsfield numbers and a calibration standard scanned simultaneously allows bone density to be expressed in terms of the equivalent concentration of the material of the standard. Regions of interest within the vertebral bodies are selected: circular, elliptical, rectangular or other chosen areas are selected to include all the cancellous bone just inside the cortex. The relationship between the observed CT number and the true attenuation coefficient is subject to short- and longterm variation, so that it is necessary to scan the patient and a calibration standard simultaneously. Recently, simple standards with fewer components based on suspensions of calcium hydroxyapatite in plastic have been adopted. Comparison between the standard and the Hounsfield numbers of the trabecular region of the vertebral bodies allows bone density to be expressed in terms of the equivalent concentration of the material of the standard.
Investigators reported the prediction of vertebral body compressive strength using QCT. In 1985, McBroom et al. [20] showed a strong positive correlation between QCT and apparent density of the vertebral trabecular bone but could find only suggestive, not quite significant, correlations between QCT and the vertebral body compressive strength. Cann et al. [21] showed that QCT evaluation of vertebral trabecular bone mineral density is a useful tool for determining the patients with increased risk of vertebral fracture. The positive correlations between QCT and vertebral body compressive strength in cadaver studies are 0.72-0.74 [22,23].
The biggest source of error in X-ray CT systems is fat within the bone marrow: accuracy errors of up to 30%. The accuracy can be improved by carrying out scans at two different potentials (dual energy techniques); typically, 80 and 120 kVp are used. Kalender et al. [24] claim an accuracy error of 5% *in vitro*, but errors *in vivo* are likely to be larger. The effective radiation doses equivalent for QCT are 0.3 mSv for single energy techniques and 1 mSv for dual energy techniques, respectively [25].
In quantitative computed tomography (QCT), a thin transverse slice through the body is imaged. Under appropriate conditions, the image can be quantified to give a measure of volumetric bone mineral density (vBMD) (mg/cm3), and cancellous bone can be measured independently of surrounding cortical bone and aortic calcification. Developments have been concentrated in two directions: the construction of special equipment using a radionuclide source for measurements of the forearm, and the adaptation of X-ray CT machines installed for general radiology to measure vBMD. The attraction of the technique is that cancellous bone can be examined separately from cortical bone. It also gives a true
A dedicated forearm scanner was first described in the mid 1970s [16,17]. The photon source is 125I and is mounted in a gantry with a sodium iodide scintillation detector. A linear scan is performed at each of 48 angular positions. Computer reconstruction generates an image in which a region of interest in the cancellous bone of the distal ulna is selected. Since 1980s, QCT has been used as a means for non-invasive quantitative determination of bone mineral
A lateral plane projection scan is necessary for precise slice positioning through the centers of the vertebrae. Comparison between the CT Hounsfield numbers and a calibration standard scanned simultaneously allows bone density to be expressed in terms of the equivalent concentration of the material of the standard. Regions of interest within the vertebral bodies are selected: circular, elliptical, rectangular or other chosen areas are selected to include all the cancellous bone just inside the cortex. The relationship between the observed CT number and the true attenuation coefficient is subject to short- and longterm variation, so that it is necessary to scan the patient and a calibration standard simultaneously. Recently, simple standards with fewer components based on suspensions of calcium hydroxyapatite in plastic have been adopted. Comparison between the standard and the Hounsfield numbers of the trabecular region of the vertebral bodies allows bone density to be expressed in terms of the equivalent concentration of the material of the
Investigators reported the prediction of vertebral body compressive strength using QCT. In 1985, McBroom et al. [20] showed a strong positive correlation between QCT and apparent density of the vertebral trabecular bone but could find only suggestive, not quite significant, correlations between QCT and the vertebral body compressive strength. Cann et al. [21] showed that QCT evaluation of vertebral trabecular bone mineral density is a useful tool for determining the patients with increased risk of vertebral fracture. The positive correlations between QCT and vertebral body compressive strength in cadaver studies are 0.72-0.74
The biggest source of error in X-ray CT systems is fat within the bone marrow: accuracy errors of up to 30%. The accuracy can be improved by carrying out scans at two different potentials (dual energy techniques); typically, 80 and 120 kVp are used. Kalender et al. [24] claim an accuracy error of 5% *in vitro*, but errors *in vivo* are likely to be larger. The effective radiation doses equivalent for QCT are 0.3 mSv for single energy techniques and 1 mSv for
**3. Quantitative computed tomography (QCT)**
value for mineral density (mg/cm3) unlike other techniques.
of the spine [18,19].
standard.
[22,23].
dual energy techniques, respectively [25].
#### **4. Quantitative ultrasound (QUS) bone assessment method**
QUS bone assessment method has been recently introduced as an alternative for peripheral bone mass assessment, reflecting bone strength, bone density, and bone elasticity or fragility, and may be superior to aBMD by DXA [26]. The advantages of this method over Xray-based techniques, which include low cost, portability, and no radiation exposure, have encouraged the use of this method for defining a stage of development of osteoporosis and evaluating bone fracture risk.
There are several reports for assessing bone conditions *in vivo* using QUS method and apparatus. QUS devices can be classified mostly into 3 groups, related to the type of ultrasound transmission. Trabecular sound transmission is best for measuring the heel [27]. Cortical transverse transmission currently only is used in phalanx contact devices [28]. And cortical axial transmission presently is being investigated for use in phalanges, the radius, and the tibia [28]. Heel devices currently appear to have the most clinical applications, where QUS are being used and evaluated for the prediction of fracture risk, the diagnosis of osteoporosis, the initiation of osteoporosis treatment, the monitoring of osteoporosis treatment, and osteoporosis case finding. For these purposes, the recommended parameter of interest in clinical routine is a composite score, e.g., heel stiffness index or Quantitative Ultrasound Index (QUI) combining the results of broad-band ultrasound attenuation (BUA) and speed of sound (SOS), as measured in meters per second.
At the present time, there is good evidence that QUS can discriminate those with osteoporotic fractures from age-matched controls without osteoporotic fracture [29,30]. The power of heel QUS to predict fracture observed in cross-sectional studies has been confirmed prospectively in some populations as defined by sex, age, and ethnic background. This is particularly true of heel QUS and for hip and spinal fractures. However, because of methodological issues, it is difficult to compare studies. Nonetheless, it is possible to make the following generations. Using QUS of the heel, the increase in relative risk for each standard deviation decrease in stiffness index (SI) is approximately 2.0 for the hip and spine and roughly 1.5 for all fractures combined [31-41].
The evidence from studies is good that the heel QUS SI using QUS devices is predictive of hip fracture risk in Caucasian and Asian women over age 55 and of any fracture risk in Asian women over age 55. Cortical axial transmission devices have no prospectively proven clinical utility, although clinical use in adults of phalanx QUS devices using cortical transverse transmission is also limited. These results for heel QUS are roughly the same as for DXA by BMD in terms of hip and spine fracture risk per SD decrease [12,42]. Discordant results between heel QUS and DXA, which are not infrequent, are not necessarily an indication of methodological error but rather due to the independence between the 2 techniques.
Diagnosing osteoporosis using QUS is less supported by evidence and more complicated and problematic than assessing fracture risk. To start with, the T-score diagnostic criteria of - 2.5, classically used for DXA aBMD, cannot be applied to QUS without discrepancies in the numbers of women diagnosed with osteoporosis because of tremendous variations in QUS measurements by skeletal site, because different QUS devices yield different results, and because of the relatively poor correlation between heel QUS and hip/spine DXA measurements. If the prevalence of osteoporosis is defined as -2.5 SD from the mean
*Ex Vivo* and *In Vivo* Assessment of Vertebral Strength
yield strength determined.
and Vertebral Fracture Risk Assessed by Dual Energy X-Ray Absorptiometry 111
bone mineral density, this distribution is used to define bone material properties, and the FE method of analysis is used to determine structural properties of the whole or a part of the bone. This information is then used to predict risk of fracture under specified loading conditions. Specifically, the distribution of bone material properties determined noninvasively is used as input to a FE analysis of structural strength, and other parameters such as loading conditions and boundary conditions are also included in the model as needed. Using mathematical methods contained in commercially-available or specially written computer programs, the model of a bone can be incrementally loaded until failure, and the
A FE method based on data from CT has been applied to predict proximal femoral fracture [66-70]. CT-based FE method appears more predictive of femoral strength than QCT or DXA alone [66] and can predict proximal femoral fracture location [68]. Nonlinear FE method demonstrated improved predictions of femoral strength [69]. For the spine, CT-based nonlinear FE method was clinically applied to assess vertebral strength [71] and cadaver studies have been performed to evaluate the accuracy of CT-based FE method [72-77]. The cadaver studies have verified CT-based FE method predicts failure loads and fracture patterns for 10-mm-thick vertebral sections [72] and can predict *ex vivo* vertebral compressive strength better than aBMD [73,74] and QCT alone [75]. CT-based nonlinear FE method can accurately predict vertebral strength, fracture sites and distribution of minimum principal strain *ex vivo* [77]. Based on verification by the cadaver studies, FE method has been applied clinically to the assessment of chronic glucocorticoid treatment at the hip [78], as well as teriparatide and alendronate treatment for osteoporosis at the lumbar spine [79],
A study assessing vertebral fracture risk and medication effects on osteoporosis *in vivo* with CT-based nonlinear FE method showed that analyzed vertebral compressive strength had stronger discriminatory power for vertebral fracture than aBMD and vBMD, and detected alendronate effects at 3 months earlier than aBMD and vBMD [80]. The CV (coefficient of variation) for the measurement of vertebral compressive strength was 0.96% *ex vivo*. The
CT-based FE method predicts compressive bone strength accurately and is useful for assessing the risk of fracture and therapeutic effects on osteoporosis, and provides unique theories from a biomechanical perspective. This method also predicts bone strength under specified loading conditions such as those normally seen in activities of daily living [81,82].
This study was conducted at Tokyo University in Tokyo, Japan. The study protocol was
Twelve thoracolumbar (T11, T12, and L1) vertebrae with no skeletal pathologies were collected within 24 hours of death from 4 males (31, 55, 67, and 83 years old). Causes of death for the four donors were myelodysplastic syndrome, pneumonia, adult T-cell leukemia, and bladder cancer, respectively. All of the specimens were obtained at Tokyo University Hospital with the approval of the ethics committee and with informed consent. They were stored at –70 C after each step in the protocol. The vertebrae were disarticulated,
proving useful for assessing medication effects on bone strength.
**6. Assessment of vertebral strength** *ex vivo* **by DXA**
approved by the ethics committee.
effective radiation dose for assessing vertebral compressive strength is 3 mSv.
threshold for QUS, even within the same sample population, different QUS instruments and different skeletal sites generate prevalence estimates that vary as much as 10-fold, such as prevalence estimates among Caucasian women over age 65 ranging from 4 to 50% [43-46]. To overcome this dilemma, there is a need for predefined, device-specific diagnostic thresholds. One recommended system suggests calibrating QUS measurements with DXA results, the latter used as the "gold standard," so that an upper QUS threshold is set to identify osteoporosis with 90% sensitivity and a lower threshold is set to identify osteoporosis with 90% specificity [47]. Using such a system, one could identify osteoporosis with high probability in patients whose results fall below the lower threshold for QUS, where specificity exceeds 90%; between the upper and lower thresholds, the diagnosis of osteoporosis would be considered quite equivocal, so that another means of measurement, like DXA aBMD, would be highly recommended; and above the upper threshold for QUS, where the sensitivity of a value below the threshold is 90%, osteoporosis would be deemed unlikely.
Except in the case of a low-energy fractures of the hip or spine, when the fracture alone is adequate to require treatment, all currently published recommendations for the initiation of treatment for osteoporosis are based on DXA aBMD values; in no instance, to date, are the results of QUS the definitive parameter. Despite this, several studies have demonstrated high levels of correlation between heel trabecular sound transmission and aBMD at matched skeletal sites [48-50]. Moreover, both SOS and BUA, standard QUS measurements, are dependent on overall bone strength which, in turn, is related to bone density, architecture and turnover, and the extent of bone mineralization [48,50,51-56]. These factors likely work together to maintain the overall quality and strength of bone and to prevent fractures and other bone failure. QUS parameters of heel trabecular transverse transmission are highly correlated with bone strength [57-62]. Consequently, it is conceivable that QUS guidelines for treatment initiation could be created, especially if combined with the use of clinical risk factors [63]. But no randomized clinical trials have been published examining whether individuals identified as high risk for fracture by QUS respond to treatment.
#### **5. Finite element (FE) method based on data from computed tomography**
The finite element (FE) method, an advanced computer technique of structural stress analysis developed in engineering mechanics, was first introduced to orthopaedic biomechanics in 1972 to evaluate stressed in human bones [64]. Since then, this method has been used to study the mechanics of human bones [65]. In the early 1990s, the FE method of analyzing a bone for fracture risk using 3-dimensional CT data was developed.
The object of this method is to measure non-invasively the strength of an individual bone in an individual patient. This measurement can then be used to determine whether or not the bone will fracture under specified loading conditions such as those normally seen in daily living. It can also be used to estimate fracture risks under abnormal loading conditions such as occur in falling, jumping or during athletic events or heavy training regimens. This method uses the distribution of physical properties of bone measured non-invasively in an individual and mathematical analysis of that distribution to predict the risk that a bone may fracture under applied loads. The use of such methods relates to the clinical disease of osteoporosis, or in general metabolic bone diseases. In a primary application, 3-dimensional CT data acquired using a conventional CT scanner are used to determine the distribution of
threshold for QUS, even within the same sample population, different QUS instruments and different skeletal sites generate prevalence estimates that vary as much as 10-fold, such as prevalence estimates among Caucasian women over age 65 ranging from 4 to 50% [43-46]. To overcome this dilemma, there is a need for predefined, device-specific diagnostic thresholds. One recommended system suggests calibrating QUS measurements with DXA results, the latter used as the "gold standard," so that an upper QUS threshold is set to identify osteoporosis with 90% sensitivity and a lower threshold is set to identify osteoporosis with 90% specificity [47]. Using such a system, one could identify osteoporosis with high probability in patients whose results fall below the lower threshold for QUS, where specificity exceeds 90%; between the upper and lower thresholds, the diagnosis of osteoporosis would be considered quite equivocal, so that another means of measurement, like DXA aBMD, would be highly recommended; and above the upper threshold for QUS, where the sensitivity of a value below the threshold is 90%, osteoporosis would be deemed
Except in the case of a low-energy fractures of the hip or spine, when the fracture alone is adequate to require treatment, all currently published recommendations for the initiation of treatment for osteoporosis are based on DXA aBMD values; in no instance, to date, are the results of QUS the definitive parameter. Despite this, several studies have demonstrated high levels of correlation between heel trabecular sound transmission and aBMD at matched skeletal sites [48-50]. Moreover, both SOS and BUA, standard QUS measurements, are dependent on overall bone strength which, in turn, is related to bone density, architecture and turnover, and the extent of bone mineralization [48,50,51-56]. These factors likely work together to maintain the overall quality and strength of bone and to prevent fractures and other bone failure. QUS parameters of heel trabecular transverse transmission are highly correlated with bone strength [57-62]. Consequently, it is conceivable that QUS guidelines for treatment initiation could be created, especially if combined with the use of clinical risk factors [63]. But no randomized clinical trials have been published examining whether
individuals identified as high risk for fracture by QUS respond to treatment.
analyzing a bone for fracture risk using 3-dimensional CT data was developed.
**5. Finite element (FE) method based on data from computed tomography**
The finite element (FE) method, an advanced computer technique of structural stress analysis developed in engineering mechanics, was first introduced to orthopaedic biomechanics in 1972 to evaluate stressed in human bones [64]. Since then, this method has been used to study the mechanics of human bones [65]. In the early 1990s, the FE method of
The object of this method is to measure non-invasively the strength of an individual bone in an individual patient. This measurement can then be used to determine whether or not the bone will fracture under specified loading conditions such as those normally seen in daily living. It can also be used to estimate fracture risks under abnormal loading conditions such as occur in falling, jumping or during athletic events or heavy training regimens. This method uses the distribution of physical properties of bone measured non-invasively in an individual and mathematical analysis of that distribution to predict the risk that a bone may fracture under applied loads. The use of such methods relates to the clinical disease of osteoporosis, or in general metabolic bone diseases. In a primary application, 3-dimensional CT data acquired using a conventional CT scanner are used to determine the distribution of
unlikely.
bone mineral density, this distribution is used to define bone material properties, and the FE method of analysis is used to determine structural properties of the whole or a part of the bone. This information is then used to predict risk of fracture under specified loading conditions. Specifically, the distribution of bone material properties determined noninvasively is used as input to a FE analysis of structural strength, and other parameters such as loading conditions and boundary conditions are also included in the model as needed. Using mathematical methods contained in commercially-available or specially written computer programs, the model of a bone can be incrementally loaded until failure, and the yield strength determined.
A FE method based on data from CT has been applied to predict proximal femoral fracture [66-70]. CT-based FE method appears more predictive of femoral strength than QCT or DXA alone [66] and can predict proximal femoral fracture location [68]. Nonlinear FE method demonstrated improved predictions of femoral strength [69]. For the spine, CT-based nonlinear FE method was clinically applied to assess vertebral strength [71] and cadaver studies have been performed to evaluate the accuracy of CT-based FE method [72-77]. The cadaver studies have verified CT-based FE method predicts failure loads and fracture patterns for 10-mm-thick vertebral sections [72] and can predict *ex vivo* vertebral compressive strength better than aBMD [73,74] and QCT alone [75]. CT-based nonlinear FE method can accurately predict vertebral strength, fracture sites and distribution of minimum principal strain *ex vivo* [77]. Based on verification by the cadaver studies, FE method has been applied clinically to the assessment of chronic glucocorticoid treatment at the hip [78], as well as teriparatide and alendronate treatment for osteoporosis at the lumbar spine [79], proving useful for assessing medication effects on bone strength.
A study assessing vertebral fracture risk and medication effects on osteoporosis *in vivo* with CT-based nonlinear FE method showed that analyzed vertebral compressive strength had stronger discriminatory power for vertebral fracture than aBMD and vBMD, and detected alendronate effects at 3 months earlier than aBMD and vBMD [80]. The CV (coefficient of variation) for the measurement of vertebral compressive strength was 0.96% *ex vivo*. The effective radiation dose for assessing vertebral compressive strength is 3 mSv.
CT-based FE method predicts compressive bone strength accurately and is useful for assessing the risk of fracture and therapeutic effects on osteoporosis, and provides unique theories from a biomechanical perspective. This method also predicts bone strength under specified loading conditions such as those normally seen in activities of daily living [81,82].
#### **6. Assessment of vertebral strength** *ex vivo* **by DXA**
This study was conducted at Tokyo University in Tokyo, Japan. The study protocol was approved by the ethics committee.
Twelve thoracolumbar (T11, T12, and L1) vertebrae with no skeletal pathologies were collected within 24 hours of death from 4 males (31, 55, 67, and 83 years old). Causes of death for the four donors were myelodysplastic syndrome, pneumonia, adult T-cell leukemia, and bladder cancer, respectively. All of the specimens were obtained at Tokyo University Hospital with the approval of the ethics committee and with informed consent. They were stored at –70 C after each step in the protocol. The vertebrae were disarticulated,
*Ex Vivo* and *In Vivo* Assessment of Vertebral Strength
(DPX; Lunar, Madison, WI, USA).
**8. Discussion**
**7. Assessment of vertebral fracture risk** *in vivo* **by DXA**
Windows version 5.0 software (SAS Institute, Cary, NC, USA).
0.759 ± 0.207 g/cm2 (Mann-Whitney *U* test, *p* = 0.0255).
written informed consent in accordance with the Declaration of Helsinki.
and Vertebral Fracture Risk Assessed by Dual Energy X-Ray Absorptiometry 113
This study was conducted at Tokyo Metropolitan Geriatric Medical Center in Tokyo, Japan. The study protocol was approved by the ethics committee and each participant provided
The inclusion criteria included ambulatory postmenopausal Japanese women aged between 49 and 85 years old. Exclusion criteria included women with any disorders of bone and mineral metabolism other than postmenopausal osteoporosis, those who had any recent or current treatment with the potential to alter bone turnover or bone metabolism. Vertebral fracture was diagnosed based on lateral spine radiography. Radiographic vertebral fracture was defined if either the anterior or central height was 20% less than posterior height. A total of 123 eligible participants were enrolled in this cross-sectional study. For all participants, aBMD of the anteroposterior (AP) lumbar spine (L2-4) were measured by DXA
Logistic regression analysis was performed to estimate risk factors for vertebral fracture. L2- 4 aBMD was assessed using sensitivity and specificity curves to determine the optimal cutoff point as the vertebral fracture threshold. For each statistical analysis, differences were considered significant at *p*<0.05. Statistical analysis was performed using StatView for
The 123 women enrolled in the *in vivo* clinical study had a mean age of 71.8 ±7.4 years, mean height of 149.4 ±5.6 cm, and mean weight of 50.2 ±7.4 kg. Measured L2-4 aBMD was 0.816 ±0.191 g/cm2. Subjects were classified on the basis of prior vertebral fracture. Among the 123 women, 75 subjects did not have any vertebral fractures (nonfracture group) and 48 subjects already had vertebral fractures (fracture group). The average aBMD of the nonfracture group was 0.860 ± 0.166 g/cm2, which was greater than that of the fracture group at
Vertebral fractures were present in 39.0% of the total study population. Among the fracture group, vertebral fractures spontaneously developed in 29 women (spontaneous fracture group) and were caused by trauma in 19 women (traumatic fracture group). Among the 19 subjects in the traumatic fracture group, 18 women developed fracture following a fall from standing height, and 1 woman developed fracture following a fall down stairs. To exclude factors of trauma, 75 subjects in the nonfracture group and 29 subjects in the spontaneous fracture group were compared. aBMD (Mann-Whitney *U* test, *p*=0.0033) was significantly decreased in the spontaneous fracture group compared with the nonfracture group. Logistic regression analysis after adjustment for age and body weight revealed that aBMD reduction as risk factors associated with spontaneous vertebral fracture, the odds ratio per SD decrease was 1.83 with 1.13-3.26 of 95% confidence interval (*p*=0.0238). aBMD was also assessed by sensitivity and specificity curves. The nonfracture group and spontaneous fracture group (104 women in total) were assessed in a cross-sectional manner. The optimal point on the sensitivity and specificity curves used as the fracture threshold to predict spontaneous vertebral fractures
for aBMD was 0.816 g/cm2 with 69.0% sensitivity and 72.0% specificity (Fig. 2).
Bone strength primarily reflects bone density and bone quality, which are influenced by bone architecture, turnover, accumulation of damage, and mineralization [5]. Previous
and the discs were excised. Then the posterior elements of each vertebra were removed by cutting through the pedicles. The vertebrae were immersed in water and aBMD (g/cm2) of the vertebrae were measured by DXA (DPX; Lunar, Madison, WI, USA) in the supine position.
To assess vertebral strength, a quasi-static uniaxial compression test of each vertebra was conducted. To restrain the specimens for load testing, both upper and lower surfaces of the vertebrae were embedded in dental resin (Ostron; GC Dental Products Co., Aichi, Japan) so that the two surfaces were exactly parallel. Then the embedded specimens were placed on a mechanical testing machine (TENSILON UTM-2.5T; Orientec, Tokyo, Japan) and were compressed at a cross-head displacement rate of 0.5 mm per minute. A compression plate with a ball joint was used to apply a uniform load onto the upper surface of the specimen. The applied load was measured by a load cell (T-CLB-5-F-SR; T. S. Engineering, Kanagawa, Japan). The load was recorded using MacLab/4(AD Instruments, Castle Hill, NSW, Australia)at a sampling rate of 2 Hz. The measured vertebral strength was defined as the ultimate load achieved. Pearson's correlation analysis was used to evaluate correlations between the measured aBMD by DXA and the measured vertebral strength by mechanical testing.
The result from the *ex vivo* assessment, aBMD by DXA ranged from 0.287 to 0.705 g/cm2, while the measured vertebral strength by mechanical testing ranged from 1.54 to 4.62 kN. There were significant linear correlations between aBMD and the measured vertebral strength (r = 0.915, *p* < 0.0001) (Fig. 1).
Fig. 1. The experimentally measured vertebral strength versus aBMD measured by DXA. They were significantly correlated.
#### **7. Assessment of vertebral fracture risk** *in vivo* **by DXA**
112 Dual Energy X-Ray Absorptiometry
and the discs were excised. Then the posterior elements of each vertebra were removed by cutting through the pedicles. The vertebrae were immersed in water and aBMD (g/cm2) of the vertebrae were measured by DXA (DPX; Lunar, Madison, WI, USA) in the supine
To assess vertebral strength, a quasi-static uniaxial compression test of each vertebra was conducted. To restrain the specimens for load testing, both upper and lower surfaces of the vertebrae were embedded in dental resin (Ostron; GC Dental Products Co., Aichi, Japan) so that the two surfaces were exactly parallel. Then the embedded specimens were placed on a mechanical testing machine (TENSILON UTM-2.5T; Orientec, Tokyo, Japan) and were compressed at a cross-head displacement rate of 0.5 mm per minute. A compression plate with a ball joint was used to apply a uniform load onto the upper surface of the specimen. The applied load was measured by a load cell (T-CLB-5-F-SR; T. S. Engineering, Kanagawa, Japan). The load was recorded using MacLab/4(AD Instruments, Castle Hill, NSW, Australia)at a sampling rate of 2 Hz. The measured vertebral strength was defined as the ultimate load achieved. Pearson's correlation analysis was used to evaluate correlations between the measured aBMD by DXA and the measured vertebral strength by mechanical
The result from the *ex vivo* assessment, aBMD by DXA ranged from 0.287 to 0.705 g/cm2, while the measured vertebral strength by mechanical testing ranged from 1.54 to 4.62 kN. There were significant linear correlations between aBMD and the measured vertebral
Fig. 1. The experimentally measured vertebral strength versus aBMD measured by DXA.
position.
testing.
strength (r = 0.915, *p* < 0.0001) (Fig. 1).
They were significantly correlated.
This study was conducted at Tokyo Metropolitan Geriatric Medical Center in Tokyo, Japan. The study protocol was approved by the ethics committee and each participant provided written informed consent in accordance with the Declaration of Helsinki.
The inclusion criteria included ambulatory postmenopausal Japanese women aged between 49 and 85 years old. Exclusion criteria included women with any disorders of bone and mineral metabolism other than postmenopausal osteoporosis, those who had any recent or current treatment with the potential to alter bone turnover or bone metabolism. Vertebral fracture was diagnosed based on lateral spine radiography. Radiographic vertebral fracture was defined if either the anterior or central height was 20% less than posterior height. A total of 123 eligible participants were enrolled in this cross-sectional study. For all participants, aBMD of the anteroposterior (AP) lumbar spine (L2-4) were measured by DXA (DPX; Lunar, Madison, WI, USA).
Logistic regression analysis was performed to estimate risk factors for vertebral fracture. L2- 4 aBMD was assessed using sensitivity and specificity curves to determine the optimal cutoff point as the vertebral fracture threshold. For each statistical analysis, differences were considered significant at *p*<0.05. Statistical analysis was performed using StatView for Windows version 5.0 software (SAS Institute, Cary, NC, USA).
The 123 women enrolled in the *in vivo* clinical study had a mean age of 71.8 ±7.4 years, mean height of 149.4 ±5.6 cm, and mean weight of 50.2 ±7.4 kg. Measured L2-4 aBMD was 0.816 ±0.191 g/cm2. Subjects were classified on the basis of prior vertebral fracture. Among the 123 women, 75 subjects did not have any vertebral fractures (nonfracture group) and 48 subjects already had vertebral fractures (fracture group). The average aBMD of the nonfracture group was 0.860 ± 0.166 g/cm2, which was greater than that of the fracture group at 0.759 ± 0.207 g/cm2 (Mann-Whitney *U* test, *p* = 0.0255).
Vertebral fractures were present in 39.0% of the total study population. Among the fracture group, vertebral fractures spontaneously developed in 29 women (spontaneous fracture group) and were caused by trauma in 19 women (traumatic fracture group). Among the 19 subjects in the traumatic fracture group, 18 women developed fracture following a fall from standing height, and 1 woman developed fracture following a fall down stairs. To exclude factors of trauma, 75 subjects in the nonfracture group and 29 subjects in the spontaneous fracture group were compared. aBMD (Mann-Whitney *U* test, *p*=0.0033) was significantly decreased in the spontaneous fracture group compared with the nonfracture group. Logistic regression analysis after adjustment for age and body weight revealed that aBMD reduction as risk factors associated with spontaneous vertebral fracture, the odds ratio per SD decrease was 1.83 with 1.13-3.26 of 95% confidence interval (*p*=0.0238). aBMD was also assessed by sensitivity and specificity curves. The nonfracture group and spontaneous fracture group (104 women in total) were assessed in a cross-sectional manner. The optimal point on the sensitivity and specificity curves used as the fracture threshold to predict spontaneous vertebral fractures for aBMD was 0.816 g/cm2 with 69.0% sensitivity and 72.0% specificity (Fig. 2).
#### **8. Discussion**
Bone strength primarily reflects bone density and bone quality, which are influenced by bone architecture, turnover, accumulation of damage, and mineralization [5]. Previous
*Ex Vivo* and *In Vivo* Assessment of Vertebral Strength
micro-CT data with less radiation dose might be promising.
Disease. Osteoporos Int. 1997;7:390-406.
Organ Tech Rep Ser 1994;843:1-129.
Biomecha 1994; 9: 180-186.
1728.
osteoporotic fractures. Epidemiol Rev. 1985;7:178-208.
**9. References**
1997;87:398-403.
and Vertebral Fracture Risk Assessed by Dual Energy X-Ray Absorptiometry 115
This method assesses bone geometry and heterogeneous bone mass distribution as well as aBMD, but cannot detect microdamage and bone turnover. In clinical application, other parameters such as age and bone turnover markers should be included to assess the risk of fracture and therapeutic effects. Methods for assessing fracture risk and therapeutic effects
Prediction by FE method with a smaller element size using the data from CT scans with a thinner slice thickness and a smaller pixel size is more accurate. On the other hand, thinner CT slices lead to more radiation exposure in the clinical situation. To decrease radiation exposure as much as possible during CT scanning, optimization of the element size of the FE method was performed by assessing the accuracy of the FE method simulation [83]. With the limited resolution of currently available CT scanners, the micro-architecture of the bone cannot be precisely assessed. Micro-CT and synchrotron micro-CT visualize bone microstructure. However, obtaining micro-CT scans of a whole vertebra *in vivo* would be impossible with the currently available scanners. Also, use of thinner CT slices to obtain images leads to more radiation exposure. With future developments, FE method based on
[1] Cummings SR, Kelsey JL, Nevitt MC, O'Dowd KJ. Epidemiology of osteoporosis and
[2] Eastell R, Reid DM, Compston J, *et al*. Secondary prevention of osteoporosis: when should a non-vertebral fracture be a trigger for action? Q J Med 2001;94:575-597. [3] Wolinsky FD, Fitzgerald JF, Stump TE. The effect of hip fracture on mortality,
[5] NIH Consensus Development Panel on Osteoporosis Prevention, Diagnosis, and
[6] Kanis JA, Delmas P, Burckhardt P, Cooper C, Torgerson D. Guidelines for diagnosis and
[7] World Health Organization. Assessment of fracture risk and its application to screening
[8] Edmondston SJ, Singer KP, Day RE, *et al.* In-vitro relationships between vertebral body
[9] Cheng XG, Nicholson PH, Boonen S, *et al.* Prediction of vertebral strength in vitro by
[10] Myers BS, Arbogast KB, Lobaugh B, Harper KD, Richardson WJ, Drezner MK.
energy x-ray absorptiometry. J Bone Miner Res 1994;9:687-693.
[4] Melton LJ 3rd. Epidemiology of spinal osteoporosis. Spine 1997;22(suppl):2-11.
hospitalization, and functional status: a prospective study. Am J Public Health
Therapy. Osteoporosis prevention, diagnosis, and therapy. JAMA 2001;285:785-795.
management of osteoporosis. The European Foundation for Osteoporosis and Bone
for postmenopausal osteoporosis. Report of a WHO Study Group. World Health
density, size and compressive strength in the elderly thoracolumbar spine. Clin
spinal bone densitometry and calcaneal ultrasound. J Bone Miner Res 1997;12:1721-
Improved assessment of lumbar vertebral body strength using supine lateral dual-
on osteoporosis in the future might include other parameters as well as CT data.
Fig. 2. Sensitivity and specificity curves to determine the optimal cut-off point of aBMD measured by DXA to predict spontaneous vertebral fracture.
studies showed that aBMD explained 50-80% of vertebral strength [8-11] based on data that the correlations between aBMD and the measured vertebral strength were 0.51 to 0.80. In this study, the correlations between the measured values of aBMD and the vertebral strength were 0.915 and better than the previous studies. This *ex vivo* study showed that aBMD measurements in isolation might assess vertebral strength well.
In the treatment of osteoporosis, the target is to assess fracture risk and prevent fractures. This *in vivo* study showed that aBMD had high discriminatory power for spontaneous vertebral fracture. The cut-off value of aBMD for predicting vertebral fractures without trauma was 0.816 g/cm2, equivalent to -2.62 SD compared to young healthy Japanese women. Low trauma fractures such as a fall from a standing height are due to osteoporosis. The present assessment excluded the traumatic fracture group. Therefore, the threshold value was not for diagnosing osteoporosis, but for assessing spontaneous vertebral fracture risk.
This *ex vivo* and *in vivo* study showed that aBMD was a good parameter of vertebral strength and vertebral fracture risk. However, aBMD in isolation can only assess bone density and can not assess bone quality. Therefore, accuracy of assessing vertebral strength by aBMD is limited.
To improve accuracy of assessing vertebral strength and vertebral fracture risk, new method has been developed. CT-based nonlinear FE method can accurately predict vertebral strength, fracture sites and distribution of minimum principal strain *ex vivo* [77]. Based on verification by the cadaver studies, FE method has been applied clinically. A study assessing vertebral fracture risk and medication effects on osteoporosis *in vivo* with CT-based nonlinear FE method showed that analyzed vertebral compressive strength had stronger discriminatory power for vertebral fracture than aBMD and detected alendronate effects at 3 months earlier than aBMD [80].
This method assesses bone geometry and heterogeneous bone mass distribution as well as aBMD, but cannot detect microdamage and bone turnover. In clinical application, other parameters such as age and bone turnover markers should be included to assess the risk of fracture and therapeutic effects. Methods for assessing fracture risk and therapeutic effects on osteoporosis in the future might include other parameters as well as CT data.
Prediction by FE method with a smaller element size using the data from CT scans with a thinner slice thickness and a smaller pixel size is more accurate. On the other hand, thinner CT slices lead to more radiation exposure in the clinical situation. To decrease radiation exposure as much as possible during CT scanning, optimization of the element size of the FE method was performed by assessing the accuracy of the FE method simulation [83]. With the limited resolution of currently available CT scanners, the micro-architecture of the bone cannot be precisely assessed. Micro-CT and synchrotron micro-CT visualize bone microstructure. However, obtaining micro-CT scans of a whole vertebra *in vivo* would be impossible with the currently available scanners. Also, use of thinner CT slices to obtain images leads to more radiation exposure. With future developments, FE method based on micro-CT data with less radiation dose might be promising.
#### **9. References**
114 Dual Energy X-Ray Absorptiometry
Fig. 2. Sensitivity and specificity curves to determine the optimal cut-off point of aBMD
studies showed that aBMD explained 50-80% of vertebral strength [8-11] based on data that the correlations between aBMD and the measured vertebral strength were 0.51 to 0.80. In this study, the correlations between the measured values of aBMD and the vertebral strength were 0.915 and better than the previous studies. This *ex vivo* study showed that
In the treatment of osteoporosis, the target is to assess fracture risk and prevent fractures. This *in vivo* study showed that aBMD had high discriminatory power for spontaneous vertebral fracture. The cut-off value of aBMD for predicting vertebral fractures without trauma was 0.816 g/cm2, equivalent to -2.62 SD compared to young healthy Japanese women. Low trauma fractures such as a fall from a standing height are due to osteoporosis. The present assessment excluded the traumatic fracture group. Therefore, the threshold value was
This *ex vivo* and *in vivo* study showed that aBMD was a good parameter of vertebral strength and vertebral fracture risk. However, aBMD in isolation can only assess bone density and can not assess bone quality. Therefore, accuracy of assessing vertebral strength by aBMD is
To improve accuracy of assessing vertebral strength and vertebral fracture risk, new method has been developed. CT-based nonlinear FE method can accurately predict vertebral strength, fracture sites and distribution of minimum principal strain *ex vivo* [77]. Based on verification by the cadaver studies, FE method has been applied clinically. A study assessing vertebral fracture risk and medication effects on osteoporosis *in vivo* with CT-based nonlinear FE method showed that analyzed vertebral compressive strength had stronger discriminatory power for vertebral fracture than aBMD and detected alendronate effects at 3
not for diagnosing osteoporosis, but for assessing spontaneous vertebral fracture risk.
measured by DXA to predict spontaneous vertebral fracture.
limited.
months earlier than aBMD [80].
aBMD measurements in isolation might assess vertebral strength well.
*Ex Vivo* and *In Vivo* Assessment of Vertebral Strength
495.
556.
2008;11:163-187.
1996;348:511-514.
Med 1997;157:629-634.
Miner Res 1995;10:341-345.
devices. London, Martin Dunitz. 1999;145-162.
women: A prospective study. BMJ 1990;301:638-641.
and Vertebral Fracture Risk Assessed by Dual Energy X-Ray Absorptiometry 117
[28] Njeh CF, Hans D, Fuerst T, Gluer CC, Genant HK (ed.), Quantitative ultrasound:
[29] Durosier C, Hans D, Krieg MA, Schott AM. Prediction and discrimination of
[30] Krieg MA, Barkmann R, Gonnelli S, *et al.* Quantitative ultrasound in the management of
[31] Porter RW, Miller CG, Grainger D, Palmer SB. Prediction of hip fracture in elderly
[32] Heaney RP, Avioli LV, Chesnut CH 3rd, Lappe J, Recker RR, Brandenburger GH.
[33] Hans D, Dargent-Molina P, Schott AM, *et al*. Ultrasonographic heel measurements to
[34] Bauer DC, Gluer CC, Cauley JA, *et al*. Broadband ultrasound attenuation predicts
[35] Pluijm SM, Graafmans WC, Bouter LM, Lips P. Ultrasound measurements for the
[36] Fujiwara S, Sone T, Yamazaki K, *et al*. Heel bone ultrasound predicts non-spine fracture
[37] Schott AM, Hans D, Duboeuf F, *et al*. Quantitative ultrasound parameters as well as
in elderly women. Results from the EPIDOS study. Bone 2005;37:858-863. [38] Gluer MG, Minne HW, Gluer CC, *et al*. Prospective identification of postmenopausal
[39] Krieg MA, Cornuz J, Ruffieux C, *et al*. Prediction of hip fracture risk by quantitative
[40] Diez-Perez, Gonzalez-Macias, Marin F, *et al*. Prediction of absolute risk of non-spinal
[41] Bauer DC, Ewing S, Cauley J, Ensrud K, Cummings S, Orwoll E. Quantitative
[42] Durosier C, Hans D, Krieg MA, Schott AM. Prediction and discrimination of
in Japanese men and women. Osteoporos int 2005;16:2107-2112.
PIOS study. J Clin Densitom 2005;8:386-395.
Miner Res 2006;21:1457-1463.
Osteoporos Int 2007;18:771-777.
2007;18:629-639.
495.
assessment of osteoporisis and bone status – Non-heel quantitative ultrasound
osteoporotic hip fracture in postmenopausal women. J Clin Densitom 2006;9:475-
osteoporosis: The 2007 ISCD official positions. J Clin Densitom Assess Skelet Health
Ultrasound velocity, through bone predicts incident vertebral deformity. J Bone
predict hip fracture in elderly women: The EPIDOS prospective study. Lancet
fractures strongly and independently of densitometry in older women. A prospective study. Study of Osteoporotic Fractures Research Group. Arch Intern
prediction of osteoporotic fractures in elderly people. Osteoporos Int 1999;9:550-
bone mineral density are better predictors of trochanteric than cervical hip fractures
osteoporotic women at high vertebral fracture risk by radiography, bone densitometry, quantitative ultrasound, and laboratory findings: results from the
ultrasound in more than 7000 Swiss women > or = 70 years of age: Comparison of three technologically different bone ultrasound devices in the SEMOF study. J Bone
fractures using clinical risk factors and heel quantitative ultrasound. Osteoporos Int
ultrasound predicts hip and non-spine fracture in men: The MrOS study.
osteoporotic hip fracture in postmenopausal women. J Clin Densitom 2006;9:475-
[11] Bjarnason K, Hassager C, Svendsen OL, Stang H, Christiansen C. Anteroposterior and
[12] Marshall D, Johnell O, Wedel H. Meta-analysis of how well measures of bone mineral density predict occurrence of osteoporotic fractures. BMJ 1996;312:1254-1259. [13] Vogel JM. Bone mineral measurement: Skylab experiment M-078. Acta Astronaut
[14] Oyster N, Smith FW. A postmortem correlation of four techniques of assessment of osteoporosis with force of bone compression. Calcif Tissue Int 1988;43:77-82. [15] Uebelhart D, Duboeuf F, Meunier PJ, Delmas PD. Lateral dual-photon absorptiometry: a
[16] Rüegsegger P, Elsasser U, Anliker M, Gnehm H, Kind H, Prader A. Quantification of bone mineralization using computed tomography. Radiology 1976;121:93-97. [17] Isherwood I, Rutherford RA, Pullan BR, Adams PH. Bone-mineral estimation by computer-assisted transverse axial tomography. Lancet 1976;2:712-715. [18] Genant HK, Boyd D, Rosenfeld D, Abols Y, Cann CE. Computed tomography. Non-
[19] Powell MR, Kolb FO, Genant HK, Cann CE, Stebler BG. Comparison of dual photon
[20] McBroom RJ, Hayes WC, Edwards WT, *et al.* Prediction of vertebral body compressive
[21] Cann CE, Genant HK, Kolb FO, Ettinger B. Quantitative computed tomography for
[22] Mosekilde L, Bentzen SM, Ortoft G, Jorgensen J. The predictive value of quantitative
[23] Eriksson SA, Isberg BO, Lindgren JU. Prediction of vertebral strength by dual photon
[24] Kalender WA, Klotz E, Suess C. Vertebral bone mineral analysis: an integrated
[25] Huda W, Morin RL. Patient doses in bone mineral densitometry. Br J Radiol
[26] Hans D, Arlot ME, Schott AM, Roux JP, Kotzki PO, Meunier PJ. Do ultrasound
prediction of vertebral fracture risk. Bone 1985;6:1-7.
approach with CT. Radiology 1987;164:419-423.
London, Martin Dunitz. 1999;109-144.
hip and forearm measurement. Osteoporosis Int 1996;6:37-42.
1975;2:129-139.
Miner Res 1990;5:525-531.
Press 1981;121-149.
A:1206-1214.
Bone 1989;10:465-470.
1989;44:243-250.
1996;69:422-425.
Excerpta Medica 1983;58-60.
lateral spinal DXA for the assessment of vertebral body strength: comparison with
new technique to measure the bone mineral density at the lumbar spine. J Bone
invasive measurements of bone mass and their clinical application. Florida, CRC
absorbtiometry and quantitative computer tomography of the lumbar spine in the same subjects. Clinical disorders of bone and mineral metabolism. Amsterdam,
fracture using quantitative computed tomography. J Bone Joint Surg 1985;67-
computed tomography for vertebral body compressive strength and ash density.
absorptiometry and quantitative computed tomography. Calcif Tissue Int
measurements on the os calcis reflect more the bone microarchitecture than the bone mass? A two-dimensional histomorphometric study. Bone 1995;16:295-300. [27] Njeh CF, Hans D, Fuerst T, Gluer CC, Genant HK (ed.), Quantitative ultrasound:
assessment of osteoporisis and bone status – Calcaneal quantitative ultrasound.
*Ex Vivo* and *In Vivo* Assessment of Vertebral Strength
fracture. Osteoporosis Int 2007;18:1651-1659.
the first decade. J Biomech 1983;16:385-409.
Phys 1997;19:742-747.
Int 1997;7:471-477.
1991;179:669-674.
Res 1998;16:300-308.
2003;33:744-750.
Osteoporos Int 2005;16:642-650.
and Vertebral Fracture Risk Assessed by Dual Energy X-Ray Absorptiometry 119
[61] Han S, Medige J, Faran K, Feng Z, Ziv I. The ability of quantitative ultrasound to predict
[62] Njeh CF, Kuo CW, Langton CM, Atrah HI, Boivin CM. Prediction of human femoral
[63] Durosier C, Hans D, Krieg MA, *et al*. Combining clinical factors and quantitative
[64] Brekelmans WA, Poort HW, Slooff TJ. A new method to analyse the mechanical
[65] Huiskes R, Chao EY. A survey of finite element analysis in orthopedic biomechanics:
[66] Cody DD, Gross GJ, Hou FJ, *et al.* Femoral strength is better predicted by finite element
[67] Keyak JH, Rossi SA, Jones KA, *et al.* Prediction of femoral fracture load using automated
[68] Keyak JH, Rossi SA, Jones KA, *et al.* Prediction of fracture location in the proximal
[69] Keyak JH. Improved prediction of proximal femoral fracture load using nonlinear finite
[70] Bessho M, Ohnishi I, Matsuyama J, *et al.* Prediction of strength and strain of the
[71] Faulkner KG, Cann CE, Hasegawa BH. Effect of bone distribution on vertebral strength:
[72] Silva MJ, Keaveny TM, Hayes WC. Computed tomography-based finite element
[73] Martin H, Werner J, Andresen R, *et al.* Noninvasive assessment of stiffness and failure load of human vertebrae from CT-data. Biomed Tech 1998;43:82-88. [74] Buckley JM, Loo K, Motherway J. Comparison of quantitative computed tomographybased measures in predicting vertebral strength. Bone 2007;40:767-774. [75] Crawford RP, Cann CE, Keaveny TM. Finite element models predict in vitro vertebral
[76] Liebschner MA, Kopperdahl DL, Rosenberg D, *et al.* Finite element modeling of the
[77] Imai K, Ohnishi I, Bessho M, *et al.* Nonlinear finite element model predicts vertebral
[78] Lian KC, Lang TF, Keyak JH, *et al.* Differences in hip quantitative computed
proximal femur by a CT-based finite element method. J Biomech 2007;40:1745-1753.
assessment with patient-specific nonlinear finite element analysis. Radiology
analysis predicts failure loads and fracture patterns for vertebral sections. J Orthop
body compressive strength better than quantitative computed tomography. Bone
tomography (QCT) measurements of bone mineral density and bone strength between glucocorticoid-treated and glucocorticoid-naive postmenopausal women.
femur using finite element models. Med Eng Phys 2001;23:657-664.
behaviour of skeletal parts. Acta Orthop Scand 1972;43:301-317.
models than QCT and DXA. J Biomech 1999;32:1013-1020.
finite element modeling. J Biomech 1998;31:125-133.
element models. Med Eng Phys 2001;23:165-173.
human thoracolumbar spine. Spine 2003;28:559-565.
bone strength and fracture site. Spine 2006;31:1789-1794.
the mechanical properties of trabecular bone under different strain rates. Med Eng
bone strength using ultrasound velocity and BMD. An *in vitro* study. Osteoporosis
ultrasound improves the detection of women both at low and high risk for hip
[43] Frost ML, Blake GM, Fogelman I. Can the WHO criteria for diagnosing osteoporosis be applied to calcaneal quantitative ultrasound? Osteoporosis Int 2000;11:321-330. [44] Faulkner KG, von Stetten E, Miller P. Discordance in patient classification using T-
[45] Damilakis J, Perisinakis K, Gourtsoyiannis N. Imaging ultrasonometry of the calcaneus:
[46] Knapp KM, Blake GM, Spector TD, Fogelman I. Can the WHO definition of
[47] Clowes JA, Peel NF, Eastell R. Device-specific thresholds to diagnose osteoporosis at the
[48] Siffert RS, Kaufman JJ. Ultrasonic bone assessment: The time has come. Bone 2007;40:5-
[49] Roux C, Fournier B, Laugier P, *et al*. Broadband ultrasound attenuation imaging: A new imaging method in osteoporosis. J Bone Miner Res 1996;11:1112-1118. [50] Hans D, Arlot ME, Schott AM, Roux JP, Kotzki PO, Meunier PJ. Do ultrasound
[52] Ammann P, Rizzoli R. Bone strength and its determinants. Osteoporosis Int 2003;14:S13-
[53] Hans D, Wu C, Njeh CF, *et al*. Ultrasound velocity of trabecular cubes reflects mainly
[54] Gluer CC. Quantitative ultrasound - it is time to focus research efforts. Bone 2007;40:9-
[55] Gluer CC, Wu CY, Genant HK. Broadband ultrasound attenuation signals depend on trabecular orientation: An *in vitro* study. Osteoporosis Int 1993;3:185-191. [56] Gluer CC, Wu CY, Jergas M, Goldstein SA, Genant HK. Three quantitative ultrasound
[57] Bouxsein ML, Coan BS, Lee SC. Prediction of the strength of the elderly proximal femur
[58] Bouxsein ML, Radloff SE. Quantitative ultrasound of the calcaneus reflects the
[59] Lochmuller EM, Burklein D, Kuhn V, *et al*. Mechanical strength of the thoracolumbar
computed tomography, and quantitative ultrasound. Bone 2002;31:77-84. [60] Hakulinen MA, Toyras J, Saarakkala S, Hirvonen J, Kroger H, Jurvelin JS. Ability of
by bone mineral density and quantitative ultrasound measurements of the heel and
mechanical properties of calcaneal trabecular bone. J Bone Miner Res 1997;12:839-
spine in the elderly: Prediction from in situ dual-energy X-ray absorptiometry, quantitative computed tomography, upper and lower limb peripheral quantitative
ultrasound backscattering to predict mechanical properties of bovine trabecular
Optimum T-score thresholds for the identification of osteoporotic subjects. Calcif
osteoporosis be applied to multisite axial transmission quantitative ultrasound?
proximal femur: An approach to interpreting peripheral bone measurements in
measurements on the os calcis reflect more the bone microarchitecture than the bone mass?: A two-dimensional histomorphometric study. Bone 1995;16:295-300. [51] Seeman E, Delmas PD. Bone quality - the material and structural basis of bone strength
scores. J Clin Densitom 1999;2:343-350.
Tissue Int 2001;68:219-224.
8.
S18.
13.
846.
tibia. Bone 1999;25:49-54.
Osteoporosis Int 2004;15:367-374.
clinical practice. Osteoporosis Int 2006;17:1293-1302.
and fragility. N Engl J Med 2006;354:2250-2261.
bone. Ultrasound Med Biol 2004;30:919-927.
bone density and elasticity. Calcif Tissue Int 1999;64:18-23.
parameters reflect bone structure. Calcif Tissue Int 1994;55:46-52.
**9**
*France*
**Dietary Protein and Bone Health**
*UMR-914 INRA-AgroParisTech, Nutrition Physiology and Ingestive Behavior, Paris,*
Dietary proteins represent 10 to 20% of energy consumption. The recommended daily minimum intake of protein and amino acids in adults is 0.8 g per kg of body weight. However, no upper limit has been identified. In industrialized countries, the main sources of protein are milk, eggs and meat. The nutritional value of protein is influenced by several factors, especially the amino acid (AA) composition, protein digestibility, protein digestion kinetics and the ability to transfer AA for protein synthesis. Diets based on either animal or vegetable products supply proteins of differing quality in differing quantities. Plant proteins are often deficient or low in some specific indispensable AAs. Soy protein is reported as a "complete" protein but its overall indispensable AA content is relatively low (~85% lower
Epidemiological data support a positive association between protein intake and bone health. Protein is the precursor of AAs used for bone matrix protein synthesis. Moreover, studies evaluating the relationship between dietary protein and bone turnover support the hypothesis that high protein intake may decrease bone resorption. IGF-1 is a key mediator of bone growth (Geusens & Boonen, 2002) and dietary protein is an important regulator of circulating IGF-1 levels (Bonjour *et al.*, 2001). However, protein is also known to be calciuric, though the origin of the increased calcium excretion is debated. According to the acid ash hypothesis, the protein-induced acid load would have a deleterious effect on bone (Barzel & Massey, 1998). Finally, some proteins, due to their amino acid content, have specific effects on bone metabolism by acting directly on bone cells or through
Even if the effect of nutrition on bone is not as dramatic as that of pharmaceuticals, it can be of great help when considering bone loss. Indeed nutrition is a lower-cost, longer-term, higher-compliance and wider-spread approach than pharmaceutical treatment. The nutritional intervention can be done alone preventively or in combination with a therapy in more severe cases. Some nutritional components such as calcium and vitamin D are recognized to have a positive effect on bone. However, the effect of protein on bone health is
This chapter reviews the literature on the subject, giving an overview of the data available and comparing the proposed mechanisms of action. Attention is drawn to protein quality and to the specific effect of some soy, collagen and milk peptides on bone metabolism.
**1. Introduction**
indirect pathways.
much more debated.
than milk) (Wilson & Wilson, 2006).
Anne Blais, Emilien Rouy and Daniel Tomé
### **Dietary Protein and Bone Health**
Anne Blais, Emilien Rouy and Daniel Tomé
*UMR-914 INRA-AgroParisTech, Nutrition Physiology and Ingestive Behavior, Paris, France*
#### **1. Introduction**
120 Dual Energy X-Ray Absorptiometry
[79] Keaveny TM, Donley DW, Hoffmann PF, *et al.* Effects of teriparatide and alendronate on
[80] Imai K, Ohnishi I, Matsumoto T, Yamamoto S, Nakamura K. Assessment of vertebral
[81] Matsumoto T, Ohnishi I, Bessho M, Imai K, Ohashi S, Nakamura K. Prediction of
[82] Bessho M, Ohnishi I, Matsumoto T, *et al.* Prediction of proximal femur strength using a
with osteoporosis. J Bone Miner Res 2007;22:149-157.
Osteoporosis Int 2009;20:801-810.
element model. Spine 2008;33:27-32.
2009;34:1464-1469.
vertebral strength as assessed by finite element modeling of QCT scans in women
fracture risk and therapeutic effects of alendronate in postmenopausal women using a quantitative computed tomography-based nonlinear finite element method.
vertebral strength under loading conditions occurring in activities of daily living using a computed tomography-based nonlinear finite element method. Spine
CT-based nonlinear finite element method: differences in predicted fracture load and site with changing load and boundary conditions. Bone 2009;45:226-231. [83] Imai K, Ohnishi I, Yamamoto S, Nakamura K. *In vivo* assessment of lumbar vertebral
strength in elderly women using computed tomography-based nonlinear finite
Dietary proteins represent 10 to 20% of energy consumption. The recommended daily minimum intake of protein and amino acids in adults is 0.8 g per kg of body weight. However, no upper limit has been identified. In industrialized countries, the main sources of protein are milk, eggs and meat. The nutritional value of protein is influenced by several factors, especially the amino acid (AA) composition, protein digestibility, protein digestion kinetics and the ability to transfer AA for protein synthesis. Diets based on either animal or vegetable products supply proteins of differing quality in differing quantities. Plant proteins are often deficient or low in some specific indispensable AAs. Soy protein is reported as a "complete" protein but its overall indispensable AA content is relatively low (~85% lower than milk) (Wilson & Wilson, 2006).
Epidemiological data support a positive association between protein intake and bone health. Protein is the precursor of AAs used for bone matrix protein synthesis. Moreover, studies evaluating the relationship between dietary protein and bone turnover support the hypothesis that high protein intake may decrease bone resorption. IGF-1 is a key mediator of bone growth (Geusens & Boonen, 2002) and dietary protein is an important regulator of circulating IGF-1 levels (Bonjour *et al.*, 2001). However, protein is also known to be calciuric, though the origin of the increased calcium excretion is debated. According to the acid ash hypothesis, the protein-induced acid load would have a deleterious effect on bone (Barzel & Massey, 1998). Finally, some proteins, due to their amino acid content, have specific effects on bone metabolism by acting directly on bone cells or through indirect pathways.
Even if the effect of nutrition on bone is not as dramatic as that of pharmaceuticals, it can be of great help when considering bone loss. Indeed nutrition is a lower-cost, longer-term, higher-compliance and wider-spread approach than pharmaceutical treatment. The nutritional intervention can be done alone preventively or in combination with a therapy in more severe cases. Some nutritional components such as calcium and vitamin D are recognized to have a positive effect on bone. However, the effect of protein on bone health is much more debated.
This chapter reviews the literature on the subject, giving an overview of the data available and comparing the proposed mechanisms of action. Attention is drawn to protein quality and to the specific effect of some soy, collagen and milk peptides on bone metabolism.
Dietary Protein and Bone Health 123
The effect of protein intake on IGF-1 level could involve Calcium-Sensing Receptor (CaSR) which was also reported to be implicated in the calciuric effect of protein (Conigrave *et al.*, 2008). *In vitro* investigations showed that CaSR is expressed in hepatocytes (Canaff *et al.*, 2001). The activation of CaSR in the liver by AA could explain the increased IGF-1 level due to protein consumption. Moreover, osteoblasts also express CaSR and produce IGF-1 under its regulation. CaSR also upregulates numerous bone forming mechanisms and inhibits bone resorption (Marie, 2010). Aromatic AA supplementation induces an increase in IGF-1 while branched chain AAs have no effect (Dawson-Hughes *et al.*, 2007). This is consistent with the findings of Conigrave *et al*., that aromatic AAs are the most potent activators of
There is no doubt that protein induces an increase of IGF-1 level. By doing so and according to what is known of IGF-1, protein is likely to have an anabolic effect on bone. The proposed mechanism would be that a high protein intake (especially one rich in aromatic AAs) would activate CaSR in hepatocytes and osteoblasts and stimulate IGF-1 production, which would
Bone turnover markers are measured in blood or urine; they give information on bone formation or on bone resorption. Both types of markers are needed in a study in order to
The recent meta-analysis of Darling *et al*. concludes that there is no clear evidence of an effect of protein intake on bone markers (Darling *et al.*, 2009). Some studies reported an increase in bone resorption while bone formation remained stable (Kerstetter *et al.*, 1999; Roughead *et al.*, 2003; Zwart *et al.*, 2005). However, in one of them, the authors report an increase of hydroxyproline (a bone resorption marker) that could be related to the protein source: a collagen-rich meat (Roughead *et al.*, 2003). On the other hand, three studies considering both formation and resorption markers concluded that there was a positive effect of protein on bone. One of them observed a decrease of desoxypyridinoline (a bone resorption marker) with stable formation, indicating a positive decoupling (Hunt *et al.*, 2009). In the second study, both formation and resorption markers increased in a proteinsupplemented group compared to control during 6 months of exercise. According to the authors, bone formation will increase faster than bone resorption over time (Ballard *et al.*, 2005). In the final study, a one year so**y** protein supplementation only increased bone formation markers while resorption markers remained at the same level (Arjmandi *et al.*, 2005). Finally, two intervention studies reported a decrease of resorption with protein supplementation but provided no data on bone formation (Dawson-Hughes *et al.*, 2004; Ince
Urinary calcium excretion is the major pathway for body calcium loss, along with that of feces and sweat. A high urinary calcium level has been linked to increased bone loss
CaSR, while branched chain AAs are the less potent ones (Conigrave *et al.*, 2000).
in turn exert an anabolic effect on bone.
**3. Protein and calcium metabolism**
**2.3 Bone turnover markers**
assess bone turnover.
*et al.*, 2004).
**3.1 Urine calcium**
(Giannini *et al.*, 2003).
#### **2. Action of protein on bone**
The mechanism underlying the debate on the effect of protein on bone is the hormonal anabolic effect. It is based on the increase of bone anabolic hormone IGF-1 with the consumption of protein.
#### **2.1 Bone protein metabolism and turnover**
Type I collagen is the major structural protein distributed throughout the whole body accounting for 25% of total body protein and for 80% of total conjunctive tissue in humans. The different factors involved in bone strength include not only mineral density, crystal characteristics, micro architecture, geometry and morphology (size and shape), but also protein matrix and collagen fiber quality. Cortical bone strength, whose main role is to protect bone integrity, is influenced not only by porosity, presence of micro damage and mineral composition, but also by the orientation of collagen fibers, extent and nature of collagen cross-linking and number and composition of cement lines (Burr, 2002; Currey, 2001; Seeman & Delmas, 2006; Turner, 2006). Collagen is an important component of bone, being the main extra cellular matrix protein for calcification and playing a role in osteoblast differentiation (Takeuchi *et al.*, 1996, 1997).
Markers of bone resorption, which measure the release of peptide resulting of the degradation of mature modified type 1 collagen into the serum and urine, are commonly used to evaluate the relationship between dietary protein and bone turnover. However, to better understand the effect of feeding on bone collagen turnover, a better knowledge of the physiology of bone collagen synthesis is needed (Babraj Smith 2005).
#### **2.2 Hormonal anabolic effect**
Protein has an anabolic effect on bone metabolism by upregulating Insulin-like Growth Factor 1 (IGF-1). IGF-1 is a peptide hormone mainly produced by the liver under the action of the Growth Hormone. A small percentage of IGF-1 is also produced locally by different cell types throughout the body, including osteoblasts (Ohlsson *et al.*, 2009). There is good evidence of the anabolic effect of IGF-1 on bone and muscle, especially when considering age-related disorders (Perrini *et al.*, 2010). Studies measuring IGF-1 levels under different protein intakes consistently report that a high protein intake increases IGF-1. This fact has been shown over a six-month period in young exercising subjects (Ballard *et al.*, 2005) and evidence also exists in older populations of both sexes (Arjmandi *et al.*, 2005; Dawson-Hughes *et al.*, 2004; Hunt *et al.*, 2009).
There is some evidence that this hormonal response is based on the quality of the protein ingested. Dawson-Hughes and colleagues showed that a fivefold increase of aromatic AA intake (phenylalanine and histidine) for 24 days induced an increase of IGF-1 level. On the other hand, a fivefold increase of branched chain AA (leucine and isoleuine) induced no effect on IGF-1, indicating that the protein intake effect on IGF-1 is probably AA-dependent (Dawson-Hughes *et al.*, 2007). Moreover, two studies comparing milk and soy protein reported that soy induces a greater increase of IGF-1 (Arjmandi *et al.*, 2003; Khalil *et al.*, 2002). However, it is important to note that the soy protein contained isoflavone in both studies, thus the effect on IGF-1 could be attributed to this molecule rather than to the AA content.
The mechanism underlying the debate on the effect of protein on bone is the hormonal anabolic effect. It is based on the increase of bone anabolic hormone IGF-1 with the
Type I collagen is the major structural protein distributed throughout the whole body accounting for 25% of total body protein and for 80% of total conjunctive tissue in humans. The different factors involved in bone strength include not only mineral density, crystal characteristics, micro architecture, geometry and morphology (size and shape), but also protein matrix and collagen fiber quality. Cortical bone strength, whose main role is to protect bone integrity, is influenced not only by porosity, presence of micro damage and mineral composition, but also by the orientation of collagen fibers, extent and nature of collagen cross-linking and number and composition of cement lines (Burr, 2002; Currey, 2001; Seeman & Delmas, 2006; Turner, 2006). Collagen is an important component of bone, being the main extra cellular matrix protein for calcification and playing a role in osteoblast
Markers of bone resorption, which measure the release of peptide resulting of the degradation of mature modified type 1 collagen into the serum and urine, are commonly used to evaluate the relationship between dietary protein and bone turnover. However, to better understand the effect of feeding on bone collagen turnover, a better knowledge of the
Protein has an anabolic effect on bone metabolism by upregulating Insulin-like Growth Factor 1 (IGF-1). IGF-1 is a peptide hormone mainly produced by the liver under the action of the Growth Hormone. A small percentage of IGF-1 is also produced locally by different cell types throughout the body, including osteoblasts (Ohlsson *et al.*, 2009). There is good evidence of the anabolic effect of IGF-1 on bone and muscle, especially when considering age-related disorders (Perrini *et al.*, 2010). Studies measuring IGF-1 levels under different protein intakes consistently report that a high protein intake increases IGF-1. This fact has been shown over a six-month period in young exercising subjects (Ballard *et al.*, 2005) and evidence also exists in older populations of both sexes (Arjmandi *et al.*, 2005; Dawson-
There is some evidence that this hormonal response is based on the quality of the protein ingested. Dawson-Hughes and colleagues showed that a fivefold increase of aromatic AA intake (phenylalanine and histidine) for 24 days induced an increase of IGF-1 level. On the other hand, a fivefold increase of branched chain AA (leucine and isoleuine) induced no effect on IGF-1, indicating that the protein intake effect on IGF-1 is probably AA-dependent (Dawson-Hughes *et al.*, 2007). Moreover, two studies comparing milk and soy protein reported that soy induces a greater increase of IGF-1 (Arjmandi *et al.*, 2003; Khalil *et al.*, 2002). However, it is important to note that the soy protein contained isoflavone in both studies, thus the effect on IGF-1 could be attributed to this molecule rather than to the AA
physiology of bone collagen synthesis is needed (Babraj Smith 2005).
**2. Action of protein on bone**
**2.1 Bone protein metabolism and turnover**
differentiation (Takeuchi *et al.*, 1996, 1997).
**2.2 Hormonal anabolic effect**
Hughes *et al.*, 2004; Hunt *et al.*, 2009).
content.
consumption of protein.
The effect of protein intake on IGF-1 level could involve Calcium-Sensing Receptor (CaSR) which was also reported to be implicated in the calciuric effect of protein (Conigrave *et al.*, 2008). *In vitro* investigations showed that CaSR is expressed in hepatocytes (Canaff *et al.*, 2001). The activation of CaSR in the liver by AA could explain the increased IGF-1 level due to protein consumption. Moreover, osteoblasts also express CaSR and produce IGF-1 under its regulation. CaSR also upregulates numerous bone forming mechanisms and inhibits bone resorption (Marie, 2010). Aromatic AA supplementation induces an increase in IGF-1 while branched chain AAs have no effect (Dawson-Hughes *et al.*, 2007). This is consistent with the findings of Conigrave *et al*., that aromatic AAs are the most potent activators of CaSR, while branched chain AAs are the less potent ones (Conigrave *et al.*, 2000).
There is no doubt that protein induces an increase of IGF-1 level. By doing so and according to what is known of IGF-1, protein is likely to have an anabolic effect on bone. The proposed mechanism would be that a high protein intake (especially one rich in aromatic AAs) would activate CaSR in hepatocytes and osteoblasts and stimulate IGF-1 production, which would in turn exert an anabolic effect on bone.
#### **2.3 Bone turnover markers**
Bone turnover markers are measured in blood or urine; they give information on bone formation or on bone resorption. Both types of markers are needed in a study in order to assess bone turnover.
The recent meta-analysis of Darling *et al*. concludes that there is no clear evidence of an effect of protein intake on bone markers (Darling *et al.*, 2009). Some studies reported an increase in bone resorption while bone formation remained stable (Kerstetter *et al.*, 1999; Roughead *et al.*, 2003; Zwart *et al.*, 2005). However, in one of them, the authors report an increase of hydroxyproline (a bone resorption marker) that could be related to the protein source: a collagen-rich meat (Roughead *et al.*, 2003). On the other hand, three studies considering both formation and resorption markers concluded that there was a positive effect of protein on bone. One of them observed a decrease of desoxypyridinoline (a bone resorption marker) with stable formation, indicating a positive decoupling (Hunt *et al.*, 2009). In the second study, both formation and resorption markers increased in a proteinsupplemented group compared to control during 6 months of exercise. According to the authors, bone formation will increase faster than bone resorption over time (Ballard *et al.*, 2005). In the final study, a one year so**y** protein supplementation only increased bone formation markers while resorption markers remained at the same level (Arjmandi *et al.*, 2005). Finally, two intervention studies reported a decrease of resorption with protein supplementation but provided no data on bone formation (Dawson-Hughes *et al.*, 2004; Ince *et al.*, 2004).
#### **3. Protein and calcium metabolism**
#### **3.1 Urine calcium**
Urinary calcium excretion is the major pathway for body calcium loss, along with that of feces and sweat. A high urinary calcium level has been linked to increased bone loss (Giannini *et al.*, 2003).
Dietary Protein and Bone Health 125
The effect of protein on calcium absorption could also be conditioned by the calcium intake itself. Two studies tested the effect of protein intake on intestinal calcium absorption at different levels of calcium intake. One observed a positive effect of protein for 675 mg Ca/d and no effect at 1510 mg Ca/d (Hunt *et al.*, 2009) whereas an other study observed a
Moreover, Kerstetter et al. repeatedly showed that, in the context of low protein intake, there is both an increase of PTH and a decrease in calcium absorption (Kerstetter *et al.*, 1997, 1998, 2006). Those two findings seem incompatible because PTH is known to increase calcium absorption (Heaney, 2007). However, Conigrave *et al*., as a part of his theory on amino acid-sensing CaSR, proposes another mechanism not involving PTH. CaSR is also present in the gastro-intestinal tract, where it promotes gastric acid secretion directly and indirectly through gastrin release (Conigrave & Brown, 2006). High-protein diets induce the acidification of the intestinal content which helps calcium salts to dissolve into Ca2+, thereby
According to the acid-ash theory, some foods such as protein would induce a metabolic acidosis whereas other foods such as vegetables and fruits would counteract this process. The acidosis resulting from a high protein diet would be deleterious to bone and induce
Dietary proteins, and more specifically their sulphur AA content, are part of the acidforming nutrients. Consequently, high protein consumption would lead to metabolic acidosis (Frassetto *et al.*, 1998; Kerstetter *et al.*, 2006). pH is closely regulated in the body and any increase of the acid load is buffered through different pathways. The main one is through lung excretion of CO2, followed by kidney acid excretion. Bone has also been proposed to take part by releasing Ca2+ (Lemann *et al.*, 2003). This last pathway is debated because the lungs' and the kidneys' buffering capacity are considered to be sufficient to compensate for the protein-induced acidosis (Bonjour, 2005). The idea that an acidic diet is deleterious for bone is supported by observational studies linking acid production to a reduced BMD (New *et al.*, 2004; Rahbar *et al.*, 2009; Wynn *et al.*, 2008). Intervention studies on this topic usually correct the acid diet of the subjects by supplementing them with alkalinizing molecules. The results consistently show an improvement of bone health with the supplementation. This improvement occurs at the level of urinary calcium excretion (Frassetto *et al.*, 2005; Lemann *et al.*, 1991), bone turnover markers (Dawson-Hughes *et al.*, 2009; He *et al.*, 2010; Maurer *et al.*, 2003) or Bone Mineral Density (BMD) (Domrongkitchaiporn *et al.*, 2002). Despite all this evidence of the beneficial effect of alkalinizing the diet, it should be noted that according to a recent meta-analysis, no causal link could be established between dietary acid load and bone disease (Fenton *et al.*, 2011). It is also important to consider that the findings supporting the acid-ash hypothesis are based
negative effect at 871 mg/d and no effect at 1346 mg/d.
**3.3 The acid-ash theory**
bone loss.
facilitating its absorption even if PTH level is low (Conigrave *et al.*, 2008).
on an alkalinisation of the diet and are not directly related to protein intake.
A decrease of pH is probably one of the main activators of bone-resorbing osteoclasts and an inhibitor of bone matrix deposition by osteoblasts, which explains the acid-induced bone reabsorption (Arnett, 2008). Extrapolation of these results to the clinical level is risky as it compares the *in vitro* pH variation to a high protein-induced acidosis. Indeed, the pH
Numerous studies have reported a positive linear relationship between dietary protein and urinary calcium (Kerstetter & Allen, 1990; Kerstetter *et al.*, 2003; Whiting *et al.*, 1997). The rise in urinary calcium with protein intake has also been observed in more recent studies (Ceglia *et al.*, 2009; Hunt *et al.*, 2009; Jajoo *et al.*, 2006). In one study, no effect of protein on calciuria was observed (Roughead *et al.*, 2003). The mechanism underlying the effect of protein on calcium excretion is debated. First, a high protein diet induces an increase of glomerular filtration rate which will exacerbate any increased calcium excretion (Kerstetter *et al.*, 1998). According to the acid ash hypothesis, the calciuric effect of a high protein intake would be caused by the protein-induced acid load. A meta-analysis by Fenton *et al*. of over 25 studies concluded that changes in acid excretion modulate calcium excretion (Fenton *et al.*, 2008). Other studies showed that urinary calcium could be reduced by decreasing the acid load (L. Frassetto *et al.*, 2005; Lemann *et al.*, 1991). The systemic acidosis generated by the protein load inhibits TRPV5 calcium channel expression in the kidney. Because TRPV5 participates in renal calcium reabsorption, its inhibition increases urinary calcium (Nijenhuis *et al.*, 2006).
Another mechanism for the protein-induced calcium excretion was proposed by Conigrave and colleagues, involving the extracellular CaSR (Conigrave *et al.*, 2008). CaSR found on cells of the parathyroid glands and in the renal tubules are sensitive not only to calcium, but also to some AAs (Conigrave *et al.*, 2000). Through this double sensitivity, a link is formed between protein and calcium metabolism. According to these findings, AAs would activate the CaSRs, reducing parathormone (PTH) production and decreasing renal calcium reabsorption, leading eventually to an increased calcium excretion (Conigrave *et al.*, 2008). Indeed, PTH level has been shown to be influenced by protein intake (Kerstetter *et al.*, 1997, 1998, 2006).
Urinary calcium loss due to a high protein diet was attributed to enhanced bone resorption (Kerstetter & Allen, 1990); however, this view has been discussed as the link between protein and bone resorption is unclear (Bonjour, 2005; M. Thorpe & Evans, 2011). Urinary calcium excretion is mainly related to bone as the main calcium storage compartment. However, an increase of the excreted calcium could also be related to an increase in calcium absorption through the diet. Hence the effect of protein on intestinal calcium absorption should also be considered.
#### **3.2 Calcium absorption**
While many studies claim that high-protein diets cause calciuria, there is no clear relationship between protein intake and calcium absorption. Some studies report a positive effect of protein on calcium absorption (Cao *et al.*, 2011; Hunt *et al.*, 2009; Kerstetter *et al.*, 1998, 2005). One study of 13 women observed an increase in calcium absorption that explained 93% of the calcium excretion (Kerstetter *et al.*, 2005). Another study reported an increased calciuria and a non-significant increase in calcium absorption with protein intake. But the net difference between calcium absorption and excretion with high or low protein intake remained stable (Cao *et al.*, 2011). However, other studies observed a decrease of calcium absorption with protein (Dawson-Hughes & Harris, 2002; Heaney, 2000), or no effect at all (Roughead *et al.*, 2003). It should be pointed out that all studies showing a positive effect were short term studies (2 to 7 weeks) whereas the two studies with a negative effect were long-term studies (years). It is thus possible that protein increases calcium absorption at first and that this effect is reversed in the long run.
Numerous studies have reported a positive linear relationship between dietary protein and urinary calcium (Kerstetter & Allen, 1990; Kerstetter *et al.*, 2003; Whiting *et al.*, 1997). The rise in urinary calcium with protein intake has also been observed in more recent studies (Ceglia *et al.*, 2009; Hunt *et al.*, 2009; Jajoo *et al.*, 2006). In one study, no effect of protein on calciuria was observed (Roughead *et al.*, 2003). The mechanism underlying the effect of protein on calcium excretion is debated. First, a high protein diet induces an increase of glomerular filtration rate which will exacerbate any increased calcium excretion (Kerstetter *et al.*, 1998). According to the acid ash hypothesis, the calciuric effect of a high protein intake would be caused by the protein-induced acid load. A meta-analysis by Fenton *et al*. of over 25 studies concluded that changes in acid excretion modulate calcium excretion (Fenton *et al.*, 2008). Other studies showed that urinary calcium could be reduced by decreasing the acid load (L. Frassetto *et al.*, 2005; Lemann *et al.*, 1991). The systemic acidosis generated by the protein load inhibits TRPV5 calcium channel expression in the kidney. Because TRPV5 participates in renal calcium reabsorption, its inhibition increases urinary calcium (Nijenhuis *et al.*, 2006). Another mechanism for the protein-induced calcium excretion was proposed by Conigrave and colleagues, involving the extracellular CaSR (Conigrave *et al.*, 2008). CaSR found on cells of the parathyroid glands and in the renal tubules are sensitive not only to calcium, but also to some AAs (Conigrave *et al.*, 2000). Through this double sensitivity, a link is formed between protein and calcium metabolism. According to these findings, AAs would activate the CaSRs, reducing parathormone (PTH) production and decreasing renal calcium reabsorption, leading eventually to an increased calcium excretion (Conigrave *et al.*, 2008). Indeed, PTH level has been shown to be influenced by protein intake (Kerstetter *et al.*, 1997,
Urinary calcium loss due to a high protein diet was attributed to enhanced bone resorption (Kerstetter & Allen, 1990); however, this view has been discussed as the link between protein and bone resorption is unclear (Bonjour, 2005; M. Thorpe & Evans, 2011). Urinary calcium excretion is mainly related to bone as the main calcium storage compartment. However, an increase of the excreted calcium could also be related to an increase in calcium absorption through the diet. Hence the effect of protein on intestinal calcium absorption
While many studies claim that high-protein diets cause calciuria, there is no clear relationship between protein intake and calcium absorption. Some studies report a positive effect of protein on calcium absorption (Cao *et al.*, 2011; Hunt *et al.*, 2009; Kerstetter *et al.*, 1998, 2005). One study of 13 women observed an increase in calcium absorption that explained 93% of the calcium excretion (Kerstetter *et al.*, 2005). Another study reported an increased calciuria and a non-significant increase in calcium absorption with protein intake. But the net difference between calcium absorption and excretion with high or low protein intake remained stable (Cao *et al.*, 2011). However, other studies observed a decrease of calcium absorption with protein (Dawson-Hughes & Harris, 2002; Heaney, 2000), or no effect at all (Roughead *et al.*, 2003). It should be pointed out that all studies showing a positive effect were short term studies (2 to 7 weeks) whereas the two studies with a negative effect were long-term studies (years). It is thus possible that protein increases
calcium absorption at first and that this effect is reversed in the long run.
1998, 2006).
should also be considered.
**3.2 Calcium absorption**
The effect of protein on calcium absorption could also be conditioned by the calcium intake itself. Two studies tested the effect of protein intake on intestinal calcium absorption at different levels of calcium intake. One observed a positive effect of protein for 675 mg Ca/d and no effect at 1510 mg Ca/d (Hunt *et al.*, 2009) whereas an other study observed a negative effect at 871 mg/d and no effect at 1346 mg/d.
Moreover, Kerstetter et al. repeatedly showed that, in the context of low protein intake, there is both an increase of PTH and a decrease in calcium absorption (Kerstetter *et al.*, 1997, 1998, 2006). Those two findings seem incompatible because PTH is known to increase calcium absorption (Heaney, 2007). However, Conigrave *et al*., as a part of his theory on amino acid-sensing CaSR, proposes another mechanism not involving PTH. CaSR is also present in the gastro-intestinal tract, where it promotes gastric acid secretion directly and indirectly through gastrin release (Conigrave & Brown, 2006). High-protein diets induce the acidification of the intestinal content which helps calcium salts to dissolve into Ca2+, thereby facilitating its absorption even if PTH level is low (Conigrave *et al.*, 2008).
#### **3.3 The acid-ash theory**
According to the acid-ash theory, some foods such as protein would induce a metabolic acidosis whereas other foods such as vegetables and fruits would counteract this process. The acidosis resulting from a high protein diet would be deleterious to bone and induce bone loss.
Dietary proteins, and more specifically their sulphur AA content, are part of the acidforming nutrients. Consequently, high protein consumption would lead to metabolic acidosis (Frassetto *et al.*, 1998; Kerstetter *et al.*, 2006). pH is closely regulated in the body and any increase of the acid load is buffered through different pathways. The main one is through lung excretion of CO2, followed by kidney acid excretion. Bone has also been proposed to take part by releasing Ca2+ (Lemann *et al.*, 2003). This last pathway is debated because the lungs' and the kidneys' buffering capacity are considered to be sufficient to compensate for the protein-induced acidosis (Bonjour, 2005). The idea that an acidic diet is deleterious for bone is supported by observational studies linking acid production to a reduced BMD (New *et al.*, 2004; Rahbar *et al.*, 2009; Wynn *et al.*, 2008). Intervention studies on this topic usually correct the acid diet of the subjects by supplementing them with alkalinizing molecules. The results consistently show an improvement of bone health with the supplementation. This improvement occurs at the level of urinary calcium excretion (Frassetto *et al.*, 2005; Lemann *et al.*, 1991), bone turnover markers (Dawson-Hughes *et al.*, 2009; He *et al.*, 2010; Maurer *et al.*, 2003) or Bone Mineral Density (BMD) (Domrongkitchaiporn *et al.*, 2002). Despite all this evidence of the beneficial effect of alkalinizing the diet, it should be noted that according to a recent meta-analysis, no causal link could be established between dietary acid load and bone disease (Fenton *et al.*, 2011). It is also important to consider that the findings supporting the acid-ash hypothesis are based on an alkalinisation of the diet and are not directly related to protein intake.
A decrease of pH is probably one of the main activators of bone-resorbing osteoclasts and an inhibitor of bone matrix deposition by osteoblasts, which explains the acid-induced bone reabsorption (Arnett, 2008). Extrapolation of these results to the clinical level is risky as it compares the *in vitro* pH variation to a high protein-induced acidosis. Indeed, the pH
Dietary Protein and Bone Health 127
men. The study lasted 28 days and the authors reported a decrease in BMC in the supplemented subjects whereas BMC in control subjects remained stable (Zwart *et al.*, 2005). It should be noted that energy intake was different between the two groups due to an energy-free placebo. Another study supplemented hospitalized elderly men and women for 38 days with 20.4g of protein. Along with other positive effects (lower complication rate, shorter hospital stay), the authors reported a decrease in BMD loss with the protein supplement (Tkatch *et al.*, 1992). Finally, a year-long study supplementing 62 postmenopausal women with 25g of soy protein observed no effect on total hip BMD or
Fracture risk is the ultimate clinical outcome when considering bone. A longitudinal study of 32 050 postmenopausal women observed a reduced relative risk of hip fracture when protein intake was increased (Munger *et al.*, 1999). Another one observed that a reduction in wrist fracture is positively associated with the consumption frequency of high protein food in 1865 perimenopausal women (D. L. Thorpe *et al.*, 2008a). Finally, in a case-control study on 1167 cases of hip fractures and 1334 controls of both sexes aged 50-89, the odds ratio for
According to these three studies, protein would have a beneficial effect on bone, but some other studies found conflicting results (Dargent-Molina *et al.*, 2008; Feskanich *et al.*, 1996; Meyer *et al.*, 1997). However, it should be noted that the negative relationship between protein and fracture risk was significant only in the lowest quartiles of calcium intake (<400 mg/d) and that no association was observed for higher calcium intake (Dargent-Molina *et al.*, 2008; Meyer *et al.*, 1997). Pooling four studies in a meta-analysis, Darling found no significant effect of protein on the risk ratio of fracture (Darling *et al.*, 2009). To our knowledge, it seems that as long as calcium intake is sufficient there is a protective effect of
The question of the effect of protein on fracture healing has also been addressed in some randomised controlled trials. All four trials were conducted by the same group and reported that a protein supplement improved the patients' condition after a low-trauma femoral neck fracture (Delmi *et al.*, 1990; Tkatch *et al.*, 1992) or hip fracture (Chevalley *et al.*, 2008; Schurch *et al.*, 1998). All the clinical trials had very similar design: patients were about 80 years old and the protein supplement given was 20g. According to the two studies on femoral neck fracture, patients taking the protein supplementation had better clinical outcomes and reduced rate of complication and mortality during the hospital stay and 6 months after (Delmi *et al.*, 1990; Tkatch *et al.*, 1992). After hip fracture, the protein supplement increased IGF-1, attenuating bone loss at 6 months and shortening hospital stay (Schurch *et al.*, 1998). The IGF-1 increase is independent of the type of protein given (casein, whey protein or whey protein and amino acids) (Chevalley *et al.*, 2010). The beneficial effect of protein is probably linked to both the
Each protein bares some information in the form of its chain of amino acids. This information can relate to specific effects of some proteins. When protein goes through the
increased IGF-1 level and the correction of the patient nutritional state.
**5. Effect of specific protein source on bone**
hip fracture decreased with the increase of protein intake (Wengreen *et al.*, 2004).
BMC (Arjmandi *et al.*, 2005).
protein on bone that lowers the fracture risk.
**4.2 Fracture risk**
variation induced by protein intake is likely to be too small to induce any effect in the bone micro-environment (Bonjour, 2005).
Although they lead to different conclusions, the acid-ash theory and the hormonal anabolic theory are not mutually exclusive. A dual-pathway model for the effect of protein on bone has been proposed (M. Thorpe & Evans, 2011). Although it has been recently criticized in a meta-analysis (Fenton *et al.*, 2011), it seems that there is a beneficial effect of diet alkalinisation. This can be achieved by lowering acidic or by increasing alkaline food consumption. When considering protein, the second method is more beneficial. Indeed, even if protein is an acid nutrient, it is also an activator of bone anabolic hormones and its consumption should not be lowered. This is especially true in populations already consuming a low protein diet such as the elderly. The acid load resulting from the protein consumption should be compensated by an increased consumption of alkalinizing fruits and vegetables.
Some diets designed to promote weight loss rely on high protein consumption for a quick effect on weight. It is hard to estimate what is the effect of such diets on bone. Indeed, hyperproteic diets imply also potentially confounding effect such as spontaneous caloric restriction and a possible lack of micronutrients. Moreover, the wide variety of hyperproteic diets and the potential lack of compliance of the subjects further increase the difficulty of designing an adequate study on the subject.
#### **4. Protein and bone parameters**
#### **4.1 Bone Mineral Density**
BMD is not the only determinant of skeletal fragility; the spatial distribution of the bone mass (as cortical and trabecular bone) and the intrinsic material properties of bone are also major components (Bouxsein & Seeman, 2009). Measuring BMD or Bone Mineral Content (BMC) is the easiest way of directly assessing bone strength in humans.
Most observation studies report a positive association between protein and bone density (Hannan *et al.*, 2000; Promislow *et al.*, 2002) or between protein and BMD change (Dawson-Hughes & Harris, 2002; Vatanparast *et al.*, 2007). Specific studies showing a positive correlation between protein and bone density cover a broad population: premenopausal women (Teegarden *et al.*, 1998), postmenopausal women (Devine *et al.*, 2005; Ilich *et al.*, 2003; Rapuri *et al.*, 2003; M. Thorpe *et al.*, 2008b), men (Whiting *et al.*, 2002) and children (Alexy *et al.*, 2005; Chevalley *et al.*, 2008).
On the other hand, two studies on premenopausal women concluded that protein intake had no relation with BMD (Beasley *et al.*, 2010; Mazess & Barden, 1991). One study reported a negative association with BMC (Metz *et al.*, 1993). Finally, a case-control study compared 134 osteoporotic women with 137 controls and identified total protein intake as a risk factor for osteoporosis. A meta-analysis by Darling and colleagues conclude that there is no evidence of a negative effect of protein intake on bone when looking at observation studies; in fact there is probably a small positive effect of protein on bone (Darling *et al.*, 2009).
There are very few interventional studies comparing high vs. low protein intake. One of them focused on the effect of an AA and carbohydrate supplement during bed rest on 13
variation induced by protein intake is likely to be too small to induce any effect in the bone
Although they lead to different conclusions, the acid-ash theory and the hormonal anabolic theory are not mutually exclusive. A dual-pathway model for the effect of protein on bone has been proposed (M. Thorpe & Evans, 2011). Although it has been recently criticized in a meta-analysis (Fenton *et al.*, 2011), it seems that there is a beneficial effect of diet alkalinisation. This can be achieved by lowering acidic or by increasing alkaline food consumption. When considering protein, the second method is more beneficial. Indeed, even if protein is an acid nutrient, it is also an activator of bone anabolic hormones and its consumption should not be lowered. This is especially true in populations already consuming a low protein diet such as the elderly. The acid load resulting from the protein consumption should be compensated by an increased consumption of alkalinizing fruits and
Some diets designed to promote weight loss rely on high protein consumption for a quick effect on weight. It is hard to estimate what is the effect of such diets on bone. Indeed, hyperproteic diets imply also potentially confounding effect such as spontaneous caloric restriction and a possible lack of micronutrients. Moreover, the wide variety of hyperproteic diets and the potential lack of compliance of the subjects further increase the difficulty of
BMD is not the only determinant of skeletal fragility; the spatial distribution of the bone mass (as cortical and trabecular bone) and the intrinsic material properties of bone are also major components (Bouxsein & Seeman, 2009). Measuring BMD or Bone Mineral Content
Most observation studies report a positive association between protein and bone density (Hannan *et al.*, 2000; Promislow *et al.*, 2002) or between protein and BMD change (Dawson-Hughes & Harris, 2002; Vatanparast *et al.*, 2007). Specific studies showing a positive correlation between protein and bone density cover a broad population: premenopausal women (Teegarden *et al.*, 1998), postmenopausal women (Devine *et al.*, 2005; Ilich *et al.*, 2003; Rapuri *et al.*, 2003; M. Thorpe *et al.*, 2008b), men (Whiting *et al.*, 2002) and children (Alexy *et*
On the other hand, two studies on premenopausal women concluded that protein intake had no relation with BMD (Beasley *et al.*, 2010; Mazess & Barden, 1991). One study reported a negative association with BMC (Metz *et al.*, 1993). Finally, a case-control study compared 134 osteoporotic women with 137 controls and identified total protein intake as a risk factor for osteoporosis. A meta-analysis by Darling and colleagues conclude that there is no evidence of a negative effect of protein intake on bone when looking at observation studies; in fact there is probably a small positive effect of protein on bone (Darling *et al.*, 2009).
There are very few interventional studies comparing high vs. low protein intake. One of them focused on the effect of an AA and carbohydrate supplement during bed rest on 13
(BMC) is the easiest way of directly assessing bone strength in humans.
micro-environment (Bonjour, 2005).
designing an adequate study on the subject.
**4. Protein and bone parameters**
**4.1 Bone Mineral Density**
*al.*, 2005; Chevalley *et al.*, 2008).
vegetables.
men. The study lasted 28 days and the authors reported a decrease in BMC in the supplemented subjects whereas BMC in control subjects remained stable (Zwart *et al.*, 2005). It should be noted that energy intake was different between the two groups due to an energy-free placebo. Another study supplemented hospitalized elderly men and women for 38 days with 20.4g of protein. Along with other positive effects (lower complication rate, shorter hospital stay), the authors reported a decrease in BMD loss with the protein supplement (Tkatch *et al.*, 1992). Finally, a year-long study supplementing 62 postmenopausal women with 25g of soy protein observed no effect on total hip BMD or BMC (Arjmandi *et al.*, 2005).
#### **4.2 Fracture risk**
Fracture risk is the ultimate clinical outcome when considering bone. A longitudinal study of 32 050 postmenopausal women observed a reduced relative risk of hip fracture when protein intake was increased (Munger *et al.*, 1999). Another one observed that a reduction in wrist fracture is positively associated with the consumption frequency of high protein food in 1865 perimenopausal women (D. L. Thorpe *et al.*, 2008a). Finally, in a case-control study on 1167 cases of hip fractures and 1334 controls of both sexes aged 50-89, the odds ratio for hip fracture decreased with the increase of protein intake (Wengreen *et al.*, 2004).
According to these three studies, protein would have a beneficial effect on bone, but some other studies found conflicting results (Dargent-Molina *et al.*, 2008; Feskanich *et al.*, 1996; Meyer *et al.*, 1997). However, it should be noted that the negative relationship between protein and fracture risk was significant only in the lowest quartiles of calcium intake (<400 mg/d) and that no association was observed for higher calcium intake (Dargent-Molina *et al.*, 2008; Meyer *et al.*, 1997). Pooling four studies in a meta-analysis, Darling found no significant effect of protein on the risk ratio of fracture (Darling *et al.*, 2009). To our knowledge, it seems that as long as calcium intake is sufficient there is a protective effect of protein on bone that lowers the fracture risk.
The question of the effect of protein on fracture healing has also been addressed in some randomised controlled trials. All four trials were conducted by the same group and reported that a protein supplement improved the patients' condition after a low-trauma femoral neck fracture (Delmi *et al.*, 1990; Tkatch *et al.*, 1992) or hip fracture (Chevalley *et al.*, 2008; Schurch *et al.*, 1998). All the clinical trials had very similar design: patients were about 80 years old and the protein supplement given was 20g. According to the two studies on femoral neck fracture, patients taking the protein supplementation had better clinical outcomes and reduced rate of complication and mortality during the hospital stay and 6 months after (Delmi *et al.*, 1990; Tkatch *et al.*, 1992). After hip fracture, the protein supplement increased IGF-1, attenuating bone loss at 6 months and shortening hospital stay (Schurch *et al.*, 1998). The IGF-1 increase is independent of the type of protein given (casein, whey protein or whey protein and amino acids) (Chevalley *et al.*, 2010). The beneficial effect of protein is probably linked to both the increased IGF-1 level and the correction of the patient nutritional state.
#### **5. Effect of specific protein source on bone**
Each protein bares some information in the form of its chain of amino acids. This information can relate to specific effects of some proteins. When protein goes through the
Dietary Protein and Bone Health 129
results because of the broad variety of foods in these groups. In future research, it would be more accurate to consider an estimation of the sulphate content of each protein-containing
Milk contains many bioactive factors, including growth hormones, enzymes, antimicrobials, anti-inflammatory agents, transporters and peptide or nonpeptide hormones. Milk, because it contains bioactive molecules, extends beyond applications in infant nutrition and was
Clinical trials showed that MBP supplementation increased BMD and decreased bone resorption biomarkers in healthy women (Aoe *et al.*, 2001; Uenishi *et al.*, 2007; Yamamura *et al.*, 2002), menopausal women (Aoe *et al.*, 2005), and healthy older women (Aoyagi *et al.*, 2010). In particular, MBP suppresses osteoclast-mediated bone resorption and leads to reduced osteoclast number in animal studies (Morita *et al.*, 2008). Morita *et al*. reported that the protein fraction responsible for the observed activities of MBP is the bovine angiogenin
However, not all studies have shown a beneficial effect of MBP. In one study, 84 healthy young women were divided into three groups receiving placebo, whole milk, or milk containing 40 mg MBP for 8 months. Compared with the baseline values, total BMD significantly increased in all groups. There was a significant decrease of bone resorption marker N-teleopeptides of type-I collagen (NTx) while bone formation remained stable in
HPLC analysis of the major proteins in the MBP fraction identified the presence of the
Lactoferrin (LF) is an 80 kDa iron-binding glycoprotein of the transferrin family. This molecule has been demonstrated to inhibit *in vitro* osteoclast-mediated bone resorption (Lorget *et al.*, 2002). LF was also demonstrated to have *in vitro* anabolic, differentiating and anti-apoptotic effects on osteoblasts, and to inhibit osteoclastogenesis. Moreover *in vivo* local injection of LF above the hemicalvaria increases bone formation and bone area in adult mice (Cornish *et al.*, 2004). LF has a role in host non-specific defense (Gahr *et al.*, 1991; Legrand *et al.*, 2004). In addition to its direct antimicrobial effects, LF is believed to modulate the inflammatory process mainly by preventing the release of inflammatory cytokines which induce recruitment and
Investigations of our group and others established that LF at physiological concentrations can stimulate proliferation of primary osteoblasts and osteoblastic-cell lines and increase osteoblast differentiation (Blais *et al.*, 2009; Cornish, 2004; Takayama & Mizumachi, 2008). Studies using 3-week culture of primary rat osteoblasts show that LF dose-dependently increases the number of nodules and the area of mineralized bone formed (Cornish, 2004). Our *in vitro* experiments demonstrated that LF could directly act on bone cells. Bovine LF
food rather than focusing only on a crude animal-vegetable distinction.
considered as a possible source of factors with anabolic effects on bone.
which acts as a bone resorption-inhibitory protein (Morita *et al.*, 2008).
glycoprotein lactoferrin in most of these fractions (Naot *et al.*, 2005).
activation of immune cells at inflammatory sites (Legrand *et al.*, 2005).
**5.2 Milk proteins**
**5.2.1 Milk Basic Proteins (MBP)**
both milk groups. (Zou *et al.*, 2009).
**5.2.2 Lactoferrin**
intestinal barrier, 50% are completely degraded as simple amino acids, 40% are partly degraded as peptides and 10% remain intact (Mallegol *et al.*, 2005). This observation means that half of the digested protein reaching blood is still bearing some information in the form of the AA chain. Much research has tried to investigate what might be the best source of protein for bone health.
#### **5.1 Animal vs. vegetable protein**
Some observational studies considered the nature of protein source when measuring the effect on bone. One of them studied the relationships between the animal/vegetable protein ratio and bone parameters on 1035 postmenopausal women, showing that the ratio is positively associated with bone loss and hip fracture risk (Sellmeyer *et al.*, 2001). However, the fact that a ratio was used in this study instead of the absolute values has been criticized (Bonjour, 2005; M. Thorpe & Evans, 2011). Other observational studies provide conflicting results. BMD was positively associated with animal protein intake and negatively to vegetable intake in one study (Promislow *et al.*, 2002) but the opposite association was found when considering bone ultrasound attenuation (Weikert *et al.*, 2005). Low levels of both protein types have been associated with deleterious effects on bone: low animal protein is related to bone loss (Hannan *et al.*, 2000) and low vegetable protein is related to low BMD (Beasley *et al.*, 2010). These results suggest that a minimum intake of both proteins is required regardless of the source. Finally, when considering fracture risk, a positive relationship was found with both animal (Dargent-Molina *et al.*, 2008; Feskanich *et al.*, 1996; Meyer *et al.*, 1997) and vegetable protein (Munger *et al.*, 1999; D. L. Thorpe *et al.*, 2008a). It should be noted that the studies finding a positive relationship between animal protein and fracture risk also found a positive relationship with total protein. On the other hand, those finding a positive relationship between vegetable protein and fracture risk reported a negative one with total protein. Hence the results for animal and vegetable protein are not obtained in the same context.
The intervention studies comparing animal and vegetable protein always focus on specific types of protein, usually soy and casein. Hence it is the specific effect of those proteins that is evaluated and not the one of the animal and vegetable food groups. The only whole-diet intervention study was based on meat or protein-rich vegetables such as nuts and grains. Only urinary parameters were measured and urinary calcium was similar for the two diets (Massey & Kynast-Gales, 2001). More studies of this type are needed to address this issue.
Mechanistically, the reason for differentiating animal from vegetable protein is their differing sulphur AA content and the consequent modulation of potential acid load (Sellmeyer *et al.*, 2001). The influence of sulphate content on protein efficiency has been emphasized in a cross-sectional study on 161 postmenopausal women. The results show a positive association of total protein intake with lumbar spine and total hip BMD; however, in lumbar spine this benefit is suppressed by the sulphur-containing AAs (M. Thorpe *et al.*, 2008b). The authors conclude that an excessive consumption of sulphur AAs is likely to be deleterious for bone. However, as underlined by Massey, the assumption that animal protein contains more sulphur AA than vegetable protein is not always true. As an example, potential acidity from sulphur AA in milk or beef is lower than that of whole wheat or white rice (Massey, 2003). Hence, even if sulphur AAs were proven to have a negative effect on bone, the extrapolation to animal and vegetable food groups is likely to give incorrect results because of the broad variety of foods in these groups. In future research, it would be more accurate to consider an estimation of the sulphate content of each protein-containing food rather than focusing only on a crude animal-vegetable distinction.
#### **5.2 Milk proteins**
128 Dual Energy X-Ray Absorptiometry
intestinal barrier, 50% are completely degraded as simple amino acids, 40% are partly degraded as peptides and 10% remain intact (Mallegol *et al.*, 2005). This observation means that half of the digested protein reaching blood is still bearing some information in the form of the AA chain. Much research has tried to investigate what might be the best source of
Some observational studies considered the nature of protein source when measuring the effect on bone. One of them studied the relationships between the animal/vegetable protein ratio and bone parameters on 1035 postmenopausal women, showing that the ratio is positively associated with bone loss and hip fracture risk (Sellmeyer *et al.*, 2001). However, the fact that a ratio was used in this study instead of the absolute values has been criticized (Bonjour, 2005; M. Thorpe & Evans, 2011). Other observational studies provide conflicting results. BMD was positively associated with animal protein intake and negatively to vegetable intake in one study (Promislow *et al.*, 2002) but the opposite association was found when considering bone ultrasound attenuation (Weikert *et al.*, 2005). Low levels of both protein types have been associated with deleterious effects on bone: low animal protein is related to bone loss (Hannan *et al.*, 2000) and low vegetable protein is related to low BMD (Beasley *et al.*, 2010). These results suggest that a minimum intake of both proteins is required regardless of the source. Finally, when considering fracture risk, a positive relationship was found with both animal (Dargent-Molina *et al.*, 2008; Feskanich *et al.*, 1996; Meyer *et al.*, 1997) and vegetable protein (Munger *et al.*, 1999; D. L. Thorpe *et al.*, 2008a). It should be noted that the studies finding a positive relationship between animal protein and fracture risk also found a positive relationship with total protein. On the other hand, those finding a positive relationship between vegetable protein and fracture risk reported a negative one with total protein. Hence the results for animal and vegetable protein are not
The intervention studies comparing animal and vegetable protein always focus on specific types of protein, usually soy and casein. Hence it is the specific effect of those proteins that is evaluated and not the one of the animal and vegetable food groups. The only whole-diet intervention study was based on meat or protein-rich vegetables such as nuts and grains. Only urinary parameters were measured and urinary calcium was similar for the two diets (Massey & Kynast-Gales, 2001). More studies of this type are needed to address this issue. Mechanistically, the reason for differentiating animal from vegetable protein is their differing sulphur AA content and the consequent modulation of potential acid load (Sellmeyer *et al.*, 2001). The influence of sulphate content on protein efficiency has been emphasized in a cross-sectional study on 161 postmenopausal women. The results show a positive association of total protein intake with lumbar spine and total hip BMD; however, in lumbar spine this benefit is suppressed by the sulphur-containing AAs (M. Thorpe *et al.*, 2008b). The authors conclude that an excessive consumption of sulphur AAs is likely to be deleterious for bone. However, as underlined by Massey, the assumption that animal protein contains more sulphur AA than vegetable protein is not always true. As an example, potential acidity from sulphur AA in milk or beef is lower than that of whole wheat or white rice (Massey, 2003). Hence, even if sulphur AAs were proven to have a negative effect on bone, the extrapolation to animal and vegetable food groups is likely to give incorrect
protein for bone health.
**5.1 Animal vs. vegetable protein**
obtained in the same context.
Milk contains many bioactive factors, including growth hormones, enzymes, antimicrobials, anti-inflammatory agents, transporters and peptide or nonpeptide hormones. Milk, because it contains bioactive molecules, extends beyond applications in infant nutrition and was considered as a possible source of factors with anabolic effects on bone.
#### **5.2.1 Milk Basic Proteins (MBP)**
Clinical trials showed that MBP supplementation increased BMD and decreased bone resorption biomarkers in healthy women (Aoe *et al.*, 2001; Uenishi *et al.*, 2007; Yamamura *et al.*, 2002), menopausal women (Aoe *et al.*, 2005), and healthy older women (Aoyagi *et al.*, 2010). In particular, MBP suppresses osteoclast-mediated bone resorption and leads to reduced osteoclast number in animal studies (Morita *et al.*, 2008). Morita *et al*. reported that the protein fraction responsible for the observed activities of MBP is the bovine angiogenin which acts as a bone resorption-inhibitory protein (Morita *et al.*, 2008).
However, not all studies have shown a beneficial effect of MBP. In one study, 84 healthy young women were divided into three groups receiving placebo, whole milk, or milk containing 40 mg MBP for 8 months. Compared with the baseline values, total BMD significantly increased in all groups. There was a significant decrease of bone resorption marker N-teleopeptides of type-I collagen (NTx) while bone formation remained stable in both milk groups. (Zou *et al.*, 2009).
HPLC analysis of the major proteins in the MBP fraction identified the presence of the glycoprotein lactoferrin in most of these fractions (Naot *et al.*, 2005).
#### **5.2.2 Lactoferrin**
Lactoferrin (LF) is an 80 kDa iron-binding glycoprotein of the transferrin family. This molecule has been demonstrated to inhibit *in vitro* osteoclast-mediated bone resorption (Lorget *et al.*, 2002). LF was also demonstrated to have *in vitro* anabolic, differentiating and anti-apoptotic effects on osteoblasts, and to inhibit osteoclastogenesis. Moreover *in vivo* local injection of LF above the hemicalvaria increases bone formation and bone area in adult mice (Cornish *et al.*, 2004). LF has a role in host non-specific defense (Gahr *et al.*, 1991; Legrand *et al.*, 2004). In addition to its direct antimicrobial effects, LF is believed to modulate the inflammatory process mainly by preventing the release of inflammatory cytokines which induce recruitment and activation of immune cells at inflammatory sites (Legrand *et al.*, 2005).
Investigations of our group and others established that LF at physiological concentrations can stimulate proliferation of primary osteoblasts and osteoblastic-cell lines and increase osteoblast differentiation (Blais *et al.*, 2009; Cornish, 2004; Takayama & Mizumachi, 2008). Studies using 3-week culture of primary rat osteoblasts show that LF dose-dependently increases the number of nodules and the area of mineralized bone formed (Cornish, 2004). Our *in vitro* experiments demonstrated that LF could directly act on bone cells. Bovine LF
Dietary Protein and Bone Health 131
inhibition in order to improve bone loss (Malet *et al.*, 2011). Recently a clinical trial of 38 healthy postmenopausal women randomized into placebo or ribonuclease-enriched-LF (R-ELF) groups evaluated bone health status over a period of 12 months. The authors demonstrated that R-ELF supplementation reduced bone resorption and increased osteoblastic bone formation; however BMD was not evaluated (Bharadwaj *et al.*, 2009).
In conclusion, LF has a positive effect on bone health and might be useful in pathological states of reduced bone density. The molecular mechanisms are not fully understood but our studies suggest that dietary bLF supplementation can have a beneficial effect on postmenopausal bone loss not only via a direct effect but also by modulating immune function. The development of pharmaceutical or nutriceutical compounds that are based on
Collagen has a unique triple helix configuration with a repeating sequence (Gly-X-Y)n, with X and Y being mostly proline and hydroxyproline (Hyp) (Bos *et al.*, 1999; Ramshaw *et al.*, 1998). Some studies suggest that a hydrolyzed collagen-enriched diet improves bone collagen metabolism and BMD. Oral administration of Hydrolyzed Collagen (HC) increased BMC and BMD in rats and mice fed a calcium or protein deficient diet (Koyama *et al.*, 2001;
Oral administration of HC was demonstrated to increase the quantity of type I collagen and proteoglycans in the bone matrix of ovariectomized rats (Nomura *et al.*, 2005). Moreover, in patients with osteoporosis, oral intake of HC with calcitonin had a stronger inhibitory effect on bone resorption than calcitonin alone (Adam *et al.*, 1996). Oesser *et al*. demonstrated the intestinal absorption and cartilage accumulation of collagen-derived peptides (Oesser *et al.*, 1999). It has been generally assumed that collagen–rich diets interact with the bone matrix. Indeed, collagen-derived di- and tripeptides rich in hydroxyproline such as Hyp, Pro-Hyp, Pro-Hyp-Gly or Gly-Pro-Val have been detected in human blood following the ingestion of HC (Iwai *et al.*, 2005). The PEPT1 proton-dependent transporter assures the transport of Pro-
A study of Ohara *et al*. compared quantity and structures of food-derived gelatin hydrolysates in human blood from fish scale, fish skin and pork skin type I collagen in a single blind crossover study (Ohara *et al.*, 2007). Amounts of free Hyp and Hyp-containing peptide were measured over a 24-h period. Hyp-containing peptides comprised approximately 30% of all detected Hyp. However, efficiency of HC ingestion depends not only on collagen origin but also on the molecular size of the HC. Collagen needs to be
Our *in vivo* studies indicate that ingestion of HC diet induced the growth of the external diameter of the bone cortical zone in OVX mice (Guillerminet *et al.*, 2010, 2011). The increased cortical area was correlated with a significant increase in the femur external
LF will require a better understanding of LF's mechanism of action on bone.
**5.3 Collagen**
Wu *et al.*, 2004).
**5.3.1 Ingestion of collagen**
**5.3.2** *In vivo* **studies**
Hyp across the intestinal barrier (Aito-Inoue *et al.*, 2007).
hydrolysed to be able to interact with bone metabolism.
(bLF) at low physiological concentrations (5 µg/ml) stimulates growth and activity of osteoblastic MC3T3-E1 cells and primary culture of murine osteoblast bone cells (Blais *et al.*, 2009). Low density lipoprotein receptor-related protein 1 and 2, which are present on osteoblastic cells, have been shown to be partially responsible for LF's mitogenic effect in osteoblasts (Grey *et al.*, 2004). Moreover, Grey *et al*., showed that LF is able to protect osteoblastic cells from apoptosis induced by serum withdrawal (Grey *et al.*, 2006).
LF action on osteoclasts is strikingly different since it produces an important arrest of osteoclastogenesis (Cornish, 2004; Lorget *et al.*, 2002). However, LF does not modulate mature ostoeclast activity. bLF at a concentration ranging from 10 to 1000 µg/ml was found to inhibit pre-osteoclastic established RAW cell growth. These results were confirmed in mixed primary culture of murine bone cells (Blais *et al.*, 2009). In contrast, there was no effect of LF on bone resorption when tested on isolated mature osteoclasts, or in organ cultures that can detect mature osteoclast activity (Grey *et al.*, 2006).
*In vivo* bone effects of LF were first studied using local injection of LF above the hemicalvaria which increased bone formation and bone area in adult mice (Cornish, 2004). Few recent studies using ovariectomized (OVX) rodents as a model for post menopausal bone loss measured the effect of dietary supplementation on bone (Blais *et al.*, 2009; Guo *et al.*, 2009; Malet *et al.*, 2011). LF administered orally to OVX rats for three months protected them against the OVX-induced reduction of bone volume and BMD and increased the parameters of mechanical strength, increased bone formation and reduced bone resorption (Guo *et al.*, 2009). Our studies using OVX mice demonstrated that the dietary bLF transfer into peripheral blood. LF supplementation increases BMD and bone strength compared to the OVX control group. This study supports a direct effect of LF on bone cells (Blais *et al.*, 2009).
Recent animal studies demonstrated that estrogen deficiency causes bone loss by mechanisms associated with inflammatory and oxidative processes (Grassi *et al.*, 2007; Lean *et al.*, 2003; Muthusami *et al.*, 2005). TNFis one of the cytokines responsible for the augmented osteoclastogenesis (Suda *et al.*, 1999). Indeed, ovariectomy causes an increase in TNF production from T-cells which in turn increases macrophage colony-stimulating factor and RANKL levels, leading to osteoclastogenesis (Cornish *et al.*, 2004; Suda *et al.*, 1999). Moreover the presence of increased levels of TNFwas reported in the bone marrow of OVX animals and in blood cells of postmenopausal women (Oh *et al.*, 2007; Shanker *et al.*, 1994). Postmenopausal osteoporosis should also be regarded as the result of an inflammatory process (Weitzmann & Pacifici, 2007). It has been shown that bLF plays a role in host non-specific defense and modulates the inflammatory process mainly by preventing the release of inflammatory cytokines that induce recruitment and activation of immune cells at inflammatory sites (Debbabi *et al.*, 1998; Legrand *et al.*, 2006). Indeed, bLF enriched diet ingestion can reduce release of pro-inflammatory cytokines and increase anti-inflammatory cytokine production. Our studies showed that bLF ingestion decreases bone loss and bone resorption markers. A decreased TNFmRNA expression associated with a TNFproduction inhibition on peripheral T-lymphocytes was observed with a bLF supplementation in OVX mice. Furthermore, bLF can prevent lymphocyte activation and cytokine release in the bone micro-environment. Production and release of TNFwere strongly down-regulated by LF. These immune modulations were spatially and temporally correlated with reduced bone loss. We suggested that bLF modulates the inflammatory process via specific TNF
(bLF) at low physiological concentrations (5 µg/ml) stimulates growth and activity of osteoblastic MC3T3-E1 cells and primary culture of murine osteoblast bone cells (Blais *et al.*, 2009). Low density lipoprotein receptor-related protein 1 and 2, which are present on osteoblastic cells, have been shown to be partially responsible for LF's mitogenic effect in osteoblasts (Grey *et al.*, 2004). Moreover, Grey *et al*., showed that LF is able to protect
LF action on osteoclasts is strikingly different since it produces an important arrest of osteoclastogenesis (Cornish, 2004; Lorget *et al.*, 2002). However, LF does not modulate mature ostoeclast activity. bLF at a concentration ranging from 10 to 1000 µg/ml was found to inhibit pre-osteoclastic established RAW cell growth. These results were confirmed in mixed primary culture of murine bone cells (Blais *et al.*, 2009). In contrast, there was no effect of LF on bone resorption when tested on isolated mature osteoclasts, or in organ
*In vivo* bone effects of LF were first studied using local injection of LF above the hemicalvaria which increased bone formation and bone area in adult mice (Cornish, 2004). Few recent studies using ovariectomized (OVX) rodents as a model for post menopausal bone loss measured the effect of dietary supplementation on bone (Blais *et al.*, 2009; Guo *et al.*, 2009; Malet *et al.*, 2011). LF administered orally to OVX rats for three months protected them against the OVX-induced reduction of bone volume and BMD and increased the parameters of mechanical strength, increased bone formation and reduced bone resorption (Guo *et al.*, 2009). Our studies using OVX mice demonstrated that the dietary bLF transfer into peripheral blood. LF supplementation increases BMD and bone strength compared to the OVX control group. This study supports a direct effect of LF on bone cells (Blais *et al.*,
Recent animal studies demonstrated that estrogen deficiency causes bone loss by mechanisms associated with inflammatory and oxidative processes (Grassi *et al.*, 2007; Lean *et al.*, 2003; Muthusami *et al.*, 2005). TNFis one of the cytokines responsible for the augmented osteoclastogenesis (Suda *et al.*, 1999). Indeed, ovariectomy causes an increase in TNF production from T-cells which in turn increases macrophage colony-stimulating factor and RANKL levels, leading to osteoclastogenesis (Cornish *et al.*, 2004; Suda *et al.*, 1999). Moreover the presence of increased levels of TNFwas reported in the bone marrow of OVX animals and in blood cells of postmenopausal women (Oh *et al.*, 2007; Shanker *et al.*, 1994). Postmenopausal osteoporosis should also be regarded as the result of an inflammatory process (Weitzmann & Pacifici, 2007). It has been shown that bLF plays a role in host non-specific defense and modulates the inflammatory process mainly by preventing the release of inflammatory cytokines that induce recruitment and activation of immune cells at inflammatory sites (Debbabi *et al.*, 1998; Legrand *et al.*, 2006). Indeed, bLF enriched diet ingestion can reduce release of pro-inflammatory cytokines and increase anti-inflammatory cytokine production. Our studies showed that bLF ingestion decreases bone loss and bone resorption markers. A decreased TNFmRNA expression associated with a TNFproduction inhibition on peripheral T-lymphocytes was observed with a bLF supplementation in OVX mice. Furthermore, bLF can prevent lymphocyte activation and cytokine release in the bone micro-environment. Production and release of TNFwere strongly down-regulated by LF. These immune modulations were spatially and temporally correlated with reduced bone loss. We suggested that bLF modulates the inflammatory process via specific TNF
osteoblastic cells from apoptosis induced by serum withdrawal (Grey *et al.*, 2006).
cultures that can detect mature osteoclast activity (Grey *et al.*, 2006).
2009).
inhibition in order to improve bone loss (Malet *et al.*, 2011). Recently a clinical trial of 38 healthy postmenopausal women randomized into placebo or ribonuclease-enriched-LF (R-ELF) groups evaluated bone health status over a period of 12 months. The authors demonstrated that R-ELF supplementation reduced bone resorption and increased osteoblastic bone formation; however BMD was not evaluated (Bharadwaj *et al.*, 2009).
In conclusion, LF has a positive effect on bone health and might be useful in pathological states of reduced bone density. The molecular mechanisms are not fully understood but our studies suggest that dietary bLF supplementation can have a beneficial effect on postmenopausal bone loss not only via a direct effect but also by modulating immune function. The development of pharmaceutical or nutriceutical compounds that are based on LF will require a better understanding of LF's mechanism of action on bone.
#### **5.3 Collagen**
Collagen has a unique triple helix configuration with a repeating sequence (Gly-X-Y)n, with X and Y being mostly proline and hydroxyproline (Hyp) (Bos *et al.*, 1999; Ramshaw *et al.*, 1998). Some studies suggest that a hydrolyzed collagen-enriched diet improves bone collagen metabolism and BMD. Oral administration of Hydrolyzed Collagen (HC) increased BMC and BMD in rats and mice fed a calcium or protein deficient diet (Koyama *et al.*, 2001; Wu *et al.*, 2004).
#### **5.3.1 Ingestion of collagen**
Oral administration of HC was demonstrated to increase the quantity of type I collagen and proteoglycans in the bone matrix of ovariectomized rats (Nomura *et al.*, 2005). Moreover, in patients with osteoporosis, oral intake of HC with calcitonin had a stronger inhibitory effect on bone resorption than calcitonin alone (Adam *et al.*, 1996). Oesser *et al*. demonstrated the intestinal absorption and cartilage accumulation of collagen-derived peptides (Oesser *et al.*, 1999). It has been generally assumed that collagen–rich diets interact with the bone matrix. Indeed, collagen-derived di- and tripeptides rich in hydroxyproline such as Hyp, Pro-Hyp, Pro-Hyp-Gly or Gly-Pro-Val have been detected in human blood following the ingestion of HC (Iwai *et al.*, 2005). The PEPT1 proton-dependent transporter assures the transport of Pro-Hyp across the intestinal barrier (Aito-Inoue *et al.*, 2007).
A study of Ohara *et al*. compared quantity and structures of food-derived gelatin hydrolysates in human blood from fish scale, fish skin and pork skin type I collagen in a single blind crossover study (Ohara *et al.*, 2007). Amounts of free Hyp and Hyp-containing peptide were measured over a 24-h period. Hyp-containing peptides comprised approximately 30% of all detected Hyp. However, efficiency of HC ingestion depends not only on collagen origin but also on the molecular size of the HC. Collagen needs to be hydrolysed to be able to interact with bone metabolism.
#### **5.3.2** *In vivo* **studies**
Our *in vivo* studies indicate that ingestion of HC diet induced the growth of the external diameter of the bone cortical zone in OVX mice (Guillerminet *et al.*, 2010, 2011). The increased cortical area was correlated with a significant increase in the femur external
Dietary Protein and Bone Health 133
Taken together, the results indicate that hydrolyzed collagen modulates bone formation and mineralization of the bone matrix by stimulating osteoblast growth and differentiation while reducing osteoclast differentiation. These effects lead to growth of the external diameter of
Soy contains isoflavones able to bind to estrogen receptors (Folman & Pope, 1969). They have received considerable interest as a possible alternative to conventional Hormone Replacement Therapy (HRT). However, the efficiency of phytoestrogens such as soy
Epidemiological studies suggest that populations with high soy intake (such as Asian populations) have a lower incidence of osteoporotic fractures (Adlercreutz & Mazur, 1997; Schwartz *et al.*, 1999). Asian women typically consume about 20g of soy daily which provides 40 mg of isoflavones (Chen *et al.*, 1999; Ho *et al.*, 2003). However, lower rates of fracture in these populations may not be fully attributed to soy consumption as ethnic related variation in fracture rates can also be explained by differences in bone structure
Many animal studies show that soy protein and/or its isoflavones have positive effects on bone mineral density (BMD) (Arjmandi *et al.*, 1998a, 1998b). However, clinical trial results ranged from no significant changes (Alekel *et al.*, 2000; Dalais *et al.*, 1998; Gallagher *et al.*, 2004; Kreijkamp-Kaspers *et al.*, 2004; Potter *et al.*, 1998) to a slight increase (Chiechi *et al.*, 2002; Lydeking-Olsen *et al.*, 2004; Potter *et al.*, 1998) in BMD. The bone protective effects of
The major isoflavones in soy foods include genistein and diadzein. Genistein 2 has one-third of the potency of estradiol 1 when it interacts with estrogen receptor-b (ER-b), and onethousandth of the potency of estradiol 1 when it interacts with ER-a. Hence Genistein 2 can induce a small estradiol-like response in bone tissues (Adlercreutz & Mazur, 1997; Zhou *et al.*, 2003). Another isoflavone, called equol, is not present in soybean but is a metabolic product of the biotransformation of diadzein by gut bacteria (Setchell *et al.*, 2002). 80% of the Asian population are equol producers (Fujimoto *et al.*, 2008; Morton *et al.*, 2002). In contrast, as few as 25% of individuals in North America and Europe are able to make S-equol (Lampe
A number of reviews describe the effects of dietary soy and isoflavones on bone (Jackson *et al.*, 2011; Messina, 2010; Reinwald & Weaver, 2010). Among the studies exploring the effect of isoflavone-containing food on BMD in postmenopausal women, few report a relationship between soy consumption and the risk of bone fracture. A clinical trial conducted by Marini *et al*. found that in postmenopausal osteopenic Italians receiving 54mg/day genistein for two years, spinal BMD increased by 5.8% (n=150), whereas it decreased in the placebo group by 6.3% (n=154). Similar effects were reported for the hip (Marini *et al.*, 2007). However,
the cortical zone.
(Bouxsein, 2011).
*et al.*, 1998).
**5.4.1 Types of isoflavones**
**5.4.2 Isoflavone and bone fracture**
**5.4 Isoflavone-containing soy protein**
isoflavone on bone is still to be proven.
soy and/or its isoflavones are at best inconclusive.
diameter, without modification of the size of the medullar area. Therefore, the increased size of the cortical area was induced by a periosteal apposition of bone on the mouse femur. Due to this increase in bone size, the ultimate strength of femurs of OVX-mice ingesting HC was significantly greater than the control OVX mice. The increase of the external diameter was also related to a higher level of bone ALP during the first month of HC ingestion. However, the effect was transient; after three months no significant ALP increase was reported. Moreover, HC ingestion was able to increase the bone non-mineral content. There was no significant modification of Young's modulus but bone stiffness increased. Assuming that the stiffness of bone is correlated to the amount of type I collagen present (Burr, 2002; Mann *et al.*, 2001), and since some previous studies showed an increase of type I collagen and proteoglycan excretion for mice fed hydrolyzed collagen, we can propose that HC ingestion increases type I collagen formation in mouse bone.
#### **5.3.3** *In vitro* **studies**
The *in vitro* results obtained with primary tissue culture of murine bone cells confirmed that HC was able to stimulate cell growth and ALP activity. In our studies, all the tested collagens were able to increase osteoblast activity but the 2kDA porcine HC was the most efficient *in vitro* (Guillerminet *et al.*, 2010). Similar observations were also reported with osteoblasts grown on collagen type I films compared to a plastic support with an improvement in various bone markers including increased ALP activity and an accelerated and uniform mineralization of the bone matrix (Lynch *et al.*, 1995). Moreover, our work using the *in vitro* BD BioCoatTM OsteologicTM bone cell culture system showed that PCH-N hydrolyzed collagen did not modify osteoclast growth but reduced osteoclast differentiation (Guillerminet *et al.*, 2010). This effect, combined with increased osteoblast activity is likely to modulate bone turnover leading to the growth of the external diameter of cortical bone.
#### **5.3.4 Mechanism of action**
Several potential mechanisms can be proposed to explain the influence of HC-derived peptides on bone metabolism. Some results have suggested that ingestion of type I hydrolyzed collagen leads to the production and absorption of collagen-derived peptides similar to peptides released from type I collagen *in situ* during bone resorption. Those peptides also act on bone cell metabolism (Adam *et al.*, 1996). Osteoblast activity involves three steps including proliferation, matrix protein synthesis (type I collagen and proteoglycans) and mineralization of the bone matrix (Owen *et al.*, 1991; Quarles *et al.*, 1992; Stein & Lian, 1993). Several hormones and cytokines can modulate osteoblast and osteoclast differentiation and activity. The cytokine TGF-β which is stored in a latent form in the bone matrix, and secreted during the bone resorption phase, is believed to exert such an effect (Oreffo *et al.*, 1989). TGF-β stimulates type I collagen and proteoglycan production while inhibiting that of hydroxyapatite. Interestingly, the type I collagen-derived peptide DGEA (asparagine, glycine, glutamine and alanine), was shown to interact with α2β1 integrin located on the osteoblast cell membrane. This interaction leads to inhibition of TGF-β and consequently bone matrix protein synthesis (Oesser *et al.*, 1999; Takeuchi *et al.*, 1996, 1997; Xiao *et al.*, 1998). Moreover, Hyp is an aromatic AA, and an increase of its concentration can, as suggested previously, increase IGF-1 levels which consequently attenuates bone loss.
diameter, without modification of the size of the medullar area. Therefore, the increased size of the cortical area was induced by a periosteal apposition of bone on the mouse femur. Due to this increase in bone size, the ultimate strength of femurs of OVX-mice ingesting HC was significantly greater than the control OVX mice. The increase of the external diameter was also related to a higher level of bone ALP during the first month of HC ingestion. However, the effect was transient; after three months no significant ALP increase was reported. Moreover, HC ingestion was able to increase the bone non-mineral content. There was no significant modification of Young's modulus but bone stiffness increased. Assuming that the stiffness of bone is correlated to the amount of type I collagen present (Burr, 2002; Mann *et al.*, 2001), and since some previous studies showed an increase of type I collagen and proteoglycan excretion for mice fed hydrolyzed collagen, we can propose that HC ingestion
The *in vitro* results obtained with primary tissue culture of murine bone cells confirmed that HC was able to stimulate cell growth and ALP activity. In our studies, all the tested collagens were able to increase osteoblast activity but the 2kDA porcine HC was the most efficient *in vitro* (Guillerminet *et al.*, 2010). Similar observations were also reported with osteoblasts grown on collagen type I films compared to a plastic support with an improvement in various bone markers including increased ALP activity and an accelerated and uniform mineralization of the bone matrix (Lynch *et al.*, 1995). Moreover, our work using the *in vitro* BD BioCoatTM OsteologicTM bone cell culture system showed that PCH-N hydrolyzed collagen did not modify osteoclast growth but reduced osteoclast differentiation (Guillerminet *et al.*, 2010). This effect, combined with increased osteoblast activity is likely to modulate bone turnover leading to the growth of the external diameter of cortical bone.
Several potential mechanisms can be proposed to explain the influence of HC-derived peptides on bone metabolism. Some results have suggested that ingestion of type I hydrolyzed collagen leads to the production and absorption of collagen-derived peptides similar to peptides released from type I collagen *in situ* during bone resorption. Those peptides also act on bone cell metabolism (Adam *et al.*, 1996). Osteoblast activity involves three steps including proliferation, matrix protein synthesis (type I collagen and proteoglycans) and mineralization of the bone matrix (Owen *et al.*, 1991; Quarles *et al.*, 1992; Stein & Lian, 1993). Several hormones and cytokines can modulate osteoblast and osteoclast differentiation and activity. The cytokine TGF-β which is stored in a latent form in the bone matrix, and secreted during the bone resorption phase, is believed to exert such an effect (Oreffo *et al.*, 1989). TGF-β stimulates type I collagen and proteoglycan production while inhibiting that of hydroxyapatite. Interestingly, the type I collagen-derived peptide DGEA (asparagine, glycine, glutamine and alanine), was shown to interact with α2β1 integrin located on the osteoblast cell membrane. This interaction leads to inhibition of TGF-β and consequently bone matrix protein synthesis (Oesser *et al.*, 1999; Takeuchi *et al.*, 1996, 1997; Xiao *et al.*, 1998). Moreover, Hyp is an aromatic AA, and an increase of its concentration can, as suggested previously, increase IGF-1 levels which consequently attenuates bone loss.
increases type I collagen formation in mouse bone.
**5.3.3** *In vitro* **studies**
**5.3.4 Mechanism of action**
Taken together, the results indicate that hydrolyzed collagen modulates bone formation and mineralization of the bone matrix by stimulating osteoblast growth and differentiation while reducing osteoclast differentiation. These effects lead to growth of the external diameter of the cortical zone.
#### **5.4 Isoflavone-containing soy protein**
Soy contains isoflavones able to bind to estrogen receptors (Folman & Pope, 1969). They have received considerable interest as a possible alternative to conventional Hormone Replacement Therapy (HRT). However, the efficiency of phytoestrogens such as soy isoflavone on bone is still to be proven.
Epidemiological studies suggest that populations with high soy intake (such as Asian populations) have a lower incidence of osteoporotic fractures (Adlercreutz & Mazur, 1997; Schwartz *et al.*, 1999). Asian women typically consume about 20g of soy daily which provides 40 mg of isoflavones (Chen *et al.*, 1999; Ho *et al.*, 2003). However, lower rates of fracture in these populations may not be fully attributed to soy consumption as ethnic related variation in fracture rates can also be explained by differences in bone structure (Bouxsein, 2011).
Many animal studies show that soy protein and/or its isoflavones have positive effects on bone mineral density (BMD) (Arjmandi *et al.*, 1998a, 1998b). However, clinical trial results ranged from no significant changes (Alekel *et al.*, 2000; Dalais *et al.*, 1998; Gallagher *et al.*, 2004; Kreijkamp-Kaspers *et al.*, 2004; Potter *et al.*, 1998) to a slight increase (Chiechi *et al.*, 2002; Lydeking-Olsen *et al.*, 2004; Potter *et al.*, 1998) in BMD. The bone protective effects of soy and/or its isoflavones are at best inconclusive.
#### **5.4.1 Types of isoflavones**
The major isoflavones in soy foods include genistein and diadzein. Genistein 2 has one-third of the potency of estradiol 1 when it interacts with estrogen receptor-b (ER-b), and onethousandth of the potency of estradiol 1 when it interacts with ER-a. Hence Genistein 2 can induce a small estradiol-like response in bone tissues (Adlercreutz & Mazur, 1997; Zhou *et al.*, 2003). Another isoflavone, called equol, is not present in soybean but is a metabolic product of the biotransformation of diadzein by gut bacteria (Setchell *et al.*, 2002). 80% of the Asian population are equol producers (Fujimoto *et al.*, 2008; Morton *et al.*, 2002). In contrast, as few as 25% of individuals in North America and Europe are able to make S-equol (Lampe *et al.*, 1998).
#### **5.4.2 Isoflavone and bone fracture**
A number of reviews describe the effects of dietary soy and isoflavones on bone (Jackson *et al.*, 2011; Messina, 2010; Reinwald & Weaver, 2010). Among the studies exploring the effect of isoflavone-containing food on BMD in postmenopausal women, few report a relationship between soy consumption and the risk of bone fracture. A clinical trial conducted by Marini *et al*. found that in postmenopausal osteopenic Italians receiving 54mg/day genistein for two years, spinal BMD increased by 5.8% (n=150), whereas it decreased in the placebo group by 6.3% (n=154). Similar effects were reported for the hip (Marini *et al.*, 2007). However,
Dietary Protein and Bone Health 135
protein-induced IGF-1 without the negative effect of the acid load by compensating the diet
Dietary protein quality adds complexity to the protein debate. It has been hypothesized that animal protein would be more deleterious to bone than vegetal protein. However, studies show no real difference between those two protein sources. Similarly, long-term observational studies support a benefit of traditional soy food consumption on bone health, but no conclusive evidence supports the hypothesis that this is due to the isoflavones. On the other hand, some peptides obtained from protein digestion have been shown to be helpful to prevent bone loss. Recent results indicate that HC could be of potential interest for nutritional intervention in the prevention of bone loss. Moreover, LF has been reported to have a positive effect on bone health and might be useful in pathological states of reduced bone density. The molecular mechanisms are not fully understood but our studies suggest that dietary bLF supplementation can have a beneficial effect on postmenopausal bone loss
Adam, M.; Spacek, P.; Hulejova, H.; Galianova, A. & Blahos, J. (1996). [Postmenopausal
Adlercreutz, H. & Mazur, W. (1997). Phyto-oestrogens and Western diseases. *Ann Med*, Vol.
Aito-Inoue, M.; Lackeyram, D.; Fan, M.Z.; Sato, K. & Mine, Y. (2007). Transport of a
Alekel, D.L.; Germain, A.S.; Peterson, C.T.; Hanson, K.B.; Stewart, J.W. & Toda, T. (2000).
Alexy, U.; Remer, T.; Manz, F.; Neu, C.M. & Schoenau, E. (2005). Long-term protein intake
Aoe, S.; Koyama, T.; Toba, Y.; Itabashi, A. & Takada, Y. (2005). A controlled trial of the effect
Aoyagi, Y.; Park, H.; Park, S.; Yoshiuchi, K.; Kikuchi, H.; Kawakami, H.; Morita, Y.; Ono, A.
controlled trial. *Int Dairy J*, Vol. 20, No. (Mar 2010), pp. 724-730
osteoporosis. Treatment with calcitonin and a diet rich in collagen proteins]. *Cas Lek*
tripeptide, Gly-Pro-Hyp, across the porcine intestinal brush-border membrane. *J*
Isoflavone-rich soy protein isolate attenuates bone loss in the lumbar spine of perimenopausal women. *Am J Clin Nutr*, Vol. 72, No. 3, (Sep 2000), pp. 844-852 Alekel, D.L.; Van Loan, M.D.; Koehler, K.J.; Hanson, L.N.; Stewart, J.W.; Hanson, K.B.;
Kurzer, M.S. & Peterson, C.T. (2009). The soy isoflavones for reducing bone loss (SIRBL) study: a 3-y randomized controlled trial in postmenopausal women. *Am J*
and dietary potential renal acid load are associated with bone modeling and remodeling at the proximal radius in healthy children. *Am J Clin Nutr*, Vol. 82, No.
of milk basic protein (MBP) supplementation on bone metabolism in healthy menopausal women. *Osteoporos Int*, Vol. 16, No. 12, (Dec 2005), pp. 2123-2128 Aoe, S.; Toba, Y.; Yamamura, J.; Kawakami, H.; Yahiro, M.; Kumegawa, M.; Itabashi, A. &
Takada, Y. (2001). Controlled trial of the effects of milk basic protein (MBP) supplementation on bone metabolism in healthy adult women. *Biosci Biotechnol*
& Shephard RJ (2010). Interactive effects of milk basic protein supplements and habitual physical activity on bone health in older women: A 1-year randomized
not only by acting on bone cells but also by modulating immune function.
*Cesk*, Vol. 135, No. 3, (Jan 31 1996), pp. 74-78
*Pept Sci*, Vol. 13, No. 7, (Jul 2007), pp. 468-474
*Clin Nutr*, Vol. 91, No. 1, (Jan 2009), pp. 218-230
*Biochem*, Vol. 65, No. 4, (Apr 2001), pp. 913-918
29, No. 2, (Apr 1997), pp. 95-120
5, (Nov 2005), pp. 1107-1114
with adequate alkalinizing foods.
**7. References**
recently published long-term trials do not confirm these results; only the trial conducted by Alekel *et al*. reports a modest effect at the femoral neck with 120mg/d isoflavone but no effect with 80mg/d (Alekel *et al.*, 2009).
In contrast clinical trials investigating associations between soy-food intake and BMD in Japanese or Chinese healthy postmenopausal women report that higher isoflavone consumption is associated with lower risk of bone fracture (Ho *et al.*, 2003; Ikeda *et al.*, 2006; Kaneki *et al.*, 2001). Analysis of fracture incidence in the Shanghai cohort (Zhang *et al.*, 2005) and of hip fracture in the Singapore cohort (Koh *et al.*, 2009) shows in both studies one-third reductions in fracture risk when comparing high- with low-soy consumers.
#### **5.4.3 Factors modulating the effect of soy on bone health**
The effectiveness of dietary adaptation of western populations which rarely consumed soy must be considered. East Asian participants in epidemiological studies did not require an adaptation period or an interruption of life-long dietary habits like a western population would. Hence the observation cannot be extrapolated from one population to another.
It may be less difficult to determine bone effects following a life-long intake of traditional foods compared with intermittent intakes of soy. Traditional soy foods are a complex blend of isoflavones, protein, lipids, vitamins, minerals and other bioactive compounds that may act individually and/or synergistically to exert healthy effects. Supplements included in western diets provide quantities of individual soy components. Types of whole soy food consumed (fermented vs. nonfermented) and/or ethnicity (equol producers) may also affect outcome interpretation of soy bone effects.
Long-term observational studies in Asian populations support a benefit of traditional soy food consumption on bone health in this population. The health effects of soy-bean phytoestrogens in non-Asian postmenopausal women are promising. No conclusive evidence supports that the isoflavones from the sources studied do have beneficial effects on bone health. More researches are needed to clarify the role of dietary phytoestrogen in osteoporosis prevention.
#### **6. Conclusion**
Protein acts on bone metabolism at different levels and through different mechanisms. There is little evidence that a high-protein diet will increase bone loss. Protein is well-known to be calciuric, yet there are conflicting data on whether the excreted calcium comes from an increase of calcium absorption or from bone resorption. The direct effects of protein on bone turnover markers and BMD seem to be positive when considering observational studies, but interventional studies do not provide significant outcomes to conclude. Finally, when considering fracture rate, there seems to be a small positive effect of protein on bone as long as calcium levels remain adequate.
Two mechanisms are proposed to explain the action of protein on bone: the acid-ash theory and the hormonal anabolic effect through IGF-1 and CaSR. The hormonal anabolic mechanism supports the fact that protein is beneficial to bone by increasing IGF-1. On the other hand, the acid-ash theory considers that the acid load due to protein consumption is harmful to bone. If both mechanisms occur at the same time, it is possible to benefit from the
recently published long-term trials do not confirm these results; only the trial conducted by Alekel *et al*. reports a modest effect at the femoral neck with 120mg/d isoflavone but no
In contrast clinical trials investigating associations between soy-food intake and BMD in Japanese or Chinese healthy postmenopausal women report that higher isoflavone consumption is associated with lower risk of bone fracture (Ho *et al.*, 2003; Ikeda *et al.*, 2006; Kaneki *et al.*, 2001). Analysis of fracture incidence in the Shanghai cohort (Zhang *et al.*, 2005) and of hip fracture in the Singapore cohort (Koh *et al.*, 2009) shows in both studies one-third
The effectiveness of dietary adaptation of western populations which rarely consumed soy must be considered. East Asian participants in epidemiological studies did not require an adaptation period or an interruption of life-long dietary habits like a western population would. Hence the observation cannot be extrapolated from one population to another.
It may be less difficult to determine bone effects following a life-long intake of traditional foods compared with intermittent intakes of soy. Traditional soy foods are a complex blend of isoflavones, protein, lipids, vitamins, minerals and other bioactive compounds that may act individually and/or synergistically to exert healthy effects. Supplements included in western diets provide quantities of individual soy components. Types of whole soy food consumed (fermented vs. nonfermented) and/or ethnicity (equol producers) may also affect
Long-term observational studies in Asian populations support a benefit of traditional soy food consumption on bone health in this population. The health effects of soy-bean phytoestrogens in non-Asian postmenopausal women are promising. No conclusive evidence supports that the isoflavones from the sources studied do have beneficial effects on bone health. More researches are needed to clarify the role of dietary phytoestrogen in
Protein acts on bone metabolism at different levels and through different mechanisms. There is little evidence that a high-protein diet will increase bone loss. Protein is well-known to be calciuric, yet there are conflicting data on whether the excreted calcium comes from an increase of calcium absorption or from bone resorption. The direct effects of protein on bone turnover markers and BMD seem to be positive when considering observational studies, but interventional studies do not provide significant outcomes to conclude. Finally, when considering fracture rate, there seems to be a small positive effect of protein on bone as long
Two mechanisms are proposed to explain the action of protein on bone: the acid-ash theory and the hormonal anabolic effect through IGF-1 and CaSR. The hormonal anabolic mechanism supports the fact that protein is beneficial to bone by increasing IGF-1. On the other hand, the acid-ash theory considers that the acid load due to protein consumption is harmful to bone. If both mechanisms occur at the same time, it is possible to benefit from the
reductions in fracture risk when comparing high- with low-soy consumers.
**5.4.3 Factors modulating the effect of soy on bone health**
effect with 80mg/d (Alekel *et al.*, 2009).
outcome interpretation of soy bone effects.
osteoporosis prevention.
as calcium levels remain adequate.
**6. Conclusion**
protein-induced IGF-1 without the negative effect of the acid load by compensating the diet with adequate alkalinizing foods.
Dietary protein quality adds complexity to the protein debate. It has been hypothesized that animal protein would be more deleterious to bone than vegetal protein. However, studies show no real difference between those two protein sources. Similarly, long-term observational studies support a benefit of traditional soy food consumption on bone health, but no conclusive evidence supports the hypothesis that this is due to the isoflavones. On the other hand, some peptides obtained from protein digestion have been shown to be helpful to prevent bone loss. Recent results indicate that HC could be of potential interest for nutritional intervention in the prevention of bone loss. Moreover, LF has been reported to have a positive effect on bone health and might be useful in pathological states of reduced bone density. The molecular mechanisms are not fully understood but our studies suggest that dietary bLF supplementation can have a beneficial effect on postmenopausal bone loss not only by acting on bone cells but also by modulating immune function.
#### **7. References**
Dietary Protein and Bone Health 137
Cao, J.J.; Johnson, L.K. & Hunt, J.R. (2011). A diet high in meat protein and potential renal
Ceglia, L.; Harris, S.S.; Abrams, S.A.; Rasmussen, H.M.; Dallal, G.E. & Dawson-Hughes, B.
Chen, Z.; Zheng, W.; Custer, L.J.; Dai, Q.; Shu, X.O.; Jin, F. & Franke, A.A. (1999). Usual
Chevalley, T.; Bonjour, J.P.; Ferrari, S. & Rizzoli, R. (2008). High-protein intake enhances the
Chevalley, T.; Hoffmeyer, P.; Bonjour, J.P. & Rizzoli, R. (2010). Early serum IGF-I response to
Chiechi, L.M.; Secreto, G.; D'Amore, M.; Fanelli, M.; Venturelli, E.; Cantatore, F.; Valerio, T.;
Conigrave, A.D. & Brown, E.M. (2006). Taste receptors in the gastrointestinal tract. II. L-
Conigrave, A.D.; Quinn, S.J. & Brown, E.M. (2000). L-amino acid sensing by the extracellular
Cornish, J. (2004). Lactoferrin promotes bone growth. *Biometals*, Vol. 17, No. 3, (Jun 2004),
Cornish, J.; Callon, K.E.; Naot, D.; Palmano, K.P.; Banovic, T.; Bava, U.; Watson, M.; Lin,
Currey, J. (2001). Sacrificial bonds heal bone. *Nature*, Vol. 414, No. 6865, (Dec 13 2001), pp. 699 Dalais, F.S.; Rice, G.E.; Wahlqvist, M.L.; Grehan, M.; Murkies, A.L.; Medley, G.; Ayton, R. &
Dargent-Molina, P.; Sabia, S.; Touvier, M.; Kesse, E.; Breart, G.; Clavel-Chapelon, F. &
homeostasis. *Annu Rev Nutr*, Vol. 28, No. 2008), pp. 131-155
276, No. 6, (Feb 9 2001), pp. 4070-4079
*Cancer*, Vol. 33, No. 1, 1999), pp. 82-87
Vol. 23, No. 1, (Jan 2008), pp. 131-142
Vol. 29, No. 1, (Feb 2010), pp. 78-83
Vol. 145, No. 9, (Sep 2004), pp. 4366-4374
*Climacteric*, Vol. 1, No. 2, (Jun 1998), pp. 124-129
*Miner Res*, Vol. 23, No. 12, (Dec 2008), pp. 1915-1922
4, (Aug 30 2002), pp. 295-300
4814-4819
pp. 331-335
women. *J Nutr*, Vol. 141, No. 3, (Mar 2011), pp. 391-397
*Clin Endocrinol Metab*, Vol. 94, No. 2, (Feb 2009), pp. 645-653
intracellular calcium mobilization and stimulation of bile flow. *J Biol Chem*, Vol.
acid load increases fractional calcium absorption and urinary calcium excretion without affecting markers of bone resorption or formation in postmenopausal
(2009). Potassium bicarbonate attenuates the urinary nitrogen excretion that accompanies an increase in dietary protein and may promote calcium absorption. *J*
dietary consumption of soy foods and its correlation with the excretion rate of isoflavonoids in overnight urine samples among Chinese women in Shanghai. *Nutr*
positive impact of physical activity on BMC in prepubertal boys. *J Bone Miner Res*,
oral protein supplements in elderly women with a recent hip fracture. *Clin Nutr*,
Laselva, G. & Loizzi, P. (2002). Efficacy of a soy rich diet in preventing postmenopausal osteoporosis: the Menfis randomized trial. *Maturitas*, Vol. 42, No.
amino acid sensing by calcium-sensing receptors: implications for GI physiology. *Am J Physiol Gastrointest Liver Physiol*, Vol. 291, No. 5, (Nov 2006), pp. G753-761 Conigrave, A.D.; Brown, E.M. & Rizzoli, R. (2008). Dietary protein and bone health: roles of
amino acid-sensing receptors in the control of calcium metabolism and bone
Ca2+-sensing receptor. *Proc Natl Acad Sci U S A*, Vol. 97, No. 9, (Apr 25 2000), pp.
J.M.; Tong, P.C.; Chen, Q.; Chan, V.A.; Reid, H.E.; Fazzalari, N.; Baker, H.M.; Baker, E.N.; Haggarty, N.W.; Grey, A.B. & Reid, I.R. (2004). Lactoferrin is a potent regulator of bone cell activity and increases bone formation in vivo. *Endocrinology*,
Strauss, B.J. (1998). Effects of dietary phytoestrogens in postmenopausal women.
Boutron-Ruault, M.C. (2008). Proteins, dietary acid load, and calcium and risk of postmenopausal fractures in the E3N French women prospective study. *J Bone*
Arjmandi, B.H.; Birnbaum, R.; Goyal, N.V.; Getlinger, M.J.; Juma, S.; Alekel, L.; Hasler, C.M.;
Arjmandi, B.H.; Getlinger, M.J.; Goyal, N.V.; Alekel, L.; Hasler, C.M.; Juma, S.; Drum, M.L.;
Arjmandi, B.H.; Lucas, E.A.; Khalil, D.A.; Devareddy, L.; Smith, B.J.; McDonald, J.; Arquitt,
Arnett, T.R. (2008). Extracellular pH regulates bone cell function. *J Nutr*, Vol. 138, No. 2, (Feb
Ballard, T.L.; Clapper, J.A.; Specker, B.L.; Binkley, T.L. & Vukovich, M.D. (2005). Effect of
Barzel, U.S. & Massey, L.K. (1998). Excess dietary protein can adversely affect bone. *J Nutr*,
Beasley, J.M.; Ichikawa, L.E.; Ange, B.A.; Spangler, L.; LaCroix, A.Z.; Ott, S.M. & Scholes, D.
Bharadwaj, S.; Naidu, A.G.; Betageri, G.V.; Prasadarao, N.V. & Naidu, A.S. (2009). Milk
Bonjour, J.P. (2005). Dietary protein: an essential nutrient for bone health. *J Am Coll Nutr*,
Bonjour, J.P.; Ammann, P.; Chevalley, T. & Rizzoli, R. (2001). Protein intake and bone
Bos, K.J.; Rucklidge, G.J.; Dunbar, B. & Robins, S.P. (1999). Primary structure of the helical domain of porcine collagen X. *Matrix Biol*, Vol. 18, No. 2, (Apr 1999), pp. 149-153 Bouxsein, M.L. (2011). Bone structure and fracture risk: do they go arm in arm? *J Bone Miner*
Bouxsein, M.L. & Seeman, E. (2009). Quantifying the material and structural determinants of bone strength. *Best Pract Res Clin Rheumatol*, Vol. 23, No. 6, (Dec 2009), pp. 741-753 Burr, D.B. (2002). The contribution of the organic matrix to bone's material properties. *Bone*,
Canaff, L.; Petit, J.L.; Kisiel, M.; Watson, P.H.; Gascon-Barre, M. & Hendy, G.N. (2001).
Extracellular calcium-sensing receptor is expressed in rat hepatocytes. coupling to
growth. *Can J Appl Physiol*, Vol. 26 Suppl, No. 2001), pp. S153-166
rats. *Am J Clin Nutr*, Vol. 68, No. 6 Suppl, (Dec 1998b), pp. 1358S-1363S Arjmandi, B.H.; Khalil, D.A.; Smith, B.J.; Lucas, E.A.; Juma, S.; Payton, M.E. & Wild, R.A.
*Nutr*, Vol. 68, No. 6 Suppl, (Dec 1998a), pp. 1364S-1368S
women. *Nutr J*, Vol. 4, No. 2005), pp. 8
Vol. 128, No. 6, (Jun 1998), pp. 1051-1053
*Clin Nutr*, Vol. 81, No. 6, (Jun 2005), pp. 1442-1448
*Metab*, Vol. 296, No. 6, (Jun 2009), pp. E1281-1288
Vol. 24, No. 6 Suppl, (Dec 2005), pp. 526S-536S
*Res*, Vol. 26, No. 7, (Jul 2011), pp. 1389-1391
Vol. 31, No. 1, (Jul 2002), pp. 8-11
*Am J Clin Nutr*, Vol. 91, No. 5, (May 2010), pp. 1311-1316
1048-1054
2008), pp. 415S-418S
Drum, M.L.; Hollis, B.W. & Kukreja, S.C. (1998a). Bone-sparing effect of soy protein in ovarian hormone-deficient rats is related to its isoflavone content. *Am J Clin*
Hollis, B.W. & Kukreja, S.C. (1998b). Role of soy protein with normal or reduced isoflavone content in reversing bone loss induced by ovarian hormone deficiency in
(2003). Soy protein has a greater effect on bone in postmenopausal women not on hormone replacement therapy, as evidenced by reducing bone resorption and urinary calcium excretion. *J Clin Endocrinol Metab*, Vol. 88, No. 3, (Mar 2003), pp.
A.B.; Payton, M.E. & Mason, C. (2005). One year soy protein supplementation has positive effects on bone formation markers but not bone density in postmenopausal
protein supplementation during a 6-mo strength and conditioning program on insulin-like growth factor I and markers of bone turnover in young adults. *Am J*
(2010). Is protein intake associated with bone mineral density in young women?
ribonuclease-enriched lactoferrin induces positive effects on bone turnover markers in postmenopausal women. *Osteoporos Int*, Vol. 20, No. 9, (Sep 2009), pp. 1603-1611 Blais, A.; Malet, A.; Mikogami, T.; Martin-Rouas, C. & Tome, D. (2009). Oral bovine
lactoferrin improves bone status of ovariectomized mice. *Am J Physiol Endocrinol*
intracellular calcium mobilization and stimulation of bile flow. *J Biol Chem*, Vol. 276, No. 6, (Feb 9 2001), pp. 4070-4079
Dietary Protein and Bone Health 139
Fujimoto, K.; Tanaka, M.; Hirao, Y.; Nagata, Y.; Mori, M.; Miyanaga, N.; Akaza, H. & Kim,
Gahr, M.; Speer, C.P.; Damerau, B. & Sawatzki, G. (1991). Influence of lactoferrin on the
Gallagher, J.C.; Satpathy, R.; Rafferty, K. & Haynatzka, V. (2004). The effect of soy protein isolate on bone metabolism. *Menopause*, Vol. 11, No. 3, (May-Jun 2004), pp. 290-298 Geusens, P.P. & Boonen, S. (2002). Osteoporosis and the growth hormone-insulin-like
Giannini, S.; Nobile, M.; Dalle Carbonare, L.; Lodetti, M.G.; Sella, S.; Vittadello, G.; Minicuci,
Grassi, F.; Tell, G.; Robbie-Ryan, M.; Gao, Y.; Terauchi, M.; Yang, X.; Romanello, M.; Jones,
*Proc Natl Acad Sci U S A*, Vol. 104, No. 38, (Sep 18 2007), pp. 15087-15092 Grey, A.; Banovic, T.; Zhu, Q.; Watson, M.; Callon, K.; Palmano, K.; Ross, J.; Naot, D.; Reid,
Grey, A.; Zhu, Q.; Watson, M.; Callon, K. & Cornish, J. (2006). Lactoferrin potently inhibits
Guillerminet, F.; Beaupied, H.; Fabien-Soule, V.; Tome, D.; Benhamou, C.L.; Roux, C. &
Guillerminet, F.; Fabien-Soule, V.; Even, P.C.; Tome, D.; Benhamou, C.L.; Roux, C. & Blais,
Guo, H.Y.; Jiang, L.; Ibrahim, S.A.; Zhang, L.; Zhang, H.; Zhang, M. & Ren, F.Z. (2009).
He, F.J.; Marciniak, M.; Carney, C.; Markandu, N.D.; Anand, V.; Fraser, W.D.; Dalton, R.N.;
in mild hypertensives. *Hypertension*, Vol. 55, No. 3, (Mar 2010), pp. 681-688 Heaney, R.P. (2000). Dietary protein and phosphorus do not affect calcium absorption. *Am J*
Heaney, R.P. (2007). Vitamin D endocrine physiology. *J Bone Miner Res*, Vol. 22 Suppl 2, No.
ovariectomized rats. *J Nutr*, Vol. 139, No. 5, (May 2009), pp. 958-964 Hannan, M.T.; Tucker, K.L.; Dawson-Hughes, B.; Cupples, L.A.; Felson, D.T. & Kiel, D.P.
growth factor axis. *Horm Res*, Vol. 58 Suppl 3, No. 2002), pp. 49-55
11, No. 3, 2008), pp. 252-257
2003), pp. 209-213
(Sep 2004), pp. 2268-2278
2504-2512
(Dec 2007), pp. V25-27
251, No. 1-2, (Jun 7 2006), pp. 96-102
*Bone*, Vol. 46, No. 3, (Mar 2010), pp. 827-834
*Clin Nutr*, Vol. 72, No. 3, (Sep 2000), pp. 758-761
ovariectomized C3H/HeN mice. *Osteoporos Int*, In Press
Vol. 49, No. 5, (May 1991), pp. 427-433
W.J. (2008). Age-stratified serum levels of isoflavones and proportion of equol producers in Japanese and Korean healthy men. *Prostate Cancer Prostatic Dis*, Vol.
function of human polymorphonuclear leukocytes and monocytes. *J Leukoc Biol*,
N. & Crepaldi, G. (2003). Hypercalciuria is a common and important finding in postmenopausal women with osteoporosis. *Eur J Endocrinol*, Vol. 149, No. 3, (Sep
D.P.; Weitzmann, M.N. & Pacifici, R. (2007). Oxidative stress causes bone loss in estrogen-deficient mice through enhanced bone marrow dendritic cell activation.
I.R. & Cornish, J. (2004). The low-density lipoprotein receptor-related protein 1 is a mitogenic receptor for lactoferrin in osteoblastic cells. *Mol Endocrinol*, Vol. 18, No. 9,
osteoblast apoptosis, via an LRP1-independent pathway. *Mol Cell Endocrinol*, Vol.
Blais, A. (2010). Hydrolyzed collagen improves bone metabolism and biomechanical parameters in ovariectomized mice: an in vitro and in vivo study.
A. (2011). Hydrolyzed collagen improves bone status and prevents bone loss in
Orally administered lactoferrin preserves bone mass and microarchitecture in
(2000). Effect of dietary protein on bone loss in elderly men and women: the Framingham Osteoporosis Study. *J Bone Miner Res*, Vol. 15, No. 12, (Dec 2000), pp.
Kaski, J.C. & MacGregor, G.A. (2010). Effects of potassium chloride and potassium bicarbonate on endothelial function, cardiovascular risk factors, and bone turnover
Darling, A.L.; Millward, D.J.; Torgerson, D.J.; Hewitt, C.E. & Lanham-New, S.A. (2009).
Dawson-Hughes, B. & Harris, S.S. (2002). Calcium intake influences the association of
Dawson-Hughes, B.; Harris, S.S.; Palermo, N.J.; Castaneda-Sceppa, C.; Rasmussen, H.M. &
Dawson-Hughes, B.; Harris, S.S.; Rasmussen, H.; Song, L. & Dallal, G.E. (2004). Effect of
Debbabi, H.; Dubarry, M.; Rautureau, M. & Tome, D. (1998). Bovine lactoferrin induces both
Delmi, M.; Rapin, C.H.; Bengoa, J.M.; Delmas, P.D.; Vasey, H. & Bonjour, J.P. (1990). Dietary
Devine, A.; Dick, I.M.; Islam, A.F.; Dhaliwal, S.S. & Prince, R.L. (2005). Protein consumption
Domrongkitchaiporn, S.; Pongskul, C.; Sirikulchayanonta, V.; Stitchantrakul, W.; Leeprasert,
Fenton, T.R.; Tough, S.C.; Lyon, A.W.; Eliasziw, M. & Hanley, D.A. (2011). Causal assessment
Frassetto, L.; Morris, R.C., Jr. & Sebastian, A. (2005). Long-term persistence of the urine
Frassetto, L.A.; Todd, K.M.; Morris, R.C., Jr. & Sebastian, A. (1998). Estimation of net
protein contents. *Am J Clin Nutr*, Vol. 68, No. 3, (Sep 1998), pp. 576-583
Hill's epidemiologic criteria for causality. *Nutr J*, Vol. 10, No. 2011), pp. 41 Feskanich, D.; Willett, W.C.; Stampfer, M.J. & Colditz, G.A. (1996). Protein consumption and bone fractures in women. *Am J Epidemiol*, Vol. 143, No. 5, (Mar 1 1996), pp. 472-479 Folman, Y. & Pope, G.S. (1969). Effect of norethisterone acetate, dimethylstilboestrol, genistein
immature mice. *J Endocrinol*, Vol. 44, No. 2, (Jun 1969), pp. 213-218
*Endocrinol Metab*, Vol. 90, No. 2, (Feb 2005), pp. 831-834
tubular acidosis. *Kidney Int*, Vol. 62, No. 6, (Dec 2002), pp. 2160-2166 Fenton, T.R.; Eliasziw, M.; Lyon, A.W.; Tough, S.C. & Hanley, D.A. (2008). Meta-analysis of
women. *J Clin Endocrinol Metab*, Vol. 89, No. 3, (Mar 2004), pp. 1169-1173 Dawson-Hughes, B.; Harris, S.S.; Rasmussen, H.M. & Dallal, G.E. (2007). Comparative
in humans. *Osteoporos Int*, Vol. 18, No. 7, (Jul 2007), pp. 955-961
*Nutr*, Vol. 90, No. 6, (Dec 2009), pp. 1674-1692
Vol. 75, No. 4, (Apr 2002), pp. 773-779
335, No. 8696, (Apr 28 1990), pp. 1013-1016
*Nutr*, Vol. 81, No. 6, (Jun 2005), pp. 1423-1428
1, (Jan 2009), pp. 96-102
1998), pp. 283-293
2008), pp. 1159-1166
Dietary protein and bone health: a systematic review and meta-analysis. *Am J Clin*
protein intake with rates of bone loss in elderly men and women. *Am J Clin Nutr*,
Dallal, G.E. (2009). Treatment with potassium bicarbonate lowers calcium excretion and bone resorption in older men and women. *J Clin Endocrinol Metab*, Vol. 94, No.
dietary protein supplements on calcium excretion in healthy older men and
effects of oral aromatic and branched-chain amino acids on urine calcium excretion
mucosal and systemic immune response in mice. *J Dairy Res*, Vol. 65, No. 2, (May
supplementation in elderly patients with fractured neck of the femur. *Lancet*, Vol.
is an important predictor of lower limb bone mass in elderly women. *Am J Clin*
V.; Ongphiphadhanakul, B.; Radinahamed, P. & Rajatanavin, R. (2002). Bone histology and bone mineral density after correction of acidosis in distal renal
the quantity of calcium excretion associated with the net acid excretion of the modern diet under the acid-ash diet hypothesis. *Am J Clin Nutr*, Vol. 88, No. 4, (Oct
of dietary acid load and bone disease: a systematic review & meta-analysis applying
and coumestrol on uptake of [3H]oestradiol by uterus, vagina and skeletal muscle of
calcium-lowering effect of potassium bicarbonate in postmenopausal women. *J Clin*
endogenous noncarbonic acid production in humans from diet potassium and
Dietary Protein and Bone Health 141
Kerstetter, J.E.; O'Brien, K.O. & Insogna, K.L. (2003). Dietary protein, calcium metabolism,
Kerstetter, J.E.; Wall, D.E.; O'Brien, K.O.; Caseria, D.M. & Insogna, K.L. (2006). Meat and soy
Khalil, D.A.; Lucas, E.A.; Juma, S.; Smith, B.J.; Payton, M.E. & Arjmandi, B.H. (2002). Soy
Koh, W.P.; Wu, A.H.; Wang, R.; Ang, L.W.; Heng, D.; Yuan, J.M. & Yu, M.C. (2009). Gender-
Kreijkamp-Kaspers, S.; Kok, L.; Grobbee, D.E.; de Haan, E.H.; Aleman, A.; Lampe, J.W. &
Lean, J.M.; Davies, J.T.; Fuller, K.; Jagger, C.J.; Kirstein, B.; Partington, G.A.; Urry, Z.L. &
Legrand, D.; Elass, E.; Carpentier, M. & Mazurier, J. (2005). Lactoferrin: a modulator of
Legrand, D.; Elass, E.; Carpentier, M. & Mazurier, J. (2006). Interactions of lactoferrin with cells
Legrand, D.; Elass, E.; Pierce, A. & Mazurier, J. (2004). Lactoferrin and host defence: an
Lemann, J., Jr.; Bushinsky, D.A. & Hamm, L.L. (2003). Bone buffering of acid and base in humans. *Am J Physiol Renal Physiol*, Vol. 285, No. 5, (Nov 2003), pp. F811-832 Lemann, J., Jr.; Pleuss, J.A.; Gray, R.W. & Hoffmann, R.G. (1991). Potassium administration
adults [corrected]. *Kidney Int*, Vol. 39, No. 5, (May 1991), pp. 973-983 Lorget, F.; Clough, J.; Oliveira, M.; Daury, M.C.; Sabokbar, A. & Offord, E. (2002).
trial. *Eur J Nutr*, Vol. 43, No. 4, (Aug 2004), pp. 246-257
*Biochem Biophys Res Commun*, Vol. 296, No. 2, (Aug 16 2002), pp. 261-266 Lydeking-Olsen, E.; Beck-Jensen, J.E.; Setchell, K.D. & Holm-Jensen, T. (2004). Soymilk or
Lynch, M.P.; Stein, J.L.; Stein, G.S. & Lian, J.B. (1995). The influence of type I collagen on the
bone loss. *J Clin Invest*, Vol. 112, No. 6, (Sep 2003), pp. 915-923
Health Study. *Am J Epidemiol*, Vol. 170, No. 7, (Oct 1 2009), pp. 901-909 Koyama, Y.; Hirota, A.; Mori, H.; Takahara, H.; Kuwaba, K.; Kusubata, M.; Matsubara, Y.;
*Vitaminol (Tokyo)*, Vol. 47, No. 1, (Feb 2001), pp. 84-86
pp. 584S-592S
2006), pp. 1890-1895
1998), pp. 335-339
pp. 2549-2559
Vol. 17, No. 3, (Jun 2004), pp. 225-229
(Sep 2002), pp. 2605-2608
and skeletal homeostasis revisited. *Am J Clin Nutr*, Vol. 78, No. 3 Suppl, (Sep 2003),
protein affect calcium homeostasis in healthy women. *J Nutr*, Vol. 136, No. 7, (Jul
protein supplementation increases serum insulin-like growth factor-I in young and old men but does not affect markers of bone metabolism. *J Nutr*, Vol. 132, No. 9,
specific associations between soy and risk of hip fracture in the Singapore Chinese
Kasugai, S.; Itoh, M. & Irie, S. (2001). Ingestion of gelatin has differential effect on bone mineral density and body weight in protein undernutrition. *J Nutr Sci*
van der Schouw, Y.T. (2004). Effect of soy protein containing isoflavones on cognitive function, bone mineral density, and plasma lipids in postmenopausal women: a randomized controlled trial. *Jama*, Vol. 292, No. 1, (Jul 7 2004), pp. 65-74 Lampe, J.W.; Karr, S.C.; Hutchins, A.M. & Slavin, J.L. (1998). Urinary equol excretion with a
soy challenge: influence of habitual diet. *Proc Soc Exp Biol Med*, Vol. 217, No. 3, (Mar
Chambers, T.J. (2003). A crucial role for thiol antioxidants in estrogen-deficiency
immune and inflammatory responses. *Cell Mol Life Sci*, Vol. 62, No. 22, (Nov 2005),
involved in immune function. *Biochem Cell Biol*, Vol. 84, No. 3, (Jun 2006), pp. 282-290
overview of its immuno-modulating and anti-inflammatory properties. *Biometals*,
reduces and potassium deprivation increases urinary calcium excretion in healthy
Lactoferrin reduces in vitro osteoclast differentiation and resorbing activity.
progesterone for prevention of bone loss--a 2 year randomized, placebo-controlled
development and maintenance of the osteoblast phenotype in primary and
Ho, S.C.; Woo, J.; Lam, S.; Chen, Y.; Sham, A. & Lau, J. (2003). Soy protein consumption and
Hunt, J.R.; Johnson, L.K. & Fariba Roughead, Z.K. (2009). Dietary protein and calcium
Ikeda, Y.; Iki, M.; Morita, A.; Kajita, E.; Kagamimori, S.; Kagawa, Y. & Yoneshima, H. (2006).
Ilich, J.Z.; Brownbill, R.A. & Tamborini, L. (2003). Bone and nutrition in elderly women:
Ince, B.A.; Anderson, E.J. & Neer, R.M. (2004). Lowering dietary protein to U.S. Recommended
Jackson, R.L.; Greiwe, J.S. & Schwen, R.J. (2011). Emerging evidence of the health benefits of
Jajoo, R.; Song, L.; Rasmussen, H.; Harris, S.S. & Dawson-Hughes, B. (2006). Dietary acid-
Kaneki, M.; Hodges, S.J.; Hosoi, T.; Fujiwara, S.; Lyons, A.; Crean, S.J.; Ishida, N.; Nakagawa,
Kerstetter, J.E. & Allen, L.H. (1990). Dietary protein increases urinary calcium. *J Nutr*, Vol.
Kerstetter, J.E.; Caseria, D.M.; Mitnick, M.E.; Ellison, A.F.; Gay, L.F.; Liskov, T.A.; Carpenter,
Kerstetter, J.E.; Mitnick, M.E.; Gundberg, C.M.; Caseria, D.M.; Ellison, A.F.; Carpenter, T.O.
Kerstetter, J.E.; O'Brien, K.O.; Caseria, D.M.; Wall, D.E. & Insogna, K.L. (2005). The impact of
Kerstetter, J.E.; O'Brien, K.O. & Insogna, K.L. (1998). Dietary protein affects intestinal calcium absorption. *Am J Clin Nutr*, Vol. 68, No. 4, (Oct 1998), pp. 859-865
women. *J Clin Endocrinol Metab*, Vol. 90, No. 1, (Jan 2005), pp. 26-31
(Oct 2003), pp. 835-842
432-448
(Jun 2006), pp. 224-230
(Apr 2001), pp. 315-321
pp. 1052-1055
120, No. 1, (Jan 1990), pp. 134-136
*Nutr*, Vol. 66, No. 5, (Nov 1997), pp. 1188-1196
Vol. 89, No. 5, (May 2009), pp. 1357-1365
*Nutr*, Vol. 136, No. 5, (May 2006), pp. 1323-1328
*Clin Nutr*, Vol. 57, No. 4, (Apr 2003), pp. 554-565
*Chem*, Vol. 53, No. 16, (Aug 10 2005), pp. 6531-6536
bone mass in early postmenopausal Chinese women. *Osteoporos Int*, Vol. 14, No. 10,
interact to influence calcium retention: a controlled feeding study. *Am J Clin Nutr*,
Intake of fermented soybeans, natto, is associated with reduced bone loss in postmenopausal women: Japanese Population-Based Osteoporosis (JPOS) Study. *J*
protein, energy, and calcium as main determinants of bone mineral density. *Eur J*
dietary allowance levels reduces urinary calcium excretion and bone resorption in young women. *J Clin Endocrinol Metab*, Vol. 89, No. 8, (Aug 2004), pp. 3801-3807 Iwai, K.; Hasegawa, T.; Taguchi, Y.; Morimatsu, F.; Sato, K.; Nakamura, Y.; Higashi, A.;
Kido, Y.; Nakabo, Y. & Ohtsuki, K. (2005). Identification of food-derived collagen peptides in human blood after oral ingestion of gelatin hydrolysates. *J Agric Food*
S-equol, an estrogen receptor beta agonist. *Nutr Rev*, Vol. 69, No. 8, (Aug 2011), pp.
base balance, bone resorption, and calcium excretion. *J Am Coll Nutr*, Vol. 25, No. 3,
M.; Takechi, M.; Sano, Y.; Mizuno, Y.; Hoshino, S.; Miyao, M.; Inoue, S.; Horiki, K.; Shiraki, M.; Ouchi, Y. & Orimo, H. (2001). Japanese fermented soybean food as the major determinant of the large geographic difference in circulating levels of vitamin K2: possible implications for hip-fracture risk. *Nutrition*, Vol. 17, No. 4,
T.O. & Insogna, K.L. (1997). Increased circulating concentrations of parathyroid hormone in healthy, young women consuming a protein-restricted diet. *Am J Clin*
& Insogna, K.L. (1999). Changes in bone turnover in young women consuming different levels of dietary protein. *J Clin Endocrinol Metab*, Vol. 84, No. 3, (Mar 1999),
dietary protein on calcium absorption and kinetic measures of bone turnover in
Dietary Protein and Bone Health 143
Munger, R.G.; Cerhan, J.R. & Chiu, B.C. (1999). Prospective study of dietary protein intake
Muthusami, S.; Ramachandran, I.; Muthusamy, B.; Vasudevan, G.; Prabhu, V.;
Naot, D.; Grey, A.; Reid, I.R. & Cornish, J. (2005). Lactoferrin--a novel bone growth factor.
New, S.A.; MacDonald, H.M.; Campbell, M.K.; Martin, J.C.; Garton, M.J.; Robins, S.P. &
Nomura, Y.; Oohashi, K.; Watanabe, M. & Kasugai, S. (2005). Increase in bone mineral
Oesser, S.; Adam, M.; Babel, W. & Seifert, J. (1999). Oral administration of (14)C labeled
Oh, S.M.; Lee, S.H.; Lee, B.J.; Pyo, C.W.; Yoo, N.K.; Lee, S.Y.; Kim, J. & Choi, S.Y. (2007). A
Ohara, H.; Matsumoto, H.; Ito, K.; Iwai, K. & Sato, K. (2007). Comparison of quantity and
Ohlsson, C.; Mohan, S.; Sjogren, K.; Tivesten, A.; Isgaard, J.; Isaksson, O.; Jansson, J.O. &
Oreffo, R.O.; Mundy, G.R.; Seyedin, S.M. & Bonewald, L.F. (1989). Activation of the bone-
Owen, T.A.; Aronow, M.S.; Barone, L.M.; Bettencourt, B.; Stein, G.S. & Lian, J.B. (1991).
Perrini, S.; Laviola, L.; Carreira, M.C.; Cignarelli, A.; Natalicchio, A. & Giorgino, F. (2010).
(Jan 1999), pp. 147-152
Vol. 360, No. 1-2, (Oct 2005), pp. 81-86
*Clin Med Res*, Vol. 3, No. 2, (May 2005), pp. 93-101
*Nephrol*, Vol. 17, No. 3, (Mar 2006), pp. 617-626
*Nutrition*, Vol. 21, No. 11-12, (Nov-Dec 2005), pp. 1120-1126
(C57/BL). *J Nutr*, Vol. 129, No. 10, (Oct 1999), pp. 1891-1895
*Immunol*, Vol. 179, No. 9, (Nov 1 2007), pp. 5686-5692
No. 4, (Feb 21 2007), pp. 1532-1535
205, No. 3, (Jun 2010), pp. 201-210
1496-1504
*Rev*, Vol. 30, No. 5, (Aug 2009), pp. 494-535
*Commun*, Vol. 158, No. 3, (Feb 15 1989), pp. 817-823
and risk of hip fracture in postmenopausal women. *Am J Clin Nutr*, Vol. 69, No. 1,
Subramaniam, V.; Jagadeesan, A. & Narasimhan, S. (2005). Ovariectomy induces oxidative stress and impairs bone antioxidant system in adult rats. *Clin Chim Acta*,
Reid, D.M. (2004). Lower estimates of net endogenous non-carbonic acid production are positively associated with indexes of bone health in premenopausal and perimenopausal women. *Am J Clin Nutr*, Vol. 79, No. 1, (Jan 2004), pp. 131-138 Nijenhuis, T.; Renkema, K.Y.; Hoenderop, J.G. & Bindels, R.J. (2006). Acid-base status
determines the renal expression of Ca2+ and Mg2+ transport proteins. *J Am Soc*
density through oral administration of shark gelatin to ovariectomized rats.
gelatin hydrolysate leads to an accumulation of radioactivity in cartilage of mice
distinct role of neutrophil lactoferrin in RelA/p65 phosphorylation on Ser536 by recruiting TNF receptor-associated factors to IkappaB kinase signaling complex. *J*
structures of hydroxyproline-containing peptides in human blood after oral ingestion of gelatin hydrolysates from different sources. *J Agric Food Chem*, Vol. 55,
Svensson, J. (2009). The role of liver-derived insulin-like growth factor-I. *Endocr*
derived latent TGF beta complex by isolated osteoclasts. *Biochem Biophys Res*
Pleiotropic effects of vitamin D on osteoblast gene expression are related to the proliferative and differentiated state of the bone cell phenotype: dependency upon basal levels of gene expression, duration of exposure, and bone matrix competency in normal rat osteoblast cultures. *Endocrinology*, Vol. 128, No. 3, (Mar 1991), pp.
The GH/IGF1 axis and signaling pathways in the muscle and bone: mechanisms underlying age-related skeletal muscle wasting and osteoporosis. *J Endocrinol*, Vol.
passaged rat calvarial osteoblasts: modification of expression of genes supporting cell growth, adhesion, and extracellular matrix mineralization. *Exp Cell Res*, Vol. 216, No. 1, (Jan 1995), pp. 35-45
Malet, A.; Bournaud, E.; Lan, A.; Mikogami, T.; Tome, D. & Blais, A. (2011). Bovine
Mallegol, J.; van Niel, G. & Heyman, M. (2005). Phenotypic and functional characterization
Mann, V.; Hobson, E.E.; Li, B.; Stewart, T.L.; Grant, S.F.; Robins, S.P.; Aspden, R.M. &
Marie, P.J. (2010). The calcium-sensing receptor in bone cells: a potential therapeutic target
Marini, H.; Minutoli, L.; Polito, F.; Bitto, A.; Altavilla, D.; Atteritano, M.; Gaudio, A.;
Massey, L.K. (2003). Dietary animal and plant protein and human bone health: a whole
Massey, L.K. & Kynast-Gales, S.A. (2001). Diets with either beef or plant proteins reduce risk
Maurer, M.; Riesen, W.; Muser, J.; Hulter, H.N. & Krapf, R. (2003). Neutralization of Western
in humans. *Am J Physiol Renal Physiol*, Vol. 284, No. 1, (Jan 2003), pp. F32-40 Mazess, R.B. & Barden, H.S. (1991). Bone density in premenopausal women: effects of age,
Messina, M. (2010). Insights gained from 20 years of soy research. *J Nutr*, Vol. 140, No. 12,
Metz, J.A.; Anderson, J.J. & Gallagher, P.N., Jr. (1993). Intakes of calcium, phosphorus, and
Meyer, H.E.; Pedersen, J.I.; Loken, E.B. & Tverdal, A. (1997). Dietary factors and the
Morita, Y.; Matsuyama, H.; Serizawa, A.; Takeya, T. & Kawakami, H. (2008). Identification
Morton, M.S.; Arisaka, O.; Miyake, N.; Morgan, L.D. & Evans, B.A. (2002). Phytoestrogen
women. *Am J Clin Nutr*, Vol. 58, No. 4, (Oct 1993), pp. 537-542
*Epidemiol*, Vol. 145, No. 2, (Jan 15 1997), pp. 117-123
*Bone*, Vol. 42, No. 2, (Feb 2008), pp. 380-387
*Nutr*, Vol. 132, No. 10, (Oct 2002), pp. 3168-3171
foods approach. *J Nutr*, Vol. 133, No. 3, (Mar 2003), pp. 862S-865S
*J Am Diet Assoc*, Vol. 101, No. 3, (Mar 2001), pp. 326-331
Vol. 53, No. 1, (Jan 1991), pp. 132-142
(Dec 2010), pp. 2289S-2295S
modulation. *Bone*, Vol. 48, No. 5, (May 1 2011), pp. 1028-1035
in osteoporosis. *Bone*, Vol. 46, No. 3, (Mar 2010), pp. 571-576
216, No. 1, (Jan 1995), pp. 35-45
No. 7, (Apr 2001), pp. 899-907
pp. 11-16
847
passaged rat calvarial osteoblasts: modification of expression of genes supporting cell growth, adhesion, and extracellular matrix mineralization. *Exp Cell Res*, Vol.
lactoferrin improves bone status of ovariectomized mice via immune function
of intestinal epithelial exosomes. *Blood Cells Mol Dis*, Vol. 35, No. 1, (Jul-Aug 2005),
Ralston, S.H. (2001). A COL1A1 Sp1 binding site polymorphism predisposes to osteoporotic fracture by affecting bone density and quality. *J Clin Invest*, Vol. 107,
Mazzaferro, S.; Frisina, A.; Frisina, N.; Lubrano, C.; Bonaiuto, M.; D'Anna, R.; Cannata, M.L.; Corrado, F.; Adamo, E.B.; Wilson, S. & Squadrito, F. (2007). Effects of the phytoestrogen genistein on bone metabolism in osteopenic postmenopausal women: a randomized trial. *Ann Intern Med*, Vol. 146, No. 12, (Jun 19 2007), pp. 839-
of calcium oxalate precipitation in patients with a history of calcium kidney stones.
diet inhibits bone resorption independently of K intake and reduces cortisol secretion
dietary intake, physical activity, smoking, and birth-control pills. *Am J Clin Nutr*,
protein, and physical-activity level are related to radial bone mass in young adult
incidence of hip fracture in middle-aged Norwegians. A prospective study. *Am J*
of angiogenin as the osteoclastic bone resorption-inhibitory factor in bovine milk.
concentrations in serum from Japanese men and women over forty years of age. *J*
Dietary Protein and Bone Health 145
Shanker, G.; Sorci-Thomas, M. & Adams, M.R. (1994). Estrogen modulates the expression of
THP-1 cells. *Lymphokine Cytokine Res*, Vol. 13, No. 6, (Dec 1994), pp. 377-382 Stein, G.S. & Lian, J.B. (1993). Molecular mechanisms mediating proliferation/differentiation
Suda, T.; Takahashi, N.; Udagawa, N.; Jimi, E.; Gillespie, M.T. & Martin, T.J. (1999).
Takayama, Y. & Mizumachi, K. (2008). Effect of bovine lactoferrin on extracellular matrix
Takeuchi, Y.; Nakayama, K. & Matsumoto, T. (1996). Differentiation and cell surface
Takeuchi, Y.; Suzawa, M.; Kikuchi, T.; Nishida, E.; Fujita, T. & Matsumoto, T. (1997).
Teegarden, D.; Lyle, R.M.; McCabe, G.P.; McCabe, L.D.; Proulx, W.R.; Michon, K.; Knight,
Thorpe, D.L.; Knutsen, S.F.; Beeson, W.L.; Rajaram, S. & Fraser, G.E. (2008a). Effects of meat
Thorpe, M. & Evans, E.M. (2011). Dietary protein and bone health: harmonizing conflicting
Thorpe, M.; Mojtahedi, M.C.; Chapman-Novakofski, K.; McAuley, E. & Evans, E.M. (2008b).
Tkatch, L.; Rapin, C.H.; Rizzoli, R.; Slosman, D.; Nydegger, V.; Vasey, H. & Bonjour, J.P.
the proximal femur. *J Am Coll Nutr*, Vol. 11, No. 5, (Oct 1992), pp. 519-525 Turner, C.H. (2006). Bone strength: current concepts. *Ann N Y Acad Sci*, Vol. 1068, No. (Apr
Uenishi, K.; Ishida, H.; Toba, Y.; Aoe, S.; Itabashi, A. & Takada, Y. (2007). Milk basic protein
Vatanparast, H.; Bailey, D.A.; Baxter-Jones, A.D. & Whiting, S.J. (2007). The effects of dietary
women. *Osteoporos Int*, Vol. 18, No. 3, (Mar 2007), pp. 385-390
calcium intake. *J Nutr*, Vol. 137, No. 12, (Dec 2007), pp. 2674-2679
*Endocr Rev*, Vol. 14, No. 4, (Aug 1993), pp. 424-442
1999), pp. 345-357
(Jan 2008), pp. 226-230
No. 7, (Feb 16 1996), pp. 3938-3944
(Nov 14 1997), pp. 29309-29316
pp. 564-572
(Jan 2008b), pp. 80-85
2006), pp. 429-446
*Clin Nutr*, Vol. 68, No. 3, (Sep 1998), pp. 749-754
theories. *Nutr Rev*, Vol. 69, No. 4, (Apr 2011), pp. 215-230
tumor necrosis factor alpha mRNA in phorbol ester-stimulated human monocytic
interrelationships during progressive development of the osteoblast phenotype.
Modulation of osteoclast differentiation and function by the new members of the tumor necrosis factor receptor and ligand families. *Endocr Rev*, Vol. 20, No. 3, (Jun
calcification by human osteoblast-like cells. *Biosci Biotechnol Biochem*, Vol. 72, No. 1,
expression of transforming growth factor-beta receptors are regulated by interaction with matrix collagen in murine osteoblastic cells. *J Biol Chem*, Vol. 271,
Differentiation and transforming growth factor-beta receptor down-regulation by collagen-alpha2beta1 integrin interaction is mediated by focal adhesion kinase and its downstream signals in murine osteoblastic cells. *J Biol Chem*, Vol. 272, No. 46,
A.P.; Johnston, C.C. & Weaver, C.M. (1998). Dietary calcium, protein, and phosphorus are related to bone mineral density and content in young women. *Am J*
consumption and vegetarian diet on risk of wrist fracture over 25 years in a cohort of peri- and postmenopausal women. *Public Health Nutr*, Vol. 11, No. 6, (Jun 2008a),
A positive association of lumbar spine bone mineral density with dietary protein is suppressed by a negative association with protein sulfur. *J Nutr*, Vol. 138, No. 1,
(1992). Benefits of oral protein supplementation in elderly patients with fracture of
increases bone mineral density and improves bone metabolism in healthy young
protein on bone mineral mass in young adults may be modulated by adolescent
Potter, S.M.; Baum, J.A.; Teng, H.; Stillman, R.J.; Shay, N.F. & Erdman, J.W., Jr. (1998). Soy
Promislow, J.H.; Goodman-Gruen, D.; Slymen, D.J. & Barrett-Connor, E. (2002). Protein
Quarles, L.D.; Yohay, D.A.; Lever, L.W.; Caton, R. & Wenstrup, R.J. (1992). Distinct
Rahbar, A.; Larijani, B.; Nabipour, I.; Mohamadi, M.M.; Mirzaee, K. & Amiri, Z. (2009).
Ramshaw, J.A.; Shah, N.K. & Brodsky, B. (1998). Gly-X-Y tripeptide frequencies in collagen:
Rapuri, P.B.; Gallagher, J.C. & Haynatzka, V. (2003). Protein intake: effects on bone mineral
Reinwald, S. & Weaver, C.M. (2010). Soy components vs. whole soy: are we betting our bones on a long shot? *J Nutr*, Vol. 140, No. 12, (Dec 2010), pp. 2312S-2317S Roughead, Z.K.; Johnson, L.K.; Lykken, G.I. & Hunt, J.R. (2003). Controlled high meat diets
Schurch, M.A.; Rizzoli, R.; Slosman, D.; Vadas, L.; Vergnaud, P. & Bonjour, J.P. (1998).
Schwartz, A.V.; Kelsey, J.L.; Maggi, S.; Tuttleman, M.; Ho, S.C.; Jonsson, P.V.; Poor, G.;
Seeman, E. & Delmas, P.D. (2006). Bone quality--the material and structural basis of bone strength and fragility. *N Engl J Med*, Vol. 354, No. 21, (May 25 2006), pp. 2250-2261 Sellmeyer, D.E.; Stone, K.L.; Sebastian, A. & Cummings, S.R. (2001). A high ratio of dietary
Setchell, K.D.; Brown, N.M. & Lydeking-Olsen, E. (2002). The clinical importance of the
*Am J Epidemiol*, Vol. 155, No. 7, (Apr 1 2002), pp. 636-644
population. *Bone*, Vol. 45, No. 5, (Nov 2009), pp. 876-881
women. *J Nutr*, Vol. 133, No. 4, (Apr 2003), pp. 1020-1026
1375S-1379S
pp. 683-692
pp. 86-91
(Jun 2003), pp. 1517-1525
10, (May 15 1998), pp. 801-809
*Osteoporos Int*, Vol. 9, No. 3, 1999), pp. 242-253
*Clin Nutr*, Vol. 73, No. 1, (Jan 2001), pp. 118-122
132, No. 12, (Dec 2002), pp. 3577-3584
protein and isoflavones: their effects on blood lipids and bone density in postmenopausal women. *Am J Clin Nutr*, Vol. 68, No. 6 Suppl, (Dec 1998), pp.
consumption and bone mineral density in the elderly : the Rancho Bernardo Study.
proliferative and differentiated stages of murine MC3T3-E1 cells in culture: an in vitro model of osteoblast development. *J Bone Miner Res*, Vol. 7, No. 6, (Jun 1992),
Relationship among dietary estimates of net endogenous acid production, bone mineral density and biochemical markers of bone turnover in an Iranian general
a context for host-guest triple-helical peptides. *J Struct Biol*, Vol. 122, No. 1-2, 1998),
density and the rate of bone loss in elderly women. *Am J Clin Nutr*, Vol. 77, No. 6,
do not affect calcium retention or indices of bone status in healthy postmenopausal
Protein supplements increase serum insulin-like growth factor-I levels and attenuate proximal femur bone loss in patients with recent hip fracture. A randomized, double-blind, placebo-controlled trial. *Ann Intern Med*, Vol. 128, No.
Sisson de Castro, J.A.; Xu, L.; Matkin, C.C.; Nelson, L.M. & Heyse, S.P. (1999). International variation in the incidence of hip fractures: cross-national project on osteoporosis for the World Health Organization Program for Research on Aging.
animal to vegetable protein increases the rate of bone loss and the risk of fracture in postmenopausal women. Study of Osteoporotic Fractures Research Group. *Am J*
metabolite equol-a clue to the effectiveness of soy and its isoflavones. *J Nutr*, Vol.
Weikert, C.; Walter, D.; Hoffmann, K.; Kroke, A.; Bergmann, M.M. & Boeing, H. (2005). The
Wengreen, H.J.; Munger, R.G.; West, N.A.; Cutler, D.R.; Corcoran, C.D.; Zhang, J. & Sassano,
residents of Utah. *J Bone Miner Res*, Vol. 19, No. 4, (Apr 2004), pp. 537-545 Whiting, S.J.; Anderson, D.J. & Weeks, S.J. (1997). Calciuric effects of protein and potassium
Wilson, J. & Wilson, G.J. (2006). Contemporary issues in protein requirements and
Wu, J.; Fujioka, M.; Sugimoto, K.; Mu, G. & Ishimi, Y. (2004). Assessment of effectiveness of
Wynn, E.; Lanham-New, S.A.; Krieg, M.A.; Whittamore, D.R. & Burckhardt, P. (2008). Low
Xiao, G.; Wang, D.; Benson, M.D.; Karsenty, G. & Franceschi, R.T. (1998). Role of the alpha2-
Zhang, X.; Shu, X.O.; Li, H.; Yang, G.; Li, Q.; Gao, Y.T. & Zheng, W. (2005). Prospective cohort
Zou, Z.Y.; Lin, X.M.; Xu, X.R.; Xu, R.; Ma, L.; Li, Y. & Wang, M.F. (2009). Evaluation of milk
Zwart, S.R.; Davis-Street, J.E.; Paddon-Jones, D.; Ferrando, A.A.; Wolfe, R.R. & Smith, S.M.
women. *Arch Intern Med*, Vol. 165, No. 16, (Sep 12 2005), pp. 1890-1895 Zhou, S.; Turgeman, G.; Harris, S.E.; Leitman, D.C.; Komm, B.S.; Bodine, P.V. & Gazit, D.
young women. *Eur J Nutr*, Vol. 48, No. 5, (Aug 2009), pp. 301-306
weightlessness. *J Appl Physiol*, Vol. 99, No. 1, (Jul 2005), pp. 134-140
rats. *J Bone Miner Metab*, Vol. 22, No. 6, 2004), pp. 547-553
No. (Nov 2007), pp. 360-375
pp. 402-409
pp. 7-27
(Jul 2008), pp. 1349-1354
relation between dietary protein, calcium and bone health in women: results from the EPIC-Potsdam cohort. *Ann Nutr Metab*, Vol. 49, No. 5, (Sep-Oct 2005), pp. 312-318 Weitzmann, M.N. & Pacifici, R. (2007). T cells: unexpected players in the bone loss induced
by estrogen deficiency and in basal bone homeostasis. *Ann N Y Acad Sci*, Vol. 1116,
N.E. (2004). Dietary protein intake and risk of osteoporotic hip fracture in elderly
bicarbonate but not of sodium chloride or phosphate can be detected acutely in adult women and men. *Am J Clin Nutr*, Vol. 65, No. 5, (May 1997), pp. 1465-1472 Whiting, S.J.; Boyle, J.L.; Thompson, A.; Mirwald, R.L. & Faulkner, R.A. (2002). Dietary
protein, phosphorus and potassium are beneficial to bone mineral density in adult men consuming adequate dietary calcium. *J Am Coll Nutr*, Vol. 21, No. 5, (Oct 2002),
consumption for resistance trained athletes. *J Int Soc Sports Nutr*, Vol. 3, No. 2006),
oral administration of collagen peptide on bone metabolism in growing and mature
estimates of dietary acid load are positively associated with bone ultrasound in women older than 75 years of age with a lifetime fracture. *J Nutr*, Vol. 138, No. 7,
integrin in osteoblast-specific gene expression and activation of the Osf2 transcription factor. *J Biol Chem*, Vol. 273, No. 49, (Dec 4 1998), pp. 32988-32994 Yamamura, J.; Aoe, S.; Toba, Y.; Motouri, M.; Kawakami, H.; Kumegawa, M.; Itabashi, A. &
Takada, Y. (2002). Milk basic protein (MBP) increases radial bone mineral density in healthy adult women. *Biosci Biotechnol Biochem*, Vol. 66, No. 3, (Mar 2002), pp. 702-704
study of soy food consumption and risk of bone fracture among postmenopausal
(2003). Estrogens activate bone morphogenetic protein-2 gene transcription in mouse mesenchymal stem cells. *Mol Endocrinol*, Vol. 17, No. 1, (Jan 2003), pp. 56-66
basic protein supplementation on bone density and bone metabolism in Chinese
(2005). Amino acid supplementation alters bone metabolism during simulated
### *Edited by Abdellah El Maghraoui*
The World Health Organization (WHO) has established dual-energy x-ray absorptiometry (DXA) as the best densitometric technique for assessing bone mineral density (BMD) in postmenopausal women and has based the definitions of osteopenia and osteoporosis on its results. DXA enables accurate diagnosis of osteoporosis, estimation of fracture risk and monitoring of patients undergoing treatment. Additional features of DXA include measurement of BMD at multiple skeletal sites, vertebral fracture assessment and body composition assessment, including fat mass and lean soft tissue mass of the whole body and the segments. This book contains reviews and original studies about DXA and its different uses in clinical practice (diagnosis of osteoporosis, monitoring of BMD measurement) and in medical research in several situations (e.g. assessment of morphological asymmetry in athletes, estimation of resting energy expenditure, assessment of vertebral strength and vertebral fracture risk, or study of dry bones such as the ulna).
ISBN 978-953-307-877-9
ISBN 978-953-51-6742-6
Dual Energy X-Ray Absorptiometry
Dual Energy X-Ray
Absorptiometry
*Edited by Abdellah El Maghraoui*
Photo by Rost-9D / iStock | doab | 2025-04-07T04:13:03.531170 | 20-4-2021 17:18 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffbd80dd-28a6-46cd-af21-a432f1dad4e5",
"url": "https://mts.intechopen.com/storage/books/602/authors_book/authors_book.pdf",
"author": "",
"title": "Dual Energy X-Ray Absorptiometry",
"publisher": "IntechOpen",
"isbn": "9789533078779",
"section_idx": 2
} |
ffc124f2-9afe-4345-bf9e-9e0b37138f4f.0 | *Edited by Taner Yonar*
Increasing population and environmental pollution are the main stress on freshwater sources. On the other hand, freshwater needs of human being increase dramatically every day. From agriculture to industry and from household to recreation, we need freshwater. In the near future, saltwater and brackish water bodies may be the main source of freshwater for our planet. Desalination phenomena are now being implemented with increasing interest. The book on desalination provides a valuable scientific contribution on freshwater production from saltwater sources. In this book, necessary theoretical knowledge and experimental results of different desalination processes are presented.
ISBN 978-953-51-3363-6
Desalination
Photo by Supersmario / iStock
## Desalination
**DESALINATION**
| doab | 2025-04-07T04:13:03.558118 | 1-12-2023 17:13 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc124f2-9afe-4345-bf9e-9e0b37138f4f",
"url": "https://mts.intechopen.com/storage/books/5768/authors_book/authors_book.pdf",
"author": "",
"title": "Desalination",
"publisher": "IntechOpen",
"isbn": "9789535133643",
"section_idx": 0
} |
ffc124f2-9afe-4345-bf9e-9e0b37138f4f.1 | **DESALINATION**
#### **Desalination**
http://dx.doi.org/10.5772/65209 Edited by Taner Yonar
#### **Contributors**
Tianlong Deng, Long Li, Yafei Guo, Syed Zaidi, Farid Fadhillah, Abel Rouboa, Hunkyun Pak, Albert S. Kim, Hyeon-Ju Kim, André Pedral Sampaio De Sena, Iqbal Ahmed, Khaled S. Balkhair, Muhammad H Albeiruttye, Amer AhmedJamil Shaiban, Dionysis Assimacopoulos, George Arampatzis, Avraam Kartalidis, Peiwen Li, Ben Xu, Penghua Guo, Javier Miguel Ochando Pulido, OrevaOghene Aliku, Coskun Aydiner
#### **© The Editor(s) and the Author(s) 2017**
The moral rights of the and the author(s) have been asserted.
All rights to the book as a whole are reserved by INTECH. The book as a whole (compilation) cannot be reproduced, distributed or used for commercial or non-commercial purposes without INTECH's written permission. Enquiries concerning the use of the book should be directed to INTECH rights and permissions department (permissions@intechopen.com).
Violations are liable to prosecution under the governing Copyright Law.
Individual chapters of this publication are distributed under the terms of the Creative Commons Attribution 3.0 Unported License which permits commercial use, distribution and reproduction of the individual chapters, provided the original author(s) and source publication are appropriately acknowledged. If so indicated, certain images may not be included under the Creative Commons license. In such cases users will need to obtain permission from the license holder to reproduce the material. More details and guidelines concerning content reuse and adaptation can be foundat http://www.intechopen.com/copyright-policy.html.
#### **Notice**
Statements and opinions expressed in the chapters are these of the individual contributors and not necessarily those of the editors or publisher. No responsibility is accepted for the accuracy of information contained in the published chapters. The publisher assumes no responsibility for any damage or injury to persons or property arising out of the use of any materials, instructions, methods or ideas contained in the book.
First published in Croatia, 2017 by INTECH d.o.o. eBook (PDF) Published by IN TECH d.o.o. Place and year of publication of eBook (PDF): Rijeka, 2019. IntechOpen is the global imprint of IN TECH d.o.o. Printed in Croatia
Legal deposit, Croatia: National and University Library in Zagreb
Additional hard and PDF copies can be obtained from orders@intechopen.com
Desalination Edited by Taner Yonar p. cm. Print ISBN 978-953-51-3363-6 Online ISBN 978-953-51-3364-3 eBook (PDF) ISBN 978-953-51-4689-6
| doab | 2025-04-07T04:13:03.558255 | 1-12-2023 17:13 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc124f2-9afe-4345-bf9e-9e0b37138f4f",
"url": "https://mts.intechopen.com/storage/books/5768/authors_book/authors_book.pdf",
"author": "",
"title": "Desalination",
"publisher": "IntechOpen",
"isbn": "9789535133643",
"section_idx": 1
} |
ffc124f2-9afe-4345-bf9e-9e0b37138f4f.2 | We are IntechOpen, the world's leading publisher of Open Access books Built by scientists, for scientists
3,650+
Open access books available
114,000+
International authors and editors
118M+
Downloads
151 Countries delivered to Our authors are among the
Top 1% most cited scientists
12.2%
Contributors from top 500 universities
Selection of our books indexed in the Book Citation Index in Web of Science™ Core Collection (BKCI)
## Interested in publishing with us? Contact book.department@intechopen.com
Numbers displayed above are based on latest data collected. For more information visit www.intechopen.com
## **Meet the editor**
Dr. Taner Yonar is an associate professor at Uludag University, Engineering Faculty, Environmental Engineering Department. He received his BSc degree in Environmental Engineering Department, Uludag University, in 1996. Dr. Yonar received his MSc (1999) and PhD (2005) degrees in Environmental Technology from Uludag University, Institute of Sciences. He did his postdoctoral
research in the UK, at Newcastle University, Chemical Engineering and Advanced Materials Department (2011). Dr. Yonar teaches graduate- and undergraduate-level courses in Environmental Engineering on water and wastewater treatment and advanced treatment technologies and works on advanced oxidation, membrane processes, and electrochemical processes. He is the author of over 65 research papers.
Contents
**Preface VII**
Chapter 1 **Membrane Thermodynamics for Osmotic Phenomena 1**
Chapter 2 **Pulsating Flow Effects on Hydrodynamics in a Desalination**
Chapter 3 **Desalination: A Means of Increasing Irrigation Water Sources**
Chapter 4 **Application of Multilayer Thin Film Technology in Desalination**
Chapter 5 **Phase Equilibria and Phase Separation of the Aqueous Solution**
Chapter 6 **Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produced**
> Coskun Aydiner, Derya Y. Koseoglu Imer, Salim Oncel, Esra Can Dogan, Ali Oguzhan Narci, Serif Cakmak, Tugba Nur Yilmaz, Emin
**Membrane Technologies: The Case of Olive Mill Effluents 127**
Javier Miguel Ochando-Pulido and Antonio Martinez-Ferez
Armando A. Soares, João Silva, Eliseu Monteiro and Abel Rouboa
Albert S. Kim and Heyon‐Ju Kim
**Membrane Filled with Spacers 27**
**for Sustainable Crop Production 47**
Syed Javaid Zaidi and Farid Fadhillah
**System Containing Lithium Ions 81** Long Li, Yafei Guo and Tianlong Deng
Ender Celebi and Yasemin Melek Tilki
Chapter 7 **On the Purification of Agro-Industrial Wastewater by**
OrevaOghene Aliku
**Membrane 63**
**Drinking Water 99**
## Contents
#### **Preface XI**
Chapter 6 **Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produced Drinking Water 99** Coskun Aydiner, Derya Y. Koseoglu Imer, Salim Oncel, Esra Can Dogan, Ali Oguzhan Narci, Serif Cakmak, Tugba Nur Yilmaz, Emin Ender Celebi and Yasemin Melek Tilki
Chapter 7 **On the Purification of Agro-Industrial Wastewater by Membrane Technologies: The Case of Olive Mill Effluents 127** Javier Miguel Ochando-Pulido and Antonio Martinez-Ferez
Preface
The need for clean water is sharply increasing day to day. The main reasons for this need are increasing population and increasing pollution of clean water sources. Although the preven‐ tion of pollution is a top priority for the protection of clean water resources, saltwater re‐ sources have also gained importance in meeting the increased water need. On the other hand, salt is the main input of many industrial processes. For this reason, many industrial wastewaters are discharged with high salt content. High saline wastewater also negatively affects both the receiving environment and the soil. Because of these reasons, desalination processes have become very important for the saltwater bodies both to be a potential water
In this book, necessary theoretical knowledge and experimental results of different desalina‐ tion processes are presented. Prof. Kim and Dr. Kim discussed the thermodynamic ensem‐ bles and associated energy function using statistical variables. Similarly, Dr. Rouboa et al. investigated the effects of a pulsating flow with a profile of a heartbeat on the hydrodynam‐ ics of feed channels of a desalination membrane filled with spacers in zigzag arrangements and transverse to the flow. Dr. Aliku's chapter provides a valuable review on the effects of salinity on soil and crop growth and yield and some possible methods of desalinization of water and soil resources for optimum utilization in a crop production system. Dr. Zaidi re‐ viewed recent activities in the field of layer-by-layer assembly particularly used in prepar‐ ing membrane for desalination. Specifically, thermodynamics and phase equilibria of lithium containing saltwater are presented by Prof. Deng et al. Dr. Aydiner et al. studied six different commercially available membranes for the investigation of the effects of different operating conditions including cross flow rate, membrane type and pore size, solution tem‐ peratures and membrane trans-temperature differences on dissolved ion rejections, and per‐ meate flux of direct contact membrane distillation process. Dr.Ochando-Pulido and Dr. Martinez-Ferez presented the use of membrane and ion adsorption technologies for the pu‐ rification of olive oil mill wastewater containing high amount of organics, oil, and phenolic compounds. A novel water and solute full separation process using solar thermal energy is introduced by Dr. Xu, Dr. Li and Dr. Guo. Dr.Arampatzis, Dr. Kartalidis, and Dr. Assimaco‐ poulos presented water-energy nexus with emphasis in remote areas such as the islands of Syros and Patmos. Dr.Ahmed, Dr. Shaiban, Dr. Balkhair, and Dr. Albeiruttye reviewed the history and future outlook of ultrafiltration/microfiltration (UF/MF) membrane for desalina‐ tion water pretreatment. The chapter by Dr. Sena describes an interesting case study (NER‐ EU project) providing for the rational and intelligent use of contaminated and hypersaline water of the marine farms, from the production area of fish and seafood, in order to neutral‐ ize all pathogens and produce pure water, with energy cogeneration. Finally, Dr. Pak gives
the detailed results and cost of multieffect solar still for small-scale desalination.
source and to minimize the negative effects on the environment.
## Preface
Chapter 8 **Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum Impact to**
Chapter 9 **Responding to Water Challenges Through Desalination: Energy**
George Arampatzis, Avraam Kartalidis and Dionysis
Chapter 10 **Importance and Significance of UF/MF Membrane Systems in**
Chapter 11 **NEREU Project: Construction of a Plasma Reactor for Reform of Greenhouse Gases for Treatment of Wastewater of the**
Chapter 12 **Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination 247**
Iqbal Ahmed, Khaled S. Balkhair, Muhammad H. Albeiruttye and
**Environment 143**
**VI** Contents
**Considerations 165**
Assimacopoulos
Ben Xu, Peiwen Li and Penghua Guo
**Desalination Water Treatment 187**
Amer Ahmed Jamil Shaiban
**Marine Farms 225** André Pedral S. de Sena
Pak Hunkyun
The need for clean water is sharply increasing day to day. The main reasons for this need are increasing population and increasing pollution of clean water sources. Although the preven‐ tion of pollution is a top priority for the protection of clean water resources, saltwater re‐ sources have also gained importance in meeting the increased water need. On the other hand, salt is the main input of many industrial processes. For this reason, many industrial wastewaters are discharged with high salt content. High saline wastewater also negatively affects both the receiving environment and the soil. Because of these reasons, desalination processes have become very important for the saltwater bodies both to be a potential water source and to minimize the negative effects on the environment.
In this book, necessary theoretical knowledge and experimental results of different desalina‐ tion processes are presented. Prof. Kim and Dr. Kim discussed the thermodynamic ensem‐ bles and associated energy function using statistical variables. Similarly, Dr. Rouboa et al. investigated the effects of a pulsating flow with a profile of a heartbeat on the hydrodynam‐ ics of feed channels of a desalination membrane filled with spacers in zigzag arrangements and transverse to the flow. Dr. Aliku's chapter provides a valuable review on the effects of salinity on soil and crop growth and yield and some possible methods of desalinization of water and soil resources for optimum utilization in a crop production system. Dr. Zaidi re‐ viewed recent activities in the field of layer-by-layer assembly particularly used in prepar‐ ing membrane for desalination. Specifically, thermodynamics and phase equilibria of lithium containing saltwater are presented by Prof. Deng et al. Dr. Aydiner et al. studied six different commercially available membranes for the investigation of the effects of different operating conditions including cross flow rate, membrane type and pore size, solution tem‐ peratures and membrane trans-temperature differences on dissolved ion rejections, and per‐ meate flux of direct contact membrane distillation process. Dr.Ochando-Pulido and Dr. Martinez-Ferez presented the use of membrane and ion adsorption technologies for the pu‐ rification of olive oil mill wastewater containing high amount of organics, oil, and phenolic compounds. A novel water and solute full separation process using solar thermal energy is introduced by Dr. Xu, Dr. Li and Dr. Guo. Dr.Arampatzis, Dr. Kartalidis, and Dr. Assimaco‐ poulos presented water-energy nexus with emphasis in remote areas such as the islands of Syros and Patmos. Dr.Ahmed, Dr. Shaiban, Dr. Balkhair, and Dr. Albeiruttye reviewed the history and future outlook of ultrafiltration/microfiltration (UF/MF) membrane for desalina‐ tion water pretreatment. The chapter by Dr. Sena describes an interesting case study (NER‐ EU project) providing for the rational and intelligent use of contaminated and hypersaline water of the marine farms, from the production area of fish and seafood, in order to neutral‐ ize all pathogens and produce pure water, with energy cogeneration. Finally, Dr. Pak gives the detailed results and cost of multieffect solar still for small-scale desalination.
I would like to give my special thanks to Ms. Romina Rovan for enabling me to publish this book. I want to thank my wife Gonca Yonar and my children Burak and Beril for their sup‐ port throughout my life.
> **Dr. Taner Yonar** Associate Professor Uludag University Engineering Faculty, Environmental Engineering Department Bursa, Turkey
**Chapter 1**
**Membrane Thermodynamics for Osmotic Phenomena**
In this chapter, we briefly review the thermodynamic ensembles and associated energy functions using the seven thermodynamic variables. The energy E, the entropy S, and the system volume V are used to derive the temperature T and pressure P. The chemical potential μ is derived as the change of the system energy with respect to the number of matters N in the isobaric-isothermal environment. A dilute solution is defined as a homogeneous mixture of solvent and inert solutes, where the total number and volume of solutes are much smaller than those of the solvent. Gibbs free energy of the dilute solution is used to rigorously derive the osmotic pressure by equilibrating chemical potentials of solutes and solvent. Nonequilibrium of the filtration systems is reviewed by introducing the irreversible thermodynamic model with Onsager's reciprocal theorem. Direct applications of the irreversible thermodynamic model are currently limited due to the absence of the exact nonequilibrium statistical mechanics. We hope this chapter, containing a review of statistical mechanics, related to membrane separations and osmosis phenomena, helps researchers and especially graduate students, who seek an in-depth understanding of membrane separation from the theoretical statistical phys-
Keywords: membrane thermodynamics, statistical mechanics, thermodynamic ensemble, Gibbs energy function, chemical potential, weak solution, osmotic pressure, Fick's
A membrane is a selective barrier between two phases, i.e., a thin layer of material that separates solute and solvent materials when a driving force is applied across it. On membrane surfaces, flows of different thermodynamic phases are introduced and maintained quasisteady with respect to time. These separation processes require driving forces mainly for
> © 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Albert S. Kim and Heyon‐Ju Kim
http://dx.doi.org/10.5772/intechopen.68406
Abstract
1. Introduction
Additional information is available at the end of the chapter
ics as applied to chemical and environmental engineering.
law, solution-diffusion model, thermodynamic irreversible model
## **Membrane Thermodynamics for Osmotic Phenomena**
Albert S. Kim and Heyon‐Ju Kim
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.68406
#### Abstract
I would like to give my special thanks to Ms. Romina Rovan for enabling me to publish this book. I want to thank my wife Gonca Yonar and my children Burak and Beril for their sup‐
Engineering Faculty, Environmental Engineering Department
**Dr. Taner Yonar** Associate Professor Uludag University
Bursa, Turkey
port throughout my life.
VIII Preface
In this chapter, we briefly review the thermodynamic ensembles and associated energy functions using the seven thermodynamic variables. The energy E, the entropy S, and the system volume V are used to derive the temperature T and pressure P. The chemical potential μ is derived as the change of the system energy with respect to the number of matters N in the isobaric-isothermal environment. A dilute solution is defined as a homogeneous mixture of solvent and inert solutes, where the total number and volume of solutes are much smaller than those of the solvent. Gibbs free energy of the dilute solution is used to rigorously derive the osmotic pressure by equilibrating chemical potentials of solutes and solvent. Nonequilibrium of the filtration systems is reviewed by introducing the irreversible thermodynamic model with Onsager's reciprocal theorem. Direct applications of the irreversible thermodynamic model are currently limited due to the absence of the exact nonequilibrium statistical mechanics. We hope this chapter, containing a review of statistical mechanics, related to membrane separations and osmosis phenomena, helps researchers and especially graduate students, who seek an in-depth understanding of membrane separation from the theoretical statistical physics as applied to chemical and environmental engineering.
Keywords: membrane thermodynamics, statistical mechanics, thermodynamic ensemble, Gibbs energy function, chemical potential, weak solution, osmotic pressure, Fick's law, solution-diffusion model, thermodynamic irreversible model
## 1. Introduction
A membrane is a selective barrier between two phases, i.e., a thin layer of material that separates solute and solvent materials when a driving force is applied across it. On membrane surfaces, flows of different thermodynamic phases are introduced and maintained quasisteady with respect to time. These separation processes require driving forces mainly for
mass transfer as gradients of physical quantities associated with the thermodynamic, flowing phases.
The driving forces for membrane separations described above include gradients of the hydraulic pressure, solute concentration, solution temperature, and external electromagnetic field. In statistical mechanics, there are seven primary variables used to explain macroscopic thermodynamic phenomena. An ensemble is made using a set of three selected variables, and a specific energy function of the ensemble is described in terms of the three independent variables. As the membrane separations are coupled phenomena of momentum, mass, and heat transfer, a holistic understanding of statistical mechanics can significantly enhance design, analysis, and optimization of the membrane processes. In this chapter, we explain ensembles and energy functions in statistical mechanics, represent the osmotic pressure using Gibbs energy function of a weak (dilute) solution, and apply statistical laws to explain the separation
Statistical mechanics is the microscopic version of thermodynamics [31]. Macroscopic quantities, dealt within thermodynamics, can be fundamentally obtained at the microscopic level in statistical mechanics. There are seven variables in thermodynamics, which are energy E, entropy S, temperature T, pressure P, volume V, number of molecules N, and chemical potential μ. An ensemble is defined as a set, in which three independent variables are used to define a specific form of an energy and the other four variables are represented as functions of the three master variables. For example, the elementary microcanonical ensemble has P, V, S,
Consider two boxes in contact containing a certain number of particles in equilibrium, forming a closed system. Then, entropy S of the total system has its maximum value for a given system
Since the energy is an additive scalar, the total energy of the entire system is the sum of the
The total entropy can be similarly expressed, knowing that the entropy is a function of the
Since the entropy is already maximized in the equilibrium state, it is independent of the energy
S ¼ Smax ð1Þ
Membrane Thermodynamics for Osmotic Phenomena http://dx.doi.org/10.5772/intechopen.68406 3
E ¼ E<sup>1</sup> þ E<sup>2</sup> ð2Þ
Smax ¼ S1ðE1Þ þ S2ðE2Þ ð3Þ
phenomena using a solution-diffusion model [30].
2.1. Primary macroscopic quantities
2.1.1. Temperature
energy, E, i.e.,
energies:
energy:
variation, i.e.,
2. Thermodynamics to statistical mechanics
and μ, represented as functions of three master variables of N, V, and E.
Separation implies collecting masses of the same particles/molecules in specific spatial locations, which is strongly correlated to diffusion phenomena. Nature tends to move from a higher energy to a lower energy state, or equivalently highly ordered to randomly disordered phases. The diffusion of solutes in a free (solvent) medium is a spontaneous tendency, which must be well understood to analyze the separation phenomena. On the other hand, if the solutes are spatially confined by permeable interfaces, through which only solvent molecules can pass, the solvents try to move from their high- to low-concentration sides. This phenomena is called osmosis, which is equivalent to the solvent moving from low- to high-concentration regions of solutes. Note that in the solute diffusion and the solvent osmosis processes, mass transfer phenomena are from high- to low-concentration regions of the transferring masses. In this light, diffusion and osmosis can be treated equivalently as energy-minimizing and entropy-increasing phenomena of solutes and solvents, respectively.
Most pressure-driven membrane separations aim to produce clean water (solvent) from concentrated solutions. These include ultrafiltration (UF) and microfiltration (MF) for particulate removal and reverse osmosis (RO) and nanofiltration (NF) for ion removal. UF uses a finely porous membrane, which is usually antisymmetric, having a mean pore diameter between 1 and 100 nm. UF aims to separate water and microsolutes from macromolecules and colloids [1–3]. MF uses porous membranes to separate suspended particles with diameters between 0.1 and 10 µm [1, 3, 4]. MF's filtration capacity is therefore between UF and conventional filtration methods. Based on the particle size, dominant diffusion mechanisms of these particulate matters include Brownian diffusion [5–9] and shear-induced diffusion [10–12]. Ballistic motion of non-Brownian particles (usually bigger than 10 µm) in MF can be treated as dynamics of inelastic granules. RO is a desalting process for water production using nonporous membranes that are permeable to water but essentially impermeable to salt. A pressurized feed stream containing dissolved salts contacts the feed-side of the RO membrane, and salt-depleted water is withdrawn as a low-pressure permeate stream [13–16]. NF membranes have lower rejection ratio, i.e., 20–80% of sodium chloride, than that of RO, typically greater than 98–99%. NF resides therefore between UF and RO in terms of salt rejection capability. High hydraulic pressure is an essential component for RO and NF to overcome the osmotic pressure of seawater and brackish water, respectively [17–21].
The concentration (equivalently, osmotic pressure) gradient can be used, however, as a driving force for forward osmosis (FO) and pressure-retarded osmosis (PRO) processes. FO extracts a solvent from the low- to high-concentration sides of the solutes in order to equilibrate the concentrations [22–25]. PRO utilizes the extra gains of hydraulic pressure due to the amount of the transferred solvent for power generation [26–29]. Both pressure-driven and osmosis-driven processes aim to achieve a high flux, i.e., a large amount of water produced per unit time per unit membrane surface area. In order to achieve a steady high flux, increases and decreases in the osmotic pressure gradient need to be prevented in the pressure- and osmosis-driven membrane processes, respectively. Therefore, maintaining a stable osmotic pressure is a primary issue in both types of processes.
The driving forces for membrane separations described above include gradients of the hydraulic pressure, solute concentration, solution temperature, and external electromagnetic field. In statistical mechanics, there are seven primary variables used to explain macroscopic thermodynamic phenomena. An ensemble is made using a set of three selected variables, and a specific energy function of the ensemble is described in terms of the three independent variables. As the membrane separations are coupled phenomena of momentum, mass, and heat transfer, a holistic understanding of statistical mechanics can significantly enhance design, analysis, and optimization of the membrane processes. In this chapter, we explain ensembles and energy functions in statistical mechanics, represent the osmotic pressure using Gibbs energy function of a weak (dilute) solution, and apply statistical laws to explain the separation phenomena using a solution-diffusion model [30].
## 2. Thermodynamics to statistical mechanics
Statistical mechanics is the microscopic version of thermodynamics [31]. Macroscopic quantities, dealt within thermodynamics, can be fundamentally obtained at the microscopic level in statistical mechanics. There are seven variables in thermodynamics, which are energy E, entropy S, temperature T, pressure P, volume V, number of molecules N, and chemical potential μ. An ensemble is defined as a set, in which three independent variables are used to define a specific form of an energy and the other four variables are represented as functions of the three master variables. For example, the elementary microcanonical ensemble has P, V, S, and μ, represented as functions of three master variables of N, V, and E.
#### 2.1. Primary macroscopic quantities
#### 2.1.1. Temperature
mass transfer as gradients of physical quantities associated with the thermodynamic, flowing
Separation implies collecting masses of the same particles/molecules in specific spatial locations, which is strongly correlated to diffusion phenomena. Nature tends to move from a higher energy to a lower energy state, or equivalently highly ordered to randomly disordered phases. The diffusion of solutes in a free (solvent) medium is a spontaneous tendency, which must be well understood to analyze the separation phenomena. On the other hand, if the solutes are spatially confined by permeable interfaces, through which only solvent molecules can pass, the solvents try to move from their high- to low-concentration sides. This phenomena is called osmosis, which is equivalent to the solvent moving from low- to high-concentration regions of solutes. Note that in the solute diffusion and the solvent osmosis processes, mass transfer phenomena are from high- to low-concentration regions of the transferring masses. In this light, diffusion and osmosis can be treated equivalently as energy-minimizing and
Most pressure-driven membrane separations aim to produce clean water (solvent) from concentrated solutions. These include ultrafiltration (UF) and microfiltration (MF) for particulate removal and reverse osmosis (RO) and nanofiltration (NF) for ion removal. UF uses a finely porous membrane, which is usually antisymmetric, having a mean pore diameter between 1 and 100 nm. UF aims to separate water and microsolutes from macromolecules and colloids [1–3]. MF uses porous membranes to separate suspended particles with diameters between 0.1 and 10 µm [1, 3, 4]. MF's filtration capacity is therefore between UF and conventional filtration methods. Based on the particle size, dominant diffusion mechanisms of these particulate matters include Brownian diffusion [5–9] and shear-induced diffusion [10–12]. Ballistic motion of non-Brownian particles (usually bigger than 10 µm) in MF can be treated as dynamics of inelastic granules. RO is a desalting process for water production using nonporous membranes that are permeable to water but essentially impermeable to salt. A pressurized feed stream containing dissolved salts contacts the feed-side of the RO membrane, and salt-depleted water is withdrawn as a low-pressure permeate stream [13–16]. NF membranes have lower rejection ratio, i.e., 20–80% of sodium chloride, than that of RO, typically greater than 98–99%. NF resides therefore between UF and RO in terms of salt rejection capability. High hydraulic pressure is an essential component for RO and NF to overcome the osmotic pressure of
The concentration (equivalently, osmotic pressure) gradient can be used, however, as a driving force for forward osmosis (FO) and pressure-retarded osmosis (PRO) processes. FO extracts a solvent from the low- to high-concentration sides of the solutes in order to equilibrate the concentrations [22–25]. PRO utilizes the extra gains of hydraulic pressure due to the amount of the transferred solvent for power generation [26–29]. Both pressure-driven and osmosis-driven processes aim to achieve a high flux, i.e., a large amount of water produced per unit time per unit membrane surface area. In order to achieve a steady high flux, increases and decreases in the osmotic pressure gradient need to be prevented in the pressure- and osmosis-driven membrane processes, respectively. Therefore, maintaining a stable osmotic pressure is a pri-
entropy-increasing phenomena of solutes and solvents, respectively.
seawater and brackish water, respectively [17–21].
mary issue in both types of processes.
phases.
2 Desalination
Consider two boxes in contact containing a certain number of particles in equilibrium, forming a closed system. Then, entropy S of the total system has its maximum value for a given system energy, E, i.e.,
$$\mathcal{S} = \mathcal{S}\_{\text{max}} \tag{1}$$
Since the energy is an additive scalar, the total energy of the entire system is the sum of the energies:
$$E = E\_1 + E\_2 \tag{2}$$
The total entropy can be similarly expressed, knowing that the entropy is a function of the energy:
$$\mathcal{S}\_{\text{max}} = \mathcal{S}\_1(E\_1) + \mathcal{S}\_2(E\_2) \tag{3}$$
Since the entropy is already maximized in the equilibrium state, it is independent of the energy variation, i.e.,
4 Desalination
$$\frac{\text{dS}}{\text{dE}\_1} = \frac{\text{dS}\_1}{\text{dE}\_1} + \frac{\text{dS}\_2}{\text{dE}\_2} \frac{\text{dE}\_2}{\text{dE}\_1} = \frac{\text{dS}\_1}{\text{dE}\_1} - \frac{\text{dS}\_2}{\text{dE}\_2} = 0 \tag{4}$$
hence, we obtain
$$\frac{\text{d}S\_1}{\text{d}E\_1} = \frac{\text{d}S\_2}{\text{d}E\_2} \tag{5}$$
The derivative of the entropy S with respect to its energy E is used to define temperature as follows:
$$\frac{\text{d}S}{\text{d}E} \equiv \frac{1}{T} \quad \left( \rightarrow \frac{1}{k\_B T} \right) \tag{6}$$
In the original definition, the magnitude of the temperature is too high so Boltzmann's constant kB is introduced as shown in the parenthesis of Eq. (6). Temperature T is now represented in terms of the Kelvin unit. Substitution of Eq. (6) into Eq. (5) for each box provides
$$T\_1 = T\_2 \tag{7}$$
where, without losing generality, ð Þ ∂E=∂V <sup>S</sup> can be interpreted as the isentropic (i.e., of constant entropy) energy density inside the body volume V. One can operate the dot product by the
!¼ Fn ¼ � <sup>∂</sup><sup>E</sup>
and dividing both sides of Eq. (10) by the area A gives the conceptual definition of the
<sup>P</sup> ¼ � <sup>∂</sup><sup>E</sup> ∂V
Here, pressure definition can be extended from the normal force per unit area to the energy density in magnitude. Because energy E is a scalar quantity, the direction of the force vector
In the previous section, we used three thermodynamic variables of energy E, entropy S, and
∂V
S
S
A ð10Þ
Membrane Thermodynamics for Osmotic Phenomena http://dx.doi.org/10.5772/intechopen.68406 5
ð11Þ
! on the left side of Eq. (9) to have
Figure 1. External forces applied to the surface of a body.
does not need to be considered in the pressure calculation.
volume V to generally define temperature T at a constant volume:
2.2. Ensembles and energy functions
2.2.1. Internal energy in microcanonical ensemble
F ! � n
normal vector n
pressure:
as a condition for the equilibrium. It is worth noting that the internal energy E and entropy S are the basic thermodynamic quantities, and the temperature is a derived variable proportional to the variation of E with respect to S (specifically, in the microcanonical ensemble).
#### 2.1.2. Pressure
In fluid mechanics, pressure is often defined as the ratio of applied force per unit surface area of an object [32]:
$$P = \frac{\langle F\_n \rangle}{\text{Area}} \tag{8}$$
where 〈Fn〉 is the mean normal component of the force vector F ! applied to the object's surface area. A conservative force can be represented as a negative gradient of the total energy E ¼ K þ U, as a sum of kinetic energy K and potential energy U. Suppose the applied force causes an infinitesimal change in the volume of the body from V to V þ δV as shown in Figure 1. Then, the compressed volume is equal to the surface area multiplied by the thickness variation, i.e., δV ¼ A δs, which is in general, A ¼n ! �∇V. Using the chain rule, one can represent the normal component of the applied force as a product of the energy density and the total surface area, which is
$$
\overrightarrow{F} = -\nabla E = -\left(\frac{\partial E}{\partial V}\right)\_S \nabla V \tag{9}
$$
Figure 1. External forces applied to the surface of a body.
dS dE<sup>1</sup>
hence, we obtain
4 Desalination
as follows:
2.1.2. Pressure
of an object [32]:
surface area, which is
<sup>¼</sup> <sup>d</sup>S<sup>1</sup> dE<sup>1</sup> þ dS<sup>2</sup> dE<sup>2</sup> dE<sup>2</sup> dE<sup>1</sup>
> <sup>¼</sup> <sup>d</sup>S<sup>2</sup> dE<sup>2</sup>
The derivative of the entropy S with respect to its energy E is used to define temperature
<sup>T</sup> ! <sup>1</sup>
In the original definition, the magnitude of the temperature is too high so Boltzmann's constant kB is introduced as shown in the parenthesis of Eq. (6). Temperature T is now represented in
as a condition for the equilibrium. It is worth noting that the internal energy E and entropy S are the basic thermodynamic quantities, and the temperature is a derived variable proportional to the variation of E with respect to S (specifically, in the microcanonical ensemble).
In fluid mechanics, pressure is often defined as the ratio of applied force per unit surface area
<sup>P</sup> <sup>¼</sup> 〈Fn〉
area. A conservative force can be represented as a negative gradient of the total energy E ¼ K þ U, as a sum of kinetic energy K and potential energy U. Suppose the applied force causes an infinitesimal change in the volume of the body from V to V þ δV as shown in Figure 1. Then, the compressed volume is equal to the surface area multiplied by the thickness
sent the normal component of the applied force as a product of the energy density and the total
∂V
S
¼ �∇<sup>E</sup> ¼ � <sup>∂</sup><sup>E</sup>
where 〈Fn〉 is the mean normal component of the force vector F
F !
variation, i.e., δV ¼ A δs, which is in general, A ¼n
kBT
T<sup>1</sup> ¼ T<sup>2</sup> ð7Þ
Area <sup>ð</sup>8<sup>Þ</sup>
! �∇V. Using the chain rule, one can repre-
∇V ð9Þ
applied to the object's surface
!
dS<sup>1</sup> dE<sup>1</sup>
dS <sup>d</sup><sup>E</sup> � <sup>1</sup>
terms of the Kelvin unit. Substitution of Eq. (6) into Eq. (5) for each box provides
<sup>¼</sup> <sup>d</sup>S<sup>1</sup> dE<sup>1</sup> � <sup>d</sup>S<sup>2</sup> dE<sup>2</sup>
¼ 0 ð4Þ
ð5Þ
ð6Þ
where, without losing generality, ð Þ ∂E=∂V <sup>S</sup> can be interpreted as the isentropic (i.e., of constant entropy) energy density inside the body volume V. One can operate the dot product by the normal vector n ! on the left side of Eq. (9) to have
$$
\overrightarrow{F} \cdot \overrightarrow{n} = F\_n = -\left(\frac{\partial E}{\partial V}\right)\_S A \tag{10}
$$
and dividing both sides of Eq. (10) by the area A gives the conceptual definition of the pressure:
$$P = -\left(\frac{\partial E}{\partial V}\right)\_S \tag{11}$$
Here, pressure definition can be extended from the normal force per unit area to the energy density in magnitude. Because energy E is a scalar quantity, the direction of the force vector does not need to be considered in the pressure calculation.
#### 2.2. Ensembles and energy functions
#### 2.2.1. Internal energy in microcanonical ensemble
In the previous section, we used three thermodynamic variables of energy E, entropy S, and volume V to generally define temperature T at a constant volume:
$$T = \left(\frac{\partial E}{\partial \mathcal{S}}\right)\_V \tag{12}$$
dE ¼ dð Þ� TS S dT � P dV þ μ<sup>k</sup> dNk ð18Þ
dA ¼ �S dT � P dV þ μ<sup>k</sup> dNk ð19Þ
dE ¼ T dS � dð Þþ PV V dP þ μ<sup>k</sup> dNk ð21Þ
dH ¼ T dS þ V dP þ μ<sup>k</sup> dNk ð22Þ
H Sð Þ¼ ;P;{Nk} E þ PV ð23Þ
� � � Nk <sup>d</sup>μ<sup>k</sup> <sup>ð</sup>24<sup>Þ</sup>
X
Membrane Thermodynamics for Osmotic Phenomena http://dx.doi.org/10.5772/intechopen.68406
A ¼ E � ST ð20Þ
<sup>k</sup> is omitted for simplic-
7
� ��
where the notation of the summation over the molecular species k,
We add dð Þ PV in the both sides of the above equation and obtain
2.2.4. Thermodynamic potential in grand canonical ensemble
where A is the Helmholtz free energy defined as
2.2.3. Enthalpy in isentropic-isobaric ensemble
energy function.
have
where
Nk dμ<sup>k</sup> to have
Subtracting d μkNk
ity. The total derivative, dðTSÞ, is subtracted from both sides of Eq. (18) to have
If a thermodynamic system is completely described using T, V, and {Nk} (for k ¼ 1; 2…), this ensemble is called canonical, and the Helmholtz free energy, AðT;V;{Nk}Þ, is the representative
Similar to how we derived the Helmholtz free energy, we start from the infinitesimal difference of the internal energy E of Eq. (16) using the mathematical identity of P dV ¼ dð Þ� PV V dP to
is defined as the enthalpy as a function of S, P, and {Nk}. Eq. (22) indicates that the enthalpy is independent of T unlike other energy functions (see the next sections for detailed discussion).
To have an ensemble that is independent of the number of particles, one can start from the infinitesimal change of Helmholtz free energy and use the identity of μ<sup>k</sup> dNk ¼ d μkNk
dA ¼ �S dT � P dV þ d μkNk
� � from each side of Eq. (24) gives
and pressure P at a constant entropy:
$$P = -\left(\frac{\partial E}{\partial V}\right)\_S \tag{13}$$
Because the derivative operand of both Eqs. (12) and (13) is the internal energy E, the total derivative of E can be written in terms of T and P:
$$\mathbf{d}E = \left(\frac{\partial E}{\partial \mathbf{S}}\right)\_V \mathbf{dS} + \left(\frac{\partial E}{\partial V}\right)\_S \mathbf{d}V = T\,\mathbf{dS} - P\,\mathbf{d}V\tag{14}$$
which indicates that E is an exact function of S and V, i.e., E ¼ E Sð Þ ;V .
If the system consists of different molecular species, i.e., k ¼ 1; 2;…;nK, where nK is the total number of species, then the total molecule number N is the sum of the number of molecules of all the species, i.e.,
$$N = \sum\_{k} N\_{k} = N\_{1} + N\_{2} + \dots + N\_{n\_{\text{K}}} \tag{15}$$
where, for example, N<sup>2</sup> is the total molecule number of species 2. Then, the infinitesimal change of E includes the effect of the particle exchange, using the chemical potential μk, as
$$\mathbf{d}E = T\,\mathbf{d}S - P\,\mathbf{d}V + \sum\_{k} \mu\_{k}\,\mathbf{d}N\_{k} \tag{16}$$
In a closed system, the molecule numbers of multiple species can change simultaneously, keeping the total molecule number invariant. If the two systems in contact are at an equilibrium and molecules in the boxes can be exchanged, then the change of energy as per the number of exchanged molecules is equivalent to the chemical potential of the species. From Eq. (16), we can represent an extended version E as an exact function of S, V, and Nk:
$$E = E(\mathbf{S}, V, \{\mathbf{N}\_k\}) \tag{17}$$
If a thermodynamic system is completely controlled by the three variables of N, V, and S, the system is said to be a microcanonical ensemble.
#### 2.2.2. Helmholtz free energy in canonical ensemble
Since the temperature is a more convenient variable to measure than the entropy S, one can use the mathematical identity of T dS ¼ dð Þ� TS S dT to rewrite Eq. (16) as
$$\mathbf{d}E = \mathbf{d}(TS) - \mathbf{S}\,\mathbf{d}T - P\,\mathbf{d}V + \mu\_k \,\mathbf{dN}\_k \tag{18}$$
where the notation of the summation over the molecular species k, X <sup>k</sup> is omitted for simplicity. The total derivative, dðTSÞ, is subtracted from both sides of Eq. (18) to have
$$\mathbf{dA} = -\mathbf{S}\,\mathbf{d}T - P\,\mathbf{d}V + \mu\_k \,\mathbf{dN}\_k \tag{19}$$
where A is the Helmholtz free energy defined as
$$A = E - ST\tag{20}$$
If a thermodynamic system is completely described using T, V, and {Nk} (for k ¼ 1; 2…), this ensemble is called canonical, and the Helmholtz free energy, AðT;V;{Nk}Þ, is the representative energy function.
#### 2.2.3. Enthalpy in isentropic-isobaric ensemble
Similar to how we derived the Helmholtz free energy, we start from the infinitesimal difference of the internal energy E of Eq. (16) using the mathematical identity of P dV ¼ dð Þ� PV V dP to have
$$\mathbf{d}E = T\,\mathbf{d}S - \,\mathbf{d}(PV) + V\,\mathbf{d}P + \mu\_k \,\mathbf{d}N\_k\tag{21}$$
We add dð Þ PV in the both sides of the above equation and obtain
$$\mathbf{d}H = T\,\mathbf{dS} + V\,\mathbf{d}P + \mu\_k\,\mathbf{dN}\_k\tag{22}$$
where
<sup>T</sup> <sup>¼</sup> <sup>∂</sup><sup>E</sup> ∂S � �
<sup>P</sup> ¼ � <sup>∂</sup><sup>E</sup> ∂V � �
Because the derivative operand of both Eqs. (12) and (13) is the internal energy E, the total
∂E ∂V � �
If the system consists of different molecular species, i.e., k ¼ 1; 2;…;nK, where nK is the total number of species, then the total molecule number N is the sum of the number of molecules of
where, for example, N<sup>2</sup> is the total molecule number of species 2. Then, the infinitesimal change of E includes the effect of the particle exchange, using the chemical potential μk, as
In a closed system, the molecule numbers of multiple species can change simultaneously, keeping the total molecule number invariant. If the two systems in contact are at an equilibrium and molecules in the boxes can be exchanged, then the change of energy as per the number of exchanged molecules is equivalent to the chemical potential of the species. From
If a thermodynamic system is completely controlled by the three variables of N, V, and S, the
Since the temperature is a more convenient variable to measure than the entropy S, one can use
the mathematical identity of T dS ¼ dð Þ� TS S dT to rewrite Eq. (16) as
k
<sup>d</sup><sup>E</sup> <sup>¼</sup> <sup>T</sup> <sup>d</sup><sup>S</sup> � <sup>P</sup> <sup>d</sup><sup>V</sup> <sup>þ</sup><sup>X</sup>
Eq. (16), we can represent an extended version E as an exact function of S, V, and Nk:
S
and pressure P at a constant entropy:
all the species, i.e.,
6 Desalination
derivative of E can be written in terms of T and P:
system is said to be a microcanonical ensemble.
2.2.2. Helmholtz free energy in canonical ensemble
<sup>d</sup><sup>E</sup> <sup>¼</sup> <sup>∂</sup><sup>E</sup> ∂S � �
V dS þ
which indicates that E is an exact function of S and V, i.e., E ¼ E Sð Þ ;V .
<sup>N</sup> <sup>¼</sup> <sup>X</sup> k
V
S
dV ¼ T dS � P dV ð14Þ
μ<sup>k</sup> dNk ð16Þ
Nk ¼ N<sup>1</sup> þ N<sup>2</sup> þ ⋯ þ NnK ð15Þ
E ¼ E Sð Þ ;V;{Nk} ð17Þ
ð12Þ
ð13Þ
$$H(S, P, \{\mathcal{N}\_k\}) = E + PV \tag{23}$$
is defined as the enthalpy as a function of S, P, and {Nk}. Eq. (22) indicates that the enthalpy is independent of T unlike other energy functions (see the next sections for detailed discussion).
#### 2.2.4. Thermodynamic potential in grand canonical ensemble
To have an ensemble that is independent of the number of particles, one can start from the infinitesimal change of Helmholtz free energy and use the identity of μ<sup>k</sup> dNk ¼ d μkNk � �� Nk dμ<sup>k</sup> to have
$$\mathbf{d}A = -S\,\mathbf{d}T - P\,\mathbf{d}V + \mathbf{d}\left(\mu\_k N\_k\right) - N\_k \,\mathbf{d}\mu\_k \tag{24}$$
Subtracting d μkNk � � from each side of Eq. (24) gives
$$\mathbf{d}\Phi = -\mathbf{S}\,\mathbf{d}T - P\,\mathbf{d}V + \mathbf{N}\_k\,\mathbf{d}\mu\_k \tag{25}$$
The thermodynamic potential is generally derived as <sup>Φ</sup> <sup>¼</sup> <sup>A</sup> �<sup>X</sup>
bles are summarized in Table 1.
2.3. Gibbs energy and anisothermal equilibrium
Table 1. Specific ensembles and associated energy functions.
thermodynamic variables change as follows:
2.3.1. Thermodynamics variables: extensive and intensive
• Additive (extensive): N ! 2N, V ! 2V, S ! 2S, and E ! 2E
and chemical potential remain invariant, and they are called nonadditive.
• Nonadditive (intensive): T ! T, P ! P, and μ<sup>k</sup> ! μ<sup>k</sup>
Gð¼ A þ PVÞ is homogeneous such as Eq. (32), Φ can be further simplified to
Ensemble Energy functions and relationships
Canonical (NVT) Helmholtz energy A Nð Þ¼ ;V;T E � TS
Isothermal-isobaric (NPT) Gibbs energy G Nð Þ¼ ;P;T A þ PV ¼ μN
Isentropic-isobaric (NPS) Enthalpy H Nð Þ¼ ;P;S E þ PV
Grand canonical (μVT) Thermodynamic potential Φ μ;V;<sup>T</sup> � � <sup>¼</sup> <sup>A</sup> �<sup>X</sup>
Microcanonical (NVS) Internal energy E Nð Þ ;V;S
transformation from the previous section. If and only if the Gibbs energy function
If the molecular interactions are strong, then Eq. (32) requires an extra coupling term proportional to NiNj, and Eq. (26) should be revisited as a general definition for Φ (see Section 1.3 for details). Dependences of the energy functions on thermodynamic variables in specific ensem-
Consider a thermodynamic system in equilibrium, shown in Figure 2. The system is made by adding two identical systems, which are now in contact with each other. In this case, the seven
As expected, the number of particles, volume, entropy, and energy are doubled by adding the two identical systems, and they are called additive. On the other hand, temperature, pressure,
The independence of the temperature to the system size can be understood using its basic definition of Eq. (12) as the change ratio of E to S as they are additive quantities. The pressure is defined in Eq. (13) as the negative ratio of changes of E to V. The chemical potential,
k
Membrane Thermodynamics for Osmotic Phenomena http://dx.doi.org/10.5772/intechopen.68406
Φ ¼ A � G ¼ �PV ð33Þ
dE ¼ TdS � PdV þ μdN
dA ¼ �SdT � PdV þ μdN
dΦ ¼ �SdT � PdV � Ndμ
dG ¼ �SdT þ VdP þ μdN
dH ¼ TdS þ VdP þ μdN
μkNk using the Legendre
9
k μkNk
where
$$\Phi(T, V, \{\mu\_k\}) = A - \mu\_k \mathcal{N}\_k \tag{26}$$
is defined as the thermodynamic potential, varying with respect to T, V, and μk. An ensemble described using μ, V, and T is called a grand canonical ensemble. The thermodynamic potential is further derived such that Φ ¼ �PV if the thermodynamic system is homogeneous.
#### 2.2.5. Gibbs energy in isothermal-isobaric ensemble
Finally, we replace P dV in the infinitesimal change of A in Eq. (19) by dð Þ� PV V dP to have
$$\mathbf{d}\mathbf{G} = -\mathbf{S}\,\mathbf{d}T + V\,\mathbf{d}P + \mu\_k \,\mathbf{d}N\_k\tag{27}$$
where
$$(G(T, P, \{N\_k\}) = A + PV = E - TS + PV \tag{28}$$
is defined as the Gibbs free energy varying with respect to T, P, and {Nk}. Now we assume that G is a homogeneous (i.e., linear) function of Nk such that G∝Nk. In this case, the chemical potential of species k is represented in terms of T and P only as
$$
\mu\_k = \left(\frac{\partial G}{\partial \mathbf{N}\_k}\right)\_{T,P} = \mu\_k(T,P) \tag{29}
$$
For the fixed number of particles, the infinitesimal change of the total chemical potential is
$$\mathbf{d}\mu = \frac{1}{N}\,\mathbf{d}G = -\overline{S}\,\mathrm{d}T + \overline{V}\,\mathrm{d}P\tag{30}$$
where S ¼ S=N and V ¼ V=N are the entropy and the volume per molecule, respectively, of the entire system. In practice, it is often convenient to use the entropy and energy per mole of molecules in engineering applications, but for basic study here we will keep using quantities divided by the number of molecules. For species k, we have the representation of the infinitesimal change in the chemical potential of species k:
$$\mathbf{d}\mu\_k = -\overline{\mathbf{S}}\_k \mathbf{d}T + \overline{V}\_k \mathbf{d}P \tag{31}$$
Keeping the homogeneity assumption, the Gibbs energy function is written as a sum of products of the chemical potentials and the particle numbers:
$$G = \sum\_{k} \left(\frac{\partial G}{\partial \mathbf{N}\_k}\right)\_{T,P} N\_k = \sum\_{k} \mu\_k(T,P) N\_k \tag{32}$$
The thermodynamic potential is generally derived as <sup>Φ</sup> <sup>¼</sup> <sup>A</sup> �<sup>X</sup> k μkNk using the Legendre transformation from the previous section. If and only if the Gibbs energy function Gð¼ A þ PVÞ is homogeneous such as Eq. (32), Φ can be further simplified to
$$
\Phi = A - G = -PV\tag{33}
$$
If the molecular interactions are strong, then Eq. (32) requires an extra coupling term proportional to NiNj, and Eq. (26) should be revisited as a general definition for Φ (see Section 1.3 for details). Dependences of the energy functions on thermodynamic variables in specific ensembles are summarized in Table 1.
Table 1. Specific ensembles and associated energy functions.
dΦ ¼ �S dT � P dV þ Nk dμ<sup>k</sup> ð25Þ
<sup>Φ</sup> <sup>T</sup>;V;{μk} � � <sup>¼</sup> <sup>A</sup> � <sup>μ</sup>kNk <sup>ð</sup>26<sup>Þ</sup>
dG ¼ �S dT þ V dP þ μ<sup>k</sup> dNk ð27Þ
G Tð Þ¼ ;P;{Nk} A þ PV ¼ E � TS þ PV ð28Þ
¼ μkð Þ T;P ð29Þ
<sup>N</sup> <sup>d</sup><sup>G</sup> ¼ �<sup>S</sup> <sup>d</sup><sup>T</sup> <sup>þ</sup> <sup>V</sup> <sup>d</sup><sup>P</sup> <sup>ð</sup>30<sup>Þ</sup>
dμ<sup>k</sup> ¼ �Sk dT þ Vk dP ð31Þ
μkð Þ T;P Nk ð32Þ
is defined as the thermodynamic potential, varying with respect to T, V, and μk. An ensemble described using μ, V, and T is called a grand canonical ensemble. The thermodynamic potential is further derived such that Φ ¼ �PV if the thermodynamic system is homogeneous.
Finally, we replace P dV in the infinitesimal change of A in Eq. (19) by dð Þ� PV V dP to have
is defined as the Gibbs free energy varying with respect to T, P, and {Nk}. Now we assume that G is a homogeneous (i.e., linear) function of Nk such that G∝Nk. In this case, the chemical
T;P
For the fixed number of particles, the infinitesimal change of the total chemical potential is
where S ¼ S=N and V ¼ V=N are the entropy and the volume per molecule, respectively, of the entire system. In practice, it is often convenient to use the entropy and energy per mole of molecules in engineering applications, but for basic study here we will keep using quantities divided by the number of molecules. For species k, we have the representation of the infinites-
Keeping the homogeneity assumption, the Gibbs energy function is written as a sum of
Nk <sup>¼</sup> <sup>X</sup> k
T;P
where
8 Desalination
where
2.2.5. Gibbs energy in isothermal-isobaric ensemble
potential of species k is represented in terms of T and P only as
imal change in the chemical potential of species k:
products of the chemical potentials and the particle numbers:
<sup>G</sup> <sup>¼</sup> <sup>X</sup> k
∂G ∂Nk � �
<sup>μ</sup><sup>k</sup> <sup>¼</sup> <sup>∂</sup><sup>G</sup> ∂Nk � �
<sup>d</sup><sup>μ</sup> <sup>¼</sup> <sup>1</sup>
#### 2.3. Gibbs energy and anisothermal equilibrium
#### 2.3.1. Thermodynamics variables: extensive and intensive
Consider a thermodynamic system in equilibrium, shown in Figure 2. The system is made by adding two identical systems, which are now in contact with each other. In this case, the seven thermodynamic variables change as follows:
As expected, the number of particles, volume, entropy, and energy are doubled by adding the two identical systems, and they are called additive. On the other hand, temperature, pressure, and chemical potential remain invariant, and they are called nonadditive.
The independence of the temperature to the system size can be understood using its basic definition of Eq. (12) as the change ratio of E to S as they are additive quantities. The pressure is defined in Eq. (13) as the negative ratio of changes of E to V. The chemical potential,
Enthalpy H Sð Þ ;P;f g Nk is often used to characterize mass transfer phenomena under an isobaric-isentropic environment between two different temperatures, allowing volume expansion or compression. H is mainly used to link two temperature-dependent quantities such as equilibrium constants of chemical reactions in the NPT ensemble because it does not vary
Consider two heterogeneous systems in equilibrium. This is similar to the case shown in Figure 2, but boxes 1 and 2 are not thermodynamically identical. In each box, the internal energy is fully represented using Ni, Vi, and Si of box i for i ¼ 1 and 2. Assume their volumes do not change so that dVi ¼ 0. We express the infinitesimal change of the entropy from
> � μi Ti
In equilibrium, the total entropy S ¼ S<sup>1</sup> þ S<sup>2</sup> must be already maximized, having a constant
and <sup>∂</sup>S<sup>2</sup> ∂N<sup>2</sup>
<sup>¼</sup> <sup>μ</sup>2ð Þ <sup>T</sup>2;P<sup>2</sup> T2
for an isothermal environment. Note that Eqs. (38) and (39) consist of only intensive thermodynamic quantities. The chemical potential can be readily derived using Eq. (29) if the Gibbs
∂S<sup>2</sup> ∂N<sup>2</sup> <sup>¼</sup> <sup>∂</sup>S<sup>1</sup> ∂N<sup>1</sup>
¼ � <sup>μ</sup><sup>2</sup> T2
dNi ð34Þ
Membrane Thermodynamics for Osmotic Phenomena http://dx.doi.org/10.5772/intechopen.68406
ð Þ¼� N � N<sup>1</sup> 1 ð35Þ
� <sup>∂</sup>S<sup>2</sup> ∂N<sup>2</sup>
μ1ðP1Þ ¼ μ2ðP2Þ ð39Þ
ð36Þ
11
ð37Þ
ð38Þ
<sup>d</sup>Si <sup>¼</sup> <sup>d</sup>Ei Ti
<sup>¼</sup> <sup>∂</sup> ∂N<sup>1</sup>
<sup>¼</sup> <sup>0</sup> <sup>¼</sup> <sup>∂</sup>S<sup>1</sup> ∂N<sup>1</sup> þ ∂N<sup>2</sup> ∂N<sup>1</sup>
As the internal energy of each box, Ei, is kept invariant in Eq. (34), we derive
¼ � <sup>μ</sup><sup>1</sup> T1
> μ1ð Þ T1;P<sup>1</sup> T1
If the total number of particles Nð Þ ¼ N<sup>1</sup> þ N<sup>2</sup> is constant, we simply derive
∂N<sup>2</sup> ∂N<sup>1</sup>
∂S ∂N<sup>1</sup>
Substitution of Eq. (37) into (36) gives
which is simplified, if T<sup>1</sup> ¼ T2, to
energy is known.
<sup>¼</sup> <sup>∂</sup>Smax ∂N<sup>1</sup>
> ∂S<sup>1</sup> ∂N<sup>1</sup>
with T.
Eq. (16) as
value Smax:
2.3.2. Anisothermal equilibrium
Figure 2. A closed system consisting of two identical boxes in a thermal equilibrium. The outer boundaries (gray) insulate mass and energy transfer from the environment, and each box has the same T, P, and N. If the central wall is removed, then the two identical systems are combined.
interpreted as the ratio of the internal energy change with respect to creation/disappearance of a molecule, must be independent of the number of molecules. Among the seven master variables in thermodynamics, additive quantities are E, S, N, and V, called extensive, and nonadditive ones are T, P, and μk, called intensive. Note that the intensive quantities are defined as ratios of extensive quantities.
In the previous sections, we reviewed the five standard ensembles with their energy functions derived from three independent variables as
Among these energy functions, E, A, Φ, and H depend on at least one extensive variable, S or V. Gibbs energy function is the only one that depends on two intensive variables, T and P. Although G basically varies with Nk, if the system is homogeneous, the chemical potential μkð Þ T;P is independent to the number of particles Nk. In many engineering applications dealing with mass transfer phenomena, temperature and pressure are often maintained as (pseudo-) constants. Molecules and particles translate spatially from one location to other, or are converted to another species (i.e., created or annihilated through physical and chemical reactions). In this light, the Gibbs energy G Tð Þ ;P;f g Nk is the most convenient representation of the system undergoing mass and/or heat transfer in the isobaric and isothermal environment. Enthalpy H Sð Þ ;P;f g Nk is often used to characterize mass transfer phenomena under an isobaric-isentropic environment between two different temperatures, allowing volume expansion or compression. H is mainly used to link two temperature-dependent quantities such as equilibrium constants of chemical reactions in the NPT ensemble because it does not vary with T.
#### 2.3.2. Anisothermal equilibrium
interpreted as the ratio of the internal energy change with respect to creation/disappearance of a molecule, must be independent of the number of molecules. Among the seven master variables in thermodynamics, additive quantities are E, S, N, and V, called extensive, and nonadditive ones are T, P, and μk, called intensive. Note that the intensive quantities are defined as
Figure 2. A closed system consisting of two identical boxes in a thermal equilibrium. The outer boundaries (gray) insulate mass and energy transfer from the environment, and each box has the same T, P, and N. If the central wall is removed,
In the previous sections, we reviewed the five standard ensembles with their energy functions
Among these energy functions, E, A, Φ, and H depend on at least one extensive variable, S or V. Gibbs energy function is the only one that depends on two intensive variables, T and P. Although G basically varies with Nk, if the system is homogeneous, the chemical potential μkð Þ T;P is independent to the number of particles Nk. In many engineering applications dealing with mass transfer phenomena, temperature and pressure are often maintained as (pseudo-) constants. Molecules and particles translate spatially from one location to other, or are converted to another species (i.e., created or annihilated through physical and chemical reactions). In this light, the Gibbs energy G Tð Þ ;P;f g Nk is the most convenient representation of the system undergoing mass and/or heat transfer in the isobaric and isothermal environment.
ratios of extensive quantities.
then the two identical systems are combined.
10 Desalination
• Internal energy E Sð Þ ;V;f g Nk
• Enthalpy H Sð Þ ;P;f g Nk
• Gibbs energy G Tð Þ ;P;f g Nk
• Helmholtz energy A Tð Þ ;V;f g Nk
• Thermodynamic potential Φ T;V; μ<sup>k</sup>
derived from three independent variables as
Consider two heterogeneous systems in equilibrium. This is similar to the case shown in Figure 2, but boxes 1 and 2 are not thermodynamically identical. In each box, the internal energy is fully represented using Ni, Vi, and Si of box i for i ¼ 1 and 2. Assume their volumes do not change so that dVi ¼ 0. We express the infinitesimal change of the entropy from Eq. (16) as
$$\mathbf{dS}\_{i} = \frac{\mathbf{d}E\_{i}}{T\_{i}} - \frac{\mu\_{i}}{T\_{i}} \,\mathbf{dN}\_{i} \tag{34}$$
If the total number of particles Nð Þ ¼ N<sup>1</sup> þ N<sup>2</sup> is constant, we simply derive
$$\frac{\partial N\_2}{\partial N\_1} = \frac{\partial}{\partial N\_1}(N - N\_1) = -1\tag{35}$$
In equilibrium, the total entropy S ¼ S<sup>1</sup> þ S<sup>2</sup> must be already maximized, having a constant value Smax:
$$\frac{\partial S}{\partial N\_1} = \frac{\partial S\_{\text{max}}}{\partial N\_1} = 0 = \frac{\partial S\_1}{\partial N\_1} + \frac{\partial N\_2}{\partial N\_1} \frac{\partial S\_2}{\partial N\_2} = \frac{\partial S\_1}{\partial N\_1} - \frac{\partial S\_2}{\partial N\_2} \tag{36}$$
As the internal energy of each box, Ei, is kept invariant in Eq. (34), we derive
$$\frac{\partial \mathbf{S}\_1}{\partial \mathbf{N}\_1} = -\frac{\mu\_1}{T\_1} \quad \text{and} \quad \frac{\partial \mathbf{S}\_2}{\partial \mathbf{N}\_2} = -\frac{\mu\_2}{T\_2} \tag{37}$$
Substitution of Eq. (37) into (36) gives
$$\frac{\mu\_1(T\_1, P\_1)}{T\_1} = \frac{\mu\_2(T\_2, P\_2)}{T\_2} \tag{38}$$
which is simplified, if T<sup>1</sup> ¼ T2, to
$$
\mu\_1(P\_1) = \mu\_2(P\_2) \tag{39}
$$
for an isothermal environment. Note that Eqs. (38) and (39) consist of only intensive thermodynamic quantities. The chemical potential can be readily derived using Eq. (29) if the Gibbs energy is known.
### 3. Dilute solution
#### 3.1. Chemical potentials
Now we consider a dilute (or weak) solution, in which the number of dissolved molecules in the solvent is much less than that of the solvent molecules. Without losing generality for environmental engineering purposes, we set water as the solvent. Gibbs free energy of the weak solution of a single solute species is [31]
$$\mathcal{G} = \mathcal{N}\mu\_0(P, T) + nk\_\mathcal{B}T\ln\left(\frac{n}{\mathcal{N}e}\right) + n\psi(P, T) \tag{40}$$
3.2.1. Using solvent chemical potential
be equal to each other from Eq. (39):
pressure difference ΔP is calculated as
and finally denoted as
using
1
In this thermodynamic environment, the chemical potentials of water in the two boxes should
We assume that the pressure difference is small enough to use the weak solution approximation without drastic thermodynamic changes but large enough to maintain the balance
between the two boxes. Then, we expand μ0ð Þ P2;T around P<sup>1</sup> using Taylor's series
μ0ð Þ P2;T ≃μ0ð Þþ P1;T
T
<sup>Δ</sup><sup>P</sup> <sup>¼</sup> kBT <sup>N</sup>Δ<sup>x</sup>
where ΔP ¼ P<sup>2</sup> � P<sup>1</sup> and Δx ¼ x<sup>2</sup> � x<sup>1</sup> are differences of pressure and solute mole fraction, respectively, between box 1 and 2. Using Eq. (31), the fundamental representation of the infinitesimal chemical potential, we replaced ∂μ0=∂P with the volume per solvent, V=N. Then, the
where ni and Ci ð Þ ¼ ni=NAV are the (absolute) number and the mole concentration of solutes in box i for i ¼ 1 and 2, NA is Avogadro's number, and R is the universal gas constant. Eq. (49) is called the van't Hoff equation,<sup>1</sup> which resembles the ideal gas law [33]. If the solution contains
Jacobus H. van't Hoff received the first Nobel Prize in Chemistry in 1901 for the discovery of osmotic pressure in
multiple species of solutes, Eq. (49) can be easily extended to
solutions. https://www.nobelprize.org/nobel\_prizes/chemistry/laureates/1901/
<sup>V</sup> <sup>¼</sup> <sup>R</sup><sup>T</sup> <sup>Δ</sup><sup>n</sup>
at a fixed temperature T. We substitute Eq. (46) into Eq. (45) to obtain
∂μ<sup>0</sup> ∂P μ<sup>w</sup>;<sup>1</sup> ¼ μ<sup>w</sup>;<sup>2</sup> ð44Þ
Membrane Thermodynamics for Osmotic Phenomena http://dx.doi.org/10.5772/intechopen.68406 13
ΔP ¼ ð Þ x<sup>2</sup> � x<sup>1</sup> kBT ð47Þ
Δπ ¼ RTΔC ð49Þ
Δn ¼ NΔx ¼ Nx<sup>2</sup> � Nx<sup>1</sup> ¼ n<sup>2</sup> � n<sup>1</sup> ð50Þ
ΔC ¼ Δn=NA ¼ C<sup>2</sup> � C<sup>1</sup> ð51Þ
ΔP ð46Þ
NAV <sup>ð</sup>48<sup>Þ</sup>
<sup>μ</sup><sup>0</sup> <sup>P</sup><sup>1</sup>;<sup>T</sup> � <sup>x</sup>1kBT <sup>¼</sup> <sup>μ</sup>0ð Þ� <sup>P</sup>2;<sup>T</sup> <sup>x</sup>2kBT <sup>ð</sup>45<sup>Þ</sup>
∂μ<sup>0</sup> ∂P
T
where N and n are the numbers of solvent and solute molecules, respectively, μ<sup>0</sup> is the chemical potential of the pure solvent, and ψð Þ P;T is an arbitrary function for the chemical potential of the pure solute. Euler's number e ¼ 2:71828218… in the denominator of the logarithmic function on the right-hand side of Eq. (40) stems from Starling's formula, used for entropy calculations: ln n! ≃ n ln n � n ¼ n lnð Þ n=e
If the weak solution contains multiple species of solutes, then the Gibbs energy function is generalized as
$$G = N\mu\_0(P, T) + k\_\mathcal{B}T \sum\_i \eta\_i \ln\left(\frac{n\_i}{\mathcal{N}e}\right) + \sum\_i n\_i \psi\_i(P, T) \tag{41}$$
One can easily calculate the chemical potentials for the solvent μ<sup>w</sup> and solute μ<sup>s</sup> as partial derivatives of G in Eq. (41) with respect to N and n, respectively. The former and latter are
$$
\mu\_w = \frac{\partial G}{\partial \mathbf{N}} = \mu\_0(P, T) - k\_B T \mathbf{x} \tag{42}
$$
and
$$
\mu\_s = \frac{\partial G}{\partial n} = \psi(P, T) + k\_B T \text{ lnx} \tag{43}
$$
respectively, where x ð Þ ¼ n=N is the number (or mole) fraction of solute molecules to solvent molecules. In a dilute solution, x ≪ 1.
#### 3.2. Osmotic pressure
Let's consider an isothermal system consisting of two boxes (1 and 2) of the same size in contact. Box 1 (and 2) has the solute mole fraction x<sup>1</sup> (and x2) and pressure P<sup>1</sup> (and P2). Since the total system is in isothermal equilibrium, the two boxes have the same temperature: T<sup>1</sup> ¼ T<sup>2</sup> ¼ T.
#### 3.2.1. Using solvent chemical potential
3. Dilute solution
12 Desalination
generalized as
and
3.1. Chemical potentials
weak solution of a single solute species is [31]
entropy calculations: ln n! ≃ n ln n � n ¼ n lnð Þ n=e
molecules. In a dilute solution, x ≪ 1.
3.2. Osmotic pressure
T<sup>1</sup> ¼ T<sup>2</sup> ¼ T.
G ¼ Nμ0ð Þþ P;T kBT
<sup>μ</sup><sup>w</sup> <sup>¼</sup> <sup>∂</sup><sup>G</sup>
<sup>μ</sup><sup>s</sup> <sup>¼</sup> <sup>∂</sup><sup>G</sup>
Now we consider a dilute (or weak) solution, in which the number of dissolved molecules in the solvent is much less than that of the solvent molecules. Without losing generality for environmental engineering purposes, we set water as the solvent. Gibbs free energy of the
where N and n are the numbers of solvent and solute molecules, respectively, μ<sup>0</sup> is the chemical potential of the pure solvent, and ψð Þ P;T is an arbitrary function for the chemical potential of the pure solute. Euler's number e ¼ 2:71828218… in the denominator of the logarithmic function on the right-hand side of Eq. (40) stems from Starling's formula, used for
If the weak solution contains multiple species of solutes, then the Gibbs energy function is
One can easily calculate the chemical potentials for the solvent μ<sup>w</sup> and solute μ<sup>s</sup> as partial derivatives of G in Eq. (41) with respect to N and n, respectively. The former and latter are
respectively, where x ð Þ ¼ n=N is the number (or mole) fraction of solute molecules to solvent
Let's consider an isothermal system consisting of two boxes (1 and 2) of the same size in contact. Box 1 (and 2) has the solute mole fraction x<sup>1</sup> (and x2) and pressure P<sup>1</sup> (and P2). Since the total system is in isothermal equilibrium, the two boxes have the same temperature:
ni ln ni Ne � �
þ<sup>X</sup> i niψ<sup>i</sup>
<sup>∂</sup><sup>N</sup> <sup>¼</sup> <sup>μ</sup>0ð Þ� <sup>P</sup>;<sup>T</sup> kBTx <sup>ð</sup>42<sup>Þ</sup>
<sup>∂</sup><sup>n</sup> <sup>¼</sup> <sup>ψ</sup>ð Þþ <sup>P</sup>;<sup>T</sup> kBT ln<sup>x</sup> <sup>ð</sup>43<sup>Þ</sup>
X i
Ne � �
þ nψð Þ P;T ð40Þ
ð Þ P;T ð41Þ
<sup>G</sup> <sup>¼</sup> <sup>N</sup>μ0ð Þþ <sup>P</sup>;<sup>T</sup> nkBT ln <sup>n</sup>
In this thermodynamic environment, the chemical potentials of water in the two boxes should be equal to each other from Eq. (39):
$$
\mu\_{w,1} = \mu\_{w,2} \tag{44}
$$
$$
\mu\_0(P\_1, T) - \mathbf{x}\_1 k\_B T = \mu\_0(P\_2, T) - \mathbf{x}\_2 k\_B T \tag{45}
$$
We assume that the pressure difference is small enough to use the weak solution approximation without drastic thermodynamic changes but large enough to maintain the balance between the two boxes. Then, we expand μ0ð Þ P2;T around P<sup>1</sup> using Taylor's series
$$
\mu\_0(P\_2, T) \simeq \mu\_0(P\_1, T) + \left(\frac{\partial \mu\_0}{\partial P}\right)\_T \Delta P \tag{46}
$$
at a fixed temperature T. We substitute Eq. (46) into Eq. (45) to obtain
$$
\left(\frac{\partial \mu\_0}{\partial P}\right)\_T \Delta P = (\mathbf{x}\_2 - \mathbf{x}\_1)k\_B T \tag{47}
$$
where ΔP ¼ P<sup>2</sup> � P<sup>1</sup> and Δx ¼ x<sup>2</sup> � x<sup>1</sup> are differences of pressure and solute mole fraction, respectively, between box 1 and 2. Using Eq. (31), the fundamental representation of the infinitesimal chemical potential, we replaced ∂μ0=∂P with the volume per solvent, V=N. Then, the pressure difference ΔP is calculated as
$$
\Delta P = k\_{\text{B}} T \frac{N \Delta x}{V} = \mathcal{R} T \frac{\Delta n}{N\_A V} \tag{48}
$$
and finally denoted as
$$
\Delta \pi = \mathcal{R} T \Delta \mathcal{C} \tag{49}
$$
using
$$
\Delta \mathbf{n} = \mathbf{N} \Delta \mathbf{x} = \mathbf{N} \mathbf{x}\_2 - \mathbf{N} \mathbf{x}\_1 = n\_2 - n\_1 \tag{50}
$$
$$
\Delta \mathbf{C} = \Delta \mathbf{n} / \mathbf{N}\_A = \mathbf{C}\_2 - \mathbf{C}\_1 \tag{51}
$$
where ni and Ci ð Þ ¼ ni=NAV are the (absolute) number and the mole concentration of solutes in box i for i ¼ 1 and 2, NA is Avogadro's number, and R is the universal gas constant. Eq. (49) is called the van't Hoff equation,<sup>1</sup> which resembles the ideal gas law [33]. If the solution contains multiple species of solutes, Eq. (49) can be easily extended to
<sup>1</sup> Jacobus H. van't Hoff received the first Nobel Prize in Chemistry in 1901 for the discovery of osmotic pressure in solutions. https://www.nobelprize.org/nobel\_prizes/chemistry/laureates/1901/
$$
\Delta \pi = \sum\_{i} \left( \mathbf{C}\_{i,2} - \mathbf{C}\_{i,1} \right) \mathcal{R}T = \mathcal{R}T \Delta \mathbf{C} \tag{52}
$$
<sup>Δ</sup><sup>P</sup> <sup>¼</sup> <sup>n</sup><sup>1</sup> V Δx x1
Eq. (55):
where
where
which reduces to the identical result of Eq. (49):
kBT <sup>¼</sup> <sup>n</sup>1=<sup>N</sup> x1
NΔx
<sup>V</sup> kBT <sup>¼</sup> <sup>1</sup> �
The same result can be obtained in a slightly more mathematical way by directly using
dlnx dx
Δx
�ð∂ψ½ Þ ∂p <sup>T</sup>
<sup>¼</sup> kBT �ð∂ψ½ Þ ∂p <sup>T</sup>
<sup>Δ</sup>ðlnx<sup>Þ</sup> <sup>⋍</sup> dln<sup>x</sup>
<sup>Δ</sup>ðlnxÞ ¼ <sup>Δ</sup>ln<sup>x</sup>
is used. If Δx is finite, a similar approximation can be suggested:
Δπ <sup>¼</sup> <sup>R</sup><sup>T</sup>
�NA ð∂ψ½ Þ ∂p <sup>T</sup>
dx
Δx Δx
〈x〉ln <sup>¼</sup> <sup>Δ</sup><sup>x</sup>
is the logarithmic average of the solute mole fraction across the membrane interior. Employing Eq. (58) and Δx=x ¼ ΔC=C, we confirm that the osmotic pressure of the dilute concentration is
In this section, we mathematically proved that the osmotic pressure (of Eqs. (49), (60), and (66)) is valid for dilute solution consisting of weakly interacting molecules. Without losing generality, the absolute value of the osmotic pressure can be expressed as (similar to the ideal gas law)
Finally, it is worth noting that in Eq. (58), the negative sign of the partial derivative indicates that the gradients of solvent and solute concentrations have opposite signs. If the middle wall between the two boxes in Figure 3 is partially removed, then solvent and solutes will diffuse in opposite
<sup>x</sup> <sup>¼</sup> <sup>R</sup><sup>T</sup> <sup>n</sup>
NAV ΔC
Δx
<sup>Δ</sup><sup>x</sup> <sup>¼</sup> <sup>Δ</sup><sup>x</sup>
<sup>¼</sup> <sup>Δ</sup><sup>x</sup> 〈x〉ln
<sup>Δ</sup><sup>P</sup> <sup>¼</sup> kBT
Δn
<sup>V</sup> kBT <sup>¼</sup> <sup>Δ</sup>n=NA
Δπ ¼ RTΔC ð60Þ
<sup>Δ</sup><sup>x</sup> <sup>ð</sup>61<sup>Þ</sup>
Membrane Thermodynamics for Osmotic Phenomena http://dx.doi.org/10.5772/intechopen.68406
<sup>x</sup> <sup>ð</sup>62<sup>Þ</sup>
<sup>x</sup> <sup>ð</sup>63<sup>Þ</sup>
<sup>Δ</sup>ln<sup>x</sup> <sup>ð</sup>65<sup>Þ</sup>
π ¼ CRT ð67Þ
<sup>C</sup> <sup>¼</sup> <sup>R</sup>TΔ<sup>C</sup> <sup>ð</sup>66<sup>Þ</sup>
ð64Þ
<sup>V</sup> <sup>R</sup><sup>T</sup> <sup>ð</sup>59<sup>Þ</sup>
15
where <sup>Δ</sup><sup>C</sup> <sup>¼</sup> <sup>X</sup> <sup>i</sup> Ci;<sup>2</sup> � Ci;<sup>1</sup> � � is, in general, the difference of total mole concentration of solutes. If the total mass concentration of multiple species is known, then it should be carefully converted to total mole concentration using molecular weights of the contained species. The underlying assumptions of the van't Hoff equation (49) are summarized as follows:
#### 3.2.2. Using solute chemical potential
If the solvent chemical potentials of boxes 1 and 2 are equal, then the solute chemical potentials should be also the same:
$$
\mu\_{s,1} = \mu\_{s,2} \tag{53}
$$
which leads to
$$
\psi\left(P\_1, T\right) + k\_B T \ln \mathbf{x}\_1 = \psi(P\_2, T) + k\_B T \ln \mathbf{x}\_2 \tag{54}
$$
Using the same approximation for the pressure difference, we derive
$$-\left(\frac{\partial\psi}{\partial P}\right)\_T \Delta P = k\_B T \Delta (\text{lnx}) \tag{55}$$
$$=k\_{\mathbb{B}}T\ln\left(\frac{\chi\_2}{\chi\_1}\right)\tag{56}$$
where Δlnx ¼ lnx<sup>2</sup> � lnx<sup>1</sup> is the logarithmic difference between concentrations in two boxes. Eq. (56) can further be approximated as follows:
$$-k\_B T \ln\left(\frac{\mathbf{x}\_2}{\mathbf{x}\_1}\right) = -k\_B T \ln\left(1 + \frac{\Delta\mathbf{x}}{\mathbf{x}\_1}\right) \simeq -k\_B T \frac{\Delta\mathbf{x}}{\mathbf{x}\_1} \tag{57}$$
We treat the negative derivative of ψ with respect to P as the volume per each solute molecule, i.e.,
$$-\left(\frac{\partial\psi}{\partial P}\right)\_T = \frac{V}{n\_1} \sim \frac{V}{n\_2} \sim \frac{V}{\bar{n}}\tag{58}$$
where n ¼ ðn<sup>1</sup> þ n2Þ=2, implicitly assuming N ≫ ni ≫ Δn for i ¼ 1; 2. The pressure difference is then calculated as
Membrane Thermodynamics for Osmotic Phenomena http://dx.doi.org/10.5772/intechopen.68406 15
$$
\Delta P = \frac{n\_1}{V} \frac{\Delta \mathbf{x}}{\mathbf{x}\_1} k\_\mathbf{B} T = \frac{n\_1/N}{\mathbf{x}\_1} \frac{N \Delta \mathbf{x}}{V} k\_\mathbf{B} T = \mathbf{1} \cdot \frac{\Delta n}{V} k\_\mathbf{B} T = \frac{\Delta n/N\_A}{V} \mathcal{R} T \tag{59}
$$
which reduces to the identical result of Eq. (49):
$$
\Delta \pi = \mathcal{R} T \Delta \mathcal{C} \tag{60}
$$
The same result can be obtained in a slightly more mathematical way by directly using Eq. (55):
$$
\Delta P = \frac{k\_B T}{- (\partial \psi [\partial p])\_T} \left(\frac{\text{dlnx}}{\text{dx}}\right) \Delta x \tag{61}
$$
$$\mathbf{x} = \frac{k\_B T}{- (\partial \boldsymbol{\psi}[\partial p])\_T} \frac{\Delta \mathbf{x}}{\mathbf{x}} \tag{62}$$
where
Δπ <sup>¼</sup> <sup>X</sup> i
1. The solute concentration is much smaller than the solvent concentration.
Using the same approximation for the pressure difference, we derive
Eq. (56) can further be approximated as follows:
�kBTln <sup>x</sup><sup>2</sup>
x1 � �
> � <sup>∂</sup><sup>ψ</sup> ∂P � �
T ¼ V n1 � V n2 � V
� <sup>∂</sup><sup>ψ</sup> ∂P � �
T
<sup>¼</sup> kBTln <sup>x</sup><sup>2</sup>
where Δlnx ¼ lnx<sup>2</sup> � lnx<sup>1</sup> is the logarithmic difference between concentrations in two boxes.
¼ �kBTln 1 þ
We treat the negative derivative of ψ with respect to P as the volume per each solute
where n ¼ ðn<sup>1</sup> þ n2Þ=2, implicitly assuming N ≫ ni ≫ Δn for i ¼ 1; 2. The pressure difference is
x1 � �
Δx x1 � �
<sup>⋍</sup> � kBT <sup>Δ</sup><sup>x</sup> x1
where <sup>Δ</sup><sup>C</sup> <sup>¼</sup> <sup>X</sup>
14 Desalination
<sup>i</sup> Ci;<sup>2</sup> � Ci;<sup>1</sup>
3.2.2. Using solute chemical potential
should be also the same:
which leads to
molecule, i.e.,
then calculated as
the van't Hoff equation (49) are summarized as follows:
2. Temperature gradient between the two boxes is zero.
Ci;<sup>2</sup> � Ci;<sup>1</sup>
If the total mass concentration of multiple species is known, then it should be carefully converted to total mole concentration using molecular weights of the contained species. The underlying assumptions of
3. The Gibbs free energy of a dilute solution is described using the weak solution approach.
If the solvent chemical potentials of boxes 1 and 2 are equal, then the solute chemical potentials
� � is, in general, the difference of total mole concentration of solutes.
� �R<sup>T</sup> <sup>¼</sup> <sup>R</sup>TΔ<sup>C</sup> <sup>ð</sup>52<sup>Þ</sup>
μ<sup>s</sup>;<sup>1</sup> ¼ μ<sup>s</sup>;<sup>2</sup> ð53Þ
ΔP ¼ kBTΔðlnxÞ ð55Þ
n¯ <sup>ð</sup>58<sup>Þ</sup>
ð56Þ
ð57Þ
<sup>ψ</sup> <sup>P</sup><sup>1</sup>;<sup>T</sup> � � <sup>þ</sup> kBT lnx<sup>1</sup> <sup>¼</sup> <sup>ψ</sup>ð Þþ <sup>P</sup>2;<sup>T</sup> kBTln <sup>x</sup><sup>2</sup> <sup>ð</sup>54<sup>Þ</sup>
$$
\Delta(\text{lnx}) \simeq \left(\frac{\text{dlnx}}{\text{d}x}\right) \Delta x = \frac{\Delta x}{x} \tag{63}
$$
is used. If Δx is finite, a similar approximation can be suggested:
$$
\Delta(\text{lnx}) = \left(\frac{\Delta \text{lnx}}{\Delta \text{x}} \Delta \text{x}\right) = \frac{\Delta \text{x}}{\langle \text{x} \rangle\_{\text{ln}}} \tag{64}
$$
where
$$
\langle \mathbf{x} \rangle\_{\ln} = \frac{\Delta \mathbf{x}}{\Delta \ln \mathbf{x}} \tag{65}
$$
is the logarithmic average of the solute mole fraction across the membrane interior. Employing Eq. (58) and Δx=x ¼ ΔC=C, we confirm that the osmotic pressure of the dilute concentration is
$$
\Delta \pi = \frac{\mathcal{R}T}{-N\_A \left(\partial \psi[\partial p]\right)\_\Gamma} \frac{\Delta \mathbf{x}}{\mathbf{x}} = \mathcal{R}T \left(\frac{\mathbf{n}}{N\_A V}\right) \frac{\Delta \mathbf{C}}{\mathbf{C}} = \mathcal{R}T \Delta \mathbf{C} \tag{66}
$$
In this section, we mathematically proved that the osmotic pressure (of Eqs. (49), (60), and (66)) is valid for dilute solution consisting of weakly interacting molecules. Without losing generality, the absolute value of the osmotic pressure can be expressed as (similar to the ideal gas law)
$$
\pi = \mathcal{CRT} \tag{67}
$$
Finally, it is worth noting that in Eq. (58), the negative sign of the partial derivative indicates that the gradients of solvent and solute concentrations have opposite signs. If the middle wall between the two boxes in Figure 3 is partially removed, then solvent and solutes will diffuse in opposite
where Cw is the concentration of water dissolved in the membrane. Assuming that the dissolved water in the membrane material can be treated as a Henrian solution, the chemical potential of
μ<sup>w</sup> ¼ constant þ RTlnCw ð69Þ
Membrane Thermodynamics for Osmotic Phenomena http://dx.doi.org/10.5772/intechopen.68406 17
Δμ<sup>w</sup> ð70Þ
dCs þ VwΔP ð71Þ
dCs ¼ �VwΔπ ð72Þ
Δμ<sup>w</sup> ¼ Vwð Þ ΔP � Δπ ð73Þ
Jw ¼ Að Þ ΔP � Δπ ð74Þ
ð75Þ
the pure water (in the membrane solvent) is
In this model, the underlying assumptions are:
Substitution of Eq. (69) into Eq. (68) gives
pressure difference, i.e., ΔP ¼ Δπ, which gives
Substitution of Eq. (73) into Eq. (70) provides
Then, Eq. (71) is simplified to
written as
2. The solution is considered as Henrian for water.
1. The water and solute molecules dissolve into a membrane material.
3. The feed and permeate streams are immiscible with the membrane.
Jw ¼ �Dw
Δμ<sup>w</sup> ¼
ð ∂μ<sup>w</sup> ∂Cs
Cw RT
ð ∂μ<sup>w</sup> ∂Cs
dμ<sup>w</sup>
where Δμ<sup>w</sup> is the transmembrane difference of μ<sup>w</sup> at a constant temperature T, which can be
where Cs is the solute concentration. In the previous section, we proved that the chemical potential difference between two subsystems should vanish in the isothermal equilibrium (i.e., ΔT ¼ 0): Δμ<sup>w</sup> ¼ 0. In this case, the transmembrane pressure difference is equal to the osmotic
which is the governing equation of solvent transport through the membrane as a medium in which water and solutes can dissolve. Here, A is the water permeability through the membrane:
> <sup>A</sup> <sup>¼</sup> DwCwVw RTδ<sup>m</sup>
<sup>d</sup><sup>x</sup> <sup>≈</sup> DwCw RTδ<sup>m</sup>
Figure 3. Osmotic pressure schematic: two boxes separated by the semipermeable wall in the isothermal environment.
directions. This should be treated in principle as a binary diffusion of two species (i.e., solvent and solute) by exchanging their positions.
#### 4. Solution-diffusion model revisited
#### 4.1. Solvent (water) transport
For pressure-driven membrane processes such as RO and NF, the applied pressure should overcome the osmotic pressure difference across the membrane. In feed and permeate solutions, salts are dissolved as solutes in the solvent water. The hydraulic pressure generates solvent flow through the membrane, which may contribute to solute transport through the membrane surface. Water molecules, however, dissolve as solutes in the membrane material (as solvent). Due to the high density of the membrane, water molecules can migrate via diffusion from a higher concentration region to a lower concentration region. This normal diffusion is reversed by applying hydraulic pressure to the feed solution with a high concentration such as seawater. Water permeation through a RO membrane can be pictured as diffusion driven by the external hydraulic pressure, which allows us to neglect convective transport of solutes through the membrane. The phenomenological phase of water in the membrane leads to solute transport as Fickian, which is also closely related to the osmotic pressure gradient between two subsystems. The above-mentioned mechanisms are included in solution-diffusion model, proposed by Lonsdale et al. [30]. An extensive overview of RO models can be found elsewhere [34–39].
Consider a semipermeable membrane of thickness δm, with high and low concentrations on two sides. The solvent flux through the membrane is assumed to be Fickian [40, 41]:
$$J\_w = -D\_w \frac{\mathbf{d} \mathbf{C}\_w}{\mathbf{d} \mathbf{x}} \tag{68}$$
where Cw is the concentration of water dissolved in the membrane. Assuming that the dissolved water in the membrane material can be treated as a Henrian solution, the chemical potential of the pure water (in the membrane solvent) is
$$
\mu\_w = \text{constant} + \mathcal{R}T \text{lnC}\_w \tag{69}
$$
In this model, the underlying assumptions are:
Substitution of Eq. (69) into Eq. (68) gives
directions. This should be treated in principle as a binary diffusion of two species (i.e., solvent
Figure 3. Osmotic pressure schematic: two boxes separated by the semipermeable wall in the isothermal environment.
For pressure-driven membrane processes such as RO and NF, the applied pressure should overcome the osmotic pressure difference across the membrane. In feed and permeate solutions, salts are dissolved as solutes in the solvent water. The hydraulic pressure generates solvent flow through the membrane, which may contribute to solute transport through the membrane surface. Water molecules, however, dissolve as solutes in the membrane material (as solvent). Due to the high density of the membrane, water molecules can migrate via diffusion from a higher concentration region to a lower concentration region. This normal diffusion is reversed by applying hydraulic pressure to the feed solution with a high concentration such as seawater. Water permeation through a RO membrane can be pictured as diffusion driven by the external hydraulic pressure, which allows us to neglect convective transport of solutes through the membrane. The phenomenological phase of water in the membrane leads to solute transport as Fickian, which is also closely related to the osmotic pressure gradient between two subsystems. The above-mentioned mechanisms are included in solution-diffusion model, proposed by Lonsdale et al. [30]. An extensive overview of RO models
Consider a semipermeable membrane of thickness δm, with high and low concentrations on
dCw
<sup>d</sup><sup>x</sup> <sup>ð</sup>68<sup>Þ</sup>
two sides. The solvent flux through the membrane is assumed to be Fickian [40, 41]:
Jw ¼ �Dw
and solute) by exchanging their positions.
4.1. Solvent (water) transport
16 Desalination
can be found elsewhere [34–39].
4. Solution-diffusion model revisited
$$J\_w = -D\_w \frac{\mathbb{C}\_w}{\mathcal{R}T} \frac{\mathbf{d}\mu\_w}{\mathbf{d}\mathbf{x}} \approx \frac{D\_w \mathbb{C}\_w}{\mathcal{R}T \delta\_m} \Delta \mu\_w \tag{70}$$
where Δμ<sup>w</sup> is the transmembrane difference of μ<sup>w</sup> at a constant temperature T, which can be written as
$$
\Delta \mu\_w = \begin{cases}
\frac{\partial \mu\_w}{\partial \mathbf{C}\_s} \, \mathbf{d} \mathbf{C}\_s + \overline{V}\_w \Delta P
\end{cases} \tag{71}
$$
where Cs is the solute concentration. In the previous section, we proved that the chemical potential difference between two subsystems should vanish in the isothermal equilibrium (i.e., ΔT ¼ 0): Δμ<sup>w</sup> ¼ 0. In this case, the transmembrane pressure difference is equal to the osmotic pressure difference, i.e., ΔP ¼ Δπ, which gives
$$\int \frac{\partial \mu\_w}{\partial \mathbf{C}\_s} \, d\mathbf{C}\_s = -\overline{V}\_w \Delta \pi \,\tag{72}$$
Then, Eq. (71) is simplified to
$$
\Delta \mu\_w = \overline{V}\_w (\Delta P - \Delta \pi) \tag{73}
$$
Substitution of Eq. (73) into Eq. (70) provides
$$J\_w = A(\Delta P - \Delta \pi) \tag{74}$$
which is the governing equation of solvent transport through the membrane as a medium in which water and solutes can dissolve. Here, A is the water permeability through the membrane:
$$A = \frac{D\_w \mathbb{C}\_w \overline{V}\_w}{\mathcal{R}T \delta\_m} \tag{75}$$
which is a characteristic value of the membrane. It is challenging to predict or measure the diffusion coefficient Dw and dissolved concentration Cw of water molecules in the membrane material. Therefore, the water permeability A is often experimentally estimated by a linear regression plot of Jw versus ΔP using fresh water as a solute-free feed solution.
#### 4.2. Solute transport
The solute transport through the membrane is also assumed to follow Fick's law:
$$J\_s = -D\_s \frac{d\mathbf{C}\_s}{d\mathbf{x}} \simeq D\_s \frac{d\mathbf{C}\_s}{\delta\_\mathbf{m}}\tag{76}$$
be used [46–48]. To the best of our knowledge, the NESM still burgeons in pure theoretical physics. Therefore, general solutions for irreversible engineering processes are barely found. The minimum condition for us to use equilibrium filtration theory is that the filtration system is already in a steady state, in which no physical quantities vary explicitly with respect to time,
∂½ �
where [ ] can hold any variables associated to the filtration system. The steady state is, in principle, far away from static equilibrium. Note that the osmotic pressure is derived from a pure equilibrium state, especially for the isobaric-isothermal ensemble. This implies that the solution-diffusion model becomes less accurate if fluid flows in the membrane channels are
To investigate the intrinsically nonequilibrium filtration processes, the irreversible thermodynamic models were developed using the Onsager2 reciprocal theorem [49]. Kedem and Katchalsky represented the local dissipation rate of free energy per unit volume as dissipation
<sup>k</sup> � ∇ �μ<sup>k</sup>
entropy must increase and therefore the dissipation rate is positive-definite, i.e., φ > 0. The
where Jv and JD are the total volumetric flux and the solute velocity relative to the solvent
where Lij are coupling coefficients of the phenomenological fluxes, Jv and JD. In order to satisfy
Lars Onsager received the Nobel Prize in Chemistry in 1968 for the discovery of the reciprocal relations in the fundamental thermodynamics of irreversible processes. http://www.nobelprize.org/nobel\_prizes/chemistry/laureates/1968/
function for isothermal, nonelectrolyte systems in a steady state [50]:
<sup>φ</sup> <sup>¼</sup> <sup>X</sup><sup>n</sup> k¼1 J !
!
<sup>∂</sup><sup>t</sup> <sup>¼</sup> <sup>0</sup> <sup>ð</sup>97<sup>Þ</sup>
Membrane Thermodynamics for Osmotic Phenomena http://dx.doi.org/10.5772/intechopen.68406 19
� � <sup>&</sup>gt; <sup>0</sup> <sup>ð</sup>80<sup>Þ</sup>
<sup>k</sup>. In the irreversible (i.e., nonequilibrium) process,
φ ¼ JvΔP þ JDΔπ ð81Þ
Jv ¼ L11ΔP þ L12Δπ ð82Þ JD ¼ L21ΔP þ L22Δπ ð83Þ
L11; L<sup>22</sup> > 0 ð84Þ
i.e., mathematically,
fast enough or almost turbulent.
for species k, having a constant flux J
dissipation function for RO is
and
2
velocity, respectively. One can write
Eq. (81), the following two conditions must be met
where Cs and Ds are the concentration and diffusivity of solutes dissolved in the membrane, respectively, and ΔCs is the solute concentration difference across the membrane interior. Similarly to Cw, Cs is hard to measure. Therefore, ΔCs is assumed to be proportional to that between membrane surfaces ΔCm. The partition coefficient K is then defined as
$$K = \frac{\Delta \mathbb{C}\_s}{\Delta \mathbb{C}\_m} \quad (<1) \tag{77}$$
where ΔCm is often approximated as the difference between feed concentration Cf and permeate concentration Cp in the RO processes. Substitution of Eq. (77) into Eq. (76) provides
$$J\_s = B\Delta\mathbb{C}\_m \tag{78}$$
where
$$B = \frac{D\_s K}{\delta\_m} \tag{79}$$
is the solute permeability through the membrane. Note that B conventionally has the same dimension as Jw. Similarly to A, B can be macroscopically measured by independent experiments, providing a Js versus ΔCm graph. The slope of the graph, estimated using linear regression, is equal to B.
#### 5. Thermodynamic irreversibility of filtration
A thermodynamic system has three types: open, closed, and isolated. In the open system, mass and heat can pass in and out of the system in contact with the environment. Only heat can be transferred between the closed system and the environment, and neither mass nor heat can be exchanged in the isolated system. Rigorously saying, all the filtration processes are open systems, having entering feed streams to be treated. Temperature gradients across the membrane are often negligible in the pressure- or osmosis-driven filtration processes, but significant in thermal membrane processes such as membrane distillation processes [42–45]. To address the open filtration processes, nonequilibrium statistical mechanics (NESM) should be used [46–48]. To the best of our knowledge, the NESM still burgeons in pure theoretical physics. Therefore, general solutions for irreversible engineering processes are barely found. The minimum condition for us to use equilibrium filtration theory is that the filtration system is already in a steady state, in which no physical quantities vary explicitly with respect to time, i.e., mathematically,
$$\frac{\partial \mathbf{J}}{\partial t} = \mathbf{0} \tag{97}$$
where [ ] can hold any variables associated to the filtration system. The steady state is, in principle, far away from static equilibrium. Note that the osmotic pressure is derived from a pure equilibrium state, especially for the isobaric-isothermal ensemble. This implies that the solution-diffusion model becomes less accurate if fluid flows in the membrane channels are fast enough or almost turbulent.
To investigate the intrinsically nonequilibrium filtration processes, the irreversible thermodynamic models were developed using the Onsager2 reciprocal theorem [49]. Kedem and Katchalsky represented the local dissipation rate of free energy per unit volume as dissipation function for isothermal, nonelectrolyte systems in a steady state [50]:
$$\phi = \sum\_{k=1}^{n} \overrightarrow{f}\_k \cdot \nabla \left( -\mu\_k \right) > 0 \tag{80}$$
for species k, having a constant flux J ! <sup>k</sup>. In the irreversible (i.e., nonequilibrium) process, entropy must increase and therefore the dissipation rate is positive-definite, i.e., φ > 0. The dissipation function for RO is
$$
\Phi = \mathbf{J}\_v \Delta \mathbf{P} + \mathbf{J}\_D \Delta \pi \tag{81}
$$
where Jv and JD are the total volumetric flux and the solute velocity relative to the solvent velocity, respectively. One can write
$$J\_v = L\_{11} \Delta P + L\_{12} \Delta \pi \tag{82}$$
$$J\_D = L\_{21} \Delta P + L\_{22} \Delta \pi \tag{83}$$
where Lij are coupling coefficients of the phenomenological fluxes, Jv and JD. In order to satisfy Eq. (81), the following two conditions must be met
$$L\_{11}, L\_{22} > 0\tag{84}$$
and
which is a characteristic value of the membrane. It is challenging to predict or measure the diffusion coefficient Dw and dissolved concentration Cw of water molecules in the membrane material. Therefore, the water permeability A is often experimentally estimated by a linear
> dCs <sup>d</sup><sup>x</sup> <sup>⋍</sup> Ds
where Cs and Ds are the concentration and diffusivity of solutes dissolved in the membrane, respectively, and ΔCs is the solute concentration difference across the membrane interior. Similarly to Cw, Cs is hard to measure. Therefore, ΔCs is assumed to be proportional to that between
where ΔCm is often approximated as the difference between feed concentration Cf and perme-
<sup>B</sup> <sup>¼</sup> DsK δm
is the solute permeability through the membrane. Note that B conventionally has the same dimension as Jw. Similarly to A, B can be macroscopically measured by independent experiments, providing a Js versus ΔCm graph. The slope of the graph, estimated using linear
A thermodynamic system has three types: open, closed, and isolated. In the open system, mass and heat can pass in and out of the system in contact with the environment. Only heat can be transferred between the closed system and the environment, and neither mass nor heat can be exchanged in the isolated system. Rigorously saying, all the filtration processes are open systems, having entering feed streams to be treated. Temperature gradients across the membrane are often negligible in the pressure- or osmosis-driven filtration processes, but significant in thermal membrane processes such as membrane distillation processes [42–45]. To address the open filtration processes, nonequilibrium statistical mechanics (NESM) should
<sup>K</sup> <sup>¼</sup> <sup>Δ</sup>Cs ΔCm
ate concentration Cp in the RO processes. Substitution of Eq. (77) into Eq. (76) provides
ΔCs δm
ð Þ < 1 ð77Þ
Js ¼ BΔCm ð78Þ
ð76Þ
ð79Þ
regression plot of Jw versus ΔP using fresh water as a solute-free feed solution.
The solute transport through the membrane is also assumed to follow Fick's law:
Js ¼ �Ds
membrane surfaces ΔCm. The partition coefficient K is then defined as
4.2. Solute transport
18 Desalination
where
regression, is equal to B.
5. Thermodynamic irreversibility of filtration
<sup>2</sup> Lars Onsager received the Nobel Prize in Chemistry in 1968 for the discovery of the reciprocal relations in the fundamental thermodynamics of irreversible processes. http://www.nobelprize.org/nobel\_prizes/chemistry/laureates/1968/
$$L\_{11}L\_{22} \ge L\_{12}L\_{21} = L\_{12}^2 \tag{85}$$
indicates that the perfect rejection can be achieved if L<sup>22</sup> � jL12j ¼ 0 in addition to σ ¼ 0. In membrane separations, the perfect rejection is related not only to the thermodynamic state of the membrane surface, but also to the specific membrane materials having salt rejecting
<sup>∂</sup><sup>σ</sup> ¼ �Δπ <sup>ð</sup>92<sup>Þ</sup>
Membrane Thermodynamics for Osmotic Phenomena http://dx.doi.org/10.5772/intechopen.68406 21
<sup>∂</sup><sup>σ</sup> � <sup>Δ</sup><sup>P</sup> <sup>ð</sup>94<sup>Þ</sup>
<sup>∂</sup><sup>σ</sup> ¼ �½Δ<sup>P</sup> <sup>þ</sup> Δπ� þ Δπσ <sup>ð</sup>93<sup>Þ</sup>
Furthermore, variations of Jv and Js with respect to σ can be investigated by calculating
1 L<sup>11</sup> ∂Jv
∂Js
∂Js
<sup>∂</sup><sup>σ</sup> � σΔπ <sup>¼</sup> <sup>1</sup>
which is valid for an arbitrary σ between 0 and 1. Eq. (92) indicates that Jv monotonously decreases with respect to σ. If the filtration system is in a transient, nonequilibrium state far from the pure static equilibrium, the volumetric flux Jv must be higher than that in the quasiequilibrium state. The left-hand side of Eq. (93) is �ΔP at σ ¼ 1 and �½ΔP þ Δπ� at
Overall, σ can be physically interpreted not only as the filtration coefficient, but also as the equilibrium coefficient. When σ ! 1, the effect of the osmotic pressure difference reaches its maximum of the quasiequilibrium state, but the zero solute flux is not automatically guaranteed. The perfect rejection is achieved if the additional condition L<sup>22</sup> ¼ jL12j is satisfied, which is, however, independent of σ. The difference of σ indicates how much the filtration system is phenomenologically close to the static equilibrium. On the other side, if σ ! 0, then the filtration system can be in a steady state, but it is far from the static equilibrium. Jv and Js approach their theoretical maximum values, and the solute transport is significantly influenced by convection. Although σ is a fundamentally and practically important parameter, to the best of our knowledge, there are no standard theories to directly predict σ. This is because the irreversible thermodynamic model relaxes the equilibrium restriction, but the
This work was financially supported by the R&D project of "Infrastructure Establishment of Thermal Energy Conversion and Desalination using Seawater Thermal Energy(2/3)" (PES9060)
from Korea Research Institute of Ships and Ocean Engineering (KRISO).
L<sup>11</sup> ∂Jv
1 CsL<sup>11</sup>
1 CsL<sup>11</sup>
σ ¼ 0: as σ decreases, the magnitude of ∂Js=∂σ increases.
NESM has not been fully developed yet.
Acknowledgements
and substitution of Eq. (92) into Eq. (93) gives
capabilities.
where L<sup>12</sup> ¼ L21.
After some theoretical steps, Kedem and Katchalsky [50] derived
$$J\_v = L\_{11} \left(\Delta P - \sigma \Delta \pi\right) \tag{86}$$
$$J\_s = \overline{\overline{\mathbb{C}}}\_s (1 - \sigma) J\_v + a\Delta\pi \tag{87}$$
where Cs is the logarithmic average of concentrations on the two membrane sides, σ ¼ �L12=L<sup>11</sup> assuming L<sup>11</sup> > 0 and L<sup>12</sup> < 0, and
$$
\omega = \overline{\mathsf{C}}\_{s} \frac{L\_{11} L\_{22} - L\_{12}^2}{L\_{11}} = \overline{\mathsf{C}}\_{s} (L\_{22} - \sigma |L\_{12}|) \tag{88}
$$
Here, σ is defined as the "filtration coefficient," representing the solute rejection property. Kedem and Katchalsky [50] interpret the physical meaning of σ as follows: when σ ¼ 1:0, the membrane is completely impermeable to solute and rejection is 100%, and when σ ¼ 0:0, the membrane is completely permeable to solute and rejection is zero. It is worth noting that the irreversible thermodynamic theory includes the solution-diffusion model as a special case. If σ ¼ 1:0, then Eqs. (86) and (87) reduce to
$$J\_v = L\_{11} \left( \Delta P - \Delta \pi \right) \to A \left( \Delta P - \Delta \pi \right) \tag{89}$$
$$J\_s = \overline{\mathbb{C}}\_s (L\_{22} - |L\_{12}|) \Delta \pi \to B \,\Delta \mathbb{C} \tag{90}$$
where L<sup>11</sup> ¼ A and B ¼ ωΔπ=ΔC, assuming the osmotic pressure is linearly proportional to the solute concentration.
In our opinion, σ ¼ 1 can be interpreted in a different way. Because the unity σ in Eqs. (86) and (87) indicates that the effect of Δπ is maximized, the thermodynamic state of the membrane surface is quite close to the static equilibrium state. The solvent flux can be considered as the barometric diffusion of water as Jv increases with ΔP, overcoming Δπ across the membrane. The solute flux in this case is purely Fickian, which is dominated by only ωΔπð Þ ∝ΔC in Eq. (87). The limiting value of σ ! 1, however, does not guarantee the perfect rejection of solutes because it does not satisfy Js ¼ 0 in Eq. (90). Knowing L<sup>12</sup> < 0, one can rewrite Eq. (85) to give
$$|L\_{22} - |L\_{12}| \ge \frac{L\_{12}^2}{L\_{11}} + L\_{12} = \sigma^2 L\_{11} + L\_{12} = L\_{11}\sigma(\sigma - 1) \tag{91}$$
which indicates that the solute flux Js may vanish if σ ¼ 0 or 1. Here, we have to discard σ ¼ 0 because Js in Eq. (87) reaches its maximum at σ ¼ 0. Then, the condition σ ¼ 1 applied to Eq. (91) must be only a necessary condition for Js ¼ 0. The inequality relationship in Eq. (91) indicates that the perfect rejection can be achieved if L<sup>22</sup> � jL12j ¼ 0 in addition to σ ¼ 0. In membrane separations, the perfect rejection is related not only to the thermodynamic state of the membrane surface, but also to the specific membrane materials having salt rejecting capabilities.
Furthermore, variations of Jv and Js with respect to σ can be investigated by calculating
$$\frac{1}{L\_{11}}\frac{\partial l\_v}{\partial \sigma} = -\Delta \pi \tag{92}$$
$$\frac{1}{\overline{\mathbb{C}}\_{s}L\_{11}}\frac{\partial l\_{s}}{\partial \sigma} = -[\Delta P + \Delta \pi] + \Delta \pi \sigma \tag{93}$$
and substitution of Eq. (92) into Eq. (93) gives
<sup>L</sup><sup>11</sup> <sup>L</sup><sup>22</sup> <sup>≥</sup> <sup>L</sup>12L<sup>21</sup> <sup>¼</sup> <sup>L</sup><sup>2</sup>
where Cs is the logarithmic average of concentrations on the two membrane sides,
12
Here, σ is defined as the "filtration coefficient," representing the solute rejection property. Kedem and Katchalsky [50] interpret the physical meaning of σ as follows: when σ ¼ 1:0, the membrane is completely impermeable to solute and rejection is 100%, and when σ ¼ 0:0, the membrane is completely permeable to solute and rejection is zero. It is worth noting that the irreversible thermodynamic theory includes the solution-diffusion model as a special case. If
where L<sup>11</sup> ¼ A and B ¼ ωΔπ=ΔC, assuming the osmotic pressure is linearly proportional to the
In our opinion, σ ¼ 1 can be interpreted in a different way. Because the unity σ in Eqs. (86) and (87) indicates that the effect of Δπ is maximized, the thermodynamic state of the membrane surface is quite close to the static equilibrium state. The solvent flux can be considered as the barometric diffusion of water as Jv increases with ΔP, overcoming Δπ across the membrane. The solute flux in this case is purely Fickian, which is dominated by only ωΔπð Þ ∝ΔC in Eq. (87). The limiting value of σ ! 1, however, does not guarantee the perfect rejection of solutes because it does not satisfy Js ¼ 0 in Eq. (90). Knowing L<sup>12</sup> < 0, one can rewrite Eq. (85)
<sup>L</sup>11L<sup>22</sup> � <sup>L</sup><sup>2</sup>
L<sup>11</sup>
After some theoretical steps, Kedem and Katchalsky [50] derived
ω ¼ Cs
σ ¼ �L12=L<sup>11</sup> assuming L<sup>11</sup> > 0 and L<sup>12</sup> < 0, and
σ ¼ 1:0, then Eqs. (86) and (87) reduce to
L<sup>22</sup> � jL12j ≥
L2 12 L<sup>11</sup>
<sup>þ</sup> <sup>L</sup><sup>12</sup> <sup>¼</sup> <sup>σ</sup><sup>2</sup>
which indicates that the solute flux Js may vanish if σ ¼ 0 or 1. Here, we have to discard σ ¼ 0 because Js in Eq. (87) reaches its maximum at σ ¼ 0. Then, the condition σ ¼ 1 applied to Eq. (91) must be only a necessary condition for Js ¼ 0. The inequality relationship in Eq. (91)
solute concentration.
to give
where L<sup>12</sup> ¼ L21.
20 Desalination
<sup>12</sup> ð85Þ
Jv ¼ L<sup>11</sup> ð Þ ΔP � σΔπ ð86Þ
Js ¼ Csð Þ 1 � σ Jv þ ωΔπ ð87Þ
Jv ¼ L<sup>11</sup> ð Þ! ΔP � Δπ A ð Þ ΔP � Δπ ð89Þ
Js ¼ Csð Þ L<sup>22</sup> � jL12j Δπ ! B ΔC ð90Þ
¼ Csð Þ L<sup>22</sup> � σjL12j ð88Þ
L<sup>11</sup> þ L<sup>12</sup> ¼ L11σ σð Þ � 1 ð91Þ
$$\frac{1}{\overline{\mathbb{C}}\_{s}L\_{11}}\frac{\partial l\_{s}}{\partial \sigma} - \sigma \Delta \pi = \frac{1}{L\_{11}}\frac{\partial l\_{v}}{\partial \sigma} - \Delta P \tag{94}$$
which is valid for an arbitrary σ between 0 and 1. Eq. (92) indicates that Jv monotonously decreases with respect to σ. If the filtration system is in a transient, nonequilibrium state far from the pure static equilibrium, the volumetric flux Jv must be higher than that in the quasiequilibrium state. The left-hand side of Eq. (93) is �ΔP at σ ¼ 1 and �½ΔP þ Δπ� at σ ¼ 0: as σ decreases, the magnitude of ∂Js=∂σ increases.
Overall, σ can be physically interpreted not only as the filtration coefficient, but also as the equilibrium coefficient. When σ ! 1, the effect of the osmotic pressure difference reaches its maximum of the quasiequilibrium state, but the zero solute flux is not automatically guaranteed. The perfect rejection is achieved if the additional condition L<sup>22</sup> ¼ jL12j is satisfied, which is, however, independent of σ. The difference of σ indicates how much the filtration system is phenomenologically close to the static equilibrium. On the other side, if σ ! 0, then the filtration system can be in a steady state, but it is far from the static equilibrium. Jv and Js approach their theoretical maximum values, and the solute transport is significantly influenced by convection. Although σ is a fundamentally and practically important parameter, to the best of our knowledge, there are no standard theories to directly predict σ. This is because the irreversible thermodynamic model relaxes the equilibrium restriction, but the NESM has not been fully developed yet.
#### Acknowledgements
This work was financially supported by the R&D project of "Infrastructure Establishment of Thermal Energy Conversion and Desalination using Seawater Thermal Energy(2/3)" (PES9060) from Korea Research Institute of Ships and Ocean Engineering (KRISO).
## Author details
Albert S. Kim<sup>1</sup> \* and Heyon-Ju Kim2
\*Address all correspondence to: albertsk@hawaii.edu
1 Civil and Environmental Engineering, University of Hawaii at Manoa, Honolulu, United States of America
[11] Acrivos A, Batchelor GK, Hinch EJ, Koch DL, Mauri R. Longitudinal shear-induced diffusion of spheres in a dilute suspension. Journal of Fluid Mechanics. 1992 Jul;240
Membrane Thermodynamics for Osmotic Phenomena http://dx.doi.org/10.5772/intechopen.68406 23
[12] Kim AS, Liu Y. Irreversible chemical potential and shear-induced diffusion in crossflow filtration. Industrial & Engineering Chemistry Research. 2008;47(15):5611–5614
[13] Sablani S, Goosen M, Al-Belushi R, Wilf M. Concentration polarization in ultrafiltration and reverse osmosis: A critical review. Desalination. 2001 Dec;141(3):269–289. Available
[14] Lee KP, Arnot TC, Mattia D. A review of reverse osmosis membrane materials for desalination—Development to date and future potential. Journal of Membrane Science. 2011 Mar;370(1–2):1–22. Available from: http://dx.doi.org/10.1016/j.memsci.2010.12.
[15] Kang Gd, Cao Ym. Development of antifouling reverse osmosis membranes for water treatment: A review. Water Research. 2012 Mar;46(3):584–600. Available from: http://dx.
[16] Pérez-González A, Urtiaga AM, Ibáñez R, Ortiz I. State of the art and review on the treatment technologies of water reverse osmosis concentrates. Water Research. 2012
[17] Rautenbach R, Linn T. High-pressure reverse osmosis and nanofiltration, a ''zero discharge'' process combination for the treatment of waste water with severe fouling/scaling potential. Desalination. 1996 Jun;105(1–2):63–70. Available from: http://dx.doi.org/10.
[18] Avlonitis SA, Kouroumbas K, Vlachakis N. Energy consumption and membrane replacement cost for seawater RO desalination plants. Desalination. 2003 Aug;157(1–3):151–158.
[19] Mazlan NM, Peshev D, Livingston AG. Energy consumption for desalination — A comparison of forward osmosis with reverse osmosis, and the potential for perfect membranes. Desalination. 2016 Jan;377:138–151. Available from: http://dx.doi.org/10.1016/j.
[20] Werber JR, Deshmukh A, Elimelech M. Can batch or semi-batch processes save energy in reverse-osmosis desalination? Desalination. 2017 Jan;402:109–122. Available from: http://
[21] Lin S, Elimelech M. Kinetics and energetics trade-off in reverse osmosis desalination with different configurations. Desalination. 2017 Jan;401:42–52. Available from: http://dx.doi.
[22] Cath T, AA Childress, Elimelech M. Forward osmosis: Principles, applications, and recent developments. Journal of Membrane Science. 2006 Sep;281(1–2):70–87. Available from:
Available from: http://dx.doi.org/10.1016/S0011-9164(03)00395-3.
Feb;46(2):267–283. Available from: http://dx.doi.org/10.1016/j.watres.2011.10.046.
(1):651. Available from: http://dx.doi.org/10.1017/S0022112092000247.
from: http://dx.doi.org/10.1016/S0011-9164(01)85005-0.
doi.org/10.1016/j.watres.2011.11.041.
1016/0011-9164(96)00059-8.
desal.2015.08.011.
dx.doi.org/10.1016/j.desal.2016.09.028.
http://dx.doi.org/10.1016/j.memsci.2006.05.048.
org/10.1016/j.desal.2016.09.008.
036.
2 Seawater Utilization Plant Research Center, Korea Research Institute of Ships and Ocean Engineering, Gangwon-do, Republic of Korea
## References
[11] Acrivos A, Batchelor GK, Hinch EJ, Koch DL, Mauri R. Longitudinal shear-induced diffusion of spheres in a dilute suspension. Journal of Fluid Mechanics. 1992 Jul;240 (1):651. Available from: http://dx.doi.org/10.1017/S0022112092000247.
Author details
\* and Heyon-Ju Kim2
Engineering, Gangwon-do, Republic of Korea
dx.doi.org/10.1016/j.jwpe.2014.04.003.
\*Address all correspondence to: albertsk@hawaii.edu
1 Civil and Environmental Engineering, University of Hawaii at Manoa, Honolulu, United
2 Seawater Utilization Plant Research Center, Korea Research Institute of Ships and Ocean
[1] Cheryan M. Microfiltration and Ultrafiltration Handbook. New York, NY: Technomic
[2] Shi X, Tal G, Hankins NP, Gitis V. Fouling and cleaning of ultrafiltration membranes: A review. Journal of Water Process Engineering. 2014 Apr;1:121–138. Available from: http://
[3] Zeman LJ, Zydney AL. Microfiltration and Ultrafiltration: Principles and Applications.
[4] Ho CC, Zydney AL. A Combined pore blockage and cake filtration model for protein fouling during microfiltration. Journal of Colloid and Interface Science. 2000 Dec;232
[5] Einstein A. {Ü}ber die von der molekularkinetischen Theorie der W{ä}rme geforderte Bewegung von in ruhenden Fl{ü}ssigkeiten suspendierten Teilchen. Annals d Physik.
[6] Einstein A. On the theory of the Brownian movement. Annals d Physik. 1906;19(4):371–381 [7] Einstein A. Zur Theorie der Brownschen Bewegung. Annals d Physik. 1906;324(2):371–
[8] Chandrasekhar S. Stochastic problems in physics and astronomy. Reviews of Modern Physics. 1943;15(1):1–89. Available from: http://prola.aps.org/abstract/RMP/v15/i1/p1\_1.
[9] Chandrasekhar S, Elbert DD. On orthogonal functions which satisfy four boundary conditions.III.TABLES for use in fourier-bessel type expansions. Astrophysical Journal
[10] Leighton D, Acrivos A. The shear-induced migration of particles in concentrated suspensions. Journal of Fluid Mechanics. 1987 Sep;181(1):415. Available from: http://dx.doi.org/
(2):389–399. Available from: http://dx.doi.org/10.1006/jcis.2000.7231.
1905;322(8):549–560. Available from:/journal/112477420/abstract.
381. Available from:/journal/112501299/abstract.
Supplement. 1958;3:453–458.
10.1017/S0022112087002155.
Albert S. Kim<sup>1</sup>
22 Desalination
States of America
References
Presse; 1998.
M. Dekker; 1996.
[23] Lutchmiah K, Verliefde ARD, Roest K, Rietveld LC, Cornelissen ER. Forward osmosis for application in wastewater treatment: A review. Water Research. 2014 Jul;58:179–197. Available from: http://dx.doi.org/10.1016/j.watres.2014.03.045.
[36] Soltanieh M, Gill WN. Review of reverse osmosis membranes and transport models. Chemical Engineering Communications. 1981 Nov;12(4–6):279–363. Available from:
Membrane Thermodynamics for Osmotic Phenomena http://dx.doi.org/10.5772/intechopen.68406 25
[37] Potts DE, Ahlert RC, Wang SS. A critical review of fouling of reverse osmosis membranes. Desalination. 1981 Mar;36(3):235–264. Available from: http://dx.doi.org/10.1016/S0011-
[38] Wijmans JG, Baker RW. The solution-diffusion model: a review. Journal of Membrane Science. 1995 Nov;107(1–2):1–21. Available from: http://dx.doi.org/10.1016/0376-7388(95)
[39] Paul D. Reformulation of the solution-diffusion theory of reverse osmosis. Journal of Membrane Science. 2004 Oct;241(2):371–386. Available from: http://dx.doi.org/10.1016/j.
[40] Fick A. Neue Ausstellung an dem Begriffe des endosmotischen Aequivalentes. Annalen der Physik und Chemie. 1854;168(6):333–335. Available from: http://dx.doi.org/10.1002/
[41] Fick A. On liquid diffusion. Journal of Membrane Science. 1995 Mar;100(1):33–38. Avail-
[42] Khayet M, Matsuura T. Membrane Distillation: Principles and Applications. New York:
[43] Kim AS. A two-interface transport model with pore-size distribution for predicting the performance of direct contact membrane distillation (DCMD). Journal of Membrane
[44] Kim AS. Cylindrical cell model for direct contact membrane distillation (dcmd) of densely packed hollow fibers. Journal of Membrane Science. 2014;455:168–186. Available
[45] Kim AS, Ki SJ, Kim HJ. Research perspective of membrane distillation: Multi-scale and multi-physics phenomena. Desalination and Water Treatment. 2017;58:351–359. Available
[46] Staverman AJ. Non-equilibrium thermodynamics of membrane processes. Transactions of the Faraday Society. 1952;48:176. Available from: http://dx.doi.org/10.1039/
[47] de Groot SR, Mazur P, King AL. Non-equilibrium thermodynamics. American Journal of Physics. 1963 Jul;31(7):558–559. Available from: http://dx.doi.org/10.1119/1.1969680.
[48] De Groot SR, Mazur P. Non-equilibrium Thermodynamics. New York, Dover Publica-
able from: http://dx.doi.org/10.1016/0376-7388(94)00230-V.
from: http://dx.doi.org/10.1016/j.memsci.2013.12.067.
from: http://dx.doi.org/10.5004/dwt.2017.11423.
http://dx.doi.org/10.1080/00986448108910843.
9164(00)88642-7.
memsci.2004.05.026.
andp.18541680612.
Elsevier; 2011.
TF9524800176.
tions, Inc.; 2013.
Science. 2013;428:410–424.
00102-I.
[36] Soltanieh M, Gill WN. Review of reverse osmosis membranes and transport models. Chemical Engineering Communications. 1981 Nov;12(4–6):279–363. Available from: http://dx.doi.org/10.1080/00986448108910843.
[23] Lutchmiah K, Verliefde ARD, Roest K, Rietveld LC, Cornelissen ER. Forward osmosis for application in wastewater treatment: A review. Water Research. 2014 Jul;58:179–197.
[24] Chekli L, Phuntsho S, Kim JE, Kim J, Choi JY, Choi JS, et al. A comprehensive review of hybrid forward osmosis systems: Performance, applications and future prospects. Journal of Membrane Science. 2016 Jan;497:430–449. Available from: http://dx.doi.org/10.1016/j.
[25] Cai Y, Hu XM. A critical review on draw solutes development for forward osmosis. Desalination. 2016 Aug;391:16–29. Available from: http://dx.doi.org/10.1016/j.desal.2016.
[26] Achilli A, Cath TY, Childress AE. Power generation with pressure retarded osmosis: An experimental and theoretical investigation. Journal of Membrane Science. 2009 Nov;343
[27] Achilli A, Childress AE. Pressure retarded osmosis: From the vision of Sidney Loeb to the first prototype installation — Review. Desalination. 2010 Oct;261(3):205–211. Available
[28] Yip NY, Elimelech M. Performance limiting effects in power generation from salinity gradients by pressure retarded osmosis. Environmental Science & Technology. 2011
[29] Helfer F, Lemckert C, Anissimov YG. Osmotic power with pressure retarded osmosis: Theory, performance and trends —A review. Journal of Membrane Science. 2014 Mar;453:
[30] Lonsdale HK, Merten U, Riley RL. Transport properties of cellulose acetate osmotic membranes. Journal of Applied Polymer Science. 1965 Apr;9(4):1341–1362. Available
[31] Landau LD, Lifshitz EM. Statistical physics. Oxford, Pergamon Press, 1991. Translated
[32] Landau LD, Lifshitz EM. Fluid mechanics. Oxford, Pergamon Press (translated from the
[33] Mulder M. Basic Principles of Membrane Technology. Dordrecht, The Netherlands:
[34] Spiegler KS, Kedem O. Thermodynamics of hyperfiltration (reverse osmosis): criteria for efficient membranes. Desalination. 1966 Dec;1(4):311–326. Available from: http://dx.doi.
[35] Lee CH. Theory of reverse osmosis and some other membrane permeation operations. Journal of Applied Polymer Science. 1975 Jan;19(1):83–95. Available from: http://dx.doi.
Dec;45(23):10273–10282. Available from: http://dx.doi.org/10.1021/es203197e.
337–358. Available from: http://dx.doi.org/10.1016/j.memsci.2013.10.053.
(1–2):42–52. Available from: http://dx.doi.org/10.1016/j.memsci.2009.07.006.
from: http://dx.doi.org/10.1016/j.desal.2010.06.017.
from: http://dx.doi.org/10.1002/app.1965.070090413.
from the Russian by J.B. Sykes and M.J. Kearsley.
Russian by J.B. Sykes and W.H. Reid). 1987.
Kluwer Academic Publishers; 1996.
org/10.1016/S0011-9164(00)80018-1.
org/10.1002/app.1975.070190107.
Available from: http://dx.doi.org/10.1016/j.watres.2014.03.045.
memsci.2015.09.041.
03.021.
24 Desalination
[49] Onsager L. Reciprocal relations in irreversible processes. I. Physical Review. 1931 Feb;37 (4):405–426. Available from: http://dx.doi.org/10.1103/PhysRev.37.405.
**Chapter 2**
**Pulsating Flow Effects on Hydrodynamics in a**
A previously developed and validated two-dimensional computational fluid dynamics (CFD) model to study the hydrodynamics in a desalination membrane filled with spacers in zig-zag arrangements has been further developed to include the effects of a pulsating flow with the profile of a heartbeat. Numerical solutions were obtained with Fluent for pulsating laminar flows in channels filled with four different spacers and four lengths of cells. Hydrodynamics was investigated for unsteady state, using a characteristic function of a heartbeat, in order to study the influence of temporal variation in the hydrodynamic behavior. The results show the velocities distribution, streamlines, pressure drop and the wall shear stress on the impermeable wall of the membrane, for Reynolds numbers up to 100. The reduction in the distance between the filaments of the spacers, leads to the appearance of more active recirculation zones that can promote mass transfer and decreasing concentrations layers. On the other hand, this reduction increases the pressure drop and consequently the energy expended in the process. Further, the characteristic function of heartbeat demonstrates promising results, with regard to the energy consumption in the
Keywords: desalination, membrane, reverse osmosis, zig-zag spacers, heartbeat
The planet has not only freshwater, but from the enormous volume of water available on our planet, only 3% is not salty. This resource is limited and finite, vital for the existence of life on
> © 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
**Desalination Membrane Filled with Spacers**
Armando A. Soares, João Silva, Eliseu Monteiro and
Additional information is available at the end of the chapter
process and optimization of the recirculation zones.
earth and for the economic and social development [1].
http://dx.doi.org/10.5772/intechopen.68777
Abel Rouboa
Abstract
flow profile
1. Introduction
[50] Kedem O, Katchalsky A. Thermodynamic analysis of the permeability of biological membranes to non-electrolytes. Biochimica et Biophysica Acta. 1958 Jan;27:229–246. Available from: http://dx.doi.org/10.1016/0006-3002(58)90330-5.
## **Pulsating Flow Effects on Hydrodynamics in a Desalination Membrane Filled with Spacers**
Armando A. Soares, João Silva, Eliseu Monteiro and
## Abel Rouboa
[49] Onsager L. Reciprocal relations in irreversible processes. I. Physical Review. 1931 Feb;37
[50] Kedem O, Katchalsky A. Thermodynamic analysis of the permeability of biological membranes to non-electrolytes. Biochimica et Biophysica Acta. 1958 Jan;27:229–246.
(4):405–426. Available from: http://dx.doi.org/10.1103/PhysRev.37.405.
Available from: http://dx.doi.org/10.1016/0006-3002(58)90330-5.
26 Desalination
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.68777
#### Abstract
A previously developed and validated two-dimensional computational fluid dynamics (CFD) model to study the hydrodynamics in a desalination membrane filled with spacers in zig-zag arrangements has been further developed to include the effects of a pulsating flow with the profile of a heartbeat. Numerical solutions were obtained with Fluent for pulsating laminar flows in channels filled with four different spacers and four lengths of cells. Hydrodynamics was investigated for unsteady state, using a characteristic function of a heartbeat, in order to study the influence of temporal variation in the hydrodynamic behavior. The results show the velocities distribution, streamlines, pressure drop and the wall shear stress on the impermeable wall of the membrane, for Reynolds numbers up to 100. The reduction in the distance between the filaments of the spacers, leads to the appearance of more active recirculation zones that can promote mass transfer and decreasing concentrations layers. On the other hand, this reduction increases the pressure drop and consequently the energy expended in the process. Further, the characteristic function of heartbeat demonstrates promising results, with regard to the energy consumption in the process and optimization of the recirculation zones.
Keywords: desalination, membrane, reverse osmosis, zig-zag spacers, heartbeat flow profile
## 1. Introduction
The planet has not only freshwater, but from the enormous volume of water available on our planet, only 3% is not salty. This resource is limited and finite, vital for the existence of life on earth and for the economic and social development [1].
© 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The exploitation of natural water resources combined with the increase in the world population, the changes in the life style, the inefficient use of water and its contamination, among other factors, leads to scarcity of natural freshwater to respond the necessities of the population [2]. Due to these factors, the demand for potable water, has led to an increasing need to find new alternative sources of drinking water.
concentration gradient between the bulk fluid and the membrane surface, and reducing foul-
Pulsating Flow Effects on Hydrodynamics in a Desalination Membrane Filled with Spacers
http://dx.doi.org/10.5772/intechopen.68777
29
The hydrodynamics inside the feed channel using computational fluid dynamics (CFD) has been investigated awhile. Cao et al. [14] investigated, trough CFD, the effect of three different cylindrical spacers arrangements: spacers touching the same channel wall, spacers touching opposite channel walls and spacers suspended in the channel. The flow simulations, in the spacer-filled channels, showed the distribution of velocity and turbulent kinetic energy as well as the detailed flow patterns. It was found that the presence of the spacer in the membrane is responsible for the local shear stress distribution on the membrane surface and improves the production of the eddy activities. It is also found that spacer configuration, geometry and position in the channel are crucial in the distribution of eddies and high shear stress on the membrane surface. The investigation also suggests that transverse spacer cylinders suspended in the channel are more desirable. Furthermore, decrease in the distance between transverse spacer cylinders can produce more active eddies and reduce the distance between shear stress peaks and consequently improve the mass transfer at the membrane surfaces. Besides that, it was demonstrated that the pressure drop is significantly increased by reduction in the distance between transverse cylinders which increase the operating costs. Schwinger et al. [15] perform CFD studies on how the use of spacers in membrane systems can affect its hydrodynamics. The flow fluid was studied in three different spacer configurations, the cavity, zig-zag, and submerged spacers. Were considered, in a cell, a single filament adjacent to the wall and centered in the channel for Reynolds numbers from 90 to 768 for different values of the length of the mesh and filament diameter. The results obtained show that the flow around the filament increases the wall shear stress and promotes the formation of large recirculation zones behind the filaments. Furthermore, under the same conditions, identical Reynolds number and filament diameter, a single filament adjacent to a membrane wall produced a larger recirculation zone than a single filament in the center of the channel. According to the authors, the utilization of the spacer in a SWM should enhance the mass transfer and reduce the accumulation of solute in the membrane surface while maintaining a low pressure loss along the channel. Saeed et al. [16] used a CFD tool to investigate the flow patterns associated with fluids within the membrane module. The effects on flow patterns through a spacer filled RO membrane with the secondary structure of the membranes (feed spacer filaments) at various angles with the inlet flow are analyzed. The results revealed that the alignment of the filaments in the direction of flow has a great influence on the generation of recirculation streams in channels filled with spacers. The optimization of the orientation of the filaments can lead to desirable recirculation patterns inside the modulus, and therefore, may increase the performance of the membrane. Sousa et al. [17] studied the hydrodynamics in a desalination membrane. CFD techniques were used to study the hydrodynamics of feed channels of a desalination membrane filled with elliptical spacers in zig-zag arrangements and transverse in relation to the flow, for four spacers and four cell lengths. They used Re values between 10 and 300 and the inlet velocity profiles are considered fully developed. The results showed that for Re values above 100, are formed recirculation zones downstream of the spacers. When the spacers are closer, there is a formation of more active recirculation zones. The authors state that these can increase the mass transfer and together with the pressure drop, there may be an
ing phenomenon [7].
Desalination provides a good alternative to reduce the problem of the scarcity of drinking water, once it provides clean water which otherwise would not be accessible for agricultural, industrial and services supply [3, 4].
Membrane systems are widely used in water treatment processes. Depending on the pore size, membrane processes can be classified as microfiltration (MF), ultrafiltration (UF), nanofiltration (NF) and reverse osmosis (RO). The MF and UF are used to remove small colloidal particles and bacteria. For removal of viruses and large molecules such as proteins, the UF is only used. The NF and RO are capable of removing the smallest particles and components, like salts dissolved in water. In the case of RO, it is possible to remove almost all the components dissolved in the water [4].
The RO is the most used membrane separation method. This is applied at chemical, textile, petrochemical, electrochemical and paper industry, as well as in the treatment of municipal wastewaters [5]. But, the major application of RO process is the desalination of brackish and salty water to obtain potable water [6]. This is due to its versatility to operate over a wide range of feed water salinities, lower intake and pumping costs, minimal space requirements and operation at ambient temperatures [7]. Spiral wound modules (SWMs) have been largely used in industrial applications to obtain potable water, due to their low cost and high membrane area to volume ratio [8]. The performance of a SWM is affected by many factors such as leaf geometry (number and dimensions), spacers, fouling propensity, cleaning ability and operating conditions [9]. Despite several benefits of membrane technology over other separation techniques, the phenomena inherent to any membrane separation method are concentration polarization and fouling. As the filtration process occurs though the semipermeable membrane, the rejected impurities begin to accumulate on the membrane surface. After a certain period of time, a concentration layer is formed on the membrane surface due to the solute. This phenomenon, in which, the concentration of solute is higher in the vicinity of the membrane as compared with the observed in the fluid, is known as concentration polarization [10, 11]. This phenomenon can significantly affect the performance of the membrane, for example, the permeate water quality and productively are affected by fouling deposited on the membrane surface. On the other hand, from the literature it is well known that spacers are essential to separate the sheets of membranes in both spiral wound modules based thin rectangular channels (slits). Spacers are also important in the optimization of the mass transfer enhancement and pressure loss minimization [12, 13].
One important way by which the efficiency of module can be enhanced is to modify the hydrodynamics in membrane channel, which can be done by improving feed channel spacers design and using the unsteady state to induce a temporal variation in the feed velocity. This can prevent the accumulation of solutes near the membrane surface, thereby lowering the concentration gradient between the bulk fluid and the membrane surface, and reducing fouling phenomenon [7].
The exploitation of natural water resources combined with the increase in the world population, the changes in the life style, the inefficient use of water and its contamination, among other factors, leads to scarcity of natural freshwater to respond the necessities of the population [2]. Due to these factors, the demand for potable water, has led to an increasing need to
Desalination provides a good alternative to reduce the problem of the scarcity of drinking water, once it provides clean water which otherwise would not be accessible for agricultural,
Membrane systems are widely used in water treatment processes. Depending on the pore size, membrane processes can be classified as microfiltration (MF), ultrafiltration (UF), nanofiltration (NF) and reverse osmosis (RO). The MF and UF are used to remove small colloidal particles and bacteria. For removal of viruses and large molecules such as proteins, the UF is only used. The NF and RO are capable of removing the smallest particles and components, like salts dissolved in water. In the case of RO, it is possible to remove almost all the components dissolved in the
The RO is the most used membrane separation method. This is applied at chemical, textile, petrochemical, electrochemical and paper industry, as well as in the treatment of municipal wastewaters [5]. But, the major application of RO process is the desalination of brackish and salty water to obtain potable water [6]. This is due to its versatility to operate over a wide range of feed water salinities, lower intake and pumping costs, minimal space requirements and operation at ambient temperatures [7]. Spiral wound modules (SWMs) have been largely used in industrial applications to obtain potable water, due to their low cost and high membrane area to volume ratio [8]. The performance of a SWM is affected by many factors such as leaf geometry (number and dimensions), spacers, fouling propensity, cleaning ability and operating conditions [9]. Despite several benefits of membrane technology over other separation techniques, the phenomena inherent to any membrane separation method are concentration polarization and fouling. As the filtration process occurs though the semipermeable membrane, the rejected impurities begin to accumulate on the membrane surface. After a certain period of time, a concentration layer is formed on the membrane surface due to the solute. This phenomenon, in which, the concentration of solute is higher in the vicinity of the membrane as compared with the observed in the fluid, is known as concentration polarization [10, 11]. This phenomenon can significantly affect the performance of the membrane, for example, the permeate water quality and productively are affected by fouling deposited on the membrane surface. On the other hand, from the literature it is well known that spacers are essential to separate the sheets of membranes in both spiral wound modules based thin rectangular channels (slits). Spacers are also important in the optimization of the mass transfer enhance-
One important way by which the efficiency of module can be enhanced is to modify the hydrodynamics in membrane channel, which can be done by improving feed channel spacers design and using the unsteady state to induce a temporal variation in the feed velocity. This can prevent the accumulation of solutes near the membrane surface, thereby lowering the
find new alternative sources of drinking water.
ment and pressure loss minimization [12, 13].
industrial and services supply [3, 4].
water [4].
28 Desalination
The hydrodynamics inside the feed channel using computational fluid dynamics (CFD) has been investigated awhile. Cao et al. [14] investigated, trough CFD, the effect of three different cylindrical spacers arrangements: spacers touching the same channel wall, spacers touching opposite channel walls and spacers suspended in the channel. The flow simulations, in the spacer-filled channels, showed the distribution of velocity and turbulent kinetic energy as well as the detailed flow patterns. It was found that the presence of the spacer in the membrane is responsible for the local shear stress distribution on the membrane surface and improves the production of the eddy activities. It is also found that spacer configuration, geometry and position in the channel are crucial in the distribution of eddies and high shear stress on the membrane surface. The investigation also suggests that transverse spacer cylinders suspended in the channel are more desirable. Furthermore, decrease in the distance between transverse spacer cylinders can produce more active eddies and reduce the distance between shear stress peaks and consequently improve the mass transfer at the membrane surfaces. Besides that, it was demonstrated that the pressure drop is significantly increased by reduction in the distance between transverse cylinders which increase the operating costs. Schwinger et al. [15] perform CFD studies on how the use of spacers in membrane systems can affect its hydrodynamics. The flow fluid was studied in three different spacer configurations, the cavity, zig-zag, and submerged spacers. Were considered, in a cell, a single filament adjacent to the wall and centered in the channel for Reynolds numbers from 90 to 768 for different values of the length of the mesh and filament diameter. The results obtained show that the flow around the filament increases the wall shear stress and promotes the formation of large recirculation zones behind the filaments. Furthermore, under the same conditions, identical Reynolds number and filament diameter, a single filament adjacent to a membrane wall produced a larger recirculation zone than a single filament in the center of the channel. According to the authors, the utilization of the spacer in a SWM should enhance the mass transfer and reduce the accumulation of solute in the membrane surface while maintaining a low pressure loss along the channel. Saeed et al. [16] used a CFD tool to investigate the flow patterns associated with fluids within the membrane module. The effects on flow patterns through a spacer filled RO membrane with the secondary structure of the membranes (feed spacer filaments) at various angles with the inlet flow are analyzed. The results revealed that the alignment of the filaments in the direction of flow has a great influence on the generation of recirculation streams in channels filled with spacers. The optimization of the orientation of the filaments can lead to desirable recirculation patterns inside the modulus, and therefore, may increase the performance of the membrane. Sousa et al. [17] studied the hydrodynamics in a desalination membrane. CFD techniques were used to study the hydrodynamics of feed channels of a desalination membrane filled with elliptical spacers in zig-zag arrangements and transverse in relation to the flow, for four spacers and four cell lengths. They used Re values between 10 and 300 and the inlet velocity profiles are considered fully developed. The results showed that for Re values above 100, are formed recirculation zones downstream of the spacers. When the spacers are closer, there is a formation of more active recirculation zones. The authors state that these can increase the mass transfer and together with the pressure drop, there may be an increase in process efficiency [17]. Amokrane et al. [18] used a model based on CFD which combines the fluid dynamics and the mass transfer in SWMs with zig-zag spacers, to investigate the impact of new spacers design. They affirm that zig-zag arrangement appears to have more advantages in the performance of RO membranes, in comparison, with submerged spacers. The results highlight that the incorporation of new spacer designs, such as elliptic and oval shapes, generates lower pressure drop (Δp) compared to conventional geometries. However, concentration polarization layer thickness, induced by elliptic and oval spacers, is higher compared to that of a circular spacer, while the mass transfer coefficient obtained with circular spacer is higher [18]. Al-Bastaki and Abbas [19] used a cyclic feed flow to improve the performance of membrane modules. The operation in cyclic mode generates flow instabilities, which in turn disturbs the fouling layer and increases the permeation rates. The cyclic operation can, however, need additional devices such as pulsation generator, oscillating pistons, and pneumatic valves. Lau et al. [20] performed a tri-dimensional study of the unsteady flow in order to optimize the spacer mesh angle for the SWM feed spacer. The results show that the presence of unsteady flow can significantly interrupt the development of the concentration of solute in the membrane surface. Filaments with different geometries and different angles may lead to different degrees of disruption in the concentration of the membrane layer. The filaments can be optimized in terms of concentration polarization and power consumption.
In the present investigation, a previously developed and validated two-dimensional CFD model to study the hydrodynamics in a desalination membrane filled with spacers in zig-zag arrangements [17], is further developed to include a characteristic function of a heartbeat within a cell of a semipermeable desalination membrane, in order to study the global and local impact of hydrodynamic behavior on the velocity, pressure drop, and wall shear stress for different spacer arrangements. The numerical approach used allows uncoupling of several phenomena and contributes to a better understanding of the mechanisms through which the shape of the flow profiles can affect membrane performance.
#### 2. Governing equations and boundary conditions
An incompressible Newtonian fluid was used to simulate the bi-dimensional unsteady state flow in a feed channel of a SWM with zig-zag spacers (Figure 1). The governing equations are the continuity and Navier-Stokes equations. For a bi-dimensional flow, the governing equations can be written as:
Continuity equation
$$
\frac{\partial v\_x}{\partial x} + \frac{\partial v\_y}{\partial y} = 0,\tag{1}
$$
and
yy—Momentum equation
2.1. Inlet velocity profiles
∂ ∂x
directions are vx and vy, respectively.
<sup>ð</sup>vxvyÞ þ <sup>∂</sup>
∂y ðv2 yÞ
In the present study, the Reynolds number is defined as
Figure 1. Schematic representation of a membrane cell with elliptical zig-zag filaments.
expected to have a small impact on your flow solution.
¼ � <sup>∂</sup><sup>p</sup> ∂y þ 2μ ∂ ∂y
where x and y are the spatial coordinates, ρ is the fluid density, μ is the dynamic viscosity, p is the static pressure, and g is the gravitational acceleration. The velocity components in x and y
Re <sup>¼</sup> <sup>ρ</sup> v h
In order to obtain a better numerical approach taking into account the flow periodicity, it was necessary to determine the velocity profile at the inlet and outlet of the cell [21]. As far as our knowledge in previous investigations, for example Refs. [8, 17, 22], it was considered a fully developed velocity profile at inlet. However, the velocity profiles between the filaments are not fully developed because there is no enough distance between filaments to allow the full development of the fluid flow. Therefore, for a given average velocity, the inlet velocity profile was obtained using the following procedure: for the value of Re studied, the corresponding average velocity is imposed at the inlet boundary. On the outlet boundary was considered the outflow option present in Fluent. With this condition, it is assumed that there are no diffusive fluxes of all variables at the exit in the flow direction, or a zero diffusion flux at the exit is
For the case of impermeable walls, when the solution is achieved, the boundary condition at the inlet is replaced by the solution obtained for the velocity profile at the outlet boundary. This process was repeated until the differences between the inlet and outlet profiles were less than 0.15%. Similarly, in order to verify the permeability influence on the flow, a study was also done for a semipermeable membrane. To this, the bottom wall was considered permeable, with a constant flow rate of permeate. The value of permeate velocity (vp) was 0.1 mm/s, value taken
where h is the height of the cell and v is the average inlet velocity in a cell, see Figure 1.
∂vy ∂y
Pulsating Flow Effects on Hydrodynamics in a Desalination Membrane Filled with Spacers
þ μ ∂ ∂x
∂vx ∂y þ ∂vy ∂x
http://dx.doi.org/10.5772/intechopen.68777
<sup>μ</sup> <sup>ð</sup>3<sup>Þ</sup>
þ ρgy ð2bÞ
31
ρ ∂vy ∂t þ
xx—Momentum equation
$$\rho \left( \frac{\partial v\_x}{\partial t} + \frac{\partial}{\partial x} (v\_x^2) + \frac{\partial}{\partial y} (v\_x v\_y) \right) = -\frac{\partial p}{\partial x} + 2\mu \frac{\partial}{\partial x} \left( \frac{\partial v\_x}{\partial x} \right) + \mu \frac{\partial}{\partial y} \left( \frac{\partial v\_x}{\partial y} + \frac{\partial v\_y}{\partial x} \right) + \rho g\_x \tag{2a}$$
and
increase in process efficiency [17]. Amokrane et al. [18] used a model based on CFD which combines the fluid dynamics and the mass transfer in SWMs with zig-zag spacers, to investigate the impact of new spacers design. They affirm that zig-zag arrangement appears to have more advantages in the performance of RO membranes, in comparison, with submerged spacers. The results highlight that the incorporation of new spacer designs, such as elliptic and oval shapes, generates lower pressure drop (Δp) compared to conventional geometries. However, concentration polarization layer thickness, induced by elliptic and oval spacers, is higher compared to that of a circular spacer, while the mass transfer coefficient obtained with circular spacer is higher [18]. Al-Bastaki and Abbas [19] used a cyclic feed flow to improve the performance of membrane modules. The operation in cyclic mode generates flow instabilities, which in turn disturbs the fouling layer and increases the permeation rates. The cyclic operation can, however, need additional devices such as pulsation generator, oscillating pistons, and pneumatic valves. Lau et al. [20] performed a tri-dimensional study of the unsteady flow in order to optimize the spacer mesh angle for the SWM feed spacer. The results show that the presence of unsteady flow can significantly interrupt the development of the concentration of solute in the membrane surface. Filaments with different geometries and different angles may lead to different degrees of disruption in the concentration of the membrane layer. The filaments can be optimized in terms of concentration polarization and power consumption.
In the present investigation, a previously developed and validated two-dimensional CFD model to study the hydrodynamics in a desalination membrane filled with spacers in zig-zag arrangements [17], is further developed to include a characteristic function of a heartbeat within a cell of a semipermeable desalination membrane, in order to study the global and local impact of hydrodynamic behavior on the velocity, pressure drop, and wall shear stress for different spacer arrangements. The numerical approach used allows uncoupling of several phenomena and contributes to a better understanding of the mechanisms through which the
An incompressible Newtonian fluid was used to simulate the bi-dimensional unsteady state flow in a feed channel of a SWM with zig-zag spacers (Figure 1). The governing equations are the continuity and Navier-Stokes equations. For a bi-dimensional flow, the governing equa-
<sup>∂</sup><sup>y</sup> <sup>¼</sup> <sup>0</sup>, <sup>ð</sup>1<sup>Þ</sup>
∂vx ∂y þ ∂vy ∂x
þ ρgx ð2aÞ
∂vx ∂x þ ∂vy
¼ � <sup>∂</sup><sup>p</sup> ∂x þ 2μ ∂ ∂x
∂vx ∂x þ μ ∂ ∂y
shape of the flow profiles can affect membrane performance.
2. Governing equations and boundary conditions
tions can be written as:
xx—Momentum equation
∂ ∂x ðv2 <sup>x</sup>Þ þ <sup>∂</sup> ∂y ðvxvyÞ
ρ ∂vx ∂t þ
Continuity equation
30 Desalination
yy—Momentum equation
$$
\rho \left( \frac{\partial v\_y}{\partial t} + \frac{\partial}{\partial x} (v\_x v\_y) + \frac{\partial}{\partial y} (v\_y^2) \right) = -\frac{\partial p}{\partial y} + 2\mu \frac{\partial}{\partial y} \left( \frac{\partial v\_y}{\partial y} \right) + \mu \frac{\partial}{\partial x} \left( \frac{\partial v\_x}{\partial y} + \frac{\partial v\_y}{\partial x} \right) + \rho g\_y \tag{2b}
$$
where x and y are the spatial coordinates, ρ is the fluid density, μ is the dynamic viscosity, p is the static pressure, and g is the gravitational acceleration. The velocity components in x and y directions are vx and vy, respectively.
In the present study, the Reynolds number is defined as
$$Re = \frac{\rho \text{ } \overline{v} \text{ } h}{\mu} \tag{3}$$
where h is the height of the cell and v is the average inlet velocity in a cell, see Figure 1.
Figure 1. Schematic representation of a membrane cell with elliptical zig-zag filaments.
#### 2.1. Inlet velocity profiles
In order to obtain a better numerical approach taking into account the flow periodicity, it was necessary to determine the velocity profile at the inlet and outlet of the cell [21]. As far as our knowledge in previous investigations, for example Refs. [8, 17, 22], it was considered a fully developed velocity profile at inlet. However, the velocity profiles between the filaments are not fully developed because there is no enough distance between filaments to allow the full development of the fluid flow. Therefore, for a given average velocity, the inlet velocity profile was obtained using the following procedure: for the value of Re studied, the corresponding average velocity is imposed at the inlet boundary. On the outlet boundary was considered the outflow option present in Fluent. With this condition, it is assumed that there are no diffusive fluxes of all variables at the exit in the flow direction, or a zero diffusion flux at the exit is expected to have a small impact on your flow solution.
For the case of impermeable walls, when the solution is achieved, the boundary condition at the inlet is replaced by the solution obtained for the velocity profile at the outlet boundary. This process was repeated until the differences between the inlet and outlet profiles were less than 0.15%. Similarly, in order to verify the permeability influence on the flow, a study was also done for a semipermeable membrane. To this, the bottom wall was considered permeable, with a constant flow rate of permeate. The value of permeate velocity (vp) was 0.1 mm/s, value taken from the literature [7, 8]. Since there is permeability, part of the flow is lost through the permeable wall and, thus for the conservation of mass principle be satisfied, the permeate flow rate has to be added to the outflow to obtain the inlet velocity profile. That is,
$$
\upsilon\_y(0, y) = \upsilon\_y(L, y) \tag{4a}
$$
3. Results and discussion
3.1. Inlet velocity profiles
case, therefore they are not shown in Figure 4.
and F4; (e) L = 6 mm and F4; (f) L = 8 mm and F4; (g) L = 10 mm and F4.
each configuration, are the same.
In this study, two-dimensional simulations using different feed channel geometries filled with spacers (different cells lengths and different shape of the elliptical filaments) were performed. We studied the impact on the velocity field, streamlines, average wall shear stress (WSS), and pressure drop for the cells lengths, L = 4, 6, 8, and 10 mm (Figure 3) and height of 0.7 mm. Four
Pulsating Flow Effects on Hydrodynamics in a Desalination Membrane Filled with Spacers
http://dx.doi.org/10.5772/intechopen.68777
33
Since for normal operation of the SWM, the velocity profiles inside the cell cannot be fully developed due to the filaments periodicity, so that the inlet profiles used in this study were determinate by the aforementioned procedure. As example, Figure 4 shows the profiles determined for the four elliptical filaments with L = 4 and Re = 100, for the membrane with impermeable walls. Since permeable velocity is 0.1 mm/s, the inlet velocity profiles are practically unaltered by the permeability of the membrane when compared with the impermeable
Figure 3. Cell filled with zig-zag spacers for (a) L = 4 mm and F1; (b) L = 4 mm and F2; (c) L = 4 mm and F3; (d) L = 4 mm
The width profiles decrease from F1 to F4 corresponding to the increase of flow obstruction by the spacers (Figure 3). Note that for the impermeable walls, the inlet and outlet profiles, for
elliptical filaments configurations (F1, F2, F3, and F4) were also tested.
$$
\upsilon\_x(0, y) = \upsilon\_x(L, y) + \frac{L}{h - h\_f} \upsilon\_p(L, y) \tag{4b}
$$
where the second term on the right-hand side of Eq. (4b) is the velocity profile in the outlet boundary obtained for the case of impermeable walls, when the imposed inlet flow rate is equal to the permeate flow rate.
The time-dependent velocity at the inlet was obtained from Doppler ultrasound images for a heartbeat, in the abdominal aorta, with a period of T = 1.53 s. Figure 2 shows the normalized inlet velocity as function of time, vh(t), obtained from a polynomial fit of the Doppler data. Once given, the normalized velocity of the heartbeat was necessary to make an adjustment thereof, in order to encompass the typical velocities of a desalination membrane. Thus, the inlet velocity profile is given by
$$
\upsilon\_x(0, y, t) = \upsilon\_x(0, y) \cdot \upsilon\_h(t) \tag{5}
$$
Then the normalized velocity vh(t) was multiplied by the velocity profiles, already determined for steady-state flow and so we obtained the inlet velocity profiles used in the present study, Eq. (5). For the impermeable membrane vxð0, yÞ ¼ vxðL, yÞ.
The no-slip condition was applied at the wall boundaries.
Note that in the presence of permeability, it is needed to differentiate inlet and outlet profiles because some fluid is permeated through the permeable wall, see Eq. (4b).
Figure 2. Normalized velocity for a heartbeat with a period T = 1.53 s.
## 3. Results and discussion
from the literature [7, 8]. Since there is permeability, part of the flow is lost through the permeable wall and, thus for the conservation of mass principle be satisfied, the permeate flow
where the second term on the right-hand side of Eq. (4b) is the velocity profile in the outlet boundary obtained for the case of impermeable walls, when the imposed inlet flow rate is
The time-dependent velocity at the inlet was obtained from Doppler ultrasound images for a heartbeat, in the abdominal aorta, with a period of T = 1.53 s. Figure 2 shows the normalized inlet velocity as function of time, vh(t), obtained from a polynomial fit of the Doppler data. Once given, the normalized velocity of the heartbeat was necessary to make an adjustment thereof, in order to encompass the typical velocities of a desalination membrane. Thus, the
Then the normalized velocity vh(t) was multiplied by the velocity profiles, already determined for steady-state flow and so we obtained the inlet velocity profiles used in the present study,
Note that in the presence of permeability, it is needed to differentiate inlet and outlet profiles
h � hf
vyð0, yÞ ¼ vyðL, yÞ ð4aÞ
vxð0, y, tÞ ¼ vxð0, yÞ � vhðtÞ ð5Þ
vpðL, yÞ ð4bÞ
rate has to be added to the outflow to obtain the inlet velocity profile. That is,
equal to the permeate flow rate.
32 Desalination
inlet velocity profile is given by
Eq. (5). For the impermeable membrane vxð0, yÞ ¼ vxðL, yÞ. The no-slip condition was applied at the wall boundaries.
Figure 2. Normalized velocity for a heartbeat with a period T = 1.53 s.
because some fluid is permeated through the permeable wall, see Eq. (4b).
vxð0, yÞ ¼ vxðL, yÞ þ <sup>L</sup>
In this study, two-dimensional simulations using different feed channel geometries filled with spacers (different cells lengths and different shape of the elliptical filaments) were performed. We studied the impact on the velocity field, streamlines, average wall shear stress (WSS), and pressure drop for the cells lengths, L = 4, 6, 8, and 10 mm (Figure 3) and height of 0.7 mm. Four elliptical filaments configurations (F1, F2, F3, and F4) were also tested.
Figure 3. Cell filled with zig-zag spacers for (a) L = 4 mm and F1; (b) L = 4 mm and F2; (c) L = 4 mm and F3; (d) L = 4 mm and F4; (e) L = 6 mm and F4; (f) L = 8 mm and F4; (g) L = 10 mm and F4.
#### 3.1. Inlet velocity profiles
Since for normal operation of the SWM, the velocity profiles inside the cell cannot be fully developed due to the filaments periodicity, so that the inlet profiles used in this study were determinate by the aforementioned procedure. As example, Figure 4 shows the profiles determined for the four elliptical filaments with L = 4 and Re = 100, for the membrane with impermeable walls. Since permeable velocity is 0.1 mm/s, the inlet velocity profiles are practically unaltered by the permeability of the membrane when compared with the impermeable case, therefore they are not shown in Figure 4.
The width profiles decrease from F1 to F4 corresponding to the increase of flow obstruction by the spacers (Figure 3). Note that for the impermeable walls, the inlet and outlet profiles, for each configuration, are the same.
Figure 4. Inlet velocity profiles for F1, F2, F3, and F4, with L = 4 mm and Re = 100.
#### 3.2. Velocity field
Figure 5 illustrates the velocity field for Re = 100 and L = 4 mm when F1 and F4 are used for both impermeable and semipermeable cases. A comparison between impermeable and semipermeable cases shows that the maximum velocity decreases by 0.2 and 3.1% for F1 and F4, respectively. This difference can be explained by the smaller useful area of permeable wall for F1 spacer, see Ref. [21]. Note that the average velocity in all cell is higher for F1 than for F4 because the inlet area is bigger for F1 spacer, however the average inlet velocity is the same for the two spacers.
the t = T/8 and its value is 0.297 m/s. For both L = 10 mm and L = 4 mm, from T/2, looking only at the Figures 5 and 6, it appears that from that instant, the velocity remains constant. A more detailed analysis indicates that there are fluctuations, although small, and for t = 5T/8 is when
Figure 5. Velocity field during a period at t = T/8, T/4, 3T/8, T/2, 5T/8, 3T/4, 7T/8, and T, for L= 4 mm: (a) F1 and vp = 0 mm/
Pulsating Flow Effects on Hydrodynamics in a Desalination Membrane Filled with Spacers
http://dx.doi.org/10.5772/intechopen.68777
35
Figure 7 shows the average velocity inside the cell for both cases, impermeable and semipermeable membrane, for a cell with length L = 4 mm, for all instants and spacers analyzed. The results presented are in concordance with Figure 5. The average velocity decreases from F1 to F4. It is also noteworthy that the highest average velocity occurs at instant T/8 undergoing a decrement up to the minimum value of average velocity at 5T/8. After 5T/8, the velocity remains almost constant but with small fluctuations in concordance with inlet velocity
A comparison between average velocities for impermeable and semipermeable cases also shows that the biggest difference is of 4.49% and occurs at 3T/8, for F4. The smallest difference between the two cases is verified for F2 at t = T/2, with a velocity decrease of 0.40%. On the other hand, for L = 10 mm, the corresponding biggest difference is of 4.13% for F4 at T/4 and
the velocity takes the lowest value, equal to 0.213 m/s.
s; (b) F1 and vp = 0.1 mm/s; (c) F4 and vp = 0 mm/s; (d) F4 and vp = 0.1 mm/s.
smallest difference is of 1.71% for F1 at 7T/4, see Figure 8.
3.3. Average velocity inside the cell
(Figures 2, 5, and 6).
From Figure 5, it can be seen that the well-known velocity behavior, that is, for each instant, the maximum velocity inside the membrane is located in the zones between the filament and the adjacent wall, for example Refs. [17, 23, 24]. This behavior is explained by the channel narrowing due to the spacers, with the consequent velocity increase to ensure the mass conservation. From general analysis of Figure 5, we can see that the maximum of velocity occurs at T/8. From this instant, velocity decreases until reaches approximately half of the period (T/2) and after that remains almost constant until the start of a new cycle. A more detailed analysis reveals that the permeability has a very little impact in the velocity. For instance, at 5T/8, velocity achieves the minimum value of 0.213 m/s for F1 to vp = 0 mm/s while for vp = 0.1 mm/s is 0.212 m/s. For F4, this value is 0.214 m/s for vp = 0 mm/s and 0.208 m/s for vp = 0.1 mm/s. Figure 6 shows the velocity field for F1, L = 10 mm and impermeable membrane.
With increasing distance between the spacers, it is verified that for each instant, the average velocity within the cell decreases with increases of L, this observation are in line with previous studies (e.g. Ref. [17]). On the other hand, as in previous cases, the maximum velocity occurs at Pulsating Flow Effects on Hydrodynamics in a Desalination Membrane Filled with Spacers http://dx.doi.org/10.5772/intechopen.68777 35
Figure 5. Velocity field during a period at t = T/8, T/4, 3T/8, T/2, 5T/8, 3T/4, 7T/8, and T, for L= 4 mm: (a) F1 and vp = 0 mm/ s; (b) F1 and vp = 0.1 mm/s; (c) F4 and vp = 0 mm/s; (d) F4 and vp = 0.1 mm/s.
the t = T/8 and its value is 0.297 m/s. For both L = 10 mm and L = 4 mm, from T/2, looking only at the Figures 5 and 6, it appears that from that instant, the velocity remains constant. A more detailed analysis indicates that there are fluctuations, although small, and for t = 5T/8 is when the velocity takes the lowest value, equal to 0.213 m/s.
#### 3.3. Average velocity inside the cell
3.2. Velocity field
34 Desalination
the two spacers.
membrane.
Figure 5 illustrates the velocity field for Re = 100 and L = 4 mm when F1 and F4 are used for both impermeable and semipermeable cases. A comparison between impermeable and semipermeable cases shows that the maximum velocity decreases by 0.2 and 3.1% for F1 and F4, respectively. This difference can be explained by the smaller useful area of permeable wall for F1 spacer, see Ref. [21]. Note that the average velocity in all cell is higher for F1 than for F4 because the inlet area is bigger for F1 spacer, however the average inlet velocity is the same for
Figure 4. Inlet velocity profiles for F1, F2, F3, and F4, with L = 4 mm and Re = 100.
From Figure 5, it can be seen that the well-known velocity behavior, that is, for each instant, the maximum velocity inside the membrane is located in the zones between the filament and the adjacent wall, for example Refs. [17, 23, 24]. This behavior is explained by the channel narrowing due to the spacers, with the consequent velocity increase to ensure the mass conservation. From general analysis of Figure 5, we can see that the maximum of velocity occurs at T/8. From this instant, velocity decreases until reaches approximately half of the period (T/2) and after that remains almost constant until the start of a new cycle. A more detailed analysis reveals that the permeability has a very little impact in the velocity. For instance, at 5T/8, velocity achieves the minimum value of 0.213 m/s for F1 to vp = 0 mm/s while for vp = 0.1 mm/s is 0.212 m/s. For F4, this value is 0.214 m/s for vp = 0 mm/s and 0.208 m/s for vp = 0.1 mm/s. Figure 6 shows the velocity field for F1, L = 10 mm and impermeable
With increasing distance between the spacers, it is verified that for each instant, the average velocity within the cell decreases with increases of L, this observation are in line with previous studies (e.g. Ref. [17]). On the other hand, as in previous cases, the maximum velocity occurs at Figure 7 shows the average velocity inside the cell for both cases, impermeable and semipermeable membrane, for a cell with length L = 4 mm, for all instants and spacers analyzed. The results presented are in concordance with Figure 5. The average velocity decreases from F1 to F4. It is also noteworthy that the highest average velocity occurs at instant T/8 undergoing a decrement up to the minimum value of average velocity at 5T/8. After 5T/8, the velocity remains almost constant but with small fluctuations in concordance with inlet velocity (Figures 2, 5, and 6).
A comparison between average velocities for impermeable and semipermeable cases also shows that the biggest difference is of 4.49% and occurs at 3T/8, for F4. The smallest difference between the two cases is verified for F2 at t = T/2, with a velocity decrease of 0.40%. On the other hand, for L = 10 mm, the corresponding biggest difference is of 4.13% for F4 at T/4 and smallest difference is of 1.71% for F1 at 7T/4, see Figure 8.
3.4. Streamlines
results for impermeable case are shown.
Figures 9 and 10 show the streamlines for L = 4 mm and the four types of spacers. Since the streamlines' differences for impermeable and semipermeable cases are insignificant, only the
Pulsating Flow Effects on Hydrodynamics in a Desalination Membrane Filled with Spacers
http://dx.doi.org/10.5772/intechopen.68777
37
Figure 8. Average velocity during a period for L = 10 mm with vp = 0 and 0.1 mm/s, for all spacers.
The recirculation zones are visible in all instants analyzed for all studied cell geometries. These zones are generally located downstream of the filaments, however, under specific conditions, small upstream recirculation zones can occur as shown in Figure 9 (a) at t = T/8 and Figure 9 (d) at t = T/8, T/4, and 5T/8. At t = T/8, the eddies occupy the largest area, since this instant is when the average velocity have the maximum value (Figures 7 and 9). From T/8 until T/2, the flow is in deceleration; consequently, the velocity decreases and the recirculation zones are reduced. After t = T/2, the velocity suffers small oscillations. These oscillations are responsible for the oscillations of the eddies length. This behavior is most easily seen in Figure 9 (a). When t = T, it starts a new cycle and the fluid enters into a new stage of acceleration, leading to an increase in velocity which in turn causes a new increase in the recirculation zones. That behavior was observed in all geometrical configurations represented in Figures 9 and 10. Comparing the behavior of the different spacers, it should be noted that there is an increase in the area of eddies from F1 to F4. This increase can be explained by the shape of the spacer filaments. Note that F4 is the filament with biggest perpendicular area to the flow. In the present study, the distance between filaments is not sufficiently small to inhibit the growth of eddies. Figure 10 shows the effects of the
cell length, L, on the eddy length for the F4 configuration and maximum average velocity.
better performance, which are in line with literature, for example, Saeed et al. [16].
We can infer that the length between filaments has no influence on the eddies formation. Therefore, for all conditions studied, L does not affect the length of the recirculation zones.
The variation of the length eddies during a pulsatile flow modeled by a heartbeat shape combined with spacer arrangement may lead to better performance of the desalination membrane. Thus, the formation of desirable flow patterns within the membrane can lead to its
Figure 6. Velocity field during a period at t = T/8, T/4, 3T/8, T/2, 5T/8, 3T/4, 7T/8, and T, for L = 10 mm, F1 and vp = 0 mm/s.
Figure 7. Average velocity during a period for L = 4 mm with vp = 0 and 0.1 mm/s, for all spacers.
Pulsating Flow Effects on Hydrodynamics in a Desalination Membrane Filled with Spacers http://dx.doi.org/10.5772/intechopen.68777 37
Figure 8. Average velocity during a period for L = 10 mm with vp = 0 and 0.1 mm/s, for all spacers.
#### 3.4. Streamlines
Figure 6. Velocity field during a period at t = T/8, T/4, 3T/8, T/2, 5T/8, 3T/4, 7T/8, and T, for L = 10 mm, F1 and vp = 0 mm/s.
36 Desalination
Figure 7. Average velocity during a period for L = 4 mm with vp = 0 and 0.1 mm/s, for all spacers.
Figures 9 and 10 show the streamlines for L = 4 mm and the four types of spacers. Since the streamlines' differences for impermeable and semipermeable cases are insignificant, only the results for impermeable case are shown.
The recirculation zones are visible in all instants analyzed for all studied cell geometries. These zones are generally located downstream of the filaments, however, under specific conditions, small upstream recirculation zones can occur as shown in Figure 9 (a) at t = T/8 and Figure 9 (d) at t = T/8, T/4, and 5T/8. At t = T/8, the eddies occupy the largest area, since this instant is when the average velocity have the maximum value (Figures 7 and 9). From T/8 until T/2, the flow is in deceleration; consequently, the velocity decreases and the recirculation zones are reduced. After t = T/2, the velocity suffers small oscillations. These oscillations are responsible for the oscillations of the eddies length. This behavior is most easily seen in Figure 9 (a). When t = T, it starts a new cycle and the fluid enters into a new stage of acceleration, leading to an increase in velocity which in turn causes a new increase in the recirculation zones. That behavior was observed in all geometrical configurations represented in Figures 9 and 10. Comparing the behavior of the different spacers, it should be noted that there is an increase in the area of eddies from F1 to F4. This increase can be explained by the shape of the spacer filaments. Note that F4 is the filament with biggest perpendicular area to the flow. In the present study, the distance between filaments is not sufficiently small to inhibit the growth of eddies. Figure 10 shows the effects of the cell length, L, on the eddy length for the F4 configuration and maximum average velocity.
We can infer that the length between filaments has no influence on the eddies formation. Therefore, for all conditions studied, L does not affect the length of the recirculation zones.
The variation of the length eddies during a pulsatile flow modeled by a heartbeat shape combined with spacer arrangement may lead to better performance of the desalination membrane. Thus, the formation of desirable flow patterns within the membrane can lead to its better performance, which are in line with literature, for example, Saeed et al. [16].
to the fluctuation in Δp/L until t = T, when starts a new cycle which will lead to a further increase in Δp/L. However, Δp/L has a reverse behavior to that observed for velocity, i.e., Δp/L increases from F1 to F4. This increase can be explained by the geometric shape of the spacer filaments, that is, F4 is the spacer with bigger transverse area to the flow and consequently have the geometry with higher resistance to the flow, for that reason is needed a bigger spent
Pulsating Flow Effects on Hydrodynamics in a Desalination Membrane Filled with Spacers
http://dx.doi.org/10.5772/intechopen.68777
39
On the other hand, vp appears to have a small influence in Δp/L. The higher effect is verified for F4, when t = T/4, with a difference between the impermeable and semipermeable cases of 7.24%. The smallest influence is verified for F2, when t = 5T/8, with a difference of 2.45%
Figure 12 shows the pressure drop per unit of length (Δp/L) in a cell of the desalination membrane under impermeable and semipermeable (vp = 0.1 mm/s) conditions for all spacers. From Figure 12, it is possible to infer that with the increase of the distance between the
of energy to overcome the resistance forces, due to the higher value of Δp/L.
Figure 12. Pressure drop per unit length for L = 10 mm with vp = 0 and 0.1 mm/s, for all spacers.
Figure 11. Pressure drop per unit length for L = 4 mm with vp = 0 and 0.1 mm/s, for all spacers.
between the two cases under study.
Figure 9. Streamlines during a period T at t = T/8, T/4, 3T/8, T/2, 5T/8, 3T/4, 7T/8 and T for vp = 0 mm/s and L = 4 mm: (a) F1; (b) F2; (c) F3 and (d) F4.
Figure 10. Streamlines at t = T/8 for F4, vp = 0 mm/s and all values of L.
#### 3.5. Pressure drop per length unit
Figure 11 shows the pressure drop per unit of length (Δp/L) in a cell membrane of desalination to the both cases, impermeable and semipermeable (vp = 0.1 mm/s) for all spacers and L = 4 mm.
The maximum value of Δp/L occurs at T/8 when the average velocity has its higher value. From there, the flow goes to a deceleration phase, and thus, there is a decrease in Δp/L until t = 5T/8. As already mentioned, from that moment, there are small variations in the velocity, which lead to the fluctuation in Δp/L until t = T, when starts a new cycle which will lead to a further increase in Δp/L. However, Δp/L has a reverse behavior to that observed for velocity, i.e., Δp/L increases from F1 to F4. This increase can be explained by the geometric shape of the spacer filaments, that is, F4 is the spacer with bigger transverse area to the flow and consequently have the geometry with higher resistance to the flow, for that reason is needed a bigger spent of energy to overcome the resistance forces, due to the higher value of Δp/L.
On the other hand, vp appears to have a small influence in Δp/L. The higher effect is verified for F4, when t = T/4, with a difference between the impermeable and semipermeable cases of 7.24%. The smallest influence is verified for F2, when t = 5T/8, with a difference of 2.45% between the two cases under study.
Figure 12 shows the pressure drop per unit of length (Δp/L) in a cell of the desalination membrane under impermeable and semipermeable (vp = 0.1 mm/s) conditions for all spacers. From Figure 12, it is possible to infer that with the increase of the distance between the
Figure 11. Pressure drop per unit length for L = 4 mm with vp = 0 and 0.1 mm/s, for all spacers.
Figure 12. Pressure drop per unit length for L = 10 mm with vp = 0 and 0.1 mm/s, for all spacers.
3.5. Pressure drop per length unit
Figure 10. Streamlines at t = T/8 for F4, vp = 0 mm/s and all values of L.
F1; (b) F2; (c) F3 and (d) F4.
38 Desalination
Figure 11 shows the pressure drop per unit of length (Δp/L) in a cell membrane of desalination to the both cases, impermeable and semipermeable (vp = 0.1 mm/s) for all spacers and L = 4 mm. The maximum value of Δp/L occurs at T/8 when the average velocity has its higher value. From there, the flow goes to a deceleration phase, and thus, there is a decrease in Δp/L until t = 5T/8. As already mentioned, from that moment, there are small variations in the velocity, which lead
Figure 9. Streamlines during a period T at t = T/8, T/4, 3T/8, T/2, 5T/8, 3T/4, 7T/8 and T for vp = 0 mm/s and L = 4 mm: (a)
filaments there is a decrease in Δp/L. This phenomenon can be explained by the increased distance between filaments, that is, as the spacers are further apart, these offer a smaller resistance to the flow which is reflected in a smaller value of Δp/L (see e.g. Ref. [17]).
WSS/L only decreases from F1 to F3. The maximum values are verified at t = T/8, when the average velocity reaches its maximum values (Figure 5). From that instant WSS/L decreases in concordance with the average velocity. From T/2, the wall shear stress also presents small
Pulsating Flow Effects on Hydrodynamics in a Desalination Membrane Filled with Spacers
http://dx.doi.org/10.5772/intechopen.68777
41
For studied conditions, vp appears to have small influence in WSS/L. The higher effect is verified for F4, at t = T/2, with a difference between the impermeable and semipermeable cases of 5.59%. The smaller influence occurs for F2, at t = T/2, with a difference between the two cases
Figure 14 presents, for L = 10 mm, the average wall shear stress per unit length in the upper wall (impermeable wall) of a desalination membrane to the impermeable and semipermeable cases for conditions studied. In general, comparing Figures 13 and 14, the behavior of WSS/L is similar, with the exception of that for cell length (L = 10 mm), both impermeable and semipermeable cases show the same behavior of WSS/L with the type of spacers filaments, that is, for
The WSS/L for L = 10 mm also shows that permeability causes a decrease in WSS/L. The higher effect is verified for F4, at t = T/8, with a difference between the impermeable and semipermeable cases of 4.80%. The lesser influence is verified for F1, at t = 5T/8, with a difference between
Comparing the results obtained by Cao et. al. [14] and Sousa et al. [17] with the present results, one can infer that the use of inlet flux with heartbeat profiles can benefit the membrane performance because the concentration polarization may be reduced, improving the mass transfer, once it dynamically changes the boundary layer concentrations and prevents clogging
Figure 14. Average wall shear stress per unit length for L = 10 mm with vp = 0 and 0.1 mm/s, for all spacers.
all instant, analyzed WSS/L decreases from F1 to F4.
fluctuations.
of 0.57%.
the two cases of 2.30%.
of the membrane surface.
The behaviors and conclusions seen in Figure 11, on the pressure drop per unit length, are also observed for this case, where L = 10 mm and for the other values of L not shown here. The main distinction between Figures 11 and 12, which is worth noting, is the difference in the values of Δp/L for the different filaments. For instance for T/8 and L = 4 mm, the difference of Δp/L between F1 and F4 is about five times higher than that for L = 10 mm, under the same conditions.
The results also show that vp have a small influence on Δp/L. The differences observed are due to the shape of the filaments. The higher relative difference is verified for F2, when t = T/2, with a difference between the impermeable and semipermeable cases of 4.95%. The lowest relative difference is verified for F4, at t = 5T/8, with a difference of 1.01% between the two cases under study.
In general, an energy gain is expected when an inlet heartbeat function is used, compared to sinusoidal functions because in heartbeat function the flow keeps small fluctuations, at least, during half cycle, so in this case the pressure gradients developed during a cycle are smaller than that for a sinusoidal function.
#### 3.6. Wall shear stresses per unit length
Figure 13 presents, for L = 4 mm, the average wall shear stress per unit length (WSS/L) in the upper wall (impermeable wall) of a desalination membrane for impermeable (vp = 0 mm/s) and semipermeable cases (vp = 0.1 mm/s). From Figure 13, one can see that for vp = 0.1 mm/s, the values of WSS/L decrease from F1 to F4, at all instants analyzed. For the impermeable case,
Figure 13. Average wall shear stress per unit length for L = 4 mm with vp = 0 and 0.1 mm/s, for all spacers.
WSS/L only decreases from F1 to F3. The maximum values are verified at t = T/8, when the average velocity reaches its maximum values (Figure 5). From that instant WSS/L decreases in concordance with the average velocity. From T/2, the wall shear stress also presents small fluctuations.
filaments there is a decrease in Δp/L. This phenomenon can be explained by the increased distance between filaments, that is, as the spacers are further apart, these offer a smaller
The behaviors and conclusions seen in Figure 11, on the pressure drop per unit length, are also observed for this case, where L = 10 mm and for the other values of L not shown here. The main distinction between Figures 11 and 12, which is worth noting, is the difference in the values of Δp/L for the different filaments. For instance for T/8 and L = 4 mm, the difference of Δp/L between F1 and F4 is about five times higher than that for L = 10 mm, under the same
The results also show that vp have a small influence on Δp/L. The differences observed are due to the shape of the filaments. The higher relative difference is verified for F2, when t = T/2, with a difference between the impermeable and semipermeable cases of 4.95%. The lowest relative difference is verified for F4, at t = 5T/8, with a difference of 1.01% between the two cases under
In general, an energy gain is expected when an inlet heartbeat function is used, compared to sinusoidal functions because in heartbeat function the flow keeps small fluctuations, at least, during half cycle, so in this case the pressure gradients developed during a cycle are smaller
Figure 13 presents, for L = 4 mm, the average wall shear stress per unit length (WSS/L) in the upper wall (impermeable wall) of a desalination membrane for impermeable (vp = 0 mm/s) and semipermeable cases (vp = 0.1 mm/s). From Figure 13, one can see that for vp = 0.1 mm/s, the values of WSS/L decrease from F1 to F4, at all instants analyzed. For the impermeable case,
Figure 13. Average wall shear stress per unit length for L = 4 mm with vp = 0 and 0.1 mm/s, for all spacers.
resistance to the flow which is reflected in a smaller value of Δp/L (see e.g. Ref. [17]).
conditions.
40 Desalination
study.
than that for a sinusoidal function.
3.6. Wall shear stresses per unit length
For studied conditions, vp appears to have small influence in WSS/L. The higher effect is verified for F4, at t = T/2, with a difference between the impermeable and semipermeable cases of 5.59%. The smaller influence occurs for F2, at t = T/2, with a difference between the two cases of 0.57%.
Figure 14 presents, for L = 10 mm, the average wall shear stress per unit length in the upper wall (impermeable wall) of a desalination membrane to the impermeable and semipermeable cases for conditions studied. In general, comparing Figures 13 and 14, the behavior of WSS/L is similar, with the exception of that for cell length (L = 10 mm), both impermeable and semipermeable cases show the same behavior of WSS/L with the type of spacers filaments, that is, for all instant, analyzed WSS/L decreases from F1 to F4.
The WSS/L for L = 10 mm also shows that permeability causes a decrease in WSS/L. The higher effect is verified for F4, at t = T/8, with a difference between the impermeable and semipermeable cases of 4.80%. The lesser influence is verified for F1, at t = 5T/8, with a difference between the two cases of 2.30%.
Comparing the results obtained by Cao et. al. [14] and Sousa et al. [17] with the present results, one can infer that the use of inlet flux with heartbeat profiles can benefit the membrane performance because the concentration polarization may be reduced, improving the mass transfer, once it dynamically changes the boundary layer concentrations and prevents clogging of the membrane surface.
Figure 14. Average wall shear stress per unit length for L = 10 mm with vp = 0 and 0.1 mm/s, for all spacers.
### 4. Conclusions
In this chapter, CFD techniques were used to study the hydrodynamics inside the feeding channel of a semipermeable membrane in spiral wound configuration. The main goal of this research was to investigate the effects of a pulsating flow with a profile of a heartbeat on the hydrodynamics of feed channels of a desalination membrane filled with spacers in zig-zag arrangements and transverse to the flow. Both membrane impermeable and semipermeable cases were analyzed. For the semipermeable case, we considered the membrane bottom wall with a typical permeation rate of 0.1 mm/s. The comparison between the two cases shows that the permeability has no significant influence on hydrodynamics. On the other hand, the permeability implies that there is a slight decrease in average velocity, Δp and WSS.
Due to the existence of the filaments in the cell, the velocity profiles are not fully developed. Thus, depending on spacer type (F1, F2, F3, or F4) and cell length, there is a different characteristic velocity profile. Analyzing the obtained velocity fields is possible to confirm that the maximum velocity occurs in the membrane region between the filament and the opposite wall. For F3 and F4, the shortest distance between consecutive filaments (L = 4 mm) provides eddies that sweep practically the entire membrane wall downstream of the filament, at t = T/8. The imposed eddies fluctuation by the pulsatile inflow can promote a reduction in the concentration polarization effect and reduce the concentration of solute at the membrane surface, which consequently increases the mass transfer. The reduction in the distance between the filaments increases WSS/L. The fluctuation in WSS/L can dynamically change the boundary layer concentrations and prevent clogging of the membrane surface. On the other hand, decreasing the distance between the filaments increases Δp/L and this should increase the energy consumption. Thus, a combination of the distance between filaments, pulsatile inflow and Δp/L is important to optimize separation process.
Author details
**Acknowledgements**
UID/SEM/04252/2013
Greek letters
Armando A. Soares1,2, João Silva1,2, Eliseu Monteiro2,3 and Abel Rouboa1,2,4\*
This work was supported by the Portuguese Foundation for Science and Technology under
x Longitudinal rectangular coordinate (m) y Transverse rectangular coordinate (m)
ρ Fluid density (kg m�<sup>3</sup>
µ Dynamic viscosity (Pa s)
Pulsating Flow Effects on Hydrodynamics in a Desalination Membrane Filled with Spacers 17
http://dx.doi.org/10.5772/intechopen.68777
43
Pulsating Flow Effects on Hydrodynamics in a Desalination Membrane Filled with Spacers
)
1 UTAD, University of Trás-os-Montes and Alto Douro, Vila Real, Portugal
3 C3i - Interdisciplinary Center for Research and Innovation, Polytechnic Institute of
4 Mechanical Engineering and Mechanics Department, University of Pennsylvania,
nation. 2006;196(1-3):1–12. DOI: 10.1016/j.desal.2005.12.009
2005;182(1):111–122. DOI: 10.1016/j.desal.2005.03.011
[1] Tsiourtis NX. Desalination and the environment. Desalination. 2001;141(3):223–236. DOI:
[2] Baalousha H. Desalination status in the Gaza Strip and its environmental impact. Desali-
[3] Marcovecchio MG, Mussati SF, Aguirre PA, Nicolás JS. Optimization of hybrid desalination processes including multi stage flash and reverse osmosis systems. Desalination.
[4] Fritzmann C, Löwenberg J, Wintgens T, Melin T. State-of-the-art of reverse osmosis desalination. Desalination. 2007;216(1-3):1–76. DOI: 10.1016/j.desal.2006.12.009
[5] Bódalo-Santoyo A, Gomez-Carrasco JL, Gomez-Gomez E, Máximo-Martín MF, Hidalgo-Montesinos AM. Spiral-wound membrane reverse osmosis and the treatment of industrial effluents. Desalination. 2004;160(2):151–158. DOI: 10.1016/S0011-9164(04)90005-7
\*Address all correspondence to: rouboa@utad.pt
.
Portalegre, Portalegre, Portugal
Philadelphia, Pennsylvania, USA
10.1016/s0011-9164(01)85001-3
References
2 CIENER-INEGI/University of Porto, Porto, Portugal
In summary, this study suggests that the inter-filaments length combined with the flow variation characteristic of a heartbeat can control the development of concentration polarization, and thus reduce the probability of fouling and energy consumption in the process, through an optimization of the recirculation zones.
#### Nomenclature
## **Acknowledgements**
This work was supported by the Portuguese Foundation for Science and Technology under UID/SEM/04252/2013 .
## Author details
4. Conclusions
16 Desalination
42 Desalination
In this chapter, CFD techniques were used to study the hydrodynamics inside the feeding channel of a semipermeable membrane in spiral wound configuration. The main goal of this research was to investigate the effects of a pulsating flow with a profile of a heartbeat on the hydrodynamics of feed channels of a desalination membrane filled with spacers in zig-zag arrangements and transverse to the flow. Both membrane impermeable and semipermeable cases were analyzed. For the semipermeable case, we considered the membrane bottom wall with a typical permeation rate of 0.1 mm/s. The comparison between the two cases shows that the permeability has no significant influence on hydrodynamics. On the other hand, the
Due to the existence of the filaments in the cell, the velocity profiles are not fully developed. Thus, depending on spacer type (F1, F2, F3, or F4) and cell length, there is a different characteristic velocity profile. Analyzing the obtained velocity fields is possible to confirm that the maximum velocity occurs in the membrane region between the filament and the opposite wall. For F3 and F4, the shortest distance between consecutive filaments (L = 4 mm) provides eddies that sweep practically the entire membrane wall downstream of the filament, at t = T/8. The imposed eddies fluctuation by the pulsatile inflow can promote a reduction in the concentration polarization effect and reduce the concentration of solute at the membrane surface, which consequently increases the mass transfer. The reduction in the distance between the filaments increases WSS/L. The fluctuation in WSS/L can dynamically change the boundary layer concentrations and prevent clogging of the membrane surface. On the other hand, decreasing the distance between the filaments increases Δp/L and this should increase the energy consumption. Thus, a combination of the distance
permeability implies that there is a slight decrease in average velocity, Δp and WSS.
between filaments, pulsatile inflow and Δp/L is important to optimize separation process.
optimization of the recirculation zones.
F Filament h Cell height (m) hf Filament height (m)
L Cell length (m) p Static pressure (Pa) Δp Pressure drop (Pa) Re Reynolds number
g Gravitational acceleration (m s�<sup>2</sup>
v Average inlet velocity (m s�<sup>1</sup>
vp Permeate velocity (m s�<sup>1</sup>
vh(t) Normalized inlet velocity
vx Velocity component in x-direction (m s�<sup>1</sup>
vy Velocity component in y-direction (m s�<sup>1</sup>
Nomenclature
In summary, this study suggests that the inter-filaments length combined with the flow variation characteristic of a heartbeat can control the development of concentration polarization, and thus reduce the probability of fouling and energy consumption in the process, through an
)
)
)
)
)
Armando A. Soares1,2, João Silva1,2, Eliseu Monteiro2,3 and Abel Rouboa1,2,4\*
\*Address all correspondence to: rouboa@utad.pt
1 UTAD, University of Trás-os-Montes and Alto Douro, Vila Real, Portugal
2 CIENER-INEGI/University of Porto, Porto, Portugal
3 C3i - Interdisciplinary Center for Research and Innovation, Polytechnic Institute of Portalegre, Portalegre, Portugal
4 Mechanical Engineering and Mechanics Department, University of Pennsylvania, Philadelphia, Pennsylvania, USA
### References
[6] Wangnick K. IDAWorldwide Desalting Plants Inventory. Report No. 17 ed. Gnarrenburg, Germany: Wangnick Consulting GMBH; 2002
[19] Al-Bastaki N, Abbas A. Use of fluid instabilities to enhance membrane performance: a review. Desalination. 2001;136(1-3):255–262. DOI: 10.1016/S0011-9164(01)00188-6
Pulsating Flow Effects on Hydrodynamics in a Desalination Membrane Filled with Spacers
http://dx.doi.org/10.5772/intechopen.68777
45
[20] Lau KK, Bakar AMZ, Ahmad AL, Murugesan T. Feed spacer mesh angle: 3D modeling, simulation and optimization based on unsteady hydrodynamic in spiral wound membrane channel. Journal of Membrane Science. 2009;343(1):16–33. DOI: 10.1016/j.
[21] Silva JC, Soares AA, Rouboa A. A numerical study of the hydrodynamics in feed channels of spiral-wound membrane modules. MEFTE. 2014; 227–232. In: APMTAC, editor. MEFTE 2014—V Conferência Nacional de Mecânica dos Fluidos, Termodinâmica e
[22] Subramani A, Kim S, Hoek E. Pressure, flow, and concentration profiles in open and spacer-filled membrane channels. Journal of Membrane Science. 2006;277(1):7–17. DOI:
[23] Fimbres-Weihs GA, Wiley DE. Review of 3D CFD modeling of flow and mass transfer in narrow spacer-filled channels in membrane modules. Chemical Engineering and Processing: Process Intensification. 2010;49(7):759–781. DOI: 10.1016/j.cep.2010.01.007 [24] Radu AI, Bergwerff L, Loosdrecht MCM, Picioreanu C. A two-dimensional mechanistic model for scaling in spiral wound membrane systems. Chemical Engineering Journal.
Energia; 11-12 Set 2014; Porto, Portugal. Porto: FEUP; 2014. p. 227–232
memsci.2009.07.001
10.1016/j.memsci.2005.10.021
2014;241(1):77–91. DOI: 10.1016/j.cej.2013.12.021
[19] Al-Bastaki N, Abbas A. Use of fluid instabilities to enhance membrane performance: a review. Desalination. 2001;136(1-3):255–262. DOI: 10.1016/S0011-9164(01)00188-6
[6] Wangnick K. IDAWorldwide Desalting Plants Inventory. Report No. 17 ed. Gnarrenburg,
[7] Shakaib M. Pressure and concentration gradients in membrane feed channels: Numerical and experimental investigations [thesis]. Karachi, Pakistan: University of Engineering and Technology; 2008. p. 271. Available from: http://prr.hec.gov.pk/Thesis/353S.pdf [8] Geraldes V, Semião V, Pinho MN. Flow management in nanofiltration spiral wound modules with ladder-type spacers. Journal of Membrane Science. 2002;203(1.2):87–102.
[9] Schwinge J, Neal PR, Wiley DE, Fletcher DF, Fane AG. Spiral wound modules and spacers: Review and analysis. Journal of Membrane Science. 2004;242(1-2):129–153. DOI:
[10] Baker RW. Membrane Technology and Applications. 2nd ed. Menlo Park, CA: John Wiley
[11] Kim S, Hoek EMV. Modeling concentration polarization in reverse osmosis processes.
[12] Zimmerer CC, Kottke V. Effects of spacer geometry on pressure drop,mass transfer, mixing behavior, and residence time distribution. Desalination. 1996;104(1-2):129–134.
[13] Ahmad AL, Lau KK. Impact of different spacer filaments geometries on 2D unsteady hydrodynamics and concentration polarization in spiral wound membrane. Journal of
[14] Cao Z, Wiley DE, Fane AG. CFD simulations of net-type turbulence promoters in a narrow channel. Journal of Membrane Science. 2001;185(2):157–176. DOI: 10.1016/S0376-
[15] Schwinge J, Wiley DE, Fletcher DF. Simulation of the flow around spacer filaments between channel walls. 1. Hydrodynamics. Industrial & Engineering Chemistry
[16] Saeed A, Vuthaluru R, Yang Y, Vuthaluru HB. Effect of feed spacer arrangement on flow dynamics through spacer filled membranes. Desalination. 2012;285:163–169. DOI:
[17] Sousa P, Soares A, Monteiro E, Rouboa A. A CFD study of the hydrodynamics in a desalination membrane filled with spacers. Desalination. 2014;349:22–30. DOI: 10.1016/j.
[18] Amokrane M, Sadaouia D, Dudeckb M, Koutsouc CP. New spacer designs for the performance improvement of the zigzag spacer configuration in spiral-wound membrane modules. Desalination and Water Treatment. 2015;57(12):5266–5274. DOI: 10.1080/19443994.2
Membrane Science. 2006;286(1-2):77–92. DOI: 10.1016/j.memsci.2005.06.056
Desalination. 2005;186(1-3):111–128. DOI: 10.1016/j.desal.2005.05.017
Germany: Wangnick Consulting GMBH; 2002
& Sons, Ltd; 2004. p. 538. DOI: 10.1002/0470020393
Research. 2002;41(12):2977–2987. DOI: 10.1021/ie010588y
DOI: 10.1016/s0376-7388(01)00753-0
DOI: 10.1016/0011-9164(96)00035-5
7388(00)00643-8
10.1016/j.desal.2011.09.050
desal.2014.06.019
015.1022003
10.1016/j.memsci.2003.09.031
44 Desalination
**Chapter 3**
**Desalination: A Means of Increasing Irrigation Water**
Globally, water resources for agricultural production have been on the decline. This is associated with increase in water demand over limited resources and poor quality water that adversely affects crop quality and yield and deteriorates soil properties. Even though soil salinity has been affecting agriculture for thousands of years, significant research has been conducted only in the past 100 years. Desalination, which is the process of reduc‐ ing the salt content in water to an acceptable level, could be an alternative for improving water quality, thereby increasing water sources and reducing the competition among various users of water. Thus, desalination could lead to improved crop quality, improved crop yield, enhanced all‐year round crop production, and as such become an important
**Keywords:** salinity, reverse osmosis, forward osmosis, irrigation water, crop production
As more than 60% of rainfall, the primary source of water for agriculture is lost to evapotrans‐ piration [1], with the continuous increase in human population and its resultant increase in water demand which is expected to nearly double its size in the next 50 years, the exploitation of the available water resources and the advent of climate change with its global warming effect on available water for crop production, the search for new, sustainable and drought‐ proof water resources is inevitable [2]. He further stated that since agricultural activities con‐ sume more than 60% of the total water demand, using treated wastewater for irrigation can reduce depletion of groundwater significantly. In Refs. [1, 3], it was stated that water‐scarce countries especially the Middle East countries located in the arid and semi‐arid zones will have to rely more on the use of non‐conventional irrigation water resources such as saline
> © 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
**Sources for Sustainable Crop Production**
Additional information is available at the end of the chapter
tool for effective agricultural water management.
http://dx.doi.org/10.5772/intechopen.69312
OrevaOghene Aliku
**Abstract**
**1. Introduction**
## **Desalination: A Means of Increasing Irrigation Water Sources for Sustainable Crop Production**
OrevaOghene Aliku
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.69312
#### **Abstract**
Globally, water resources for agricultural production have been on the decline. This is associated with increase in water demand over limited resources and poor quality water that adversely affects crop quality and yield and deteriorates soil properties. Even though soil salinity has been affecting agriculture for thousands of years, significant research has been conducted only in the past 100 years. Desalination, which is the process of reduc‐ ing the salt content in water to an acceptable level, could be an alternative for improving water quality, thereby increasing water sources and reducing the competition among various users of water. Thus, desalination could lead to improved crop quality, improved crop yield, enhanced all‐year round crop production, and as such become an important tool for effective agricultural water management.
**Keywords:** salinity, reverse osmosis, forward osmosis, irrigation water, crop production
## **1. Introduction**
As more than 60% of rainfall, the primary source of water for agriculture is lost to evapotrans‐ piration [1], with the continuous increase in human population and its resultant increase in water demand which is expected to nearly double its size in the next 50 years, the exploitation of the available water resources and the advent of climate change with its global warming effect on available water for crop production, the search for new, sustainable and drought‐ proof water resources is inevitable [2]. He further stated that since agricultural activities con‐ sume more than 60% of the total water demand, using treated wastewater for irrigation can reduce depletion of groundwater significantly. In Refs. [1, 3], it was stated that water‐scarce countries especially the Middle East countries located in the arid and semi‐arid zones will have to rely more on the use of non‐conventional irrigation water resources such as saline
© 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
aquifers to partly alleviate water scarcity. Although, the present freshwater resources may soon be insufficient to meet the growing demand for food [4], most of these drought‐proof water resources contain dissolved solids and chemicals such as salts. The application of these water resources for irrigation purposes often result to the detrimental effect of salinization of soils, environmental degradation and low crop yield.
sufficient salts at the root zone to impair crop growth [7]. Also, Corwin et al. [14] noted salin‐ ity as one of the most significant soil properties influencing cotton yield in a response model. In Ref. [15], it was reported that the emergence of sunflower and maize was affected by salin‐ ity and that the higher the salinity, the lower the leaf area and the dry matter production.
Desalination: A Means of Increasing Irrigation Water Sources for Sustainable Crop Production
http://dx.doi.org/10.5772/intechopen.69312
49
Desalination describes a range of processes which are used to reduce the amount of dissolved solids in water [16]. Also, Nofal [17] defined desalination as the removal of excess salt and other minerals from water in order to get fresh water suitable for drinking water, animal consumption and irrigation purposes. It is used to produce clean water from water sources containing dissolved chemicals, and in most cases, water sources are salty, producing fresh water from sea water or brackish water [16]. They further stated that natural waters may be classified approximately according to their total dissolved solid (TDS) values as listed in **Table 1**. Desalination is a water saving alternative to brackish water irrigation even though its diffusion as a viable method of water treatment has been limited by high costs and concern about the lack of plant nutrients in desalinated water [17]. In Ref. [4], it was also confirmed that desalination not only separates the undesirable salts from the water but also removes ions that are essential to plant growth. Although, a recent report concludes that the costs of desalination remain prohibitively expensive for full use by irrigated agriculture [18], for high value cash crops like green‐house vegetables and flowers, its use may be economically fea‐ sible [4]. According to Smith and Shaw [16], low‐cost methods of desalination by distillation
Due to the impact of climate change which has led to uncertainty in the amount and duration of rainfall for crop production, 69% of global water supply is being channelled for irrigation purpose [19]. As a result, present fresh water resources may soon be insufficient to meet the growing demand for food [4]. Although, at present, sea water desalination provides 1% of the world's drinking water, desalinized water is increasingly considered a source of water for agriculture [4]. In some countries, farmers have already adopted the use of desalinized
**2.1. Desalination**
are also available.
Source: Smith and Shawerji [16].
**Table 1.** Classes of natural waters.
**2.2. Prospects of desalinized water for agriculture**
Sweet waters 0–1000 Brackish waters 1000–5000 Moderately saline waters 5000–10,000 Severely saline waters 10,000–30,000 Seawater More than 30,000
**Type of water Total dissolved solids (mg/L)**
Salinization is one of the land degradation processes rendering millions of hectares of land unproductive for crop cultivation. It was stated in Ref. [5] that salinization is one of the most serious land degradation problems facing the world. According to El‐Swaify [6], salin‐ ity is when an 'excessive' amount or concentration of soluble salts occurs in the soil, either naturally or as a result of mismanaged irrigation water. Although, he further reported that salt‐affected soils are most abundant in arid regions worldwide, the extent of saline soils is variable [7], whereas Yan et al. [5] stated that soil salinity vary in time and space. Salts are often introduced into soil and water systems via the use of excessive inorganic fertilizers which are leached or washed away as runoff into underground water bodies used for irriga‐ tion purposes. According to El‐Swaify [6], salts in soil and irrigation water may be either natu‐ rally present as products of geochemical weathering of rocks and parent materials or derived directly from sea water flooding, spray or intrusion into groundwater sources and/or caused by irrigation mismanagement, particularly when internal soil drainage is impeded. Due to the presence of salts, most saline lands are virtually uncultivated in the dry season because of strong salinity and lack of water in good quality and quantity [7]. According to Gleick [8], almost half of the human population suffers insufficient access to portable water, and water scarcity in agriculture has been considered to be a global crisis [9].
Hence, desalination, which is any process that removes salt from water [10] to produce desali‐ nized water, is increasingly considered a source of water for agriculture [4]. Even though soil salinity has been affecting agriculture for thousands of years, significant research has been conducted only in the past 100 years [11]. Thus, this review highlights some of the effects of salinity on soil and crop growth and yield, and some possible methods of desalinization of water and soil resources for optimum utilization in a crop production system.
## **2. Effects of salinity on soil and crop**
Different salts, cations and anions vary in their effects on plants and soils, and as such differ‐ ences in ionic compositions of soil solutions and waters with similar electrical conductivity values may lead to dissimilar effects [6]. Salinity may adversely affect soil structure and other physical properties, and this could finally be transmitted to crop growth and development. For instance, the breakdown of soil structure can exacerbate salt effects on crops through increased surface crusting, germination inhibition and reduced permeability, porosity and aeration [6]. In Ref. [12], it was reported that soil infiltration rate was greatly affected by sodicity and electrolyte concentration of the irrigation water. In Ref. [13], it was reported that increasing salinity and sodicity resulted in a progressively smaller, more stressed microbial community which was less metabolically efficient. Saline soils have been reported to contain sufficient salts at the root zone to impair crop growth [7]. Also, Corwin et al. [14] noted salin‐ ity as one of the most significant soil properties influencing cotton yield in a response model. In Ref. [15], it was reported that the emergence of sunflower and maize was affected by salin‐ ity and that the higher the salinity, the lower the leaf area and the dry matter production.
#### **2.1. Desalination**
aquifers to partly alleviate water scarcity. Although, the present freshwater resources may soon be insufficient to meet the growing demand for food [4], most of these drought‐proof water resources contain dissolved solids and chemicals such as salts. The application of these water resources for irrigation purposes often result to the detrimental effect of salinization of
Salinization is one of the land degradation processes rendering millions of hectares of land unproductive for crop cultivation. It was stated in Ref. [5] that salinization is one of the most serious land degradation problems facing the world. According to El‐Swaify [6], salin‐ ity is when an 'excessive' amount or concentration of soluble salts occurs in the soil, either naturally or as a result of mismanaged irrigation water. Although, he further reported that salt‐affected soils are most abundant in arid regions worldwide, the extent of saline soils is variable [7], whereas Yan et al. [5] stated that soil salinity vary in time and space. Salts are often introduced into soil and water systems via the use of excessive inorganic fertilizers which are leached or washed away as runoff into underground water bodies used for irriga‐ tion purposes. According to El‐Swaify [6], salts in soil and irrigation water may be either natu‐ rally present as products of geochemical weathering of rocks and parent materials or derived directly from sea water flooding, spray or intrusion into groundwater sources and/or caused by irrigation mismanagement, particularly when internal soil drainage is impeded. Due to the presence of salts, most saline lands are virtually uncultivated in the dry season because of strong salinity and lack of water in good quality and quantity [7]. According to Gleick [8], almost half of the human population suffers insufficient access to portable water, and water
Hence, desalination, which is any process that removes salt from water [10] to produce desali‐ nized water, is increasingly considered a source of water for agriculture [4]. Even though soil salinity has been affecting agriculture for thousands of years, significant research has been conducted only in the past 100 years [11]. Thus, this review highlights some of the effects of salinity on soil and crop growth and yield, and some possible methods of desalinization of
Different salts, cations and anions vary in their effects on plants and soils, and as such differ‐ ences in ionic compositions of soil solutions and waters with similar electrical conductivity values may lead to dissimilar effects [6]. Salinity may adversely affect soil structure and other physical properties, and this could finally be transmitted to crop growth and development. For instance, the breakdown of soil structure can exacerbate salt effects on crops through increased surface crusting, germination inhibition and reduced permeability, porosity and aeration [6]. In Ref. [12], it was reported that soil infiltration rate was greatly affected by sodicity and electrolyte concentration of the irrigation water. In Ref. [13], it was reported that increasing salinity and sodicity resulted in a progressively smaller, more stressed microbial community which was less metabolically efficient. Saline soils have been reported to contain
soils, environmental degradation and low crop yield.
48 Desalination
scarcity in agriculture has been considered to be a global crisis [9].
**2. Effects of salinity on soil and crop**
water and soil resources for optimum utilization in a crop production system.
Desalination describes a range of processes which are used to reduce the amount of dissolved solids in water [16]. Also, Nofal [17] defined desalination as the removal of excess salt and other minerals from water in order to get fresh water suitable for drinking water, animal consumption and irrigation purposes. It is used to produce clean water from water sources containing dissolved chemicals, and in most cases, water sources are salty, producing fresh water from sea water or brackish water [16]. They further stated that natural waters may be classified approximately according to their total dissolved solid (TDS) values as listed in **Table 1**. Desalination is a water saving alternative to brackish water irrigation even though its diffusion as a viable method of water treatment has been limited by high costs and concern about the lack of plant nutrients in desalinated water [17]. In Ref. [4], it was also confirmed that desalination not only separates the undesirable salts from the water but also removes ions that are essential to plant growth. Although, a recent report concludes that the costs of desalination remain prohibitively expensive for full use by irrigated agriculture [18], for high value cash crops like green‐house vegetables and flowers, its use may be economically fea‐ sible [4]. According to Smith and Shaw [16], low‐cost methods of desalination by distillation are also available.
#### **2.2. Prospects of desalinized water for agriculture**
Due to the impact of climate change which has led to uncertainty in the amount and duration of rainfall for crop production, 69% of global water supply is being channelled for irrigation purpose [19]. As a result, present fresh water resources may soon be insufficient to meet the growing demand for food [4]. Although, at present, sea water desalination provides 1% of the world's drinking water, desalinized water is increasingly considered a source of water for agriculture [4]. In some countries, farmers have already adopted the use of desalinized
**Table 1.** Classes of natural waters.
brackish water for crop production. For instance, Mechell and Lesikar [20] reported that ∼22% of water desalinated in Spain are used for agricultural irrigation purposes, whereas an Australian survey found that 53% of the population envisioned desalinated water usage for irrigation of vegetables as highly likely.
#### **2.3. Benefits of desalination**
Desalination is a water saving alternative to brackish water irrigation [17]. By implication, it could increase the possible sources of water for irrigation, and as such enhance sustain‐ able all‐year round crop production. According to Ref. [4], the low level of salinity of desali‐ nized water is an extra benefit, because the salts [especially Sodium (Na<sup>+</sup> ) and Chlorine (Cl− )] damage soils, stunt plant growth and harm the environment. Hence, desalinized water could improve the quality of irrigation water thereby reducing the possibilities of the incidence of soil salinity with its consequent adverse effect on crop growth and yield via its deteriorat‐ ing effects on soil properties. Furthermore, desalination could increase the size of land area for cultivation, the number of crops (including salt sensitive crops) cultivated, improve crop quality, increase crop productivity and increase the broad band of water use for other pur‐ poses [17]. Desalination has been reported to improve farmers' income [17].
energy cost low. However, thermal technologies have rarely been used for brackish water desali‐ nation, because of the high cost involved [10]. According to Refs. [6, 21], thermal technologies are grouped into three major large scale processes, i.e., multi‐stage flash distillation (MSF), multi‐ effect distillation (MED) and vapour compression distillation (VCD). They stated that solar dis‐ tillation, which is another thermal technology, is typically used for very small production rates.
Desalination: A Means of Increasing Irrigation Water Sources for Sustainable Crop Production
http://dx.doi.org/10.5772/intechopen.69312
51
**Process Total dissolved solid value (mg/L)**
Ion exchange (not described here) 500–1000 Electrodialysis 500–3000 Reverse osmosis (standard membranes) 500–5000 Reverse osmosis (high‐resistance membranes) Over 5000 Distillation Over 30,000
**Table 2.** Suitability of desalination process based on the total dissolved solids.
This process of distillation involves the use of several (multi‐stage) chambers [10]. According to Ref. [21], this process sends the pre‐treated saline water through multiple chambers as illustrated in **Figure 1** [22]. In the MSF process, each successive stage of the plant operates at progressively lower pressures. In Ref. [21], it was explained that the pre‐treated saline water is heated and compressed to a high temperature and high pressure, and the pressure is reduced as the water progressively passes through the chambers, causing the water to rapidly boil. In other words, the pre‐treated water is first heated under high pressure as it is passed into the first 'flash chamber', where the pressure is released, causing the water to boil rapidly, result‐ ing in sudden evaporation or 'flashing', which continues in each successive stage, because the pressure at each stage is lower than that of the previous stage [10]. The vapour produced by the flashing is then condensed on a heat exchanger tubing that runs through each stage and collected as fresh water. Generally, only a small percentage of the pre‐treated saline water is
The MED process has been used since the late 1950s and the early 1960s [10]. According to Ref. [21], the MED employs the same principles as the MSF process except that instead of using multiple chambers of a single vessel, MED uses successive vessels (**Figure 2**), i.e., MED occurs in a series of vessels, using the principles of evaporation and condensation at reduced ambient pressure [21]. Here, water is produced by a series of evaporator vessels at progres‐ sively lower pressures. Water boils at lower temperatures as pressure decreases, such that the water vapour of the first vessel serves as the heating medium for the second, and so on [10]. According to Ref. [21], the multiple vessels make the MED process more efficient, while [10]
stated that the more the vessels, the higher the performance ratio of the MED.
*2.5.1. Multi‐stage flash distillation (MSF)*
Source: Smith and Shawerji [16].
converted into vapour and condensed [10].
*2.5.2. Multi‐effect distillation (MED)*
#### **2.4. Techniques in desalination**
According to Refs. [10, 21], techniques used in a desalination process essentially separates saline water into two parts, hence, two streams of water are produced.
It is often associated with electrical generation plants, from which both electricity and waste heat are available [16]. Some of these desalination methods could be relatively expensive, whereas others such as desalination by distillation could be low‐cost methods. According to Refs. [10, 21], the two major types of technologies used for desalination can be broadly classified into thermal technologies (multi‐stage flash distillation, multi‐effect distillation and vapour compression distillation) and membrane technologies (electrodialysis/electrodialysis reversal and reverse osmosis), with reverse osmosis, and distillation followed by condensa‐ tion being two main desalination methods [16]. In Ref. [10], it was stated that both technolo‐ gies need energy to operate and produce fresh water. However, the most appropriate method can be selected on the basis of the total dissolved solids (TDS) value of the raw water (**Table 2**).
#### **2.5. Thermal technologies**
These technologies involve the eating of saline water and collecting the condensed vapour distil‐ late to produce pure water [10]. In Ref. [21], it was reported that thermal distillation technologies are widely used in the Middle East, primarily because the region's petroleum reserves keep
**Table 2.** Suitability of desalination process based on the total dissolved solids.
energy cost low. However, thermal technologies have rarely been used for brackish water desali‐ nation, because of the high cost involved [10]. According to Refs. [6, 21], thermal technologies are grouped into three major large scale processes, i.e., multi‐stage flash distillation (MSF), multi‐ effect distillation (MED) and vapour compression distillation (VCD). They stated that solar dis‐ tillation, which is another thermal technology, is typically used for very small production rates.
#### *2.5.1. Multi‐stage flash distillation (MSF)*
brackish water for crop production. For instance, Mechell and Lesikar [20] reported that ∼22% of water desalinated in Spain are used for agricultural irrigation purposes, whereas an Australian survey found that 53% of the population envisioned desalinated water usage for
Desalination is a water saving alternative to brackish water irrigation [17]. By implication, it could increase the possible sources of water for irrigation, and as such enhance sustain‐ able all‐year round crop production. According to Ref. [4], the low level of salinity of desali‐
damage soils, stunt plant growth and harm the environment. Hence, desalinized water could improve the quality of irrigation water thereby reducing the possibilities of the incidence of soil salinity with its consequent adverse effect on crop growth and yield via its deteriorat‐ ing effects on soil properties. Furthermore, desalination could increase the size of land area for cultivation, the number of crops (including salt sensitive crops) cultivated, improve crop quality, increase crop productivity and increase the broad band of water use for other pur‐
According to Refs. [10, 21], techniques used in a desalination process essentially separates
(b) Concentrate or brine, which has salt and mineral concentrations higher than that of
It is often associated with electrical generation plants, from which both electricity and waste heat are available [16]. Some of these desalination methods could be relatively expensive, whereas others such as desalination by distillation could be low‐cost methods. According to Refs. [10, 21], the two major types of technologies used for desalination can be broadly classified into thermal technologies (multi‐stage flash distillation, multi‐effect distillation and vapour compression distillation) and membrane technologies (electrodialysis/electrodialysis reversal and reverse osmosis), with reverse osmosis, and distillation followed by condensa‐ tion being two main desalination methods [16]. In Ref. [10], it was stated that both technolo‐ gies need energy to operate and produce fresh water. However, the most appropriate method can be selected on the basis of the total dissolved solids (TDS) value of the raw water (**Table 2**).
These technologies involve the eating of saline water and collecting the condensed vapour distil‐ late to produce pure water [10]. In Ref. [21], it was reported that thermal distillation technologies are widely used in the Middle East, primarily because the region's petroleum reserves keep
) and Chlorine (Cl−
)]
nized water is an extra benefit, because the salts [especially Sodium (Na<sup>+</sup>
poses [17]. Desalination has been reported to improve farmers' income [17].
saline water into two parts, hence, two streams of water are produced.
(a) Treated water that has low concentrations of salts and minerals.
irrigation of vegetables as highly likely.
**2.3. Benefits of desalination**
50 Desalination
**2.4. Techniques in desalination**
the pre‐treated water.
**2.5. Thermal technologies**
This process of distillation involves the use of several (multi‐stage) chambers [10]. According to Ref. [21], this process sends the pre‐treated saline water through multiple chambers as illustrated in **Figure 1** [22]. In the MSF process, each successive stage of the plant operates at progressively lower pressures. In Ref. [21], it was explained that the pre‐treated saline water is heated and compressed to a high temperature and high pressure, and the pressure is reduced as the water progressively passes through the chambers, causing the water to rapidly boil. In other words, the pre‐treated water is first heated under high pressure as it is passed into the first 'flash chamber', where the pressure is released, causing the water to boil rapidly, result‐ ing in sudden evaporation or 'flashing', which continues in each successive stage, because the pressure at each stage is lower than that of the previous stage [10]. The vapour produced by the flashing is then condensed on a heat exchanger tubing that runs through each stage and collected as fresh water. Generally, only a small percentage of the pre‐treated saline water is converted into vapour and condensed [10].
#### *2.5.2. Multi‐effect distillation (MED)*
The MED process has been used since the late 1950s and the early 1960s [10]. According to Ref. [21], the MED employs the same principles as the MSF process except that instead of using multiple chambers of a single vessel, MED uses successive vessels (**Figure 2**), i.e., MED occurs in a series of vessels, using the principles of evaporation and condensation at reduced ambient pressure [21]. Here, water is produced by a series of evaporator vessels at progres‐ sively lower pressures. Water boils at lower temperatures as pressure decreases, such that the water vapour of the first vessel serves as the heating medium for the second, and so on [10]. According to Ref. [21], the multiple vessels make the MED process more efficient, while [10] stated that the more the vessels, the higher the performance ratio of the MED.
**Figure 1.** An illustration of the multi‐stage flash distillation (MSF) process (Source: Buros, 1990).
**Figure 2.** A schematic diagram of a multi‐effect distillation (MED) process (Source: [22]).
#### *2.5.3. Vapour compression distillation (VCD)*
The VCD can function independently or in combination with other thermal distillation pro‐ cesses such as the MED [10, 21]. According to Ref. [23], the heat for evaporating the pre‐ treated saline water comes from the compression of vapour, rather than the direct exchange of heat from steam produced in a boiler (**Figure 3**). It usually involves the use of a mechanical compressor to generate heat for evaporation [10]. Vapour compression distillation unit are commonly used to produce fresh water for small‐ to medium‐scale purposes such as resorts, hotels and industrial applications [21].
plastic covering is used to transmit radiant energy and also to allow water vapour to condense on its interior surface before it is collected as fresh water. Alike VCD, solar desalination is
Desalination: A Means of Increasing Irrigation Water Sources for Sustainable Crop Production
http://dx.doi.org/10.5772/intechopen.69312
53
According to Ref. [21], there are several membrane treatment processes, including reverse osmosis, nanofiltration, ultrafiltration and microfiltration. These processes involve the use of a barrier, which is a membrane, and a driving force. The membranes contain pores which dif‐ fer in sizes according to the type of process (**Figure 5**). It was explained in Ref. [21] that mem‐ brane technologies often require that the water undergo chemical and physical pre‐treatment
generally used for small‐scale operations [21].
**Figure 4.** An example of a solar still distillation process (Source: [22]).
**Figure 3.** An example of a vapour compression distillation (VCD) process (Source: [22]).
**2.6. Membrane technologies**
#### *2.5.4. Solar distillation*
This involves the use of solar energy for water desalination as shown in **Figure 4**. Also, Buros [21] stated that although the designs of solar distillation units vary greatly, the basic prin‐ ciples are the same. They explained that the sun provides the energy to evaporate the saline water, and the water vapour formed from the evaporation process then condenses on a clear glass covering before it is collected as fresh water in the condensate trough. The clear glass or Desalination: A Means of Increasing Irrigation Water Sources for Sustainable Crop Production http://dx.doi.org/10.5772/intechopen.69312 53
**Figure 3.** An example of a vapour compression distillation (VCD) process (Source: [22]).
**Figure 4.** An example of a solar still distillation process (Source: [22]).
plastic covering is used to transmit radiant energy and also to allow water vapour to condense on its interior surface before it is collected as fresh water. Alike VCD, solar desalination is generally used for small‐scale operations [21].
#### **2.6. Membrane technologies**
*2.5.3. Vapour compression distillation (VCD)*
hotels and industrial applications [21].
*2.5.4. Solar distillation*
52 Desalination
The VCD can function independently or in combination with other thermal distillation pro‐ cesses such as the MED [10, 21]. According to Ref. [23], the heat for evaporating the pre‐ treated saline water comes from the compression of vapour, rather than the direct exchange of heat from steam produced in a boiler (**Figure 3**). It usually involves the use of a mechanical compressor to generate heat for evaporation [10]. Vapour compression distillation unit are commonly used to produce fresh water for small‐ to medium‐scale purposes such as resorts,
**Figure 1.** An illustration of the multi‐stage flash distillation (MSF) process (Source: Buros, 1990).
**Figure 2.** A schematic diagram of a multi‐effect distillation (MED) process (Source: [22]).
This involves the use of solar energy for water desalination as shown in **Figure 4**. Also, Buros [21] stated that although the designs of solar distillation units vary greatly, the basic prin‐ ciples are the same. They explained that the sun provides the energy to evaporate the saline water, and the water vapour formed from the evaporation process then condenses on a clear glass covering before it is collected as fresh water in the condensate trough. The clear glass or According to Ref. [21], there are several membrane treatment processes, including reverse osmosis, nanofiltration, ultrafiltration and microfiltration. These processes involve the use of a barrier, which is a membrane, and a driving force. The membranes contain pores which dif‐ fer in sizes according to the type of process (**Figure 5**). It was explained in Ref. [21] that mem‐ brane technologies often require that the water undergo chemical and physical pre‐treatment
**Figure 5.** An illustration of the range of nominal membrane pore sizes for reverse osmosis (RO), nanofiltration (NF), ultrafiltration (UF), and microfiltration (MF) (Source: [23]).
to limit blockage by debris and scale formation on the membrane surfaces. The general char‐ acteristics of membrane processes are presented in **Table 3**. Membrane technologies can be subdivided into two broad categories: electrodialysis/electrodialysis reversal (ed/edr) and reverse osmosis (RO) [10]. According to Ref. [21], the driving force used in electrodialysis or electrodialysis reversal is an electrical potential, whereas that used in reverse osmosis is a pressure gradient.
#### *2.6.1. Electrodialysis and electrodialysis reversal (ED/EDR)*
This is a voltage‐driven membrane process in which an electrical potential is used to move salts through a membrane, leaving fresh water behind as product water [10]. In Ref. [21], it was explained that the membrane used for ED/EDR are built in such a way that they only allow passage of either positively or negatively charged ions, but not both. Here, ionic mol‐ ecules, such as sodium, chloride, calcium and carbonate in saline water, that are known to cause adverse effects on soil and crop productivity are removed from the treated water as the cations are attracted to the negative electrode, whereas the anions are attracted to the positive electrode while passing through selected membranes. According to Ref. [10], the membranes are usually arranged in an alternate pattern, with anion‐selective membrane followed by a cation‐selective membrane. He further explained that during this process, the salt content of the water channel is diluted, while concentrated solutions are formed at the electrodes. Concentrated and diluted solutions are created in the spaces between the alternating mem‐ branes, and these spaces bound by two membranes are called cells [10]. The pre‐treated saline
water passes through all the cells simultaneously to provide a continuous flow of desalinated water and a steady stream of concentrate from the stack [10]. Although the ED was originally conceived as a seawater desalination process, it has generally been used for brackish water
According to Refs. [10, 21], the EDR functions in a similar way as the ED. However, El‐Swaify [6] explained that the only exception to the EDR operating on the same general principle as the ED unit is that both the product and the concentrate channels are identical in the EDR, whereas Buros [21] also explained that the polarity or charge of the electrodes is switched periodically in the reverse process. Immediately following reversal, the product water is removed until the lines are flushed out and the desired water quality restored [10]. They explained that the reversal in flow of ions helps to remove scaling, slimes and other debris from the membranes
before they accumulate in large amount, thus extending the system's operating life.
desalination [10].
**Membrane process**
**Membrane driving force**
pressure difference or vacuum in open vessels
pressure difference
pressure difference
pressure difference
difference
**Table 3.** General characteristics of membrane processes.
Microfiltration Hydrostatic
Ultrafiltration Hydrostatic
Nanofiltration Hydrostatic
Reverse osmosis Hydrostatic
Dialysis Concentration
Electrodialysis Electromotive force
Source: Metcalf and Eddy [22].
**Typical separation mechanism**
Sieve + solution/ diffusion + exclusion
Solution/ diffusion + exclusion
Ion exchange with selective membranes **Operating structure (pore**
(2–50 nm)
Micropores (<2 nm)
**Typical operating range (μm)**
Desalination: A Means of Increasing Irrigation Water Sources for Sustainable Crop Production
0.08–2.0 Water +
0.005–0.2 Water + small molecules
0.001–0.01 Water +
Dense (<2 nm) 0.0001– 0.001 Water +
**Permeate description**
dissolved solutes
very small molecules, ionic solutes
very small molecules, ionic solutes
– Water + small molecules
– Water + ionic solutes
**Typical constituents**
55
**removed**
http://dx.doi.org/10.5772/intechopen.69312
TSS, turbidity, protozoan oocysts and cysts, some bacteria and viruses
Macromolecules, colloids, most bacteria, some viruses, proteins
Small molecules, some hardness, viruses
Macromolecules, colloids, most bacteria, some viruses, proteins
Ionized salt ions
Very small molecules, colour, hardness, sulfates, nitrate, sodium, other ions
**size)**
Sieve Macropores (>50 nm)
Sieve Mesopores
Diffusion Mesopores
(2–50 nm)
Micropores (<2 nm)
**Table 3.** General characteristics of membrane processes.
to limit blockage by debris and scale formation on the membrane surfaces. The general char‐ acteristics of membrane processes are presented in **Table 3**. Membrane technologies can be subdivided into two broad categories: electrodialysis/electrodialysis reversal (ed/edr) and reverse osmosis (RO) [10]. According to Ref. [21], the driving force used in electrodialysis or electrodialysis reversal is an electrical potential, whereas that used in reverse osmosis is a
**Figure 5.** An illustration of the range of nominal membrane pore sizes for reverse osmosis (RO), nanofiltration (NF),
This is a voltage‐driven membrane process in which an electrical potential is used to move salts through a membrane, leaving fresh water behind as product water [10]. In Ref. [21], it was explained that the membrane used for ED/EDR are built in such a way that they only allow passage of either positively or negatively charged ions, but not both. Here, ionic mol‐ ecules, such as sodium, chloride, calcium and carbonate in saline water, that are known to cause adverse effects on soil and crop productivity are removed from the treated water as the cations are attracted to the negative electrode, whereas the anions are attracted to the positive electrode while passing through selected membranes. According to Ref. [10], the membranes are usually arranged in an alternate pattern, with anion‐selective membrane followed by a cation‐selective membrane. He further explained that during this process, the salt content of the water channel is diluted, while concentrated solutions are formed at the electrodes. Concentrated and diluted solutions are created in the spaces between the alternating mem‐ branes, and these spaces bound by two membranes are called cells [10]. The pre‐treated saline
pressure gradient.
54 Desalination
*2.6.1. Electrodialysis and electrodialysis reversal (ED/EDR)*
ultrafiltration (UF), and microfiltration (MF) (Source: [23]).
water passes through all the cells simultaneously to provide a continuous flow of desalinated water and a steady stream of concentrate from the stack [10]. Although the ED was originally conceived as a seawater desalination process, it has generally been used for brackish water desalination [10].
According to Refs. [10, 21], the EDR functions in a similar way as the ED. However, El‐Swaify [6] explained that the only exception to the EDR operating on the same general principle as the ED unit is that both the product and the concentrate channels are identical in the EDR, whereas Buros [21] also explained that the polarity or charge of the electrodes is switched periodically in the reverse process. Immediately following reversal, the product water is removed until the lines are flushed out and the desired water quality restored [10]. They explained that the reversal in flow of ions helps to remove scaling, slimes and other debris from the membranes before they accumulate in large amount, thus extending the system's operating life.
#### *2.6.2. Reverse osmosis*
In relation to thermal processes, reverse osmosis is a relatively new process that was com‐ mercialized in the 1970s [10, 24]. Currently, it is the most widely used method for desalination in the United States [10]. This process of desalination uses a pressure gradient as the driv‐ ing force to move high pressure pre‐treated saline water through a membrane that prevents the salt ions from passing, thus, yielding the product water stream and a concentrated brine stream as shown in **Figure 6**, respectively [10, 21]. In other words, reverse osmosis utilizes hydraulic pressure to offset osmotic pressure and induces mass transport of water across a semi‐permeable membrane [25]. This is simply applying pressure (in excess of the osmotic pressure) to the saline water [16]. Osmotic pressure (π) is calculated using the Van't Hoff equation:
$$
\pi \quad \text{= MRT} \tag{1}
$$
explained that unlike nanofiltration, which is a membrane process that is used for the removal of divalent salt ions such as calcium, magnesium and sulphate, reverse osmosis is used for the
Desalination: A Means of Increasing Irrigation Water Sources for Sustainable Crop Production
According to Ref. [26], following mass balance equations are commonly used to describe reverse osmosis and nanofiltration membrane process performance. Equation (2) indicates
is concentrate flow rate (m<sup>3</sup>
is feedwater solute concentration, units of mass per volume (mg/L); *C*p is permeate
Forward osmosis is used to describe the use of osmosis as a salt‐water separation mecha‐ nism through an engineered membrane. It is an emerging membrane treatment process that belongs to the class of osmotically driven membrane processes [25]. It was first presented by Cath et al. [27] and could also be called direct osmosis. Unlike reverse osmosis where pressure is applied to the pre‐treated saline water and a low salinity permeate is produced, forward osmosis involves a semi‐permeable membrane which separates a high osmotic pressure 'draw'
/d); *ρ*<sup>f</sup>
**Figure 7.** Dissected view of a spiral reverse osmosis membrane element (Source: [22]).
+ *Q<sup>c</sup> ρc* (2)
/d) and *ρ*<sup>c</sup>
http://dx.doi.org/10.5772/intechopen.69312
57
+ *Q<sup>c</sup> Cc* (3)
is concentrate solute concentra‐
is density of
is density of feedwater; *Q*p is permeate flow rate
removal of sodium and chloride.
mass balance for water flow:
where *Q*<sup>f</sup>
where *C*<sup>f</sup>
*2.6.3. Forward osmosis*
concentrate.
(m<sup>3</sup>
*Qf ρ<sup>f</sup>* = *Qp ρ<sup>p</sup>*
Equation (3) describes mass balance for solute flux:
*Qf C<sup>f</sup>* = *Qp C<sup>p</sup>*
solute concentration, units of mass per volume (mg/L) and *C*<sup>c</sup>
is feedwater flow rate (m<sup>3</sup>
/d); *ρ*p is density of permeate; *Q*<sup>c</sup>
tion, units of mass per volume (mg/L).
where *M* is the molar concentration of dissolved species; *R* is the ideal gas constant and *T* is the temperature on the Kelvin scale.
According to Ref. [10], high pressure pumps supply the pressures between the range of 150 psi for slightly brackish water to 800–1000 psi for salt water, to enable the water to pass through the membrane and have the salt rejected. It is worthy to note that the membrane is easily torn and needs to be supported carefully [16]. Due to the fact that the membrane of the reverse osmosis process consists of small pores, the salt water needs to be filtered first to remove particles which might damage the membranes, while chemical additives may be added to prevent biological growth and scaling [16, 21]. This is very important as the mem‐ brane surfaces must remain clean [10].
The individual spiral reverse osmosis membrane element through which the high pressure pre‐treated saline water flows are constructed in a concentric spiral pattern that allow alter‐ nating layers of pre‐treated water and brine spacing, reverse osmosis membrane and a porous product water carrier (**Figure 7**) [21]. The porous product water carrier allows the fresh water to flow into the centre of the membrane element to be collected in the product water tube. According to Ref. [10], the reverse osmosis processes are used for desalinating brackish water (TDS > 1500 mg/L) and seawater. Although membrane desalination processes using reverse osmosis or nanofiltration are diffusion‐controlled membrane processes [25], also, Krishna [10]
**Figure 6.** Basic components of a reverse osmosis membrane treatment process (Source: [21]).
Desalination: A Means of Increasing Irrigation Water Sources for Sustainable Crop Production http://dx.doi.org/10.5772/intechopen.69312 57
**Figure 7.** Dissected view of a spiral reverse osmosis membrane element (Source: [22]).
explained that unlike nanofiltration, which is a membrane process that is used for the removal of divalent salt ions such as calcium, magnesium and sulphate, reverse osmosis is used for the removal of sodium and chloride.
According to Ref. [26], following mass balance equations are commonly used to describe reverse osmosis and nanofiltration membrane process performance. Equation (2) indicates mass balance for water flow:
$$Q\_{\uparrow}\rho\_{f} = Q\_{\rho}\rho\_{\rho} + |Q\_{c}\rho\_{c}|\tag{2}$$
where *Q*<sup>f</sup> is feedwater flow rate (m<sup>3</sup> /d); *ρ*<sup>f</sup> is density of feedwater; *Q*p is permeate flow rate (m<sup>3</sup> /d); *ρ*p is density of permeate; *Q*<sup>c</sup> is concentrate flow rate (m<sup>3</sup> /d) and *ρ*<sup>c</sup> is density of concentrate.
Equation (3) describes mass balance for solute flux:
$$\mathbf{Q}\_{\mathbf{f}} \mathbf{C}\_{\mathbf{f}} = \mathbf{Q}\_{\mathbf{p}} \mathbf{C}\_{\mathbf{p}} + \mathbf{Q}\_{\mathbf{c}} \mathbf{C}\_{\mathbf{c}} \tag{3}$$
where *C*<sup>f</sup> is feedwater solute concentration, units of mass per volume (mg/L); *C*p is permeate solute concentration, units of mass per volume (mg/L) and *C*<sup>c</sup> is concentrate solute concentra‐ tion, units of mass per volume (mg/L).
#### *2.6.3. Forward osmosis*
**Figure 6.** Basic components of a reverse osmosis membrane treatment process (Source: [21]).
In relation to thermal processes, reverse osmosis is a relatively new process that was com‐ mercialized in the 1970s [10, 24]. Currently, it is the most widely used method for desalination in the United States [10]. This process of desalination uses a pressure gradient as the driv‐ ing force to move high pressure pre‐treated saline water through a membrane that prevents the salt ions from passing, thus, yielding the product water stream and a concentrated brine stream as shown in **Figure 6**, respectively [10, 21]. In other words, reverse osmosis utilizes hydraulic pressure to offset osmotic pressure and induces mass transport of water across a semi‐permeable membrane [25]. This is simply applying pressure (in excess of the osmotic pressure) to the saline water [16]. Osmotic pressure (π) is calculated using the Van't Hoff
*π* = *MRT* (1)
where *M* is the molar concentration of dissolved species; *R* is the ideal gas constant and *T* is
According to Ref. [10], high pressure pumps supply the pressures between the range of 150 psi for slightly brackish water to 800–1000 psi for salt water, to enable the water to pass through the membrane and have the salt rejected. It is worthy to note that the membrane is easily torn and needs to be supported carefully [16]. Due to the fact that the membrane of the reverse osmosis process consists of small pores, the salt water needs to be filtered first to remove particles which might damage the membranes, while chemical additives may be added to prevent biological growth and scaling [16, 21]. This is very important as the mem‐
The individual spiral reverse osmosis membrane element through which the high pressure pre‐treated saline water flows are constructed in a concentric spiral pattern that allow alter‐ nating layers of pre‐treated water and brine spacing, reverse osmosis membrane and a porous product water carrier (**Figure 7**) [21]. The porous product water carrier allows the fresh water to flow into the centre of the membrane element to be collected in the product water tube. According to Ref. [10], the reverse osmosis processes are used for desalinating brackish water (TDS > 1500 mg/L) and seawater. Although membrane desalination processes using reverse osmosis or nanofiltration are diffusion‐controlled membrane processes [25], also, Krishna [10]
*2.6.2. Reverse osmosis*
56 Desalination
equation:
the temperature on the Kelvin scale.
brane surfaces must remain clean [10].
Forward osmosis is used to describe the use of osmosis as a salt‐water separation mecha‐ nism through an engineered membrane. It is an emerging membrane treatment process that belongs to the class of osmotically driven membrane processes [25]. It was first presented by Cath et al. [27] and could also be called direct osmosis. Unlike reverse osmosis where pressure is applied to the pre‐treated saline water and a low salinity permeate is produced, forward osmosis involves a semi‐permeable membrane which separates a high osmotic pressure 'draw' solution from the pre‐treated saline water with relatively lower salinity and osmotic pressure. Here, water is drawn across the membrane by natural osmosis, restricting the passage of salts at the membrane surface. In Ref. [25], it explained that when equal volumes of a dilute feed solution and a concentrated draw solution are separated by a semi‐permeable membrane, water flows into the concentrated draw solution, which has a higher osmotic pressure. This flow continues until chemical equilibrium is reached. The increase in water column height in the high osmotic pressure chamber at equilibrium equates to the difference in osmotic pres‐ sure between the dilute and concentrated solutions. Thus, forward osmosis uses the osmotic pressure differential (*Δπ*) across the membrane, rather than the hydraulic pressure differential as in reverse osmosis, as the driving force for transport of water through the membrane. The transport of water in forward osmosis is described in Eq. (4):
$$J\_W = K\_\eta (\sigma \Delta \pi - \Delta P) \tag{4}$$
the existing hydraulic structures. As irrigated agriculture does not require the strict standards that apply for drinking‐water requirements, opportunities appear to exist for the adoption of high‐quality desalinated water, and in this way, the final cost of a cubic metre of irrigation
Desalination: A Means of Increasing Irrigation Water Sources for Sustainable Crop Production
http://dx.doi.org/10.5772/intechopen.69312
59
**Figure 8.** Estimated energy consumption for desalination processes (Source: [29]).
Salinity arises from various natural and human‐induced processes and is a major phenom‐ enon that deteriorates soil properties, thus limiting the potentials of soils for sustainable crop production. Desalinated water is usually of high quality and can have less negative impact on soils and crops in comparison with direct use of brackish water. Thus, water desalination could have positive impacts on agriculture and the environment, such as increasing water availability and recycling poor‐quality water. The use of osmotic and distillation mechanisms to recover high quality water from wastewater effluents and saline waters could be high‐tech demanding especially when considering desalination of large volume of water for irrigation
Although, the use of low‐tech distillation methods could be easily adopted by peasant farm‐ ers in rural communities, the use of reverse osmosis has been said to be the most suitable for irrigated agriculture. As some of the processes involved in desalinizing saline water for sustainable crop production could be expensive, it could also be cost‐effective, owing to the
water can be reduced [29].
and other forms of utilization.
**3. Conclusions**
where *J*W = water flux; *Δπ is* differential osmotic pressure across the membrane; *K*W is water permeability coefficient of the membrane; *σ* is reflection coefficient (a measure of the relative permeability of a particular membrane to a particular solute) and *ΔP* is differential applied pressure across the membrane.
Past research has shown that forward osmosis membranes are good barriers to a broad range of contaminants, including bacteria, protozoa, viruses and other dissolved organic and inor‐ ganic constituents in contaminated water [27]. Also, in comparison to other desalination pro‐ cesses such as the multi‐stage flash, multi‐effect distillation and reverse osmosis, McGinnis and Elimelech [28] estimated that the forward osmosis has relatively lowest relative energy consumption (**Figure 8**). The authors estimated that forward osmosis with a thermally decom‐ posing draw solution [such as in the forward osmosis low temperature distillation (FO‐LT) process which incorporates the use of low‐quality heat for thermal decomposition of the draw solution and recovery using distillation columns] would use less than one‐third the work energy of reverse osmosis for desalination.
#### **2.7. Application of desalinated water in irrigated agriculture**
According to Ref. [1], the amount of fresh groundwater or agricultural activities is negligible and exists only in some locations. He further stated that desalination of brackish and saline water seems to be promising, especially in the absence of any other alternative. In spite of this, the cost of desalinated water are still too high for full use of this resource in irrigated agricul‐ ture, with the exception of intensive horticulture or high‐value cash crops, such as vegetables and flowers grown in greenhouses [29]. In Refs. [1, 29], reverse osmosis was reported to be the preferred desalination technology for agricultural uses because of the cost reductions driven by improvements in membranes in recent years. An example of countries that have adopted the application of desalinated water for irrigated agriculture is Spain. According to Ref. [30], Spain has more than 300 treatment plants with most of the plants processing brackish water, and located in coastal areas or within 60 km of the sea. It was also noted in Ref. [29] that small and medium size brackish water desalination plants, with a capacity of less than 1000 m3 /d (11.6 L/s), are common because they adapt better to individual farmer requirements and to
Desalination: A Means of Increasing Irrigation Water Sources for Sustainable Crop Production http://dx.doi.org/10.5772/intechopen.69312 59
**Figure 8.** Estimated energy consumption for desalination processes (Source: [29]).
the existing hydraulic structures. As irrigated agriculture does not require the strict standards that apply for drinking‐water requirements, opportunities appear to exist for the adoption of high‐quality desalinated water, and in this way, the final cost of a cubic metre of irrigation water can be reduced [29].
#### **3. Conclusions**
solution from the pre‐treated saline water with relatively lower salinity and osmotic pressure. Here, water is drawn across the membrane by natural osmosis, restricting the passage of salts at the membrane surface. In Ref. [25], it explained that when equal volumes of a dilute feed solution and a concentrated draw solution are separated by a semi‐permeable membrane, water flows into the concentrated draw solution, which has a higher osmotic pressure. This flow continues until chemical equilibrium is reached. The increase in water column height in the high osmotic pressure chamber at equilibrium equates to the difference in osmotic pres‐ sure between the dilute and concentrated solutions. Thus, forward osmosis uses the osmotic pressure differential (*Δπ*) across the membrane, rather than the hydraulic pressure differential as in reverse osmosis, as the driving force for transport of water through the membrane. The
where *J*W = water flux; *Δπ is* differential osmotic pressure across the membrane; *K*W is water permeability coefficient of the membrane; *σ* is reflection coefficient (a measure of the relative permeability of a particular membrane to a particular solute) and *ΔP* is differential applied
Past research has shown that forward osmosis membranes are good barriers to a broad range of contaminants, including bacteria, protozoa, viruses and other dissolved organic and inor‐ ganic constituents in contaminated water [27]. Also, in comparison to other desalination pro‐ cesses such as the multi‐stage flash, multi‐effect distillation and reverse osmosis, McGinnis and Elimelech [28] estimated that the forward osmosis has relatively lowest relative energy consumption (**Figure 8**). The authors estimated that forward osmosis with a thermally decom‐ posing draw solution [such as in the forward osmosis low temperature distillation (FO‐LT) process which incorporates the use of low‐quality heat for thermal decomposition of the draw solution and recovery using distillation columns] would use less than one‐third the work
According to Ref. [1], the amount of fresh groundwater or agricultural activities is negligible and exists only in some locations. He further stated that desalination of brackish and saline water seems to be promising, especially in the absence of any other alternative. In spite of this, the cost of desalinated water are still too high for full use of this resource in irrigated agricul‐ ture, with the exception of intensive horticulture or high‐value cash crops, such as vegetables and flowers grown in greenhouses [29]. In Refs. [1, 29], reverse osmosis was reported to be the preferred desalination technology for agricultural uses because of the cost reductions driven by improvements in membranes in recent years. An example of countries that have adopted the application of desalinated water for irrigated agriculture is Spain. According to Ref. [30], Spain has more than 300 treatment plants with most of the plants processing brackish water, and located in coastal areas or within 60 km of the sea. It was also noted in Ref. [29] that small and medium size brackish water desalination plants, with a capacity of less than 1000 m3
(11.6 L/s), are common because they adapt better to individual farmer requirements and to
*<sup>W</sup>* = *KW*(*σΔπ* − *ΔP* ) (4)
/d
transport of water in forward osmosis is described in Eq. (4):
*J*
energy of reverse osmosis for desalination.
**2.7. Application of desalinated water in irrigated agriculture**
pressure across the membrane.
58 Desalination
Salinity arises from various natural and human‐induced processes and is a major phenom‐ enon that deteriorates soil properties, thus limiting the potentials of soils for sustainable crop production. Desalinated water is usually of high quality and can have less negative impact on soils and crops in comparison with direct use of brackish water. Thus, water desalination could have positive impacts on agriculture and the environment, such as increasing water availability and recycling poor‐quality water. The use of osmotic and distillation mechanisms to recover high quality water from wastewater effluents and saline waters could be high‐tech demanding especially when considering desalination of large volume of water for irrigation and other forms of utilization.
Although, the use of low‐tech distillation methods could be easily adopted by peasant farm‐ ers in rural communities, the use of reverse osmosis has been said to be the most suitable for irrigated agriculture. As some of the processes involved in desalinizing saline water for sustainable crop production could be expensive, it could also be cost‐effective, owing to the fact that desalination could save water for agricultural production, increase the amount and types of crops grown, the area of land cultivated and as such improve the quality of crop yield and farmers' income.
[11] Qadir M, Wichelns D, Oster J, Jacobsen SE, Basra MA, Choukr‐Allah R. Sustainable Management of Saline Waters and Salt‐Affected Soils for Agriculture. Proceedings of the Second Bridging Workshop, 15‐18 November, 2009. The International Center for Agricultural Research in Dry Areas (Aleppo, Syria), and the International Water
Desalination: A Means of Increasing Irrigation Water Sources for Sustainable Crop Production
http://dx.doi.org/10.5772/intechopen.69312
61
[12] Agassi M, Shainberg I, Morin J. Effect of electrolyte concentration and soil sodic‐ ity o infiltration rate and crust formation. Soil Science Society of America Journal.
[13] Rietz DN, Haynes RJ. Effects of irrigation‐induced salinity and sodicity on soil and
[14] Corwin DL, Lesch SM, Shouse PJ, Soppe R, Ayars JE. Identifying soil properties that influence cotton yield using soil sampling directed by apparent soil electrical conductiv‐
[15] Katerji N, van Hoom JW, Hamdy A, Karam F, Mastrorilli M. Effect of salinity on emergence and on water stress and early seedling growth of sunflower and maize.
[16] Smith M, Shaw R. Desalination. WEDC Loughborough University Leicestershire LE11 3TU UK [Internet]. 1996. Available form: www.lboro.ac.uk/departments/cv/wedc/ [Assessed: 15
[17] Nofal I. Desal brackish water for agriculture. The regional water knowledge sharing
[18] Beltran JM, Koo‐Oshima S. Water desalination for agricultural applications. In: Proceedings of the FAO Expert Consultation on Water Desalination for Agricultural Applications; 26‐27 April 2004. Rome (Italy). FAO Land and Water Discussion Paper,
[19] Meerganz von Medeazza G. Water desalination as a long‐term solution to alleviate global fresh water scarcity? A North‐South Approach. Desalination, 2004; **169**: 287‐301
[20] Mechell JK, Lesikar B. Desalination methods for producing drinking water. Texas A&M
[21] Buros OK. The ABC's of Desalting. Topsfield, Massachusetts: International Desalination
[22] Metcalf and Eddy. Wastewater Engineering: Treatment and Reuse. 4th ed. New York:
[23] Buros OK. "The ABCs of Desalting", International Desalination Association, 2000;
[24] CH2M H. Assessment of Osmotic Mechanisms Pairing Desalination Concentrate and Wastewater Treatment. Austin, Texas: Texas Water Development Board; 2011. 159 p
microbial activity. Soilless Biology and Biochemistry. 2003;**35**(6):845‐854
Management Institute (Sri Lanka), 2010; 115pp.
ity. Agronomy Journal. 2003;**95**(2):352‐364
Agricultural Water Management. 1994;**26**(1‐2):81‐91
forum. 12‐14 May, 2015, 2015; Sharm El Sheikh, Egypt
Agrilife Extension Communications. E‐249, 04‐10, 2010; 8pp.
Topsfield, Massachusetts, USA, 2nd Edition, 31pp.
1981;**45**(5):848‐851
November 2016.]
2006; no. 5, 60 p.
Association; 1990
McGraw‐Hill; 2003
## **Author details**
OrevaOghene Aliku
Address all correspondence to: orevaoghenealiku@gmail.com
Department of Agronomy, University of Ibadan, Ibadan, Nigeria
### **References**
[11] Qadir M, Wichelns D, Oster J, Jacobsen SE, Basra MA, Choukr‐Allah R. Sustainable Management of Saline Waters and Salt‐Affected Soils for Agriculture. Proceedings of the Second Bridging Workshop, 15‐18 November, 2009. The International Center for Agricultural Research in Dry Areas (Aleppo, Syria), and the International Water Management Institute (Sri Lanka), 2010; 115pp.
fact that desalination could save water for agricultural production, increase the amount and types of crops grown, the area of land cultivated and as such improve the quality of crop yield
[1] Baalousha H. Desalination status in the Gaza Strip and its environmental impact.
[2] Texas Water Development Board. The Future of Desalination in Texas: Biennial Report on Seawater Desalination. Austin, Texas: Texas Water Development Board; 2010
[3] Zaide M. Drought and arid land water management: Government Focal Points. CSD‐ 16/17 National Report, Ministry of National Infrastructure, Israel; 2016; 11pp.
[4] Yermiyahu U, Tal A, Ben‐Gal A, Bar‐Tal A, Tarchitzky J, Laha O. Rethinking desalinated
[5] Yan N, Marschner P, Cao W, Zuo C, Qin W. Influence of salinity and water content on soil microorganisms. International Soil and Water Conservation Research. 2015;**3**:316‐323
[6] El‐Swaify SA. Soil and water salinity. In: Silva JA, Uchida R, editors. From: Plant Nutrient Management in Hawaii's Soils, Approaches for Tropical and Subtropical Agriculture. Manoa: College of Tropical Agriculture and Human Resources, University of Hawaii;
[7] Lantin RS, Quijano CC, Reyes RY, Neue HU. Rice and problem soils in the Philippines and the humid tropics: Past development and strategies for the 21st Century. Philippine
[8] Gleick PH. The World's Water 2002‐2003: The Biennial Report on Freshwater Resources.
[9] Postel S. Pillar of Sand: Can the Irrigation Miracle Last? Washington, DC: World Watch;
[10] Krishna HJ. Introduction to Desalination Technologies. Austin, Texas: Texas Water
water quality and agriculture. Environmental Science. 2007;**318**:920‐921
Address all correspondence to: orevaoghenealiku@gmail.com
Department of Agronomy, University of Ibadan, Ibadan, Nigeria
and farmers' income.
60 Desalination
**Author details**
OrevaOghene Aliku
**References**
Desalination. 2006;**196**:1‐12
2000. pp. 151‐158
1999
Journal of Crop Science. 1990;**15**(1):41‐47
Washington, DC: Island Press; 2002
Development Board; 2004. p. 7
[25] AWWA. Reverse Osmosis and Nanofiltration. American Water Works Association (AWWA) Manual of Water Supply Practices, Second Edition, M46, Quincy Avenue, Denver, USA, 2007; 240 pp.
**Chapter 4**
**Application of Multilayer Thin Film Technology in**
Membrane‐based desalination is the fastest growing technology in the area of desalina‐ tion. Reverse osmosis (RO) and nanofiltration (NF) have been established in the last couple of decades; meanwhile, forward osmosis (FO) has begun to find its own place in the field of desalination. Typical commercial polyamide (PA) thin film composite (TFC) membrane has been mostly used in those membrane processes, but it has no drawback. Recently, a versatile, robust technique in preparing ultra‐thin films, so‐called layer‐ by‐layer assembly (LbL), was adopted in fabrication of desalination membrane. This chapter highlights the most important literatures in the application of LbL assembly for preparing RO, NF and FO membranes, the obstacles and future works, which are
United Nations Environment Programme reported that two out of every three people will live in water‐stressed areas by the year 2025. In 2008, 450 million people in 29 countries suffered from water shortages [1]. In addition to that, many more people have been suffering from consuming contaminated water that may cost human lives [2]. These problems just come from domestic uses, which account for 5% of the total water consumption. Meanwhile, 75% of total global water consumption comes from agricultural uses and balance for industrial uses. So far, ground water represents about 90% of the world's readily available freshwater
> © 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
**Keywords:** layer‐by‐layer, dip‐LbL, spray‐LbL, spin‐LbL, reverse osmosis,
**Desalination Membrane**
Syed Javaid Zaidi and Farid Fadhillah
http://dx.doi.org/10.5772/intechopen.68375
**Abstract**
**1. Introduction**
Additional information is available at the end of the chapter
essential for those who wish to work in the field.
nanofiltration, forward osmosis, surface modification
## **Application of Multilayer Thin Film Technology in Desalination Membrane**
Syed Javaid Zaidi and Farid Fadhillah
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.68375
#### **Abstract**
[25] AWWA. Reverse Osmosis and Nanofiltration. American Water Works Association (AWWA) Manual of Water Supply Practices, Second Edition, M46, Quincy Avenue,
[26] Loeb S, Sourirajan S. Sea water demineralization by means of an osmotic membrane.
[27] Cath TY, Childress AE, Elimelech M. Forward osmosis: Principles, applications, and
[28] McGinnis RL, Elimelech M. Energy requirements of ammonia–carbon dioxide forward
[29] FAO. Water Desalination for Agricultural Applications. Proceedings of the FAO Expert Consultation on Water Desalination for Agricultural Applications, FAO Land and Water
[30] FAO. La desalación de agua de mar en la agricultura: situación actual y perspectivas futuras,
recent developments. Journal of Membrane Science. 2006;**281**:70‐87
Denver, USA, 2007; 240 pp.
62 Desalination
Advances in Chemistry Series. 1963;**38**:117
discussion paper 5, Rome, 2006; 60 pp.
by J. A. Medina. Draft report. Rome, 2003. 33 pp.
osmosis desalination. Desalination. 2007;**207**:370‐382
Membrane‐based desalination is the fastest growing technology in the area of desalina‐ tion. Reverse osmosis (RO) and nanofiltration (NF) have been established in the last couple of decades; meanwhile, forward osmosis (FO) has begun to find its own place in the field of desalination. Typical commercial polyamide (PA) thin film composite (TFC) membrane has been mostly used in those membrane processes, but it has no drawback. Recently, a versatile, robust technique in preparing ultra‐thin films, so‐called layer‐ by‐layer assembly (LbL), was adopted in fabrication of desalination membrane. This chapter highlights the most important literatures in the application of LbL assembly for preparing RO, NF and FO membranes, the obstacles and future works, which are essential for those who wish to work in the field.
**Keywords:** layer‐by‐layer, dip‐LbL, spray‐LbL, spin‐LbL, reverse osmosis, nanofiltration, forward osmosis, surface modification
### **1. Introduction**
United Nations Environment Programme reported that two out of every three people will live in water‐stressed areas by the year 2025. In 2008, 450 million people in 29 countries suffered from water shortages [1]. In addition to that, many more people have been suffering from consuming contaminated water that may cost human lives [2]. These problems just come from domestic uses, which account for 5% of the total water consumption. Meanwhile, 75% of total global water consumption comes from agricultural uses and balance for industrial uses. So far, ground water represents about 90% of the world's readily available freshwater
resources, and 1.5 billion people depend on it for their drinking water. Freshwater resources are so unevenly distributed that makes most people lack access to freshwater supply [1].
in colloidal and interfacial science [8]. After various testing and proofing, particularly for different multilayer precursors [9–13], the systematic way was then reported in 1997 and became the most‐cited article in the field of chemistry for 10 years (1998–2008) [14]. The classic approach to assemble PEM thin films is by alternately dipping the substrate into
Application of Multilayer Thin Film Technology in Desalination Membrane
http://dx.doi.org/10.5772/intechopen.68375
65
Although LbL assembly was initially invented by making use of electrostatic interaction between the two oppositely charge polyelectrolytes or colloidal particles, nowadays, it can also be formed via donor/acceptor [15, 16], hydrogen bonding [17, 18], covalent bonds [19–21] and stereo‐complex formation [22, 23]. The precise structure of each layer depends on a set of control parameters such as polyelectrolyte concentration, adsorption times, ionic strength
Sr‐LbL was introduced by Schlenoff [27] by employing poly(diallyl dimethyl ammonium chloride) (PDADMAC) and poly(styrene sulfonate) (PSS) on silicon wafer. Using this tech‐ nique, similar film structures and properties to d‐LbL film can be achieved in shorter deposi‐ tion time. It was reported that a fully automatic system of this technique can fabricate the film of the same quality as a d‐LbL film 25 times faster [28–30]. The main drawbacks of Sr‐LbL are still relatively slow polyelectrolyte assembly coupled with inefficient use of polymer solu‐ tion, which is about 99% of the polymer solution that is rinsed off during film preparation. However, if the size of the substrate is the concern, then, this technique is more suitable than d‐LbL [27]. Sr‐LbL can also be used to conformally coat individual fibres within a textured
SA‐LbL was introduced by Hong *et al.* in 2001 [31, 32] and Chiarelli *et al.* in the same year [33]. Hong *et al.* successfully fabricated very smooth thin layers with controllable thickness that comprised of the combination of nanoparticles and polyelectrolyte. The major difference
**Figure 1.** A schematic diagram of dip‐layer‐by‐layer assembly (adapted from Ref. [14]).
two oppositely charged polyelectrolytes as can be seen in **Figure 1**.
[24], pH [25] or temperature [26].
surface of hydrophobic textile [29].
**2.3. Spin LbL (SA‐LbL)**
**2.2. Spray‐LbL (Sr‐LbL)**
Desalination has been considered as the only method that can provide freshwater for drinking from traditional sources such as ground water, sea water and other saline aquifers [3]. These traditional sources of water account for 97.5% of all water on the Earth. Thus, capturing even a tiny portion of that water will have a significant impact on reducing water scarcity [3]. As generally known, the main desalination technology can be divided into two main categories, that is thermal‐based and membrane‐based desalination technology. Membrane technology, particularly RO, has been a dominant technology in the field of desalination membrane. Due to significant improvement of RO process in the last four decades, it has been projected to be the leading desalination technology in general and is mostly used in almost all areas except in the countries having readily available fossil fuels [4].
Since Cadotte developed thin film composite (TFC) membrane polyamide (PA) RO membrane using interfacial polymerization in 1980 [5], it has been mostly used in desalination membrane. However, commercial TFC RO still faces a major problem, particularly bio‐fouling. This is a result of hydrophobic and rough nature of the PA membrane itself and also partly due to the fabrication technique. The method is quite powerful to prepare thin films but lacks fine control over surface properties.
In the middle of the 1990s, a rediscovery of the so‐called layer‐by‐layer method has opened a new paradigm in the preparation of an ultra‐thin layer. LbL assembly offers nano‐level con‐ trol over several surface properties such as thickness, surface composition, surface roughness and so on. Not to mention, the flexibility in terms of material opened the room for improve‐ ment in terms of chemistry of the membrane, a field that has long been considered mature [6].
Therefore, the purpose of this chapter is to review recent activities in the field of LbL assem‐ bly, particularly those used in preparing the membrane for desalination. At the beginning, we will briefly highlight some important aspects about LbL assembly in general. We will then focus on method of preparation, some important results, drawbacks and future challenges related to the application of LbL in desalination membrane. In addition to that, we will also highlight some recent works related to the application of this method in the modification of commercially available membranes used in the field of desalination. We limit our discussion to the application of this membrane in separating or removing ions of salt only because there have been many reported works as well in the application of LbL NF, for instance, for remov‐ ing organic contaminants from water, etc.
## **2. Layer‐by‐layer**
#### **2.1. Dip‐LbL (d‐LbL)**
The root of LbL assembly might be traced back to 1966 when Iler introduced a novel technique in which colloidal oppositely charged particles can be assembled into layer‐ by‐layer films [7]. However, it was only after Decher reintroduced a similar technique for polyelectrolytes multilayer (PEM) assembly that the technique became very popular in colloidal and interfacial science [8]. After various testing and proofing, particularly for different multilayer precursors [9–13], the systematic way was then reported in 1997 and became the most‐cited article in the field of chemistry for 10 years (1998–2008) [14]. The classic approach to assemble PEM thin films is by alternately dipping the substrate into two oppositely charged polyelectrolytes as can be seen in **Figure 1**.
Although LbL assembly was initially invented by making use of electrostatic interaction between the two oppositely charge polyelectrolytes or colloidal particles, nowadays, it can also be formed via donor/acceptor [15, 16], hydrogen bonding [17, 18], covalent bonds [19–21] and stereo‐complex formation [22, 23]. The precise structure of each layer depends on a set of control parameters such as polyelectrolyte concentration, adsorption times, ionic strength [24], pH [25] or temperature [26].
#### **2.2. Spray‐LbL (Sr‐LbL)**
resources, and 1.5 billion people depend on it for their drinking water. Freshwater resources are so unevenly distributed that makes most people lack access to freshwater supply [1].
Desalination has been considered as the only method that can provide freshwater for drinking from traditional sources such as ground water, sea water and other saline aquifers [3]. These traditional sources of water account for 97.5% of all water on the Earth. Thus, capturing even a tiny portion of that water will have a significant impact on reducing water scarcity [3]. As generally known, the main desalination technology can be divided into two main categories, that is thermal‐based and membrane‐based desalination technology. Membrane technology, particularly RO, has been a dominant technology in the field of desalination membrane. Due to significant improvement of RO process in the last four decades, it has been projected to be the leading desalination technology in general and is mostly used in almost all areas except in
Since Cadotte developed thin film composite (TFC) membrane polyamide (PA) RO membrane using interfacial polymerization in 1980 [5], it has been mostly used in desalination membrane. However, commercial TFC RO still faces a major problem, particularly bio‐fouling. This is a result of hydrophobic and rough nature of the PA membrane itself and also partly due to the fabrication technique. The method is quite powerful to prepare thin films but lacks fine control
In the middle of the 1990s, a rediscovery of the so‐called layer‐by‐layer method has opened a new paradigm in the preparation of an ultra‐thin layer. LbL assembly offers nano‐level con‐ trol over several surface properties such as thickness, surface composition, surface roughness and so on. Not to mention, the flexibility in terms of material opened the room for improve‐ ment in terms of chemistry of the membrane, a field that has long been considered mature [6]. Therefore, the purpose of this chapter is to review recent activities in the field of LbL assem‐ bly, particularly those used in preparing the membrane for desalination. At the beginning, we will briefly highlight some important aspects about LbL assembly in general. We will then focus on method of preparation, some important results, drawbacks and future challenges related to the application of LbL in desalination membrane. In addition to that, we will also highlight some recent works related to the application of this method in the modification of commercially available membranes used in the field of desalination. We limit our discussion to the application of this membrane in separating or removing ions of salt only because there have been many reported works as well in the application of LbL NF, for instance, for remov‐
The root of LbL assembly might be traced back to 1966 when Iler introduced a novel technique in which colloidal oppositely charged particles can be assembled into layer‐ by‐layer films [7]. However, it was only after Decher reintroduced a similar technique for polyelectrolytes multilayer (PEM) assembly that the technique became very popular
the countries having readily available fossil fuels [4].
over surface properties.
64 Desalination
ing organic contaminants from water, etc.
**2. Layer‐by‐layer**
**2.1. Dip‐LbL (d‐LbL)**
Sr‐LbL was introduced by Schlenoff [27] by employing poly(diallyl dimethyl ammonium chloride) (PDADMAC) and poly(styrene sulfonate) (PSS) on silicon wafer. Using this tech‐ nique, similar film structures and properties to d‐LbL film can be achieved in shorter deposi‐ tion time. It was reported that a fully automatic system of this technique can fabricate the film of the same quality as a d‐LbL film 25 times faster [28–30]. The main drawbacks of Sr‐LbL are still relatively slow polyelectrolyte assembly coupled with inefficient use of polymer solu‐ tion, which is about 99% of the polymer solution that is rinsed off during film preparation. However, if the size of the substrate is the concern, then, this technique is more suitable than d‐LbL [27]. Sr‐LbL can also be used to conformally coat individual fibres within a textured surface of hydrophobic textile [29].
#### **2.3. Spin LbL (SA‐LbL)**
SA‐LbL was introduced by Hong *et al.* in 2001 [31, 32] and Chiarelli *et al.* in the same year [33]. Hong *et al.* successfully fabricated very smooth thin layers with controllable thickness that comprised of the combination of nanoparticles and polyelectrolyte. The major difference
**Figure 1.** A schematic diagram of dip‐layer‐by‐layer assembly (adapted from Ref. [14]).
between d‐LbL and SA‐LbL is the way polyelectrolytes are deposited on the substrate. As it is generally known, in the case of d‐LbL, the substrate is immersed in polyelectrolyte solutions, while in the case of SA‐LbL, only a small amount of polyelectrolyte is injected onto the spin‐ ning substrate.
**LbL film substrate Main result Testing condition Ref**
/SO4 2− = 9.9
showed selectivity of Cl−
= 96%, NaCl =
2− = 92.3;
2 − = 96%;
showed
SO4
/SO4 2− = 30 0.1 M for each of NaCl, MgCl<sup>2</sup>
0.1 M for each of KCl and K<sup>2</sup>
membrane active area: 2 cm2
1000 ppm for each of NaCl,
active area: 1.5 cm2
; cross flow cell.
Membrane active area: 2 cm2
; P = 4.8 bars; membrane
, 1000 ppm of SO4
, 1000 ppm of SO4
4.8 bars; membrane active area: 1.5
SO4
membrane active area: 23.75 cm2
; cross flow cell.
4.8 bars; membrane active area: 1.5
, dead end cell
cm2
Application of Multilayer Thin Film Technology in Desalination Membrane
end cell
MgSO4
cm2
1000 of Cl−
2− = 150 0.1 F for each of KCl and K<sup>2</sup>
1000 of Cl−
1000 ppm Na<sup>2</sup>
cm2
/
NaSO4, membrane active area: 4.53
http://dx.doi.org/10.5772/intechopen.68375
and
SO4 ,
; cross flow cell.
, dead
2− ; P =
SO4
.
2−; P =
, P = 2 bars [40]
[35]
67
[38]
[39]
[42]
[47]
[48]
(PAH/PSS)60 showed selectivity of
(PAH/PSS)4.5 showed individual
.day; rejection of SO<sup>4</sup>
/SO4
/SO4 2− = 360
(PSS/PDADMAC)4.5 showed WF = 1.6
/(PAH/PSSMA)<sup>1</sup>
.h, rejection of Na<sup>2</sup>
.day; rejection of SO<sup>4</sup>
/SO4 2− = 32
**Table 1.** Selected work for NF prepared from polyelectrolyte multilayer membrane using LbL assembly.
/SO4 2− = 15
/Mg2+ = 15.1; Cl−
Na<sup>+</sup>
SO4 2− = 7
(PSS/PDADMAC)‐porous alumina (PSS/PDADMAC)4.5 showed WF = 2.4 m3 /m2
> m3 /m2
(PAH/PAA)2.5 Selectivity of Cl−
PAN/PET = polyacrylonitrile/polyether terephthalate.
PSSMA = poly (4‐styrenesulfonic acid‐co‐maleic acid).
(PAH/PSS)<sup>5</sup>
rejection of MgSO<sup>4</sup>
selectivity Cl−
Selectivity of Cl−
selectivity Cl−
(PAH/PSS)<sup>1</sup>
= 91.4%
**Figure 2.** A rejection model of multi‐bipolar membrane (adapted from Ref. [35]).
WF = 28.6 L/m<sup>2</sup>
29% and selectivity Cl−
(PAH/PSS)‐PAN/PET with 20–200
(PAH/PSS)‐porous alumina with
(PAH/PSS)‐porous alumina with
Support: Porous alumina with 20
After heat‐induced crosslinking
(PDADMAC/PSS)‐PES UF 50 kDa
(PAH/PSS)/(PAH/PSSMA)**‐**PAN
PES = polyethersulfone.
nm pore size1
20 nm pore size
20 nm pore size
(PAH/PSS)<sup>5</sup>
nm pore size
at 115°C
MWCO<sup>2</sup>
UF<sup>3</sup>
1
2
3
It was reported that the SA‐LbL film is several times thicker than d‐LbL film for the same number of bilayers, which is due to different adsorption mechanisms. In d‐LbL, the poly‐ mer chain will be adsorbed on the substrate under influence of electrostatic force and then followed by chain rearrangement. Meanwhile, in SA‐LbL, due to high‐speed spinning, the adsorption, the rearrangement of polyelectrolyte chains and water removal occur simulta‐ neously. Quick water removal increases polyelectrolyte concentration in a very short time and also removes screening effect by water molecules. This in turn promotes faster adsorp‐ tion and stronger electrostatic force; hence, more polymer chains are adsorbed within short time and result in a thicker film than d‐LbL film [31].
Air shear force that occurs due to the relative movement between spinning substrate and air enhances the planarization of multilayer film, significantly reduces the surface roughness of the film and enhances mechanical integrity [34]. The above features are indirect evidences that SA‐LbL film has a highly ordered internal structure [31], and experimentally, it has been revealed by neutron reflectivity study. Because a highly ordered internal structure can be obtained in much shorter time than conventional d‐LbL or Sr‐LbL, SA‐LbL is considered to be more "technologically friendly" [34].
## **3. Layer‐by‐layer application**
#### **3.1. LbL for separation membrane**
#### *3.1.1. Nanofiltration*
Due to its charged characteristic, polyelectrolyte LbL membrane has been found to be promis‐ ing for NF application. There have been many works done and are ongoing in this area. Some representative works are listed in **Table 1**. We selected the best performance in terms of flux and selectivity of Cl− /SO4 2− from each work so they can be compared to each other. All those reported works used d‐LbL approach unless it is stated differently.
The earliest notable work in d‐LbL NF membrane perhaps was done by Krasemann and Tieke in 1999 [35]. They suggested the mechanism of salt rejection by LbL membrane similar to that of bipolar membrane as can be seen in **Figure 2**. However, some fundamental studies showed that PEM films do not exhibit an internal structure as originally projected and commonly depicted. There is significant intertwine between polyelectrolytes as a consequence of charge compensation. In d‐LbL assembly, this charge compensation is most likely attained since the contact time between adjacent layers is sufficient to achieve it [36]. In accordance with this statement, our work using SA‐LbL assembly also showed no significant increase in rejection
1 PAN/PET = polyacrylonitrile/polyether terephthalate.
2 PES = polyethersulfone.
between d‐LbL and SA‐LbL is the way polyelectrolytes are deposited on the substrate. As it is generally known, in the case of d‐LbL, the substrate is immersed in polyelectrolyte solutions, while in the case of SA‐LbL, only a small amount of polyelectrolyte is injected onto the spin‐
It was reported that the SA‐LbL film is several times thicker than d‐LbL film for the same number of bilayers, which is due to different adsorption mechanisms. In d‐LbL, the poly‐ mer chain will be adsorbed on the substrate under influence of electrostatic force and then followed by chain rearrangement. Meanwhile, in SA‐LbL, due to high‐speed spinning, the adsorption, the rearrangement of polyelectrolyte chains and water removal occur simulta‐ neously. Quick water removal increases polyelectrolyte concentration in a very short time and also removes screening effect by water molecules. This in turn promotes faster adsorp‐ tion and stronger electrostatic force; hence, more polymer chains are adsorbed within short
Air shear force that occurs due to the relative movement between spinning substrate and air enhances the planarization of multilayer film, significantly reduces the surface roughness of the film and enhances mechanical integrity [34]. The above features are indirect evidences that SA‐LbL film has a highly ordered internal structure [31], and experimentally, it has been revealed by neutron reflectivity study. Because a highly ordered internal structure can be obtained in much shorter time than conventional d‐LbL or Sr‐LbL, SA‐LbL is considered to be
Due to its charged characteristic, polyelectrolyte LbL membrane has been found to be promis‐ ing for NF application. There have been many works done and are ongoing in this area. Some representative works are listed in **Table 1**. We selected the best performance in terms of flux
The earliest notable work in d‐LbL NF membrane perhaps was done by Krasemann and Tieke in 1999 [35]. They suggested the mechanism of salt rejection by LbL membrane similar to that of bipolar membrane as can be seen in **Figure 2**. However, some fundamental studies showed that PEM films do not exhibit an internal structure as originally projected and commonly depicted. There is significant intertwine between polyelectrolytes as a consequence of charge compensation. In d‐LbL assembly, this charge compensation is most likely attained since the contact time between adjacent layers is sufficient to achieve it [36]. In accordance with this statement, our work using SA‐LbL assembly also showed no significant increase in rejection
2− from each work so they can be compared to each other. All those
time and result in a thicker film than d‐LbL film [31].
more "technologically friendly" [34].
**3. Layer‐by‐layer application**
**3.1. LbL for separation membrane**
/SO4
reported works used d‐LbL approach unless it is stated differently.
*3.1.1. Nanofiltration*
and selectivity of Cl−
ning substrate.
66 Desalination
3 PSSMA = poly (4‐styrenesulfonic acid‐co‐maleic acid).
**Table 1.** Selected work for NF prepared from polyelectrolyte multilayer membrane using LbL assembly.
**Figure 2.** A rejection model of multi‐bipolar membrane (adapted from Ref. [35]).
by doubling the number of layers, even though it is well known that the contact between layers is very short in this type of LbL assembly. This is evidence that even in short time, charge compensation is attained to some extent. It is also evidence that a bipolar mechanism may not be suitable, and this type of membrane rather follows a typical mechanism for NF or RO membrane [37]. It was also investigated that by changing the outermost layer, the anion flux changes significantly suggest that only uncompensated charges lie near the film/solution interface that have significant effect in ion rejection [38, 39].
As previously discussed, swelling of the membrane can result in higher flux due to expansion of the pores. For NF application, in which the membrane is exposed to relatively low salt concentration solution, swelling may not cause serious problems, but in the case of FO or RO, swelling degree can be very high and significantly decreases the salt rejection. Several attempts have been done to enhance LbL film stability, for example, by using crosslinking. Glutaraldehyde (GA) is the mostly used chemical crosslinker as can be seen in **Table 2**. With this crosslinking, LbL membrane can maintain its structure and performance and showed quite comparable performance with the commercial membrane [50]. Besides GA, UV light can also be employed to further increase the extent of crosslinking [51]. The mechanism of GA and UV light crosslinking can be seen in **Figure 4**. Some polyelectrolytes that contain carbonyl and amine functional group can be crosslinked simply by heating at 180°C, and this results in a
So far, the preparation of LbL membrane is mainly from polyelectrolytes. Recently, the researcher has started utilizing the technique for fabrication of LbL membrane using meth‐ ylphenylene diamine (MPD) and trimesoyl chloride (TMC), two most commonly used
**LbL film substrate LbL assembly Result/testing condition4 Ref.**
SWFR of 2.8 mM
SWFR of 4.86 mM
(PAH/PSS)<sup>3</sup>
= 73.5 L/m2
= 19 L/m2
L/m2
m2
at 0.5 M = 32.65 atm.
**Table 2.** The work progress for FO prepared from polyelectrolyte multilayer membrane using LbL assembly.
NaCl, dead end cell
Dip 10 layers of molecular LbL showed WF of 33 L/
.h and SWFR = 0.94 mM
m2
SWFR = 2.1 mM
= 15 L/m2
showed WF of 105.4 L/m<sup>2</sup>
Application of Multilayer Thin Film Technology in Desalination Membrane
http://dx.doi.org/10.5772/intechopen.68375
69
with UV exposure = 2 h showed WF
inner deposited layers showed WF
showed water flux of 28 L/
with MMT = 1 wt% showed WF
with SG = 1 wt% showed WF = 77.9
, FS = pure water, cross flow system, unless it is mentioned
.h and SWFR of 0.09 mM/DS: 0.5 M
(PAH/PSS)3‐3 showed WF of 35.5 L/m<sup>2</sup>
(PAH/PSS)2.5 showed WF = 42 L/m<sup>2</sup>
.h and SWFR = 0.8 mM
showed WF = 40.5 L/m<sup>2</sup>
.h and SWFR = 0.6 mM
.h and SWFR of 0.74 mM/dead end cell
.h and SWFR of 1.7 mM/DS: 0.5 M NaCl
= 0.84 mM / FS: 10 mM NaCl
.h and
.h and
.h and SWFR
.h and
[50]
[49]
[55]
[51]
[56]
[41]
[57]
[45]
[46]
[53]
polyamide‐like layer which has better stability and lower swelling degree [52].
(PAH/PSS)‐PAN Dip, crosslinked using GA (PAH/PSS)<sup>3</sup>
(PAH/PSS)‐hollow fibre
(CHI/PAA)−(PVA+MMT−
(MPD/TMC)–(PEI/PAA
PES
TA+ LiCl)
coated PAN)
4
Dip, crosslinked using GA, double‐skinned layer
Dip, crosslinked using GA, embedded silver nanoparticle
Dip, crosslinked using combined GA and UV light
(PAH/PAA‐PSS)‐PEI Dip, crosslinked using GA (PAH/PAA‐PSS)<sup>3</sup>
(PAH/PSS)−(PAN+SG) Dip (PAH/PSS)<sup>3</sup>
All reported default testing condition is DS = 0.5 M MgCl<sup>2</sup>
differently. πNaCl at 0.5 M = 22.74 atm and *π* MgCl <sup>2</sup>
Dip (PAH/PSS)<sup>6</sup>
Semi‐dynamic (PAH/PSS)<sup>2</sup>
Dip (CHI/PAA)<sup>3</sup>
It is widely known that ion rejection is one of the most importance performance indicators of a membrane. Aside employing highly charged polyelectrolytes which can improve rejection from the outermost layer, using dynamic LbL assembly seems quite helpful in this aspect. In this type of LbL assembly, polyelectrolyte solution is forced to pass the support layer and leaves behind the polyelectrolyte on the surface of the support layer. Dynamic LbL produces a more compact and tighter LbL membrane, which in turn increases the rejection [40]. There is also so‐called semi‐dynamic LbL assembly, which involves the dynamic introduction/ replacement of solution into the fibre lumen by syringes followed by static contact for desired time. This technique is quite useful for hollow fibre membrane as the support [41].
In addition to ion rejection, flux is also considered as the most important performance indicator. Many membranes have high flux due to swelling which in turn will reduce the rejection. From investigation of several frequently used PEM pairs, it was reported that solution fluxes decrease in the following order: PDADMAC/PSS > Poly(allylamine hydrochloride) PAH/PSS > PDADMAC/ Poly(acrylic acid) PAA >> PAH/PAA [42]. PDADMAC/PSS, for instance, permits higher flux than PAH/PSS due to higher degree of swelling [43]. In addition to swelling, highly porous support can also contribute to higher flux because it helps in reducing membrane resistance. Several methods have been used to produce highly porous support such as using electrospun nanofiber [44] or incorporating nanoparticles such as montmorillonite (MMT) or silica gel (SG). For example, poly‐ vinyl alcohol (PVA) mixed with 1 wt% montmorillonite (MMT) can achieve porosity of 80% [45]. Likewise, PAN support mixed with 1 wt% SG can also achieve porosity of 80% [46].
#### *3.1.2. Forward osmosis*
The application of LbL assembly for NF is quite promising as can be seen in the previous sec‐ tion. There have been many works done and are ongoing in that area; meanwhile, the use of LbL assembly for fabricating FO is still at its infancy. The main results from FO mode of testing of sev‐ eral LbL FO membranes can be seen in **Table 2**. It is important to note that the information given in **Table 2** is only part of the works that showed the best performance in terms of water flux (WF) and salt to water flux ratio (SWFR) with active layer facing draw solution (ALDS) testing mode. It is known that ALDS mode shows higher initial water flux compared to ALFS mode, but it is also prone to internal fouling because foulants enter the porous support easily. This type of fouling in fact is much more difficult to be cleaned. In order to overcome this problem, a double‐skinned design was proposed [49]. Double‐skinned layer approach uses active layer deposited on both surfaces with the support layer, which will be in between as can be seen in **Figure 3**. Hence, it will prevent the foulants from entering the membrane from the support layer side.
As previously discussed, swelling of the membrane can result in higher flux due to expansion of the pores. For NF application, in which the membrane is exposed to relatively low salt concentration solution, swelling may not cause serious problems, but in the case of FO or RO, swelling degree can be very high and significantly decreases the salt rejection. Several attempts have been done to enhance LbL film stability, for example, by using crosslinking. Glutaraldehyde (GA) is the mostly used chemical crosslinker as can be seen in **Table 2**. With this crosslinking, LbL membrane can maintain its structure and performance and showed quite comparable performance with the commercial membrane [50]. Besides GA, UV light can also be employed to further increase the extent of crosslinking [51]. The mechanism of GA and UV light crosslinking can be seen in **Figure 4**. Some polyelectrolytes that contain carbonyl and amine functional group can be crosslinked simply by heating at 180°C, and this results in a polyamide‐like layer which has better stability and lower swelling degree [52].
by doubling the number of layers, even though it is well known that the contact between layers is very short in this type of LbL assembly. This is evidence that even in short time, charge compensation is attained to some extent. It is also evidence that a bipolar mechanism may not be suitable, and this type of membrane rather follows a typical mechanism for NF or RO membrane [37]. It was also investigated that by changing the outermost layer, the anion flux changes significantly suggest that only uncompensated charges lie near the film/solution
It is widely known that ion rejection is one of the most importance performance indicators of a membrane. Aside employing highly charged polyelectrolytes which can improve rejection from the outermost layer, using dynamic LbL assembly seems quite helpful in this aspect. In this type of LbL assembly, polyelectrolyte solution is forced to pass the support layer and leaves behind the polyelectrolyte on the surface of the support layer. Dynamic LbL produces a more compact and tighter LbL membrane, which in turn increases the rejection [40]. There is also so‐called semi‐dynamic LbL assembly, which involves the dynamic introduction/ replacement of solution into the fibre lumen by syringes followed by static contact for desired
In addition to ion rejection, flux is also considered as the most important performance indicator. Many membranes have high flux due to swelling which in turn will reduce the rejection. From investigation of several frequently used PEM pairs, it was reported that solution fluxes decrease in the following order: PDADMAC/PSS > Poly(allylamine hydrochloride) PAH/PSS > PDADMAC/ Poly(acrylic acid) PAA >> PAH/PAA [42]. PDADMAC/PSS, for instance, permits higher flux than PAH/PSS due to higher degree of swelling [43]. In addition to swelling, highly porous support can also contribute to higher flux because it helps in reducing membrane resistance. Several methods have been used to produce highly porous support such as using electrospun nanofiber [44] or incorporating nanoparticles such as montmorillonite (MMT) or silica gel (SG). For example, poly‐ vinyl alcohol (PVA) mixed with 1 wt% montmorillonite (MMT) can achieve porosity of 80% [45].
time. This technique is quite useful for hollow fibre membrane as the support [41].
Likewise, PAN support mixed with 1 wt% SG can also achieve porosity of 80% [46].
prevent the foulants from entering the membrane from the support layer side.
The application of LbL assembly for NF is quite promising as can be seen in the previous sec‐ tion. There have been many works done and are ongoing in that area; meanwhile, the use of LbL assembly for fabricating FO is still at its infancy. The main results from FO mode of testing of sev‐ eral LbL FO membranes can be seen in **Table 2**. It is important to note that the information given in **Table 2** is only part of the works that showed the best performance in terms of water flux (WF) and salt to water flux ratio (SWFR) with active layer facing draw solution (ALDS) testing mode. It is known that ALDS mode shows higher initial water flux compared to ALFS mode, but it is also prone to internal fouling because foulants enter the porous support easily. This type of fouling in fact is much more difficult to be cleaned. In order to overcome this problem, a double‐skinned design was proposed [49]. Double‐skinned layer approach uses active layer deposited on both surfaces with the support layer, which will be in between as can be seen in **Figure 3**. Hence, it will
*3.1.2. Forward osmosis*
68 Desalination
interface that have significant effect in ion rejection [38, 39].
So far, the preparation of LbL membrane is mainly from polyelectrolytes. Recently, the researcher has started utilizing the technique for fabrication of LbL membrane using meth‐ ylphenylene diamine (MPD) and trimesoyl chloride (TMC), two most commonly used
4 All reported default testing condition is DS = 0.5 M MgCl<sup>2</sup> , FS = pure water, cross flow system, unless it is mentioned differently. πNaCl at 0.5 M = 22.74 atm and *π* MgCl <sup>2</sup> at 0.5 M = 32.65 atm.
**Table 2.** The work progress for FO prepared from polyelectrolyte multilayer membrane using LbL assembly.
**Figure 3.** A conceptual illustration of doubled‐skin LbL membrane (adapted from Ref. [49]).
PSF supports showed salt rejection of around 81% with water permeability of 0.35 L/m<sup>2</sup>
active area of 13.85 cm2
with active area of 42 cm2
88% and water permeability of 0.22 L/m2
[59].
**Figure 5.** A schematic diagram for m‐LbL process (adapted from Ref. [54]).
tested at a pressure of 20 bars and NaCl concentration of 2000 ppm using cross flow cells with
From the above literature review, almost all RO and FO LbL membranes were tested at low salt concentrations. Meanwhile, RO and FO membranes are normally used to treat water with rather high salt concentrations. There has been an issue with LbL film stability when it is exposed to such solution. Some PEMs are quite stable even without crosslinking. For instance, it was reported that 35 layers of PAH/PAA were successfully deposited on PSF support using SA‐LbL and tested at pressure of 48.6 bars and NaCl concentration of 15,000 ppm using cross flow cells
apply SA‐LbL for fabricating RO membrane [60]. Besides that, the use of Sr‐LbL to fabricate RO membrane was also investigated, for example, assembly of clay (laponite (LAP)) and polyelec‐ trolyte multilayers on the top of PSF support. (PAH/PAA) (PAH/LAP) deposited at pH 5 and tested at pressure of 18 bars, with NaCl concentration of 10,000 ppm using dead end permeation
As previously mentioned, the use of m‐LbL for RO membrane was also studied (see **Figure 5**). Using m‐LbL, the membrane performance can be finely tuned by simply varying the number of layers. In general, as the number of layers increases, the flux decreases and rejection increases as can be seen in **Figure 6**. For instance, with 10 layers of m‐LbL only, salt rejection of around 96% coupled with flux of 82% higher than IP TFC PA membrane was attained. In addition to this remarkable result, the surface roughness of this m‐LbL membrane was only 3.4 nm, which was much smoother than surface roughness of IP PA TFC (i.e. 45.1 nm in this study) [54].
Recently, the same group of researchers investigated the role of interaction between m‐LbL as active layer and the support layer, that is, PAN by showing the performance difference
between hydrolyzed PAN (HPAN) and non‐hydrolyzed PAN (see **Figure 7**) [62].
cells showed salt rejection of 89% with water permeability of 2.82 x 10−13 m2
. The test showed very stable performance with salt rejection of around
.h.bar. This work was considered the first attempt to
Application of Multilayer Thin Film Technology in Desalination Membrane
http://dx.doi.org/10.5772/intechopen.68375
/Pa.s. [61].
.h.bar
71
**Figure 4.** A schematic mechanism of crosslinking (A) using GA linker and (B) using UV light (adapted from Ref. [51]).
monomers for PA. This method is known as molecular layer‐by‐layer assembly (m‐LbL) [53]. The mechanism of m‐LbL is shown in **Figure 5**. Compared to several typical LbL FO and commercial FO membranes, m‐LbL FO membrane showed superior performance for both ALDS and ALFS modes. Having said that, m‐LbL FO membrane still has problems with stability and irreversible defect, especially after exposing it to very high salt concentration solution, for example, at 2 M NaCl. This instability is most likely due to relatively weak inter‐ action between m‐LbL film and the support. One of the approaches to enhance this interac‐ tion is by employing interlayers that will be discussed in more detail in the RO section.
#### *3.1.3. Reverse osmosis*
The use of LbL assembly for fabricating RO membrane is one of the most challenging appli‐ cations because a typical RO membrane is tested at relatively high salt concentration up to 32,000 ppm with high operating pressure. Many polyelectrolyte multilayer films are not sta‐ ble when exposed to such conditions.
The earliest work in this field to the author's knowledge was done by Jin *et al.* in 2003 [58]. Polyvinyl Amine (PVAm)/Polyvinyl sulphate (PVS) was used in their work and deposited on PAN/PET support using d‐LbL assembly. Membrane performance was examined at a pressure of 40 bars, with NaCl concentration of 584.4 ppm using dead end cells with mem‐ brane active area of 36.8 cm2 . The result showed salt rejection of 93.5% and permeability of around 0.11 L/m2 .h.bar. Similar work using thermally induced PAH/PAA deposited onto Application of Multilayer Thin Film Technology in Desalination Membrane http://dx.doi.org/10.5772/intechopen.68375 71
**Figure 5.** A schematic diagram for m‐LbL process (adapted from Ref. [54]).
**Figure 4.** A schematic mechanism of crosslinking (A) using GA linker and (B) using UV light (adapted from Ref. [51]).
**Figure 3.** A conceptual illustration of doubled‐skin LbL membrane (adapted from Ref. [49]).
monomers for PA. This method is known as molecular layer‐by‐layer assembly (m‐LbL) [53]. The mechanism of m‐LbL is shown in **Figure 5**. Compared to several typical LbL FO and commercial FO membranes, m‐LbL FO membrane showed superior performance for both ALDS and ALFS modes. Having said that, m‐LbL FO membrane still has problems with stability and irreversible defect, especially after exposing it to very high salt concentration solution, for example, at 2 M NaCl. This instability is most likely due to relatively weak inter‐ action between m‐LbL film and the support. One of the approaches to enhance this interac‐ tion is by employing interlayers that will be discussed in more detail in the RO section.
The use of LbL assembly for fabricating RO membrane is one of the most challenging appli‐ cations because a typical RO membrane is tested at relatively high salt concentration up to 32,000 ppm with high operating pressure. Many polyelectrolyte multilayer films are not sta‐
The earliest work in this field to the author's knowledge was done by Jin *et al.* in 2003 [58]. Polyvinyl Amine (PVAm)/Polyvinyl sulphate (PVS) was used in their work and deposited on PAN/PET support using d‐LbL assembly. Membrane performance was examined at a pressure of 40 bars, with NaCl concentration of 584.4 ppm using dead end cells with mem‐
. The result showed salt rejection of 93.5% and permeability of
.h.bar. Similar work using thermally induced PAH/PAA deposited onto
*3.1.3. Reverse osmosis*
70 Desalination
ble when exposed to such conditions.
brane active area of 36.8 cm2
around 0.11 L/m2
PSF supports showed salt rejection of around 81% with water permeability of 0.35 L/m<sup>2</sup> .h.bar tested at a pressure of 20 bars and NaCl concentration of 2000 ppm using cross flow cells with active area of 13.85 cm2 [59].
From the above literature review, almost all RO and FO LbL membranes were tested at low salt concentrations. Meanwhile, RO and FO membranes are normally used to treat water with rather high salt concentrations. There has been an issue with LbL film stability when it is exposed to such solution. Some PEMs are quite stable even without crosslinking. For instance, it was reported that 35 layers of PAH/PAA were successfully deposited on PSF support using SA‐LbL and tested at pressure of 48.6 bars and NaCl concentration of 15,000 ppm using cross flow cells with active area of 42 cm2 . The test showed very stable performance with salt rejection of around 88% and water permeability of 0.22 L/m2 .h.bar. This work was considered the first attempt to apply SA‐LbL for fabricating RO membrane [60]. Besides that, the use of Sr‐LbL to fabricate RO membrane was also investigated, for example, assembly of clay (laponite (LAP)) and polyelec‐ trolyte multilayers on the top of PSF support. (PAH/PAA) (PAH/LAP) deposited at pH 5 and tested at pressure of 18 bars, with NaCl concentration of 10,000 ppm using dead end permeation cells showed salt rejection of 89% with water permeability of 2.82 x 10−13 m2 /Pa.s. [61].
As previously mentioned, the use of m‐LbL for RO membrane was also studied (see **Figure 5**). Using m‐LbL, the membrane performance can be finely tuned by simply varying the number of layers. In general, as the number of layers increases, the flux decreases and rejection increases as can be seen in **Figure 6**. For instance, with 10 layers of m‐LbL only, salt rejection of around 96% coupled with flux of 82% higher than IP TFC PA membrane was attained. In addition to this remarkable result, the surface roughness of this m‐LbL membrane was only 3.4 nm, which was much smoother than surface roughness of IP PA TFC (i.e. 45.1 nm in this study) [54].
Recently, the same group of researchers investigated the role of interaction between m‐LbL as active layer and the support layer, that is, PAN by showing the performance difference between hydrolyzed PAN (HPAN) and non‐hydrolyzed PAN (see **Figure 7**) [62].
**Figure 6.** (a) Water flux (Jw, filled symbols) and NaCl rejection (unfilled symbols) as functions of a number of layers. (b) Normalized flux as a function of a number of bilayers (membrane was tested at NaCl concentration of 2000 ppm and pressure of 15.5 bars; adapted from Ref. [54]).
**3.2. Membrane modification**
PEI‐HPAN (right figure, unfilled symbol) (adapted from Ref. [62]).
In this section, the use of LbL assembly to modify the surface of existing RO, NF or FO membranes will be discussed. One of the most important surface modification purposes is to protect the membrane from being directly exposed to harsh environment such as chlorine or oxidants content in water, foulants and pH of the water at which the membrane life will be shortened. There have been tremendous works in the modification of commercially available membranes using different approaches such as coating, blending, incorporating nanomateri‐ als, functionalization, grafting, etc. [63]. This section is devoted only for membrane modifi‐ cation using LbL assembly, which is mostly a sort of coating technique that can be simply a
**Figure 8.** m‐LbL PA TFC performance (represented by a circle for water flux and triangle for rejection). PAN support (left figure, filled symbol), HPAN (left figure, unfilled symbol), PEI/PAA‐HPAN support (right figure, filled symbol) and
Application of Multilayer Thin Film Technology in Desalination Membrane
http://dx.doi.org/10.5772/intechopen.68375
73
The idea behind modification of the RO membrane using LbL is basically to alter surface prop‐ erties that can reduce fouling tendency for example by reducing the roughness and hydro‐ phobicity. The study showed that additional LbL films improve, significantly, the smoothness of the surface, improve salt rejection but sacrifice the flux. However, as far as fouling is con‐ cerned, this surface modification is very promising as no flux decrease was noticed when the
Recently, graphene oxide (GO), a single‐sheet functionalized graphene with oxygen‐rich functional group, has attracted attention from researchers in the field of water treatment. This material provides fast water transport, hydrophilicity as well as excellent chemical sta‐ bility. The main objectives of employing GO or GO/polyelectrolyte are to improve chlorine resistance and reduce the bio‐fouling without sacrificing the flux (see **Figure 9** for the sche‐ matic illustration of GO‐modified membrane) [65, 66]. Biofouling test using bovine serum albumin (BSA) showed that GO‐modified PA TFC membrane can maintain the flux constant after approximately 6 h of testing; meanwhile, pristine PA TFC membrane kept showing
physical process or chemically bonded to the membrane surface.
membrane was exposed to foulants containing water [64].
**Figure 7.** A conceptual design of m‐LbL TFC PA membrane with interlayer (adapted from Ref. [62]).
As can be seen in **Figure 8**, deposition of m‐LbL PA on the top of HPAN resulted in much higher rejection than that on the top of PAN. This improvement occurred due to more uni‐ form active layers deposited on HPAN as a result of better interaction between carboxylate group of HPAN with amine group of MPD. Even though the interaction was better, in gen‐ eral, the active layer was still found to be insufficiently dense and selective towards NaCl so the rejection was still lower than IP PA TFC. The extent of uniformity then was further increased by additional interlayers. PEI/PAA interlayer on the top of HPAN showed by far the best performance. This was due to more carboxylate groups provided by PAA as the outermost layer prior to the deposition of MPD/TMC. Carboxylate groups of PAA will serve as seeding sites for subsequent deposition of MPD/TMC. The membrane with PEI/PAA interlayer showed remarkable results, that is, rejection of 98.7% coupled with water flux of 20.7 L/m2 .h which is 75% higher than IP PA TFC (the membrane was tested at pressure of 15.5 bars and NaCl concentration of 2000 ppm). This is indeed a very promising result in the field of RO.
Application of Multilayer Thin Film Technology in Desalination Membrane http://dx.doi.org/10.5772/intechopen.68375 73
**Figure 8.** m‐LbL PA TFC performance (represented by a circle for water flux and triangle for rejection). PAN support (left figure, filled symbol), HPAN (left figure, unfilled symbol), PEI/PAA‐HPAN support (right figure, filled symbol) and PEI‐HPAN (right figure, unfilled symbol) (adapted from Ref. [62]).
#### **3.2. Membrane modification**
**Figure 7.** A conceptual design of m‐LbL TFC PA membrane with interlayer (adapted from Ref. [62]).
20.7 L/m2
the field of RO.
pressure of 15.5 bars; adapted from Ref. [54]).
72 Desalination
As can be seen in **Figure 8**, deposition of m‐LbL PA on the top of HPAN resulted in much higher rejection than that on the top of PAN. This improvement occurred due to more uni‐ form active layers deposited on HPAN as a result of better interaction between carboxylate group of HPAN with amine group of MPD. Even though the interaction was better, in gen‐ eral, the active layer was still found to be insufficiently dense and selective towards NaCl so the rejection was still lower than IP PA TFC. The extent of uniformity then was further increased by additional interlayers. PEI/PAA interlayer on the top of HPAN showed by far the best performance. This was due to more carboxylate groups provided by PAA as the outermost layer prior to the deposition of MPD/TMC. Carboxylate groups of PAA will serve as seeding sites for subsequent deposition of MPD/TMC. The membrane with PEI/PAA interlayer showed remarkable results, that is, rejection of 98.7% coupled with water flux of
**Figure 6.** (a) Water flux (Jw, filled symbols) and NaCl rejection (unfilled symbols) as functions of a number of layers. (b) Normalized flux as a function of a number of bilayers (membrane was tested at NaCl concentration of 2000 ppm and
.h which is 75% higher than IP PA TFC (the membrane was tested at pressure of
15.5 bars and NaCl concentration of 2000 ppm). This is indeed a very promising result in
In this section, the use of LbL assembly to modify the surface of existing RO, NF or FO membranes will be discussed. One of the most important surface modification purposes is to protect the membrane from being directly exposed to harsh environment such as chlorine or oxidants content in water, foulants and pH of the water at which the membrane life will be shortened. There have been tremendous works in the modification of commercially available membranes using different approaches such as coating, blending, incorporating nanomateri‐ als, functionalization, grafting, etc. [63]. This section is devoted only for membrane modifi‐ cation using LbL assembly, which is mostly a sort of coating technique that can be simply a physical process or chemically bonded to the membrane surface.
The idea behind modification of the RO membrane using LbL is basically to alter surface prop‐ erties that can reduce fouling tendency for example by reducing the roughness and hydro‐ phobicity. The study showed that additional LbL films improve, significantly, the smoothness of the surface, improve salt rejection but sacrifice the flux. However, as far as fouling is con‐ cerned, this surface modification is very promising as no flux decrease was noticed when the membrane was exposed to foulants containing water [64].
Recently, graphene oxide (GO), a single‐sheet functionalized graphene with oxygen‐rich functional group, has attracted attention from researchers in the field of water treatment. This material provides fast water transport, hydrophilicity as well as excellent chemical sta‐ bility. The main objectives of employing GO or GO/polyelectrolyte are to improve chlorine resistance and reduce the bio‐fouling without sacrificing the flux (see **Figure 9** for the sche‐ matic illustration of GO‐modified membrane) [65, 66]. Biofouling test using bovine serum albumin (BSA) showed that GO‐modified PA TFC membrane can maintain the flux constant after approximately 6 h of testing; meanwhile, pristine PA TFC membrane kept showing
studies have been done to investigate long‐term stability of those membranes and even to the best of our knowledge most of studies were done under very soft conditions such as low salt concentration, typically 1000–2000 ppm. Thus, testing in harsh conditions for example at NaCl concentration of 32,000 ppm for longer duration must be done particularly for those LbL membranes intended to be used in RO or FO applications. The result of this study will definitely drive the research in finding the best protocol to create the most stable
Application of Multilayer Thin Film Technology in Desalination Membrane
http://dx.doi.org/10.5772/intechopen.68375
75
Also, from literature study, we found that most of the works were done using traditional dip‐LbL. This method is quite difficult to be brought to an industrial scale as the adsorption process limited by diffusion of polyelectrolyte onto the support surface is a time‐consuming process. Several researchers have migrated to other LbL assemblies such as dynamic LbL, semi‐dynamic LbL, Sr‐LbL and SA‐LbL that utilize external force to speed up the adsorption
process. Surely, more extensive works are urgently required for those LbL assemblies.
One of the advantages of LbL assembly that has not been thoroughly observed was the flex‐ ibility and applicability of this technique to create ultra‐thin films from various materials. To the best of our knowledge, most of the works are still focused on polyelectrolyte. It is true some researchers have started introducing some inorganic nanomaterials such as silver nanoparticles, graphene oxide and clay, but still there is huge space available to do research in terms of membrane material. There were many polymers that have never been investigated because no appropriate technology was applicable to prepare ultra‐thin films using those polymers. After the rediscovery of LbL, the door has opened. Using LbL assembly, one can create the film either from the polymer itself or from the monomers as in the case of m‐LbL. One can also combine organic and inorganic materials with nano‐level control easily to fab‐ ricate highly resistant membranes towards chlorine and foulants and, at the same time, give high flux and high rejection or selectivity for instance. Tailoring the film properties is one of the strength and advantages of LbL assembly that has not been deeply investigated but for
As a new emerging technology, LbL must still go a long journey; it is going to face many chal‐ lenges in the future. However, with all of its strength, versatility and robustness, we believe that LbL membrane one day will dominate the desalination membrane just as IP PA TFC did
2 Chemical Engineering Department, Al Imam Mohammad Ibn Saud Islamic University
LbL membrane.
preparing separation membrane.
in the last couple of decades.
\* and Farid Fadhillah<sup>2</sup>
1 Center for Advanced Materials, Qatar University, Qatar
\*Address all correspondence to: szaidi@qu.edu.qa
(IMSIU), Riyadh, Kingdom of Saudi Arabia
**Author details**
Syed Javaid Zaidi<sup>1</sup>
**Figure 9.** A schematic illustration of a multilayer GO on PA TFC membrane (adapted from Ref. [65]).
linear flux reduction even after 12 h of testing. This result showed that GO‐modified PA TFC improves the fouling resistance but does not completely remove the fouling as indicated by flux decrease for the first 6 h of the experiment [65].
## **4. Conclusion and future work**
IP PA TFC membrane has been extensively used in the area of RO, NF and FO. However, the nature of the process as well as the properties of the polyamide creates several problems such as chlorine and fouling resistance that are likely difficult to solve as long as the same fabrication technique is used. This is because the IP itself does not provide fine control over the film properties, not to mention its limited applicability only to few types of polymers such as polyamide and polycarbonate. Meanwhile, polyamide itself is naturally hydropho‐ bic which causes severe fouling problems and has weak resistance against chlorine and oxi‐ dants. Relatively new technology, LbL assembly, offers flexibility and great control over the film properties which are the main keys to overcome aforementioned problems associated with PA TFC.
Based on literature study, most of the work in LbL desalination membrane still focuses in employing polyelectrolyte to form an active layer. Meanwhile, it has been investigated that many polyelectrolytes are highly hydrophilic in nature, they swell a lot when exposed to water and even swelling can be aggravated by salt infiltration [67] while those are two main compounds in saline water. Some attempts have been done to maintain stability of polyelectrolyte such as heat‐induced, chemical or UV light crosslinking. However, only few studies have been done to investigate long‐term stability of those membranes and even to the best of our knowledge most of studies were done under very soft conditions such as low salt concentration, typically 1000–2000 ppm. Thus, testing in harsh conditions for example at NaCl concentration of 32,000 ppm for longer duration must be done particularly for those LbL membranes intended to be used in RO or FO applications. The result of this study will definitely drive the research in finding the best protocol to create the most stable LbL membrane.
Also, from literature study, we found that most of the works were done using traditional dip‐LbL. This method is quite difficult to be brought to an industrial scale as the adsorption process limited by diffusion of polyelectrolyte onto the support surface is a time‐consuming process. Several researchers have migrated to other LbL assemblies such as dynamic LbL, semi‐dynamic LbL, Sr‐LbL and SA‐LbL that utilize external force to speed up the adsorption process. Surely, more extensive works are urgently required for those LbL assemblies.
One of the advantages of LbL assembly that has not been thoroughly observed was the flex‐ ibility and applicability of this technique to create ultra‐thin films from various materials. To the best of our knowledge, most of the works are still focused on polyelectrolyte. It is true some researchers have started introducing some inorganic nanomaterials such as silver nanoparticles, graphene oxide and clay, but still there is huge space available to do research in terms of membrane material. There were many polymers that have never been investigated because no appropriate technology was applicable to prepare ultra‐thin films using those polymers. After the rediscovery of LbL, the door has opened. Using LbL assembly, one can create the film either from the polymer itself or from the monomers as in the case of m‐LbL. One can also combine organic and inorganic materials with nano‐level control easily to fab‐ ricate highly resistant membranes towards chlorine and foulants and, at the same time, give high flux and high rejection or selectivity for instance. Tailoring the film properties is one of the strength and advantages of LbL assembly that has not been deeply investigated but for preparing separation membrane.
As a new emerging technology, LbL must still go a long journey; it is going to face many chal‐ lenges in the future. However, with all of its strength, versatility and robustness, we believe that LbL membrane one day will dominate the desalination membrane just as IP PA TFC did in the last couple of decades.
## **Author details**
linear flux reduction even after 12 h of testing. This result showed that GO‐modified PA TFC improves the fouling resistance but does not completely remove the fouling as indicated by
**Figure 9.** A schematic illustration of a multilayer GO on PA TFC membrane (adapted from Ref. [65]).
IP PA TFC membrane has been extensively used in the area of RO, NF and FO. However, the nature of the process as well as the properties of the polyamide creates several problems such as chlorine and fouling resistance that are likely difficult to solve as long as the same fabrication technique is used. This is because the IP itself does not provide fine control over the film properties, not to mention its limited applicability only to few types of polymers such as polyamide and polycarbonate. Meanwhile, polyamide itself is naturally hydropho‐ bic which causes severe fouling problems and has weak resistance against chlorine and oxi‐ dants. Relatively new technology, LbL assembly, offers flexibility and great control over the film properties which are the main keys to overcome aforementioned problems associated
Based on literature study, most of the work in LbL desalination membrane still focuses in employing polyelectrolyte to form an active layer. Meanwhile, it has been investigated that many polyelectrolytes are highly hydrophilic in nature, they swell a lot when exposed to water and even swelling can be aggravated by salt infiltration [67] while those are two main compounds in saline water. Some attempts have been done to maintain stability of polyelectrolyte such as heat‐induced, chemical or UV light crosslinking. However, only few
flux decrease for the first 6 h of the experiment [65].
**4. Conclusion and future work**
with PA TFC.
74 Desalination
Syed Javaid Zaidi<sup>1</sup> \* and Farid Fadhillah<sup>2</sup>
\*Address all correspondence to: szaidi@qu.edu.qa
1 Center for Advanced Materials, Qatar University, Qatar
2 Chemical Engineering Department, Al Imam Mohammad Ibn Saud Islamic University (IMSIU), Riyadh, Kingdom of Saudi Arabia
## **References**
[1] Executive Summary of Vital Water Graphics: An overview of the state of the world's Fresh and Marine Waters, 2nd Edition, 2008. Available from: http://new.unep.org/dewa/vitalwa‐ ter/article186.html
[15] Shimazaki Y, Mitsuishi M, Ito S, Yamamoto M. Preparation and characterization of the layer‐by‐layer deposited ultrathin film based on the charge‐transfer interaction in
Application of Multilayer Thin Film Technology in Desalination Membrane
http://dx.doi.org/10.5772/intechopen.68375
77
[16] Shimazaki Y, Mitsuishi M, Ito S, Yamamoto M. Preparation of the layer‐by‐layer deposited ultrathin film based on the charge‐transfer interaction. Langmuir. 1997;**13**(6):1385‐1387
[17] DeLongchamp DM, Hammond PT. Highly ion conductive poly(ethylene oxide)‐based solid polymer electrolytes from hydrogen bonding layer‐by‐layer assembly. Langmuir.
[18] Lee H, Mensire R, Cohen RE, Rubner MF. Strategies for hydrogen bonding based layer‐by‐layer assembly of poly(vinyl alcohol) with weak polyacids. Macromolecules. 2012;**45**(1):347‐355
[19] Tian Y, He Q, Tao C, Li J. Fabrication of fluorescent nanotubes based on layer‐by‐layer
[20] Liang Z, Dzienis KL, Xu J, Wang Q. Covalent layer‐by‐layer assembly of conjugated polymers and CdSe nanoparticles: multilayer structure and photovoltaic properties.
[21] Bechler SL, Lynn DM. Reactive polymer multilayers fabricated by covalent layer‐by‐ layer assembly: 1,4‐conjugate addition‐based approaches to the design of functional bio‐
[22] Hamada K, Serizawa T, Kitayama T, Fujimoto N, Hatada K, Akashi M. Stepwise ste‐ reocomplex assembly of isotactic poly(methyl methacrylate) and syndiotactic poly(alkyl
[23] Serizawa T, Hamada K, Kitayama T, Katsukawa K, Hatada K, Akashi M. Stepwise assembly of isotactic poly(methyl methacrylate) and syndiotactic poly(methacrylic acid)
[24] Helm C, Lösche M, Möhwald H, Decher G, Schmitt J. Fine‐Tuning of the film thickness of ultrathin multilayer films composed of consecutively alternating layers of anionic and cationic polyelectrolytes. In: Trends in Colloid and Interface Science VI. Berlin/
[25] Shiratori SS, Rubner MF. pH‐dependent thickness behavior of sequentially adsorbed
[26] Büscher K, Graf K, Ahrens H, Helm CA. Influence of adsorption conditions on the struc‐
[27] Schlenoff JB, Dubas ST, Farhat T. Sprayed polyelectrolyte multilayers. Langmuir.
[28] Nogueira GM, Banerjee D, Cohen RE, Rubner MF. Spray‐layer‐by‐layer assembly can more rapidly produce optical‐quality multistack heterostructures. Langmuir. 2011;
layers of weak polyelectrolytes. Macromolecules. 2000;**33**(11):4213‐4219
ture of polyelectrolyte multilayers. Langmuir. 2002;**18**(9):3585‐3591
organic solvents. Langmuir. 1998;**14**(10):2768‐2773
assembly via covalent bond. Langmuir. 2005;**22**(1):360‐362
Advanced Functional Materials. 2006;**16**(4):542‐548
interfaces. Biomacromolecules. 2012;**13**(5):1523‐1532
on a substrate. Langmuir. 2000;**16**(18):7112‐7115
Heidelberg: Springer; 1992. pp. 160‐164.
2000;**16**(26):9968‐9969
**27**(12):7860‐7.
methacrylate)s on surfaces. Langmuir. 2001;**17**(18):5513‐5519
2004;**20**(13):5403‐5411
[15] Shimazaki Y, Mitsuishi M, Ito S, Yamamoto M. Preparation and characterization of the layer‐by‐layer deposited ultrathin film based on the charge‐transfer interaction in organic solvents. Langmuir. 1998;**14**(10):2768‐2773
**References**
76 Desalination
ter/article186.html
factsheets/fs391/en/
**21**(6):569‐594
Membrane Science. 1993;**83**:81‐150
poly(Allylamine). Langmuir. 1993;**9**(2):481‐486
Macromolecules. 1993;**26**(25):7058‐7063
Advanced Materials. 1997;**9**(1):61‐65.
Thin Solid Films. 1996;**284**(285):220‐223
1994;**10**(11):4232‐4236.
1997;**277**(5330):1232‐1237
[1] Executive Summary of Vital Water Graphics: An overview of the state of the world's Fresh and Marine Waters, 2nd Edition, 2008. Available from: http://new.unep.org/dewa/vitalwa‐
[2] Dringking Water, Fact sheet, 2016, available from: http://www.who.int/mediacentre/
[3] Shannon MA, Bohn PW, Elimelech M, Georgiadis JG, Marinas BJ, Mayes AM. Science and technology for water purification in the coming decades. Nature. 2008;**452**(7185):301‐310
[4] Greenlee LF, Lawler DF, Freeman BD, Marrot B, Moulin P. Reverse osmosis desalination: Water sources, technology, and today's challenges. Water Research. 2009;**43**(9):2317‐2348
[5] Cadotte JE. Interfacially synthesized reverse osmosis membrane [Internet]. Google
[6] Petersen RJ. Composite reverse osmosis and nanofiltration membranes. Journal of
[7] Iler RK. Multilayers of colloidal particles. Journal of Colloid and Interface Science. 1966;
[8] Decher G, Hong JD, Schmitt J. Buildup of ultrathin multilayer films by a self‐assembly process: III. Consecutively alternating adsorption of anionic and cationic polyelectro‐
[9] Lvov Y, Decher G, Mohwald H. Assembly, structural characterization, and ther‐ mal‐behavior of layer‐by‐layer deposited ultrathin films of poly(vinyl sulfate) and
[10] Schmitt J, Grunewald T, Decher G, Pershan PS, Kjaer K, Losche M. Internal structure of layer‐by‐layer adsorbed polyelectrolyte ilms‐a neutran and X‐ray reflectivity study.
[11] Lvov Y, Haas H, Decher G, Mohwald H, Mikhailov A, Mtchedlishvily B, et al. Successive deposition of alternate layers of polyelectrolytes and a charged virus. Langmuir.
[12] Schmitt J, Decher G, Dressick WJ, Brandow SL, Geer RE, Shashidhar R, et al. Metal nanoparticle/polymer superlattice films: Fabrication and control of layer structure.
[13] Sukhorukov GB, M◌hwald H, Decher G, Lvov YM. Assembly of polyelectrolyte multi‐ layer films by consec1utively alternating adsorption of polynucleotides and polycations.
[14] Decher G. Fuzzy nanoassemblies: Toward layered polymeric multicomposites. Science.
Patents, 1981. Available from: https://www.google.com/patents/US4277344
lytes on charged surfaces. Thin Solid Films. 1992;**210‐211**(Part 2):831‐835
[29] Krogman KC, Zacharia NS, Schroeder S, Hammond PT. Automated process for improved uniformity and versatility of layer‐by‐layer deposition. Langmuir. 2007;**23**(6):3137‐3141
[43] Hong SU, Ouyang L, Bruening ML. Recovery of phosphate using multilayer polyelectro‐ lyte nanofiltration membranes. Journal of Membrane Science. 2009;**327**(1‐2):2‐5
Application of Multilayer Thin Film Technology in Desalination Membrane
http://dx.doi.org/10.5772/intechopen.68375
79
[44] Ritcharoen W, Supaphol P, Pavasant P. Development of polyelectrolyte multilayer‐coated electrospun cellulose acetate fiber mat as composite membranes. European Polymer
[45] Pardeshi P, Mungray AA. Synthesis, characterization and application of novel high flux FO membrane by layer‐by‐layer self‐assembled polyelectrolyte. Journal of Membrane
[46] Lee J‐Y, Qi S, Liu X, Li Y, Huo F, Tang CY. Synthesis and characterization of silica gel–poly‐ acrylonitrile mixed matrix forward osmosis membranes based on layer‐by‐layer assem‐ bly. Separation and Purification Technology. 2014;**124**:207‐216. Available from: http:// www.scopus.com/inward/record.url?eid=2‐s2.0‐84894071531&partnerID=tZOtx3y1 [47] Stair JL, Harris JJ, Bruening ML. Enhancement of the ion‐transport selectivity of lay‐ ered polyelectrolyte membranes through crosslinking and hybridization. Chemistry of
[48] Malaisamy R, Bruening ML. High‐flux nanofiltration membranes prepared by adsorp‐ tion of multilayer polyelectrolyte membranes on polymeric supports. Langmuir. 2005;**21**
[49] Qi S, Qiu CQ, Zhao Y, Tang CY. Double‐skinned forward osmosis membranes based on layer‐by‐layer assembly‐FO performance and fouling behavior. Journal of Membrane
[50] Qiu C, Qi S, Tang CY. Synthesis of high flux forward osmosis membranes by chemically crosslinked layer‐by‐layer polyelectrolytes. Journal of Membrane Science. **381**(1‐2):74‐80
[51] Duong PHH, Zuo J, Chung T‐S. Highly crosslinked layer‐by‐layer polyelectrolyte FO membranes: Understanding effects of salt concentration and deposition time on FO per‐
[52] Harris JJ, DeRose PM, Bruening ML. Synthesis of passivating, nylon‐like coatings through cross‐linking of ultrathin polyelectrolyte films. Journal of the American Chemical Society. 1999;**121**(9):1978‐1979. Available from: http://dx.doi.org/10.1021/ja9833467 [53] Kwon S‐B, Lee JS, Kwon SJ, Yun S‐T, Lee S, Lee J‐H. Molecular layer‐by‐layer assembled forward osmosis membranes. Journal of Membrane Science. 2015;**488**:111‐120
[54] Gu J‐E, Lee S, Stafford CM, Lee JS, Choi W, Kim B‐Y, et al. Molecular layer‐by‐layer assembled thin‐film composite membranes for water desalination. Advanced Materials.
2013;**25**(34):4778‐4782. Available from: http://dx.doi.org/10.1002/adma.201302030 [55] Liu X, Qi S, Li Y, Yang L, Cao B, Tang CY. Synthesis and characterization of novel anti‐ bacterial silver nanocomposite nanofiltration and forward osmosis membranes based on layer‐by‐layer assembly. Water Research. 2013;**47**(9):3081‐3092. Available from: http:// www.scopus.com/inward/record.url?eid=2‐s2.0‐84876699477&partnerID=tZOtx3y1
formance. Journal of Membrane Science. 2013; **427**:411‐21.
Journal. 2008;**44**(12):3963‐3968
Materials. 2001;**13**(8):2641‐2648
(23):10587‐10592
Science. 2012; **405**:20‐9
Science. 2014;**453**:202‐211
[43] Hong SU, Ouyang L, Bruening ML. Recovery of phosphate using multilayer polyelectro‐ lyte nanofiltration membranes. Journal of Membrane Science. 2009;**327**(1‐2):2‐5
[29] Krogman KC, Zacharia NS, Schroeder S, Hammond PT. Automated process for improved uniformity and versatility of layer‐by‐layer deposition. Langmuir. 2007;**23**(6):3137‐3141
[30] Izquierdo A, Ono SS, Voegel JC, Schaaf P, Decher G. Dipping versus spraying: Exploring the deposition conditions for speeding up layer‐by‐layer assembly. Langmuir. 2005;**21**(16):
[31] Cho J, Char K, Hong JD, Lee KB. Fabrication of highly ordered multilayer films using a
[32] Lee S‐S, Hong J‐D, Kim CH, Kim K, Koo JP, Lee K‐B. Layer‐by‐layer deposited mul‐ tilayer assemblies of ionene‐type polyelectrolytes based on the spin‐coating method.
[33] Chiarelli PA, Johal MS, Casson JL, Roberts JB, Robinson JM, Wang HL. Controlled fabri‐ cation of polyelectrolyte multilayer thin films using spin‐assembly. Advanced Materials.
[34] Kharlampieva E, Kozlovskaya V, Chan J, Ankner JF, Tsukruk VV. Spin‐assisted layer‐by‐ layer assembly: Variation of stratification as studied with neutron reflectivity. Langmuir.
[35] Krasemann L, Tieke B. Selective ion transport across self‐assembled alternating multi‐ layers of cationic and anionic polyelectrolytes. Langmuir. 1999;**16**(2):287‐290
[36] Farhat TR, Schlenoff JB. Ion transport and equilibria in polyelectrolyte multilayers. Langmuir. 2001;**17**(4):1184‐1192. Available from: http://dx.doi.org/10.1021/la001298+
[37] Fadhillah F, Zaidi SMJ, Khan Z, Khaled MM, Rahman F, Hammond PT. Development of polyelectrolyte multilayer thin film composite membrane for water desalination applica‐
[38] Harris JJ, Stair JL, Bruening ML. Layered polyelectrolyte films as selective, ultrathin bar‐
[39] Stanton BW, Harris JJ, Miller MD, Bruening ML. Ultrathin, multilayered polyelectrolyte
[40] Deng H‐Y, Xu Y‐Y, Zhu B‐K, Wei X‐Z, Liu F, Cui Z‐Y. Polyelectrolyte membranes prepared by dynamic self‐assembly of poly (4‐styrenesulfonic acid‐co‐maleic acid) sodium salt (PSSMA) for nanofiltration (I). Journal of Membrane Science. 2008;**323**(1):
[41] Liu C, Shi L, Wang R. Enhanced hollow fiber membrane performance via semi‐dynamic layer‐by‐layer polyelectrolyte inner surface deposition for nanofiltration and forward
[42] Hong SU, Malaisamy R, Bruening ML. Optimization of flux and selectivity in Cl‐/SO42‐ separations with multilayer polyelectrolyte membranes. Journal of Membrane Science.
osmosis applications. Reactive and Functional Polymers. 2015;**86**:154‐160
riers for anion transport. Chemistry of Materials. 2000;**12**(7):1941‐1946
films as nanofiltration membranes. Langmuir. 2003;**19**(17):7038‐7042
spin self‐assembly method. Advanced Materials. 2001;**13**(14):1076‐1078
Macromolecules. 2001;**34**(16):5358‐5360
2001;**13**(15):1167‐1171
2009;**25**(24):14017‐14024
tion. Desalination. 2013;**318**:19‐24
125‐133
2006;**283**(1‐2):366‐372
7558‐7567
78 Desalination
[56] Liu C, Fang W, Chou S, Shi L, Fane AG, Wang R. Fabrication of layer‐by‐layer assem‐ bled FO hollow fiber membranes and their performances using low concentration draw solutions. Desalination. 2013;**308**:147‐153. Available from: http://www.scopus.com/ inward/record.url?eid=2‐s2.0‐84870728544&partnerID=tZOtx3y1
**Chapter 5**
**Phase Equilibria and Phase Separation of the Aqueous**
Brines including seawater, concentrated seawater after desalinization, salt lake, oil/gas water, and well bitter are widely distributed around the world. In order to promote the comprehensive utilization and effective protection of the valuable chemical resources existing in brines such as freshwater, lithium, sodium, potassium, and magnesium salts, the systematic foundation and application foundation research including phase equilibria and thermodynamic properties for the salt-water electrolyte solution are essential,
especially for solid lithium salts and their aqueous solution systems.
Keywords: thermodynamics, phase equilibria, aqueous solution, lithium salts
Lithium is the lightest alkali metal, which plays a growing role in numerous processes such as rechargeable batteries, thermonuclear fusion, medical drugs, lubricant greases, ceramic, glasses, dyes, adhesives, and electrode welding [1–19]. Lithium is a critical energy material and a strategic resource for the twenty-first century. Consequently, the market demands for lithium resources are
Lithium naturally occurs in compound forms because of its high reactivity. Economic concentrations of lithium are found in brines, minerals, and clays in various parts of the world. Brines and high-grade lithium ores are the present sources for all commercial lithium production. The global lithium reserve is estimated at 14.0 megatons [22], which is 74.5 megatons of lithium carbonate equivalent. Lithium reserves are mainly distributed in South America, China, and Australia [23].
> © 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
**Solution System Containing Lithium Ions**
Long Li, Yafei Guo and Tianlong Deng
http://dx.doi.org/10.5772/intechopen.68363
Abstract
1. Introduction
1.1. Lithium resources situation
increasing around the world [20, 21].
Additional information is available at the end of the chapter
## **Phase Equilibria and Phase Separation of the Aqueous Solution System Containing Lithium Ions**
Long Li, Yafei Guo and Tianlong Deng
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.68363
#### Abstract
[56] Liu C, Fang W, Chou S, Shi L, Fane AG, Wang R. Fabrication of layer‐by‐layer assem‐ bled FO hollow fiber membranes and their performances using low concentration draw solutions. Desalination. 2013;**308**:147‐153. Available from: http://www.scopus.com/
[57] Cui Y, Wang H, Wang H, Chung T‐S. Micro‐morphology and formation of layer‐by‐ layer membranes and their performance in osmotically driven processes. Chemical
[58] Jin W, Toutianoush A, Tieke B. Use of polyelectrolyte layer‐by‐layer assemblies as nano‐
[59] Park J, Park J, Kim SH, Cho J, Bang J. Desalination membranes from pH‐controlled and thermally‐crosslinked layer‐by‐layer assembled multilayers. Journals of Materials Chemistry. 2010;**20**(11):2085‐2091. Available from: http://dx.doi.org/10.1039/B918921A [60] Fadhillah F, Zaidi SMJ, Khan Z, Khaled MM, Hammond PT, Javaid Zaidi SM, et al. Reverse osmosis desalination membrane formed from weak polyelectrolytes by spin assisted layer by layer technique. Desalination and Water Treatment. 2011;**34**(1‐3):44‐49. Available from: http://www.tandfonline.com/doi/abs/10.5004/dwt.2011.2856%5Cnpapers2://publication/
[61] Kovacs JR, Liu C, Hammond PT. Spray layer‐by‐layer assembled clay composite thin films as selective layers in reverse osmosis membranes. ACS Applied Materials and Interfaces. 2015;**7**(24):13375‐13383. Available from: http://www.scopus.com/inward/
[62] Gu J‐E, Lee JS, Park S‐H, Kim IT, Chan EP, Kwon Y‐N, et al. Tailoring interlayer struc‐ ture of molecular layer‐by‐layer assembled polyamide membranes for high separation performance. Applied Surface Science. 2015;**356**:659‐667. Available from: http://www.
[63] Rana D, Matsuura T. Surface modifications for antifouling membranes. Chemical Reviews. 2010;**110**(4):2448‐2471. Available from: http://dx.doi.org/10.1021/cr800208y [64] Ishigami T, Amano K, Fujii A, Ohmukai Y, Kamio E, Maruyama T, et al. Fouling reduc‐ tion of reverse osmosis membrane by surface modification via layer‐by‐layer assembly. Separation Purification Technology. 2012;**99**:1‐7. Available from: http://www.sciencedi‐
[65] Choi W, Choi J, Bang J, Lee J‐H. Layer‐by‐layer assembly of graphene oxide nanosheets on polyamide membranes for durable reverse‐osmosis applications. ACS Applied Material Interfaces. 2013;**5**(23):12510‐12519. Available from: http://pubs.acs.org/doi/
[66] Hu M, Zheng S, Mi B. Organic fouling of graphene oxide membranes and its implica‐ tions for membrane fouling control in engineered osmosis. Environmental Science & Technology. 2016;**50**(2):685‐693. Available from: http://www.scopus.com/inward/record.
[67] Dubas ST, Schlenoff JB. Swelling and smoothing of polyelectrolyte multilayers by salt. Langmuir. 2001;**17**(25):7725‐7727. Available from: http://dx.doi.org/10.1021/la0112099
record.url?eid=2‐s2.0‐84932643621&partnerID=tZOtx3y1
sciencedirect.com/science/article/pii/S0169433215019327
rect.com/science/article/pii/S1383586612004133
url?eid=2‐s2.0‐84955322151&partnerID=tZOtx3y1
filtration and reverse osmosis membranes. Langmuir. 2003;**19**(7):2550‐2553
inward/record.url?eid=2‐s2.0‐84870728544&partnerID=tZOtx3y1
Engineering Science. **101**:13‐26
80 Desalination
doi/10.5004/dwt.2011.2856
abs/10.1021/am403790s
Brines including seawater, concentrated seawater after desalinization, salt lake, oil/gas water, and well bitter are widely distributed around the world. In order to promote the comprehensive utilization and effective protection of the valuable chemical resources existing in brines such as freshwater, lithium, sodium, potassium, and magnesium salts, the systematic foundation and application foundation research including phase equilibria and thermodynamic properties for the salt-water electrolyte solution are essential, especially for solid lithium salts and their aqueous solution systems.
Keywords: thermodynamics, phase equilibria, aqueous solution, lithium salts
## 1. Introduction
#### 1.1. Lithium resources situation
Lithium is the lightest alkali metal, which plays a growing role in numerous processes such as rechargeable batteries, thermonuclear fusion, medical drugs, lubricant greases, ceramic, glasses, dyes, adhesives, and electrode welding [1–19]. Lithium is a critical energy material and a strategic resource for the twenty-first century. Consequently, the market demands for lithium resources are increasing around the world [20, 21].
Lithium naturally occurs in compound forms because of its high reactivity. Economic concentrations of lithium are found in brines, minerals, and clays in various parts of the world. Brines and high-grade lithium ores are the present sources for all commercial lithium production. The global lithium reserve is estimated at 14.0 megatons [22], which is 74.5 megatons of lithium carbonate equivalent. Lithium reserves are mainly distributed in South America, China, and Australia [23].
Generally, lithium is obtained from two major resources: the lithium mineral ores including spodumene and petalite ores and the containing-lithium brine resources including seawater, underground water and salt lake brine [24–27]. Currently, the former is well exploited while the latter is being developed by industries with relatively low efficiency. Nonetheless, more than 60% of the total lithium amount exists in sea water and brines [24]. Therefore, a great potential exists for obtaining lithium from aqueous sources, if an efficient lithium recovery technology can be developed. Separation and extraction of lithium from either sea water or brine is carried out on a semi-industrial scale and industrial scale in the USA from salt lakes [28], in Japan from thermal water [29, 30], in Israel from the Dead Sea [31], and in China from underground brines and salt lakes [22].
Category Lithium-containing systems Temperature (K) Reference LiCl Hþ, Li<sup>þ</sup>, Mg2<sup>þ</sup>//Cl�–H2O 273.15 [36]
Li2SO4 Liþ, K<sup>þ</sup>//SO4
Li2CO3 Liþ, Naþ, K<sup>þ</sup>//CO3
Li2B4O7 Liþ//Cl�, SO4
LiBO2 Li<sup>þ</sup>//SO4
<sup>2</sup>�–H2O
<sup>2</sup>�–H2O
Liþ, Mg2<sup>þ</sup>//SO4
Liþ, K<sup>þ</sup>//SO4
Liþ//Cl�, SO4
Liþ, Mg2<sup>þ</sup>//SO4
Liþ, Kþ//Cl�, CO3
Liþ, Naþ//Cl�, CO3
Liþ, Kþ//Cl�, CO3
Liþ, Naþ, K<sup>þ</sup>//CO3
Liþ, Mg2<sup>þ</sup>//SO4
Liþ, K<sup>þ</sup>//CO3
Liþ, K<sup>þ</sup>//CO3
Liþ, Na<sup>þ</sup>//CO3
Li<sup>þ</sup>//CO3
<sup>2</sup>�, B4O7
<sup>2</sup>�, B4O7
<sup>2</sup>�, B4O7
<sup>2</sup>�, BO2
<sup>2</sup>�, BO2
�–H2O
<sup>2</sup>�, B4O7
<sup>2</sup>�, B4O7
Liþ, Naþ, Kþ//Cl�, B4O7
Liþ, Naþ, K<sup>þ</sup>//CO3
Liþ//Cl�, BO2
Liþ//Cl�, BO2
Liþ//Cl�, SO4
Li<sup>þ</sup>//SO42, CO3
Liþ//Cl�, BO2
Li<sup>þ</sup>//SO4
<sup>2</sup>�, BO2
<sup>2</sup>�, BO2
Table 1. Stable phase equilibria of lithium-containing salt-water systems.
<sup>2</sup>�, B4O7
<sup>2</sup>�–H2O
<sup>2</sup>�–H2O
<sup>2</sup>�–H2O
<sup>2</sup>�–H2O
<sup>2</sup>�, B4O7
<sup>2</sup>�–H2O
<sup>2</sup>�–H2O
<sup>2</sup>�–H2O
Liþ, K<sup>þ</sup>, Mg2<sup>þ</sup>//SO4
Liþ, K<sup>þ</sup>, Mg2<sup>þ</sup>//Cl�, SO4
Liþ, Naþ, Kþ, Mg2<sup>þ</sup>//SO4
Liþ, Naþ, Kþ, Mg2þ//Cl�–H2O 298.15 [37] Liþ, Mg2þ//Cl�–H2O 298.15 [38] Liþ, NH4þ//Cl�–H2O 273.15,298.15, 323.15 [39] Liþ, Rb<sup>þ</sup>, Mg2<sup>þ</sup>//Cl�–H2O 323.15 [40] Liþ, Naþ, Kþ, Sr2þ//Cl�–H2O 298.15 [41] Liþ, Mg2þ//Cl�–H2O 288.15 [42]
Phase Equilibria and Phase Separation of the Aqueous Solution System Containing Lithium Ions
298.15 [43]
http://dx.doi.org/10.5772/intechopen.68363
83
298.15 [54] 298.15 [55] 298.15 [56] 298.15 [57] 298.15 [58] 288.15 [59] 298.15 [60]
<sup>2</sup>�–H2O 298.15 [44]
<sup>2</sup>�–H2O 288.15 [47]
<sup>2</sup>�–H2O 308.15 [48]
<sup>2</sup>�–H2O 288.15 [42]
<sup>2</sup>�–H2O 298.15 [49]
<sup>2</sup>�–H2O 298.15 [50]
<sup>2</sup>�–H2O 298.15 [51]
<sup>2</sup>�–H2O 298.15 [52]
<sup>2</sup>�–H2O 288.15 [53]
<sup>2</sup>�–H2O 288.15 [61]
�–H2O 288.15, 298.15 [62]
�–H2O 288.15, 298.15 [63]
�–H2O 308.15 [64]
�–H2O 323.15 [67]
�–H2O 298.15 [65]
�–H2O 288.15, 298.15, 308.15 [66]
<sup>2</sup>�–H2O 298.15 [45]
<sup>2</sup>�–H2O 298.15 [46]
In addition, the most important physical and chemical processes occurring in brines are evaporation, concentration, crystallization, precipitation, dissolution, and phase transformation. It is obvious that phase equilibria and thermodynamics can explain above phenomenon and even guide those processes effectively. Therefore, it is particularly meaningful to engage the research on phase equilibria and thermodynamics properties of lithium-containing aqueous solution systems for describing the geochemical evolution of containing lithium brines and exploiting valuable lithium resources.
## 2. Phase equilibria of lithium-containing salt-water systems
Brines, including seawater, concentrated seawater after desalinization, salt lake, oil/gas water, and well bitter are all complex multi-component salt-water systems, whose study and application is mainly in reference to the solubility of salts in the water and the solid-liquid equilibrium rule. Hence, solid-liquid phase equilibria form the basis for salt-water systems, which in turn are used in the chemical industry for the separation of lithium [32].
#### 2.1. Stable phase equilibria of lithium-containing salt-water systems
Early in the 1960s, Soviet scholars had conducted research on stable phase equilibria of lithiumcontaining salt-water systems [33–35]. The solubilities of the systems (Liþ, Na<sup>þ</sup>, Mg2<sup>þ</sup>//Cl� – H2O), (Liþ, Na<sup>þ</sup>, Mg2<sup>þ</sup>//SO4 <sup>2</sup>� – H2O), (Li<sup>þ</sup>, Naþ, K<sup>þ</sup>//SO4 2- – H2O) in the temperature range from 288.15 to 373.15 K were determined. Three types of double salts containing lithium, 2Li2SO4�Na2SO4 �K2SO4, Li2SO4�3Na2SO4�12H2O, and Li2SO4�K2SO4 were found for the first time and the physicochemical properties of these were measured. All these studies could provide the solubility data to extract lithium resources in sea water and other brines.
Researchers in China also studied stable phase equilibria of complex system (Liþ, Naþ, Kþ, Mg2<sup>þ</sup>//Cl�, CO3 <sup>2</sup>�, SO4 <sup>2</sup>�, borate–H2O) and its subsystems. Some research results are shown in Table 1. In addition, depending on the aquatic chemical types, the lithium-containing salt-water systems can be divided into lithium-containing chloride system, sulfate system, carbonate system, and borate system.
In order to apply the hydride salting-out effect to the separation of lithium and magnesium, phase equilibria of the quaternary system Hþ, Liþ, Mg2þ//Cl� – H2O and its subsystems at
Table 1. Stable phase equilibria of lithium-containing salt-water systems.
Generally, lithium is obtained from two major resources: the lithium mineral ores including spodumene and petalite ores and the containing-lithium brine resources including seawater, underground water and salt lake brine [24–27]. Currently, the former is well exploited while the latter is being developed by industries with relatively low efficiency. Nonetheless, more than 60% of the total lithium amount exists in sea water and brines [24]. Therefore, a great potential exists for obtaining lithium from aqueous sources, if an efficient lithium recovery technology can be developed. Separation and extraction of lithium from either sea water or brine is carried out on a semi-industrial scale and industrial scale in the USA from salt lakes [28], in Japan from thermal water [29, 30], in Israel from the Dead Sea [31], and in China
In addition, the most important physical and chemical processes occurring in brines are evaporation, concentration, crystallization, precipitation, dissolution, and phase transformation. It is obvious that phase equilibria and thermodynamics can explain above phenomenon and even guide those processes effectively. Therefore, it is particularly meaningful to engage the research on phase equilibria and thermodynamics properties of lithium-containing aqueous solution systems for describing the geochemical evolution of containing lithium brines and exploiting
Brines, including seawater, concentrated seawater after desalinization, salt lake, oil/gas water, and well bitter are all complex multi-component salt-water systems, whose study and application is mainly in reference to the solubility of salts in the water and the solid-liquid equilibrium rule. Hence, solid-liquid phase equilibria form the basis for salt-water systems, which in turn
Early in the 1960s, Soviet scholars had conducted research on stable phase equilibria of lithiumcontaining salt-water systems [33–35]. The solubilities of the systems (Liþ, Na<sup>þ</sup>, Mg2<sup>þ</sup>//Cl� – H2O),
to 373.15 K were determined. Three types of double salts containing lithium, 2Li2SO4�Na2SO4 �K2SO4, Li2SO4�3Na2SO4�12H2O, and Li2SO4�K2SO4 were found for the first time and the physicochemical properties of these were measured. All these studies could provide the solubility data to
Researchers in China also studied stable phase equilibria of complex system (Liþ, Naþ, Kþ,
Table 1. In addition, depending on the aquatic chemical types, the lithium-containing salt-water systems can be divided into lithium-containing chloride system, sulfate system, carbonate sys-
In order to apply the hydride salting-out effect to the separation of lithium and magnesium, phase equilibria of the quaternary system Hþ, Liþ, Mg2þ//Cl� – H2O and its subsystems at
<sup>2</sup>�, borate–H2O) and its subsystems. Some research results are shown in
2- – H2O) in the temperature range from 288.15
2. Phase equilibria of lithium-containing salt-water systems
are used in the chemical industry for the separation of lithium [32].
2.1. Stable phase equilibria of lithium-containing salt-water systems
<sup>2</sup>� – H2O), (Li<sup>þ</sup>, Naþ, K<sup>þ</sup>//SO4
extract lithium resources in sea water and other brines.
<sup>2</sup>�, SO4
from underground brines and salt lakes [22].
valuable lithium resources.
82 Desalination
(Liþ, Na<sup>þ</sup>, Mg2<sup>þ</sup>//SO4
Mg2<sup>þ</sup>//Cl�, CO3
tem, and borate system.
273.15, 293.15, and 313.15 K had been researched [36, 68]. Phase distribution of the quaternary system was confirmed and the salting-out effect was investigated preliminarily, which provides the physical chemistry foundation for lithium-preparation technique in brines.
Furthermore, the phase equilibria and phase diagram of various salt-water systems containing lithium had been reported heavily. However, the experimental temperature was almost focused on 298.15 K. With the maturity of research techniques and the development of instruments, phase equilibria of multi-component systems at multiple temperatures should be the research focus point in the future. It is worth mentioning that the structures of borate are complicated and have diversified aggregation forms because of the changes of pH, boron contents, types of coexisting ions, and the concentration conditions in the brines. Most studies on the borate-type salt lake brine are mainly for lithium tetraborate. Our group has made excellent progress on phase equilibria of salt-water system containing different species of lithium borates [62–67].
#### 2.2. Metastable phase equilibria of lithium-containing salt-water systems
In the process of seawater and salt lake brine evaporation, the metastable phenomenon is ubiquitous. Because of the conditions of temperature, wind speed, and humidity in the natural environment, brine systems are in a metastable state. Early in the eighteenth century, Van't Hoff had already found that some phase regions disappeared and some enlarged in the stable phase diagram when the salt-water system was in the process of simulating evaporation [32]. So it is useful for the metastable phase diagram to extract the products which cannot be obtained in the stable phase diagram [69].
Some of metastable phase equilibria of lithium-containing salt-water systems are shown in Table 2. A lot of lithium-containing systems of chloride, sulfate, carbonate, and borate were researched. The concentration of lithium salt becomes higher in the sulfate-type salt lake brine in the final evaporation period. In our group, metastable phase equilibria of the
lithium-containing sulfate system (Liþ, Naþ, Kþ, Mg2þ//Cl�, SO4
gained based on the principles of thermodynamics [99].
at different temperature were determined, which has great help for industrial production
Category Lithium-containing systems Temperature (K) Reference
<sup>2</sup>�–H2O 308.15 [81]
http://dx.doi.org/10.5772/intechopen.68363
85
Phase Equilibria and Phase Separation of the Aqueous Solution System Containing Lithium Ions
<sup>2</sup>�–H2O 308.15 [82]
<sup>2</sup>�–H2O 273.15 [83]
<sup>2</sup>�–H2O 308.15 [86]
<sup>2</sup>�–H2O 288.15 [89]
<sup>2</sup>�–H2O 298.15 [60]
<sup>2</sup>�–H2O 288.15 [91]
<sup>2</sup>�–H2O 288.15 [92]
<sup>2</sup>�–H2O 288.15 [93]
<sup>2</sup>�–H2O 288.15 [94]
<sup>2</sup>�–H2O 298.15 [95]
<sup>2</sup>�–H2O 273.15 [96]
<sup>2</sup>�–H2O 288.15 [98]
<sup>2</sup>�–H2O 273.15 [97]
<sup>2</sup>�–H2O 298.15 [90]
<sup>2</sup>�–H2O 288.15, 308.15 [84]
<sup>2</sup>�–H2O 308.15, 348.15 [85]
<sup>2</sup>�–H2O 308.15 [48]
<sup>2</sup>�–H2O 288.15, 323.15 [87]
<sup>2</sup>�–H2O 308.15 [88]
Liþ,Mg<sup>2</sup>
Liþ,Mg<sup>2</sup>
Liþ,Kþ//Cl�,SO4
Liþ,Na<sup>þ</sup>//SO4
Liþ,Na<sup>þ</sup>//SO4
Liþ//Cl�,SO4
Liþ,K<sup>þ</sup>//SO4
Liþ,K<sup>þ</sup>//SO4
Liþ,Kþ//Cl�,CO3
Liþ,Na<sup>þ</sup>//SO4
Liþ,Naþ,K<sup>þ</sup>//CO3
Liþ,Kþ//Cl�,SO4
Liþ,K<sup>þ</sup>//SO4
Liþ//Cl�,CO3
Liþ,K<sup>þ</sup>//CO3
Liþ,K<sup>þ</sup>//CO3
Liþ,K<sup>þ</sup>//CO3
Li2CO3 Liþ,Na<sup>þ</sup>//SO4
Li2B4O7 Liþ,Naþ,Kþ//Cl�,B4O7
Liþ,Naþ,Kþ//Cl�,SO4
<sup>2</sup>�,CO3
<sup>2</sup>�,B4O7
<sup>2</sup>�,B4O7
<sup>2</sup>�,B4O7
<sup>2</sup>�,B4O7
<sup>2</sup>�,SO4
<sup>2</sup>�,B4O7
Table 2. Metastable phase equilibria of lithium-containing salt-water systems.
<sup>2</sup>�,B4O7
<sup>2</sup>�,B4O7
<sup>2</sup>�,B4O7
//Cl�,SO4
//Cl�,SO4
3. Thermodynamics of lithium salts and their aqueous solution systems
In the long-term production activities and scientific practice, it is troublesome for some production process of the new technology, new processes, or new product development, because there is no data and phase diagram of the relevant salt-water system. In addition, experimental determination on solubilities of multi-component systems is a complex work and it is virtually impossible for researchers to investigate all the salt-water systems. However, it is well-known that phase diagram is a geometry description of phase relation in the system under the condition of thermodynamic equilibrium. In theory, the phase diagram should be able to be
and the comprehensive utilization of lithium-containing sulfate system salt lakes.
2- – H2O) and its subsystems
Table 2. Metastable phase equilibria of lithium-containing salt-water systems.
273.15, 293.15, and 313.15 K had been researched [36, 68]. Phase distribution of the quaternary system was confirmed and the salting-out effect was investigated preliminarily, which pro-
Furthermore, the phase equilibria and phase diagram of various salt-water systems containing lithium had been reported heavily. However, the experimental temperature was almost focused on 298.15 K. With the maturity of research techniques and the development of instruments, phase equilibria of multi-component systems at multiple temperatures should be the research focus point in the future. It is worth mentioning that the structures of borate are complicated and have diversified aggregation forms because of the changes of pH, boron contents, types of coexisting ions, and the concentration conditions in the brines. Most studies on the borate-type salt lake brine are mainly for lithium tetraborate. Our group has made excellent progress on phase equilibria of salt-water system containing different species of
In the process of seawater and salt lake brine evaporation, the metastable phenomenon is ubiquitous. Because of the conditions of temperature, wind speed, and humidity in the natural environment, brine systems are in a metastable state. Early in the eighteenth century, Van't Hoff had already found that some phase regions disappeared and some enlarged in the stable phase diagram when the salt-water system was in the process of simulating evaporation [32]. So it is useful for the metastable phase diagram to extract the products which cannot be
Some of metastable phase equilibria of lithium-containing salt-water systems are shown in Table 2. A lot of lithium-containing systems of chloride, sulfate, carbonate, and borate were researched. The concentration of lithium salt becomes higher in the sulfate-type salt lake brine in the final evaporation period. In our group, metastable phase equilibria of the
Category Lithium-containing systems Temperature (K) Reference LiCl Liþ,Naþ,Mg2þ//Cl�–H2O 308.15 [70]
> Liþ,Naþ,Ca2þ//Cl�–H2O 288.15 [71] Liþ,Kþ//Cl�–H2O 298.15 [72] Liþ,Kþ//Cl�–H2O 323.15 [73] Liþ,Kþ,Rbþ//Cl�–H2O 298.15 [74] Liþ,Kþ,Rbþ//Cl�–H2O 323.15 [75]
> > <sup>2</sup>�–H2O 298.15 [76]
<sup>2</sup>�–H2O 273.15 [77]
<sup>2</sup>�–H2O 323.15 [79]
<sup>2</sup>�–H2O 323.15 [80]
<sup>2</sup>�–H2O 263.15 [78]
vides the physical chemistry foundation for lithium-preparation technique in brines.
2.2. Metastable phase equilibria of lithium-containing salt-water systems
lithium borates [62–67].
84 Desalination
obtained in the stable phase diagram [69].
Li2SO4 Liþ,Mg2þ//Cl�,SO4
Liþ,Naþ//Cl�,SO4
Liþ,Mg2<sup>þ</sup>//SO4
Liþ,Mg2//Cl�,SO4
Liþ,Naþ,Mg2<sup>þ</sup>//SO4
lithium-containing sulfate system (Liþ, Naþ, Kþ, Mg2þ//Cl�, SO4 2- – H2O) and its subsystems at different temperature were determined, which has great help for industrial production and the comprehensive utilization of lithium-containing sulfate system salt lakes.
#### 3. Thermodynamics of lithium salts and their aqueous solution systems
In the long-term production activities and scientific practice, it is troublesome for some production process of the new technology, new processes, or new product development, because there is no data and phase diagram of the relevant salt-water system. In addition, experimental determination on solubilities of multi-component systems is a complex work and it is virtually impossible for researchers to investigate all the salt-water systems. However, it is well-known that phase diagram is a geometry description of phase relation in the system under the condition of thermodynamic equilibrium. In theory, the phase diagram should be able to be gained based on the principles of thermodynamics [99].
As to the classical electrolyte theory, Debye-Hückel theory is only suitable for the dilution solution with a concentration below 0.1 m (molality) and it is unusable to solve the thermodynamic behaviors and to predictive the dissolution equilibria for the complex salt lake brine systems [100]. Pitzer theory [101], which was developed on the basis of Debye-Hückel ioninteraction theory, characterizing thermodynamics properties of electrolyte solution with brief and terse form is widely used either in geochemical behaviors of natural waters and mineral deposits or in the predictions of solubility of salt-water systems. A series of calculated expressions for the activity coefficient and osmotic coefficient of any electrolytes in multi-component systems were proposed by Pitzer [102]. After measuring the thermodynamic parameters, such as osmotic coefficient, activity coefficient, heat of dissolution, heat of dilution, heat of mixing, and specific heat, it is easy to calculate and fit the relative model-parameter theoretically and solubility on the basis of Pitzer and its extended ion-interaction model to promote the development of theory and practice, such as the new field of calculating phase diagram and its application [32].
determined at 298.15 K in the total ionic strength range from 0.05 to 6.0 mol�kg�<sup>1</sup> with the electromotive force method using a lithium-selective electrode and Ag/AgCl electrode by Wang et al. [108]. The activity coefficients of the experiment were compared with which calculated by
Phase Equilibria and Phase Separation of the Aqueous Solution System Containing Lithium Ions
The process of chemical reaction, dissolution, dilution, and mixing are often accompanied with heat changes. Reaction heat data, such as dissolution heat, dilution heat, mixing heat, and special heat, were determined critically by the calorimetry technology, which can work out thermodynamic enthalpy, entropy, Gibbs free energy, and thermodynamic equilibrium constant. Hence, the calorimetric method became the research highlights to measure the thermo-
There are some thermodynamic parameters measured by the calorimetric method. In 1965, Wu [109] and Wood [110] researched heats of mixing of a variety of aqueous containing lithium solutions of the same ionic strength at 298.15 K. Some of them were in the different concentration. The concentration dependence of the heats of mixing indicated that likecharged ion pairs were important contributors to the heat of mixing. The enthalpies of dilution of lithium in the range 0.1–1.0 m had been measured at 303.15 K with a microcalorimeter by Leung et al. [111]. The relative apparent enthalpies of these solutions had been determined with the aid of an extended form of Debye-Hückel limiting law. Enthalpies of solution of Li2SO4 and Li2SO4�H2O in water at 298.15 K were investigated [112]. The molar enthalpies of
lithium sulfate had been calculated by assuming that the excess water in the sample was present as an aqueous saturated solution. There were presented preliminary specific heat
The standard molar enthalpy of the formation of some lithium borates were determined
Meanwhile, the thermodynamic properties in solution system contained lithium borates had been widely researched. Zhang et al. [117] determined the molar heat capacities of the aqueous Li2B4O7 solution at a concentration of 0.0187 mol�kg�<sup>1</sup> in the temperature range from 80 to 355 K by a precision automated adiabatic calorimetry. The enthalpies of dilution for the aqueous Li2B4O7 solutions from 0.0212 to 2.1530 mol�kg�<sup>1</sup> at 298.15 K have been measured [118]. The relative apparent molar enthalpies and relative partial molar enthalpies of the solvent and solute for the aqueous Li2B4O7 system were also calculated. The thermodynamic properties of the binary aqueous system Li2B4O7–H2O were represented with the extended Pitzer ion-interaction model. And the enthalpies of dilution, ΔdilHm, have been also measured for the LiCl– Li2B4O7–H2O system at T ¼ 298.15 K [119]. A suitable microcalorimetric method was used to obtain the better data of the enthalpies of dilution for the ternary system LiCl–Li2B4O7–H2O at a low concentration. The relative apparent molar enthalpies have been determined and the
, ΔfH<sup>m</sup> θ
<sup>θ</sup> (LiBO2�2H2O) ¼ �(1627.46 � 0.90) kJ�mol�<sup>1</sup>
(Li3B5O8(OH)2 (II)) ¼ �(4723.8 � 4.2) kJ�mol�<sup>1</sup>
.
<sup>∞</sup> (Li2SO4) ¼ �(30502 � 170)
, ΔfH<sup>m</sup> θ
, and
(Li3B5O8(OH)2 (I)) ¼
. The value for the monohydrate
http://dx.doi.org/10.5772/intechopen.68363
87
(LiB5O8�5H2O) ¼ �(5130.25 � 4.05)
, ΔfH<sup>m</sup> θ
the Pitzer equation with the known parameters of experimental osmotic coefficients.
dynamic properties (dissolution heat, dilution heat, mixing heat, and capacity heat).
solution extrapolated to infinite dilution at 298.15 K were ΔsolH<sup>m</sup>
, ΔfH<sup>m</sup> θ
(Li4[B5O13(OH)2]�3H2O) ¼ � (7953.8 � 6.6) kJ�mol�<sup>1</sup>
<sup>∞</sup> (Li2SO4�H2O) ¼ �(17899 � 152) J�mol�<sup>1</sup>
<sup>θ</sup> (Li2B4O7�3H2O) ¼ �(4290.86 � 3.31) kJ�mol�<sup>1</sup>
<sup>J</sup>�mol�<sup>1</sup> and <sup>Δ</sup>solH<sup>m</sup>
, ΔfH<sup>m</sup>
�(4724.1 � 4.2) kJ�mol�<sup>1</sup>
kJ�mol�<sup>1</sup>
ΔfH<sup>m</sup> θ
capacities of lithium sulfate solution.
[113–116]. The results were ΔfH<sup>m</sup>
(LiBO2�8H2O) ¼ �(3397.00 � 0.94) kJ�mol�<sup>1</sup>
At present, the domestic and foreign research methods of electrolyte solution of thermodynamic properties are mainly isopiestic method, electromotive force method, calorimetric method, conductivity method, hygrometry, density method, and so on. The isopiestic method and electromotive force method are the most common experimental method to be widely used in measuring the thermodynamic properties such as permeability and activity coefficient. They complement each other. The basic property of matter which is the change of energy can be measured directly by the calorimetric method. In recent years, the calorimetric method is widely implemented in the research of solution thermodynamic properties [103].
In 1992, Yao et al. [104] measured the osmotic and activity coefficients of aqueous mixtures of LiCl and MgCl2 in the range of low concentration to near crystallization limits by the isopiestic method. The Pitzer single-salt parameters and the mixed parameters were calculated by the osmotic and activity coefficients, which were applicative for Pitzer's equation. The predicted solubilities for the system studied using Pitzer's approach were shown to be in reasonable agreement with experimental results from references. The osmotic coefficients of aqueous mixtures of Li2SO4 and MgSO4 had been reported from 1.4 to 13.5 molkg<sup>1</sup> at 298 K using the isopiestic method by Zhang et al [105]. In the ranges of 0.2–8.7 and 0.6–12.7 molkg<sup>1</sup> the osmotic coefficients of Li2SO4 and MgSO4 were also reported, respectively. The predicted solubilities for this system using Pitzer's approach showed good agreement with experimental results. Yang et al. [106] measured isopiestic molalities and water activities for the Li2B4O7– LiCl–H2O system at 298.15 K using an improved isopiestic apparatus, and the two types of osmotic coefficients were calculated and compared. Pitzer's primary model with minor modifications, in combination with the chemical equilibria, was used to represent the experimental data for the complex Li2B4O7–LiCl–H2O system.
Based on the principle of isopiestic method, the osmotic and activity coefficients can be obtained. However, it might be difficult to get these coefficients with the strict thermodynamics calculation in the mixed electrolyte solutions. The activity coefficients of KCl and LiCl in KCl–LiCl aqueous mixtures have been studied at 298.15 K in ionic strength range of 0.1–4.0 molkg<sup>1</sup> by Li et al. [107]. The data were fitted to Pitzer's equation using the regression method, Pitzer parameters were obtained. The activity confidents of aqueous LiCl in the system LiCl–MgCl2–H2O were determined at 298.15 K in the total ionic strength range from 0.05 to 6.0 mol�kg�<sup>1</sup> with the electromotive force method using a lithium-selective electrode and Ag/AgCl electrode by Wang et al. [108]. The activity coefficients of the experiment were compared with which calculated by the Pitzer equation with the known parameters of experimental osmotic coefficients.
As to the classical electrolyte theory, Debye-Hückel theory is only suitable for the dilution solution with a concentration below 0.1 m (molality) and it is unusable to solve the thermodynamic behaviors and to predictive the dissolution equilibria for the complex salt lake brine systems [100]. Pitzer theory [101], which was developed on the basis of Debye-Hückel ioninteraction theory, characterizing thermodynamics properties of electrolyte solution with brief and terse form is widely used either in geochemical behaviors of natural waters and mineral deposits or in the predictions of solubility of salt-water systems. A series of calculated expressions for the activity coefficient and osmotic coefficient of any electrolytes in multi-component systems were proposed by Pitzer [102]. After measuring the thermodynamic parameters, such as osmotic coefficient, activity coefficient, heat of dissolution, heat of dilution, heat of mixing, and specific heat, it is easy to calculate and fit the relative model-parameter theoretically and solubility on the basis of Pitzer and its extended ion-interaction model to promote the development of theory and practice, such as the new field of calculating phase diagram and its application [32].
At present, the domestic and foreign research methods of electrolyte solution of thermodynamic properties are mainly isopiestic method, electromotive force method, calorimetric method, conductivity method, hygrometry, density method, and so on. The isopiestic method and electromotive force method are the most common experimental method to be widely used in measuring the thermodynamic properties such as permeability and activity coefficient. They complement each other. The basic property of matter which is the change of energy can be measured directly by the calorimetric method. In recent years, the calorimetric method is
In 1992, Yao et al. [104] measured the osmotic and activity coefficients of aqueous mixtures of LiCl and MgCl2 in the range of low concentration to near crystallization limits by the isopiestic method. The Pitzer single-salt parameters and the mixed parameters were calculated by the osmotic and activity coefficients, which were applicative for Pitzer's equation. The predicted solubilities for the system studied using Pitzer's approach were shown to be in reasonable agreement with experimental results from references. The osmotic coefficients of aqueous mixtures of Li2SO4 and MgSO4 had been reported from 1.4 to 13.5 molkg<sup>1</sup> at 298 K using the isopiestic method by Zhang et al [105]. In the ranges of 0.2–8.7 and 0.6–12.7 molkg<sup>1</sup> the osmotic coefficients of Li2SO4 and MgSO4 were also reported, respectively. The predicted solubilities for this system using Pitzer's approach showed good agreement with experimental results. Yang et al. [106] measured isopiestic molalities and water activities for the Li2B4O7– LiCl–H2O system at 298.15 K using an improved isopiestic apparatus, and the two types of osmotic coefficients were calculated and compared. Pitzer's primary model with minor modifications, in combination with the chemical equilibria, was used to represent the experimental
Based on the principle of isopiestic method, the osmotic and activity coefficients can be obtained. However, it might be difficult to get these coefficients with the strict thermodynamics calculation in the mixed electrolyte solutions. The activity coefficients of KCl and LiCl in KCl–LiCl aqueous mixtures have been studied at 298.15 K in ionic strength range of 0.1–4.0 molkg<sup>1</sup> by Li et al. [107]. The data were fitted to Pitzer's equation using the regression method, Pitzer parameters were obtained. The activity confidents of aqueous LiCl in the system LiCl–MgCl2–H2O were
widely implemented in the research of solution thermodynamic properties [103].
data for the complex Li2B4O7–LiCl–H2O system.
86 Desalination
The process of chemical reaction, dissolution, dilution, and mixing are often accompanied with heat changes. Reaction heat data, such as dissolution heat, dilution heat, mixing heat, and special heat, were determined critically by the calorimetry technology, which can work out thermodynamic enthalpy, entropy, Gibbs free energy, and thermodynamic equilibrium constant. Hence, the calorimetric method became the research highlights to measure the thermodynamic properties (dissolution heat, dilution heat, mixing heat, and capacity heat).
There are some thermodynamic parameters measured by the calorimetric method. In 1965, Wu [109] and Wood [110] researched heats of mixing of a variety of aqueous containing lithium solutions of the same ionic strength at 298.15 K. Some of them were in the different concentration. The concentration dependence of the heats of mixing indicated that likecharged ion pairs were important contributors to the heat of mixing. The enthalpies of dilution of lithium in the range 0.1–1.0 m had been measured at 303.15 K with a microcalorimeter by Leung et al. [111]. The relative apparent enthalpies of these solutions had been determined with the aid of an extended form of Debye-Hückel limiting law. Enthalpies of solution of Li2SO4 and Li2SO4�H2O in water at 298.15 K were investigated [112]. The molar enthalpies of solution extrapolated to infinite dilution at 298.15 K were ΔsolH<sup>m</sup> <sup>∞</sup> (Li2SO4) ¼ �(30502 � 170) <sup>J</sup>�mol�<sup>1</sup> and <sup>Δ</sup>solH<sup>m</sup> <sup>∞</sup> (Li2SO4�H2O) ¼ �(17899 � 152) J�mol�<sup>1</sup> . The value for the monohydrate lithium sulfate had been calculated by assuming that the excess water in the sample was present as an aqueous saturated solution. There were presented preliminary specific heat capacities of lithium sulfate solution.
The standard molar enthalpy of the formation of some lithium borates were determined [113–116]. The results were ΔfH<sup>m</sup> <sup>θ</sup> (LiBO2�2H2O) ¼ �(1627.46 � 0.90) kJ�mol�<sup>1</sup> , ΔfH<sup>m</sup> θ (LiBO2�8H2O) ¼ �(3397.00 � 0.94) kJ�mol�<sup>1</sup> , ΔfH<sup>m</sup> θ (LiB5O8�5H2O) ¼ �(5130.25 � 4.05) kJ�mol�<sup>1</sup> , ΔfH<sup>m</sup> <sup>θ</sup> (Li2B4O7�3H2O) ¼ �(4290.86 � 3.31) kJ�mol�<sup>1</sup> , ΔfH<sup>m</sup> θ (Li3B5O8(OH)2 (I)) ¼ �(4724.1 � 4.2) kJ�mol�<sup>1</sup> , ΔfH<sup>m</sup> θ (Li3B5O8(OH)2 (II)) ¼ �(4723.8 � 4.2) kJ�mol�<sup>1</sup> , and ΔfH<sup>m</sup> θ (Li4[B5O13(OH)2]�3H2O) ¼ � (7953.8 � 6.6) kJ�mol�<sup>1</sup> .
Meanwhile, the thermodynamic properties in solution system contained lithium borates had been widely researched. Zhang et al. [117] determined the molar heat capacities of the aqueous Li2B4O7 solution at a concentration of 0.0187 mol�kg�<sup>1</sup> in the temperature range from 80 to 355 K by a precision automated adiabatic calorimetry. The enthalpies of dilution for the aqueous Li2B4O7 solutions from 0.0212 to 2.1530 mol�kg�<sup>1</sup> at 298.15 K have been measured [118]. The relative apparent molar enthalpies and relative partial molar enthalpies of the solvent and solute for the aqueous Li2B4O7 system were also calculated. The thermodynamic properties of the binary aqueous system Li2B4O7–H2O were represented with the extended Pitzer ion-interaction model. And the enthalpies of dilution, ΔdilHm, have been also measured for the LiCl– Li2B4O7–H2O system at T ¼ 298.15 K [119]. A suitable microcalorimetric method was used to obtain the better data of the enthalpies of dilution for the ternary system LiCl–Li2B4O7–H2O at a low concentration. The relative apparent molar enthalpies have been determined and the relationships between apparent molar enthalpies and ionic strength at different molar fractions of Li2B4O7 were obtained. Li [120] measured the heats of dilution and heat capacities of eutectic point solution system Li2B4O7–Li2SO4–LiCl–H2O and subsystems Li2B4O7–Li2SO4– H2O and Li2B4O7–LiCl–H2O to cover the ionic strength range from 19 to 0.1 at 298.15 K. The data of the heat of dilution were extrapolated to infinite dilution by use of the Debye-Hückel limiting law to obtain relative apparent molar enthalpies.
Acknowledgements
Author details
Tianjin, PR China
References
Long Li<sup>1</sup>
Financial supports from the National Natural Science of China (21276194, 21306136, U1407113, U1607123), the Chinese Postdoctoral Science Foundation (2016M592827), the Laboratory Foundation of Chinese Universities (SY2015018), and the Training Program of Yangtze Scholars and
Phase Equilibria and Phase Separation of the Aqueous Solution System Containing Lithium Ions
http://dx.doi.org/10.5772/intechopen.68363
89
Innovative Research Team in Universities of China ([2013]373) are acknowledged.
1 Tianjin Key Laboratory of Marine Resources and Chemistry, College of Chemical Engineering and Materials Science, Tianjin University of Science and Technology,
2 College of Chemistry and Materials Science, Northwest University, Xi'an, PR China
Chemical Education. 2003;80:1015–1020. DOI: 10.1021/ed080p1015
[1] Treptow RS. Lithium batteries: A practical application of chemical principles. Journal of
[2] Lu L, Han X, Li J, Hua J, Ouyang M. A review on the key issues for lithium-ion battery management in electric vehicles. Journal of Power Sources. 2013;226:272–288. DOI: 10.
[3] Song JY, Wang YY, Wan CC. Review of gel-type polymer electrolytes for lithium-ion batteries. Journal of Power Sources. 1999;77:183–197. DOI: 10.1016/S0378-7753(98)00193-1
[4] Delgado MA, Sánchez MC, Valencia C, Franco JM, Gallegos C. Relationship among microstructure, rheology and processing of a lithium lubricating grease. Chemical Engi-
[5] Delgado MA, Valencia C, Sánchez MC, Franco JM, Gallegos C. Thermorheological behaviour of a lithium lubricating grease. Tribology Letters. 2006;23:47–54. DOI: 10.1007/s11249-
[6] Monmaturapoj N, Lawita P, Thepsuwan W. Characterisation and properties of lithium disilicate glass ceramics in the SiO2-Li2O-K2O-Al2O3 system for dental applications. Advances in Materials Science & Engineering. 2013;2013:680–685. DOI: 10.1155/2013/
neering Research & Design. 2005;83:1085–1092. DOI: 10.1205/cherd.04311
, Yafei Guo1,2\* and Tianlong Deng1
\*Address all correspondence to: guoyafei@tust.edu.cn
1016/j.jpowsour.2012.10.060
006-9109-5
763838
In our group, the heat capacities of aqueous solution systems (Li2B4O7–H2O) m ¼ 0.00415– 0.4208 mol�kg�<sup>1</sup> �at T ¼ 298.15, 308.15 and 323.15 K were determined experimentally using the Setaram BT 2.15 microcalorimeter [121]. On the basis of experimental data, the apparent molar heat capacities at different concentrations and temperatures were calculated, and the relationship equations between apparent molar heat capacity and solution concentration of lithium tetraborate at 298.15, 308.15 and 323.15 K were obtained. On the other hand, the Pitzer single salt parameters of lithium tetraborate at different temperatures were fitted on the basis of the Pitzer ion-interaction theory of the electrolytes on the apparent molar heat capacity.
So far, the thermodynamic parameters of lithium salts are still scarce. The Pitzer single salt parameters and the mixing ion-interaction parameters at different temperatures have not been established yet. So, more works on the thermodynamics parameters of lithium salts and their aqueous solution systems at multi-temperatures are essential.
## 4. Conclusion
With the gradually increasing demands of lithium salt resources as the lithium energy battery, to exploit the lithium-containing brine resources including seawater, concentrated seawater after desalinization, salted lake, oil/gas field water, and well bitter is essential. Therefore, studies on phase equilibria and phase separation of the lithium-containing brine systems are significant to guide the comprehensive utilization of those lithium-containing brine resources around the world. In this chapter, the following three main aspects were discussed: Firstly, the stable and metastable phase equilibria of lithium-containing salt-water multi-systems including the different types of chloride, sulfate, carbonate, and borate brines at different temperatures were summarized. Secondly, a series of valuable thermodynamic properties of standard molar enthalpy of formation for solid lithium salts such as LiCl�H2O, Li2SO4�H2O, LiBO2�2H2O, LiBO2�8H2O, Li2B4O7�3H2O, Li3B5O8(OH)2, and Li4[B5O13(OH)2]�3H2O, and the thermodynamic properties (dissolution heat, dilution heat, mixing heat, and capacity heat) for their relatively aqueous solutions were obtained combined by the isothermal dissolution equilibrium method, isopiestic method, and adiabatic calorimetry, and so on. Thirdly, on the basis of classical Debye-Hückel electrolyte theory, the extended modern electrolyte model, developed by Pitzer KS and his coworkers to express the activity coefficient and osmotic coefficient of any electrolytes in multi-component lithium-containing systems, was successfully used to obtain a series parameters such as the model parameter fitting, thermodynamic property calculation, and the solubility prediction.
## Acknowledgements
relationships between apparent molar enthalpies and ionic strength at different molar fractions of Li2B4O7 were obtained. Li [120] measured the heats of dilution and heat capacities of eutectic point solution system Li2B4O7–Li2SO4–LiCl–H2O and subsystems Li2B4O7–Li2SO4– H2O and Li2B4O7–LiCl–H2O to cover the ionic strength range from 19 to 0.1 at 298.15 K. The data of the heat of dilution were extrapolated to infinite dilution by use of the Debye-Hückel
In our group, the heat capacities of aqueous solution systems (Li2B4O7–H2O) m ¼ 0.00415–
Setaram BT 2.15 microcalorimeter [121]. On the basis of experimental data, the apparent molar heat capacities at different concentrations and temperatures were calculated, and the relationship equations between apparent molar heat capacity and solution concentration of lithium tetraborate at 298.15, 308.15 and 323.15 K were obtained. On the other hand, the Pitzer single salt parameters of lithium tetraborate at different temperatures were fitted on the basis of the
So far, the thermodynamic parameters of lithium salts are still scarce. The Pitzer single salt parameters and the mixing ion-interaction parameters at different temperatures have not been established yet. So, more works on the thermodynamics parameters of lithium salts and their
With the gradually increasing demands of lithium salt resources as the lithium energy battery, to exploit the lithium-containing brine resources including seawater, concentrated seawater after desalinization, salted lake, oil/gas field water, and well bitter is essential. Therefore, studies on phase equilibria and phase separation of the lithium-containing brine systems are significant to guide the comprehensive utilization of those lithium-containing brine resources around the world. In this chapter, the following three main aspects were discussed: Firstly, the stable and metastable phase equilibria of lithium-containing salt-water multi-systems including the different types of chloride, sulfate, carbonate, and borate brines at different temperatures were summarized. Secondly, a series of valuable thermodynamic properties of standard molar enthalpy of formation for solid lithium salts such as LiCl�H2O, Li2SO4�H2O, LiBO2�2H2O, LiBO2�8H2O, Li2B4O7�3H2O, Li3B5O8(OH)2, and Li4[B5O13(OH)2]�3H2O, and the thermodynamic properties (dissolution heat, dilution heat, mixing heat, and capacity heat) for their relatively aqueous solutions were obtained combined by the isothermal dissolution equilibrium method, isopiestic method, and adiabatic calorimetry, and so on. Thirdly, on the basis of classical Debye-Hückel electrolyte theory, the extended modern electrolyte model, developed by Pitzer KS and his coworkers to express the activity coefficient and osmotic coefficient of any electrolytes in multi-component lithium-containing systems, was successfully used to obtain a series parameters such as the model parameter fitting, thermodynamic
Pitzer ion-interaction theory of the electrolytes on the apparent molar heat capacity.
�at T ¼ 298.15, 308.15 and 323.15 K were determined experimentally using the
limiting law to obtain relative apparent molar enthalpies.
aqueous solution systems at multi-temperatures are essential.
property calculation, and the solubility prediction.
0.4208 mol�kg�<sup>1</sup>
88 Desalination
4. Conclusion
Financial supports from the National Natural Science of China (21276194, 21306136, U1407113, U1607123), the Chinese Postdoctoral Science Foundation (2016M592827), the Laboratory Foundation of Chinese Universities (SY2015018), and the Training Program of Yangtze Scholars and Innovative Research Team in Universities of China ([2013]373) are acknowledged.
## Author details
Long Li<sup>1</sup> , Yafei Guo1,2\* and Tianlong Deng1
\*Address all correspondence to: guoyafei@tust.edu.cn
1 Tianjin Key Laboratory of Marine Resources and Chemistry, College of Chemical Engineering and Materials Science, Tianjin University of Science and Technology, Tianjin, PR China
2 College of Chemistry and Materials Science, Northwest University, Xi'an, PR China
## References
[7] Huang S, Cao P, Wang C, Huang Z, Gao W. Fabrication of a high-strength lithium disilicate glass-ceramic in a complex glass system. Journal of Asian Ceramic Societies. 2013;1:46–52. DOI: 10.1016/j.jascer.2013.02.007
[20] Vikström H, Davidsson S, Höök M. Lithium availability and future production outlooks. Applied Energy. 2013;110:252–266. DOI: 10.1016/j.apenergy.2013.04.005
Phase Equilibria and Phase Separation of the Aqueous Solution System Containing Lithium Ions
http://dx.doi.org/10.5772/intechopen.68363
91
[21] Sonoc A, Jeswiet J. A review of lithium supply and demand and a preliminary investigation of a room temperature method to recycle lithium ion batteries to recover lithium and other materials. Procedia Cirp. 2014;15:289–293. DOI: 10.1016/j.procir.2014.06.006 [22] Hao H, Liu Z, Zhao F, Geng Y, Sarkis J. Material flow analysis of lithium in China.
[23] Swain B. Recovery and recycling of lithium: A review. Separation and Purification
[24] Grosjean C, Miranda PH, Perrin M, Poggi P. Assessment of world lithium resources and consequences of their geographic distribution on the expected development of the electric vehicle industry. Renewable & Sustainable Energy Reviews. 2012;16:1735–1744.
[25] Chan LH, Leeman WP, Plank T. Lithium isotopic composition of marine sediments. Geochemistry Geophysics Geosystems. 2006;3613:107–109. DOI: 10.1029/2005GC001202
[26] Chitrakar R, Makita Y, Ooi K, Sonoda A. Synthesis of iron-doped manganese oxides with an ion-sieve property: lithium adsorption from Bolivian brine. Industrial & Engi-
[27] Kesler SE, Gruber PW, Medina PA, Keoleian GA, Everson MP, Wallington TJ. Global lithium resources: relative importance of pegmatite, brine and other deposits. Ore Geol-
[28] Marinsky, JA, Marcus Y, editors. Ion Exchange and Solvent Extraction: A Series of
[29] Yanagase K, Yoshinaga T, Kawano K, Matsuoka T. The recovery of lithium from geothermal water in the Hatchobaru area of Kyushu, Japan. Bulletin of the Chemical Society
[30] Yoshinaga T, Kawano K, Imoto H. Basic study on lithium recovery from lithium containing solution. Bulletin of the Chemical Society of Japan. 1986;59:1207–1213. DOI:
[31] Epstein JA, Feist EM, Zmora J, Marcus Y. Extraction of lithium from the dead sea.
[32] Deng TL, Zhou H, Chen X. Salt-water System Phase Diagrams and Applications. Bei-
Hydrometallurgy. 1981;6:269–275. DOI: 10.1016/0304-386X(81)90044-X
neering Chemistry Research. 2014;53:3682–3688. DOI: 10.1021/ie4043642
ogy Reviews. 2012;48:55–69. DOI: 10.1016/j.oregeorev.2012.05.006
Advances. Vol. 12. Boca Raton: CRC Press; 1995. p. 93
of Japan. 1983;56:2490–2498. DOI: 10.1246/bcsj.56.2490
jing: Chemical Industry Press; 2013. 11, pp. 203–205
[33] Lepshkov JH, Vopapeva NW. Inorganic Salts Phim. 1958;3:278 [34] Lepshkov JH, Vopapeva NW. Inorganic Salts Phim. 1961;6:1961 [35] Lepshkov JH, Vopapeva NW. Inorganic Salts Phim. 1962;7:1699
Resources Policy. 2017;51:100–106. DOI: 10.1016/j.resourpol.2016.12.005
Technology. 2017;172:388–403. DOI: 10.1016/j.seppur.2016.08.031
DOI: 10.1016/j.rser.2011.11.023
10.1246/bcsj.59.1207
[20] Vikström H, Davidsson S, Höök M. Lithium availability and future production outlooks. Applied Energy. 2013;110:252–266. DOI: 10.1016/j.apenergy.2013.04.005
[7] Huang S, Cao P, Wang C, Huang Z, Gao W. Fabrication of a high-strength lithium disilicate glass-ceramic in a complex glass system. Journal of Asian Ceramic Societies.
[8] Palacios-Bereche R, Gonzales R, Nebra SA. Exergy calculation of lithium bromide–water solution and its application in the exergetic evaluation of absorption refrigeration systems LiBr-H2O. International Journal of Energy Research. 2012;36:166–181. DOI:
[9] Hollands KGT. The regeneration of lithium chloride brine in a solar still for use in solar air conditioning. Solar Energy. 1963;7:39–43. DOI: 10.1016/0038-092X(63)90003-3 [10] Tobishima SI, Okada T. Effects of quinoneimine dyes on lithium cycling efficiency for LiClO4-propylene carbonate. Journal of Applied Electrochemistry. 1985;15:901–906.
[11] Kimura K, Tanaka M, Kitazawa S, Shono T. Highly lithium-selective crown ether dyes for extraction photometry. Chemistry Letters. 1985;8:1239–1240. DOI: 10.1246/cl.1985.1239 [12] Dunner DL, Fieve RR. Clinical factors in lithium carbonate prophylaxis failure. Archives of General Psychiatry. 1974;30:229–233. DOI: 10.1001/archpsyc.1974.01760080077013 [13] Millard MJ, Kurtis KE. Effects of lithium nitrate admixture on early-age cement hydration. Cement & Concrete Research. 2008;38:500–510. DOI: 10.1016/j.cemconres.2007.11.009 [14] Tyrsa VE, Burtseva LP. Generation of thermonuclear energy by fusing hydrogen and
lithium atoms. Technical Physics. 2003;48:807–812. DOI: 10.1134/1.1593184
[15] Blink JA, Hogam WJ, Hovingh J, Meier ER, Pitts JH. High-yield lithium-injection fusionenergy (HYLIFE) reactor. Journal of Agricultural & Food Chemistry. 2014;62:4276–4284.
[16] Mansfield DK, Strachan JD, Bell MG, Scott SD, Budny R, Marmar ES, et al. Enhanced performance of deuterium–tritium-fueled supershots using extensive lithium conditioning in the Tokamak Fusion Test Reactor. Physics of Plasmas (1994-Present) 1995;2:4252–
[17] Karki K, Epstein E, Cho JH, Jia Z, Li T, Picraux ST, et al. Lithium-assisted electrochemical welding in silicon nanowire battery electrodes. Nano Letters 2012;12:1392–1397. DOI:
[18] Swann AC, Koslow SH, Katz MM, Maas JW, Javaid J, Secunda SK, et al. Lithium carbonate treatment of mania. Cerebrospinal fluid and urinary monoamine metabolites and treatment outcome. Archives of General Psychiatry. 1987;44:345–354. DOI: 10.1001/
[19] Marocho SM, Ozcan M, Amaral R, Bottino MA, Valandro LF. Effect of resin cement type on the microtensile bond strength to lithium disilicate ceramic and dentin using different test assemblies. Journal of Adhesive Dentistry. 2013;15:361–368. DOI: 10.3290/j.jad.a28624
2013;1:46–52. DOI: 10.1016/j.jascer.2013.02.007
10.1002/er.1790
90 Desalination
DOI: 10.1007/BF00614366
DOI: 10.2172/6124368
10.1021/nl204063u
4256. DOI: 10.1063/1.871050
archpsyc.1987.01800160057008
[36] Hu KY, Chai WQ. A study of phase equilibria in the quaternary aqueous salt system Hþ, Liþ, Mg2þ//Cl�– H2O at 0�C. Acta Chimica Sinica. 1965 31:189–197
[49] Zeng Y, Yin HA, Tang ML, Wang LS. Studies on the phase diagram and liquid physico-
[50] Deng TL, Yin HA, Tang ML. A study of solubilities and physicochemistry properties of
[51] Deng TL, Yin HA, Tang ML. Experimental and predictive phase equilibrium of the Liþ,
[52] Zeng Y, Yin HA, Tang ML, Feng XX. A study of the phase equilibrium for quinary
[53] Sang SH, Tang ML, Yin HA, Zeng Y. A study on the phase equilibrium of the quaternary system K2CO3–Na2CO3–Li2CO3–H2O at 288 K. Chinese Journal of Applied Chemistry.
[54] Song PS, Du XH. A study of solubilities and physicochemistry properties of equilibrium solutions in the quaternary system Li2B4O7–Li2SO4–LiCl–H2O at 25�C. Chinese Science
[55] Song PS, Fu HG. Solubilities and properties of solution in the reciprocal system Li<sup>þ</sup>, Mg2<sup>þ</sup>/
[56] Zeng Y, Tang ML, Yin HA, Wang LS. An Experimental study on the equilibrium phase
[57] Zeng Y, Xiao X, Yin HA. A study on the phase equilibrium and solution properties of the
[58] Yin HA, Hao LF, Zeng Y, Tang ML. Studies on the phase equilibrium and physicochem-
[59] Yin HA, Sang SH, Tang ML Zeng Y. Equilibrium of quaternary system Liþ, K<sup>þ</sup>//CO3
[60] Deng TL. Phase equilibrium for the aqueous containing lithium, sodium, potassium, chloride, and borate ions at 298.15 K. Journal of Chemical & Engineering Data.
[61] Sang SH, Yin HA, Tang ML. (Liquid þ solid) phase equilibria in the quinary system Li<sup>þ</sup>
2-–H2O at 288 K. Journal of Chemical Industry and Engineering (China).
2-, B4O7
diagram and solution properties of the ternary systems Li<sup>þ</sup>/CO3
ical properties of solutions for the quinary system Liþ, Na<sup>þ</sup>//CO3
2004;5:464–467. DOI: 10.3321/j.issn:0438-1157.2004.03.001
2- <sup>þ</sup> B4O7
Journal of Chemical Engineering of Chinese Universities. 2003;17:1–5
Journal of Mineralogy & Petrology. 1999;19:89–92
ing of Chinese Universities. 2002;16:591–596
2004;47:1295–1299. DOI: 10.1021/je049975f
þ Na<sup>þ</sup> þ K<sup>þ</sup> þ CO3
2005;36:1557–1562
2-–H2O at 25�C. Chinese Journal of Inorganic Chemistry. 1991;7:344–348
2-–H2O system at 298.15 K. Journal of Chemical & Engineering Data.
Phase Equilibria and Phase Separation of the Aqueous Solution System Containing Lithium Ions
–H2O at 298.15 K. Chemical Journal of Chinese Univer-
2-, B4O7
2-, B4O7
2-–H2O at 298 K. Journal of Chemical Engineer-
2- <sup>þ</sup> H2O at 288 K. Physical Inorganic Chemistry.
2-–H2O at 298 K.
2-–H2O at 298 K.
2-,
2-–H2O at 298 K. Chemical
2-–H2O at 298
93
, CO3
http://dx.doi.org/10.5772/intechopen.68363
chemical properties for the quinary system, Liþ, Naþ, K<sup>þ</sup>/CO3
equilibrium solutions in the reciprocal quaternary system Liþ, Kþ/Cl-
K. Chemical Research in Chinese Universities. 2000;21:1572–1574
2-, Cl-
Engineering(China). 1999;27:45–47, 57
2002;47:26–29. DOI: 10.1021/je000368b
Naþ//Cl-
, CO3
system Liþ, Naþ, K<sup>þ</sup>//CO3
sities. 2003;24:968–972
Bulletin. 1986;3:209–213
quaternary system Liþ, K<sup>þ</sup>//CO3
2004;21:509–511
2-, SO4
B4O7
B4O7
[49] Zeng Y, Yin HA, Tang ML, Wang LS. Studies on the phase diagram and liquid physicochemical properties for the quinary system, Liþ, Naþ, K<sup>þ</sup>/CO3 2-–H2O at 298 K. Chemical Engineering(China). 1999;27:45–47, 57
[36] Hu KY, Chai WQ. A study of phase equilibria in the quaternary aqueous salt system Hþ,
[37] Zhang FX, Guo ZJ, Chen PH Ma JH, Chen YS. Studies on the solubility and the isothermal evaporation process in the quinary system Li, Na, K, Mg/Cl-H2O at 25�C. Chemical
[38] Zhang FX, Guo ZJ, Chen PH, Ma JH, Chen YS. A study on the ternary system LiCl-MgCl2-H2O at 25�C. Journal of Northwest University (Natural Science Edition).
[39] Ouyang HT, Zeng DW, Zhou HY, Han H, Yao Y. Solubility of the ternary system LiCl– NH4Cl–H2O. Journal of Chemical & Engineering Data. 2011;56:1096–1104. DOI: 10.1021/
[40] Yu XD, Jiang DB, Tan Q, Zeng Y. Solid–liquid equilibrium in the aqueous system containing the chlorides of lithium, rubidium and magnesium at 323 K. Fluid Phase
[41] Meng LZ, Gruszkiewicz MS, Deng TL, Guo YF, Li D. Isothermal evaporation process simulation using the Pitzer model for the quinary system LiCl þ NaCl þ KCl þ SrCl2 þ H2O at 298.15 K. Industrial & Engineering Chemistry Research. 2015;54:8311–8318. DOI:
[42] Wang SQ, Guo YF, Li DC. Solid-liquid phase equilibria in the ternary systems (LiCl þ MgCl2 þ H2O) and (Li2SO4 þ MgSO4 þ H2O) at 288.15 K. Journal of Chemical &
[44] Fang CH, Li B, Li J, Wang QZ, Song PS. Studies on the phase diagram and solution
[45] Sun B, Li B, Fang CH, Du XH, Song PS. Studies on the phase diagram and solution
[46] Li B, Sun B, Fang CH, Du XH, Song PS. Studies on the phase diagram for the quinary
[47] Wang SQ, Guo YF, Li DC, Tang P, Deng TL. Experimental determination and modelling of the solubility phase diagram of the ternary system (Li2SO4 þ K2SO4 þ H2O) at 288.15
[48] Liu YH, Guo YF, Wang XK, Deng TL. Phase equilibria in the ternary system (LiCl þ Li2SO4þ H2O) at T¼308.15 K and p¼0.1 MPa: Experimental data and predictions using the Pitzer model. Fluid Phase Equilibria. 2015;391:85–89. DOI: 10.1016/j.fluid.2015.02.009
K. Thermochimica Acta. 2015;601:75–81. DOI: 10.1016/j.tca.2015.01.003
2--H2O at 25�C. Acta Physico-Chimica Sinica.1994;10:536–542
2--H2O at 25�C. Acta Chimica Sinica. 1997;55:545–552
<sup>2</sup>�-H2O
2--H2O at 25�C. Acta Chimica
2--H2O at 25�C. Journal of Salt
[43] Li B, Wang QZ, Li J, Fang C, Song P. A study on the ternary systems Liþ, Kþ/SO4
Liþ, Mg2þ//Cl�– H2O at 0�C. Acta Chimica Sinica. 1965 31:189–197
Research in Chinese Universities. 1987;8:301–305
Equilibria. 2014;367:63–68. DOI: 10.1016/j.fluid.2014.01.037
Engineering Data. 2015;60:821–827. DOI: 10.1021/je500946w
properties for the quaternary system Liþ, K<sup>þ</sup>, Mg2<sup>þ</sup>/SO4
properties for the quinary system, Liþ, Kþ, Mg2þ/Cl�, SO4
1988;18:75–78
10.1021/acs.iecr.5b01897
and Liþ, Mg2<sup>þ</sup>/SO4
Sinica. 1994;52:954–959
Lake Research. 1995;3:50–56.
system Liþ, Naþ, K<sup>þ</sup>, Mg2<sup>þ</sup>/SO4
je101056t
92 Desalination
[62] Gao DL, Wang Q, Guo YF, Yu XP, Wang SQ, Deng TL. Solid-liquid phase equilibria in the aqueous ternary system Li2SO4 þ LiBO2 þ H2O at T ¼ 288.15 and 298.15 K. Fluid Phase Equilibria. 2014;371:121–124. DOI: 10.1016/j.fluid.2014.03.019
[74] Yu XD, Zeng Y, Yang JY. Sold-liquid isothermal evaporation metastable phase equilibria in the aqueous quaternary system LiCl þ KCl þ RbCl þ H2O at 298.15 K. Journal of
Phase Equilibria and Phase Separation of the Aqueous Solution System Containing Lithium Ions
http://dx.doi.org/10.5772/intechopen.68363
95
[75] Li ZQ, YU XD, Yin QH, Zeng Y. Thermodynamics metastable phase equilibria of aqueous quaternary system LiCl þ KCl þ RbCl þ H2O at 323.15 K. Fluid Phase Equilibria.
[76] Guo ZZ, Liu ZQ, Chen JQ. The metastable phase equilibrium in the system Li<sup>þ</sup>, Mg2<sup>þ</sup>//
[77] Wang SQ, Deng TL. Metastable phase equilibria of the reciprocal quaternary system containing lithium, sodium, chloride, and sulphate ions at 273.15 K. Journal of Chemical
[78] Li ZY, Deng TL, Liao MX. Solid-liquid metastable equilibria in the quaternary system Li2SO4 þ MgSO4 þ Na2SO4 þ H2O at T ¼ 263.15 K. Fluid Phase Equilibria.
[79] Deng TL, Yin HJ, Guo YF. Metastable phase equilibrium in the aqueous ternary system (Li2SO4 þ MgSO4þ H2O) at 323.15 K. Journal of Chemical & Engineering Data.
[80] Meng LZ, Yu XP, Li D, Deng TL. Solid-liquid metastable equilibria of the reciprocal quaternary system (LiCl þ MgCl2 þ Li2SO4 þ MgSO4 þ H2O) at 323.15 K. Journal of
[81] Gao J, Deng TL. Metastable phase equilibrium in the aqueous quaternary system (LiCl þ MgCl2 þ Li2SO4 þ MgSO4þ H2O) at 308.15 K. Journal of Chemical & Engineering Data.
[82] Liu YH, Deng TL Song PS. Metastable phase equilibrium of the reciprocal quaternary system LiCl þ KCl þ Li2SO4 þ K2SO4 þ H2O at 308.15 K. Journal of Chemical &
[83] Wang SQ, Guo YF, Deng TL. Solubility prediction for the reciprocal quaternary system
Materials Research. 2012;549:437–440. DOI: 10.4028/www.scientific.net/AMR.549.437
[84] Guo YF, Liu YH, Wang Q, Lin CX, Wang SQ, Deng TL. Phase equilibria and phase diagrams for the aqueous ternary system (Na2SO4 þ Li2SO4 þ H2O) at (288 and 308) K. Journal of Chemical & Engineering Data. 2013;58:2763–2767. DOI: 10.1021/je4004146 [85] Guo YF, Gao DL, Han HJ, Shen DL, Wang SQ, Deng TL. Metastable phase equilibria in the aqueous ternary system (Na2SO4 þ Li2SO4 þ H2O) at 308.15 and 348.15 K. Fluid
[86] Liu YH, Guo YF, Yu XP. Solid-liquid metastable phase equilibria in the five-component system (Li þ Na þ K þ Cl þ SO4 þ H2O) at 308.15 K. Journal of Chemical & Engineering
2-–H2O) at 273.15 K using Pitzer ion-interaction model. Advanced
Chemical & Engineering Data. 2011;56:4627–4632. DOI: 10.1021/je200563j
Engineering Data. 2011;56:1139–1147. DOI: 10.1021/je1010888
Phase Equilibria. 2013;358:56–59. DOI: 10.1016/j.fluid.2013.07.048
Chemical & Engineering Data. 2012;57:172–132. DOI: 10.1021/je200880m
2-–H2O at 25�. Acta Chimica Sinica. 1991;49:937–943
& Engineering Data. 2010;55:4213–4215. DOI: 10.1021/je1003669
2013;358:131–136. DOI: 10.1016/j.fluid.2013.08.011
2010;2931:42–46. DOI: 10.1016/j.fluid.2010.02.016
2011;56:3585–3588. DOI: 10.1021/je800482y
2011;56:1452–1458. DOI: 10.1021/je1008526
, SO4
(Liþ, Mg2þ//Cl-
Data. 2014;59:1685–1691
Cl- , SO4
[74] Yu XD, Zeng Y, Yang JY. Sold-liquid isothermal evaporation metastable phase equilibria in the aqueous quaternary system LiCl þ KCl þ RbCl þ H2O at 298.15 K. Journal of Chemical & Engineering Data. 2012;57:172–132. DOI: 10.1021/je200880m
[62] Gao DL, Wang Q, Guo YF, Yu XP, Wang SQ, Deng TL. Solid-liquid phase equilibria in the aqueous ternary system Li2SO4 þ LiBO2 þ H2O at T ¼ 288.15 and 298.15 K. Fluid
[63] Gao DL, Guo YF, Yu XP, Wang SQ, Deng TL. Solubilities, densities, and refractive indices in the salt-water ternary system (LiCl þ LiBO2þ H2O) at T ¼ 288.15 K and 298.15 K and p ¼ 0.1 MPa. Journal of Chemical & Engineering Data. 2015;60:2594–2599.
[64] Zhang N, Guo YF, Liu YH, Wang SQ, Deng TL. Thermodynamic phase equilibria of the aqueous ternary system (LiCl þ LiBO2 þ H2O) at 308 K: Experimental data and predictions using the Pitzer model. Journal of Chemical Engineering of Japan. 2016;49:324–
[65] Cao LN, Li L, Zhang N, Guo YF, Deng TL. Phase equilibria of quaternary system LiCl– LiBO2–Li2SO4–H2O at 298.15 K. Chinese Journal of Chemical Engineering. 2016;67:
[66] Guo, YF, Li L, Cao LN, Yu XP, Wang SQ, Deng TL. Solubilities, densities and refractive indices in the aqueous quaternary system of lithium sulphate, lithium metaborate, and lithium carbonate at 288.15, 298.15, 308.15 K and 0.1 MPa. Journal of Chemical &
[67] Li L, Guo YF, Zhang SS, Sheng MM, Deng TL. Phase equilibria in the aqueous ternary systems (LiCl þ LiBO2 þ H2O) and (Li2SO4þ LiBO2 þ H2O) at 323.15 K and 0.1 MPa.
[68] Hu KY, Chai WQ, Liu DG. A study of phase equilibria in the quaternary aqueous salt system Hþ, Liþ, Mg2þ//Cl- – H2O at 0�, 20�, 40�. Liu Dagang's Science Works Collection.
[69] Zhang BJ, Sun B, Deng TL. Progresses on meta-stable phase equilibria of salt-water system. Journal of Salt and Chemical Industry. 2007;36:39–34. DOI: 10.3969/j.issn.1673-
[71] Deng TL, Li D. Solid-liquid metastable equilibria in the quaternary system (NaCl þ LiCl þ CaCl2 þ H2O) at 288.15 K. Journal of Chemical & Engineering Data. 2008;53:2488–
[72] Zeng Y, Yu XD, Zhang JQ, Li LG. Metastable phase equilibria in the aqueous ternary system containing potassium, lithium and chloride ions at 298.15 K. Advanced Materials Research. 2011;233–235:1619–1622. DOI: 10.4028/www.scientific.net/AMR.233-
[73] Chang WH, Zeng Y, Yan SY, Zhang ZH. Metastable phase equilibrium in the ternary system LiCl þ KCl þ H2O at 323 K. Journal of Salt & Chemical Industry. 2012;41:26–28
–H2O
–H2O at 35�. [Dissertation]. Chengdu University
[70] Zhang BJ. Metastable phase equilibrium of quaternary system Liþ, Naþ, Mg2þ//Cl-
Phase Equilibria. 2014;371:121–124. DOI: 10.1016/j.fluid.2014.03.019
DOI: 10.1021/acs.jced.5b00121
94 Desalination
331. DOI: 10.1252/jcej.15we045
1117–1122. DOI: 10.11949/j.issn.0438-1157.20151204
Engineering Data. DOI: 10.1021/acs.jced.6b00777
Beijing: Science Press; 1977. pp. 162–183
and ternary system Naþ(Kþ), Mg2þ//Cl-
6850.2007.01.014
of Technology; 2007
235.1619
2492. DOI: 10.1021/je8000798
Fluid Phase Equilibria. DOI: 10.1016/j.fluid.2016.12.013
[87] Bu BH, Li L, Zhang N, Guo YF, Wang SQ, Sun LY, Deng TL..Solid–liquid metastable phase equilibria for the ternary system (Li2SO4 þ K2SO4 þ H2O) at 288.15 and 323.15 K, p ¼ 0.1 MPa.Fluid Phase Equilibria. 2015;402:78–82. DOI: 10.1016/j.fluid.2015.05.044
[101] Pitzer KS. Thermodynamics of electrolytes I: Theoretical basis and general equations.
Phase Equilibria and Phase Separation of the Aqueous Solution System Containing Lithium Ions
http://dx.doi.org/10.5772/intechopen.68363
97
[102] Pitzer KS. Activity Coefficients in Electrolyte Solutions. 2nd ed. Boca Raton: The Chem-
[103] Deng TL, Li J, Li L, Zhang SS, Guo YF. Application of calorimetry to the study of thermodynamic properties of borate systems. Journal of Tianjing University of Science
[104] Yao Y, Sun B, Song PS, Zhang Z, Wang RL, Chen JQ. Thermodynamics of aqueous electrolyte solution—Isopiestic determination of osmotic and activity coefficients in
[105] Zhang Z, Yao Y, Song PS, Chen JQ. Isopiestic determination of the osmotic and activity coefficients of aqueous mixtures of Li2SO4 and MgSO4. Acta Physico-Chimica Sinica.
[106] Yang JM, Yao Y, Xia QY, Li BH, Xu SY, Yang JY. Isopiestic determination of the osmotic coefficients and Pitzer model representation for the Li2B4O7 þ LiCl þ H2O system at T ¼ 298.15 K. Journal of Solution Chemistry. 2008;37:377–389. DOI:
[107] Li J, Song PS, Yao Y, Wang RL. Thermodynamic properties of KCl–LiCl–H2O at 25�.
[108] Wang DB, Song PS, Yang JZ. Thermodynamics of mixture of boric acid with lithium borate and chloride. Chinese Journal of Chemistry. 1994;12:97–104. DOI: 10.1002/
[109] Wu YC, Smith MB, Young TF. Heats of mixing of electrolytes having common ions. Journal of Physical Chemistry. 1965;69:1868–1872. DOI: 10.1021/j100890a014
[110] Wood RH, Smith RW. Heats of mixing of aqueous electrolytes. I. Concentration dependence of 1-1 electrolytes. Journal of Physical Chemistry. 2002;69:2974–2979. DOI:
[111] Leung WH, Millero FJ. The enthalpy of formation of magnesium sulphate ion pairs.
[112] Apelblat A. Enthalpy of solution of lithium sulphate and lithium sulphate monohydrate in water at 298.15 K. Journal of Chemical Thermodynamics. 1985;17:769–773. DOI:
[113] M Bán, Madarász J, Bombicz P, Pokol G, Gál S. Thermal and structural study on the lattice compound 1, 4-diammoniumbutane bis (theophyllinate). Thermochimica Acta.
[114] Li J, Li B, Gao SY. Thermochemistry of hydrated lithium borates. Journal of Chemical
Journal of Solution Chemistry. 1975;4:145–159. DOI: 10.1007/BF00649155
Journal Physical Chemistry. 1973;77:268–277. DOI: 10.1021/j100621a026
& Technology. 2016; 31:6–12. DOI: 10.13364/j.issn.1672-6510.20150098
LiCl–MgCl2–H2O at 25�. Acta Chimica Sinica. 1992;50:839–848
ical Rubber Company Press; 1991. pp. 75–153
1993;9:366–373
cjoc.19940120201
10.1021/j100893a025
10.1016/0021-9614(85)90108-9
2004;420:105–109. DOI: 10.1016/j.tca.2003.12.033
Thermodynamics. 1998;30:681–688. DOI: 10.1006/jcht.1997.0324
10.1007/s10953-007-9244-0
Acta Physico-Chimica Sinica. 1992;8:94–99
[101] Pitzer KS. Thermodynamics of electrolytes I: Theoretical basis and general equations. Journal Physical Chemistry. 1973;77:268–277. DOI: 10.1021/j100621a026
[87] Bu BH, Li L, Zhang N, Guo YF, Wang SQ, Sun LY, Deng TL..Solid–liquid metastable phase equilibria for the ternary system (Li2SO4 þ K2SO4 þ H2O) at 288.15 and 323.15 K, p ¼ 0.1 MPa.Fluid Phase Equilibria. 2015;402:78–82. DOI: 10.1016/j.fluid.2015.05.044 [88] Li L, Bu BH, Zhang N, Guo YF, Deng TL. Metastable phase equilibria for the ternary aqueous system of lithium sulphate and potassium sulphate at T ¼ 308.15 K: Experimental data and prediction using Pitzer model. Russian Journal of Inorganic Chemistry.
[89] Sang SH, Yin HA, Zeng Y, Liu FY. Study on metastable equilibria in quaternary system
[90] Yan SY, Yin HA, Liang Q, Tang ML, Wang LM. An experimental study on metastable
[91] Sang SH, Yu HY, Cai DZ. Study on metastable equilibria in quaternary system Liþ,
[92] Sang SH, Yin HA, Tang ML. (Liquid þ solid) phase equilibria in the quinary system Li<sup>þ</sup>
[93] Zeng Y, Shao M. Liquid-solid metastable equilibria in the quinary system Liþ, K<sup>þ</sup>//Cl-
[94] Sang SH, Yin HA, Lei NF. (Solid þ liquid) metastable equilibria in quaternary system Li2SO4 þ K2SO4 þ Li2B4O7 þ K2B4O7 þ H2O at 288 K. Journal of Chemical Thermody-
[95] Yan SY, Chen T, Liang Q, Yin HA, Zeng Y. A study on metastable equilibrium in the
[96] Zeng Y, Lin XF, Meng S. Study on metastable phase equilibrium of quaternary system
[97] Qu SD. Study on the metastable equilibrium of the quinary system Liþ, K<sup>þ</sup>//CO3
[98] Sang SH, Zhang X, Zhao XP, Xiao LJ. Metastable equilibria in the quinary system of Liþ,
[99] Gao SY, Song PS, Xia SP, Zheng MP. Salt Lake Chemistry: A New Type Salt Lake of
[100] Song PS, Yao Y, Li J. Thermodynamics and phase equilibria of the salt lake brine system
, CO3
2-–H2O at 288 K. Acta Chimica Sinica. 2006;64:2247–2253
, CO3
2- <sup>þ</sup> H2O at 288 K. Journal of Chemical & Engineering Data.
2-–H2O at 288 K. Chinese Journal of Inorganic Chemistry. 2005;21:
2-–H2O at 288.15 K. Journal of Chemical & Engineering Data. 2006;51:219–
2-, B4O7
2-–H2O at 273 K. Chemical Engineering (2008) 36:47–50
2-–H2O at 273 K [Dissertation]. Chengdu University of Technology; 2008
2-–H2O at 288 K. Acta Chimica Sinica.2010;68:476–480
2-–H2O at 298 K. Chemical
2-–H2O at 298 K. Chinese Journal of
,
2-,
2016;61:1169–1174. DOI: 10.1134/S0036023616090138
phase equilibria in the quaternary system Liþ, Kþ//Cl-
2-, CO3
Research and Application. 2008;20:166–168
2- <sup>þ</sup> B4O7
2005;50:1555–1558. DOI: 10.1002/chin.200548015
namics. 2006;38:173–178. DOI: 10.1016/j.jct.2005.04.002
Liþ, Na<sup>þ</sup>//SO4
96 Desalination
Naþ//SO42-, B4O7
þ Na<sup>þ</sup> þ K<sup>þ</sup> þCO3
2-, B4O7
Liþ, K<sup>þ</sup>//CO3
2-, B4O7
2-, SO4
SO4
K<sup>þ</sup>//CO3
222. DOI: 10.1021/je050337m
quaternary reciprocal system of Liþ/Cl-
Inorganic Chemistry. 2006;26:95–97
2-, B4O7
2-, B4O7
Boron and Lithium. Beijing: Science Press; 2007
at 25�. Progress in Chemistry. 2000;12:255–267
1316–1320
SO4
[115] Li P, Liu ZH. Hydrothermal synthesis, characterization, and thermodynamic properties of a new lithium borate, Li3B5O8(OH)2. Journal of Chemical & Engineering Data. 2010;55:2682–2686. DOI: 10.1021/je900947d
**Chapter 6**
**Marmara Seawater Desalination by Membrane**
**Distillation: Direct Consumption Assessment of**
Drinking water was produced from Marmara seawater by membrane distillation (MD). The best operating conditions were determined by batch experiments as: 0.45 μm PTFE, 30°C distillate temperature and temperature difference, and 270–360 L/h cross‐flow rates in feed‐distillate. Seawater desalination was carried out with 99.93% solute rejection and
desalinated water contained no organic carbon, turbidity, and nitrate, it seemed to be very suitable for immediate service with quality of 7.3 pH, clear, odor‐free, 76.0 μS/cm, 47.1 mg TDS/L, <0.001 color, and 0.01 mg boron/L. The product water lacked of vital cations, espe‐
healthy development. A holistic management approach towards satisfying specific water quality requirements in direct service of MD effluents to human consumption was pro‐ posed that jointly included in injecting into urban potable water, adding appropriate chem‐ icals into the effluent, and mixing effluents with raw or concentrated seawater (1:250/1:1000
**Keywords:** seawater desalination, drinking water production, membrane distillation,
The data of United Nations indicate that the rate of increase in water usage has been higher than twice the rate of increase in population over the past century. It is estimated that there
> © 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
for Marmara seawater) or brackish natural waters under hygienic precautions.
product water quality, human consumption suitability
h permeate flux in 66% concentration ratio by lab‐scale pilot system. Since the
, Ca2+, Mg2+ that are essentials for promoting osmotic balanced body liquid and
**Produced Drinking Water**
Serif Cakmak, Tugba Nur Yilmaz,
http://dx.doi.org/10.5772/intechopen.68653
**Abstract**
17.2 L/m2
cially Na+
**1. Introduction**
, K<sup>+</sup>
Coskun Aydiner, Derya Y. Koseoglu Imer,
Emin Ender Celebi and Yasemin Melek Tilki
Additional information is available at the end of the chapter
Salim Oncel, Esra Can Dogan, Ali Oguzhan Narci,
## **Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produced Drinking Water**
Coskun Aydiner, Derya Y. Koseoglu Imer, Salim Oncel, Esra Can Dogan, Ali Oguzhan Narci, Serif Cakmak, Tugba Nur Yilmaz, Emin Ender Celebi and Yasemin Melek Tilki
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.68653
#### **Abstract**
[115] Li P, Liu ZH. Hydrothermal synthesis, characterization, and thermodynamic properties of a new lithium borate, Li3B5O8(OH)2. Journal of Chemical & Engineering Data.
[116] Li P, Liu ZH. Synthesis, structure and thermodynamic property of a new lithium borate:
[117] Zhang ZH, Tao ZC, Yin GY. Heat capacities and thermodynamic functions of the aqueous Li2B4O7 solution in the temperature range from 80 K to 355 K. Journal of Chemical &
[118] Yin GY, Yao Y, Jiao BJ, Chen SP, Gao SL. Enthalpies of dilution of aqueous Li2B4O7 solutions at 298.15 K and application of ion-interaction model. Thermochimica Acta.
[119] Yin GY, Yao Y, Liu ZH. Enthalpies of dilution of aqueous mixed solutions of LiCl and Li2B4O7 at 298.15 K. Journal of Thermal Analysis and Calorimetry. 2009;95:377–380.
[120] Li JC, Zhai ZX, Zeng ZM, Wang LY. Thermo-chemistry of solution of salt lakes(IV): Heats of dilution, heat capacities and apparent molal enthalpies of Li2B4O7—Li2SO4—
[121] Li L, Zhang SS, Liu YH, Guo YF, Deng TL. Heat capacities and the ion-interaction of lithium tetraborate aqueous solution system. Chemical Journal of Chinese Universities.
LiCl—H2O system at 298.15K. Journal of Salt Lake Research. 1993;1:34–38
Li4[B8O13(OH)2].3H2O. Chinese Journal of Chemistry. 2012;33:847–853
Engineering Data. 2007;52:866–870. DOI: 10.1021/je060462d
2005;435:125–128. DOI: 10.1016/j.tca.2005.04.007
2016;37:349–353. DOI: 10.7503/cjcu20150594
DOI: 10.1007/s10973-008-9239-0
2010;55:2682–2686. DOI: 10.1021/je900947d
98 Desalination
Drinking water was produced from Marmara seawater by membrane distillation (MD). The best operating conditions were determined by batch experiments as: 0.45 μm PTFE, 30°C distillate temperature and temperature difference, and 270–360 L/h cross‐flow rates in feed‐distillate. Seawater desalination was carried out with 99.93% solute rejection and 17.2 L/m2 h permeate flux in 66% concentration ratio by lab‐scale pilot system. Since the desalinated water contained no organic carbon, turbidity, and nitrate, it seemed to be very suitable for immediate service with quality of 7.3 pH, clear, odor‐free, 76.0 μS/cm, 47.1 mg TDS/L, <0.001 color, and 0.01 mg boron/L. The product water lacked of vital cations, espe‐ cially Na+ , K<sup>+</sup> , Ca2+, Mg2+ that are essentials for promoting osmotic balanced body liquid and healthy development. A holistic management approach towards satisfying specific water quality requirements in direct service of MD effluents to human consumption was pro‐ posed that jointly included in injecting into urban potable water, adding appropriate chem‐ icals into the effluent, and mixing effluents with raw or concentrated seawater (1:250/1:1000 for Marmara seawater) or brackish natural waters under hygienic precautions.
**Keywords:** seawater desalination, drinking water production, membrane distillation, product water quality, human consumption suitability
## **1. Introduction**
The data of United Nations indicate that the rate of increase in water usage has been higher than twice the rate of increase in population over the past century. It is estimated that there
© 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
will be 1.8 billion people living in areas with scarce water resources by 2025, while two‐thirds of the global population reside in regions of water stress due to use, growth, and climate change [1]. Potable water production has also become a worldwide concern, for many com‐ munities, increasing of industrial processes, population growth, climate change, over exploi‐ tation of ground water and nearby river systems as well as demand exceed of conventional available water resources [2–7]. For firmly preservation and sustaining stable development of life on earth at present situation, there needs to use plentiful salty water to produce freshwater supplies capable of meeting the increasing demand [7, 8]. Therefore, countries are progres‐ sively turning to desalination technologies as a solution to obtain direct potable water from seawater [4, 6, 9].
et al. [27] used different nanofiber microfiltration membranes produced from PVDF material to obtain drinking water through the air gap MD process. This was the first time that an elec‐ trospun nanofiber membrane was used in MD. At temperature differences ranging between 25
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce...
approach could potentially compete with other conventional desalination systems. Fard et al. [28] investigated bench‐scale performance of DCMD process using flat sheet PTFE membrane under various conditions of inlet flow rates, temperatures, and salinity composition. A perme‐
cold stream sides. The rate of salt rejection during the conducted tests was really high with 99.9% and virtually independent of any operational parameters studied. It has been seen that DCMD is a viable and effective technology, which can produce high quality distillate in a con‐ sistent way from a very high salinity feed even with dramatic differences in quality compared to other methods of desalination like RO and MSF. Bouguecha et al. [29] attempted to run a DCMD pilot plant powered by solar energy using collectors plus PV panels under actual weather conditions in Jeddah, KSA, throughout two selected sunny days. It had been aimed to assess how the operating parameters affected the process performance in which the trans‐ membrane temperature difference (Δ*T*) and the fluid mass flow rates (hot and cold) consti‐ tuted the most remarkable operating parameters. A maximum permeate flux of 8.87 L/m2
achieved at a Δ*T* of 60.5°C. According to their findings, it seems that the DCMD is a promising solution for the desalination of brackish water and also for seawater, particularly in distant
In this paper, six different commercially available membranes were used to investigate the effects of different operating conditions including cross‐flow rate, membrane type and pore size, solution temperatures, and membrane trans‐temperature differences on dissolved ions rejections and permeate flux of DCMD process. Some characteristics of membranes such as roughness and wettability were additionally tested to more comprehensively understand the performances. At suitable operation conditions determined based on batch experimental runs, the process was operated along a 30‐h period of time in which raw seawater was concentrated at approximately 70% to examine the direct usability of the MD output water as drinking water. The effluent quality established for direct supply to human consumptions was evalu‐ ated in terms of appropriateness to maximum allowable concentrations in the national stan‐ dard and international guidelines, and a general framework for pragmatic solutions toward practical water quality management applications was proposed for facilitation to serve the
Raw seawater used as feed stream in all membrane distillation experiments was collected from Muallimköy Coast of Marmara Sea beside the coastal city of İzmit/Kocaeli in Turkey. It was taken from under one meter of seawater level and then pre‐filtered through roughing filtra‐ tion to remove large particles. The detailed characteristics of the seawater are given in **Table 1**.
places and/or whenever affordable low temperature sources are accessible.
MD desalination effluents to the human drinking directly.
**2. Materials and methods**
**2.1. Seawater characterization**
h could be generated at a different temperature of 50°C between hot and
h. They also reported that this new
http://dx.doi.org/10.5772/intechopen.68653
h was
101
and 83°C, they obtained flux values between 5 and 28 kg/m2
ate flux of 35.6 L/m2
Commercial desalination technologies are mainly categorized under two main categories as pressure‐driven membrane separation by reverse osmosis (RO) and thermal distillation by multi‐stage flash (MSF) and multi‐effect distillation (MED) [4, 10]. In spite of their growing popularity and improved technological applications in seawater desalination, available plants have potential negative impacts on coastal environments and marine ecosystems, which can be characterized with concentrate and chemical discharges to the sea, limiting effects on land use, salting of groundwater, noise pollution, air pollutants discharges to the atmosphere and high energy consumptions [11–17]. Total capacity of seawater desalination plants in the world about 50% is operated by RO‐based membrane technology [4]. The decline in RO performance resulting from unavoidable membrane fouling, which needs expensive pretreatment, higher operation pressures, and frequent cleaning with chemicals which give harm to membranes, impairs the quality of permeate and accelerate membrane replacement, which pushes up the cost of water treatment and energy consumption [6, 18].
Membrane distillation (MD) could be an alternative to RO desalination process to overcome its negative impacts. MD is not highly affected by salt concentration in saline feed solutions, and hence, this technology can achieve good quality distillate with minimal brine discharge [19, 20]. MD, a thermal integrated membrane process, incorporates transporting vapor through microporous hydrophobic membranes, and its operation is based on the principle of vapor‐liquid equilibrium as a basis for molecular separation. This process uses a gradient of temperature between the two sides of a porous membrane in order to establish a differ‐ ence in vapor pressure that actually drives process [5, 19, 21–25]. Its four different configura‐ tions that consist of the type of the condensing design have been proposed, but direct contact membrane distillation (DCMD) is the most widely preferred technology because the step of condensation is performed within the membrane module, which brings about a simple mode of operation with no need of external condensers [19, 22, 24]. Additionally, more convenient membranes in MD process became available. Several materials such as polytetrafluoroethyl‐ ene (PTFE), polypropylene (PP), polyethylene (PE), and polyvinylidene fluoride (PVDF) are used for producing hydrophobic MD membrane for desalination uses [24, 25]. The different commercial membranes have been used in several recent pilot plant studies, but there are still some problems that have not yet been clearly solved to use MD process under real scale [19].
In practice, the use of MD for desalination is still largely limited to pilot‐scale studies. The com‐ mercial launch of MD desalination technology will first require certain technical issues to be resolved, such as its high energy consumption and the wetting of membrane pores [26]. Feng et al. [27] used different nanofiber microfiltration membranes produced from PVDF material to obtain drinking water through the air gap MD process. This was the first time that an elec‐ trospun nanofiber membrane was used in MD. At temperature differences ranging between 25 and 83°C, they obtained flux values between 5 and 28 kg/m2 h. They also reported that this new approach could potentially compete with other conventional desalination systems. Fard et al. [28] investigated bench‐scale performance of DCMD process using flat sheet PTFE membrane under various conditions of inlet flow rates, temperatures, and salinity composition. A perme‐ ate flux of 35.6 L/m2 h could be generated at a different temperature of 50°C between hot and cold stream sides. The rate of salt rejection during the conducted tests was really high with 99.9% and virtually independent of any operational parameters studied. It has been seen that DCMD is a viable and effective technology, which can produce high quality distillate in a con‐ sistent way from a very high salinity feed even with dramatic differences in quality compared to other methods of desalination like RO and MSF. Bouguecha et al. [29] attempted to run a DCMD pilot plant powered by solar energy using collectors plus PV panels under actual weather conditions in Jeddah, KSA, throughout two selected sunny days. It had been aimed to assess how the operating parameters affected the process performance in which the trans‐ membrane temperature difference (Δ*T*) and the fluid mass flow rates (hot and cold) consti‐ tuted the most remarkable operating parameters. A maximum permeate flux of 8.87 L/m2 h was achieved at a Δ*T* of 60.5°C. According to their findings, it seems that the DCMD is a promising solution for the desalination of brackish water and also for seawater, particularly in distant places and/or whenever affordable low temperature sources are accessible.
In this paper, six different commercially available membranes were used to investigate the effects of different operating conditions including cross‐flow rate, membrane type and pore size, solution temperatures, and membrane trans‐temperature differences on dissolved ions rejections and permeate flux of DCMD process. Some characteristics of membranes such as roughness and wettability were additionally tested to more comprehensively understand the performances. At suitable operation conditions determined based on batch experimental runs, the process was operated along a 30‐h period of time in which raw seawater was concentrated at approximately 70% to examine the direct usability of the MD output water as drinking water. The effluent quality established for direct supply to human consumptions was evalu‐ ated in terms of appropriateness to maximum allowable concentrations in the national stan‐ dard and international guidelines, and a general framework for pragmatic solutions toward practical water quality management applications was proposed for facilitation to serve the MD desalination effluents to the human drinking directly.
## **2. Materials and methods**
will be 1.8 billion people living in areas with scarce water resources by 2025, while two‐thirds of the global population reside in regions of water stress due to use, growth, and climate change [1]. Potable water production has also become a worldwide concern, for many com‐ munities, increasing of industrial processes, population growth, climate change, over exploi‐ tation of ground water and nearby river systems as well as demand exceed of conventional available water resources [2–7]. For firmly preservation and sustaining stable development of life on earth at present situation, there needs to use plentiful salty water to produce freshwater supplies capable of meeting the increasing demand [7, 8]. Therefore, countries are progres‐ sively turning to desalination technologies as a solution to obtain direct potable water from
Commercial desalination technologies are mainly categorized under two main categories as pressure‐driven membrane separation by reverse osmosis (RO) and thermal distillation by multi‐stage flash (MSF) and multi‐effect distillation (MED) [4, 10]. In spite of their growing popularity and improved technological applications in seawater desalination, available plants have potential negative impacts on coastal environments and marine ecosystems, which can be characterized with concentrate and chemical discharges to the sea, limiting effects on land use, salting of groundwater, noise pollution, air pollutants discharges to the atmosphere and high energy consumptions [11–17]. Total capacity of seawater desalination plants in the world about 50% is operated by RO‐based membrane technology [4]. The decline in RO performance resulting from unavoidable membrane fouling, which needs expensive pretreatment, higher operation pressures, and frequent cleaning with chemicals which give harm to membranes, impairs the quality of permeate and accelerate membrane replacement, which pushes up the
Membrane distillation (MD) could be an alternative to RO desalination process to overcome its negative impacts. MD is not highly affected by salt concentration in saline feed solutions, and hence, this technology can achieve good quality distillate with minimal brine discharge [19, 20]. MD, a thermal integrated membrane process, incorporates transporting vapor through microporous hydrophobic membranes, and its operation is based on the principle of vapor‐liquid equilibrium as a basis for molecular separation. This process uses a gradient of temperature between the two sides of a porous membrane in order to establish a differ‐ ence in vapor pressure that actually drives process [5, 19, 21–25]. Its four different configura‐ tions that consist of the type of the condensing design have been proposed, but direct contact membrane distillation (DCMD) is the most widely preferred technology because the step of condensation is performed within the membrane module, which brings about a simple mode of operation with no need of external condensers [19, 22, 24]. Additionally, more convenient membranes in MD process became available. Several materials such as polytetrafluoroethyl‐ ene (PTFE), polypropylene (PP), polyethylene (PE), and polyvinylidene fluoride (PVDF) are used for producing hydrophobic MD membrane for desalination uses [24, 25]. The different commercial membranes have been used in several recent pilot plant studies, but there are still some problems that have not yet been clearly solved to use MD process under real scale [19].
In practice, the use of MD for desalination is still largely limited to pilot‐scale studies. The com‐ mercial launch of MD desalination technology will first require certain technical issues to be resolved, such as its high energy consumption and the wetting of membrane pores [26]. Feng
cost of water treatment and energy consumption [6, 18].
seawater [4, 6, 9].
100 Desalination
#### **2.1. Seawater characterization**
Raw seawater used as feed stream in all membrane distillation experiments was collected from Muallimköy Coast of Marmara Sea beside the coastal city of İzmit/Kocaeli in Turkey. It was taken from under one meter of seawater level and then pre‐filtered through roughing filtra‐ tion to remove large particles. The detailed characteristics of the seawater are given in **Table 1**.
**2.2. MD membranes**
**2.3. Experimental setup**
brane area of 140 cm<sup>2</sup>
**Membranes Material Membrane**
**Table 2.** Various properties of MD membranes used.
PTFE Polytetrafluoro ethylene
PVDF Polyvinylidene fluoride
membranes, respectively.
**<sup>b</sup>**Not available.
**a**
a cooling system, is maintained at a low temperature.
**thickness (μm)**
values.
Six different types of flat sheet hydrophobic microfiltration membrane with pore sizes of 0.2, 0.45, and 1.0 μm that were made of PTFE and PVDF materials (Membrane Solutions Inc., China) were used in the experiments. The characteristics and intrinsic performances of the membranes are presented in **Table 2**. The membrane thicknesses were provided by the manufacturer. The results of contact angle measurement below 90° indicate that the PVDF membranes demonstrated lower hydrophobicity than the PTFE membranes. Liquid entry pressure (LEP) parameter also shows that the PTFE membranes are more suitable for the MD processing of seawater due to its lower wettability and higher entry pressure
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce...
http://dx.doi.org/10.5772/intechopen.68653
103
The experimental setup of cross‐flow lab‐scale pilot DCMD system is schematically shown in **Figure 1**. The setup was composed of two thermostatic cycles, that is, feed and permeate, which were connected to a membrane module made by kestamid having an effective mem‐
compressing the rectangular MD membrane. Connected to a heating resistant, the feed flow side is kept under high temperature; on the other hand, the permeate flow side, connected to
During the experimental investigations, rough‐filtrated raw seawater from the Marmara Sea was preheated to intended temperatures and circulated through one side of the membrane (feed side), while de‐ionized water was circulated through the other side of the membrane (distillate or permeate side) in a simultaneous way in a counter‐current flow mode. The condensation of permeate vapor that was diffused across the membrane occurred in the cold distillate side. The temperatures of feed and permeate streams were controlled at desirable levels using a heater and chiller, respectively. The spacer placed in the feed flow side only to promote turbulence
> **Nominal pore size (μm)**
The values given by manufacturer that measured at trans‐membrane pressures of 0.2 and 1.0 bar for PTFE and PVDF
. The compartment cells of the module consisted of two machined parts
**Contact angle (±10°)**
0.45 126 81.1 0.333 1.0 123 131.4 0.417
0.45 81 47.5 22.2–36.0 1.0 84 22.5 –**<sup>b</sup>**
160 ± 40 0.22 121 121.3 0.250
100 ± 10 0.22 68 58 6.0–9.0
**Liquid entry pressure (±2.5 kPa)**
**Water flux<sup>a</sup>**
**m2 h)** **(m3 /**
a Units of all parameters are mg/L except for temperature (°C), electrical conductivity (μS/cm), UVA254 (1/cm), SUVA (L/ mg.m) and color (1/cm).
b Seawaters I and II are raw seawater samples that were used in batch and continuous MD experiments, respectively.
**Table 1.** Characteristics of Marmara seawater.
#### **2.2. MD membranes**
**Parametersa Raw seawater**
NH<sup>4</sup>
102 Desalination
HCO<sup>3</sup>
NO<sup>3</sup>
CO<sup>3</sup>
SO<sup>4</sup>
Color
a
b
mg.m) and color (1/cm).
**Table 1.** Characteristics of Marmara seawater.
**Ib IIb Average**
*T* 24.1 24.2 24.15 ± 0.05 pH 8.69 8.62 8.66 ± 0.04 *Ec* 40,400 41,500 40,950 ± 550 TDS 25, 250 25, 772 25,511 ± 261 TOC 4.3 5.1 4.7 ± 0.4 DOC 4.1 4.9 4.5 ± 0.4 UVA254 0.050 0.040 0.045 ± 0.005 SUVA 1.220 0.816 1.018 ± 0.202 Na+ 8880 8940 8910 ± 30 K<sup>+</sup> 443 422 433 ± 11
<sup>+</sup> <0.1 <0.1 <0.1 Ba2+ 0.030 0.029 0.0295 ± 0.0005 Ca2+ 395 404 400 ± 5 Mg2+ 696 692 694 ± 2 Mn2+ 0.018 0.019 0.0185 ± 0.0005 Sr2+ 5.50 5.70 5.60 ± 0.10 B 2.95 3.05 3.00 ± 0.05 Si 0.89 0.91 0.90 ± 0.01 Fe 0.007 0.007 0.007 ± 0.000 Cl<sup>−</sup> 13,996 14,481 14,239 ± 243
<sup>−</sup> 300 303 302 ± 2
<sup>−</sup> 0.5 0.5 0.5 ± 0.0
2− 0.0 30 15.0 ± 15.0
2− 2314 2440 2377 ± 63 TKN 0.6 0.7 0.65 ± 0.05 Total nitrogen 1.1 1.2 1.15 ± 0.05 Alkalinity 300 333 317 ± 17 Total hardness 4986 5236 5111 ± 125
436 nm 0.0300 0.0300 0.0300 ± 0.0000 525 nm 0.0020 0.0100 0.0060 ± 0.0040 620 nm 0.0030 0.0020 0.0025 ± 0.0005
Units of all parameters are mg/L except for temperature (°C), electrical conductivity (μS/cm), UVA254 (1/cm), SUVA (L/
Seawaters I and II are raw seawater samples that were used in batch and continuous MD experiments, respectively.
Six different types of flat sheet hydrophobic microfiltration membrane with pore sizes of 0.2, 0.45, and 1.0 μm that were made of PTFE and PVDF materials (Membrane Solutions Inc., China) were used in the experiments. The characteristics and intrinsic performances of the membranes are presented in **Table 2**. The membrane thicknesses were provided by the manufacturer. The results of contact angle measurement below 90° indicate that the PVDF membranes demonstrated lower hydrophobicity than the PTFE membranes. Liquid entry pressure (LEP) parameter also shows that the PTFE membranes are more suitable for the MD processing of seawater due to its lower wettability and higher entry pressure values.
#### **2.3. Experimental setup**
The experimental setup of cross‐flow lab‐scale pilot DCMD system is schematically shown in **Figure 1**. The setup was composed of two thermostatic cycles, that is, feed and permeate, which were connected to a membrane module made by kestamid having an effective mem‐ brane area of 140 cm<sup>2</sup> . The compartment cells of the module consisted of two machined parts compressing the rectangular MD membrane. Connected to a heating resistant, the feed flow side is kept under high temperature; on the other hand, the permeate flow side, connected to a cooling system, is maintained at a low temperature.
During the experimental investigations, rough‐filtrated raw seawater from the Marmara Sea was preheated to intended temperatures and circulated through one side of the membrane (feed side), while de‐ionized water was circulated through the other side of the membrane (distillate or permeate side) in a simultaneous way in a counter‐current flow mode. The condensation of permeate vapor that was diffused across the membrane occurred in the cold distillate side. The temperatures of feed and permeate streams were controlled at desirable levels using a heater and chiller, respectively. The spacer placed in the feed flow side only to promote turbulence
**a** The values given by manufacturer that measured at trans‐membrane pressures of 0.2 and 1.0 bar for PTFE and PVDF membranes, respectively. **<sup>b</sup>**Not available.
**Table 2.** Various properties of MD membranes used.
*<sup>R</sup>rm<sup>s</sup>* <sup>=</sup> <sup>√</sup>
brane surfaces.
hardness, Na<sup>+</sup>
**2.5. Analytical procedure**
, K<sup>+</sup> , NH<sup>4</sup> +
of Hach HQ440 d by *4500‐NH*<sup>3</sup>
*titrimetric method*.
Na+ , K<sup>+</sup>
*for the Examination of Water and Wastewater*" [32].
\_\_\_\_\_\_\_\_\_\_ ∑ (*zcu* − *zav*)
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce...
The average difference in height among the five highest peaks and the five lowest val‐ leys, *R*Z, was calculated in relation to the mean plane, on which the image data have a minimum variance [30, 31]. The cross‐sectional morphologies of fouled membranes were monitored by scanning electron microscopy (SEM) (Philips XL30 SFEG) to visualize the fouling formations on the membrane surface. All the SEM observations were performed at 5 kV using Au‐coated membrane specimens. Energy dispersive X‐ray spectrometry (EDX) analysis of the components found on the membrane surface after the fouling was also done with EDX detector of the same device. The hydrophilicity or wettability of the membrane surfaces was analyzed with contact angle (°) measurements according to sessile‐drop technique using a goniometry instrument (Attension Theta Lite Optical Tensiometer) [31]. In the analyses, 2 μL of pure water in tight syringe was manually dropped on the membrane surface. By means of the software processing the measure‐ ment data obtained from the camera of the device, the results were determined as the averages of contact angles at both sides of drops fall on five arbitrary places of the mem‐
Water quality analyses in both MD streams were carried out based upon the parameters that included in temperature, pH, conductivity, turbidity, total dissolved solids TDS, total organic carbon (TOC), dissolved organic carbon (DOC), ultraviolet absorbance at 254 nm (UVA254),
color at 436, 525, and 620 nm wavelengths that were measured according to "*Standard Methods*
Temperature, conductivity, TDS and pH were analyzed with the desktop multi‐param‐ eter with Hach HQ440 d (Hach‐Lange GmBH). TOC was measured at 750°C by carbon analyzer equipped with a high pressure NDIR detector (Hach Lange IL550 TOC‐TN) in which *5310 B‐a high temperature catalytic oxidation method* was applied. DOC analyses were conducted on the samples filtered by Whatman syringe filter 0.45 μm using the TOC ana‐ lyzer apparatus. UVA254 was measured by *5910 B‐UV‐absorbing organic constituents method,* and SUVA was calculated with values of UVA254 and DOC. TKN was analyzed by *4500‐ N*org*B‐macro‐Kjeldahl method*, and ammonia was determined using the measurement probe
hardness were determined in accordance with *2320B‐titration method* and *2340C‐EDTA*
Optima 7000 DV with *3120‐B inductively coupled plasma mass spectrometer* (ICP‐MS) (Perkin‐ Elmer SCIEX Instruments, Canada). Si and Sr2+ were measured using flame atomic absorp‐ tion spectrometer (Perkin Elmer 1100). Prior to the analyses, samples were filtered through a
, Ba2+, Ca2+, Mg2+, Mn2+, B, and Fe concentrations were measured by Perkin‐Elmer ELAN
specific ultraviolet absorbance (SUVA), total Kjeldahl nitrogen (TKN), NH<sup>4</sup>
, Ba2+, Ca2+, Mg2+, Mn2+, Sr2+, B, Si, Fe, Cl−
2
\_\_\_\_\_\_\_\_\_\_ *<sup>p</sup>* (2)
http://dx.doi.org/10.5772/intechopen.68653
105
+
, HCO<sup>3</sup> − , NO<sup>3</sup> − , CO<sup>3</sup>
*E‐ammonia‐selective electrode method*. Alkalinity and total
, alkalinity, total
2− and
2−, SO<sup>4</sup>
**Figure 1.** Schematic flow diagram of experimental MD setup.
and to support the membrane was obtained from Sterlitech Inc., USA. In the batch experiments, the storage tanks for the feed and distillate streams in the system were kept at equal volumes of five L. In the continuous experiments operated at concentration mode of the feed seawater of 12 L, the system was worked for 6 hours per day in a period of 5 consecutive days. During these experimental runs, the membrane active layers were cleaned in‐place inside the module at the end of each operation day by means of using 1 L solutions in each one of the applications, which were operated by a flushing order of first applying "*1% HCl + distilled water*", and then followed by "*1% NaOH + distilled water*" except for the first operation day.
#### **2.4. Surface morphology of membranes**
The surface morphologies of clean and fouled membranes were marked by means of atomic force microscopy (AFM) to assess the vertically distribution of the fouling on the top layer or membrane surface roughness. After each membrane sample was let dry in air, the visual observations were provided using NanoScope IV AFM system (Digital Instruments, USA) operated in contact mode. The mean roughness (*R*<sup>a</sup> ), the root mean square (*R*rms) of the average height of membrane surface peaks, and the mean difference in height among the five highest peaks and the five lowest valleys (*R*<sup>z</sup> ) were identified in order to compare the roughness of clean and fouled membranes. Due to its more representative findings for the surface foulings, the analyses were evaluated only via the variations of the measured *R*<sup>z</sup> values of the specimens. In the AFM analyses, the value of mean roughness (*R*<sup>a</sup> ) stands for the mean value of surface in relation to centre plane and was calculated using the following equation [30]:
$$\mathcal{R}\_a = \frac{1}{P} \sum\_{\nu=0}^{P} \left| z\_{ca} - z\_{av} \right| \tag{1}$$
where *z*av is the average of the *z* values in the particular area, *z*cuis the present value of *z*, and *p* refers to the number of points in a certain area. The root mean square of *z* values (*R*rms) was calculated using the equation given below [30]:
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce... http://dx.doi.org/10.5772/intechopen.68653 105
$$R\_{rms} = \sqrt{\frac{\sum \left(z\_{in} - z\_{sr}\right)^2}{p}} \tag{2}$$
The average difference in height among the five highest peaks and the five lowest val‐ leys, *R*Z, was calculated in relation to the mean plane, on which the image data have a minimum variance [30, 31]. The cross‐sectional morphologies of fouled membranes were monitored by scanning electron microscopy (SEM) (Philips XL30 SFEG) to visualize the fouling formations on the membrane surface. All the SEM observations were performed at 5 kV using Au‐coated membrane specimens. Energy dispersive X‐ray spectrometry (EDX) analysis of the components found on the membrane surface after the fouling was also done with EDX detector of the same device. The hydrophilicity or wettability of the membrane surfaces was analyzed with contact angle (°) measurements according to sessile‐drop technique using a goniometry instrument (Attension Theta Lite Optical Tensiometer) [31]. In the analyses, 2 μL of pure water in tight syringe was manually dropped on the membrane surface. By means of the software processing the measure‐ ment data obtained from the camera of the device, the results were determined as the averages of contact angles at both sides of drops fall on five arbitrary places of the mem‐ brane surfaces.
#### **2.5. Analytical procedure**
and to support the membrane was obtained from Sterlitech Inc., USA. In the batch experiments, the storage tanks for the feed and distillate streams in the system were kept at equal volumes of five L. In the continuous experiments operated at concentration mode of the feed seawater of 12 L, the system was worked for 6 hours per day in a period of 5 consecutive days. During these experimental runs, the membrane active layers were cleaned in‐place inside the module at the end of each operation day by means of using 1 L solutions in each one of the applications, which were operated by a flushing order of first applying "*1% HCl + distilled water*", and then followed
The surface morphologies of clean and fouled membranes were marked by means of atomic force microscopy (AFM) to assess the vertically distribution of the fouling on the top layer or membrane surface roughness. After each membrane sample was let dry in air, the visual observations were provided using NanoScope IV AFM system (Digital Instruments, USA)
average height of membrane surface peaks, and the mean difference in height among the
roughness of clean and fouled membranes. Due to its more representative findings for the surface foulings, the analyses were evaluated only via the variations of the measured *R*<sup>z</sup>
the mean value of surface in relation to centre plane and was calculated using the following
*<sup>P</sup>* ∑ *i*=0 *p*
where *z*av is the average of the *z* values in the particular area, *z*cuis the present value of *z*, and *p* refers to the number of points in a certain area. The root mean square of *z* values (*R*rms) was
values of the specimens. In the AFM analyses, the value of mean roughness (*R*<sup>a</sup>
), the root mean square (*R*rms) of the
) stands for
) were identified in order to compare the
by "*1% NaOH + distilled water*" except for the first operation day.
operated in contact mode. The mean roughness (*R*<sup>a</sup>
five highest peaks and the five lowest valleys (*R*<sup>z</sup>
*<sup>R</sup><sup>a</sup>* <sup>=</sup> \_\_1
calculated using the equation given below [30]:
**2.4. Surface morphology of membranes**
**Figure 1.** Schematic flow diagram of experimental MD setup.
equation [30]:
104 Desalination
Water quality analyses in both MD streams were carried out based upon the parameters that included in temperature, pH, conductivity, turbidity, total dissolved solids TDS, total organic carbon (TOC), dissolved organic carbon (DOC), ultraviolet absorbance at 254 nm (UVA254), specific ultraviolet absorbance (SUVA), total Kjeldahl nitrogen (TKN), NH<sup>4</sup> + , alkalinity, total hardness, Na<sup>+</sup> , K<sup>+</sup> , NH<sup>4</sup> + , Ba2+, Ca2+, Mg2+, Mn2+, Sr2+, B, Si, Fe, Cl− , HCO<sup>3</sup> − , NO<sup>3</sup> − , CO<sup>3</sup> 2−, SO<sup>4</sup> 2− and color at 436, 525, and 620 nm wavelengths that were measured according to "*Standard Methods for the Examination of Water and Wastewater*" [32].
Temperature, conductivity, TDS and pH were analyzed with the desktop multi‐param‐ eter with Hach HQ440 d (Hach‐Lange GmBH). TOC was measured at 750°C by carbon analyzer equipped with a high pressure NDIR detector (Hach Lange IL550 TOC‐TN) in which *5310 B‐a high temperature catalytic oxidation method* was applied. DOC analyses were conducted on the samples filtered by Whatman syringe filter 0.45 μm using the TOC ana‐ lyzer apparatus. UVA254 was measured by *5910 B‐UV‐absorbing organic constituents method,* and SUVA was calculated with values of UVA254 and DOC. TKN was analyzed by *4500‐ N*org*B‐macro‐Kjeldahl method*, and ammonia was determined using the measurement probe of Hach HQ440 d by *4500‐NH*<sup>3</sup> *E‐ammonia‐selective electrode method*. Alkalinity and total hardness were determined in accordance with *2320B‐titration method* and *2340C‐EDTA titrimetric method*.
Na+ , K<sup>+</sup> , Ba2+, Ca2+, Mg2+, Mn2+, B, and Fe concentrations were measured by Perkin‐Elmer ELAN Optima 7000 DV with *3120‐B inductively coupled plasma mass spectrometer* (ICP‐MS) (Perkin‐ Elmer SCIEX Instruments, Canada). Si and Sr2+ were measured using flame atomic absorp‐ tion spectrometer (Perkin Elmer 1100). Prior to the analyses, samples were filtered through a 0.45 μm pore‐size membrane filter, and pH values of the samples were adjusted to <4.0 using HNO<sup>3</sup> . HCO<sup>3</sup> 2−and CO<sup>3</sup> 2− were analyzed by *2320‐B titrimetric method*, and SO<sup>4</sup> 2− and Cl‐were measured by *4500‐SO*<sup>4</sup> 2−*‐E turbidity and 4500‐Cl*<sup>−</sup> *‐D potentiometric methods*, respectively. NO<sup>3</sup> 2− was determined by *4500‐NO*<sup>3</sup> 2−*‐C* spectrophotometric method, while color measurements at 436, 525, and 620 nm wavelengths were conducted by *2120C* spectrophotometric method for making possible the time‐dependent calculations of MD water fluxes, the measurements of osmolalities of the feed and permeate solutions were executed at definite time inter‐ vals in duplicate for each data point using the Advanced Osmometer instrument (Model 3250—Advanced Instruments Inc., USA) by the method of freezing point depression.
#### **2.6. MD performance analysis**
The technical performance of MD was analyzed using not only the permeate flux values but the rejection values of the parameters as well, which are given in national and international guidelines for water reuse in drinking water quality. The flux was calculated from the time‐ dependent variations between the pretest and posttest values of TDS and osmolality as an explicit indicator for solute ions transports from seawater to distillate stream. So, after TDS and osmolality at both sides were individually measured for definite time intervals, *V* was calculated from the differences of sequential time points according to TDS/osmolarity‐based mass balances calculations. From the average values of *V* volume calculated based on TDS and osmolality data measured at each time point, permeate water fluxes were totally obtained for the whole time scale of relevant experiment using Eq. (3).
$$J = \frac{1}{A} \frac{dV}{dt} \tag{3}$$
The experimental performance levels were determined based on the passing water volume, the seawater concentration, and the water flux parameters. Initial experiments were per‐ formed with a 30°C difference between the solution temperatures, which are the generally recommended level of difference in the literature [21], by maintaining seawater and perme‐ ate water at temperatures of 55 and 25°C, respectively, and by using seawater and permeate water flow rates of 270 and 360 L/h, respectively. In general, available polymeric materials for manufacturing hydrophobic membranes suitable for MD are, typically, PP, PVDF, and PTFE
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce...
With PTFE membranes, water flux values for 0.22, 0.45, and 1.0 μm sized pores were deter‐
PTFE membranes with 0.22, 0.45, and 1.0 μm pore sizes were determined as 4.3, 7.9, and 9.8 mg/L, respectively. **Figure 2** shows permeate flux and solute concentration for the MD membranes with different materials (PTFE and PVDF) and pore sizes (0.22, 0.45, and 1 μm)
An evaluation of the permeate solute ion concentrations with the three different pore sizes revealed that increasing pore size led to higher concentrations in the permeate. Based on the observed water flux performances, it was determined that the PTFE membrane with 0.45 μm pore was the most suitable type of membrane for obtaining potable water from seawater. The
Distillate contamination resulting from wetting of membrane pore is among the chief fac‐ tors, which impede the broader application of the MD technology. In order to avoid pore wetting, it is necessary that the membrane material be hydrophobic with as high a contact angle as possible. Moreover, it is recommended that the membrane has a comparatively small maximum pore size [34]. The hydrophobic membrane can obstruct the penetration of liq‐ uid through surface tension force, but not that of vapor. Consequently, water vapor will be capable of passing from the hot solution side of higher vapor pressure to the cold distillate side of lower vapor pressure [35]. **Figure 3** shows measurement results of contact angle (θ°)
) for surface characterizations of clean (θ<sup>c</sup>
Results of AFM and SEM evaluations performed before and after filtration on the active lay‐ ers of the MD membranes are shown in **Figure 4**. The high pore sizes used in the experiments led to greater accumulation of inorganic pollutants inside the pores than on the surface. In addition, the surface fouling layer was observed to be soluble. Consequently, no significant differences were observed on the SEM images. An evaluation of the EDX results revealed that, depending on the structure of polymers used for producing the membranes, and the added chemicals, the membrane surfaces contained the elements boron (B), fluorine (F), carbon (C), nitrogen (N), and oxygen (O). It was observed that the fouling of the surface was largely local and lacked a large variety of ions, and when a prevalent distribution of ions was actu‐
under the effect of temperature [36]. Inorganic foulant materials were not observed on the membrane surfaces. It is hence possible to say that no different elements were observed on
ally present, this was found to be due to the crystallized precipitates (CaCO<sup>3</sup>
PTFE (0.45 μm) membrane was hence chosen for the next stage.
h, respectively. With PVDF membranes, these values were
h, respectively. TDS values in output water for the
http://dx.doi.org/10.5772/intechopen.68653
107
and *R*z,c) and fouled (θ<sup>f</sup>
, CaSO<sup>4</sup>
) formed
[19, 21].
mined as 27.7, 40.7, and 28.5 L/m2
under batch operating conditions.
and surface roughness (*R*<sup>z</sup>
and *R*z,f) membranes.
determined as 11.3, 19.4, and 29.6 L/m2
where *J* is the permeate flux; *A* is the effective membrane filtration area; *V* is the total permeate volume; and *t* is the filtration time.
The rejection performances in MD process for each water quality parameter were calculated from the water quality analyses results using Eq. (4):
$$R(\%) = \left(1 - \frac{C\_r}{C\_\gamma}\right) \times 100\tag{4}$$
where *R*, *Cp ,* and *Cf* are the rejection, concentration in the permeate, and concentration in the feed stream, respectively [33].
#### **3. Results and discussion**
#### **3.1. The effects of operating conditions on MD performance**
#### *3.1.1. Membrane type and size*
The first part of the study was performed under batch experimental runs by using different membranes to examine the effect of different pore sizes and materials on the MD performance. The experimental performance levels were determined based on the passing water volume, the seawater concentration, and the water flux parameters. Initial experiments were per‐ formed with a 30°C difference between the solution temperatures, which are the generally recommended level of difference in the literature [21], by maintaining seawater and perme‐ ate water at temperatures of 55 and 25°C, respectively, and by using seawater and permeate water flow rates of 270 and 360 L/h, respectively. In general, available polymeric materials for manufacturing hydrophobic membranes suitable for MD are, typically, PP, PVDF, and PTFE [19, 21].
0.45 μm pore‐size membrane filter, and pH values of the samples were adjusted to <4.0 using
at 436, 525, and 620 nm wavelengths were conducted by *2120C* spectrophotometric method for making possible the time‐dependent calculations of MD water fluxes, the measurements of osmolalities of the feed and permeate solutions were executed at definite time inter‐ vals in duplicate for each data point using the Advanced Osmometer instrument (Model
The technical performance of MD was analyzed using not only the permeate flux values but the rejection values of the parameters as well, which are given in national and international guidelines for water reuse in drinking water quality. The flux was calculated from the time‐ dependent variations between the pretest and posttest values of TDS and osmolality as an explicit indicator for solute ions transports from seawater to distillate stream. So, after TDS and osmolality at both sides were individually measured for definite time intervals, *V* was calculated from the differences of sequential time points according to TDS/osmolarity‐based mass balances calculations. From the average values of *V* volume calculated based on TDS and osmolality data measured at each time point, permeate water fluxes were totally obtained
> *A* \_\_\_ *dV*
where *J* is the permeate flux; *A* is the effective membrane filtration area; *V* is the total permeate
The rejection performances in MD process for each water quality parameter were calculated
The first part of the study was performed under batch experimental runs by using different membranes to examine the effect of different pore sizes and materials on the MD performance.
*C* \_*p*
are the rejection, concentration in the permeate, and concentration in the
3250—Advanced Instruments Inc., USA) by the method of freezing point depression.
2−*‐E turbidity and 4500‐Cl*<sup>−</sup>
for the whole time scale of relevant experiment using Eq. (3).
*J* = \_\_1
from the water quality analyses results using Eq. (4):
**3.1. The effects of operating conditions on MD performance**
*<sup>R</sup>*(%) <sup>=</sup> (<sup>1</sup> <sup>−</sup>
volume; and *t* is the filtration time.
*,* and *Cf*
**3. Results and discussion**
*3.1.1. Membrane type and size*
feed stream, respectively [33].
where *R*, *Cp*
2− were analyzed by *2320‐B titrimetric method*, and SO<sup>4</sup>
2− and Cl‐were
2−
*‐D potentiometric methods*, respectively. NO<sup>3</sup>
*dt* (3)
*Cf*) <sup>×</sup> <sup>100</sup> (4)
2−*‐C* spectrophotometric method, while color measurements
HNO<sup>3</sup>
106 Desalination
. HCO<sup>3</sup>
measured by *4500‐SO*<sup>4</sup>
2−and CO<sup>3</sup>
was determined by *4500‐NO*<sup>3</sup>
**2.6. MD performance analysis**
With PTFE membranes, water flux values for 0.22, 0.45, and 1.0 μm sized pores were deter‐ mined as 27.7, 40.7, and 28.5 L/m2 h, respectively. With PVDF membranes, these values were determined as 11.3, 19.4, and 29.6 L/m2 h, respectively. TDS values in output water for the PTFE membranes with 0.22, 0.45, and 1.0 μm pore sizes were determined as 4.3, 7.9, and 9.8 mg/L, respectively. **Figure 2** shows permeate flux and solute concentration for the MD membranes with different materials (PTFE and PVDF) and pore sizes (0.22, 0.45, and 1 μm) under batch operating conditions.
An evaluation of the permeate solute ion concentrations with the three different pore sizes revealed that increasing pore size led to higher concentrations in the permeate. Based on the observed water flux performances, it was determined that the PTFE membrane with 0.45 μm pore was the most suitable type of membrane for obtaining potable water from seawater. The PTFE (0.45 μm) membrane was hence chosen for the next stage.
Distillate contamination resulting from wetting of membrane pore is among the chief fac‐ tors, which impede the broader application of the MD technology. In order to avoid pore wetting, it is necessary that the membrane material be hydrophobic with as high a contact angle as possible. Moreover, it is recommended that the membrane has a comparatively small maximum pore size [34]. The hydrophobic membrane can obstruct the penetration of liq‐ uid through surface tension force, but not that of vapor. Consequently, water vapor will be capable of passing from the hot solution side of higher vapor pressure to the cold distillate side of lower vapor pressure [35]. **Figure 3** shows measurement results of contact angle (θ°) and surface roughness (*R*<sup>z</sup> ) for surface characterizations of clean (θ<sup>c</sup> and *R*z,c) and fouled (θ<sup>f</sup> and *R*z,f) membranes.
Results of AFM and SEM evaluations performed before and after filtration on the active lay‐ ers of the MD membranes are shown in **Figure 4**. The high pore sizes used in the experiments led to greater accumulation of inorganic pollutants inside the pores than on the surface. In addition, the surface fouling layer was observed to be soluble. Consequently, no significant differences were observed on the SEM images. An evaluation of the EDX results revealed that, depending on the structure of polymers used for producing the membranes, and the added chemicals, the membrane surfaces contained the elements boron (B), fluorine (F), carbon (C), nitrogen (N), and oxygen (O). It was observed that the fouling of the surface was largely local and lacked a large variety of ions, and when a prevalent distribution of ions was actu‐ ally present, this was found to be due to the crystallized precipitates (CaCO<sup>3</sup> , CaSO<sup>4</sup> ) formed under the effect of temperature [36]. Inorganic foulant materials were not observed on the membrane surfaces. It is hence possible to say that no different elements were observed on
**Figure 2.** Permeate flux (a) and solute concentration (b) for MD membranes with different materials (PTFE and PVDF) and pore sizes (0.22, 0.45 and 1.0 μm) under batch operating conditions (seawater/permeate: 55/25°C (ΔT: 30°C) and 270/360 L/h cross‐flow rates at seawater/permeate streams).
the membranes at the end of the process. The presence of such low and dilute levels of fouling during the experiments clearly indicated that under effective pass vapor pressure and low levels of pollution, the MD membrane is fairly efficient in providing drinking water from seawater.
*3.1.2. Cross‐flow rate*
(PTFE‐0.45 μm and PVDF‐1 μm).
*3.1.3. Solutions' temperatures*
secutively reached 23.4, 29.9, 40.7, and 43.4 L/m2
The effect of the flow rates in membrane channel during the MD process was evaluated by applying flow rates with different Reynolds values (90/120, 180/240, 270/360, and 360/480 L/h seawater/permeate water flow velocity values) through a PTFE membrane at a fixed mem‐ brane trans‐temperature difference of 30±0.5°C. It was observed that increasing flow rate led to a higher volume of water passing from the seawater to the permeate flow, which resulted in an increase in permeate flux performances. With increasing flow rates, water flux levels con‐
**Figure 4.** AFM (10 μm × 10 μm) images and SEM (1.4 mm × 0.9 mm) microphotographs of fouled membranes
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce...
http://dx.doi.org/10.5772/intechopen.68653
109
as 5.3, 6.0, 7.9, and 9.75 mg/L, respectively, and output water with low solute ions was obtained in all flow velocity experiments. **Figure 5** shows permeate flux and salt concentration for dif‐ ferent cross‐flow rates (L/h) at seawater/permeate water under batch operating conditions.
Under the turbulence regime, similar performance values were observed in the 270/360 L/h and 360/480 L/h seawater/distilled water flow rate pairs. Taking into account that increasing flow rates might increase costs in practice, suitable flow rates were determined as 270 L/h
In the next part of the study, experiments were conducted with 0.45 μm PTFE membranes, a fixed membrane trans‐temperature difference value (30°C), a seawater and permeate water
(*Re*av: 4320) for seawater and 360 L/h (*Re*av: 4222) for permeate water.
h. Output water TDS values were determined
**Figure 3.** Measurement results of contact angle (θ°) and surface roughness (*R*<sup>z</sup> ) for surface characterizations of clean (θ<sup>c</sup> and *R*z,c) and fouled (θ<sup>f</sup> and *R*z,f) membranes.
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce... http://dx.doi.org/10.5772/intechopen.68653 109
**Figure 4.** AFM (10 μm × 10 μm) images and SEM (1.4 mm × 0.9 mm) microphotographs of fouled membranes (PTFE‐0.45 μm and PVDF‐1 μm).
#### *3.1.2. Cross‐flow rate*
the membranes at the end of the process. The presence of such low and dilute levels of fouling during the experiments clearly indicated that under effective pass vapor pressure and low levels of pollution, the MD membrane is fairly efficient in providing drinking water from
**Figure 2.** Permeate flux (a) and solute concentration (b) for MD membranes with different materials (PTFE and PVDF) and pore sizes (0.22, 0.45 and 1.0 μm) under batch operating conditions (seawater/permeate: 55/25°C (ΔT: 30°C) and
) for surface characterizations of clean (θ<sup>c</sup>
**Figure 3.** Measurement results of contact angle (θ°) and surface roughness (*R*<sup>z</sup>
and *R*z,f) membranes.
270/360 L/h cross‐flow rates at seawater/permeate streams).
seawater.
108 Desalination
and *R*z,c) and fouled (θ<sup>f</sup>
The effect of the flow rates in membrane channel during the MD process was evaluated by applying flow rates with different Reynolds values (90/120, 180/240, 270/360, and 360/480 L/h seawater/permeate water flow velocity values) through a PTFE membrane at a fixed mem‐ brane trans‐temperature difference of 30±0.5°C. It was observed that increasing flow rate led to a higher volume of water passing from the seawater to the permeate flow, which resulted in an increase in permeate flux performances. With increasing flow rates, water flux levels con‐ secutively reached 23.4, 29.9, 40.7, and 43.4 L/m2 h. Output water TDS values were determined as 5.3, 6.0, 7.9, and 9.75 mg/L, respectively, and output water with low solute ions was obtained in all flow velocity experiments. **Figure 5** shows permeate flux and salt concentration for dif‐ ferent cross‐flow rates (L/h) at seawater/permeate water under batch operating conditions.
Under the turbulence regime, similar performance values were observed in the 270/360 L/h and 360/480 L/h seawater/distilled water flow rate pairs. Taking into account that increasing flow rates might increase costs in practice, suitable flow rates were determined as 270 L/h (*Re*av: 4320) for seawater and 360 L/h (*Re*av: 4222) for permeate water.
#### *3.1.3. Solutions' temperatures*
In the next part of the study, experiments were conducted with 0.45 μm PTFE membranes, a fixed membrane trans‐temperature difference value (30°C), a seawater and permeate water
permeate flow temperature of 30°C, these values were increased to 31.3, 48.0, and 63.0 L/m2
trans‐temperature differences.
30°C temperatures at permeate).
and CaSO<sup>4</sup>
result in CaCO<sup>3</sup>
respectively. At a permeate flow temperature of 15°C, the output water TDS values for mem‐ brane trans‐temperature differences of 20, 30, and 40°C were determined as 6.5, 6.8, and 7.4 mg/L, respectively. On the other hand, at a permeate flow temperature of 30°C, the output water TDS values for membrane trans‐temperature differences of 20, 30, and 40°C were determined as 6.75, 8.3, and 10.3 mg/L, respectively. In all temperature difference experiments, the output water con‐ tained low levels of dissolved solids. **Figure 7** shows permeate flux and effective permeate flux per unit of membrane trans‐temperature difference and total solutes concentration for varying
**Figure 6.** Permeate flux (a) and solute concentration (b) for different temperatures (°C) at seawater/permeate solutions under batch operating conditions (PTFE—0.45 μm, 270/360 L/h cross‐flow rates at seawater/permeate, ΔT: 30°C).
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce...
http://dx.doi.org/10.5772/intechopen.68653
As described in the literature, continuous operation at higher seawater temperature may
active layer, as well as the pore spaces within the membrane [36]. Based on this consideration, 30°C can be evaluated as the suitable trans‐difference among the solutions for the continuous
**Figure 7.** Permeate flux and effective permeate flux per unit of trans‐temperature difference (a) and solute concentration (b) for varying trans‐temperature differences (PTFE—0.45 μm, 270/360 L/h cross‐flow rates at seawater/permeate, 15 and
precipitates that foul/clog the pore entrances on the membrane
h,
111
**Figure 5.** Permeate flux (a) and solute concentration (b) for different cross‐flow rates (L/h) at seawater/permeate water under batch operating conditions (PTFE—0.45 μm and seawater/permeate: 55/25°C (ΔT: 30°C)).
flow rates of 270 and 360 L/h, respectively, and distilled water temperatures of 15, 20, 25, and 30°C in order to determine the effect of solution temperature on process performance. The performance results of the experiments were evaluated at seawater‐permeate water operation temperatures of 45–15, 50–20, 55–25, and 60–30°C based on the water flux and output water TDS parameters.
It was determined that the water flux values of the process gradually increased in parallel with increasing solution temperature, as 40.3, 32.7, 36.8, 40.7, and 48.0 L/m2 h, respectively. Total solute concentrations in MD output water were determined as 7.2, 7.5, 8.0, and 8.7 mg/L, respectively, and it was clearly observed that the output water had low dissolved solids. **Figure 6** shows permeate flux and salt concentration for different temperatures at seawater/ permeate solutions under batch operating conditions.
The CaCO<sup>3</sup> and CaSO<sup>4</sup> compounds, which have lower solubility at high temperatures, can cause a serious reduction in water permeation performances of MD membranes by fouling occurring on the membrane surface during the processing. A previous study reported that increasing the input feed flow temperature from 80 to 90°C led to a fourfold decrease in flow level [36]. At membrane, considering the possibility of Ca2+ precipitation that might result continuous operation and constant membrane trans‐temperature difference, the most suitable was to obtain high flux outputs to continually operate the membranes at fixed permeate water and seawater flow temperatures of 30 and 60°C, respectively.
#### *3.1.4. Membrane trans‐temperature difference*
It was observed that increasing membrane trans‐temperature difference (20, 30, and 40°C) led to an increase in the volume of water passing from the seawater to the permeate water. At a permeate flow temperature of 15°C, the water flux values for membrane trans‐temperature dif‐ ferences of 20, 30, and 40°C were determined as 17.0, 32.7, and 53.3 L/m2 h, respectively. At a
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce... http://dx.doi.org/10.5772/intechopen.68653 111
**Figure 6.** Permeate flux (a) and solute concentration (b) for different temperatures (°C) at seawater/permeate solutions under batch operating conditions (PTFE—0.45 μm, 270/360 L/h cross‐flow rates at seawater/permeate, ΔT: 30°C).
**Figure 5.** Permeate flux (a) and solute concentration (b) for different cross‐flow rates (L/h) at seawater/permeate water
flow rates of 270 and 360 L/h, respectively, and distilled water temperatures of 15, 20, 25, and 30°C in order to determine the effect of solution temperature on process performance. The performance results of the experiments were evaluated at seawater‐permeate water operation temperatures of 45–15, 50–20, 55–25, and 60–30°C based on the water flux and output water
It was determined that the water flux values of the process gradually increased in parallel
Total solute concentrations in MD output water were determined as 7.2, 7.5, 8.0, and 8.7 mg/L, respectively, and it was clearly observed that the output water had low dissolved solids. **Figure 6** shows permeate flux and salt concentration for different temperatures at seawater/
cause a serious reduction in water permeation performances of MD membranes by fouling occurring on the membrane surface during the processing. A previous study reported that increasing the input feed flow temperature from 80 to 90°C led to a fourfold decrease in flow level [36]. At membrane, considering the possibility of Ca2+ precipitation that might result continuous operation and constant membrane trans‐temperature difference, the most suitable was to obtain high flux outputs to continually operate the membranes at fixed permeate water
It was observed that increasing membrane trans‐temperature difference (20, 30, and 40°C) led to an increase in the volume of water passing from the seawater to the permeate water. At a permeate flow temperature of 15°C, the water flux values for membrane trans‐temperature dif‐
compounds, which have lower solubility at high temperatures, can
h, respectively.
h, respectively. At a
under batch operating conditions (PTFE—0.45 μm and seawater/permeate: 55/25°C (ΔT: 30°C)).
with increasing solution temperature, as 40.3, 32.7, 36.8, 40.7, and 48.0 L/m2
permeate solutions under batch operating conditions.
and seawater flow temperatures of 30 and 60°C, respectively.
ferences of 20, 30, and 40°C were determined as 17.0, 32.7, and 53.3 L/m2
and CaSO<sup>4</sup>
*3.1.4. Membrane trans‐temperature difference*
TDS parameters.
110 Desalination
The CaCO<sup>3</sup>
permeate flow temperature of 30°C, these values were increased to 31.3, 48.0, and 63.0 L/m2 h, respectively. At a permeate flow temperature of 15°C, the output water TDS values for mem‐ brane trans‐temperature differences of 20, 30, and 40°C were determined as 6.5, 6.8, and 7.4 mg/L, respectively. On the other hand, at a permeate flow temperature of 30°C, the output water TDS values for membrane trans‐temperature differences of 20, 30, and 40°C were determined as 6.75, 8.3, and 10.3 mg/L, respectively. In all temperature difference experiments, the output water con‐ tained low levels of dissolved solids. **Figure 7** shows permeate flux and effective permeate flux per unit of membrane trans‐temperature difference and total solutes concentration for varying trans‐temperature differences.
As described in the literature, continuous operation at higher seawater temperature may result in CaCO<sup>3</sup> and CaSO<sup>4</sup> precipitates that foul/clog the pore entrances on the membrane active layer, as well as the pore spaces within the membrane [36]. Based on this consideration, 30°C can be evaluated as the suitable trans‐difference among the solutions for the continuous
**Figure 7.** Permeate flux and effective permeate flux per unit of trans‐temperature difference (a) and solute concentration (b) for varying trans‐temperature differences (PTFE—0.45 μm, 270/360 L/h cross‐flow rates at seawater/permeate, 15 and 30°C temperatures at permeate).
operation in which seawater or feed stream would be heated up to max 60°C. Taking into account the fact that energy costs represent the most important obstacle standing in the way of the rising uses of MD process compared to RO seawater desalination plants, it was com‐ prehended that 30°C and some below or above differences would render further possible to arrive the optimal cost solutions in on‐land MD plants due to lower heat energy consumption per unit of potable water to be produced.
study using continuously operated MD process for producing drinking water from seawater made use of a PTFE membrane (0.22μm). In this study, where the feed temperature was 60°C and the permeate temperature was 20°C, the process flux level was reported to decrease from
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce...
procedures, the same membrane was successfully reused in the DCMD system [22]. In another
Although the literature describes the use of various different acid and base derivatives for cleaning membranes, HCl use is common in MD [36, 38]. HCl is a quite effective chemical
synthetic seawater, it was observed that citric acid and NaOH completely remedied for both the flux and the membrane's hydrophobicity [39]. In another study, washing MD membrane
Within the context of the study experiments, the membrane cleaning at the end of day I was performed using only distilled water. On the following days, the cleaning was done using 1% HCl for 20 min, then 1% NaOH for 20 min, and lastly distilled water for 20 min. After flushing, the pH‐conductivity values of the acid and base washing solutions changed from 2.05–51.1 and 12.7–52.9 to 2.04–44.9 and 12.71–46.6 mS/cm, respectively. Based on these results, it can be said that no significant changes were observed in the pH values. By the end of the day IV of operation, the changes in the conductivity of the acid and base washing solutions were 11.9 and 12.1%, respectively. For 1 month of a continuous cleaning at the same conditions per the applied operation cycle, it was calculated that the amounts of acid and base lost will reach to about 89.3 and 91.0%, respectively. With regard to the membrane flushing step that will be applied in a real‐scale facility, it appears a necessary to renew the 1% acid and base solutions
By maintaining seawater TDS value within an interval from 26,400 to a max of 78,900 mg/L throughout the processing (66,800 mg/L at the end of the operation), it was ensured that the DCMD system could be permanently operated with steady‐state concentrating conditions at a constant temperature difference. In the operations, where distilled water is kept within the system for 24 h, conductivity values of the permeate flow at the end of days I, II, III, and IV were always below the threshold level of 500 μS/cm, being 14, 78, 117, and 375 μS/cm, respectively, while the TDS values for the days I, II, III, and IV permeates were 7, 49, 54, and 179 mg/L, respectively (*data were not shown*). Thanks to the distilled water replace at the begin‐ ning of day V, it was noted that these values decreased to 76.0 μS/cm and 47.1 mg TDS/L as adjacent values to the values at the end of the first 2 days. The qualities of concentrated sea‐ water and potable water produced at the end of continuous operation are given in **Table 3**,
According to **Table 3**, the output water obtained was in very good quality in which the reten‐ tions of dissolved organics, boron, silicium, and iron as well as anions and cations were at above 99%. Too low color presence at 525 nm wavelength was observed resulting mostly probably from the trace level amounts of dissolved organics. Based on all results obtained
h following a 1‐month operation. Further, following membrane washing
h was observed following the 70 h opera‐
http://dx.doi.org/10.5772/intechopen.68653
113
[36]. In a study being used
fouling, bringing the flux
23.76 to 14.36 L/m2
values to their original level [38].
at certain time scales.
together with MD rejection performances.
study using PTFE membranes, a flux level of 25 L/m2
tion of the DCMD process with a temperature difference of 60°C [37].
in cleaning of scaling and fouling from basic salts such as CaCO<sup>3</sup>
with HCl resolved the reduction in flux values caused by CaCO<sup>3</sup>
#### **3.2. MD processing of Marmara seawater**
Under controlled membrane cleaning conditions per day cycle where 1% HCl, 1% NaOH and distilled water were consecutively used for flushing the active layer of the membrane while inside the module, and the permeate flow was replaced with distilled water within a suitable time frame, the continuous processing experiments were performed to identify the design flow that would serve as a basis for the field application of the DCMD. In the study, the DCMD was operated in order to concentrate seawater flow to obtain daily concentration performances of 63–66%, and at the end of the 5‐day operation period, a 60% volumetric concentration ratio was reached. To ensure that the continuous operation performance for the periods shown in **Figure 8** remained consistent, the initial volume of the feed flow was completed (i.e. brought to the desired level) every day by adding raw seawater.
At the end of days I, II, III, IV, and V, the permeate water fluxes were determined as 38.9, 31.1, 22.6, 16.7, and 15.5 L/m2 h, respectively. It was found that the flux changes in the first 4‐day period were generally characterized by a consistent decrease, while the flux value on day V was very close to the flux value of day IV (despite the fact that the permeate was changed with distilled water). Based on this observation, it is possible to say that the water flux performance almost reached a constant operation level by the end of the day V.
An MD optimization study using PP membrane operated under conditions of ΔP = 0.355 × 10−5 Pa, Q<sup>f</sup> = 73.6 L/h and Qp = 18.8 L/h determined an MD process flux of 4.192 L/m2 h [24]. A previous
**Figure 8.** MD process performance under continuous operating condition ((a): permeate flux, (b): water recovery ratio).
study using continuously operated MD process for producing drinking water from seawater made use of a PTFE membrane (0.22μm). In this study, where the feed temperature was 60°C and the permeate temperature was 20°C, the process flux level was reported to decrease from 23.76 to 14.36 L/m2 h following a 1‐month operation. Further, following membrane washing procedures, the same membrane was successfully reused in the DCMD system [22]. In another study using PTFE membranes, a flux level of 25 L/m2 h was observed following the 70 h opera‐ tion of the DCMD process with a temperature difference of 60°C [37].
operation in which seawater or feed stream would be heated up to max 60°C. Taking into account the fact that energy costs represent the most important obstacle standing in the way of the rising uses of MD process compared to RO seawater desalination plants, it was com‐ prehended that 30°C and some below or above differences would render further possible to arrive the optimal cost solutions in on‐land MD plants due to lower heat energy consumption
Under controlled membrane cleaning conditions per day cycle where 1% HCl, 1% NaOH and distilled water were consecutively used for flushing the active layer of the membrane while inside the module, and the permeate flow was replaced with distilled water within a suitable time frame, the continuous processing experiments were performed to identify the design flow that would serve as a basis for the field application of the DCMD. In the study, the DCMD was operated in order to concentrate seawater flow to obtain daily concentration performances of 63–66%, and at the end of the 5‐day operation period, a 60% volumetric concentration ratio was reached. To ensure that the continuous operation performance for the periods shown in **Figure 8** remained consistent, the initial volume of the feed flow was
At the end of days I, II, III, IV, and V, the permeate water fluxes were determined as 38.9, 31.1,
period were generally characterized by a consistent decrease, while the flux value on day V was very close to the flux value of day IV (despite the fact that the permeate was changed with distilled water). Based on this observation, it is possible to say that the water flux performance
An MD optimization study using PP membrane operated under conditions of ΔP = 0.355 × 10−5
**Figure 8.** MD process performance under continuous operating condition ((a): permeate flux, (b): water recovery ratio).
h, respectively. It was found that the flux changes in the first 4‐day
h [24]. A previous
completed (i.e. brought to the desired level) every day by adding raw seawater.
= 73.6 L/h and Qp = 18.8 L/h determined an MD process flux of 4.192 L/m2
almost reached a constant operation level by the end of the day V.
per unit of potable water to be produced.
**3.2. MD processing of Marmara seawater**
22.6, 16.7, and 15.5 L/m2
Pa, Q<sup>f</sup>
112 Desalination
Although the literature describes the use of various different acid and base derivatives for cleaning membranes, HCl use is common in MD [36, 38]. HCl is a quite effective chemical in cleaning of scaling and fouling from basic salts such as CaCO<sup>3</sup> [36]. In a study being used synthetic seawater, it was observed that citric acid and NaOH completely remedied for both the flux and the membrane's hydrophobicity [39]. In another study, washing MD membrane with HCl resolved the reduction in flux values caused by CaCO<sup>3</sup> fouling, bringing the flux values to their original level [38].
Within the context of the study experiments, the membrane cleaning at the end of day I was performed using only distilled water. On the following days, the cleaning was done using 1% HCl for 20 min, then 1% NaOH for 20 min, and lastly distilled water for 20 min. After flushing, the pH‐conductivity values of the acid and base washing solutions changed from 2.05–51.1 and 12.7–52.9 to 2.04–44.9 and 12.71–46.6 mS/cm, respectively. Based on these results, it can be said that no significant changes were observed in the pH values. By the end of the day IV of operation, the changes in the conductivity of the acid and base washing solutions were 11.9 and 12.1%, respectively. For 1 month of a continuous cleaning at the same conditions per the applied operation cycle, it was calculated that the amounts of acid and base lost will reach to about 89.3 and 91.0%, respectively. With regard to the membrane flushing step that will be applied in a real‐scale facility, it appears a necessary to renew the 1% acid and base solutions at certain time scales.
By maintaining seawater TDS value within an interval from 26,400 to a max of 78,900 mg/L throughout the processing (66,800 mg/L at the end of the operation), it was ensured that the DCMD system could be permanently operated with steady‐state concentrating conditions at a constant temperature difference. In the operations, where distilled water is kept within the system for 24 h, conductivity values of the permeate flow at the end of days I, II, III, and IV were always below the threshold level of 500 μS/cm, being 14, 78, 117, and 375 μS/cm, respectively, while the TDS values for the days I, II, III, and IV permeates were 7, 49, 54, and 179 mg/L, respectively (*data were not shown*). Thanks to the distilled water replace at the begin‐ ning of day V, it was noted that these values decreased to 76.0 μS/cm and 47.1 mg TDS/L as adjacent values to the values at the end of the first 2 days. The qualities of concentrated sea‐ water and potable water produced at the end of continuous operation are given in **Table 3**, together with MD rejection performances.
According to **Table 3**, the output water obtained was in very good quality in which the reten‐ tions of dissolved organics, boron, silicium, and iron as well as anions and cations were at above 99%. Too low color presence at 525 nm wavelength was observed resulting mostly probably from the trace level amounts of dissolved organics. Based on all results obtained from the continuous operation, it can be suggested for successful field operations of DCMD seawater desalination plants that the distillate stream or product output water should not be remained for a longer operating time than about 10–12 h in the product water storage tank. Certainly, more accurate heuristic knowledge for safe plant operation would be ensured by means of taking into account the performance information to be retrieved from the data of the pilot and/or field scale installations.
**3.3. Health impacts assessments of produced drinking water**
the steadily growing demand for water [40].
respectively [43].
Color
mg.m) and color (1/cm).
Almost 75 million people across the world acquire drinking water from the sea by treating seawater in desalination plants, and it is expected that this number will go up as a result of
**\***Units of all parameters are mg/L except for temperature (°C), electrical conductivity (μS/cm), UVA254 (1/cm), SUVA (L/
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce...
http://dx.doi.org/10.5772/intechopen.68653
115
**Table 3.** Quality of potable water produced by continuous MD operation together with MD rejection performances.
**Parameters***\** **MD concentrated seawater MD permeate (output water)** *R* **(%)**
436 nm 0.003 0.000 100.00 525 nm 0.004 0.001 75.00 620 nm 0.001 0.000 100.00
Seawater abounds with certain ions like calcium, magnesium, sodium, chloride, and iodine but has a lower content of essential ions such as zinc, copper, chromium, and manganese. As a result of seawater desalination applications, the amount of almost all of these ions and minerals, which are essential to the human health and to the agricultural productivity, is dramatically reduced in drinking water [14, 40]. Some essential metals like Cu, Mn, and Zn are required for normal body growth and function [41]. For healthy body, calcium and magnesium are both essential elements, and they play protective roles in the human body development [14, 40, 42]. Calcium, which is a primary constituent of the skeletal structure, is also important for different key physical functions. Due to decreases in bone mass and mineral content, calcium deficiency results in an increased risk of fractures. Deficiency of magnesium, an essential element for physiological processes such as mineralization and skeletal development, cardiac excitability and vascular tone, contractility, reactivity, and growth, can give rise to the pathophysiology of hypertension [40, 43]. Depending on the beneficial health effects rendered by intake of Ca and Mg through drinking water, it has been recommended by World Health Organization (2007) that the minimum and optimum Ca and Mg levels in drinking water should be 20 and 40–80 mg/L and 10 and 20–30 mg/L,
Otherwise, it is also likely that high quantities of these metals will inflict harms for human health. High concentrations of Cd, Cr, Ni, and Pb are regarded to be extremely toxic for humans and aquatic organisms [41]. Furthermore, demineralized water is more aggressive to piping, and therefore, additional risks could be posed through exposure to extracted trace elements such as lead and cadmium. According to certain studies, the use of demineralized water for cooking escalates the loss of essential mineral content of foods, bringing about det‐ rimental effects on health [14, 42]. When the quality parameter values were evaluated in the output water obtained during this study, produced water is low in ions such as Ca, Mg, Na, Cl and essential ions like Zn, Cu, Cr, and Mn. In addition to this, its low solute content of 47.1 mg TDS/L paves the way for corrosive effects to the metal distribution piping and thus the risks
mg.m) and color (1/cm).
from the continuous operation, it can be suggested for successful field operations of DCMD seawater desalination plants that the distillate stream or product output water should not be remained for a longer operating time than about 10–12 h in the product water storage tank. Certainly, more accurate heuristic knowledge for safe plant operation would be ensured by means of taking into account the performance information to be retrieved from the data of the
**Parameters***\** **MD concentrated seawater MD permeate (output water)** *R* **(%)** *T* 24.2 24.3 – pH 8.54 7.30 – *Ec* 108, 200 76.0 99.93 TDS 66, 800 47.1 99.93 TOC 20.5 0.08 99.61 DOC 18.4 0.06 99.67 UVA254 0.160 0.001 99.38 SUVA 0.870 1.667 – Na+ 20, 445 0.114 99.99 K<sup>+</sup> 1143 0.005 99.99
<sup>+</sup> <0.1 0.0 – Ba2+ 0.058 <0.01 – Ca2+ 1097 3.5 99.68 Mg2+ 1770 1.4 99.92 Mn2+ 0.044 0.0 100.00 Sr2+ 15.20 0.02 99.87 B 4.39 0.007 99.84 Si 1.22 0.008 99.34 Fe 0.017 0.0 100.00 Cl<sup>−</sup> 46, 236 12.6 99.97
<sup>−</sup> 244 0.58 99.76
<sup>−</sup> 1.3 0.0 100.00
2 − 112 0.0 100.00
2 − 7050 7.2 99.90 TKN 1.8 0.0 100.00 Total nitrogen 3.1 0.0 100.00 Alkalinity 356 0.58 99.84 Total hardness 16,035 27.1 99.83
pilot and/or field scale installations.
NH<sup>4</sup>
114 Desalination
HCO<sup>3</sup>
NO<sup>3</sup>
CO<sup>3</sup>
SO<sup>4</sup>
**Table 3.** Quality of potable water produced by continuous MD operation together with MD rejection performances.
#### **3.3. Health impacts assessments of produced drinking water**
Almost 75 million people across the world acquire drinking water from the sea by treating seawater in desalination plants, and it is expected that this number will go up as a result of the steadily growing demand for water [40].
Seawater abounds with certain ions like calcium, magnesium, sodium, chloride, and iodine but has a lower content of essential ions such as zinc, copper, chromium, and manganese. As a result of seawater desalination applications, the amount of almost all of these ions and minerals, which are essential to the human health and to the agricultural productivity, is dramatically reduced in drinking water [14, 40]. Some essential metals like Cu, Mn, and Zn are required for normal body growth and function [41]. For healthy body, calcium and magnesium are both essential elements, and they play protective roles in the human body development [14, 40, 42]. Calcium, which is a primary constituent of the skeletal structure, is also important for different key physical functions. Due to decreases in bone mass and mineral content, calcium deficiency results in an increased risk of fractures. Deficiency of magnesium, an essential element for physiological processes such as mineralization and skeletal development, cardiac excitability and vascular tone, contractility, reactivity, and growth, can give rise to the pathophysiology of hypertension [40, 43]. Depending on the beneficial health effects rendered by intake of Ca and Mg through drinking water, it has been recommended by World Health Organization (2007) that the minimum and optimum Ca and Mg levels in drinking water should be 20 and 40–80 mg/L and 10 and 20–30 mg/L, respectively [43].
Otherwise, it is also likely that high quantities of these metals will inflict harms for human health. High concentrations of Cd, Cr, Ni, and Pb are regarded to be extremely toxic for humans and aquatic organisms [41]. Furthermore, demineralized water is more aggressive to piping, and therefore, additional risks could be posed through exposure to extracted trace elements such as lead and cadmium. According to certain studies, the use of demineralized water for cooking escalates the loss of essential mineral content of foods, bringing about det‐ rimental effects on health [14, 42]. When the quality parameter values were evaluated in the output water obtained during this study, produced water is low in ions such as Ca, Mg, Na, Cl and essential ions like Zn, Cu, Cr, and Mn. In addition to this, its low solute content of 47.1 mg TDS/L paves the way for corrosive effects to the metal distribution piping and thus the risks by exposure to extracted trace elements like lead and cadmium into the pipe will increase. The allowable limit values for these ions in the relevant national and international drinking water standards were presented in **Table 4** together with the other drinking water quality parameters specific to the produced water.
As described in the literature, to fulfill the desirable values for these parameters under limit concentrations, one possible solution involves making certain additions to the distilled prod‐ uct water by taking into account hygienic requirements. This will make it possible to supply drinking water that does not pose any problems with regard to public health. It is possible to overcome mineral deficiencies directly through addition of minerals at the MD desalination plant or locally by means of chemical injection at specified locations into water distribution systems by also considering hygienic requirements. In an alternative way, quality standards can be reached by mixing high‐quality and low‐quality water [47], which lead to the sig‐
> nificant savings since remineralization facilities and chemicals are no longer needed [40]. The most affordable option to increase the dissolved content of specific ions in desalinated water is blending remineralized desalinated water with treated brackish groundwater or treated seawater [48]. When the MD output water produced in this study is evaluated in terms of physicochemical water quality for directly human consumption, it was found to be much lower than necessary levels especially in terms of dissolved minerals concentrations, which also give corrosive properties to the demineralized water. Based also on the literature knowl‐ edge and land experiences, a general management approach for field‐scale pragmatic solu‐ tions toward to directly use the MD desalination effluents as drinking water was holistically proposed in **Figure 9**; as being the first, adding into existing urban potable water at desirable ratios; the second, injecting appropriate chemicals into effluent by completely mixing; and the third, mixing effluents with raw/concentrated seawaters (1:250/1:1000 for Marmara seawater)
> **Table 4.** Maximum allowable concentrations of specific water quality parameters in use of produced water as drinking
**Parameters\* Permeate TSE 266 [44] WHO [45] EPA [46]**
TKN 0.0 – – – Total Nitrogen 0.0 – – – Alkalinity 0.58 – – – Total hardness 27.1 – 150–500 mg/L –
525 nm 0.001 – – –
2 − 7.2 250 mg/l 500 mg/l 250 mg/l
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce...
http://dx.doi.org/10.5772/intechopen.68653
117
As the third distinctive solution, the MD product water can be readily mixed with seawater concentrate at a ratio of 1/1000 or with raw seawater at a ratio of 1/250 as shown in **Table 5** after MD desalination of raw seawater. In the calculations, dissolved solutes concentration increased to 180–200 mg/L in the mix ratios of 1/500 and about 1/200 for concentrated and raw seawaters, respectively. But, Ca2+ and Mg2+ concentration levels increased by 1.7 and 3.5 times although they could not fulfill the desired levels. Low level of total hardness, which is reached to ∼60 mg/L, needs to be still increased to the desired standard levels by applying carbonation to the produced water. Hence, the obtained mixtures would also still lack of the desired levels of minerals and ions to be targeted for healthy drinking water production. For this reason, it is deduced that stand‐alone use of this solution for an absolute success would not be sufficient. In addition, the practical ways of the mixing applications can be evaluated as an insignificant improving factor for decreasing of the negative effects of the concentrated seawater discharges that importantly affect the sea life
or clean brackish natural waters under hygienic precautions.
near discharge points.
SO<sup>4</sup>
Color
water supply.
436 nm 0.000
620 nm 0.000
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce... http://dx.doi.org/10.5772/intechopen.68653 117
by exposure to extracted trace elements like lead and cadmium into the pipe will increase. The allowable limit values for these ions in the relevant national and international drinking water standards were presented in **Table 4** together with the other drinking water quality
As described in the literature, to fulfill the desirable values for these parameters under limit concentrations, one possible solution involves making certain additions to the distilled prod‐ uct water by taking into account hygienic requirements. This will make it possible to supply drinking water that does not pose any problems with regard to public health. It is possible to overcome mineral deficiencies directly through addition of minerals at the MD desalination plant or locally by means of chemical injection at specified locations into water distribution systems by also considering hygienic requirements. In an alternative way, quality standards can be reached by mixing high‐quality and low‐quality water [47], which lead to the sig‐
**Parameters\* Permeate TSE 266 [44] WHO [45] EPA [46]** *T* 24.3 – – – pH 7.30 6.5–9.5 6.5–8.5 6.5–8.5 *Ec* 76.0 2500 μS/cm 250 μS/cm –
TDS 47.1 – – 500 mg/l
TOC 0.08 – – – DOC 0.06 – – – UVA254 0.001 – – – SUVA 1.667 – – – Na+ 0.114 200 mg/l 200 mg/l 20 mg/l K<sup>+</sup> 0.005 – – –
<sup>+</sup> 0.0 – – – Ba2+ <0.01 – 0.7 mg/l 2 mg/l Ca2+ 3.5 – – – Mg2+ 1.4 – – –
Mn2+ 0.0 50 μg/l 0.4 mg/l 0.05 mg/l
<sup>−</sup> 0.58 – – –
2 − 0.0 – – –
<sup>−</sup> 0.0 50 mg/l 50 mg/l 10 mg/l
Sr2+ 0.02 – – – B 0.007 1 mg/l 0.5 mg/l – Si 0.008 – – – Fe 0.0 200 μg/l 0.3 mg/l 0.3 mg/l Cl<sup>−</sup> 12.6 250 mg/l 250 mg/l 250 mg/l
parameters specific to the produced water.
NH<sup>4</sup>
116 Desalination
HCO<sup>3</sup>
NO<sup>3</sup>
CO<sup>3</sup>
**Table 4.** Maximum allowable concentrations of specific water quality parameters in use of produced water as drinking water supply.
nificant savings since remineralization facilities and chemicals are no longer needed [40]. The most affordable option to increase the dissolved content of specific ions in desalinated water is blending remineralized desalinated water with treated brackish groundwater or treated seawater [48]. When the MD output water produced in this study is evaluated in terms of physicochemical water quality for directly human consumption, it was found to be much lower than necessary levels especially in terms of dissolved minerals concentrations, which also give corrosive properties to the demineralized water. Based also on the literature knowl‐ edge and land experiences, a general management approach for field‐scale pragmatic solu‐ tions toward to directly use the MD desalination effluents as drinking water was holistically proposed in **Figure 9**; as being the first, adding into existing urban potable water at desirable ratios; the second, injecting appropriate chemicals into effluent by completely mixing; and the third, mixing effluents with raw/concentrated seawaters (1:250/1:1000 for Marmara seawater) or clean brackish natural waters under hygienic precautions.
As the third distinctive solution, the MD product water can be readily mixed with seawater concentrate at a ratio of 1/1000 or with raw seawater at a ratio of 1/250 as shown in **Table 5** after MD desalination of raw seawater. In the calculations, dissolved solutes concentration increased to 180–200 mg/L in the mix ratios of 1/500 and about 1/200 for concentrated and raw seawaters, respectively. But, Ca2+ and Mg2+ concentration levels increased by 1.7 and 3.5 times although they could not fulfill the desired levels. Low level of total hardness, which is reached to ∼60 mg/L, needs to be still increased to the desired standard levels by applying carbonation to the produced water. Hence, the obtained mixtures would also still lack of the desired levels of minerals and ions to be targeted for healthy drinking water production. For this reason, it is deduced that stand‐alone use of this solution for an absolute success would not be sufficient. In addition, the practical ways of the mixing applications can be evaluated as an insignificant improving factor for decreasing of the negative effects of the concentrated seawater discharges that importantly affect the sea life near discharge points.
**Figure 9.** Holistic management approach for satisfying specific water quality requirements in long‐term use of MD effluents demineralized from seawater as drinking water.
**4. Conclusions**
raw seawaters.
SO<sup>4</sup>
Total hardness
Color
membrane trans‐temperature difference of 30°C.
was operated with steady‐state permeate flux of 17.21 L/m2
In this chapter, suitable operating conditions for continuous DCMD processing of seawater were first determined based on water flux and solutes rejection performances. Thereafter, by relative long‐term MD desalination in seawater concentration mode at lab pilot‐scale system, the water qualities of MD product water and seawater concentrate were investigated in light of physi‐ cochemical parameters specific to the seawater characteristics. In the final, it was elaborately examined whether the MD produced water is suitable for direct human use as drinking water. Within the scope of the DCMD experiments under batch conditions, it was determined that the best operating conditions involved the use of hydrophobic PTFE membrane with 0.45 μm pore diameter; a flow velocity for seawater and permeate (distilled water) of 270 and 360 L/h, respectively (corresponding to mean Re numbers of 4320 and 4222); and flow temperatures for seawater and permeate streams of 60 and 30°C, respectively, which are associated with a
**Table 5.** Changing values of output water quality parameters according to the ratios of mixings with concentrated or
**1/100 1/250 1/500 1/1000 1/20 1/50 1/100 1/250**
http://dx.doi.org/10.5772/intechopen.68653
119
**Mixing ratio with Parameters Units Permeate Concentrated seawater Raw seawater**
2 − mg/L 7.2 78 35 21 14 123 53 30 15 TKN mg/L 0.0 0.02 0.01 0.0 0.0 0.0 0.0 0.0 0.0 Alkalinity mg/L 0.58 4 2 1 1 16 7 4 2
436 nm 1/cm 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 525 nm 0.001 0.001 0.001 0.001 0.001 0.0 0.0 0.0 0.001 620 nm 0.000 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001
mg/L 27.1 187 91 59 43 275 126 77 42
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce...
By means of 30 h MD processing of rough‐filtrated raw seawater from the Marmara Sea that
seawater concentrating level reached to 66% in the constant temperature difference, the produc‐ tion of MD output water below threshold level of 500 μS/cm was continually carried out with‐ out replacing intrinsic distillated water of distillate stream. After its replacement was applied, ultimate product water was supplied with 76.0 μS/cm conductivity and 47.1 mg/L dissolved solids. A replacement time of about 10–12 h for initial clean distillate in the output water stor‐ age tank would be sufficient for field‐scale operations of DCMD seawater desalination plants. In case the MD product water is to be used as drinking water, it will be necessary to ensure that the dissolved minerals that are essential for a healthy life are found in the water in at least the minimum recommended levels, and to prevent the various trace elements (Cu, Fe, and Mn)
h and solutes retentions of >99% at
**Table 5.** Changing values of output water quality parameters according to the ratios of mixings with concentrated or raw seawaters.
## **4. Conclusions**
**Mixing ratio with Parameters Units Permeate Concentrated seawater Raw seawater**
effluents demineralized from seawater as drinking water.
HCO<sup>3</sup>
118 Desalination
NO<sup>3</sup>
CO<sup>3</sup>
T °C 24.3 24.3 24.3 24.3 24.3 24.3 24.3 24.3 24.3 pH – 7.30 7.31 7.30 7.30 7.30 7.3 7.3 7.3 7.3 Ec μS/cm 76.0 1157 508 292 184 2092 882 479 222 TDS mg/L 47.1 715 314 181 114 1254 530 289 134 TOC mg/L 0.0 0.2 0.1 0.0 0.0 0.2 0.1 0.0 0.0 DOC mg/L 0.0 0.2 0.1 0.0 0.0 0.2 0.1 0.0 0.0 UVA254 1/cm 0.004 0.01 0.0 0.0 0.0 0.01 0.0 0.0 0.0 SUVA L/mg.m 1.67 1.66 1.67 1.67 1.67 1.64 1.66 1.66 1.33 Na+ mg/L 0.114 205 82 41 21 444 178 89 36 K<sup>+</sup> mg/L 0.01 4 2 1 0.0 22 9 4 2 Ca2+ mg/L 3.5 14 8 6 5 23 11 7 4 Mg2+ mg/L 1.4 19 8 5 3 36 15 8 4 Mn2+ mg/L 0.001 0.001 0.001 0.001 0.001 0.002 0.001 0.001 0.001 Sr2+ mg/L 0.02 0.17 0.08 0.05 0.04 0.29 0.13 0.07 0.04 B mg/L 0.01 0.05 0.02 0.02 0.01 0.21 0.09 0.05 0.02 Si mg/L 0.02 0.03 0.02 0.02 0.02 0.07 0.04 0.03 0.02 Fe mg/L 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Cl<sup>−</sup> mg/L 12.6 475 197 105 59 712 292 152 66
**Figure 9.** Holistic management approach for satisfying specific water quality requirements in long‐term use of MD
<sup>−</sup> mg/L 0.58 3 2 1 1 16 7 4 2
<sup>−</sup> mg/L 0.0 0.01 0.01 0.0 0.0 0.0 0.0 0.0 0.0
2 − mg/L 0.0 1 0.0 0.0 0.0 0.0 0.0 0.0 0.0
**1/100 1/250 1/500 1/1000 1/20 1/50 1/100 1/250**
In this chapter, suitable operating conditions for continuous DCMD processing of seawater were first determined based on water flux and solutes rejection performances. Thereafter, by relative long‐term MD desalination in seawater concentration mode at lab pilot‐scale system, the water qualities of MD product water and seawater concentrate were investigated in light of physi‐ cochemical parameters specific to the seawater characteristics. In the final, it was elaborately examined whether the MD produced water is suitable for direct human use as drinking water.
Within the scope of the DCMD experiments under batch conditions, it was determined that the best operating conditions involved the use of hydrophobic PTFE membrane with 0.45 μm pore diameter; a flow velocity for seawater and permeate (distilled water) of 270 and 360 L/h, respectively (corresponding to mean Re numbers of 4320 and 4222); and flow temperatures for seawater and permeate streams of 60 and 30°C, respectively, which are associated with a membrane trans‐temperature difference of 30°C.
By means of 30 h MD processing of rough‐filtrated raw seawater from the Marmara Sea that was operated with steady‐state permeate flux of 17.21 L/m2 h and solutes retentions of >99% at seawater concentrating level reached to 66% in the constant temperature difference, the produc‐ tion of MD output water below threshold level of 500 μS/cm was continually carried out with‐ out replacing intrinsic distillated water of distillate stream. After its replacement was applied, ultimate product water was supplied with 76.0 μS/cm conductivity and 47.1 mg/L dissolved solids. A replacement time of about 10–12 h for initial clean distillate in the output water stor‐ age tank would be sufficient for field‐scale operations of DCMD seawater desalination plants.
In case the MD product water is to be used as drinking water, it will be necessary to ensure that the dissolved minerals that are essential for a healthy life are found in the water in at least the minimum recommended levels, and to prevent the various trace elements (Cu, Fe, and Mn) that can have toxic effects on living beings above certain levels or cause damage to the process‐ ing lines. Possible management options recommended in the literature to remedy these issues include the direct addition of minerals to final waters, the addition of chemicals to specific loca‐ tions on process lines, and the blending of demineralized water with treated brackish ground‐ water/seawater. However, stand‐alone generalization of each one of these practical solutions would not be made possible for all the challenges to be encountered in all application varieties of seawater desalination plants. In this frame, there need to develop unique approaches ori‐ ented on novel pragmatic solutions toward to the direct use of MD demineralized effluents as drinking water. In this respect, under an integrated conservative approach, a general manage‐ ment framework toward satisfying the specific water quality requirements in long‐term use of MD effluents was proposed. With an aim of fulfilling the deficiencies of minerals and ions to be targeted for healthy drinking water supply, the developed holistic approach is jointly dependent on injections to urban water distribution systems at desirable ratios, mixings with raw/concentrated seawaters (1:250/1:1000 for Marmara seawater) or brackish natural waters under hygienic precautions, and additions in sufficient amounts of appropriate chemicals by completely mixing.
## **Acknowledgements**
Authors thanks the Scientific and Technological Research Council of Turkey, TUBITAK due to the financial support provided by a national project (*no: 111 Y279*) which was entitled as "Techno‐economic analysis of drinking water production from seawater by forward osmosis/ direct contact membrane distillation integrated membrane system".
**Author details**
Greek symbols
Coskun Aydiner<sup>1</sup>
Istanbul, Turkey
Kocaeli, Turkey
, Serif Cakmak<sup>1</sup>
University, Kocaeli, Turkey
University, Istanbul, Turkey
Narci4
\*, Derya Y. Koseoglu Imer2,3, Salim Oncel<sup>1</sup>
, Emin Ender Celebi<sup>1</sup>
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce...
1 Department of Environmental Engineering, Engineering Faculty, Gebze Technical
2 Department of Environmental Engineering, Engineering Faculty, Istanbul Technical
3 National Research Center on Membrane Technologies, Istanbul Technical University,
4 Department of Environmental Engineering, Engineering Faculty, Kocaeli University,
, Tugba Nur Yilmaz<sup>1</sup>
\*Address all correspondence to: aydiner@gtu.edu.tr
PTFE polytetrafluoroethylene
PVDF polyvinylidene fluoride Q<sup>p</sup> flow of permeate side (L/h) Q<sup>f</sup> flow of feed side (L/h)
Rz,c surface roughness of clean membrane Rz,f surface roughness of fouled membrane
SEM scanning electron microscopy SUVA specific ultraviolet absorbance
UVA254 ultraviolet absorbance at 254 nm WHO world health organization
θ<sup>c</sup> contact angle of clean membrane (°) θ<sup>f</sup> contact angle of fouled membrane (°)
TDS total dissolved solids TKN total Kjeldahl nitrogen TOC total organic carbon
PV photovoltaic
R rejection (%) Re Reynolds number RO reverse osmosis
, Esra Can Dogan<sup>4</sup>
and Yasemin Melek Tilki<sup>1</sup>
http://dx.doi.org/10.5772/intechopen.68653
121
, Ali Oguzhan
## **Nomenclature**
## **Author details**
that can have toxic effects on living beings above certain levels or cause damage to the process‐ ing lines. Possible management options recommended in the literature to remedy these issues include the direct addition of minerals to final waters, the addition of chemicals to specific loca‐ tions on process lines, and the blending of demineralized water with treated brackish ground‐ water/seawater. However, stand‐alone generalization of each one of these practical solutions would not be made possible for all the challenges to be encountered in all application varieties of seawater desalination plants. In this frame, there need to develop unique approaches ori‐ ented on novel pragmatic solutions toward to the direct use of MD demineralized effluents as drinking water. In this respect, under an integrated conservative approach, a general manage‐ ment framework toward satisfying the specific water quality requirements in long‐term use of MD effluents was proposed. With an aim of fulfilling the deficiencies of minerals and ions to be targeted for healthy drinking water supply, the developed holistic approach is jointly dependent on injections to urban water distribution systems at desirable ratios, mixings with raw/concentrated seawaters (1:250/1:1000 for Marmara seawater) or brackish natural waters under hygienic precautions, and additions in sufficient amounts of appropriate chemicals by
Authors thanks the Scientific and Technological Research Council of Turkey, TUBITAK due to the financial support provided by a national project (*no: 111 Y279*) which was entitled as "Techno‐economic analysis of drinking water production from seawater by forward osmosis/
direct contact membrane distillation integrated membrane system".
AFM atomic force microscopy Cf concentration in the feed Cp concentration in the permeate DCMD direct contact membrane distillation
DOC dissolved organic carbon
LEP liquid entry pressure MED multi‐effect distillation MD membrane distillation MSF multi‐stage flash distillation
PE polyethylene PP polypropylene
EDX energy dispersive X‐ray spectrometry
completely mixing.
120 Desalination
**Nomenclature**
**Acknowledgements**
Coskun Aydiner<sup>1</sup> \*, Derya Y. Koseoglu Imer2,3, Salim Oncel<sup>1</sup> , Esra Can Dogan<sup>4</sup> , Ali Oguzhan Narci4 , Serif Cakmak<sup>1</sup> , Tugba Nur Yilmaz<sup>1</sup> , Emin Ender Celebi<sup>1</sup> and Yasemin Melek Tilki<sup>1</sup>
\*Address all correspondence to: aydiner@gtu.edu.tr
1 Department of Environmental Engineering, Engineering Faculty, Gebze Technical University, Kocaeli, Turkey
2 Department of Environmental Engineering, Engineering Faculty, Istanbul Technical University, Istanbul, Turkey
3 National Research Center on Membrane Technologies, Istanbul Technical University, Istanbul, Turkey
4 Department of Environmental Engineering, Engineering Faculty, Kocaeli University, Kocaeli, Turkey
## **References**
[1] UN Water Scarcity Report [Internet]. 2016. Available from: http://www.un.org/waterfor‐ lifedecade/scarcity.shtml [Accessed: 25 March 2016]
[12] Lattemann S, Höpner T. Environmental impact and impact assessment of seawater
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce...
http://dx.doi.org/10.5772/intechopen.68653
123
[13] Al‐Agha MR, Mortaja RSh. Desalination in the Gaza Strip: Drinking water supply and environmental impact. Desalination. 2015;**173**:157‐171. DOI: 10.1016/j.desal.2004.06.212
[14] Cotruvo JA. Desalination guidelines development for drinking water: Background. WHO [Internet]. 2004. Available from: http://www.who.int/water\_sanitation\_health/
[15] Sadhwania JJ, Vezaa JM, Santana C. Case studies on environmental impact of seawater
[16] Liu T, Sheu H, Tseng C. Environmental impact assessment of seawater desalination plant under the framework of integrated coastal management. Desalination. 2013;**326**:10‐18.
[17] Manawi YM, Khraisheh M, Fard AK, Benyahia F, Adham S. Effect of operational param‐ eters on distillate flux in direct contact membrane distillation (DCMD): Comparison between experimental and model predicted performance. Desalination. 2014;**336**:110‐
[18] Choi YH, Kweon JH, Kim DI, Lee S. Evaluation of various pretreatment for particle and inorganic fouling control on performance of SWRO. Desalination. 2009;**247**:137‐147.
[19] Francis L, Ghaffour N, Alsaadi AS, Nunes SP, Amy GL. Performance evaluation of the DCMD desalination process under bench scale and large scale module operating condi‐ tions. Journal of Membrane Science. 2014;**455**:103‐112. DOI: 10.1016/j.memsci.2013.12.033
[20] Naidu G, Jeong S, Choi Y, Jang E, Hwang T, Vigneswaran S. Application of vacuum mem‐ brane distillation for small scale drinking water production. Desalination. 2014;**354**:53‐
[21] Al–Obaidania S, Curcio E, Macedonio F, Profio G.D, Al‐Hinai H, Drioli E. Potential of membrane distillation in seawater desalination: Thermal efficiency, sensitivity study and cost estimation. Journal of Membrane Science. 2008;**323**:85‐98. DOI: 10.1016/j.
[22] He K, Hwang HJ, Woo MW, Moon IS. Production of drinking water from saline water by direct contact membrane distillation (DCMD). Journal of Industrial and Engineering
[23] Winter D, Koschikowski J, Wieghaus M. Desalination using membrane distillation: Experimental studies on full scale spiral wound modules. Journal of Membrane Science.
[24] Boubakri A, Hafiane A, Bouguecha SAT. Direct contact membrane distillation: Capability to desalt raw water. Arabian Journal of Chemistry. DOI: http://dx.doi.org/10.1016/j.
Chemistry. 2011;**17**:41‐48. DOI: 10.1016/j.jiec.2010.10.007
2011;**375**:104‐112. DOI: 10.1016/j.memsci.2011.03.030
desalination. Desalination. 2008;**220**:1‐15. DOI: 10.1016/j.desal.2007.03.009
desalination. Desalination. 2015;**185**: 1‐8. DOI: 10.1016/j.desal.2005.02.072
dwq/nutdesalination.pdf [Accessed: 29 May 2016]
DOI: 10.1016/j.desal.2013.07.003
120. DOI: 10.1016/j.desal.2014.01.003
DOI: 10.1016/j.desal.2008.12.019
61. DOI: 10.1016/j.desal.2014.09.026
memsci.2008.06.006
arabjc.2014.02.010 (in press)
[12] Lattemann S, Höpner T. Environmental impact and impact assessment of seawater desalination. Desalination. 2008;**220**:1‐15. DOI: 10.1016/j.desal.2007.03.009
**References**
122 Desalination
renene.2007.03.019
jhazmat.2011.06.056
desal.2010.08.035
desal.2014.05.032
DOI: 10.1016/j.watres.2009.03.010
[1] UN Water Scarcity Report [Internet]. 2016. Available from: http://www.un.org/waterfor‐
[2] Lorain O, Hersant B, Persin P, Grasmick A, Brunard N, Espenan J.M. Ultrafiltration membrane pre‐treatment benefits for reverse osmosis process in seawater desalt‐ ing. Quantification in terms of capital investment cost and operating cost reduction.
[3] Gilau AM, Small MJ. Designing cost‐effective seawater reverse osmosis system under optimal energy options. Renewable Energy. 2008;**33**:617‐630. DOI: 10.1016/j.
[4] Grenlee LF, Lawler DF, Freeman BD, Marrot B, Moulin P. Reverse osmosis desalination: Water sources, technology, and today's challenges. Water Research. 2009;**43**:2317‐2348.
[5] Yarlagadda S, Gude VG, Camacho LM, Pinappu S, Deng S. Potable water recovery from As, U, and F contaminated ground waters by direct contact membrane distil‐ lation process. Journal of Hazardous Materials. 2011;**192**:1388‐1394. DOI: 10.1016/j.
[6] Melián‐Martel N, Sadhwani JJ, Malamis S, Ochsenkühn‐Petropoulou M. Structural and chemical characterization of long‐term reverse osmosis membrane fouling in a full scale
[7] Jansen AE, Assink JW, Hanemaaijer JH, Medevoort JV, Sonsbeek EV. Development and pilot testing of full‐scale membrane distillation modules for deployment of waste heat.
[8] Frioui S, Oumeddour R. Investment and production costs of desalination plants by semi‐ empirical method. Desalination. 2008;**223**:457‐463. DOI: 10.1016/j.desal.2007.01.180
[9] Matsuura T, Rana D, Qutaishat MR, Singh G. Recent advances in membrane science and technology in seawater desalination—With technology development in the middle east and Singapore. In: UNESCO‐EOLLS Joint Committee, editors. Water and Wastewater Treatment Technologies. Encyclopedia of Life Support Systems (EOLLS), developed under the Auspices of the UNESCO [Internet]. Oxford, UK: EOLLS Publishers; 2011.
[10] Mezher T, Fath H, Abbas Z, Khaled A. Techno‐economic assessment and environmen‐ tal impacts of desalination technologies. Desalination. 2011;**266**:263‐273. DOI: 10.1016/j.
[11] Fuentes‐Bargues JL. Analysis of the process of environmental impact assessment for seawater desalination plants in Spain. Desalination. 2014;**347**:166‐174. DOI: 10.1016/j.
desalination plant. Desalination. 2012;**305**:44‐53. DOI: 10.1016/j.desal.2012.08.011
Desalination. 2013;**323**:55‐65. DOI: 10.1016/j.desal.2012.11.030
Available from: http://www.eolls.net. [Accessed: 12 April 2016]
lifedecade/scarcity.shtml [Accessed: 25 March 2016]
Desalination. 2007; **203**:277‐285. DOI: 10.1016/j.desal.2006.02.022
[25] Guillen‐Burrieza E, Thomas R, Mansoor B, Johnson D, Hilal N, Arafat H. Effect of dry out on the fouling of PVDF and PTFE membranes under conditions simulating intermittent seawater membrane distillation (SWMD). Journal of Membrane Science. 2013;**438**:126‐ 139. DOI: 10.1016/j.memsci.2013.03.014
[37] Shirazi MMA, Kargari A, Tabatabaei M. Evaluation of commercial PTFE membranes in desalination by direct contact membrane distillation. Chemical Engineering and
Marmara Seawater Desalination by Membrane Distillation: Direct Consumption Assessment of Produce...
http://dx.doi.org/10.5772/intechopen.68653
125
[38] Gryta M. Polyphosphates used for membrane scaling inhibition during water desali‐ nation by membrane distillation. Desalination. 2012;**285**:170‐176. DOI: 10.1016/j.
[39] Curcio E, Ji X, Profio GD, Sulaiman AO, Fontananova E, Drioli E. Membrane distilla‐ tion operated at high seawater concentration factors: Role of the membrane on CaCO<sup>3</sup> scaling in presence of humic acid. Journal of Membrane Science. 2010;**346**:263‐269. DOI:
[40] Avni N, Eben‐Chaime M, Oron G. Optimizing desalinated sea water blending with other sources to meet magnesium requirements for potable and irrigation waters. Water
[41] Khan K, Lu Y, Khan H, Zakir S, Ihsanullah, Khan S, Khan A.A, Wei L, Wang T. Health risks associated with heavy metals in the drinking water of Swat, northern Pakistan. Journal of Environmental Sciences. 2013;**25**(10):2003‐2013. DOI: 10.1016/S1001‐0742(12)60275‐
[42] Verma KC, Kushwaha AS. Contemporary issue demineralization of drinking water: Is it prudent?. Medical Journal Armed Forces India. 2014;**70**:377‐379. DOI: 10.1016/j.
[43] Abtahi M, Yaghmaeian K, Mohebbi MR, Koulivand A, Rafiee M, Jahangiri‐rad M, Jorfi S, Saeedi R, Oktaie S. An innovative drinking water nutritional quality index (DWNQI) for assessing drinking water contribution to intakes of dietary elements: A national and sub‐national study in Iran. Ecological Indicators. 2016;**60**:367‐376. DOI: 10.1016/j.
[44] MEU (Republic of Turkey Ministry of Environment and Urbanisation). Legislation on waters to be used for human consumption. Official Journal date and number: 17 Feb
[45] WHO. Guidelines for Drinking Water Quality Third Edition Incorporating the First and Second Addenda. Vol. 1—Recommendations WHO Guidelines for the Safe Use of
[46] USEPA. Guidelines for Water Reuse. EPA 822‐S‐12‐001. Washington, DC: U.S. Edition of
[47] Yang Sh‐L, Sun Y‐H, Yeh W‐G. Optimization of regional water distribution system with blending requirements. Journal of Water Resources Planning and Management.
[48] Withers A. Options for recarbonation, remineralisation and disinfection for desalination
Wastewater. Geneva: World Health Organization; 2008. ISBN 978 92 4 154761 1
the Drinking Water Standards and Health Advisories; 2012
2000;**126**(4):229‐235. DOI: 10.1061/(ASCE)0733‐9496(2000)126:4(229)
plants. Desalination. 2005;**179**:11‐24. DOI: 10.1016/j.desal.2004.11.051
Processing. 2014;**76**:16‐25. DOI: 10.1016/j.cep.2013.11.010
Research. 2013;**47**:2164‐2176. DOI: 10.1016/j.watres.2013.01.018
desal.2011.09.051
60277
mjafi.2013.11.011
ecolind.2015.07.004
2005—25730. Ankara, Turkey; 2005
10.1016/j.memsci.2009.09.044
[37] Shirazi MMA, Kargari A, Tabatabaei M. Evaluation of commercial PTFE membranes in desalination by direct contact membrane distillation. Chemical Engineering and Processing. 2014;**76**:16‐25. DOI: 10.1016/j.cep.2013.11.010
[25] Guillen‐Burrieza E, Thomas R, Mansoor B, Johnson D, Hilal N, Arafat H. Effect of dry out on the fouling of PVDF and PTFE membranes under conditions simulating intermittent seawater membrane distillation (SWMD). Journal of Membrane Science. 2013;**438**:126‐
[26] Duong HC, Cooper P, Nelemans B, Cath TY, Nghiem LD. Optimising thermal efficiency of direct contact membrane distillation by brine recycling for small‐scale seawater desal‐
[27] Feng C, Khulbe KC, Matsuura T, Gopal R, Kaur S, Ramakrishna S, Khayet M. Production of drinking water from saline water by air‐gap membrane distillation using polyvinyli‐ dene fluoride nanofiber membrane. Journal of Membrane Science. 2008;**311**:1‐6. DOI:
[28] Fard AK, Manawi YM, Rhadfi T, Mahmoud KA, Khraisheh M, Benyahia F. Synoptic analysis of direct contact membrane distillation performance in Qatar: A case study.
[29] Bouguecha ST, Aly SE, Al‐Beirutty MH, Hamdi MM, Boubakri A. Solar driven DCMD: Performance evaluation and thermal energy efficiency. Chemical Engineering Research
[30] Kaya Y, Aydiner C, Barlas H, Keskinler B. Nanofiltration of single and mixture solu‐ tions containing anionics and nonionic surfactants below their critical micelle con‐ centrations (CMCs). Journal of Membrane Science. 2006;**282**:401‐412. DOI: 10.1016/j.
[31] Aydiner C. A novel approach based on distinction of actual and pseudo resistances in membrane fouling: "Pseudo resistance" concept and its implementation in nanofiltra‐ tion of single solutions. Journal of Membrane Science. 2010;**361**:96‐112. DOI: 10.1016/j.
[32] APHA, AWWA, WEF. Standard Methods for the Examination of Water and Wastewater. In: Eaton AD, Clesceri LS, Rice EW, Greenberg AE, editors. 21st ed. Centennial Edition,
[33] Dogan EC, Yasar A, Sen U, Aydiner C. Water recovery from treated urban wastewa‐ ter by ultrafiltration and reverse osmosis for landscape irrigation. Urban Water Journal.
[34] Yang X, Wang R, Shi L, Fane A.G, Debowski M. Performance improvement of PVDF hollow fiber‐based membrane distillation process. Journal of Membrane Science.
[35] Hsu ST, Cheng KT, Chiou JS. Seawater desalination by direct contact membrane distilla‐
[36] Warsinger DM, Swaminathan J, Guillen‐Burrieza E, Arafat HA, Lienhard VJH. Scaling and fouling in membrane distillation for desalination applications: A review.
tion. Desalination. 2002;**143**:279‐287. DOI: 10.1016/S0011‐9164(02)00266‐7
Desalination. 2015;**356**:294‐313. DOI: 10.1016/j.desal.2014.06.031
Washington DC: American Public Health Association Publication; 2005
2016;**13**:553‐568. DOI: 10.1080/1573062X.2014.992917
2011;**369**:437‐447. DOI: 10.1016/j.memsci.2010.12.020
ination. Desalination. 2015;**374**:1‐9. DOI: 10.1016/j.desal.2015.07.009
Desalination. 2015;**360**:97‐107. DOI: 10.1016/j.desal.2015.01.016
and Design. 2015;**100**:331‐340. DOI: 10.1016/j.cherd.2015.05.044
139. DOI: 10.1016/j.memsci.2013.03.014
124 Desalination
10.1016/j.memsci.2007.12.026
memsci.2006.05.047
memsci.2010.06.003
**Chapter 7**
**On the Purification of Agro-Industrial Wastewater by**
The olive oil production is one of the main industrial activities in the Mediterranean Basin: Italy, Portugal, Greece, and Northern African countries—Syria, Algeria, Turkey, Morocco, Tunisia, Libya, Lebanon, and Egypt. Also, France, Serbia and Montenegro, Macedonia, Cyprus, Turkey, Israel, and Jordan produce a considerable annual yield. Moreover, it is an emergent agro-food industry in China, the USA, Australia, the Middle East, and China, which is expected to develop a considerable production potential. Hence, the treatment of olive mill effluents is a task of global concern. In this context, advanced separation technologies comprising membranes and adsorption resins have been a breakthrough in terms of advanced separation and purification technologies, but many aspects are still in development or under investigation. In this chapter, a focus on the use of membrane and ion adsorption technologies for the purification of these wastewaters will be given. The effect of different factors comprising the type of membrane, i.e., ultrafiltration, nanofiltration, and reverse osmosis; the type of adsorbent (waste material, resins); and the operating conditions will be addressed. Conventional treatments are not able to abate the high concentration of dissolved species present in these effluents. The
use of these technologies can be a feasible solution if properly engineered.
**Keywords:** olive mill effluents, membranes, adsorption, ion exchange, wastewater
© 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
**Membrane Technologies: The Case of Olive Mill**
**Effluents**
Javier Miguel Ochando-Pulido and
http://dx.doi.org/10.5772/intechopen.68401
Additional information is available at the end of the chapter
Antonio Martinez-Ferez
**Abstract**
treatment
## **On the Purification of Agro-Industrial Wastewater by Membrane Technologies: The Case of Olive Mill Effluents**
Javier Miguel Ochando-Pulido and Antonio Martinez-Ferez
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.68401
#### **Abstract**
The olive oil production is one of the main industrial activities in the Mediterranean Basin: Italy, Portugal, Greece, and Northern African countries—Syria, Algeria, Turkey, Morocco, Tunisia, Libya, Lebanon, and Egypt. Also, France, Serbia and Montenegro, Macedonia, Cyprus, Turkey, Israel, and Jordan produce a considerable annual yield. Moreover, it is an emergent agro-food industry in China, the USA, Australia, the Middle East, and China, which is expected to develop a considerable production potential. Hence, the treatment of olive mill effluents is a task of global concern. In this context, advanced separation technologies comprising membranes and adsorption resins have been a breakthrough in terms of advanced separation and purification technologies, but many aspects are still in development or under investigation. In this chapter, a focus on the use of membrane and ion adsorption technologies for the purification of these wastewaters will be given. The effect of different factors comprising the type of membrane, i.e., ultrafiltration, nanofiltration, and reverse osmosis; the type of adsorbent (waste material, resins); and the operating conditions will be addressed. Conventional treatments are not able to abate the high concentration of dissolved species present in these effluents. The use of these technologies can be a feasible solution if properly engineered.
**Keywords:** olive mill effluents, membranes, adsorption, ion exchange, wastewater treatment
© 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
## **1. Introduction**
The scarcity of water particularly concerns agricultural irrigation, which demands more than 70% of the total water consumption worldwide. In this scenario, however, there is a big potential to use regenerated wastewater for irrigation purposes. Reuse of used regenerated wastewater could be a very positive solution regarding environmental and economic impacts.
Jordan, as well as the USA, the Middle East, and China, with important production capacities
On the Purification of Agro-Industrial Wastewater by Membrane Technologies: The Case of Olive Mill Effluents
http://dx.doi.org/10.5772/intechopen.68401
129
The treatment of the olive mill effluents includes wastewater from the washing of the olives (OWW), as well as olive mill wastewater (OMW-3, for three phases of mills), wastewater from olive oil washing (OMW-2), as well as from other activities in the facility, including cleaning and sanitation. OWW includes high concentration of suspended solids (mainly, peel, pulp, ground, branches, and leaves debris) dragged during the olive fruit washing process but low concentration of dissolved organic matter—depending on the water flow exchange rate in the washing machines during the fruit cleaning procedure usually lower than the standardized limits to discharge the effluent on superficial land
OMW is characterized by strong odor nuisance, acid pH, intensive violet-dark color, and high saline toxicity, exhibiting considerable electroconductivity (EC) values [16]. Uncontrolled disposal of these effluents represents an environmental hazard, causing soil contamination, underground leakage, and water body pollution. Due to this presence of high COD load including refractory compounds, as well as fats and lipids, direct discharge of these wastewaters to the municipal sewage treatment plants is forbidden. Legal limits are set to prevent difficulties to keep the biomasses alive on which the municipal sewer
Discharge of untreated OMW to the ground fields and superficial waters bodies is currently prohibited in Spain, whereas in Italy as well as in other European countries, only partial discharge on suitable terrains is allowed; otherwise, in Portugal OMW can be stored and used for irrigation of arbustive cultures under control manner. Straight discharge of OMW has been reported to cause strong odor nuisance, soil contamination, plants growth inhibition, underground leaks, water body pollution, and hindrance of self-purification processes, as
The two-phase system appears to be more ecological, thus has been strongly promoted in Spain, and is now being implemented in Portugal and Greece. Nevertheless, the three-phase system is still surviving in other countries where scarcity of financial support has not permitted the technological switch. Considering that in the two-phase extraction water injection is only practiced in the final vertical centrifugation step, the volume of liquid effluent derived from the decanting process (OMW-2) is reduced by one third on average if compared to the amount required for the three-phase system. Moreover, much of the organic matter remains in the solid waste, which contains more humidity than the pomace from the three-phase system (60–70% in two-phase systems versus 30–45% in three-phase ones, OMW-3) and hence
In this chapter, a focus on the use of membrane and adsorption technologies for the purification of OMW will be given. The effect of different factors comprising the type of membrane, that is, microfiltration (MF), ultrafiltration (UF), nanofiltration (NF), and reverse osmosis (RO), the type of adsorbent (waste material, resins), and the operating conditions will be addressed. The problem of membrane fouling for these systems will also be covered.
well as severe impacts to the aquatic fauna and to the ecological status [16–20].
OMW-2 exhibits lower pollutants degree, too (**Table 1**).
nowadays.
suitable for the purpose.
treatment plants rely.
In the last decades, new advanced separation technologies, less intensive in terms of specific energy consumption than conventional separation procedures and "greener" regarding the minor use of chemicals and reagents to achieve the desired separation, have been more and more implemented. Concretely, membrane technology, adsorption, and ion-exchange (IE) processes can take the lead for these purposes.
Adsorption technology implies relatively low operation costs and long service life because adsorption materials can be regenerated and reused or raw materials can be used. Indeed, adsorption is attractive for its simplicity of design, avoidance of toxic solvents, and minimization of the transformation of the target substances [1]. On another hand, ion exchange is considered a green and simple technology, capable of achieving high removal efficiencies and targeted selectivity, with application for species from wastewaters [2].
Otherwise, membrane technology is modular, easy to design, thus easily scalable to the industry, requires low maintenance, and is environmentally friendly, providing high purifying capacity and selectivity [3–10]. There has been a significant boost in the use of membranes for a plethora of applications and particularly in the field of water and wastewater treatments in the last years. This impulse has been a result of the new membrane materials, modules designs, and optimization of the operating conditions, in specific those for the minimization of fouling issues.
In fact, membrane processes are becoming increasingly used in purifying processes for water and groundwater, to replace classic separation processes, as well as for the reclamation of effluents of different origins, especially those by-produced in agro-industries [6–15].
Among agro-industrial effluents, olive mill wastewater (OMW), generated during the production of olive oil in factories called "mills," is one of the most hardly polluted, depending on the procedure used, reaching chemical oxygen demand (COD) which values up to 100,000 mg/L. The volumes of these effluents have increased markedly in the last decades. Currently, average-sized modern olive oil mills generate several tenths of cubic meters of OMW daily, which raises several millions of cubic meters a year. This can be explained by two linked facts: the increase in the demand of olive oil worldwide due to its health-promoting features (nutritional, antioxidant, anti-inflammatory, cosmetic) and the adoption of continuous production procedures as a response to cope with that demand. Now, the affected countries are not only those in the Mediterranean region, where these industries are ancestral and represent an important sector of the industrial economy, i.e., Spain, Italy, Portugal, Greece, and Northern African countries—Syria, Algeria, Turkey, Morocco, Tunisia, Libya, Lebanon, and Egypt, but also France, Serbia and Montenegro, Macedonia, Cyprus, Turkey, Israel, and Jordan, as well as the USA, the Middle East, and China, with important production capacities nowadays.
**1. Introduction**
processes can take the lead for these purposes.
impacts.
128 Desalination
of fouling issues.
The scarcity of water particularly concerns agricultural irrigation, which demands more than 70% of the total water consumption worldwide. In this scenario, however, there is a big potential to use regenerated wastewater for irrigation purposes. Reuse of used regenerated wastewater could be a very positive solution regarding environmental and economic
In the last decades, new advanced separation technologies, less intensive in terms of specific energy consumption than conventional separation procedures and "greener" regarding the minor use of chemicals and reagents to achieve the desired separation, have been more and more implemented. Concretely, membrane technology, adsorption, and ion-exchange (IE)
Adsorption technology implies relatively low operation costs and long service life because adsorption materials can be regenerated and reused or raw materials can be used. Indeed, adsorption is attractive for its simplicity of design, avoidance of toxic solvents, and minimization of the transformation of the target substances [1]. On another hand, ion exchange is considered a green and simple technology, capable of achieving high removal efficiencies and
Otherwise, membrane technology is modular, easy to design, thus easily scalable to the industry, requires low maintenance, and is environmentally friendly, providing high purifying capacity and selectivity [3–10]. There has been a significant boost in the use of membranes for a plethora of applications and particularly in the field of water and wastewater treatments in the last years. This impulse has been a result of the new membrane materials, modules designs, and optimization of the operating conditions, in specific those for the minimization
In fact, membrane processes are becoming increasingly used in purifying processes for water and groundwater, to replace classic separation processes, as well as for the reclamation of
Among agro-industrial effluents, olive mill wastewater (OMW), generated during the production of olive oil in factories called "mills," is one of the most hardly polluted, depending on the procedure used, reaching chemical oxygen demand (COD) which values up to 100,000 mg/L. The volumes of these effluents have increased markedly in the last decades. Currently, average-sized modern olive oil mills generate several tenths of cubic meters of OMW daily, which raises several millions of cubic meters a year. This can be explained by two linked facts: the increase in the demand of olive oil worldwide due to its health-promoting features (nutritional, antioxidant, anti-inflammatory, cosmetic) and the adoption of continuous production procedures as a response to cope with that demand. Now, the affected countries are not only those in the Mediterranean region, where these industries are ancestral and represent an important sector of the industrial economy, i.e., Spain, Italy, Portugal, Greece, and Northern African countries—Syria, Algeria, Turkey, Morocco, Tunisia, Libya, Lebanon, and Egypt, but also France, Serbia and Montenegro, Macedonia, Cyprus, Turkey, Israel, and
effluents of different origins, especially those by-produced in agro-industries [6–15].
targeted selectivity, with application for species from wastewaters [2].
The treatment of the olive mill effluents includes wastewater from the washing of the olives (OWW), as well as olive mill wastewater (OMW-3, for three phases of mills), wastewater from olive oil washing (OMW-2), as well as from other activities in the facility, including cleaning and sanitation. OWW includes high concentration of suspended solids (mainly, peel, pulp, ground, branches, and leaves debris) dragged during the olive fruit washing process but low concentration of dissolved organic matter—depending on the water flow exchange rate in the washing machines during the fruit cleaning procedure usually lower than the standardized limits to discharge the effluent on superficial land suitable for the purpose.
OMW is characterized by strong odor nuisance, acid pH, intensive violet-dark color, and high saline toxicity, exhibiting considerable electroconductivity (EC) values [16]. Uncontrolled disposal of these effluents represents an environmental hazard, causing soil contamination, underground leakage, and water body pollution. Due to this presence of high COD load including refractory compounds, as well as fats and lipids, direct discharge of these wastewaters to the municipal sewage treatment plants is forbidden. Legal limits are set to prevent difficulties to keep the biomasses alive on which the municipal sewer treatment plants rely.
Discharge of untreated OMW to the ground fields and superficial waters bodies is currently prohibited in Spain, whereas in Italy as well as in other European countries, only partial discharge on suitable terrains is allowed; otherwise, in Portugal OMW can be stored and used for irrigation of arbustive cultures under control manner. Straight discharge of OMW has been reported to cause strong odor nuisance, soil contamination, plants growth inhibition, underground leaks, water body pollution, and hindrance of self-purification processes, as well as severe impacts to the aquatic fauna and to the ecological status [16–20].
The two-phase system appears to be more ecological, thus has been strongly promoted in Spain, and is now being implemented in Portugal and Greece. Nevertheless, the three-phase system is still surviving in other countries where scarcity of financial support has not permitted the technological switch. Considering that in the two-phase extraction water injection is only practiced in the final vertical centrifugation step, the volume of liquid effluent derived from the decanting process (OMW-2) is reduced by one third on average if compared to the amount required for the three-phase system. Moreover, much of the organic matter remains in the solid waste, which contains more humidity than the pomace from the three-phase system (60–70% in two-phase systems versus 30–45% in three-phase ones, OMW-3) and hence OMW-2 exhibits lower pollutants degree, too (**Table 1**).
In this chapter, a focus on the use of membrane and adsorption technologies for the purification of OMW will be given. The effect of different factors comprising the type of membrane, that is, microfiltration (MF), ultrafiltration (UF), nanofiltration (NF), and reverse osmosis (RO), the type of adsorbent (waste material, resins), and the operating conditions will be addressed. The problem of membrane fouling for these systems will also be covered.
any pretreatment [5–12]. Therefore, adequate and optimally designed pretreatment processes on each particular feedstock, in other words, pretreatment tailoring of membrane processes, must
On the Purification of Agro-Industrial Wastewater by Membrane Technologies: The Case of Olive Mill Effluents
http://dx.doi.org/10.5772/intechopen.68401
131
Canepa et al. [21] conducted a study on OMW-3 (average COD = 90 g/L) treatment based on a combined UF membrane followed by adsorption and RO process at pilot scale. The UF membrane was made of polysulfone, the permeate was treated with adsorbing polymers, and the resins eluate with RO polypiperazine-amide membranes. A rapid decline of the UF permeate flux was observed by the authors, and a daily washing was necessary. The fact that there was a flux decline of 80% after 20 h of UF and 70% after 15 h for the RO membrane was a capital handicap for the feasibility of the proposed process. UF was also examined by Borsani and Ferrando [22]. They proposed a pretreatment consisting of oil removal and suspended solid settling, achieving a final 50% organic matter removal with respect to the raw OMW-3 (COD = 70 g/L). Turano et al. [23] proposed centrifugation as pretreatment, simple, and mechanical operation which does not need addition of chemicals and is available in the production line, before UF—polysulfone, molecular weight cutoff (MWCO) 17 kDa—for the treatment of OMW-3. The centrifugation pretreatment removed 80% of the suspended solid concentration, and 90% COD reduction was achieved at the outlet of the integrated process. Paraskeva et al. [24] examined the treatment/fractionation of OMW-3 by a combination of polypropylene filter screening (80 μm), UF (zirconia, mean pore size 100 nm) followed by NF (polymeric, 200 Da) and RO. In particular, UF removed suspended solids and solid fat/lipid
Stoller and Chianese [14] studied the purification of olive washing wastewater (OWW) to comply with municipal sewer discharge standards in Italy. In contrast with OMW, this effluent presents moderate organic pollutant load but high concentration of suspended solids. The authors proposed a coagulation-flocculation pretreatment with polyelectrolytes (aluminum sulfate (AS) or aluminum hydroxide (AH)) before UF and NF in series (composite thin-film spiral-wound membranes). Both pretreatment processes yielded similar COD and biological
subsequent membrane-in-series process after flocculation with AS. The same procedure was applied by these researchers to OMW-3 [8, 10]. In this case, they also compared the results with two additional pretreatments, heterogeneous photocatalysis with titania nanocatalysts and aerobic digestion. They noted that major pollutant removal by the pretreatment, e.g., COD, is not sufficient to state the suitability of the adopted pretreatment for the downstream membrane operation, but the fact that the pollutant particle size is shifted away from the size
On another hand, Akdemir and Ozer [25] applied pH adjustment (acidic or alkaline) and cartridge filtration (20 μm) as pretreatment before UF (MWCO 30–100 kDa) and reported the best performance for the 100 kDa UF membrane, pointing for 1 bar as the critical pressure to reduce fouling issues. Also, Coskun et al. [26] examined UF as part of an integral membrane process for OMW-3 reclamation, but did not consider the entity of fouling in the performance
of the membrane pores is relevant to enhance steady-state flux.
) rejection efficiencies, but higher productivity was observed in the
be developed in order to maximize productivity and minimize fouling.
**2.1. Microfiltration and ultrafiltration membranes**
components (90%).
oxygen demand (BOD<sup>5</sup>
OWW: olive washing wastewater; OMW-P, OMW-3 and OMW-2: olive mill wastewater from batch press process as well as three-phase and two-phase continuous extraction procedures, respectively; OMET-2: mixture of all effluents produced in the olive mill, including OWW, OMW, and from other activities in the facility (e.g., cleaning and sanitation); COD: chemical oxygen demand; BOD<sup>5</sup> : biological oxygen demand; EC: conductivity.
**Table 1.** Average physicochemical composition of the different types of olive mill effluents [28].
Conventional physicochemical treatments are not effective to eliminate the significant salinity of these wastewaters, reflected in their high electric conductivity values. These treatments are not able to abate the high concentration of dissolved species present in these effluents, which present hazardous salinity levels per guideline established by the FAO for irrigation purposes or discharge into public waterways. The use of membrane and adsorption technologies can be a feasible solution if properly engineered.
## **2. Membrane processes for olive mill effluent purification**
Membrane technology, in particular pressure-driven membrane processes comprising MF, UF, NF, and RO, can offer a series of advantages in contrast with other separation processes, making them very promising and environmentally friendly for the purification and remediation of OMW: needless of chemicals and reagents, e.g., solvents—to accomplish the desired separation and concentration; minor costs (capex and opex) and specific energetic requirements than most conventional separation alternatives, upon significant purification potentiality, selectivity, and recovery factors; and also easy scaling to the industry, low space necessities as they are modular, design simplicity, easy operation, and low maintenance needs [5–10].
Fouling mechanisms are very important to fully understand what is taking place between the membrane and the effluent, in view of the adoption and implementation of adequate decisions for the successful design of the membrane plant. This comprises the setup of specifically tailored pretreatment process and optimized operating conditions. Irreversible fouling arises quickly on the membranes due to the high concentration of pollutants when wastewater is purified without any pretreatment [5–12]. Therefore, adequate and optimally designed pretreatment processes on each particular feedstock, in other words, pretreatment tailoring of membrane processes, must be developed in order to maximize productivity and minimize fouling.
#### **2.1. Microfiltration and ultrafiltration membranes**
Conventional physicochemical treatments are not effective to eliminate the significant salinity of these wastewaters, reflected in their high electric conductivity values. These treatments are not able to abate the high concentration of dissolved species present in these effluents, which present hazardous salinity levels per guideline established by the FAO for irrigation purposes or discharge into public waterways. The use of membrane and adsorption technologies can be
OWW: olive washing wastewater; OMW-P, OMW-3 and OMW-2: olive mill wastewater from batch press process as well as three-phase and two-phase continuous extraction procedures, respectively; OMET-2: mixture of all effluents produced in the olive mill, including OWW, OMW, and from other activities in the facility (e.g., cleaning and sanitation);
: biological oxygen demand; EC: conductivity.
**Parameter OMW-P OMW-3 OMET-2 OMW-2 OWW** pH 4.5 5.4 7.2 4.9 6.3 Moisture, % 93.0 93.4 99.4 99.3 99.7 Total solids, % 12.0 6.6 0.59 0.6 0.27 Organic matter, % 10.5 5.8 0.39 0.49 0.10 Ashes, % 1.5 0.9 0.21 0.11 0.17
, mg/L 90.0 42.0 0.29 0.79 0.50 COD, mg/L 180.0 151.4 7.1 7.8 0.8 Total phenols, mg/L 2,400 921.0 86.0 157.0 4.0 EC, mS·cm−1 9.0 7.9 1.9 1.3 0.9
Membrane technology, in particular pressure-driven membrane processes comprising MF, UF, NF, and RO, can offer a series of advantages in contrast with other separation processes, making them very promising and environmentally friendly for the purification and remediation of OMW: needless of chemicals and reagents, e.g., solvents—to accomplish the desired separation and concentration; minor costs (capex and opex) and specific energetic requirements than most conventional separation alternatives, upon significant purification potentiality, selectivity, and recovery factors; and also easy scaling to the industry, low space necessities as they are modular, design simplicity, easy operation, and
Fouling mechanisms are very important to fully understand what is taking place between the membrane and the effluent, in view of the adoption and implementation of adequate decisions for the successful design of the membrane plant. This comprises the setup of specifically tailored pretreatment process and optimized operating conditions. Irreversible fouling arises quickly on the membranes due to the high concentration of pollutants when wastewater is purified without
**2. Membrane processes for olive mill effluent purification**
**Table 1.** Average physicochemical composition of the different types of olive mill effluents [28].
a feasible solution if properly engineered.
COD: chemical oxygen demand; BOD<sup>5</sup>
BOD<sup>5</sup>
130 Desalination
low maintenance needs [5–10].
Canepa et al. [21] conducted a study on OMW-3 (average COD = 90 g/L) treatment based on a combined UF membrane followed by adsorption and RO process at pilot scale. The UF membrane was made of polysulfone, the permeate was treated with adsorbing polymers, and the resins eluate with RO polypiperazine-amide membranes. A rapid decline of the UF permeate flux was observed by the authors, and a daily washing was necessary. The fact that there was a flux decline of 80% after 20 h of UF and 70% after 15 h for the RO membrane was a capital handicap for the feasibility of the proposed process. UF was also examined by Borsani and Ferrando [22]. They proposed a pretreatment consisting of oil removal and suspended solid settling, achieving a final 50% organic matter removal with respect to the raw OMW-3 (COD = 70 g/L). Turano et al. [23] proposed centrifugation as pretreatment, simple, and mechanical operation which does not need addition of chemicals and is available in the production line, before UF—polysulfone, molecular weight cutoff (MWCO) 17 kDa—for the treatment of OMW-3. The centrifugation pretreatment removed 80% of the suspended solid concentration, and 90% COD reduction was achieved at the outlet of the integrated process. Paraskeva et al. [24] examined the treatment/fractionation of OMW-3 by a combination of polypropylene filter screening (80 μm), UF (zirconia, mean pore size 100 nm) followed by NF (polymeric, 200 Da) and RO. In particular, UF removed suspended solids and solid fat/lipid components (90%).
Stoller and Chianese [14] studied the purification of olive washing wastewater (OWW) to comply with municipal sewer discharge standards in Italy. In contrast with OMW, this effluent presents moderate organic pollutant load but high concentration of suspended solids. The authors proposed a coagulation-flocculation pretreatment with polyelectrolytes (aluminum sulfate (AS) or aluminum hydroxide (AH)) before UF and NF in series (composite thin-film spiral-wound membranes). Both pretreatment processes yielded similar COD and biological oxygen demand (BOD<sup>5</sup> ) rejection efficiencies, but higher productivity was observed in the subsequent membrane-in-series process after flocculation with AS. The same procedure was applied by these researchers to OMW-3 [8, 10]. In this case, they also compared the results with two additional pretreatments, heterogeneous photocatalysis with titania nanocatalysts and aerobic digestion. They noted that major pollutant removal by the pretreatment, e.g., COD, is not sufficient to state the suitability of the adopted pretreatment for the downstream membrane operation, but the fact that the pollutant particle size is shifted away from the size of the membrane pores is relevant to enhance steady-state flux.
On another hand, Akdemir and Ozer [25] applied pH adjustment (acidic or alkaline) and cartridge filtration (20 μm) as pretreatment before UF (MWCO 30–100 kDa) and reported the best performance for the 100 kDa UF membrane, pointing for 1 bar as the critical pressure to reduce fouling issues. Also, Coskun et al. [26] examined UF as part of an integral membrane process for OMW-3 reclamation, but did not consider the entity of fouling in the performance of the membranes. Zirehpour et al. [27] examined a lab-made UF membrane for the purification of OMW-3, preceded by a MF membrane (50, 5, and 0.2 μm). The commercial UF membrane provided higher flux than the lab-made one, but the antifouling properties and rejection efficiency of the latter were better.
On another hand, Ochando-Pulido et al. [28, 29] tested a polymeric UF membrane for the reclamation of both OWW and OMW-2, separately or mixed. As pretreatments, three different processes were examined: (i) first, gridding (cut size equal to 300 μm) of the raw effluent was carried out in order to remove the coarse particles; (ii) then, pH-temperature (T) flocculation (pH-T F) was performed by adding HNO<sup>3</sup> (70% w/w) under continuous stirring (320 rpm); and (iii) the supernatant phase of the pH-T F process after the separation of the mud was either directly conducted to the UF unit and thereby referred to as OMWW-F or further pretreated by photocatalysis (PhC) under ultraviolet irradiation (UV) with lab-made ferromagnetic-core TiO2 nanoparticles (UV/TiO<sup>2</sup> PhC), thus named OMWW-F/PhC. Finally, the differently pretreated streams were driven to the UF and NF membrane units. The labmade ferromagnetic photocatalyst nanoparticles were produced in three consecutive steps, following the procedure fully reported elsewhere [30]. Briefly, in the first place, magnetite was synthesized by a sol-gel process in a spinning disk reactor, using tetraethylorthosilicate and tetraisopropoxide as precursors. Moreover, threshold flux-based methods, as previously pointed by Stoller [9], were applied for to optimize the design of the plant and control of the operation.
**2.2. Nanofiltration and reverse osmosis membranes**
Jb
permeate flux at any time t [29].
applicable as well for the design and control of the membrane plant.
Jb
**Figure 1.** Operation strategies for membranes: above (curve line) or upon boundary flux (J<sup>b</sup>
Jb
On the Purification of Agro-Industrial Wastewater by Membrane Technologies: The Case of Olive Mill Effluents
The specific selectivity toward small solutes and the lower energy consumption of NF membranes have boosted their use as tertiary treatment in integrated wastewater treatment processes. On the other hand, RO membranes permit complying with the most stringent regulations for public health and environment protection. Both types of membranes are already applied in the management of industrial effluents of very different sectors, such as stainless steel [40], energy cogeneration [41], nuclear power [42], textile and tannery [43–47], coking [48, 49], carwash [50], pulp and paper [34, 51], pharmaceutical [52], and agro-food industries, such as dairy [13], tomato [12], and olive oil [5–10, 24], among others. However, again, the problem of fouling is always present in the treatment of any kind of effluent by NF and RO membranes. The critical, threshold, and boundary flux theories are, nonetheless,
) (straight line) conditions; Jp:
http://dx.doi.org/10.5772/intechopen.68401
133
Paraskeva et al. [24] used NF after MF/UF for the purification of OMW-3, achieving flux values up to 100–120 L/h and 95% phenol removal. Moreover, better efficiency was achieved by applying RO after NF, ensuring a significant conductivity, salinity, and turbidity decrease and nearly 30 L/h flux and 75–80% recovery of the initial feed volume. However, the permeate flux data regarding fouling was not discussed by the authors. The post-treated effluent was suitable for disposal in aquatic receptors or for irrigation purposes, and the inorganic part (including nitrogen, phosphorus, magnesium, potassium, and metal traces) and the organic fraction (hydrocarbons, nitrogenous compounds, organic acids, polyalcohols) may be potential plant nutrients, mixed with other inorganic or organic fertilizers, for instance, manure or sludge from biological treatments of other sorts of wastes. Zirehpour et al. (2012) evaluated
The first studies on the experimental and theoretical behavior of fouling in MF and UF membranes were performed by Belfort et al., Field and Aimar [31, 32]. The first theoretical model that shed light on membrane transport phenomena of colloidal particles was elaborated by the research groups of Field et al. [3, 32], Bacchin et al. [33], and Mänttäri and Nystörm, [34]. The concept of the critical flux, as the flux which can be successfully attained by a given membrane without incurring in fouling formation during operation time, was explained and proven by these scientists for MF membranes. The insight into the different fouling typologies was addressed by Bacchin et al. and Ognier et al. Fouling was explained because of local conditions that may trigger liquid/gel phases over the membrane layer and in the membrane pores caused by concentration polarization profiles [35, 36]. Later, this concept was also extended to UF and NF membranes [4] and complemented with the concept of the threshold flux. In many cases, it was observed that it was not possible to completely inhibit fouling during the operation of some liquid-liquid membrane systems, as is the case of wastewater [4–7, 37]. The threshold flux divides a low fouling region from a high fouling region. Stoller et al. [38] worked on a reliable methodology to convert previously measured critical flux data into threshold flux values in his studies on OMW-3 purification with polymeric NF membranes. Stoller underlined the key necessity to develop quick and reliable methods capable to allow the estimation of threshold flux conditions and for the conversion of previously available critical flux measurement data into threshold flux data. Finally, both concepts were merged by Stoller and Ochando [39] into a new concept, the boundary flux. The knowledge of real-time boundary flux values is a key factor to design stable control systems for membrane processes (**Figure 1**).
On the Purification of Agro-Industrial Wastewater by Membrane Technologies: The Case of Olive Mill Effluents http://dx.doi.org/10.5772/intechopen.68401 133
**Figure 1.** Operation strategies for membranes: above (curve line) or upon boundary flux (J<sup>b</sup> ) (straight line) conditions; Jp: permeate flux at any time t [29].
#### **2.2. Nanofiltration and reverse osmosis membranes**
of the membranes. Zirehpour et al. [27] examined a lab-made UF membrane for the purification of OMW-3, preceded by a MF membrane (50, 5, and 0.2 μm). The commercial UF membrane provided higher flux than the lab-made one, but the antifouling properties and
On another hand, Ochando-Pulido et al. [28, 29] tested a polymeric UF membrane for the reclamation of both OWW and OMW-2, separately or mixed. As pretreatments, three different processes were examined: (i) first, gridding (cut size equal to 300 μm) of the raw effluent was carried out in order to remove the coarse particles; (ii) then, pH-temperature (T)
(320 rpm); and (iii) the supernatant phase of the pH-T F process after the separation of the mud was either directly conducted to the UF unit and thereby referred to as OMWW-F or further pretreated by photocatalysis (PhC) under ultraviolet irradiation (UV) with lab-made
the differently pretreated streams were driven to the UF and NF membrane units. The labmade ferromagnetic photocatalyst nanoparticles were produced in three consecutive steps, following the procedure fully reported elsewhere [30]. Briefly, in the first place, magnetite was synthesized by a sol-gel process in a spinning disk reactor, using tetraethylorthosilicate and tetraisopropoxide as precursors. Moreover, threshold flux-based methods, as previously pointed by Stoller [9], were applied for to optimize the design of the plant and control of
The first studies on the experimental and theoretical behavior of fouling in MF and UF membranes were performed by Belfort et al., Field and Aimar [31, 32]. The first theoretical model that shed light on membrane transport phenomena of colloidal particles was elaborated by the research groups of Field et al. [3, 32], Bacchin et al. [33], and Mänttäri and Nystörm, [34]. The concept of the critical flux, as the flux which can be successfully attained by a given membrane without incurring in fouling formation during operation time, was explained and proven by these scientists for MF membranes. The insight into the different fouling typologies was addressed by Bacchin et al. and Ognier et al. Fouling was explained because of local conditions that may trigger liquid/gel phases over the membrane layer and in the membrane pores caused by concentration polarization profiles [35, 36]. Later, this concept was also extended to UF and NF membranes [4] and complemented with the concept of the threshold flux. In many cases, it was observed that it was not possible to completely inhibit fouling during the operation of some liquid-liquid membrane systems, as is the case of wastewater [4–7, 37]. The threshold flux divides a low fouling region from a high fouling region. Stoller et al. [38] worked on a reliable methodology to convert previously measured critical flux data into threshold flux values in his studies on OMW-3 purification with polymeric NF membranes. Stoller underlined the key necessity to develop quick and reliable methods capable to allow the estimation of threshold flux conditions and for the conversion of previously available critical flux measurement data into threshold flux data. Finally, both concepts were merged by Stoller and Ochando [39] into a new concept, the boundary flux. The knowledge of real-time boundary flux values is a key factor to design stable control systems for
(70% w/w) under continuous stirring
PhC), thus named OMWW-F/PhC. Finally,
rejection efficiency of the latter were better.
ferromagnetic-core TiO2
membrane processes (**Figure 1**).
the operation.
132 Desalination
flocculation (pH-T F) was performed by adding HNO<sup>3</sup>
nanoparticles (UV/TiO<sup>2</sup>
The specific selectivity toward small solutes and the lower energy consumption of NF membranes have boosted their use as tertiary treatment in integrated wastewater treatment processes. On the other hand, RO membranes permit complying with the most stringent regulations for public health and environment protection. Both types of membranes are already applied in the management of industrial effluents of very different sectors, such as stainless steel [40], energy cogeneration [41], nuclear power [42], textile and tannery [43–47], coking [48, 49], carwash [50], pulp and paper [34, 51], pharmaceutical [52], and agro-food industries, such as dairy [13], tomato [12], and olive oil [5–10, 24], among others. However, again, the problem of fouling is always present in the treatment of any kind of effluent by NF and RO membranes. The critical, threshold, and boundary flux theories are, nonetheless, applicable as well for the design and control of the membrane plant.
Paraskeva et al. [24] used NF after MF/UF for the purification of OMW-3, achieving flux values up to 100–120 L/h and 95% phenol removal. Moreover, better efficiency was achieved by applying RO after NF, ensuring a significant conductivity, salinity, and turbidity decrease and nearly 30 L/h flux and 75–80% recovery of the initial feed volume. However, the permeate flux data regarding fouling was not discussed by the authors. The post-treated effluent was suitable for disposal in aquatic receptors or for irrigation purposes, and the inorganic part (including nitrogen, phosphorus, magnesium, potassium, and metal traces) and the organic fraction (hydrocarbons, nitrogenous compounds, organic acids, polyalcohols) may be potential plant nutrients, mixed with other inorganic or organic fertilizers, for instance, manure or sludge from biological treatments of other sorts of wastes. Zirehpour et al. (2012) evaluated the performance of a NF membrane at the end of a treatment line comprising MF and UF for OMW-3 purification. A specific arrangement of the integrated membrane system was found to be the UF membrane followed by two-step NF membranes in series, the first NF step providing high flux while the second one providing high rejection. Analysis of the fouling behavior of the used membranes was performed, per the recovery ratio and degree of the total flux loss during volume reduction factor experiments.
and minimizes the transformation of the target substances [1]. On the contrary, it can present limited purification efficiency and the loss of capacity depending on the feed solution [54].
On the Purification of Agro-Industrial Wastewater by Membrane Technologies: The Case of Olive Mill Effluents
Among the adsorption methodologies tested with OMW, there are some studies using socalled biosorbents (those of biological origin) as alternatives to other expensive adsorbents. Achak et al. [55] investigated the performance of a low-cost biosorbent made of banana peel. They examined various parameters including adsorbent dosage, pH, and contact time. After optimization, adsorption of phenols from 60 to 88% was achieved upon the equilibrium reached after 3 h of contact time. These authors also tested low-cost biosorbent, wheat bran, which is inexpensive and an available biomaterial. This biosorbent material yielded adsorption efficiencies toward phenolic compounds up to 67%, and the equilibrium was reached in 4 h of contact time. Moreover, both materials were found to provide better performances
Another biosorbent, olive stone biomass, by-product of olive oil industry, was addressed by as adsorbent for iron from secondary-treated OMW [18, 56, 57]. The OMW effluent was pretreated by an advanced oxidation process based on Fenton's reaction, which uses iron-based catalysts. The equilibrium adsorption capacity was found to increase when the particle size decreased (from < 1 to 4.8 mm). The percentage of iron adsorption increased from 30 to 70% when the initial concentration of biomass increased from 25 to 125 g/L. These authors demonstrated that direct reuse of olive pitches or a simple washing with cold and hot water can be sufficient for the adsorption process, which was fast and spontaneous within the first 10–20 min. The experimental data supports both pseudo-first and pseudo-second order models. Also, the adsorption capacity became incremented with the temperature, pin pointing for an endothermic adsorption process. In addition, the values of thermodynamic parameters of the process were reported: the activation energy (Ea = 8.04 kJ mol−1), the standard enthalpy
= 91.4 J mol−1K−1). Furthermore, column regeneration studies were carried out for various adsorption-desorption cycles [56], and the column performance was confirmed to permit multiple service and regeneration cycles, thus enabling the treatment of considerably large volumes of wastewater. A negligible loss in bed height and olive stone mass was observed. After adsorption, the olive stone biosorbent could be used as a biofuel for domestic or industrial uses. The present process was highlighted to be environmentally friendly and capable to reduce the iron load from other different effluents, providing an affordable technology for
Dried *Azolla caroliniana*, a freshwater aquatic fern, and granular activated carbon (AFC-LS activated with phosphoric acid) were studied to evaluate their adsorption and desorption capacities by Ena et al. [58]. Granular activated carbon (GAC) showed better efficiency, although total phenols were more than twice as concentrated in Azolla as those found in the GAC powder. Other studies with activated carbon, in this case in powdered form, for OMW treatment were performed by Sabbah et al. [59], after sand filtration, achieving 95% removal of the phenolic compounds present in OMW. Also, Azzam et al. [60] examined powdered activated carbon in one-stage, two-stage, and three-stage countercurrent adsorption system series in stirred batch vessels. The three-stage countercurrent adsorption process could reduce
the phenol concentration in 96%, and the equilibrium was reached within 2–3 h.
= −19.51 kJ mol−1), and the standard entropy
http://dx.doi.org/10.5772/intechopen.68401
135
upon alkaline pH environments.
= 8.86 kJ mol−1), the standard free energy (∆G<sup>0</sup>
these small- and medium-scale industries.
(∆H<sup>0</sup>
(∆S<sup>0</sup>
Di Lecce et al. [53] used polyamide thin-film composite spiral-wound NF at pilot scale for the fractionation of OMW-3 after MF. The purified NF permeate obtained presented COD and phenolic loads in the range of the standards for discharge into surface waters, with a rejection of 98% for COD, dry matter, and phenolic compounds. However, again, the performance of the NF membrane regarding its fouling was not reported.
On another hand, Ochando-Pulido et al. [28, 29] tested polymeric NF and RO membranes for the reclamation of OWW and OMW-2. The NF membrane was downstream of an UF one, and the effluent was initially pretreated by flocculation and/or heterogeneous photocatalysis (pH-T F and/or UV/TiO<sup>2</sup> PhC). This treatment sequence helped reducing the required membrane area for both membranes, and the applied UV/TiO<sup>2</sup> photocatalysis process enhanced the productivity and longevity of both membranes, achieving a final treated effluent stream compatible for irrigation. Moreover, the threshold flux theory was applied successfully to determine the threshold operating conditions for the process. In a similar line, Stoller [10] worked during 3 years continuously with a NF membrane for the treatment of OMW-3 with the goal to check for the reliability of threshold flux optimization methodologies for the fouling control on the used membrane in the long run. The author reported the successful operation of the NF membrane for the used period of pilot-scale work. Stoller highlighted that this was possible if adequately targeted fouling minimization control methods were carried out, based on the estimation of the critical flux for the optimal operation of the plant. These theories were successfully applied to this system, which could be explained by the threshold flux model.
Ochando-Pulido et al. [6, 7] studied the reclamation of OMW-2 (Spain) by RO, with the goal of "closing the loop," as is the trend of the current "circular economy." the quality to recirculate the final effluent to the olive washing machines of the manufacture process to finally close the loop. The effluent was firstly treated by means of an advanced oxidation process (AOP) which consisted in homogeneous catalysis with Fenton's reagents [16]. The authors noticed the importance of an operating variable of the RO process, the recirculation of a fraction of the permeate. The adoption of this operating sequence permitted enhancing the permeate production of the membrane, which was also stabilized, and resulted as well in a steady rejection of the target species, comprising suspended solids, phenols, and iron removal total rejection, as high as 99.4 and 98.2% of the COD and conductivity, respectively.
## **3. Adsorption processes for olive mill effluent purification**
Adsorption technology has several advantages such as simple handling, relatively low operation costs, and long service lifetime because adsorption materials can be regenerated and reused. Also, adsorption is attractive for its simplicity of design, avoids toxic solvents, and minimizes the transformation of the target substances [1]. On the contrary, it can present limited purification efficiency and the loss of capacity depending on the feed solution [54].
the performance of a NF membrane at the end of a treatment line comprising MF and UF for OMW-3 purification. A specific arrangement of the integrated membrane system was found to be the UF membrane followed by two-step NF membranes in series, the first NF step providing high flux while the second one providing high rejection. Analysis of the fouling behavior of the used membranes was performed, per the recovery ratio and degree of
Di Lecce et al. [53] used polyamide thin-film composite spiral-wound NF at pilot scale for the fractionation of OMW-3 after MF. The purified NF permeate obtained presented COD and phenolic loads in the range of the standards for discharge into surface waters, with a rejection of 98% for COD, dry matter, and phenolic compounds. However, again, the performance of
On another hand, Ochando-Pulido et al. [28, 29] tested polymeric NF and RO membranes for the reclamation of OWW and OMW-2. The NF membrane was downstream of an UF one, and the effluent was initially pretreated by flocculation and/or heterogeneous photocatalysis
productivity and longevity of both membranes, achieving a final treated effluent stream compatible for irrigation. Moreover, the threshold flux theory was applied successfully to determine the threshold operating conditions for the process. In a similar line, Stoller [10] worked during 3 years continuously with a NF membrane for the treatment of OMW-3 with the goal to check for the reliability of threshold flux optimization methodologies for the fouling control on the used membrane in the long run. The author reported the successful operation of the NF membrane for the used period of pilot-scale work. Stoller highlighted that this was possible if adequately targeted fouling minimization control methods were carried out, based on the estimation of the critical flux for the optimal operation of the plant. These theories were successfully applied to this system, which could be explained by the threshold flux model.
Ochando-Pulido et al. [6, 7] studied the reclamation of OMW-2 (Spain) by RO, with the goal of "closing the loop," as is the trend of the current "circular economy." the quality to recirculate the final effluent to the olive washing machines of the manufacture process to finally close the loop. The effluent was firstly treated by means of an advanced oxidation process (AOP) which consisted in homogeneous catalysis with Fenton's reagents [16]. The authors noticed the importance of an operating variable of the RO process, the recirculation of a fraction of the permeate. The adoption of this operating sequence permitted enhancing the permeate production of the membrane, which was also stabilized, and resulted as well in a steady rejection of the target species, comprising suspended solids, phenols, and iron removal
total rejection, as high as 99.4 and 98.2% of the COD and conductivity, respectively.
Adsorption technology has several advantages such as simple handling, relatively low operation costs, and long service lifetime because adsorption materials can be regenerated and reused. Also, adsorption is attractive for its simplicity of design, avoids toxic solvents,
**3. Adsorption processes for olive mill effluent purification**
PhC). This treatment sequence helped reducing the required mem-
photocatalysis process enhanced the
the total flux loss during volume reduction factor experiments.
the NF membrane regarding its fouling was not reported.
brane area for both membranes, and the applied UV/TiO<sup>2</sup>
(pH-T F and/or UV/TiO<sup>2</sup>
134 Desalination
Among the adsorption methodologies tested with OMW, there are some studies using socalled biosorbents (those of biological origin) as alternatives to other expensive adsorbents. Achak et al. [55] investigated the performance of a low-cost biosorbent made of banana peel. They examined various parameters including adsorbent dosage, pH, and contact time. After optimization, adsorption of phenols from 60 to 88% was achieved upon the equilibrium reached after 3 h of contact time. These authors also tested low-cost biosorbent, wheat bran, which is inexpensive and an available biomaterial. This biosorbent material yielded adsorption efficiencies toward phenolic compounds up to 67%, and the equilibrium was reached in 4 h of contact time. Moreover, both materials were found to provide better performances upon alkaline pH environments.
Another biosorbent, olive stone biomass, by-product of olive oil industry, was addressed by as adsorbent for iron from secondary-treated OMW [18, 56, 57]. The OMW effluent was pretreated by an advanced oxidation process based on Fenton's reaction, which uses iron-based catalysts. The equilibrium adsorption capacity was found to increase when the particle size decreased (from < 1 to 4.8 mm). The percentage of iron adsorption increased from 30 to 70% when the initial concentration of biomass increased from 25 to 125 g/L. These authors demonstrated that direct reuse of olive pitches or a simple washing with cold and hot water can be sufficient for the adsorption process, which was fast and spontaneous within the first 10–20 min. The experimental data supports both pseudo-first and pseudo-second order models. Also, the adsorption capacity became incremented with the temperature, pin pointing for an endothermic adsorption process. In addition, the values of thermodynamic parameters of the process were reported: the activation energy (Ea = 8.04 kJ mol−1), the standard enthalpy (∆H<sup>0</sup> = 8.86 kJ mol−1), the standard free energy (∆G<sup>0</sup> = −19.51 kJ mol−1), and the standard entropy (∆S<sup>0</sup> = 91.4 J mol−1K−1). Furthermore, column regeneration studies were carried out for various adsorption-desorption cycles [56], and the column performance was confirmed to permit multiple service and regeneration cycles, thus enabling the treatment of considerably large volumes of wastewater. A negligible loss in bed height and olive stone mass was observed. After adsorption, the olive stone biosorbent could be used as a biofuel for domestic or industrial uses. The present process was highlighted to be environmentally friendly and capable to reduce the iron load from other different effluents, providing an affordable technology for these small- and medium-scale industries.
Dried *Azolla caroliniana*, a freshwater aquatic fern, and granular activated carbon (AFC-LS activated with phosphoric acid) were studied to evaluate their adsorption and desorption capacities by Ena et al. [58]. Granular activated carbon (GAC) showed better efficiency, although total phenols were more than twice as concentrated in Azolla as those found in the GAC powder. Other studies with activated carbon, in this case in powdered form, for OMW treatment were performed by Sabbah et al. [59], after sand filtration, achieving 95% removal of the phenolic compounds present in OMW. Also, Azzam et al. [60] examined powdered activated carbon in one-stage, two-stage, and three-stage countercurrent adsorption system series in stirred batch vessels. The three-stage countercurrent adsorption process could reduce the phenol concentration in 96%, and the equilibrium was reached within 2–3 h.
Another adsorption material used is adsorption resins. Novel ion-exchange (IE) resins developed over the past few decades have promoted this technology as a suitable separation and purification process for wastewater treatment and fractionation purposes. In this sense, IE is considered a green and simple technology, capable of achieving high removal efficiencies and targeted selectivity [2].
Víctor-Ortega et al. [67] studied the removal of phenols from synthetic olive mill aqueous solutions with two different anionic resins: a strong-base resin (Amberlyst A26) and a weak-base one (Amberlite IRA-67). The effect of the phenolic load in the raw effluent and the recirculation time were studied. The equilibrium data were modeled with Langmuir, Freundlich, and Temkin isotherms, and the best correlated one was found to be Langmuir isotherm for both resins. Kinetics of the IE process was examined with the pseudo-first order, pseudo-second order, and intraparticle diffusion models. Both second-order and intraparticle diffusion models could describe the IE mechanism accurately. On the other hand, results revealed that phenol uptake is a spontaneous process for Amberlyst A26 anionic resin (ΔG° = −1.55 kJ mol−1), whereas it was found to be non-spontaneous for Amberlite IRA-67 (ΔG° = 3.06 kJ mol−1). Finally, Amberlyst A26 resin was confirmed to be considerably more efficient (80–98%) than Amberlite IRA-67 for the potential removal of phenols from olive mill
On the Purification of Agro-Industrial Wastewater by Membrane Technologies: The Case of Olive Mill Effluents
http://dx.doi.org/10.5772/intechopen.68401
137
In addition to this, Víctor-Ortega et al. [68] examined a continuous-flow IE process for the purification of OMW-2. They compared the performance of strong-base and weak-base anionic resins. They found that the pH of the feed affected the IE process. This permitted to achieve up to 98% removal of phenols upon an increment of the pH up to 7 with the strongbase resin. The efficiency was noted to be maintained constant at higher pH value. Otherwise, with the other resin, a similar behavior was observed, and the phenolic concentration could be removed up to 57% at pH around 7. With the aim to predict the performance of the resins bed break breakthrough curves, the experimental data were fitted to various models for varying effluent concentrations (5–100 mg/L) under the optimal simulated operating conditions. The authors found an enhancement of the IE efficacy with major concentration of phenols in the feed. For the different model equations tested, Thomas model yielded utmost accuracy, above that of Yoon-Nelson and Clark models. To sum up, and as a very important fact, column regeneration studies showed that almost 100% phenol recovery efficiencies were ensured. The proposed IE process led to a phenol solution susceptible to be concentrated and used in food,
cosmetic, or pharmaceutical sectors and a purified effluent for irrigation purposes.
(OMW, for three phases of mills), and wastewater from olive oil washing (OOW).
The production of olive oil is becoming global, and thus the treatment of the derived effluents from this industry includes wastewater from olive washing (OWW), olive mill wastewater
Conventional physicochemical treatments are not effective to eliminate the significant salinity of OME, reflected in their high electric conductivity values. In this context, advanced separation technologies comprising membranes and adsorption resins have been a breakthrough in terms of advanced separation and purification technologies, but many aspects are still in
In this chapter, a focus on the use of membrane and ion adsorption technologies for the purification of OMW is reported. The effect of different factors comprising the type of membrane, i.e., ultrafiltration (UF), nanofiltration (NF), and reverse osmosis (RO); the type of
industrial effluents.
**4. Conclusions**
development or under investigation.
IE technology is very attractive because of its relative simplicity of application as well as due to the low cost and the effectiveness to remove target species from wastewaters, particularly from diluted solutions. IE resins have also found an increasing application in the drinking water treatment sector over the last few decades, especially when there is a high concentration of natural organic matter (NOM) since high percentages on the removal efficiency of NOM by IE process are found (Comstock and Boyer) [61]. The efficient use of IE depends on several operating factors such as the contact time, operating temperature, pH, flow rate, initial pollutant concentration, and resin characteristics [62].
Canepa et al. [21] used a resin-bed column for the final purification of OMW-3 previously treated by UF, and performed good regarding the retention of polyphenols, and could treat around 3000 L until saturation.
Mineral resins, such as zeolites, have been reported to be useful mineral materials for reducing the phenolic compound concentration from OMW compared to other substrates, such as clay soil and bentonite. The regeneration of zeolite was easy after treatment either by simple settling or light centrifugation procedures. Moreover, this material could be easily regenerated by a very interesting eco-friendly technique known as low-temperature ashing (LTA) [63].
On another hand, Amberlite XAD16, a nonpolar resin, was used by Scoma et al. [64] as the adsorbent phase and ethanol as biocompatible desorbing phase. This process could remove until the 60% of the phenolic compounds from OMW. Ferri et al. [65] studied the adsorption and desorption efficiency of five resins with different physical properties (Amberlite XAD4, XAD7, XAD16, IRA96, and ISOLUTE ENV+) toward an aqueous solution of 10 typical phenolic compounds occurring in OMW. The desorption solvents tested were water, methanol, and ethanol under basic and acidified conditions. IRA96 polar resin reached the highest phenol adsorption (76%). However, the maximum desorption ratios achieved (60%) was achieved with ENV + resin and ethanol as the desorbing phase.
Zagklis et al. [66] tested the fractionation of OMW-3 exiting a previous membrane process comprising NF and RO by a three-step resin process. The nonionic XAD4, XAD16, and XAD7HP resins were selected. The proposed three-step resin process enabled at high extent the separation of phenols from carbohydrates, which hinder further concentration of phenols from the RO concentrate. After this, vacuum distillation was implemented for the concentration of phenols, facilitated by the carbohydrate removal and the change of solvent from water to ethanol. The final product had phenol concentration of 378 g/L in gallic acid equivalents, from an initial content in the raw OMW of 2.64 g/L. Kaleh and Geißen [54] studied the behavior of 16 commercial resins, and they demonstrated that the reduction and selective uptake of phenols from OMW are feasible by choosing the appropriate sorbents, conditions, and pretreatment.
Víctor-Ortega et al. [67] studied the removal of phenols from synthetic olive mill aqueous solutions with two different anionic resins: a strong-base resin (Amberlyst A26) and a weak-base one (Amberlite IRA-67). The effect of the phenolic load in the raw effluent and the recirculation time were studied. The equilibrium data were modeled with Langmuir, Freundlich, and Temkin isotherms, and the best correlated one was found to be Langmuir isotherm for both resins. Kinetics of the IE process was examined with the pseudo-first order, pseudo-second order, and intraparticle diffusion models. Both second-order and intraparticle diffusion models could describe the IE mechanism accurately. On the other hand, results revealed that phenol uptake is a spontaneous process for Amberlyst A26 anionic resin (ΔG° = −1.55 kJ mol−1), whereas it was found to be non-spontaneous for Amberlite IRA-67 (ΔG° = 3.06 kJ mol−1). Finally, Amberlyst A26 resin was confirmed to be considerably more efficient (80–98%) than Amberlite IRA-67 for the potential removal of phenols from olive mill industrial effluents.
In addition to this, Víctor-Ortega et al. [68] examined a continuous-flow IE process for the purification of OMW-2. They compared the performance of strong-base and weak-base anionic resins. They found that the pH of the feed affected the IE process. This permitted to achieve up to 98% removal of phenols upon an increment of the pH up to 7 with the strongbase resin. The efficiency was noted to be maintained constant at higher pH value. Otherwise, with the other resin, a similar behavior was observed, and the phenolic concentration could be removed up to 57% at pH around 7. With the aim to predict the performance of the resins bed break breakthrough curves, the experimental data were fitted to various models for varying effluent concentrations (5–100 mg/L) under the optimal simulated operating conditions. The authors found an enhancement of the IE efficacy with major concentration of phenols in the feed. For the different model equations tested, Thomas model yielded utmost accuracy, above that of Yoon-Nelson and Clark models. To sum up, and as a very important fact, column regeneration studies showed that almost 100% phenol recovery efficiencies were ensured. The proposed IE process led to a phenol solution susceptible to be concentrated and used in food, cosmetic, or pharmaceutical sectors and a purified effluent for irrigation purposes.
## **4. Conclusions**
Another adsorption material used is adsorption resins. Novel ion-exchange (IE) resins developed over the past few decades have promoted this technology as a suitable separation and purification process for wastewater treatment and fractionation purposes. In this sense, IE is considered a green and simple technology, capable of achieving high removal efficien-
IE technology is very attractive because of its relative simplicity of application as well as due to the low cost and the effectiveness to remove target species from wastewaters, particularly from diluted solutions. IE resins have also found an increasing application in the drinking water treatment sector over the last few decades, especially when there is a high concentration of natural organic matter (NOM) since high percentages on the removal efficiency of NOM by IE process are found (Comstock and Boyer) [61]. The efficient use of IE depends on several operating factors such as the contact time, operating temperature, pH, flow rate, initial
Canepa et al. [21] used a resin-bed column for the final purification of OMW-3 previously treated by UF, and performed good regarding the retention of polyphenols, and could treat
Mineral resins, such as zeolites, have been reported to be useful mineral materials for reducing the phenolic compound concentration from OMW compared to other substrates, such as clay soil and bentonite. The regeneration of zeolite was easy after treatment either by simple settling or light centrifugation procedures. Moreover, this material could be easily regenerated by a very interesting eco-friendly technique known as low-temperature ashing
On another hand, Amberlite XAD16, a nonpolar resin, was used by Scoma et al. [64] as the adsorbent phase and ethanol as biocompatible desorbing phase. This process could remove until the 60% of the phenolic compounds from OMW. Ferri et al. [65] studied the adsorption and desorption efficiency of five resins with different physical properties (Amberlite XAD4, XAD7, XAD16, IRA96, and ISOLUTE ENV+) toward an aqueous solution of 10 typical phenolic compounds occurring in OMW. The desorption solvents tested were water, methanol, and ethanol under basic and acidified conditions. IRA96 polar resin reached the highest phenol adsorption (76%). However, the maximum desorption ratios achieved (60%) was achieved
Zagklis et al. [66] tested the fractionation of OMW-3 exiting a previous membrane process comprising NF and RO by a three-step resin process. The nonionic XAD4, XAD16, and XAD7HP resins were selected. The proposed three-step resin process enabled at high extent the separation of phenols from carbohydrates, which hinder further concentration of phenols from the RO concentrate. After this, vacuum distillation was implemented for the concentration of phenols, facilitated by the carbohydrate removal and the change of solvent from water to ethanol. The final product had phenol concentration of 378 g/L in gallic acid equivalents, from an initial content in the raw OMW of 2.64 g/L. Kaleh and Geißen [54] studied the behavior of 16 commercial resins, and they demonstrated that the reduction and selective uptake of phenols from OMW are feasible by choosing the appropriate sorbents, conditions, and
cies and targeted selectivity [2].
136 Desalination
around 3000 L until saturation.
(LTA) [63].
pretreatment.
pollutant concentration, and resin characteristics [62].
with ENV + resin and ethanol as the desorbing phase.
The production of olive oil is becoming global, and thus the treatment of the derived effluents from this industry includes wastewater from olive washing (OWW), olive mill wastewater (OMW, for three phases of mills), and wastewater from olive oil washing (OOW).
Conventional physicochemical treatments are not effective to eliminate the significant salinity of OME, reflected in their high electric conductivity values. In this context, advanced separation technologies comprising membranes and adsorption resins have been a breakthrough in terms of advanced separation and purification technologies, but many aspects are still in development or under investigation.
In this chapter, a focus on the use of membrane and ion adsorption technologies for the purification of OMW is reported. The effect of different factors comprising the type of membrane, i.e., ultrafiltration (UF), nanofiltration (NF), and reverse osmosis (RO); the type of adsorbent (waste material, resins); and the operating conditions will be addressed. The problem of membrane fouling for these systems will also be covered. The use of membrane and adsorption technologies can be a feasible solution if properly engineered, to comply with the guidelines established, e.g., by the Food and Agricultural Organization (FAO) for irrigation purposes or discharge into public waterways.
[10] Stoller M. A three-year long experience of effective fouling inhibition by threshold flux based optimization methods on a NF membrane module for olive mill wastewater
On the Purification of Agro-Industrial Wastewater by Membrane Technologies: The Case of Olive Mill Effluents
http://dx.doi.org/10.5772/intechopen.68401
139
[11] Bódalo A, Gómez JL, Gómez E, Máximo F, Hidalgo AM. Application of reverse osmosis to reduce pollutants present in industrial wastewater. Desalination. 2003;**155**:101-108 [12] Iaquinta M, Stoller M. Optimization of a nanofiltration membrane for tomato industry
[13] Luo J, Ding L, Wan Y, Jaffrin MY. Threshold flux for shear-enhanced nanofiltration: Experimental observation in dairy wastewater treatment. Journal of Membrane Science.
[14] Stoller M, Chianese A. Optimization of membrane batch processes by means of the
[15] Stoller M, Bravi M. Critical flux analyses on differently pretreated olive vegetation waste
[16] Martínez NL, Hodaifa G, Rodríguez VS, Giménez JA, Ochando J. Degradation of organic matter in olive-oil mill wastewater through homogeneous Fenton-like reaction. Chemical
[17] Danellakis D, Ntaikou I, Kornaros M, Dailianis S. Olive oil mill wastewater toxicity in the marine environment: Alterations of stress indices in tissues of mussel *Mytilus*
[18] Hodaifa G, Ben Driss Alami G, Ochando-Pulido JM, Víctor-Ortega MD, Martinez-Ferez A. Kinetic and thermodynamic parameters of iron adsorption onto olive stones.
[19] Karaouzas I, Skoulikidis NT, Giannakou U, Albanis TA. Spatial and temporal effects of olive mill wastewaters to stream macroinvertebrates and aquatic ecosystems status.
[20] Ntougias S, Gaitis F, Katsaris P, Skoulika S, Iliopoulos N, Zervakis GI. The effects of olives harvest period and production year on olive mill wastewater properties – Evaluation of Pleurotus strains as bioindicators of the effluent's toxicity. Chemosphere. 2013;**92**:399-405
[21] Canepa P, Marignetti N, Rognoni U, Calgari S. Olive mills wastewater treatment by
[22] Borsani R, Ferrando B. Ultrafiltration plant for olive vegetation waters by polymeric
[23] Turano E, Curcio S, De Paola MG, Calabrò V, Iorio G. An integrated centrifugation– ultrafiltration system in the treatment of olive mill wastewater. Journal of Membrane
[24] Paraskeva CA, Papadakis VG, Tsarouchi E, Kanellopoulou DG, Koutsoukos PG. Membrane processing for olive mill wastewater fractionation. Desalination. 2007;**213**:218-229
combined membrane processes. Water Research. 1988;**22**(12):1491-1494
membrane batteries. Desalination. 1996;**108**:281-286
water streams: some case studies. Desalination. 2010;**250**:578-582
*galloprovincialis*. Aquatic Toxicology. 2011;**101**(2):358-366
treatment. Membranes. 2013;**32**:37-42
2012;**409**:276-284
wastewater treatment. Desalination. 2009;**245**:314-320
critical flux theory. Desalination. 2006;**191**:62-70
Engineering Journal. 2011;**173**:503-510
Ecological Engineering. 2014;**73**:270-275
Water Research. 2011;**45**(19):6334-6346
Science. 2002;**206**:519-531
## **Author details**
Javier Miguel Ochando-Pulido\* and Antonio Martinez-Ferez
\*Address all correspondence to: jmochandop@gmail.com
Department of Chemical Engineering, University of Granada, Granada, Spain
## **References**
[10] Stoller M. A three-year long experience of effective fouling inhibition by threshold flux based optimization methods on a NF membrane module for olive mill wastewater treatment. Membranes. 2013;**32**:37-42
adsorbent (waste material, resins); and the operating conditions will be addressed. The problem of membrane fouling for these systems will also be covered. The use of membrane and adsorption technologies can be a feasible solution if properly engineered, to comply with the guidelines established, e.g., by the Food and Agricultural Organization (FAO) for irrigation
purposes or discharge into public waterways.
Javier Miguel Ochando-Pulido\* and Antonio Martinez-Ferez
Department of Chemical Engineering, University of Granada, Granada, Spain
[1] Soto ML, Moure A, Domínguez H, Parajó JC. Recovery, concentration and purification of phenolic compounds by adsorption: A review. Journal of Food Engineering. 2011;
[2] Caetano M, Valderrama C, Farran A, Cortina JL. Phenol removal from aqueous solution by adsorption and ion exchange mechanisms onto polymeric resins. Journal of Colloid
[3] Field RW, Wu D, Howell JA, Gupta BB. Critical flux concept for microfiltration fouling.
[4] Field RW, Pearce GK. Critical, sustainable and threshold fluxes for membrane filtration with water industry applications. Advances in Colloid and Interface Science. 2011;**164**: 38-44
[5] Le-Clech P, Chen V, Fane TAG. Fouling in membrane bioreactors used in wastewater
[6] Ochando-Pulido JMS, Rodriguez-Vives JM, Martinez-Ferez A. The effect of permeate recirculation on the depuration of pretreated olive mill wastewater through reverse
[7] Ochando-Pulido JM, Hodaifa G, Rodriguez-Vives S, Martinez-Ferez A. Impacts of operating conditions on reverse osmosis performance of pretreated olive mill wastewa-
[8] Stoller M. On the effect of flocculation as pretreatment process and particle size distri-
[9] Stoller M. Effective fouling inhibition by critical flux based optimization methods on a NF membrane module for olive mill wastewater treatment. Chemical Engineering
bution for membrane fouling reduction. Desalination. 2009;**240**:209-217
\*Address all correspondence to: jmochandop@gmail.com
and Interface Science. 2009;**338**:402-409.
Journal of Membrane Science. 1995;**100**:259-272
treatment. Journal of Membrane Science. 2006;**284**(1-2):17-53
osmosis membranes. Desalination. 2012;**286**:145-154
ter. Water Research. 2012;**46**(15):4621-4632
Journal. 2011;**168**:1140-1148
**Author details**
138 Desalination
**References**
**105**:1-27
[25] Akdemir EO, Ozer A. Investigation of two ultrafiltration membranes for treatment of olive oil mill wastewater. Desalination. 2009;**249**:660-666
[41] Samodurov AN, Lysenko SE, Gromov SL, Panteleev AA, Fedoseeva EB. The use of reverse osmosis technology for water treatment in power engineering. Thermal
On the Purification of Agro-Industrial Wastewater by Membrane Technologies: The Case of Olive Mill Effluents
http://dx.doi.org/10.5772/intechopen.68401
141
[42] Epimakhov VN, Oleinik MS, Moskvin LN. Reverse osmosis filtration based water treatment and special water purification for nuclear power systems. Atomic Energy. 2004;**96**(4):
[43] Ellouze E, Tahri N, Ben Amar R. Enhancement of textile wastewater treatment process
[44] Gönder ZB, Arayici S, Barlas H. Advanced treatment of pulp and paper mill wastewater by nanofiltration process: Effects of operating conditions on membrane fouling.
[45] Liu M, Lü Z, Chen Z, Yu S, Gao C. Comparison of reverse osmosis and nanofiltration membranes in the treatment of biologically treated textile effluent for water reuse.
[46] Stoller M, Sacco O, Sannino D, Chianese A. Successful integration of membrane technologies in a conventional purification process of tannery wastewater streams. Membranes.
[47] Srisukphun T, Chiemchaisri C, Urase T, Yamamoto K. Foulant interaction and RO productivity in textile wastewater reclamation plan. Desalination. 2010;**250**:845-849 [48] Korzenowski C, Minhalma M, Bernardes AM, Ferreira JZ, de Pinho MN. Nanofiltration for the treatment of coke plant ammoniacal wastewaters. Separation and Purification
[49] Yin N, Yang G, Zhong Z, Xing W. Separation of ammonium salts from coking wastewater with nanofiltration combined with diafiltration. Desalination. 2011;**268**:233-237
[50] Boussu K, Kindts C, Vandecasteele C, Van der Bruggen B. Applicability of nanofiltration in the carwash industry. Separation and Purification Technology. 2007;**54**(2):139-146
[51] Pizzichini M, Russo C, Di Meo C. Purification of pulp and paper wastewater with membrane technology for water reuse in a closed loop. Desalination. 2005;**178**:351-359
[52] Wei X, Wang Z, Fan F, Wang J, Wang S. Advanced treatment of a complex pharmaceutical wastewater by nanofiltration: Membrane foulant identification and cleaning.
[53] Di Lecce G, Cassano A, Bendini A, Conidi C, Giorno L, Gallina T. Characterization of olive mill wastewater fractions treatment by integrated membrane process. Journal of Science
[54] Kaleh Z, Geißen SU. Selective isolation of valuable biophenols from olive mill wastewa-
[55] Achak M, Hafidi A, Mandia L, Ouazzani N. Removal of phenolic compounds from olive mill wastewater by adsorption onto wheat bran. Desalination and Water Treatment.
ter. Journal of Environmental Chemical Engineering. 2016;**4**:373-384
Engineering. 2006;**53**(6):439-443
Desalination. 2011;**281**:372-378
Technology. 2011;**76**(3):303-307
Desalination. 2010;**251**:167-175
2014;**52**:2875-2885
of Food and Agriculture. 2014;**94**:2935-2942
2013;**3**:126-135
using nanofiltration. Desalination. 2012;**286**:16-23
Separation and Purification Technology. 2011;**76**(3):292-302
234-240
[41] Samodurov AN, Lysenko SE, Gromov SL, Panteleev AA, Fedoseeva EB. The use of reverse osmosis technology for water treatment in power engineering. Thermal Engineering. 2006;**53**(6):439-443
[25] Akdemir EO, Ozer A. Investigation of two ultrafiltration membranes for treatment of
[26] Coskun T, Debik E, Demir NM. Treatment of olive mill wastewaters by nanofiltration
[27] Zirehpour A, Jahanshahi M, Rahimpour A. Unique membrane process integration for olive oil mill wastewater purification. Separation and Purification Technology.
[28] Ochando-Pulido JM, Hodaifa G, Victor-Ortega MD, Rodriguez-Vives S, Martinez-Ferez A. Effective treatment of olive mill effluents from two-phase and three-phase extraction processes by batch membranes in series operation upon threshold conditions. Journal of
[29] Ochando-Pulido JM, Verardo V, Segura-Carretero A, Martinez-Ferez A. Technical optimization of an integrated UF/NF pilot plant for conjoint batch treatment of two-
[30] Ochando-Pulido JM, Stoller M. Boundary flux optimization of a nanofiltration membrane module used for the treatment of olive mill wastewater from a two-phase extrac-
[31] Belfort G, Davis RH, Zydney AL. The behavior of suspensions and macromolecular solutions in crossflow microfiltration. Journal Membrane Science. 1994;**96**:1-58
[32] Field RW, Aimar P. Ideal limiting fluxes in ultrafiltration: Comparison of various
[33] Bacchin P, Aimar P, Sanchez V. Influence of surface interaction on transfer during
[34] Mänttäri M, Nystörm M. Critical flux in NF of high molar mass polysaccharides and effluents from the paper industry. Journal of Membrane Science. 2000;**170**:257-273
[35] Bacchin P, Aimar P, Field RW. Critical and sustainable fluxes: Theory, experiments and
[36] Ognier S, Wisniewski C, Grasmick A. Membrane bioreactor fouling in sub-critical filtration conditions: A local critical flux concept. Journal of Membrane Science.
[37] Espinasse B, Bacchin P, Aimar P. On an experimental method to measure critical flux in
[38] Stoller M, Bravi M, Chianese A. Threshold flux measurements of a nanofiltration membrane module by critical flux data conversion. Desalination. 2012;**315**:142-148
[39] Stoller M, Ochando Pulido JM. The boundary flux handbook: A comprehensive database of critical and threshold flux values for membrane practitioners. Netherlands: Elsevier; 2015
[40] Lee J, Kwon T, Moon I. Performance of polyamide reverse osmosis membranes for steel
phase olives and olive oil washing wastewaters. Desalination. 2015;**364**:82-89
tion process. Separation and Purification Technology. 2014;**130**:124-131
theoretical relationships. Journal of Membrane Science. 1993;**80**:107-115
colloid ultrafiltration. Journal of Membrane Science. 1996;**115**:49-63
applications. Journal of Membrane Science. 2006;**281**:42-69
ultrafiltration. Desalination. 2002;**146**:91-96
wastewater reuse. Desalination. 2006;**189**:309-322
olive oil mill wastewater. Desalination. 2009;**249**:660-666
Hazardous Material. 2013;**263**(1):168-176
2012;**96**:124-131
140 Desalination
2004;**229**:171-177
and reverse osmosis membranes. Desalination. 2010;**259**:65-70
[56] Hodaifa G, Ochando-Pulido JM, Alami SBD, Rodriguez-Vives S, Martinez-Ferez A. Iron removal from liquid effluents by olive stones on adsorption column: Breakthrough curves. Ecological Engineering. 2014;**73**:270-275
**Chapter 8**
**Solar Thermal‐Driven Desalination Pursuing Products**
**of Pure Water and Salts and Leaving Minimum Impact**
Desalination, removal of salt and other minerals from seawater, brackish water, and wastewater, is becoming a promising solution for providing the increasing need of freshwater. It is highly desirable that environmentally friendly renewable energy resources be utilized for water treatment to minimize the consumption of fossil fuels. Given that most desalination systems can directly use thermal energy, concentrated solar thermal energy is very suitable for application to the water treatment. To avoid the potential negative impacts from disposing the concentrates, recovery of important minerals from concentrates to achieve zero discharge is a promising option. The recent technology development on solar thermal energy storages has shown that sea salts are very promising materials for large-scale thermal energy storage. Hence, a full separation of salts and water in desalination process becomes a necessity in advanced water treatment technologies, which should be achieved in an economically feasible way. Literature review and studies about innovative concept of full separation desalination system will be presented in this study. A full separation device integrated with conventional multieffect distillation or multistage flashing water treatment system will be introduced into the system design to enhance the water productivity and thermal efficiency.
Keywords: desalination, full separation multieffect distillation (FSMED), concentrating
The population growth, urbanization, and climate change lead to increase in household and industrial uses of more freshwater, while natural freshwater reserves cannot meet the demands. Over one-third of the population lives in water-stressed countries, and by 2025 this
> © 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
**to Environment**
Abstract
Ben Xu, Peiwen Li and Penghua Guo
http://dx.doi.org/10.5772/intechopen.68702
Additional information is available at the end of the chapter
solar power (CSP), thermal energy storage (TES)
1. Introduction to desalination process
## **Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum Impact to Environment**
Ben Xu, Peiwen Li and Penghua Guo
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.68702
#### Abstract
[56] Hodaifa G, Ochando-Pulido JM, Alami SBD, Rodriguez-Vives S, Martinez-Ferez A. Iron removal from liquid effluents by olive stones on adsorption column: Breakthrough
[57] Martínez Nieto L, Ben Driss Alami S, Hodaifa G, Faur C, Rodríguez S, Giménez JA, Ochando J. Adsorption of iron on crude olive stones. Industrial Crops and Products.
[58] Ena A, Pintucci C, Carlozzi P. The recovery of polyphenols from olive mill waste using two adsorbing vegetable matrices. Journal of Biotechnology. 2012;**157**:573-577
[59] Sabbah I, Marsook T, Basheer S. The effect of pretreatment on anaerobic activity of olive mill wastewater using batch and continuous systems. Process Biochemistry.
[60] Azzam MOJ, Al-Malah K, Al-Gazzawi Z, Al-Omari SA. Dynamic treatment response of olive mills wastewater using series of adsorption steps. Clean – Soil, Air, Water.
[61] Comstock SEH, Boyer TH. Combined magnetic ion exchange and cation exchange for removal of DOC and hardness. Chemical Engineering Journal. 2014;**241**:366-375
[62] Víctor-Ortega MD, Ochando-Pulido JM, Hodaifa G, Martinez-Ferez A. Final purification of synthetic olive oil mill wastewater treated by chemical oxidation using ion exchange: Study of operating parameters. Chemical Engineering Process. Process Intensification.
[63] Santi CA, Cortes S, D'Acqui LP, Sparvoli E, Pushparaj B. Reduction of organic pollutants in olive mill wastewater by using different mineral substrates as adsorbents.
[64] Scoma A, Bertin L, Zanaroli G, Fraraccio S, Fava F. A physicochemical-biotechnological approach for an integrated valorization of olive mill wastewater. Bioresources
[65] Ferri F, Bertin L, Scoma A, Marchetti L, Fava F. Recovery of low molecular weight phenols through solid-phase extraction. Chemical Engineering Journal. 2011;**166**:994-1001
[66] Zagklis DP, Vavouraki AI, Kornaros ME, Paraskeva C. Purification of olive mill wastewater phenols through membrane filtration and resin adsorption/desorption. Journal of
[67] Víctor-Ortega MD, Ochando-Pulido JM, Martínez-Férez A. Phenols removal from industrial effluents through novel polymeric resins: Kinetics and equilibrium studies.
[68] Víctor-Ortega MD, Ochando-Pulido JM, Martínez-Férez A. Performance and modeling of continuous ion exchange processes for phenols recovery from olive mill wastewater.
curves. Ecological Engineering. 2014;**73**:270-275
Bioresources Technology. 2008;**99**:1945-1951
Technology. 2011;**102**:10273-10279
Hazardous Materials. 2015;**285**:69-76
Separation and Purification Technology. 2016a;**160**:136-144
Process Safety and Environment Protection. 2016b;**100**:242-251
2010;**32**:467-471
142 Desalination
2004;**39**:1947-1951
2010;**38**:822-830
2014;**85**:241-247
Desalination, removal of salt and other minerals from seawater, brackish water, and wastewater, is becoming a promising solution for providing the increasing need of freshwater. It is highly desirable that environmentally friendly renewable energy resources be utilized for water treatment to minimize the consumption of fossil fuels. Given that most desalination systems can directly use thermal energy, concentrated solar thermal energy is very suitable for application to the water treatment. To avoid the potential negative impacts from disposing the concentrates, recovery of important minerals from concentrates to achieve zero discharge is a promising option. The recent technology development on solar thermal energy storages has shown that sea salts are very promising materials for large-scale thermal energy storage. Hence, a full separation of salts and water in desalination process becomes a necessity in advanced water treatment technologies, which should be achieved in an economically feasible way. Literature review and studies about innovative concept of full separation desalination system will be presented in this study. A full separation device integrated with conventional multieffect distillation or multistage flashing water treatment system will be introduced into the system design to enhance the water productivity and thermal efficiency.
Keywords: desalination, full separation multieffect distillation (FSMED), concentrating solar power (CSP), thermal energy storage (TES)
### 1. Introduction to desalination process
The population growth, urbanization, and climate change lead to increase in household and industrial uses of more freshwater, while natural freshwater reserves cannot meet the demands. Over one-third of the population lives in water-stressed countries, and by 2025 this
figure is projected to rise to two-thirds [1]. Water with sufficient quantity and good quality for household uses and industrial applications is critical to health and well-being, as well as the opportunities to achieve human and economic development. Desalination, removal of salt and other minerals from seawater, brackish water, and wastewater, is a promising solution to grow the supply for fresh water. Presently, many arid areas have to rely on desalination to provide major quantities of safe water [2].
According to the International Desalination Association (IDA), there are 18,426 desalination plants in operation in more than 150 countries in June 2015, producing about 86.8 million cubic meters of water per day [3]. This number is continuously growing as the need for fresh water supply grows. Figure 1 shows the world desalination plants per geographical area, and over half of these plants are in the Middle East.
A simplified process sequence of a typical desalination plant is provided in Figure 2. A brief introduction of the source intake, pretreatment, and post-treatment is given below, whereas the detailed description of the desalination process will be given in the next section.
Basically, there are two types of intake facilities to access reliable water sources: subsurface and open ocean intakes [6]. Proper design of the intake facility would not only protect downstream equipment and reduce adverse effects on aquatic life but also be beneficial to system productivity and capital/operating costs. The choice of the intake facility depends on the geographical conditions and desalination technology employed. For example, the raw feed water obtained from the intake well (belong to subsurface intake) is pretreated via slow filtration which is beneficial to the pretreatment process. However, the intake well is usually used for small-scale desalination plant due to their small productivity [5].
In addition to dissolved solids, the raw feed water generally contains other impurities such as silt, algae, bacteria, and even small aquatic life. The pretreatment process, involving some filtration and physical-chemical processes, is mainly used to improve the quality of the raw feed water to meet the requirement of different desalination technologies. Generally, the membrane desalination requires a higher degree of pretreatment than distillation technologies. To ensure that the product water from the desalination process meets statutory water quality standards for public health and protection of the water distribution system, post-treatment of the product water is necessary. This process may involve pH adjustment, disinfection, boron removal, and addition of minerals and corrosion inhibitors.
Despite the tremendous improvements in conventional water treatment technology, desalination industry is still facing several practical challenges on high-energy consumption, using either membrane-based technologies or thermal-based phase change approaches. Therefore, it is highly desirable that environmentally friendly renewable energy resources should be utilized for water treatment to minimize the consumption of fossil fuels. More details on this issue can be found in Section 3.
Another key concern associated with the desalination is the potential negative environmental impacts caused by the concentrated discharges to the environment, such as the adverse effects on water and sediment quality, aquatic lives, and the functioning of ecosystems. The concentrate is generated as a side product of the desalination process, which contains most of the
minerals and contaminants of the source water and pretreatment additives in concentrated form. Based on a survey of concentrated disposal methods in 203 desalination plants in USA [7], it turns out that 87% of all 203 desalination plants are using surface water discharge
Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum…
http://dx.doi.org/10.5772/intechopen.68702
145
Figure 2. Typical sequence of desalination treatment and distribution processes [5].
Figure 1. World desalination plants per geographical area [4].
Figure 1. World desalination plants per geographical area [4].
figure is projected to rise to two-thirds [1]. Water with sufficient quantity and good quality for household uses and industrial applications is critical to health and well-being, as well as the opportunities to achieve human and economic development. Desalination, removal of salt and other minerals from seawater, brackish water, and wastewater, is a promising solution to grow the supply for fresh water. Presently, many arid areas have to rely on desalination to provide
According to the International Desalination Association (IDA), there are 18,426 desalination plants in operation in more than 150 countries in June 2015, producing about 86.8 million cubic meters of water per day [3]. This number is continuously growing as the need for fresh water supply grows. Figure 1 shows the world desalination plants per geographical area, and over
A simplified process sequence of a typical desalination plant is provided in Figure 2. A brief introduction of the source intake, pretreatment, and post-treatment is given below, whereas
Basically, there are two types of intake facilities to access reliable water sources: subsurface and open ocean intakes [6]. Proper design of the intake facility would not only protect downstream equipment and reduce adverse effects on aquatic life but also be beneficial to system productivity and capital/operating costs. The choice of the intake facility depends on the geographical conditions and desalination technology employed. For example, the raw feed water obtained from the intake well (belong to subsurface intake) is pretreated via slow filtration which is beneficial to the pretreatment process. However, the intake well is usually used for small-scale
In addition to dissolved solids, the raw feed water generally contains other impurities such as silt, algae, bacteria, and even small aquatic life. The pretreatment process, involving some filtration and physical-chemical processes, is mainly used to improve the quality of the raw feed water to meet the requirement of different desalination technologies. Generally, the membrane desalination requires a higher degree of pretreatment than distillation technologies. To ensure that the product water from the desalination process meets statutory water quality standards for public health and protection of the water distribution system, post-treatment of the product water is necessary. This process may involve pH adjustment, disinfection, boron
Despite the tremendous improvements in conventional water treatment technology, desalination industry is still facing several practical challenges on high-energy consumption, using either membrane-based technologies or thermal-based phase change approaches. Therefore, it is highly desirable that environmentally friendly renewable energy resources should be utilized for water treatment to minimize the consumption of fossil fuels. More details on this issue
Another key concern associated with the desalination is the potential negative environmental impacts caused by the concentrated discharges to the environment, such as the adverse effects on water and sediment quality, aquatic lives, and the functioning of ecosystems. The concentrate is generated as a side product of the desalination process, which contains most of the
the detailed description of the desalination process will be given in the next section.
major quantities of safe water [2].
144 Desalination
half of these plants are in the Middle East.
desalination plant due to their small productivity [5].
removal, and addition of minerals and corrosion inhibitors.
can be found in Section 3.
Figure 2. Typical sequence of desalination treatment and distribution processes [5].
minerals and contaminants of the source water and pretreatment additives in concentrated form. Based on a survey of concentrated disposal methods in 203 desalination plants in USA [7], it turns out that 87% of all 203 desalination plants are using surface water discharge and sanitary sewer discharge methods, while the rest part includes deep-well injection, evaporation ponds, and spray irrigation methods, unfortunately none of them has any zero-liquid discharge technology applied.
Obviously, most of the desalination brines are disposed to the sea or the sewer lines, and this will lead to adverse effects to the environment. To avoid the potential negative impacts from disposing the concentrates, exploring the options of recovering the dissolved salts from the desalination effluent to achieve zero-liquid discharge is necessary. The existing method to recover salts from desalination is solar pond approach which requires extensive areas of land. To better understand all the pros and cons, a brief review of various desalination technologies will be provided in the next section, then a novel water and solute full separation process using solar thermal energy will be introduced. Unlike the existing process which uses crystallizer, the proposed approach is a once-through process involving zero recirculation and zero-liquid discharge. It is expected that the proposed full separation thermal-driven desalination process in this work can address problems of concentrated discharge disposal, solute recovery, and high-energy consumption.
## 2. Classification of desalination technologies
Currently, most of the desalination plants have been implemented in large scales; there are more than 15,000 desalination plants installed by 2010 in the world, with a production capacity of 65 million m3 /day for both domestic consumption and industrial water production. There are a lot of different desalination technologies, some of them have already been fully developed at large scales, whereas others are still in pilot scales for demonstration or laboratory scales for research and development. Figure 3 provides a list of common desalination technologies.
Figure 3. A list of contemporary desalination technologies.
Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum…
http://dx.doi.org/10.5772/intechopen.68702
147
Figure 4. Schematic of RO [8].
Basically, desalination process can be categorized as two major ones: thermal desalination and mechanical desalination. Thermal desalination utilizes the heat from combustion, power block, or even renewable energy to evaporate seawater. Vapor compression can be combined with thermal and mechanical desalination, which has the capability of increasing volumes and efficiency of the whole process. Thermal desalination has three classifications, which include filtration, evaporation, and crystallization. Mechanical desalination is discussed in this section and thermal desalination will be introduced in the next section.
Reverse osmosis (RO) is a membrane separation process that recovers water from a saline solution pressurized to a point greater than the osmotic pressure of the solution. The saline water is fed to the porous membranes at high pressures. The hydrophilic membranes allow only water to pass through it, as shown in Figure 4. This technique requires high-pressure pumps and costly membranes; also the membrane is susceptible to fouling and needs frequent replacement, which results in high cost of maintenance. Incorporation of energy recovery system reduces specific energy consumption and product cost but increases the capital cost. The product cost is significantly affected by the price of electricity in this technique.
Pressurizing the saline water accounts for most of the energy consumed in RO. The osmotic pressure required to separate water from the brine is related to the salt concentration; therefore,
Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum… http://dx.doi.org/10.5772/intechopen.68702 147
Figure 3. A list of contemporary desalination technologies.
and sanitary sewer discharge methods, while the rest part includes deep-well injection, evaporation ponds, and spray irrigation methods, unfortunately none of them has any zero-liquid
Obviously, most of the desalination brines are disposed to the sea or the sewer lines, and this will lead to adverse effects to the environment. To avoid the potential negative impacts from disposing the concentrates, exploring the options of recovering the dissolved salts from the desalination effluent to achieve zero-liquid discharge is necessary. The existing method to recover salts from desalination is solar pond approach which requires extensive areas of land. To better understand all the pros and cons, a brief review of various desalination technologies will be provided in the next section, then a novel water and solute full separation process using solar thermal energy will be introduced. Unlike the existing process which uses crystallizer, the proposed approach is a once-through process involving zero recirculation and zero-liquid discharge. It is expected that the proposed full separation thermal-driven desalination process in this work can address problems of concentrated discharge disposal, solute recovery, and
Currently, most of the desalination plants have been implemented in large scales; there are more than 15,000 desalination plants installed by 2010 in the world, with a production capacity of
lot of different desalination technologies, some of them have already been fully developed at large scales, whereas others are still in pilot scales for demonstration or laboratory scales for research and development. Figure 3 provides a list of common desalination technologies.
Basically, desalination process can be categorized as two major ones: thermal desalination and mechanical desalination. Thermal desalination utilizes the heat from combustion, power block, or even renewable energy to evaporate seawater. Vapor compression can be combined with thermal and mechanical desalination, which has the capability of increasing volumes and efficiency of the whole process. Thermal desalination has three classifications, which include filtration, evaporation, and crystallization. Mechanical desalination is discussed in this section
Reverse osmosis (RO) is a membrane separation process that recovers water from a saline solution pressurized to a point greater than the osmotic pressure of the solution. The saline water is fed to the porous membranes at high pressures. The hydrophilic membranes allow only water to pass through it, as shown in Figure 4. This technique requires high-pressure pumps and costly membranes; also the membrane is susceptible to fouling and needs frequent replacement, which results in high cost of maintenance. Incorporation of energy recovery system reduces specific energy consumption and product cost but increases the capital cost.
Pressurizing the saline water accounts for most of the energy consumed in RO. The osmotic pressure required to separate water from the brine is related to the salt concentration; therefore,
The product cost is significantly affected by the price of electricity in this technique.
/day for both domestic consumption and industrial water production. There are a
discharge technology applied.
146 Desalination
high-energy consumption.
65 million m3
2. Classification of desalination technologies
and thermal desalination will be introduced in the next section.
Figure 4. Schematic of RO [8].
and separate the relatively pure vapor for subsequent condensation and use. Membrane separation systems usually drive high-pressure pumps that overcome osmotic pressure differentials or create electric fields that drive electro-migration of ions in solution. Figure 6 summa-
Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum…
http://dx.doi.org/10.5772/intechopen.68702
149
MSF is based on the principle of heating the fluid at certain pressure and then flashing it at lower pressure to form vapor, as shown in Figure 7. This vapor is collected and condensed which gives purified water. Here, the brine water is fed through a series of feed water heaters, to recover the energy from flashed steam, and then fed to an unfired boiler or a heat exchanger. The brine water gains maximum heat at unfired boiler and then flashed in several stages with decreasing pressure, each stage giving out some amount of steam. The difference of pressure between subsequent stages is the main factor influencing steam production in each stage. Highly concentrated brine is discharged from the last stage. The main problem of MSF process is the low-performance ratio which causes lower efficiency; however, it has lower scaling problems than that of MED due to relatively simpler design [10]. The number of MSF plants grows since its conception. And the cost of MSF equipment has been reduced by 50% in the last 20 years; however, this is accompanied with an increasing unit size [11]. Therefore, MSF units are economical with large capacities. In case of solar-coupled MSF system, the low-pressure steam can be formed by circulating water through the field of parabolic trough
An optimization study indicates that there is significant declining trend of product water cost with increasing top brine temperature (TBT). For a 30-stage MSF plant product, water cost is 1.15 \$/m<sup>3</sup> for 347 K TBT, and that for 377 K TBT is 0.95 \$/m<sup>3</sup> [13]. However, the TBT corresponding to the minimum product cost cannot be achieved due to the problem of scaling at high TBT. The area of technical optimization for MSF can be new corrosion-resistant alloys
The governing principle for MED is to boil inlet seawater or brine in different evaporation effects, as shown in Figure 8. In the first effect, heat given by steam from waste heat source or solar collector is used to vaporize seawater. The generated vapor passes to the next effect. This vapor loses its latent heat to evaporate a part of seawater fed in the next effect, and so on. Flow schemes for MED systems include forward feeding, backward feeding, parallel feeding, and parallel feeding with cross flow. In forward feed, the direction of brine flow and steam flow is same and all the feed seawater is sent to first effect; in backward feeding, the direction of brine flow is opposite to steam flow and the seawater is firstly introduced into the last effect. Backward feed scheme makes more sense thermodynamically, but the first effect receives highest concentration brine at a high temperature. This escalates the scaling problems; therefore, this scheme is avoided. In parallel feed scheme, feed is equally divided and distributed to different
effects. MED units typically operate below 120�C TBT, to avoid the problem of scaling.
rizes all these methods.
collectors.
and corrosion and scale-inhibiting techniques.
3.1.2. Multieffect distillation (MED)
3.1.1. Multistage flash (MSF)
Figure 5. Principle of vapor compression [9].
RO has been considered as a practical approach for brackish water, due to the fact that brackish water only requires low to intermediate pressure range (10–15 bar) for the RO process, which is only one-fifth of the pressure for seawater RO process.
To improve the system efficiency, vapor compression can be added to a multieffect distillation (MED) process, as shown in Figure 5. The reuse of vapor is the key for vapor compression process, where the vapor is generated from the distiller after recompression. The first module in Figure 5 can be heated up by utilizing the recovered heat from the partially recompressed vapor on stage. The vapor can be compressed either by a mechanical compressor or by a steam ejector, which can be categorized as mechanical vapor compression (MVC) and thermal vapor compression (TVC), respectively,
For TVC, motive steam (in ejector) at higher pressure is withdrawn from another process, for example, a steam power cycle or an industrial process steam. MVC is useful for small- or medium-scale desalination plants. MVC units typically can generate fresh water up to about 3000 m3 /day, while TVC units have much more capacity, with daily fresh water generation of 36,000 m3 . Most of MVC systems have only one stage, while TVC systems usually have several stages. This difference arises from the fact that the pressure and temperature increased by the compressor and its capacity are limited.
## 3. Thermal-driven desalination systems and the application of molten salts as heat transfer fluid in concentrating solar power plants
#### 3.1. Thermal-driven desalination systems
Conventional thermal-driven desalination technologies are broadly classified into two major categories: filtration and evaporation. The thermal distillation systems will heat saline water and separate the relatively pure vapor for subsequent condensation and use. Membrane separation systems usually drive high-pressure pumps that overcome osmotic pressure differentials or create electric fields that drive electro-migration of ions in solution. Figure 6 summarizes all these methods.
### 3.1.1. Multistage flash (MSF)
RO has been considered as a practical approach for brackish water, due to the fact that brackish water only requires low to intermediate pressure range (10–15 bar) for the RO process, which is
To improve the system efficiency, vapor compression can be added to a multieffect distillation (MED) process, as shown in Figure 5. The reuse of vapor is the key for vapor compression process, where the vapor is generated from the distiller after recompression. The first module in Figure 5 can be heated up by utilizing the recovered heat from the partially recompressed vapor on stage. The vapor can be compressed either by a mechanical compressor or by a steam ejector, which can be categorized as mechanical vapor compression (MVC) and thermal vapor
For TVC, motive steam (in ejector) at higher pressure is withdrawn from another process, for example, a steam power cycle or an industrial process steam. MVC is useful for small- or medium-scale desalination plants. MVC units typically can generate fresh water up to about
stages. This difference arises from the fact that the pressure and temperature increased by the
Conventional thermal-driven desalination technologies are broadly classified into two major categories: filtration and evaporation. The thermal distillation systems will heat saline water
3. Thermal-driven desalination systems and the application of molten
salts as heat transfer fluid in concentrating solar power plants
/day, while TVC units have much more capacity, with daily fresh water generation of
. Most of MVC systems have only one stage, while TVC systems usually have several
only one-fifth of the pressure for seawater RO process.
compression (TVC), respectively,
Figure 5. Principle of vapor compression [9].
compressor and its capacity are limited.
3.1. Thermal-driven desalination systems
3000 m3
148 Desalination
36,000 m3
MSF is based on the principle of heating the fluid at certain pressure and then flashing it at lower pressure to form vapor, as shown in Figure 7. This vapor is collected and condensed which gives purified water. Here, the brine water is fed through a series of feed water heaters, to recover the energy from flashed steam, and then fed to an unfired boiler or a heat exchanger. The brine water gains maximum heat at unfired boiler and then flashed in several stages with decreasing pressure, each stage giving out some amount of steam. The difference of pressure between subsequent stages is the main factor influencing steam production in each stage. Highly concentrated brine is discharged from the last stage. The main problem of MSF process is the low-performance ratio which causes lower efficiency; however, it has lower scaling problems than that of MED due to relatively simpler design [10]. The number of MSF plants grows since its conception. And the cost of MSF equipment has been reduced by 50% in the last 20 years; however, this is accompanied with an increasing unit size [11]. Therefore, MSF units are economical with large capacities. In case of solar-coupled MSF system, the low-pressure steam can be formed by circulating water through the field of parabolic trough collectors.
An optimization study indicates that there is significant declining trend of product water cost with increasing top brine temperature (TBT). For a 30-stage MSF plant product, water cost is 1.15 \$/m<sup>3</sup> for 347 K TBT, and that for 377 K TBT is 0.95 \$/m<sup>3</sup> [13]. However, the TBT corresponding to the minimum product cost cannot be achieved due to the problem of scaling at high TBT. The area of technical optimization for MSF can be new corrosion-resistant alloys and corrosion and scale-inhibiting techniques.
#### 3.1.2. Multieffect distillation (MED)
The governing principle for MED is to boil inlet seawater or brine in different evaporation effects, as shown in Figure 8. In the first effect, heat given by steam from waste heat source or solar collector is used to vaporize seawater. The generated vapor passes to the next effect. This vapor loses its latent heat to evaporate a part of seawater fed in the next effect, and so on. Flow schemes for MED systems include forward feeding, backward feeding, parallel feeding, and parallel feeding with cross flow. In forward feed, the direction of brine flow and steam flow is same and all the feed seawater is sent to first effect; in backward feeding, the direction of brine flow is opposite to steam flow and the seawater is firstly introduced into the last effect. Backward feed scheme makes more sense thermodynamically, but the first effect receives highest concentration brine at a high temperature. This escalates the scaling problems; therefore, this scheme is avoided. In parallel feed scheme, feed is equally divided and distributed to different effects. MED units typically operate below 120�C TBT, to avoid the problem of scaling.
Figure 6. Thermal desalination technologies.
them to use cheap heat sources such as waste heat or solar collectors. For example, an existing diesel waste heat-coupled MED plant's energy consumption is 2 kWh/Ton for water
Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum…
http://dx.doi.org/10.5772/intechopen.68702
151
In membrane distillation process, feed water is heated and then allowed to flow through the porous hydrophobic membrane. The high pressure or electrical potential is applied on water vapor to produce fresh water from saline water [16]. The vapor pressure difference across membrane causes water vapor molecules to flow and it is condensed on the other side of the membrane. MD has the following characteristics: high porosity, hydrophobic and low thermal conductivity; the membrane thickness should be reduced and to maintain high pore size for
The following are the operating parameters which effects distillate yield like flow rate, thickness of air gap, membrane thickness, porosity, long-term operation, thermal conductivity of membranes, and porosity. It is still a big challenge to develop high and efficient membrane technology, since MD has basic challenges to meet to compete with MSF, MED, and RO. In
3.2. Application of molten salts as heat transfer fluid in concentrating solar power plants
tion process is very energy consuming. For instance, to produce 1000 m<sup>3</sup>
Despite the tremendous improvements in conventional desalination technology, the desalina-
water in a year, nearly 10,000 tons of oil is required [16]. More detailed energy consumption
/day of desalinated
pumps [15].
the increase of flux.
3.1.3. Membrane distillation (MD)
Figure 8. Schematic of triple-effect MED [14].
addition, MD system consumes a lot of energy [17, 18].
data for different desalination processes are listed in Table 1.
Figure 7. Schematic of MSF [12].
And the minimum brine temperature is determined by temperature difference between the last effect and ambient temperature necessary for heat transfer. MED system is usually combined with mechanical vapor compression or thermal vapor compression to use a part of steam from any of the effects and increase its pressure. MED systems can operate at low TBT, thus enabling Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum… http://dx.doi.org/10.5772/intechopen.68702 151
Figure 8. Schematic of triple-effect MED [14].
them to use cheap heat sources such as waste heat or solar collectors. For example, an existing diesel waste heat-coupled MED plant's energy consumption is 2 kWh/Ton for water pumps [15].
### 3.1.3. Membrane distillation (MD)
And the minimum brine temperature is determined by temperature difference between the last effect and ambient temperature necessary for heat transfer. MED system is usually combined with mechanical vapor compression or thermal vapor compression to use a part of steam from any of the effects and increase its pressure. MED systems can operate at low TBT, thus enabling
Figure 6. Thermal desalination technologies.
150 Desalination
Figure 7. Schematic of MSF [12].
In membrane distillation process, feed water is heated and then allowed to flow through the porous hydrophobic membrane. The high pressure or electrical potential is applied on water vapor to produce fresh water from saline water [16]. The vapor pressure difference across membrane causes water vapor molecules to flow and it is condensed on the other side of the membrane. MD has the following characteristics: high porosity, hydrophobic and low thermal conductivity; the membrane thickness should be reduced and to maintain high pore size for the increase of flux.
The following are the operating parameters which effects distillate yield like flow rate, thickness of air gap, membrane thickness, porosity, long-term operation, thermal conductivity of membranes, and porosity. It is still a big challenge to develop high and efficient membrane technology, since MD has basic challenges to meet to compete with MSF, MED, and RO. In addition, MD system consumes a lot of energy [17, 18].
#### 3.2. Application of molten salts as heat transfer fluid in concentrating solar power plants
Despite the tremendous improvements in conventional desalination technology, the desalination process is very energy consuming. For instance, to produce 1000 m<sup>3</sup> /day of desalinated water in a year, nearly 10,000 tons of oil is required [16]. More detailed energy consumption data for different desalination processes are listed in Table 1.
thermal energy storages has shown that sea salts are very promising materials for large-scale thermal energy storage [28–31]. At present, salt mixtures have been considered as promising HTF candidates due to their thermal stability at high temperatures [32]. So far, alkali nitrate and nitrite mixtures are the two most successful HTF candidates. Solar salt (KNO3 40 wt%- NaNO3 60 wt%), Hitec (NaNO3 7 wt%-KNO3 53 wt%-NaNO2 40 wt%), and HitecXL (NaNO3 7 wt%-KNO3 45 wt%-Ca(NO3)2 48 wt%) are three commercialized alkali nitrate and nitrite mixtures. Other nitric salt-based HTFs include Sandia Mix (NaNO3 9–18 wt%-KNO3 40–52 wt %-LiNO3 13–21 wt.%-Ca(NO3)2 20–27 wt%) developed by Bradshaw and Brosseau [33] at the Sandia National Laboratories and SS-500 (NaNO3 6 wt%-KNO3 23 wt%-LiNO3 8 wt%-CsNO3 44 wt%-Ca(NO3)2 19 wt%) developed by Halotechnics Inc. [34]. Reddy et al. [35] developed a mixture of alkali-fluoride and carbonate salt (LiF-NaF-K2CO3) as the HTF with a working temperature range of 400–900�C. Li et al. [31, 36–38] recently developed a chloride salt mixture (NaCl 7.5 wt%-KCl 23.9 wt%-ZnCl2 68.6 wt%), its melting temperature is about 850�C, while its viscosity is only 0.325 Pa s at 300�C. The cost of this ternary chloride salt mixture is expected
Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum…
http://dx.doi.org/10.5772/intechopen.68702
153
Consequently, a full separation of salts and water in desalination process needs to be developed, the byproducts of which can be utilized as the HTF in CSP plants. Furthermore, such a full separation desalination system is expected to become a necessity in advanced water
As discussed in Section 1, the disposal of high-temperature-concentrated brine is a big concern to the environment. It has been clearly documented that the discharged high-temperature concentrates can cause a lasting change in species composition and abundance in the discharge site and affect aquatic life [21]. To avoid the potential adverse effects from disposing of the concentrates, the recovery of important minerals from concentrates to achieve zero discharge is a promising option. Concentrates from desalting process contain inorganic salts and other compounds that may be purified for commercial applications. Section 3.2 introduced a ternary halide salts as a potential candidate of thermal storage material; this may dramatically change the technology for desalination and water treatment. Therefore, a full separation of salts and water for desalination becomes necessary in new water technologies. A proposed full separation multieffect distillation (FSMED) desalination system can accommodate the demands of using high-temperature thermal energy to effectively separate salts and water at 100%,
To achieve 100% water extraction, an innovative concept of FSMED system using solar thermal energy is proposed. In the proposed FSMED desalination system, a full separation tank (FST)
4. Heat and mass transfer analysis of a novel full separation system using solar thermal energy from concentrated solar power (CSP)
to be below 1 \$/kg.
4.1. Motivations
4.2. Concepts
treatment technologies in an economically feasible way.
resulting in simultaneous collection of pure water and dry salts.
Table 1. Energy consumption of seawater desalination process [19].
Due to high expenses of the conventional energy resources, renewable energy sources can provide alternatives. Since most desalination systems can directly use thermal energy, concentrated solar thermal energy is very suitable for water treatment. Furthermore, the potential negative environmental impact from the desalination process is the discharge of high concentrated brine to the environment [20]. To avoid the potential negative impacts from disposing of the concentrates, the recovery of important minerals from concentrates to achieve zero discharge is a promising option. Recent development on concentrating solar thermal power generation has shown that the sea salts are very promising materials for large-scale thermal energy storage in concentrating solar power (CSP) plants. This may dramatically change the technology for desalination and water treatment because the salts may be collected for better value, rather than disposed.
A typical CSP plant consists of a solar collection system and a traditional power block. Solar energy is concentrated in the collection system, and the heat from direct normal insulation (DNI) is collected by a heat transfer fluid (HTF). The HTF circulates to the power block and transfers the collected heat to generate steam. The steam drives a turbine to produce electricity in a steam cycle [21–23]. Four most commonly used solar-concentrating methods are parabolic trough collectors, linear Fresnel reflectors, solar power tower collectors, and parabolic dish collectors. Parabolic trough and solar power tower are the two solar-concentrating methods mostly used in current CSP plants. HTF is the most important part in CSP plant beside energy storage media, because it can transfer the heat from the solar field to the power block. Currently, several typical HTFs are used in commercialized CSP plants, including air, water/steam, synthetic oils, organics, and molten salts [24, 25]. A good candidate of HTF in CSP should have the following characteristics: low freezing point, good thermal stability, low viscosity, high specific heat capacity, as well as acceptably low chemical corrosion rate, while low cost is another key criterion for industrial applications [26, 27].
Concentrates from desalting process contain inorganic salts and other compounds that may be purified for commercial value. For example, the recent technology development on solar thermal energy storages has shown that sea salts are very promising materials for large-scale thermal energy storage [28–31]. At present, salt mixtures have been considered as promising HTF candidates due to their thermal stability at high temperatures [32]. So far, alkali nitrate and nitrite mixtures are the two most successful HTF candidates. Solar salt (KNO3 40 wt%- NaNO3 60 wt%), Hitec (NaNO3 7 wt%-KNO3 53 wt%-NaNO2 40 wt%), and HitecXL (NaNO3 7 wt%-KNO3 45 wt%-Ca(NO3)2 48 wt%) are three commercialized alkali nitrate and nitrite mixtures. Other nitric salt-based HTFs include Sandia Mix (NaNO3 9–18 wt%-KNO3 40–52 wt %-LiNO3 13–21 wt.%-Ca(NO3)2 20–27 wt%) developed by Bradshaw and Brosseau [33] at the Sandia National Laboratories and SS-500 (NaNO3 6 wt%-KNO3 23 wt%-LiNO3 8 wt%-CsNO3 44 wt%-Ca(NO3)2 19 wt%) developed by Halotechnics Inc. [34]. Reddy et al. [35] developed a mixture of alkali-fluoride and carbonate salt (LiF-NaF-K2CO3) as the HTF with a working temperature range of 400–900�C. Li et al. [31, 36–38] recently developed a chloride salt mixture (NaCl 7.5 wt%-KCl 23.9 wt%-ZnCl2 68.6 wt%), its melting temperature is about 850�C, while its viscosity is only 0.325 Pa s at 300�C. The cost of this ternary chloride salt mixture is expected to be below 1 \$/kg.
Consequently, a full separation of salts and water in desalination process needs to be developed, the byproducts of which can be utilized as the HTF in CSP plants. Furthermore, such a full separation desalination system is expected to become a necessity in advanced water treatment technologies in an economically feasible way.
## 4. Heat and mass transfer analysis of a novel full separation system using solar thermal energy from concentrated solar power (CSP)
### 4.1. Motivations
Due to high expenses of the conventional energy resources, renewable energy sources can provide alternatives. Since most desalination systems can directly use thermal energy, concentrated solar thermal energy is very suitable for water treatment. Furthermore, the potential negative environmental impact from the desalination process is the discharge of high concentrated brine to the environment [20]. To avoid the potential negative impacts from disposing of the concentrates, the recovery of important minerals from concentrates to achieve zero discharge is a promising option. Recent development on concentrating solar thermal power generation has shown that the sea salts are very promising materials for large-scale thermal energy storage in concentrating solar power (CSP) plants. This may dramatically change the technology for desalination and water treatment because the salts may be collected for better
90 252–327 2.5–4 70 462–567 2.5–4
70 235–294 1.2–1.8 60 294–394 1.2–1.8
Thermal energy (kJ/kg) Electric power (kWh/m<sup>3</sup>
)
Desalination process Max seawater temperature (�C) Energy consumption
MSF 120 184–222 2.5–4
MED 80 176–231 1.2–1.8
RO (without energy recover) – – 7–8 RO (with energy recovery) – – 5–6
Table 1. Energy consumption of seawater desalination process [19].
A typical CSP plant consists of a solar collection system and a traditional power block. Solar energy is concentrated in the collection system, and the heat from direct normal insulation (DNI) is collected by a heat transfer fluid (HTF). The HTF circulates to the power block and transfers the collected heat to generate steam. The steam drives a turbine to produce electricity in a steam cycle [21–23]. Four most commonly used solar-concentrating methods are parabolic trough collectors, linear Fresnel reflectors, solar power tower collectors, and parabolic dish collectors. Parabolic trough and solar power tower are the two solar-concentrating methods mostly used in current CSP plants. HTF is the most important part in CSP plant beside energy storage media, because it can transfer the heat from the solar field to the power block. Currently, several typical HTFs are used in commercialized CSP plants, including air, water/steam, synthetic oils, organics, and molten salts [24, 25]. A good candidate of HTF in CSP should have the following characteristics: low freezing point, good thermal stability, low viscosity, high specific heat capacity, as well as acceptably low chemical corrosion rate, while low cost is
Concentrates from desalting process contain inorganic salts and other compounds that may be purified for commercial value. For example, the recent technology development on solar
value, rather than disposed.
152 Desalination
another key criterion for industrial applications [26, 27].
As discussed in Section 1, the disposal of high-temperature-concentrated brine is a big concern to the environment. It has been clearly documented that the discharged high-temperature concentrates can cause a lasting change in species composition and abundance in the discharge site and affect aquatic life [21]. To avoid the potential adverse effects from disposing of the concentrates, the recovery of important minerals from concentrates to achieve zero discharge is a promising option. Concentrates from desalting process contain inorganic salts and other compounds that may be purified for commercial applications. Section 3.2 introduced a ternary halide salts as a potential candidate of thermal storage material; this may dramatically change the technology for desalination and water treatment. Therefore, a full separation of salts and water for desalination becomes necessary in new water technologies. A proposed full separation multieffect distillation (FSMED) desalination system can accommodate the demands of using high-temperature thermal energy to effectively separate salts and water at 100%, resulting in simultaneous collection of pure water and dry salts.
#### 4.2. Concepts
To achieve 100% water extraction, an innovative concept of FSMED system using solar thermal energy is proposed. In the proposed FSMED desalination system, a full separation tank (FST) is integrated to a conventional forward feed MED water treatment system to enhance the water productivity and thermal efficiency, as shown in Figure 9. The main advantage of forward feed system is to be able to operate at high top brine temperature. Although the mean temperature of heat addition in forward feed is lower as compared to backward feed, it is better from the operational point of view. In backward feed arrangement, the highest brine temperature occurs at the highest brine salinity, thereby causing severe scaling problems [39].
The heat source for the FST is from concentrated solar thermal system. The air feed to the FST is preheated by the exhaust air/steam mixture to achieve heat recovery. The FST receives concentrated brine from the last effect of the MED. The concentrated brine from the MED is atomized into tiny droplets and fully evaporated in the FST due to the effective convective heat transfer between water droplets and hot airflow. The brine vaporizes into steam leaving behind salt down to the bottom of the FST, resulting in the simultaneous collection of pure water and dry salts. The hot air/steam mixture leaves from the top of the FST and passes through the first effect of MED to heat the feed seawater. There is no need to retain the brine for an internal convective heat transfer process, which eliminates the scaling problem. The only section subject to a hot environment is the pipe introducing brine to the nozzles and sprayers, which can be insulated and maintained at low temperature. Therefore, this novel system can operate effectively at either very high temperatures using primary heat or at medium to low temperatures using exhaust heat.
#### 4.3. Heat and mass transfer analysis
The proposed process is a closed-loop system which increases the nonlinearity and complexity of resulting system of equations, unlike conventional systems which are open loop. The mass and energy balance analysis for the FSMED system in steady-state operation will be introduced in this section. Figure 10 shows the energy balance of the FSMED system.
According to Figure 10, the input energy involves the energy from the heat source and the feed seawater, as shown in Eq. (1)
$$E\_{\rm in} = m\_{\rm air} \cdot \mathbb{C}\_{p,\rm air} \cdot (T\_{\rm hs} - T\_{n,\rm out}) + \left[\beta \cdot m\_{\rm s,n} (h\_{\rm gs,hs} - h\_{\rm gs,n})\right] + m\_{\rm csw} \cdot h\_{\rm sw,amb} \tag{1}$$
<sup>E</sup>out <sup>¼</sup> <sup>1</sup> � <sup>β</sup> � � � ms,n � hf s,n <sup>þ</sup>X<sup>n</sup>�<sup>1</sup>
Figure 10. Energy balance of the FSMED system [40].
Figure 9. Concept of the FSMED desalination system.
the feed seawater heater.
i¼1
ms,i � hf s,i þ msalt � Cp, salt � ð Þþ TFST � Tamb ð Þ� mcsw � msw hfsw;WH ð2Þ
where the subscript WH means the thermal properties are obtained under the temperature of
Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum…
http://dx.doi.org/10.5772/intechopen.68702
155
The mass flow into the system is the cooling seawater, and the total mass flow out involves the rejected seawater, collected salts, and the pure water. The mass flow rates, salinity of different streams, and total energy consumption under different temperature of hot air going into the
where mair is the mass flow rate of the dry air and mcsw is the mass flow rate of the cooling seawater. The subscript n denotes the number of the effects, then Tn,out and hgs,n represent the parameters in the last effect. The subscript hs means the parameters are obtained under the temperature of the heat source. β is the bleed steam fraction, which indicates the ratio of mass flow of bleed steam to the mass flow of steam from which it is extracted. To reduce the thermal energy consumption, a small fraction of steam (steam 21 in Figure 9) from the last effect is bled and mixed with the air steam (steam 13 in Figure 9), and then this mixture is heated up and blown into the FST. The bleed fraction is a very important parameter which influences other crucial parameter like salinity of brine and energy requirement.
The output energy is carried out by the saturated water from each effect, collected salts and the seawater rejected from condenser, as listed in Eq. (2)
Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum… http://dx.doi.org/10.5772/intechopen.68702 155
Figure 9. Concept of the FSMED desalination system.
is integrated to a conventional forward feed MED water treatment system to enhance the water productivity and thermal efficiency, as shown in Figure 9. The main advantage of forward feed system is to be able to operate at high top brine temperature. Although the mean temperature of heat addition in forward feed is lower as compared to backward feed, it is better from the operational point of view. In backward feed arrangement, the highest brine temperature occurs
The heat source for the FST is from concentrated solar thermal system. The air feed to the FST is preheated by the exhaust air/steam mixture to achieve heat recovery. The FST receives concentrated brine from the last effect of the MED. The concentrated brine from the MED is atomized into tiny droplets and fully evaporated in the FST due to the effective convective heat transfer between water droplets and hot airflow. The brine vaporizes into steam leaving behind salt down to the bottom of the FST, resulting in the simultaneous collection of pure water and dry salts. The hot air/steam mixture leaves from the top of the FST and passes through the first effect of MED to heat the feed seawater. There is no need to retain the brine for an internal convective heat transfer process, which eliminates the scaling problem. The only section subject to a hot environment is the pipe introducing brine to the nozzles and sprayers, which can be insulated and maintained at low temperature. Therefore, this novel system can operate effectively at either very high temperatures using primary heat or at medium to low
The proposed process is a closed-loop system which increases the nonlinearity and complexity of resulting system of equations, unlike conventional systems which are open loop. The mass and energy balance analysis for the FSMED system in steady-state operation will be intro-
According to Figure 10, the input energy involves the energy from the heat source and the feed
where mair is the mass flow rate of the dry air and mcsw is the mass flow rate of the cooling seawater. The subscript n denotes the number of the effects, then Tn,out and hgs,n represent the parameters in the last effect. The subscript hs means the parameters are obtained under the temperature of the heat source. β is the bleed steam fraction, which indicates the ratio of mass flow of bleed steam to the mass flow of steam from which it is extracted. To reduce the thermal energy consumption, a small fraction of steam (steam 21 in Figure 9) from the last effect is bled and mixed with the air steam (steam 13 in Figure 9), and then this mixture is heated up and blown into the FST. The bleed fraction is a very important parameter which influences other
The output energy is carried out by the saturated water from each effect, collected salts and the
<sup>þ</sup> <sup>m</sup>csw � <sup>h</sup>sw;amb <sup>ð</sup>1<sup>Þ</sup>
duced in this section. Figure 10 shows the energy balance of the FSMED system.
Ein ¼ mair � Cp, air � ð Þþ Ths � Tn,out β � ms,n hgs,hs � hgs,n
crucial parameter like salinity of brine and energy requirement.
seawater rejected from condenser, as listed in Eq. (2)
at the highest brine salinity, thereby causing severe scaling problems [39].
temperatures using exhaust heat.
154 Desalination
seawater, as shown in Eq. (1)
4.3. Heat and mass transfer analysis
Figure 10. Energy balance of the FSMED system [40].
$$E\_{\rm out} = \left(1 - \beta\right) \cdot m\_{\rm s,n} \cdot h\_{\rm fs,n} + \sum\_{i=1}^{n-1} m\_{\rm s,i} \cdot h\_{\rm fs,i} + m\_{\rm sall} \cdot \mathbb{C}\_{p,\rm sall} \cdot \left(T\_{\rm FST} - T\_{\rm amb}\right) + \left(m\_{\rm csw} - m\_{\rm sw}\right) \cdot h\_{\rm fsw,WH} \tag{2}$$
where the subscript WH means the thermal properties are obtained under the temperature of the feed seawater heater.
The mass flow into the system is the cooling seawater, and the total mass flow out involves the rejected seawater, collected salts, and the pure water. The mass flow rates, salinity of different streams, and total energy consumption under different temperature of hot air going into the FST and varying bleed steam fraction can be calculated through detailed energy and mass balance analysis of each component of the FSMED system, which is not shown here for the sake of brevity.
An example of the calculated energy consumption for FSMED system with 3, 4, and 5 as the total number of effects is shown in Figure 11, in which the heat source temperature varies from 150 to 400�C. The bleed steam fractions for 3, 4, and 5 effects are 0.6, 0.57, and 0.55, respectively. It is apparent that a five-effect system is the least energy consuming. The minimum energy consumption for five-effect forward feed FSMED desalination system is estimated to be about 350 kJ/kg of seawater, which compares very well to thermal energy requirements of existing MED plant which require 394 kJ/kg of seawater [19]. More detailed operating data for the five-effect FSMED system are listed in Table 2.
study mainly focuses on the evaporation of a single water droplet in hot air to make a
Table 2. Calculated results for a FSMED system with five effects under different heat source temperature [40].
Heat source temperature (�C) 150 200 250 300 350 400 Mass flow rate of dry air (kg/s) 3.844 2.291 1.600 1.209 0.957 0.781 Mass flow rate of seawater (kg/s) 6.160 5.716 5.515 5.401 5.327 5.276 Mass flow rate of pure water (kg/s) 0.965 0.965 0.965 0.965 0.965 0.965 Energy input by heat source (kWh/1000 Gal of seawater) 407.6 390.4 382.5 378.1 375.2 373.2
Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum…
A simplified non-equilibrium vaporization model is adopted to describe the movement and evaporation behavior of a single water droplet in the FST. The model is described based on the
The droplet size change is related to the heat/mass transfer that determines the evaporation of the droplet. The conservation equations and mass transfer equation are described in the
> 3 πr 3 sρl Cp,l dTs
Bm <sup>¼</sup> <sup>Y</sup>Fs � YF<sup>∞</sup> 1 � YFs
¼ �G ð3Þ
http://dx.doi.org/10.5772/intechopen.68702
157
G ¼ 2πrsρD � Sh � Lnð1 þ BmÞ ð5Þ
dt <sup>þ</sup> GL <sup>ð</sup>4<sup>Þ</sup>
ð6Þ
d dt 4 3 πr 3 sρl
<sup>2</sup>πrpλð Þ <sup>T</sup><sup>∞</sup> � Ts Nu <sup>¼</sup> <sup>4</sup>
The droplet velocity and position are evaluated using the following equations:
where Bm is the Spalding mass transfer number, and is given by
preliminary assessment.
following assumptions:
Mass conservation equation:
Energy conservation equation:
Mass transfer equation:
following.
1. No temperature gradient in the droplet;
2. Thermal properties are uniform in the droplet;
Parameters Value
4. Vapor mass fraction in the hot airflow is 0.
3. Droplet maintains a spherical shape during falling;
#### 4.4. Preliminary assessment of water droplet evaporation path in FST
To achieve 100% water extraction, detailed knowledge of the water droplet behaviors in the FST is essential to the design optimization of the FSMED desalination system. It is thus necessary to determine the lifetime and trajectory of tiny water drops as a function of the drop size, the ambient temperature, and the spray-injection parameters. Given that the analysis of vaporization of saline droplet swarm is complex (involving crystallization process), the present
Figure 11. Thermal energy requirement of FSMED desalination system versus hot air and heating steam temperature.
Table 2. Calculated results for a FSMED system with five effects under different heat source temperature [40].
study mainly focuses on the evaporation of a single water droplet in hot air to make a preliminary assessment.
A simplified non-equilibrium vaporization model is adopted to describe the movement and evaporation behavior of a single water droplet in the FST. The model is described based on the following assumptions:
The droplet size change is related to the heat/mass transfer that determines the evaporation of the droplet. The conservation equations and mass transfer equation are described in the following.
Mass conservation equation:
FST and varying bleed steam fraction can be calculated through detailed energy and mass balance analysis of each component of the FSMED system, which is not shown here for the
An example of the calculated energy consumption for FSMED system with 3, 4, and 5 as the total number of effects is shown in Figure 11, in which the heat source temperature varies from 150 to 400�C. The bleed steam fractions for 3, 4, and 5 effects are 0.6, 0.57, and 0.55, respectively. It is apparent that a five-effect system is the least energy consuming. The minimum energy consumption for five-effect forward feed FSMED desalination system is estimated to be about 350 kJ/kg of seawater, which compares very well to thermal energy requirements of existing MED plant which require 394 kJ/kg of seawater [19]. More detailed operating data for
To achieve 100% water extraction, detailed knowledge of the water droplet behaviors in the FST is essential to the design optimization of the FSMED desalination system. It is thus necessary to determine the lifetime and trajectory of tiny water drops as a function of the drop size, the ambient temperature, and the spray-injection parameters. Given that the analysis of vaporization of saline droplet swarm is complex (involving crystallization process), the present
Figure 11. Thermal energy requirement of FSMED desalination system versus hot air and heating steam temperature.
sake of brevity.
156 Desalination
the five-effect FSMED system are listed in Table 2.
4.4. Preliminary assessment of water droplet evaporation path in FST
$$\frac{d}{dt}\left(\frac{4}{3}\pi r\_s^3 \rho\_l\right) = -G\tag{3}$$
Energy conservation equation:
$$2\pi r\_p \overline{\lambda} (T\_\simeq - T\_s) \text{Nu} = \frac{4}{3} \pi r\_s^3 \rho\_l \mathbb{C}\_{p,l} \frac{dT\_s}{dt} + \text{GL} \tag{4}$$
Mass transfer equation:
$$G = 2\pi r\_s \overline{\rho} \overline{D} \cdot \text{Sh} \cdot \text{Ln} (1 + B\_m) \tag{5}$$
where Bm is the Spalding mass transfer number, and is given by
$$B\_m = \frac{Y\_{\rm fs} - Y\_{\rm F\circ}}{1 - Y\_{\rm F\circ}} \tag{6}$$
The droplet velocity and position are evaluated using the following equations:
$$-\frac{du\_x}{dt} = \frac{\Im \mathbf{C}\_D}{8r\_s} \left(\frac{\overline{\rho}}{\rho\_l}\right) u\_r \cdot u\_x \tag{7}$$
velocity overshadows that of the lifetime decrease, so the falling distance increased slightly with the initial injection velocity. For the cases of θ ¼ 20� and 30�, because the increase of vertical component of the injection velocity always has greater influence than the lifetime decrease, the falling distance increases with the initial injection velocity. The stopping distance decreases with the injection angle due to the decrease of the horizontal component of the initial
Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum…
http://dx.doi.org/10.5772/intechopen.68702
159
Figure 13. Variation of the falling and stopping distance with the injection velocity and angle.
Desalination is a promising solution for providing the increasing need of freshwater, due to the increasing growth of population and climate change. Despite the tremendous improvements in conventional water treatment technology, the desalination process is very energy consuming. Therefore, it is highly desirable that environmentally friendly renewable energy resources should be utilized for water treatment. Since most desalination systems can directly use thermal energy, concentrated solar thermal energy is very suitable for water treatment. Furthermore, recent developments on CSP have shown that the sea salts are very promising materials as HTF in thermal energy storage system, which in fact provides an alternative to desalination and water treatment by collecting the salts through full separation from seawater
To achieve 100% water extraction, an innovative concept of FSMED desalination system using solar thermal energy was proposed in the present study. Compared with conventional MED, the FSMED technology will overcome the limitation of operating temperature for water desalination. The concentrated brine from the MED is atomized into tiny droplets and fully evaporated in the FST by using the high-temperature heat from concentrated solar thermal energy system, resulting in simultaneous collection of pure water and dry salts. A simplified non-equilibrium vaporization model is used to investigate the lifetime and trajectory of a tiny water droplet in the FST as a function of the droplet size, the ambient temperature, and the
and brackish water, instead of disposing the high concentrate to the environment.
injection velocity.
5. Concluding remarks
$$-\frac{du\_y}{dt} = \frac{3C\_D}{8r\_s} \left(\frac{\overline{\rho}}{\rho\_l}\right) u\_r \cdot \left(u\_y - u\_{\approx}\right) - g\frac{\rho\_l - \overline{\rho}}{\rho\_l} \tag{8}$$
The lifetime of the water droplet in the FST under different diameters and temperatures is compared in Figure 12. The droplet lifetime increases with an increase in the droplet diameter but decreases with an increase in the ambient temperate. In addition, the lifetime decreases with the ambient temperature increase in a nonlinear way, with a gradual decrease in the slope.
To confine the vaporization within the FST with certain size, the variations of the falling distance in the vertical direction and stopping distance in the horizontal direction were calculated for a water droplet with 400 μm in diameter under different injection parameters, as shown in Figure 13.
For the case of θ ¼ 0�, the falling distance decreases with the increase in the initial injection velocity. This is because the lifetime of the droplet decreases with the increase of the injection velocity. For the case of θ ¼ 10�, when the injection velocity is below 15 m/s, the influence of the lifetime decrease on the falling distance is larger than that of vertical component of the injection velocity. Hence, the falling distance decreased slightly with the initial injection velocity. When the injection velocity exceeds 15 m/s, the effect of vertical component of the injection
Figure 12. Comparison of the lifetime under different droplet diameters and ambient temperatures.
Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum… http://dx.doi.org/10.5772/intechopen.68702 159
Figure 13. Variation of the falling and stopping distance with the injection velocity and angle.
velocity overshadows that of the lifetime decrease, so the falling distance increased slightly with the initial injection velocity. For the cases of θ ¼ 20� and 30�, because the increase of vertical component of the injection velocity always has greater influence than the lifetime decrease, the falling distance increases with the initial injection velocity. The stopping distance decreases with the injection angle due to the decrease of the horizontal component of the initial injection velocity.
#### 5. Concluding remarks
� dux
� duy
shown in Figure 13.
158 Desalination
dt <sup>¼</sup> <sup>3</sup>CD 8rs
dt <sup>¼</sup> <sup>3</sup>CD 8rs
> ρ ρl
ρ ρl
ur � uy � u<sup>∞</sup> � <sup>g</sup>
The lifetime of the water droplet in the FST under different diameters and temperatures is compared in Figure 12. The droplet lifetime increases with an increase in the droplet diameter but decreases with an increase in the ambient temperate. In addition, the lifetime decreases with the ambient temperature increase in a nonlinear way, with a gradual decrease in the slope.
To confine the vaporization within the FST with certain size, the variations of the falling distance in the vertical direction and stopping distance in the horizontal direction were calculated for a water droplet with 400 μm in diameter under different injection parameters, as
For the case of θ ¼ 0�, the falling distance decreases with the increase in the initial injection velocity. This is because the lifetime of the droplet decreases with the increase of the injection velocity. For the case of θ ¼ 10�, when the injection velocity is below 15 m/s, the influence of the lifetime decrease on the falling distance is larger than that of vertical component of the injection velocity. Hence, the falling distance decreased slightly with the initial injection velocity. When the injection velocity exceeds 15 m/s, the effect of vertical component of the injection
Figure 12. Comparison of the lifetime under different droplet diameters and ambient temperatures.
ur � ux ð7Þ
ð8Þ
ρ<sup>l</sup> � ρ ρl
> Desalination is a promising solution for providing the increasing need of freshwater, due to the increasing growth of population and climate change. Despite the tremendous improvements in conventional water treatment technology, the desalination process is very energy consuming. Therefore, it is highly desirable that environmentally friendly renewable energy resources should be utilized for water treatment. Since most desalination systems can directly use thermal energy, concentrated solar thermal energy is very suitable for water treatment. Furthermore, recent developments on CSP have shown that the sea salts are very promising materials as HTF in thermal energy storage system, which in fact provides an alternative to desalination and water treatment by collecting the salts through full separation from seawater and brackish water, instead of disposing the high concentrate to the environment.
> To achieve 100% water extraction, an innovative concept of FSMED desalination system using solar thermal energy was proposed in the present study. Compared with conventional MED, the FSMED technology will overcome the limitation of operating temperature for water desalination. The concentrated brine from the MED is atomized into tiny droplets and fully evaporated in the FST by using the high-temperature heat from concentrated solar thermal energy system, resulting in simultaneous collection of pure water and dry salts. A simplified non-equilibrium vaporization model is used to investigate the lifetime and trajectory of a tiny water droplet in the FST as a function of the droplet size, the ambient temperature, and the
spray-injection parameters. The relationship of the water droplet size and falling distance with the hot air-steam temperature, and initial injection/spray parameters is investigated and presented. Results from the study provide important guidance to the design of such a water treatment system. It is expected that the proposed full separation thermal-driven desalination process can address problems of concentrated discharge disposal, solute recovery, and highenergy consumption.
[6] Goosen MA, Mahmoudi H, Ghaffour N. Today's and future challenges in applications of renewable energy technologies for desalination. Critical Reviews in Environmental
Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum…
http://dx.doi.org/10.5772/intechopen.68702
161
[7] Mickley MC. Membrane Concentrate Disposal: Practices and Regulation. Desalination and Water Purification Research Program Report, US Bureau of Reclamation. 2001 [8] Available from: http://resource2.rockyview.ab.ca/science8/Sci8%20U1%20MixFlow/Sci8U1
[9] Available from: http://www.arabwatersource.com/water-blog/category/desalination, July 9,
[10] Van der Bruggen B, Vandecasteele C. Distillation vs. membrane filtration: Overview of process evolutions in seawater desalination. Desalination. 2002;143(3):207–218
[11] Borsani R, Rebagliati S. Fundamentals and costing of MSF desalination plants and
[12] Multiple Stage Flash. Available from: http://www.sidem-desalination.com/en/Process/
[13] Fiorini P, Sciubba E. Thermoeconomic analysis of a MSF desalination plant. Desalination.
[14] Multiple Effect Distillation. Available from: http://www.sidem-desalination.com/en/Pro-
[15] Ophir A, Lokiec F. Advanced MED process for most economical sea water desalination.
[16] Sharon H, Reddy KS. A review of solar energy driven desalination technologies. Renew-
[17] Kesieme UK, Milne N, Aral H, Cheng CY, Duke M. Economic analysis of desalination technologies in the context of carbon pricing, and opportunities for membrane distilla-
[18] Ali MT, Fath HE, Armstrong PR. A comprehensive techno-economical review of indirect solar desalination. Renewable and Sustainable Energy Reviews. 2011;15(8):4187–4199 [19] Deng RY, Xie L, Lin H, Liu J, Han W. Integration of thermal energy and seawater
[20] Laspidou C, Hadjibiros K, Gialis S. Minimizing the environmental impact of sea brine disposal by coupling desalination plants with solar saltworks: A case study for Greece.
[21] Zhang HL, Baeyens J, Degrève J, Cacères G. Concentrated solar power plants: Review and design methodology. Renewable and Sustainable Energy Reviews. 2013;22:466–481
[22] Barlev D, Vidu R, Stroeve P. Innovation in concentrated solar power. Solar Energy Mate-
comparison with other technologies. Desalination. 2005;182(1):29–37
Science and Technology. 2014;44:929–999
%20Lessons/Sci8U1L4.html
MSF/. [Accessed: March 12, 2016]
cess/MED/. [Accessed March 12, 2016]
able and Sustainable Energy Reviews. 2015;41:1080–1118
Desalination. 2005;182(1):187–198
tion. Desalination. 2013;323:66–74
Water. 2010;2(1):75–84
desalination. Energy. 2010;35:4368–4374
rials and Solar Cells. 2011;95:2703–2725
2005;182(1):39–51
2014
## Acknowledgements
The authors are grateful to the support of the New Faculty Supporting Program (NFSP) from the University of Texas Rio Grande Valley and the University of Arizona under its program of Water, Environmental and Energy Solutions (WEES).
## Author details
Ben Xu<sup>1</sup> , Peiwen Li<sup>2</sup> \* and Penghua Guo<sup>3</sup>
\*Address all correspondence to: peiwen@email.arizona.edu
1 Department of Mechanical Engineering, The University of Texas Rio Grande Valley, Edinburg, TX, USA
2 Department of Aerospace and Mechanical Engineering, The University of Arizona, Tucson, AZ, USA
3 School of Energy and Power Engineering, Xi'an Jiaotong University, Xi'an, Shanxi, China
## References
[6] Goosen MA, Mahmoudi H, Ghaffour N. Today's and future challenges in applications of renewable energy technologies for desalination. Critical Reviews in Environmental Science and Technology. 2014;44:929–999
spray-injection parameters. The relationship of the water droplet size and falling distance with the hot air-steam temperature, and initial injection/spray parameters is investigated and presented. Results from the study provide important guidance to the design of such a water treatment system. It is expected that the proposed full separation thermal-driven desalination process can address problems of concentrated discharge disposal, solute recovery, and high-
The authors are grateful to the support of the New Faculty Supporting Program (NFSP) from the University of Texas Rio Grande Valley and the University of Arizona under its program of
1 Department of Mechanical Engineering, The University of Texas Rio Grande Valley,
2 Department of Aerospace and Mechanical Engineering, The University of Arizona, Tucson,
3 School of Energy and Power Engineering, Xi'an Jiaotong University, Xi'an, Shanxi, China
[1] Elimelech M, Phillip WA. The future of seawater desalination: Energy, technology, and
[2] Liyanaarachchi S, Shu L, Muthukumaran S, Jegatheesan V, Baskaran K. Problems in seawater industrial desalination processes and potential sustainable solutions: A review.
[3] IDA. Available from: http://idadesal.org/desalination-101/desalination-by-the-numbers.
[5] WHO. Desalination for safe water supply: Guidance for the health and environmental
Reviews in Environmental Science and Bio/Technology. 2014;13:203–214
aspects applicable to desalination. In: World Health Organization. 2007
energy consumption.
160 Desalination
Author details
Edinburg, TX, USA
, Peiwen Li<sup>2</sup>
Ben Xu<sup>1</sup>
AZ, USA
References
Acknowledgements
Water, Environmental and Energy Solutions (WEES).
\* and Penghua Guo<sup>3</sup>
\*Address all correspondence to: peiwen@email.arizona.edu
the environment. Science. 2011;333:712–717
June 30, 2015. [Accessed: July 12, 2016]
[4] IDA. 2014. The IDA Desalination Yearbook 2014–2015
[23] Vignarooban K, Xu X, Arvay A, Hsu K, Kannan AM. Heat transfer fluids for concentrating solar power systems—a review. Applied Energy. 2015;146:383–396
[36] Li P, Molina E, Wang K, Xu X, Dehghani G, Kohli A, Hao Q, Teja AS. Thermal and transport properties of NaCl-KCl-ZnCl2 eutectic salts for new generation high tempera-
Solar Thermal‐Driven Desalination Pursuing Products of Pure Water and Salts and Leaving Minimum…
http://dx.doi.org/10.5772/intechopen.68702
163
[37] Vignarooban K, Xu X, Wang K, Molina EE, Li P, Gervasio D, Kannan AM. Vapor pressure and corrosivity of ternary metal-chloride molten-salt based heat transfer fluids for use in
[38] Xu B, Li PW, Chan C. Application of phase change materials for thermal energy storage in concentrated solar thermal power plants: A review to recent developments. Applied
[39] Al-Mutaz IS, Wazeer I. Development of a steady-state mathematical model for MEE-TVC
[40] Peri A. Development and analysis of a novel thermal driven water-solute separation
ture heat transfer fluids. Journal of Solar Energy Engineering. 2016;138(5):054501
concentrating solar power systems. Applied Energy. 2015;159:206–213
Energy. 2015;160:286–307
desalination plants. Desalination. 2014;351:9–18
process [Master thesis]. University of Arizona; 2015
[36] Li P, Molina E, Wang K, Xu X, Dehghani G, Kohli A, Hao Q, Teja AS. Thermal and transport properties of NaCl-KCl-ZnCl2 eutectic salts for new generation high temperature heat transfer fluids. Journal of Solar Energy Engineering. 2016;138(5):054501
[23] Vignarooban K, Xu X, Arvay A, Hsu K, Kannan AM. Heat transfer fluids for concentrat-
[24] acio J, Wetzel T. Assessment of liquid metal technology status and research paths for their use as efficient heat transfer fluids in solar central receiver systems. Solar Energy.
[25] Tian Y, Zhao CY. A review of solar collectors and thermal energy storage in solar thermal
[26] Cordaro JG, Rubin NC, Bradshaw RW. Multicomponent molten salt mixtures based on nitrate/nitrite anions. ASME Journal of Solar Energy Engineering. 2011;133:011014 [27] Bauer T, Laing D, Tamme R. Recent progress in alkali nitrate/nitrite developments for solar thermal power applications. Molten Salts Chemistry and Technology. Trondheim,
[28] Wang K, Molina E, Dehghani G, Xu B, Li P, Hao Q, Lucas P, Kassaee MH, Jeter SM, Teja AS. Experimental investigation to the properties of eutectic salts by NaCl-KCl-ZnCl2 for application as high temperature heat transfer fluids. In: ASME 2014 8th International Conference on Energy Sustainability, Paper ES-FuelCell 2014-6578, American Society of Mechanical
[29] Vignarooban K, Pugazhendhi P, Tucker C, Gervasio D, Kannan A. Corrosion resistance of Hastelloys in molten metal-chloride heat-transfer fluids for concentrating solar power
[30] Li PW, Chan CL, Hao Q, Deymier PA, Muralidharan K, Gervasio DF, Momayez M, Jeter SM, Teja AS, Kannan AM. Halide and oxy-halide eutectic systems for high performance high temperature heat transfer fluids. SunShot Concentrating Solar Power Program
[31] Xu B, Li P, Chan C. Application of phase change materials for thermal energy storage in concentrated solar thermal power plants: A review to recent developments. Applied
[32] Peng Q, Wei X, Ding J, Yang J, Yang X. High-temperature thermal stability of molten salt
[33] Bradshaw RW, Brosseau DA. U.S. Patent No. 7,588,694. Washington, DC: U.S. Patent and
[34] Raade JW, Padowitz D. Development of molten salt heat transfer fluid with low melting point and high thermal stability. Journal of Solar Energy Engineering. 2011;133(3):031013
[35] Wang T, Mantha D, Reddy RG. Novel high thermal stability LiF-Na2CO3-K2CO3 eutectic ternary system for thermal energy storage applications. Solar Energy Materials and Solar
materials. International Journal of Energy Research. 2008;32(12):1164–1174
ing solar power systems—a review. Applied Energy. 2015;146:383–396
applications. Applied Energy. 2013;104:538–553
Engineers; Boston, MA, June 29-July 2, 2014
applications. Solar Energy. 2014;103:62–69
2013;93:11–22
162 Desalination
Norway. June 2011
Review. 2013:85–86
Energy. 2015;160:286–307
Trademark Office; 2009
Cells. 2015;140:366–375
**Chapter 9**
**Responding to Water Challenges Through**
Desalination technology and reverse osmosis in particular, is used by several island authorities in Greece to address water scarcity. However, this is a highly energy-intensive technique, requiring the consumption of significant quantities of fossil fuels. The case of Syros island is presented to demonstrate the strong water-energy link in the operation of desalination plants. The use of renewable energy sources as a means for reducing water cost from desalination is also discussed. A simple algorithm to calculate estimating water costs with renewable energy sources (RES) is presented and is applied in the
Keywords: desalination, reverse osmosis, RES-powered desalination, hybrid energy
Economic advancement and prosperity depend on the significant resources of energy and water, the links between which are equally complex and important: energy production necessitates significant quantities of water, and water supply requires great amounts of energy. The study of the "water-energy nexus" demands a holistic view of the production and consumption chains of each resource [1] presents a view of the complexity of the water–energy nexus),
Energy requirements along the stages of the water chain depend on the quantity and quality of water and usually increase over time [3]. Energy intensity will increase with growth in water demand, as abstraction of water is carried out from deeper boreholes with lengthy interbasin
> © 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
focusing on the core operational components, i.e., infrastructure and technologies [2].
**Desalination: Energy Considerations**
George Arampatzis, Avraam Kartalidis and
Additional information is available at the end of the chapter
island of Patmos and in Hermoupolis, Syros island.
systems, water scarcity, water-energy nexus
http://dx.doi.org/10.5772/intechopen.69956
Dionysis Assimacopoulos
Abstract
1. Introduction
## **Responding to Water Challenges Through Desalination: Energy Considerations**
George Arampatzis, Avraam Kartalidis and Dionysis Assimacopoulos
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.69956
#### Abstract
Desalination technology and reverse osmosis in particular, is used by several island authorities in Greece to address water scarcity. However, this is a highly energy-intensive technique, requiring the consumption of significant quantities of fossil fuels. The case of Syros island is presented to demonstrate the strong water-energy link in the operation of desalination plants. The use of renewable energy sources as a means for reducing water cost from desalination is also discussed. A simple algorithm to calculate estimating water costs with renewable energy sources (RES) is presented and is applied in the island of Patmos and in Hermoupolis, Syros island.
Keywords: desalination, reverse osmosis, RES-powered desalination, hybrid energy systems, water scarcity, water-energy nexus
### 1. Introduction
Economic advancement and prosperity depend on the significant resources of energy and water, the links between which are equally complex and important: energy production necessitates significant quantities of water, and water supply requires great amounts of energy. The study of the "water-energy nexus" demands a holistic view of the production and consumption chains of each resource [1] presents a view of the complexity of the water–energy nexus), focusing on the core operational components, i.e., infrastructure and technologies [2].
Energy requirements along the stages of the water chain depend on the quantity and quality of water and usually increase over time [3]. Energy intensity will increase with growth in water demand, as abstraction of water is carried out from deeper boreholes with lengthy interbasin
© 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
water transfers, strict water-quality standards, and limitations on wastewater discharge. Despite all this, there are opportunities for improving energy efficiency. According to the European Innovation Partnership for water [4], there is an important potential for increasing the efficiency of energy use in water supply systems by using low-energy technologies in water treatment and wastewater cleaning processes. Also, the use of renewable resources is beneficial for the total energy efficiency of these systems.
The sources of water (in terms of type, quality, and seasonal availability) and the end-use characteristics (such as consumption patterns and infrastructure) play an important role on the energy intensity of the water supply chain. An increase of 35% in energy consumption is expected by 2035 due to the impacts of climate change on water resources [5], the increased competition among water users, and the "loop effect" of the energy-water interlinks. This might cause 85% increase in the water consumption and thus extra energy requirements are needed for water distribution and treatment [6, 7]. As these interlinkages are established, the quantification of the energy-water relationship is a priority issue, particularly in water-scarce regions.
In water-scarce areas, the main supply management methods are (1) water transports from areas with abundant water sources to areas with water scarcity; (2) intensive use of the available resources; and (3) use of nonconventional water resources (e.g., wastewater reuse, desalination). All these supply methods cause an increase in the energy consumption for water supply and as a consequence, the cost of water is higher either for consumers or governments. Thus, any improvements in the energy use and the efficiency of water supply may lead to profits for the environment as well as important reductions in the cost of water.
In this chapter, the water–energy nexus is presented with emphasis in remote areas. The island of Syros, located in the Cyclades island complex, is used as a case study to demonstrate the links between energy consumption and water production and the need to increase energy-use efficiency in the isolated water and electrical energy systems of the Greek islands. The desalination option with RES is explored through modeling, and as a case study, Hermoupolis in the island of Syros and Patmos island is selected. The rich history of all the desalination methods that have been applied in the Greek islands is also described in brief.
absent in most of the islands. These facts intensify the problems of water scarcity and create
In each island, the specific conditions shape the approach used to address water deficiency problems. The current practices for domestic water supply according to Ref. [10] are: (a) desalination of sea or brackish water, (b) water transfer from the mainland at high cost
Complex Number of islands Population Population water balance [11]
), (c) dams, and (d) boreholes. The state officials regard desalination as an
Responding to Water Challenges Through Desalination: Energy Considerations
http://dx.doi.org/10.5772/intechopen.69956
167
water deficiencies in the local areas (Table 1).
Figure 1. Map of Greek islands and complexes.
Table 1. Island complexes in Greece [8].
appropriate long-term settlement to the water-shortage problem.
Dodecanese 24 191,084 Negative Ionian 17 211,954 Positive Cyclades 28 118,000 Negative North Aegean 14 202,360 Negative Sporades 7 16,792 Negative Saronic 7 63,467 Negative
(about 10 €/m<sup>3</sup>
## 2. The water system of the Greek islands
Greece has the longest coastline in Europe, approximately 14,000 km. The country has about 2500 islands (Figure 1), with a total area of 21,580 km<sup>2</sup> . The islanders are estimated to be 1,633,433 [8].
The climate on the Greek islands is typical Mediterranean (dry summers and wet winters), with relatively low precipitation (less than 400 mm per year, especially for the Aegean islands [9]). The storage, surface, or groundwater of sufficient quality and quantity is impossible on some islands due to the low precipitation and the geological formations. Furthermore, there is a peak in water demand on the summer period (due to tourism and irrigation), the water losses are high due to the leakages in the old distribution networks, the aquifers are degraded due to over-pumping, and above all, an integrated water management system is Responding to Water Challenges Through Desalination: Energy Considerations http://dx.doi.org/10.5772/intechopen.69956 167
Figure 1. Map of Greek islands and complexes.
water transfers, strict water-quality standards, and limitations on wastewater discharge. Despite all this, there are opportunities for improving energy efficiency. According to the European Innovation Partnership for water [4], there is an important potential for increasing the efficiency of energy use in water supply systems by using low-energy technologies in water treatment and wastewater cleaning processes. Also, the use of renewable resources is beneficial
The sources of water (in terms of type, quality, and seasonal availability) and the end-use characteristics (such as consumption patterns and infrastructure) play an important role on the energy intensity of the water supply chain. An increase of 35% in energy consumption is expected by 2035 due to the impacts of climate change on water resources [5], the increased competition among water users, and the "loop effect" of the energy-water interlinks. This might cause 85% increase in the water consumption and thus extra energy requirements are needed for water distribution and treatment [6, 7]. As these interlinkages are established, the quantification
In water-scarce areas, the main supply management methods are (1) water transports from areas with abundant water sources to areas with water scarcity; (2) intensive use of the available resources; and (3) use of nonconventional water resources (e.g., wastewater reuse, desalination). All these supply methods cause an increase in the energy consumption for water supply and as a consequence, the cost of water is higher either for consumers or governments. Thus, any improvements in the energy use and the efficiency of water supply may lead to
In this chapter, the water–energy nexus is presented with emphasis in remote areas. The island of Syros, located in the Cyclades island complex, is used as a case study to demonstrate the links between energy consumption and water production and the need to increase energy-use efficiency in the isolated water and electrical energy systems of the Greek islands. The desalination option with RES is explored through modeling, and as a case study, Hermoupolis in the island of Syros and Patmos island is selected. The rich history of all the desalination methods
Greece has the longest coastline in Europe, approximately 14,000 km. The country has about
The climate on the Greek islands is typical Mediterranean (dry summers and wet winters), with relatively low precipitation (less than 400 mm per year, especially for the Aegean islands [9]). The storage, surface, or groundwater of sufficient quality and quantity is impossible on some islands due to the low precipitation and the geological formations. Furthermore, there is a peak in water demand on the summer period (due to tourism and irrigation), the water losses are high due to the leakages in the old distribution networks, the aquifers are degraded due to over-pumping, and above all, an integrated water management system is
. The islanders are estimated to be
of the energy-water relationship is a priority issue, particularly in water-scarce regions.
profits for the environment as well as important reductions in the cost of water.
that have been applied in the Greek islands is also described in brief.
2. The water system of the Greek islands
1,633,433 [8].
2500 islands (Figure 1), with a total area of 21,580 km<sup>2</sup>
for the total energy efficiency of these systems.
166 Desalination
absent in most of the islands. These facts intensify the problems of water scarcity and create water deficiencies in the local areas (Table 1).
In each island, the specific conditions shape the approach used to address water deficiency problems. The current practices for domestic water supply according to Ref. [10] are: (a) desalination of sea or brackish water, (b) water transfer from the mainland at high cost (about 10 €/m<sup>3</sup> ), (c) dams, and (d) boreholes. The state officials regard desalination as an appropriate long-term settlement to the water-shortage problem.
Table 1. Island complexes in Greece [8].
## 3. Energy-intensive water desalination
Desalination was first introduced into Greek islands during 1960s, based on solar still collectors. Five systems of various sizes were installed until 1973, without great success, mainly due to operational breakdowns, poor maintenance [12], and conflicts of interest with the water tanker owners. The success in operation of these desalination units would result in a significant decline in the quantities of water they transferred to the islands.
In 1969, a multistage-flash system was installed in Syros, powered by fuel oil. The high cost of fuel in addition to mechanical problems led to the shift toward RO [13]. The next desalination attempt took place in the island of Corfu in 1977. A reverse electrodialysis (ED) plant with a daily capacity of 15,000 m3 was installed for the treatment of low-salinity brackish water (up to 2000 ppm). The operation of this desalination plant ceased due to functional problems happened after few years [14]. Its specific energy consumption was 1.7 kWh/m<sup>3</sup> [15].
Before the Syros RO installations, the first RO units for the public community water supply were installed in 1981–1982 in the islands of Ithaca and Mykonos. The reasons for the domination of RO are: (1) compact design, (2) relatively low water needs, (3) modular operation to meet seasonal and diurnal variations in demand, (4) easy operation, (5) relatively low-energy consumption comparing to other desalination methods, and (6) fast installation (2–3 months).
Figure 3. Desalination facilities in Greece, 1981-2014 ([13, 16–22]; TEMAK SA, personal communication).
Responding to Water Challenges Through Desalination: Energy Considerations
http://dx.doi.org/10.5772/intechopen.69956
169
Figure 4. Specific energy consumption of Greek desalination plants (Syros WSSC, personal communication, 2014;
TEMAK SA, personal communication, 2014).
Currently, the total capacity of the installed desalination units (for public use in the islands) is about 60,000 m3 /day spread over 39 islands (Figure 2) with 9000 m3 /day of brackish water feed, and 51,000 m3 /day of seawater feed. Very small islands of Pserimos and Rho are not included in the map of Figure 2. Most desalination units are located on islands, which are not connected to the mainland electricity grid (31 out of 39 islands), and the rest are installed on islands that are interconnected to the main grid, through underwater cables. The rate of development of new desalination units was quite low until the early 2000s, but rose very fast prior to the Athens Olympic Games in 2004 (Figure 3). In the coming years, almost five more islands are about to install and operate desalination plants.
Figure 2. Desalination units in the Greek islands ([13, 16–22]; TEMAK SA, personal communication).
3. Energy-intensive water desalination
about 60,000 m3
168 Desalination
feed, and 51,000 m3
cant decline in the quantities of water they transferred to the islands.
islands are about to install and operate desalination plants.
Desalination was first introduced into Greek islands during 1960s, based on solar still collectors. Five systems of various sizes were installed until 1973, without great success, mainly due to operational breakdowns, poor maintenance [12], and conflicts of interest with the water tanker owners. The success in operation of these desalination units would result in a signifi-
In 1969, a multistage-flash system was installed in Syros, powered by fuel oil. The high cost of fuel in addition to mechanical problems led to the shift toward RO [13]. The next desalination attempt took place in the island of Corfu in 1977. A reverse electrodialysis (ED) plant with a daily capacity of 15,000 m3 was installed for the treatment of low-salinity brackish water (up to 2000 ppm). The operation of this desalination plant ceased due to functional problems hap-
Before the Syros RO installations, the first RO units for the public community water supply were installed in 1981–1982 in the islands of Ithaca and Mykonos. The reasons for the domination of RO are: (1) compact design, (2) relatively low water needs, (3) modular operation to meet seasonal and diurnal variations in demand, (4) easy operation, (5) relatively low-energy consumption comparing to other desalination methods, and (6) fast installation (2–3 months). Currently, the total capacity of the installed desalination units (for public use in the islands) is
/day spread over 39 islands (Figure 2) with 9000 m3
Figure 2. Desalination units in the Greek islands ([13, 16–22]; TEMAK SA, personal communication).
included in the map of Figure 2. Most desalination units are located on islands, which are not connected to the mainland electricity grid (31 out of 39 islands), and the rest are installed on islands that are interconnected to the main grid, through underwater cables. The rate of development of new desalination units was quite low until the early 2000s, but rose very fast prior to the Athens Olympic Games in 2004 (Figure 3). In the coming years, almost five more
/day of seawater feed. Very small islands of Pserimos and Rho are not
/day of brackish water
pened after few years [14]. Its specific energy consumption was 1.7 kWh/m<sup>3</sup> [15].
Figure 3. Desalination facilities in Greece, 1981-2014 ([13, 16–22]; TEMAK SA, personal communication).
Figure 4. Specific energy consumption of Greek desalination plants (Syros WSSC, personal communication, 2014; TEMAK SA, personal communication, 2014).
The specific energy consumption for a number of desalination units installed on the islands is presented in Figure 4. These data were acquired through the desalination manufactures and local water companies (TEMAK SA, personal communication, 2014; Culligan Hellas SA, personal communication on technical characteristics of existing desalination units, 2014). Both high pressure and booster pumps are included in these data.
A promising mechanical vapor compression (MVC) desalination system with the exploitation of wind energy (330 kW) was built in the island of Symi in 2009. The installed system was not successful due to important technical problems that caused very low water production. This system was stand-alone, with specific energy consumption of about 14.5 kWh/m3 [25]. Its operation ceased in 2011 due to a fire and it was never repaired. An autonomous
of 20 kWp photovoltaics as power source. The unit provides potable water to a small army
The urban water supply in Syros is covered mainly by desalinated water. The island power grid is autonomous and the electrical energy supply relies on diesel fuel transferred from the mainland. Currently, there are 13 RO units distributed in five regions (Figure 5). According to the Syros Water Company (Syros Water Supply and Sewerage Company [26] personal communication) and the Hellenic Electricity Distribution Network Operator [27], more than 11% of the electricity produced in the island is used for potable water production and the total desalination power demand, when plants are running at full capacity is 5.2% of the conven-
These plants require 2.08 MW of power and the cost of energy is the primary operating cost for the Water Supply Company. The total operational cost is estimated at about €1.2–1.6/m3
contribution of energy to the water cost is €0.7 m3 (about 45%), with an average energy cost of
The following indicators will be used to assess the energy-for-water nexus: (a) energy consumption per volume of water sold, (b) energy cost per volume of water sold, and (c) energy
. The
5. The case of Syros island: comparing the operation and energy
/day was installed in 2013 on the very small island of Strogyli with the use
Responding to Water Challenges Through Desalination: Energy Considerations
http://dx.doi.org/10.5772/intechopen.69956
171
RO unit of 8 m<sup>3</sup>
€0.086/kWh [12].
Figure 5. Desalination units in Syros.
costs of desalination
tional installed power production units.
camp.
### 4. Desalination with renewable energy sources
Renewable energy powered desalination offers noteworthy benefits in comparison to use conventional energy supply. Although desalination systems with renewable energy have increased installation costs, they have lower operating costs due to the fact that costly fuel oil used in most islands is avoided. During the summer period, higher wind speed and solar radiation occur while water demand is also higher, and thus these have been identified as possible energy sources for supporting desalination systems in general and in the Greek islands especially.
Annual mean wind speed in most areas varies from 5 to 7 m/s, but it can be much higher in some areas. Annual solar radiation varies from 1400 to 1700 kWh/m2 , and in the islands of Milos, Nisyros, Kimolos, and Thira have high-enthalpy geothermal fields. Since the installation of solar stills, several attempts have been made to use renewable energy for desalination, both on pilot and commercial systems. The Greek state has promoted desalination systems powered by renewable energy by prioritizing the licensing procedure for the installation of the renewable systems.
A desalination unit with geothermal energy as an energy source with capacity of 80 m3 /day using multi-effect distillation (MED) technology was installed in the framework of a European project and operated effectively for demonstrations in the island of Kimolos in 1997–1998 [23]. The water production cost was estimated at € 1.7/m<sup>3</sup> . However, at the end of the project, the desalination unit was abandoned. A very small RO unit with a capacity of 4.8 m3 /day, powered by a stand-alone 15 kW wind turbine, was installed as a pilot system in Therasia in 1997.
Hydriada, a floating RO desalination unit (80 m3 /day) with power supplied by a wind turbine and photovoltaics, was launched in 2007. It was a promising prototype designed to meet the potable water needs of 300 inhabitants in a small and arid island. Iraklia, a suitable island in Cyclades was selected as a pilot area. Early technical problems emerged and were compounded by the indifference of the local authority and the high maintenance costs, and thus the project was abandoned [24].
In the island of Milos, a successful desalination unit was constructed during 2007–2009 using wind energy supply. Three similar RO units with a total capacity of 3360 m3 /day and very low specific energy consumption (approximately 3.5 kWh/m3 ) were powered by a 850 kW wind turbine. A supervisory control and a data acquisition system were installed to assist and to optimize the operation of the RO units and the wind turbines as both the RO unit and the wind turbine are connected into the island's autonomous power grid. Optimization is accomplished by the use of operation modes depending on climate conditions, the potable water demand forecasting, and the level of the water tanks. One more important innovation of this project was its ownership status: the wind turbine and the desalination unit are private investments, and they work under contract agreement with the island municipality for supplying water.
A promising mechanical vapor compression (MVC) desalination system with the exploitation of wind energy (330 kW) was built in the island of Symi in 2009. The installed system was not successful due to important technical problems that caused very low water production. This system was stand-alone, with specific energy consumption of about 14.5 kWh/m3 [25]. Its operation ceased in 2011 due to a fire and it was never repaired. An autonomous RO unit of 8 m<sup>3</sup> /day was installed in 2013 on the very small island of Strogyli with the use of 20 kWp photovoltaics as power source. The unit provides potable water to a small army camp.
## 5. The case of Syros island: comparing the operation and energy costs of desalination
The urban water supply in Syros is covered mainly by desalinated water. The island power grid is autonomous and the electrical energy supply relies on diesel fuel transferred from the mainland. Currently, there are 13 RO units distributed in five regions (Figure 5). According to the Syros Water Company (Syros Water Supply and Sewerage Company [26] personal communication) and the Hellenic Electricity Distribution Network Operator [27], more than 11% of the electricity produced in the island is used for potable water production and the total desalination power demand, when plants are running at full capacity is 5.2% of the conventional installed power production units.
These plants require 2.08 MW of power and the cost of energy is the primary operating cost for the Water Supply Company. The total operational cost is estimated at about €1.2–1.6/m3 . The contribution of energy to the water cost is €0.7 m3 (about 45%), with an average energy cost of €0.086/kWh [12].
The following indicators will be used to assess the energy-for-water nexus: (a) energy consumption per volume of water sold, (b) energy cost per volume of water sold, and (c) energy
Figure 5. Desalination units in Syros.
The specific energy consumption for a number of desalination units installed on the islands is presented in Figure 4. These data were acquired through the desalination manufactures and local water companies (TEMAK SA, personal communication, 2014; Culligan Hellas SA, personal communication on technical characteristics of existing desalination units, 2014). Both
Renewable energy powered desalination offers noteworthy benefits in comparison to use conventional energy supply. Although desalination systems with renewable energy have increased installation costs, they have lower operating costs due to the fact that costly fuel oil used in most islands is avoided. During the summer period, higher wind speed and solar radiation occur while water demand is also higher, and thus these have been identified as possible energy sources for supporting desalination systems in general and in the Greek islands especially.
Annual mean wind speed in most areas varies from 5 to 7 m/s, but it can be much higher in some
Nisyros, Kimolos, and Thira have high-enthalpy geothermal fields. Since the installation of solar stills, several attempts have been made to use renewable energy for desalination, both on pilot and commercial systems. The Greek state has promoted desalination systems powered by renewable energy by prioritizing the licensing procedure for the installation of the renewable systems. A desalination unit with geothermal energy as an energy source with capacity of 80 m3
using multi-effect distillation (MED) technology was installed in the framework of a European project and operated effectively for demonstrations in the island of Kimolos in 1997–1998 [23].
and photovoltaics, was launched in 2007. It was a promising prototype designed to meet the potable water needs of 300 inhabitants in a small and arid island. Iraklia, a suitable island in Cyclades was selected as a pilot area. Early technical problems emerged and were compounded by the indifference of the local authority and the high maintenance costs, and thus the project
In the island of Milos, a successful desalination unit was constructed during 2007–2009 using
turbine. A supervisory control and a data acquisition system were installed to assist and to optimize the operation of the RO units and the wind turbines as both the RO unit and the wind turbine are connected into the island's autonomous power grid. Optimization is accomplished by the use of operation modes depending on climate conditions, the potable water demand forecasting, and the level of the water tanks. One more important innovation of this project was its ownership status: the wind turbine and the desalination unit are private investments, and they
desalination unit was abandoned. A very small RO unit with a capacity of 4.8 m3
wind energy supply. Three similar RO units with a total capacity of 3360 m3
work under contract agreement with the island municipality for supplying water.
by a stand-alone 15 kW wind turbine, was installed as a pilot system in Therasia in 1997.
, and in the islands of Milos,
. However, at the end of the project, the
) were powered by a 850 kW wind
/day) with power supplied by a wind turbine
/day
/day, powered
/day and very low
high pressure and booster pumps are included in these data.
4. Desalination with renewable energy sources
areas. Annual solar radiation varies from 1400 to 1700 kWh/m2
The water production cost was estimated at € 1.7/m<sup>3</sup>
specific energy consumption (approximately 3.5 kWh/m3
Hydriada, a floating RO desalination unit (80 m3
was abandoned [24].
170 Desalination
consumption per capita. These analyses take into account only the desalination plant of Syros capital, Hermoupolis, for which the required input data are available. For comparison purposes, similar results are presented for the two major water supply systems in Greece (Table 2), the Athens, and the Thessaloniki water supply systems.
a pattern, as seen in Figure 7. Similar ratios between water systems also apply for the case of the per capita energy demand for water supply. Syros has a value 14 times than that of Athens
Responding to Water Challenges Through Desalination: Energy Considerations
http://dx.doi.org/10.5772/intechopen.69956
173
Figure 7. Energy cost per volume of water sold in the water supply systems of Hermoupolis, Athens, and
Figure 8. Energy consumption per capita in the water supply systems of Hermoupolis, Athens, and Thessaloniki [27–30].
and more than twice that of Thessaloniki (Figure 8).
Thessaloniki [27–30].
#### 5.1. Assessment indicators
The energy consumption per volume of water sold is presented in Figure 6 for three water energy systems. The Athens water system has a value 0.75 kWh/m<sup>3</sup> , Thessaloniki's water system has twice that, and Hermoupolis about 15 times. The cost of energy also follows such
Table 2. Data used for the calculation of the assessment indicators.
Figure 6. Energy consumption per volume of water sold in the water supply systems of Hermoupolis, Athens, and Thessaloniki [27–30].
a pattern, as seen in Figure 7. Similar ratios between water systems also apply for the case of the per capita energy demand for water supply. Syros has a value 14 times than that of Athens and more than twice that of Thessaloniki (Figure 8).
consumption per capita. These analyses take into account only the desalination plant of Syros capital, Hermoupolis, for which the required input data are available. For comparison purposes, similar results are presented for the two major water supply systems in Greece (Table 2),
The energy consumption per volume of water sold is presented in Figure 6 for three water
system has twice that, and Hermoupolis about 15 times. The cost of energy also follows such
Hermoupolis DEYA Syrou
) 1.05 385.5 89.7
/year) 66.6 74.9 66.3
Energy cost (€/kWh) 0.086 0.089 No data (assumed same as Athens)
Figure 6. Energy consumption per volume of water sold in the water supply systems of Hermoupolis, Athens, and
) 0.89 337 66.3
, Thessaloniki's water
Athens EYDAP Thessaloniki EYATH
the Athens, and the Thessaloniki water supply systems.
energy systems. The Athens water system has a value 0.75 kWh/m<sup>3</sup>
Population served 13,400 4,500,000 1,000,000
Energy consumption (GWh) 10 251 119 Reference year or period 2010–2013 2009 2008
Data sources [27] [28, 29] [29, 30]
Table 2. Data used for the calculation of the assessment indicators.
5.1. Assessment indicators
Water production (hm<sup>3</sup>
Annual per capita consumption (m<sup>3</sup>
Water billed (hm<sup>3</sup>
172 Desalination
Thessaloniki [27–30].
Figure 7. Energy cost per volume of water sold in the water supply systems of Hermoupolis, Athens, and Thessaloniki [27–30].
Figure 8. Energy consumption per capita in the water supply systems of Hermoupolis, Athens, and Thessaloniki [27–30].
#### 5.2. Monthly energy consumption
Water consumption in Syros is high during summer, due to the tourism. For instance, consumption during August is around 85% higher than in February (Figure 9). The same pattern applies to the energy consumption of the desalination. However, the desalination process is significantly more efficient in the summer period (Figure 10). Specific energy for the Hermoupolis plant is 8.1 kWh/m<sup>3</sup> in August and 9.7 kWh/m<sup>3</sup> in February, a reduction of 16.5%. This reduction can be credited to three main factors which are as follows: (i) the units operate continuously during summer, with less start-and-stop cycles, (ii) the temperature of the seawater during summer period is high, affecting properties such as viscosity and lowering the energy demand of the high-pressure pump [31], and (iii) as the Hermoupolis desalination plant comprises several desalination units, each one with different capacities and specific energies, making extensive use of the unit with the lower specific energy also lowers the specific energy of the desalination plant as
Responding to Water Challenges Through Desalination: Energy Considerations
http://dx.doi.org/10.5772/intechopen.69956
175
6. Satisfying energy needs by using RES: a design methodology
mix contribution comes from lignite, natural gas, and RES.
water demand is the same for every day (Qdaily) of the month.
is obtained through the grid.
6.1. Design methodology
6.1.1. Water demand
6.1.2. Desalination unit
Desalination capacity (Qcap in m<sup>3</sup>
energy demand (Sp in kWh/m<sup>3</sup>
A method for assessing the desalination system with RES is presented in the following section. In most cases, the desalination systems are interconnected with the electrical grid of the island, where the energy mix is dominated by diesel or heavy oil and a small percentage is covered by RES or other sources. Some islands are interconnected with the mainland and thus the energy
The renewable sources considered in the present study are solar and wind energy. Energy demand for desalination is covered by the produced renewable energy and the excess is supplied to the grid. If the renewable energy is not sufficient to cover the demand, then energy
The optimal configuration is the one that minimizes the water cost from the investor point of view. A configuration is defined by a set of design parameters such as the desalination capacity, the photovoltaic installed power, etc. A methodology, using hourly simulation, was developed to identify the optimum configuration. The islands to be examined are Syros and
Water demand can be derived from real historical data, from average monthly values or from other sources. The developed methodology requires daily water demand values throughout the year. However, if only monthly average data are available, then it can be assumed that
and water salinity. Thus, the average power that is needed during normal operation is as
/day) is one of the design parameters. The specific
) is a specification parameter of a given desalination unit
a whole.
Patmos.
follows:
Figure 9. Monthly water production and energy consumption for the Hermoupolis desalination plant (average values for 2010–2013) [27].
Figure 10. Specific energy consumption of the desalination units in the Hermoupolis desalination plant (average values for 2010–2013) [27].
pattern applies to the energy consumption of the desalination. However, the desalination process is significantly more efficient in the summer period (Figure 10). Specific energy for the Hermoupolis plant is 8.1 kWh/m<sup>3</sup> in August and 9.7 kWh/m<sup>3</sup> in February, a reduction of 16.5%. This reduction can be credited to three main factors which are as follows: (i) the units operate continuously during summer, with less start-and-stop cycles, (ii) the temperature of the seawater during summer period is high, affecting properties such as viscosity and lowering the energy demand of the high-pressure pump [31], and (iii) as the Hermoupolis desalination plant comprises several desalination units, each one with different capacities and specific energies, making extensive use of the unit with the lower specific energy also lowers the specific energy of the desalination plant as a whole.
## 6. Satisfying energy needs by using RES: a design methodology
A method for assessing the desalination system with RES is presented in the following section. In most cases, the desalination systems are interconnected with the electrical grid of the island, where the energy mix is dominated by diesel or heavy oil and a small percentage is covered by RES or other sources. Some islands are interconnected with the mainland and thus the energy mix contribution comes from lignite, natural gas, and RES.
The renewable sources considered in the present study are solar and wind energy. Energy demand for desalination is covered by the produced renewable energy and the excess is supplied to the grid. If the renewable energy is not sufficient to cover the demand, then energy is obtained through the grid.
The optimal configuration is the one that minimizes the water cost from the investor point of view. A configuration is defined by a set of design parameters such as the desalination capacity, the photovoltaic installed power, etc. A methodology, using hourly simulation, was developed to identify the optimum configuration. The islands to be examined are Syros and Patmos.
#### 6.1. Design methodology
#### 6.1.1. Water demand
5.2. Monthly energy consumption
174 Desalination
2010–2013) [27].
for 2010–2013) [27].
Water consumption in Syros is high during summer, due to the tourism. For instance, consumption during August is around 85% higher than in February (Figure 9). The same
Figure 9. Monthly water production and energy consumption for the Hermoupolis desalination plant (average values for
Figure 10. Specific energy consumption of the desalination units in the Hermoupolis desalination plant (average values
Water demand can be derived from real historical data, from average monthly values or from other sources. The developed methodology requires daily water demand values throughout the year. However, if only monthly average data are available, then it can be assumed that water demand is the same for every day (Qdaily) of the month.
#### 6.1.2. Desalination unit
Desalination capacity (Qcap in m<sup>3</sup> /day) is one of the design parameters. The specific energy demand (Sp in kWh/m<sup>3</sup> ) is a specification parameter of a given desalination unit and water salinity. Thus, the average power that is needed during normal operation is as follows:
$$P\_{\rm des} = Sp \cdot \frac{Q\_{\rm cap}}{24 \,\mathrm{hr}} (\mathrm{kW}) \tag{1}$$
where GTðt<sup>Þ</sup> is the mean hour solar radiation for every hour in (kWh/m<sup>2</sup>
The total installed power of the wind turbines is also a design parameter (PWEC in kW). The energy produced by the wind energy conversion (WEC) system in a given hour t, is given by:
where PWECð Þ V tð Þ is the power produced by the wind energy system for wind speed V tð Þ. In the case of a single wind turbine, the power from the WEC system is given by the power curve
The scheme concerning the energy flows is quite simple. There is no sale of extra electrical
The renewable energy consumed by the RO unit is the sum of the wind turbine energy and
The energy produced from renewable sources when the RO unit is not operating is as follows:
The energy difference among the available renewable energy and energy needed for desalina-
If ENETð Þt ≥ 0, then the renewable energy is not fully exploited by the RO unit. Therefore, the
If ENETð Þt < 0, then there are additional energy needs and they are going to be covered from
Due to technical limitations or legislation, there is an upper limit to the amount of energy that can be infused into the grid, depending on the relation between the conventional and RES energy production in the island. Thus, the energy that can be infused to the grid will be
energy to the grid, but net metering method is applied with annual clearance.
photovoltaic energy produced during the operating time of the unit:
total amount of renewable energy that can be infused into the grid is.
EWECðtÞ ¼ ηoef f � PWECð Þ� V tð Þ 1 hr ðkWhÞ (5)
Responding to Water Challenges Through Desalination: Energy Considerations
ERESavðÞ¼ t ðEWECð Þþ t EPVð Þt Þ � OP tð Þ (6)
ErestðÞ¼ t EWECð Þþ t EPVð Þ� t ERESavð Þt (7)
ENETðÞ¼ t ERESavð Þ� t EDESð Þt (8)
ETogridavðÞ¼ t ENETð Þþ t Erestð Þt (9)
EFromgridðÞ¼j t ENETð Þj t (10)
test condition reference radiation (1000 W/m<sup>2</sup>
6.1.4. Wind energy conversion system model
overall efficiency respectively.
of the wind turbine.
tion is as follows:
the grid.
6.1.5. Energy management scheme
), GSTC is the standard
177
), ηinv, and ηoef f are the inverter efficiency and
http://dx.doi.org/10.5772/intechopen.69956
The hours of operation of the desalination unit per day can be derived as follows:
$$T\_{op} = \min\left(\frac{Q\_{daily}}{Q\_{cap}/24 \text{ hr}}, 24 \text{ hr}\right) \text{(hr)}\tag{2}$$
The min operator denotes that hours of operation needed for each day cannot be more than 24 hr. The desalination might not be able to fully cover the demand and in this case, there is a water demand deficit. The operating hours of the desalination unit is a critical factor for the exploitation of the renewable sources. In case of solar energy, the best strategy corresponds to the operation of the unit in the period around noon (Figure 11). In case of wind energy, then the optimum period depends on the micro scale of the location of the power plant (in most cases, the afternoon is the windiest time).
The hourly distribution of energy requirements for desalination is given by Eq. (3):
$$E\_{DES}(t) = OP(t) \cdot P\_{DES}(\text{kWh}) \tag{3}$$
where OP tð Þ is the desalination production distribution, representing the percentage of operation for each hour t of the year (OP tð Þ is 100% when the desalination unit is in operation for an hour without breaks, 0% when the desalination is not operating at all, and a value between 0 and 100% in other cases).
Figure 11. Reverse osmosis operation schedule (example).
#### 6.1.3. Photovoltaic model
The installed power of the photovoltaic (PV) system is a design parameter, denoted by PPV (kWp) and the energy produced by a PV array, in a given hour t, is given by:
$$E\_{\rm PV}(t) = \eta\_{inv} \cdot \eta\_{\rm c\!f\!f\!f} \cdot P\_{\rm PV} \cdot \frac{\overline{G}\_{\rm T}(t)}{G\_{\rm STC}} \cdot 1 \text{ hr} \quad (\text{kWh}) \tag{4}$$
where GTðt<sup>Þ</sup> is the mean hour solar radiation for every hour in (kWh/m<sup>2</sup> ), GSTC is the standard test condition reference radiation (1000 W/m<sup>2</sup> ), ηinv, and ηoef f are the inverter efficiency and overall efficiency respectively.
#### 6.1.4. Wind energy conversion system model
Pdes ¼ Sp �
The hours of operation of the desalination unit per day can be derived as follows:
The hourly distribution of energy requirements for desalination is given by Eq. (3):
Top <sup>¼</sup> min Qdaily
cases, the afternoon is the windiest time).
0 and 100% in other cases).
176 Desalination
6.1.3. Photovoltaic model
Figure 11. Reverse osmosis operation schedule (example).
Qcap
Qcap=24 hr, 24 hr !
The min operator denotes that hours of operation needed for each day cannot be more than 24 hr. The desalination might not be able to fully cover the demand and in this case, there is a water demand deficit. The operating hours of the desalination unit is a critical factor for the exploitation of the renewable sources. In case of solar energy, the best strategy corresponds to the operation of the unit in the period around noon (Figure 11). In case of wind energy, then the optimum period depends on the micro scale of the location of the power plant (in most
where OP tð Þ is the desalination production distribution, representing the percentage of operation for each hour t of the year (OP tð Þ is 100% when the desalination unit is in operation for an hour without breaks, 0% when the desalination is not operating at all, and a value between
The installed power of the photovoltaic (PV) system is a design parameter, denoted by PPV
GTðtÞ GSTC
� 1 hr ðkWhÞ (4)
(kWp) and the energy produced by a PV array, in a given hour t, is given by:
EPVðtÞ ¼ ηinv � ηoef f � PPV �
24 hrðkW<sup>Þ</sup> (1)
EDESðtÞ ¼ OPðtÞ � PDESðkWhÞ (3)
ðhrÞ (2)
The total installed power of the wind turbines is also a design parameter (PWEC in kW). The energy produced by the wind energy conversion (WEC) system in a given hour t, is given by:
$$E\_{WEC}(t) = \eta\_{\rm{eff}} \cdot P\_{WEC}(V(t)) \cdot \mathbf{1} \text{ hr} \quad (\text{kWh}) \tag{5}$$
where PWECð Þ V tð Þ is the power produced by the wind energy system for wind speed V tð Þ. In the case of a single wind turbine, the power from the WEC system is given by the power curve of the wind turbine.
#### 6.1.5. Energy management scheme
The scheme concerning the energy flows is quite simple. There is no sale of extra electrical energy to the grid, but net metering method is applied with annual clearance.
The renewable energy consumed by the RO unit is the sum of the wind turbine energy and photovoltaic energy produced during the operating time of the unit:
$$E\_{RESav}(t) = \left(E\_{WEC}(t) + E\_{PV}(t)\right) \cdot OP(t) \tag{6}$$
The energy produced from renewable sources when the RO unit is not operating is as follows:
$$E\_{\rm rest}(t) = E\_{\rm WEC}(t) + E\_{PV}(t) - E\_{\rm RESAv}(t) \tag{7}$$
The energy difference among the available renewable energy and energy needed for desalination is as follows:
$$E\_{NET}(t) = E\_{RESw}(t) - E\_{DES}(t) \tag{8}$$
If ENETð Þt ≥ 0, then the renewable energy is not fully exploited by the RO unit. Therefore, the total amount of renewable energy that can be infused into the grid is.
$$E\_{Togridv}(t) = E\_{NET}(t) + E\_{rest}(t) \tag{9}$$
If ENETð Þt < 0, then there are additional energy needs and they are going to be covered from the grid.
$$E\_{Frough}(t) = |E\_{NET}(t)|\tag{10}$$
Due to technical limitations or legislation, there is an upper limit to the amount of energy that can be infused into the grid, depending on the relation between the conventional and RES energy production in the island. Thus, the energy that can be infused to the grid will be
$$E\_{Togrid}(t) = \min\{E\_{Togridw}(t), E\_{gridS}(t)\}\tag{11}$$
EgridS is the RES margin of the autonomous grid system.
The annual energy demand for the desalination is
$$E\_{\text{DESat}} = \sum\_{t=1}^{8760} E\_{\text{DES}}(t) \tag{12}$$
6.2. Case study
6.2.1. Model inputs
6.2.1.1. Water demand
The installation of renewable energy systems for covering water demand will be examined for the city of Hermoupolis in the island of Syros and for the island of Patmos. As detailed above, Hermoupolis has high energy needs for water production; therefore, the introduction of RES technologies could be beneficial in minimizing the cost of the water. The island of Patmos still covers its water needs by water transports and local drills. However, two units of 600 m<sup>3</sup> each
Responding to Water Challenges Through Desalination: Energy Considerations
Monthly water production for Hermoupolis is given in Figure 9. Maximum water demand for Patmos, according to a recent public tender, was estimated at 236,000 m3 per year, but not all of the demand will be covered by the desalination. Thus, an estimation can be derived from the amount of water that was transferred to the island by tanker ships, according to Special
tion data, we can assume that the profile is the same as the previous years (monthly data exists). Also, due to the rebound effect on desalination, the water produced by desalination can be up to 20% more than water transferred. However, the case of covering all water needs
The data needed for the simulation are the wind speed and the solar radiation as hourly time series for 1 year. For Hermoupolis, the average wind speed at wind turbine height is 6.5 m/s
Patmos is an important site of cultural heritage, the wind energy option is not examined. The
Total cost has been estimated at 2.082 M€ using data from the Syros Water Service [32]. The installation cost of the reverse osmosis unit in Patmos installation was estimated at 0.739 M€. The specific cost of the photovoltaic system was calculated for a fixed price of 1200 €/kWp and
• Labor is a fixed cost, in a small desalination plant estimated at LB ¼ 25000 € per year and per person. When the annual water production is low, the labor cost is an important factor in the water cost. One person was employed for the RO unit of Patmos and four persons
and the total solar radiation at the photovoltaic inclination is 1847 kWh/m<sup>2</sup>
.
the installation cost of the wind turbine system is estimated at 1500 €/kW.
These costs are either fixed or dependent on the water production:
. Due to the lack of monthly distribu-
http://dx.doi.org/10.5772/intechopen.69956
179
. As the island of
have already been installed and are planned to be operational in the near future.
Secretariat for Water. For 2014, this value was 68,654 m3
through desalination will be examined.
annual solar radiation is 1935 kWh/m<sup>2</sup>
6.2.1.4. Operation and maintenance cost
for the unit of Hermoupolis.
6.2.1.2. Meteorological data
6.2.1.3. Installation cost
and the renewable energy collected in one year is
$$E\_{RESam} = \sum\_{t=1}^{8760} \left( E\_{WEC}(t) + E\_{PV}(t) \right) \tag{13}$$
#### 6.1.6. Economic evaluation
Economic evaluation will allow the identification of the most profitable and optimal configurations with the use of water cost as the main decision parameter. The levelized water cost (WC) per cubic meter can be calculated as:
$$WC = \frac{(IC\_{EN} + IC\_{RO}) \cdot R + OM + EC}{WP} \tag{14}$$
where: ICEN is the installation cost of the power system.
ICRO is the reverse osmosis installation cost.
R is the annuity factor (n is the life time of the investment and i the interest rate):
$$R = \frac{i}{1 - (1 + i)^{-n}} \tag{15}$$
OM is the annual operation and maintenance cost that contains consumables (CN), (filters, spare parts), chemicals for posttreatment and pretreatment (CHM), membrane replacements (MR) and labor (LB):
$$\text{COM} = \text{CN} + \text{CHM} + \text{MR} + \text{LB} \tag{16}$$
EC is the annual cost of energy and ECSP is the specific cost of energy (€/kWh):
$$EC = EC\_{SP} \cdot \max\left( \left( \sum\_{t=1}^{8760} E\_{Frongrid}(t) - \sum\_{t=1}^{8760} E\_{Togrid}(t) \right), 0 \right) \tag{17}$$
WP is the annual water production:
$$WP = \sum\_{t=1}^{8760} \frac{Q\_{cap}}{24 \text{ hr}} OP(t) \tag{18}$$
#### 6.2. Case study
ETogridðÞ¼ <sup>t</sup> min ETogridavð Þ<sup>t</sup> , EgridSð Þ<sup>t</sup> � � (11)
EDESð Þt (12)
ð Þ EWECð Þþ t EPVð Þt (13)
WP (14)
<sup>1</sup> � ð Þ <sup>1</sup> <sup>þ</sup> <sup>i</sup> �<sup>n</sup> (15)
OM ¼ CN þ CHM þ MR þ LB (16)
ETogridð Þt
, 0
24 hr OP tð Þ (18)
(17)
8760
t¼1
EgridS is the RES margin of the autonomous grid system.
EDESan <sup>¼</sup> <sup>X</sup>
8760
t¼1
ERESan <sup>¼</sup> <sup>X</sup>
8760
t¼1
Economic evaluation will allow the identification of the most profitable and optimal configurations with the use of water cost as the main decision parameter. The levelized water cost
WC <sup>¼</sup> <sup>ð</sup>ICEN <sup>þ</sup> ICROÞ � <sup>R</sup> <sup>þ</sup> OM <sup>þ</sup> EC
R is the annuity factor (n is the life time of the investment and i the interest rate):
EC is the annual cost of energy and ECSP is the specific cost of energy (€/kWh):
8760
t¼1
WP <sup>¼</sup> <sup>X</sup> 8760
t¼1
Qcap
OM is the annual operation and maintenance cost that contains consumables (CN), (filters, spare parts), chemicals for posttreatment and pretreatment (CHM), membrane
EFromgridð Þ� <sup>t</sup> <sup>X</sup>
!
!
<sup>R</sup> <sup>¼</sup> <sup>i</sup>
The annual energy demand for the desalination is
and the renewable energy collected in one year is
(WC) per cubic meter can be calculated as:
where: ICEN is the installation cost of the power system.
replacements (MR) and labor (LB):
WP is the annual water production:
EC <sup>¼</sup> ECSP � max <sup>X</sup>
ICRO is the reverse osmosis installation cost.
6.1.6. Economic evaluation
178 Desalination
The installation of renewable energy systems for covering water demand will be examined for the city of Hermoupolis in the island of Syros and for the island of Patmos. As detailed above, Hermoupolis has high energy needs for water production; therefore, the introduction of RES technologies could be beneficial in minimizing the cost of the water. The island of Patmos still covers its water needs by water transports and local drills. However, two units of 600 m<sup>3</sup> each have already been installed and are planned to be operational in the near future.
### 6.2.1. Model inputs
### 6.2.1.1. Water demand
Monthly water production for Hermoupolis is given in Figure 9. Maximum water demand for Patmos, according to a recent public tender, was estimated at 236,000 m3 per year, but not all of the demand will be covered by the desalination. Thus, an estimation can be derived from the amount of water that was transferred to the island by tanker ships, according to Special Secretariat for Water. For 2014, this value was 68,654 m3 . Due to the lack of monthly distribution data, we can assume that the profile is the same as the previous years (monthly data exists). Also, due to the rebound effect on desalination, the water produced by desalination can be up to 20% more than water transferred. However, the case of covering all water needs through desalination will be examined.
#### 6.2.1.2. Meteorological data
The data needed for the simulation are the wind speed and the solar radiation as hourly time series for 1 year. For Hermoupolis, the average wind speed at wind turbine height is 6.5 m/s and the total solar radiation at the photovoltaic inclination is 1847 kWh/m<sup>2</sup> . As the island of Patmos is an important site of cultural heritage, the wind energy option is not examined. The annual solar radiation is 1935 kWh/m<sup>2</sup> .
#### 6.2.1.3. Installation cost
Total cost has been estimated at 2.082 M€ using data from the Syros Water Service [32]. The installation cost of the reverse osmosis unit in Patmos installation was estimated at 0.739 M€. The specific cost of the photovoltaic system was calculated for a fixed price of 1200 €/kWp and the installation cost of the wind turbine system is estimated at 1500 €/kW.
#### 6.2.1.4. Operation and maintenance cost
These costs are either fixed or dependent on the water production:
• Labor is a fixed cost, in a small desalination plant estimated at LB ¼ 25000 € per year and per person. When the annual water production is low, the labor cost is an important factor in the water cost. One person was employed for the RO unit of Patmos and four persons for the unit of Hermoupolis.
• Chemicals cost is variable and depends on the water production. The specific cost of the chemicals is estimated from 0.02 to 0.05 \$/m<sup>3</sup> [33], but in some cases it can be as high as 0.23 \$/m<sup>3</sup> [34]. For the Greek islands a value of 0.065 €/m3 is proposed [35]. Thus, the chemical costs are CH <sup>¼</sup> WP � <sup>0</sup>:065€=m3.
When installing photovoltaic units in the island of Patmos, water cost reaches a minimum of
taic power plant of 2.3 MW The reduction in water cost is 0.1€/m3 (the same as is in Patmos). On the other hand, when installing wind turbines, the water cost can be as low as 1.12 €/m<sup>3</sup>
total wind turbine installed power of 1750 kW. In Figure 14, the water cost is presented versus
cost component is different. Energy contributes 61% of the total water cost (Figure 12).
Figure 14. Water cost versus wind and photovoltaic install power for Hermoupolis. One technology each time.
The installation of photovoltaics minimizes water cost to the value of 1.16 €/m<sup>3</sup>
Water cost in the city of Hermoupolis is estimated at 1.26 €/m<sup>3</sup>
, which can be achieved when the installed power of the photovoltaic system is
Responding to Water Challenges Through Desalination: Energy Considerations
, but the contribution of each
http://dx.doi.org/10.5772/intechopen.69956
for a photovol-
for
181
about 1.52 €/m3
about 200 kWp (Figure 13).
the installed power of each technology.
Figure 13. Water cost versus photovoltaic install power for Patmos.
#### 6.2.1.5. Other inputs
The specific energy for Patmos is assumed to beSp <sup>¼</sup> <sup>5</sup>:5 kWh=m<sup>3</sup> and for Hermoupolis, a monthly specific energy is used (Figure 10). The flushing duration will be 15 min (f m ¼ 0:25 hr). The cost of electrical energy for Hermoupolis is 0.086 €/kWh according to local data, and for Patmos a value of 0.1 €/kWh was assumed.
#### 6.2.2. Results
The water cost in the island of Patmos, using conventional energy sources, is estimated at WC <sup>¼</sup> <sup>1</sup>:62€=m3. Figure 12 presents the cost breakdown graph, where it is obvious that all cost components have approximately the same participation in water cost. If the RO unit is sized to produce water to cover all the demand of Patmos, then the water cost drops to WC <sup>¼</sup> <sup>1</sup>:154€=m3.
Figure 12. Water cost breakdown into the basic components for Patmos and Hermoupolis.
When installing photovoltaic units in the island of Patmos, water cost reaches a minimum of about 1.52 €/m3 , which can be achieved when the installed power of the photovoltaic system is about 200 kWp (Figure 13).
Water cost in the city of Hermoupolis is estimated at 1.26 €/m<sup>3</sup> , but the contribution of each cost component is different. Energy contributes 61% of the total water cost (Figure 12).
The installation of photovoltaics minimizes water cost to the value of 1.16 €/m<sup>3</sup> for a photovoltaic power plant of 2.3 MW The reduction in water cost is 0.1€/m3 (the same as is in Patmos). On the other hand, when installing wind turbines, the water cost can be as low as 1.12 €/m<sup>3</sup> for total wind turbine installed power of 1750 kW. In Figure 14, the water cost is presented versus the installed power of each technology.
Figure 13. Water cost versus photovoltaic install power for Patmos.
• Chemicals cost is variable and depends on the water production. The specific cost of the chemicals is estimated from 0.02 to 0.05 \$/m<sup>3</sup> [33], but in some cases it can be as high as 0.23 \$/m<sup>3</sup> [34]. For the Greek islands a value of 0.065 €/m3 is proposed [35]. Thus, the
• Membrane cost is the most difficult to estimate, due to the fact that the life span of a membrane depends on many parameters. Proper use of the reverse osmosis system can produce water of potable quality for 5 years or more. In literature and in feasibility studies, the life span of the membrane varies from 3 to 5 years. Water production in the islands is limited and the periodical operation of the reverse osmosis can incur a specific replacement cost that is three times higher. For this reason, membrane cost
The specific energy for Patmos is assumed to beSp <sup>¼</sup> <sup>5</sup>:5 kWh=m<sup>3</sup> and for Hermoupolis, a monthly specific energy is used (Figure 10). The flushing duration will be 15 min (f m ¼ 0:25 hr). The cost of electrical energy for Hermoupolis is 0.086 €/kWh according to local data, and for
The water cost in the island of Patmos, using conventional energy sources, is estimated at WC <sup>¼</sup> <sup>1</sup>:62€=m3. Figure 12 presents the cost breakdown graph, where it is obvious that all cost components have approximately the same participation in water cost. If the RO unit is sized to produce water to cover all the demand of Patmos, then the water cost drops to
Figure 12. Water cost breakdown into the basic components for Patmos and Hermoupolis.
• Consumables and other costs: will be taken equal to CN <sup>¼</sup> WP � <sup>0</sup>:04€=m<sup>3</sup> [16].
[36]. In this study, membrane replacement cost will be
, which agrees with data from real plants in similar areas MR ¼
chemical costs are CH <sup>¼</sup> WP � <sup>0</sup>:065€=m3.
varies from 0.04 to 0.34 €/m<sup>3</sup>
Patmos a value of 0.1 €/kWh was assumed.
an average 0.15 €/m<sup>3</sup>
WP � <sup>0</sup>:15€=m3[17].
6.2.1.5. Other inputs
180 Desalination
6.2.2. Results
WC <sup>¼</sup> <sup>1</sup>:154€=m3.
Figure 14. Water cost versus wind and photovoltaic install power for Hermoupolis. One technology each time.
water. Both at the pilot and commercial levels, efforts have been made to improve the performance of desalination technology. Economic considerations and technical limitations are the
Responding to Water Challenges Through Desalination: Energy Considerations
http://dx.doi.org/10.5772/intechopen.69956
183
The primary challenge in the desalination practice and specifically in the reverse osmosis technology is the reduction of the energy consumption of the units. As energy is a very high cost factor for desalination, various methods for lowering the specific energy consumption can help decrease the cost of water. Another important option is the use of cheaper energy sources. Renewable energy technologies can offer a feasible way to reduce the pressure on energy systems for water production, and thus developing renewable-energy-powered desalination plants has become a priority in the Greek energy and water sectors. Energy storage is not a practical solution due to its high investment cost and the low lifetime of battery systems, but the net metering method can lower water cost by up to 13.5%. As a state of the art technology, hybrid systems, combined with operation optimization methods, provide appropriate solutions for cheaper energy; however, wind farms and photovoltaic arrays often face disapproval, especially in the tourism-dependent islands, as local inhabitants find that the local esthetic value is reduced. Environmental reasons and land-use conflicts may also prevent the integration of renewable-energy-powered RO systems. To construct efficiently operated, optimal RO desalination powered by renewable energy
Water deficiency in most of the arid islands is treated with water transfers at the expense of the central government. The operation and maintenance cost of the desalination units is subsidized. In an era of economic crisis, this condition may not be feasible and there is a pressing need to move on toward sustainable solutions. Private sector participation, though not always socially acceptable by the locals, has to be examined. Moreover, as water and energy resources are due to face increasing pressure over the next few decades, the evaluation of trade-offs and the encouragement of cross-sector planning will be critical for further development and for
The Greek experience indicates that constructing and operating RO units is made possible by state support (in the form of subsidies). The energy intensity of RO units remains high. However, RO technology, although with high cost, is reliable and offers potable water of adequate quality and quantity. Therefore, technical and operation policy challenges (including securing adequate funding) should be addressed so that the implementation of RO technology will further enhance
Environmental and Energy Management Research Unit, School of Chemical Engineering,
most important reasons that such efforts are not always successful.
systems, the state should promote pilot systems through the relevant studies.
energy efficiency and will be a part of an integrated water management system.
George Arampatzis\*, Avraam Kartalidis and Dionysis Assimacopoulos
\*Address all correspondence to: arab@chemeng.ntua.gr
National Technical University of Athens, Athens, Greece
their sustainable management.
Author details
Figure 15. Water cost versus wind and photovoltaic installed power for Hermoupolis when both technologies exists simultaneously.
When both photovoltaics and wind turbines are installed simultaneously, water cost has an absolute minimum value at 1.09 €/m<sup>3</sup> , which is 0.17 €/m<sup>3</sup> less than the case with conventional energy. This value can be achieved for various installed power combinations of each technology. The lower values for both the photovoltaics and wind turbine-installed power (which consequently gives the lower installation costs) are for 1250kWp and 1250 kW, respectively (Figure 15).
In both cases, an increase in water demand during the next years has not been taken into account; even so, the installation of renewable energy technologies will be beneficial and will make water cheaper due to higher utilization of the desalination units and due to cheap energy sources.
#### 7. Discussion and conclusions
The practice of desalination has been identified as a promising solution to the water scarcity problem of the Greek islands (Cyclades and Dodecanese). The early efforts during the 1960s with the use of solar distillation technology were not successful, as the technical know-how was very limited and maintenance was inadequate. Also, the operational costs (labor) were very high due to the lack of automation technologies. The advances in desalination technology, combined with the need to meet water demand in the rapidly developing tourist areas, enhanced the prospects of desalination as a supplementary or even primary source of potable water. Both at the pilot and commercial levels, efforts have been made to improve the performance of desalination technology. Economic considerations and technical limitations are the most important reasons that such efforts are not always successful.
The primary challenge in the desalination practice and specifically in the reverse osmosis technology is the reduction of the energy consumption of the units. As energy is a very high cost factor for desalination, various methods for lowering the specific energy consumption can help decrease the cost of water. Another important option is the use of cheaper energy sources. Renewable energy technologies can offer a feasible way to reduce the pressure on energy systems for water production, and thus developing renewable-energy-powered desalination plants has become a priority in the Greek energy and water sectors. Energy storage is not a practical solution due to its high investment cost and the low lifetime of battery systems, but the net metering method can lower water cost by up to 13.5%. As a state of the art technology, hybrid systems, combined with operation optimization methods, provide appropriate solutions for cheaper energy; however, wind farms and photovoltaic arrays often face disapproval, especially in the tourism-dependent islands, as local inhabitants find that the local esthetic value is reduced. Environmental reasons and land-use conflicts may also prevent the integration of renewable-energy-powered RO systems. To construct efficiently operated, optimal RO desalination powered by renewable energy systems, the state should promote pilot systems through the relevant studies.
Water deficiency in most of the arid islands is treated with water transfers at the expense of the central government. The operation and maintenance cost of the desalination units is subsidized. In an era of economic crisis, this condition may not be feasible and there is a pressing need to move on toward sustainable solutions. Private sector participation, though not always socially acceptable by the locals, has to be examined. Moreover, as water and energy resources are due to face increasing pressure over the next few decades, the evaluation of trade-offs and the encouragement of cross-sector planning will be critical for further development and for their sustainable management.
The Greek experience indicates that constructing and operating RO units is made possible by state support (in the form of subsidies). The energy intensity of RO units remains high. However, RO technology, although with high cost, is reliable and offers potable water of adequate quality and quantity. Therefore, technical and operation policy challenges (including securing adequate funding) should be addressed so that the implementation of RO technology will further enhance energy efficiency and will be a part of an integrated water management system.
## Author details
When both photovoltaics and wind turbines are installed simultaneously, water cost has an
Figure 15. Water cost versus wind and photovoltaic installed power for Hermoupolis when both technologies exists
energy. This value can be achieved for various installed power combinations of each technology. The lower values for both the photovoltaics and wind turbine-installed power (which consequently gives the lower installation costs) are for 1250kWp and 1250 kW, respectively
In both cases, an increase in water demand during the next years has not been taken into account; even so, the installation of renewable energy technologies will be beneficial and will make water cheaper due to higher utilization of the desalination units and due to cheap energy sources.
The practice of desalination has been identified as a promising solution to the water scarcity problem of the Greek islands (Cyclades and Dodecanese). The early efforts during the 1960s with the use of solar distillation technology were not successful, as the technical know-how was very limited and maintenance was inadequate. Also, the operational costs (labor) were very high due to the lack of automation technologies. The advances in desalination technology, combined with the need to meet water demand in the rapidly developing tourist areas, enhanced the prospects of desalination as a supplementary or even primary source of potable
, which is 0.17 €/m<sup>3</sup> less than the case with conventional
absolute minimum value at 1.09 €/m<sup>3</sup>
7. Discussion and conclusions
(Figure 15).
simultaneously.
182 Desalination
George Arampatzis\*, Avraam Kartalidis and Dionysis Assimacopoulos
\*Address all correspondence to: arab@chemeng.ntua.gr
Environmental and Energy Management Research Unit, School of Chemical Engineering, National Technical University of Athens, Athens, Greece
## References
[1] Marsh DM. The water-energy nexus: A comprehensive analysis in the context of New South Wales [dissertation]. Sydney: University of Technology Sydney; 2008
[16] Cuenca J. Report on water desalination status in the Mediterranean Countries. Murcia:
Responding to Water Challenges Through Desalination: Energy Considerations
http://dx.doi.org/10.5772/intechopen.69956
185
[17] Dagkalidis A. Greek sectorial studies—study on desalination. Athens: Piraeus Bank; 2009 [18] Dalmira A. Desalination Unit in Schoinousa [Internet]. [Updated: 2013]. Available from: www.koinignomi.gr/news/politiki/politiki-kyklades/2013/12/27/monada-afalatosis-gia-ti-
[19] Kampourakis N. Deyath Project Report 2011-12 [Internet]. 2013. Available from: www. thira.gr/nomika-proswpa/deya-thiras/16-dimos/nomika-proswpa-synthesi/deya-thiras.
[20] Imerisia. New Desalination Unit from Temak [Internet]. Available from: www.imerisia.
[21] Sychem. Desalination of Seawater in Paros Island [Internet]. 2011. Available from: www. sychem.gr/el/erga/afalatosi-yperdiithisi-erga/item/112-afalatwsh-thalassinoy-neroy-ston-
[22] Ventouris E. Kimolos Desalination Plant [Internet]. 2013. Available from: www.gokimolos.
[23] Karytsas C, Alexandrou V, Boukis. The Kimolos Geothermal Desalination Project [Internet]. 2002 . Available from: pangea.stanford.edu/ERE/db/IGAstandard/record\_detail.
[24] Tsamopoulos M. Cyclades: Desalination System is Rotten in Iraklia [Internet]. 2014. Available from: www.protothema.gr/greece/article/400961/kuklades-katarrei-sustima-
[25] Bundschuh J, Hoinkis J, editors. Renewable Energy Applications for Freshwater Produc-
[26] Syros Water Supply and Sewerage Company. DEYA Syrou Energy Requirements [Inter-
[27] Hellenic Electricity Distribution Network Operator S.A. Informative Report of Energy Production on Non-Interconnected Islands [Internet]. June 2012. Available from: www.deddie. gr/en/miniaia-deltia-ape-kai-thermikis-paragwgis-sta-mi-diasundedemena-nisia/2012 [28] Athens Water Supply and Sewerage Company. Annual Report & Annual Bulletin [Internet]. [Updated: 2011]. Available from: www.eydap.gr/userfiles/Presentations/ependutes/
[29] Georgalas L. Approach of Water-Energy Relation in the Water Supply-Sewage System [Internet]. 2010. Available from: www.nomosphysis.org.gr/articles.php?artid=4237&lang
[30] Thessaloniki Water Supply and Sewerage Company. Annual Bulletin [Internet]. 2011.
Available from: www.eyath.gr/misc/ETISIO\_DELTIO\_2009.zip
gr/article.asp?catid=26519&subid=2&pubid=7421123&tag=9464#
gr/el/component/content/article/15-news/68-dimos-kimolou.html
IMIDA. Murcia: IMIDA; 2002
shoinoysa.html
html?layout=blog
dhmo-paroy
php?id=5365
net]. 2014
=1&catpid=2
afalatosis-stin-irakleia-
tion. Leiden: CRC Press/Balkema; 2012
enimerotika-deltia/etisio\_deltio\_2011.pdf
[16] Cuenca J. Report on water desalination status in the Mediterranean Countries. Murcia: IMIDA. Murcia: IMIDA; 2002
References
184 Desalination
10.1016/j.enpol.2011.08.013
orities/water-energy-nexus
Bank; 2013
Committees/CommitteePresentations.cfm?c=10
Kingdom and New York, NY, USA: C. U. Press; 2014
english/climatology/climatology\_html
Available from: wfd.ypeka.gr
atwsis-anaduomenes-taseis-1
Corfu. 2005
[1] Marsh DM. The water-energy nexus: A comprehensive analysis in the context of New
[2] Scott CA, Pierce AS, Pasqualtti JM, Jones LA, Montz EB, Hoover HJ. Policy and institutional dimensions of the water–energy nexus. Energy Policy. 2011;39(10):6622–6630. DOI:
[3] Schumacher L. Water-Energy Nexus [Internet]. 2007. Available from: www.naruc.org/
[4] EIP-Water. Water-Energy Nexus [Internet]. 2015. Available from: www.eip-water.eu/pri-
[5] IPCC, editor. Climate change 2014: Impacts, adaptation, and vulnerability. Part A: Global and sectoral aspects. Contribution of working Group II to the fifth assessment report of the intergovernmental panel on climate change (Report No. 978-1-107-05807-1). United
[6] International Energy Agency, editor. World Energy Outlook 2012. Paris, France: IEA; 2012 [7] Rodriguez DJ, Delgado A, DeLaquil P, Sohns A. Thirsty Energy. Washington, DC: World
[8] Hellenic Statistical Authority. Announcement of the Results of the 2011 Population Census for the Resident Population [Internet]. 2012. Available from: www.statistics.gr/portal/ page/portal/ESYE/BUCKET/General/A1602\_SAM01\_DT\_DC\_00\_2011\_02\_F\_EN.pdf [9] Hellenic National Meteorological Service. 2015. Available from: www.hnms.gr/hnms/
[10] Nokas H, editors. Estimation and comparative analysis of water production cost in the Cyclades Islands. Syros Island: Water Directorate, Prefecture of Cyclades; 2002
[11] Special Secretariat for Water. Integrated River Basin Management Plans [Internet]. 2013.
[12] Deligiannis E-E, Belesiotis V. Solar application in desalination: The Greek Islands experi-
[13] Vakondios G. Potable Water Production With Desalination [Internet]. 2013. Available from: www.edeya.gr/2013-09-23-10-58-06/2013-09-23-11-11-03/uliko/hmerides-1/texnologies-afal
[14] Dafnis I. Local reverse osmosis units of Corfu municipality. In: Potable Water: Quality Control, Legislation, Problems of Quality and Water Management in Corfu Prefecture;
[15] Arnold J. Operating experiences of a 15,000 cubic meter per day municipal desalting plant at Corfu, Greece. Desalination. 1979;30(9):145–153. DOI: 10.1016/S0011-9164(00)88442-8
ment. Desalination. 1995;100(8)27-34. DOI: 10.1016/0011-9164(96)00006-9
South Wales [dissertation]. Sydney: University of Technology Sydney; 2008
[31] Water Reuse Association – Desalination Committee, editor. Seawater Desalination Power Consumption. USA: Water Reuse Association; 2011
**Chapter 10**
**Importance and Significance of UF/MF Membrane**
The proposed chapter addresses a comprehensive overview of the history and future outlook of ultrafiltration/microfiltration (UF/MF) membrane for desalination water pretreatment. Known theories on UF/MF membrane formation from phase inversion (Dr/wet) systems can be prolonged to define the consequences of high or low molecular weight additives. Also, direct material reengineering and surface modification for high-performance antifouling of UF/MF membranes are also highlighted. Before the modern final polymeric film, the characterization techniques, particularly molecular weight cut-off, pore size, pore size distribution, and microbiological activity classification, on to the UF/MF membrane surface were presented, respectively. Lab scale to commercial scale UF/MF membrane configuration and market size of UF/MF membranes for pretreatment desalination are described. The significance of UF/MF provided here as an unconventional approach for desalination water pretreatment is in contrast with the current conventionally used technologies. The recent development made in the integration of established desalination processes, such as spiral wound reverse osmosis (SWRO), multi-stage flash (MSF), multi-effect distillation (MED), electrodialysis (ED) desalination, and UF pretreatment, is addressed. Finally, the influence of UF/MF on desalination water pretreatment step on to the energy cost of desalination
**Keywords:** desalination, microfiltration, post-treatment, ultrafiltration
© 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
**Systems in Desalination Water Treatment**
Iqbal Ahmed, Khaled S. Balkhair, Muhammad H. Albeiruttye and Amer Ahmed Jamil Shaiban
http://dx.doi.org/10.5772/intechopen.68694
process system is discussed.
**Abstract**
Additional information is available at the end of the chapter
## **Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment**
Iqbal Ahmed, Khaled S. Balkhair, Muhammad H. Albeiruttye and Amer Ahmed Jamil Shaiban
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.68694
#### **Abstract**
[31] Water Reuse Association – Desalination Committee, editor. Seawater Desalination Power
[32] Wilf M. Fundamentals of RO-NF technology. In: International Conference on Desalina-
[33] Vince F, Marechal F, Aoustin E, Breant P. Multi-objective optimization of RO desalination
[34] Lamei A, Van der Zaag P, Von Munch E. Basic cost equations to estimate unit production costs for RO desalination and long-distance piping to supply water to tourism-dominated arid coastal regions of Egypt. Desalination. 2008;225(1-3):1–12. DOI: 10.1016/j.desal.2007.08.
[35] Karagiannis IC, Soldatos PG. Current status of water desalination in the Aegean Islands.
[36] Avlonitis S, Kouroumbas K, Vlachakis N. Energy consumption and membrane replacement cost for seawater RO desalination plants. Desalination. 2003;157(1–3, 1 August
plants. Desalination . 2008;222(1-3):96–118. DOI: 10.1016/j.desal.2007.02.064
Consumption. USA: Water Reuse Association; 2011
Desalination. 2007;203(6). DOI: 10.1016/j.desal.2006.04.006
2003):151–158. DOI: 10.1016/S0011-9164(03)00395-3
tion Costing; Limassol, Cyprus. December 2004
003
186 Desalination
The proposed chapter addresses a comprehensive overview of the history and future outlook of ultrafiltration/microfiltration (UF/MF) membrane for desalination water pretreatment. Known theories on UF/MF membrane formation from phase inversion (Dr/wet) systems can be prolonged to define the consequences of high or low molecular weight additives. Also, direct material reengineering and surface modification for high-performance antifouling of UF/MF membranes are also highlighted. Before the modern final polymeric film, the characterization techniques, particularly molecular weight cut-off, pore size, pore size distribution, and microbiological activity classification, on to the UF/MF membrane surface were presented, respectively. Lab scale to commercial scale UF/MF membrane configuration and market size of UF/MF membranes for pretreatment desalination are described. The significance of UF/MF provided here as an unconventional approach for desalination water pretreatment is in contrast with the current conventionally used technologies. The recent development made in the integration of established desalination processes, such as spiral wound reverse osmosis (SWRO), multi-stage flash (MSF), multi-effect distillation (MED), electrodialysis (ED) desalination, and UF pretreatment, is addressed. Finally, the influence of UF/MF on desalination water pretreatment step on to the energy cost of desalination process system is discussed.
**Keywords:** desalination, microfiltration, post-treatment, ultrafiltration
© 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
## **1. Introduction**
#### **1.1. An early history of UF membrane**
Almost every chemical process involves at least one separation or purification step, and the chemical industry has developed a range of separation techniques to facilitate recovery of the required products. In recent years, membranes and membrane separation techniques have grown from laboratory tool to an industrial process with considerable technical and commercial impact. Instantly, the membrane processes are faster, more efficient, and economical than conventional separation techniques, particularly for desalination water treatment. Among all the membranes process, UF/MF membranes have the largest variety of applications in various industries, because it is a separation technology of high efficiency and low energy consumption [1].
The permeation of water by a thin sheet of animal bladders (diaphragm) was introduced by Abey's Nollet in 1755 (France), and the phenomenon of water permeation was named as osmosis [2]. Later on, Dr. Adolf Eugen Fick from Germany has introduced diffusion law and was developed the first high-pressure synthetic membrane made from nitrocellulose in 1855 [3]. After 50 years, Dr. Bechhold from Germany developed first low-pressure cellulosic membranes (collodion), which is prepared by impregnating filter paper with glacial acetic acid. The first such low-pressure membranes were produced in 1907, and Dr. Bechhold revealed the term "Ultrafilter" collodion membranes. Since for Bechhold's original membranes were introduced and applied protein solutions by forcing at several pressures against to the atmospheric pressure through his collodion membranes [4]. After Dr. Bechhold breakthrough in UF/MF membrane, there has been continuous effort to develop improved UF membranes, which have resulted in many diverse types of such membranes [5]. Further early developments, principally Zsigmondy and Bachmann [6] and Ferry [7] improved on Bechhold's membrane fabrication method, Elford developed graded porosities UF membranes having the properties "Gradocol" [8]. By the early 1930s, microporous cellulosic membranes such as cellulose nitrate, cellulose di/triacetate were commercially available [9, 10]. With further growth of UF/MF synthetic cellulosic membrane during 1950s, the synthetic membrane technology was commonly available for removal of bacteria, virus, dextrin, protein from water, in addition to salinated water cleansing, respectively. But, the little flux was the main drawback of such Bechhold's type membranes. These deficiencies put together are too costly and practically inappropriate. The period of cost-effective feasible membrane advancement, which was started in the late 1950s and prolongs to this date, may be divided into two time periods. The first generation was from 1959 to 1970 of cellulose acetate integral asymmetric membranes, and the second generation started from 1971 to 1984 of noncellulosic asymmetric membranes [10, 11].
thin-film-composite (TFC) polymeric membranes started during 1963, initiated by a research institute and one of its first employees, Peter S. Francis [11, 12]. A significant discovery was made in the art of thin-film-composite membranes by Cadotte in 1970 with the beginning of large-scale commercial polymeric membrane. Cadotte invented two innovative techniques of TFC membrane based on interfacial polymerization and solution coating methods [12]. During 1970–1990, the researchers developed important methods of membrane materials synthesis, membrane fabrication process, membrane geometry, separation, and purification processing techniques. Also during 1985, the synthetic commercial membrane entered into a new era, and by the end of 1990, the MF, UF/dialysis, nanofiltration (NF), RO, ED, and gas separation membranes technology have grown steadily at the industrial level. Moreover, by the end of 19th century, the advancement of membrane growth has enhanced performance, steadiness, and provided lower operating costs, making membranes the preferred technology in the water treatment industry as well as in the food and pharmaceutical industries [1, 10, 13].
**Table 1.** Key historical development of UF membrane from Market laboratory - scale to commercial scale [10, 11].
**Inventor Development Year** Bechold Prepares collodion membranes of graded pore
Filter GmbH Commercializes ultrafiltration membranes 1926
Reid and Breton Selection of cellulosic material for membrane making
Amicon by Koch Market laboratory-scale UF membranes develop
Abcor by Koch Installs commercial tubular UF plant (electro coat paint)
Abcor by Koch First commercially significant ceramic membrane
polysulfone, PVDF membranes
Amicon by Koch Make first UF hollow fiber membrane 1967
Romicon by Koch Introduces hollow fiber capillary UF plants 1973 Abcor by Koch Commercializes spiral wound UF modules 1980
Loeb and Surirajan Cellulosic acetate integral-asymmetric membranes
ultrafilter
Zsibmondy and Bachmann
size measure bubble point and use the term
Patent collodion filter (German Patent 329-060) 1918
1906
http://dx.doi.org/10.5772/intechopen.68694
189
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
1959
1960
1966
1969
1988
UF/MF membranes [14, 15] have a porous barrier structure that retains components by a sieving mechanism and used for separation of solutes from the feed solution. The water flux in MF and UF is proportional to the applied pressure (Darcy's law). A pressure gradient (0.5–5 bar) across the membrane transports the solvent through, while the substances larger than the pores are rejected. It is agreed that the important process involved in UF/MF is one of sieve action, tricky by adsorption and other consequences are emerging from the unusually large ratio of
**1.2. Theoretical background of UF/MF**
The significant development in artificial membrane technology began in the 1960s and to be headed toward the growth of new UF/MF membranes materials. These events allowed for wider industrial and commercial application of UF/MF [10, 11]. In 1962 Loeb and Surirajan developed a new method of polymeric membrane fabrication, called dry/wet phase inversion process. After the breakthrough of phase inversion the history of the synthetic membrane were entirely changed which was active properties regarding mechanical strength, membrane morphologies, and ten times higher performance than the earlier membrane. **Table 1** shows the first contribution to UF membrane development [10, 11]. The beginning of
**Table 1.** Key historical development of UF membrane from Market laboratory - scale to commercial scale [10, 11].
thin-film-composite (TFC) polymeric membranes started during 1963, initiated by a research institute and one of its first employees, Peter S. Francis [11, 12]. A significant discovery was made in the art of thin-film-composite membranes by Cadotte in 1970 with the beginning of large-scale commercial polymeric membrane. Cadotte invented two innovative techniques of TFC membrane based on interfacial polymerization and solution coating methods [12]. During 1970–1990, the researchers developed important methods of membrane materials synthesis, membrane fabrication process, membrane geometry, separation, and purification processing techniques. Also during 1985, the synthetic commercial membrane entered into a new era, and by the end of 1990, the MF, UF/dialysis, nanofiltration (NF), RO, ED, and gas separation membranes technology have grown steadily at the industrial level. Moreover, by the end of 19th century, the advancement of membrane growth has enhanced performance, steadiness, and provided lower operating costs, making membranes the preferred technology in the water treatment industry as well as in the food and pharmaceutical industries [1, 10, 13].
#### **1.2. Theoretical background of UF/MF**
**1. Introduction**
188 Desalination
**1.1. An early history of UF membrane**
asymmetric membranes [10, 11].
Almost every chemical process involves at least one separation or purification step, and the chemical industry has developed a range of separation techniques to facilitate recovery of the required products. In recent years, membranes and membrane separation techniques have grown from laboratory tool to an industrial process with considerable technical and commercial impact. Instantly, the membrane processes are faster, more efficient, and economical than conventional separation techniques, particularly for desalination water treatment. Among all the membranes process, UF/MF membranes have the largest variety of applications in various industries, because it is a separation technology of high efficiency and low energy consumption [1].
The permeation of water by a thin sheet of animal bladders (diaphragm) was introduced by Abey's Nollet in 1755 (France), and the phenomenon of water permeation was named as osmosis [2]. Later on, Dr. Adolf Eugen Fick from Germany has introduced diffusion law and was developed the first high-pressure synthetic membrane made from nitrocellulose in 1855 [3]. After 50 years, Dr. Bechhold from Germany developed first low-pressure cellulosic membranes (collodion), which is prepared by impregnating filter paper with glacial acetic acid. The first such low-pressure membranes were produced in 1907, and Dr. Bechhold revealed the term "Ultrafilter" collodion membranes. Since for Bechhold's original membranes were introduced and applied protein solutions by forcing at several pressures against to the atmospheric pressure through his collodion membranes [4]. After Dr. Bechhold breakthrough in UF/MF membrane, there has been continuous effort to develop improved UF membranes, which have resulted in many diverse types of such membranes [5]. Further early developments, principally Zsigmondy and Bachmann [6] and Ferry [7] improved on Bechhold's membrane fabrication method, Elford developed graded porosities UF membranes having the properties "Gradocol" [8]. By the early 1930s, microporous cellulosic membranes such as cellulose nitrate, cellulose di/triacetate were commercially available [9, 10]. With further growth of UF/MF synthetic cellulosic membrane during 1950s, the synthetic membrane technology was commonly available for removal of bacteria, virus, dextrin, protein from water, in addition to salinated water cleansing, respectively. But, the little flux was the main drawback of such Bechhold's type membranes. These deficiencies put together are too costly and practically inappropriate. The period of cost-effective feasible membrane advancement, which was started in the late 1950s and prolongs to this date, may be divided into two time periods. The first generation was from 1959 to 1970 of cellulose acetate integral asymmetric membranes, and the second generation started from 1971 to 1984 of noncellulosic
The significant development in artificial membrane technology began in the 1960s and to be headed toward the growth of new UF/MF membranes materials. These events allowed for wider industrial and commercial application of UF/MF [10, 11]. In 1962 Loeb and Surirajan developed a new method of polymeric membrane fabrication, called dry/wet phase inversion process. After the breakthrough of phase inversion the history of the synthetic membrane were entirely changed which was active properties regarding mechanical strength, membrane morphologies, and ten times higher performance than the earlier membrane. **Table 1** shows the first contribution to UF membrane development [10, 11]. The beginning of
UF/MF membranes [14, 15] have a porous barrier structure that retains components by a sieving mechanism and used for separation of solutes from the feed solution. The water flux in MF and UF is proportional to the applied pressure (Darcy's law). A pressure gradient (0.5–5 bar) across the membrane transports the solvent through, while the substances larger than the pores are rejected. It is agreed that the important process involved in UF/MF is one of sieve action, tricky by adsorption and other consequences are emerging from the unusually large ratio of pore length to pore diameter in all pressure-driven membranes. The principle of the pure water flux into the UF/MF membranes is due to the capillary pore diffusion model and the mean pore radius can be calculated by Hagan-Poiseuille equation [15, 16],
$$\mathcal{F}^\* = \left(\frac{8J\_\mu \Delta x \, H\_p}{\mathcal{E}}\right)^{\frac{1}{2}}\tag{1}$$
such as thickness, pore size, pore radius, pore volume, pore density, and tortuosity, respectively. However, Sakai successfully investigated the qualitative attempts to account for tortu-
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
is present water content, which approximates the pore volume in the membrane.
*ε* . *dp* <sup>2</sup> .∆*PT* \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_
When Eq. (3) is used to calculate the mean pore radius, it becomes complicated because of the introduction of tortuosity, and the result is larger than that obtained from Eq. (1). Unfortunately, it is impossible to determine the pore sizes of asymmetric membranes with the aid of Hagen Poiseuille equation (3), which makes it necessary to use, for this purpose, the data of more complicated methods [23]. Among all above methods, the most informative are the means of electron microscopy, gas pycnometer, which give the possibility to determine
**2. Significant development of low-fouling UF/MF membranes lab scale to**
Synthetic polymeric membranes can be divided into hydrophobic and hydrophilic classifications, and structure can also be classified. Structural classification is critical because it is the structure which determines the separation mechanisms and the membrane application [10, 28, 29]. Membranes can be further classified as symmetric or asymmetric [10, 29]. The symmetric membranes can be porous, cylindrical porous, and homogeneous (nonporous). The asymmetric membranes can be porous, microporous with top layer, and composite that is consisting of a porous substrate with a dense top layer. The thickness of the top layer in asymmetric membranes is in the range of 0.1–0.5 μm and is supported on a porous sub-layer with a thickness of about 50–150 μm [10, 11, 29]. The development of pressure-driven membrane technology was began after Loeb-Surirajan and Riley et al. [30, 31] and Cadott discovered
i. Ability to fabricate particular selective membrane that has high permeation, essentially ultrathin, dense layer, surface barrier layer, integrally supported by a thick, porous,
ii. Ability to form polymeric membranes into compact, high-surface-area, economical mem-
iii. TFC or ultra thin film (UTF) able to be laminated on microporous or porous support layer.
Modern membrane technology began in late 1990s, the development of polymeric membrane chemistry and processing techniques are used in membrane fabrication. With the developments
⁄∆*<sup>a</sup>* > 1 (4)
http://dx.doi.org/10.5772/intechopen.68694
191
<sup>32</sup> <sup>∆</sup> *<sup>x</sup>* . *<sup>μ</sup>* ) (5)
osity (τ) in the pore model. Tortuosity τ perhaps described the ratio of pore as,
*<sup>p</sup>* = (
pore sizes and pore size distributions (PSD) of asymmetric membranes [18, 24–27].
*<sup>τ</sup>* <sup>=</sup> Mean pathh length \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_ Membrane thickness <sup>=</sup> *<sup>L</sup>* ⁄∆*<sup>x</sup>* <sup>=</sup> *<sup>W</sup><sup>C</sup>*
about the three following significant developments [12],
spongy structure and was able to produce at commercial scale.
*J*
where *Wc*
**commercial scale**
brane configuration.
whereas \_\_ *r* is a mean pore radius, *Jμ* shows pure water viscosity, *∆x* is membrane thickness, *Hp* is hydraulic pressure, and *ε* shows membrane porosity.
However,
$$H\_p = \,\_{l\_p}/\text{map} \tag{2}$$
whereas, *Jp* shows water permeation, and TMP presents transmembrane pressure (∆P).
Combine Eq. (2) into Eq. (1),
$$\tau = \left(\frac{8\,\mathrm{J}\_{\mu}\Delta x\,\mathrm{TMP}\,\tau}{\varepsilon\,\mathrm{J}\_{\mathrm{p}}}\right)^{\frac{1}{2}}\tag{3}$$
The leading theory of fluid flow through UF/MF membrane in an ideal condition such as consistently sized pores in the membrane, negligible concentration polarization, no fouling, respectively. **Figure 1** summarized the theory of transport phenomenon and fluid dynamics of an UF membrane. Also the pore size, some other factors such as interactions between UF feed components and membrane matrix play a significant role in the transport through the membrane [10, 17–19].
Sakai [19] reported that the Eq. (3) is directly relating to the membrane structure. Sakai and co-worker also indicated that Verniory et al. improved the UF/MF membrane transport parameters. Among membranologist, it has been widely proven that UF/MF membranes flux and rejection are depended upon their structure. Nakao [19–22] has reported that in the case of known relation between flux and rejection, the membrane structure can be characterized
**Figure 1.** Predicting flux from pore statistics using Hagen-Poiseuille Equation [17].
such as thickness, pore size, pore radius, pore volume, pore density, and tortuosity, respectively. However, Sakai successfully investigated the qualitative attempts to account for tortuosity (τ) in the pore model. Tortuosity τ perhaps described the ratio of pore as, *<sup>τ</sup>* <sup>=</sup> Mean pathh length \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_ Membrane thickness <sup>=</sup> *<sup>L</sup>* ⁄∆*<sup>x</sup>* <sup>=</sup> *<sup>W</sup><sup>C</sup>*
pore length to pore diameter in all pressure-driven membranes. The principle of the pure water flux into the UF/MF membranes is due to the capillary pore diffusion model and the mean pore
> \_\_1 2
> > \_\_1 2
*<sup>p</sup>* ⁄*TMP* (2)
*r* is a mean pore radius, *Jμ* shows pure water viscosity, *∆x* is membrane thickness,
(1)
(3)
*r* = ( 8 *J <sup>μ</sup>* <sup>∆</sup> *<sup>x</sup> Hp* \_\_\_\_\_\_\_\_\_\_\_\_\_ *<sup>ε</sup>* )
whereas, *Jp* shows water permeation, and TMP presents transmembrane pressure (∆P).
8 *J <sup>μ</sup>* <sup>∆</sup> *<sup>x</sup> TMP <sup>τ</sup>* \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_ *ε J <sup>p</sup>* )
The leading theory of fluid flow through UF/MF membrane in an ideal condition such as consistently sized pores in the membrane, negligible concentration polarization, no fouling, respectively. **Figure 1** summarized the theory of transport phenomenon and fluid dynamics of an UF membrane. Also the pore size, some other factors such as interactions between UF feed components and membrane matrix play a significant role in the transport through the
Sakai [19] reported that the Eq. (3) is directly relating to the membrane structure. Sakai and co-worker also indicated that Verniory et al. improved the UF/MF membrane transport parameters. Among membranologist, it has been widely proven that UF/MF membranes flux and rejection are depended upon their structure. Nakao [19–22] has reported that in the case of known relation between flux and rejection, the membrane structure can be characterized
*<sup>r</sup>* <sup>=</sup> (
**Figure 1.** Predicting flux from pore statistics using Hagen-Poiseuille Equation [17].
radius can be calculated by Hagan-Poiseuille equation [15, 16],
*Hp* is hydraulic pressure, and *ε* shows membrane porosity.
\_\_
*Hp* = *<sup>J</sup>*
Combine Eq. (2) into Eq. (1),
membrane [10, 17–19].
\_\_
whereas \_\_
190 Desalination
However,
$$
\tau = \frac{\text{Mean path length}}{\text{Membrance thickness}} = \text{ } \%\_{\text{lx}} = \text{ } \%\_{\text{ha}} > 1\tag{4}
$$
where *Wc* is present water content, which approximates the pore volume in the membrane.
$$J\_p = \left(\frac{\varepsilon \cdot d\_p^2 \,\Delta P\_r}{32\,\Delta \ge \mu}\right) \tag{5}$$
When Eq. (3) is used to calculate the mean pore radius, it becomes complicated because of the introduction of tortuosity, and the result is larger than that obtained from Eq. (1). Unfortunately, it is impossible to determine the pore sizes of asymmetric membranes with the aid of Hagen Poiseuille equation (3), which makes it necessary to use, for this purpose, the data of more complicated methods [23]. Among all above methods, the most informative are the means of electron microscopy, gas pycnometer, which give the possibility to determine pore sizes and pore size distributions (PSD) of asymmetric membranes [18, 24–27].
## **2. Significant development of low-fouling UF/MF membranes lab scale to commercial scale**
Synthetic polymeric membranes can be divided into hydrophobic and hydrophilic classifications, and structure can also be classified. Structural classification is critical because it is the structure which determines the separation mechanisms and the membrane application [10, 28, 29]. Membranes can be further classified as symmetric or asymmetric [10, 29]. The symmetric membranes can be porous, cylindrical porous, and homogeneous (nonporous). The asymmetric membranes can be porous, microporous with top layer, and composite that is consisting of a porous substrate with a dense top layer. The thickness of the top layer in asymmetric membranes is in the range of 0.1–0.5 μm and is supported on a porous sub-layer with a thickness of about 50–150 μm [10, 11, 29]. The development of pressure-driven membrane technology was began after Loeb-Surirajan and Riley et al. [30, 31] and Cadott discovered about the three following significant developments [12],
Modern membrane technology began in late 1990s, the development of polymeric membrane chemistry and processing techniques are used in membrane fabrication. With the developments in polymeric membrane materials, manufacturing technologies, and water treatment processing systems have made this technology an efficient, economical for water treatments, and competitive with traditional water treatment methods [10, 15, 29]. Recently, UF/MF membranes have been succeeded for a range of industrial applications [32]. Each application enforces precise specifications on the membrane material and membrane structure. The revolution in understanding the origin of these structural elements of Loeb and Surirajan phase inversion process was obtained by Wienk et al. and Wu et al. [33, 34],
The latest development revealed "next generation" of membrane materials for UF/MF focus onto:
> Commercial scale polymeric UF/MF membrane systems run in a continuous cross-flow mode, where backwash and cleaning operation interchanges with the normal operation. The decrease in the flux for pure water from cycle to cycle, because of fouling, the flux decline within a period due to concentration polarization, and the average flux under steady state level. The latter is also decreasing from cycle to cycle, suggests irreversible solute adsorption or fouling [41–43]. And of the solute retained on a membrane surface leads to increasing permeate flow resistance at the membrane wall region. Strategies to minimize the effect of fouling can be divided into two groups: avoidance and remediation. The remediation is to clean up by the cleanup process, is usually done by chemical cleaning at regular times [44], and this is necessary for all membrane processes in nearly all applications. However, large differences in the cleaning frequency can be found, ranging from daily to yearly, depending on the concentration of foulant and the pretreatment. A large number of cleaning agents are commercially available. The choice of optimal product depends on feed characteristics. Acid cleaning
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
http://dx.doi.org/10.5772/intechopen.68694
193
used to remove adsorbed organics. Nearly, all cleaning products contain detergents. Another remediation technique often used in UF/MF is backwashing or back pulsing. A short pulse of water or air from the permeate side to the feed side efficiently removes all fouls blocking the membrane pores [45]. This principle is often applied in a dead-end or semi-dead-end filtration. It is possible to avoid fouling by using adequate pretreatments, such as coagulation
Recently, UF/MF membranes become an innovative and dominant technology and have been extensively used in many areas, including wastewater treatment, protein separation, dialysis, and dairy industry [47]. However, the most common applications of UF/MF in downstream processing are protein concentration (i.e., solvent removal), buffer exchange and desalting, virus removal and clarification [41]. Since, the improvements of UF/MF technology to make membrane water treatments economically competitive with traditional water treatment methods
, whereas alkaline cleaning is
is suitable for the removal of precipitated salts, such as CaCO<sup>3</sup>
**Figure 2.** Typical cross-flow process with asymmetric membrane [17].
precipitation, or slow sand filtration [10, 15, 29, 46].
**2.1. Material selection for UF/MF fabrication**
Also, the development is involved in both organic and interpenetrating multiphase structures with excellent transport properties, agreement to allow membranes with superior chemical/ thermal stability, fouling resistance, organic solvent resistance, and unusually high permselectivities and permeabilities [35]. Such kind of polymeric membranes may well circumvent many of these limitations. Similarly, recent developments in UF/MF membrane module design, including rotational membrane devices and cycled flow fluid management for fouling control, use of low-cost refractory monoliths as membrane supports, and use of electric potentials to minimize the fouling rate on to the membrane surface [24]. Today, almost 98% of cross-flow membrane systems installations use polymeric UF/MF membranes [13, 36].
Principally, almost all commercial membrane lifespan faces two serious issues due to natural phenomena during separation (solutes) and purification (fluxes/permeate) such as concentration polarization (solutes) and fouling. A typically asymmetric polymeric membrane as shown in **Figure 2** has random pore sizes. Therefore, concentration polarization plugging up on membrane pores and fouling is happened due to the living microorganism adhesion, gel layer formation, and solute adhesion at the membrane surface (see **Figure 1**) [37, 38]. Several researchers and manufacturers have revealed that the natural phenomenon is responsible for restricting the permeate flux during cross-flow (i.e., permeation followed by cleaning). Throughout the early stage of filtration process within a cross-flow rotation, concentration polarization is one of the prime causes for flux reductions [39, 40].
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment http://dx.doi.org/10.5772/intechopen.68694 193
**Figure 2.** Typical cross-flow process with asymmetric membrane [17].
in polymeric membrane materials, manufacturing technologies, and water treatment processing systems have made this technology an efficient, economical for water treatments, and competitive with traditional water treatment methods [10, 15, 29]. Recently, UF/MF membranes have been succeeded for a range of industrial applications [32]. Each application enforces precise specifications on the membrane material and membrane structure. The revolution in understanding the origin of these structural elements of Loeb and Surirajan phase inversion
i. Thermodynamic calculation and kinetics of phase separation of polymer/solvent/
ii. The role of additives, both high- and low-molecular weight, on membrane formation.
The latest development revealed "next generation" of membrane materials for UF/MF focus
ii. Innovative processing of polymers for membranes, particular neatness of membrane
iii. Superior functional polymer membranes, qualifying the integration of active barrier
iv. Groundwork of nanoparticles mixed matrix membranes for the synergistic allying of dif-
Also, the development is involved in both organic and interpenetrating multiphase structures with excellent transport properties, agreement to allow membranes with superior chemical/ thermal stability, fouling resistance, organic solvent resistance, and unusually high permselectivities and permeabilities [35]. Such kind of polymeric membranes may well circumvent many of these limitations. Similarly, recent developments in UF/MF membrane module design, including rotational membrane devices and cycled flow fluid management for fouling control, use of low-cost refractory monoliths as membrane supports, and use of electric potentials to minimize the fouling rate on to the membrane surface [24]. Today, almost 98% of
cross-flow membrane systems installations use polymeric UF/MF membranes [13, 36].
Principally, almost all commercial membrane lifespan faces two serious issues due to natural phenomena during separation (solutes) and purification (fluxes/permeate) such as concentration polarization (solutes) and fouling. A typically asymmetric polymeric membrane as shown in **Figure 2** has random pore sizes. Therefore, concentration polarization plugging up on membrane pores and fouling is happened due to the living microorganism adhesion, gel layer formation, and solute adhesion at the membrane surface (see **Figure 1**) [37, 38]. Several researchers and manufacturers have revealed that the natural phenomenon is responsible for restricting the permeate flux during cross-flow (i.e., permeation followed by cleaning). Throughout the early stage of filtration process within a cross-flow rotation, concentration
iii. Theory behind macrovoids porous and nodular structures formation.
i. Well-distinct configuration as 'tailored' membrane materials.
structure with 'customized' approach of interactions.
polarization is one of the prime causes for flux reductions [39, 40].
ferent functions by different polymeric materials.
process was obtained by Wienk et al. and Wu et al. [33, 34],
nonsolvent.
industrialized.
onto:
192 Desalination
Commercial scale polymeric UF/MF membrane systems run in a continuous cross-flow mode, where backwash and cleaning operation interchanges with the normal operation. The decrease in the flux for pure water from cycle to cycle, because of fouling, the flux decline within a period due to concentration polarization, and the average flux under steady state level. The latter is also decreasing from cycle to cycle, suggests irreversible solute adsorption or fouling [41–43]. And of the solute retained on a membrane surface leads to increasing permeate flow resistance at the membrane wall region. Strategies to minimize the effect of fouling can be divided into two groups: avoidance and remediation. The remediation is to clean up by the cleanup process, is usually done by chemical cleaning at regular times [44], and this is necessary for all membrane processes in nearly all applications. However, large differences in the cleaning frequency can be found, ranging from daily to yearly, depending on the concentration of foulant and the pretreatment. A large number of cleaning agents are commercially available. The choice of optimal product depends on feed characteristics. Acid cleaning is suitable for the removal of precipitated salts, such as CaCO<sup>3</sup> , whereas alkaline cleaning is used to remove adsorbed organics. Nearly, all cleaning products contain detergents. Another remediation technique often used in UF/MF is backwashing or back pulsing. A short pulse of water or air from the permeate side to the feed side efficiently removes all fouls blocking the membrane pores [45]. This principle is often applied in a dead-end or semi-dead-end filtration. It is possible to avoid fouling by using adequate pretreatments, such as coagulation precipitation, or slow sand filtration [10, 15, 29, 46].
#### **2.1. Material selection for UF/MF fabrication**
Recently, UF/MF membranes become an innovative and dominant technology and have been extensively used in many areas, including wastewater treatment, protein separation, dialysis, and dairy industry [47]. However, the most common applications of UF/MF in downstream processing are protein concentration (i.e., solvent removal), buffer exchange and desalting, virus removal and clarification [41]. Since, the improvements of UF/MF technology to make membrane water treatments economically competitive with traditional water treatment methods [10, 29, 32], the use of these membranes has increased exponentially for the downstream process of sea water desalination (SWD). Also, the UF/MF membranes employed in SWD have gained significant attention as these methods are efficient in removing the turbidity, the particles, and the microorganisms present in wastewater [10, 29]. With the improvements in this technology to make membrane for separation and purifications economically competitive with traditional separation methods [10, 17, 45], the use of these membranes has increased exponentially.
It has been established that the porous structure and hydrophilicity of UF/MF play crucial roles in membrane manufacturing processes [29, 34, 48]. A suitable porous membrane should be excellent in permeability, hydrophilicity, and chemical resistance to the feed streams. An asymmetric membrane is a good option for high permeability. Thus, currently, much effort is being devoted to improve the performance of the existing membranes regarding anti-fouling properties, high mechanical strength, and excellent chemical resistance. To make a porous or microporous membrane, some mineral or ceramic membranes have been developed. However, polymeric membranes are yet mostly used [49, 50]. Therefore, different polymeric materials have been used for UF/MF membranes and investigated at lab scale to be commercial with changing results, as not all of them produce membranes with suitable performances [50]. Nevertheless, since the first membrane cellulosic and noncellulosic materials were described by Reid and Breton in late 1959 [51], numerous materials have been developed to improve the capacity and performance of membranes filtration [11, 17, 29]. For a given treatment stream, a particular polymeric membrane material can be selected from an assortment of candidates. Till now, there are more than 130 materials (cellulosic, noncellulosic polymers, composite, and inorganic) that have been used to manufacture membranes [11, 17]. The range of materials from which it is possible to create some form of artificial membrane structure is extensive. Each year, number of research papers in polymer and membrane science present many new examples of materials that demonstrate semi-permeable qualities at some scale. However, only a very limited number of these potential candidates make it to the commercial environment [1, 52–54]. Very few materials possess the structural and chemical properties necessary to render them suitable for application in industrial scale membrane processes. **Table 2** shows the various hydrophilic and hydrophobic polymers used for membrane production at lab scale to commercial scale [11, 17, 29, 53]. Furthermore, typical commercial hydrophilic co-polymers are made of polyethylene oxide (PEO), crystallizable polyamide (PA), nylon or aliphatic polyamide (PA6 and PA66), polyurethanes (PU), and polyester (PET). These materials can be used to make a hydrophobic polymer more hydrophilic. Hydrophobic polymeric materials such as PC, PSF, PES, PVDF, PI, PEI, Ar.PA, polyether ether ketone (PEEK), and PAN, are also conventional polymeric materials for the preparation of UF/MF membranes [10, 17, 29, 55].
Among all those materials, CA, PSF, PES, SPES, PAN, and PVDF are the most commonly used polymers for UF/MF membranes at lab scale to commercial scale [1, 17, 29, 52]. Generally, PSF, PES, and polycarbonate (PC), respectively are produced by aromatic bisphenol intermediates such as bisphenol-A-PSF, tetramethyl bisphenol-A polysulfone (TM-PSF), bisphenol-B polysulfone (PSF), and their modified form such as sulfonated polysulfone (SPSF), and sulfonated polyethersulfone (SPES) have been used extensively to fabricate UF/MF [11, 17, 54]. **Figure 3** shows several types of aromatic bisphenol intermediates used for PES, PSF, and polycarbonate (PC) preparation [56]. Typically, all these thermoplastic base materials can easily be dissolved
**Figure 3.** Chemical structure of aromatic bisphenol intermediates of PSf, PES, and PC [55].
**Hydrophilic polymers**
**Hydrophobic polymers**
Poly(viny1 alcohol) PVAL Cellulose and its derivative Poly(viny1 chloride) PVC Cellulose acetate CA Polyamide PA Cellulose triacetate CTA Poly(acry1ic acid) PAA Cellulose acetate butyrate CAB
Polyacrylonitrile PAN Cellulose nitrate CN Poly(viny1 acetate) PVAC Cellulose propionate CP Poly(viny1 butyral) PVB Ethyl cellulose EC Poly (p-hydroxystyrene) PHS Carboxymethyl cellulose CMC
Polysulfone PSF Polytetrafluoroethylene PTFE Polyethersulfone PES Polyethylene PE Poly (vinylidene fluoride) PVDF Silicone Si Polycarbonate PC Polyphenylene oxide PPO Polypropylene PP Polyphenylene sulfide PPS poly(methyl methacrylate) PMMA Polystyrene PS
**Table 2.** Commercial available hydrophilic and hydrophobic polymers for membrane production [11, 17, 29, 53].
propionate
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
CAP
http://dx.doi.org/10.5772/intechopen.68694
195
Poly(ethy1ene oxide) PEOX Cellulose acetate
**Table 2.** Commercial available hydrophilic and hydrophobic polymers for membrane production [11, 17, 29, 53].
[10, 29, 32], the use of these membranes has increased exponentially for the downstream process of sea water desalination (SWD). Also, the UF/MF membranes employed in SWD have gained significant attention as these methods are efficient in removing the turbidity, the particles, and the microorganisms present in wastewater [10, 29]. With the improvements in this technology to make membrane for separation and purifications economically competitive with traditional separation methods [10, 17, 45], the use of these membranes has increased exponentially.
It has been established that the porous structure and hydrophilicity of UF/MF play crucial roles in membrane manufacturing processes [29, 34, 48]. A suitable porous membrane should be excellent in permeability, hydrophilicity, and chemical resistance to the feed streams. An asymmetric membrane is a good option for high permeability. Thus, currently, much effort is being devoted to improve the performance of the existing membranes regarding anti-fouling properties, high mechanical strength, and excellent chemical resistance. To make a porous or microporous membrane, some mineral or ceramic membranes have been developed. However, polymeric membranes are yet mostly used [49, 50]. Therefore, different polymeric materials have been used for UF/MF membranes and investigated at lab scale to be commercial with changing results, as not all of them produce membranes with suitable performances [50]. Nevertheless, since the first membrane cellulosic and noncellulosic materials were described by Reid and Breton in late 1959 [51], numerous materials have been developed to improve the capacity and performance of membranes filtration [11, 17, 29]. For a given treatment stream, a particular polymeric membrane material can be selected from an assortment of candidates. Till now, there are more than 130 materials (cellulosic, noncellulosic polymers, composite, and inorganic) that have been used to manufacture membranes [11, 17]. The range of materials from which it is possible to create some form of artificial membrane structure is extensive. Each year, number of research papers in polymer and membrane science present many new examples of materials that demonstrate semi-permeable qualities at some scale. However, only a very limited number of these potential candidates make it to the commercial environment [1, 52–54]. Very few materials possess the structural and chemical properties necessary to render them suitable for application in industrial scale membrane processes. **Table 2** shows the various hydrophilic and hydrophobic polymers used for membrane production at lab scale to commercial scale [11, 17, 29, 53]. Furthermore, typical commercial hydrophilic co-polymers are made of polyethylene oxide (PEO), crystallizable polyamide (PA), nylon or aliphatic polyamide (PA6 and PA66), polyurethanes (PU), and polyester (PET). These materials can be used to make a hydrophobic polymer more hydrophilic. Hydrophobic polymeric materials such as PC, PSF, PES, PVDF, PI, PEI, Ar.PA, polyether ether ketone (PEEK), and PAN, are also conventional polymeric materials for the preparation
Among all those materials, CA, PSF, PES, SPES, PAN, and PVDF are the most commonly used polymers for UF/MF membranes at lab scale to commercial scale [1, 17, 29, 52]. Generally, PSF, PES, and polycarbonate (PC), respectively are produced by aromatic bisphenol intermediates such as bisphenol-A-PSF, tetramethyl bisphenol-A polysulfone (TM-PSF), bisphenol-B polysulfone (PSF), and their modified form such as sulfonated polysulfone (SPSF), and sulfonated polyethersulfone (SPES) have been used extensively to fabricate UF/MF [11, 17, 54]. **Figure 3** shows several types of aromatic bisphenol intermediates used for PES, PSF, and polycarbonate (PC) preparation [56]. Typically, all these thermoplastic base materials can easily be dissolved
of UF/MF membranes [10, 17, 29, 55].
194 Desalination
**Figure 3.** Chemical structure of aromatic bisphenol intermediates of PSf, PES, and PC [55].
in aprotic solvents and produce membranes with excellent thermal, hydrolytic, and mechanical stability properties in both hot and wet environments. **Figures 4** and **5** show chemical structures of CA, CTA [56, 57], and thermoplastic polymers [17, 58–62], and **Table 3** summarized the leading manufacturer of polymer used for membrane fabrication. Almost more than 90% of membrane manufacturer are producing PSF, PES, TM-PSF, SPSF, SPES, PVDF, and PAN membrane in a wide range of UF/MF applications. To prepare membranes for the liquid separation processes using repeated applications with either hot water or sterilization to keep the membrane clean [11, 17, 63]. Sulfonated polyethersulfone membrane has been customized to be drastically more hydrophilic than standard PES, PSF membranes [29, 42, 58–60, 62, 63]. It is biocompatible and has highest opposition to fouling by hydrophobic compounds such as fats, lipids, anti-foams, and other similar highly fouling substances [1, 10, 32, 42, 60, 64].
**Manufacturer Product name Trade mark WA TM Tg MW**
**BASF** Blend GF45 1010 G9 0.65% 15,500 216 np
**SOLVEY** Udel PSU P-1700 0.30% 2480 174 np
**Solvey Solef** PVDF 6020 0.04% 1700 −40 np
Udel PSU P-1720 0.30% 2480 174 np Udel PSU P-3500 LCD 0.30% 2480 174 np Udel PSU GF-110 0.29% 3720 179 np Udel PSU GF-120 0.29% 3720 180 np Udel PSU GF-130 0.20% 8690 181 np Veradel PES 3000 P 0.50% np np np Veradel PES 201 NT 0.50% 2100 np np
PESU E 1010 NAT 0.80% 2650 222 np PESU E 2010 SW Q31 0.80% 2650 225 np PESU-CF30 E 2010 C6 0.60% 22,000 225 np PESU-GF20 E 2010 G4 0.60% 7300 225 np PESU-GF20 E 2010 G4 MR 0.60% 7301 225 np PESU-GF30 E 2010 G6 0.60% 9800 225 np PESU E 2010 HC 0.80% 2650 225 np PESU E 2010 MR 0.80% 2650 225 np PESU E 2020 P 1% 2650 2250 4800 PESU E 2020 P SR 1% np np 55,000 PESU E 2020 P SR 1% np np 55,001 PESU E 3010 0.80% 2650 228 np PESU E 3010 MR 0.80% 2650 228 np PESU E 6020 P 1% 2650 225 75,000 PESU E 7020 P 1% np 225 92,000 PESU+PTFE KR 4113 1.50% 11,000 225 np PPSU P 3010 1.20% 2270 220 np PPSU P 3010 MR 1.20% 2270 221 np PSU S 2010 0.80% 2550 187 np PSU S 2010 G4 np np np np PSU-GF30 S 2010 G6 0.60% 8900 187 np PSU S 3010 0.80% 2550 187 52,000 PSU S 3010 MR 0.80% 2600 187 np PSU S 6010 0.80% 2550 187 60,000
**% MPa oC g/mol**
http://dx.doi.org/10.5772/intechopen.68694
197
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
**Figure 4.** Chemical structure of commercial CA and CTA [56, 57].
**Figure 5.** Chemical structure of hydrophobic PSF, PES and modified hydrophilic SPSf, SPES [58, 59].
in aprotic solvents and produce membranes with excellent thermal, hydrolytic, and mechanical stability properties in both hot and wet environments. **Figures 4** and **5** show chemical structures of CA, CTA [56, 57], and thermoplastic polymers [17, 58–62], and **Table 3** summarized the leading manufacturer of polymer used for membrane fabrication. Almost more than 90% of membrane manufacturer are producing PSF, PES, TM-PSF, SPSF, SPES, PVDF, and PAN membrane in a wide range of UF/MF applications. To prepare membranes for the liquid separation processes using repeated applications with either hot water or sterilization to keep the membrane clean [11, 17, 63]. Sulfonated polyethersulfone membrane has been customized to be drastically more hydrophilic than standard PES, PSF membranes [29, 42, 58–60, 62, 63]. It is biocompatible and has highest opposition to fouling by hydrophobic compounds such as fats, lipids, anti-foams, and other similar highly fouling substances [1, 10, 32, 42, 60, 64].
**Figure 4.** Chemical structure of commercial CA and CTA [56, 57].
196 Desalination
**Figure 5.** Chemical structure of hydrophobic PSF, PES and modified hydrophilic SPSf, SPES [58, 59].
via functionalization and macromolecule immobilization methods for membrane surface
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
http://dx.doi.org/10.5772/intechopen.68694
199
Morao et al. [69] also reported a comprehensive literature work on surface modification plans that include ion-beam irradiation [70]. Mulder, Pinnau and Freeman said plasma treatment or grafting [29, 66] and UV-induced grafting [67] on a polymeric membrane surface. Several other authors also reported chemical sulfone enrichment [60, 71], chemical dehydrofluorination by alkaline solution [71–73], coating temperature-sensitive polymeric brushes [74], and grafting with pH- and ionic-strength-sensitive polymeric brushes [10, 29, 72, 74]. Moreover, researchers has been used several other techniques such as, the irradiation-induced grafting [75], physical adsorption of water-soluble polymers [76], a formation of Langmuir-Blodgett films [77], thermal grafting of a hydrophilic polymeric surface coating [78, 79], and photografting with UV irradiation [68], respectively. All these surface modification techniques are usually applied on hydrophobic-casted surface and these are complicated and expensive and require at least one additional step in the membrane preparation process [68]. Besides, the physical techniques, all the methods mentioned above allow the membrane surface to be modified without affecting the bulk properties too much when appropriate conditions (the modification time) are selected. Technically the modification of thermoplastic or thermoset types of UF/MF membrane surface by using physical techniques revealed are easy, economical, green, and improved surface properties. However treating polymeric membrane surface by physical method resulting unstable
Typically, several methods have been widely used to fabricate synthetic membranes using all mentioned polymers (See **Table 2**). The principle of synthesis is to transform the polymeric material using a convenient method to achieve a polymeric membrane structure with a significant morphology for a distinct separation. The techniques that are being employed for the preparation of artificial membrane are phase inversion, stretching of films, irradiation and etching of films, track-etching, sintering of powders, sol–gel process, microfabrication vapor deposition, and coating [11, 29, 33, 49, 81]. The ultimate morphology of the membrane film or fibers will deviate significantly based on the properties of the selected polymeric materials and the operational conditions. The majority of membranes at the lab to commercial scale are prepared by controlled phase inversion process, i.e., dry/dry, dry/wet, and wet/wet; yet, preferably, the dry/wet phase inversion techniques are the most useable process currently [1, 29, 33, 49, 81]. Consequently, polymeric membranes are formed by the so-called phase inversion techniques, which include the main steps of Refs. [33, 49, 81] are:
**a.** Dissolving the polymer together with pore-forming additives (either organic or inorganic)
**b.** Flat sheet casting the resulting solution as a thin film on the surface of nonwoven PET
in an aprotic solvent (preferably NMP, DMAc, DMF, DMSO, acetone, etc.).
**d.** Dipping the spun hollow fiber or cast film in a polymer nonsolvent bath.
**c.** The solution is spinning the self-support as a geometry of capillary or hollow fiber.
modification.
fabric.
and mechanical strength drawback [80–82].
**Table 3.** Membrane polymer manufacturer.
#### *2.1.1. Surface chemistry and choice of additives*
The significance of all pressure-driven membranes particularly UF/MF rely on the properties of their surfaces. As reviewed above and has been proven that the thermoplastic polymers specifically PSF, PES, SPES, PVDF, and PAN are dominant materials for membrane making at retail level [10, 17, 63]. However, surface contamination which may lead to deterioration in membrane performance is also known to be governed by the membrane surface properties and obstacle in membrane performance. Therefore, the membronologist has been paid much attention to the membrane surface modification and were identified theoretical and phenomenological reasons behind the hydrophobic reasons of thermoplastic polymers [17, 29, 32, 64, 65]. Also during 50 years, membronologist has developed very innovative methods to modify the hydrophobic membrane surface into hydrophilic. Zeman and Zydney [66] have reported that almost 50% of all MF and UF membranes traded by 1996 were surface-modified. Moreover, many numbers of high cited research manuscripts and books have been published regarding membrane surface modification techniques. Mittal1 has compiled highly cited work in several volumes entitled "Polymer Surface Modification: Relevance to Adhesion." Pinnau and Freeman [65] edited a book of a membrane-based symposium entitled "Advanced Materials for Membrane Separations" [62]. In this book, relevant topics covered in the 2001 ACS Symposium summarize recent advances in various research areas for development of novel materials used in membrane separations. Benham and Kinstle [67] edited ACS symposium entitled "Chemical Reactions on Polymers," the topics included as one of the greatest active fields in polymer science because of its unique ability to produce specialty polymers with desirable chemical and physical properties through modification of readily available polymers [58]. Xu and co-authors [68] published a very comprehensive book on surface engineering of polymer membranes, squeezes those processes which alter membrane surfaces to improve their in-service performance. The book shows the basics of the surface design of UF/MF polymeric membranes, together with membrane surface modification, to minimize fouling, to modulate hydrophilic and hydrophobic. Also, improve biocompatibility, act as a diffusion barrier, provide bio- or chemical functionalities, mimic a biomembrane, fabricate nanostructures, or directly improve the esthetic appearance of the membrane surface. Xu and co-authors [68] also described general techniques of surface modification of membranes via functionalization and macromolecule immobilization methods for membrane surface modification.
Morao et al. [69] also reported a comprehensive literature work on surface modification plans that include ion-beam irradiation [70]. Mulder, Pinnau and Freeman said plasma treatment or grafting [29, 66] and UV-induced grafting [67] on a polymeric membrane surface. Several other authors also reported chemical sulfone enrichment [60, 71], chemical dehydrofluorination by alkaline solution [71–73], coating temperature-sensitive polymeric brushes [74], and grafting with pH- and ionic-strength-sensitive polymeric brushes [10, 29, 72, 74]. Moreover, researchers has been used several other techniques such as, the irradiation-induced grafting [75], physical adsorption of water-soluble polymers [76], a formation of Langmuir-Blodgett films [77], thermal grafting of a hydrophilic polymeric surface coating [78, 79], and photografting with UV irradiation [68], respectively. All these surface modification techniques are usually applied on hydrophobic-casted surface and these are complicated and expensive and require at least one additional step in the membrane preparation process [68]. Besides, the physical techniques, all the methods mentioned above allow the membrane surface to be modified without affecting the bulk properties too much when appropriate conditions (the modification time) are selected. Technically the modification of thermoplastic or thermoset types of UF/MF membrane surface by using physical techniques revealed are easy, economical, green, and improved surface properties. However treating polymeric membrane surface by physical method resulting unstable and mechanical strength drawback [80–82].
*2.1.1. Surface chemistry and choice of additives*
**Table 3.** Membrane polymer manufacturer.
Sumikaexcel PES
**Sumitomo chemicals**
198 Desalination
np: not provided.
The significance of all pressure-driven membranes particularly UF/MF rely on the properties of their surfaces. As reviewed above and has been proven that the thermoplastic polymers specifically PSF, PES, SPES, PVDF, and PAN are dominant materials for membrane making at retail level [10, 17, 63]. However, surface contamination which may lead to deterioration in membrane performance is also known to be governed by the membrane surface properties and obstacle in membrane performance. Therefore, the membronologist has been paid much attention to the membrane surface modification and were identified theoretical and phenomenological reasons behind the hydrophobic reasons of thermoplastic polymers [17, 29, 32, 64, 65]. Also during 50 years, membronologist has developed very innovative methods to modify the hydrophobic membrane surface into hydrophilic. Zeman and Zydney [66] have reported that almost 50% of all MF and UF membranes traded by 1996 were surface-modified. Moreover, many numbers of high cited research manuscripts and books have been published regarding membrane surface modification techniques. Mittal1 has compiled highly cited work in several volumes entitled "Polymer Surface Modification: Relevance to Adhesion." Pinnau and Freeman [65] edited a book of a membrane-based symposium entitled "Advanced Materials for Membrane Separations" [62]. In this book, relevant topics covered in the 2001 ACS Symposium summarize recent advances in various research areas for development of novel materials used in membrane separations. Benham and Kinstle [67] edited ACS symposium entitled "Chemical Reactions on Polymers," the topics included as one of the greatest active fields in polymer science because of its unique ability to produce specialty polymers with desirable chemical and physical properties through modification of readily available polymers [58]. Xu and co-authors [68] published a very comprehensive book on surface engineering of polymer membranes, squeezes those processes which alter membrane surfaces to improve their in-service performance. The book shows the basics of the surface design of UF/MF polymeric membranes, together with membrane surface modification, to minimize fouling, to modulate hydrophilic and hydrophobic. Also, improve biocompatibility, act as a diffusion barrier, provide bio- or chemical functionalities, mimic a biomembrane, fabricate nanostructures, or directly improve the esthetic appearance of the membrane surface. Xu and co-authors [68] also described general techniques of surface modification of membranes
**Manufacturer Product name Trade mark WA TM Tg MW**
3600P 0.43% 2550
**% MPa oC g/mol**
4100P np np np np 4800P np np np np 5200P np np np np 5003P np np np np
> Typically, several methods have been widely used to fabricate synthetic membranes using all mentioned polymers (See **Table 2**). The principle of synthesis is to transform the polymeric material using a convenient method to achieve a polymeric membrane structure with a significant morphology for a distinct separation. The techniques that are being employed for the preparation of artificial membrane are phase inversion, stretching of films, irradiation and etching of films, track-etching, sintering of powders, sol–gel process, microfabrication vapor deposition, and coating [11, 29, 33, 49, 81]. The ultimate morphology of the membrane film or fibers will deviate significantly based on the properties of the selected polymeric materials and the operational conditions. The majority of membranes at the lab to commercial scale are prepared by controlled phase inversion process, i.e., dry/dry, dry/wet, and wet/wet; yet, preferably, the dry/wet phase inversion techniques are the most useable process currently [1, 29, 33, 49, 81]. Consequently, polymeric membranes are formed by the so-called phase inversion techniques, which include the main steps of Refs. [33, 49, 81] are:
### **2.2. Comprehensive characterization of UF/MF membrane**
The most significant methodological problem in engineering and fabricating of pressuredriven membranes is "tailoring" of membrane surface function into hydrophilic or hydrophobic together with anti-fouling properties, uniform structure, and pore forming properties for the selective sieving process [62, 81, 83]. Therefore, the characterization of surface chemistry is an essential tool in membrane science and technology, because it is well known that membrane performance depends not only on feed hydrodynamics and steric hindrances but also on membrane surface (hydrophobicity, hydrophilicity, and membrane surface charge) and membrane-solute(s)-solvent chemical interactions. Moreover, without characterization, the synthesized pressure-driven membrane cannot predict their properties mainly pore size or molecular weight cut-off (MWCO); thus, it is only a tool for membranologist to emphasize the significance of prepared polymeric membranes. Nevertheless, to tailor UF/MF membranes and eventually use the most important polymeric membrane for a final application, most important membrane has to determine the attribute of MWCO, pore size distribution, using independent characterization methods [10, 11, 29]. Among these are:
qualitatively determined by computing the contact angle of a water drop deposited onto the surface, which has commonly been used to assess the wettability and interfacial energy of the substrate surfaces [76]. **Figure 6** summarized several types and most suitable hydrophobic and hydrophilic thermoplastic polymers for membrane constructions [17, 58–61]. In **Figure 6**, among all polymers, sulfonated polymers are hydrophilic form of PSF, and PES and their contact angle almost 50% lower than virgin polymer [85]. Typically, contact angle measurements should be made inside an enclosed chamber to stop airborne particles and establish an equilibrium vapor pressure of the liquid tested, which is especially preferable when the test liquid is volatile. It has been observed that evaporation can cause the fluid front to retract and that a retreating or an average contact angle is recorded unintentionally. However, the inherent inaccuracy of the direct measurement technique and the use of liquids with high boiling points make the enclosed chamber unnecessary in many
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
http://dx.doi.org/10.5772/intechopen.68694
201
Therefore, the contact angle of the surface of the membrane can be measured at ambient conditions using the contact angle to water the produced membranes was measured by the sessile drop method [85–87]. Sample coupons at horizontal dimension (2.5\*6 ± 0.5 cm) should be prepared by cutting the membranes sheet at random locations. Then, the sample should be placed on a glass plate (active surface of the membrane must be upward) and fixed with a double-sided tape. A drop of double-distilled water (5 μl) should be placed on the surface using a microsyringe (Hamilton Company, Reno, NV). The position of the moving bed can be adjusted manually so that the water drop will be fitted to the scale when projected on the screen. After a fixed deposition time (10 sec), the image will be recorded. To measure the contact angle, the height-width method can be used. For the reason that this approach gives reliable data for contact angles less than 90°. The contact angle should be measured at five dif-
ferent spots on each membrane sample so as to ensure reproducibility of data.
**Figure 6.** Membrane hydraulic resistance against pressure with respect to time.
cases.
#### *2.2.1.1. Contact angle*
As reviewed above that the surface hydrophilicity of selected polymer is a major component to predict the membrane affinity with permeate and anti-fouling properties of UF/MF membranes. Hydrophobic polymeric membranes have nonpolar groups and lower surface free energy; this can prevent contact with water and can push out the water molecules adjoining it [36, 75, 84]. The majority of membranologists and manufacturers are using the hydrophobic polymer as a base polymer for UF/MF fabrication. The relative hydrophobicity and hydrophilicity of membrane surface can be qualitatively determined by computing the contact angle of a water drop deposited onto the surface, which has commonly been used to assess the wettability and interfacial energy of the substrate surfaces [76]. **Figure 6** summarized several types and most suitable hydrophobic and hydrophilic thermoplastic polymers for membrane constructions [17, 58–61]. In **Figure 6**, among all polymers, sulfonated polymers are hydrophilic form of PSF, and PES and their contact angle almost 50% lower than virgin polymer [85]. Typically, contact angle measurements should be made inside an enclosed chamber to stop airborne particles and establish an equilibrium vapor pressure of the liquid tested, which is especially preferable when the test liquid is volatile. It has been observed that evaporation can cause the fluid front to retract and that a retreating or an average contact angle is recorded unintentionally. However, the inherent inaccuracy of the direct measurement technique and the use of liquids with high boiling points make the enclosed chamber unnecessary in many cases.
**2.2. Comprehensive characterization of UF/MF membrane**
independent characterization methods [10, 11, 29]. Among these are:
(b) Membrane equilibrium water content (EWC) and porosity
i. Surface functionalization analysis
(b) Fourier transform infrared (FTIR)
(b) Molecular weight cut-off (MWCO) (c) Pore size and pore size distribution
*2.2.1. Surface functionalization analysis*
(a) Contact angle
200 Desalination
(c) X-ray diffraction ii. Structural analysis
(a) AFM and SEM/EDX
iii. Performance evaluation (a) Membrane affinity test
(d) Solute rejection curve
*2.2.1.1. Contact angle*
The most significant methodological problem in engineering and fabricating of pressuredriven membranes is "tailoring" of membrane surface function into hydrophilic or hydrophobic together with anti-fouling properties, uniform structure, and pore forming properties for the selective sieving process [62, 81, 83]. Therefore, the characterization of surface chemistry is an essential tool in membrane science and technology, because it is well known that membrane performance depends not only on feed hydrodynamics and steric hindrances but also on membrane surface (hydrophobicity, hydrophilicity, and membrane surface charge) and membrane-solute(s)-solvent chemical interactions. Moreover, without characterization, the synthesized pressure-driven membrane cannot predict their properties mainly pore size or molecular weight cut-off (MWCO); thus, it is only a tool for membranologist to emphasize the significance of prepared polymeric membranes. Nevertheless, to tailor UF/MF membranes and eventually use the most important polymeric membrane for a final application, most important membrane has to determine the attribute of MWCO, pore size distribution, using
As reviewed above that the surface hydrophilicity of selected polymer is a major component to predict the membrane affinity with permeate and anti-fouling properties of UF/MF membranes. Hydrophobic polymeric membranes have nonpolar groups and lower surface free energy; this can prevent contact with water and can push out the water molecules adjoining it [36, 75, 84]. The majority of membranologists and manufacturers are using the hydrophobic polymer as a base polymer for UF/MF fabrication. The relative hydrophobicity and hydrophilicity of membrane surface can be Therefore, the contact angle of the surface of the membrane can be measured at ambient conditions using the contact angle to water the produced membranes was measured by the sessile drop method [85–87]. Sample coupons at horizontal dimension (2.5\*6 ± 0.5 cm) should be prepared by cutting the membranes sheet at random locations. Then, the sample should be placed on a glass plate (active surface of the membrane must be upward) and fixed with a double-sided tape. A drop of double-distilled water (5 μl) should be placed on the surface using a microsyringe (Hamilton Company, Reno, NV). The position of the moving bed can be adjusted manually so that the water drop will be fitted to the scale when projected on the screen. After a fixed deposition time (10 sec), the image will be recorded. To measure the contact angle, the height-width method can be used. For the reason that this approach gives reliable data for contact angles less than 90°. The contact angle should be measured at five different spots on each membrane sample so as to ensure reproducibility of data.
**Figure 6.** Membrane hydraulic resistance against pressure with respect to time.
#### *2.2.1.2. Fourier transform infrared (FTIR)*
The analysis of polymeric membrane films in transmission using FTIR is particularly a consistent method for assessing the significant properties of polymers due to additives for achieving quantitative information on modified polymer used for the membrane [53, 88]. There are different methods of polymer sample handling used with FTIR, together with diffuse reflectance (DRIFTS), attenuated total reflection (ATR), and correct specular reflectance/reflection absorption. Established methods of sample holding are more helpful than others for particular specimen types. To analyze and achieve a high-quality spectrum of a sample, it is important to know which handling technique works best for the sample type. Since, FTIR scanning is too responsive, therefore, before analyzing membrane samples make sure the sample should be free of contamination. Therefore, before scanning membrane sample should be clean with 1:1 alcohol/DI-water and then dry in nitrogen chamber at 50–60°C (subject to base polymer properties) for 24 h. Typically, almost all ATR accessory equipped with a Zinc selenide (ZnSe) and diamond crystal at a nominal incident angle of 45°, yielding to 12 internal reflections at the membranes surface with 16 scans at 4 cm−1 resolution and ratio to background spectra recorded in the air at abscissa 350–4000 cm−1 [17].
using the contact mode of imaging. The surface roughness morphology at 2D and 3D topographic images can be extracted. The roughness analysis of mean roughness (Ra), the root mean square of data (Rz), and the average difference in the height between the five highest peaks and the five lowest valleys (Ry), as well as regarding the diameter of the nodules [27] of prepared membrane and additives such as BSA, PEG, and PVP, etc. Z is described as the
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
http://dx.doi.org/10.5772/intechopen.68694
203
Membrane imaging with scanning electron microscopy (SEM), transmission electron microscope (TEM), and field-emission scanning electron microscopes (FESEM) has become standard method to investigate membrane morphology or structure and the mechanism of membrane formation [91, 92]. However, all these microscopic samples required one additional step of the metal coating before scanning, but a metalization of the membrane sample is necessary to yield high-quality images and for the visualization of the pore structures [93]. Kim et al. [91], have revealed that polymeric membrane coating is essential at low voltage for conductivity and high voltage for contrast. Also, Kim et al. [91], have revealed that FESEM has proven essential tools for membranologist to examining both surface morphology of finely porous membranes and the fouling process in membrane research. Several researchers [26, 27, 94–96] has combine AFM with SEM together with solute transport data and were determined UF/ MF membrane pore size and pore size distribution. Earlier researchers used AFM to study the UF/MF membrane surface structures [96–99] and to interpret the mechanisms of fouling
To analyze polymeric membrane with SEM/FESEM, the sample should be washed with ion-
phology, the membrane samples must be ruptured in liquid nitrogen [46]. The sample needs to be electroconductive for current, and it can be done by coating them with a very lean layer of about 1.5–2.0 nm using gold or gold/palladium, platinum metals. Also, polymeric membranes sample must be capable of upholding at high vacuum and should not change the space. Polymers, metals, and crystals are usually little problematic and keep their structure
Equilibrium moisture content or water uptake or water swelling is considered to be an essential tool to characterization factor as it indirectly shows the degree of hydrophilicity or hydrophobicity of a membrane [17, 100–102]. EWC is directly related to the porosity, and UF/MF membrane and the porosity play a vital role not only in also characterizing in important in membrane application. Therefore, some researchers have described the porosity and numbers of techniques have been suggested to estimate membrane porosity [35, 103–108]. Some of them give information on the overall membrane porosity. Nevertheless, the porosity of the membrane can be measured by the gravimetric method, and by apparent density method as reported several researchers [11, 100–107]. According to this approach, the average membrane porosity is determined as the overall void fraction, calculated as the volume of the pores divided by the total volume of the membrane. Perfectly dried membrane samples were weighed with a precision balance. Membrane samples, then, immersed in pure water for
C. Also, to get an efficient cross-sectional mor-
difference between the highest and lowest points within the given area (nm).
in membrane processing [85].
in the SEM [89].
ized water and dried in vacuum oven at 50–60<sup>o</sup>
*2.2.2.2. Membrane equilibrium water content (EWC) and porosity*
#### *2.2.1.3. X-ray diffraction (XRD) techniques*
X-ray diffraction method is nondestructive analytical techniques, which reveal information about the crystal structure, chemical composition, and physical properties of materials. Also with the help of XRD, the memranologist can be identified unknown material during membrane modification particularly while we use nanoparticles during dope solution preparation and membrane fabrication and can be compared its crystal structure to that of a standard database [83]. XRD can also be used to identify the presence of multiple phases where different crystalline compounds coexist [89].
#### *2.2.2. Structural analysis*
#### *2.2.2.1. AFM and electron microscopy*
Atomic force microscopy (AFM) is a comparatively a vital tool for membranologist to understand the surface roughness of membrane. As compared to scanning electron microscopy (SEM), AFM is capable of producing images both in air and liquid without conducting the membrane. Besides, the samples do not require any chemical etching or conductive metal coating. However, due to membrane surface roughness, it is often difficult to obtain a pore size distribution. Nevertheless, since the discovery of AFM, it has been used broadly for investigating the surface roughness of UF/MF membranes [26]. Moreover, a significant attribute of AFM is its tendency to measure force interactions as a function to look into surface separation distance [90]. All AFM measurements can be carried out using multimode AFM with a Nanoscope IIIa controller and contact mode OTR8 silicon nitride probes (Olympus, from Japan, Bruker, USA, Agilent, USA). Before scan polymeric membrane samples, the probes should be wiped in argon plasma [90]. Membrane samples can be fixed to the metal sample discs using epoxy resin. Measurements should be carried out under ambient conditions using the contact mode of imaging. The surface roughness morphology at 2D and 3D topographic images can be extracted. The roughness analysis of mean roughness (Ra), the root mean square of data (Rz), and the average difference in the height between the five highest peaks and the five lowest valleys (Ry), as well as regarding the diameter of the nodules [27] of prepared membrane and additives such as BSA, PEG, and PVP, etc. Z is described as the difference between the highest and lowest points within the given area (nm).
*2.2.1.2. Fourier transform infrared (FTIR)*
202 Desalination
recorded in the air at abscissa 350–4000 cm−1 [17].
*2.2.1.3. X-ray diffraction (XRD) techniques*
ent crystalline compounds coexist [89].
*2.2.2.1. AFM and electron microscopy*
*2.2.2. Structural analysis*
The analysis of polymeric membrane films in transmission using FTIR is particularly a consistent method for assessing the significant properties of polymers due to additives for achieving quantitative information on modified polymer used for the membrane [53, 88]. There are different methods of polymer sample handling used with FTIR, together with diffuse reflectance (DRIFTS), attenuated total reflection (ATR), and correct specular reflectance/reflection absorption. Established methods of sample holding are more helpful than others for particular specimen types. To analyze and achieve a high-quality spectrum of a sample, it is important to know which handling technique works best for the sample type. Since, FTIR scanning is too responsive, therefore, before analyzing membrane samples make sure the sample should be free of contamination. Therefore, before scanning membrane sample should be clean with 1:1 alcohol/DI-water and then dry in nitrogen chamber at 50–60°C (subject to base polymer properties) for 24 h. Typically, almost all ATR accessory equipped with a Zinc selenide (ZnSe) and diamond crystal at a nominal incident angle of 45°, yielding to 12 internal reflections at the membranes surface with 16 scans at 4 cm−1 resolution and ratio to background spectra
X-ray diffraction method is nondestructive analytical techniques, which reveal information about the crystal structure, chemical composition, and physical properties of materials. Also with the help of XRD, the memranologist can be identified unknown material during membrane modification particularly while we use nanoparticles during dope solution preparation and membrane fabrication and can be compared its crystal structure to that of a standard database [83]. XRD can also be used to identify the presence of multiple phases where differ-
Atomic force microscopy (AFM) is a comparatively a vital tool for membranologist to understand the surface roughness of membrane. As compared to scanning electron microscopy (SEM), AFM is capable of producing images both in air and liquid without conducting the membrane. Besides, the samples do not require any chemical etching or conductive metal coating. However, due to membrane surface roughness, it is often difficult to obtain a pore size distribution. Nevertheless, since the discovery of AFM, it has been used broadly for investigating the surface roughness of UF/MF membranes [26]. Moreover, a significant attribute of AFM is its tendency to measure force interactions as a function to look into surface separation distance [90]. All AFM measurements can be carried out using multimode AFM with a Nanoscope IIIa controller and contact mode OTR8 silicon nitride probes (Olympus, from Japan, Bruker, USA, Agilent, USA). Before scan polymeric membrane samples, the probes should be wiped in argon plasma [90]. Membrane samples can be fixed to the metal sample discs using epoxy resin. Measurements should be carried out under ambient conditions Membrane imaging with scanning electron microscopy (SEM), transmission electron microscope (TEM), and field-emission scanning electron microscopes (FESEM) has become standard method to investigate membrane morphology or structure and the mechanism of membrane formation [91, 92]. However, all these microscopic samples required one additional step of the metal coating before scanning, but a metalization of the membrane sample is necessary to yield high-quality images and for the visualization of the pore structures [93]. Kim et al. [91], have revealed that polymeric membrane coating is essential at low voltage for conductivity and high voltage for contrast. Also, Kim et al. [91], have revealed that FESEM has proven essential tools for membranologist to examining both surface morphology of finely porous membranes and the fouling process in membrane research. Several researchers [26, 27, 94–96] has combine AFM with SEM together with solute transport data and were determined UF/ MF membrane pore size and pore size distribution. Earlier researchers used AFM to study the UF/MF membrane surface structures [96–99] and to interpret the mechanisms of fouling in membrane processing [85].
To analyze polymeric membrane with SEM/FESEM, the sample should be washed with ionized water and dried in vacuum oven at 50–60<sup>o</sup> C. Also, to get an efficient cross-sectional morphology, the membrane samples must be ruptured in liquid nitrogen [46]. The sample needs to be electroconductive for current, and it can be done by coating them with a very lean layer of about 1.5–2.0 nm using gold or gold/palladium, platinum metals. Also, polymeric membranes sample must be capable of upholding at high vacuum and should not change the space. Polymers, metals, and crystals are usually little problematic and keep their structure in the SEM [89].
#### *2.2.2.2. Membrane equilibrium water content (EWC) and porosity*
Equilibrium moisture content or water uptake or water swelling is considered to be an essential tool to characterization factor as it indirectly shows the degree of hydrophilicity or hydrophobicity of a membrane [17, 100–102]. EWC is directly related to the porosity, and UF/MF membrane and the porosity play a vital role not only in also characterizing in important in membrane application. Therefore, some researchers have described the porosity and numbers of techniques have been suggested to estimate membrane porosity [35, 103–108]. Some of them give information on the overall membrane porosity. Nevertheless, the porosity of the membrane can be measured by the gravimetric method, and by apparent density method as reported several researchers [11, 100–107]. According to this approach, the average membrane porosity is determined as the overall void fraction, calculated as the volume of the pores divided by the total volume of the membrane. Perfectly dried membrane samples were weighed with a precision balance. Membrane samples, then, immersed in pure water for 24 h and weighed again. The overall porosity (ε) was calculated using the following formula (Eq. (6)):
$$
\varepsilon = \frac{(M\_{\text{out}} \cdot M\_{\text{dry}})}{A \,\Delta\_{\text{x}}} \times 100\tag{6}
$$
state after about 6 to 8 h. The resistance of the membrane, *Rm*, can be examined from the slope
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
Following equations can be used to calculate the membrane resistance and membrane com-
*<sup>w</sup>* <sup>=</sup> \_\_\_\_\_\_\_ <sup>∆</sup>*Wp*
Where *Rm* shows membrane hydrodynamic resistance and can be determined by the slope of the water vs transmembrane pressure (*∆P*) difference graph, *πm* is the applied pressure on the
is the osmotic pressure, and *∆W<sup>p</sup>*
difference, ∆t is an interval of time, A is the membrane active surface area, and ρ is the density
Technically, UF/MF is a pressure-driven process designed to remove large macromolecule (>0.001 μm for UF and >0.1 for MF) from selected solution (see **Figure 8a**). MWCO is a pore characteristic of membranes and is related to rejection for a given molecular weight of a solute and the value frequently used by membrane manufacturers to described their porous UF/MF membranes. Dextran, PVP, PEG, and BSA of a range of molecular weights are often used to rate the MWCO of UF/MF membranes [11, 17, 29]. For example, a membrane that can remove dissolved solids with molecular weights of 150+ has a molecular weight cut-off of 150. **Figure 8b** [17] shows a retention vs. molecular mass curve and membranes with this particular MWCO
, where ∆ *P* = *π<sup>m</sup>* - *π<sup>b</sup>* (10)
*<sup>ρ</sup><sup>A</sup>* <sup>∆</sup> *<sup>t</sup>* <sup>∆</sup> *<sup>P</sup>* (11)
http://dx.doi.org/10.5772/intechopen.68694
205
is the permeate weight. The
of water flux vs. (*∆P*) graph, and **Figure 7** shows *Rm* vs. flux with respect to time.
*<sup>w</sup>* <sup>=</sup> \_\_\_ ∆*P Rm*
paction against pressure [11, 102]:
*J*
membrane (driving force), *π<sup>b</sup>*
of permeate.
*J*
*2.2.3.2. Molecular weight cut-off (MWCO)*
**Figure 7.** Classification of solute separation and MWCO of UF/MF [17].
where *M*wet is the wet membranes, *M*dry is the weight of dry membranes, The volumetric porosity,
$$\varepsilon = \frac{\mathbf{V}\_{\text{e}}}{\mathbf{V}\_{\text{w}}} = \frac{\mathbf{V}\_{\text{w}} \cdot \mathbf{V}\_{\text{d}}}{\mathbf{V}\_{\text{w}}} = \mathbf{1} \cdot \frac{\mathbf{V}\_{\text{d}}}{\mathbf{V}\_{\text{i}}} \tag{7}$$
$$
\varepsilon = 1 \cdot \frac{\rho\_m}{\rho\_p} \tag{8}
$$
where *Ve* is the empty volume of the membrane, *Vd* is the dry volume of the membrane, *Vw* is the wet volume of the membrane, *ρm* shows membrane density, and *ρ<sup>p</sup>* is the density of base polymer used for UF/MF membranes fabrication.
Porosity via pycnometer calculation as follows,
Porosity via pycnometer calculation as follows,\\
$$
\varepsilon = \frac{\left[M\_{dry} + P\_l \cdot \left(P\_l + M\_{wo}\right)\right]}{\rho\_w V\_t} \tag{9}
$$
Where *M*dry is the weight of dry membrane, *Pl* is the weight of pycnometer totally filled with the pure liquid, such as pure water, kerosene, toluene, etc., and *Pl* + *M*wet is the weight of pycnometer filled with liquid and membrane. The density of membrane shows *ρm* [17, 103]
#### *2.2.3. Performance evaluation*
#### *2.2.3.1. Membrane affinity test*
Typically, the primary characterization of any membrane is their performance and rest of all above characterization are depended on and to support the performance. Theoretically, the permeate (flux) of any substance per unit driving force which is "transmembrane pressure" is directly proportional to the permeability of the material. Before evaluating UF/MF membrane for protein or waste water treatment, it is critical to examine the membrane affinity test of the pure water. Thus, in membrane affinity, the hydraulic resistance (*Rm*) is one of the important properties to study the membrane compatibility with clean water against transmembrane pressure (*∆P*) with respect to time [11, 95]. To characterize the membrane hydraulic resistance (*Rm*) and compaction, the pure water (de-ionized water) flux of membranes can calculate at different transmembrane pressures (*∆P*) later than compaction. Yasuda and Tsai [100], revealed that after earlier time, the initial flux would be started after the pressurization of a test cell. The de-ionized water permeates usually dropped severely in the starting and flattened steady state after about 6 to 8 h. The resistance of the membrane, *Rm*, can be examined from the slope of water flux vs. (*∆P*) graph, and **Figure 7** shows *Rm* vs. flux with respect to time.
Following equations can be used to calculate the membrane resistance and membrane compaction against pressure [11, 102]:
$$J\_w = \frac{\Delta P}{R\_n}, \text{ where } \Delta P = \pi\_w \text{--}\pi\_b \tag{10}$$
$$J\_w = \frac{\Delta W\_p}{\rho A \Delta t \,\Delta P} \tag{11}$$
Where *Rm* shows membrane hydrodynamic resistance and can be determined by the slope of the water vs transmembrane pressure (*∆P*) difference graph, *πm* is the applied pressure on the membrane (driving force), *π<sup>b</sup>* is the osmotic pressure, and *∆W<sup>p</sup>* is the permeate weight. The difference, ∆t is an interval of time, A is the membrane active surface area, and ρ is the density of permeate.
#### *2.2.3.2. Molecular weight cut-off (MWCO)*
24 h and weighed again. The overall porosity (ε) was calculated using the following formula
*A* ∆*<sup>x</sup>*
<sup>=</sup> <sup>V</sup><sup>w</sup> \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_ - <sup>V</sup><sup>d</sup> V<sup>w</sup>
> \_\_\_\_\_\_\_\_ m ρp
\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_ *ρ<sup>m</sup> Vt*
nometer filled with liquid and membrane. The density of membrane shows *ρm* [17, 103]
Typically, the primary characterization of any membrane is their performance and rest of all above characterization are depended on and to support the performance. Theoretically, the permeate (flux) of any substance per unit driving force which is "transmembrane pressure" is directly proportional to the permeability of the material. Before evaluating UF/MF membrane for protein or waste water treatment, it is critical to examine the membrane affinity test of the pure water. Thus, in membrane affinity, the hydraulic resistance (*Rm*) is one of the important properties to study the membrane compatibility with clean water against transmembrane pressure (*∆P*) with respect to time [11, 95]. To characterize the membrane hydraulic resistance (*Rm*) and compaction, the pure water (de-ionized water) flux of membranes can calculate at different transmembrane pressures (*∆P*) later than compaction. Yasuda and Tsai [100], revealed that after earlier time, the initial flux would be started after the pressurization of a test cell. The de-ionized water permeates usually dropped severely in the starting and flattened steady
<sup>=</sup> <sup>1</sup>- \_\_\_\_\_\_\_\_ <sup>V</sup><sup>d</sup> Vt
where *M*wet is the wet membranes, *M*dry is the weight of dry membranes,
Ve V<sup>w</sup> × 100 (6)
is the dry volume of the membrane, *Vw* is
is the weight of pycnometer totally filled with
is the density of base
+ *M*wet is the weight of pyc-
(7)
(8)
(9)
(Eq. (6)):
204 Desalination
where *Ve*
The volumetric porosity,
*<sup>ε</sup>* <sup>=</sup> (*M*wet - *<sup>M</sup>*dry) \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_
*<sup>ε</sup>* <sup>=</sup> <sup>1</sup>- <sup>ρ</sup>
polymer used for UF/MF membranes fabrication. Porosity via pycnometer calculation as follows,
Where *M*dry is the weight of dry membrane, *Pl*
*2.2.3. Performance evaluation*
*2.2.3.1. Membrane affinity test*
is the empty volume of the membrane, *Vd*
*<sup>ε</sup>* <sup>=</sup> [*M*dry <sup>+</sup> *Pl* - ( *Pl* <sup>+</sup> *<sup>M</sup>*wet) ]
the pure liquid, such as pure water, kerosene, toluene, etc., and *Pl*
the wet volume of the membrane, *ρm* shows membrane density, and *ρ<sup>p</sup>*
*ε* = \_\_\_
Technically, UF/MF is a pressure-driven process designed to remove large macromolecule (>0.001 μm for UF and >0.1 for MF) from selected solution (see **Figure 8a**). MWCO is a pore characteristic of membranes and is related to rejection for a given molecular weight of a solute and the value frequently used by membrane manufacturers to described their porous UF/MF membranes. Dextran, PVP, PEG, and BSA of a range of molecular weights are often used to rate the MWCO of UF/MF membranes [11, 17, 29]. For example, a membrane that can remove dissolved solids with molecular weights of 150+ has a molecular weight cut-off of 150. **Figure 8b** [17] shows a retention vs. molecular mass curve and membranes with this particular MWCO
**Figure 7.** Classification of solute separation and MWCO of UF/MF [17].
**Figure 8.** A typical rejection curve, variable shows UF membranes with and without (0%) additives [17].
would be appropriated for applications. Theoretically, UF and MF have the same chemistry and same phenomenon except for pore size difference. **Figure 8a** [17] shows a pore size or MWCO differences between UF/MF membranes. Typically, UF membrane has pore sizes in the range from 10 to 1000 Å and is capable of retaining species in the molecular weight range of 300 to 500,000 Daltons, while MF pore size is usually almost 100 times higher than UF. Typically, the value of MWCO is defined as the molecular weight (MW) which is rejected by 90%. The molecular weight has a linear relationship with the pore radius or pore size of a membrane [10, 11, 25–27, 108]. In general, the MWCO of a membrane is determined by the identification of an inert solute, which has the lowest molecular weight and has a solute rejection of 80–100% in steady state UF experiments. The permeate solutes (bovine serum albumin or polyethylene glycol or PVP) concentration can be determined using UV-vis spectrophotometer at a wavelength of 280 nm. Theoretically, solute rejection performance for UF/MF membranes is usually illustrated as the observed amount of rejection, and it can be defined as,
$$\text{Solute rejection (R)} = \left[1 \text{-} \frac{\text{C}\_p}{\text{C}\_f}\right] \times 100\tag{12}$$
Whereas, *Cp* and C<sup>f</sup> are the solute concentration in the flux and feed, respectively. In UF/MF macromolecular solutions, the concentration polarization phenomenon always happened. As a result of this phenomenon, the solute concentration at the membrane surface *Cs* is higher than for UF/MF membranes; therefore, it is expressed as the true rejection, *Rt* , and it can be evaluated as:
$$R\_r = \begin{bmatrix} 1 \cdot \frac{C\_r}{C\_n} \end{bmatrix} \tag{13}$$
The value of *Cm* can be calculated using the boundary layer resistance (BLR) model for the
*<sup>w</sup>* <sup>=</sup> \_\_\_\_\_\_ <sup>∆</sup>*<sup>P</sup>*
where *Rb1* is the hydrodynamic resistance of the boundary layer to the flow of solvent defined
∫
0 *δ*
If the relationship of transmembrane pressure as a function of solute concentration is known,
Typically, the average pore size, pore volume, pore size distribution, pore density, pore geometry, and properties of UF/MF porous membranes are imperative. Also, it is vital to the development of new kind of membrane [19, 21, 23] but also a great help in selecting and using membrane correctly and fast in the application. Kaneko [106] and Nakao [21, 22] has comprehensively reviewed and described various kind of pore size and pore size distribution methods. Sakai was also discussed and provided a general description of the concept of pore size and pore size distribution method for UF/MF. Calvo and co-authors [104] have reported comprehensive reviews on almost all pore size and pore size distribution methods used for UF/MF. However, recently among all the methods, the membranologist is using true solute rejection data and MWCO (solute rejection curve) for pore size and pore size distribution
The variations of solute rejection with solute molecular diameter yield and shaped curve as illustrated in **Figure 9** (MWCO) and **Figure 10** (log-normal). **Figure 9** suggests that the relation between solute rejection and solute diameter is described by log-normal probability function as reported by several authors [22–27] and yields a straight line on a log-normal probability graph as shown in **Figure 10**. If the solute rejection correlates with solute diameter
by the log-normal probability function, this relationship can be expressed as:
√ \_\_\_ 2*π* ∫
∞ *y e*-*u*<sup>2</sup> <sup>⁄</sup><sup>2</sup>
<sup>=</sup> *<sup>π</sup>* \_\_\_\_\_\_\_\_ *<sup>b</sup> μ<sup>O</sup> Rb*<sup>1</sup>
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
*<sup>w</sup>* <sup>=</sup> *<sup>π</sup><sup>m</sup>* - *<sup>π</sup>* \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_ *<sup>b</sup> μ<sup>O</sup> Rb*<sup>1</sup>
*<sup>μ</sup><sup>O</sup>*(*Rm*+*Rb*1) (14)
http://dx.doi.org/10.5772/intechopen.68694
*Ps* (*x*)-<sup>1</sup> *dx* (16)
*du* (17)
(15)
207
permeate flux. In BLR, the permeate flux is expressed as the following Eq. (14):
*J*
*J*
*Rb*<sup>1</sup> =
*2.2.3.3. Pore size and pore size distribution*
then the value of *Cm* can be determined from Eqs. (14) and (15).
Insert Eq. (10) in Eq. (14)
as:
method.
*2.2.3.4. Solute rejection curve*
*Rt* <sup>=</sup> \_\_\_1
The value of *Cm* can be calculated using the boundary layer resistance (BLR) model for the permeate flux. In BLR, the permeate flux is expressed as the following Eq. (14):
$$J\_{\nu} = \frac{\Delta P}{\mu\_{O(\mathbb{R}\_{\ast} \times \mathbb{R}\_{\ast})}} \tag{14}$$
Insert Eq. (10) in Eq. (14)
$$J\_w = \frac{\pi\_w \cdot \pi\_b}{\mu\_o \cdot R\_{\text{st}}} = \frac{\pi\_b}{\mu\_o \cdot R\_{\text{st}}} \tag{15}$$
where *Rb1* is the hydrodynamic resistance of the boundary layer to the flow of solvent defined as:
$$R\_{b1} = \int\_0^b P\_s(\mathbf{x})^{-1} d\_x \tag{16}$$
If the relationship of transmembrane pressure as a function of solute concentration is known, then the value of *Cm* can be determined from Eqs. (14) and (15).
#### *2.2.3.3. Pore size and pore size distribution*
would be appropriated for applications. Theoretically, UF and MF have the same chemistry and same phenomenon except for pore size difference. **Figure 8a** [17] shows a pore size or MWCO differences between UF/MF membranes. Typically, UF membrane has pore sizes in the range from 10 to 1000 Å and is capable of retaining species in the molecular weight range of 300 to 500,000 Daltons, while MF pore size is usually almost 100 times higher than UF. Typically, the value of MWCO is defined as the molecular weight (MW) which is rejected by 90%. The molecular weight has a linear relationship with the pore radius or pore size of a membrane [10, 11, 25–27, 108]. In general, the MWCO of a membrane is determined by the identification of an inert solute, which has the lowest molecular weight and has a solute rejection of 80–100% in steady state UF experiments. The permeate solutes (bovine serum albumin or polyethylene glycol or PVP) concentration can be determined using UV-vis spectrophotometer at a wavelength of 280 nm. Theoretically, solute rejection performance for UF/MF membranes is usually
**Figure 8.** A typical rejection curve, variable shows UF membranes with and without (0%) additives [17].
1⁻ *Cp* \_\_\_
macromolecular solutions, the concentration polarization phenomenon always happened. As a result of this phenomenon, the solute concentration at the membrane surface *Cs* is higher than for
*Cp* \_\_\_
are the solute concentration in the flux and feed, respectively. In UF/MF
*Cf*] <sup>×</sup> <sup>100</sup> (12)
*Cm*] (13)
, and it can be evaluated as:
illustrated as the observed amount of rejection, and it can be defined as,
UF/MF membranes; therefore, it is expressed as the true rejection, *Rt*
Solute rejection (*R*) <sup>=</sup> [
*Rt* = [1⁻
Whereas, *Cp*
206 Desalination
and C<sup>f</sup>
Typically, the average pore size, pore volume, pore size distribution, pore density, pore geometry, and properties of UF/MF porous membranes are imperative. Also, it is vital to the development of new kind of membrane [19, 21, 23] but also a great help in selecting and using membrane correctly and fast in the application. Kaneko [106] and Nakao [21, 22] has comprehensively reviewed and described various kind of pore size and pore size distribution methods. Sakai was also discussed and provided a general description of the concept of pore size and pore size distribution method for UF/MF. Calvo and co-authors [104] have reported comprehensive reviews on almost all pore size and pore size distribution methods used for UF/MF. However, recently among all the methods, the membranologist is using true solute rejection data and MWCO (solute rejection curve) for pore size and pore size distribution method.
#### *2.2.3.4. Solute rejection curve*
The variations of solute rejection with solute molecular diameter yield and shaped curve as illustrated in **Figure 9** (MWCO) and **Figure 10** (log-normal). **Figure 9** suggests that the relation between solute rejection and solute diameter is described by log-normal probability function as reported by several authors [22–27] and yields a straight line on a log-normal probability graph as shown in **Figure 10**. If the solute rejection correlates with solute diameter by the log-normal probability function, this relationship can be expressed as:
$$R\_{\iota} = \frac{1}{\sqrt{2\pi}} \int\_{u}^{y} e^{-u^{\ast}} \, d\_{\iota} \tag{17}$$
(18), a straight line in the form of solute yield among R (solute separation in %) and *Ds* (solute
) = *A* + *B*(ln *D<sup>s</sup>*
Where, *A* and *B* are the intercept and the slope, respectively. From this log-normal plot, mean solute size (*Ds*) can be calculated as *Ds* corresponding to *R* =ˆ50% can be determined from the ratio of *Ds* at *R* =ˆ84.13% and 50% [26, 27, 106]. By ignoring the dependence of solute separation on the steric and hydrodynamic interaction between solute and pore sizes [25, 26], the
sidered to be the same as that of solute mean size and solute geometric standard deviation. From Eqs. (18) and (19) and the tables of statistics that were presented by Michaels [25], μ
*<sup>s</sup>* = (ln *D<sup>s</sup>*
)*Rt*
To minimize the operational in desalination process, beach well-intake pretreatment is used [10, 14, 28, 32, 36, 46, 109]. However, this is not always precisely promising and not liable to headway due to operating cost. Thus, the pressure-driven membrane processes such as MF, UF, and NF are now the new drift in deceitful n RO or MSF pretreatment systems [46]. Microfiltration (MF) is an open technique for the removal of suspended solids and for lowering the silt density index (SDI). Energy consumption in MF is relatively small so that the total costs for the MF pretreatment are comparable to the beach well intake [83, 84, 109–112]. Whereas, the cost for a corresponding conventional pretreatment is more than double. MF provides an RO feed water of high quality, with (slightly) lower COD/BOD, and smaller SD1 in comparison to the untreated seawater, although there is a significant influence on the feed water quality [111–114]. Good quality seawater may be used for large SWRO plants with a
To minimize the energy cost, scaling inhibitor, and fouling lessening, respectively for the useful of the RO, and MSF treatment practice of MF and UF optimizes only the pretreatment given lower capital and operating costs, or on a wider variety of sources [83]. Besides, the implementation of NF as a pretreatment, on the other hand, will lead to a breakthrough in the application of RO or MSF because it has implications for the desalination process itself and not only on the quality of the feed water. Turbidity, microorganisms, and hardness are removed in the NF unit, as well as a fraction of the dissolved salts [14, 109]. The worldwide market for UF membranes has moving parallel with RO, grown to nearly \$3.3 billion in 2016 from \$3.1 billion in 2015. The market is expected to increase at a 5-year compound annual
<sup>=</sup>0.5 = -*A*/
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
) (19)
http://dx.doi.org/10.5772/intechopen.68694
209
) of the membrane can be con-
*B* (20)
*<sup>B</sup>* (21)
diameter) on a log-normal,
are evaluated as:
*<sup>μ</sup>* <sup>=</sup> ln *<sup>D</sup>*¯
*<sup>σ</sup><sup>g</sup>* <sup>=</sup> (*lnDs*)*Rt*=0.84
minimal pretreatment and at relatively low cost.
**2.3. UF/MF membrane configuration and components**
and σg
*R*(*Rt*
mean pore size (*Ds*) and the geometric standard deviation (σg
**Figure 9.** A typical rejection curve on log normal probability graph, variable shows PES UF membranes with LiBr (1%-5%) and without additives (0% LiBr) [17].
**Figure 10.** Recent status of commercial HF membranes [63, 108].
$$y = \frac{\ln D\_s \cdot \ln \vec{D}\_s}{\ln \sigma\_x} = \frac{\ln D\_s \cdot \mu}{\ln \sigma\_x} \tag{18}$$
Where *D<sup>s</sup>* is the solute diameter, *μ* is the geometric mean diameter of solute at *Rt* = 50%, and *σ<sup>g</sup>* is the geometric standard deviation about the mean diameter. According to the Eqs. (17) and (18), a straight line in the form of solute yield among R (solute separation in %) and *Ds* (solute diameter) on a log-normal,
$$R\{R\_{\natural}\} = A + B\{\ln D\_{\circ}\} \tag{19}$$
Where, *A* and *B* are the intercept and the slope, respectively. From this log-normal plot, mean solute size (*Ds*) can be calculated as *Ds* corresponding to *R* =ˆ50% can be determined from the ratio of *Ds* at *R* =ˆ84.13% and 50% [26, 27, 106]. By ignoring the dependence of solute separation on the steric and hydrodynamic interaction between solute and pore sizes [25, 26], the mean pore size (*Ds*) and the geometric standard deviation (σg ) of the membrane can be considered to be the same as that of solute mean size and solute geometric standard deviation. From Eqs. (18) and (19) and the tables of statistics that were presented by Michaels [25], μ and σg are evaluated as:
$$\mu = \ln D\_s = \left(\ln D\_s\right)\_{\mathbb{R}^{40.5}} = -A|B\tag{20}$$
$$
\sigma\_{\gamma} = \frac{\left(\ln D\_{\cdot}\right)\_{\mathbb{R}\_{\text{ess}}} \cdot \left(\ln D\right)\_{\mathbb{R}\_{\text{ess}}}}{2} = \frac{1}{B} \tag{21}
$$
#### **2.3. UF/MF membrane configuration and components**
*<sup>y</sup>* <sup>=</sup> *lnD<sup>s</sup>* <sup>⁻</sup> ln *<sup>D</sup>*¯¯ \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_ *<sup>s</sup>*
**Figure 10.** Recent status of commercial HF membranes [63, 108].
(1%-5%) and without additives (0% LiBr) [17].
208 Desalination
Where *D<sup>s</sup>*
ln *σ<sup>g</sup>*
is the solute diameter, *μ* is the geometric mean diameter of solute at *Rt*
is the geometric standard deviation about the mean diameter. According to the Eqs. (17) and
**Figure 9.** A typical rejection curve on log normal probability graph, variable shows PES UF membranes with LiBr
<sup>=</sup> ln *<sup>D</sup><sup>s</sup>* <sup>⁻</sup> *<sup>μ</sup>* \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_ ln *σ<sup>g</sup>*
(18)
= 50%, and *σ<sup>g</sup>*
To minimize the operational in desalination process, beach well-intake pretreatment is used [10, 14, 28, 32, 36, 46, 109]. However, this is not always precisely promising and not liable to headway due to operating cost. Thus, the pressure-driven membrane processes such as MF, UF, and NF are now the new drift in deceitful n RO or MSF pretreatment systems [46]. Microfiltration (MF) is an open technique for the removal of suspended solids and for lowering the silt density index (SDI). Energy consumption in MF is relatively small so that the total costs for the MF pretreatment are comparable to the beach well intake [83, 84, 109–112]. Whereas, the cost for a corresponding conventional pretreatment is more than double. MF provides an RO feed water of high quality, with (slightly) lower COD/BOD, and smaller SD1 in comparison to the untreated seawater, although there is a significant influence on the feed water quality [111–114]. Good quality seawater may be used for large SWRO plants with a minimal pretreatment and at relatively low cost.
To minimize the energy cost, scaling inhibitor, and fouling lessening, respectively for the useful of the RO, and MSF treatment practice of MF and UF optimizes only the pretreatment given lower capital and operating costs, or on a wider variety of sources [83]. Besides, the implementation of NF as a pretreatment, on the other hand, will lead to a breakthrough in the application of RO or MSF because it has implications for the desalination process itself and not only on the quality of the feed water. Turbidity, microorganisms, and hardness are removed in the NF unit, as well as a fraction of the dissolved salts [14, 109]. The worldwide market for UF membranes has moving parallel with RO, grown to nearly \$3.3 billion in 2016 from \$3.1 billion in 2015. The market is expected to increase at a 5-year compound annual growth rate (CAGR) of 6.9% from 2016 to 2021, increasing to nearly \$4.6 billion in 2021 [115, 116]. Thus, the world demand of UF is parallel moving with RO membranes. As described above that majority of membrane manufacturer and supplier currently prefer hollow fiber membrane components for UF and MF application. Recently, MF/UF membrane systems have enjoyed exceptional growth, producing high-quality feed waters to downstream RO systems and as stand-alone filtration methods for a variety of applications. Therefore, the majority of fresh water treatment plants are preferred low-pressure hollow fiber membranes elements configuration [63, 113, 114]. **Figure 10** summarized the recent status of hollow fiber membrane elements as compared to SW setup and flat sheet elements [63, 108]. Several types of HF elements configuration are available commercially depend upon types of water treatment or other applications. **Figure 11** summarized the HF elements for a configuration such bundles, using a simple technique or spiral winding techniques that better shell flow distribution and lower mass transfer resistance [108, 111, 117]. Fiber crimping also has been introduced to reduce mass transfer resistances. The introduction of fluid into the shell region of these devices has received significant attention as it can impact performance significantly [108].
Commercially, available in microfiltration and ultrafiltration, and the tubular membranes operate in tangential, or cross-flow, a configuration where process fluid is pumped along the membrane surface in a full action. Several membrane manufacturer and suppliers provide several offers of strong choice that are effortless to control and clean, serving many industrial and municipal applications [63, 108].The major demand of HF element configuration is usually the strict quality requirement of pharmaceutical, food industry. Besides, HF is also suitable for wastewater treatment, gas separation, etc. **Figure 11** also shows several types of tubular membrane configuration, and **Figure 12** revealed top seller of UF/MF membrane configuration. Tubular membrane elements easy procedure, high suspended solids, and concentrate product competent and often to greater end-point concentration degrees without plugging, making them ideal for recovering wastewaters, and clarifying juices [63, 108, 111].
#### **2.4. Commercial UF/MF materials**
Currently, several membrane configurations are available commercially. **Table 4** lists some of the commercially available membranes for UF [1, 17, 63]. The membranes itself must satisfy some mechanical, hydrodynamic, and economic requirements [17, 63]. Automatic membrane condition means the ability to provide the necessary physical support for the membrane including the capacity to undergo the required pressure drop and any back flushing. Hydrodynamic membrane requirement means minimizing pressure drop through the device and thus reduce pumping costs [52, 118]. The solute mass transfer is optimized, and thus, concentration polarization is reduced. Particulate plugging and dead spots for clean design are minimized. Economic membrane requirements reduce membrane packing densities, manufacturing costs, permit easy access for cleaning/replacement and provide sufficient chemical resistance [45, 52, 63, 110, 115]. Recently, Millipore Corporation from the USA has introduced their new generation of UF membranes. These membranes are based on new casting processes that provide void-free membranes and high process flux with enhanced product retention and significantly increased mechanical resistance. These membranes have been commercialized under the trade names of Biomax (void-free PES family) and Ultracel
(void-free composite regenerated cellulose family). Applied membrane Inc from the USA is a leading supplier of RO and UF membranes systems and water filtration elements under the AMI label for over 25 years. Previously, FILMTEC produced a broad range of CA thin composite membranes ranges for RO. Dow/FILMTEC is currently commercializing UF membrane series M-U4040 PES and MU2540 PAN for pharmaceutical, food beverage, and wastewater
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
http://dx.doi.org/10.5772/intechopen.68694
211
**Figure 11.** Commercial available hollow fiber membrane elements configuration.
treatment [17, 52, 63].
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment http://dx.doi.org/10.5772/intechopen.68694 211
**Figure 11.** Commercial available hollow fiber membrane elements configuration.
growth rate (CAGR) of 6.9% from 2016 to 2021, increasing to nearly \$4.6 billion in 2021 [115, 116]. Thus, the world demand of UF is parallel moving with RO membranes. As described above that majority of membrane manufacturer and supplier currently prefer hollow fiber membrane components for UF and MF application. Recently, MF/UF membrane systems have enjoyed exceptional growth, producing high-quality feed waters to downstream RO systems and as stand-alone filtration methods for a variety of applications. Therefore, the majority of fresh water treatment plants are preferred low-pressure hollow fiber membranes elements configuration [63, 113, 114]. **Figure 10** summarized the recent status of hollow fiber membrane elements as compared to SW setup and flat sheet elements [63, 108]. Several types of HF elements configuration are available commercially depend upon types of water treatment or other applications. **Figure 11** summarized the HF elements for a configuration such bundles, using a simple technique or spiral winding techniques that better shell flow distribution and lower mass transfer resistance [108, 111, 117]. Fiber crimping also has been introduced to reduce mass transfer resistances. The introduction of fluid into the shell region of these devices has received significant attention as it can impact performance significantly [108].
Commercially, available in microfiltration and ultrafiltration, and the tubular membranes operate in tangential, or cross-flow, a configuration where process fluid is pumped along the membrane surface in a full action. Several membrane manufacturer and suppliers provide several offers of strong choice that are effortless to control and clean, serving many industrial and municipal applications [63, 108].The major demand of HF element configuration is usually the strict quality requirement of pharmaceutical, food industry. Besides, HF is also suitable for wastewater treatment, gas separation, etc. **Figure 11** also shows several types of tubular membrane configuration, and **Figure 12** revealed top seller of UF/MF membrane configuration. Tubular membrane elements easy procedure, high suspended solids, and concentrate product competent and often to greater end-point concentration degrees without plugging, making them ideal for recovering wastewaters, and clarifying juices [63, 108, 111].
Currently, several membrane configurations are available commercially. **Table 4** lists some of the commercially available membranes for UF [1, 17, 63]. The membranes itself must satisfy some mechanical, hydrodynamic, and economic requirements [17, 63]. Automatic membrane condition means the ability to provide the necessary physical support for the membrane including the capacity to undergo the required pressure drop and any back flushing. Hydrodynamic membrane requirement means minimizing pressure drop through the device and thus reduce pumping costs [52, 118]. The solute mass transfer is optimized, and thus, concentration polarization is reduced. Particulate plugging and dead spots for clean design are minimized. Economic membrane requirements reduce membrane packing densities, manufacturing costs, permit easy access for cleaning/replacement and provide sufficient chemical resistance [45, 52, 63, 110, 115]. Recently, Millipore Corporation from the USA has introduced their new generation of UF membranes. These membranes are based on new casting processes that provide void-free membranes and high process flux with enhanced product retention and significantly increased mechanical resistance. These membranes have been commercialized under the trade names of Biomax (void-free PES family) and Ultracel
**2.4. Commercial UF/MF materials**
210 Desalination
(void-free composite regenerated cellulose family). Applied membrane Inc from the USA is a leading supplier of RO and UF membranes systems and water filtration elements under the AMI label for over 25 years. Previously, FILMTEC produced a broad range of CA thin composite membranes ranges for RO. Dow/FILMTEC is currently commercializing UF membrane series M-U4040 PES and MU2540 PAN for pharmaceutical, food beverage, and wastewater treatment [17, 52, 63].
**Figure 12.** Top seller of UF/MF membrane configuration [17, 63].
American Membrane Corporation (AMC) was incorporated in Michigan USA in 2001 and established its first subsidiary in China in 2002. AMC has launched the Accupor membrane which is a highly microporous membrane composed of modified hydrophilic PES that is a tough, durable, and temperature-resistant aromatic polymer. This membrane is specifically designed for biological, analytical, electronic, pharmaceutical, beverage, and sterilizing filtration applications [17, 63]. Nitto Denko and Asahi Kasei commercialized Hydracap (hydrophilic PES), UF AP series (Hydrophilic PAN), low-pressure UF hollow fiber membranes for wastewater, pharmaceutical, and food industry. Recently, KOCH has supplied a series of UF HFK-131/138 (PES) and HF8H-72-PMPW (PSF) spiral wound and hollow fiber membranes for wastewater treatment [17, 52, 63]. PALL has introduced BTS highly asymmetric membranes; these membranes are cast from PES and PSF by a unique process exclusive to another supplier. The "cut off" layer in the BTS membranes is only about 10 μm thick, vs. traditional membranes with cut-off layers of about 100–125 μm thickness. This difference in thickness gives the BTS highly asymmetric membrane significantly higher flow with much lower pressure drop for pharmaceutical purposes. Besides that Polymer has also launched a very high-performance asymmetric PSF hollow fiber membrane with uppermost permeation
Note: Polyethersulfone (PES), polysulfone (PSF), polyvinylidene fluoride (PVDF), polyacrylonitrile (PAN), cellulose
**Manufacturer Brand Material & module MWCO (kDa) Flux (L.m−1.h−1) Application**
Nitto Denko, JP Hydracap PES (HF) 150 51–128 at 3 bar Wastewater PALL, USA BTS PES (Cassette) 0.5–10 187.2 at 3 bar Pharmaceutical
Polymem polymem PSF (HF) 6 313.2 at bar Protein
Synder, Canada PES 100 PES (SW) 70 51 at 3 bar Gelatin
acetate (CA), spiral wound (SW), hollow fiber (HF), Japan (JP), note provided (NP).
**Table 4.** Commercially available membranes for ultrafiltration.
Trisep, Canada UE10 PES (SW) 10 2.1 at 2–3 bar Dairy and Food
HFU, HFS, HSU PVDF (HF) 150–200 2.6–8 at 3 bar RO feed water &
CA& PES 5–10 35–97.2 at 1 bar Protein
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
M-series PAN (HF) 0.03–1 μm 4.5–18.2 at 9.3 bar Oil water
purification
213
http://dx.doi.org/10.5772/intechopen.68694
separation
& food
purification
separation
Wastewater
During 50 years, UF/MF membrane technology has experienced a quick growth as it relates to most applications such as pharmaceutical, food, and beverage industries. Also, theoretically, it has been proven that UF/MF is the best post-treatment process (upstream) for RO, eliminating from the feed water nearly all of the possible constituents liable of desalinating membranes fouling. The 30 years of commercialization have viewed new materials of UF/MF membranes
rate at very low pressure for Protein Purification.
**3. Conclusion**
Millipore, USA Biomex &
membrane element
Toray, Switzerland Amicon
Note: Polyethersulfone (PES), polysulfone (PSF), polyvinylidene fluoride (PVDF), polyacrylonitrile (PAN), cellulose acetate (CA), spiral wound (SW), hollow fiber (HF), Japan (JP), note provided (NP).
**Table 4.** Commercially available membranes for ultrafiltration.
tough, durable, and temperature-resistant aromatic polymer. This membrane is specifically designed for biological, analytical, electronic, pharmaceutical, beverage, and sterilizing filtration applications [17, 63]. Nitto Denko and Asahi Kasei commercialized Hydracap (hydrophilic PES), UF AP series (Hydrophilic PAN), low-pressure UF hollow fiber membranes for wastewater, pharmaceutical, and food industry. Recently, KOCH has supplied a series of UF HFK-131/138 (PES) and HF8H-72-PMPW (PSF) spiral wound and hollow fiber membranes for wastewater treatment [17, 52, 63]. PALL has introduced BTS highly asymmetric membranes; these membranes are cast from PES and PSF by a unique process exclusive to another supplier. The "cut off" layer in the BTS membranes is only about 10 μm thick, vs. traditional membranes with cut-off layers of about 100–125 μm thickness. This difference in thickness gives the BTS highly asymmetric membrane significantly higher flow with much lower pressure drop for pharmaceutical purposes. Besides that Polymer has also launched a very high-performance asymmetric PSF hollow fiber membrane with uppermost permeation rate at very low pressure for Protein Purification.
## **3. Conclusion**
**Manufacturer Brand Material & module MWCO (kDa) Flux (L.m−1.h−1) Application**
American Membrane Corporation (AMC) was incorporated in Michigan USA in 2001 and established its first subsidiary in China in 2002. AMC has launched the Accupor membrane which is a highly microporous membrane composed of modified hydrophilic PES that is a
DOW/FLMTEC SFP & SFD PVDF, (HF) 0.03 μm 40–90 at 2.5 bar RO feed water &
AsahiKasei, JP AP PAN (HF) 69 16 Pharmaceutical
KOCH, USA HFK PES, 5–10 24–53 at 3–4 bar Wastewater
L series PES, PSF, PVDF
**Figure 12.** Top seller of UF/MF membrane configuration [17, 63].
(tubular
CA, PES, PSF, PVDF, PAN (Tubular)
M-U4040 PES (SW) 10 4.5–18.2 at 3.2 bar Pharmaceutical
AC 120R01 PES (HF) 15 112 at 4.1 bar Pharmaceutical
HF8H7235PMPW PSF 100 32 at 3 bar PVC separation
5 27–45 Food &
2–200 NP MBR,
MU2540 PAN (SW) 20 50 at 3.5 bar Wastewater
& food
Wastewater
& food
& food
Wastewater
wastewater, Pharmaceutical & food
Applied Membranes (USA)
212 Desalination
AMC (USA/ China)
Luxx Ultratech,
PCI, USA CA, ES, PU, FP,
AN, FP
USA
During 50 years, UF/MF membrane technology has experienced a quick growth as it relates to most applications such as pharmaceutical, food, and beverage industries. Also, theoretically, it has been proven that UF/MF is the best post-treatment process (upstream) for RO, eliminating from the feed water nearly all of the possible constituents liable of desalinating membranes fouling. The 30 years of commercialization have viewed new materials of UF/MF membranes employed at ever lower pressures (1–2 bar) and with remarkable performance. However, UF/ MF technology still has some limits, and yet membrane fouling is a destructive obstacle for UF/MF membranes. Material customization and membrane configuration (modules) for selective application and molecular transport modeling are efficient tools to minimize the fouling phenomenon, which is used in the advancement of membrane technology. Also, UF/MF system design continues to progress, and current movements in MF and UF in certain are a feedback to industry pressures for a basic MF/UF system design that holds a diversity (not clear) of membrane components and system configurations. Consequently, the remarkable UF/MF market growth leads to continuously increasing demand particularly in feed water treatment for SWD and wastewater treatment process and without doubt, UF is recently becoming a reasonable feed water system for RO in an extensive field of raw water quality.
**Technical words Abbreviations Chemical compounds and**
Fourier transform infrared FTIR Polyethylenimine (PEI) or
Molecular weight cut-off MWCO Aromatic polyamide Ar.PA
Equilibrium water content EWC Polyether ether ketone PEEK attenuated total reflection ATR Polyacrylonitrile PAN Atomic force microscope AFM Cellulose acetate CA Scan electron microscopy SEM Cellulose triacetate CTA Energy-dispersive X-ray EDX Trimethyl polysulfone TM-PSF
TEM Sulfonated
Kilo Dalton KDA N-Methyl-2-pyrrolidone NMP Ultraviolet UV Dimethylacetamide DMAc Silt density index SDI Dimethylformamide DMF Carbon oxygen demand COD Dimethyl sulfoxide DMSO Biological oxygen demand BOD Bovine serum albumin BSA Annual growth rate CAGR Polyethylene glycol PEG Hollow fiber HF Polyvinylpyrrolidone PVP
Molecular weight MW Aliphatic polyamide
Field emission scanning electron microscopic
transmission electron microscope
Difference between the highest and lowest points **polymers**
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
(Nylon)
FE-SEM Sulfonated polysulfone SPSF
**Equations**
Mean pore radius ¯¯*r* pure water viscosity *Jμ* Membrane thickness ∆x or ∆a Hydraulic pressure *Hp* Membrane porosity *ε* Water permeation *Jp* Transmembrane pressure TMP (∆P) Tortuosity τ Water content *Wc* Pore volume *Pv* Pore size distributions PSD Mean roughness Ra Root mean square of data *Rz* Highest peaks *Ry*
Dry membranes *Mdry* Empty volume *Ve* Dry volume *Vd* Wet volume *V<sup>w</sup>* Membrane density *ρ<sup>m</sup>* Density of base polymer *ρ<sup>p</sup>* Pure liquid *Pl* Wet membranes *Mwet*
polyethersulfone
*Z* Wet membranes *Mwet*
polyaziridine
**Abbreviations**
http://dx.doi.org/10.5772/intechopen.68694
215
PA6 and PA66
PEI
SPES
## **Acknowledgements**
I gratefully acknowledge the Centre of Excellence in Desalination Technology, Department of Mechanical Engineering King Abdulaziz University and for the kind moral support by providing me platform thus enables me to carry out this chapter.
## **Appendix**
employed at ever lower pressures (1–2 bar) and with remarkable performance. However, UF/ MF technology still has some limits, and yet membrane fouling is a destructive obstacle for UF/MF membranes. Material customization and membrane configuration (modules) for selective application and molecular transport modeling are efficient tools to minimize the fouling phenomenon, which is used in the advancement of membrane technology. Also, UF/MF system design continues to progress, and current movements in MF and UF in certain are a feedback to industry pressures for a basic MF/UF system design that holds a diversity (not clear) of membrane components and system configurations. Consequently, the remarkable UF/MF market growth leads to continuously increasing demand particularly in feed water treatment for SWD and wastewater treatment process and without doubt, UF is recently becoming a
I gratefully acknowledge the Centre of Excellence in Desalination Technology, Department of Mechanical Engineering King Abdulaziz University and for the kind moral support by
**polymers**
UF/MF Polysulfone PSU or PSF
SWRO Zinc selenide ZnSe
**Abbreviations**
reasonable feed water system for RO in an extensive field of raw water quality.
providing me platform thus enables me to carry out this chapter.
**Technical words Abbreviations Chemical compounds and**
Sea water desalination SWD Calcium carbonate CaCO<sup>3</sup>
Multi-stage flash MSF Polyethylene oxide PEO Multi-effect distillation MED Polyamide PA Electrodialysis ED Polyurethanes PU Reverse osmosis RO Polyethylene terephthalate PET Nanofiltration NF Polycarbonate PC
Thin film composite layer TFC Polyethersulfone PES Ultra thin film UTF Polyvinylidene fluoride PVDF Pore size distributions PSD Polyimide PI
**Acknowledgements**
**Appendix**
214 Desalination
See Appendix Table.
Spiral wound reverse
osmosis
Ultrafiltration/ Microfiltration
[5] Allegrezza AE, Burke ET, Winchester. Composite Ultrafiltration Membranes. 1989; United
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
http://dx.doi.org/10.5772/intechopen.68694
217
[6] Zsigmondy R, Bachmann W, Über near filter. Zeitschrift für anorganische und allgeme-
[7] Ferry JD. Ultrafilter membranes, and ultrafiltration. Chemical Reviews. 1936;**18**:373-453. [8] Elford EJ. Principles governing the preparation of membranes having graded porosities. The properties of "Gradocol" membranes as ultrafilters. Trans. Faraday Soc. 1937;**33**:1094
[9] Elford WJ, Grabar P, Ferry JD. Graded Collodion membranes for bacteriological studies. Practical aspects of the mechanism determining the character of the membrane, and the roles of particular solvent constituents. British Journal of Experimental Pathology.
[10] Baker RW. Membrane Technology and Applications. John Wiley & Sons Ltd; 2004,
[11] Lloyd, DR. Materials Science of Synthetic Membrane, Vol. 269. ACS Symposium Series.
[12] Cadotte JE, Petersen BJ. Synthetic membranes. Vol. 153**.** ACS Symposium Series. Washington, 1981. Chapter 21: pp. 305-326. DOI: 10.1021/bk-1981-0153.ch021
[13] Wilf I. New membrane research and development achievements. Desalination and
[14] Bruggen BV, Vandecasteele C, Gestel TV, Doyenb W, Leysenb RA. Review of pressuredriven membrane processes in wastewater treatment and drinking water production.
[15] Boddeker, KW. Liquid Separations with Membranes: An Introduction to Barrier
[16] Gooch, JW. Hagen-Poiseuille equation. Encyclopedic Dictionary of Polymers, Springer.
[17] Ahmed I. High-performance ultrafiltration polyethersulfone membrane. Ph.D. Thesis.
[18] Ahmed A, Idris A, Noordin MY, Rajput R. High-performance ultrafiltration membranes prepared by the application of modified microwave irradiation technique. Industrial &
[19] Sakai K. Determination of pore size and pore size distribution 2. Dialysis membranes.
[20] Verniory A, Du Bois R, Decoodt P, Gassee JP, Lambert PP. Measurement of the permeability of biological membranes application to the glomerular wall. Journal of General
Washington, DC: American Chemical Society: 1985
Interference. Berlin Heidelberg: Springer-Verlag; 2008
2011;**1**:355-355. DOI: 10.1007/978-1-4419-6247-8\_5752
University Technology Malaysia; Malaysia. 2009
Engineering Chemistry Research. 2011;**50**:2272-2283
Journal of Membrane Science. 1994;**96**:91-130
Physiology. 1973;**62**:489-507.
ine Chemie (ZAAC). 1918; **103**:1-256. DOI: 10.1002/zaac.19181030107
States Patent 4,824,568.
1935;**16**:583-587
Water Reuse, 2000;**10**:28-33
Environmental Progress. 2003;**22**:46-56
England
## **Author details**
Iqbal Ahmed1,2\*, Khaled S. Balkhair<sup>1</sup> , Muhammad H. Albeiruttye<sup>1</sup> and Amer Ahmed Jamil Shaiban<sup>1</sup>
\*Address all correspondence to: iqbalmoujdin@gmail.com
1 Center of Excellence in Desalination Technology, King Abdulaziz University, Jeddah, Saudi Arabia
2 Department of Mechanical Engineering, King Abdulaziz University, Jeddah, Saudi Arabia
## **References**
**Author details**
Membrane hydrodynamic
Membrane active surface
Solute concentration in
Solute concentration at membrane surface
Geometric standard
An interval of time *∆t*
resistance
216 Desalination
area
the flux
deviation
Shaiban<sup>1</sup>
Arabia
**References**
Iqbal Ahmed1,2\*, Khaled S. Balkhair<sup>1</sup>
\*Address all correspondence to: iqbalmoujdin@gmail.com
**Technical words Abbreviations Chemical compounds and**
Osmotic pressure *π<sup>b</sup>* Permeate weight *∆W<sup>p</sup>*
Hydrodynamic resistance Rb<sup>1</sup> Solute diameter *D<sup>s</sup>* Geometric mean diameter *μ* Solute rejection *Rt*
Slope *B* Boundary layer resistance BLR
, Muhammad H. Albeiruttye<sup>1</sup>
**polymers**
*Cp* Solute concentration in the feed
*Rm* Applied pressure π<sup>m</sup>
*A* Density of permeate *ρ*
*Cs or Cm* True rejection *Rt*
*σ<sup>g</sup>* Intercept *A*
1 Center of Excellence in Desalination Technology, King Abdulaziz University, Jeddah, Saudi
2 Department of Mechanical Engineering, King Abdulaziz University, Jeddah, Saudi Arabia
[1] Nunes SP, Peinemann KV. Membrane Technology in the Chemical Industry. 2nd ed.
[2] Nollet JA. Investigations on the causes for the ebullition of liquids. Journal of Membrane
[3] Fick A. Ueber difusion. Annalen der Physik. 2006:**170**:59-86. DOI: 0.1002/andp.18551700105. [4] Bechhold H. Kolloidstudien mit der filtration method. Zeitschrift für Elektrochemie und Angewandte Physikalische Chemie. 1907;**13**:509-544. DOI: 10.1002/bbpc.19070133207
WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim; 2006, Chapter 1-3.
Science. 1995;**100**:1-3. DOI: 10.1016/0376-7388(94)00224-M
and Amer Ahmed Jamil
*Cf*
**Abbreviations**
[21] Nakao S, Kimura S. Models of membrane transport phenomena and their applications for ultrafiltration data. Journal of Chemical Engineering of Japan. 1982;**15**:220
metal plating wastewater treatment: A review. Journal of Water Processes Engineering.
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
http://dx.doi.org/10.5772/intechopen.68694
219
[36] Koch Membrane Systems, Inc. Application Bulletin, An Overview of Membrane Technology and Theory. December 1, 2011. Available from: http://www.kochmembrane.
[37] Zheng QZ, Wang P, Yang YN, Cui DJ. The relationship between porosity and kinetics parameter of membrane formation in PSF ultrafiltration membrane. Journal of
[38] Zhang H, Li H, Yang F, Wang T, Gao J, Jiang T. Preparation of cross-linked polyvinyl alcohol nanospheres and the synthesization of low-fouling modified membranes.
[39] Fan LH, Harris JL, Roddick FA, Booker NA. Influence of the characteristics of natural organic matter on the fouling of microfiltration membranes. Water Research. 2001;**35**:4455-4463.
[40] Sun W, Liu J, Chu H, Dong B. Pretreatment and membrane hydrophilic modification to reduce membrane fouling. Membranes. 2013;**3**:226-241. DOI: 10.3390/membranes3030226
[41] Wan X, Ghosh R, Cui Z. Fractionation of proteins using ultrafiltration: Developments and challenges. Developments in Chemical Engineering & Mineral Processing.
[42] Wang H, Yang L, Zhao X, Yu TQ, Du. Improvement of hydrophilicity and blood compatibility on polyethersulfone membrane by blending sulfonated polyethersulfone. Chinese
[43] Wang X, Husson SM, Qian XS, Wickramasinghe R. Inverse colloidal crystal microfiltra-
[44] Zhao X, Liu C. Irreversible fouling control of PVDF ultrafiltration membrane with "fouled surface" for mimetic sewage treatment. Royal Society of Chemistry Advances. 2016;**6**:94184
[45] Arnal JM, García-Fayos B, Sancho M. Membrane Cleaning, Chapter 3, Expanding Issues
[46] Futselaar H, Blankert B, Spanjer T, Spenkelink F. Ultrafiltration used as pre-treatment for SWRO desalination: Dynamic coagulant control under extreme conditions. Desalination
[47] Yu H, Cao Y, Kang G, Liu Z, Kuang W, Liu J, Zhou M. Enhancing the antifouling properties of polysulfone ultrafiltration membranes by the grafting of poly(ethylene glycol) derivatives via surface amidation reactions. Journal of Applied Polymer Science.
[48] Ahmed I, Ani I, Rajput R. Performance of microwave-synthesized dual solvent dope solution and lithium bromide additives on poly(ethersulfone) membranes. Journal of
Chemical Technology & Biotechnology. 2011; **87**:177-188. DOI: 10.1002/jctb.2676
and Water Treatment. 2013;**51**:407-415. DOI: 10.1080/19443994.2012.699259
com/PDFs/Application-Bulletins/KMS-Membrane-Theory.aspx
Separation and Purification Technology. 2011;**77**:162.
Journal of Chemical Engineering. 2009;**17**:324-329.
in Desalination. InTechOpen; 2011, DOI: 10.5772/19760
2015;**132**:41870. DOI: 10.1002/APP.41870
tion membranes. Journal of Membrane Science. 2010;**365**:302-310
2016;**9**:78-110.
2005;**13**:121-136
Membrane Science. 2006;**286**:7-11
metal plating wastewater treatment: A review. Journal of Water Processes Engineering. 2016;**9**:78-110.
[36] Koch Membrane Systems, Inc. Application Bulletin, An Overview of Membrane Technology and Theory. December 1, 2011. Available from: http://www.kochmembrane. com/PDFs/Application-Bulletins/KMS-Membrane-Theory.aspx
[21] Nakao S, Kimura S. Models of membrane transport phenomena and their applications for ultrafiltration data. Journal of Chemical Engineering of Japan. 1982;**15**:220
[22] Nakao, S. Determination of pore size and pore size distribution 3. Filtration membranes.
[23] Cherkasov AN, Polotsky AE. Critical particle-to-pore size ratio in ultrafiltration. Journal
[24] Michaels AS. Membranes, membrane processes, and their applications: Needs, unsolved
[25] Michaels AS. Analysis, and prediction of sieving curves for ultrafiltration membranes: A universal correlation?. Separation Science and Technology. 1980;**15**:1305-1322.
[26] Singh S, Khulbe KC, Matsuura T, Ramamurthy P. Membrane characterization by solute transport and atomic force microscopy Journal of Membrane Science. 1998;1**42**:111-127.
[27] Idris A, Mat Zain N, Noordin MY. Synthesis, characterization, and performance of asymmetric polyethersulfone (PES) ultrafiltration membranes with polyethylene glycol
[28] Winston Ho WS, Sirkar KK. Membrane Hand Book. Springer Science+Business Media
[29] Mulder, M. Basic Principles of Membrane Technology. 2nd ed. Dordrecht, The Netherlands:
[30] Riley RL, Lonsdale HK, Lyons CR, Merten U. Preparation of ultrathin reverse osmosis membranes and the attainment of theoretical salt rejection. Journal of Applied Polymer
[31] Riley RL, Lonsdale HK, Lyons CR, Merten U. Preparation of ultrathin reverse osmosis membranes and the attainment of theoretical salt rejection. Journal of Applied Polymer
[32] Basile SA, Nunes UP. Advanced Membrane Science and Technology for Sustainable
[33] Wienk IM, Boom RM, Beerlage MAM, Bulte AMW, Smolders CA, Strathmann H. Recent advances in the formation of phase inversion membranes made from amorphous or
[34] Wu T, Zhou B, Zhu T, Shi I, Xu Z, Hua C, Wang J. Facile and low-cost approach towards a PVDF ultrafiltration membrane with enhanced hydrophilicity and antifouling performance via graphene oxide/water-bath coagulation. Royal Society of Chemistry
[35] Hosseini SS, Bringas E, Tan NR, Ortiz I, Chahramani M, Shahmirzadi MAA. Recent progress in the development of high-performance polymeric membranes and materials for
semi-crystalline polymers. Journal of Membrane Science. 1996;**113**:361-371
Energy. Woodhead Publishing Limited, Cambridge, UK
Advances. 2015;**5**:7880. DOI: 10.1039/c4ra13476a
of different molecular weights as additives. Desalination. 2007;**207**:324-339.
problems, and challenges of the 1990's. Desalination. 1990;**77**:5-34
Journal of Membrane Science 1994;**96**:131-165.
of Membrane Science. 1995;**106**:161-166
New York, 1992
218 Desalination
Kluwer Academic;1996.
Science. 1967;**11**:2143
Science. 1967;**2**:2143-2158.
[49] Munari S, Bottino A, Camera Roda G, Annell GC. Preparation of ultrafiltration membranes. State of the art. Desalination. 1990;**77**:85-100
[64] Khulbe KC, Feng C, Matsuura T. The art of surface modification of synthetic polymeric
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
http://dx.doi.org/10.5772/intechopen.68694
221
[65] Pinnau I, Freeman BD. Advanced Materials for Membrane Separations. Vol. 876, ACS
[66] Zeman L, Wales M. Steric rejection of polymeric solutes by membranes with uniform
[67] Benham JL, Kinstle JF. Chemical Reactions on Polymers. ACS Symposium Series.
[68] Xu Z, Huang X, Wan L. Surface Engineering of Polymer Membranes. Springer, Zhejiang
[69] Morao A, Escobar IC, Pessoa de Amorim MT, Lopes A, Gonc IC. Postsynthesis modification of a cellulose acetate ultrafiltration membrane for applications in water and waste-
[70] Good K, Escobar I, Xu X, Coleman M., Ponting M. Effects of ion beam irradiation on the transport properties and fouling control of commercial membranes. Desalination.
[71] Ma H, Bowman CN, Davis RH. Membrane fouling reduction by back pulsing and sur-
[72] Pozniak G, Krakewska B, Trochimczuk W. Urease immobilized on modified polysulfone
[73] Bottino A, Capannelli G, Monticelli O, Piaggio P. Poly(vinylidene fluoride) with improved functionalization for membrane production. Journal of Membrane Science. 2000;**166**:23-29.
[74] Bordawekar M, Escobar I, Lipscomb G. Application of a temperature-sensitive polymer brush to reduce fouling on membranes. Separation Science and Technology. 2009;**44**:
[75] Hegazy E, Elrehim HAA, Khalifa NA, Atwa SM, Shawky HA. Anionic/cationic membranes obtained by a radiation grafting method for use in waste water treatment.
[76] Jonsson C, Jonsson A. Influence of the membrane material on the adsorptive fouling of
[77] Kim KJ, Fane AG, Fell CJD. The effect of Langmuir–Blodgett layer pretreatment on the performance of ultrafiltration membranes. Journal of Membrane Science. 1989;**43**:187-204.
[78] Stengaard, FF. Characteristics, and performance of new types of ultrafiltration mem-
[79] Jang H, Song DH, Kim IC, Kwon YN. Fouling control through the hydrophilic surface modification of poly(vinylidene fluoride) membranes. Journal of Applied Polymer
ultrafiltration membranes. Journal of Membrane Science. 1995;**10**:79-88
branes with chemically modified surfaces. Desalination. 1988;**70**:207-224.
membranes. Journal of Applied Polymer Science. 2010;**115**:855-895
American Chemical Society: Washington, DC; 1988.
water treatment. Environmental Progress. 2005;**24**:367-382
face modification. Journal of Membrane Science. 2000;**173**:191-200.
membrane—Preparation and properties. Biomaterials. 1995;**16**:129-134.
University Press, Hangzhou, China; 2007
Polymer International. 1997;**43**:321-332.
Science. 2015;**132**:41712. DOI: 10.1002/APP.41712
2002;**146**:259-264.
3369-3391
Symposium Series. Washington, DC: American Chemical Society; 2004.
pore size distribution. Separation Science and Technology. 1981;**16**:275-290.
[64] Khulbe KC, Feng C, Matsuura T. The art of surface modification of synthetic polymeric membranes. Journal of Applied Polymer Science. 2010;**115**:855-895
[49] Munari S, Bottino A, Camera Roda G, Annell GC. Preparation of ultrafiltration mem-
[50] Ahmad AL, Abdulkarim AA, Ooi BS, Ismail S. Recent development in additives modifications of polyethersulfone membrane for flux enhancement. Chemical Engineering
[51] Reid CE, Breto EJ. Water and ion flow across cellulosic membranes. Journal of Applied
[52] Mohanty K, Purkait MK. Membrane Technologies and Applications (Edited), CRC Press,
[53] Kesting, R, Irvine. Synthetic Polymeric Membranes: A Structural Perspective. Wiley-
[54] AWWA. Microfiltration and Ultrafiltration Membranes for Drinking Water, Manual of
[55] Matsuura, T. Synthetic Membranes, and Membrane Separation Process, Chapter 5. CRC
[56] Sigma-Aldrich, Cellulose acetate (CAS Number: 9012-09-3), 2012. Available from: http://
[57] Sigma-Aldrich. Cellulose acetate (CAS Number: 9004-355-7), 2004. Available from: http://www.sigmaaldrich.com/catalog/search?term=Cellulose+acetate&interface=Prod uct%20Name&N=0+&mode=mode%20matchpartialmax&lang=en®ion=SA&focus=
[58] Kenichi I, Hiroki I, Shouichi Y, C/O Nitto Electric Ind. Co. Ltd. Sulfonated polysulfone composite semipermeable membranes. EP 0165077 B2. 1998; US4818387. 1989.
[59] Li L, Zhang M, Rong M, Ruan W. Studies on the transformation process of PVDF from α to β phase by stretching. Royal Society of Chemistry Advances. 2014;**4**:3938-3943. DOI:
[60] Kim IC, Choi JG, Tak TM. Sulfonated polyethersulfone by heterogeneous method and its membrane performances. Journal of Applied Polymer Science. 1999;**74**:2046-2055
[61] Huang L, Liao M, Yang X, Gong H, Ma L, Zhaoa Y, Huang K. Bisphenol analogs differently affect human islet polypeptide amyloid formation. Royal Society of Chemistry
[62] Li HB, Shi WY, Zhang YF, Liu DQ, Liu XF. Effects of additives on the morphology and performance of PPTA/PVDF in situ blend UF membrane. Polymers. 2014;**6**:1846-1861.
[63] Pearce G. Introduction to membranes: Manufacturers' comparison: Part 1. Filtration &
www.sigmaaldrich.com/catalog/product/sial/22199?lang=en®ion=SA
branes. State of the art. Desalination. 1990;**77**:85-100
Journal. 2013;**223**:246-267. DOI: 10.1016/j.cej.2013.02.130
Taylor & Francis Group, New York, USA. Chapter 2, 2012.
Polymer Science. 1959;**1**:133-143.
220 Desalination
Interscience; 2nd edition, 1985. pp. 237-261
Water Supply Practices-M53. 1st ed. 2005.
10.1039/C3RA45134H
DOI: 10.3390/polym6061846
Separation. 2007;**44**:36-38.
Press, Taylor & Francis Group, Florida, USA; 1994
productN=0%20220003048%20219853286%20219853231
Advances. 2016;**6**:7239. DOI: 10.1039/c5ra21792j
[80] Fang Y, Xu ZK, Wu J. Surface Modification of Membranes. "Encyclopedia of Membrane Science and Technology". John Wiley & Sons Singapore Pte. Ltd.; 2013. DOI: 10.1002/ 9781118522318.emst061
[94] Fritzsche AK, Arerado AR, Moore MD, Hara CO. The surface structure and morphology of polyacrylonitrile membranes by atomic force microscope. Journal of Membrane
Importance and Significance of UF/MF Membrane Systems in Desalination Water Treatment
http://dx.doi.org/10.5772/intechopen.68694
223
[95] Fritzsche AK, Arevalo AR, Connolly AF, Moore MD, Elings V, Wu CM. The structure and morphology of skin of polyethersulfone ultrafiltration membranes: A comparative atomic force microscope and scanning electron microscope study. Journal of Applied
[96] Bottino A, Capannelli G, Grosso A, Monticelli O, Cavalleri O, Rolandi R, Soria R. Surface characterization of ceramic membranes by atomic force microscopy. Journal of
[97] Chahboun A, Coratger R, Ajustron F, Beauvillain J, Aimar A, Sanchez V. Comparative study of micro- and ultrafiltration membranes using STM, AFM, and SEM techniques.
[98] Dietz P, Hansma PK, Inacker O, Lehmann HD, Herrmann KH. Surface pore structures of micro and ultrafiltration membranes imaged with the atomic force microscope. Journal
[99] Miwa T, Yamaki M, Yoshimura H, Ebina S, Nagayama K. Artifacts in atomic force microscopy images of fine particle and protein two-dimensional crystals as evaluated with scanning electron microscopy and simulations. Japanese Journal of Applied Physics. 1992;**31**:L1495.
[100] Kim YW, Kim JJ, Kim YH. Surface characterization of biocompatible polysulfone membranes modified with poly(ethylene glycol) derivatives. Korean Journal of Chemical
[101] Abaticchio P, Boitino A, Roda GC, Capannell G, Munari S. Characterization of ultrafil-
[102] Mahendran R, Malaisamy R, Mohan DR. Cellulose acetate and polyethersulfone blend ultrafiltration membranes. Part I: Preparation and characterizations. Polym. Adv.
[103] Chakrabarty B, Ghoshal AK, Purkait MK. Effect of molecular weight of PEG on membrane morphology and transport properties. Journal of Membrane Science.
[104] Calvo JI, Bottino A, Prádanos P, Palacio L, Hernández A. Porosity, Membrane Materials, Characterization, and Module Design. Encyclopedia of Membrane Science and Technology. John Wiley & Sons, Inc; 2013. DOI: 10.1002/9781118522318.emst065
[105] Yasuda H, Tsai JT. Pore size of microporous polymer membranes. Journal of Applied
[106] Kaneko K. Determination of pore size and pore size distribution 1. Adsorbents and
Science. 1993;**81**:109-120
Polymer Science. 1992;**45**:1945-1956.
Membrane Science. 1994;**95**:289-296.
Ultramicroscopy. 1992;**41**:235-244.
Engineering. 2003;**20**:1158-1165.
Polymer Science. 1974;**18**:805-819
2008;**309**:209-221
tration polymeric. Desalination. 1990;**78**:235-255
Technol. 2004;**15**:149-157. DOI: 0.1002/pat.417
catalysts. Journal of Membrane Science. 1994;**96**:59-89
of Membrane Science. 1992;**65**:101-111.
[94] Fritzsche AK, Arerado AR, Moore MD, Hara CO. The surface structure and morphology of polyacrylonitrile membranes by atomic force microscope. Journal of Membrane Science. 1993;**81**:109-120
[80] Fang Y, Xu ZK, Wu J. Surface Modification of Membranes. "Encyclopedia of Membrane Science and Technology". John Wiley & Sons Singapore Pte. Ltd.; 2013. DOI: 10.1002/
[81] Strathmann H, Koch K. The formation mechanism of phase inversion membranes.
[82] Pham VA, Santerre P, Matsuura T, Narbaitz RM. Application of surface modifying macromolecules in polyethersulfone membranes: Influence on PES surface chemistry and physical properties. Journal of Applied Polymer Science. 1999;**73**:1366-1378. DOI:
[83] Gao W, Liang H, Ma J, Han M, Chen Z, Han Z. Membrane fouling control in ultrafiltration technology for drinking water production: A review. Desalination. 2011;**72**:1-8.
[84] Mehta A, Zydney AL. Permeability and selectivity analysis for ultrafiltration mem-
[85] Goddard JM, Hotchkiss JH. Polymer surface modification for the attachment of bioactive
[86] O'Grady BR. Sulfonated polyethersulfone as a new platform for thin film composite membranes. Master Thesis. University of Connecticut (UConn), United States. 2013
[87] Atae-Allah C, Cabrerizo-Vílchez M, Gomez-Lopera JF, Holgado-Terriza1 JA, Roman-Rold R, Luque-Escamilla PL. Measurement of surface tension and contact angle using
[88] Bai H, Wang X, Zhou Y, Zhan L. Preparation and characterization of poly(vinylidene fluoride) composite membranes blended with nano-crystalline cellulose. Progress in
[89] Battaglia S. Quantitative X-ray diffraction analysis on thin samples deposited on cellulose membrane filters. X-Ray Spectrometry. 1985;**14**:16-19. DOI: 10.1002/xrs.1300140105
[90] Mohammad AW, Hilal N, Pei LY. Atomic force microscopy as a tool for asymmetric poly-
[91] Kim KJ, Dickson MR, Chen V, Fane AG. Applications of field emission scanning electron microscopy to polymer membrane research. Micron and Microscopica Acta.
[92] Bowen WR, Hilal N, Lovitt RW, Wright CJ. Atomic force microscopy as a tool for the
[93] Manzak A, Yıldız Y, Tutkun O. Characterization of Polymer Inclusion Membrane Containing Aliquat 336 as a Carrier. Membrane Water Treatment. 2015;**6**:5-102. DOI:
meric membrane characterization. Sains Malaysiana. 2011;**40**:237-244
membrane technologist. Microscopy and Analysis. 1998;**68**:13-16.
http://dx.doi.org/10.12989/mwt.2015.6.2.095
entropic edge detection. Measurement Science & Technology. 2001;**12**:288-298
10.1002/(SICI)1097-4628(19990822)73:8<1363::AID-APP3>3.0.CO;2-P
branes. Journal of Membrane Science. 2005;**249**:245-249.
compounds. Progress in Polymer Science. 2007;**32**:698-725.
Natural Science: Materials International. 2012;**22**:250-257
9781118522318.emst061
222 Desalination
1992;**23**:259-271
Desalination. 1977;**21**:241-249.
[107] Youm KH, Kim WS. Prediction of intrinsic pore properties of ultrafiltration membrane by solute rejection curves: Effect of operating conditions on pore properties. Journal of Chemical Engineering of Japan. 1981;**24**:1-7.
**Chapter 11**
**NEREU Project: Construction of a Plasma Reactor for**
The predatory exploitation techniques used for the supply of protein resources have been systematically causing the decrease in seafood. The solution was sought in the form of marine farms, for the purpose of production of fish and seafood. Brazil created an incentive production of seafood in order to increase seafood production. In 1998, the state governments joined the project, encouraging the creation of marine farms without an assessment of the impact on the environment. In 2005, after several records of seafood production in Santa Catarina started an epidemic of white spot shrimp, decimating several fishing farms. In Bahia came the disease lethargic crab, which simply decimated almost 90% of crabs those states. In 2010, the State Government of Bahia invested in research to combat this degradation of mangroves; the project had to in essence deal with the brown mare and help save the fishing industry of the state. It was based on these principles that Nereu project emerged, which provides for the rational and intelligent use of contaminated and hypersaline water of the marine farms from the production area of fish and seafood, in order to neutralize all pathogens and produce pure water with
**Keywords:** colloidal carbon, desalination, hydrogen, methane, stirling engines
The constant negative action of man over nature, promoting deforestation, releasing waste without treatment and chemicals in the water, grounding of the mangroves that are nurseries for many species, and overfishing causing the decrease of fish stocks and leading the fisherman to stop fishing and look for other activities. The crisis in the fisheries sector caused by the
> © 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
**Reform of Greenhouse Gases for Treatment of**
**Wastewater of the Marine Farms**
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.69313
André Pedral S. de Sena
**Abstract**
energy cogeneration.
**1. Introduction**
## **NEREU Project: Construction of a Plasma Reactor for Reform of Greenhouse Gases for Treatment of Wastewater of the Marine Farms**
André Pedral S. de Sena
[107] Youm KH, Kim WS. Prediction of intrinsic pore properties of ultrafiltration membrane by solute rejection curves: Effect of operating conditions on pore properties. Journal of
[108] García NP, Rodriguez J, Vigo F, Chester S. UF Membranes Autopsies: An Approach to Hollow Fibre Membranes Surface The International Desalination Association World Congress on Desalination and Water Reuse 2015/San Diego, CA, USA, REF:
[109] Ebrahim S, Bou-Hamed S, Abdel-Jawad M, Burney N. Microfiltration system as a pretreatment for RO units: Technical and economic assessment. Desalination. 1997:**109**:
[110] Zeman LJ, Zydney AL. Microfiltration and ultrafiltration: Principles and applications.
[112] Lazarova V, Gallego S, García Molina V, Rougé P. Problems of operation and main reasons for failure of membranes in tertiary treatment systems. Water Science and
[113] Rautenbach, R, Linn T, A1-Gobaisi DMK. Present and future pretreatment concepts– Strategies for reliable and low-maintenance reverse osmosis seawater desalination.
[114] Strathmann H. Electrodialytic Membrane Processes and their Practical Application,
[115] BQ Magazine. Available from: http://www.bq-magazine.com/industries/2015/02/water-
[117] Porcelli N, Judd S. Chemical cleaning of potable water membranes: A review. Separation
[118] Straub AP, Deshmukh A, Elimelech M. Pressure-retarded osmosis for power generation from salinity gradients: Is it viable? Energy & Environmental Science. 2016;**9**:31-48
[111] Cheryan M. Ultrafiltration Handbook. Lancaster (PA): Technomic; 1986
Technology. 2008;**57**:1777-1784. DOI: 10.2166/wst.2008.241.
[116] Bcc Research, 2016. Available from: http://www.bccresearch.com
Studies in Environmental Science. 1994;**59**:495-533
and Purification Technology. 2010;**71**:137-143.
Chemical Engineering of Japan. 1981;**24**:1-7.
IDAWC15- PEÑA
New York: Marcel Dekker; 1996.
Desalination. 1997;**110**;97-106
scarcity-in-the-middle-east
165-175
224 Desalination
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.69313
#### **Abstract**
The predatory exploitation techniques used for the supply of protein resources have been systematically causing the decrease in seafood. The solution was sought in the form of marine farms, for the purpose of production of fish and seafood. Brazil created an incentive production of seafood in order to increase seafood production. In 1998, the state governments joined the project, encouraging the creation of marine farms without an assessment of the impact on the environment. In 2005, after several records of seafood production in Santa Catarina started an epidemic of white spot shrimp, decimating several fishing farms. In Bahia came the disease lethargic crab, which simply decimated almost 90% of crabs those states. In 2010, the State Government of Bahia invested in research to combat this degradation of mangroves; the project had to in essence deal with the brown mare and help save the fishing industry of the state. It was based on these principles that Nereu project emerged, which provides for the rational and intelligent use of contaminated and hypersaline water of the marine farms from the production area of fish and seafood, in order to neutralize all pathogens and produce pure water with energy cogeneration.
**Keywords:** colloidal carbon, desalination, hydrogen, methane, stirling engines
## **1. Introduction**
The constant negative action of man over nature, promoting deforestation, releasing waste without treatment and chemicals in the water, grounding of the mangroves that are nurseries for many species, and overfishing causing the decrease of fish stocks and leading the fisherman to stop fishing and look for other activities. The crisis in the fisheries sector caused by the
© 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
uncontrolled exploitation of the coastal zone with the consequent reduction of marine stocks makes aquaculture an important food production alternative, significantly contributing to food security and poverty alleviation on the planet [1, 2].
the marine farms established in these states generate waste, which cause the problem, generating conditions for the proliferation of pathogens in the environment, often resistant to the immune system of animals due to the indiscriminate use of antibiotics in effluent ponds and crops laboratories. The DNAs of the *Exophiala cf psycrofila* and *Nimaviridae* were found in the existing mangrove in the region of *Jaguaribe*, fishing town in the state of *Bahia*. This organic material was present in a phenomenon known as brown tide and was caused by the discharge
NEREU Project: Construction of a Plasma Reactor for Reform of Greenhouse Gases...
http://dx.doi.org/10.5772/intechopen.69313
227
It is problematic and with the need to clean the environment of this waste, the principles of Nereu project were built, which provides for the rational and intelligent use of contaminated water and hypersaline of the marine farms, in order to neutralize all pathogens and produce
Despite the changes in development models that have taken place throughout history, natural resources have always been exploited without any concern for the sustainability of the system, often exhausting these resources to the extreme. The current model of development has generated global environmental impacts, putting the biosphere at risk. Development models that revert to the situation or at least minimize impacts on the biosphere are urgently
According to some data, world mariculture expanded rapidly from 16.7 million tons in 2004 to 22.8 million tons in 2014, an increase of 35.8%, whereas the fishing sector declined in this period from 83.8 to 69.9 million tons. In 2004, mariculture accounted for 16.6% of the world's fish production, which increased to 30.1% in 2014. In contrast to the increasing decline in fisheries, seafood production is seen as an alternative to maintain the world demand for aquatic products. If current development is maintained over the next 15 years, 32% of the world's
In Brazil, mariculture has been progressively expanding, mainly since the 1990s due to government programs, which is represented by bivalve mollusks in the South and Southeast
Brazil produced 91,405 t of mariculture products in 2014, of which 71,000 t are shrimp and 20,405 t represented by mollusks. The total value of mariculture production is US\$876.8 million,
Thus, mariculture is a deliberate attempt by humans to modify and manipulate naturecontrolled food relationships. It can be observed that all productive activities are impacting the environment and that in aquaculture, these impacts can be classified into three sets: those from the environment (exogenous); those resulting from aquaculture (endogenous),
In addition, it is emphasized that the environmental impacts caused by aquaculture will
of waste from marine farms who had settled in this region [4, 6–9].
total production of marine fish will come from mariculture [1, 4–7, 9].
regions and by shrimps in the North and Northeast regions [4].
of which US\$256.8 million is represented by mollusks [3, 4, 8–10].
and those caused by aquaculture on the environment [8–10].
almost always be more negative than positive.
pure water with energy cogeneration.
**2. Farm of shrimp: the problem**
needed.
The aquaculture is a deliberate attempt by humans to modify and manipulate the trophic relationships controlled by nature. Thus, it can be seen that all production activities are influencing the environment and in aquaculture, such impacts can be classified into three groups: those from the environment (exogenous), those resulting from aquaculture owned (endogenous), and those caused by the aquaculture on the environment. Moreover, it is noteworthy that the environmental impacts caused by aquaculture will always be more negative than positive [2, 3].
Thus, fish farming and shellfish generate waste in the form of particles and inorganic and organic soluble material. The nature and quantity of effluent vary with the type of cultivation, crop species, and intensification level. In general, feces, excreta, and pseudo-feces are the main sources of inorganic and organic nutrients [4].
The bacteria, fungi, and protozoa participate in the degradation process of organic matter, and release of toxic gases such as sulphur become unavoidable. The ideal of sustainable aquaculture, which is consistent with the economic, social, ecological, spatial, and cultural contexts [4, 5], is at an early stage in Brazil, where mistakes are being evaluated and studied so that new management techniques may be suitable for targeting a smaller effect on the ecosystem balance.
This concern about the environmental impact was not taken into account by the Federal Government of Brazil, which despite the warnings given by CONAMA [6] decided, through the Ministries of Agriculture, Environment and Development of Industry and Trade in 1997, to create an incentive to the fishing and production of seafood through the creation of marine farms. In 1998, the state governments began to join the project, and the states of *Bahia, Santa Catarina, Ceara*, and *Espirito Santo* invested heavily in the project for installation of farms mainly shrimp, oysters, and scallops. Through tax incentives, public funding, and federal funding, there was an incentive to create these marine farms, but there was no evaluation on the impact of aquaculture in the environment [4–9].
Between 2000 and 2005, the record increase in the seafood production was achieved. A total of 75% of this production aimed at export [4, 6–9]. However, in 2005, it began a productive decline, beginning with Santa Catarina, southern Brazil, which started an epidemic of white spot syndrome in the shrimp. It is a disease that affects shrimps and is one of the four identified in Brazil contained in the OIE list of aquatic animals. The WSSV—"White Spot Syndrome Virus"—is a persistent infection, with high mortality in shrimp cultivation. This disease began to spread along the Brazilian coast, currently reaching the state of *Ceara*, with production loss of nearly 80% of shrimps. Almost concurrent appearance of this disease arose in the mangroves of *Espirito Santo* and *Bahia*, the disease lethargic crab (DCL). It was found that the causative agent of DCL is the pathogenic fungus *Exophiala cf psycrofila* that simply decimated almost 90% of crabs in these states [3, 4, 9].
In 2015, the disease proliferates reaching other states, decimating the mangrove areas and ending with the crab. After several analyses of the environmental agencies, it was found that the marine farms established in these states generate waste, which cause the problem, generating conditions for the proliferation of pathogens in the environment, often resistant to the immune system of animals due to the indiscriminate use of antibiotics in effluent ponds and crops laboratories. The DNAs of the *Exophiala cf psycrofila* and *Nimaviridae* were found in the existing mangrove in the region of *Jaguaribe*, fishing town in the state of *Bahia*. This organic material was present in a phenomenon known as brown tide and was caused by the discharge of waste from marine farms who had settled in this region [4, 6–9].
It is problematic and with the need to clean the environment of this waste, the principles of Nereu project were built, which provides for the rational and intelligent use of contaminated water and hypersaline of the marine farms, in order to neutralize all pathogens and produce pure water with energy cogeneration.
## **2. Farm of shrimp: the problem**
uncontrolled exploitation of the coastal zone with the consequent reduction of marine stocks makes aquaculture an important food production alternative, significantly contributing to
The aquaculture is a deliberate attempt by humans to modify and manipulate the trophic relationships controlled by nature. Thus, it can be seen that all production activities are influencing the environment and in aquaculture, such impacts can be classified into three groups: those from the environment (exogenous), those resulting from aquaculture owned (endogenous), and those caused by the aquaculture on the environment. Moreover, it is noteworthy that the environmental impacts caused by aquaculture will always be more negative than positive [2, 3]. Thus, fish farming and shellfish generate waste in the form of particles and inorganic and organic soluble material. The nature and quantity of effluent vary with the type of cultivation, crop species, and intensification level. In general, feces, excreta, and pseudo-feces are the
The bacteria, fungi, and protozoa participate in the degradation process of organic matter, and release of toxic gases such as sulphur become unavoidable. The ideal of sustainable aquaculture, which is consistent with the economic, social, ecological, spatial, and cultural contexts [4, 5], is at an early stage in Brazil, where mistakes are being evaluated and studied so that new management techniques may be suitable for targeting a smaller effect on the ecosystem
This concern about the environmental impact was not taken into account by the Federal Government of Brazil, which despite the warnings given by CONAMA [6] decided, through the Ministries of Agriculture, Environment and Development of Industry and Trade in 1997, to create an incentive to the fishing and production of seafood through the creation of marine farms. In 1998, the state governments began to join the project, and the states of *Bahia, Santa Catarina, Ceara*, and *Espirito Santo* invested heavily in the project for installation of farms mainly shrimp, oysters, and scallops. Through tax incentives, public funding, and federal funding, there was an incentive to create these marine farms, but there was no evaluation on
Between 2000 and 2005, the record increase in the seafood production was achieved. A total of 75% of this production aimed at export [4, 6–9]. However, in 2005, it began a productive decline, beginning with Santa Catarina, southern Brazil, which started an epidemic of white spot syndrome in the shrimp. It is a disease that affects shrimps and is one of the four identified in Brazil contained in the OIE list of aquatic animals. The WSSV—"White Spot Syndrome Virus"—is a persistent infection, with high mortality in shrimp cultivation. This disease began to spread along the Brazilian coast, currently reaching the state of *Ceara*, with production loss of nearly 80% of shrimps. Almost concurrent appearance of this disease arose in the mangroves of *Espirito Santo* and *Bahia*, the disease lethargic crab (DCL). It was found that the causative agent of DCL is the pathogenic fungus *Exophiala cf psycrofila* that simply
In 2015, the disease proliferates reaching other states, decimating the mangrove areas and ending with the crab. After several analyses of the environmental agencies, it was found that
food security and poverty alleviation on the planet [1, 2].
main sources of inorganic and organic nutrients [4].
the impact of aquaculture in the environment [4–9].
decimated almost 90% of crabs in these states [3, 4, 9].
balance.
226 Desalination
Despite the changes in development models that have taken place throughout history, natural resources have always been exploited without any concern for the sustainability of the system, often exhausting these resources to the extreme. The current model of development has generated global environmental impacts, putting the biosphere at risk. Development models that revert to the situation or at least minimize impacts on the biosphere are urgently needed.
According to some data, world mariculture expanded rapidly from 16.7 million tons in 2004 to 22.8 million tons in 2014, an increase of 35.8%, whereas the fishing sector declined in this period from 83.8 to 69.9 million tons. In 2004, mariculture accounted for 16.6% of the world's fish production, which increased to 30.1% in 2014. In contrast to the increasing decline in fisheries, seafood production is seen as an alternative to maintain the world demand for aquatic products. If current development is maintained over the next 15 years, 32% of the world's total production of marine fish will come from mariculture [1, 4–7, 9].
In Brazil, mariculture has been progressively expanding, mainly since the 1990s due to government programs, which is represented by bivalve mollusks in the South and Southeast regions and by shrimps in the North and Northeast regions [4].
Brazil produced 91,405 t of mariculture products in 2014, of which 71,000 t are shrimp and 20,405 t represented by mollusks. The total value of mariculture production is US\$876.8 million, of which US\$256.8 million is represented by mollusks [3, 4, 8–10].
Thus, mariculture is a deliberate attempt by humans to modify and manipulate naturecontrolled food relationships. It can be observed that all productive activities are impacting the environment and that in aquaculture, these impacts can be classified into three sets: those from the environment (exogenous); those resulting from aquaculture (endogenous), and those caused by aquaculture on the environment [8–10].
In addition, it is emphasized that the environmental impacts caused by aquaculture will almost always be more negative than positive.
These impacts arise mainly from the use of resources, such as space, water, raw materials, and food, reduction of biodiversity, as well as from the production of organic and inorganic waste (e.g., excreta production, introduction of microorganisms, pathogens and parasites in the environment, accumulation of residues of cultivated organisms, and release of antibiotics in the effluent of nurseries and crop laboratories) [3–5, 10].
Therefore, it can be observed that the negative impacts can be of two types: direct, for example, by the introduction of exotic genetic material into the environment and indirect, due to loss of habitat and ecological niche.
Any alteration of the physical, chemical, and biological properties of the environment caused by any form of matter or energy resulting from human activities that directly or indirectly affect the health, safety, and well-being of the population shall be considered an environmental impact, which include social and economic activities; biota, aesthetic, and sanitary conditions of the environment and the quality of environmental resources. These changes need to be quantified, since they present relative variations, which can be positive or negative, big, or small [8, 10]. The different aquaculture modalities can generate diverse environmental impacts, depending mainly on the cultivation system (closed, semi-open, and open systems), aquaculture (freshwater or marine), of the species used and especially the density and quantity of production. Due to the many variables that may influence the generation or identification of such impacts, and because it is a relatively recent activity in Brazil, few conclusive studies have been published on the possible environmental impacts caused by aquaculture, especially by mariculture. Nevertheless, in any form of production, the impact on the environment occurs through three processes: the consumption of natural resources, the process of transformation, and the generation of final products (waste) [3].
of the marine community caused by the accumulation of organic matter in the sediment has already been studied by several researchers. The ideal of sustainable aquaculture, which is consistent with the economic, social, ecological, spatial, and cultural contexts, is at a final stage in Brazil, where errors are being evaluated and studied so that new management tech-
NEREU Project: Construction of a Plasma Reactor for Reform of Greenhouse Gases...
http://dx.doi.org/10.5772/intechopen.69313
229
This concern about the environmental impact was not taken into account by the Federal Government of Brazil, which, despite the warnings given by CONAMA, decided, through the Ministries of Agriculture, Environment and Development of Industry and Commerce, to create in 1997 a project to encourage fishing and seafood production, through the creation of marine farms, in order to increase the production of seafood, mainly crustaceans and mollusks. In 1998, the state governments began to join the project, and the states of *Bahia, Santa Catarina, Ceara*, and *Espirito Santo* invested heavily in the project for the installation of farms mainly shrimp, oysters, and scallops. Through fiscal incentives, public funding, and federal funds, there was an incentive to create marine farms in these states, but there was no prior and direct evaluation of the impact of the implementation of aquaculture in these states and
Between 2000 and 2005, records of seafood production were achieved. A total of 75% of this production was for export [4, 10]. But in 2005, a productive decline began, beginning in Santa Catarina, southern Brazil, where an epidemic of shrimp of the white spot started. It is a disease that affects shrimps and is one of the four identified in Brazil that are listed in the OIE in aquatic animals. It is of viral ethology. WSSV—"White Spot Syndrome Virus" is a persistent infection throughout the animal's life and has a high mortality in shrimp cultures.
niques can have a lower effect on the ecosystem balance [3, 8, 10].
**Figure 1.** Sample containers of contaminated water discharged to ocean [8].
in the environment [4, 6–10].
The main environmental impacts caused by aquaculture (including fish farming and shrimp farming) are conflicts with the use of water bodies, sedimentation and obstruction of water flows, over nutrition and eutrophication, discharge of pond effluents, and pollution by chemical residues used in the different stages of cultivation [2, 5].
Thus, fish, crustacean and mollusk crops generate waste in the form of particles and organic and inorganic soluble material. The nature and quantity of the effluent vary with the type of farm, cultivated species, and level of intensification. In general, feces, excreta, and pseudofeces are the major sources of inorganic and organic nutrients [5].
The water column is the container for the dissolved material, while the greater proportion of solid waste precipitates into the sediment below the growing area or in the immediate vicinity of the farm. This almost immediate deposition of daily organic matter in the sediment leads to an onset of eutrophication with the decrease of the oxygen rate dissolved in the medium. The level of water contamination of the shrimp farms in the Jaguaripe/Bahia—Brazil can be demonstrated by **Figure 1**, which contains the photo of containers of water samples, which were collected by federal agents of the Ministry of the Environment.
Bacteria, fungi, and protozoa participate in the process of degradation of organic matter, and the release of toxic gases such as sulphur become unavoidable. In this sense, the alteration
**Figure 1.** Sample containers of contaminated water discharged to ocean [8].
These impacts arise mainly from the use of resources, such as space, water, raw materials, and food, reduction of biodiversity, as well as from the production of organic and inorganic waste (e.g., excreta production, introduction of microorganisms, pathogens and parasites in the environment, accumulation of residues of cultivated organisms, and release of antibiotics
Therefore, it can be observed that the negative impacts can be of two types: direct, for example, by the introduction of exotic genetic material into the environment and indirect, due to
Any alteration of the physical, chemical, and biological properties of the environment caused by any form of matter or energy resulting from human activities that directly or indirectly affect the health, safety, and well-being of the population shall be considered an environmental impact, which include social and economic activities; biota, aesthetic, and sanitary conditions of the environment and the quality of environmental resources. These changes need to be quantified, since they present relative variations, which can be positive or negative, big, or small [8, 10]. The different aquaculture modalities can generate diverse environmental impacts, depending mainly on the cultivation system (closed, semi-open, and open systems), aquaculture (freshwater or marine), of the species used and especially the density and quantity of production. Due to the many variables that may influence the generation or identification of such impacts, and because it is a relatively recent activity in Brazil, few conclusive studies have been published on the possible environmental impacts caused by aquaculture, especially by mariculture. Nevertheless, in any form of production, the impact on the environment occurs through three processes: the consumption of natural resources, the process of transformation, and the generation of final products (waste) [3]. The main environmental impacts caused by aquaculture (including fish farming and shrimp farming) are conflicts with the use of water bodies, sedimentation and obstruction of water flows, over nutrition and eutrophication, discharge of pond effluents, and pollution by chemi-
Thus, fish, crustacean and mollusk crops generate waste in the form of particles and organic and inorganic soluble material. The nature and quantity of the effluent vary with the type of farm, cultivated species, and level of intensification. In general, feces, excreta, and pseudo-
The water column is the container for the dissolved material, while the greater proportion of solid waste precipitates into the sediment below the growing area or in the immediate vicinity of the farm. This almost immediate deposition of daily organic matter in the sediment leads to an onset of eutrophication with the decrease of the oxygen rate dissolved in the medium. The level of water contamination of the shrimp farms in the Jaguaripe/Bahia—Brazil can be demonstrated by **Figure 1**, which contains the photo of containers of water samples, which
Bacteria, fungi, and protozoa participate in the process of degradation of organic matter, and the release of toxic gases such as sulphur become unavoidable. In this sense, the alteration
in the effluent of nurseries and crop laboratories) [3–5, 10].
cal residues used in the different stages of cultivation [2, 5].
feces are the major sources of inorganic and organic nutrients [5].
were collected by federal agents of the Ministry of the Environment.
loss of habitat and ecological niche.
228 Desalination
of the marine community caused by the accumulation of organic matter in the sediment has already been studied by several researchers. The ideal of sustainable aquaculture, which is consistent with the economic, social, ecological, spatial, and cultural contexts, is at a final stage in Brazil, where errors are being evaluated and studied so that new management techniques can have a lower effect on the ecosystem balance [3, 8, 10].
This concern about the environmental impact was not taken into account by the Federal Government of Brazil, which, despite the warnings given by CONAMA, decided, through the Ministries of Agriculture, Environment and Development of Industry and Commerce, to create in 1997 a project to encourage fishing and seafood production, through the creation of marine farms, in order to increase the production of seafood, mainly crustaceans and mollusks. In 1998, the state governments began to join the project, and the states of *Bahia, Santa Catarina, Ceara*, and *Espirito Santo* invested heavily in the project for the installation of farms mainly shrimp, oysters, and scallops. Through fiscal incentives, public funding, and federal funds, there was an incentive to create marine farms in these states, but there was no prior and direct evaluation of the impact of the implementation of aquaculture in these states and in the environment [4, 6–10].
Between 2000 and 2005, records of seafood production were achieved. A total of 75% of this production was for export [4, 10]. But in 2005, a productive decline began, beginning in Santa Catarina, southern Brazil, where an epidemic of shrimp of the white spot started. It is a disease that affects shrimps and is one of the four identified in Brazil that are listed in the OIE in aquatic animals. It is of viral ethology. WSSV—"White Spot Syndrome Virus" is a persistent infection throughout the animal's life and has a high mortality in shrimp cultures. Animals that recover from the infection are persistent carriers of the virus. There are no known biological vectors. This virus has a type of flagellum and its genetic material is DNA (being genetically more stable) and is part of the *Nimaviridae* family, decimating several fishing grounds. This disease began to extend along the Brazilian coast, currently reaching the state of *Ceara*, with productive loss of almost 80% of the prawns [4, 5, 9].
**3. Nereu project**
the new demands.
for consumption centres' [11–13].
The oceans and seas are vital to sustain life and to provide for the various human activities
NEREU Project: Construction of a Plasma Reactor for Reform of Greenhouse Gases...
http://dx.doi.org/10.5772/intechopen.69313
231
In this context, it is verified that the activity of fishing and gathering of seafood stands out due to its importance in the production of food and world economy. But in order to reach the levels of production and productivity demanded today, fishing has destroyed most of the
The oceans have strategic relevance for maintaining the dignity of human life and for the economy, which converges to the challenge of integrated management of seafood production and management. Environmental resources, especially ocean resources, are currently in a vulnerable position of degradation and scarcity. Several authors argue that the main factors of the maritime crisis are the environmental pollution and the unbridled increase of the world population, without the policies on territorial planning and environment to adequately meet
The degradation of ocean water quality is a reflection of the current pattern of consumption and lack of care with the oceans. In addition, the Industrial Revolution, in which the concern to increase fish production, without considering the externalities, resulted in the degradation and waste of marine resources. In this crisis scenario, countless losses can be raised, such as loss of ecological balance, loss of biological diversity, climatic imbalance, and profound changes in the water cycle, which affect society and the environment leading to reduced or depleted resources. Thus, today, the world population has established a situation of imbalance between the spatial pattern of availability in the oceans and the spatial pattern of demand
With the emergence of aquaculture, the first attempt was made to reduce the pressure on fishing banks, but the lack of an environmental policy to treat wastewater has increased the degradation of the oceans. To address this degradation caused by the seafood farms that affected several Brazilian states, the Ministry of the Environment analyzed the environmental issue in order to build an environmental awareness. In this sense, with the introduction of new sustainable development concepts, a new cycle began, based on the elaboration and implementation of environmental policies, in the search for negotiation and understanding
Advances can be verified with the Green Protocol (Federal Law: institutional device of introduction of the environmental variable as a relevant criterion in economic policy decisions and project financing) and other devices of environmental dimension inserted in the decisions of public policies. Specific measures for the rational management of irrigated areas and collec-
From these premises, the Nereu project was conceived by associating two Chiron and Prometheus projects, which combine three basic conditions for the production of water and energy, with food production based on a clean production system. A conception was used
between preservation of the environment and production processes.
tion for the use of water are already established by law.
and are the primordial elements for the entire global economy.
oceans' fishing grounds due to the phenomenon of overfishing.
Almost concomitant with the appearance of this disease, in the mangroves of *Espirito Santo* and *Bahia*, appear the disease of the lethargic crab (DCL), it was found that the agent that causes DCL is the pathogenic fungus *Exophiala cf psycrofila*, which simply decimated almost 90% of the crabs in these states [3].
By 2015, this disease spread to other states, decimating the mangrove areas and drastically reducing the population of crabs. After several analyses of environmental agencies, it was verified that marine farms, implanted in these states, generate residues that provoke the problem, generating conditions for the proliferation of pathogens in the environment, often resistant to the immune system of the animals, due to the indiscriminate use of antibiotics from nurseries and crop laboratories.
The DNAs of *Exophiala cf psycrofila* and *Nimaviridae* were found in the mangrove in the region of *Jaguaripe*, a fishing town in the State of *Bahia*—Brazil. This organic material was present in a phenomenon known as brown tide and was caused by the discharge of tailings from the fishing farms that had settled in this region (see **Figure 2**).
Thus, from the studies of the discharge effluents from tanks of shrimp farms in the *Jaguaripe*— *Bahia/*Brazil region, promoted by the Ministries of Fisheries and Environment, a map of the destruction and contamination of the region known as the coast of the *Dende* was created that raised the points of impact and associated the degradation of the aquatic environment to the effluents discarded by these farms, which influenced native marine life, causing the degradation and destruction of native species [8].
**Figure 2.** Shrimp farm: region of *Jaguaripe*—*Bahia*, Brazil [8].
## **3. Nereu project**
Animals that recover from the infection are persistent carriers of the virus. There are no known biological vectors. This virus has a type of flagellum and its genetic material is DNA (being genetically more stable) and is part of the *Nimaviridae* family, decimating several fishing grounds. This disease began to extend along the Brazilian coast, currently reaching the
Almost concomitant with the appearance of this disease, in the mangroves of *Espirito Santo* and *Bahia*, appear the disease of the lethargic crab (DCL), it was found that the agent that causes DCL is the pathogenic fungus *Exophiala cf psycrofila*, which simply decimated almost
By 2015, this disease spread to other states, decimating the mangrove areas and drastically reducing the population of crabs. After several analyses of environmental agencies, it was verified that marine farms, implanted in these states, generate residues that provoke the problem, generating conditions for the proliferation of pathogens in the environment, often resistant to the immune system of the animals, due to the indiscriminate use of antibiotics from
The DNAs of *Exophiala cf psycrofila* and *Nimaviridae* were found in the mangrove in the region of *Jaguaripe*, a fishing town in the State of *Bahia*—Brazil. This organic material was present in a phenomenon known as brown tide and was caused by the discharge of tailings from the
Thus, from the studies of the discharge effluents from tanks of shrimp farms in the *Jaguaripe*— *Bahia/*Brazil region, promoted by the Ministries of Fisheries and Environment, a map of the destruction and contamination of the region known as the coast of the *Dende* was created that raised the points of impact and associated the degradation of the aquatic environment to the effluents discarded by these farms, which influenced native marine life, causing the degrada-
state of *Ceara*, with productive loss of almost 80% of the prawns [4, 5, 9].
90% of the crabs in these states [3].
230 Desalination
nurseries and crop laboratories.
tion and destruction of native species [8].
**Figure 2.** Shrimp farm: region of *Jaguaripe*—*Bahia*, Brazil [8].
fishing farms that had settled in this region (see **Figure 2**).
The oceans and seas are vital to sustain life and to provide for the various human activities and are the primordial elements for the entire global economy.
In this context, it is verified that the activity of fishing and gathering of seafood stands out due to its importance in the production of food and world economy. But in order to reach the levels of production and productivity demanded today, fishing has destroyed most of the oceans' fishing grounds due to the phenomenon of overfishing.
The oceans have strategic relevance for maintaining the dignity of human life and for the economy, which converges to the challenge of integrated management of seafood production and management. Environmental resources, especially ocean resources, are currently in a vulnerable position of degradation and scarcity. Several authors argue that the main factors of the maritime crisis are the environmental pollution and the unbridled increase of the world population, without the policies on territorial planning and environment to adequately meet the new demands.
The degradation of ocean water quality is a reflection of the current pattern of consumption and lack of care with the oceans. In addition, the Industrial Revolution, in which the concern to increase fish production, without considering the externalities, resulted in the degradation and waste of marine resources. In this crisis scenario, countless losses can be raised, such as loss of ecological balance, loss of biological diversity, climatic imbalance, and profound changes in the water cycle, which affect society and the environment leading to reduced or depleted resources. Thus, today, the world population has established a situation of imbalance between the spatial pattern of availability in the oceans and the spatial pattern of demand for consumption centres' [11–13].
With the emergence of aquaculture, the first attempt was made to reduce the pressure on fishing banks, but the lack of an environmental policy to treat wastewater has increased the degradation of the oceans. To address this degradation caused by the seafood farms that affected several Brazilian states, the Ministry of the Environment analyzed the environmental issue in order to build an environmental awareness. In this sense, with the introduction of new sustainable development concepts, a new cycle began, based on the elaboration and implementation of environmental policies, in the search for negotiation and understanding between preservation of the environment and production processes.
Advances can be verified with the Green Protocol (Federal Law: institutional device of introduction of the environmental variable as a relevant criterion in economic policy decisions and project financing) and other devices of environmental dimension inserted in the decisions of public policies. Specific measures for the rational management of irrigated areas and collection for the use of water are already established by law.
From these premises, the Nereu project was conceived by associating two Chiron and Prometheus projects, which combine three basic conditions for the production of water and energy, with food production based on a clean production system. A conception was used that interrelated CH4 reform through a systematic pyrolysis in a cold plasma reactor to obtain H2 whose thermal cycle was used for water desalination and energy generation combining with the production of totally sterilized rock salt (see **Figure 3**).
With this conception, the Nereu<sup>1</sup> project has the following structure:
In view of these premises, the Nereu project seeks to meet the main characteristics of the Apollo Research Group projects, that is, to have a simple automation in their constructive characteristics, guaranteeing their interchangeability and maintenance, as well as their low automation costs. In addition, the Nereu Project units were conceived and developed in modular systems, proto-units, to be placed in reduced and limited physical spaces; from its constructive concept, the Nereu project can expand its productive capacity according to the needs.
**3.1. Chiron project**
**Figure 4.** Process flowchart [12, 13].
rubber industry.
between molecules [12, 14, 15].
*3.1.1. Plasma generation reactor for H2*
nied by the emission of CO, CO<sup>2</sup>
In the specific case of pyrolysis of CH<sup>4</sup>
The desalination process is a complex process because the salts are strongly bound to the water molecules, which makes conventional water treatment processes inefficient. Therefore, it is necessary to make use of other processes capable of breaking the forces of attraction
hydrogen, which can be used as fuel, and colloidal carbon, with many uses in the cement and
Conventional hydrogen and black carbon production processes are based on the incomplete combustion of hydrocarbons, subject to low yields and emission of high levels of pollutants. Due to incomplete combustion in the presence of oxygen, the process is inevitably accompa-
The reactor of the Chiron project was designed to produce colloidal carbon and hydrogen, through thermal plasma, although using electric energy and an inert gas in the process have
by thermal plasma, the products formed include
NEREU Project: Construction of a Plasma Reactor for Reform of Greenhouse Gases...
http://dx.doi.org/10.5772/intechopen.69313
233
, and volatile hydrocarbons. In addition, the maxi-
*—Unit 01*
, SO<sup>2</sup>
mum reaction temperature achievable is from 1500 to 2000 K [11–15].
, NO<sup>x</sup>
**Figure 3.** Diagram of the Nereu project.
<sup>1</sup> The Nereus project is a tribute to Nereus (Greek: swim), in Greek mythology, is a primitive sea god, depicted as an elderly character - the old man of the sea. He is known for his virtues and his wisdom. Pindar celebrates his fairness, hence his epithets "true," "benevolent," "without lying or oblivion".
<sup>2</sup> The Chiron project is a tribute to Chiron (Greek: hand), in Greek mythology, was a centaur, considered superior by his own peers. Chiron was intelligent, civilized and kind, and celebrated for his knowledge and skill in medicine; he was professor of Heracles and other demigods.
<sup>3</sup> The Prometheus project is a tribute to Prometheus (in Greek: foresight), in Greek mythology is a titan; he was a defender of humanity, known for his astute intelligence, responsible for stealing the fire of Hestia and giving the Mortals.
NEREU Project: Construction of a Plasma Reactor for Reform of Greenhouse Gases... http://dx.doi.org/10.5772/intechopen.69313 233
**Figure 4.** Process flowchart [12, 13].
#### **3.1. Chiron project**
that interrelated CH4
With this conception, the Nereu<sup>1</sup>
metheus reactors and produce O<sup>2</sup>
H2
232 Desalination
1
2
3
• Chiron2
• Prometheus<sup>3</sup>
reform through a systematic pyrolysis in a cold plasma reactor to obtain
using a modular and interchangeable structure [12, 13].
whose thermal cycle was used for water desalination and energy generation combining
plasma reactor for the separation of colloidal carbon and hydrogen, which will feed the Pro-
aquaculture systems, whether having a high content of dissolved solutes, with the purpose of sterilizing, purifying, and neutralizing pathogens present in process waste water, through a reactor of hydrogen. This process desalinates water by generating the superheated steam which is sent to a set of the stirling system, without affecting the environment and having by-products of the process the generation of clean energy, gem salt, and demineralized and
In view of these premises, the Nereu project seeks to meet the main characteristics of the Apollo Research Group projects, that is, to have a simple automation in their constructive characteristics, guaranteeing their interchangeability and maintenance, as well as their low automation costs. In addition, the Nereu Project units were conceived and developed in modular systems, proto-units, to be placed in reduced and limited physical spaces; from its constructive concept,
The Nereus project is a tribute to Nereus (Greek: swim), in Greek mythology, is a primitive sea god, depicted as an elderly character - the old man of the sea. He is known for his virtues and his wisdom. Pindar celebrates his fairness,
The Chiron project is a tribute to Chiron (Greek: hand), in Greek mythology, was a centaur, considered superior by his own peers. Chiron was intelligent, civilized and kind, and celebrated for his knowledge and skill in medicine; he was
The Prometheus project is a tribute to Prometheus (in Greek: foresight), in Greek mythology is a titan; he was a defender of humanity, known for his astute intelligence, responsible for stealing the fire of Hestia and giving the Mortals.
the Nereu project can expand its productive capacity according to the needs.
project has the following structure:
project: the purpose of this project is to carry out pyrolysis of CH4 through a cold
project: it provides for the rational and intelligent use of reuse water from
with the production of totally sterilized rock salt (see **Figure 3**).
deionized water (water pure; see **Figure 4**) [12, 13].
hence his epithets "true," "benevolent," "without lying or oblivion".
professor of Heracles and other demigods.
**Figure 3.** Diagram of the Nereu project.
The desalination process is a complex process because the salts are strongly bound to the water molecules, which makes conventional water treatment processes inefficient. Therefore, it is necessary to make use of other processes capable of breaking the forces of attraction between molecules [12, 14, 15].
#### *3.1.1. Plasma generation reactor for H2 —Unit 01*
In the specific case of pyrolysis of CH<sup>4</sup> by thermal plasma, the products formed include hydrogen, which can be used as fuel, and colloidal carbon, with many uses in the cement and rubber industry.
Conventional hydrogen and black carbon production processes are based on the incomplete combustion of hydrocarbons, subject to low yields and emission of high levels of pollutants. Due to incomplete combustion in the presence of oxygen, the process is inevitably accompanied by the emission of CO, CO<sup>2</sup> , SO<sup>2</sup> , NO<sup>x</sup> , and volatile hydrocarbons. In addition, the maximum reaction temperature achievable is from 1500 to 2000 K [11–15].
The reactor of the Chiron project was designed to produce colloidal carbon and hydrogen, through thermal plasma, although using electric energy and an inert gas in the process have several advantages in its application that are compensatory, as for example: increased efficiency, reduction of the reaction time, low cost of the equipment, favoring variety of productive size, and making possible its industrial application. The use of a noble gas as a plasma gas avoids the formation of CO<sup>2</sup> in the process and is only necessary to initiate plasma formation and was initially replaced by H2 , and then by CH4 , another important factor is that the plasma process enabled the use of different sources of raw material, besides CH4 gas, butane and propane, but for this project, we chose to use the gases from landfill gas. Thus, the production of hydrogen as an energy source and that of colloidal carbon as an industrial input becomes attractive and compensates the energy and operational expenditure of the system [11–13, 15, 16].
#### *3.1.1.1. The plasma system*
The Chiron project has a thermal plasma system, which is composed of a plasma arc torch, and a reactor or pyrolysis chamber designed to retain most of the heat flux from the plasma jet. The inert gas argon (Ar) was used as a plasma gas and was introduced in the torch by a controlled flow [12, 13, 15, 16].
The electric arc between the cathode and anode was initiated by a high-voltage discharge and a variable current. The formation of the plasma arc occurs between the cathode and the anode (cooled by water). The gas to be degraded, CH4 , was introduced into the pyrolysis chamber with controlled flow [12–14, 16].
The torch consisted of a cone-shaped cathode and an anode rivet, channel-shaped. The electric arc established between the electrodes (cathode and anode) was initiated by a high-voltage discharge. The gas is heated, ionized, and emerged from the torch as a plasma jet. CH<sup>4</sup> injected into the reactor was heated and passed then into the liner, by contacting the plasma and suffering degradation [11, 12, 16].
and high degree of ionization, chemical reactions are accelerated by the power supply, and
(■) and efficiency of H<sup>2</sup>
plasma. However, the existence of a collision process between the existing elements in the plasma
occur as a result of impacts between the molecules, atoms, ions, electrons, and protons [11–15].
The usage of Unit 01—reactor pyrolysis—has several advantages such as high energy density existing in the plasma which allows the construction of a compact reactor. Due to the plasma characteristics, a continuous flow is possible, enabling the formation of carbonaceous materials of different qualities and providing no pollution to obtain two commercial products (H<sup>2</sup>
A revolutionary constructive architecture for the treatment of methane, the reactor provides
into H2
residence time in the reaction zone and also a longer residence time inside the
and hydrogen flow, important for the heat exchange between the
molecule can be dissociated into one or more parts to collide with another particle. Various collision processes may occur inside the plasma reactor, including the recombination
emissions or other contaminant gases, once the system is deoxygen-
NEREU Project: Construction of a Plasma Reactor for Reform of Greenhouse Gases...
http://dx.doi.org/10.5772/intechopen.69313
235
(▲).
and other molecules (CH3
is broken down mainly by the heat from the arc
. Different types of crashes happen in a
, and CH) may also
, CH2
molecules. This feature is important for a
and C. In addition to presenting a favorable
moreover, there is no CO<sup>2</sup>
**Figure 5.** Percentage of degradation of CH<sup>4</sup>
In the Unit 01 plasma pyrolysis reactor, CH<sup>4</sup>
plasma state of this nature; dissociation of CH4
of occurring particles forming a molecule [16].
reactor, enabling minimizing the ionization of CH<sup>4</sup>
and C), both of high purity [12, 13].
high efficiency in the conversion of CH<sup>4</sup>
flow of the mixture of CH<sup>4</sup>
catalytically assists the decomposition reaction of CH4
ated [12, 13, 16].
A CH<sup>4</sup>
a greater CH4
#### *3.1.1.2. The process*
The values of degradation of CH4 and the efficiency of H<sup>2</sup> production depend directly on the CH4 flow; these results do not depend on the argon flow. Important to note that the amount of CH4 converted is related to the total flow of CH<sup>4</sup> , relating to what has been transformed into solid and gaseous products. For example, when the flow of CH<sup>4</sup> was adjusted to 0.001 m3 /min, 59.4% of the gas was transformed into solids and the remainder 40.6% by gaseous products (see **Figure 5**) [12–16].
The proportional reduction in the conversion relates to the constant electric power applied to the torch; since the available power is controlled, the CH4 conversion is also limited and any additional amount of CH4 remains in the system as a nondegraded fraction. To correct this problem, argon flow was increased in the plasma torch, which caused the increase in plasma jet temperature and consequent increase in system power [11, 12].
The decomposition of CH4 to produce hydrogen and colloidal carbon is quite promising, if aggregated with other activities that use their products. As the CH<sup>4</sup> conversion function is endothermic, the thermal plasma is very efficient because in these conditions of high temperatures
**Figure 5.** Percentage of degradation of CH<sup>4</sup> (■) and efficiency of H<sup>2</sup> (▲).
several advantages in its application that are compensatory, as for example: increased efficiency, reduction of the reaction time, low cost of the equipment, favoring variety of productive size, and making possible its industrial application. The use of a noble gas as a plasma gas
process enabled the use of different sources of raw material, besides CH4 gas, butane and propane, but for this project, we chose to use the gases from landfill gas. Thus, the production of hydrogen as an energy source and that of colloidal carbon as an industrial input becomes attractive and compensates the energy and operational expenditure of the system [11–13, 15, 16].
The Chiron project has a thermal plasma system, which is composed of a plasma arc torch, and a reactor or pyrolysis chamber designed to retain most of the heat flux from the plasma jet. The inert gas argon (Ar) was used as a plasma gas and was introduced in the torch by a
The electric arc between the cathode and anode was initiated by a high-voltage discharge and a variable current. The formation of the plasma arc occurs between the cathode and the anode
The torch consisted of a cone-shaped cathode and an anode rivet, channel-shaped. The electric arc established between the electrodes (cathode and anode) was initiated by a high-voltage
into the reactor was heated and passed then into the liner, by contacting the plasma and suf-
and the efficiency of H<sup>2</sup>
59.4% of the gas was transformed into solids and the remainder 40.6% by gaseous products
The proportional reduction in the conversion relates to the constant electric power applied to
problem, argon flow was increased in the plasma torch, which caused the increase in plasma
thermic, the thermal plasma is very efficient because in these conditions of high temperatures
remains in the system as a nondegraded fraction. To correct this
to produce hydrogen and colloidal carbon is quite promising, if
flow; these results do not depend on the argon flow. Important to note that the amount of
discharge. The gas is heated, ionized, and emerged from the torch as a plasma jet. CH<sup>4</sup>
, and then by CH4
in the process and is only necessary to initiate plasma formation
, another important factor is that the plasma
, was introduced into the pyrolysis chamber
production depend directly on the
was adjusted to 0.001 m3
conversion is also limited and any
conversion function is endo-
, relating to what has been transformed into
injected
/min,
avoids the formation of CO<sup>2</sup>
234 Desalination
*3.1.1.1. The plasma system*
controlled flow [12, 13, 15, 16].
with controlled flow [12–14, 16].
fering degradation [11, 12, 16].
The values of degradation of CH4
*3.1.1.2. The process*
(see **Figure 5**) [12–16].
additional amount of CH4
The decomposition of CH4
CH4
CH4
(cooled by water). The gas to be degraded, CH4
converted is related to the total flow of CH<sup>4</sup>
solid and gaseous products. For example, when the flow of CH<sup>4</sup>
the torch; since the available power is controlled, the CH4
jet temperature and consequent increase in system power [11, 12].
aggregated with other activities that use their products. As the CH<sup>4</sup>
and was initially replaced by H2
and high degree of ionization, chemical reactions are accelerated by the power supply, and moreover, there is no CO<sup>2</sup> emissions or other contaminant gases, once the system is deoxygenated [12, 13, 16].
In the Unit 01 plasma pyrolysis reactor, CH<sup>4</sup> is broken down mainly by the heat from the arc plasma. However, the existence of a collision process between the existing elements in the plasma catalytically assists the decomposition reaction of CH4 . Different types of crashes happen in a plasma state of this nature; dissociation of CH4 and other molecules (CH3 , CH2 , and CH) may also occur as a result of impacts between the molecules, atoms, ions, electrons, and protons [11–15].
A CH<sup>4</sup> molecule can be dissociated into one or more parts to collide with another particle. Various collision processes may occur inside the plasma reactor, including the recombination of occurring particles forming a molecule [16].
The usage of Unit 01—reactor pyrolysis—has several advantages such as high energy density existing in the plasma which allows the construction of a compact reactor. Due to the plasma characteristics, a continuous flow is possible, enabling the formation of carbonaceous materials of different qualities and providing no pollution to obtain two commercial products (H<sup>2</sup> and C), both of high purity [12, 13].
A revolutionary constructive architecture for the treatment of methane, the reactor provides a greater CH4 residence time in the reaction zone and also a longer residence time inside the reactor, enabling minimizing the ionization of CH<sup>4</sup> molecules. This feature is important for a high efficiency in the conversion of CH<sup>4</sup> into H2 and C. In addition to presenting a favorable flow of the mixture of CH<sup>4</sup> and hydrogen flow, important for the heat exchange between the flows and good distribution of current lines coupled with good mixing between the flows, it is important to provide a homogenous treatment of particles of carbon [13, 15].
Thus, in the pyrolysis gas plasma, the energy required for reaction is supplied to the elements in the reaction chamber of the reactor by a high-energy electrical discharge. The electrical discharge at atmospheric pressure generates thermal plasma with high temperatures that can completely dissociate CH4 molecules. The molecular decomposition in the absence of oxygen generates hydrogen and carbon, which do not bother catalysts, for there is the formation of CO and CO<sup>2</sup> [12, 13].
The pyrolysis of CH4 does not produce CO or CO<sup>2</sup> ; the carbon in CH4 molecule does not react with the oxygen because it does not exist in the process. What happens in the process is the breaking of the molecules; resulting in isolated black carbon. The process allows a simple purification of hydrogen, since the solid carbon may be extracted by physical separation of the flow [12].
heating (= constant volume), isothermal expansion, and isochronous cooling. This is the idealized cycle (valid for perfect gases), which diverges from the actual cycle measured by instruments. Nevertheless, it is very close to the so-called Carnot Cycle, which establishes the
NEREU Project: Construction of a Plasma Reactor for Reform of Greenhouse Gases...
http://dx.doi.org/10.5772/intechopen.69313
237
The Stirling engine surprises for its simplicity, because they are of external combustion and are used in the process to generate energy, since they can work with more than one source of heat and thus are projected, receiving heat of the cooling water of the pyrolysis reactor, of the salt leaving the bottom of the salinization unit at a temperature of approximately 1000 K of the steam leaving the desalination unit and even of heating of solar origin (see **Figure 7**). Because they are very simple machines with easy maintenance, easy operation, quiet operation, and low vibration, these peculiarities made possible their use in a distributed way throughout the entire production chain, reducing the cost of energy and the need for large generation units. The use of the stirling engine for power generation is the best way to reduce costs and increase energy production. Thus, the option to use the stirling engine as driver for power generation units is due to its low cost of manufacturing, the ease of working with small
maximum theoretical limit of thermal machines performance.
**Figure 6.** Oxygen unit [13].
**Figure 7.** Prometheus unit [12, 13].
#### *3.1.2. Oxygen unit*
Oxygen is the most known and widespread of the gases, being the most abundant element in the earth's crust, about 46.6% by weight. To meet the oxygen needs of the Prometheus project, a little production plant of O<sup>2</sup> was created in this unit, taking advantage of the project's power generation. Thus, the O<sup>2</sup> production system followed a PSA ("Pressure Swing Adsorption") process, which is a classical mechanical O<sup>2</sup> production system [16].
Pressure swing adsorption—PSA allows the viability of obtaining oxygen at various on-site scales and its low operating cost was the main drivers of this process.
In this system, there is an oxygen production plant, which is installed next to the consumer (on site). This technology allows the separation and concentration of oxygen (approximately 95% O<sup>2</sup> ) by subjecting the ambient air to a zeolite molecular sieve under low pressure (3–6 bar) for a period sufficient to adsorb carbon monoxide, steam, CO<sup>2</sup> , and almost all of the nitrogen present in the air [13].
The process uses two metal vats containing molecular sieve (zeolite) in antiparallel, through six valves, which retains nitrogen from the air at high pressure, which will be released at low pressure, and allows the oxygen to cross the adsorbent bed as final product.
Also, there must exist in the process a filtration and drying step of the compressed air that is admitted in the sieves, protecting the zeolite from contamination so as not to compromise the efficiency of the sieve.
Thus, this is an alternative of lower cost and better energy efficiency, if compared to the cryogenic process, but produces a gas with a lower oxygen percentage, that is, a minimum purity of 92% (see **Figure 6**)
#### *3.1.3. Power generators—using stirling engine*
This type of engine operates with a thermodynamic cycle composed of four phases and executed in two-stroke piston: isothermal (= constant temperature) compression, isochronous
**Figure 6.** Oxygen unit [13].
flows and good distribution of current lines coupled with good mixing between the flows, it
Thus, in the pyrolysis gas plasma, the energy required for reaction is supplied to the elements in the reaction chamber of the reactor by a high-energy electrical discharge. The electrical discharge at atmospheric pressure generates thermal plasma with high temperatures that can completely dissociate CH4 molecules. The molecular decomposition in the absence of oxygen generates hydrogen and carbon, which do not bother catalysts, for there is the formation of
with the oxygen because it does not exist in the process. What happens in the process is the breaking of the molecules; resulting in isolated black carbon. The process allows a simple purification of hydrogen, since the solid carbon may be extracted by physical separation of the flow [12].
Oxygen is the most known and widespread of the gases, being the most abundant element in the earth's crust, about 46.6% by weight. To meet the oxygen needs of the Prometheus project,
Pressure swing adsorption—PSA allows the viability of obtaining oxygen at various on-site
In this system, there is an oxygen production plant, which is installed next to the consumer (on site). This technology allows the separation and concentration of oxygen (approximately
The process uses two metal vats containing molecular sieve (zeolite) in antiparallel, through six valves, which retains nitrogen from the air at high pressure, which will be released at low
Also, there must exist in the process a filtration and drying step of the compressed air that is admitted in the sieves, protecting the zeolite from contamination so as not to compromise the
Thus, this is an alternative of lower cost and better energy efficiency, if compared to the cryogenic process, but produces a gas with a lower oxygen percentage, that is, a minimum purity
This type of engine operates with a thermodynamic cycle composed of four phases and executed in two-stroke piston: isothermal (= constant temperature) compression, isochronous
) by subjecting the ambient air to a zeolite molecular sieve under low pressure (3–6 bar)
; the carbon in CH4
was created in this unit, taking advantage of the project's power
production system followed a PSA ("Pressure Swing Adsorption")
production system [16].
molecule does not react
, and almost all of the nitrogen
is important to provide a homogenous treatment of particles of carbon [13, 15].
does not produce CO or CO<sup>2</sup>
scales and its low operating cost was the main drivers of this process.
pressure, and allows the oxygen to cross the adsorbent bed as final product.
for a period sufficient to adsorb carbon monoxide, steam, CO<sup>2</sup>
CO and CO<sup>2</sup>
236 Desalination
The pyrolysis of CH4
*3.1.2. Oxygen unit*
95% O<sup>2</sup>
a little production plant of O<sup>2</sup>
process, which is a classical mechanical O<sup>2</sup>
generation. Thus, the O<sup>2</sup>
present in the air [13].
efficiency of the sieve.
of 92% (see **Figure 6**)
*3.1.3. Power generators—using stirling engine*
[12, 13].
heating (= constant volume), isothermal expansion, and isochronous cooling. This is the idealized cycle (valid for perfect gases), which diverges from the actual cycle measured by instruments. Nevertheless, it is very close to the so-called Carnot Cycle, which establishes the maximum theoretical limit of thermal machines performance.
The Stirling engine surprises for its simplicity, because they are of external combustion and are used in the process to generate energy, since they can work with more than one source of heat and thus are projected, receiving heat of the cooling water of the pyrolysis reactor, of the salt leaving the bottom of the salinization unit at a temperature of approximately 1000 K of the steam leaving the desalination unit and even of heating of solar origin (see **Figure 7**).
Because they are very simple machines with easy maintenance, easy operation, quiet operation, and low vibration, these peculiarities made possible their use in a distributed way throughout the entire production chain, reducing the cost of energy and the need for large generation units. The use of the stirling engine for power generation is the best way to reduce costs and increase energy production. Thus, the option to use the stirling engine as driver for power generation units is due to its low cost of manufacturing, the ease of working with small
**Figure 7.** Prometheus unit [12, 13].
generating units, and the distribution of several generating units throughout the production unit. Thus, Stirling engines are used along the production line, using the generated steam and heat of the gem salt, which is extracted by the lower nozzle of the Prometheus unit.
Thus, the reactor receives the salinized water mixed with all the solutes of the production of mariculture, which would become a contaminant very dangerous for the environment, mainly to the mangroves; although the concentration varies according to the source, the productive stage and the type of production and the relationship between the most abundant constituents is practically constant. Among these constituents are chlorides, sulphates,
NEREU Project: Construction of a Plasma Reactor for Reform of Greenhouse Gases...
http://dx.doi.org/10.5772/intechopen.69313
239
The reactor configuration used in plant is the type of rotor under pressure with an arrangement in which the saline solution is introduced into the chamber through a spray system; the spraying occurs on a vortex hydrogen plasma vaporizing it. One container captures the salts deposited in the bottom in the form of rock salt. The superheated steam is sent to Unit 03 in
The production of energy, water, and food in sufficient quantity and at competitive prices is important for economic development. At present, there is great concern about the environmental aspects of production in these segments, with emphasis on the use of renewable sources that do not degrade the environment and guarantee the productive continuity for
Therefore, it is no secret that fishery supplies are being affected by predatory fishing and discharges of chemical and organic waste. The disorderly growth of the world's population, coupled with the intensified demands of protein needs, has been draining the oceans. But, the solution found by mankind through aquaculture apparently solved the question of productive continuity and protein reserves. What is the problem with this activity? The vast majority of the water available for swimming pools is not suitable for return to the oceans and also is not suitable for the demands of aquaculture. These tailings are dumped in the oceans and can create areas of marine desertification in dump areas, which would lead to famine in many communities.
Over the years, the conviction that infrastructure in the world should be based on a mix of productive sources and self-sustainable systems, composing production triad—sustainability environment. In the same way, it offers great opportunities for expansion of new business initiatives and integrated production projects; they are relatively new and suffer the discrimination of the need for productive specialization to guarantee the maximum profit of the capi-
As the production of energy, food, and water treatment are important elements for the economic development of societies and the maintenance of the quality of life and health of families; in recent years, emphasis has been placed on the relationship between production and the environmental issue, considering possible reductions in the availability of nonrenewable
Thus, in order to evaluate these investments, there are indicators of analysis capable of assisting in the perception of the behavior between risk and return, the Nereu Project would have a
iodides, bacteria, fungi, protozoans among others.
order to generate energy (see **Figure 7**) [12, 13].
**4. The results analysis**
future generations.
tal employed.
natural resources.
#### **3.2. Prometheus project**
The need to obtain drinking water, increasing demand and neglect of much use by the population are the few factors because of which water scarcity begin to emerge in places where hitherto was not a problem. Water, unlike other natural resources, cannot be replaced, must be preserved, and distributed properly to avoid its loss.
Desalination is a technique that has been used for thousands of years in places where we cannot get enough fresh water. It is considered the future alternative to meet the needs of living beings, since 97.2% of the planet's water is salty or brackish. It is currently underutilized due to the high cost of the process, since it requires a lot of energy and sophisticated materials [12, 13].
Since the sixteenth century the desalination of sea water was used in vessels. Land desalination began in the 18th century and began to play an important role in the late 1940s and the early 1950s, especially in countries where potable water is scarce as in the Arab Gulf, USA, Caribbean and some areas of North America.
The composition of the salinized water varies according to the source, although the concentration of salts varies from place to place, the ratio between the most abundant constituents is practically constant. Among these constituents are sodium chloride, bicarbonates, calcium sulphate, and magnesium sulphate [12, 13].
Another alternative to obtain water is to reuse it, by replacing a source of drinking water with another of inferior quality where such substitution is possible, supplying the less restrictive demands made, releasing water of better quality for the noblest uses, such as domestic supplies.
The reuse of water in the urban sector can be accomplished by appropriate treatment of sewage and its reuse for potable purposes (indirect reuse) or non-potable (irrigation, fire reserve, dust control, decorative aquatic systems, etc.). In the industrial sector, the reuse of water is a frequent practice, since it reduces production costs. They are often used in cooling towers, boilers, civil construction, and maintenance of industrial facilities and within industrial processes. The recharge of aquifers is another form of water reuse [12, 13].
The Prometheus project uses a systematic rationalization in the use of water and energy, basing its process on a continuous control of what is generated and what is consumed. Thus, the project provides for the rational and intelligent use of wastewater from industrial processes, whether it has low oxygenation content and a high presence of dissolved salts, in order to desalinate and neutralize dissolved solutes in the water.
In Counterpart, there is to the production of electricity by the use of H2 and associated thermogeneration cycles, at low cost and with high profitability, constructed in a modular and interchangeable way, without affecting the organizational structure and the environment, and having by-products of the process generation of clean energy, gem salt and demineralized and deionized water [12, 13].
Thus, the reactor receives the salinized water mixed with all the solutes of the production of mariculture, which would become a contaminant very dangerous for the environment, mainly to the mangroves; although the concentration varies according to the source, the productive stage and the type of production and the relationship between the most abundant constituents is practically constant. Among these constituents are chlorides, sulphates, iodides, bacteria, fungi, protozoans among others.
The reactor configuration used in plant is the type of rotor under pressure with an arrangement in which the saline solution is introduced into the chamber through a spray system; the spraying occurs on a vortex hydrogen plasma vaporizing it. One container captures the salts deposited in the bottom in the form of rock salt. The superheated steam is sent to Unit 03 in order to generate energy (see **Figure 7**) [12, 13].
## **4. The results analysis**
generating units, and the distribution of several generating units throughout the production unit. Thus, Stirling engines are used along the production line, using the generated steam and
The need to obtain drinking water, increasing demand and neglect of much use by the population are the few factors because of which water scarcity begin to emerge in places where hitherto was not a problem. Water, unlike other natural resources, cannot be replaced, must
Desalination is a technique that has been used for thousands of years in places where we cannot get enough fresh water. It is considered the future alternative to meet the needs of living beings, since 97.2% of the planet's water is salty or brackish. It is currently underutilized due to the high cost of the process, since it requires a lot of energy and sophisticated materials [12, 13]. Since the sixteenth century the desalination of sea water was used in vessels. Land desalination began in the 18th century and began to play an important role in the late 1940s and the early 1950s, especially in countries where potable water is scarce as in the Arab Gulf, USA,
The composition of the salinized water varies according to the source, although the concentration of salts varies from place to place, the ratio between the most abundant constituents is practically constant. Among these constituents are sodium chloride, bicarbonates, calcium
Another alternative to obtain water is to reuse it, by replacing a source of drinking water with another of inferior quality where such substitution is possible, supplying the less restrictive demands made, releasing water of better quality for the noblest uses, such as domestic supplies. The reuse of water in the urban sector can be accomplished by appropriate treatment of sewage and its reuse for potable purposes (indirect reuse) or non-potable (irrigation, fire reserve, dust control, decorative aquatic systems, etc.). In the industrial sector, the reuse of water is a frequent practice, since it reduces production costs. They are often used in cooling towers, boilers, civil construction, and maintenance of industrial facilities and within industrial pro-
The Prometheus project uses a systematic rationalization in the use of water and energy, basing its process on a continuous control of what is generated and what is consumed. Thus, the project provides for the rational and intelligent use of wastewater from industrial processes, whether it has low oxygenation content and a high presence of dissolved salts, in order to
generation cycles, at low cost and with high profitability, constructed in a modular and interchangeable way, without affecting the organizational structure and the environment, and having by-products of the process generation of clean energy, gem salt and demineralized
and associated thermo-
cesses. The recharge of aquifers is another form of water reuse [12, 13].
In Counterpart, there is to the production of electricity by the use of H2
desalinate and neutralize dissolved solutes in the water.
and deionized water [12, 13].
heat of the gem salt, which is extracted by the lower nozzle of the Prometheus unit.
be preserved, and distributed properly to avoid its loss.
Caribbean and some areas of North America.
sulphate, and magnesium sulphate [12, 13].
**3.2. Prometheus project**
238 Desalination
The production of energy, water, and food in sufficient quantity and at competitive prices is important for economic development. At present, there is great concern about the environmental aspects of production in these segments, with emphasis on the use of renewable sources that do not degrade the environment and guarantee the productive continuity for future generations.
Therefore, it is no secret that fishery supplies are being affected by predatory fishing and discharges of chemical and organic waste. The disorderly growth of the world's population, coupled with the intensified demands of protein needs, has been draining the oceans. But, the solution found by mankind through aquaculture apparently solved the question of productive continuity and protein reserves. What is the problem with this activity? The vast majority of the water available for swimming pools is not suitable for return to the oceans and also is not suitable for the demands of aquaculture. These tailings are dumped in the oceans and can create areas of marine desertification in dump areas, which would lead to famine in many communities.
Over the years, the conviction that infrastructure in the world should be based on a mix of productive sources and self-sustainable systems, composing production triad—sustainability environment. In the same way, it offers great opportunities for expansion of new business initiatives and integrated production projects; they are relatively new and suffer the discrimination of the need for productive specialization to guarantee the maximum profit of the capital employed.
As the production of energy, food, and water treatment are important elements for the economic development of societies and the maintenance of the quality of life and health of families; in recent years, emphasis has been placed on the relationship between production and the environmental issue, considering possible reductions in the availability of nonrenewable natural resources.
Thus, in order to evaluate these investments, there are indicators of analysis capable of assisting in the perception of the behavior between risk and return, the Nereu Project would have a distinct characteristic for all desalination projects. To carry out this technical-economic evaluation, it is necessary to make a hybrid analysis, taking into account four basic characteristics of this project, which are as follows:
In spite of the Nereu project, a miniaturized version of the Chiron project and the equipment of the Prometheus project were coupled to its structure, and the average project yield
NEREU Project: Construction of a Plasma Reactor for Reform of Greenhouse Gases...
http://dx.doi.org/10.5772/intechopen.69313
241
The average yield in the project stems from the main constraint adopted in the operationalization and implementation of the Nereu project. In other words, due to budget constraints, it was only possible to use recycled materials, scrap, and/or products recovered from old
Thus, the great problem generated by the filtration of the shrimp production ponds and dumped in the pool 03AD, without having a real destination for this residue, was solved with the Nereu project, which desalinized and sterilized this residue, producing pure water and gem salt, totally free of any pathogenic element. At this point, the Nereu project stands out in relation to the other desalination processes, because while the other systems consume energy, the Nereu project produces its own energy to maintain the system and to be commercialized
When analyzing the feasibility of desalination projects, one should not only consider the question of return on investment, but also the impact on the entire productive and social context, especially for locations lacking these resources. It is therefore of utmost importance not to dwell on a simplistic reading of numbers and results. The Nereu project was a predetermined research funded by an investor with limited resources, with a limited value of US\$35,000.00
of the generation of residues, nor greenhouse gas emanations, in the operation of the system was verified. Thus, the combination of the processes added to a system of financial control would enable the recovery of the investment in a period of 10 months. This was made possible
Throughout the Nereu project's operating time, it was possible to operate at low production costs and it was observed that, based on the standards of taxes and input costs in Brazil, the project has a low maintenance cost, despite being fully developed with scrap. In this way, the average cost of the plant is low per hour of operation. The cost of each product, including taxes and operating costs, can be deployed as follows: pure water: US\$0.12 per 1 l/h; electric power: US\$0.037 kWh; Carbon: US\$1.46 per 1 kg/h; hydrogen: US\$1.46 per 1 kg/h; and rock
With the diversity of by-products generated, the Nereu project had a multiplicity of revenues that allowed the recovery of investments within 6 months, and an investment profit of around 10.6%, despite being a basic small plant, built from scraps. These results reflect the reality of the Brazilian market, incidental taxes and charges, labor costs, and employed automation. These values will change according to the economic, fiscal, and equipment conditions present
, the nonexistence
was 79.6%.
irons.
in the energy market.
**4.2. Economic resulting**
salts: US\$ 0.028 per 1 kg/h.
in each country.
to prove that it was technically and economically viable.
With the architecture of the project working with the thermal cycle of the H<sup>2</sup>
by the generation of several marketable by-products that have added value.
It is worth noting that the Nereu project aimed to reduce the risks of contamination in the oceans caused by aquiculture tailings, supported by a demand from CONAMA, which provided for the eradication of pathogenic releases to coastal areas.
The project, in an experimental phase, was built in a proto-unit system in the region of the *Dende* Coast, in the city of *Jaguaripe*–*Bahia*, Brazil, operating for 8 months, in laboratory scale, with the support of *Funil* Farm, which invested in materials and assembly, with the purpose of neutralizing the pathogenic elements of the pool 03AD, with volume of 3.5 m<sup>3</sup> , purge tank used to deposit filtration wastes from other pools. Thus, the Nereu project operated to make the cleaning residual water from the existing filtration system.
#### **4.1. Techniques resulting**
The Nereu project used its two reactors and seven generators with stirling engine to treat waste water from *Funil* Farm, generating energy and producing food. To perform this procedure, the Chiron Project was miniaturized to work with limited flow rates to meet the needs of the Prometheus project. Thus, the project was conceived as a pilot plant with the following technical characteristics: total flow: 100 l/h; CH<sup>4</sup> flow rate: 400 l/h; and average concentration of salts: 35 g/l.
It was verified that the Nereu project consumes a total of 25.4 kWh of energy to meet the energetic demands of desalination and decontamination of the waste water, and the reactor consumption, in maximum efficiency, for CH<sup>4</sup> pyrolysis, is 7.2 kWh, with advantages over other techniques, as the process does not release CO<sup>2</sup> .
The percentage of degradation of CH4 reaches the maximum value of 98.8% for an applied power of 7.2 kWh. In order to cope with the energy needs of the system, seven stirling engines of the Alpha type have been adapted which have been developed for other purposes and reused for the project. Thus, it was possible to adapt them for power generation, producing a total of 40 kWh of energy, which has its stepped drive, being connected as the energy needs are presented. The Nereu project consumed 54.4% of the energy produced, and 45.6% was made available to the electricity grid of the farm.
The C produced by the degradation of CH4 in thermal plasma has hydrophobic characteristics, is amorphous and of high purity, presenting added value and can be commercialized, for companies producing carbon fibber and carbon nanotubes.
In spite of the Nereu project, a miniaturized version of the Chiron project and the equipment of the Prometheus project were coupled to its structure, and the average project yield was 79.6%.
The average yield in the project stems from the main constraint adopted in the operationalization and implementation of the Nereu project. In other words, due to budget constraints, it was only possible to use recycled materials, scrap, and/or products recovered from old irons.
Thus, the great problem generated by the filtration of the shrimp production ponds and dumped in the pool 03AD, without having a real destination for this residue, was solved with the Nereu project, which desalinized and sterilized this residue, producing pure water and gem salt, totally free of any pathogenic element. At this point, the Nereu project stands out in relation to the other desalination processes, because while the other systems consume energy, the Nereu project produces its own energy to maintain the system and to be commercialized in the energy market.
#### **4.2. Economic resulting**
distinct characteristic for all desalination projects. To carry out this technical-economic evaluation, it is necessary to make a hybrid analysis, taking into account four basic characteristics
• Technique: development of new technologies; reducing the emission of greenhouse gases into the atmosphere; producing water and energy for regions lacking these resources.
• Economic: the low cost of implementation, the low maintenance cost, and diversification
• Environmental: carbon sequestration, eliminating greenhouse gases, and producing water
It is worth noting that the Nereu project aimed to reduce the risks of contamination in the oceans caused by aquiculture tailings, supported by a demand from CONAMA, which pro-
The project, in an experimental phase, was built in a proto-unit system in the region of the *Dende* Coast, in the city of *Jaguaripe*–*Bahia*, Brazil, operating for 8 months, in laboratory scale, with the support of *Funil* Farm, which invested in materials and assembly, with the purpose
used to deposit filtration wastes from other pools. Thus, the Nereu project operated to make
The Nereu project used its two reactors and seven generators with stirling engine to treat waste water from *Funil* Farm, generating energy and producing food. To perform this procedure, the Chiron Project was miniaturized to work with limited flow rates to meet the needs of the Prometheus project. Thus, the project was conceived as a pilot plant with the following technical
It was verified that the Nereu project consumes a total of 25.4 kWh of energy to meet the energetic demands of desalination and decontamination of the waste water, and the reactor
power of 7.2 kWh. In order to cope with the energy needs of the system, seven stirling engines of the Alpha type have been adapted which have been developed for other purposes and reused for the project. Thus, it was possible to adapt them for power generation, producing a total of 40 kWh of energy, which has its stepped drive, being connected as the energy needs are presented. The Nereu project consumed 54.4% of the energy produced, and 45.6% was
tics, is amorphous and of high purity, presenting added value and can be commercialized, for
.
flow rate: 400 l/h; and average concentration of salts: 35 g/l.
reaches the maximum value of 98.8% for an applied
in thermal plasma has hydrophobic characteris-
pyrolysis, is 7.2 kWh, with advantages over
, purge tank
of neutralizing the pathogenic elements of the pool 03AD, with volume of 3.5 m<sup>3</sup>
of this project, which are as follows:
240 Desalination
of income and production.
**4.1. Techniques resulting**
characteristics: total flow: 100 l/h; CH<sup>4</sup>
The percentage of degradation of CH4
consumption, in maximum efficiency, for CH<sup>4</sup>
made available to the electricity grid of the farm.
companies producing carbon fibber and carbon nanotubes.
The C produced by the degradation of CH4
other techniques, as the process does not release CO<sup>2</sup>
and energy without generating contaminant residues.
vided for the eradication of pathogenic releases to coastal areas.
the cleaning residual water from the existing filtration system.
When analyzing the feasibility of desalination projects, one should not only consider the question of return on investment, but also the impact on the entire productive and social context, especially for locations lacking these resources. It is therefore of utmost importance not to dwell on a simplistic reading of numbers and results. The Nereu project was a predetermined research funded by an investor with limited resources, with a limited value of US\$35,000.00 to prove that it was technically and economically viable.
With the architecture of the project working with the thermal cycle of the H<sup>2</sup> , the nonexistence of the generation of residues, nor greenhouse gas emanations, in the operation of the system was verified. Thus, the combination of the processes added to a system of financial control would enable the recovery of the investment in a period of 10 months. This was made possible by the generation of several marketable by-products that have added value.
Throughout the Nereu project's operating time, it was possible to operate at low production costs and it was observed that, based on the standards of taxes and input costs in Brazil, the project has a low maintenance cost, despite being fully developed with scrap. In this way, the average cost of the plant is low per hour of operation. The cost of each product, including taxes and operating costs, can be deployed as follows: pure water: US\$0.12 per 1 l/h; electric power: US\$0.037 kWh; Carbon: US\$1.46 per 1 kg/h; hydrogen: US\$1.46 per 1 kg/h; and rock salts: US\$ 0.028 per 1 kg/h.
With the diversity of by-products generated, the Nereu project had a multiplicity of revenues that allowed the recovery of investments within 6 months, and an investment profit of around 10.6%, despite being a basic small plant, built from scraps. These results reflect the reality of the Brazilian market, incidental taxes and charges, labor costs, and employed automation. These values will change according to the economic, fiscal, and equipment conditions present in each country.
#### **4.3. Environmental results**
The development of projects based on environmental preservation attitudes is part of the goal plan of any research project prepared by GPAp. The goal proposed in any project of the research group is in the search for the development of actions with emphasis on sustainable development projects, the insertion of needy communities, in generation of new projects, always focused on the protection of the environment.
It should be noted that the project's application potential is directly associated with its uniqueness, simplicity, portability, modularity, and constructive speed, so the Nereu project is fully applicable to work in consortium with other desalination systems, producing energy to sustain them and treat the waste from the filtration without the environment being attacked, as it works as an alternative to take advantage of the wastewater potential without the need for large engineering projects and civil works for its implementation. In addition, the generation of energy is justified by the thermal availability of the process, compared to the conventional production of separate thermal and electromechanical energy, which, from this point of view, has led to the search for the most appropriate technology in order to provide the highest possible energy efficiency, therefore the option to use stirling engines to power the generators, due to its low cost of production and ease
NEREU Project: Construction of a Plasma Reactor for Reform of Greenhouse Gases...
http://dx.doi.org/10.5772/intechopen.69313
243
In the process, it was possible to determine that the most relevant costs in the project come from four basic factors: energy, pure water, natural gas, and oxygen, which have a much greater influence than other costs. In this way, the Nereu project prioritized to minimize costs in three of the four items, which were energy, oxygen and water, directly impacting the final
As the Nereu project produces pure water, this productive cost is minimized by using the water that is provided by the Chiron subproject. With this advantage in mind, stirling engines are used for energy production, since they can be introduced at any stage of the production system in order to generate energy with any existing thermal potential, minimizing costs, and dependence on external energy to the project, stirling engines have low production and
ANA National Water Agency
HDI Human Development Index MMA Ministry of the Environment
CONAMA National Council for the Environment MPA Ministry of Fisheries and Aquaculture
IBGE Brazilian Institute of Geography and Statistics
BR Brazil
Ar Argon
CH4 Methane H2 Hydrogen NG Natural gas O<sup>2</sup> Oxygen
CO Carbon monoxide
of maintenance.
cost of the project.
maintenance costs.
**Nomenclature**
Contained in the thinking of sustainability is the possibility of change in the relationship between the individual and the environment. From the normative point of view, it is the necessary adoption of attitudes more coherent with the demands of environmental conservation. Thus, the development of an education strategy that allows better relations with the environment is undoubtedly one of the most strategic and effective actions to achieve the objectives proposed by the notion of sustainability.
With these bases, the Nereu project was built with a sustainable development procedure, so as not to generate any residue and to use only what one cannot take advantage of. Using natural resources such as hypersaline waters, greenhouse gases and scrap, sequestering C and generating pure water. The main characterization of sustainable development is the assurance of environmental concern. As, the plasma reactor did not generate CO<sup>2</sup> , decomposed CH4 , with C sequestration, which can be sold or used in the production of shrimp.
The salt produced by the Nereu project was sent for testing at the university. It was verified that there were no pathogens present in the product. This had a very relevant chemical balance of chlorides, iodides, sulphides, and carbonates in the mixture, similar to the rock salt that is produced in mines in Europe and Asia. As the salt produced was processed from wastewater treatments, despite being completely free of pathogens and contaminants, under Brazilian laws, the rock salt could only be sold for animal feed.
Thus, serious environmental problems from desalination plants, such as high energy consumption, greenhouse gases, and waste generation were solved in the Nereu Project, which processed and neutralized greenhouse gases, generated its own energy for desalination, and treated the waste from the desalination process.
## **5. Final considerations**
The Nereu project combined two different guideline projects, combining a plasma reactor for H2 generation, a steam desalination unit and a turbo-generator of electrical energy in a totally clean system. This was created to work with desalination residues. In this way, a system of treatment of contaminated hypersaline waters of any origin has been developed. To achieve these goals, the Nereu project has adopted a sustainable development rule and, consequently, the nonproduction of waste that damages the environment.
The Nereu project, when it unified the two projects, created a totally high sustainable and closed process, ensuring that the most critical points in Prometheus, which is continuity of energy supply and combustion for thermal energy generation, were fully met by Chiron.
It should be noted that the project's application potential is directly associated with its uniqueness, simplicity, portability, modularity, and constructive speed, so the Nereu project is fully applicable to work in consortium with other desalination systems, producing energy to sustain them and treat the waste from the filtration without the environment being attacked, as it works as an alternative to take advantage of the wastewater potential without the need for large engineering projects and civil works for its implementation. In addition, the generation of energy is justified by the thermal availability of the process, compared to the conventional production of separate thermal and electromechanical energy, which, from this point of view, has led to the search for the most appropriate technology in order to provide the highest possible energy efficiency, therefore the option to use stirling engines to power the generators, due to its low cost of production and ease of maintenance.
In the process, it was possible to determine that the most relevant costs in the project come from four basic factors: energy, pure water, natural gas, and oxygen, which have a much greater influence than other costs. In this way, the Nereu project prioritized to minimize costs in three of the four items, which were energy, oxygen and water, directly impacting the final cost of the project.
As the Nereu project produces pure water, this productive cost is minimized by using the water that is provided by the Chiron subproject. With this advantage in mind, stirling engines are used for energy production, since they can be introduced at any stage of the production system in order to generate energy with any existing thermal potential, minimizing costs, and dependence on external energy to the project, stirling engines have low production and maintenance costs.
### **Nomenclature**
**4.3. Environmental results**
242 Desalination
always focused on the protection of the environment.
environmental concern. As, the plasma reactor did not generate CO<sup>2</sup>
Brazilian laws, the rock salt could only be sold for animal feed.
the nonproduction of waste that damages the environment.
treated the waste from the desalination process.
**5. Final considerations**
H2
C sequestration, which can be sold or used in the production of shrimp.
proposed by the notion of sustainability.
The development of projects based on environmental preservation attitudes is part of the goal plan of any research project prepared by GPAp. The goal proposed in any project of the research group is in the search for the development of actions with emphasis on sustainable development projects, the insertion of needy communities, in generation of new projects,
Contained in the thinking of sustainability is the possibility of change in the relationship between the individual and the environment. From the normative point of view, it is the necessary adoption of attitudes more coherent with the demands of environmental conservation. Thus, the development of an education strategy that allows better relations with the environment is undoubtedly one of the most strategic and effective actions to achieve the objectives
With these bases, the Nereu project was built with a sustainable development procedure, so as not to generate any residue and to use only what one cannot take advantage of. Using natural resources such as hypersaline waters, greenhouse gases and scrap, sequestering C and generating pure water. The main characterization of sustainable development is the assurance of
The salt produced by the Nereu project was sent for testing at the university. It was verified that there were no pathogens present in the product. This had a very relevant chemical balance of chlorides, iodides, sulphides, and carbonates in the mixture, similar to the rock salt that is produced in mines in Europe and Asia. As the salt produced was processed from wastewater treatments, despite being completely free of pathogens and contaminants, under
Thus, serious environmental problems from desalination plants, such as high energy consumption, greenhouse gases, and waste generation were solved in the Nereu Project, which processed and neutralized greenhouse gases, generated its own energy for desalination, and
The Nereu project combined two different guideline projects, combining a plasma reactor for
The Nereu project, when it unified the two projects, created a totally high sustainable and closed process, ensuring that the most critical points in Prometheus, which is continuity of energy supply and combustion for thermal energy generation, were fully met by Chiron.
generation, a steam desalination unit and a turbo-generator of electrical energy in a totally clean system. This was created to work with desalination residues. In this way, a system of treatment of contaminated hypersaline waters of any origin has been developed. To achieve these goals, the Nereu project has adopted a sustainable development rule and, consequently,
, decomposed CH4
, with
## **Author details**
André Pedral S. de Sena
Address all correspondence to: apedral@gmail.com
Universidade Federal da Bahia / UFBa, Salvador, Brazil
## **References**
[1] FAO—Food and Agriculture Organization of the United Nations. The State of the World Fisheries and Aquaculture 2012. New York: FAO; 2012. p. 345
[8] Conama – Conselho Nacional De Meio Ambiente (National Environmental Council). Relatório de Maricultura e Regras de Implantação (Mariculture Report and Deployment Rules) [Internet]. 2015 [Updated: 2015]. Available from: http://www.mma.gov.br/port/
NEREU Project: Construction of a Plasma Reactor for Reform of Greenhouse Gases...
http://dx.doi.org/10.5772/intechopen.69313
245
[9] MPA – Ministério de Pesca e Aquicultura (Ministry of Fisheries and Aquaculture). Guia de Orientação para A Regularização da Aquicultura em Águas da União (Guidance for the Regularization of Aquaculture in the Waters of the Union). 1st ed. Brasília: MPA; 2015
[10] Dubois S, Marin-Leal JC, Ropert M, Lefebvre S. Effects of oyster farming on macrofaunal assemblages associated with Lanice conchilega tubeworm populations: A trophic analy-
[11] Rogdakis ED, Antonakos GD, Koronaki IP. Thermodynamic analysis and experimental investigation of a Solo V161 Stirling cogeneration unit. (Original Research Article).
[12] de Sena APS, Coutinho OA; Lima JR, Ailton de SS. Neptune project: A reactor of cold plasma to treat seawater for use in the food production. In: Terrell J, editor. Water
[13] Chiron project: Reform of methane from a cold plasma reactor with hydrogen production for desalination and energy cogeneration. Energy Conversion and Management.
[14] Kim KS, Seo JH, Nam JS, Ju WT, Hong SH. Production of hydrogen and black carbon by methane decomposition using DC-RF hybrid thermal plasmas. IEEE Transactions on
[15] Zhanghua WU, Guoyao YU, Limin Z, Wei D, Ercang L. Development of a 3 kW doubleacting thermoacoustic stirling electric generator. Applied Energy. 2014;**136**(1):866-872
[16] Gallego JS, Barrault J, Batiot-Dupeyrat C, Mondragón F. Production of hydrogen and
perovskite.
MWCNTs by methane decomposition over catalysts originated from LaNiO<sup>3</sup>
conama/relatórios/relatorioanual.pdf [Accessed: May 23, 2016]
sis using natural stable isotopes. Aquaculture. 2007; **271**:336-349
Filtration System. 1st ed. New York: Nova Publishers; 2016. pp. 27-70
Energy. 2012;**45**(1):503-511
Plasma Science. 2005;**33**:813-823
Catalysis Today. 2010;**149**:365-371
2014;**85**:933-943
**Author details**
CH3 , CH2
244 Desalination
**References**
27-72
André Pedral S. de Sena
& M. O'Keefe; 2011
Address all correspondence to: apedral@gmail.com
Universidade Federal da Bahia / UFBa, Salvador, Brazil
GPAp Apollo Research Group
, CH Methane decomposition chains
C Carbon black CO<sup>2</sup> Carbon dioxide
HC's Hydrocarbons NO<sup>x</sup> Nitrous oxide SO<sup>2</sup> Sulfur dioxide
[1] FAO—Food and Agriculture Organization of the United Nations. The State of the World
[2] Hall SJ, Delaporte A, Phillips MJ, Beveridge MCM. Blue Frontiers: Managing the Environmental Costs of Aquaculture. The WorldFish Center. 1st ed. Penang, Malaysia: Beveridge
[3] Van Kouwen F, Dieprink C, Schot P, Wassen N. Applicability of decision support systems for integrated coastal zone management. Coastal Management. 2008;**36**:19-34 [4] Boscardin NR. *A produção aquícola brasileira* (Brazilian aquaculture production). In: Ostrensky A, Borghetti JR, Soto D. (Orgs.), editors. Aquicultura no Brasil: o desafio é crescer (Aquaculture in Brazil: the challenge is to grow). 1st ed. Brasilia: UNB; 2014. pp.
[5] Vita R, Marin A, Jimenez-Brinquis B, Cesar A, Marin-Guirao L, Borredat M. Aquaculture of Bluefin tuna in the Mediterranean: Evaluation of organic particulate wastes.
[6] Plano Nacional de Recursos Hídricos - quinquênio 2014/2019 (National Water Resources
[7] ANA (Water National Agency). Relatório Avaliativo dos Recursos Hídricos do Brazil - 2014 (Evaluation Report of the Brazilian Water Resources - 2014). 1st ed. Brasilia: ANA; 2015.p. 274
Fisheries and Aquaculture 2012. New York: FAO; 2012. p. 345
Aquaculture Environment Interact. 2010;**69**:57-70
Plan - 2014/2019). 1st ed. Brasilia: ANA; 2014. p. 476
**Chapter 12**
unit is
effec-
) in
**Low-Cost Multi-Effect Solar Still: Alternative**
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.68365
Pak Hunkyun
**Abstract**
tive area. 9kg/m<sup>2</sup>
for speeding up implementation.
heat, personal, individual
**1. Introduction**
**Appropriate Technology for Personal Desalination**
Multi effect solar still (MES) has a stack of multiple layers for evaporation and condensation. The latent heat dissipated during condensation at the front layers are repeatedly recycled for evaporation at the back layers to increase overall desalination productivity. Despite of high efficiency and long history, MES has not been widely used yet, because of relative high cost. In this chapter, newly designed MES is introduced. Since it has low cost, light weight material and simple structure, it could be easily mass even at less developed country. The cost of production for a 1 m<sup>2</sup>
expected to be less than 300 USD. Structural features are introduced with experimental result which was outdoor tested with homemade lab prototype with 0.219 m<sup>2</sup>
Seoul, Korea, which is close to WHO's recommended minimal daily water supply for individuals (7.5~15 liters). For more practical implementation, further development on prototype and production process should be made as well as long term outdoor test under actual climate it would be used. Worldwide collaboration would be necessary
Various desalination methods have been developed since clean water is one of the most vital resources for human life. Generally, the unit cost of fresh water production tends to decrease dramatically, as the total scale of production has increased [1]. Therefore, mega-desalination
> © 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
**Keywords:** solar still, multi-effect solar still, desalination, low cost, appropriate technology, water purification, mass production, small scale, renewable energy, solar
per day of fresh water was obtained at sunny day (19.5MJ/m<sup>2</sup>
## **Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination**
## Pak Hunkyun
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.68365
#### **Abstract**
Multi effect solar still (MES) has a stack of multiple layers for evaporation and condensation. The latent heat dissipated during condensation at the front layers are repeatedly recycled for evaporation at the back layers to increase overall desalination productivity. Despite of high efficiency and long history, MES has not been widely used yet, because of relative high cost. In this chapter, newly designed MES is introduced. Since it has low cost, light weight material and simple structure, it could be easily mass even at less developed country. The cost of production for a 1 m<sup>2</sup> unit is expected to be less than 300 USD. Structural features are introduced with experimental result which was outdoor tested with homemade lab prototype with 0.219 m<sup>2</sup> effective area. 9kg/m<sup>2</sup> per day of fresh water was obtained at sunny day (19.5MJ/m<sup>2</sup> ) in Seoul, Korea, which is close to WHO's recommended minimal daily water supply for individuals (7.5~15 liters). For more practical implementation, further development on prototype and production process should be made as well as long term outdoor test under actual climate it would be used. Worldwide collaboration would be necessary for speeding up implementation.
**Keywords:** solar still, multi-effect solar still, desalination, low cost, appropriate technology, water purification, mass production, small scale, renewable energy, solar heat, personal, individual
### **1. Introduction**
Various desalination methods have been developed since clean water is one of the most vital resources for human life. Generally, the unit cost of fresh water production tends to decrease dramatically, as the total scale of production has increased [1]. Therefore, mega-desalination
© 2017 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
plants have become a popular trend in modern civilization. However, smaller scale desalination is still important for many applications: (i) It could supply water in remote areas where a water supply line is hardly available. (ii) It would be useful for people who cannot afford large investment for a water plant. (iii) It could also be useful as a temporal water source for many people, such as nomads, campers, sailors, or survivors of various disasters.
Therefore, substantial improvement of productivity should be made on solar still for more practical applications. Multi-effect solar still (MES) was developed to overcome this low productivity problem, by recycling wasted latent heat repeatedly [8–27]. As shown in **Figure 1**, in MES, multiple layers of evaporating wicks and condensing surface are stacked together [14]. Saline water flowing through first evaporation wick is partially evaporated, while the condensed saline water is drained out. The water vapour condensed on the first condensation surface would be collected as fresh water. During the condensation process, dissipated latent heat is reused to evaporate saline water running on the second evaporation wick. By repeating this process on multiple layers, the overall water productivity is very much increased. The Tanaka's group proved both theoretically and experimentally that MES can produce over 10–20 L of fresh water per day with a square metre effective area [8], which is comparable
Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination
http://dx.doi.org/10.5772/intechopen.68365
249
Though MES has more than half century history [2, 3] and performance verification, it is not widely used on practical application yet because of relatively high cost. Metal plates covered with fabric, glass cover, metal or wooden frames and airtight vapour sealants were commonly used materials for traditional MES system. Though they enable rigid and effective MES structure, but still they are relatively expensive and heavy material, which increase overall system
**Figure 1.** Typical structure of MES. Saline water in upper troughs (114, 124, 134, 144) flows through evaporation wicks (112, 122, 132, 142) and condensed on the condensation surfaces (121, 131, 141) to be collected through the collectors
amount to that WHO recommended.
cost.
(123, 133, 143).
Thanks to modern technology development, efficient small-scale desalinators became frequently used. Portable reverse osmosis water purifiers powered by electric or fuel engine are commonly used instruments for such purposes. However, such instruments are still too expensive, hard to supply energy source or difficult to maintain for many people in the world, who cannot afford substantial budget, resource or education. For such cases, the desalination method is appropriate technology, which is affordable to the actual users in do-it-yourself (DIY) style and would be a more realistic option than the expensive high-technology solution.
Solar stills are one of the simplest ways of desalination. In solar still, saline water is evaporated by solar heat and condensed to become fresh water. Though it is old, primitive tool, but is still useful for many people. It can be simply made out of low-cost common material, such as plastic sheet or bottles. Therefore, it could be easily installed and operated with doit-yourself manner. Various solar stills were made and used for long history [2–6]. However, the extreme low productivity has limited its application only to emergency or temporal water source. Considering that the latent heat of water vapour is about 540 cal/g or 2.26 kJ/g and assuming solar energy per unit area as 1 kJ/s m<sup>2</sup> (equivalent to 1 sun, which is nominal full sun intensity on bright clear day on earth), we may get about 1.6 kg of distilled water with 1 m<sup>2</sup> solar still in an hour. However, even this is only a theoretical maximum value. In reality, only few cups of fresh water could be obtained daily, since there is additional energy loss by vapour leakage or heat dissipation. Though several cups of water should be valuable to save thirsty sufferer, but it might not be enough for sustainable use in everyday life. Of course, we can enlarge the area of solar still to increase water production. However, in many cases, sunny land is also finite resource to be occupied only by solar still.
As shown in **Table 1**, WHO, the world health organization, recommended 2.5–3 L of drinking water per day per person for survival, including the basic hygiene and cooking purpose, at least 7.5–15 L per day per person is required [7].
**Table 1.** Minimal amount of fresh water required for a human living, recommended by the World Health Organization.
Therefore, substantial improvement of productivity should be made on solar still for more practical applications. Multi-effect solar still (MES) was developed to overcome this low productivity problem, by recycling wasted latent heat repeatedly [8–27]. As shown in **Figure 1**, in MES, multiple layers of evaporating wicks and condensing surface are stacked together [14].
plants have become a popular trend in modern civilization. However, smaller scale desalination is still important for many applications: (i) It could supply water in remote areas where a water supply line is hardly available. (ii) It would be useful for people who cannot afford large investment for a water plant. (iii) It could also be useful as a temporal water source for
Thanks to modern technology development, efficient small-scale desalinators became frequently used. Portable reverse osmosis water purifiers powered by electric or fuel engine are commonly used instruments for such purposes. However, such instruments are still too expensive, hard to supply energy source or difficult to maintain for many people in the world, who cannot afford substantial budget, resource or education. For such cases, the desalination method is appropriate technology, which is affordable to the actual users in do-it-yourself (DIY) style and would be a more realistic option than the expensive high-technology solution. Solar stills are one of the simplest ways of desalination. In solar still, saline water is evaporated by solar heat and condensed to become fresh water. Though it is old, primitive tool, but is still useful for many people. It can be simply made out of low-cost common material, such as plastic sheet or bottles. Therefore, it could be easily installed and operated with doit-yourself manner. Various solar stills were made and used for long history [2–6]. However, the extreme low productivity has limited its application only to emergency or temporal water source. Considering that the latent heat of water vapour is about 540 cal/g or 2.26 kJ/g and
sun intensity on bright clear day on earth), we may get about 1.6 kg of distilled water with 1
As shown in **Table 1**, WHO, the world health organization, recommended 2.5–3 L of drinking water per day per person for survival, including the basic hygiene and cooking purpose, at
Survival (drinking and food) 2.5–3 lpd Depends on climate and individual
Basic hygiene practices 2–6 lpd Depends on social and cultural
Basic cooking needs 3–6 lpd Depends on food type, social and
**Table 1.** Minimal amount of fresh water required for a human living, recommended by the World Health Organization.
Total 7.5–15 lpd Lpd: Litres per day
solar still in an hour. However, even this is only a theoretical maximum value. In reality, only few cups of fresh water could be obtained daily, since there is additional energy loss by vapour leakage or heat dissipation. Though several cups of water should be valuable to save thirsty sufferer, but it might not be enough for sustainable use in everyday life. Of course, we can enlarge the area of solar still to increase water production. However, in many cases, sunny
(equivalent to 1 sun, which is nominal full
physiology
cultural norms
norms
many people, such as nomads, campers, sailors, or survivors of various disasters.
assuming solar energy per unit area as 1 kJ/s m<sup>2</sup>
least 7.5–15 L per day per person is required [7].
land is also finite resource to be occupied only by solar still.
**Type of need Quantity Comments**
m<sup>2</sup>
248 Desalination
Saline water flowing through first evaporation wick is partially evaporated, while the condensed saline water is drained out. The water vapour condensed on the first condensation surface would be collected as fresh water. During the condensation process, dissipated latent heat is reused to evaporate saline water running on the second evaporation wick. By repeating this process on multiple layers, the overall water productivity is very much increased. The Tanaka's group proved both theoretically and experimentally that MES can produce over 10–20 L of fresh water per day with a square metre effective area [8], which is comparable amount to that WHO recommended.
Though MES has more than half century history [2, 3] and performance verification, it is not widely used on practical application yet because of relatively high cost. Metal plates covered with fabric, glass cover, metal or wooden frames and airtight vapour sealants were commonly used materials for traditional MES system. Though they enable rigid and effective MES structure, but still they are relatively expensive and heavy material, which increase overall system cost.
**Figure 1.** Typical structure of MES. Saline water in upper troughs (114, 124, 134, 144) flows through evaporation wicks (112, 122, 132, 142) and condensed on the condensation surfaces (121, 131, 141) to be collected through the collectors (123, 133, 143).
**Figure 1** represents typical inner structure of MES. It may include upper saline water trough, vapour condensing plates, wicks where the saline water evaporates, condensed fresh water collecting guide trough. The structures are packed in sun light passing windows and housing.
For higher productivity in MES, intervals between each evaporation wick and condensing surface should be minimized [9, 11]. However, if the distance is too short, the layers could touch each other, so that water droplet on the condensing surface can move to evaporation wick, which causes loss of distilled water. Similarly, if the saline water overflow into condensing surface, serious contamination of produced water could be occurred. Therefore, the wick and condensing surface should be securely isolated by air gap. Despite air gaps between the layers, there is always a chance of contact within layers because of impact, vibration, and gravitational deformation. To minimize such deformation, various approaches were made. Rigid metallic plates were commonly used for stable condensing layer. However, it resulted in increase of cost and system weight. Spacers were placed between the layers [4]. However, water droplet may flow back to wick along the spacer if too much spacers were applied, which may reduce productivity. The layers were placed vertically to avoid gravitational deformation [8–13]. However, it may increase discrepancy between solar incident angle to the system window, which reduces solar energy usage. Reflecting mirrors [12, 24] or tilted window [8, 9, 13, 27] could be placed in front of the vertical layers to maximize solar intensity. However, it may increase the system cost.
To solve this high-cost problem, new MES was designed and tested. To reduce production cost, alternative light-weight and low-cost material was used instead of such expensive materials. Structure of each component was also designed for easier mass production, which may reduce cost. In the next section, the materials and structure of the low-cost MES would be described.
> other hands, tight vapour sealing and thermal insulation are essential for assuring high productivity. In this system, sealing, spacers and inner frames altogether are constructed in one
> **Figure 2.** The origami structure of wick/condenser layer. Each part of the structure would have role of the evaporation wick (A), condensing surface (B), saline water feeding trough (or pocket) (C), condensed fresh water guiding channel (D)
Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination
http://dx.doi.org/10.5772/intechopen.68365
251
Vertically elongated spacers are connected to upper and lower parts of the frame. The spacer
It has a role of internal framework structure standing for each layer of evaporation wick/condensation layer film. The spacer layers and wick/condensing films are alternately stacked as shown in **Figure 4**. Though the film is too flexible to be stand alone, the film-spacer combined
The frame holds each vertical spacer in right position and orientation. To avoid a layer's film touching the next layer film, it is important that each spacer on a layer should be exactly
layer is stiff enough to be stand. The stiffness increases by stacking multiple layers.
structure, the spacer layer. Its structural features are shown in **Figure 3** [14].
and draining guide which separate the fresh water and waste water (F).
layer has following roles.
*Role 1: Internal framework for the MES*
*Role 2: Fixture for the vertical spacers*
superposed on top of the next layer's spacer.
## **2. Structure and material of the low-cost MES**
#### **2.1. Design of the wick/condenser laminated film layer**
In this work, flexible thin plastic film laminated to black fabric wick was used instead of stainless plate covered with fabric wick [14, 15]. The laminated film is obviously low-cost, light-weight alternative than metallic plates. It could be produced by roll-to-roll process, which is mass production favourable. The laminated film was folded like origami as shown in **Figure 2** [14].
Evaporation wick, condensation surface, upper saline water trough, condensed water guiding channel and concentrated saline water draining guide could be included in this single origami structure. Mixing the fresh and saline water at the end of draining guide could be avoided by folding the edge of the film. This simple structure and process may enable mass production and reduce the production cost.
#### **2.2. Design of the spacer layer**
Flexible wick/condensing layer is vulnerable for deformation which causes mixing back the fresh and saline water. Therefore, specially designed spacers are required for this system. In Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination http://dx.doi.org/10.5772/intechopen.68365 251
**Figure 2.** The origami structure of wick/condenser layer. Each part of the structure would have role of the evaporation wick (A), condensing surface (B), saline water feeding trough (or pocket) (C), condensed fresh water guiding channel (D) and draining guide which separate the fresh water and waste water (F).
other hands, tight vapour sealing and thermal insulation are essential for assuring high productivity. In this system, sealing, spacers and inner frames altogether are constructed in one structure, the spacer layer. Its structural features are shown in **Figure 3** [14].
Vertically elongated spacers are connected to upper and lower parts of the frame. The spacer layer has following roles.
#### *Role 1: Internal framework for the MES*
**Figure 1** represents typical inner structure of MES. It may include upper saline water trough, vapour condensing plates, wicks where the saline water evaporates, condensed fresh water collecting guide trough. The structures are packed in sun light passing windows and housing. For higher productivity in MES, intervals between each evaporation wick and condensing surface should be minimized [9, 11]. However, if the distance is too short, the layers could touch each other, so that water droplet on the condensing surface can move to evaporation wick, which causes loss of distilled water. Similarly, if the saline water overflow into condensing surface, serious contamination of produced water could be occurred. Therefore, the wick and condensing surface should be securely isolated by air gap. Despite air gaps between the layers, there is always a chance of contact within layers because of impact, vibration, and gravitational deformation. To minimize such deformation, various approaches were made. Rigid metallic plates were commonly used for stable condensing layer. However, it resulted in increase of cost and system weight. Spacers were placed between the layers [4]. However, water droplet may flow back to wick along the spacer if too much spacers were applied, which may reduce productivity. The layers were placed vertically to avoid gravitational deformation [8–13]. However, it may increase discrepancy between solar incident angle to the system window, which reduces solar energy usage. Reflecting mirrors [12, 24] or tilted window [8, 9, 13, 27] could be placed in front of the vertical layers to maximize solar intensity. However, it
To solve this high-cost problem, new MES was designed and tested. To reduce production cost, alternative light-weight and low-cost material was used instead of such expensive materials. Structure of each component was also designed for easier mass production, which may reduce cost. In the next section, the materials and structure of the low-cost MES would be described.
In this work, flexible thin plastic film laminated to black fabric wick was used instead of stainless plate covered with fabric wick [14, 15]. The laminated film is obviously low-cost, light-weight alternative than metallic plates. It could be produced by roll-to-roll process, which is mass production favourable. The laminated film was folded like origami as shown in **Figure 2** [14].
Evaporation wick, condensation surface, upper saline water trough, condensed water guiding channel and concentrated saline water draining guide could be included in this single origami structure. Mixing the fresh and saline water at the end of draining guide could be avoided by folding the edge of the film. This simple structure and process may enable mass production
Flexible wick/condensing layer is vulnerable for deformation which causes mixing back the fresh and saline water. Therefore, specially designed spacers are required for this system. In
may increase the system cost.
250 Desalination
and reduce the production cost.
**2.2. Design of the spacer layer**
**2. Structure and material of the low-cost MES**
**2.1. Design of the wick/condenser laminated film layer**
It has a role of internal framework structure standing for each layer of evaporation wick/condensation layer film. The spacer layers and wick/condensing films are alternately stacked as shown in **Figure 4**. Though the film is too flexible to be stand alone, the film-spacer combined layer is stiff enough to be stand. The stiffness increases by stacking multiple layers.
#### *Role 2: Fixture for the vertical spacers*
The frame holds each vertical spacer in right position and orientation. To avoid a layer's film touching the next layer film, it is important that each spacer on a layer should be exactly superposed on top of the next layer's spacer.
**Figure 3.** Structure of the spacer layers. The vertical spacers are connected to the perimeter frame. Small pictures in circles represent cross section of the spacers (332) which touches the condensing surface (310) laminated to the wick (320).
Maintaining regular spacer orientation is also necessary. As shown in small circles in **Figure 3**, the cross sections of the spacers were designed to have sharp angles (less than 90°) towards the contacting plane of the vapour condensing surface. Though most of the condensed water droplets drip down straight on the condensing film, some droplet could stagger out while confronting minor defects on the film surface. Staggering motion may increase chance of the droplet hit the spacer. Once touching the spacer, the droplet would be trapped and flow along the sharpangled crevice by surface tensional force. As long as it was trapped in the crevice, it would not flow back into the saline water running wick. Therefore, it is important that the spacer to be oriented to maintain the crevice located on condensation surface.
*Role 3: Gasket for vapor and heat*
checkerboard).
process and reduces production cost.
*Role 4: Guide of the condensed fresh water*
produced.
The frame also has a role of gasket and heat insulator which minimize heat and vapour leakage from the system. Since it is made out of soft material, foam board, leakage could be avoided by simply pressing the stacks without using additional sealant. It simplifies assembly
**Figure 4.** Alternately stacked wick (coloured black)/condenser (coloured grey) film and the spacer layer (coloured as
Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination
http://dx.doi.org/10.5772/intechopen.68365
253
Bottom part of the frame (**Figure 3**) is slightly tilted to the upper part of the frame (**Figure 3**). The upper part of the frame has role of a cross beam, holding saline water containing trough (or pocket) in **Figure 2**. Bottom part of the frame has role of condensed water dripping guide. There are multiple grooves on the bottom frame contacting the condensing film. These grooves are for guiding the collected water to be easily flow into the condensed water-guiding channel. Without grove, condensed water could overflow towards the wick, if too much fresh water
For lab prototype, 5-mm thick polystyrene foam board was hand carved with a knife to make the spacing layer, as shown in **Figure 5**. While carving, edges of the spacers were shaped to be 60° angle.
In case of mass production, the foam board could be cut by die press machine (commonly called as Thompson die cutter, a press with pre-shaped knife for cutting soft sheet). While pressing the knife into the foam board, edges of the cutting perimeter are automatically slightly collapsed as round shape, which makes sharp angle at the contact point between the spacer and condensing surface.
Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination http://dx.doi.org/10.5772/intechopen.68365 253
**Figure 4.** Alternately stacked wick (coloured black)/condenser (coloured grey) film and the spacer layer (coloured as checkerboard).
#### *Role 3: Gasket for vapor and heat*
Maintaining regular spacer orientation is also necessary. As shown in small circles in **Figure 3**, the cross sections of the spacers were designed to have sharp angles (less than 90°) towards the contacting plane of the vapour condensing surface. Though most of the condensed water droplets drip down straight on the condensing film, some droplet could stagger out while confronting minor defects on the film surface. Staggering motion may increase chance of the droplet hit the spacer. Once touching the spacer, the droplet would be trapped and flow along the sharpangled crevice by surface tensional force. As long as it was trapped in the crevice, it would not flow back into the saline water running wick. Therefore, it is important that the spacer to be
**Figure 3.** Structure of the spacer layers. The vertical spacers are connected to the perimeter frame. Small pictures in circles represent cross section of the spacers (332) which touches the condensing surface (310) laminated to the wick
For lab prototype, 5-mm thick polystyrene foam board was hand carved with a knife to make the spacing layer, as shown in **Figure 5**. While carving, edges of the spacers were shaped to
In case of mass production, the foam board could be cut by die press machine (commonly called as Thompson die cutter, a press with pre-shaped knife for cutting soft sheet). While pressing the knife into the foam board, edges of the cutting perimeter are automatically slightly collapsed as round shape, which makes sharp angle at the contact point between the
oriented to maintain the crevice located on condensation surface.
be 60° angle.
(320).
252 Desalination
spacer and condensing surface.
The frame also has a role of gasket and heat insulator which minimize heat and vapour leakage from the system. Since it is made out of soft material, foam board, leakage could be avoided by simply pressing the stacks without using additional sealant. It simplifies assembly process and reduces production cost.
#### *Role 4: Guide of the condensed fresh water*
Bottom part of the frame (**Figure 3**) is slightly tilted to the upper part of the frame (**Figure 3**). The upper part of the frame has role of a cross beam, holding saline water containing trough (or pocket) in **Figure 2**. Bottom part of the frame has role of condensed water dripping guide. There are multiple grooves on the bottom frame contacting the condensing film. These grooves are for guiding the collected water to be easily flow into the condensed water-guiding channel. Without grove, condensed water could overflow towards the wick, if too much fresh water produced.
**Figure 5.** Hand carved spacers.
#### **2.3. Design of the saline water distributer to each wick**
Since the evaporation process on each layer repeatedly recycles the latent heat dissipated from the front layer, energy usage of each layer is different. Generally, the layer closely facing sun has more energy for distillation. Therefore, distribution of saline water in proper amount to each layer is critical for high productivity. The Tanaka group did intensive study on how to optimize water distribution, both theoretically and experimentally [8, 11]. They used multiple capillaries with different length to control amount of saline water feed to each layer. Longer capillary reduces flow rate, while short one leads higher flow rate. This is a very useful method, but further simple structure was used in this chapter. A piece of cotton fabric was vertically cut with different width as shown in **Figure 6**.
**Figure 6.** Structure of the low-cost water feeding distributer to each wick.
(up).
**Figure 7.** Cross section of the triple layer air-gap polycarbonate window (bottom) and extended polystyrene foam board
Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination
http://dx.doi.org/10.5772/intechopen.68365
255
One end of the piece was immersed in saline water feeding container, and other shredded ends were inserted on the saline water pockets (**Figure 2**) of each layer. Saline water flows along the cotton fabric due to capillary force. Since the flow rate may proportional to the width of the fabric, allocated ratio of saline water to each layer could be simply controlled by controlling shred width of the fabric.
#### **2.4. Durability and maintenance of the system**
Heat loss should be minimized for higher productivity in MES. Especially for the current structure, it is important to insulate the front and back sides of the layer stacks, since the sides of stacks are partially insulated already by the frame of spacer layer. Thick (comparable to 10 cm) extended polystyrene foam plate or commercially named as "Isopink" could be placed on the back side of the layer stack. It is commonly used as a construction material for building insulation. For the front window, air-gap window, such as 18 mm thickness, triple layer air-gap polycarbonate could be placed (**Figure 7**). This window material, commercially called Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination http://dx.doi.org/10.5772/intechopen.68365 255
**Figure 6.** Structure of the low-cost water feeding distributer to each wick.
**2.3. Design of the saline water distributer to each wick**
was vertically cut with different width as shown in **Figure 6**.
controlling shred width of the fabric.
**Figure 5.** Hand carved spacers.
254 Desalination
**2.4. Durability and maintenance of the system**
Since the evaporation process on each layer repeatedly recycles the latent heat dissipated from the front layer, energy usage of each layer is different. Generally, the layer closely facing sun has more energy for distillation. Therefore, distribution of saline water in proper amount to each layer is critical for high productivity. The Tanaka group did intensive study on how to optimize water distribution, both theoretically and experimentally [8, 11]. They used multiple capillaries with different length to control amount of saline water feed to each layer. Longer capillary reduces flow rate, while short one leads higher flow rate. This is a very useful method, but further simple structure was used in this chapter. A piece of cotton fabric
One end of the piece was immersed in saline water feeding container, and other shredded ends were inserted on the saline water pockets (**Figure 2**) of each layer. Saline water flows along the cotton fabric due to capillary force. Since the flow rate may proportional to the width of the fabric, allocated ratio of saline water to each layer could be simply controlled by
Heat loss should be minimized for higher productivity in MES. Especially for the current structure, it is important to insulate the front and back sides of the layer stacks, since the sides of stacks are partially insulated already by the frame of spacer layer. Thick (comparable to 10 cm) extended polystyrene foam plate or commercially named as "Isopink" could be placed on the back side of the layer stack. It is commonly used as a construction material for building insulation. For the front window, air-gap window, such as 18 mm thickness, triple layer air-gap polycarbonate could be placed (**Figure 7**). This window material, commercially called
**Figure 7.** Cross section of the triple layer air-gap polycarbonate window (bottom) and extended polystyrene foam board (up).
Lexan, is frequently used for greenhouse construction. At the outdoor test, described in the next section, no significant degradation was observed, after leaving the prototype exposed to outdoor climate for more than half year. Though no further long-term stability test was made with this system yet, I expect certain level of environmental durability, since both Lexan and Isopink are already market proven as construction material.
Commercial polycarbonate windows are well known for toughness and UV resistance. Therefore, it might be stable for long period. If polycarbonate window is not available, stacks of glass plated with air gap, which has even higher UV resistance, could be used. However, extended polystyrene foam board is somewhat vulnerable to UV light. Though core parts of the system are protected from direct UV light, external case could be damaged for long period. For such case, the housing could be covered with other UV-blocking materials, such as paint or clay. The wicks are protected by UV-resistant window. However, long-term exposure under strong sunlight with continuous water dripping may lead black colour fading out. For that case, the front-most layer could be replaced with the inner layer periodically, since the inner layers are more protected from the UV light, and all the layers are designed to be exactly of same shape. Assuming that the system has 10 layers, and the front-most layer should be replaced after every 2 years, one system can be used for 20 years [23]. For this purpose, the system should have simple structure for assemble and disassemble. Window, stack of the layers and insulating back plates could be assembled by using common string or tie through holes, premade on each component. Unskilled users would be able to do the maintenance without special tool.
#### **2.5. Installation of the system**
Earlier MES models were commonly positioned to receive maximum solar energy during the year. With this structure, the evaporation/condensation layer stack was naturally tilted from the gravitational field. However, gravitational deformation of layers could be caused at the MES. It may cause contact of fresh water droplet to the wick or failure of the system.
**Figure 8.** MES system installed tilted 45°.
**Figure 9.** One square metre effective area MES prototype with 10 layers, whose weight is 15 kg.
Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination
http://dx.doi.org/10.5772/intechopen.68365
257
As the structure proposed in this chapter, such failure would be minimized because of the spacer layers. Therefore, the system could be placed towards sun side with tilted angle (**Figure 8**).
Following the structure described in this section, dry weight of the 1 m<sup>2</sup> system with 10 layers may not exceed 20 kg, since it has only light-weight material. (Actually, I got 15 kg mockup system, with the window, the 10-layer stack and insulator layer, as shown in **Figure 9**.) Therefore, the user could easily carry and install in DIY manner.
#### **2.6. Expected cost and further cost down**
Production cost for the described MES system is estimated in **Table 2** [15]. Material cost could be much lower, if we have massive economy of scale. Further cost down could be made, if lower cost alternative materials are available nearby. For example, multi-layer (air gap) polycarbonate window is relatively expensive material. It could be replaced with plane window and transparent air-bubble wrap which are inexpensive packing materials for wrapping Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination http://dx.doi.org/10.5772/intechopen.68365 257
**Figure 8.** MES system installed tilted 45°.
Lexan, is frequently used for greenhouse construction. At the outdoor test, described in the next section, no significant degradation was observed, after leaving the prototype exposed to outdoor climate for more than half year. Though no further long-term stability test was made with this system yet, I expect certain level of environmental durability, since both Lexan and
Commercial polycarbonate windows are well known for toughness and UV resistance. Therefore, it might be stable for long period. If polycarbonate window is not available, stacks of glass plated with air gap, which has even higher UV resistance, could be used. However, extended polystyrene foam board is somewhat vulnerable to UV light. Though core parts of the system are protected from direct UV light, external case could be damaged for long period. For such case, the housing could be covered with other UV-blocking materials, such as paint or clay. The wicks are protected by UV-resistant window. However, long-term exposure under strong sunlight with continuous water dripping may lead black colour fading out. For that case, the front-most layer could be replaced with the inner layer periodically, since the inner layers are more protected from the UV light, and all the layers are designed to be exactly of same shape. Assuming that the system has 10 layers, and the front-most layer should be replaced after every 2 years, one system can be used for 20 years [23]. For this purpose, the system should have simple structure for assemble and disassemble. Window, stack of the layers and insulating back plates could be assembled by using common string or tie through holes, premade on each component. Unskilled users would be able to do the
Earlier MES models were commonly positioned to receive maximum solar energy during the year. With this structure, the evaporation/condensation layer stack was naturally tilted from the gravitational field. However, gravitational deformation of layers could be caused at the
As the structure proposed in this chapter, such failure would be minimized because of the spacer layers. Therefore, the system could be placed towards sun side with tilted angle
may not exceed 20 kg, since it has only light-weight material. (Actually, I got 15 kg mockup system, with the window, the 10-layer stack and insulator layer, as shown in **Figure 9**.)
Production cost for the described MES system is estimated in **Table 2** [15]. Material cost could be much lower, if we have massive economy of scale. Further cost down could be made, if lower cost alternative materials are available nearby. For example, multi-layer (air gap) polycarbonate window is relatively expensive material. It could be replaced with plane window and transparent air-bubble wrap which are inexpensive packing materials for wrapping
system with 10 layers
MES. It may cause contact of fresh water droplet to the wick or failure of the system.
Following the structure described in this section, dry weight of the 1 m<sup>2</sup>
Therefore, the user could easily carry and install in DIY manner.
**2.6. Expected cost and further cost down**
Isopink are already market proven as construction material.
maintenance without special tool.
**2.5. Installation of the system**
(**Figure 8**).
256 Desalination
**Figure 9.** One square metre effective area MES prototype with 10 layers, whose weight is 15 kg.
(and also, save the foam board material), if made with hand one by one. However, for mass
Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination
http://dx.doi.org/10.5772/intechopen.68365
259
The spacer layers were alternately sandwiched with the evaporation/condensation film, while the condensation surface facing the sharp edge of the spacer. Ten identical spacer-film layers were made and stacked. The stack was sandwiched between 1.8-cm triple layer air-gap polycarbonate (Lexan) window and 10 cm polystyrene foam board, by cable tie or box tape. Box tape is useful for short-term usage but may not be a good choice for long-term test, since it would be deteriorated by UV. Common ropes are more recommended. A piece of shred cotton fabric was used for the capillary water distributer. Gutter was placed to collect all the
ers. It was placed in relatively non-shaded place in Seoul, Korea (North 37° 34′, East 126° 58′),
Ambient and inner temperature was measured by thermocouples. For inner temperature measurement, position between first and second layers as well as position backside of the last layer was chosen. Two water collecting beakers were placed on top of load cells to measure weight of fresh and concentrated water drained. A pyranometer was placed as same tilting angle (45°) as the prototype, for measuring solar irradiation. All the data were automatically
**Figures 13**–**15** show the result on a sunny day (On 7 July 2014, the daily accumulated solar
day or 0.46 kg/MJ. On a partially cloudy day (30 July 2014, the daily accumulated solar
), it was reduced to 5.7 kg/L per m<sup>2</sup>
active area (27 cm width and 81 cm height) with 10 multi-effect lay-
). It shows that it can produce fresh water about 9 kg/L
day or 0.41 kg/MJ
draining fresh water and concentrated waste water in separate beakers (**Figure 11**).
45° tilted from the vertical position, facing south (**Figure 12**).
The prototype has 0.219 m<sup>2</sup>
collected by computer every minute.
intensity was measured as 19.5 MJ/m<sup>2</sup>
intensity was measured as 13.7 MJ/ m<sup>2</sup>
**Figure 11.** Collecting fresh water from each layer by a gutter.
per m<sup>2</sup>
(**Figures 16**–**18**).
production, cutting out from one large sheet by die press would be easier.
**Table 2.** Estimated production cost of the MES system.
parcels. The material and design could be modified reflecting the best condition where the system would be produced and used.
## **3. Lab prototype and its outdoor test result**
A MES prototype was handmade using the structure described in previous section for testing its practical feasibility.
Black fabric laminated with plastic film was cut and folded to make the evaporation/condensation layer. Five-millimetre thick polystyrene foam board was used to make the spacer layer. It could be either carved out from one large sheet or assembled from small parts of the frames and thin spacers (**Figure 10**). It may be easier to make the layer by assembling small parts
**Figure 10.** Hand carved and assembled spacer layer (A) and die pressed spacer layer (B).
(and also, save the foam board material), if made with hand one by one. However, for mass production, cutting out from one large sheet by die press would be easier.
**Item Expected cost Comment**
Housing 50 USD Window and insulator housing Inner parts 180 USD Stacks with 10 layers of spacers,
Flow controller 30 USD Feeding rate controller, hoses and
parcels. The material and design could be modified reflecting the best condition where the
A MES prototype was handmade using the structure described in previous section for testing
Black fabric laminated with plastic film was cut and folded to make the evaporation/condensation layer. Five-millimetre thick polystyrene foam board was used to make the spacer layer. It could be either carved out from one large sheet or assembled from small parts of the frames and thin spacers (**Figure 10**). It may be easier to make the layer by assembling small parts
Processing cost 40 USD Labour and instrumental cost
Total 300 USD Model for 1 m<sup>2</sup>
**Table 2.** Estimated production cost of the MES system.
**3. Lab prototype and its outdoor test result**
**Figure 10.** Hand carved and assembled spacer layer (A) and die pressed spacer layer (B).
system would be produced and used.
its practical feasibility.
258 Desalination
wick/condensing film
effective area, which
supplies about 10 kg/day fresh water
container
The spacer layers were alternately sandwiched with the evaporation/condensation film, while the condensation surface facing the sharp edge of the spacer. Ten identical spacer-film layers were made and stacked. The stack was sandwiched between 1.8-cm triple layer air-gap polycarbonate (Lexan) window and 10 cm polystyrene foam board, by cable tie or box tape. Box tape is useful for short-term usage but may not be a good choice for long-term test, since it would be deteriorated by UV. Common ropes are more recommended. A piece of shred cotton fabric was used for the capillary water distributer. Gutter was placed to collect all the draining fresh water and concentrated waste water in separate beakers (**Figure 11**).
The prototype has 0.219 m<sup>2</sup> active area (27 cm width and 81 cm height) with 10 multi-effect layers. It was placed in relatively non-shaded place in Seoul, Korea (North 37° 34′, East 126° 58′), 45° tilted from the vertical position, facing south (**Figure 12**).
Ambient and inner temperature was measured by thermocouples. For inner temperature measurement, position between first and second layers as well as position backside of the last layer was chosen. Two water collecting beakers were placed on top of load cells to measure weight of fresh and concentrated water drained. A pyranometer was placed as same tilting angle (45°) as the prototype, for measuring solar irradiation. All the data were automatically collected by computer every minute.
**Figures 13**–**15** show the result on a sunny day (On 7 July 2014, the daily accumulated solar intensity was measured as 19.5 MJ/m<sup>2</sup> ). It shows that it can produce fresh water about 9 kg/L per m<sup>2</sup> day or 0.46 kg/MJ. On a partially cloudy day (30 July 2014, the daily accumulated solar intensity was measured as 13.7 MJ/ m<sup>2</sup> ), it was reduced to 5.7 kg/L per m<sup>2</sup> day or 0.41 kg/MJ (**Figures 16**–**18**).
**Figure 11.** Collecting fresh water from each layer by a gutter.
**Figure 14.** Solar irradiation measured at same day and location in **Figure 13**.
**Figure 15.** Temperature of ambient (blue, the lower curve at the range between 12:00 to 16:00 PM), backside of the last layer (red, the middle curve at the range between 12:00 to 16:00 PM) and backside of the first front layer (black, the upper
Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination
http://dx.doi.org/10.5772/intechopen.68365
261
curve at the range between 12:00 to 16:00 PM) measured at same day and location in **Figure 13**.
**Figure 12.** Experimental setup of the prototype MES outdoor test.
**Figure 13.** Flow rate (normalized to unit area, 1m2) of the fresh (red, the lower curve at the range after 15:00 PM) and wasted (black, the upper curve at the range after 15:00 PM) water verses time by the MES, measured on a sunny day, July 7th, at Seoul Korea.
Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination http://dx.doi.org/10.5772/intechopen.68365 261
**Figure 14.** Solar irradiation measured at same day and location in **Figure 13**.
**Figure 12.** Experimental setup of the prototype MES outdoor test.
July 7th, at Seoul Korea.
260 Desalination
**Figure 13.** Flow rate (normalized to unit area, 1m2) of the fresh (red, the lower curve at the range after 15:00 PM) and wasted (black, the upper curve at the range after 15:00 PM) water verses time by the MES, measured on a sunny day,
**Figure 15.** Temperature of ambient (blue, the lower curve at the range between 12:00 to 16:00 PM), backside of the last layer (red, the middle curve at the range between 12:00 to 16:00 PM) and backside of the first front layer (black, the upper curve at the range between 12:00 to 16:00 PM) measured at same day and location in **Figure 13**.
**Figure 16.** Flow rate (normalized to unit area, 1m2) of the fresh (red, the lower curve at the range after 16:00 PM) and wasted (black, the upper curve at the range after 15:00 PM)water verses time by the MES, measured on a partially cloudy day, July 30th, at Seoul, Korea.
This is somewhat smaller number than previous result reported from other group, who showed
**Figure 18.** Temperature of ambient (blue, the lower curve at the range between 15:00 to 18:00 PM), backside of the last layer (red, the middle curve at the range between 15:00 to 18:00 PM) and backside of the first front layer (black, the upper
curve at the range between 15:00 to 18:00 PM) measured at same day and location in **Figure 16**.
mized yet, it would be a good starting point, proving the practical feasibility. There are still many ways that remain to improve productivity. One example is to optimize and control saline water feeding rate. The Tanaka group showed that proper saline water feeding rate is very critical for high productivity [8, 12]. Too much feeding may cause bad productivity, because the saline water would be drained before it gains enough energy for evaporation. Too low feeding rate may cause dry out of the wick, so the system itself could be deteriorated by salt crystal. It was recommended that ratio between fresh water and concentrated water should be around 1:1. As shown in both **Figure 13** and **Figure 16**, this ratio was not controlled well enough yet, especially from around 15:00 O'clock. Better control on water flow may increase productivity. Further optimization on the feeding water distribution towards each layer would also lead higher productivity. Though shred fabric distributer, described in the previous section, was applied on this
prototype, there are still room for further optimization of the shred width for each layer.
To avoid salt contamination, saline water in the wick should not flow into condensation layer. Fortunately, in most cases, it is automatically avoided since capillary force in the wick captures
**4. Miscellaneous comments on the experiment**
**4.1. Checking salt contamination of the condensed water**
day was obtained [4]. However, considering that the system is not well opti-
Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination
http://dx.doi.org/10.5772/intechopen.68365
263
that 10–20 kg/m<sup>2</sup>
**Figure 17.** Solar irradiation measured at same day and location in **Figures 16**.
Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination http://dx.doi.org/10.5772/intechopen.68365 263
**Figure 18.** Temperature of ambient (blue, the lower curve at the range between 15:00 to 18:00 PM), backside of the last layer (red, the middle curve at the range between 15:00 to 18:00 PM) and backside of the first front layer (black, the upper curve at the range between 15:00 to 18:00 PM) measured at same day and location in **Figure 16**.
This is somewhat smaller number than previous result reported from other group, who showed that 10–20 kg/m<sup>2</sup> day was obtained [4]. However, considering that the system is not well optimized yet, it would be a good starting point, proving the practical feasibility. There are still many ways that remain to improve productivity. One example is to optimize and control saline water feeding rate. The Tanaka group showed that proper saline water feeding rate is very critical for high productivity [8, 12]. Too much feeding may cause bad productivity, because the saline water would be drained before it gains enough energy for evaporation. Too low feeding rate may cause dry out of the wick, so the system itself could be deteriorated by salt crystal. It was recommended that ratio between fresh water and concentrated water should be around 1:1. As shown in both **Figure 13** and **Figure 16**, this ratio was not controlled well enough yet, especially from around 15:00 O'clock. Better control on water flow may increase productivity. Further optimization on the feeding water distribution towards each layer would also lead higher productivity. Though shred fabric distributer, described in the previous section, was applied on this prototype, there are still room for further optimization of the shred width for each layer.
#### **4. Miscellaneous comments on the experiment**
**Figure 16.** Flow rate (normalized to unit area, 1m2) of the fresh (red, the lower curve at the range after 16:00 PM) and wasted (black, the upper curve at the range after 15:00 PM)water verses time by the MES, measured on a partially cloudy
**Figure 17.** Solar irradiation measured at same day and location in **Figures 16**.
day, July 30th, at Seoul, Korea.
262 Desalination
#### **4.1. Checking salt contamination of the condensed water**
To avoid salt contamination, saline water in the wick should not flow into condensation layer. Fortunately, in most cases, it is automatically avoided since capillary force in the wick captures the saline water. However, if too much saline water flows into the system abruptly, which could not be kept within the fabric, of course it may overflow into the condensation layer.
construction, which would be available on most part of the world. All the materials, processes and structures could be modified by the developers reflecting the industrial condition of each one. Because of this industrial flexibility, MES could be useful self-producible desalination
Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination
http://dx.doi.org/10.5772/intechopen.68365
265
Low-cost multi-effect solar still could be mostly useful for supplying fresh water for individuals or family who cannot access to public water work. Residents near salty or contaminated water source, remote island or seashore could use this tool. However, it could be useful for public or national level also, not only limited to the individual level. It could be a part of social infrastructure: for example, it can replace highway divider or fences in desert or bridge over sea. Fresh water supplied by MES could be used to cultivate plants along the road. It can be used for fence/wall around buildings or districts to provide fresh water for planting or for citizens. MES modules would also be made in form of tile or curtain walls of buildings. Similar to that photovoltaics module could be part of external surface of buildings (building integrated photovoltaics: BIPV), MES could be part of building (building integrated desalination: BIDSAL), which can produce fresh water for the residents [15]. MES could also be useful as public stockpile against natural disaster or terrorism, in case of existing water line
The MES system introduced in this chapter has both aspects of individuality and publicity. The system could be used as personal water supplier. It could be made, operated and maintained personally without assistance of high-tech industry. However, on the other hands, the system could be more easily produced with lower cost and widely implemented, if mass production industry supports. Once mass produced in low cost, it could be a practical solution to
Not only the production and application, research and development of the MES has both individual and public aspects. Initially, I started to develop the low-cost MES as a private project. Because of relatively low material cost and simple measuring instrument, it was an executable project for even an individual. Therefore, I expect that developing MES is relatively easily accessible subject to many other researchers in the world, regardless of their financial status or infrastructure. Every future users, producers or developers of MES in the world may be in different condition: climate; social necessity; industrial level and raw material cost. Therefore, they may need to develop their own optimized system, reflecting their own specific status. Meanwhile, they can share the new findings with others, for worldwide collaboration.
The structure and material introduced in this chapter are just an initial example of low-cost MES. There are many steps that remain before implementation. Further optimization and structural improvement should be made. Practical size prototype should be developed. Longterm stability should be proven by outdoor test under actual climate where the product to be used. Mass production process should be established, reflecting the industrial condition
the potable water deficiency problem for large part of the world.
tool for many countries.
malfunctioning.
**6. Suggestion for future work**
**5.1. Potential applications of the low-cost MES**
To check any trace of salt contamination, simple method could be used [15]. Diluted basic solution, such as aqueous sodium hydroxide solution, could be mixed with the saline water, which would be desalinated by the system. Since even small trace of hydroxide ion increases pH of the solution, it can be easily noticed if the distillates contaminated by salt. Of course, evaporable pH altering material, such as hydrochloric acid, should not be used for such purpose, since it could be evaporated at the wick to change pH of the condensed water.
#### **4.2. Operation of the MES in cold weather**
With good insulation and tight seal, inner temperature of the MES was observed to be more than 70–80°C under sun, even below freezing point ambient temperature. This is high enough temperature for water evaporation. However, it is generally not recommended to use it under freezing climate. Though the temperature inside of MES is high under sun, it may freeze during night, which may damage the system if repeated for a longer period. In addition to that, the drained water could be frozen at the outlet of MES at cold weather. At the winter, I found icicles underneath the prototype, which mixed back the purified and concentrated water.
## **5. Mass production in less-developed country and potential application of the MES**
Most of the other competing, small-scale solar desalination instrument requires relatively high-end technology and expensive facility to produce each component, such as nano-porous membrane for reverse osmosis, photovoltaic modules, vacuum tubes etc. It makes less-developed societies difficult to self-supply the tool by themselves. However, MES prototype introduced in this chapter was handmade with elementary tools and materials, such as hand knife, polystyrene foam board and fabrics laminated on plastic film. In other word, it can be relatively easily prepared and studied with minimum budget. In addition to that, it is also easy to be mass produced, not only in advanced countries but also in less-developed countries. Therefore, MES supplying chain could be easily made at the actual countries, where the lowcost small-scale desalination is necessary, regardless of its industrial level.
Wick/condensation layer could be mass produced by roll-to-roll lamination of fabric on plastic film. Lamination process itself is not a high-tech engineering. It would be affordable to most of the underdeveloped countries. However, if it is still too difficult to be manufactured by local industry, laminated material could be produced in more-developed region and then transported to the local producers who do not have expensive machineries. Since the film is light weight and small volume, it is easy to be transported. By the local maker, the sheet material could be cut, folded and assembled with relatively low-labour cost.
Similar approach could be made to the spacer layers. It could be mass produced with foam board by die-cutting press. Any other type of cutters could be used, if the press is not available. Other raw materials, such as foam insulators and windows, are common material for construction, which would be available on most part of the world. All the materials, processes and structures could be modified by the developers reflecting the industrial condition of each one. Because of this industrial flexibility, MES could be useful self-producible desalination tool for many countries.
#### **5.1. Potential applications of the low-cost MES**
the saline water. However, if too much saline water flows into the system abruptly, which could not be kept within the fabric, of course it may overflow into the condensation layer.
To check any trace of salt contamination, simple method could be used [15]. Diluted basic solution, such as aqueous sodium hydroxide solution, could be mixed with the saline water, which would be desalinated by the system. Since even small trace of hydroxide ion increases pH of the solution, it can be easily noticed if the distillates contaminated by salt. Of course, evaporable pH altering material, such as hydrochloric acid, should not be used for such pur-
With good insulation and tight seal, inner temperature of the MES was observed to be more than 70–80°C under sun, even below freezing point ambient temperature. This is high enough temperature for water evaporation. However, it is generally not recommended to use it under freezing climate. Though the temperature inside of MES is high under sun, it may freeze during night, which may damage the system if repeated for a longer period. In addition to that, the drained water could be frozen at the outlet of MES at cold weather. At the winter, I found icicles underneath the prototype, which mixed back the purified and concentrated water.
**5. Mass production in less-developed country and potential application**
Most of the other competing, small-scale solar desalination instrument requires relatively high-end technology and expensive facility to produce each component, such as nano-porous membrane for reverse osmosis, photovoltaic modules, vacuum tubes etc. It makes less-developed societies difficult to self-supply the tool by themselves. However, MES prototype introduced in this chapter was handmade with elementary tools and materials, such as hand knife, polystyrene foam board and fabrics laminated on plastic film. In other word, it can be relatively easily prepared and studied with minimum budget. In addition to that, it is also easy to be mass produced, not only in advanced countries but also in less-developed countries. Therefore, MES supplying chain could be easily made at the actual countries, where the low-
Wick/condensation layer could be mass produced by roll-to-roll lamination of fabric on plastic film. Lamination process itself is not a high-tech engineering. It would be affordable to most of the underdeveloped countries. However, if it is still too difficult to be manufactured by local industry, laminated material could be produced in more-developed region and then transported to the local producers who do not have expensive machineries. Since the film is light weight and small volume, it is easy to be transported. By the local maker, the sheet mate-
Similar approach could be made to the spacer layers. It could be mass produced with foam board by die-cutting press. Any other type of cutters could be used, if the press is not available. Other raw materials, such as foam insulators and windows, are common material for
cost small-scale desalination is necessary, regardless of its industrial level.
rial could be cut, folded and assembled with relatively low-labour cost.
pose, since it could be evaporated at the wick to change pH of the condensed water.
**4.2. Operation of the MES in cold weather**
**of the MES**
264 Desalination
Low-cost multi-effect solar still could be mostly useful for supplying fresh water for individuals or family who cannot access to public water work. Residents near salty or contaminated water source, remote island or seashore could use this tool. However, it could be useful for public or national level also, not only limited to the individual level. It could be a part of social infrastructure: for example, it can replace highway divider or fences in desert or bridge over sea. Fresh water supplied by MES could be used to cultivate plants along the road. It can be used for fence/wall around buildings or districts to provide fresh water for planting or for citizens. MES modules would also be made in form of tile or curtain walls of buildings. Similar to that photovoltaics module could be part of external surface of buildings (building integrated photovoltaics: BIPV), MES could be part of building (building integrated desalination: BIDSAL), which can produce fresh water for the residents [15]. MES could also be useful as public stockpile against natural disaster or terrorism, in case of existing water line malfunctioning.
## **6. Suggestion for future work**
The MES system introduced in this chapter has both aspects of individuality and publicity. The system could be used as personal water supplier. It could be made, operated and maintained personally without assistance of high-tech industry. However, on the other hands, the system could be more easily produced with lower cost and widely implemented, if mass production industry supports. Once mass produced in low cost, it could be a practical solution to the potable water deficiency problem for large part of the world.
Not only the production and application, research and development of the MES has both individual and public aspects. Initially, I started to develop the low-cost MES as a private project. Because of relatively low material cost and simple measuring instrument, it was an executable project for even an individual. Therefore, I expect that developing MES is relatively easily accessible subject to many other researchers in the world, regardless of their financial status or infrastructure. Every future users, producers or developers of MES in the world may be in different condition: climate; social necessity; industrial level and raw material cost. Therefore, they may need to develop their own optimized system, reflecting their own specific status. Meanwhile, they can share the new findings with others, for worldwide collaboration.
The structure and material introduced in this chapter are just an initial example of low-cost MES. There are many steps that remain before implementation. Further optimization and structural improvement should be made. Practical size prototype should be developed. Longterm stability should be proven by outdoor test under actual climate where the product to be used. Mass production process should be established, reflecting the industrial condition of the producers. Supply chain of raw materials should be established. Local or worldwide distribution of the product should be done with proper instruction of operation and maintenance to the final users. These tasks could be done with international collaboration, especially including the groups in the country where the system has to be implemented. Main purpose of writing this chapter is to suggest active collaboration all over the world. Collaboration between private, public or international supported groups would be helpful for MES implementation. I will also be very happy to be part of the collaboration. Furthermore, it would be helpful to mitigate potable water deficiency problem in the world. Beauty of solar still would be that it may quench thirst of anybody, whether he or she is a drinker or a developer.
[3] Defoe C Ginnings. Multiple effect solar still. US Patent 2445350A; 1948
Mechanical, Automobile and Production Engineering. 2012;**2**:166-173
sion-coupled solar still. Desalination. 2002;**150**:131-144
reflector. Desalination. 2007;**216**:139-146
still. Solar Energy. 1995;**54**:317-325
Sustainable Energy Reviews. 2013;**17**:248-259
B. 2010;**6**:607-613
633-644
[4] Biswas A, Ruby. Distillation of water by solar energy. VSRD International Journal of
Low-Cost Multi-Effect Solar Still: Alternative Appropriate Technology for Personal Desalination
http://dx.doi.org/10.5772/intechopen.68365
267
[5] Qiblawey HM, Banat F. Solar thermal desalination technologies. Desalination. 2008;**220**:
[6] Sivakumaran R, Judhesh P. Improvement techniques in performance and productivity of solar stills. International Journal of Engineering Science and Computing. 2016;**6**:2441-2450
[7] World Health Organization (WHO). How Much Water Is Needed in Emergencies [Internet]. [Updated: July 2013]. Available from: http://www.who.int/water\_sanitation\_health/publications/2011/WHO\_TN\_09\_How\_much\_water\_is\_needed.pdf?ua=1 http://www.who.int/ water\_sanitation\_health/emergencies/qa/emergencies\_qa5/en/ [Accessed: 2 January 2017]
[8] Tanaka H, Nosoko T, Nagata T. Experimental study of basin-type, multiple-effect, diffu-
[9] Tanaka H, Nakatake Y, Watanabe K. Parametric study on a vertical multiple-effect diffusion-type solar still coupled with a heat-pipe solar collector. Desalination. 2004;**171**:243-255
[10] Tanaka H, Nakatake Y, Tanaka M. Indoor experiments of the vertical multiple-effect diffusion-type solar still coupled with a heat-pipe solar collector. Desalination. 2005;**177**:291-302
[11] Tanaka H, Nakatake Y. Factors influencing the productivity of a multiple-effect diffusion-type solar still coupled with a flat plate reflector. Desalination. 2005;**186**:299-310 [12] Tanaka H, Nakatake Y. Improvement of the tilted wick solar still by using a flat plate
[13] Park C-D, Lim B-J, Tanaka H. Development of seawater distiller utilizing waste heat of portable electric generators. Transactions of the Korean Society of Mechanical Engineers
[14] Pak H. Evaporative saline-water desalination apparatus using solar energy. Patent appli-
[15] Pak H. Building integrated solar desalination (BIDSAL): Preliminary works with mul-
[16] Toyama S, Aragaki T, Salah HM, Murase K, Sando M. Simulation of a multieffect solar still and the static characteristics. Journal of Chemical Engineering of Japan. 1987;**20**:473-478
[17] Adhikari RS, Kumar A, Sootha GD. Simulation studies on a multi-stage stacked try solar
[18] Adhikari RS, Kumar A, Sodha GD. Thermal performance of a multi-effect diffusion solar
[19] Rajaseenivasan T, Kalidasa Murugavel K, Elango T, Samuel Hansen R. A review of different methods to enhance the productivity of the multi-effect solar still. Renewable and
cation number (S. Korea) KR 10-2015-0045121, (PCT) WO2016159625A1
tiple effect solar still. Energy Proceedia. 2014;**62**:639-646
still. International Journal of Energy Research. 1991;**15**:769-779
## **Acknowledgements**
I appreciate Dr. Park, Chang-Dae in Korea Institute of Machinery and Materials (KIMM), who first introduced MES to me. The subject introduced in this chapter was strongly motivated by his work. I also appreciate Mr. Jin, Byungjin in On-test who helped to equip the measurement system. I appreciate to Korea Water Forum Secretariat (KWFS), since the prize money from 2015 World Water Challenge at 7th World Water Forum was financially helpful on doing this project.
Since it was a privately supported project, I owe many things to my family. My parents-inlaw, Mr. Ahn, Jae Chul and Ms. Lee, Young Sook, kindly encouraged me to install the prototype on their rooftop. I was lucky enough to have sunshine test field at the middle of Seoul, where the non-shaded space is rare because of many tall skyscrapers. I developed the origami structure of wick/condensation film while playing origami with my 2-year-old son, Pak, Jongmin I sincerely hope he lives in better future where less people suffer thirst in the world.
## **Author details**
Pak Hunkyun
Address all correspondence to: hunkyunpak@gmail.com
Electronics and Telecommunications Research Institute, Daejeon, Republic of Korea
## **References**
[3] Defoe C Ginnings. Multiple effect solar still. US Patent 2445350A; 1948
of the producers. Supply chain of raw materials should be established. Local or worldwide distribution of the product should be done with proper instruction of operation and maintenance to the final users. These tasks could be done with international collaboration, especially including the groups in the country where the system has to be implemented. Main purpose of writing this chapter is to suggest active collaboration all over the world. Collaboration between private, public or international supported groups would be helpful for MES implementation. I will also be very happy to be part of the collaboration. Furthermore, it would be helpful to mitigate potable water deficiency problem in the world. Beauty of solar still would be that it may quench thirst of anybody, whether he or she is a drinker or a developer.
I appreciate Dr. Park, Chang-Dae in Korea Institute of Machinery and Materials (KIMM), who first introduced MES to me. The subject introduced in this chapter was strongly motivated by his work. I also appreciate Mr. Jin, Byungjin in On-test who helped to equip the measurement system. I appreciate to Korea Water Forum Secretariat (KWFS), since the prize money from 2015 World Water Challenge at 7th World Water Forum was financially helpful on doing this
Since it was a privately supported project, I owe many things to my family. My parents-inlaw, Mr. Ahn, Jae Chul and Ms. Lee, Young Sook, kindly encouraged me to install the prototype on their rooftop. I was lucky enough to have sunshine test field at the middle of Seoul, where the non-shaded space is rare because of many tall skyscrapers. I developed the origami structure of wick/condensation film while playing origami with my 2-year-old son, Pak, Jongmin I sincerely hope he lives in better future where less people suffer thirst in the world.
Electronics and Telecommunications Research Institute, Daejeon, Republic of Korea
[1] Ghaffour N, Missimer TM, Amy GL. Technical review and evaluation of economics of water desalination: Current and future challenges for better sustainability. Desalination.
[2] Talbert SG, et al., editors. Manual on Solar Distillation of Saline Water. Office of Saline Water, United States Department of the Interior; 1970. p. 263 (PDF file available from: http://www.appropedia.org/Solar\_distillation#cite\_note-20 [Accessed: 2 January 2017])
**Acknowledgements**
project.
266 Desalination
**Author details**
Address all correspondence to: hunkyunpak@gmail.com
Pak Hunkyun
**References**
2013;**309**:197-207
[20] Tanaka H, Nosoko T, Nagata T. A highly productive basin-type-multiple-effect coupled
[21] Estahbanati MRK, Feilizadeh M, Jafarpur K, Feilizadeh M, Rahimpour MR. Experimental investigation of a multi-effect active solar still: The effect of the number of stages. Applied
[22] Huang B-J, Chong T-L, Chang H-S, Wu P-H, Kao Y-C. Solar distillation system based on multiple-effect diffusion type still. Journal of Sustainable Development of Energy, Water
[23] Tiwari GN, Lawrence SA, Gupta SP. Analytical study of multi-effect solar still. Energy
[24] Goff PL, Goff JL, Jeday MR. Development of a rugged design of a high efficiency multi-
[25] Xiong J, Xie G, Zheng H. Experimental and numerical study on a new multi-effect solar still with enhanced condensation surface. Energy Conversion and Management.
[26] Gräter F, Dürrbeck M, Rheinländer J. Multi-effect still for hybrid solar/fossil desalination
[27] Park C-D, Lim B-J, Chung K-Y, Lee S-S, Kim Y-M. Experimental evaluation of hybrid
solar still. Desalination. 2000;**130**:279-293
and Environment System. 2014;**2**:41-50
Conversion and Management. 1989;**29**:259-263
stage solar still. Desalination. 1991;**82**:153-163
of sea and brackish water. Desalination. 2001;**138**:111-119
solar still using waste heat. Desalination. 2016;**379**:1-9
Energy. 2015;**137**:46-55
268 Desalination
2013;**73**:176-185
## *Edited by Taner Yonar*
Increasing population and environmental pollution are the main stress on freshwater sources. On the other hand, freshwater needs of human being increase dramatically every day. From agriculture to industry and from household to recreation, we need freshwater. In the near future, saltwater and brackish water bodies may be the main source of freshwater for our planet. Desalination phenomena are now being implemented with increasing interest. The book on desalination provides a valuable scientific contribution on freshwater production from saltwater sources. In this book, necessary theoretical knowledge and experimental results of different desalination processes are presented.
Photo by Supersmario / iStock
Desalination
Desalination | doab | 2025-04-07T04:13:03.558448 | 1-12-2023 17:13 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc124f2-9afe-4345-bf9e-9e0b37138f4f",
"url": "https://mts.intechopen.com/storage/books/5768/authors_book/authors_book.pdf",
"author": "",
"title": "Desalination",
"publisher": "IntechOpen",
"isbn": "9789535133643",
"section_idx": 2
} |
ffc27b79-9210-4e73-a820-6f98a7ec95e7 | # SHIFTING THE PARADIGMS FOR SUSTAINABLE WILDMEAT USE IN TROPICAL AND SUB-TROPICAL REGIONS
EDITED BY : Nathalie Van Vliet and Robert Nasi PUBLISHED IN : Frontiers in Ecology and Evolution
#### Frontiers eBook Copyright Statement
The copyright in the text of individual articles in this eBook is the property of their respective authors or their respective institutions or funders. The copyright in graphics and images within each article may be subject to copyright of other parties. In both cases this is subject to a license granted to Frontiers. The compilation of articles constituting this eBook is the property of Frontiers.
Each article within this eBook, and the eBook itself, are published under the most recent version of the Creative Commons CC-BY licence. The version current at the date of publication of this eBook is CC-BY 4.0. If the CC-BY licence is updated, the licence granted by Frontiers is automatically updated to the new version.
When exercising any right under the CC-BY licence, Frontiers must be attributed as the original publisher of the article or eBook, as applicable.
Authors have the responsibility of ensuring that any graphics or other materials which are the property of others may be included in the CC-BY licence, but this should be checked before relying on the CC-BY licence to reproduce those materials. Any copyright notices relating to those materials must be complied with.
Copyright and source acknowledgement notices may not be removed and must be displayed in any copy, derivative work or partial copy which includes the elements in question.
All copyright, and all rights therein, are protected by national and international copyright laws. The above represents a summary only. For further information please read Frontiers' Conditions for Website Use and Copyright Statement, and the applicable CC-BY licence.
ISSN 1664-8714 ISBN 978-2-88966-238-8 DOI 10.3389/978-2-88966-238-8
#### About Frontiers
Frontiers is more than just an open-access publisher of scholarly articles: it is a pioneering approach to the world of academia, radically improving the way scholarly research is managed. The grand vision of Frontiers is a world where all people have an equal opportunity to seek, share and generate knowledge. Frontiers provides immediate and permanent online open access to all its publications, but this alone is not enough to realize our grand goals.
#### Frontiers Journal Series
The Frontiers Journal Series is a multi-tier and interdisciplinary set of open-access, online journals, promising a paradigm shift from the current review, selection and dissemination processes in academic publishing. All Frontiers journals are driven by researchers for researchers; therefore, they constitute a service to the scholarly community. At the same time, the Frontiers Journal Series operates on a revolutionary invention, the tiered publishing system, initially addressing specific communities of scholars, and gradually climbing up to broader public understanding, thus serving the interests of the lay society, too.
### Dedication to Quality
Each Frontiers article is a landmark of the highest quality, thanks to genuinely collaborative interactions between authors and review editors, who include some of the world's best academicians. Research must be certified by peers before entering a stream of knowledge that may eventually reach the public - and shape society; therefore, Frontiers only applies the most rigorous and unbiased reviews.
Frontiers revolutionizes research publishing by freely delivering the most outstanding research, evaluated with no bias from both the academic and social point of view. By applying the most advanced information technologies, Frontiers is catapulting scholarly publishing into a new generation.
#### What are Frontiers Research Topics?
Frontiers Research Topics are very popular trademarks of the Frontiers Journals Series: they are collections of at least ten articles, all centered on a particular subject. With their unique mix of varied contributions from Original Research to Review Articles, Frontiers Research Topics unify the most influential researchers, the latest key findings and historical advances in a hot research area! Find out more on how to host your own Frontiers Research Topic or contribute to one as an author by contacting the Frontiers Editorial Office: researchtopics@frontiersin.org
# SHIFTING THE PARADIGMS FOR SUSTAINABLE WILDMEAT USE IN TROPICAL AND SUB-TROPICAL REGIONS
Topic Editors:
Nathalie Van Vliet, Center for International Forestry Research (CIFOR), Indonesia Robert Nasi, Center for International Forestry Research (CIFOR), Indonesia
In Tropical and sub-tropical Range States, wildmeat is an important source of nutrition and income, but current extraction levels of vulnerable taxa are considered unsustainable. As such, wildmeat use is often seen as problematic for wildlife conservation. From a development perspective, balancing the nutritional needs of people who depend on wildmeat with biodiversity conservation is the greatest challenge. But why can't wildmeat use be seen as an ally for conservation?
Most analysis of wildmeat use have framed the problem around a rather simplistic paradigm where wildmeat use is unsustainable and should therefore be reduced or stopped to ensure wildlife conservation. Indeed, until the early start of this century most research efforts have been rooted in the biological disciplines, focused on quantifying the magnitude of the trade and measuring its level of destruction on wildlife species and ecosystems. This most often led to the institution of prohibitive policies intended for the protection of the wild resources, such as separating people from wildlife, expanding tightly-managed protected area networks, blanket criminalization of wild meat hunting, and increasing enforcement and interdiction measures. More recently, based on the elucidation of the role of wild meat in human livelihoods, some practitioners defend the idea that consumptive uses of wildlife are the only way to save it in the long run.
Citation: Van Vliet, N., Nasi, R., eds. (2020). Shifting the Paradigms for Sustainable Wildmeat Use in Tropical and Sub-Tropical Regions. Lausanne: Frontiers Media SA. doi: 10.3389/978-2-88966-238-8
# Table of Contents
Pedro Mayor, Hani R. El Bizri, Thais Q. Morcatty, Kelly Moya, Samantha Solis and Richard E. Bodmer
Anders Henrik Sirén and Kalle Parvinen
*88 Frameworks Regulating Hunting for Meat in Tropical Countries Leave the Sector in the Limbo*
Nathalie van Vliet, André Pinassi Antunes, Pedro de Araujo Lima Constantino, Juanita Gómez, Dídac Santos-Fita and Eugenio Sartoretto
David S. Wilkie, Michelle Wieland and John R. Poulsen
*136 Impact of Low-Intensity Hunting on Game Species in and Around the Kanuku Mountains Protected Area, Guyana*
Matthew T. Hallett, Anouska A. Kinahan, Rayon McGregor, Thadaigh Baggallay, Timothy Babb, Howard Barnabus, Asaph Wilson, Fernando M. Li, Wesley W. Boone and Brittany A. Bankovich
# "Bushmeat Crisis" and "Cultural Imperialism" in Wildlife Management? Taking Value Orientations Into Account for a More Sustainable and Culturally Acceptable Wildmeat Sector
#### Nathalie van Vliet 1,2 \*
#### Edited by:
*Divya Karnad, Ashoka University, India*
#### Reviewed by:
*Pratheesh C. Mammen, Institute for Climate Change Studies, India Ambika Aiyadurai, Indian Institute of Technology Gandhinagar, India*
#### \*Correspondence:
*Nathalie van Vliet vanvlietnathalie@yahoo.com*
#### Specialty section:
*This article was submitted to Conservation, a section of the journal Frontiers in Ecology and Evolution*
> Received: *03 May 2018* Accepted: *09 July 2018* Published: *02 August 2018*
#### Citation:
*van Vliet N (2018) "Bushmeat Crisis" and "Cultural Imperialism" in Wildlife Management? Taking Value Orientations Into Account for a More Sustainable and Culturally Acceptable Wildmeat Sector. Front. Ecol. Evol. 6:112. doi: 10.3389/fevo.2018.00112* *<sup>1</sup> Center for International Forestry Research, Bogor, Indonesia, <sup>2</sup> Graduate School of Asian and African Area Studies, Kyoto University, Kyoto, Japan*
Keywords: bushmeat, wildlife value orientations, wildmeat sector, cultural imperialism, human dimensions
In tropical regions, two decades after the "Bushmeat Crisis" outcry, there is now a growing recognition of the failure of single solutions to the issue. Strict protectionist measures toward wildlife consumption through highly militarized law enforcement has proved to fail (Bennett, 2011; Wellsmith, 2011; Challender and MacMillan, 2014; Cooney et al., 2017). The development of alternative livelihoods, which was based on the hypothesis that hunting and consumption of wildmeat could be downsized if the reliance on wildlife as a source of food and income could be reduced, also evidenced several short comes (Wicander and Coad, 2015; Alves and van Vliet, 2018). More recent recommendations by the scientific community (Wilkie et al., 2016) and endorsed by the Convention on Biological Diversity now acknowledge the need for more comprehensive and context specific responses to prevent wildlife declines (CBD, 2017). While these recommendations clearly show progress in our understanding of wildlife management complexities, I argue that any approach to manage wildmeat use in tropical regions might continue to result inadequate, un-effective or un-acceptable without a mutualistic understanding of the complexity and nuance regarding the multiple connections that people maintain with wildlife and how these reflect the value orientations shared within the resource constituency. I use a humans' dimension approach to characterize human relationships with wildmeat in tropical forest areas, both in rural and urban/western contexts. Then, I analyze how the two opposed ends of the wildlife value orientations continuum are resulting in stigmas, which represent clear bottlenecks for sustainability in tropical regions. Finally, I call for a better understanding of the cultural constructions that shape beliefs, attitudes and behavior among the different beneficiaries of wildlife, taking into account local/international, rural/urban, traditional/western specificities. Indeed, considering that the mass of the funding available for wildlife conservation originates from foreign countries and is mostly executed through international institutions, claims of "cultural imperialism" may legitimately continue to arise if the complex and dynamic cultural dimensions of human-wildlife relations is not adequately analyzed and considered.
**4**
### THE COMPLEXITY AND NUANCES IN WILDLIFE VALUE ORIENTATIONS IN RURAL AND URBAN CONTEXTS
Human relationships with wildlife have existed since human kind (Alves and Albuquerque, 2018) and have shaped different value orientations toward wildlife depending on the social and cultural constructs, moral values, material realities and political dynamics characteristic of a given time, location and social group (Manfredo, 2008; Jacobs, 2009; Alves and Barboza, 2018). Different authors in human dimensions research have employed various terms to describe patterns of basic beliefs that give direction to values toward wildlife, but basically follow the "protection vs. use" (Vaske and Manfredo, 2011) or the "mutualism vs. domination" (Teel et al., 2007; Manfredo et al., 2009) continuum. Individuals with a utilitarian or domination value orientation believe wildlife should be managed for human benefit, whereas individuals with a protectionist or mutualism orientation view wildlife as part of an extended family, deserving rights and care (Manfredo et al., 2009). This bi-dimensional model, tested and proved for North American contexts, is not necessarily adapted to other cultural contexts and methodologies based on emotional prompts have been developed to identify context specific wildlife value orientations (Dayer et al., 2007).
In rural areas from tropical regions, despite changing socioecological environments, increased market access, globalization, transition to cash economies, forest degradation, erosion of cultural heritages and nutritional transitions, wildmeat remains part of the menu (Alves and van Vliet, 2018). Rural people in tropical contexts usually maintain a utilitarian link to wildlife, but the degree of utilitarianism varies according to the context. Households more dependent on wildlife products will develop more utilitarian values than those who make a living out of wildlife based eco-tourism (Novelli et al., 2006). Similarly, households that highly depend on wildlife as a source of food (e.g., hunter-gatherer vs. sedentary agro-pastoralists) will have a more utilitarian orientation toward wildlife (Dounias and Froment, 2011). Poor households, who are usually highly dependent on wildmeat, are associated with more utilitarian attitudes toward wildlife and acutely perceive wildlife costs (e.g., crop raiding, dangerous encounters, etc.), particularly women who are more involved in agricultural and gathering activities (Bragagnolo et al., 2016; Rickenbach et al., 2017). Concern for safety or damage is indeed a mayor dimension shaping the domination orientation, with social factors as diverse as religious affiliation, ethnicity and cultural beliefs all shaping humanwildlife conflict intensity (Dickman, 2010).
However, qualifying rural wildlife value orientations as merely utilitarian or domination oriented would be simplistic and fail to elucidate the complex, nuanced and varied relations that humans have with animals, and that animals have with humans around the world (Hovorka, 2017). In rural contexts, the use of wildlife serves multiple purposes depending on the specificities of each context, but usually include an important role as a source of food, a strategy to reduce costs in crop production, a source of income, a source of medicine, as a means to strengthen social bounds, or as part of a wider system of interconnected socio-physical relationships and identity (Nasi et al., 2008; Fischer et al., 2013; El Bizri et al., 2015; van Vliet et al., 2015b; Ichikawa et al., 2016; Alves and van Vliet, 2018). Reducing the relationship with wildlife to a materialistic relationship erases the possibility to understand the pluralistic value orientations that persist and reproduce in rural contexts. The spiritualism/religious dimension, which could be interpreted as eco-centric (Rose, 2001) is clearly elucidated in buddhism communities living around the Khao Yai National Park and Kui Buri National Park in Thailand (Tanakanjana and Saranet, 2007) or among the Monpa villagers in Tawang district, India, who avoid hunting for religious/spiritual reasons (Aiyadurai et al., 2010). Some traditional people who live in wilderness areas continue to view themselves as elements of nature, asserting spiritual values to wildlife that are reproduced by myths, rituals, taboos, and totems (Jimoh et al., 2012; Golden and Comaroff, 2015). Based on case studies from 33 countries, Bhagwat and Rutte (2006), showed that several communities across the globe believe in sacred areas, which are left relatively untouched. The cultural and ceremonial values of wildmeat are translated in how ritual feasts rely on visual and culinary consistency (e.g., bushmeat used in circumcision ceremonies in Gabon (van Vliet and Mbazza, 2011); festival foods among the Kichwa in Ecuador (Sirén, 2012); Mishmi tribe rituals in India (Aiyadurai et al., 2010); communal rituals among the Chakhesang (Naro et al., 2015). Familiarity, identity and taste for wildmeat are among the values that our nervous systems shape by starving for the familiar flavors and aromas of wildmeat and rejecting the more unusual tastes (Rose, 2001; Aiyadurai et al., 2010; van Vliet and Mbazza, 2011). For most hunters the motivation is not merely to satisfy hunger but also to meet a desire for bushmeat (the socalled "meat hunger" by Dounias and Ichikawa, 2017). Wildmeat consumption promotes a sense of "groundedness," security and identity, whose value is difficult to capture in materialistic terms (Jepson and Canney, 2003). Food preferences and habits are formed in large part through childhood experiences and actually persist throughout the course of an individual's life, helping to maintain memories and strengthen connections with traditional origins and territory (van Vliet et al., 2015c). The importance of hunting for cultural prestige is also a reality in many contemporary societies. In Kenya, for example, young men kill lions to earn social recognition, and there is a strong link between adherence to a local evangelical religion and the propensity to kill lions (Hazzah et al., 2009). Either through collective sharing or through the reciprocity logic, bushmeat sharing contributes to strengthen social bonds and reproduce cultural identity (van Vliet et al., 2015c; Lupo and Schmitt, 2017). Even in modern indigenous semi-urban communities in the Amazon, the consumption of wildmeat in positive social contexts results in a positive association between wildmeat consumption, emotional well-being and collective happiness (van Vliet et al., 2015c).
Value orientations toward wildlife probably differ substantially between small to medium sized towns flourishing in wilderness areas and the larger cities in which extinction of experience of wildmeat and wildlife might already be a reality, as evidenced in temperate regions from Europe and the United states (Cox and Gaston, 2018). However, for urban contexts in tropical forest areas, there is a lack of available data to generalize this assumption. With wild landscapes experiencing growing urbanization, new behaviors toward hunting and wildmeat consumption are gradually shaping, for example with the development of urban and peri-urban hunting patterns (Parry et al., 2014; van Vliet et al., 2015a) and the consumption of wildmeat becoming more associated to specific social events or considered as a delicacy or a source of prestige (Morsello et al., 2015; Shairp et al., 2016). In larger towns, urban lifestyles reduce daily interactions with nature as observed in temperate regions (Van Velsor and Nilon, 2006; Ballouard et al., 2011; Soga and Gaston, 2016; Cox and Gaston, 2018) and urban value orientations are likely to become more protectionist with strong emotional attachments to individual animals as already observed in Australia (Miller, 2003). While, available evidence has shown that protectionist orientations are much more prevalent in Western cultures than in other cultures (Novelli et al., 2006; Crudge et al., 2016), through globalization, TV, advertisement, conservation lobbies and social media, Western value orientations toward wildlife are increasingly spread beyond their geographic boundaries (as already evidenced in Kuala Lumpur by Baharuddin, 2013). How new behaviors toward wildmeat consumption actually evidence changes in beliefs and values toward wildlife is a key question that needs urgent attention from a human dimensions perspective. Currently, data available regarding social values toward wildlife, bushmeat, and the environment in urban contexts from tropical forest regions is mostly anecdotal, theoretical, or outdated. In Africa alone, which will see it's urban population increase to 62% by 2050 (World Health Organization Centre for Health Development, 2010), a better understanding of human/wildlife relations along the rural-urban continuum appears to be an evident necessity.
### FROM CONTINUUM TO STIGMAS AND CONFLICT OVER WILDLIFE MANAGEMENT
While the relationships with wildlife are obviously complex and full of nuance, the debate has often ended in over simplifying and polarizing the opposed visions. The more the "hunterwildlife" relationship is reduced to the negative connotations of domination values, the more likely it is that protectionist behaviors are accused of "cultural imperialism" and provoke cultural backlash. With the media acting as a debate heater, these two extreme visions are becoming more difficult to reconcile.
On one hand, over the past decades, with the alarming scientific evidences of wildlife declines (Dirzo et al., 2014; Ripple et al., 2016; Benítez-López et al., 2017; van Velden et al., 2018), the protectionist orientation has gained more strength (Cooney et al., 2017). A conservation war through stricter law enforcement, militarized protection, and behavioral change approaches, are all part of the international agenda to downsize consumption of wildmeat in tropical regions at local, national and international scales (Government of the UK., 2013; Commission européenne, 2015; USAID, 2016).
On the other hand, active indigenous groups worldwide are gaining more power to voice their right to consume wildlife, including the right to trade wildmeat (Eilperin, 2013; Searles, 2016; O'Neill, 2018). The main arguments used are food sovereignty (Searles, 2016; Hoover et al., 2017), quality of the diets (Samson and Pretty, 2006; Bodirsky and Johnson, 2008; Bordeleau et al., 2016; van Vliet et al., 2017a,b), protection of cultural identities (Fischer et al., 2013), and the right for selfdetermination (Schweitzer et al., 2000). Protectionist measures are increasingly tagged with severe accusations of cultural imperialism (Neves-Graça, 2010 and cultural genocide Kingston, 2015. Recently, an international conservation organization has been accused of inadvertently facilitating serious human rights abuses against pygmy groups living in Cameroonian rainforests (Survival International, 2016). The report entitled "The human costs of conservation in Republic of Congo" (Ayari and Counsell, 2017) reached un-precedent influence on conservation business in Africa and is pushing funding agencies to foster human rightsbased approaches to conservation.
These extremes in "cultures of nature" only exacerbate conflicts over management decisions. Following the term used by Manfredo et al. (2017), the stigmatization of the debate around the use of wildmeat in tropical regions will ultimately foster a "cultural backlash" with negative impacts on both wildlife and local livelihoods. A recent paper by Verweijen and Marijnen (2018) already demonstrates that strict law enforcement and joint operations of the Congolese army and park guards in Virunga National Park, fuel, rather than mitigate, wildlife poaching and armed mobilization. Local resistance to the strict enforcement approach translates into forms of "resistance poaching" within the boundaries of the park (purposely targeting key conservation species), under the protection of armed groups. As such, the perpetuation of extreme value orientations will result in a lack of adequate policy and management responses, trapping rural/indigenous communities in a vicious cycle of illegality, un-sustainability and criminalization and leading to the continued ecological and cultural extinctions of tropical wildlife.
### CONCLUSION
I stress the need for a more careful consideration of value orientations toward wildlife not assuming attitudes in congruence with western conservation interests nor assuming that traditional /indigenous values toward wildlife are carved in stone. The challenge is to bring segmented perspectives away from hegemony, into an overall vision for conservation that is broadly inclusive of a full range of wildlife values (Manfredo et al., 2017). Taking into account both hegemonic and marginalized ideas about wildlife will reduce the likelihood for conservation abuses in postcolonial contexts (McGregor, 2005) and provide a unique opportunity to shift the paradigms in tropical wildlife management. The human stakeholders with the most to lose often have no voice in decision-making. This is why, although some conservation practitioners suggest that promoting cultural change regarding wildlife use is legitimate based on evidence-based scientific knowledge about the "bushmeat crisis" (Jepson and Canney, 2003; Dickman et al., 2015), I argue that acknowledging the disparities in power relationships, providing the necessary grounds for a fair debate and support free decision making by the legitimate constituency are all necessary steps to avoid claims of "cultural imperialism" in conservation practice. Failing to do so might increase the potential for social conflict over wildmeat management issues. In line with Hovorka (2017) I think it is crucial to embrace the richness and complexity of cross-cultural plurality and take disparate value orientations seriously without privileging any-one presumptively. In a period of unparalleled socialecological change, bringing together the differences in wildlife value orientations between local/international, rural/urban, traditional/western visions is as necessary step in radically
### REFERENCES
reconstructing a new paradigm for a sustainable and culturally respectful wildmeat sector.
### AUTHOR CONTRIBUTIONS
The author confirms being the sole contributor of this work and approved it for publication.
### FUNDING
The funding agency is CGIAR funds.
### ACKNOWLEDGMENTS
This work was carried out under the Bushmeat Research Initiative from the Forests, Trees and Agroforestry Program, CGIAR.
Synthèse. Available online at: https://ec.europa.eu/europeaid/sites/devco/files/ eu-wildlife-strategy-africa-synthesis-2015\_fr\_0.pdf
**Conflict of Interest Statement:** The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Copyright © 2018 van Vliet. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
# Hunting in Times of Change: Uncovering Indigenous Strategies in the Colombian Amazon Using a Role-Playing Game
Nicole Ponta<sup>1</sup> \*, Tina Cornioley <sup>1</sup> , Anne Dray <sup>1</sup> , Nathalie van Vliet <sup>2</sup> , Patrick O. Waeber <sup>1</sup> and Claude A. Garcia1,3
*<sup>1</sup> Forest Management and Development group, Institute of Terrestrial Ecosystems, Department of Environmental System Science, Swiss Federal Institute of Technology (ETH), Zurich, Switzerland, <sup>2</sup> Center for International Forestry Research (CIFOR), Bogor, Indonesia, <sup>3</sup> CIRAD, UR Forest and Societies, Montpellier, France*
#### Edited by:
*Enrico Di Minin, University of Helsinki, Finland*
#### Reviewed by:
*Edson Gandiwa, Chinhoyi University of Technology, Zimbabwe Hirokazu Yasuoka, Kyoto University, Japan*
> \*Correspondence: *Nicole Ponta nicole.ponta@usys.ethz.ch*
#### Specialty section:
*This article was submitted to Conservation, a section of the journal Frontiers in Ecology and Evolution*
> Received: *31 October 2018* Accepted: *29 January 2019* Published: *28 March 2019*
#### Citation:
*Ponta N, Cornioley T, Dray A, van Vliet N, Waeber PO and Garcia CA (2019) Hunting in Times of Change: Uncovering Indigenous Strategies in the Colombian Amazon Using a Role-Playing Game. Front. Ecol. Evol. 7:34. doi: 10.3389/fevo.2019.00034* Despite growing industrialization, the shift to a cash economy and natural resource overexploitation, indigenous people of the Amazon region hunt and trade wildlife in order to meet their livelihood requirements. Individual strategies, shaped by the hunters' values and expectations, are changing in response to the region's economic development, but they still face the contrasting challenges of poverty and overhunting. For conservation initiatives to be implemented effectively, it is crucial to take into account people's strategies with their underlying drivers and their adaptive capabilities within a transforming socio-economic environment. To uncover hunting strategies in the Colombian Amazon and their evolution under the current transition, we co-designed a role-playing game together with the local stakeholders. The game revolves around the tension between ecological sustainability and food security—hunters' current main concern. It simulates the mosaic of activities that indigenous people perform in the wet and dry season, while also allowing for specific hunting strategies. Socio-economic conditions change while the game unfolds, opening up to emerging alternative potential scenarios suggested by the stakeholders themselves. Do hunters give up hunting when given the opportunity of an alternative income and protein source? Do institutional changes affect their livelihoods? We played the game between October and December 2016 with 39 players—all of them hunters—from 9 different communities within the Ticoya reserve. Our results show that providing alternatives would decrease overall hunting effort, but impacts are not spatially homogenous. Legalizing trade could lead to overhunting except when market rules and competition come into place. When it comes to coupled human-nature systems, the best way forward to produce socially just and resilient conservation strategies might be to trigger an adaptive process of experiential learning and scenario exploration. The use of games as "boundary objects" can guide stakeholders through the process, eliciting the plurality of their strategies, their drivers and how outside change affects them.
Keywords: wildmeat, hunting, role-playing games, wildlife management, Colombia, indigenous, alternative livelihoods, companion modeling
## INTRODUCTION
In the tropical Anthropocene, hunting, and the trade of wildlife still play a crucial role in the livelihoods of rural communities (van Vliet, 2011; WHO and CBD, 2015; Nielsen et al., 2018). More than 150 million households in Asia, Africa and Latin America rely to some extent on wildmeat to meet their dietary requirements and support their economies (Nielsen et al., 2018). Tropical forest productivity for wildmeat is generally lower than in open habitats (Robinson and Bennett, 2000). As a result, overhunting is considered a major threat for biodiversity and for the people that depend on it as a source of food and income (Bennett and Robinson, 2000; Milner-Gulland et al., 2003; Ripple et al., 2016). Hunting may have far-reaching consequences on entire habitats, depleting species responsible for key ecosystem functions such as seed dispersion, predation and herbivory (Emmons, 1989; Wright, 2003). The effects are not homogenous on the plant community as hunters tend to target large-bodied vertebrates, which are more likely to disperse largeseeded plants (Peres, 2007; Kurten, 2013). Selective hunting, along with habitat fragmentation, has led many seemingly "pristine" tropical forests to suffer from the "half-empty forest syndrome" (Redford and Feinsinger, 2003).
While there is general consensus on the unsustainability of hunting (Fa et al., 2002; Milner-Gulland et al., 2003; van Vliet et al., 2015b; Ripple et al., 2016; Benítez-López et al., 2017), assessing the impact of harvest remains a challenge. Static sustainability indices and site comparison studies have often proven very sensitive to model parameters and are ultimately not appropriate for measuring the impact of hunting (Ling and Milner-Gulland, 2006; Levi et al., 2009; Weinbaum et al., 2013; van Vliet et al., 2015a). Sustainability studies need to acknowledge the complexity of the hunting system, its spatial and temporal heterogeneity and its inherent human component made of the evolving needs and aspirations of millions of people around the globe (van Vliet et al., 2015b).
Bringing harvest to sustainable levels means integrating both social and environmental components of management as well as their dynamic relationships in modeling efforts (Verburg, 2006). Effective and socially just conservation initiatives should be guided not only by the best available information on the resource but also by a deeper understanding of people's strategies, their drivers and their adaptive capabilities (Feintrenie et al., 2010; Garcia et al., 2010; Bennett et al., 2016). The need for more inclusive, community-based approaches to conservation practice is widely recognized; however, their implementation is still limited and performance remains well-below expectations (Berkes, 2004; Bennett et al., 2017). Centralized command and control approaches that alienate local resource users are still prevalent but tend not to be effective where weak, underfunded institutions fail in enforcing the rules and simultaneously contribute to the marginalization and poverty of rural communities (Brandon et al., 1998; Barrett et al., 2001; Andrade and Rhodes, 2012; Brockington and Wilkie, 2015). The simultaneous lack of enforcement and criminalization of hunting and trade bare the risk of encouraging hidden practices that—because of their illegality—evade institutional control and restrictions (Nasi et al., 2008; Duffy et al., 2016).
This reflects the current situation in Colombia where, despite trade prohibition, wildmeat can be found in the markets of rural as well as urban centers around the country (van Vliet et al., 2014a). Decree 2811 from 1974 allows hunting of nonprotected species outside protected areas as long as it is for the subsistence of the hunter and her or his family. For trading, independently of the scale and purpose, hunters need an official license, which is extremely complex if not impossible to get for members of rural communities (van Vliet and Gomez, 2015). The institutional definition of subsistence in the legal framework considers only food safety. Against this narrow definition, the local concept of subsistence includes other needs linked to housing, education and health, which can be covered through trading part of the wildmeat. Despite its history of strict conservations and the foreseen implementation challenges, the current political debate in Colombia is favorable to sustainable use models (van Vliet, 2016).
The Ticoya indigenous reserve, in southern Colombia, serves as an example of these processes. The reserve's local economy relies mainly on shifting cultivation. The main staple crops yucca, plantain and corn are protein-poor and people complement their diet by fishing and to a lesser extent by hunting (Eden, 1990; Maldonado, 2010; van Vliet et al., 2014b). Despite its relative remoteness, the region is undergoing significant socioeconomic changes at an exceptionally high rate. Economic development is affecting people' diets as well. Processed food products coming from Southern Brazil, the Peruvian Andes and other areas of Colombia can now be easily found in the reserve (van Vliet et al., 2014b). Because of all these factors, local communities are relying more and more on the cash economy and industrialized products. At the same time, their cultural identity and indigenous rights over land and political autonomy are increasingly acknowledged (van Vliet et al., 2018). Yet in this emerging globalized society, local communities still rely on their surrounding forest and—among other activities do hunt and trade wildlife in order to meet their livelihood requirements (Bodmer and Lozano, 2001; van Vliet et al. 2015d; Bennett et al., 2016).
To ensure food security and strengthen cultural identity, local hunters have created in 2016 Colombia's first indigenous hunters' association: Airumaküchi (van Vliet, 2016). One of Airumaküchi's first objectives is to work toward sustainable use of wildlife and it is therefore in their own interest to uncover hunters' strategies, with their driving values and aspirations. This would not only urge institutions to acknowledge their effort and adapt the legal framework for subsistence trade, but it is also changing the way hunters are perceived by society.
Given the persistence of hunting and trading in Colombia, the challenges to sustainability and the openness of the national government toward sustainable use, there is a need to better understand hunters' decisions to ensure that future conservation initiatives have the desired ecological and social outcomes. Airumaküchi and our research team co-designed a role-playing game—named TICOYA —to initiate and support a learning and collective decision-making process around hunters'
most relevant issue: how can they hunt sustainably, ensuring biodiversity conservation, and food security under current and future socio-economic conditions? We developed the game following the Companion Modeling (ComMod) approach (Étienne, 2013). We used the game to elicit hunters' strategies and their underlying reasons under different scenarios. The design of the scenarios was guided by two hypotheses:
These two hypotheses led to the formulation of four scenarios. We specifically explore what drives hunters' decisions when (1) communities are isolated and wildmeat trade is illegal, (2) communities have access to income and diet alternatives and trade is illegal, (3) communities have access to alternatives and trade is legal and finally (4) trade is legal and competition is higher due to the intrusion of external hunters. A more detailed description of the scenarios is given in the Materials and Methods section.
Games, particularly role-playing games, help stakeholders shed light on complex socio-ecological systems, their internal dynamics and feedback mechanisms and the multiplicity of perceptions (Basco-Carrera et al., 2017; Reibelt et al., 2017; Redpath et al., 2018). Confronting players with their actions and their impacts not only fosters understanding but also contributes to strategic management, supporting people to think adaptively and creatively in the face of the challenges encountered in the game that reflect those of the real world (Barreteau et al., 2011; Speelman et al., 2017). Within this context, role-playing games follow a constructivist approach that does not aim at finding definitive solutions but at triggering an adaptive process of collective learning, exploration and experimentation (Xiang, 2013; Redpath et al., 2018). Games and simulations have already been used to explore hunters' behavior in different contexts (Bousquet et al., 2001; Mathevet et al., 2007; Le Page et al., 2015; Bodonirina et al., 2018; Marrocoli et al., 2018). To our knowledge, this is the first study that looks at the effects of livelihood and policy interventions within a setting codesigned by the local hunters and the research team.
### MATERIALS AND METHODS
### Study Area
The study was conducted in the Colombian department of Amazonas, a strip of land that stretches between Brazil and Perú. The area, part of the so-called tri-frontier, stands out as a dynamic region, where people, cultures, and goods flow ceaselessly across the few roads and the many rivers (Schor and da Costa Avelino, 2017). Specifically, this work was performed within the boundaries of the Ticoya indigenous reserve and
conducted, from right to left: Puerto Esperanza, Puerto Nariño, Ticoya, Santa Teresita, San Francisco, Doce de Octubre, and Tipisca. In green, partly overlapping the Ticoya reserve, the Amacayacu National Park. One workshop was organised in the proximity of Leticia along the National Route 85.
along the Leticia-Tarapacá road (**Figure 1**). The reserve (140 623 Ha) comprises a total of 22 communities located along the banks of the Amazon, Loretoyacu, Amacayacu, Boyahuazu, and Atacuari rivers. It was established in 1990 through INCORA Resolution 021 (Ruiz, 2008) and is home to several indigenous ethnic groups, mainly Ticunas, Cocamas and Yagua—the Ticunas being the most numerous (Riaño, 2003; INEI, 2010). Amazonian indigenous groups as well as non-indigenous people (colonos and mestizos) have converged on the area during the past century, attracted by a series of economic booms (such as rubber, pelt, coca) or because they were displaced from their original settlements (Ortiz, 1984; INEI, 2010). A section of the Ticoya territory is shared with Amacayacu National Park (ANP), which covers 293,500 Ha between the Amacayacu river and the border with Peru (Franco, 2006). Most of the Ticoya reserve is forested area classified according to the rivers' flooding regimes: the varzea forest is seasonally flooded by nutrientrich water, the swamp forest by nutrient-poor waters, while the terra firme forest is never flooded (Moreno Arocha, 2014). Such a forest mosaic sustains a rich and diversified fauna which has been described mainly within the Amacayacu park (PNNA, 2006; Maldonado, 2010). Outside the park borders, information on wildlife richness and abundance originates mainly from hunters' offtakes and markets (van Vliet et al., 2014a,c; Sandrin et al., 2016). Birds represent the most diverse vertebrate group, with more than 450 species present. Among the most detected are birds of the Cracidae family such as the nocturnal curassow (Nothocrax urumutm—least concern) and Ponta et al. Hunting in Times of Change
the helmeted curassow (Pauxi Pauxi—endangered). Within the reptiles, both caimans—such as the spectacled caiman (Caiman crocodilus—least concern) and tortoises—such as the yellowfooted tortoise (Chelonoidis denticulate—vulnerable) are present in the region. The most numerous group of mammals is represented by rodents, with the lowland paca (Cuniculus paca least concern) and the black agouti (Dasyprocta fuliginosa least concern) being rather common—and hunted—species. At least three species of armadillos are present, including the giant armadillo (Priodontes maximus—vulnerable) and about 12 species of primates such as the common wooly monkey (Lagothrix lagotricha—vulnerable). Among the largest and most valued mammals are the white-lipped peccary (Tayassu pecari vulnerable), the collared peccary (Pecari tajacu—least concern), the red brocket (Mazama americana —data deficient), the gray brocket (Mazama gouazoubira—least concern) and the lowland tapir (Tapirus terrestris—vulnerable).
The main settlement of the Ticoya reserve—and second largest municipality in the department after Leticia—is Puerto Nariño. The inhabitants of the communities surrounding Puerto Nariño as well as Leticia, base their subsistence mainly on farming and fishing (Eden, 1990; Trujillo, 2008; Maldonado, 2010). Additional income is gained from the illicit trade of coca and other forest products (mainly cedar—Cedrela spp) and from the expanding tourism industry (Zárate and Ahumada, 2008). Though to a lesser extent compared to the other activities, wildmeat hunting and trade significantly contribute to the local economy as well as to the people's diets (van Vliet et al., 2015c). Irrespective of the purpose, any kind of trading is illegal and hunters run the risk of paying fines and having their catch confiscated every time they sell wildmeat. Despite the prohibition, about 43% of the catch is sold, mainly to neighbors within the same community and occasionally to restaurants, schools, army soldiers and retailers (Quiceno-Mesa et al., 2014). Animals traded are mainly mammals (60%), birds (26%), and reptiles (14%). Most hunters use rifles for hunting although other techniques such as traps, hunting with dogs and—to a minor extent—blowpipes are also used (Sandrin et al., 2016). Rifles, as well as munitions, are mostly illegally sourced but they are generally tolerated if they are not carried around in the urban centers.
### Game Development
The game used in this study has been developed jointly by the research team and the local hunters, following the ComMod approach (Etienne, 2014). ComMod is an iterative, participatory modeling approach based on the assumption that participation of the local actors in the model design benefit not only the actors themselves but also the researchers and the decision-makers.
Researchers and hunters repeatedly met in two field missions in 2016 and 2017 to (1) understand and agree upon the main issues at stake, (2) build a conceptual model identifying the most relevant components in the system and (3) develop and validate a role-playing game that allows stakeholders to discuss creatively and constructively how to address the issues identified. Although the co-design phase (1 and 2) triggers a learning process helping stakeholders to share their own perceptions in order to build a common vision of the socio-ecological system (Bodonirina et al., 2018), we report here on the last phase of the project: the implementation of the validated role-playing game in the field.
The game recreates a simplified reality covering the main dynamics related to hunting as elicited during the ComMod process (Etienne et al., 2011). All game parameters have been calibrated based on information collected during the diagnostic phase through semi-structured interviews, collective workshops and, to a lesser extent, via literature. Stakeholders participated actively to every stage of the process and developed a sense of ownership and commitment toward the objectives of the study. Mutual trust was an essential ingredient for ensuring dialogue, promoting learning, and supporting collective decision-making. The study was done in compliance with the ethical guidelines and principles outlined by the Swiss Commission for Research Partnerships with Developing Countries (Stöckli et al., 2012).
### Data Collection
We organized nine workshops at three separate locations with 40 different participants (35 men, 5 women) between October and December 2017 (see **Table S1** for more information on participants). The selection of participants was organized by our local partner, the hunting association Airumaküchi. As women are mainly garden hunters (Linares, 1976; Smith, 2005) and rarely engage in long-distance hunting, only five were part of the workshops.
All participants were experienced hunters, with 31 of 40 being members of the association. The total number of hunters active within the Ticoya reserve is not known with accuracy given that the definition of hunter is vague. Most inhabitants of the reserve hunt opportunistically while fishing and farming while only a fraction hunts regularly—except when presented with alternative income activities. The number of members of Airumaküchi about 50 at the time of the study—is not an exact representation of all active hunters but is a good proxy as the association spent significant effort in promoting its activities throughout the reserve. Moreover, many of the hunter members of Airumaküchi are those who explicitly expressed willingness to work toward a sustainable management plan. A voluntary engagement fosters the building of trust and legitimacy, essential ingredients of a ComMod approach.
Each workshop had a different set of participants, most of them from Puerto Nariño or from close communities along the Loretoyacu river. Only three participants were from another indigenous reserve from the outskirts of Leticia, on the incomplete road to Tarapacá. The communities within the reserve along the Amazon river were not included in this study as their residents rely mainly on fishing. Most of their territory is indeed seasonally flooded by the Amazon river and hunting except for birds and caimans—is negligible.
Most of the participants were Ticunas (25), while the remaining 15 participants consisted of Yagua (7), Cocamas (3), Muinane (2), Bora (1) and one was a colono who had settled in the region 34 years ago after fleeing from another department of Colombia. All participants spoke fluent Spanish and most were able to write and read, although this was not a prerequisite for playing the game.
Ponta et al. Hunting in Times of Change
Most workshops were organized in Puerto Nariño and the participants were personally informed and invited a few days before by the researchers and members of Airumaküchi. Only two workshops took place in smaller and more remote communities as this was logistically easier for all partners. Traveling expenses were covered and lunch was provided at the end of every workshop for all participants and their families. No additional compensation was given.
### Workshop Structure
Every workshop followed the same structure (**Figure 2**). First, a facilitator explained the players' targets and the rules of the game. Then, two successive game sessions were played, each under two different scenarios. Each scenario has two rounds, representing the region's two ecological seasons, characterized as low water and high water. As the workshops were conducted when the level of the Amazon river was at its lowest, the research team always started the games with the low water round. We encapsulated different time scales in one round. While for the players a round represents approximately 1 month within the respective season, and the players need to take decisions within this timeframe, ecologically each round corresponds to the entire season and animals reproduce and move accordingly. As the game progressed, players needed to allocate effort to different activities, possibly go hunting and consume or trade the wildmeat harvested. After that, animals reproduced and moved within the system. At the end of each game totaling four rounds, time was allocated for in-depth debriefings. It is at this stage that game and reality come together (Garcia et al., 2016); the players carefully reflected upon their decisions and the resulting impacts they experienced during the game and connected them with the decisions and practices in their everyday life. Generally, each workshop lasted between 4 and 5 h depending on the debriefing's depth and on the players' commitments.
### Game Structure
Each round, players have to fulfill a two-fold objective: a livelihood requirement represented by a specific protein intake (20 kg) and a budgetary (100,000 Colombian Pesos COP) target. These goals are based on the monthly energy and income requirement of a household composed of two adults, two children and one elder (WHO, 1991).
To satisfy their targets, players have at their disposal two currencies. Energy points represent their human capital, with 10 energy points given every turn to each player. Money, fake bank notes with the same denomination as the Colombian Pesos, represents financial capital. Each player receives 70,000 COP. Players have to allocate their energy budget to a combination of different activities: farming, fishing, hunting in four different territories, logging and performing a salaried job to meet their nutritional and income targets. All of them, except hunting and fishing, return immediate monetary rewards. Fishing returns a fixed amount of fish, depending on the amount of energy invested. Hunting is a risky activity and not all hunting trips are successful, the reward will depend on the prey—if any that is killed. If players choose to allocate part of their energy to hunting, they need to decide whether to hunt in a territory
closer or more distant to the village. Traveling—on foot and/or by boat—to a certain territory, spending the night in the forest and using rifles entails a certain cost. The price of hunting—as well as the energy required to perform it—depends on the season, on the territory chosen and on the game scenario (**Table 1**). The scenarios represent situations reflecting important modalities of the basic assumptions behind the study. A scenario dictates what rules and options (activities) are in play and therefore represent the context to which players will need to respond. The scenarios were developed during the design phase of the project through collective workshops and interviews.
In scenario 1, players live in a remote community and do not have access to alternative sources of protein or to any paid job. They can only perform subsistence activities (farming, fishing and hunting) and can meet their protein target only by hunting and fishing, or by buying fish or potential excess wildmeat from other players within their community. If players have hunted or fished in excess of their monthly target, they can sell the surplus to other hunters in need or to the local market. As in reality, hunting wildmeat for subsistence purposes is legal, while trading it is illegal. In the game, this means that every time a player wants to sell any excess meat, the game master throws the dice and players run the risk of being caught by the police, having their meat confiscated and being fined. In scenario 2, players live in a connected community and have access to grocery
*This last variable indicates how many times hunters can draw/hunt within each territory depending on the energy allocated. Activities reward, price and number of attempts are all based on the minimum energy required to perform each activity.* \**Paid job and Logging become available for players only from scenario 2.*
shops that sell industrial meats, and to an alternative source of income, a job offered either by the town council or by a logging company. The former refers to a job performed in town (such as construction worker), the latter one to a job in the forest. Meat trade is still illegal. Scenario 3 represents the situation where the Colombian government changes the requirements for obtaining a commercial hunting license and wildmeat trading becomes legal. This causes certain changes, such as an increase in the price of ammunition, which can now be legally obtained. In scenario 4—the final scenario—the legality of the trade has attracted commercial hunters from elsewhere to converge on the region, thereby creating competition for the local hunters and increased pressure on the animal population. In all scenarios, market prices for fish (5,000 COP/kg), wildmeat (8,000 COP/kg) and, when available, industrial meat (8,000 COP/Kg) do not change. When players trade their catches between them, they can bargain about the price.
The landscape to which the players have access (the gameboard) hosts three animal species moving and reproducing according to species-specific characteristics: the lowland paca (Cuniculus paca), the white-lipped peccary (Tajassu pecari) and the South American Tapir (Tapirus terrestris). These species have been chosen because they are popular game species and because of their different life histories. The paca is a large frugivorous rodent present across the whole Neotropics (Emmons, 2016). It can be encountered both in the forest as well as in the farming areas close to the villages and it is one of the most common species of prey caught by the local hunters (Sandrin et al., 2016). They occupy a relatively small home range and their mass ranges from 7 to 12 kg (Ojasti, 1996). The peccary has similar distribution compared to the paca but a higher weight (25–40 kg) and a much larger home range (Gottdenker and Bodmer, 1998). They are a nomadic species and move in herds of a few individuals up to a few hundreds. The tapir is a large-bodied solitary herbivore and with its 150–250 kg it represents the holy grail for hunters (Robinson and Redford, 1986). It plays a key role in the forest's dynamics as a seed disperser and predator (de Thoisy et al., 2010) and, as the peccary, it is classified as vulnerable by the International Union for Conservation of Nature (IUCN) Red List of Threatened Species (Naveda et al., 2008; Keuroghlian et al., 2013).
The animal population dynamics of these three species are modeled through an agent-based model, which runs in the back in support of the tabletop role-playing game and is used for analytical purposes by the research team. Players do not have any interaction with the computer interface, only with the gameboard, where animals are represented by wooden tokens hidden in four different bags—the hunting territories (**Figure 3**). Each token represents an animal individual, either male or female. Among the animal tokens there are also "empty" tokens, which indicate when a hunting trip was not successful, either because no animal was encountered, the target was missed, or the rifle failed. The initial probability of success based on monitoring data from the hunting association—is set at 80% for all territories. If the number of animals in the game fluctuates, so will the probability of encounter. At the suggestion of the hunters, the probability of an encounter will never surpass 80%. There will always be an incompressible uncertainty in hunting, no matter how many animals are in the game. Players do not have any prior knowledge about the animal populations in the bags and can only acquire information by hunting.
All individual animals in the game are adults, either female or male. These individuals reproduce when they are found within the same territory, and the abundance of their species is below carrying capacity. Pacas reproduce every season, i.e., twice a year. Each male can reproduce with only one female, generating one offspring with a 50% probability of being either a male or a female. Peccaries reproduce only every other season, i.e., once a year, where each male can mate with a maximum of three females, generating two offspring, each with a 65% probability of being a female. Tapirs also reproduce only every other season, but each male can mate with one female only, generating one offspring, with a 50% probability of being either
FIGURE 3 | Gameboard during low water (A) and high water (B) season and agent-based model interface during low water (C) and high water (D) season. The colors in the virtual interface represent the different game species: paca (green), peccary (blue), and tapir (red); the shapes distinguish the gender, females (circles), males (triangles).
a male or a female. Irrespective of the species, individuals die only through hunting. There is no natural mortality in our model (Healy, 2017).
In the game, hunting occurs after all players have allocated their energy budgets to the portfolio of available activities and drawn at random one token from the bag corresponding to the chosen territory. There is no ordered game turn, with players following the "first come, first served" principle. We deliberately left this rule flexible to foster discussions among the players about their practices.
When hunters encounter a female of any species, they are asked if they want to know whether the female is pregnant or not. This choice originates from previous workshops where hunters proposed—as a way to reduce the impact of hunting—to ban the killing of pregnant individuals. The research team thus introduced it in the game to stimulate discussion on the applicability of such rule. Killing a pregnant female has impacts on the game population size, structure, and composition. In the game, the consequences of killing a pregnant female translate to non-reproduction of any individuals of the respective species and territory in the current game round. This is an obviously exaggerated effect used to spur the discussion during the debriefing stage. When players encounter a peccary as part of a bigger herd (there are at least four other individuals within the same territory), they are given the option of killing more individuals in the herd, as it is likely to happen in reality. Again, for every female, they must go through the usual set of questions related to its possible pregnancy.
Once hunting is over for all territories, players have to check whether they can meet their protein target with the wildmeat and/or fish they have harvested. If they are still lacking in protein, they can buy wildmeat from other players (if these have a surplus), buy fish from the community market (always available), or, starting from scenario two onwards, they can buy processed meat from the grocery shop. The players will also need to cover their expenses (100,000 COP) using the rewards from the activities or by selling fish or wildmeat either to other players, retailers or the market. It is at this stage that the agent-based model calculates the resulting abundance and distribution of the animals for the next round, and the tokens in the bags are updated accordingly. At every stage of the game, players make their decisions individually, although communication was never formally forbidden and some individual decisions might have been affected by other players. Simplified representations of the game dynamics are depicted in **Figures 4** and **5**. For a full description, see the Overview-Design-Detail (ODD) protocol in the **Supplementary Material**.
### Statistical Analysis
Throughout the game, the research team monitored and recorded the decisions of the players as well as their implications on the abundance and distribution of the three animal species. Players had to decide the following: (1) How to allocate their energy budgets to the available activities (energy budget allocation), (2) where to go hunting (territory selection), (3) what kind of protein to consume (diet composition), and (4) which, if any, protein they wanted to sell (protein reward).
For the analysis of the **players' behavioral data**, we used a multilevel multinomial logistic regression approach. This method is suitable for the analysis of energy allocation, territory selection, diet composition and protein reward
because it accounts for the trade-off-character of these behavioral choices; selecting one behavior precludes performing another (Koster and McElreath, 2017).
containers indicate questions posed to the players, the hexagonal ones are the game's outcomes.
We fit multi-response generalized linear mixed models using a Bayesian Monte Carlo Markov Chain (MCMC) approach from the package MCMCglmm (Hadfield, 2010) for R (R Core Team, 2018). The Bayesian framework provides a more flexible alternative for the analysis of hierarchical data compared to a frequentist approach, which is not always effective for the analysis of multivariate or non-normally distributed variables (McCulloch and Searle, 2001; Bolker et al., 2009).
All our multinomial models follow a categorical (generalized Bernoulli) distribution and have response variables of K categories, with one being the reference level with which the other categories are contrasted to. For each model, we fitted three random effects: individual ID of participants, round and workshop number. Individual ID is nested within workshop because people's decisions were measured repeatedly within each workshop and no one participated to more than one workshop. We fitted round as a random effect to account not only for potential temporal autocorrelation between successive rounds but also for the decomposition of the game session into two games following each other; the game was reset to initial conditions at the beginning of round 3. This implied that round 1 of scenario 1 was more similar to round 1 of scenario 3.
For the analysis of the **animal population**, we fit three generalized linear models with a Gaussian distribution, one for each animal species, using as response variable the population growth rate, calculated as:
$$r = N(t+1)/N(t) \tag{1}$$
Where N is the number of individuals of the focal species at round t and round t+1. In this case, our only random effect is workshop.
For both the behavioral and the animal population analyses, we supplied the models with weakly informative priors for the fixed effect parameters and for the variance-covariance matrices. For each model we first ran four parallel chains and we used the Gelman-Rubin diagnostics to check whether the chains converged to the same posterior distribution, an indication that the process is not happening by chance (Gelman and Rubin, 1992). Number of iterations, thinning and burn-in period for each model were determined using diagnostic plots from the coda package (Plummer et al., 2006).
### Energy Allocation to Activities
The activity model tested whether scenario and season had an effect on how people allocated their energy to different activities. The K possible categories were as follows: hunting, farming, fishing and income activities. Paid job and logging were pulled together in the same category because players considered them both as activities providing a fixed income independently of whether it was a job performed in the town or in the forest. Hunting was also represented by only one category, with the four different territories pulled together. Farming was our reference category. As fixed effects, we included scenario, season, the interaction between the two and the money availability of players. The model was run for 300,000 iterations with a burn-in of 100,000 and thinning of 10.
#### Hunting Territory Selection
Not all players chose to go hunting at every time step. When they did, they needed to decide where. This model tested whether scenario and season had an effect on hunting territory selection. For this model, we selected only the observations in which players did allocate some energy to hunting. The categories were the four Ponta et al. Hunting in Times of Change
different hunting territories, with hunting territory 1 being the reference. As for the previous model, fixed effects were scenario, season with their interaction and money availability. The model was run for 300,000 iterations with a burn-in of 150,000 and thinning of 10.
### Choice to Kill
Within hunting, we also intended to examine players' decisions concerning pregnant females and peccaries' herds. However, in both cases the sample size was too small to fit the models.
### Protein Consumption
At the end of the round, players needed to decide how to satisfy their protein target. There were three different kinds of protein that they could choose from: wildmeat, fish and industrial meat. Fish was always locally sourced. Explanatory variables were money availability and the amount of wildmeat hunted. The latter indicated whether players had hunted no wildmeat at all ("none"), less than the 20 kg target ("little") or as much as or above their 20 kg target ("enough"). The model was run for 150,000 iterations with a burn-in of 30,000 and thinning of 10.
### Protein Sale
This model examined how players satisfied their budget target. They could gain money by performing income activities, by farming and by selling the wildmeat and the fish they collected during the round. Explanatory variables are scenario and season. The model was run for 40,000 iterations with a burn-in of 10,000 and thinning of 10.
### Animal Populations
We fitted three models with Gaussian distribution and population growth rate for paca, peccary and tapir as response variable. As explanatory variable, we used the proportion of energy dedicated to hunting by all players at every round for every workshop. Hunting territory and season seem not to play an important role as shown by DIC (Spiegelhalter et al., 2002) and by the overlapping confidence interval of the posterior probabilities. Models were run for 250,000 iterations with a burn-in of 50,000 and thinning of 10.
### RESULTS
### Energy Allocation to Activities
The probability of allocating energy to hunting or fishing in scenario 1 did not differ between low and high water (**Figure 6**). In this scenario, income activities were not available and model predictions for this activity were virtually equal to zero. In scenario 2, when alternative sources of income and protein became available, less energy was allocated to hunting compared to scenario 1, in both seasons. In scenario 2, the probability of allocating energy to hunting was higher in the high-water season. The same was true for fishing. The probability of allocating energy to income activities reached a peak of 60.2% (CI = 50.2–66%) in low water and dropped to 29% in high water (CI = 22.6–33.6%) in season 2. When trade became legal, in scenario 3, the probability of allocating energy to hunting rose again substantially in both seasons. Income activities on the
FIGURE 6 | Probability of choosing each activity at every season and scenarios with money availability held constant at the sample mean. The confidence intervals are the 95th percentile intervals, as calculated from the posterior samples of the model. In gray is the reference level (farming).
TABLE 2 | MCMC results for the multinomial regression: we report the posterior mean and the 95% credible interval for energy budget allocation model.
*Farming is not shown as it is the reference level and its probabilities are calculated by subtracting from 1 the posterior means of the other activities.*
other hand became less popular and represented less than a quarter of the energy budget in both seasons. In the 4th and final scenario, when competition became fierce, players chose again to allocate substantially more energy to income activities and less to hunting compared to scenario 3, independently of the season. Across all scenarios, fishing and farming were the least affected by the socio-economic changes occurring, except in the low water round of scenario 2 when paid job was the preferred choice (**Table 2**).
Across the nine workshops, only one player stopped hunting starting from scenario 2. He was the only non-indigenous player. Except this one case, players never stopped hunting, even when given access to alternative sources of protein and income.
### Hunting Territory Selection
The territory model examined where players who allocated energy to hunting, chose to go hunting. As shown in **Figure 7**, territory selection varied between seasons within scenarios and
FIGURE 7 | Probability of choosing each hunting territory at every season and scenarios with money availability held constant at the sample mean. The confidence intervals are the 95th percentile intervals, as calculated from the posterior samples of the model. In gray is the reference level (hunting in territory 1).
*Hunting in territory 1 is not shown as it is the reference level and its probabilities are calculated by subtracting from 1 the posterior means of hunting in the other territories.*
between scenarios. Overall, players allocated more energy to territory 2 and less energy to territories 3 and 4 compared to territory 1, our reference level (**Table 3**). If we compare between TABLE 4 | We report the posterior mean and the 95% credible interval for the consumption model.
*Consumption of fish is not shown as it is the reference level and its probabilities are calculated by subtracting from 1 the posterior means of wildmeat and industrial meat.*
seasons, independently of the scenario, territory 2 was less visited during high water compared to territory 1, while territories 3 and 4 were more visited.
In both scenarios 1 and 2, during the low water season players had a higher probability to hunt in territories 1 and 2—which are closer to the village—compared to territories 3 and 4—which are further away from the village. When an alternative source of income became available (scenario 2), the probability of visiting territories 3 and 4 substantially decreased compared to scenario 1, meaning that in scenario 2, during low water, players preferred to visit hunting territories closer to the village.
When trade became legal in scenario 3, territory 2 was still the most likely visited territory during the low water season (mean = 45.1%, CI = 36–52.2%). However, there was a 27% (CI = 20.9–31.1%) and 27.7% (CI = 19.2–32.7%) probability of choosing territory 3 in low and high water, respectively, despite the energetic and monetary effort required to reach it when the water is low. Similarly, when competition came in (scenario 4), players had a 24.8% (CI = 16.5–28.9%) chance of visiting
TABLE 5 | We report the posterior mean and the 95% credible interval for the protein sale model.
confidence intervals are the 95th percentile intervals, as calculated from the posterior samples of the model. In gray is the reference level (reward from farming).
probability of selling each protein type at every season and scenario. The
territory 4, the farthest of all, during the low water season, and the probability to hunt in this territory never rose above 10% in all other scenarios. On the other hand, the probability of choosing territory 3 in scenario 4 compared to scenario 3 decreased significantly in both seasons.
### Choice to Kill
Overall, players drew 215 times a female token from the sacks. In 80% of the cases they wanted to know whether it was pregnant while in the remaining 20% they did not want to know. Out of the 80%, 59 females were not pregnant, 61 were pregnant and 52 escaped the hunter. It is the 61 cases where females were actually pregnant that interested us. In 15% of the cases players decided to kill the female.
Players across all workshops encountered a herd of peccaries only 25 times. Twelve percentage of players did not want to kill any additional individual, 48% killed one extra individual, 20% killed two more, 12% three more and 8% four more.
However, in both cases, there was not enough variation across scenarios and money availability to be detected by the models. Although the sample size was too small to prove any trend, data suggest that if players would have been able to recognize a pregnant female during the hunt, they would, in most cases, not kill it. In the case of the peccaries it seems that most players would kill at least one additional individual but only few would go for a higher catch.
### Protein Consumption
It is clear that both the amount of money and wildmeat players had at the end of the round affected their diet (**Table 4**). When they did not hunt any wildmeat, they did not consume any and only few bought wildmeat from other players (**Figure 8**). At low levels of income, they had an 80.4% probability of consuming fish (either fished or bought). With increasing income, the *Farming is not shown as it is the reference level and its probabilities are calculated by subtracting from 1 the posterior means of hunting in the other territories.*
probability of consuming industrial meat increased to a maximum of 70.4% (CI = 49.4–84.9%), basically replacing the fish fraction.
When players did collect some wildmeat ("little")—though below their protein target—they consumed it and complemented it with fish and/or industrial meat depending on the money availability. The proportion of the protein budget covered with wildmeat ranged from 35.4% (CI = 21.8–49.7%) at low levels of income to 68.2% (CI = 59.4–72.5%) at high levels of income. Despite being constantly lower, industrial meat consumption doubled as well, rising from about 11.2% (CI = 6.8–15.8%) to 21.5% (CI = 18.6–23%) of the total protein intake.
When the wildmeat collected was equal to or higher than the target ("enough"), players mainly consumed wildmeat, independently of the money availability. In this case, wildmeat represented between 85% (CI = 74.5–90.8%) and 95.4% (CI = 93.5–96%) of the total protein intake. Despite the possibility of meeting their protein target with wildmeat only, players consumed a relatively constant amount of industrial meat, ranging between 2.9% (CI = 2.5–3.2%) and 3.2% (CI = 3.1–3.4%) along the income axis. As in all other cases, fish consumption decreased with increasing income, declining from 12–1.4%.
TABLE 6 | We report the posterior mean and the 95% credible interval for the three animal population models.
### Protein Reward
In the absence of another source of income (scenario 1), wildmeat was the players' main source of income (**Figure 9**, **Table 5**). In the presence of an alternative source of income (scenario 2–4), a smaller fraction of players' income came from selling wildmeat. Yet, when wildmeat trading is legal and hunting competition is absent (scenario 3), a substantial fraction of the players' income originated from wildmeat. Across all scenarios, the proportion of income originating from wildmeat and paid jobs fluctuated substantially compared to the relatively constant contribution of fishing and farming. Overall, relatively little money is gained through fishing.
### Animal Populations
The only process affecting the animal population was hunting, as defined per game mechanisms. Hunting can not only remove animals from the system but can also stop reproduction if a pregnant female is killed during a round. Carrying capacity could temporarily halt animal reproduction but it was never reached in any of the game workshops and therefore played no role in defining animal abundance.
The animal population models showed that the higher the proportion of energy dedicated to hunting, the lower the population growth rate (**Figure 10**, **Table 6**).
### DISCUSSION
A common concern about experimental games is to what degree what happens under experimental conditions can be extrapolated to explain real-life behavior (Jackson, 2012). For a model to be able to support discussion about the real world, it first needs to seem credible to the users (Checkland, 1995). To ensure the relevance of our game, we designed it, tested it and refined it with the actors of the socio-ecological system we wanted to represent. Gaming and the collective discussions that followed during the debriefing sessions served as triangulation. Players reported that the game represents well the constraints that they have to face in their everyday lives. It created time for selfreflection. "The game gave us the time to analyze our everyday practices and it demonstrated to us that through a game we can understand what is happening in the real world" (Workshop 2, Player ID 2.3). Additionally, most players' decisions departed from the assumptions of maximization and rationality common in economic theory. Players seldom chose the most profitable activities, despite the shortage of money. Similarly, they sold one kind of protein to buy another one at exactly the same price, even though they could have met their target with the initial protein collected. This is a clear example of the difference between the game's internal and external validity. A decision that is consistent with the rules of the game is internally valid. A decision that makes no sense in regard to the rules but has nevertheless a meaning for the players because of their desires, beliefs and intentions they brought from real life is externally valid. Interpreting choices in the game as accurate representations of social realities is risky, even when the game actions match field data (Le Page et al., 2014). More importantly, it is not actually necessary that game choices represent accurately real life behavior (Speelman et al., 2017). The value of the games we use within the ComMod approach lies in their ability to generate collective learning, to foster critical thinking and to encourage creative actions in response to the issues encountered in the game (Checkland, 1995; Le Page et al., 2014). Acknowledging that the game is a tool and not a goal (Verburg et al., 2016), the TICOYA game allowed hunters to hold such in-depth discussions.
The scenarios played in the TICOYA game represent the current situation of the communities included in the study and a glimpse of their potential future. The current reality of the communities lies along a gradient between the first two scenarios. While the context in Puerto Nariño is closer to the one described in scenario 2, some of the most remote communities along the river are still better represented by scenario 1, and most communities are somewhere in between. On the other hand, scenario 3 and 4 represent options that—though plausible—have not occurred yet. Scenario 3 represents the political objective of the hunter's association, scenario 4 a likely outcome threatening community management. When players discuss their strategies in scenarios 1 and 2, they can draw upon real life examples depending on their community of residence. When they move to scenarios 3 and 4, they build their narratives through their values' and aspirations' lenses. While we should not infer directly real-life responses based on the actions in the game, we can nonetheless use the game as a metaphor of reality to foster in-depth discussion on real-life strategies with the hunters.
Brown (2002) hypothesized that providing forest dwellers with income-generating activity and domestic meat would reduce their dependency on wild meat, in turn decreasing hunting pressure. This is supposed to be particularly relevant for South America, where intense livestock production has the potential to cover the protein demand of the Amazon basin (Rushton et al., 2005). Domestic meat, though expensive, is indeed easily found in the grocery shops of Puerto Nariño. In the game, we observed a substantial decrease in hunting effort in scenario 2 compared to scenario 1 when alternative job opportunities and industrial meat become available, supporting the aforementioned hypothesis. Participants in the game workshops confirmed that if offered a job by a logging company or by the town council, they would accept it, especially during the low water season when forest streams and ponds are dry and hunting requires more time and energy investment. Most participants have indeed taken up job opportunities in their everyday life and reduced their hunting effort. Their wages allow them to buy expensive goods that they do not produce themselves but are available in the grocery shops and have become essential in their diets, such as vegetable oil, sugar and rice. During the workshop debriefings, participants claimed that in real life, they would still dedicate some time to hunting—though mainly in territories closer to the villages as they would not have the time anymore to go for longer expeditions deeper into the forest: "If I have a job, I can work during the day and go hunting during the night, it helps saving money" (workshop 4, player ID 4.5). Game results are consistent with players' statements. In scenario 2 they allocated significantly less energy to territories 3 and 4, those farther away from the village, during both seasons.
Participants gave several reasons for not giving up hunting in their everyday lives. First, the jobs that the hunters of the Ticoya reserve have access to are generally short-term and unstable, i.e., employees might have work 1 day but not the next one, and there are often delays for the generally meager payments. After all, hunters with low levels of formal education have little prospects of finding a job. In these conditions, although risky, hunting can be much more profitable than any of the jobs they can get. In addition, hunting offers a safety net for when salary does not come on time or for when it is insufficient to cover all expenses. This reflects the classic strategy of forest dwellers, composed of a mosaic of activities that complement each other at different levels depending on the season and on specific needs (Zenteno et al., 2013). This dynamic structure is one more reason why hunting might be preferred over a regular job whose intrinsic rigidity is less compatible with the agricultural cycle (Brown and Williams, 2003). All except one participant of the game workshops engaged in slash and burn cultivation, an activity that requires high levels of labor inputs at discontinuous times.
Cultural attributes alongside taste preference and diet diversification have also been mentioned in the workshop debriefings to justify the persistence of hunting effort in real life. Participants stated that they hunt because they like the taste of wildmeat and because it allows them to vary from a fish and canned meat diet—the latter considered unhealthy and a threat to traditional practices. Previous studies suggested that wildmeat—although routinely consumed—is not the favored source of animal protein in Amazonian towns (Nardoto et al., 2011; Morsello et al., 2015). However, these studies focused on a random sample of mainly urban households. Participants in our study were all hunters of medium to very small rural settlements who have cultural connections to wildmeat consumption and hunting practices. Some players who, in the game, sold wildmeat to buy industrial meat at the exact same price, justified themselves saying that their children prefer beef and chicken meat over wildmeat. This is in accordance with a previous study on children's preferences which included some of the same rural communities (van Vliet et al., 2015a). One player added, "We abandoned our culture, we are adopting the lifestyle of mestizo people and our children do not want anything to do with our indigenous culture" (workshop 3, player ID 3.4).
Consumption is not driven by cultural attributes only but by income and prices too, with wildmeat consumption falling with the decreasing price of domestic meat and vice versa (Ayres et al., 1991; Wilkie and Godoy, 2001). Economic theory suggests that an increasing income will increase the consumption of a certain good if there are no alternatives or if it is considered a superior good (in which case its consumption would be even higher). Consumption of that specific good would follow an inverted "U"- shaped curve (Kuznets, 1955), increasing up to a tipping point in which consumers—whose income has substantially increased—switch to other products that have become affordable. In our game, an increase in income led to an increase in wildmeat consumption when wildmeat was available. Industrial meat consumption doubled when little wildmeat was available. The increase in wildmeat consumption with increasing income and despite access to alternative products such as fish and industrial meat is an indication that in the game, wildmeat is considered a superior good. However, the parallel increase of industrial meat consumption suggests that we are close to the curve's expected tipping point. Independent of the amount of wildmeat available, consumption of fish constantly decreased with increasing income, suggesting that fish is an inferior good, consumed because there are no affordable substitutes. van Vliet et al. (2015b) support our findings by showing that children have a very low preference for fish compared to other proteins especially egg and beef.
Understanding the drivers behind wildmeat consumption is essential for designing effective policies for management of wildlife hunting and trading (Schenck et al., 2006). Our results suggest that wildmeat is consumed—and thus hunted—despite the presence of affordable substitutes, although increasing incomes led to increased consumption of industrial meat. The way that players behaved in the game, their statements and their real-life examples suggest that economic development the availability of alternative sources of income and of protein—could drive the consumption of wildmeat either up or down depending on the initial level of income. Empirical case studies show both trajectories in which additional incomes and changes in market prices decreased reliance on wildmeat (Ayres et al., 1991; Wilkie and Godoy, 2001), or increased it because of better access to more effective hunting equipment (Damania et al., 2005), or had no significant effect on resource exploitation (Torell et al., 2010).
Our results also indicate a more immediate impact of income alternatives on the spatial footprint of hunting. As workshop participants themselves admitted, if they had a job they would concentrate their hunting effort in areas easily reachable from the village—especially during the dry season when accessing more remote regions by boat is not possible. This situation would put additional pressure on an area already affected by habitat fragmentation and human disturbance (e.g., slash and burn agriculture, logging and noise). The lack of hunting heterogeneity over time and space could cause local wildlife populations to deplete over time and could prevent it from being replenished from other less hunted populations (Van Vliet et al., 2010).
In the game as in reality, the risk of getting the meat confiscated when players decided to sell is very low. Out of the 40 participants, only two players had had their wildmeat confiscated within the game, while seven of them stated that they have been confiscated in real life. Most trade occurs locally within the communities and none of the participants bring the meat to the department capital where control is much stricter (van Vliet et al., 2015c). When trade became legal (scenarios 3 and 4), players carried away a guided thought experiment since, though they aspire to it, they have never experienced legal trade first hand and their narratives are not based on concrete practices but on their values and aspirations only. Within the game, hunting effort substantially increased in both seasons once trade became legal. It increased particularly in the most distant territories from the community, territories 3 and 4, during the low water season. This is an indication that despite the cost and effort, players perceived hunting as worthier than other activities. When trade was illegal (scenario 2), hunting effort was significantly lower but, although income could be easily obtained through a regular salary, wildmeat was still sold—especially during the high-water season.
Sustainable wildlife management is gaining recognition under the assumption that a more flexible framework that takes into account the rights and the knowledge of local communities would enhance both conservation and human welfare (Miller et al., 2011). Acknowledging and strengthening the engagement of the human dimension is considered crucial for effective conservation decision-making (Bennett et al., 2016). Despite a general consensus toward sustainable use, wildmeat trade is still strongly criminalized in Colombia and forest dwellers are forced to walk a fine line between subsistence hunting and illegal trade prosecutable by law (van Vliet and Gomez, 2015). Overlooking the role that wildmeat plays in the food security, family economy and cultural identity of rural communities poses a problem for communities and might well be detrimental for the wildlife itself (Nasi et al., 2008). Our results show that players sold wildmeat independently of whether the trade was legal or not. Participants confirmed that they do often sell part of their harvest, a fact also shown by a previous study that included hunters from Puerto Nariño (van Vliet et al., 2014b). The customers are generally neighbors and other community members or, more rarely, restaurants in Puerto Nariño.
A common concern among conservationists is that legalizing wildmeat trade could increase hunting pressure by legitimizing potentially unsustainable levels of hunting, leading affected populations to extirpation (Wilkie et al., 2006). Of the 40 participants, only eight said that they would not hunt more if the wildmeat trade became legal because they hunt for the subsistence of their family and they are concerned with the animals' long-term viability. The other 32 participants declared they would indeed hunt more, endorsing what happened in the game where hunting effort was higher in scenario 3 compared to scenario 2. For most, legal trade would represent an opportunity to have an extra income for everyday necessities. This shows how the concept of subsistence for rural communities has a wider meaning compared to the official definition (Law 84 of 1989, article 30), which restricts legal wildlife use to food provision for the hunter and his/her family. Only three participants mentioned commercial hunting, and the possibility of selling large quantities of wildmeat at the market.
In the scenario 4, we introduced in the game some of the changes that the legality of the trade might bring, such as higher prices for hunting equipment and external competition. Most participants, when confronted with the new circumstances, declared they were not aware of the conditions that legal trade could entail. Taxation and competition had a negative effect on the game's hunting effort, which significantly decreased, though the effect was not equal for all territories. Players reacted to competition by allocating some of their time to a paid job, which again increased in popularity compared to the previous scenario without external players. However, no particular action was taken against the competing hunters that invaded the territory—except for a few occasions in which players rushed to draw animal tokens before the external competitors. Indigenous hunters in this region have implicit norms that regulate hunting access to the forest adjacent to the communities. These rules transcend national borders and are implemented by all neighboring communities, whether they are in Colombia, Perú or Brazil. The competition issue seems to arise only when intruders come from other regions.
The game and the scenarios explored within this study have been designed according to the context of the Ticoya indigenous reserve in Colombia. However, the guiding hypotheses are relevant for most tropical and sub-tropical regions where hunting and trade of wildmeat persists (Milner-Gulland et al., 2003; Brashares et al., 2011). In the Congo basin, for example, people also hunt wildmeat for food—whether directly consumed or sold to third parties—but extraction rates are higher compared to the less studied Amazon basin (Nasi et al., 2011). While rural consumption patterns in the two basins are comparable (Nasi et al., 2011), the rates are different at the urban level. Despite increasing acknowledgment of South American wildmeat urban consumption (Parry et al., 2014; van Vliet et al., 2014a, 2015a), the volumes of trade and consumption in African cities have yet to be overcome (Wilkie et al., 2005; Mbete et al., 2011). This is also due to the smaller livestock production of many Central African countries compared to South America and the consequent lack of valuable alternative sources of protein and income (Rushton et al., 2005; Bennett et al., 2007). We expect our results to be similar—though more trade-focused in the African context given similar socio-economic conditions of the study area, such as distance to the next urban settlement. Despite differences at the market and institutional level, the forest dwellers of the tropics are facing comparable challenges and opportunities. Over the long term, we expect hunters in Colombia as well as in other tropical regions—to benefit from the increasingly available alternative livelihoods but at the same time to be lured by a bigger and more accessible market for wildmeat products. The sustainable use of wildlife has the potential to tip the balance toward long-term conservation while at the same time providing a legal source of income and protein.
The aim of this study was to explore hunters' behavior within the environment of a game setting and relate them to their everyday practices, while eliciting individual and collective values, attitudes and aspirations. Specifically, we looked at the effects of policy interventions that are already partially in place in the Amazon region (alternative livelihoods) or that are strongly demanded by local communities (trade legalization). Our results support the hypothesis that providing alternatives would indeed decrease overall hunting effort but might also focus its footprint on smaller areas, canceling the positive effects of temporal and spatial hunting heterogeneity. Legalizing trade could encourage commercial and therefore less sustainable hunting, except when increased prices and competition make it less attractive. This
### REFERENCES
is so unless communities self-organize to control practices and exclude free riders.
Ultimately, for policies to be effective, they need to take into account the coping strategies of the people they are directed to. Our game proved to be a powerful tool to this end, capable of generating a safe and inclusive environment for stakeholders to discuss pressing yet delicate issues such as illegal trade and unsustainable hunting. Within and after the game, people do not fear being explicit about what drives their actions and, eventually, the whole system.
## DATA AVAILABILITY
The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.
### AUTHOR CONTRIBUTIONS
NP co-designed the study, collected data, performed the analysis, and wrote the manuscript. TC led the statistical analysis, provided figures, contributed ideas and reviewed. AD contributed to study design and ideas, provided figures, and reviewed. NvV contributed to study design and ideas. PW contributed to study design, ideas and reviewed. CG co-designed the study, contributed to ideas and reviewed.
### FUNDING
This work was supported by ETH Research Grant ETH-02 15-1.
### ACKNOWLEDGMENTS
The authors are grateful to the hunting association Airumaküchi, CIFOR, Fundación Science International and Fundación Omacha for sharing their data and their expertise and for their support in the field. We are grateful to the hunters of Amazonas, for dedicating their time, energy and knowledge to this project. We thank John Garcia Ulloa for creating the study area map.
### SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fevo. 2019.00034/full#supplementary-material
Barreteau, O., Bousquet, F., Etienne, M., Souchère, V., and D'Aquino, P. (2011). "Companion modelling: a method of adaptive and participatory research," in Companion Modelling. A Participatory Approach to Support Sustainable Development, ed M. Étienne (Dordrecht: Springer), 21–44.
**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Copyright © 2019 Ponta, Cornioley, Dray, van Vliet, Waeber and Garcia. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
# From the Forest to the Dish: A Comprehensive Study of the Wildmeat Value Chain in Yangambi, Democratic Republic of Congo
Given the importance of wildmeat for local and national economies, understanding the
Nathalie Van Vliet\*, Jonas Muhindo, Jonas Kambale Nyumu\* and Robert Nasi
*Center for International Forestry Research, Bogor, Indonesia*
#### Edited by:
*Franco Andreone, Museo Regionale di Scienze Naturali, Italy*
#### Reviewed by:
*Francesco Maria Angelici, Independent Researcher, Rome, Italy Massimiliano Di Vittorio, Ecologia Applicata Italia s.r.l, Italy*
#### \*Correspondence:
*Nathalie Van Vliet vanvlietnathalie@yahoo.com Jonas Kambale Nyumu jonasnyumu@gmail.com*
#### Specialty section:
*This article was submitted to Conservation, a section of the journal Frontiers in Ecology and Evolution*
> Received: *14 January 2019* Accepted: *01 April 2019* Published: *24 April 2019*
#### Citation:
*Van Vliet N, Muhindo J, Nyumu JK and Nasi R (2019) From the Forest to the Dish: A Comprehensive Study of the Wildmeat Value Chain in Yangambi, Democratic Republic of Congo. Front. Ecol. Evol. 7:132. doi: 10.3389/fevo.2019.00132* structure and operations of the informal wildmeat value chains is necessary to provide recommendations for a sustainable wildmeat sector. However, the limited number of case studies available hinders our capacity to understand general patterns in wildmeat trade chains and provide recommendations that apply in different contexts. This study contributes to our understanding of wildmeat trade chains with another case study from the Yangambi landscape, in the Democratic Republic of Congo. We use a value chain approach to explore the structure and functioning of the trade and identify the main barriers to entry into the business, as well as the main levers that can be used to reverse unsustainable use. Bushmeat remains the most consumed source of meat both in the main urban area and in surrounding villages. Urban consumption generates a trade of about 103–145 tons of bushmeat per year for a human population of 37,997 inhabitants. Yangambi combines all the factors for a depletion scenario: a quasi-open access system and high levels of dependency on the resource at all levels of the chain (from consumers to hunters). Despite this, emblematic species such as the chimpanzee, buffalo, okapi, red colobus and giant pangolin are still present in the area. The trade chain follows a "redundant" structure with few barriers to participation in the sector: (1) many hunters and rights holders; (2) many traders; (3) significant demand. Hunters, on average, obtain a higher profit than traders, who bear the highest costs of transportation, fines and bribes. Reducing unsustainable trade in this context, will necessarily imply reducing the burden on natural ecosystems as the main providers of animal protein. Moreover, supporting processes to re-structure local governance systems in this post-conflict context will also support efforts to reverse unsustainable use. The differences observed in Yangambi as compared to other well-studied wildmeat trade chains illustrates that no two bushmeat market chains are alike. Recommendations to reduce unsustainable trade in urban areas need to be tailored to specific contexts, taking into consideration differences in terms of whether markets are open or underground, the length of the trade chains (from local to international trade chains), the existence and type of barriers to entry, the number and type of stakeholders involved and the factors influencing the demand–supply equilibrium.
Keywords: wildmeat, market, trade chain, structure, Congo basin
## INTRODUCTION
Wildmeat is defined as non-domesticated terrestrial mammals, birds, reptiles and amphibians harvested for food. It is part of the diet in contemporary societies, particularly in tropical and subtropical areas, where it contributes to food security, nutritional diversity and personal well-being (Alves and van Vliet, 2017). While hunting for wildmeat often occurs primarily to satisfy the needs of the family, surplus meat is traded and the income is used to purchase other food items, invest in medical care, pay school fees and purchase non-necessities (Coad et al., 2010; Luskin et al., 2014; Endamana et al., 2016; Vasco and Sirén, 2016). Where a high demand exists and no other alternatives are available as sources of income, hunters may specialize in commercial hunting and sell most of the prey (van Vliet et al., 2015b; Greengrass, 2016; Mendonça et al., 2016).
The sale of wildmeat often occurs in the informal sector, either because the legal texts do not provide a stipulation to allow commercialization (Ruas et al., 2017), or because the regulatory framework presents contradictions or gaps (van Vliet et al., 2015a), or because legal texts, often inherited from colonial times (e.g., in Central African countries), no longer represent local realities (Sartoretto et al., 2017). Understanding the structure and operations of wildmeat value chains is now recognized as necessary to provide recommendations for a sustainable wildmeat sector, given the importance of wildmeat for local and national economies (CBD, 2012). As for any other marketable forest product, value chain analysis is important to infer recommendations for improving the business environment, the horizontal and vertical linkages between actors and the marketing issues, as well as ensuring ecological, economic and social sustainability (Te Velde et al., 2006).
Information on wildmeat consumption, as well as on the biomass traded in markets is increasingly available for the tropics, in Central Africa, West Africa and the Amazon (Fa, 2007; Parry et al., 2014; van Vliet et al., 2014, 2017a,b). What is drastically limited is our understanding of the main barriers to entry into the wildmeat business and the levers that can be used for a more sustainable sector. Only a limited number of studies have described the structure and operations of wildmeat value chains, the actors involved, the direction of flows and the economic value of wildmeat species traded (Cowlishaw et al., 2005; van Vliet et al., 2015b; Lescuyer and Nasi, 2016; Nielsen et al., 2016). These studies suggest that the sale of wildmeat generates significant revenues for different stakeholders, including the hunters, retailers and traders. However, the differences observed in each context, highlight for the need of more case studies to understand the general patterns in wildmeat trade chains and explain the differences observed in each context.
Our aim is to complement existing literature on wildmeat market chains, with a case study from the Yangambi landscape, in the Democratic Republic of Congo (DRC), a region characterized by a post-conflict situation. Using participant observation, semi-structured interviews of hunters, traders and consumers, participatory mapping and market monitoring, we analyze how wildmeat reaches its point of consumption from the point of extraction, and the implications of this for local economies, food security and ecological sustainability. Our study provides a comprehensive understanding of the structure and operations of the trade, the nature of the flows, the income generated by the wildmeat trade at the level of hunters and traders, the contribution to food security and the status and trends of the resource as perceived by the users.
### MATERIALS AND METHODS
### Study Site
Yangambi is a town located in the northeast of the DRC, about 100 km West of Kisangani City in Tshopo Province (**Figure 1**). As is typical in the Congo Basin, the landscape around Yangambi is characterized by a range of land tenures combining the Yangambi Man and Biosphere Reserve (YBR), created in 1979; the Ngazi Forest Reserve, which belongs to the Institut National des Etudes et Recherches Agronomiques (INERA); a logging concession; and customary land. In practice, due to the lack of human and financial resources, both reserves (Yangambi and Ngazi) have no official management plan, their limits are contested and they are not under any specific form of management.
The climate in our study region is marked by two dry seasons (from December to mid-March and from June to July) that alternate with two rainy seasons (from April to May and from August to November). The landscape is covered by old secondary forests, semi-deciduous dense forests, young secondary forests and dense evergreen forests. The rest is covered by a mosaic of agriculture, marshy forests and agroforestry systems. While several botanic surveys carried out since colonial times have provided a good understanding of the vegetation (Jacobsen et al., 2018), the only information on mammals comes from a recent assessment of hunters' perceptions (van Vliet et al., 2018a).
The human population living around the YBR is estimated at 141 643 inhabitants based on data from the Yangambi Registry Office dating from 2016. Yangambi was originally a research campus of INERA and IFA (Institut Facultaire de sciences Agronomiques) during colonial times, where only staff and their families could live, but over the years it became a town, due to the migration of workers and people searching for job opportunities in what became an economic hub for the area. The population around the YBR can be sub-divided into three groups: (1) the urban population (37,997 inhabitants) living in the 10 districts (Bangala, Ekutsu, Likango, Lomboto, Lumumba, Lusambila, Moussa, N'Gazi, Okito, and Yaekema) of the research campus of INERA and IFA (Institut Facultaire de sciences Agronomiques), which have evolved into a town; (2) the Turumbo and Topoke populations living in villages surrounding the reserve to the south, west and north along the unpaved trails/roads (Yambau, Yawenda, Yelongo, and Weko community groups); (3) The Bamanga (Bamanga Bengamisa and Bamanga Yambuya community groups) living toward the northeast of the reserve belonging to the Bamanga and Mba tribes.
Traditional agriculture, including cultivating cassava, banana, maize, rice, cowpeas, beans and groundnuts, is the main activity in all villages around the reserve and provides basic household livelihoods. The Bamanga population is more specialized in
agriculture (particularly rice and peanuts) than the Turumbo, who practice hunting as the second most important livelihood activity. Agriculture is a complementary activity for the urban population of Yangambi, where the majority of the residents are state employees (researchers, technical and administrative staff from the INERA and IFA research centers, the Tshopo Province Administration and the YBR that falls under the Ministry of Environment). Besides hunting and fishing, families also use many non-timber forest products for food and medicinal purposes and non-wood plant products for crafts and building materials. In the villages, households keep small livestock in extensive traditional production systems (poultry, pigs, ducks, goats, sheep), used to cover exceptional expenses, donations, dowries or to solve village conflicts.
During the last three decades, basic community infrastructure (roads, housing, educational and health facilities, etc.) has deteriorated significantly. The roads are in poor condition and basic necessities are supplied mainly by canoes on the Congo River. Health establishments are insufficiently equipped, most urban and rural households have no access to drinking water and the town of Yangambi is not electrified.
### Data Collection
The methodology applied in the context of this study as approved by CIFOR's Ethics committee is based on a combination of participatory methods including: participant observation, semistructured interviews and group discussions. The actors involved in this study all participated freely giving informed consent. The objectives of this study were introduced to the competent local authorities (INERA, IFA, UNIKIS (University of Kisangani), MAB (Man and Biosphere), sector chiefs, village chiefs) in order to obtain the necessary authorizations and the institutional support required to carry out fieldwork. The working team consisted of three people: one main coordinator and two research assistants. Field work was carried out from July 2017 to March 2018.
### Participant Observation
The first stage of this study was based on participant observation and informal interviews with different stakeholders. The research team, composed of two local researchers, spent a month staying overnight in villages surrounding the reserve, visiting the various places involved in the bushmeat trade, carrying out informal discussions, and observing the dynamics of the different stakeholders involved in consumption and trade. Information on market days was obtained for the different markets from the study area and visits were organized to observe wildmeat flows, relationships among the different stakeholders involved in wildmeat trade, client choices, means of transportation to and from the market, and number of traders, among others. Participatory mapping was used to locate the different source areas and the flows of wildmeat from rural areas to town. In each of the villages, discussions were held with the chief and key members of the community (as chosen by the chief) to identify the number of active hunters participating in trade, to understand customary rights over the resources and to locate the limits of the hunting grounds. Participant observation was not only used at the beginning of the study, it was embedded in the researcher's attitude and was used as a continuous approach to elucidate issues that were not immediately obvious and gain trust from stakeholders.
### Semi-structured Interviews
Semi-structured surveys were conducted with a representative sample of different key stakeholder groups (206 hunters, 127 traders and 632 household heads) to better understand their practices and their level of dependency on hunting, wildmeat trade and wildmeat consumption. In addition, hunters were also interviewed on the status and trends of mammal species as explained in more detail below. All interviews were administered using Kobocollect <sup>R</sup> , a data collection tool for Android phones.
### Semi-structured Interviews With Traders
All traders in each market were identified through informal discussions with market authorities, participant observations, market stalls counts and informal interviews with local informants. In each market, we sampled traders according to the following rules:
Sampled traders were chosen according to availability and willingness to participate in the interview. As such, a total of 127 traders (out of the 253 identified) were interviewed (78 women and 49 men), as shown in **Table 1**.
The aim of interviewing traders was to provide qualitative and quantitative information on the socio-economic background of the trader (gender, ethnicity, main income-generating activities), general information on the trade (number of years in the activity, relations with other actors, species marketed, places of sale and supply), motivations for trading, problems encountered, potential solutions, detailed information on biomass, prices, variable costs related to the sale (e.g., transportation costs) and fixed costs (e.g., taxes, depreciation of investments). This information was used to calculate the average net profit per trader according to the following formula:
$$\begin{aligned} \text{Average net profit/week/trader} &= \left(\sum \text{average revenue/week}\right) \\ &- \sum \text{average variable costs/week} \\ &- \sum \text{averagefixed costs/week}/\text{N} \end{aligned}$$
Profit-related calculations did not include time spent on the activity, since it is often difficult to quantify, particularly when traders spend time transporting the meat from the source area to town, with unpredictable transportation means, or when traders combine different commercial activities at the same time. Thus, the calculated net profits are based on remuneration for the work.
In order to quantify the commercial flows of wildmeat, we chose a sub-sample of traders (23 out of the 127 interviewed) with whom we had established trust and visited them once a month from September to December 2017 (covering the end of the dry season and the start of the rainy season), on market day, to monitor their activity. We know that trade may vary with season and other factors, so we tried to cover different seasons and include the two peak seasons paralleling cash needs (September for the start of the school year and December for the holidays), as well as two regular months (October and November). We focused our questions on the sale of mammal species, which generally make up the bulk of bushmeat traded in urban areas. Questions included species sold, hunting method, status of the meat, provenance, purchase price, and biomass per species. Biomass was measured using two balances: a mechanical balance of 100 kg for whole carcasses and a digital balance of 5 kg for pieces. Due to the difficulty in identifying certain taxa to species level in smoked specimens, some species were combined and recorded as generic groups (e.g., small diurnal monkeys, red duikers). The quantities of biomass sold by the sub-sample of traders were extrapolated to all traders in order to calculate the total biomass sold per month (minimum and maximum) and per year in Yangambi markets, as well as the economic value of the trade in the region.
### Semi-structured Interviews With Hunters
The total number of hunters in the area was obtained through discussions with village chiefs and other key stakeholders. The information was corroborated through informal discussions with different stakeholders. Only hunters who actively participated in the trade and for whom hunting represented one of the main activities were considered. A total of 206 hunters were surveyed out of 538 identified. To ensure spatial representativeness, the sample was geographically distributed among the different villages around YBR (see **Table 2**). The sample was chosen based on availability and willingness to participate in the interview.
The semi-structured interviews with hunters included two main sections: one section on hunting activity and one exploring wildlife abundance and trends as perceived by the hunters.
The first section of the interview provided information on the socio-economic background of the hunter (gender, ethnicity, main activities, number of years of hunting experience), hunting practices (hunting tools, hunting grounds), motivations for
TABLE 2 | Number of active hunters per village/neighborhood and number of hunters sampled in this study.
hunting, problems encountered, potential solutions, data needed to calculate the net profit (frequency of hunting, average quantities sold per hunting trip, variable costs related to each hunting trip (cartridges, batteries, food), and fixed costs (taxes or other, including depreciation of investments). We used these data to calculate the average net profit per hunter using the following formula:
$$\begin{aligned} \text{Average Net profit/week/hunter} &= \left(\sum \text{average revenue/week}\right) \\ &- \sum \text{average variable costs/week} \\ &- \sum \text{average fixed costs/week} / \text{N} \end{aligned}$$
As in the case of traders, profit-related calculations for hunters did not take into account the time spent on this activity, since it is often difficult to quantify. For example, when a hunter spends 4–5 full days in a hunting camp, he may spend his time traveling, resting, cooking, eating or doing other complementary activities (fishing/collecting), besides hunting. Thus, the calculated net profit is remuneration for work.
The second section of the interview used an ethno-zoological approach to analyze the abundance and distribution of hunted species, as well as the main factors that explain the observed trends, according to their perceptions. This section included questions about date and location of last observation of each species, observed population trends for each species in the last 30 years and factors explaining such trends. The geographical location of observations was based on a detailed map of the hunting ground covered by a grid in which each cell was identified by a letter and a number. The map of the hunting ground was produced prior to the interviews through a combination of participatory mapping and the geo-location of each of the landscape features using GPS and visits to the different locations (hunting camps, hunting trails, streams, rivers, etc.) (see van Vliet et al., 2018a, for more detail).
### Semi-structured Interviews With Consumers
The purpose of the household interviews was to understand the contribution of wildmeat to food security and the levels of wildmeat consumption in comparison with other sources of protein from the wild (fish, caterpillars, etc.) or of domestic origin (poultry, beef, goat, etc.). The semi-structured interview provided information on the socio-economic background of the household (gender of household head, ethnicity, main activities), perceptions of household food security, recall of meat types consumed in the last 24 h, dietary preferences in terms of animal sources of food, consumption patterns of wildmeat (frequency, form of supply, availability and prices). A sample of 632 households was chosen to represent the total population of Yangambi (town and villages surrounding the reserve). Our household interviews on the consumption of animal products were conducted in September 2017 and are therefore not necessarily representative of consumption throughout the year.
### Group Discussions
Discussions were organized with separate groups of traders and hunters to gain more qualitative insights into how the market chain operates, how the sector has changed over time and main factors of change.
One workshop was organized in Yangambi with traders from the main markets (35 participants) and 9 group discussions with hunters (10 participants per group) were organized in the main villages (Bandele, Bangala, Bossukulu, Lokeli, Lumumba, Ngazi, Weko, Yaliboto, Yaselia) surrounding YBR. Participants were selected based on their availability and willingness to participate; efforts were made to include a range of age groups.
The aim of the group discussion with the traders was to obtain supplementary information on the stakeholders involved in the trade, the geographical location of the flows, the economic importance of the sector (number of actors involved, volumes marketed, income generated globally) and the factors limiting or driving the sector. A historical trend analysis was developed covering the period from 1995 to 2017, highlighting changes that have occurred in the market chain and the main drivers of change (e.g., changes in wildlife habitat, climate-related changes, changes in infrastructure, demographic changes (displacement, emigration, migration), changes in governance, changes in local economies and changes in hunting techniques.
The aim of the group discussion with the hunters was to supplement information on hunting grounds, hunting practices, relationships with other stakeholders in the chain, seasonality, economic importance of hunting for their community (number of hunters involved, volumes hunted, income generated globally), and the limiting or driving factors influencing hunting. A historical trend analysis was developed covering the period from 1995 to 2017, highlighting changes that have occurred in hunting activity and the main drivers of change. Moreover, participants were invited to add to the information gathered through the interviews on changes in mammal populations across the study site and the main drivers of change.
## RESULTS
### Structure of the Market Chain
We identified 845 people involved in the bushmeat trade chain as their primary activity: 253 traders and brokers (153 women and 100 men), 539 hunters and 53 women, who participate in hunting trips with their husbands.
There are three main categories of actors actively involved in the trade (**Figure 2**):
they are encountered on the road to market. Some brokers rely on other women (often with family links to the hunters) to act as intermediaries between hunters and the brokers at the informal market. These intermediaries earn a small profit when they successfully facilitate trading between hunters and brokers. Some brokers establish relationships with the hunters and supply them with ammunition and other necessities. In fact, many brokers are also traders in the markets, as described below.
There are no restaurants in Yangambi. However, a few women sell food from market stalls and include bushmeat on their menu.
### Flows of Bushmeat
The town acts as the main hub for bushmeat demand in Yangambi. The volume of smoked bushmeat varies from 2,150 to 3,036 kg/week. These volumes, if extrapolated to the whole year, range from 103 to 145 tons. Until the last decade, bushmeat from the Yangambi area was taken as far as Kisangani or along the Congo river to Kinshasa and Equateur Province. However, with increased local demand from the town and the decrease in supply, the wildmeat trade chain is now limited to the immediate area. The main supply location for Yangambi brokers is Mipila market (about 25 km north of Yangambi), a spontaneous informal market. It is located in the forest close to Ngazi village, at a crossroads of different hunting trails in Weko Forest. The trails extend 25–50 km from Yangambi. The market has been active since 1978 and serves as a gathering point for hunters. Weko Forest supplies more than 66% of traders. Another supply area is located in Lileko, to the west of Yangambi toward Bassoko (about 30 km from Yangambi) and serves as a meeting point for brokers and hunters from Monganzo Forest. Bushmeat trade occurs throughout the year with peaks in August, September and December, when households need cash to pay school fees and for holidays. Hunters state that about 34 species are hunted. The most traded species are small monkeys (38% of the biomass) and red duikers (Cephalophus spp.) (31%), blue duikers (Philantomba monticola), bush pigs (Potamochoerus porcus), and bush tailed
porcupines (Atherurus africanus) (**Table 3**). Most meat is smoked and sold. The whole carcass of small species (monkeys, rodents and blue duikers) is smoked, while larger species are cut into pieces before being smoked.
### Income Generated by the Bushmeat Trade in Yangambi
Traders sell an average of 10 kg of bushmeat per week, with a margin of about USD 0.99 per kilo. Weekly costs were calculated at approximately USD 4.6. Thus, the weekly profit that the traders can make is equivalent to USD 7.3 per week, or USD 29.3 per month), which is about 1.6 times the official minimum wage (USD 17 per month). The main items of expenditure in the sale of game are various taxes and bribes. When the trader is also the broker, then transportation costs exceed all other costs. Income from the sale of game is used for the following expenses (in decreasing order): school, health, clothing and food. The majority of traders operate to provide for the basic needs of their family, given the lack of better job opportunities. Female-headed households often get involved in bushmeat trade due to the relatively low investment required compared to other businesses. However, the sale of game is complemented by other activities in the household, because it does not generate sufficient income for the subsistence of the family.
For hunters, income varies depending on luck and skill, but can range between USD 15 to 20 a week (or a multi-day hunting trip), for a net profit of about USD 40 per month (double the 2017 official minimum income in the DRC). Commercial hunters sell 80% of their meat. The most profitable months are October to December, during the rainy season, and the least profitable are January to March, during the dry season. Hunting income is used by men to cover for their own expenses (alcohol and cigarettes) and pay for school fees, clothing, health and food for the family. The main variable costs of hunting (in decreasing order) are the price of ammunition, cables (for snare hunters), food needed in the forest and batteries.
### Importance of Bushmeat for Food Security
Starchy foods, meat/fish and vegetables form the basis of the diets in the studied location. Households consume few sources of plant protein (lentils, beans, etc.), little or no sugar, no dairy products, and few or no fried foods. A significant proportion of households report not having consumed any animal product the day before the interviews (27%), while 32% of households consumed bushmeat, 18% consumed fish and 10% consumed caterpillars. Pork and chicken were only consumed by 7 and 2% of households, respectively, on the day before the interview. More than 60% of households eat bushmeat more than once a week and about 40% of households eat fish more than once a week. Thus, even if these data correspond only to the month of September, which would lead to a bias in the contribution of caterpillars, the contribution of wildmeat and fish is clearly significant throughout the year. The perception of food security varies between the rural Turumbo sector, where <4% of households consider themselves as having poor food security, and Yangambi, where more than 50% of households consider their food security as bad or very bad. In the rural Turumbo area, most households live on hunting, fishing, farming and small livestock rearing. In these villages, 91% of households obtain bushmeat by hunting themselves and 85% of them obtain fish by fishing themselves. The opposite trend found in the town of Yangambi, where a large majority of households depend on a monthly salary to purchase food, as they have limited access to land and resources.
TABLE 3 | List of species traded as wildmeat in Yangambi.
The most consumed fish are ngolo (Clarias sp.), sela (Labeo sp.), njombo (Protopterus sp.), ndakala (Stolothrissa tanganyicae/Limnothrissa miodon), and mpoto (Distichodus sp.). The most consumed bushmeat species are small monkeys, red duikers, bush tailed porcupine, blue duiker and bush pig. The most favored meats are bushmeat (fresh rather than smoked) and fish. If bushmeat and fish were to disappear or become too expensive, chicken and pork would be the two most suitable substitute proteins. In general, when families do not eat enough of their favored food, it is because it is seasonal or unavailable. Sometimes, selling prices also explain food choices. In Yangambi, fish costs almost twice as much as bushmeat (average price: 8.1 USD/kg for fish and 4.1 USD/kg for smoked bush meat). During the sampled season (end of harvest period), only smoked caterpillars were sold, attaining a very high price (13.6 USD/kg). Bushmeat remains the most affordable source of meat. Domestic
sources of meat (eg. pork, chicken) have limited availability in local shops or markets, due to the lack of electricity for proper preservation, and, when they are available, they are expensive.
### State of the Resource
At all levels of the sector, the availability of game seems to be the main barrier to entry into the system. At least 79% of traders find it difficult to obtain the quality of bushmeat they are looking for and 81% of sellers do not find enough. For 75% of traders, the sale of bushmeat has become more difficult in the last 10 years due to the lack of wildlife and 64% of them believe that their income has decreased. For 92% of consumers, access to bushmeat has reduced over the last 10 years. Over 88% of hunters consider hunting to have become more difficult over the last 10 years.
According to hunters, mammal abundance in the study area is characterized by a steady decline for all species. Three species are thought to have drastically declined: the okapi (Okapia johnstoni), red colobus (Piliocolobus badius), and chimpanzee (Pan troglodytes). Only small species with short gestation periods, such as hyrax (Dendrohyrax sp.), cane rat (Thryonomis swinderianus) and African giant squirrel (Protoxerus stangeri) are believed to be stable. Species that are nocturnal, have cryptic behavior or prefer habitats that are not easily accessible to hunters are more likely to persist in hunting areas. Five of the 34 hunted species (classified in one of the IUCN critical categories: near threatened, vulnerable, endangered, critically endangered) are vulnerable: chimpanzee (Pan troglodytes), okapi (Okapia johnstoni), giant pangolin (Smutia gigantea), otter (Aonyx congicus), and red colobus (Piliocolobus badius). The elephant (Loxodonta africana) is the only mammal species to have disappeared from the entire study landscape over the period from 1995 to 2018, the last sighting (direct and indirect observations combined) dates from 2007. While okapi (Okapia johnstoni), forest buffalo (Syncarus caffer nanus), and red colobus (Piliocolobus badius) are still present in the north of Ngazi, these species have become very rare or locally extinct from within the boundaries of YBR.
### Drivers of Change in the Wildmeat Value Chain
Political conflicts related to successive rebellions between 1996 and 2003 are presented as being the source of many cascading effects on the social, economic and ecological systems of the study area, with significant direct and indirect impacts on wildlife. During these periods of rebellion, soldiers of the Armed Forces of the Democratic Republic of Congo (FARDC), and other armed groups (Congolese, Rwandans and Ugandans from eastern DRC) hunted for meat, but were also involved in trafficking ivory, skins and meat. The uncontrolled use of wildlife resources by armed soldiers eroded local customary governance systems. In addition, the already weakened economy of the region was significantly affected by political instability and resulted in limited transportation means, reduced exchange networks and the closure of factories, which were the only source of stable employment in the region. With population growth and the lack of production/supply of meat from domestic animals, more families became dependent on forest resources for food security. As a result, the number of hunters involved in commercial hunting and the volumes traded have increased steadily over the last 20 years. Hunting practices have also changed to maximize harvest with both day and night hunting (due to the introduction of headlamps), multi-day hunting trips, and locally manufactured guns and cartridges. The number of traders has increased from 15 traders in the 1990s to more than 200 traders today. This is principally explained by the reduction in numbers of employees and the reduction of salaries in state-owned research and outreach centers, which previously represented the main source of employment in the region.
### DISCUSSION
Our study provides an example of a bushmeat trade chain in a medium-sized catchment area around the town of Yangambi. In the studied location, the trade contributes significantly to local food security and constitutes the most affordable and available source of animal protein, as also observed in Kisangani in 2002 (van Vliet et al., 2012) or in Bangui (Fargeot et al., 2017). As opposed to what has been observed in large cities, where bushmeat consumption is consumed as a delicacy or for specific cultural reasons rather than as a necessity (Wilkie et al., 2016; Luiselli et al., 2018), in medium sized towns, such as Yangambi, bushmeat seems to remain a key component for the food security of the poor.
The bushmeat trade chain in Yangambi has a "redundant" structure, which, according to the typology described by Phelps et al. (2016), refers to a structure where the number of stakeholders is very high at all levels of the chain. In Yangambi, the number of active hunters (N = 539), the number of traders (N = 252) are very high and more than 60% of households eat bushmeat more than once a week, keeping demand at the highest level. According to Phelps et al. (2016), redundant market chains often occur in contexts where there are few barriers to participation in the sector. In Yangambi, the bushmeat sector is poorly controlled by the State, as is the case throughout the Congo Basin (Fa, 2007; Lescuyer and Nasi, 2016). At the level of hunters, the resource is quasi-open access. Indeed, the armed conflict eroded local governance structures controlling access to resources. Hunting is not limited by access to firearms, as observed in Tanzania (Nielsen et al., 2016), since hunters rely on locally made firearms and ammunition. At the level of traders, the number of stakeholders involved is not limited by ethnicity as observed in Makokou, Gabon (Okouyi, 2006), as traders belonging to different ethnic groups may engage in trading. In Leticia, Colombia, van Vliet et al. (2018b) observed that the likelihood of any person to engage in trading wildmeat was low because traders required a well-established network in order to navigate the illegality of their activity. This does not seem the case in Yangambi, as the number of traders has been increasing steadily over the decades and many of them are new to the business. The number of consumers is not limited either, since the population is increasing and no other source of meat of domestic origin is competitive with bushmeat. In fact, the only barrier to entry in Yangambi, is the availability of wildlife resources. All stakeholders agree that wildlife is becoming scarce and some vulnerable species have almost been locally depleted. Despite the persistence of some emblematic species such as the chimpanzee, buffalo, okapi, red colobus and giant pangolin, the quasi-open access system observed in Yangambi, with high levels of dependency on the resource at all levels of the chain (from consumers to hunters) is likely to jeopardize the resilience of wildlife populations in the future.
Our study confirms observations by Brown and Williams (2003) in Ghana that hunting generates high profits, while requiring low investment and risks. As such, beyond the camaraderie of hunting trips, hunting is also attractive to young men living in rural areas for financial reasons. As observed by Nielsen et al. (2016) in Tanzania, hunters gain a higher profit than traders. In Yangambi, hunting may generate an income of about USD 400/year, an amount comparable to values provided by Lescuyer and Nasi (2016) for rural areas in Cameroon. In the case of Yangambi, it is actually the trade which is currently at the limit of profitability. Transportation costs reduce net profits for brokers and traders, given the poor state of the roads, which are only accessible by bicycle or motorbike most of the year. When they move to supply areas, brokers strive to maximize the amount of meat they get for the cheapest price possible, to cover their transportation costs. In contrast to the situation in Ghana (Cowlishaw et al., 2005) and Colombia (van Vliet et al., 2015b), but, in common with Tanzania (Nielsen et al., 2016), transportation costs are not borne by the hunters, but rather by the brokers and traders, as a result of demand outstripping supply.
Wildmeat trade in Yangambi illustrates the functioning of a medium-sized catchment area, involving relatively short market chains. We hypothesize that these short market chains in emerging towns from the Congo Basin are becoming more common, as wildlife harvests no longer generate surplus to supply more distant towns and local demand increases with rising human population in those secondary urban areas. Transportation costs hinder the profitability of wildmeat trade to distant areas. In those distant towns, as wildmeat becomes scarcer, prices of this commodity rise and become less competitive as compared to other sources of meat available, and consumption of wildmeat becomes a luxury rather than a necessity. Particular attention should be given to understanding market dynamics in those emerging consumption hubs where both urban food security and ecological sustainability are at stake.
The differences observed in available market chain studies, call for tailored approaches to each context to reduce unsustainable trade to urban areas. Initiatives targeting the hunters will have little impact in a context where the profits generated by hunting are attractive as compared to other sources of livelihoods. While behavior change campaigns, as suggested by Chaves et al. (2018), might work in some contexts, we believe that patterns
### REFERENCES
of bushmeat consumption in Yangambi will not change until alternative sources of food and income for a growing population can be established. Where the dependency on the resource is so intrinsically linked to people's basic needs, there are little chances that behavioral change campaigns will have a significant impact. Strict law enforcement, although necessary, will not stop illegal behaviors, but may rather fuel retaliatory killing (Soliku and Schraml, 2018), particularly where local livelihoods have continued to expand over the protected area's territory in the absence of any sort of management for decades. In contrast, if substitutes become more available, as already observed in the neighboring town of Kisangani (van Vliet et al., 2017b), the amounts of bushmeat traded may decrease over time. With sustained peace and a structured economy to guarantee steady incomes, maintain exchange networks and incentivize investment in livestock production systems, consumers may find other alternatives to secure their nutrition. A process that supports the re-structuring of local governance systems in a postconflict context will also help reverse unsustainable use in the long run.
### AUTHOR CONTRIBUTIONS
NV developed the methodology of the research, coordinated field activities, analyzed the data and wrote the initial draft of this manuscript. JN and JM carried out the field work and contributed to data analysis. RN provided guidance all along the process and participated in the writing of the manuscript.
### FUNDING
We are grateful for the financial support from USAID and the European Union, through the Bushmeat Research Initiative from the Forest, Trees and Agroforestry Program of the CGIAR.
income of remote forest communities in Central Africa. Int. Forest. Rev. 18, 280–295. doi: 10.1505/146554816819501682
plantation-dominated landscapes of Sumatra, Indonesia. Hum. Ecol. 42, 35–45. doi: 10.1007/s10745-013-9606-8
**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Copyright © 2019 Van Vliet, Muhindo, Nyumu and Nasi. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
## Price Elasticity of Bushmeat Demand in the Greater Serengeti Ecosystem: Insights for Managing the Bushmeat Trade
Solomon Zena Walelign1,2 \*, Martin Reinhardt Nielsen<sup>1</sup> and Jette Bredhal Jakobsen1,3
*<sup>1</sup> Department of Food and Resource Economics, University of Copenhagen, Frederiksberg, Denmark, <sup>2</sup> School of Economics, College of Business and Economics, University of Gondar, Gondar, Ethiopia, <sup>3</sup> Center for Microecology, Evolution and Climate, University of Copenhagen, Copenhagen, Denmark*
Rural households across the tropics rely on bushmeat hunting to fulfill their subsistence and cash income needs. As human populations grow, and urban market demand drives commercial trade, hunting is often unsustainable, compromising community long-term food security and wildlife conservation objectives. Scarce information about the effectiveness of different intervention options hampers design of informed management strategies to reduce bushmeat hunting while simultaneously safeguarding community's food security. Here we examine the potential of interventions aimed at reducing bushmeat demand by evaluating the own- and cross-price elasticities, i.e., how consumers respond to changes in the price of bushmeat and the price of five substitutes—beef, chicken, lamb, goat, and fish. We conducted stated preference surveys, complemented by a socio-economic survey using the Poverty Environment Network protocol in 452 households in 21 villages in the Greater Serengeti Ecosystem in Tanzania. Using random intercept Poisson regression models, we find significant and elastic negative own-price elasticities of bushmeat demand and significant positive cross-price elasticities except for goat and fish. The significant (all at the 0.01 level) own-price elasticities ranges from −1.099 when bushmeat is paired with beef to −0.718 when bushmeat is paired with fish while the significant cross-price elasticities ranges from 0.128 when bushmeat is paired with beef to 0.590 when bushmeat is paired with lamb suggesting that most cross-price relations were inelastic. Variation between districts was considerable and depended on substitutes included in the model. Estimated elasticities were modified by socio-economic covariates including ethnicity, household size, household income, household Tropical Livestock Units ownership, household land ownership and distance to nearest protected area boundary, Lake Victoria and nearest road. Overall, we find mixed support for the hypothesis that interventions increasing the price of bushmeat and decreasing that of its substitutes will reduce bushmeat demand. The effectiveness of demand reducing interventions should increase if complemented by other policy interventions, e.g., interventions that increase the opportunity cost of hunting, by providing alternative income generation opportunities for hunters.
Keywords: bushmeat demand, preference experiment, price elasticities, Greater Serengeti Ecosystem, demand side policy
#### Edited by:
*Nathalie Van Vliet, Center for International Forestry Research, Indonesia*
#### Reviewed by:
*Laura Vang Rasmussen, University of British Columbia, Canada Willandia Chaves, Woodrow Wilson School of Public and International Affairs, Princeton University, United States*
\*Correspondence:
*Solomon Zena Walelign szw@ifro.ku.dk*
#### Specialty section:
*This article was submitted to Conservation, a section of the journal Frontiers in Ecology and Evolution*
> Received: *31 October 2018* Accepted: *24 April 2019* Published: *15 May 2019*
#### Citation:
*Walelign SZ, Nielsen MR and Jakobsen JB (2019) Price Elasticity of Bushmeat Demand in the Greater Serengeti Ecosystem: Insights for Managing the Bushmeat Trade. Front. Ecol. Evol. 7:162. doi: 10.3389/fevo.2019.00162*
## INTRODUCTION
Rural households across the tropics and sub-tropics rely on bushmeat hunting for subsistence and to generate cash income (Nielsen et al., 2017, 2018). However, bushmeat hunting is in many locations unsustainable (Dirzo et al., 2014; Ripple et al., 2016a; Benítez-López et al., 2017). Human population growth, technological advancement of hunting equipment and improved access to transport is driving a commercial bushmeat trade supplying urban centers of demand (Bennett and Robinson, 1999; Cawthorn and Hoffman, 2015). The resulting depletion of wildlife populations threatens both local food security and biodiversity conservation across the tropics (Harrison, 2011; Lindsey et al., 2013; Cawthorn and Hoffman, 2015; Ripple et al., 2016b). Hunting is often illegal, unregulated and unreported and most protected areas in the tropics are affected to some extent by bushmeat hunting (Schipper et al., 2008; Jones et al., 2018; Schulze et al., 2018). Hence, appropriate interventions are necessary to reduce illegal bushmeat hunting while safeguarding rural communities food security.
Interventions aiming to reduce illegal bushmeat hunting can target the supply side (i.e., hunters and other actors in the bushmeat market value chain), by increasing law enforcement or providing alternative livelihood opportunities for hunters (Moro et al., 2013; Nielsen et al., 2014). Interventions can alternatively target the demand side (i.e., consumers), by changing the purchasing habits of consumers by affecting the price of bushmeat and its substitutes to reduce demand (Rentsch and Damon, 2013). Evidence on the effect of demand-side interventions is scarce and inconclusive (Wilkie et al., 2005; van Velden et al., 2018; Veríssimo et al., 2018). Existing empirical evidence on price effects is ambiguous and appears highly context-dependent but tend to show that a price increase of bushmeat leads to decreased household bushmeat consumption while a price increase of substitute meat products leads to increased bushmeat consumption (Wilkie and Godoy, 2001; Wilkie et al., 2005; Fa and Brown, 2009; Foerster et al., 2012; Rentsch and Damon, 2013; Moro et al., 2015).
A considerable number of studies have examined bushmeat hunting in the Greater Serengeti Ecosystem (GSE) and found detrimental effects on wildlife populations including in Serengeti National Park (SNP) (Setsaas et al., 2007; Marealle et al., 2010; Strauss et al., 2015). Various interventions have been implemented to reduce bushmeat hunting in the SNP. These include strengthening law enforcement capacity and enhancing wildlife conservation awareness, promoting alternative protein and income sources (e.g., wage-earning activities), and providing veterinary care for domestic animals as well as community-level benefits including schools and health dispensaries (Moro et al., 2013, 2015; Rentsch and Damon, 2013; Kaaya and Chapman, 2017). Despite these interventions, bushmeat hunting persists and is expected to increase in the future (Rentsch and Packer, 2014) due to population growth and infrastructure development increasing market access (Dobson et al., 2010). In general, interventions aim to reduce bushmeat consumption by (i) raising local household income (i.e., assuming that bushmeat is an inferior good), (ii) increasing the price of bushmeat relative to its substitutes (i.e., other meat products), and (iii) increasing the opportunity cost of hunting. These interventions all affect the real price of bushmeat and its substitutes in one way or another. However, interventions have often been designed with a limited understanding of the effect of the price of bushmeat and its substitutes on bushmeat demand, which may explain the limited impact of interventions. Existing studies of the elasticity of bushmeat demand in the GSE have been geographically restricted to Western Serengeti and focused on a few substitute meat products (beef, fish, and daaga or chicken and fish) (Rentsch and Damon, 2013; Moro et al., 2015). However, determining how bushmeat demand responds to price changes and change in the price of a broader range of its substitutes across the wider GSE is essential to evaluate the heterogeneity in the likely effect on bushmeat hunting (Moro et al., 2015). Spatial heterogeneity may occur for example due to distance-induced differences in the availability of bushmeat and its substitutes and locationspecific culturally determined differences in the acceptability of substitutes.
Economic theory suggests that change in the price of bushmeat can affect bushmeat consumption in two different ways: (i) increased bushmeat price reduce bushmeat demand and vice versa (the law of demand), and (ii) increased bushmeat price increase bushmeat demand and vice versa (Giffen goods hypothesis) (Varian, 2010). The latter is more hypothetical but could occur because bushmeat remains cheaper (relative to other meat types) even if its price increases and people consume more bushmeat at the expense of more expensive substitutes to compensate for lost income due to the increased price of bushmeat (i.e., the income effect outweighs the substitution effect). The relationship between bushmeat demand and its price is measured by own-price elasticity indicating the responsiveness of bushmeat demand to change in its price. In addition, change in the price of substitute protein sources can affect bushmeat demand in two ways: (i) increased substitute price increase bushmeat demand and vice versa (substitute good hypothesis) and (ii) increased substitute price reduce bushmeat demand and vice versa (complementary goods hypothesis) (Varian, 2010). The latter would occur if the substitute meat types were consumed together with bushmeat, e.g., for cultural or culinary reasons which is not the case here. The relationship between bushmeat demand and the price of its substitutes (or complements) are measured through the cross-price elasticity of bushmeat demand indicating the responsiveness of bushmeat demand to changes in prices of other (meat) products.
Evaluating household level price elasticities of bushmeat demand requires as a minimum information about bushmeat demand and price. This data can be obtained through revealed preferences in conventional household surveys—i.e., observing actual consumption over time (e.g., Rentsch and Damon, 2013). However, the revealed preference approach has limitations in its application to bushmeat research for at least two reasons: (i) the illegal nature of bushmeat supply and resulting fear of repercussions, may cause households to withhold or provide incorrect information (Nuno et al., 2013) and (ii) observed prices might not change markedly during the survey period and hence not include sufficient variation to support policy development (Moro et al., 2015). Stated preference methods have shown potential to overcome problems affecting data quality when asking sensitive questions in conservation research and are thus increasingly used (e.g., Moro et al., 2013; Nielsen et al., 2014). A stated preference approach also enables generation of bushmeat demand and price data beyond the current market situation enabling predictions about the likely response of bushmeat consumers if prices change significantly in the near future. Such predictions can provide relevant insights about bushmeat market development for informed policymaking. We, therefore, used a stated preference survey to generate data about bushmeat demand in response to price. The generated data hence represents stated rather than actual market demand.
Only one study in the GSE by Moro et al. (2015) has employed a stated preference survey to assess the price elasticity of bushmeat demand. Their study was geographically limited to Western Serengeti and assessed the price effects of two potential substitutes (chicken and fish) making results less representative across the ecosystem and for all available meat types. However, a broader insight into the effects of price change is needed to support conservation intervention across the GSE that encompasses considerable cultural and livelihood strategy diversity. Here we aim to bridge this information gap by assessing the own- and cross-price elasticity of bushmeat demand in five districts bordering protected areas around the GSE (452 households in 21 communities) including the effect of five available meat substitutes (fish, lamb, beef, chicken, and goat) and considering the effect of socio-economic (e.g., income, livestock ownership) and location (district) covariates. We also test the effect of spatial covariates including minimum distance to protected area boundaries, Lake Victoria and the nearest road as reflections of access to bushmeat and availability of substitutes.
## METHODS
### Study Area
The GSE is dominated by plains hosting the last remaining great wildlife migration consisting of wildebeests (Connochaetes),
zebras (Equus quagga) and Thomson's gazelles (Eudorcas thomsonii) tracking the availability of grazing across the Ecosystem. Both the national economy and local households in the GSE depend on tourism revenue and direct use of natural resources in the GSE. Over two million people live in the nearest seven districts (Kaltenborn et al., 2011) originating from more than 27 different ethnic groups. The most prevalent tribes are Maasai, Sukuma, and Kurya. Poverty levels are high and economic development is constrained by low agricultural productivity and market access and restrictions on land utilization imposed by the protected areas (Fyumagwa et al., 2017). The human population is growing at an alarming pace increasing pressure on the ecosystem to meet the demand for food, fuelwood, construction material, water, and land (Dybas, 2011; Estes et al., 2012). Protected areas in the GSE with different level of protection includes Serengeti National Park, Ikorongo, Maswa, and Grumeti Game Reserves, Ngorongoro Conservation Area and Loilondo Game Controlled Area (see **Figure 1**). Consumptive activities (e.g., settlement, agriculture, livestock grazing, environmental goods collection, and bushmeat hunting) are not allowed in the first four areas while regulated pastoralism is permitted in the latter two areas. However, illegal bushmeat hunting is widespread across the GSE. Numerous studies have attempted to estimate the prevalence of bushmeat hunting but are constrained by the secretive nature of this trade. Hence, estimates vary considerably across studies ranging over 32% (Loibooki et al., 2002), 9% (Knapp et al., 2010) and 18% (Nuno et al., 2013) of households in Western Serengeti—the latter estimate based on the unmatched-count technique. Rentsch and Packer (2014) estimate an annual offtake of 97,796–140,615 wildebeests threatening conservation objectives.
### Data Collection
The sampling unit is the household defined as a group of people (both family and non-family members) living under the same roof sharing labor and income (PEN, 2007). Households were selected using a three-stage stratified sampling strategy. In the first stage, 21 villages in the five districts, Meatu, Bariadi, Serengeti, Tarime, and Ngorongoro districts (see **Figure 1**) were purposively selected to encompass the variation in biophysical, socio-economic and administrative characteristics of the GSE. The districts differed markedly in precipitation, soil characteristics, human population density, ethnic composition and level of development (**Table 1**).
Communities were selected in clusters of three villages within each district at an increasing distance from the nearest protected area boundary. In the second stage, forty households were selected in each community (a total of 840 households) using stratified random sampling based on participatory wealth ranking (Grandin, 1988) of all households residing in the community according to an updated village register. Wealth ranking was conducted by a focus group consisting of 6–8 community members knowledgeable about local affairs grouping all households into three wealth categories (rich, intermediate, and poor) based on locally relevant and agreed criteria. From this sample 10, 20, and 10 households were randomly selected from the rich, intermediate and poor group, respectively, along with a contingency sample of three wealthy, four intermediate, and three poor households, which was used as a replacement in case of attrition. All households in the sample were subjected to four quarterly household surveys each producing detailed records of all cash and subsistence income the past 1 or 3 months depending on income source following the Poverty Environment Network questionnaire protocol (see Jiao et al., 2018). In the third stage, a sub-sample of approximately 21 households per village was presented with the stated preference survey. The choice of a sub-sample was based on cost, time and respondent fatigue implications in relation to other survey objectives (Walelign et al., 2019). The final sample in the analysis of own and cross-price elasticities consists of 452 households. Sample weights reflecting the inverse of the probability that a household was included based on the sampling strategy and the wealth rank distribution in the village was used in the analysis.
### Designing the Choice Experiments
Development of the experimental design followed a threestep process. First, Focus Group Discussions (FGDs) were conducted in one village in each of four districts representative of the diversity of the GSE to understand meat consumption behavior, identify substitute protein sources, and determine price variations of bushmeat and the substitutes. The four villages are Mwamtimba, Gibaso, Oloipiri and Nyiberekera-Isenye in Bariadi, Tarime, Ngorongoro and Serengeti districts, respectively. Households in the GSE consume a wide variety of protein sources including various bushmeat species, beef, goat, lamb, chicken, larger fish that typically are tilapia, dagaa (small dried fish), milk, and eggs as well as vegetable-based protein sources. Milk, egg and vegetables are typically from own production or can be sourced locally very cheap (Lowassa et al., 2012). Dagaa is similarly, frequently consumed and very cheap. These goods are therefore not seen as substitutes for meat or larger fish, and policy intervention aiming to change the price of these goods may thus not significantly affect bushmeat consumption. We, therefore, focused on five meat types namely beef, lamb, goat, chicken, and larger fish the price of which is considered more likely to influence bushmeat demand and therefore constitutes better targets for policy interventions. FGDs aimed to establish common units of measurement across meat types in the level of processing (raw or sundried, with or without bones) and part often traded. Visual aids were developed consisting of pictures showing selected meat types, units and their size to ensure respondents common frame of reference. Three prices were obtained for each meat types: (1) the prices at which the household would consume the meat type as the main protein source (hereafter the minimum price) keeping the price of other meat sources at the current level, (2) the prevailing current market price (hereafter current price), and (3) the expected price 10 years from now assuming prevailing inflation rates due to population growth and resulting resource scarcity and as indicated by the focus groups (hereafter maximum price; **Table 2**).
Secondly, the price of bushmeat was paired with the price of one substitute at a time to develop a total of five experiments: bushmeat vs. beef, bushmeat vs. chicken, bushmeat vs. lamb,
*Source: Project documents available in the AfricanBioServices data repository.*
TABLE 2 | List of substitute meat types and characteristics i.e. measurement unit, description, minimum, current, and maximum prices in TSH.
*Values in square brackets are the range of the values mentioned during FGDs.*
bushmeat vs. goat meat and bushmeat vs. fish. This design was selected to reduce the cognitive burden on respondents as a design combining all meat types proved challenging to comprehend for the respondents during pre-testing. Each household was subjected to a stated preference questionnaire with two randomly selected meat types. A full factorial design (using the three prices; the minimum, current and maximum prices) was generated for each experiment resulting in nine combinations (3<sup>2</sup> ). The nine combinations were divided into three random blocks, and each respondent was randomly presented with two blocks representing two different meat types—i.e., each respondent was presented with twelve choice tasks (six from each experiment; see **Figure 2** for an example of a choice card).
Thirdly, the questionnaire was pilot tested in October 2016 in 64 households not part of the final sample to improve the clarity of questionnaires and update the description of scenarios with further information relevant for households demand decision and to make scenarios as credible as possible to respondents in accordance with Johnston et al. (2017). The generated choicesets were posed as an open-ended choice experiment allowing respondents to provide an answer about the quantity of meat demanded at a given combination of prices. Thus, while the approach resembles designs from the discrete choice experiment literature, the data generated are continuous in the form of a count. The questions were asked in an interview-based questionnaire survey preceded by an introductory explanation given by enumerators. The explanation included a cheap talk script and a budget reminder to minimize bias arising from the hypothetical nature of the experiments (Tonsor and Shupp, 2011; Ladenburg and Olsen, 2014). Follow-up questions were included to determine whether the respondents attend to all meat type prices presented to them (**SM1** in Supplementary Materials).
The data was collected between November 2016 and February 2017 using an ODK tablet interface, which enabled real-time data entry, and facilitated showing respondents pictures of meat types, units and their magnitudes to ensure common frame of reference (cf. above). The tablets also enabled presenting videos introducing the choice experiment. Interviews targeted the household head along with the wife (if the head was a man) whenever possible. In the rare cases where the household head was absent (estimated to be 1% of the households), the wife (if the head was a man) or the senior female household member (if the head was a woman) were interviewed alone as we believe that these individuals are in the best position to know and make decisions about meat demand on behalf of the household.
### Data Analysis
As the demand was measured as the number of pieces purchased at a given price, the nature of the outcome variable is a count. We, therefore, employed Poisson models to examine the effect of price and other covariates on bushmeat demand. The basic
Poisson regression model involves an equidispersion assumption requiring that the mean and the variance are equal (Wooldridge, 2002; Cameron and Trivedi, 2005). However, this assumption is often violated, and researchers, therefore, use alternative specifications of the general model including the negative binomial count model (Greene, 2008) and mixed effect Poisson regression models (Rabe-Hesketh and Skrondal, 2012). Mixed effect models have the advantage of being able to accommodate within-group covariance originating from the nested structure of the sampling strategy, sampling at the district and village level and each household performing several choice tasks. However, (i) the intracluster correlation (ICC) for the district level random intercept was <10%—a commonly used threshold for including random intercepts and slopes in mixed effect models (Rabe-Hesketh and Skrondal, 2012) and (ii) for the village level random intercept, either ICC was <10%, or its inclusion did not improve model performance (See **SM2** in Supplementary Materials). We, therefore, used mixed effect Poisson models with household ID as the only random intercept. Following Rabe-Hesketh and Skrondal (2012), the expected number of pieces of bushmeat demanded, λij for the i ′ th choice situation in the j ′ th household can be estimated by:
$$\lambda\_{ij} = \exp\left(\alpha + P\_{ij}\beta\_p + P\_{ij}X\_{ij}\beta\_{px} + \left(P\_{ij}D\_{ij}\right)\beta\_{pd} + \gamma\_j\right) \tag{1}$$
where, γ<sup>j</sup> is the household level intercept with a random distribution with zero mean and constant variance. Pij is a vector of the logarithmic transformed price of meat types, which implies that the average marginal effects can be interpreted as elasticities reflecting the proportional change in bushmeat demand as a result of a 1% change in the price of bushmeat or its substitute, keeping other covariates constant. PijXij is an interaction between the vector of meat type prices and a vector of socio-economic covariates. PijDij is an interaction between the vector of meat type prices and a vector of spatial variables including effect coded variables representing district. The covariates enter the model through interaction with prices, as the covariates were not included as attributes in the stated preference design. Hence, the coefficient of the interaction terms (i.e., βpx or βpd) reflects differences in responsiveness to price between socio-economic groups or locations. The average marginal effects were estimated as the average of the marginal effects for each observation which in turn were estimated as the product of the predicted values based on the model in Equation 1 and βp.
We considered different specifications of Equation 1: simple models and extended models with socio-economic and spatial covariates. This was guided by the aim of estimating the price elasticities from the simple model and determine how different covariates modify the estimated elasticities. Due to multicollinearity between the co-variates, we tested sets of covariates in different models. This implies that where multicollinearity is high (we used Variance Inflation Factor (VIF) above 10 as a threshold), we may end up with a less efficient model and exagerate or underestimate the effect of a given parameter (depending on sign). We therefore started with simple models where bushmeat demand was modeled as a function of the logarithmically transformed own-price and substitute price for individual meat types. We then extend the simple model by including socio-economic covariates. The socio-economic covariates were selected based on general theory about determinants of demand and previous empirical findings (see **SM3** in Supplementary Materials for a description of the covariates). Selected covariates include: livestock possession measured in Tropical Livestock Units (TLUs), land ownership in acre, ethnic group affiliation, household size measured in Adult Equivalent Units (AEUs), total household income in PPP converted USD, and household preference for the relevant meat type measured as stated attendance to the prices in the experiments<sup>1</sup> . To assess differences between districts, we also extend the simple model by including an effect coded district variable (in place of socio-economic covariates) so that the effect of each district can be interpreted by comparison to all other districts and not just a single reference district (Gupta, 2008). Since respondents were presented with two meat type experiments, we controlled for the order in which the experiment under consideration was presented. Control for the order was included in all three versions of the model (i.e., simple model, model with socioeconomic covariates, and the model with district variables). We did not include the demand for substitute products as independent variables in the models for two main reasons. This includes the assumption that prices are the main determinants of demand for products rather than the demand for its substitutes. And the fact that because the demand for bushmeat and its substitutes are determined simultaneously, inclusion of demand for the substitutes could entail endogeneity bias.
Spatial variables reflecting the minimum distance to the protected area boundary, to Lake Victoria and a road, were not included in models mentioned above due to multicollinearity
<sup>1</sup>Using a wealth index was not considered as it would not allow us to assess the importance of different asset types. Furthermore, including income may potentially impose an endogeneity issue. Although we cannot rule out an effect of income, we tried to avoid it by explicitly telling respondents to consider income constraints in their choices.
(i.e., mean VIF = 2902.26 and mean VIF = 2.68, with and without the distance variables, respectively, in the model extended with district variables). The effect of these variables on the elasticities was instead estimated through extracting the average marginal effect for significant elasticities for each household from the models with socio-economic covariates averaged across choice cards and regressed against the spatial variables using Ordinary Least Squares (OLS) regression. We furthermore explored non-linearity by including the level and squared terms of distance variables in the OLS models. The order in which the meat experiment under consideration was presented, was also controlled for in these models. Distances were determined as the minimum Euclidian distance. We preferred distance as a proxy for access to and availability of bushmeat and its substitutes over other spatial variables (e.g., forest cover), for two main reasons. First, the GSE is dominated by grasslands and the use of forest cover does not represent wildlife densities. Second, closeness to the source of meat and fish determines access to and availability by reducing transport costs and risk of being caught while transporting the meat in the case of bushmeat. In addition, the empirical literature on the availability of bushmeat in the GSE and elsewhere suggests that distance to protected areas is an appropriate measure of the availability of bushmeat (e.g., Brashares et al., 2011; Nuno et al., 2013).
### RESULTS
The results of the simple models, including the price of bushmeat and its substitute only, are presented in **Table 3** reflecting the sign and statistical significance of the own- and cross-price elasticities for bushmeat and individual substitutes separately. The coefficients reflect the sign and statistical significance of the attributes and the average marginal effects can be interpreted as the magnitude of elasticities reflecting the proportional change in bushmeat demand as a result of a 1% change in the price of bushmeat or its substitute, keeping other things constant. As expected, the coefficient of bushmeat price is negative and significant indicating that the desired amount of bushmeat decreases as the price increases (i.e., it has a negative own-price elasticity). The own-price elasticities represented by the average marginal effects for the statistically significant coefficients range from −1.099, when bushmeat is paired with beef, to −0.718 when bushmeat is paired with fish. The beta coefficient for the price of substitutes is as expected positive and significant (beef at the 0.1 level) indicating positive cross-price elasticities except for goat and fish where it is insignificant. The estimated average marginal effects reflecting the magnitude of cross-price elasticities range from 0.128 when bushmeat is paired with beef to 0.590 when bushmeat is paired with lamb.
In summary, bushmeat demand is largely inelastic with respect to both its own price and the price of its substitutes in the simple model meaning that 1 percent increase in the price of bushmeat or its substitutes leads to <1 percent decrease and increase in bushmeat demand, respectively. The only exception is beef where one percent increase in the price of bushmeat leads to slightly above one percent decrease in bushmeat demand.
The models controlling for the effect of socioeconomic covariates are presented in **Table 4**. The own-price elasticity of bushmeat demand increase (numerically) when controlling for socioeconomic characteristics to the extent that average marginal effects become elastic for all meat types except goat. This implies that if bushmeat price increases by 1% it leads to more than 1% decrease in bushmeat demand when the substitute is beef, chicken, lamb and fish. However, cross-price elasticities were statistically insignificant except for fish where bushmeat demand was inelastic to change in fish price. These results indicate that the inelastic feature from **Table 3**, may be caused by heterogeneity in socioeconomic groups as it applies to cross-price elasticity.
Few socioeconomic covariates had significant effects and these varied between models depending on substitutes. Bushmeat demand by higher income households was significantly less responsive to the price of substitutes when the substitutes were chicken and fish only. In other word bushmeat demand by higher-income households increased less than by poorer households as the price of substitutes increased. Households with high TLUs were significantly (at the 0.1 level) less responsive to bushmeat price when the substitutes were chicken and lamb and less responsive to substitute price when the substitute was fish. Land rich households were more responsive to bushmeat price when the substitute was goat and less responsive to substitute price when the substitute was beef. Finally, larger households were less responsive to bushmeat price when the substitutes were beef (at the 0.1 level) and fish and less responsive to substitute price when the substitute was fish. Overall, the results reveal that larger households are less responsive to own-price whereas wealthier households measured in TLU and land are less and more responsive, respectively. Larger households, more income rich households and households wealthier in TLUs, and land were less responsive to substitute price.
Maasai household's bushmeat demand was less responsive to bushmeat price when the substitutes were chicken and fish and more responsive to substitute price when the substitutes were goat (at the 0.1 level) and fish (**Table 4**). Bushmeat demand by the Sukuma was less responsive to bushmeat price when the substitute was beef and less responsive to substitute price when the substitutes were beef, goat, and fish. Bushmeat demand by the Kuria was less responsive to the price of bushmeat when the substitute was lamb, and it was irresponsive to the price of any of the substitutes. Attendance to the price of meat types in making the demand decision differed between ethnic groups. Maasai households had the largest proportion of any tribe stating not attending to the price for all meat types (**SM4** in Supplementary Materials). The model presented in **Table 4** included a variable controlling for attendance to bushmeat price and the price of the relevant substitute. Households who do not attend to bushmeat and substitute price are less responsive to change in both bushmeat price when the substitute was beef and price of substitutes when the substitute was chicken.
The results of models exploring differences between districts are presented in **Table 5**. The own-price elasticity of bushmeat demand was higher in Serengeti district compared to other districts when the substitute was chicken while it was lower in Bariadi and Tarime districts when the substitute was fish.
TABLE 3 | Coefficients of the simple mixed effect Poisson models representing the own and cross-price elasticities of bushmeat demand for individual substitute meat types.
*Values in parenthesis are standard errors. Order effect of meat experiment was controlled for in the model—not shown.*
\*\*\**,* \*\* *and* \* *reflects significance at the 0.01, 0.05, and 0.1% level respectively.*
Cross-price elasticities were higher in Serengeti district for beef, chicken and fish and in Bariadi district it was lower for beef, goat and fish. In Meatu district the cross-price elasticity was lower for beef and fish. In these models, eastern GSE districts function as a baseline. However, if setting all other districts as a baseline, we find low own-price elasticities in Loliondo Game Controlled Area (LGCA) when the substitute is chicken and goat and in Ngorongoro Conservation Area (NCA) when the substitute is chicken (at the 0.1 level) and lamb (**SM5** in Supplementary Materials). Cross-price elasticities were lower in NCA for chicken (at the 0.1 level) and higher for goat compared to other districts. These results indicate that demand responsiveness to price varies considerably between districts depending on substitutes.
We included distance to protected area boundaries and to Lake Victoria as variables in an ex-post OLS regression of elasticities as indicators of access to bushmeat and availability of substitutes, mainly fish. This approach was selected over including distance variables directly in the estimation due to multicollinearity. The results presented in **Table 6** reveal that the own-price elasticity of bushmeat demand is positively associated with distance to the nearest protected area boundary and negatively associated with the distance to Lake Victoria when the substitute was beef (based on predictions of elasticities presented in **Table 4** controlling for socioeconomic covariates). Hence responsiveness to bushmeat price is lower further from protected areas but higher further from Lake Victoria and it appears that distance to Lake Victoria exerts a higher impact on the own-price elasticity. The squared terms of these distances were also significantly positive and negative, respectively, meaning that the observed effects increase at an increasing rate as distance increase. When the substitute was chicken, the ownprice elasticity of bushmeat demand was significantly positively associated with both the level and squared terms of distance to Lake Victoria indicating that the own-price elasticity of bushmeat demand decrease with distance to Lake Victoria at an increasing rate. The own-price elasticity of bushmeat demand decreased significantly and linearly with distances to the nearest road and Lake Victoria when the substitutes were lamb and fish, respectively. Fish was the only substitute for which bushmeat demand had significant cross-price elasticity when controlling for socioeconomic covariates (i.e., **Table 4**). The fish cross-price elasticity was significantly negatively associated with distance to Lake Victoria. This relationship means that the responsiveness of bushmeat demand to fish price decrease linearly as the distance to Lake Victoria increase.
### DISCUSSION
This study has investigated the own- and cross-price elasticity of bushmeat demand to provide information for informed decisions about interventions and policies to reduce hunting by affecting the bushmeat trade that currently exerts considerable pressure on wildlife populations threatening conservation objectives in the GSE as well as in other biodiversity-rich tropical regions. Compared to studies using observed preferences, we can evaluate the implications of larger price changes because we rely on stated preferences. Including a wider geographical area of the GSE furthermore, allow us to make more general conclusions including about geographical differences.
TABLE 4 | Coefficients of the extended mixed effect Poisson models representing the own and cross-price elasticities of bushmeat demand for individual substitute meat types contingent on socioeconomic covariates.
*Values in parenthesis are standard errors. Order effect of meat experiment was controlled for in the model—not shown.*
\*\*\**,* \*\* *and* \* *reflects significance at the 0.01, 0.05, and 0.1% level respectively. The italic values represent the interaction effects of bushmeat and substitute prices with covariates included in the model.*
TABLE 5 | Coefficients of the mixed effects Poisson models representing the own and cross-price elasticities of bushmeat demand for individual substitute meat types contingent on district effects.
*Values in parenthesis are standard errors. Order effect of meat experiment was controlled for in the model—not shown.*
\*\*\**,* \*\* *and* \* *reflects significance at the 0.01, 0.05 and 0.1% level respectively. The italic values represent the interaction effects of bushmeat and substitute prices with covariates included in the model.*
### Responsiveness of Demand
We find that bushmeat demand responds negatively to changes in own-price in all five simple models without covariates and in all five extended models with socioeconomic covariates implying that bushmeat price increase will decrease demand across the GSE. This is consistent with the law of demand. Bushmeat demand is inelastic to price in four of the simple models indicating that one percent price increase leads to <1 percent decrease in bushmeat demand. However, controlling for socioeconomic covariates, four of the five extended models reveal elastic responses to bushmeat price increase. This suggests that socioeconomic covariates (i.e., household income, livestock ownership/TLU, land owned and household size) are important determinants of the responsiveness of bushmeat demand in GSE. Rentsch and Damon (2013) used an Almost Ideal Demand System analysis on revealed meat expenditure data from 131 households collected over 34 months in eight communities in Serengeti and Bunda districts implementing Seemingly Unrelated Regression models accounting for cross-equation correlations in evaluating elasticities. In their study Rentsch and Damon also found elastic uncompensated (Marshallian) ownprice elasticities (−1.122) but contrary to us found inelastic income-compensated (Hicksian) elasticities (−0.696) in an analysis including beef, fish and dagaa as substitutes. Moro et al. (2015) using a stated preference approach with separate models for each substitute (very similar to ours) on data from 200 TABLE 6 | OLS regression of estimated elasticities for individual meat types contingent on distance variables. Values in parenthesis are clustered standard errors at the village level.
*Order effect of meat experiment was controlled for in the model—not shown.*
\*\*\**,* \*\* *and* \* *reflects significance at the 0.01, 0.05 and 0.1% level respectively.*
households in six villages in Western Serengeti found inelastic own-price elasticities in both simple models (−0.657 and −0.703) and in a model controlling for socioeconomic covariates (−0.138 and −0.551) with chicken and fish as substitutes.
We also find that bushmeat demand responds positively to a price increase for three out of five substitutes (consistent with the substitute good hypothesis) but only to fish when controlling for socioeconomic covariates. Consistent with other studies in the GSE none of these cross-price effects were elastic (Rentsch and Damon, 2013; Moro et al., 2015). Moro et al. (2015) found significant cross-price elasticities when the substitute was chicken only in the simple model (+0.286) whereas the crossprice elasticities for fish were significant in both the simple model (+0.371) and in models with socioeconomic covariates (+0.734 to +0.974), albeit not elastic. Rentsch and Damon (2013) found that beef, dagaa and fish all were substitutes for bushmeat and more so in the income-compensated model but also not elastically.
Overall this indicates that initiatives targeting poachers to increase the supply–costs thereby increasing the price of bushmeat, through enhanced enforcement and severer sanctions, are more likely to effectively reduce bushmeat demand than policies aiming to make substitutes cheaper (e.g., through subsidies and extension programs). Similar conclusions were reached by the two previous studies in the GSE (Rentsch and Damon, 2013; Moro et al., 2015). However, our findings suggests that the effectiveness of supply side interventions can be optimized by designing policies to target the different substitutes across the various social, economic and spatial contexts in the ecosystem as the results shows differential effect of availability of different meat types across socio-economic groups, districts and with distance to the source of the meat type (see below for details)
A number of revealed preference studies on the role of price on meat consumption has been conducted among Amerindian communities in Bolivia. Wilkie and Godoy (2001) found more elastic own-price elasticities, particularly in the top half of the income distribution (−5.852) but similarly weak cross-price elasticities using a sample of 443 households in 42 communities (Wilkie and Godoy, 2001). Apaza et al. (2002) found less elastic own-price elasticities (-1.145) but elastic cross-price elasticities for fish (+1.464) and particularly livestock (+7.446) expanding the same sample to 510 households in 59 communities. Similar results to ours but with inelastic own-price as well as cross-price elasticities were also observed in a study in 1,208 rural and urban households in six locations across Gabon (Wilkie et al., 2005). The considerable differences in the magnitude of elasticities between the Latin American and African studies may originate from differences in purchasing power and the availability of different meat types.
As expected demand differed depending on the substitute to which it was compared. Bushmeat demand was more responsive to own-price when the substitute was beef and least responsive when the substitute was fish or goat depending on control for socioeconomic covariates (i.e., with and without). Bushmeat demand was more responsive to substitute price when the substitute was lamb and fish, with and without control for socioeconomic covariates, respectively. Hence, attempting to reduce bushmeat demand by increasing its price is theoretically likely more effective when beef is available as a substitute, while strategies working through reduced substitute price will likely be more effective when substitutes are lamb or fish. However, increasing the availability of substitutes sufficiently to reduce the price and affect bushmeat demand is complicated by the currently large price differences between bushmeat and its substitutes (Ndibalema and Songorwa, 2008; Nielsen and Meilby, 2015), the ability of Lake Victoria fish stocks to sustainably support demand (Rentsch and Damon, 2013) and environmental impacts of even higher demand for grazing land for livestock production in the GSE (Veldhuis et al., 2019). Change in the price of one meat type will likely have ecological consequences through the system by affecting demand for other meat types that are intrinsically linked and may produce negative externalities (Brashares et al., 2004; Rentsch and Damon, 2013).
### Effects of Socioeconomic Characteristics
Including socioeconomic variables in the models revealed that household income and wealth measured in livestock ownership (i.e., TLUs) and area of land owned reduced responsiveness to substitute price whereas effects on responsiveness to bushmeat price were mixed revealing lower and higher responsiveness of TLU and land-rich households, respectively. This indicates that wealthier households less readily shift to substitutes when the price of these decrease but that only land-rich households reduce bushmeat demand more than land-poor households when its price increase whereas the demand of TLU-rich households are less affected than TLU-poor households. The effect of TLU may initially seem counterintuitive but may be explained by the dominance of cattle in the TLU measure (about 74%) used mainly for milk production rather than to satisfy household meat demand, and that cattle in the GSE constitutes a source of saving and prestige more than meat (Knapp et al., 2015).
The mixed finding in relation to income and asset variables is comparable with the previous studies in the GSE that do not provide a uniform conclusion. For instance, evaluating expenditure elasticities as a measure of wealth assuming a high relation between income and expenditure due to generally low savings, Rentsch and Damon (2013) found elastic positive expenditure elasticities for bushmeat (+1.322) as well as for beef (+1.184) and fish (+1.006) indicating that consumption of these goods will increase as income (expenditure) increases. Moro et al. (2015) found that household wealth and number of household members with a job surprisingly had no significant effect on demand response to bushmeat price in Western Serengeti. Nyahongo et al. (2009) investigating bushmeat consumption frequencies in five villages in Western Serengeti found no relation with household income except in a village 80 km from the SNP.
Similar inconclusive findings are observed in other ecosystems. For instance, in Bolivia, no significant relationships were observed between bushmeat consumption and income or wealth in Amerindian households (Wilkie and Godoy, 2001; Apaza et al., 2002) although the elasticity varied from a necessity in the bottom half (+0.056) to an inferior good in the top half of the income distribution (−0.137) (Wilkie and Godoy, 2001). However, an extension of these surveys using a five-wave panel dataset from 324 households found a significant positive association between bushmeat consumption and wealth but not income and attributed this to a high degree of self-sufficiency and wealth being associated with investment in hunting technology (Godoy et al., 2010). Studies in Sub-Saharan Africa have found both negative (Albrechtsen et al., 2006), and positive relationships between bushmeat demand or consumption and income or wealth (East et al., 2005; Wilkie et al., 2005; Brashares et al., 2011; Foerster et al., 2012). Wilkie et al. (2005) for instance found increasing consumption of bushmeat (+0.169) as well as fish (+0.266), chicken (+0.262), and livestock (+0.144) with income and largest effect at the low end of the income distribution (due to the curvilinear relationship of log-transformed variables). National-wide surveys in Liberia found a considerable decrease of bushmeat consumption during the Ebola outbreak, but that wealthier households reduced their bushmeat consumption less than poorer households, that bushmeat prices remained stable and that peoples preferences for bushmeat remained the same despite its possible role as a disease vector (Ordaz-Németh et al., 2017). The stable price of bushmeat was likely explained by decreased hunting countering the lower demand. Households were, furthermore more likely to decrease bushmeat consumption if believing that Ebola could be contracted from bushmeat consumption.
In summary, our findings show that the relationship between bushmeat demand and wealth depends on the type of meat available, which is consistent with findings in the literature that show that bushmeat demand varies depending on the context, including whether it is rural or urban (Fa and Brown, 2009; Brashares et al., 2011; Luiselli et al., 2019) and likely also depending on food state and bushmeat species (East et al., 2005; Schenck et al., 2006; Ndibalema and Songorwa, 2008; Mwakatobe et al., 2012). Our results only partially support the concerns of other studies in the GSE indicating that efforts to increase household income and wealth will also increase bushmeat demand as well as demand for other protein sources. In general, there is a need for a much better understanding of what poverty is and how it relates to motivations for hunting and consuming bushmeat (Duffy et al., 2016).
Household size was also negatively associated with bushmeat demand responsiveness to own price (beef) and substitute price (beef and fish). The opposite results were observed by Moro et al. (2015) who found that household size increased own and cross-price elasticity in models where the substitute was chicken and fish, respectively. This difference may be explained by the lower variation in household size in their sample from Western Serengeti (mainly in Serengeti district), largely excluding households in Meatu and Bariadi district that tend to be significantly larger [mean AEU = 7.93, 7.72 and 5.80 for Meatu, Bariadi and Serengeti districts, respectively, P < 0.01 (ANOVA with Bonferroni multiple comparison test)]. Other studies have found negative relations between the quantity of bushmeat demanded or consumed per individual and household size (Wilkie et al., 2005; Albrechtsen et al., 2006; Godoy et al., 2010; Foerster et al., 2012), which contradicts general theory predicting higher efficiency of larger households (see Foerster et al., 2012). In this case, we expect that higher protein demands of larger households marginally override budget constraints also because demand is not measured per AEU.
The responsiveness of bushmeat demand also varied depending on household ethnicity and substitute considered. The own-price elasticity decreased when the respondent was from a Maasai household, and the substitute was chicken or fish; when the respondent was Sukuma, and the substitute was beef; and Kuria and the substitute was lamb. The cross-price elasticity for goat and fish declined when the respondent was Massai. It also declined for beef, goat and fish when the respondent was Sukuma, and for lamb when the respondent was Kuria. Moro et al. (2015) also found differences between ethnic groups in Western Serengeti in the effect of substitutes on elasticities. Culturally determined consumption preferences are likely to be important determinants of bushmeat consumption (Fa et al., 2002; East et al., 2005; Kaltenborn et al., 2005; Schenck et al., 2006; Kiffner et al., 2015). A survey of 600 households in five districts in Bunda, Meatu, Bariadi, and Tarime districts in GSE found high variability in bushmeat consumption and that the Ikoma and other inhabitants in Bunda district consumed more bushmeat than members of the Sukuma and Kurya ethnic groups (Ndibalema and Songorwa, 2008). Ceppi and Nielsen (2014) also found differences in prevalence and diversity of bushmeat consumption across a sample of 300 households from 10 ethnic groups across Tanzania. However, information about cultural and taste preferences for specific domestic as wells as wildlife species are currently insufficient to rigorously interpret these results.
### Spatial Effects on Bushmeat Demand
The responsiveness of demand to price varied between districts and in relation to distance to spatial features in the landscape. Demand was more responsive to bushmeat price in Serengeti district compared to other districts while it was less responsive in Bariadi and Tarime districts but inconsistently and depending on the substitute. Responsiveness to beef price was high in Serengeti and Meatu and low in Bariadi while responsiveness to fish price was high in both Serengeti and Bariadi. Responsiveness to goat price was low in Bariadi but high in the NCA. Previous studies have found high consumption of bushmeat in Serengeti district followed closely by Meatu and Bariadi compared to Bunda and Tarime (Ndibalema and Songorwa, 2008). By including particularly the LGCA and NCA but also other districts not considered in previous elasticity studies (Rentsch and Damon, 2013; Moro et al., 2015) our results provide new insights to the design of policies aiming to reduce bushmeat demand through interventions manipulating prices by enabling optimization of design to the population's preferences in each district adjacent to the GSE.
Preferences were also influenced by households location in relation to spatial features irrespective of districts, but the direction of influence depends on the substitute. The responsiveness to bushmeat price was lower further away from the protected areas when the substitute was beef and decreased at an increasing rate with distance to the boundary perhaps indicating a tendency to becoming a luxury good with households further from the boundary willing to pay higher prices for bushmeat. Responsiveness to bushmeat price was also higher further away from Lake Victoria and increased at an increasing rate when the substitute was beef while it was lower and decreased at an increasing rate when the substitute was chicken. Responsiveness to the price of fish also decreased linearly with distance to Lake Victoria. These trends are likely associated with culturally determined preferences of the Massai in the eastern part of the GSE who may find chicken and fish unacceptable substitutes although bushmeat consumption by the Massai is also a relatively recent development (Ceppi and Nielsen, 2014; Kiffner et al., 2015). Finally, the responsiveness of demand to bushmeat price decreased with increasing distance to a road indicating that more remote households have a higher preference for bushmeat. A number of studies have investigated the influence of roads and distance to protected areas on bushmeat consumption, trade and game depletion (Macdonald et al., 2012; Fa et al., 2015; Mavah et al., 2018). Macdonald et al. (2011) surveyed bushmeat trading points in 87 villages in Nigeria and Cameroon through 150 days and found that prices increased with distance from protected area boundaries and were also higher closer to road networks. In Gabon, a study covering 928 households in 56 villages adjacent to three newly established national parks found that bushmeat consumption decreased as distance to protected area boundaries increased (Foerster et al., 2012). In Western Serengeti, the study by Nyahongo et al. (2009) found that bushmeat consumption declined significantly with distance to the protected area boundary.
Overall our results suggest that policies aiming to reduce bushmeat consumption through manipulation of prices are likely to be most effective by targeting areas close to the boundary in more remote areas where also evidence from other studies indicates that the amount of bushmeat consumed is likely to be higher. On a larger scale, such initiatives are either more or less likely to work further away from Lake Victoria depending on local culture and acceptability of substitutes.
### Assessment of the Empirical Approach
We asked people to state the amount of meat they would purchase at different combinations of prices. As this is a hypothetical question rather than actual market transaction, it involves uncertainty about the amounts, familiarity with substitutes and own demand and is subject to hypothetical market bias. Hence, our results do not predict elasticities of actual demand but instead, reflect elasticities of stated demand. Furthermore, since bushmeat trade is illegal in the GSE, respondents may have incentives to provide strategic answers to influence policy decision in their favor. It is not clear which direction such motivations would have—i.e., whether they would increase or lower elasticities. Furthermore, previous studies using stated preference experiments in the context of bushmeat trade with actors actively involved in hunting and trading bushmeat suggests a large potential to provide information about such sensitive activities (Nielsen et al., 2014). Our design furthermore framed the experiment as a legal trade where all meat types were sold by a vendor coming to respondents household. Therefore, we do not expect the strategic element to be driving the results.
We found different elasticity estimates depending on the substitutes considered in each model. This is likely due to heterogeneity in our sample as a result of the large geographical extent and cultural diversity of the population in our study area causing different preferences, availability and familiarity with different meat types. Such differences were also observed in interviews of 2453 individuals in 27 communities across Nigeria, Togo, Niger and Burkina Faso (Luiselli et al., 2019) and may explain why most crossprice elasticities were insignificant. However, as we have tried to capture this heterogeneity by the inclusion of relevant covariates, we can discern general trends in bushmeat demand elasticities and identify differences in elasticities across socio-economic groups, locations and depending on spatial variables.
The estimated models only consider bushmeat demand while the demand for its substitutes was not included, contrary to the approach used by Rentsch and Damon (2013). However, if a household has no preference for consuming bushmeat or its substitutes at the given prices, this could have implications for estimation, reducing, the elasticity of bushmeat demand. We included a variable in each model reflecting whether respondents attended to bushmeat and substitute price in making their demand decisions to ensure that the results are not driven by such differences. However, we found that respondents not attending to the price of bushmeat in the model with beef as the substitute, indicating that they do not consume bushmeat, have significantly lower elasticity for bushmeat (cf. **Table 4**) and we, therefore, cannot exclude such effects in all models.
### CONCLUSION
We assessed the own- and cross-price elasticity of bushmeat demand for more substitutes and across a wider geographical area of the GSE than previous studies and evaluated the implications of socioeconomic differences, distances to protected area boundaries, Lake Victoria and roads and compared districts using a stated preference approach. Bushmeat demand was negatively correlated with the price of bushmeat (i.e., negative own-price elasticity) and positively correlated with the price of substitutes (i.e., positive cross-price elasticity) (particularly for fish). Demand responded elastically to the price of bushmeat indicating that a 1% increase in the price of bushmeat leads to more than 1% decrease in bushmeat demand in most models controlling for socioeconomic covariates. However, demand responded inelastically to substitute price. These results suggest that increasing the price of bushmeat by targeting poachers to increase the supply–costs likely makes policies and initiatives aiming to reduce bushmeat hunting more effective than subsidies and extension programs aiming to make substitutes cheaper. Observed differences between ethnic groups and districts provide important insights enabling optimization of program design to the population's preferences in each district adjacent to the GSE. Household income and wealth measured in TLU and land mainly reduced the cross-price elasticity of bushmeat demand but also reduced the own-price elasticity of bushmeat demand for more land-rich households. This only partially support previous findings that efforts to improve household welfare across the GSE will increase protein demand increasing the pressure on wildlife populations. Demand responsiveness to bushmeat price furthermore declined with distance to protected area boundary but increased with distance to Lake Victoria. However, most effects differed between models depending on substitute considered, in a pattern that is difficult to explain due to limited information about cultural and taste preferences for specific domestic as wells as wildlife species.
Overall our results reveal that interventions aiming to reduce bushmeat demand by affecting prices while maintaining communities food security may not meaningfully reduce demand within the realistic price range shifts in the GSE context. However, the effectiveness of demand-reducing interventions should increase if complemented by other policy interventions. These interventions should ideally provide intrinsic motivations, that can be developed into long-lasting cultures of conservation (Cetas and Yasué, 2017) by appropriately acknowledging local value orientations in relation to wildlife and bushmeat (van Vliet, 2018). Options for engendering change in consumer preferences as well as hunter behavior may include edutainment interventions if appropriately designed to achieve sufficient audience penetration (Veríssimo et al., 2018), social marketing in the form of community engagement and information campaigns (Chaves et al., 2018; Green et al., 2019; Greenfield and Veríssimo, 2019; Veríssimo, 2019), social learning (Roux et al., 2011) and environmental education (Salazar et al., 2018). Simultaneously providing alternative income generation opportunities for hunters, that ideally should be incompatible with poaching or contingent on wildlife increase, may further increase the opportunity cost of hunting but may require substantial conceptual rethinking as well as improvement in funding design, monitoring and evaluation and the use of adaptive management strategies (Wright et al., 2016; Wicander and Coad, 2018). Furthermore, given sufficient time and prevalent urbanization, cultural norms and preferences toward bushmeat consumption are likely to change and reduce the acceptability of bushmeat consumption (Luiselli et al., 2019). The question is in what state the GSE and its wildlife populations will be at that time.
### ETHICS STATEMENT
This study passed an EU Horizon 2020 ethics review procedure before commencement of the project (proposal number 641918). The ethics advisory group of the Department of Food and Resource Economics at the University of Copenhagen evaluated and approved the ethics guidelines, free, prior informed consent form (ID: GA641918). The Tanzanian Commission for Science and Technology granted permission for implementation of the study (research permit No 2017-299-NA2011-21). Finally, procedures for collecting information from human subjects were approved by the Tanzanian National Health Research Ethics Committee (ID: NIMR/HQ/R.8a/Vol. IX/2609).
### AUTHOR CONTRIBUTIONS
All the authors were involved in designing the study and choice experiment as well as writing the manuscript. SW conducted the focus group discussions and undertook data collection, management and analysis.
### ACKNOWLEDGMENTS
We are grateful for the efforts of TAWIRI enumerators and field supervisors for assisting with data collection. Permission for this study was granted through COSTECH research clearance
### REFERENCES
No 2017-299-NA2011-21. This study was financed by the European Union's Horizon 2020 research and innovation program under grant agreement No. 641918 through the AfricanBioServices project. JBJ also acknowledges support from the Danish National Research Foundation to the Center for Macroecology, Evolution, and Climate.
### SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fevo. 2019.00162/full#supplementary-material
for illegally hunted bushmeat in the Serengeti, Tanzania. Anim. Conserv. 18, 377–386. doi: 10.1111/acv.12184
protected areas? - a case study from the Serengeti Ecosystem, Tanzania. Biol. Conserv. 136, 563–570 doi: 10.1016/j.biocon.2007.01.001
**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Copyright © 2019 Walelign, Nielsen and Jakobsen. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
# Assessing the Minimum Sampling Effort Required to Reliably Monitor Wild Meat Trade in Urban Markets
Pedro Mayor 1,2,3,4 \*, Hani R. El Bizri 3,5,6,7, Thais Q. Morcatty 6,8, Kelly Moya<sup>9</sup> , Samantha Solis <sup>9</sup> and Richard E. Bodmer 2,3
<sup>1</sup> Departament de Sanitat i d'Anatomia Animals, Facultat de Veterinària, Universitat Autònoma de Barcelona, Barcelona, Spain, <sup>2</sup> Museo de Culturas Indígenas Amazónicas, Iquitos, Peru, <sup>3</sup> ComFauna, Comunidad de Manejo de Fauna Silvestre en la Amazonía y en Latinoamérica, Iquitos, Peru, <sup>4</sup> Postgraduate Program in Animal Health and Production in Amazonia, Federal Rural University of the Amazon (UFRA), Belém, Brazil, <sup>5</sup> Rede de Pesquisa em Diversidade, Conservação e Uso da Fauna da Amazônia (REDEFAUNA), Manaus, Brazil, <sup>6</sup> Mamirauá Sustainable Development Institute (IDSM), Tefé, Brazil, <sup>7</sup> School of Science and the Environment, Manchester Metropolitan University, Manchester, United Kingdom, <sup>8</sup> Oxford Wildlife Trade Research Group, Oxford Brookes University, Oxford, United Kingdom, <sup>9</sup> Universidad Nacional de la Amazonia Peruana, Iquitos, Peru
#### Edited by:
Robert NASI, Center for International Forestry Research, Indonesia
#### Reviewed by:
David S. Wilkie, Wildlife Conservation Society, United States Daniel J. Ingram, University College London, United Kingdom
#### \*Correspondence:
Pedro Mayor mayorpedro@hotmail.com; pedrogines.mayor@uab.cat
#### Specialty section:
This article was submitted to Conservation, a section of the journal Frontiers in Ecology and Evolution
> Received: 02 March 2019 Accepted: 06 May 2019 Published: 29 May 2019
#### Citation:
Mayor P, El Bizri HR, Morcatty TQ, Moya K, Solis S and Bodmer RE (2019) Assessing the Minimum Sampling Effort Required to Reliably Monitor Wild Meat Trade in Urban Markets. Front. Ecol. Evol. 7:180. doi: 10.3389/fevo.2019.00180 The trade of wild meat generates great economic returns for local communities but at a cost of increasing harvest rates of game species. Monitoring wild meat trade in urban markets is a low-cost method that can be employed to assess impacts of hunting on game populations. Nevertheless, wild meat markets are complex systems to monitor since they often vary over time, are illegal in some countries, and often vendors distrust researchers. We investigated the wild meat trade in the Belén market in Iquitos, Peru, the largest wild meat market in the Amazon, to estimate the minimum sampling effort required to obtain reliable estimates of the amounts and prices of wild meat sold. During two 12-month surveys (Sept. 2006–Aug. 2007, Sept. 2017–Aug. 2018), we conducted a total of 4,524 vendor interviews in 320 sample days. By modeling 10 possible scenarios in which sampling size and amount of meat traded varied, we calculated the accuracy and precision of different survey protocols. We found that in scenarios where the daily amount of wild meat on sale was between 40 and 650 kg, a sampling effort equal to or >2 sampling days per month provided good accuracy (>90%) and precision (>85%). However, in scenarios where wild meat traded was less frequent, or for rarer species, an effort of at least one interview per week is required. Vendor declaration of the daily amounts of meat sold was similar to the quantity on sale (accuracy = 98%), suggesting that sellers are aware of the volume of wild meat brought to market. To accurately monitor the trade of wild meat in urban markets, we recommend a minimum sampling effort, ranging from two interviews per week to two interviews per month, depending on the amount of wild meat traded; in other occasions, a punctual interview on meat sellers' perception may also be useful.
Keywords: bushmeat, wildlife trade, sampling effort, accuracy, amazon
## INTRODUCTION
Wild meat represents an important source of protein and income for local people in tropical forests in Africa, Asia, and Latin America (van Vliet and Nasi, 2008; Zapata-Ríos et al., 2009; van Holt et al., 2010; Fa et al., 2015). In many of these regions, the trade of wild meat to supply urban markets is considered a main cause of population declines of many game species (de Thoisy et al., 2005; Zapata-Ríos et al., 2009). Although the sale of wild meat by local hunters can be an important source of revenue for poor families, a greater emphasis on profits will cause a significant rise in wildlife harvest rates (Morsello et al., 2015). The observed intensification in wild meat harvest levels in many parts of the tropics has been linked to a greater commitment by indigenous and rural populations to supply city markets, in turn fuelling greater demand for wildlife products in urban areas (Ohl-Schacherer et al., 2007; Suarez et al., 2009; Fa et al., 2015; Kirkland et al., 2018).
The trade in wild meat has proved to be a very accessible and cost-effective indicator of the regional dynamics of game populations (Fa et al., 2000, 2015). Trends in the volumes of wild meat sold in urban markets can be used as evidence of hunting sustainability in the rural areas supplying the urban center (Fa et al., 2004; Morcatty and Valsecchi, 2015). For instance, in only 2 years of sampling in the Bioko Island, Africa, Albrechtesen et al. (2007) predicted unsustainable hunting in the surrounding areas through reductions of wild meat availability in urban markets. Therefore, developing statistically robust techniques to understand the wild meat trade in urban markets is essential to enable appropriate management strategies to emerge for the control of demand and supply of wild species used for food.
One of the main hindrances in monitoring the trade of wild meat in markets in most tropical countries is the fact that this is an illegal activity and, therefore, difficult to investigate directly. Thus, studies assessing the wild meat trade have usually consisted of short-term surveys, making the reliability of this information uncertain. In addition, wild meat markets have been shown to be complex systems that may change considerably over time based on supply- and demand-driven forces (McNamara et al., 2016), causing the availability of species, amount of wild meat, and their prices to vary daily, seasonally, and annually. Therefore, any survey intended to effectively monitor the wild meat trade must acknowledge this variation. However, there is still a lack of studies assessing what should be the minimum effort needed to obtain reliable estimates of the wild meat trade. To date, only one study in Africa investigated the performance of different sampling regimes from long-term data from five wild meat markets in West and Central Africa. This study showed that the accuracy and precision of samplings increased with sample size, and for the markets with the highest amounts of wild meat, these parameters started reaching an asymptote with an effort of around 28 and 35 sampling days per year (Fa et al., 2004).
In the Amazon, where the commercial route of goods depends on the distribution of rivers, the supplying of wild meat usually concenters in the largest urban markets along large rivers; surveying these markets may provide useful indicators of the status of wildlife populations at the regional scale (Fa et al. 2004). However, differently from Africa, there is still no consensus on what should be the minimum effort to reliably monitor Amazonian markets, and how this effort varies according to the amount of meat traded. In this study, we used two monitoring datasets collected from wild meat sellers in the Belén Market in Iquitos, Peru—the most important and largest open market in wildlife in the Amazon—to model the minimum effort required to obtain reliable information on the amount and trends of wild meat trade in Amazonia. We assessed the efficiency of using different sampling efforts and the sellers' perception to measure the volume and the price of wild meat traded.
### MATERIALS AND METHODS
### Study Area
The Belén Market is located in Iquitos, the largest city in the Peruvian Amazon (437,376 inhabitants). It is one of the most important Amazon markets in terms of volume of wild meat sold (Bodmer and Pezo, 2001). This market offers countless different types of goods extracted from the rainforest, from traditional medicines and pets to fresh fruits and vegetables.
In the Belén Market, wild meat is sold openly, thus making it relatively easy to track. Wild meat is typically supplied directly by intermediaries or by hunters that travel from their villages to the cities to sell their products to market vendors, household consumers, or restaurants (Bendayan, 1991). Although wild meat is not consumed daily, being secondary to the more commonly eaten poultry and fish, it is eaten as a traditional dish, where some species are considered luxury.
### Data Collection
Two 12-month surveys were conducted in the Belén market during Sept. 2006–Aug. 2007 and Sept. 2017–Aug. 2018. Before the start of the surveys, we identified vendors to interview with the help of local informants. To gather information on the volume and price of the traded meat of wild species in the market, we used informal interviews and participant observation. All informants participated voluntarily after being primed of the project's aim. Anonymity of all participants was respected.
Since vendors display their wild meat products upon openair market stalls, we could count volumes and species sold. Interviews were conducted twice daily between 6:00 a.m. and 12:00 p.m.; after midday, sales decreased substantially or sold out. The following data were recorded: date, species sold, type of meat preservation (fresh, salted fresh, salted dry, or smoked), selling price per kilogram, amount of wild meat brought by sellers at the start of the day, including the amount of wild meat displayed on the stall and stored indoor, and amount left at the end of the day. The amount of wild meat sold was calculated from the difference between the amounts on sale at the beginning minus the amount left at the end of the day. Although vendors were asked to confirm the taxa on sale, we independently verified each species. Since mammals make up over 80% of all wild meat traded in this market, we focused only on this group (Bodmer and Pezo, 2001).
In 2006–2007, we interviewed 29 vendors, a total of 2,443 interviews (203.6 ± 35.1 monthly interviews) in 182 sampling days, covering 50.1% of days per year (one sampling every 1.99 days). During 2017–2018, we performed 2,081 interviews (173.4 ± 59.3 monthly interviews), 30 vendors during 138 sampling days, 37.8% of days per year (one sampling every 2.64 days). These vendors, a large sample of all active ones in the markets, were regular sellers of wild meat who agreed to participate throughout the whole study period. Occasional vendors, those who sold only a small volume of wild meat (along with other rainforest goods), were not considered in this study.
At the end of the 2017–2018 survey we interviewed 11 of the most frequent wild meat sellers so as to obtain their opinion on the average price and average daily amount of wild meat sold year-long.
### Data Analysis
The amounts of meat (salted fresh, salted dry, and smoked) recorded per species were transformed into fresh meat using the conversion indexes proposed by Bardales-García et al. (2004). For those species for which we did not have conversion indices, we applied the index for a taxonomically related species of similar body mass. The daily price in US dollar (US\$) per kilogram of wild meat was calculated for all mammal species and all kinds of meat. To convert Peruvian Soles (S\$) into US dollars, we used the exchange rate from 10 October 2007 (S\$ 3.00 = US\$1.00) for the survey 2006–2007 and from 04 October 2018 (S\$ 3.32 = US\$1.00) for the survey 2017–2018.
We achieved 182 and 138 interview-days during the 2006– 2007 and 2017–2018 survey periods, respectively. To assess the effectiveness of different survey efforts, we modeled 10 scenarios using different sample sizes. We reduced the number of sampling days within each year-survey by using a progressive random selection of interview-days homogenously distributed along the year: 182 and 132 (maximum effort), 90 and 75 (with 2 replicates per survey), 45, 24, 12, 6, 4, and 3 (with 5 replicates per survey), and 2 interview-days (with 10 replicates per survey). We also modeled two seasonal sampling periods, consisting of interviews performed every 2 days for the months with the highest and lowest water levels of the Amazon River (Servicio de Hidrografía, 2015).
For every scenario, we calculated the average and standard deviation of the price and total amount of fresh-converted wild meat sold per day. We considered that the maximum survey effort (hereafter "reference model") was the most reliable information, and any reduction in sampling effort would bias the reference model. Bias is a reduction in the accuracy and precision of the price and amount of meat sold. Accuracy refers to the level of proximity, in percentage, of the average relative to the reference model. Precision refers to how variable estimates from different samples were compared with each other, and was estimated based on the standard deviation of the different parameters. To predict accuracy, we calculated the relative difference between the daily average in any experimental effort with respect to the reference model. Similarly, to predict precision, we calculated the relative difference of the daily standard deviation in each experimental effort compared to the reference model. Values close to 1.00 (or 100%) meant maximum accuracy or precision relative to the reference model. We considered effective sampling for those efforts that concomitantly combined accuracy and precision values higher than 90%. We also presented the amount of wild meat sold and the accuracy and precision in each sampling scenario for seven different taxa: Cuniculus paca, Pecari tajacu, Tayassu pecari, Mazama sp., Hydrochoerus hydrochaeris, Tapirus terrestris, and Lagothrix sp. Since these species presented different trading volumes in the market, they were used to assess the influence of different amounts of wild meat on the precision and accuracy obtained.
We used multiple regressions to model the relationship between sample size and accuracy or precision with the software CurveExpert 2.4 (©Copyright 2017, Daniel G. Hyams). Functions that best fitted the plots were selected by employing those with the highest correlation coefficient (r).
For interviews conducted in October 2018 on sellers' perceptions, we used a paired t-test to compare average price and daily amount of wild meat sold throughout a year by comparing records for the 11 most important sellers for amounts of wild meat sold in the reference model and their own perception.
Randomization of the survey days for building the models was performed using R-Studio version 3.3.3 (RCore Team, 2017). Results with P < 0.05 were considered significant.
### RESULTS
In 2006–2007, sellers sold a total of 220,487 kg of wild meat, an average of 663.1 ± 188.7 kg per day, at an average price of US\$ 3.82 ± 0.19 per kg. The total amount of wild meat sold in 2017– 2018 was 288,336 kg, an average 886.2 ± 399.0 kg per day, at an average price of US\$ 6.04 ± 0.33 per kg. For both years pooled, the average daily amounts of wild meat traded differed among species. Meat of Pecari tajacu and Cuniculus paca was the most traded (197.6 ± 96.0 and 190.4 ± 107.6 kg, respectively), while Lagothrix sp. had the lowest sale rate (2.8 ± 7.3 kg; P < 0.001). Tayassu pecari, Mazama sp., Hydrochoerus hydrachoerus, and Tapirus terrestris had intermediate sales rates (124.8 ± 94.5, 57.9 ± 43.3, 38.8 ± 34.0, and 22.1 ± 23.5 kg, respectively; **Figure 1**).
Accuracy and precision of the price and of the amount of wild meat sold increased proportionally to sampling effort (**Figures 2**, **3**, **Table 1**). Nevertheless, no significant differences were observed between 12 and 182 interview-days in the average accuracy for both price and amount of total wild meat sold (96.7 ± 2.1% and 96.5 ± 3.0%, respectively): an average precision of 87.7 ± 12.2% for price and 90.4 ± 9.9% for amount of meat (**Figure 2**, **Table 1**). Scenarios with lower sampling efforts, between six and two annual interviews, resulted in decreased accuracy for price (79.6 ± 5.6%) and total amount of wild meat (87.9 ± 9.7%). Similarly, we also detected a decreased precision for price (66.8 ± 22.3%) and total amount of wild meat (55.9 ± 26.3%) within this effort range. The seasonal experimental design showed similar accuracy compared to the reference model (total amount 82.6% and price 92.9%), but precision was considerably lower (total amount 77.3% and price 29.3%).
Accuracy and precision varied according to the amount of meat sold. In the case of Pecari tajacu and Cuniculus paca, which represented a daily sale between 190 and 200 kg, the
2006–2007 and 2017–2018. Solid bold lines represent the average amount of wild meat (in fresh-converted kilograms) commercialized per species and year.
accuracy with 12–45 annual interview-days remained close to 95%, and precision with 24–45 annual interview-days was higher than 90%. In scenarios with daily sales between 40 and 125 kg of wild meat (Hydrochoerus hydrachoerus and Tayassu pecari, respectively), accuracy was at least 85% and 95% at an effort of 12 and 75 annual interview-days, respectively, but precision decreased to 86% with a sampling effort of 1 monthly interview. In species with daily sales of 22 and 3 kg (Tapirus terrestris and Lagothrix sp.) accuracy was at least 90% only with 75 and 90 annual interview-days, respectively (**Figure 3**; **Table 1**).
Results of our sellers' perception interviews indicated that vendors accurately assessed 78.2 ± 96.4% of the amount of wild meat traded within our year-long survey (t10<sup>=</sup> −2.815, P = 0.018). Conversely, vendors' perception of meat available was similar to the amount of wild meat on sale at the beginning of the day, showing an accuracy of 97.7 ± 41.3% (t10<sup>=</sup> −0.452, P = 0.661). In addition, their perception of price was also similar to the average price obtained in the year-along survey: an accuracy of 98.2 ± 9.5%.
### DISCUSSION
Despite some caveats, information on the amounts and species of wild meat traded in urban markets can be used to understand the impact of hunting over large geographical areas (Fa et al., 2004; Fa, 2007), especially for the most frequently sold species. Nevertheless, since wild meat trade is forbidden in some tropical countries, this activity is excluded from official statistics. Although several studies have been recently conducted in markets of some Neotropical countries (Bodmer and Pezo, 2001; van Vliet et al., 2015), their reliability can be compromised due to the evasive behavior of meat sellers and buyers. In this context where long-term monitoring of wild meat markets can be expensive or even too risky, defining efficient and adequate minimum sampling effort has been a priority (Fa et al., 2004). In this study, we assessed the efficiency of different sampling efforts for monitoring the largest market of wild meat in the Amazon. Although the trade in wild products in urban markets is forbidden in Peru (Law No 29763), the surveillance authorities have been unable to enforce this law due to logistical
and financial limitations and the high traditional demand for wild meat.
the maximum sampling effort in two 12-month surveys conducted in 2006–2007 and 2017–2018.
Our study suggests that the optimal sampling effort would range between two weekly interviews to two sampling days per month depending on the amount of meat sold in the market. Since financial and personnel resources are often limited in research projects (Garden et al., 2007), the sampling design may be adjusted within this optimal range according to available resources. While two sampling days per month resulted in high accuracy and precision values compared to the long-term interday sampling effort for the total wild meat and for common species in the market, a minor sampling effort of one interview per month resulted in an acceptable accuracy (>90%) but a decreased precision (<90%). As also detected for African markets, the minimum effort depends on the average amount of wild meat sold (Fa et al., 2004). However, these differences are slight for the most traded species, and we suggest that in markets with sales volumes between 40 and 650 kg, including total wild meat or particular traded species, a minimum monthly effort of two interview-days should be maintained. The ability to estimate the trade of less frequent or rarer species requires a higher effort of at least 1 weekly interview. In addition, in the Amazon, any short-term seasonal experimental design showed very low precision, compromising the reliability of the data obtained. A temporally distributed sampling over the year also resulted in higher accuracy and precision for estimations of the amount of meat traded in African markets and should be employed in future studies (Fa et al., 2004).
The unique interview on the sellers' perception was considerably effective at estimating the meat available for trade (approximately 98% of accuracy). This result suggests that sellers are aware of the amount of wild meat brought daily to the market, but they do not control the volume of products they actually trade. Occasional vendors would probably show a perception farther away from reality due to the lower repetitiveness of sales events. In contrast, the sellers' perception on the price was well-adjusted to the average annual price, presenting both high accuracy and precision, probably because this parameter presents lower variability along the year.
Therefore, this sampling strategy may be useful to determine the amounts of animals removed from the forests and their prices but should be used with caution. Besides the reduced number of sampling days, the level of confidence between surveyor and seller may also influence the reliability of results. This relationship may depend on the degree of openness of the market, the regularity of the sale, and the amount of wild meat sold. The Belén Market, sampled in this study, is well-known for having being largely studied for around 20 years (see Bodmer and Pezo, 2001). In the last 10 years, we carried out several studies in this market, which allowed us to gain the confidence from some important sellers of wild meat. Nevertheless, even with this trustful relationship, we observed that some sellers distrust our purposes and fear an alliance with researchers and local governmental institutions. In hidden markets, it is expected that the wild meat trade might be more difficult to observe, increasing uncertainness and likely leading to underestimations of the amounts of wild meat sold. Illegal sellers, such as those participating in the wild meat sector, may respond hindering the truth due to fear of legal consequences. Therefore, we advocate that punctual interviews, or even long-term monitoring schemes, should be used only when trust from the sellers is obtained.
In the Amazon, where most areas have a scarcity of roads and most products are supplied through fluvial transportations, the commercial route of goods depends on the distribution of rivers, and wild meat trade usually concenters in the largest urban markets along large rivers; surveying these markets may
(E) Hydrochoerus hydrochoeris, (F) Tapirus terrestris, and (G) Lagothrix sp. (in fresh converted kilograms) in relation to the results obtained with the maximum sampling effort in two 12-month surveys conducted in 2006–2007 and 2017–2018.
TABLE 1 | Accuracy and precision (in %) of the daily amount of wild meat sold according to different experimental efforts in two 12-month surveys conducted in 2006–2007 and 2017–2018.
provide useful indicators of the status of wildlife populations at regional scale. Consequently, the long-term monitoring of Amazonian urban markets can stand as a better indicator of the regional conservation status of wildlife and is essential to anticipate management strategies that provide a response to population crisis of game species. Ultimately, the use of cost-effective and accurate tools to obtain key market indicators allows comparing annual trends in the volumes of wild meat sold for certain species. In this context, we consider that accuracy and precision values higher than 90% are acceptable.
Since wild meat trade is forbidden in most tropical rainforests, efficient and adequate sampling strategies have rarely been developed. Our study, conducted in the largest open market of wildlife in the Amazon, provides appropriate estimations of the minimum effort required to monitor wild meat trade. Since the cultural importance of the wild meat consumption is shared among almost all Amazonian countries, we believe that the minimum effort estimated here may apply to other Amazonian urban markets. Finally, a sampling effort ranging from two weekly interviews to two interviews per month homogenously distributed over the year, or a punctual interview with sellers on their perceptions may provide accurate estimates of both amount and price of wild meat, as far as a trustful relationship is attained and bearing in mind the limitations these data may have at informing trade rates.
### DATA AVAILABILITY
The datasets generated for this study can be found in Pangaea (doi.pangaea.de/10.1594/PANGAEA.898710).
### AUTHOR CONTRIBUTIONS
PM and HEB were responsible for idea conception, study design, analyses, and manuscript preparation. KM and SS contributed to the field work. TM designed the statistical analyses and participated in the manuscript preparation. RB was responsible for idea conception and study design.
### FUNDING
This work was supported by the Museo de Culturas Indígenas Amazónicas. The publication of this manuscript was possible thanks to the financial support from USAID, through the Bushmeat Research Initiative, from CIFOR and the CGIAR Forests, Trees and Agroforestry Programme. HEB received scholarship from the National Council for Scientific and Technological Development (CNPq) (grant number 201475/2017-0). TM is supported by the WCS Graduate Scholarship Program, a program of the Wildlife Conservation Society and the Christensen Conservation Leaders Scholarship,
### REFERENCES
and by the Wildlife Conservation Network Scholarship Program through the Sidney Byers Scholarship award.
### ACKNOWLEDGMENTS
We thank all the vendors of wild meat in the Belén Market of Iquitos who kindly participated in the data collection.
tortoise in the Amazon. Ecol. Soc. 20:3. doi: 10.5751/ES-07701-2 00303
**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The reviewer, DI, declared a past co-authorship with one of the authors, HEB, to the handling editor.
Copyright © 2019 Mayor, El Bizri, Morcatty, Moya, Solis and Bodmer. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
# Declining Ungulate Populations in an African Rainforest: Evidence From Local Knowledge, Ecological Surveys, and Bushmeat Records
#### Towa Olivier William Kamgaing1,2 \*, Zeun's Célestin Brice Dzefack <sup>3</sup> and Hirokazu Yasuoka<sup>1</sup>
<sup>1</sup> Center for African Area Studies, Kyoto University, Kyoto, Japan, <sup>2</sup> School for the Training of Wildife Specialists, Ministry of Forestry and Wildlife, Garoua, Cameroon, <sup>3</sup> Department of Forestry, University of Dschang, Dschang, Cameroon
Sound wildlife management requires an awareness about the trends in animal composition and abundance by all stakeholders, including local peoples. Hunters (n = 255) from two adjacent community hunting zones (CHZ) in southeast Cameroon were interviewed about the species composition of the animals killed using snare traps between 1952 and 2015–2016 and the drivers of change. The comparison of the perceived spatial and temporal trends in game composition to those from transect surveys and bushmeat records conducted in the area since the nineties evidenced the followings: (1) hunters are aware of the coarse changes in prey abundance, even for the species for which population density is difficult to estimate using more conventional survey methods; (2) in southeast Cameroon and in forests regions with similar fauna, the bay duiker (Cephalophus dorsalis) and the Peter's duiker (C. callipygus) are clearly more abundant than the white-bellied duiker (C. leucogaster) and black-fronted duiker (C. nigrifrons); (3) the two sites surveyed are at different stages of prey depletion, and (4) perception of prey composition is consistent with village-based bushmeat records and is likely to reflect more the species compositions in anthropogenic forest mosaics, where hunting is more frequent. Hunters' interviews constitute a valuable means to rapidly assess the status and trends in animal populations. However, the discrepancies between perceptions and prey composition in remote forest areas, combined with the assumption that shifting baseline syndrome is operating, highlight the need of caution when using local knowledge to generalize trends in fauna assemblages over large geographical and temporal scales.
Keywords: bushmeat hunting, Congo basin, forest duiker, local perception, species composition, snare trapping
### INTRODUCTION
The meat of wild animals (known as "bushmeat") has long constituted an important source of proteins for forest-dwelling peoples in Africa (Mendelson et al., 2003). However, in many regions, patterns of bushmeat consumption, and trade are changing rapidly, mainly because of increasing demand from urban areas and declining supply in rapidly degrading locations (Fa et al., 2002). Historically, wild animals have been hunted in African forests with diverse "traditional" tools, including crossbows, nets, spears, and snares made of vegetal materials. Wire
#### Edited by:
Nathalie Van Vliet, Center for International Forestry Research, Indonesia
#### Reviewed by:
Romain Duda, Institut Pasteur, France Miguel Farfan, University of Málaga, Spain
\*Correspondence: Towa Olivier William Kamgaing
wkamgaing@gmail.com
#### Specialty section:
This article was submitted to Conservation, a section of the journal Frontiers in Ecology and Evolution
Received: 01 November 2018 Accepted: 14 June 2019 Published: 09 July 2019
#### Citation:
Kamgaing TOW, Dzefack ZCB and Yasuoka H (2019) Declining Ungulate Populations in an African Rainforest: Evidence From Local Knowledge, Ecological Surveys, and Bushmeat Records. Front. Ecol. Evol. 7:249. doi: 10.3389/fevo.2019.00249 snares were introduced to Africa after the Second World War (Shetler, 2007; Yasuoka, 2014; Dounias, 2016). To date, the use of "traditional" hunting tools has considerably reduced in Central African rainforests whereas most productive tools, namely metallic snares and shotguns become widespread (Noss, 1998, 2000; Wilkie and Carpenter, 1999; Fa et al., 2005); (Kümpel, 2006).
The wide-spreading of cable snares and shotguns in Africa during the last decades has significantly contributed to the increase of hunting pressure although they are officially banned by forest codes in most countries (Noss, 1998; Wilkie and Carpenter, 1999; Fa et al., 2005; Fa and Brown, 2009). In southeast Cameroon, wire snares constitute the main hunting tool, providing 48 to 95% of the game (Dethier, 1995; Jeanmart, 1998; Muchaal and Ngandjui, 1999; Yasuoka, 2006; Bobo et al., 2015; Yasuoka et al., 2015; Duda et al., 2017).
Studies estimated the total harvest of wildlife in Afrotropical forests to be 1–5 million tons per annum (Wilkie and Carpenter, 1999; Fa et al., 2002). The increase in hunting puts disproportionate pressure in the vertebrate community, resulting in the alteration of his structure. Changes in vertebrate community structure can also alter many of the drivers of tree community dynamics by decreasing the abundance of largeseeded, mammal-dispersed plant species, and increasing the abundance of non-mammal-dispersed tree species (Nasi et al., 2008; Terborgh et al., 2008; Rosin and Poulsen, 2016). Mammals are the main targets of hunting and among them, rodents, and ungulates such as duikers largely constitute the most hunted taxonomic groups in terms of numbers and weight (Fa et al., 2005). Large-bodied animals with low reproductive rates are the most vulnerable to hunting and therefore, the first to be extirpated from hunting forests (Nasi et al., 2008).
Estimating the level at which hunting becomes unsustainable remains challenging. Numerous authors have focused on forest duikers to evaluate the sustainability of hunting given their importance in Central Africa (Dethier, 1995; Muchaal and Ngandjui, 1999; Fimbel et al., 2000; Yasuoka, 2006; van Vliet and Nasi, 2008; Bobo et al., 2015). The Robinson and Redford's (1991) model is the most popular used. This model calculates the sustainable harvest level for a given population based on its population density. Dung survey along linear transects is the method most commonly used to estimate the population density of forest duikers in Central Africa (Fimbel et al., 2000; Lahm, 2001; Bobo et al., 2014; Jost Robinson et al., 2016) as this indirect observation method allows rapid population estimates over large and remote areas. However, compared to nocturnal surveys, dung counts are likely to underestimate the density of forest duikers (Waltert et al., 2006; Jost Robinson et al., 2016; Kamgaing et al., 2018). Also, the difficulty in distinguishing the dungs of different species living in the same area reduces the accuracy of estimates (van Vliet et al., 2008). Mediumsized duiker species (15–25 kg) are generally pooled into "red duikers" to improve the accuracy of density estimates. Hence, the sustainability of hunting for this group of species is often evaluated by considering them as a single taxon, obscuring the differential effects of hunting on individual species (Yasuoka, 2006; Bobo et al., 2015).
Other authors have used the species composition of the animals killed by local hunters as a proxy of the sustainability of hunting (Dethier, 1995; Yasuoka et al., 2015; Fa et al., 2016). In fact, prey composition is influenced by an associated history of hunting and can serve as a good indicator of the status of the surrounding fauna and hunting levels (Dethier, 1995; Bobo et al., 2015; Taylor et al., 2015). Information is available on bushmeat species composition across many sites in Central Africa (Fa et al., 2005, 2016; Taylor et al., 2015). However, although essential to enable the development of plans for conservation, dataset over time on prey composition and abundance is rarely available for the same site.
In this study, we use interviews to investigate the spatial and temporal patterns of changes in the composition of the animals killed using snare traps in two community hunting zones (CHZ) with contrasting hunting pressure and human population densities in southeast Cameroon. We also evaluate the reliability of local knowledge as a tool to assess the status and trends of multiple wildlife species over space and time. Although numerous studies have already proved the robustness of traditional ecological knowledge as a tool for management (Gandiwa, 2012; Pan et al., 2015; Nash et al., 2016; Brittain et al., 2018), our study contributes to improving the current knowledge on the state and trends in wildlife composition by integrating the spatial and temporal scales, although a similar study has been conducted in D.R. Congo (van Vliet et al., 2018). To facilitate comparisons with previous studies, we collated available data on interviews, bushmeat records and ecological surveys that were conducted in the same area since the nineties and contrasted the patterns in animal composition between different sources and time periods.
### MATERIALS AND METHODS
### The Study Area
This study was conducted in CHZ 13 and 14, located in the northern periphery of Boumba-Bek and Nki National Parks, southeast Cameroon (**Figure 1**). The climate of the region is characterized as a four-season equatorial climate with a long dry season from December to February, and a short dry season in July-August. The major rainfall runs from September to November, and the minor rainfall is between March and June (Ekobo, 1998). The mean annual rainfall is 1,500 mm and the average temperature is 24◦C. The major vegetation type in the area is a mixture of evergreen and semi-deciduous forests (Letouzey, 1985).
The Baka and the Kunabembe (a Bantu-speaking population) are the main occupants of southeast Cameroon. Agriculture and harvest and trade of non-timber forest products are the main economic activities of both ethnic groups. The Baka have started subsistence farming several decades ago. Major food crops produced by the two ethnic groups are plantain, banana, and cassava whereas cocoa constitutes the principal cash crop, especially for the Kunabembe. Logging started in southeast Cameroon during the 1970's and gradually enabled the connection of remote villages to market areas. A logging road was built from Yokadouma to Biwala II, at the east side
of Boumba river. Thereafter, part of the residents of "ancien" villages (located in CHZ 14) have deliberately migrated to disenclaved areas beyond the northeast of CHZ 13 and have created the "nouveau" or new villages. According to two local informants in their forties, the road was opened up in CHZ 13 in 1996, first up to Gribe. Thereafter, the road reached Malea Ancien, and stopped at the Bek river, which constitute the north boundary between Boumba-Bek and Nki National Parks. In the rest of CHZ 14, the road was opened in 2002 (Yasuoka, 2006). Human population size was estimated at 4,932 and 2,053 persons inside CHZ 13 and CHZ 14, respectively (Halle, 2000). CHZ 13 and CHZ 14 cover 111,824 and 86,822 ha respectively. Human population density in CHZ 13 is about two times as high as in CHZ 14.
Community Hunting Zones (ZICGC or zones d'intéret cynégetique à gestion communautaire) have been introduced in Cameroon in 1995 (Egbe, 2001). They constitute an attempt at a community-based management to ensure the conservation of wildlife resources by local people and allowing them to legally derive benefits from the exploitation of trophies. The management of a CHZ relies on an agreement between its neighboring communities and the wildlife administration. In practice, CHZ are co-managed by professional guides selected by the wildlife administration with members of the communities, organized into a COVAREF (Comité de valorisation des ressources fauniques). As in 2019, 14 CHZ have been allocated to local peoples in southeast Cameroon.
### Data Collection
Data on hunters' perception of prey composition was collected in June–July 2015, January–February, May–June, and September– October 2016. Informants were selected based on their willingness and availability to contribute to the study. Ethical approval was not required in this study although it meets the ethical guidelines of the Social Research Association (2003). The study was prior informed. Before each interview, we stated explicitly that participation was free and that all information provided would be treated confidentially and anonymously. Verbal consent only was obtained from all informants as most speak French but could not write nor read any language. For informants under the age of 18 (4 out of 255), consent was given by their parents.
Hunters were asked to rank in order of importance the species most commonly killed by their snares in 2015–2016 and during their beginnings in hunting (1952–2009). In southeast Cameroon, most hunters get introduced to snares in their childhood. Young boys often use fiber materials to set "traps" just at the vicinity of village houses. Such activities may not be regarded as effective hunting, but rather as fun games (since children can mimic snaring as conducted by elders). We clarified to each informant that the "beginning in hunting" refers to the year when he effectively started setting snares for himself and, at a considerable distance from his village. This "subjective" qualification of hunting applies well to the settings of southeast Cameroon and probably to those of many other sites in Central Africa. Using it in the interview allowed a clearer understanding of the focus of this research by the informants i.e., hunting for livelihood (which is generally practiced from the adult age and at a certain distance in the forest, where preys are "present").
When the ranking of catch frequencies differed between the beginning in hunting and in 2015–2016, we asked the informants to cite the drivers of changes. To ensure representativeness, interviews were conducted in 9 out of 10 villages encompassing the two CHZ. Contrarily to the Kunabembe, most Baka did not know their age. We asked to those informants to cite one of their Kunabembe neighbors with whom they started primary school. We then visited the Kunabembe in question to asking his age and attributed the same age to the Baka.
In total, 255 hunters of whom 106 in CHZ 13 (43 Baka and 63 Kunabembe) and 149 in CHZ 14 (90 Baka and 59 Kunabembe) were interviewed. Informants were 16–74 years aged and had lived and hunted in the same area for most of their lives. All of them were males. Migrants were not included in our sample, guaranteeing that historical information provided on prey composition effectively reflected the context of the study area. Interviews were conducted primarily in French. Informants who could not speak French were interviewed in Baka or Kunabembe (the two local languages) with the aid of a local translator. To ensure correct communication, informants gave animal names in their respective native tongue (**Table 1**) and species were confirmed together with the data collection team using the pictures provided in the Kingdon field guide to African Mammals (Kingdon, 2015).
### Data Analysis
Villages were categorized into CHZ 13 or CHZ 14 depending on their localization. Three villages were located in CHZ 13
TABLE 1 | Species listed by hunters in southeast Cameroon as the mostly killed using snares and local names used for the interviews.
\*Individual body weights are from Kingdon (2015).
(Massea, Zoka Diba, and Bintom) and six in CHZ 14 (Gribe, Song Ancien, Gouonepoum Ancien, Malea Ancien, Zoulabot Ancien, and Ngatto Ancien). Gribe, which is located at the selvage between the two CHZ, was classified in CHZ 14 where its residents primarily carry out their hunting activities (Bobo et al., 2015).
We considered only the first four species perceived as the most common for analyses because from the fifth species cited, most informants seemed unsure. Hunters who could not remember the age at which they began hunting (21 Baka vs. 9 Kunabembe) were attributed the mean age at which other informants in their respective ethnic group started hunting (14 ± 8 years old for the Baka vs. 16 ± 8 for the Kunabembe). We used the ranking of catch frequencies by local hunters as a proxy for understanding wildlife changes through space and time. Statistical analyses were completed using the statistical software R (R Core Team, 2016). We used a chi-square test of independence (α = 0.05) to determine whether the ranking of catches significantly varied between the time period when informants effectively started snare hunting (1952–2009) and in 2015–2016, when data collection was carried out. Results of the interviews were compared with data from other interviews, bushmeat records, and mammal surveys conducted in southeast Cameroon since the nineties. We used diverse sources of information relative to large and medium-sized mammal abundance in this area (scientific papers, reports, and own unpublished data).
### RESULTS
In average, informants effectively started snaring at 14 ± 5 years old and had experienced 22 ± 11 years in snaring (range: 3–51 years). About 21, 25, 50, and 4% of the informants started snaring, respectively, between 1952–1989, 1990–1999, 2000–2009, and 2010–2014 (**Table 2**).
Overwhelming proportions of the informants (79% in CHZ13 vs. 95% in CHZ 14) reported that the current catch frequency of larger preys such as the red river hog (Potamoecherus porcus) and red duikers (Cephalophus spp.) is considerably low compared to the period between 1952 and 2009 (we excluded the informants who started hunting after 2009). However, 20% of the respondents in CHZ 13 and 5% in CHZ 14 did not perceive any substantial change in the composition of their catches over time and one informant (37 years old) in CHZ 13 reported that his
TABLE 2 | Distribution of hunters according to the period when they started snare hunting by themselves.
\*Because of low sample sizes in both CHZ, we excluded the time period [2010–2014] in Figures 2, 3.
catch frequency has actually increased as he has developed more hunting skills throughout years.
Local hunters have attributed these declines to a number of factors (**Table 3**) including logging, which might make the animals "fleeing" to remote forest areas (77% of informants in CHZ 13 vs. 67% in CHZ 14; X² = 0.39, df = 1, p = 0.533), overhunting (65 vs. 70%, X² = 0.99, df = 1, p = 0.319), demographic growth (44 vs. 7%, X² = 22.28, df = 1, p < 0.001), witchcraft (4 vs. 0%), and farming (4 vs. 6%). Two respondents in CHZ 14 (one Baka and one Kunabembe) perceived that the decline in large preys resulted also from the restriction of their hunting grounds within smaller forest blocks, following the establishment of protected areas. Nine percent of the respondents in the CHZ 13 and 11% in CHZ 14 had no opinion about the driver of prey alteration (X² = 0.2, df = 1, p = 0.666). By ethnic groups, the main drivers of decline in larger game species were also logging (76% Kunabembe; 70% Baka X² = 0.449; df = 1, p = 0.503) and overhunting (57; 80%; X² = 7.9; df = 1, p < 0.01). Compared to the semi-nomadic Baka (11%), a significantly higher proportion of Kunabembe (46%) perceived demographic growth as a major driver of prey alteration (X² = 28.2; df = 1, p < 0.001). Thirteen percent of the Baka and 6% of the Kunabembe did not know why the composition of harvested animals has changed over time (X² = 1.7, df = 1, p = 0.189).
In CHZ 13, where human population density was higher, the species perceived to be the most commonly captured by snare hunters in 1952–2009 (**Figures 2**, **3**) were the blue duiker Philantomba monticola (78%), followed by the brush-tailed porcupine Atherurus africanus (55%), bay duiker Cephalophus dorsalis (30%), and Peters duiker C. callipygus (17%). As in 2015– 2016, significantly higher proportions of informants perceived the blue duiker (93%; X² = 8.9; df = 1, p < 0.005) to be the most frequent prey, followed by the brush-tailed porcupine (71%; X² = 4.8; df = 1, p < 0.05). The Peter's duiker became the third most frequent prey (40%, X² = 11.8; df = 1, p < 0.001) overtaking the bay duiker (27%; X² = 0.1; df = 1, p = 0.791.). However, the tree pangolin Phataginus tricupsis, which was not among the main preys as in 1952–2009, became the fourth most common in 2015–2016 (24%, X² = 1.00; df = 1, p = 0.318). Although not significant, the harvest frequency of the giant pouched rat Cricetomys emini has also increased between the two periods (X² = 3.4; df = 1, p = 0.066). Fifteen percent and 13% of the hunters perceived, respectively, the red river hog and the yellowbacked duiker (Cephalophus silvicultor) to have been among the four most frequent preys in 1952–2009. However, only 3% (X² = 8.1; df = 1, p < 0.005) and 1% (X² = 10.0; df = 1, p < 0.005), respectively, perceived the same as for 2015–2016.
In CHZ 14, where hunting pressure was lower, the blue duiker (54%) was also the most common prey in 1952–2009. In contrast to CHZ 13 (where the brush-tailed porcupine had already emerged as the second main quarry), the second most common prey in CHZ 14 was the Peter's duiker (32%), followed by the bay duiker (34%), and red river hog (24%).
It is worth noting that perceived prey composition in CHZ 14 as in 2015–2016 is similar to the depiction made by hunters in the more heavily hunted CHZ 13 in 1952–2009. As in 2015–2016, higher proportions of informants in CHZ 14 perceived the blue duiker to be the most common prey (92%; X² = 51.9; df = 1, p < 0.001). The brush-tailed porcupine became the second most hunted species for a significantly higher proportion of informants (58%; X² = 46.0; df = 1, p < 0.001), followed by the bay duiker (37%; X² = 0.23; p = 0.631), and the Peters duiker (perceived as the fourth most common catch by 5 and 15% of the informants in 1952–2009 and 2015–2016, respectively; X² = 8.01; df = 1, p < 0.005). The proportion of informants who perceived the Peters duiker as the second most hunted game species has slightly decreased in 2015–2016 (26%; X² = 1.3; df = 1, p = 0.254).
Compared to CHZ 13, higher proportions of informants in CHZ 14 perceived, respectively, the red river hog (40%) and the yellow-backed duiker (21%) to be among the four most common preys in 1952–2009. However, only seven percent of the informants (X² = 56.7; df = 1, p < 0.001) perceived the red river hog to still being common in 2015–2016, whereas the yellowbacked duiker had disappeared among the four principal preys.
### DISCUSSION
The knowledge accumulated over generations by local people in the use of natural resources can provide valuable insights for sustainable management (Pan et al., 2015; Nash et al., 2016; Duda et al., 2017; Brittain et al., 2018; van Vliet et al., 2018). We analyzed the perceived state and trends in species composition of the animals killed using snare traps (here considered as a proxy for understanding wildlife changes through space and time) in two CHZ with contrasting human population densities in southeast Cameroon. Our approach combines spatial and temporal changes in prey composition as perceived by local hunters themselves with the trends from ecological and ethnographic surveys of hunted species.
### Perceived Trends in the Composition of Hunting Catches
If we assume that a perceived regression in the catch frequency of a given animal species indicates a reduction in its abundance, a reported increase in the catch frequency would not necessarily imply an increase in its population abundance, but rather a growing hunting pressure on that population.
Informants in both CHZ claimed to harvest smaller prey species such as the blue duiker, porcupine, and giant pouched rat more frequently in 2015–2016 than in 1952–2009. However, substantial declines in the catches of larger bushmeat species were reported, especially in CHZ 14 where anthropogenic pressure was lower. Among the catches, the red river hog, the yellowbacked duiker, and red duikers had the sharpest declines. The white-bellied duiker (Cephalophus leucogaster), the black-fronted duiker (Cephalophus nigrifrons), and the yellow-backed duiker are extremely scarce or have almost disappeared among the catches. These results provide further evidence that hunting in southeast Cameroon has resulted in an increase of the proportion of blue duikers killed in snare traps and a decline in the proportion of red duikers (Dethier, 1995; Jeanmart, 1998; Yasuoka et al., 2015).
TABLE 3 | Perceived drivers of declines in larger bushmeat species by ethnic group and by community hunting zone (CHZ) in Southeast Cameroon.
Total proportions exceeds 1.00 because respondents were allowed to give multiple answers.
df = 1 for all analyses. Bold values indicate significant differences between groups (α = 0.05).
Why the perceived decline of larger species is not greater in the area of higher anthropogenic pressure is an interesting question, which suggests the existence of shifting baseline syndrome (SBS) as a potential explanation of the perceived differences. SBS can be described as a psychological and social phenomenon whereby each new human generation conceives as normal or as a reference the situation in which it was raised, due to a lack of experience, knowledge, and/or memory (Pauly, 1995; Bonebrake et al., 2010). Simply put, the concept refers to the loss of perception of change which occurs when a new generation redefines new norms. Therefore, local perception of prey profiles in 1952–2009 does not reconstruct the initial state of prey assemblages/species abundance, but attempts at depicting its depletion level during that period, a level which informants might consider as normal.
Thus, declines in larger preys were more difficult to perceive in the heavily hunted CHZ 13 because during our study period, populations of larger game species had already considerably reduced, probably as a result of logging and overhunting. The more a species was rare, the more it was difficult to perceive any change in its catch frequency, especially for the younger hunters. This type of SBS is otherwise referred to "generational amnesia" (Kahn and Kellert, 2002; Papworth et al., 2009). However, in CHZ 14, where logging was relatively recent (Yasuoka, 2006) and populations of larger game species considerably more abundant (Bobo et al., 2015), wildlife declines might be more recent. Consequently, perceiving the temporal changes in the catches of larger animal species could have been somewhat easier to local peoples, including the younger hunters.
Independently to whether the arguments developed above apply to the settings of our study area or not, remembering ancient hunting activities with accuracy could be difficult, especially for the oldest informants. Memory recalls of hunting returns could also be biased, since this activity in its essence may be influenced by narratives of declines, which create "false memories" or "memory illusions" (Hyman and Pentland, 1996; Roediger et al., 1996). In conclusion, perceptions of past prey profiles, which represent what some informants think and/or relate, might constitute a relatively altered vision of the real.
Hunting may affect different species of red duikers differently. Beside the blue duiker (Philantomba monticola) and the yellowbacked duiker (Cephalophus silvicultor), four duiker species grouped as "red duikers" (Cephalophus callipygus, C. dorsalis, C. leucogaster, and C. nigrifrons) live in southeast Cameroon (Ekobo, 1998). In both CHZ, the latter two were cited as among the four main preys by <4% of the hunters during their beginnings in hunting. Only 1% of the hunters in CHZ 14 mentioned them among their most common preys in 2015– 2016 whereas they have disappeared among the main catches in CHZ 13. Kingdon (2015) argued that the white-bellied duiker is also the least commonly killed duiker in the D.R. Congo and probably the less abundant. In contrast, the two other species of red duikers (Peter's duiker and bay duiker) were perceived to have always been among the four main catches, despite the perceived retrogression in their ranking position, as reported earlier. In the two study sites, the levels of perception of Peter's duiker as being among the four top preys remained similar between the two periods (CHZ 13: 52% in 1952–2009 vs. 46% in 2015– 2016, X² = 0.0; df = 1, p = 1.0); CHZ 14: 79 vs. 78%, X² = 0.7; df = 1, p = 0.405). In CHZ 13, the bay duiker was cited as among the main preys by 73 and 42% of the informants in 1952–2009 and in 2015–2016, respectively (X² = 19.7; df = 1, p < 0.001). In CHZ 14, this game species was among the main preys of 70% of the informants in 1952–2009 and 76% in 2015– 2016 (X² = 1.1; df = 1, p = 0.298.). A possible explanation of the differences in catch frequencies is that in southeast Cameroon, the black-fronted duiker and the white-bellied duiker might be naturally less abundant as reflected in hunting catches (Dethier, 1995; Yasuoka, 2006; Bobo et al., 2015; Duda et al., 2017). Another is that these four medium-sized duiker species existed at higher densities in the past, but the effect of hunting has been more severe on the black-fronted duiker and the white-bellied duiker.
Hunting can also affect the same species differently between different sites. For example, in northeast Gabon (where the same group of red duikers living in southeast Cameroon co-exist), nocturnal transect surveys conducted in the eighties shown that among duikers, the white-bellied duiker and the black-fronted duiker were naturally the least abundant (Feer, 1988). The blue duiker was largely the most abundant, followed by the bay duiker and Peters' duiker. However, van Vliet et al. (2007) repeated the surveys after two decades and concluded that the blue duiker remained the most abundant, but the bay duiker was depleted as a result of hunting, while the Peters' duiker was still withstanding the hunting level. Similarly, interview data have shown that in Baka villages around Lomié and Messok (more densely humanpopulated areas in southeast Cameroon), contrarily to the whitebellied duiker and the black-fronted duiker, only the blue duiker, followed by the Peter's duiker and the bay duiker still appear among duiker catches (Duda et al., 2017).
### Trends From Density Estimates and Bushmeat Records
As far as we know, empirical density variations of hunted species over time have not been examined in southeast Cameroon. However, studies indicate that in Central Africa, reductions in mammal densities between unhunted and hunted sites can vary between 13 and 100% (Fimbel et al., 2000; Hart, 2000; Lahm, 2001). As for today, at least five publications of mammal density estimates are available in the Boumba-Bek and Nki area. Out of them, one was carried out in the nineties (Ekobo, 1998), two in the 2000's (Bene Bene and Nzooh-Dongmo, 2005; Nzooh Dongmo et al., 2006), and two in the 2010's (Bobo et al., 2014; Kamgaing et al., 2018).
These studies used dung surveys and estimated higher population densities for red duikers. (2.5–20.0 animals km−²), followed by the blue duiker (0.1–10.6 animals km−²), although absolute values were comparable in some cases. However, nocturnal surveys conducted in our study area (Kamgaing et al., 2018; own unpublished data) reported the converse, with density estimates of the blue duiker (59.8 animals km−²) at least six times as high as that of red duikers (9.0), brush-tailed porcupine (6.6), and tree pangolin (4.1). Jost Robinson et al. (2016) argued that such high density of the blue duiker occurred in forests where hunting is relatively recent.
To our knowledge, records of bushmeat harvests in southeast Cameroon also began in the 1990's (Dethier, 1995; Jeanmart, 1998; Muchaal and Ngandjui, 1999; Fimbel et al., 2000; Yasuoka, 2006, 2014; Bobo et al., 2015; Yasuoka et al., 2015). All these studies have clearly shown that in hunted forests, there is a spatial heterogeneity of human pressure and faunal assemblages. For example, the catch frequency of red duikers in snares can be 3–23 times as high as the catch frequency of the blue duiker, depending on human population density and the distance between the hunting area and the main road or settlement (Yasuoka, 2006). In remote zones (here located at 10–22 km from the main road), red river hogs were captured more than the blue duiker. However, in areas where hunting is more intense (<10 km from the main road), red duikers were still the most hunted species, but the blue duiker was more prevalent than the red river hog. A village-based record of animal carcasses conducted 10 years later in the same site suggested the converse (Bobo et al., 2015), although differences were more moderate (blue duikers were captured 2.4 times as much as red duikers). A possible explanation of such patterns is that in areas under low or moderate human pressure, the density of red duikers is higher than that of blue duikers whereas the reverse is likely to be observed in areas under moderate or intense hunting pressure as suggested by Yasuoka et al. (2015).
However, the hunting strategy, rather than differences in population densities may explain why the capture frequency of less abundant species (e.g., red river hog) can exceed that of more abundant species such as red duikers and blue duikers in remote forest areas. Although the wire snare technology is known to be typically non-selective (Noss, 1998, 2000; Dounias, 2000), it allows the possibility to target in preference large body-sized animals such as red river hogs in areas where the signs of their activity are abundant. Snare hunters can target larger-sized preys by increasing the number of strands of the wire snares (Yasuoka, 2014; Dounias, 2016), by decreasing the sensitivity of the trigger mechanism, and by selecting a tougher support stick. Yet, the point that hunters can target in preference larger species by adjusting the snares design does not invalidate the use of prey composition as a reliable proxy of species abundance. Because larger preys are generally the first to be targeted, populations of these wildlife species are usually lower in forests near-human settlements. In conclusion, the influence of the previously described snaring setting on the global off take and on the structure of game composition may remain marginal, since it is seldom practiced in isolated areas where hunting remains infrequent and, other medium body-sized species relatively abundant.
### Similarities and Discrepancies in Wildlife Trends Between Hunters' Perception, Ecological Surveys and Bushmeat Records
Previous studies have shown that larger species are usually sent to markets whereas most of the carcasses from smaller species are consumed in villages. Since hunting occurs both for consumption and trade in our study area, market demand can potentially affect the choice of prey (or prey composition) and species abundance. Even if this is the case for both CHZ, perceptions of prey composition in our data may still reflect species abundance,
TABLE 4 | Most common bushmeat species according to local hunters (n = 169) in CHZ 13 and CHZ 14 in 1952–1999 and according to studies conducted before 2000.
<sup>a</sup>Ekobo (1998).
<sup>b</sup>Fimbel et al. (2000).
<sup>c</sup>Dethier (1995).
\*Cephalophus callipygus, C. dorsalis, C. leucogaster, and C. nigrifrons. Data from hunters who started hunting after 1999 were excluded.
given that exclusively hunters were interviewed. If interviews were administered away from hunting sites (e.g., in bushmeat markets), game composition, and abundance could have been biased in favor of larger species. Yet, this was not the case since in both CHZ, our data reflect well the spatial and temporal increases in small body-sized preys such as giant pouched rat and tree pangolin which are usually overlooked by market data.
Local knowledge of prey composition concurs with villagebased bushmeat records (Fimbel et al., 2000) for the two species most commonly killed in snare traps before the 2000's (**Table 4**). In fact, the majority of informants (66%) all over the study area reported that as in 1952–2009, the blue duiker was the most common prey. Nearly half of the informants cited one of the red duiker species as the second (49%) or the third (51%) major game during the same period. However, Dethier (1995) reported a similar, though slightly higher share of red duikers compared to blue duikers (1.1 times lower) in village-based bushmeat data. Such similarity in the proportions of harvested red duikers and blue duikers may indicate that hunting pressure was considerably low or moderate during Dethier (1995)'s study and the density of red duikers excided (or was similar to) that of the blue duiker, as argued by Yasuoka et al. (2015).
As in 2015–2016 (**Table 5**), significantly higher proportions of hunters (94% in CHZ 13 vs. 92% in CHZ 14) perceived that the blue duiker is the most common prey, as observed in villagebased carcass records (Yasuoka, 2006; Bobo et al., 2015) and in nocturnal transect surveys (Kamgaing et al., 2018).
However, the trends derived from dung surveys and bushmeat data collected in remote forest areas call into question the insights from interviews. Opinions on the two most abundant preys both before 2000 (**Table 4**) and in 2015–2016 (**Table 5**) controvert the trends observed in most camp-based carcass records (Dethier, 1995; Fimbel et al., 2000; Yasuoka, 2006) and dung surveys (Ekobo, 1998; Bobo et al., 2014), which actually suggest that red duikers, followed by the blue duiker are the most abundant. This apparent discrepancy is understandable acknowledging that in African rainforests, dung counts are likely to underestimate the density of forest ungulates, especially for the blue duiker (Waltert et al., 2006; Viquerat et al., 2012; Jost Robinson et al., 2016; Kamgaing et al., 2018). Another reason why the blue duiker instead was perceived as the most common prey may be that hunting is more frequent in forests around human settlements, where small body-sized games (generally the most resistant to hunting) are likely to be more common following the depletion of larger preys (Koerner et al., 2016). A third reason might be that in this study, informants were not asked to rank catch frequencies by forest area, but to sort them out globally. A future investigation of the former could explicitly highlight the differences between forest areas regarding animal abundance, since most informants reported that larger prey species are more abundant in 'remote' forest areas (own-unpublished data).
Studies based on dung counts in Southeast Cameroon (**Table 6**) suggest that since the 1990's red duikers, followed by blue duikers have always been the most abundant preys among the main bushmeat species (Ekobo, 1998; Bene Bene and Nzooh-Dongmo, 2005; Nzooh Dongmo et al., 2006; Bobo et al., 2014). However, this survey method is not practicable to estimating population densities for important bushmeat species such as the brush-tailed porcupine and tree pangolin (largely because they usually hide in burrows and trees, respectively) and thus, masks their potential importance among terrestrial mammals. Actually, high proportions of hunters (71% in CHZ 13 vs. 58% in CHZ 14) agreed that as in 2015–2016, after the blue duiker, the brush-tailed porcupine was the most abundant prey (**Table 5**). If we ignore the records of the brush-tailed porcupine in the interview dataset (since transect sampling methods do not usually provide density estimates for this species), insights from hunters' perception would largely support those from nocturnal surveys and village- based carcass records that after the blue duiker, red duikers constitute the second most common prey. It is important to note that population densities provided for the brush-tailed porcupine and tree pangolin using direct observations along nocturnal line transects are likely to be underestimates for the same reasons mentioned above (own unpublished data), although both species are nocturnal (**Table 6**).
Our data on local knowledge suggest that snare hunting has potentially affected red duikers more severely than the blue duiker as expected from classic hunting models. Red duikers were the most common prey for 26% of the informants before 2000 (**Table 4**). However, only 5% of the informants in CHZ 14 and 1% in CHZ 13 have respectively, cited a species from this group as being the most commonly killed by snares in 2015– 2016 (**Table 5**). Similarly, nearly half of the informants (49%) perceived red duikers as the second predominant prey during their beginnings in hunting. However, only 33% in CHZ 14 and 20% in CHZ 13 perceived that in 2015–2016, a species of this group still constitutes the second predominant game. In contrast to the red duikers, the blue duiker has maintained its population as the most common prey. In fact, the blue duiker TABLE 5 | Most abundant bushmeat species according to snare hunters in CHZ 13 (n = 105) and in CHZ 14 (n = 148) in 2015–2016.
\* "Red duikers" are constituted of Cephalophus callipygus, C. dorsalis, C. leucogaster and C. nigrifrons.
Data from hunters who stopped hunting before 2000 were excluded.
has always been the most common prey for at least 65% of the informants in this study. As for 2015–2016, this species was the most hunted for more than 91% of the informants in both CHZ. Other small body-sized species such as the brush-tailed porcupine, tree pangolin and giant pouched rat also became more common in catches, especially in CHZ 13, where human pressure is higher.
Insights from bushmeat harvest data in more heavily hunted areas also go in that direction. For example, Duda et al. (2017) analyzed the memory recalls of bushmeat harvest in Baka villages in Lomié and Messok districts, two towns, respectively, located at about 185 and 130 km west of our study area. The data from Duda et al. (2017) show that 32 species were reported as harvested. Among them, three species contributed up to 65% of the total number of catches, first the giant pouched rat (28%), followed by the blue duiker (28%), and brush-tailed porcupine (9%). Primate species constituted 12% of the total off take. Another study led in Southwest Cameroon found that local hunters perceived the catch frequencies of rodents (brush-tailed porcupine, giant cane rat Tryonomis swinderianus, and giant pouched rat) to exceed that of the blue duiker and red duiker species (Wright and Priston, 2010). In conclusion, alterations of wildlife assemblages may be more recent and/or less severe in southeast Cameroon than in other regions like southwest Cameroon, where human population density is higher. Such patterns indicate differential levels of wildlife depletion between areas, as suggested in Gabon (Lahm, 1993; Koerner et al., 2016).
As mentioned above, the findings from Duda et al. (2017) indicate that despite their low individual body size, giant pouched rats can be captured as much as (or even more than) blue duikers (**Table 6**). In total 95% of the captures from that cavedwelling rodent was obtained by smoking individuals out of their burrows, whereas most blue duikers were killed using shotgun and snares. Although studies conducted in our research site and elsewhere in Southeast Cameroon do not report smoking out as a hunting method among local practices (Dethier, 1995; Jeanmart, 1998; Fimbel et al., 2000; Bobo et al., 2015; Yasuoka et al., 2015; but see Hagino, 2015), our field observations indicate that this technique is seldom practiced in our study site, and usually targets the brush-tailed porcupine. Smoking out of animals might emerge as a hunting method as other terrestrial/arboreal animals become relatively rare. Such insights indicate that TABLE 6 | Most abundant bushmeat species based on data from ecological and ethnographic studies conducted in southeast Cameroon since the 2000's.
<sup>a</sup>Duda et al. (2017).
<sup>b</sup>Bene Bene and Nzooh-Dongmo (2005).
<sup>c</sup>Nzooh Dongmo et al. (2006).
<sup>d</sup>Bobo et al. (2014).
<sup>e</sup>Bobo et al. (2015).
<sup>f</sup> Yasuoka et al. (2015).
Data from hunters who stopped hunting before 2000 were excluded.
"Red duikers" are constituted of Cephalophus callipygus, C. dorsalis, C. leucogaster and C. nigrifrons.
\*Numbers in brackets represent population density estimates (ind.km−<sup>2</sup> ), calculated using Distance 6.2 software (Thomas et al., 2010). Kamgaing et al. (2018) provided published data on population density estimates for duikers and unpublished data for the tree pangolin and brush-tailed porcupine.
hunters may shift their hunting techniques depending on the relative abundance of animal species, as suggested by Fa and Peres (2001).
However, untangling the effects of different hunting methods on the dynamics of game species was beyond the scope of the present study. Since our sampling approach focused on snare hunting, our analyses ignore the effects of other hunting methods such as shotgun and the contribution of other groups of prey species (e.g., arboreal primates) to the hunting profile. Nevertheless, our study still provides insightful findings on the spatial and temporal patterns in terrestrial bushmeat species in Southeast Cameroon, where wire snares remain the most widely used hunting tool and terrestrial mammals the main preys, as reported in other sites in the Congo basin (Wilkie and Carpenter, 1999; Fa and Peres, 2001; Fa et al., 2006; Dounias, 2016). We recommend analysis of the effects of hunting on prey profiles based on multiple hunting technologies, which allows reflectiveness of a more diverse range of game species.
In conclusion, local knowledge suggests a decline in the catch frequency of medium body-sized preys such as the red river hog, the yellow-backed duiker, and red duikers. This change is balanced with an increasing catch of small preys such as blue duikers, brush-tailed porcupines, tree pangolins, and giant pouched rats, which is congruent with global trends in species composition in hunting forests (Nasi et al., 2008; Fa and Brown, 2009; Koerner et al., 2016).
### Implication for Sustainable Management
Stakeholders, especially the local peoples may more actively engage in participative management if a common understanding of the threats and trends in different species is met. This study has shown that local people have an acute awareness of wildlife changes, notably the declining populations of medium body-sized preys, coupled to an increasing share of smaller preys. Furthermore, local knowledge can supplement robust information on the abundance of wildlife populations, even for the taxa for which densities are difficult to estimate using more classic survey methods such as individual species of mediumsized duikers, medium-sized monkeys, red river hog, porcupines, and pangolins, important targets of bushmeat trade in West-Central Africa.
Trends in species abundance from local knowledge and bushmeat records suggest some dichotomy regarding the protection level of different species of red duikers according to the norms of attribution of hunting quotas in Cameroon (Government of Cameroon, 1998). In fact, the categorization of animal species into different classes of protection has been set up inspired from the checklist of the Convention on International Trade of Endangered Species of Wild Fauna and Flora (CITES). Accordingly, the bay duiker and the Peters' duiker are in class B, which fix the species which are partially protected and for which hunting may be granted by a license. The black-fronted duiker and the white-bellied duiker are in class C which includes species such as the blue duiker, some rodents and small carnivores that can be hunted moderately, without a hunting license. However, as previously suggested, among the red duiker species, the whitebellied duiker and the black-fronted duiker are by far the rarest. It appears that a lower level of protection is attributed to the least abundant (and probably the most threatened) species of red duikers. This calls for the need to update this classification, at least in southeast Cameroon.
The general outlines that emerge from this study are: (1) insights from local people can rapidly help to identify coarse changes in prey abundance, even for the species for which population density is difficult to estimate using more conventional survey methods; (2) in southeast Cameroon and in forests regions with similar fauna, the bay duiker, and the Peters duiker are substantially more abundant than the white-bellied duiker and black-fronted duiker; (3) larger species are more represented in hunting catches in CHZ 14 than in CHZ 13, reflecting the spatial and temporal variations of hunting and animal abundance at the local level, as suggested in other hunting forests in Central Africa (van Vliet and Nasi, 2008; Koerner et al., 2016); and (4) the ranking of prey abundance by local hunters is likely to reflect more the faunal assemblages in anthropogenic forest mosaics, where most hunting activities are undertaken (Bobo et al., 2014) and is congruent with villagebased bushmeat records.
As far as we know, this study provides the most comprehensive survey combining spatial and temporal trends in bushmeat species in Cameroon to date. Overall, resource user-based interviews are useful to rapidly provide or supplement valuable information on wildlife population dynamics over years, especially in cases where ecological data may be absent. However, extrapolations of local perceptions over large geographical scales should be made with caution, since local knowledge is typically site specific (Gandiwa, 2012). The acute awareness of local hunters about the declining sizes of their largest prey populations suggests a potential for synergy with more effective participative management initiatives. We therefore, advocate an increase use of local knowledge to design new studies or to seek for adaptive management options, which are acceptable for local peoples and other stakeholders.
### AUTHOR CONTRIBUTIONS
HY and TK conceived the methodology and analyzed the data. TK and ZD collected the field data. TK wrote original draft. All authors have reviewed, edited, and approved the current version of the manuscript.
### ACKNOWLEDGMENTS
This research was supported by the JST-JICA SATREPS program (Projet Coméca), JSPS KAKENHI (Nos. 15H02598, 16H05661), Japan, and Institute for Agronomic Research and Rural Development (IRAD), Cameroon. Permission for field work was granted by the Ministry of Scientific Research and Innovation (MINRESI), Cameroon. We are grateful to the people of the study area for contributing this research and their warm hospitality.
### REFERENCES
Tropical Forest, eds J. G. Robinson and E. L. Bennett (New York, NY: Colombia University Press), 356–374.
**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Copyright © 2019 Kamgaing, Dzefack and Yasuoka. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
# Bioeconomic Modeling of Hunting in a Spatially Structured System With Two Prey Species
Anders Henrik Sirén1,2 \* and Kalle Parvinen3,4
1 Inti Anka Taripay, Puyo, Ecuador, <sup>2</sup> Department of Geography and Geology, University of Turku, Turku, Finland, <sup>3</sup> Department of Mathematics and Statistics, University of Turku, Turku, Finland, <sup>4</sup> Evolution and Ecology Program, International Institute for Applied Systems Analysis (IIASA), Laxenburg, Austria
Although it is well-known and documented that subsistence hunting in the tropics typically takes place in systems characterized by multiple prey species, and that are spatially structured, as hunting effort decreases with the distance from settlements and transportation routes, bioeconomic harvest models tend to be single-species and non-spatial. This paper presents a bioeconomic model that incorporates transport costs and handling costs, as well as two prey species, which interact by being hunted together. In particular, it focuses on how different parameters, corresponding to variability in ecological, socio-economic, and technological characteristics, affect two key dependent variables related to the distance from settlements, or transportation routes, namely (a) the extinction distance, i.e., the distance up to which one of the species, in some cases, becomes extirpated due to excessive hunting, and (b) the no-harvest distance, i.e., the distance beyond which no hunting takes place and the species in question persists at natural levels of abundance. Model results indicate, among other things, that the extinction distance and the no-harvest distance are piecewise smooth functions, which abruptly change slope at certain parameter values.
#### Edited by:
Robert Nasi, Center for International Forestry Research, Indonesia
#### Reviewed by:
Viorel Dan Popescu, Ohio University, United States Katharine Anne Abernethy, University of Stirling, United Kingdom
> \*Correspondence: Anders Henrik Sirén anders.siren@utu.fi
#### Specialty section:
This article was submitted to Conservation, a section of the journal Frontiers in Ecology and Evolution
> Received: 01 February 2019 Accepted: 25 June 2019 Published: 17 July 2019
#### Citation:
Sirén AH and Parvinen K (2019) Bioeconomic Modeling of Hunting in a Spatially Structured System With Two Prey Species. Front. Ecol. Evol. 7:268. doi: 10.3389/fevo.2019.00268 Keywords: extinction, transport, handling, central place foraging, bushmeat, wildlife, bioeconomic equilibrium, tropics
## BACKGROUND
Excessive hunting in tropical forests, whether for subsistence or commercial purposes, is a major threat to biodiversity as well as to the well-being for the people who depend on hunting for their livelihood (Cawthorn and Hoffman, 2015; Ripple et al., 2016). It is well-known and documented that subsistence hunting in the tropics typically takes place in systems characterized by multiple prey species, and that are spatially structured, as hunting effort decreases with the distance from settlements and transportation routes (e.g., Peres and Lake, 2003; Smith, 2003; Sirén et al., 2004; Sirén, 2012). Bioeconomic modeling has become an important tool in order to understand how different socioeconomic, technological, or institutional parameters affect wildlife harvest and abundance. Their usefulness is, however, limited by that they typically are non-spatial, i.e., do not take into account transport costs, and are based on a single prey species. Some such models do take into account either transport costs (Ling and Milner-Gulland, 2008; Sirén et al., 2013; Sirén and Parvinen, 2015; Robinson, 2016) or more than one prey species, whether two (Milner-Gulland and Mace, 1998, pp. 71–77) or multiple (Damania et al., 2005). A bioeconomic model of hunting that includes transport costs as well as more than one prey species is, however, almost absent. One such model was published by Keeling et al. (1999), but the particularity that it involves transport in an infinitively (!) large truck makes generalizing its results a bit problematic.
A related field, with different roots, is that of optimal foraging theory, which has been extensively used in anthropological studies of hunting, although it was originally borrowed from ecology (Charnov, 1976; Stephens and Krebs, 1986). Optimal foraging models often deal with choice of prey among multiple species present (Winterhalder, 1981; Hames and Vickers, 1982; Alvard, 1993). Later models have also included the transport costs for human central-place foragers (Levi et al., 2011). Optimal foraging models, however, deal primarily with explaining or predicting hunters' behavior in the short term, more than with the long-term outcomes and sustainability aspects.
The inclusion of spatial variability and multiple prey species in harvest models could have important implications for the way we think about hunted species and how they could be sustainably managed. According to the standard non-spatial bioeconomic harvest model (Clark, 1976; Milner-Gulland and Mace, 1998), the only variable whose magnitude people could adjust in order to improve sustainability and long-term benefits is hunting "effort." In real life, however, this is difficult to control, and management strategies based on spatial controls, possibly different for different species, might be more feasible. The lack of stringent theoretical harvest models that allow incorporation of such measures, however, might hamper the development of such management strategies. Moreover, in the standard model (Clark, 1976; Milner-Gulland and Mace, 1998), as well as in its spatial version (Sirén and Parvinen, 2015), extinction is impossible, because as a species gets less abundant, hunting ceases as the increased search time required makes it unprofitable. In real life, however, local extirpations do frequently occur, and one important mechanism of this is that even though the abundance of one species might get so reduced that hunting it alone would not be profitable, hunting nevertheless continues because of the presence of other species, which are more resilient to hunting (e.g., Stirnemann et al., 2018). Thus, spatial two-species models could be very helpful in order to understand the mechanisms leading to such local extirpation.
Considerable research efforts have been made in order to find out how variability in income, wealth, and general socioeconomic development affect wildlife harvest and abundance (Shively, 1997; Overman and Demmer, 1999; Wilkie and Godoy, 2001; Apaza et al., 2002; Demmer et al., 2002; Godoy et al., 2010; Foerster et al., 2012; Vasco and Sirén, 2016). The results from such studies are, however, often inconclusive or contradictory to each other, and one reason for this is that economic development tends to lead to simultaneous changes of several different parameters. This makes it difficult to empirically determine which parameter has which effect, and therefore, bioeconomic models have an important role, as they permit analyzing the effects of each parameter separately.
The purpose of this paper was to present a spatial two-species bioeconomic model, focusing on how different parameters, corresponding to variability in ecological, socioeconomic, and technological characteristics, affect two key dependent variables related to the distance from settlements or transportation routes, namely, (a) the extinction distance, i.e., the distance up to which a particular species becomes extirpated due to excessive hunting and (b) the no-harvest distance, i.e., the distance beyond which no hunting takes place and the species in question persists at natural levels of abundance (carrying capacity).
### MODEL ASSUMPTIONS
The parameters and output variables of the model are listed in **Table 1**. The model is based on a common equation of resource growth with harvesting,
$$\frac{dN}{dt} = rN\left(1 - \frac{N}{K}\right) - H,\tag{1}$$
where r is the intrinsic growth rate, N is the population size, and K is the carrying capacity. The harvest H is
$$H = q \text{SN},\tag{2}$$
where q is the catchability coefficient and S is what is usually called "effort," but we prefer the more exact term search labor. To this, finally economic parameters are added: the cost per unit of labor, c, and the market price for one unit of harvested resource, p. Thus, the profit, 5, is:
$$
\Pi = \mathfrak{p}H - \mathfrak{c}\mathbb{S} \tag{3}
$$
In this basic, non-spatial, model, originally developed by Clark (1976) for fisheries and adopted by Milner-Gulland and Mace (1998) for hunting, the only cost the hunter incurs is the time cost of searching for prey. Later models have included also the time cost of transport (Ling and Milner-Gulland, 2006; Sirén and Parvinen, 2015) and the cost of handling the prey (Sirén and Parvinen, 2015). Whereas Sirén and Parvinen (2015) expressed handling as the cost of time divided by the handling speed, we here have chosen to instead use the cost of time multiplied by variable handling time cost, th, in order to facilitate comparison with optimal foraging models, where this is the standard (Charnov, 1976; Stephens and Krebs, 1986; Levi et al., 2011). In addition to the time needed in order to pursue, shoot, and eviscerate an animal, we also include the cost of ammunition in this parameter, because that, too, is directly proportional to the number of prey hunted and has been empirically shown to have significant effects on prey choice (Sirén and Wilkie, 2016). This handling time cost could be expressed just as well in time units or in monetary units, and we have chosen to do the former. Thus, for ammunition, this corresponds to the time it takes to earn the money to buy it. Thus, whereas the total cost, C, in the standard model is simply C = cS, in the spatial model, instead, the total cost in each patch is
$$C = c \left( S + H \eta\_t + \frac{\varkappa H}{\nu\_t} \right) \tag{4}$$
where t<sup>h</sup> is the handling time, v<sup>t</sup> is the speed of transport, and x is the distance from a "central place" (corresponding to, e.g., a village, a road, or a trade point) from which hunters depart and to which they return with the hunted prey after hunting, in a one-dimensional space, consisting of an infinite number of equidistant and equally sized patches. In this model, as shown
by Sirén and Parvinen (2015), different parameter values lead to very distinct spatial patterns of resource abundance and harvest. It can be noted also that introducing the costs of handling and transport to the model renders the term "catchability coefficient" for the parameter q somewhat inadequate, because it represents no longer the probability of a certain individual animal to actually get hunted as a result of a certain amount of hunting "effort" or search labor, but only the probability to be encountered. And once encountering a prey, according to this model, the hunter still assesses, based on the expected handling and transport costs, whether it is worthwhile to actually hunt the prey in question.
We will here develop further the spatial model of Sirén and Parvinen (2015) by including not only one but two prey species, in accordance with the non-spatial two-species model of Milner-Gulland and Mace (1998, pp. 72–77). In this model, the two species interact by being harvested together, but they have no other ecological interactions. This leads to the following form for the equations of growth of each of the species and for the profit made by the harvesters:
$$\frac{dN\_1}{dt} = r\_1 N\_1 \left(1 - \frac{N\_1}{K\_1}\right) - H\_1 \tag{5}$$
$$\frac{dN\_2}{dt} = r\_2 N\_2 \left(1 - \frac{N\_2}{K\_2}\right) - H\_2 \tag{6}$$
$$\prod = \rho\_1 H\_1 + \rho\_2 H\_2 - c \left( \mathbb{S} + H\_1 t\_{\hbar\_1} + H\_2 t\_{\hbar\_2} + \frac{\mathbb{x} (H\_1 + H\_2)}{\nu\_t} \right) \tag{7}$$
The two species may be of greatly different size and mass, and the use of the same transport speed, v<sup>t</sup> , for both species therefore requires that this parameter is defined as the speed of transport per unit of mass, rather than per number of hunted prey. Accordingly, also the harvest variable, H<sup>i</sup> , must be defined not as number of hunted prey animals, but as the mass of harvested matter and the handling time, th<sup>i</sup> , scaled to the mass of harvested matter.
In the standard model, we would have always H<sup>i</sup> = qiSN<sup>i</sup> , but when the model includes the handling cost and the cost of transport, it may be that although it is profitable to have a positive search labor S, it is only beneficial to harvest one species. This occurs when the price of one species does not cover the handling and transport costs, so that
$$H\_i = \begin{cases} q\_i \text{SN}\_i, & \text{if } p\_i \ge c t\_{h\_l} + c \frac{x}{v\_l} \\ 0 & \text{otherwise} \end{cases} \tag{8}$$
At a biological (ecological) equilibrium, the populations of the two species remain constant, i.e., we have dN<sup>1</sup> dt = dN2 dt = 0. We assume an open access scenario, where many individuals harvest resources from a common resource pool in an uncoordinated and self-interested manner. Under such conditions, hunters will not hunt species that are too costly to handle or transport. Therefore
$$N\_{i} = \begin{cases} K\_{i} & \text{if } p\_{i} < c t\_{\text{hi}} + c \frac{\underline{x}}{\underline{v}\_{l}} \text{ or } \text{ S} = 0\\ K\_{i} \left(1 - \frac{q\_{i}}{r\_{i}} \text{S}\right) & \text{if } p\_{i} \ge c t\_{\text{hi}} + c \frac{\underline{x}}{\underline{v}\_{l}} \text{ and } \text{ S} < \frac{r\_{i}}{q\_{i}}\\ 0 & \text{if } p\_{i} \ge c t\_{\text{hi}} + c \frac{\underline{x}}{\underline{v}\_{l}} \text{ and } \text{ S} \ge \frac{r\_{i}}{q\_{i}} \end{cases} \tag{9}$$
The first row of Equation (9) tells us that the species N<sup>i</sup> occurs at natural densities, i.e., carrying capacity, at the distance x if either its value p<sup>i</sup> is so low that it does not make up for the inevitable costs of handling and transport or, alternatively, hunters are simply absent (S = 0). The second and third rows correspond to two situations in which the value p<sup>i</sup> is high enough so that hunting species N<sup>i</sup> is profitable at least if search costs are neglected. The third row tells us that a species is extirpated at the distance x, if its value is larger than the costs of handling and transport, and the search labor exerted by hunters exceeds a threshold determined by the species' intrinsic growth rate and the species' catchability coefficient. The middle row, finally, tells us that, in all other cases, the species in question will occur at a density larger than zero but smaller than the carrying capacity and which will be determined by the local search effort exerted by hunters (S) and the species-specific parameters carrying capacity (Ki), catchability coefficient (qi), and intrinsic growth rate (ri).
Extinction of both species at the same location is not possible in this model. According to Equation (9), species 1 will be extinct (N<sup>1</sup> = 0) if the marginal benefits are not negative, p<sup>i</sup> ≥ cth<sup>i</sup> + c x vt and search labor is large enough, S ≥ r1 q1 , where S is the search time resulting from hunting of species 2 alone. The expression for profit when species 1 is locally extirpated and only species 2 is hunted is obtained from Equation (7) by substituting H<sup>1</sup> = 0 and H<sup>2</sup> = q2SN2, and we get
$$\prod = \mathbb{S}\left[p\_2q\_2N\_2 - c\left(1 + q\_2N\_2t\_{h\_2} + \frac{\chi q\_2N\_2}{\nu\_t}\right)\right] \tag{10}$$
Next, we consider extinction of species 1 in a bioeconomic equilibrium, so that in addition to Equation (9), Q = 0 holds. Solving Q = 0 with (Equation 10) for N2, we get the zero-profit population size of species 2 when species 1 is locally extirpated:
$$N\_2 = \frac{c}{q\_2 \left(p\_2 - c\mathfrak{h}\_{\mathbb{H}\_2} - \frac{c\chi}{\nu\_t}\right)} \text{ if } p\_2 - c\mathfrak{t}\_{\mathbb{H}\_2} - \frac{c\chi}{\nu\_t} > 0 \tag{11}$$
The zero-profit population size N<sup>2</sup> obtained from Equation (11) should agree with the equilibrium population size given by the second row of Equation (9), which results in the following condition for S:
$$\frac{c}{q\_2 \left(p\_2 - c t\_{\rm h\_2} - \frac{c\chi}{\nu\_1}\right)} = K\_2 \left(1 - \frac{q\_2}{r\_2} \mathcal{S}\right). \tag{12}$$
Solving (Equation 12) with S > 0 is possible, if <sup>c</sup> q2 p2−cth<sup>2</sup> − cx vt <
K2. Solving S from Equation (12), we obtain that the amount of local search labor in a bioeconomic equilibrium, when species 1 is not present, is
$$\overline{S} = \frac{r\_2}{q\_2} \left( 1 - \frac{c}{K\_2 q\_2 \left( p\_2 - c t\_{\hbar\_2} - \frac{c\chi}{\eta\_1} \right)} \right) \tag{13}$$
For the species 1 to be extirpated, we have the condition S ≥ r1 q1 (Equation 9). Substituting (Equation 13), we obtain
$$\frac{r\_2 q\_1}{r\_1 q\_2} \left( 1 - \frac{c}{K\_2 q\_2 \left( p\_2 - c t\_{h\_2} - \frac{c \alpha}{v\_t} \right)} \right) \ge 1 \tag{14}$$
Solving (Equation 14) with equality for x, we obtain what we call the extinction distance, xe<sup>1</sup> , for species 1, meaning that species 1 is present only beyond this distance, having been extirpated by hunting at closer distances to the central place from which hunters start their hunting journeys:
$$\begin{aligned} \varkappa\_{\mathfrak{e}\_1} &= \nu\_{\mathfrak{t}} \left[ \frac{p\_2}{c} - t\_{\mathfrak{h}\_2} - \frac{1}{K\_2 q\_2 \left( 1 - \frac{r\_1 q\_2}{r\_2 q\_1} \right)} \right], \\ \text{if } p\_i &\ge \ ct\_{\mathfrak{h}\_i} + \varepsilon \frac{\varkappa\_{\mathfrak{e}\_i}}{\nu\_{\mathfrak{t}}} \text{ for both } i, \\ \text{and} & \qquad \frac{r\_1}{q\_1} < \frac{r\_2}{q\_2}. \end{aligned} \tag{15}$$
The conditions for the prices come from the third row of Equation (9) and are needed to ensure that handling and transporting both species are profitable at the distance given by the expression xe<sup>1</sup> . Together with the condition <sup>r</sup><sup>1</sup> q1 < r2 q2 , this means that the third row of Equation (9) may hold for species 1 and the second row for species 2. If either of the price conditions does not hold, the extinction distance is given by the minimum of xh<sup>1</sup> and xh<sup>2</sup> .
Analogously, the extinction distance for species 2 is
$$\begin{aligned} \varkappa\_{\mathfrak{e}\_2} &= \nu\_{\mathfrak{t}} \left[ \frac{p\_{\mathfrak{l}}}{c} - t\_{\mathfrak{h}\_{\mathfrak{l}}} - \frac{1}{K\_1 q\_1 \left( 1 - \frac{r\_2 q\_1}{r\_1 q\_2} \right)} \right], \\ \text{if } p\_{\mathfrak{i}} &\ge c t\_{\mathfrak{h}\_{\mathfrak{i}}} + c \frac{\varkappa\_{\mathfrak{e}\_i}}{\nu\_{\mathfrak{t}}} \text{ for both } i, \\ \text{and} & \quad \frac{r\_1}{q\_1} > \frac{r\_2}{q\_2}. \end{aligned} \tag{16}$$
Note that only the species with lower ratio <sup>r</sup><sup>i</sup> qi may become extirpated.
Next, we consider the distance beyond which either of the species is not harvested, so that the first row of Equation (9) holds. As no harvesting of species i occurs if the price does not cover handling and transport costs, i.e., if p<sup>i</sup> < cth<sup>i</sup> + c x vt (Equation 9), we get from solving p<sup>i</sup> = cth<sup>i</sup> + c x vt for x that species i will not be harvested further than
$$\alpha\_{\mathbf{h}\_{\mathbf{i}}} = \begin{cases} \nu\_{\mathbf{t}} \left[ \frac{\rho\_{\mathbf{i}}}{c} - t\_{\mathbf{h}\_{\mathbf{i}}} \right], & \text{if } \rho\_{\mathbf{i}} \ge c t\_{\mathbf{h}\_{\mathbf{i}}} \\ \mathbf{0}, & \text{if } \rho\_{\mathbf{i}} \le c t\_{\mathbf{h}\_{\mathbf{i}}}. \end{cases} \tag{17}$$
From now on, we assume that xh<sup>i</sup> > 0. The second row of Equation (17) corresponds to a situation in which the price does not even cover handling costs alone. The actual no-harvest distance may also be even shorter than the expression xh<sup>i</sup> given by the first row of Equation (17), because this does not take search costs into account. This is therefore a precise no-harvest distance only in the case that the other species is significantly more profitable to hunt, such that the search costs are covered by hunting for that species.
When the species are similar—but not necessarily equal—in their price and handling time, they have the same no-harvest distance. We can solve this no-harvest distance by substituting H<sup>i</sup> = qiSK<sup>i</sup> in Equation (7) and solving for x from Q = 0, i.e.,
$$S\left[p\_1q\_1K\_1 + p\_2q\_2K\_2 - c\left(1 + q\_1K\_1t\_{\mathrm{h}\_1}\right)^2\right]$$
$$\left[1 + q\_2K\_2t\_{\mathrm{h}\_2} + \frac{\varkappa\left(q\_1K\_1 + q\_2K\_2\right)}{\nu\_t}\right] = 0,\tag{18}$$
resulting in the common no-harvest distance
$$\mathbf{x}\_{\mathbf{h}} = \frac{\nu\_{\mathbf{f}}}{q\_1 K\_1 + q\_2 K\_2} \left( q\_1 K\_1 \left( \frac{p\_1}{c} - t\_{\mathbf{h}\_1} \right) + q\_2 K\_2 \left( \frac{p\_2}{c} - t\_{\mathbf{h}\_2} \right) - 1 \right), \tag{19}$$
provided that harvesting both species at that distance would be profitable without search costs: p<sup>i</sup> ≥ cth<sup>i</sup> + c xh vt , or equivalently xh<sup>i</sup> ≥ xh, for both i.
However, when the species are not similar enough in their price and handling time, it is possible that for one species, p<sup>i</sup> < cth<sup>i</sup> + c xh vt , i.e., xh<sup>i</sup> < xh, so that at the distance x<sup>h</sup> given by Equation (19), it would not be profitable to hunt species i even without search costs. In such a situation, hunting the other species is very profitable and the hunters earn better simply by neglecting species i. As the search costs are covered by hunting the other species, the species i will then have no-harvest distance given by xh<sup>i</sup> . For the other species (= j), we have at the noharvest distance
$$\mathcal{S}\left[p\_{\dot{j}}q\_{\dot{j}}\mathcal{K}\_{\dot{j}} - c\left(1 + q\_{\dot{j}}\mathcal{K}\_{\dot{j}}t\_{l\dot{j}} + \frac{\varkappa q\_{\dot{j}}\mathcal{K}\_{\dot{j}}}{\nu\_{l}}\right)\right] = 0,\tag{20}$$
which is obtained by substituting H<sup>i</sup> = 0 and H<sup>j</sup> = qjSK<sup>j</sup> in Equation (7) and setting Q = 0. The no-harvest distance for species j is then obtained by solving for x from Equation (20), resulting in
$$
\tilde{\chi}\_{\mathfrak{h}\_j} = \nu\_t \left( \frac{p\_j}{c} - t\_{\mathfrak{h}\_j} - \frac{1}{q\_j \mathcal{K}\_j} \right). \tag{21}
$$
Note that the formulas satisfy xe<sup>1</sup> < x˜h<sup>2</sup> , when <sup>r</sup><sup>1</sup> q1 < r2 q2 , because species 2 can cause the extinction of species 1 only if species 2 is harvested at that distance. Furthermore, x˜h<sup>2</sup> < xh<sup>2</sup> , which means that potential no-harvest distance x˜h<sup>2</sup> derived assuming that search costs are covered by hunting species 2 only is strictly smaller than the upper bound xh<sup>2</sup> of the extinction distance derived from the marginal benefits, neglecting search costs.
Furthermore, the common no-harvest distance x<sup>h</sup> from Equation (19) can be written as
$$\chi\_{h} = \frac{q\_{1}K\_{1}}{q\_{1}K\_{1} + q\_{2}K\_{2}} \underbrace{\left[\nu\_{t}\left(\frac{p\_{1}}{c} - t\_{h1}\right)\right]}\_{\chi\_{h\_{1}}} + \frac{q\_{2}K\_{2}}{q\_{1}K\_{1} + q\_{2}K\_{2}}$$
$$\underbrace{\left[\nu - t\left(\frac{p\_{2}}{c} - t\_{h2} - \frac{1}{q\_{2}K\_{2}}\right)\right]}\_{\tilde{\chi}\_{h\_{2}}},\tag{22}$$
so that x<sup>h</sup> is a biased average of xh<sup>1</sup> and x˜h<sup>2</sup> , and analogously, a biased average of x˜h<sup>1</sup> and xh<sup>2</sup> . Since an average of two values is always in between the two values the average is taken from, we have the relations
$$\begin{cases} \begin{aligned} \boldsymbol{\chi\_{h\_1}} \leq \boldsymbol{\chi\_{h\_1}} \leq \tilde{\boldsymbol{\chi}\_{h\_2}}, \text{ if } \boldsymbol{\chi\_{h\_1}} \leq \tilde{\boldsymbol{\chi\_{h\_2}}}\\ \boldsymbol{\chi\_{h\_1}} \geq \boldsymbol{\chi\_{h}} \geq \tilde{\boldsymbol{\chi\_{h\_2}}}, \text{ if } \boldsymbol{\chi\_{h\_1}} \geq \tilde{\boldsymbol{\chi\_{h\_2}}} \end{aligned} \text{ and} \end{cases}$$
$$\begin{cases} \boldsymbol{\chi\_{h\_2}} \leq \boldsymbol{\chi\_{h}} \leq \tilde{\boldsymbol{\chi\_{h\_1}}}, \text{ if } \boldsymbol{\chi\_{h\_2}} \leq \tilde{\boldsymbol{\chi\_{h\_1}}}\\ \boldsymbol{\chi\_{h\_2}} \geq \boldsymbol{\chi\_{h}} \geq \tilde{\boldsymbol{\chi\_{h\_1}}}, \text{ if } \boldsymbol{\chi\_{h\_2}} \geq \tilde{\boldsymbol{\chi\_{h\_1}}} \end{aligned} \tag{23}$$
In principle, we could have four different cases in which inequalities in Equation (23) hold. However, inequalities xh<sup>1</sup> < x˜h2 and xh<sup>2</sup> < x˜h<sup>1</sup> cannot hold at the same time, because then from Equation (23), we would have xh<sup>1</sup> < x<sup>h</sup> and x<sup>h</sup> < x˜h<sup>1</sup> , so that xh<sup>1</sup> < x˜h<sup>1</sup> , which leads to contradiction. Overall, we have, thus, three different cases of no-harvest distances, and in different parts of the parameter space, we have different formulas determining the no-harvest distances summarized in **Table 2**.
### MODEL RESULTS
**Figure 1** shows the two basic patterns that might occur depending on the r/q quotient of the two species, when all other parameters are equal or almost equal (here there is a minor difference of K between the two species, just in order to improve the visual presentation, avoiding that the curves for the two respective species overlap each other). When the r/q ratio is equal or similar between the two species, they coexist over the entire range of distances, depending on the parameter values (**Figure 1A**), but when their r/q ratios differ, the species with lower r/q may get extinct up to a certain distance, which we call the extinction distance. When cost-related parameters of the two species are similar (in **Figure 1** they are the same), the no-harvest distance, i.e., the distance beyond which a species is not hunted at all, however, is the same for both species, regardless of the difference in r/q quotient.
**Figure 2** shows with some more detail how different values of r and q affect extinction distances and no-harvest distances. According to Equation (14), when the parameters r and q of the two species are close to being equal, one species cannot cause the extinction of the other species, as in those regions of **Figure 2A**, where r<sup>1</sup> ≈ r<sup>2</sup> = 1, and in **Figures 2B,C**, in those regions where q<sup>1</sup> ≈ q<sup>2</sup> = 1. If there is a considerable difference in the r/q quotient between the two species, however, the species with the lower ratio <sup>r</sup><sup>i</sup> qi of growth rate and catchability may go extinct at short distances. For r1, this is illustrated in **Figure 2**A: species 1 goes extinct when r<sup>1</sup> is low, and species 2 goes extinct when r<sup>1</sup> is large. The same phenomenon occurs in **Figures 2B,C**, when q<sup>1</sup> is large, as it is species 1, which has lower ratio <sup>r</sup><sup>i</sup> qi , that goes extinct. For low q1, however, comparing the ratios <sup>r</sup><sup>i</sup> qi only is not sufficient. Especially, when q<sup>1</sup> = 0, species 1 is not harvested at all, so that the model is essentially a one-species model, in which harvesting cannot cause extinction. Consequently, if extinction of species 2 occurs for some q<sup>1</sup> < q2, it only occurs for intermediate values of q<sup>1</sup> (**Figure 2C**), and the extinction distance has a humped shape reaching a maximum at q<sup>1</sup> = q2r1 2r2 (at q<sup>1</sup> = 1 2 in **Figure 2C**). It is also possible that extinction of species 2 does not occur for any q<sup>1</sup> < q2, even though species 2 then has lower ratio <sup>r</sup><sup>i</sup> qi (**Figure 2B**). Such a situation occurs, when <sup>p</sup><sup>1</sup> <sup>c</sup> −th<sup>1</sup> −4 r2 q2r1K1 < 0. In **Figures 2B,C**, we have <sup>p</sup><sup>1</sup> <sup>c</sup> − th<sup>1</sup> − 4 r2 q2r1K1 = 3 − 4 K1 , so that in **Figure 2B** we have 3 − 4 K1 = −1 < 0, and in **Figure 2C** 3 − 4 K1 = 1 > 0. Again, since the cost-related parameters of the two species are similar (the same in **Figure 2**), the no-harvest distance still is identical (x<sup>h</sup> given by Equation 19) for both species in all these cases, and x<sup>h</sup> increases with q1, but is unaffected by r1.
In **Figure 2**, all species-specific parameters were the same for the two species, except for r<sup>i</sup> or q<sup>i</sup> . If <sup>r</sup><sup>1</sup> q1 = r2 q2 , the condition (Equation 14) for overharvested extinction is not satisfied. Therefore, it is meaningful to investigate the effect of other parameters on extinction distances only if <sup>r</sup><sup>1</sup> q1 6= r2 q2 . This is illustrated in **Figure 3**, in which we have chosen such parameters that <sup>r</sup><sup>1</sup> q1 < r2 q2 , such that species 1 is the more vulnerable species and the only one that may be driven to extinction.
#### TABLE 2 | The different cases of no-harvest distances.
FIGURE 1 | Typical cases for similar species. Population densities N1 and N2 (thick curves), harvest H1 and H2 (thick dashed curves), and search labor S (thin dashed curve) in a bioeconomic equilibrium with respect to distance x. (A) No extinction. (B) Species 1 is overharvested to extinction near the village, at distances 0 ≤ x ≤ x˜e<sup>1</sup> . In both cases harvesting becomes non-beneficial at long distances, for x ≥ xh . Parameters: K<sup>1</sup> = 1, K<sup>2</sup> = 1.05, r<sup>1</sup> = r<sup>2</sup> = 1, q<sup>2</sup> = 1, p<sup>1</sup> = p<sup>2</sup> = 4, c = 1, th<sup>1</sup> = th<sup>2</sup> = 1, vt = 1.
Increasing either one of the carrying capacities increases the overall abundance of prey and thus makes it profitable to hunt further away from the village, such that the no-harvest distance increases (**Figures 3A,B**). The carrying capacity K<sup>1</sup> of species 1 does not affect the extinction distance of the species 1 itself, <sup>d</sup> dK1 xe<sup>1</sup> = 0 (**Figure 3A**). In contrast, increasing the carrying capacity K<sup>2</sup> of species 2 leads to increased search labor and, therefore, increased extinction distance of species 1, <sup>d</sup> dK2 xe<sup>1</sup> > 0 (**Figure 3B**).
**Figures 3C,D** illustrates the effects of prices p<sup>1</sup> and p2. Increasing either one of the prices will make harvesting further away economically more profitable, and the no-harvesting distances xh, xh<sup>i</sup> , and x˜h<sup>i</sup> either increase linearly with p<sup>i</sup> or are constants (actually, all curves separating different areas in **Figures 3C,D** are straight lines). If the prices of the two respective species differ very much, the species with the lower price is not harvested at all. If the price of one species is large, then far from the village only that species is harvested. Closer to the village, both species are harvested unless species 1 is extinct. Increasing either one of the prices may cause the extinction of species 1. Increasing p<sup>1</sup> does this by increasing the profitability of hunting species 1, and increasing p<sup>2</sup> does this by increasing the search effort. An interesting feature in **Figures 3C,D** is that the region "Species 1 extinct" borders to the region "Only species 2 harvested." This implies that in a certain range of prices (see the line xh<sup>1</sup> in the intervals 1 < p<sup>1</sup> < 2 in **Figure 3C** and p<sup>2</sup> > 6 in **Figure 3D**), species 1 is hunted to extinction up to a certain distance, beyond which harvest of that species abruptly ceases, and it is present at its carrying capacity.
FIGURE 3 | Distances with respect to various parameters: The no-harvest distances xh , x˜h<sup>1</sup> , and x˜h<sup>2</sup> (thin dashed curves) separating the areas of no harvesting and some harvesting, the no-harvest distances xh<sup>1</sup> and xh<sup>2</sup> (thick dashed curves) separating the areas of only one species harvested and both species harvested (potentially leading to the extinction of one of them), and the extinction distance xe<sup>1</sup> , separating the areas where both species are successfully harvested, and where harvesting of both species results in the extinction of species 1 (thick solid curves). Species differ in r1 = 0.5 < r2 = 1. Except for the parameter displayed in the horizontal axis of each panel, the parameters are as follows: K<sup>1</sup> = K<sup>2</sup> = 1, q<sup>l</sup> = q<sup>2</sup> = 1, p<sup>1</sup> = P<sup>2</sup> = 4, c = 1, th<sup>1</sup> = th<sup>2</sup> = 1, vt = 1.
Handling times have similar effects as prices, but the patterns are reversed (**Figures 3E,F**). If handling time is too long for a species, it will not be harvested. For intermediate handling times, both species are harvested. Decreasing either one of the handling times may cause the extinction of species 1. Decreasing th1 does this by increasing the profitability of hunting species 1, and decreasing th<sup>2</sup> does this by increasing the search effort. Also, in these panels, the curves separating different areas are straight lines.
The last two panels of **Figure 3** show the effects of systemwide parameters. **Figure 3G** shows that if the cost of time c is large, then no harvesting takes place. This happens at least when th1 c > p<sup>1</sup> and th<sup>2</sup> c > p2. For low cost, the species with lower ratio <sup>r</sup><sup>i</sup> qi goes extinct. For intermediate c, both species may be harvested. **Figure 3H**, on the other hand, shows that both the extinction distance and no-harvest distance increase linearly with respect to the speed of transport, v<sup>t</sup> . Whereas the transport speed, vt , and the cost of time, c, have opposite effects, there is one additional important difference. Doubling the transport speed always leads to a doubling of the extinction distance as well as the no-harvest distance, and only at v<sup>t</sup> =0 (an unrealistic scenario with completely sessile hunters), there is no local extinction at any distance. In contrast, doubling the cost of time reduces the extinction distance, as well as the no-harvest distance, with much more than half, and at a certain level of c, the extinction distance hits zero, as does also, at an even higher c, the noharvest distance.
### DISCUSSION
This piece of research provides mathematical explanations to the commonly observed phenomenon that different species that are hunted together tend to not only differ in abundance as such but also show qualitatively different spatial patterns of abundance. Model results show a wide variety of possible scenarios when two species are hunted together, depending on how the parameter values of the two species differ from each other. Some of these results have potentially important implications for understanding the causes behind huntinginduced extirpations and practical wildlife management. For example, model results indicate that the extinction distance as well as the no-harvest distance are piecewise smooth—in relation to price or handling time (**Figures 3C–F**) even linear—functions that abruptly change slope at certain parameter values. For another part, model results suggest that even modest increases of the opportunity cost of time can have very positive effects on hunted wildlife populations, as the extinction distance is reduced with a factor larger than the factor of increase of the cost (**Figure 3G**).
All models are simplifications of reality, and it is therefore important to discuss the implications of the assumptions implicit in the model. For the single-species version of the model, Sirén and Parvinen (2015) discussed, for example, the implications (1) that it was deterministic, whereas real-life hunting involves a great deal of stochasticity; (2) that it had just one spatial dimension; (3) that it assumed that hunters have one single start- and endpoint for hunting trips; (4) that it considered travel and search as two separate activities; and (5) that it involved no animal movements. Some of these assumptions have still additional implications for the two-species model and the discussion of local extirpations.
Regarding the assumption of no dispersal or movements of animals, Novaro et al. (2000) argued that dispersal could have a key role in rebuilding animal populations depleted by hunting in tropical forests, and this was also supported by Sirén et al. (2004), who showed that, despite the dispersal rates for most major game species being relatively low, they were large enough to ensure that practically no species were permanently absent anywhere in the study area. Even the most severely depleted species, such as spider monkeys (Ateles belzebuth), were at rare occasions spotted (and killed) even very near the village. Thus, extinction according to the model should not be interpreted as a constant and complete absence of a species in real life, but rather as absence of breeding and reliance on a continuous influx due to source–sink dynamics in order to maintain a very low abundance or even just intermittent presence
Similarly, the model assumes a fixed handling time for each respective species. In reality, this can vary considerably. Any species might suddenly, by chance, appear within shooting range in front of a hunter, such that the handling time becomes minimal. At other times, the hunter might just hear the animal at a distance, requiring the hunter to carefully pursue it, without making noise that scares it away. Some species, such as large rodents, armadillos, and the white-collared peccary (Tayassu pecari) commonly take escape in burrows when stalked by dogs, and it can then be a quite lengthy procedure to kill them and recover the carcass from inside the burrow. Rather than a fixed handling time for each species, in real life, there is just a different probability for different handling times for each species. In addition, as also the cost of ammunition is included in this parameter, another cause of variability is that sometimes hunters miss the target, thus having to shoot more than once or, in worst case, wasting ammunition but failing to recover the prey. Again, therefore, the model predictions in **Figure 3** should not be taken too literally. That the model predicts that for some combinations of parameter values one of the species is not hunted does not mean that in real life it will not be hunted at all, but rather that it will be hunted in relatively small numbers.
Although the inclusion of two species is an important improvement in comparison with the single-species model, it is still a major simplification, as empirical studies indicate that tropical forest hunters tend to hunt a large number of different species, ranging from around 20 (Ohl-Schacherer et al., 2007) to around 40 (Franzen, 2006), or 60 (Sirén, 2012; Constantino, 2016). It should also be noted that in this model the two species do not interact with each other in any other way than that they are harvested together.
Some of these limitations of the model could, in principle, be resolved relatively easily. For example, including ecological interactions between the two species, such as competition for resources, would also be relatively straightforward (cf. Milner-Gulland and Mace, 1998, pp. 71–77). It would also be possible to include multiple species in the model or to introduce stochasticity. The more complex a model becomes, however, it also becomes less perspicuous, and the whole point of analytical models is to highlight some certain aspects of reality, which requires disregarding others. This two-species model can help us gain important insights into the spatial patterns of harvest and abundance and the mechanisms leading to sequential extirpations, in space and time, of certain species in harvested multispecies systems (cf. Rowcliffe et al., 2003). There is ample empirical evidence showing that animal species with certain traits (particularly, large bodied, large group living, arboreal, and diurnal) are more susceptible to extirpation due to hunting than others (e.g., Ripple et al., 2016; Abrahams et al., 2017). However, it is not well-understood how such characteristics of different species interact with each other and with socioeconomic parameters in order to produce different outcomes in terms of local extirpations of some of the species. We believe that the model presented here can help to gain a better theoretical understanding of the mechanisms leading to sequential extirpation of different game species and of the spatial distribution of the so-called "extinction envelopes" (cf. Shaffer et al., 2018) of different species.
Theoretical models should ideally always be validated by comparison with empirical data, but this involves, in this case, considerable challenges and is beyond the scope of this piece of research. In order to fully validate this model, one would need a large set of empirical data including spatially explicit data on wildlife harvest and abundance as well as trustworthy estimates of key biological and economic parameters for each hunted wildlife species. In addition, such a dataset would need to cover a wide range of variability in the opportunity cost of time, as well as within-species and between-species variabilities in price and handling time (the latter mediated by technology). Currently, however, there is no dataset available that is even close to fulfilling these criteria. It remains an open question whether it would be feasible to construct such a dataset even by pooling together data from many different case studies, conducted by different researchers in different parts of the world.
An alternative approach to validating the model, however, could be to look for cases in the real world where a hunted species qualitatively behaves like the model predicts. Such a case could be, for example, that a species with lower market price or use
### REFERENCES
value than most other prey species is hunted to extirpation up to a certain distance, but is not hunted at all beyond this distance, as predicted by **Figures 3C,D**, where the line xh<sup>1</sup> separates the area "Species 1 extinct" from "Only species 2 harvested." Another such case could be that a relatively modest increase of the price of some species that previously have not been hunted leads to extirpation of the species over a significant distance, but that this extinction distance afterward remains relatively constant despite further increases of price, as in **Figure 3C**, where the slope of line xe<sup>1</sup> is horizontal, or, analogously, the same phenomenon for reduced handling time, due to some technological improvement, as in **Figure 3E**.
Whereas we here have analyzed only the case of open access hunting, a next step will be to analyze also the social optimum case, as Sirén and Parvinen (2015) did for the onespecies version of this model, and also to analyze the economic and ecological effects of different sorts of hunting regulations and enforcement strategies (cf. Albers, 2010). Because of the huge challenges involved in collecting empirical data on the parameters and variables included in models of hunting in tropical forests (cf. Carrillo et al., 2000; Van Vliet and Nasi, 2008), practical wildlife management will have to rely more on trial and error than on prescriptions based on quantitative modeling (cf. Johannes, 1998). Analytical models like this one could be a useful support, however, when trying to figure out which management measures might be worthwhile to try out in the real world.
### AUTHOR CONTRIBUTIONS
AS formulated the basic characteristics of the model, defined the questions to be answered, and wrote the Background and Discussion sections. KP did most of the mathematical analyses, designed the figures, and wrote most of the sections Model Assumptions and Model Results.
### FUNDING
We were granted a waiver from Frontiers for 50% of the article processing fee, and CIFOR provides the remaining 50%.
Amerindians of Bolivia. Oryx 36, 382–388. doi: 10.1017/S00306053020 0073X
**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Copyright © 2019 Sirén and Parvinen. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
# Frameworks Regulating Hunting for Meat in Tropical Countries Leave the Sector in the Limbo
Nathalie van Vliet <sup>1</sup> \*, André Pinassi Antunes <sup>2</sup> , Pedro de Araujo Lima Constantino<sup>2</sup> , Juanita Gómez <sup>3</sup> \*, Dídac Santos-Fita<sup>4</sup> and Eugenio Sartoretto<sup>5</sup>
<sup>1</sup> Center for International Forestry Research, Bogor, Indonesia, <sup>2</sup> RedeFauna - Rede de Pesquisa em Diversidade, Conservação e Uso da Fauna da Amazônia, Brasilia, Brazil, <sup>3</sup> Independent Researcher, Bogotá, Colombia, <sup>4</sup> Instituto Amazônico de Agriculturas Familiares, Federal University of Pará, Belém, Brazil, <sup>5</sup> Food and Agriculture Organization of the United Nations, Rome, Italy
Despite restrictive legal frameworks, hunting for meat is a reality in tropical countries. In this policy paper, we argue that formal regulations are ill adapted to the contexts in which they should be applied and are characterized by gaps and contradictions that maintain the sector in a limbo. We use contemporary examples from Latin America and Africa described in detail in publications ranging from 2015 to 2019, to illustrate the need for legal reforms that clarify the rights to sell surplus of meat and align land tenure rights with wildlife use rights to suggest a new definition of subsistence hunting which accounts for the realities of communities from different cultural backgrounds.
#### Edited by:
Divya Karnad, Ashoka University, India
#### Reviewed by:
Edson Gandiwa, Chinhoyi University of Technology, Zimbabwe Charlotte H. Chang, National Institute for Mathematical and Biological Synthesis (NSF), United States
#### \*Correspondence:
Nathalie van Vliet vanvlietnathalie@yahoo.com Juanita Gómez plconstantino@gmail.com
#### Specialty section:
This article was submitted to Conservation, a section of the journal Frontiers in Ecology and Evolution
> Received: 30 January 2019 Accepted: 09 July 2019 Published: 02 August 2019
#### Citation:
van Vliet N, Antunes AP, Constantino PAL, Gómez J, Santos-Fita D and Sartoretto E (2019) Frameworks Regulating Hunting for Meat in Tropical Countries Leave the Sector in the Limbo. Front. Ecol. Evol. 7:280. doi: 10.3389/fevo.2019.00280 Keywords: hunting, tropical count, legal framework, wildmeat, land tenure rights, sustainable use, rural societies, subsistence use
## INTRODUCTION
Wildmeat (or Bushmeat), defined as any non-domesticated mammals, reptiles, amphibians and birds hunted for food in tropical forests (Nasi et al., 2008), continues to play a key role for the food security of contemporary rural societies in tropical countries (Alves and van Vliet, 2018). It is also practiced in relation to a diversity of socio-cultural reasons (Morsello et al., 2015; Santos-Fita et al., 2015; van Vliet, 2018; Martins and Shackleton, 2019); crop protection (Abrahams et al., 2018; Constantino, 2019a), zoo-therapeutical purposes (Santos-Fita et al., 2012; Alves and da Silva Policarpo, 2018), income (Mavah et al., 2018; Rodríguez-Ríos and García-Páez, 2018; Rogan et al., 2018), and sport (Fischer et al., 2013). Our focus in this paper is on hunting for meat, because access to adequate food is a human right globally recognized to local communities. Despite the recognition that hunting for meat significantly contributes to local livelihoods (CBD, 2016) and to local economies (Lescuyer and Nasi, 2016), this type of hunting continues to occur in a context of informality and in parallel with existing regulations. Hunting for meat is therefore often stigmatized as "illegal," without distinction from other more detrimental illegal hunting practices, for example those linked to organized crime. This, on its own, limits innovations in promoting sustainable hunting practices and offers little opportunities for signatory countries to observe the recommendations from the Convention of Biological Diversity (CBD), with regards to sustainable hunting (Coad et al., 2019).
With the global decline of wildlife worldwide, donor's attention continues to focus on efforts to reduce illegal wildlife trade through law enforcement. However, lessons learnt from practical experience show that this has limited impact, and particularly when it only focuses on militarization (Duffy et al., 2019). The reasons for failure are numerous and include corruption, lack of resources from the government to exercise sovereignty in remote locations where hunting takes place, lack of political will to prioritize law enforcement regarding wildlife crimes, lack of alternatives to replace supply, high level of dependency on wildmeat for households living in extreme poverty, and a demand which is difficult to downsize due to the high value that consumers are ready to pay (Bennett, 2011; Wellsmith, 2011; Challender and MacMillan, 2014; Cooney et al., 2017; Swan, 2017; Constantino, 2019b). In addition, the intrinsic nature of laws and regulations on hunting is also at the heart of the failure, but little is said about the urgency to reform and adapt the regulatory framework, particularly with regards to hunting for food. Indeed, the lack of clarity and ambiguity prevailing in legal texts leaves room for diverse interpretations, favors insecurities and marginalization of subsistence hunters, fuels underground markets that are difficult to control, and dilute responsibilities for sustainable use.
In this policy paper, we argue that formal regulations are ill adapted to the contexts in which they should be applied and are characterized by gaps and contradiction that maintain hunting for meat and the sale of its surplus in a limbo. We use examples from Latin America [Mexico, Brazil, Colombia, Guyana) and Africa (Congo, Gabon and Democratic Republic of Congo (DRC)] to illustrate the urgent need for legal reforms that enhance the sustainable use of wildlife resources. Three questions have guided our analysis (**Table 1**): Is hunting for food legal? Can a hunter sell a surplus of wildmeat? Does the wildlife belong to the land owner? Based on the analysis on case studies in those two regions (Van Vliet et al., 2015; Sartoretto et al., 2017; Gomez, 2018; Santos-Fita, 2018; Antunes et al., 2019; Pezzuti et al., 2019), we discuss the main reasons why current legal texts fail to address the need for more sustainable practices and further marginalize those who depend on hunting for their livelihoods. General recommendations for the improvement of current regulations in line with the principles of adaptive management are provided.
### THE BLURRY CONCEPT OF SUBSISTENCE HUNTING
The scientific literature on hunting originally distinguished two main types of hunting for meat carried out by local communities depending on the main motive for hunting: subsistence hunting and commercial hunting (Nasi et al., 2008). While the main incentive for hunting is often the need for self-consumption, hunters may sell part of the game as a source of income while keeping the rest to satisfy the food security of their families. In fact, the proportion and volumes of meat being sold varies from one context to another, making it difficult to establish simple categories. From a technical perspective, a flexible definition of subsistence hunting could include selling (mostly locally) part of the game hunted for consumption to purchase other subsistence goods (e.g., soap, gasoline, oil). However, in legal terms, the concept of subsistence hunting is defined differently and refers to different realities across our case studies. The diversity of terms used in legal frameworks attests of the difficulty to constrain the concept into a clear definition.
In Brazil, only Amerindians have the right to hunt in indigenous lands. For non-Amerindian hunters, this right is amenable to legal interpretation from a set of contradictory laws and incongruous legal concepts regarding human rights or wildlife protection. In practice, subsistence hunting is generally tolerated if intended "to quench the hunger" of a person in remote regions. As such, the concept of subsistence has been interpreted by some as restricted to the concept of "extreme necessity" (Antunes et al., 2019; Pezzuti et al., 2019). In Colombia, hunting for non-protected species, for food provision to the hunters' families is authorized under the term "subsistence hunting" (Van Vliet et al., 2015). All inhabitants may hunt for subsistence without permit in the national territory, provided there is no prohibition issued by environmental authorities. Guyana grants Amerindian villages the right to hunt for consumption as part of the "traditional rights," defined under the Amerindian Act as "any subsistence right or privilege, which is exercised sustainably in accordance with the spiritual relationship with the land" (Gomez, 2018). In Mexico, there is no clear indication of whether hunting can be practiced as part of the legally recognized "subsistence uses" or if it is subject to previous authorization by the Ministry in charge. Subsistence uses include the use of resources for direct consumption or sale, for satisfaction of basic needs, as well as those of economically dependent subjects (Santos-Fita, 2018). In an attempt to account for the spiritual dimension of subsistence hunting, the Mexican law allows communities to request a specific authorization for the use of wildlife in rituals and traditional ceremonies (Santos-Fita, 2018).
In Central Africa, the legislations of Gabon and Congo recognize customary use rights to local communities, which can take many forms, but often include all use and exploitation of timber and non-timber forest products to meet needs and requirements, including hunting (Sartoretto et al., 2017). However, due to the undistinguished regulation across all forms of hunting (commercial, sport, subsistence), it is often unclear whether current hunting restrictions (e.g., hunting seasons and gears) also apply to subsistence hunters. Moreover, since customary rights are often granted for subsistence reasons, the law limits their enjoyment solely to the satisfaction of personal or community needs. Trade in products resulting from the exercise of user rights is either prohibited, as in Congo, or restricted within the local community, as in Gabon. The DRC does not explicitly include the right to hunt among customary rights and, by doing so, excludes hunting from the regime of free exercise (Sartoretto et al., 2017). Hunting, including by local communities, is subordinated to the acquisition of a collective hunting license, which authorizes hunting "within the strict limits of their food needs."
### THE SALE OF SURPLUS MEAT: FROM LACK OF CLARITY TO COMPLEX PROCEDURES
While selling the surplus of meat after fulfilling the needs of the family is an integral part of subsistence strategies, it
TABLE 1 | Comparison of national regulations regarding the use and trade of wildmeat in Colombia, Brazil, Guyana, Mexico, Republic of Congo, Gabon, and Democratic Republic of Congo.
(Continued)
d'aménagement durable des concessions forestières
TABLE 1 | Continued
is not recognized as such by most regulations. Differences exist across countries, but a common denominator is the lack of clarity concerning the right to sell game. Currently, the sale of surplus meat is at a cross-road between being under regulated, on one hand, and over-regulated on the other. While a number of key aspects regarding hunting and trade rights remain a vacuum, some very specific instruments, probably developed in isolation from the rest of the regulatory framework, have ended up over-regulating the activity, to a point where enforcement becomes nearly impossible. Many of the regulations are inoperative, and the institutions in charge are not prepared, operationally and or financially to comply with established responsibilities.
In Brazil, it is forbidden to transport, sell and acquire eggs, larvae or specimens of fauna and by-products from hunting and harvesting or from un-authorized breeding sites (Antunes et al., 2019; Pezzuti et al., 2019). Inside indigenous lands, Amerindians have management rights over aboveground natural resources and there are no commercial legal restrictions. Commercial extensive management of wildlife by local communities is permitted in exceptional circumstances that require specific regulations currently available for few species [e.g., Melanosuchus niger (Ranzi et al., 2018)].
In Colombia, the sale of surplus falls into the category of commercial hunting, and therefore subsistence trade is not distinguished from commercial trade (Van Vliet et al., 2015). For the purpose of selling the surplus of meat, a subsistence hunter should apply for a commercial hunting license that is subjected to complex requirements including the submission of an Environmental Assessment Study (EIS). These requirements fall far from the capacities of local communities promoting illegality. Moreover, given the lack on regulations establishing hunting quotas, obtaining a commercial hunting license is impossible in practice. As such, even if wildmeat trade is not explicitly forbidden it is not allowed in practice (Van Vliet et al., 2015).
In Guyana, the Wildlife Regulations of 2013 established that any person who proposes to engage in buying or selling wildlife shall, before commencing such activities, apply for a commercial license. This license includes the commercial activities but not the collection of wildlife. Thus, a commercial license holder who will harvest the animals by himself will require a wildlife collecting license as well (Gomez, 2018).
In Mexico, despite the creation of the Wildlife Management Units (UMA), which were initially designed to ensure that sustainable resource use could be an economic opportunity as well as a conservation strategy, only sport hunting was recognized in practice. Indeed, the term subsistence hunting was implicitly associated with concepts such as "furtive," "illegal," "unregulated," or "inadequate use" and was therefore not valued as a valid option for sustainable use, management, and conservation of wild fauna under the initial strategic proposition of UMA instrument (Santos-Fita, 2018). For the case of wildmeat, only intensive or extensive breeding authorized centers (under UMA) were given the right to trade meat and other products obtained from wild species in established, official and legal markets (Pilar and Moguel, 2007).
In Central Africa, Gabon is the only country which, following a forest law reform in 2008, introduced the concept of "economic user rights" (Sartoretto et al., 2017). These are rights, recognized by the State to local communities, to market locally and without intermediaries, part of the collection of products derived from their customary use rights. Customary hunters selling game products outside their community must apply to a hunting permit and a commercial capture license. In addition, the Gabonese legislation provides that the possession and transport of the remains of species requires a certificate of origin, a zoosanitary certificate and a certificate of harvest (Sartoretto et al., 2017). Those requirements complexify processes far beyond the capacities of contemporary Gabonese hunters.
In the DRC, following the creation of two community permits for different purposes, there is a fairly clear legal framework for the marketing of game products. The commercial exploitation of wild animals and their by-products is possible only under a specific license subject to a number of conditions. The DRC also provides for a "commercial catch" permit among the special hunting permits, which can be allocated to communities (Sartoretto et al., 2017) but a "hunting ability test" is required for anyone applying for a hunting license (Sartoretto et al., 2017).
In Congo, user rights, including the right to hunt, are reserved for the satisfaction of the personal needs of their beneficiaries and the products derived from them cannot be used for commercial purposes. As such, commercial use in Congo is forbidden (Sartoretto et al., 2017).
### LAND TENURE AS A GUARANTY TO RESOURCE USE RIGHTS: FROM POSITIVE DISCRIMINATION TO FURTHER MARGINALIZATION
Land tenure and resource ownership are key elements guarantying to diverse degrees the rights of local communities over wildlife. Land tenure legislations in different tropical countries provide special rights to indigenous people and traditional ethnic groups. This, in a sense, can be interpreted as part of the exceptional rights intended to compensate for past inequities imposed on indigenous communities. In Latin America, legislations on indigenous rights have evolved, especially since the 1980s, with the rewriting of national constitutions and the ratification of the ILO Convention 169. In Guyana, titles were granted to 96 Amerindian Villages over Guyana's territory. Brazil and Peru together account for more than 75% of all indigenous lands in the Amazon (RAISG, 2016). However, many titles do not correspond to the land extent of indigenous territories, leading to claims (Dooley and Griffiths, 2014; Constantino et al., 2018). This is particularly critical because current indigenous lands might not be large enough to guarantee the sustainable subsistence hunting. Although policymakers recognize the importance of hunting for indigenous people, and participatory zoning has been encouraged, hunting is still poorly considered in land delimitation in Brazil and Peru (Constantino et al., 2018). In Mexico, biodiversity is still considered to be "property of the nation," but the UMA model implies the possibility to transfer the rights of ownership and usufruct (including sale) to legal landowners, private or community owned. In this sense, the system in place does not discriminate one group or another. However, because the UMA instrument was originally formulated with the goal of guaranteeing the legality of sport hunting and commercial wildlife farming in northern Mexico, the procedures for the establishment of UMAs were developed with the private land model in mind and are poorly adapted to communal land tenure systems, marginalizing de facto subsistence hunting by indigenous groups (Santos-Fita, 2018).
In Central Africa, one of the characteristics of the tenure systems is the coexistence of property rights based on the modern civil law system and customary land rights. The distinction between statutory rights and socially recognized customary rights is blurred in some countries and when statutory tenure rights are granted regardless of existing customary tenure rights, the resulting overlap creates conflicts and abuse (Sartoretto et al., 2017).
Efforts to protect the rights of traditional people have often left some marginalized groups behind. Indeed, the exceptional rights for indigenous groups concerning land tenure and resource use was often based on stereotypes and ended up fostering further marginalization of non-indigenous traditional groups. In Brazil, a line of interpretation understands that non-indigenous traditional groups have no granted rights to use wildlife, even in officially protected areas specifically designed to allow for sustainable use (Antunes et al., 2019; Pezzuti et al., 2019). In Guyana, subsistence use is not granted to Amerindians without title land, nor to afro-descendants, west-Indians and European descendants (Gomez, 2018). In Congo, the law recognizes customary land rights in different ways depending on whether they are local communities or indigenous peoples (e.g., pigmy groups) (Sartoretto et al., 2017). For indigenous peoples, preexisting customary tenure rights are recognized even in the absence of land titles. The law gives them the right to own, access and use the land and natural resources that they possess, occupy or traditionally use for subsistence but implementing decrees are still missing. Instead, for local communities (despite their traditional lifestyles and cosmovision), recognition of customary land rights follows a (a priori) simplified procedure, but still requires that the rights are registered to be recognized through a titling process.
## ACTIONABLE RECOMMENDATIONS
The "ample government tolerance" to hunting for food contrasts with the rather prohibitive legal frameworks and the discourse stigmatizing hunters as criminals. The question is: who benefits and who loses in the context of ill adapted and ambiguous legal frameworks? The absence of enabling legal frameworks for sustainable use is clearly detrimental to the most marginalized sections of the society and hampers the possibility to generate new knowledge and test innovative models for sustainable use.
The following recommendations are formulated to the attention of countries that are showing increased interest in recognizing the importance of hunting and the need to provide an enabling environment for sustainable use:
heterogeneous realities of the ecological and socio-cultural contexts of hunting.
### CONCLUSIONS
The analysis of available case studies describing legal frameworks concerning the use of wildlife for food highlights the need
### REFERENCES
to clarify the rights to sell surplus of meat and to reconsider sustainable use of wildlife in light of a new definition of subsistence hunting. It will also be key to support the articulation of land tenure with wildlife tenure in such a way that accounts for the realities and needs of communities from different cultural backgrounds. Without the revision of current inconsistencies, overlaps and gaps, there is little hope that investments in law enforcement will achieve tangible outputs for wildlife conservation and the livelihoods of marginalized groups.
### AUTHOR CONTRIBUTIONS
NvV had the idea of drafting this paper, collected information from different tropical countries on hunting for meat regulations, wrote a first draft and consolidated a final version based on the comments and changes made by other authors. AA, PC, and DS-F made direct contributions to the paper based on their research work in Brazil in regards to regulations on hunting for meat. DS-F also contributed from his research of this topic in Mexico. JG made direct contributions to the paper based on her research work on the legal framework on wildlife in Colombia and Guyana. ES made direct contributions to the paper based on his research work in Gabon and Congo in regards to hunting for meat regulations.
### FUNDING
This research received funds from USAID through the Bushmeat Research Initiative from the Forest Trees and Agroforestry Program, CGIAR.
from central Quintana Roo, Mexico. J. Ethnobiol. Ethnomed. 11:71. doi: 10.1186/s13002-015-0055-x
**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Copyright © 2019 van Vliet, Antunes, Constantino, Gómez, Santos-Fita and Sartoretto. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
# Hunting in French Guiana Across Time, Space and Livelihoods
Cecile Richard-Hansen<sup>1</sup> \*, Damien Davy <sup>2</sup> , Guillaume Longin<sup>3</sup> , Laurent Gaillard<sup>1</sup> , François Renoux <sup>4</sup> , Pierre Grenand<sup>5</sup> and Raphaëlle Rinaldo<sup>3</sup>
<sup>1</sup> Office National de la Chasse et de la Faune Sauvage - UMR EcoFoG, Kourou, France, <sup>2</sup> CNRS - OHM Oyapock, USR LEEISA 3456, Cayenne, France, <sup>3</sup> Parc Amazonien de Guyane, Rémire-Montjoly, France, <sup>4</sup> Independent Researcher, Paris, France, <sup>5</sup> IRD-OHM Oyapock USR 3456, Cayenne, France
Hunting sustainability in Amazonian ecosystems is a key challenge for modern stakeholders. Predictive models have evolved from first mostly biological data-based to more recent modeling including human behavior. We analyze here the hunting data collected in French Guiana through a panel of indices aiming at drawing the puzzle of parameters influencing hunting activity and impact in various socio ecological conditions across the country. Data were collected from five different study sites differing in cultural origins and remoteness from market economy, and over a 10 years period. Most indices show an impact on wildlife populations, and using a full set of indicators allowed us to better understand some underlying mechanisms that lead to a community's hunting profile. The results showed that there are noticeable differences between the study sites in the practices and the ways hunters face the changes in environment and resources availability.
#### Edited by:
Nathalie Van Vliet, Center for International Forestry Research, Indonesia
#### Reviewed by:
Pratheesh C. Mammen, Institute for Climate Change Studies, India Pedro De Araujo Lima Constantino, Independent Researcher, Brazil
\*Correspondence:
Cecile Richard-Hansen cecile.richard-hansen@ecofog.gf
#### Specialty section:
This article was submitted to Conservation, a section of the journal Frontiers in Ecology and Evolution
> Received: 31 January 2019 Accepted: 18 July 2019 Published: 08 August 2019
#### Citation:
Richard-Hansen C, Davy D, Longin G, Gaillard L, Renoux F, Grenand P and Rinaldo R (2019) Hunting in French Guiana Across Time, Space and Livelihoods. Front. Ecol. Evol. 7:289. doi: 10.3389/fevo.2019.00289 Keywords: sustainable hunting, French Guiana, indices, livelihoods, diachronic
### INTRODUCTION
Hunting sustainability in Amazonian ecosystems is a key challenge for modern stakeholders (Weinbaum et al., 2013; de Oliveira et al., 2018; Van Vliet, 2018). In temperate ecosystems, managing hunting deals most of the time with recreational activities, severe habitat loss and land use changes in highly industrialized/agricultural lands, and sometimes regulation of overpopulation issues, induced by ecosystem changes. At the opposite, neotropical wildlife managers often face the critical issue of human subsistence (Sarti et al., 2015; Van Vliet et al., 2017), with very different and poorly understood sociocultural drivers and human-animal relationships (Alves and Van Vliet, 2018; Shaffer et al., 2018a), but also occurring in a changing world.
In conservation and ecological literature, this topic peaked following the first Robinson and Redford (1991) sustainability model (Robinson and Redford, 1994; Alvard et al., 1997; Bodmer et al., 1997; Peres, 1997; Slade et al., 1998), then slowed down (Weinbaum et al., 2013) until recent researches bring new modeling approaches and insights in the late 2010s, including more and more human related dimensions. The first sustainability models were mainly based on ecological and biological data and provided a basic useful tool to detect clear overharvesting situations. They were widely used and sometimes misused when concluding to sustainability for harvest levels below the maximum potential yield, without assessing local ecologic and demographic parameters (Levi et al., 2011b). On the other hand, many studies reported that hunting considered unsustainable through the use of these indices has continued for decades with little to no evidence for prey depletion (Shaffer et al., 2018b). One of the main criticism is that they are static on-off indices, extrapolating, and inferring sustainability from spatial and temporal punctual situations (Levi et al., 2009, 2011b). They suppose closed systems, not taking into account possible animal movements in and out the catchment area. This is a major issue, as source-sink processes are thought to play a major role in animal densities balance on a larger scale (Novaro et al., 2000; Takashina and Mougi, 2015; Shaffer et al., 2018b; da Silva Chaves et al., 2019). From an operational point of view, another limitation comes also from the large amount of data needed if one wants to go further a first diagnose derived from maximum production conditions. In the absence of reliable data, uncertainty accumulates and predictions often failed (Van Vliet and Nasi, 2008). Finally, very few is known about the hunters' behaviors, choices and constraints that lead to the measured hunting profile.
The further step in modeling sustainability included spatial parameters, taking into account the variability in the distribution of habitats, of species and of hunters (Siren et al., 2004; Ohlschacherer et al., 2007; Levi et al., 2009, 2011b; Van Vliet et al., 2010; Takashina and Mougi, 2015). Such biodemographic models allowed to include gradual depletion of vulnerable species from a central foraging place, multiple or growing number of settlements (Shaffer et al., 2018b), and probability of animals moving within the area through a diffusivity parameter (Levi et al., 2009). They were successfully used to predict patterns of game densities and depletion in Guyana (Shaffer et al., 2017, 2018b).
Finally, last generation models include human behavior in the predictive sustainability. Hunter behavior is derived from predators' one, and based on optimal foraging and dietbreath theories (Rowcliffe et al., 2003; Levi et al., 2011a). Underlying hypothesis are that hunters tend to maximize their offtake considering opportunities and constraints, generally as central-place foragers. The more complex models consider a multi-preys system with various prey value (mainly protein intake), probability of encounter (declining with the distance from settlement according to species rarity, detectability and vulnerability) and kill rates, as well as the hunter's investment (increasing with distance to the settlement), and various limitations as cartridge availability and the need to come back home. These models aim at explaining the proportional representation of different species (prey profile) (Levi et al., 2011a).
Long term monitoring allows detecting the changes in game populations. Ecological studies directly compare density, abundance or biomass over places with various hunting pressure, and control "undisturbed" plots, sometimes accounting for habitat variability (Hill et al., 1997; Peres, 1997, 2000; Siren et al., 2004; Haugaasen and Peres, 2005). Other indicators of ecological changes can be based on animal performance (group size, reproductive success, and body mass), habitat impact and habitat use, providing relevant information on the population–habitat system and status, and quantitative basis for flexible management decisions (Morellet et al., 2007). In tropical areas, various indicators of species or ecosystem conservation status based on hunting data have been used (Milner-Gulland and Akçakaya, 2001; Jerozolimski and Peres, 2003; Ohl-schacherer et al., 2007; Parry et al., 2009b; Weinbaum et al., 2013; Constantino, 2016). Hunting data are theoretically easier to collect than ecological ones, and are used to build indices of impact on game population based on various models and hypothesis as reported above. Without concluding with a yes/no answer on sustainability, continuous trends of validated indicators reveal potential issues for the future, and monitoring allows to confirm or to correct the trends. Moreover, we argue here that a large set of hunting-based indicators informs not only on the impact of hunting, but helps to understand the modalities of the practice and of its changes over times and/or places (Parry et al., 2009a).
In French Guiana, hunting and wildlife management are still parsimonious and need modernization and deployment. As a piece of a "western" country dropped in tropical environment, the area hosts at the same time several indigenous populations still relying on bush meat for their subsistence, modern cities with full market access and a gradient of rural areas keeping strong cultural and alimentary relationships with the forest. For the last 10 years, the government has expressed a growing interest for developing adapted management rules, based on local and scientific knowledge (Richard-Hansen and Hansen, 2004).
This study intended in first objective to document the changes in hunting-based indicators of game-species conservation status over 10 years across several villages of French Guiana. Moreover, our panel of indicators informs not only on the wildlife status, but also on the different strategies deployed by hunters according to various contexts, or to face the variations in their natural resources, helping to better understand their underlying behaviors and incentives.
### STUDY AREA
French Guiana (2◦ 7 ′ -5◦ 44 ′′ North, and 51◦ 38′ -54◦ 35′ West), is a French overseas department, situated between Suriname and the Brazilian state of Amapá, covering ∼85,000 km<sup>2</sup> in the eastern Guiana Shield. Altitude generally ranges between 0 and 200 m above sea level (mean 140 m) with some peaks ranging from 500 to 800 m. Annual rainfall ranges from 3,600 mm (northeast) to 2,000 mm (south and west). Mean annual temperature is of 25.7◦C. Evergreen rainforest covers more than 90% of the country (FAO, 2010). In 2014, 88% of the population (250,400 people) lived in the coastal strip in human-modified areas (artificial, agricultural, and disturbed areas) covering ∼1,000 km<sup>2</sup> (ONF, 2016). Outside this area, the average population density is 0.04 people km<sup>2</sup> (INSEE, 2012). Forest logging is restricted to less than a third of the territory, under National Forest Office control. The area used to be considered as well-preserved (Taber et al., 2008), but presently suffers from persistent and growing impacts of illegal goldmining (Hammond et al., 2007; Dezécache et al., 2017), inducing habitat destruction, rivers asphyxia, and uncontrolled hunting.
Mainland French hunting laws were not suitable to the tropical ecological and sociological contexts, and consequently do not apply in French Guiana. However, conservation laws protect the most vulnerable species, and several protected areas regulate access to hunting. Since the last 10 years, things are on the move. Most recent regulations established hunting bags for some species, based on a number of pieces allowed to harvest by day and hunter. In 2018, a new law created a specific Frenchguianan hunting license, freely and systematically attributed to all residents asking for it until 2020.
Many different cultural and ethnic groups are present in French Guiana: Creole, Bushinengue (Aluku, Paamaka, Saamaka, Ndjuka), Hmong, Chinese, Brazilian, Surinamese, people coming from mainland France, and six different Amerindian groups (Wayãpi, Wayana, Teko, Kali'na, Palikur, Arawak-Lokono) coexist in the country. Owing to the first article of the French 1958 Constitution, stating equality of rights to all French citizens, a specific status cannot be recognized for native populations, and there are no indigenous reserve nor specific rights given to these populations. However, specific uses and livelihoods are recognized. In 2007 a national park was created (Parc amazonien de Guyane) in the southern part of the region, an area encompassing the historical territories of several indigenous populations of various ethnical origins, with the willingness to protect and promote both natural and cultural richness of the area. To fit with both the constitutional law and the spirit of the project, "communities of inhabitants traditionally subsisting from the forest" are allowed to hunt in the Park. These communities also benefit from exception of recent regulation law on hunting bags and seasons. This exception rule also applies in "Zones of Common Use Rights" outside the Park, as several native groups also live in the most northern area (Kali'na, Palikur, Arawak, Saamaka, Ndjuka) (Filoche et al., 2017). However, protected species remain strictly forbidden to harvest for anybody. The National Park authorities are willing to set up co-constructed management rules, taking into account traditional and cultural needs. A particularity to underline is that, as French citizens, local communities benefit from national social incomes and/or may work and have salaries, allowing the purchase of motorboats, gas, firearms and cartridges, which influences hunting patterns (Tritsch et al., 2015).
## METHODS
### Hunting Data Collection
The first hunting surveys were initiated between 2000 and 2002, to document hunting practices, quantify and describe the harvest, initiate a monitoring of its impact, and try to estimate its sustainability (Grenand, 2002). In selected study sites, hunters voluntarily shared the results of their hunting activity, recorded by local field assistants. Hunters' self-monitoring was attempted but did not succeed, because most hunters rejected or rapidly gave up, with exception of very few motivated participants. Data collected for each hunt included the number of pieces of all species harvested, the duration of the hunt (from home to home, including traveling to the hunting area), the number of hunters, and information about the practice (weapon, means of transport used). We distinguished the harvests occurring during other activities (agriculture, transport), considered as opportunistic catches. The hunt location was mapped according to the hunter's indications on a 5 × 5 km grid. A form was filled even in case of unsuccessful hunts. The harvested biomass was estimated afterwards, from a database of mean weights of animals issued from local data when possible (Richard-Hansen et al., 1999; Richard-Hansen pers. data) or from literature when not available.
Around 10 years later, from 2010 to 2015, new surveys were conducted, some of them on the same sites. Surveys conducted between 2000 and 2002 are grouped as survey 1, and those conducted in the 2010s as survey 2 (**Table 1**). Among all the study sites, we selected five for which data were the more complete in both surveys, and lasted at least several months. These study sites represent a gradient from higher subsistence and isolated conditions (Trois Sauts, the most southern site) to more rural context, with growing access to market economy (Saint-Georges de l'Oyapock). Trois Sauts, Camopi, and Elahe are located in the southern part, within the National park, are mostly inhabited by Amerindian communities (Wayãpi, Teko, and Wayana), and cannot be reached by road (**Figure 1**). Régina and Saint-Georges are mostly Créole villages, with a mixed population including Palikur (Amerindian community) and Brazilian people, in the northern part of French Guiana. Conducting hunting surveys in larger towns proved to be more challenging, and could not be addressed here.
### Data Analysis
Variables describing the hunting strategies included the mean duration of the hunting trips, the proportion of daily vs. overnight trips; the proportion of hunts using a motorized vehicle (may be car or motorboat), and the number of hunters. Variables describing the hunting results included the total number of preys and the total biomass returned in each hunt, the mean prey biomass, the proportion of species or species groups (prey profile), and in particular the Rodents/ungulates ratio. The Catch per unit effort (CPUE = biomass/hunter/hour) was calculated as an integrated indicator of returns over effort. The mean number of preys and the mean biomass collected by each hunter in a single trip were also computed. Duration and time related indicators were calculated excluding opportunistic catches, for which the spent time was not related to hunting effort. Some large species known to be the most vulnerable to hunting pressure (Robinson and Redford, 1986, 1991) were analyzed separately, or grouped together as sensitive species: the tapir (Tapirus terrestris), the black spider monkey (Ateles paniscus), the Artiodactyla (peccaries and deers: Tayassu pecari, Pecari tajacu, Mazama nemorivaga, and Mazama americana), and the large Cracid black curassow (Crax alector).
The hunting spatial pattern was estimated by the total catchment area, calculated by the number of grid cells with at least one catch. The shape of the catchment area was described by the perimeter/area ratio. For each prey, the straight-distance from the catch to the village was calculated from the centers of 5 × 5 grid cells.
All indicators reflect an underlying hypothesis regarding impact of harvest on the game population and/or an associated hunting practice, according to the different theoretical frameworks presented above (**Table 2**).
More data were collected during the second survey because a much larger effort has been involved in the program. However, we checked that sampling enclosed both rainy and dry season in each place and survey, to avoid bias related
FIGURE 1 | Map of the five study sites, and the catchments areas during surveys 1 and 2. Catchment areas were represented on different maps to avoid overlapping between sites. (A) catchment areas of Trois Sauts, Saint-Georges, and Elahe villages. (B) catchment areas of Camopi and Régina. Each square is a 5 × 5 unit with at least one kill recorded during the considered survey. White cells: catchment during survey 1 only; black cells: catchments during survey 2 only; gray cells: catchments during both surveys.
to seasonal activities, and that no indicators were correlated with monitoring effort (as the number of days of study in each village and survey) (Spearman correlation r < 0.5, p > 0.2 for all comparisons). Only one weak correlation was found, with the use of motor transportation (r = 0.6, p = 0.05), but we assumed no causality. To explore the relationships between hunting strategies and hunting returns, together with time-(two surveys) and site-scales effects, we first ran stepwise regressions to fit generalized linear models with backward selection of candidate variables, and selection criteria based on AIC. Response variables were log transformed, and Poisson regression was used on count data (number of preys), with a goodness-of-fit chi-squared test on the residual deviance. To closer examine the time effect on our indicators, mean values from surveys 1 and 2 were compared by student t-test or Wilcoxon tests, and proportions by khi square tests. Spearman rank correlation test were conducted on paired variables.
### RESULTS
### General Patterns
Five thousand nine hundred and seventy one hunts and Twenty one thousand five hundred and fifty six
kills were recorded during the whole study, in the five study sites.
We first explored the effects of some possible explanatory parameters on selected indicators as response variables. Do the number of hunters participating in a hunt, the duration of the trip, the use of motor to reach the hunting place, the date of study (survey) or the study site influence hunting results, in terms of total number of preys, total biomass, or mean prey biomass? Spearman correlation matrix between variables showed first that the number of hunters was positively correlated with the duration of the hunt (r = 0.82, p < 0.01). The total biomass harvested by hunt increased with the number of hunters (r = 0.64, p < 0.05) but not with the duration of the hunt (r = 0.2, p > 0.0.5), while the biomass per hunter is not significantly correlated. The duration of motorized hunts is significantly higher (7.5 vs. 5.7 h; p < 0.01 Student t-test), and these motorized hunting trips provided more biomass (19.3 vs. 13.1 kg; p < 0.01 Student t-test) and more preys (4.3 vs. 2.5 preys; p < 0.01 Student t-test), with preys being a little bigger (4.7 vs. 3.9 kg; p < 0.01 Student t-test). However, the biomass collected for each hunter is not different (12.5 and 12 kg/hunter of biomass, t-test p > 0.5).
Overnight trips globally provided less preys (24%) but more biomass (38%) that expected on the basis of days allocated (30% of hunting days), both proportions being significantly different from null hypothesis (χ 2 tests, p < 0.001).
General linear models including all explanatory variables (number of hunters, duration, use of motor engine, survey, and site) returned poor fitting and predictive power. According to the selected model (quasi poisson regression, full model), the most important significant positive parameters on the total number of preys was the number of hunters and the use of motor to go hunting (**Table 3**). The goodness-of-fit chi-square test on residual deviance was not significant (p > 0.5), indicating the model fits the data. Negative parameters corresponded to sites effects, mainly in Régina and Saint-Georges. The survey variable had no significant effect. The full model was selected to partly explain the total biomass returns (R <sup>2</sup> = 0.18, p < 0.001, Gaussian regression). The number of hunters and the same sites as previous analysis (Régina and Saint-Georges) positively influenced the returned biomass, while Trois Sauts site and the survey negatively did.
Mean prey biomass (logged) was negatively influenced by the survey and motor use, positively in Régina and Saint George and Elahe, and negatively in and Trois Sauts (R <sup>2</sup> = 0.21, p < 0.001). The duration of hunt has no significant influence, and was not selected in the best model after backwards selection, based on AIC values.
These first results show that general models on the whole dataset explain only a small part of the variance, and that the site effect is strong. Trois Sauts site differs the most from others, with higher number of preys but lower total and prey biomass. TABLE 3 | Variables explaining the variation of the hunting returns, in terms of (i) number of preys by hunt (quasipoisson regression, N = 3,192, Dispersion parameter taken to be 0.2899811, GOF test on residuals >0.5) (ii) total biomass by hunt (gaussian regression, N = 2,889, adjusted R-squared: 0.18, p-value: <2.2e-16) (iii) mean biomass of individual preys, N = 16,328, adjusted R-squared = 0.21, p < 0.0001).
Backward stepwise selection selected the full model for total number of preys and total biomass by hunt, and model without the duration variable for the prey biomass (AIC criterion) \*p < 0.05; \*\*p < 0.01; \*\*\*p < 0.001.
On the other hand, Régina and Saint-Georges sites generally have patterns that are more similar.
### Spatial Patterns
The size of the hunting areas vary from 200 to 2,775 km<sup>2</sup> , as estimated by the number of 25 km<sup>2</sup> grid squares with at least one catch recorded (**Table 4**), according to the survey and the study site. The shape of the main hunting area (excluding isolated square grids cf. map **Figures 1A,B**) was characterized by the perimeter/area ratio. For all study sites, it was around 0.3 (up to 0.44 for Régina), indicating an elongated shape. Rounded shapes corresponding to the same areas would have a ratio inferior to 0.1, except for Elahe.
Mean catchment distance during daily hunts is lower for birds (13.8 km) than mammals (21.7 km), and Reptiles (27.3 km) (p < 0.001 Student t-tests for paired comparisons) (**Figure 2**). Among mammals, it is the highest for Tapirus terrestris (28 km) and large primates (26 and 29 km from settlements for Ateles and Alouatta, 25 km for Cebus apella vs. 8 km for Saguinus, paired t-tests <0.001). Rodents and Artiodactyla are both harvested at around 20 km from settlements. For Rodents, the distance is higher for Agouti paca (26 km) than Dasysprocta leporina (10 km, p < 0.0001 t-test), and among Artiodactyla, it is higher for peccaries (respectively, 21 and 22 km for Pecari tajacu and Tayassu pecari) than for brocket deers (15 and 14 km for Mazama americana and Mazama nemorivaga, respectively, p < 0.0001). Among birds, largest distances of hunting are recorded for the largest species as black curassow Crax alector (24 km), trumpet birds (Psophia crepitans, 22 km), or Penelope guan (Penelope marail, 17 km), while the
Hunting in
French Guiana
TABLE 4 |Changes in various indices between two hunting surveys (S1 and S2) conducted 10 years apart in five study sites of French Guiana.
\*Catchment area is estimated on the basis of the number of grid cells (5 × 5 km) with at least one kill recorded\*p<0.05; \*\*p<0.01; \*\*\*p<0.001.
smaller species from Columbiform, Passeriform, Psittaciforms orders are collected within a ten km radius (6, 8, and 9 km, respectively). Among Reptiles, the crocodilian are harvested nearest villages (28 km) than Squamates (mainly Iguana iguana, 35 km).
### Trends Over 10 Years
Trends of most indices were compared between the two surveys period. Most indicators showed clear changes between the two surveys, but also varying according to the study site.
### Hunting Returns
The total biomass per hunt and the various hunting yields (CPUE, biomass harvested by hunter) tend to decrease everywhere between the two surveys, but less markedly in Trois Sauts (**Table 4**, Mann Whitney test). As suggested by the results of general linear models, the number of preys harvested by hunting trip does not change a lot between the two surveys, although it tend to increase in Trois Sauts and decrease in Régina. The mean biomass of preys (e.g., body size) significantly decreased in three sites, whereas no changes were observed in Camopi, and it slightly increased in Trois Sauts. The mean biomass of primates did not change, indicating no shift to smaller primate species.
### Prey Profiles
The proportion of mammals significantly decreased (Chi square test, p < 0.001 for all sites, comparing survey 1 and 2), and the proportions of birds increased (Chi square test, p < 0.001 for all sites except Elahe) in all prey profiles of survey 2 compared to survey 1 (**Table 4**). Regarding Reptiles, their proportions increased in all sites but Trois Sauts, where their proportion decreased. Among mammals, the proportion of Artiodactyla clearly decreased everywhere, while the primates' one decreased in Trois sauts and Elahe but remained stable in Camopi and Régina, and increased in Saint-Georges. In particular, the largest, most vulnerable and preferred primate species Ateles paniscus presented the same trends as the primates group considered as a whole. The mean number of sensitive species (see Methods) harvested per hunt tend to decrease, except in Trois Sauts and Camopi. The Rodents/Ungulates ratio increased in all sites.
### Hunting Strategies
Considering the hunting strategies, the mean duration of daily hunts significantly increased in all sites (**Table 3**, Mann Whitney test all p < 0.01, except for Elahé, **Table 4**). The mean number of hunters participating in a hunting trip tended to increase in Trois Sauts and Saint-Georges. The proportion of motorized access to hunting decreased in Trois Sauts and Camopi, and increased in Saint-Georges, and the proportions of overnight trips show no clear tendency.
### Spatial Pattern
The size of total catchment area increased in the three southern villages (Trois Sauts, Camopi, and Elahé) but not in the two other ones (Régina and Saint-Georges), as estimated by the number of quadrats with at least one catch recorded (**Table 4**). The map shows that the most clear spread of hunting area occurred in Trois Sauts, from which hunters clearly went farther for hunting during survey 2 (black squares indicate areas used only during survey 2, **Figures 1A,B**). In Camopi, we observe a more intensive use of middle distance areas, and some new rivers explored (**Figure 1B**). One consequence is a higher overlap of hunting areas between villages (Camopi/Trois Sauts, and Camopi/Saint-Georges de l'Oyapock) during survey 2.
Mean catchment distances tend to increase for daily hunts (but decreased in Régina), but for overnights hunts only in Camopi. The straight distance of catch of sensitive species as primates, Artiodactyla or Perissodactyla did not change markedly (**Table 4**).
### Inter-sites Differences
Many previous results pointed how indicators' changes over year differed according to study sites. A full analysis is currently under process to characterize a hunting typology in French Guiana, but we can point out here some major differences between our five study sites.
### Hunting Strategies
Overnight trips were much more common in Saint-Georges de l'Oyapock and Régina, representing around half of the records (54 and 43% of occurrences, **Table 5**). Most overnight hunts lasted 3 days in average (87% between 2 and 4 days), so this hunting practice represents up to 78% of hunting days recorded in Saint-Georges de l'Oyapock. Regarding the number of preys and the harvested biomass, overnight trips bring back a lower amount of game or biomass than expected on the basis of proportion of days allocated χ 2 tests p < 0.001), except in Elahe in which 23% of catches and 21% of biomass occurred during the 2% of time spent in overnight hunting trips. In Camopi, less preys but more biomass than expected was harvested during overnight trips (χ 2 test p < 0.001)
Voluntary daily hunts were also longer in Saint-Georges de l'Oyapock (9 h in average, **Table 5**) than in other sites (6.7 h in average, t-test, p < 0.001). Using a motor vehicle (car, boat or
#### TABLE 5 | Hunting strategies in five study sites of French Guiana.
motorcycle) to reach the hunting area is much more common in Saint-Georges and Régina (78 and 96%) than in the southern sites, with a minimum of 44% in Elahé.
Biomass harvested per hour and per hunter (CPUE) is of same magnitude for the three southern villages (Trois Sauts, Camopi, and Elahe: respectively, 2.5, 2.4, and 3 kg/hunter/hour), and clearly lower than hunting yields from Régina and Saint-Georges de l'Oyapock (6.8 and 10.2 kg/h/h, respectively).
In southern sites (Elahe, Trois Sauts, and Camopi), opportunistic harvests (i.e., performed during another activity as travel or agriculture) provide a 10–14% proportion of catches, while it is negligible in Régina and Saint-Georges (**Table 5**).
### Prey Profile
Over the whole study, mammals represent 36% of the number of animals harvested, birds 56% and Reptiles 8%. However, 86% of the biomass is mammals, and 8% birds. There are large differences between the five study sites regarding the prey profile (**Figure 3**). Mammals, and in particular ungulates clearly dominate the harvest for Saint-Georges de l'Oyapock and Régina, while birds (67% of preys in Trois Sauts) and particularly toucans (36% of preys in Trois Sauts) are the most hunted species in southern Amerindian villages. For these villages, primates are an important part of the harvest (14–27%), representing 19– 26% of the biomass. Regarding protein intake, the large amount of harvested toucans brings <3% of the biomass consumed. Tapir hunting provides 12% (Camopi) to 39% (Régina) of the biomass harvested.
The Rodent/Ungulate ratio is higher in the three southern sites (1.28, 0.96, and 0.63 in Trois Sauts, Camopi, and Elahe, respectively), than the two other ones (0.33 and 0.21 in Saint-Georges de l'Oyapock and Régina, respectively).
In all sites, overnight trips bring back a larger proportion of large primates (Ateles, Alouatta, and Cebus), large birds (Galliform), Crocodilians and Iguanas than expected on the basis of number of days spent.
### Spatial Patterns
Among study sites, Elahe has the most spatially concentrated hunting pattern, the majority of hunts taking place within 4 km, except for primates (**Figure 4**). Hunters from Trois Sauts mostly hunt within a 10 km radius, except for primates, and rodents. For Camopi, the distance is the highest for primates. The two northern sites, Régina and Saint-Georges, have the wider spatial patterns, catching all preys over 20 km away from the village. The total catchment area, estimated through the number of grid cells (5 × 5 km) with at least one catch, was the largest for Camopi with 121 grid cells (3,025 km<sup>2</sup> ), and the minimum area value is found in Elahe (**Figure 1A**).
### DISCUSSION
The main objectives of the study was to assess the hunting impacts on the animal community and to scan its evolution over a 10-year period, using hunting-based indicators according to various theoretical frames presented in the introduction. We also wanted to explore hunters' strategies in relation with hunting results, and how they changed or adapted over the years.
Concerning impact on game populations, we can underline that hunting areas and distances recorded here are generally higher in our study comparing to literature on Amazonian traditional hunting patterns. Traditional central foragers generally spread over a 10 km radius (Constantino, 2015; Shaffer et al., 2018b), while it is the smallest mean distance recorded in this study. We showed that catch distances of species known to be vulnerable as tapirs and larges primates are significantly higher than other species, which is coherent with central foraging theory predictions, and denote their progressive depletion around villages. Among larger bodied species, artiodactyla (peccaries and brocket deers) are encountered at comparatively smaller distances from central places, whereas they are considered as preferred items and vulnerable species. Despite their body mass, brocket deers are considered as to be less vulnerable and preferred species than peccaries, because they are difficult to hunt (Levi et al., 2011a), or taboos (Shepard et al., 2012), and their harvest is often considered as sustainable (Hurtado-Gonzales and Bodmer, 2004). In this study, they appear to persist closer from villages than other large game species. Peccaries are generally used as indicator species, strongly impacted by overhunting, but we previously found that their abundances in French Guiana are highly unpredictable and driven by large temporal and geographical scale process not only directly related to local hunting pressure (Richard-Hansen et al., 2018, 2019). The Rodent/Ungulate ratio is a commonly used indicator, useful because synthetic, but it is however strongly influenced by the presence of white lipped peccaries, in the environment and consequently in the harvest. It therefore has to be carefully interpreted, and considering a larger context. Birds follow the same predictable pattern, with large species being probably extirpated around villages, and smaller species harvested around them. Rodents are generally considered as second choice, but we found that the Agouti paca seems to be depleted in the vicinity of villages also. Note that Reptiles (mainly green iguanas and crocodilians) are also harvested at very large distances, which would indicate that these games are among the most depleted around villages.
Hunters' practices influence only slightly the quantitative outcomes of the hunts. The total hunting returns seem to increase with some allocated efforts, as the number of hunters and the use of motorized engines, but not the real benefits for each
hunter. The overnight trips allowed harvesting smaller number but higher biomass of preys compared to the expected quantities based on the number of allocated days, which proved they are efficient for large game harvesting. Overnight and longer hunting trips and as well as motor use allow farther hunting from the village (as they do for slash and burn agriculture, Tritsch et al., 2015), giving access to less disturbed areas with largersized and more abundant game, but requires more hunters, probably for cost sharing and/or security (Siren et al., 2012). The "benefit" is counterbalanced by the necessity of sharing the total amount harvested.
### Trends Over Time
Comparing over 10 years, many indices point out a probable diminution of game populations around all the studied sites, with clear differences between them. The most general tendency observed indicating the decreasing resources is the growing time allocated to hunting, while the total biomass brought back home generally decreased despite this increasing effort, with the noticeable exception of Trois Sauts. Although we found that using motorized vehicles or boat may increase the yields, as well as going to overnight hunting trips, hunters did not systematically increase this behavior to face the decreasing yields. At the opposite, although increasing the number of hunters does not increase the amount of biomass, we found a general tendency to go hunting in larger groups. In the northern part, this can be related to a need of sharing costs and increase security. In the southern villages, younger people tend to loose traditional knowledge because of time spent in school, and need to be accompanied by eldest for cultural transmission (P. Grenand, obs. pers.). These examples also show that the hunters' behavior is not always optimal, or that this optimization has to deal with many complex constraints. Concomitantly with longer trips, we found that hunting areas expanded, and/or partly shifted. In Trois Sauts, for example, hunters used more intensively the upriver area, which is not farther from the village, but harder to reach because of strong rapids and rocky bars on the river, increasing the effort for hunters. This probably explain that proportion and hunting distances of sensitive species of primates for example did not increase substantially. This area used to be prospected for long time by a small part of the population, leaving upriver from the falls, but increasing the population size also probably leading to a social reorganization of hunting territories (Tritsch et al., 2015).
The other expected changes when facing decreasing resources is shifting to more resilient and/or less preferred species, generally smaller ones (**Table 2**). This is illustrated in our results by the general increase of the Rodent/Ungulates ratio, and decrease of the size of preys and the proportion of mammals. Hunting rodents and smaller species is generally a second choice (Suárez et al., 1995; Levi et al., 2011a), but these species with higher reproductive rates (Robinson and Redford, 1986) are less impacted by hunting pressure and more resilient. However, when hunting turns to more recreational or commercial activity, hunters tend to focus on a few attractive species, and are not systematically shifting to smaller preys (Redford and Robinson, 1987). In Régina and Saint-Georges, ungulates are undoubtedly preferred preys, but their proportion nonetheless decreased clearly between the two surveys, while the proportion of rodents increased in Régina and not Saint-Georges.
Diet breath theory also predicts that much preferred species will be pursued with increasing effort until intense depletion (Levi et al., 2011a). Primates are a highly preferred game for Amerindian populations but are very sensitive to harvest owing to their very low reproductive rates (Bodmer, 1995; Robinson, 2000). The proportion of primates in the prey profile of Trois Sauts previously increased from 11.3 to 19.3% between 1977/78 and 95/96 (Ouhoud-Renoux, 1998b), when the hunting technology shifted from bow to firearms (Grenand, 1995), increasing their hunting efficiency. However, we observed that this proportion tend to stabilize until our first survey in 2002 (18%), and began to decrease after 10 additional years of harvesting (13% in 2012). Increasing effort can be measured by the increasing of time and/or distances allocated to reach favorite game species. The mean distance of catch of primates (and in particular for Ateles) effectively increased in Camopi, but not in other sites, and even decreased in Trois Sauts, probably in relation with a change in hunting area (see above). The mean biomass of primates did not change, indicating no shift to smaller species. These results indicate that the depletion of large primates is not achieved around Trois Sauts or Camopi, as hunters still manage to find these favorite games, but that the decline is initiated.
Our results also highlight the differences between our five study sites, and particularly between the three most traditional villages of the National park in one hand (southern sites) and the two rural ones (northern sites) in the other hand. The three autochthonous villages living for great part from subsistence hunting (Trois Sauts, Camopi, and Elahe) present some similarities in their practices and in the evolution of their practices over time. They used to hunt more birds than mammals, and larger proportion of primates and smaller birds as Toucans. The larger bird consumption is generally related to subsistence contexts (Benítez-López et al., 2017), as the larger width of the diet (Grenand, 1980, 2002). Comparing to rural sites (Régina and Saint-Georges), these hunters go hunting more often alone, with less motorized transport mean, for daily hunts on smaller distances. Harvesting animals is an everyday reality, regularly implemented during other activities as transportation, fishing, or slash-and-burn agriculture ("opportunistic catches") (Grenand, 1980). In this place, there is really no market, subsistence and proteins need are absolute for everyday meals. In this situation, hunters manage to come back with a similar amount of biomass, increasing efforts, distances, or shifting preys.
In Régina and Saint-Georges de l'Oyapock, most hunters use motor vehicles and boats to go farther and for longer hunts farther from the village, more often with especially dedicated overnight trips. The depletion in the 21st km away from the villages is probably strong, as proved by the long hunting distances recorded. However, the small proportion of primates harvested is not easy to interpret: they could have been severely depleted, but as they are not pursued as preferred game, this indicator could be non-representative in this case. Large mammals and particularly ungulates are the main preys, providing hunters higher biomass yields although they hunt in larger groups. These high yields do not reflect healthy populations, but increased efforts over very large areas, allowed by modern means and money. In these sites, hunting is still a strong tradition, and provides protein complement to low income rural populations. However, bushmeat is not a survival need anymore, and access to market modifies the relationships with natural resources in contradictory ways, providing both alternative protein source, and bushmeat trade opportunities. Hunters tend to reduce they prey choice, seeking mainly ungulates, tending to less traditional prey profiles (Redford and Robinson, 1987), but wild meat remains on the menu (Alves and Van Vliet, 2018). According to their incentives, hunters will not face the changes in their environment on the same way.
### Spatial Patterns
As often underlined, the use of space is the major key (Levi et al., 2011b; McNamara et al., 2015; Takashina and Mougi, 2015; Constantino et al., 2018; Shaffer et al., 2018b; Van Vliet et al., 2018). Most models rely on central foraging models, with a homogeneous access to the forest from the settlement. However, the size and shape of a hunting area depend both on geographical (penetrability of the environment) and economic parameters (De souza-Mazurek et al., 2000; Siren et al., 2012; Siren and Wilkie, 2015). Infrastructure expansion has been widely related to the increase of wild meat harvest and trade in the tropics (Laurance et al., 2015; Benítez-López et al., 2017). As long as there is no new access paths, hunting will be concentrated in reduced areas, constrained by physical limits. This explains in particular why the hunting areas expanded few across the years, despite decreasing
available resources. In our study, most hunting areas are already strongly enlarged, and widely elongated along major access ways. In Camopi, the social incomes together with the presence of two major rivers allowed the inhabitants, by equipping themselves by canoe and motor to increase their predation and agricultural zone (Tritsch et al., 2015). Other parameters can limit the expansion of the area. In Elahe, although the village is also located along two main rivers, hunters remain on a small territory, apparently taking few advantage of this opportunity to expand their catchment area. This is partly due to competition with other populations for access to the resources (Davy, pers. data), as the territory of this village is surrounded both by other villages (Wayana and Businenge) and illegal gold-miners. In Trois Sauts, hunting territories spread from 770 to 1,180 km<sup>2</sup> between the years 1976/1977 and 1994/1995, to face the population increase (Ouhoud-Renoux, 1998a). Methodologies used to estimate the superficies are not directly comparable, but it seems that 20 years later, their territory increased again a little. Indeed, the main changes consist in the creation of new settlements related to population increase, rather than increasing size of the hunting territories of initial villages we sampled (Davy et al., 2012). In this case, there is no competition with other groups, but topography do not allow further expansion.
We think that the hunting impact on wildlife populations around our study sites is manifest, but still concentrated around settlements and access paths. The larger, linear and ramified shapes of the catchment areas in our study sites enhance the sustainability of the harvest (Constantino, 2015; Tritsch et al., 2015), diluting the harvest on larger superficies, and providing more sources-sink exchanging systems along elongated frontiers than a circular central foraging traditional catchment area (Salas and Kim, 2002). Although the area are larger than estimated in more traditional contexts (Constantino et al., 2018), this can be viewed as positive changes through modernized practices. Allowing spatial expansion of hunting would spread the impacted area, but may increase the sustainability, as long as the total offtake does not increase (Stearman and Redford, 1995; Ouhoud-Renoux, 1998a; Grenand, 2002). Improving sustainability of hunting in isolated villages may therefore go through favoring their movements, as it was implemented in some ethnodevelopment projects (Stearman and Redford, 1995). Constantino et al. (2018) provide a full analysis of the integration of hunting and source areas in the delimitation of indigenous lands in Brazil, taking into account population growth. They propose useful scenarii to be implemented, as redistributing villages respecting distances between them to ensure functional source-sink systems. This may be discussed with communities, as it may echo a behavior currently spontaneously emerging in some places (Tritsch et al., 2015), but in French Guiana may be difficult owing to the very large hunting territories recorded.
The sustainability diagnosis of a harvest depends not only on the scale of the catchment area (Robinson and Redford, 1994), but also on the scale of the analysis (Hill and Padwe, 2000; Shaffer et al., 2017). Shaffer et al. (2018b) concluded that the WaiWai hunting may be considered as sustainable within the Indigenous reserve as a whole, because sensitive species will be extirpated from <13% of the total area of their Reserve, and persist in the catchment area. Novaro et al. (2000) estimated the size of refugia (area with unharvested population) needed to prevent overharvesting, in relation to reproductive parameter of species and proportionally of the size of catchment area. A buffer area of 15 km around the main catchment area of Camopi represent a total area of more than 11,000 km<sup>2</sup> . Proportionally to the hunted area of <3,000 km<sup>2</sup> , 73% are non-hunted and act as refugia and adjacent source for game species, which fulfill the requirements for sustainable use of most vulnerable species as spider monkey and tapir (Novaro et al., 2000). The situation is the same for Trois Sauts, insuring sustainability on large scale, but probably hardly in Elahe, because of neighboring populations, as well as Régina and Saint-Georges for which only the southern part of the catchment area is surrounded by non-hunted areas. However, even in most favorable cases, the sustainability issue has not to deal with the species conservation but more with the survival of these human communities. Finding proteins every day may however become very hard for them, who probably reached their maximal capacity of spatial extension.
### Sustainability Models and Indicators
In the literature, most models are elaborated from data from one study site or community, and assuming hunters acting with optimal behaviors. We saw here that hunters' behaviors are complex, not always optimal, and driven by many sociological, cultural, economic and geographical constraints. Comparing with Levi et al. (2011a) results and predictions, some of our results do not fit their hypothesis, and some do, depending on the socio-economic and cultural context. As predicted by Levi's model, for example, we found hunters' return rates remaining consistent even when vulnerable species were depleted around the village of Trois Sauts, but they decreased clearly in the other sites. However, they predicted that lower-value game do not expand into the diet in a depletion scenario, but we found a general increase of the rodent/ungulate ratio, of birds and smaller preys. The game value is a strong cultural parameter, not only dealing with protein amount or probability of killing. The proportion of large primates, being considered to be the most vulnerable species, is generally thought as an indicator of game depletion and hunting sustainability. However, in some places, their proportion in prey profile decreased less than ungulates 'one, probably because they are very actively researched in those communities. Many reasons may lead to kill or not kill an apparently valuable prey, as complex dietary taboos, taste of the meat, cultural or religious bans that are likely to change along the time. In French Guiana, the Busi nenge community (Maroon's descendants) has a strong demand for Ateles killing for ritual mourning ceremonies. Wayãpi had a strong relationship to living environment, and their harvest used to be controlled by the precept "No doing too much" (Grenand and Grenand, 1996). Those populations coexist in French Guiana, aside urban and rural ones, having very different constraints and incentives.
### CONCLUSION
Our results alert on the risk of relying on too few indicators to assess hunting impact, sustainability or game depletion in different contexts. Indicators sometimes give apparently contradictory conclusions, but this is without accounting hunters' strategies. For example, a stable CPUE value can mask an increasing harvest together with a proportional increasing effort, therefore not going toward sustainability but overharvest, until a tipping point. Using a full set of indicators allowed us not only to look for hunting impact and sustainability, but also on some of the underlying mechanisms that lead to a community's hunting profile. We saw that each community has a different answer to the changing conditions, according to its own values and constraints. Integrative indices are useful, but may reflect different realities. Modeling detect general trends, but human behavior are complex to model and management needs to adapt to local constraints (Renoux and de Thoisy, 2016).
Long term monitoring with hunters' participation remains the best way to understand and accompany the changes in behaviors, practices and impacts of hunting. It provides basic elements for a flexible management, which can be improved by the integration or recovering of traditional knowledge (Berkes et al., 2000). Modern tools of adaptive management include the concept of learning processes to improve the knowledge of the system functioning, in particular in situations of high uncertainty levels (Keith et al., 2011).
However, our results suggest also that hunting management in French Guiana should be considered differently in southern and northern part of the territory, owing to different hunting practices, accessibility level to the wildlife resource, gradient of modernity in the livelihoods, and former and cumulated human impacts. Moreover, we highlighted that the sustainability of highlevel harvests depends mainly of the presence and persistence of large non-impacted areas surrounding the catchment areas. Southern French Guiana remains quite preserved by its remoteness, but this relative quietness is presently severely threatened by the diffuse sprawl of illegal goldmining.
### REFERENCES
### AUTHOR CONTRIBUTIONS
PG coordinated the Silvolab Program in 2000, and co-supervised the hunting program of the Parc amazonien de Guyane from 2010 to 2012 with DD, CR-H, and RR. GL conducted most field work of this same program. FR and LG conducted hunting surveys in northern villages in surveys 1 and 2, respectively. CR-H participated to fieldwork and/or supervision of all programs, made the analysis and wrote the initial draft of the manuscript. All authors revised the manuscript text.
### FUNDING
Ministère de l'Environnement. French Ministry of Environment (Programme Ecoystemes Tropicaux, Mission pour la Création du Parc de la Guyane, DIREN Guyane, DEAL Guyane, Parc amazonien de Guyane, ONCFS), Labex CEBA, and Labex DRIIHM (OHM Oyapock, CNRS).
### ACKNOWLEDGMENTS
We are very grateful to all hunters that accepted to participate to the study, on a voluntary basis, and shared their hunting results, as well as Amerindian leaders Jacky Pawey, Thomas Palassissi, Raymond Lassouka, René Maïpouri, Laurent Pilauku, Albert Mifsud (†), Guy Barcarel, and Roger Labonté who allowed this study. Many thanks to all people that contributed to collect the data across the years, Pierre Alounawalé, Lucienne Taloekaïdoe, Henry Civette, Gérard Jean-Baptiste, Thierry Yawalou, Luc Lassouka, Yves Kouyouli, Jean-Michel Miso, René Maïpouri, Loïc Zidock, Laurent Pilauku, Jean-François Maillard, and Manfred Ulitzka. We are grateful to referees who helped enhancing the first draft of this work.
Atroari Indians in central Amazonia, Brazil. Biodivers. Conserv. 9, 579–596. doi: 10.1023/A:1008999201747
Park, Peru: landscape-scale management implications for Amazonia. J. Appl. Ecol. 46, 804–814. doi: 10.1111/j.1365-2664.2009.01661.x
**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Copyright © 2019 Richard-Hansen, Davy, Longin, Gaillard, Renoux, Grenand and Rinaldo. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
# Wildlife and Livelihoods in the Cardamom Mountains, Cambodia
Lauren Coad1,2 \*, Sotheary Lim3,4 and Lim Nuon<sup>5</sup>
*<sup>1</sup> Center for International Forestry Research, Bogor, Indonesia, <sup>2</sup> United Nations Environment-World Conservation Monitoring Centre (UNEP-WCMC), Cambridge, United Kingdom, <sup>3</sup> Fauna and Flora International, Phnom Penh, Cambodia, <sup>4</sup> Centre for Biodiversity Conservation, University of Phnom Penh, Phnom Penh, Cambodia, <sup>5</sup> Osoam Cardamom Community Center, Veal Veng, Cambodia*
This study provides a preliminary assessment of the use of wild meat and fish by rural communities in the Northern Cardamoms, Cambodia. We used a case-study approach in three villages, to identify key characteristics and drivers of wildlife use, with a view to informing the design of future larger-scale investigations of wild meat and fish use in the Cardamoms. Interviews with 41 households, conducted from August to October 2013, were used to investigate the main livelihood activities of each household, including hunting and fishing activity, key hunting techniques and hunted and consumed species. Group discussions with households and hunters were used to determine the relative importance of hunting and fishing as a livelihood activity and food source. We found that over 80% of interviewed households hunted, and similarly over 90% fished. Hunters employed a range of techniques, and caught at least 38 different mammal, bird and reptile species. However, our results suggest that arable farming is the backbone of livelihoods in these villages, providing the bulk of household incomes, and that most households are hunting to prevent crop-raiding, or opportunistically, rather than to supply the commercial trade. While households expressed a preference for wild meat, bought domestic meats and fish were eaten more frequently. A potentially lucrative commercial trade with high profits per animal exists but catches are unpredictable, and hunting is dangerous. However, as many species populations are already heavily depleted, even low hunting offtakes could have significant impacts on vulnerable species. Previous research suggests that commercial hunting which targets larger-bodied and high-value species for the international wildlife trade is mainly conducted by professional hunting groups, external to local communities. The importance of agricultural trade to local communities suggests that "wildlife-friendly farming" initiatives may help to both secure a fair and reliable price for village agricultural products, while promoting conservation of biodiversity in the Cardamoms mountains. However, due to the likely larger impacts of commercial hunting groups, declines in biodiversity are likely to continue without stricter enforcement of wildlife trade laws in Cambodia's towns and cities, and the reduction of demand for wildlife products in consumer countries.
Keywords: hunting, wildmeat, mammals, community, wildlife, trade, nutrition, household
### Edited by:
*Divya Karnad, Ashoka University, India*
#### Reviewed by:
*Edson Gandiwa, Chinhoyi University of Technology, Zimbabwe Jonah Henri Ratsimbazafy, Madagascar Primate Study and Research Group, Madagascar*
> \*Correspondence: *Lauren Coad lauren.coad@me.com*
#### Specialty section:
*This article was submitted to Conservation, a section of the journal Frontiers in Ecology and Evolution*
> Received: *23 January 2019* Accepted: *24 July 2019* Published: *27 August 2019*
#### Citation:
*Coad L, Lim S and Nuon L (2019) Wildlife and Livelihoods in the Cardamom Mountains, Cambodia. Front. Ecol. Evol. 7:296. doi: 10.3389/fevo.2019.00296*
## INTRODUCTION
The Cardamom mountains, in southwest Cambodia, represent one of the largest, most diverse, and least developed extents of lowland deciduous forest in mainland Southeast Asia (Daltry and Momberg, 2000), and are internationally recognized as a hotspot for biodiversity conservation (BirdLife International, 2019), supporting diverse and abundant populations of large bodied mammals, including leopards (Panthera pardus delacouri, Neofelis nebulosi), Asiatic black bear (Ursus thibetanus), gaur (Bos gaurus), and other large ungulates (Gray and Phan, 2011). In 2000, the first, and most recent biodiversity survey of the Cardamom Mountains revealed that they represent a disproportionately large amount of Cambodia's biodiversity; the Mountains cover about 6% of Cambodia's land area but support half of the country's known resident bird, reptile and amphibian species, and most of Cambodia's large mammal species (Daltry and Momberg, 2000). Historically, human population densities in the Cardamom mountains have been low, and until the early 1990s, the main inhabitants of the Cardamoms were the indigenous Mon Khmer Pear (also known as "Por" or "Khmer Dauem"), whose livelihoods, cultures and beliefs were, and still are intimately linked with the forest (Sarou, 2009).
However, events since the 1960's have had devastating impacts on the Cambodian people, their livelihoods and their wildlife. Initial civil conflicts (1968–1975), subsequent rule under the Khmer Rouge (KR; 1975–1979), and then conflicts between government forces and the remnant KR groups (1979–1997), claimed the lives of millions of Cambodians, and displaced millions of others, mainly into rural areas (Terry, 2002). Many areas of the mountains and fields surrounding the villages were heavily mined during the conflict, and mine clearance is still ongoing (Daltry and Momberg, 2000; pers. obs.). By 1991, an estimated 319,500–462,500 weapons were stockpiled nationally, with 136,000–200,000 soldiers and 250,000 militia trained in their use (Loucks et al., 2009). The Cardamom mountains were one of the last strongholds of the KR (Terry, 2002), and while the KR controlled the Cardamoms, thousands of Mon Khmer Pear were driven from their homes (Sarou, 2009). In the early 1980's, conflict, starvation and outbreaks of malaria forced many communities in the Northern Cardamoms to evacuate to the Thai border camps, only returning to their villages in the Cardamoms after the arrival of peace in 1999 (Terry, 2002). Armed militia were re-integrated into society after 1999, and many KR members settled in the Cardamom mountain villages.
In the 1990's there was a boom in the timber trade, and large tracts of forest were licensed to private timber and agricultural companies, resulting in rapid deforestation. Between 2001 and 2012 Cambodia lost over 14% of its forest cover; this deforestation rate was the fifth fastest in the world during this time period (Hansen et al., 2013). Lowland forests in eastern Cambodia have been cleared faster than in the less accessible western Cardamom Mountains, but as highvalue timber species become harder to find in the eastern forests, deforestation pressure is increasing in the southwest of Cambodia (Hansen et al., 2013; visualizations accessed 06/01/2019). Logging concessions and agricultural expansion have brought new road networks, providing increased access to once-remote forest; one of the most significant road expansions for the Cardamoms has been the upgrading of Route 5 between 2000 and 2005, which connects Phnom Penh with the Thai border, via Battambang. The availability of employment with logging and agricultural companies has attracted many lowland Khmer to settle in the Cardamoms, and now the vast majority of inhabitants of the Cardamom mountains originate from the lowland Khmer (Sarou, 2009).
All of these changes in the Cardamom mountains have driven an increase in the commercial wildlife trade. During the civil conflict in the 1970–90s, military training and the availability of firearms resulted in more traditional hunting methods, such as crossbows, being replaced with more efficient firearms (Drury, 2005). After the end of the conflict, many returning indigenous communities and ex-militias were reliant on hunting and collection of other NTFPs, due to the loss of their villages and livelihoods, and the unavailability of agricultural lands due to the danger of landmines (Drury, 2005). Growing affluence in China, and the influx of Vietnamese and Chinese, often with foreign timber companies, in the late 1990s then further increased the demand for wildlife products for traditional medicine. Road expansion made forest more accessible to commercial hunters, and reduced travel times to major markets (Drury, 2005). Following the conflict, limited rule of law, particularly concerning environmental management, meant that the chances of being caught or sanctioned were low. In addition to village community hunting, the potential profits to be gained from supplying the market for traditional markets drove the formation of commercial hunting gangs external to communities, often formed from ex-militia (Wutty and Simms, 2005). An influx of lowland peoples, following employment opportunities and settling agricultural lands, as well as an increase in the commercial trade in NTFPs, eroded the customs and traditions of the indigenous communities, although many retain their traditional beliefs, and many groups in the Cardamoms have lost their language through long periods of interaction with the more dominant lowland Khmer culture (Sarou, 2009).
Biodiversity impacts of the conflict are thought to have been significant, although no quantitative baseline data exist for comparison. However, interviews with hunters in NE Cambodia suggest that species abundance many have halved from 1950 to 2000 (Loucks et al., 2009). Several larger-bodied mammals have been nationally extirpated through over-hunting including the Javan Rhinoceros (Rhinoceros sondaicus), with no confirmed sightings in the Cardamoms or Cambodia since the 1980's (Daltry and Momberg, 2000) and the Indochinese tiger (Panthera tigris tigris), which was declared functionally extinct in Cambodia in 2016 (WWF, 2016). The Indochinese leopard (Panthera pardus delacouri) was found to have declined in abundance by 72% between 2009 and 2014 in the Srepok Wildlife Sanctuary, Eastern Cambodia, most probably due to widespread snare hunting, and is now under threat of national extirpation (Rostro-García et al., 2018). In 2000, of the 67 mammal species recorded in the Cardamom mountains, 26 were threatened or near-threatened according to the IUCN Red list, with the greatest threat coming from commercial hunting (Daltry and Momberg, 2000).
From the 1990s onwards, significant measures have been taken by the Cambodian government to conserve Cambodia's forests and wildlife. Cambodia's protected areas (PAs) were established under the 1993 Royal Decree on the Protection of Natural Areas, and recent estimates suggest that over 40% of Cambodia is covered by some sort of PA. PAs cover a large proportion of the Northern Cardamom Mountain forests; Phnom Samkos Wildlife Sanctuary (3,307 km<sup>2</sup> ), the Central Cardamom Mountains Protected Forest (4,010 km<sup>2</sup> ), and Mount Aural Wildlife Sanctuary (2,544 km<sup>2</sup> ) (UNEP-WCMC IUCN, 2018). Cambodia's principal wildlife legislation, the Law on Forestry (Kingdom of Cambodia, 2003) was enacted in 2003 and is overseen by the Ministry of Agriculture, Forestry and Fisheries. Under this legislation, wildlife is considered State property. Hunting which uses "dangerous means," is conducted during the closed season (which has not yet been defined) and of rare and endangered species (as categorized by separate Ministerial Declarations), is illegal. Local communities are allowed to hunt "common" wildlife using traditional methods, for "customary subsistence use" (this important term has never been clearly defined), although "common" wildlife may not be transported and traded in "an amount exceeding that necessary for customary use." This does not apply to the core-zones of PAs, under the 2008 Law on Natural Protected Areas (Kingdom of Cambodia, 2008). The uncertainties surrounding the definitions of "dangerous means," the closed season and "customary use" effectively makes most wildlife trade potentially illegal.
Wildlife trade volumes in the Cardamoms, while still significant, are thought to have reduced from a 1990s peak, partly due to reduced wildlife populations and partly due to the removal of many of the guns from general circulation by the local authorities, beginning in 2001 (Oul and Cheam, 2005). However, the use of non-selective and cheap wire and nylon snares have increased, and in 2013 over 13,000 snares were removed from the Southern Cardamom National Park by law enforcement patrols; this had increased to almost 28,000 by 2015 (Gray et al., 2018). Informant networks employed in 2005 identified both commercial village hunters and commercial external hunting gangs, and the Central Cardamoms as one of the "hotspots" for wildlife trade. Most of the wildlife traded in the Cardamoms is now likely to leave Cambodia for the international market, traveling first to traders within the Cardamoms, then to Phnom Penh along Routes 4 and 5, and finally exported to China, Vietnam and Thailand (Wutty and Simms, 2005).
While forest area and wildlife populations have reduced, wildlife still thought to be important culturally and economically to both the lowland Khmer and indigenous Mon Khmer Pear living in the Cardamoms (Daltry and Momberg, 2000; Fox, 2006; Sarou, 2009). Socio-economic surveys of over 40 villages within the Phnom Samkos Protected Area, between 2004 and 2006, found that arable crops provided the backbone of local livelihoods, with other activities including livestock rearing, market gardening, fishing, and hunting (Fox, 2006). Fishing was a daily activity for most households, providing an important source of protein. Direct questions on hunting frequency were not included in the survey due to the illegality of hunting, but researchers observed that crop raiding was common, and farmers would catch wild pig, deer and porcupine around their arable fields. Researchers also observed that Sunda pangolin (Manis javanica), Malay sun bear (Helarctos malayanus), tiger, gaur, and banteng (Bos javanicus) were sold by hunters, who were sometimes contracted by town traders. A later evaluation of livelihoods in the Central Cardamoms Protected Forest again stressed the importance of agriculture for local communities (Sarou, 2009). While 80% of households reported fishing only 15% reported hunting; however, the author noted that ranger presence in the villages was high and that when asked about hunting "in the past," over 50% of households reported hunting, which may provide a truer reflection of current hunting levels.
## STUDY AIMS
While previous studies suggest that wildlife is still part of local livelihoods in the Cardamoms, there is still limited information on the role wildlife plays a part in overall community livelihood strategies. To gain a preliminary understanding of contemporary use of wildlife by local communities in the Cardamom mountains, we used a case-study approach in three villages in the northern Cardamoms. We aimed to describe and investigate:
To this end, we employed semi-structured interviews with households and hunters, as well as group PRA techniques, situating wildlife use within the wider livelihood strategies of local communities. We use our results to build preliminary hypotheses as to the drivers of wildlife use by communities in the Cardamom mountains. We then discuss how these results and hypotheses, could be used to design further in-depth studies of wildlife use, and might inform wildlife management policies in the region.
### METHODS
### Study Area
We conducted our research in three rural villages in the Northern Cardamom Mountains, in the Battambang and Pursat provinces (**Supplementary Material S1**). We conducted an initial scoping trip in August 2013, visiting three communes (an administrative unit comprising several villages). We selected one village in each commune to work in based on their proximity to the forest and how comfortable they seemed with our presence and preliminary questions; our aim was not to create a systematic sample of all villages in the Cardamoms, but rather to form preliminary hypotheses on the use of wildlife in villages bordering the Cardamom mountains forest.
The three study villages are located directly on the northern border of the Phnom Samkos Wildlife Sanctuary, at between 180 and 400 m above sea level, at the base of the Cardamom mountains. To the north of the villages, most of the landscape is below 200 m, and habitat has been converted to farmland and rice fields. To the south, the mountains within the Phnom Samkos Wildlife Sanctuary are covered by relatively intact tropical moist broadleaf forests (the Cardamom mountains Rain Forest Ecoregion; WWF, 2019), rising to 1,700 m above sea level. Climatic conditions in the Ecoregion are relatively stable, with a rainy season from May to October, mean annual rainfall between 2,000 and 4,000 mm, and average temperatures between 29 and 32◦C. The ecoregion is thought to be home to over 100 mammals species and over 450 bird species (WWF, 2019).
All three villages had <100 households, were majority Khmer ethnic group (>95% of the population) and were evacuated in the 1990s during the Cambodian war. Villages 1 and 2 were ex-Khmer Rouge (KR) villages, and still have significant densities of land mines in the surrounding fields and forests. As defined by the Cambodian Government's National Policy Framework for Poverty Reduction, 40, 100, and 50% of households in Village 1, 2, and 3, respectively are in poverty, and eligible for state assistance (Ministry of Planning, Kingdom of Cambodia, 2012). **Supplementary Material S2** provides basic characteristics of each village, and the study timetable. The exact location and name of each village has been kept anonymous.
### Asking for Local Permission to Conduct the Study
During the initial scoping trip we presented the project to commune chiefs, asking for their permission and support to conduct the study, and providing them with our letter of introduction from the University of Phnom Penh (UPP). On arrival at each of the potential study villages we met with the village chief and gave him our letter of introduction. We discussed the interviews and PRA activities that we wished to conduct and asked his permission to carry out research interviews in the village. All villages chiefs that we asked gave their permission. We then conducted surveys in three villages (1 per commune) on three separate field trips, over 6–7 days, in September and October 2013.
### Household Interviews
We used semi-structured household interviews to familiarize ourselves with the main livelihood activities of households in each of the study villages. Interviews took ∼2 h and were conducted by two Cambodian researchers (LN and SL), with LC present. Households were selected using a systematic random sample, following Starkey (2004). We first counted the number of households in the village, and then divided this number by the aimed-for sample size of 15 households (i.e., if there were 60 household in the village, 60/15 = 4. We then visited every nth household (i.e., in our example every fourth household), using this calculation, starting at one end of the village and counting along the houses. Where household members were working in their fields, we did not attempt to interview them due to the number of land mines still present in the fields. In this case, where no-one was available in the household to be interviewed, we would then move onto the next adjacent household to the selected household.
On arrival at each household, we introduced ourselves and the aim of the project, and we asked if they would be willing to be interviewed. We emphasized that participation was completely optional, and that the identity of all villages and interviewees would stay anonymous. Everyone that we asked agreed to be interviewed. Names of interviewees or households were not recorded at any point. Details were recorded in small notebooks rather than on questionnaire sheets, to reduce the formality of the interview. We completed 14 interviews in Villages 1 and 2, and 13 interviews in Village 3, representing 19, 38, and 13% of all households in each village, respectively. In each village, the number of interviews was limited by the time available. As a token of thanks, each interviewee was given a krama (a small sarong) at the end of the interview. Interview questions are provided in **Supplementary Material S3**.
Interviewees were asked to describe their main livelihood activities, including:
The mainly illegal timber trade is a highly contentious subject in the study area, and we quickly removed the question on the use of timber from our list of questions, after observing the reaction of interviewees. Direct observation of livelihood activities was not possible due to the continued presence of land mines in the surrounding fields and forests, and responses provide us simply with the interviewee's estimate of frequencies and amounts.
Additionally, where interviewees were happy to discuss hunting, we asked how frequently the household consumed different animal species (i.e., whether a species was consumed daily, fortnightly, monthly, quarterly, or yearly). To prompt discussion, we showed the interviewee a set of 41 cards, depicting different common and rare species thought to be present in the Cardamom mountains (Daltry and Momberg, 2000; **Supplementary Material S4**). We included a card for each of the large mammal species recorded as present by Daltry and Momberg (2000). We did not include all bird and reptile species to species-level, instead choosing key species known to be hunted, and including other cards to represent key groups (i.e., turtle sp., land lizard sp. bat sp.; see **Supplementary Material S4**). We included one "wildcard" species—Javan rhinoceros—known to no longer occur in the area.
Where interviewees seemed comfortable to talk further about hunting we continued to discuss broader questions on hunting, including methods, preferred species and which species had been hunted by the household in the previous year. To prevent any potential unease in discussing the illegal wildlife trade, we did not attempt to estimate the amount of household income made from wildlife sales, and only asked follow-up questions on wildlife trade when information on the trade had been given unprompted and interviewees seemed comfortable with the discussion. We were not able to identify fish using binomial nomenclature, as we did not have a fisheries expert as part of the team, but wrote down the name of fished species in Khmer. Our analyses therefore do not attempt to break down results for fish by species.
### Hunter Interviews
We conducted a group interview with 3–4 hunters at the end of each village visit (to allow time for trust to develop). Hunters were identified through household interviews (i.e., where a member of the household that we interviewed engaged in hunting regularly), through discussions with village elders, and also through general conversations during the day between LN and village men. The opportunistic sampling strategy may mean that hunter interviews were not representative of village hunters as a whole, and our results should be taken as preliminary, with the aim of helping to develop more in-depth studies in the future.
Interviews were kept anonymous and informal in tone, conducted while sharing a meal or drink. We showed the interviewee the set of 41 species cards (predominantly mammals), and asked about the hunting techniques, frequency of capture, use and village prices for each species. In addition, we asked about more generally about the drivers of hunting in the village, the characteristics of local hunters, the level of hunting enforcement and changes in the availability of wildlife in the area.
### Participatory Rural Appraisal (PRA)
### Livelihood Analysis
We used the IUCN's Poverty Forestry Livelihood Analysis Toolkit (PFLA) Tool 6 (PROFOR, 2008) to investigate the relative importance of different cash income and non-cash income activities for rural households. Cash income refers to income brought into the household in the form of money (such as agricultural sales, income from employment, sales of NTFPs etc.). Non-cash income refers to products brought into the household, which are then used directly by the household (i.e., agricultural products which are eaten rather than sold, NTFPs used for food, construction, or medicine, etc.).
Following the completion of 14 household interviews, toward the end of our visit in each village, we invited the household to send an adult representative of the household to attend one of two half-day PRA sessions (seven interviewees in each group), one of which was conducted in the morning and the other in the afternoon of the same day. Together with participants, we wrote a list of the different livelihood activities in the village, in Khmer, onto A0 paper. We then asked each participant to distribute 50 "counters" (we used 50 pieces of corn) between the different livelihood activities so that the number of counters given to each activity illustrated the amount of income from that livelihood activity. This was done twice—once for cash-income sources (illustrating the value of the items produced/procured which were then sold) and one for non-cash income sources (the value of produced/procured items which were then used by the household, not sold). As described in the PFLA (PROFOR, 2008), this was conducted in front of the participants from other households, which may have influenced the results of each household, but also resulted in illuminating group discussions on the use and importance of different products.
During the exercise, we discussed each livelihood activity with the group. We only discussed wildlife and timber sales if the participants brought it up unprompted, because both activities are illegal. At the end of each meeting we provided participants with lunch or an evening meal and gave them each a krama to thank them for their participation.
In Villages 1 and 2, the consumption of insects did not come up in household interviews, and insects were therefore not included in any of the PRA exercises. In Village 3, grasshoppers and beetles were reported and observed to be consumed, and in we therefore added insects as a category for the PRA exercise.
### Consumption of Meat and Fish
We used a similar method to the PFLA toolkit to look at the importance of different fish and animal protein sources in the household diet. We asked participants from each household to distribute 50 pieces of corn to illustrate the importance of beef, pork, chicken, market-bought fish (from outside the village), caught fish (including crabs, shrimp and small fry), and wild meat, for feeding their household, over the course of a year. For Village 3, an additional category of "insects" was added after household interviews highlighted that insects were an important part of the diet.
### Data Analyses
We have used the results of these interviews and PRA exercises to build a first look at the livelihood activities of, and the use of wildlife by, local communities in the Cardamom mountains. Specifically we analyse and present:
In some specific cases, we have used Pearson's product moment correlations to further explore correlations between livelihood activities.
Qualitative information on household livelihood activities collected during household and hunter interviews is presented with these quantitative results, providing contextual details on individual activities, and potential explanations and hypotheses for the quantitative results.
Data were analyzed using Microsoft Excel and R computing language (R Core Team, 2016).
### RESULTS
### Hunting Frequency, Species, and Methods
Most households (83%) reported having hunted during the previous year. At least 38 species were reported to have been caught in the village territories (**Figure 1**). The five most frequently reported hunted species were monitor lizard (Varanus sp.), SE Asian porcupine (Hystrix brachyura), wild pig (Sus scrofus), Indian muntjac (Muntiacus muntjac), and turtle species (Bataguridae and Trinychidae sp) (see **Figure 1** and **Supplementary Material S4** for full species list and scientific names). Households had not observed Javan rhinoceros, tiger or wild dog (Cuon alpinus) in the forest areas surrounding the village, agreeing with the findings of Daltry and Momberg (2000). A few hunters said that tigers had been seen over 2 years ago in the more remote, mountainous sections of the northern Cardamoms, and reported catching wild dog in the past 5 years. Hunters and households reported that they did not catch clouded leopard (Neofelis nebulosi), elephant (Elephas maximus), silvered langur (Trachithecus cristatus), or Malay sun bear (Helarctos malayanus), although they were thought to exist in the surrounding forest.
The main hunting methods were snares, slingshots, dogs and homemade "pump guns" (**Supplementary Materials S5, S6**). Snares (wire cable or nylon) were used by 76% of households, generally for crop protection (other crop-raiding deterrents used are outlined in **Supplementary Material S7**). If households did not set traps it was often because there was no adult man within the household, rather than a lack of need. The main crop-raiding species were porcupine, wild pig, Indian muntjac, squirrels species, and civet species. Large herds of wild pigs, or group of macaques (Macaca nemestrina, M. fascicularis), coming through a plantation could result in high losses in agricultural yields. The two most frequently used snare types were foot snares, to catch larger animals, such as wild pig, Indian muntjac, and small cats (Prionailurus bengalensis, P. viverrinus) (although they are indiscriminate and will catch a wide range of species; **Supplementary Material S6**), and neck snares, to catch smaller animals, such as snakes and rats. In addition to crop protection, households and hunter groups reported that a few commercial hunters in each village (men who used hunting as a primary livelihood activity, and source of income) would set snares in the forest. One hunter focus group reported that these hunters would commonly set around 50–90 snares in the forest at one time and check these snares every 3 days.
Over half (56%) of households used slingshots for short hunting trips in the evening or night, when coming home from the fields, or after setting fishing rods in the river. Although the animals procured with slingshots tend to be small and of low value [birds, Slow loris (Nycticebus coucang), snakes, and civets] slingshot hunting reliably catches something for the cooking pot and is easy to fit around other livelihood activities. A similar proportion (54%) of households hunted with dogs. Although dogs can help hunters to catch a range of species (**Supplementary Material S6**), dogs are primarily trained to catch Sunda pangolins (Manis javanicus), which are traded for their scales, and attract high prices on the black market. Trained dogs track and point at pangolins, which roll into a ball when attacked and are easily picked up by the hunter.
Twenty nine percentage of households used homemade "pump guns." Interviewees explained that between 10 and 20 years ago, shotguns were readily available, a legacy from the civil war in the 1990's. Shotguns have now mainly been confiscated or hidden, but homemade guns are cheap and easy to make, and are used frequently for night hunting. Pump guns operate using loose shot rather than cartridges or bullets and use a pump-action fashioned from a bicycle pump. Manufacture is cheap and parts easy to come by. Although pump guns are effective at killing medium-sized animals (such as muntjac), hunters reported than they were not capable of killing larger mammals, such as gaur, wild pig, sambar (Cervus unicolor), and bear species.
In addition to these main hunting techniques, animals are also harvested when field are cleared and burned (generally monitor lizards, turtles and snakes), poison is used to catch small birds and tree shrew species, and leopard cats (Prionailurus bengalensis) are sometimes hunted with a crossbow.
### Fishing Activity and Methods
Almost all households (95%) fished. Crab and shrimp and small fry were caught by women, using a scoop net. While they only provide a small amount of food, crab and shrimp can be collected on the way home from working in the fields and were added to forest and farm vegetables to create a low-cost meal, the only purchased good being rice (which is not often grown in these upland villages). In larger rivers, men caught river fish and eels with a fishing line, rod or net. These fish are still quite small (5– 15 cm in length), and men reported catching between 0.5 and 1 kg per trip, or enough for 1 or 2 meals. Several men reported that rod or line fishing and hunting activity tended to coincide; men would set their lines in the evening, and then go night hunting (or "lamping") with a torch and a slingshot or pump gun. On their return from hunting they would then check and retrieve their lines. **Supplementary Material S8** provides further details on fish types and fishing techniques.
Electric shock fishing was practiced in all three villages. A car battery is used to provide a high voltage shock into the river, killing or stunning the fish (and turtles), which are then easily collected by hand. Electric shock fishing is illegal, and most respondents talked about "their neighbor" using this technique. It is therefore hard to gauge it popularity. However, as our stay in the villages progressed, and villagers became more trusting, several interviewees suggested that most households used electric shock fishing, and a few respondents suggested that over-use of electric shock fishing was one reason for the current low fishing returns.
Fishing activity varied with the season; in Village 1 which had a small river, households reported fishing more often during the wet season, when water levels were high enough for fish (rather than small fry, shrimp and crab) to be found. In Villages 2 and 3, some households reported reducing their fishing activity during the wet season months, when water levels were too high, and nets would get torn.
In the first month of the wet season (April/May) over 80% of households caught frogs 2–3 times a week in their fields and nearby ponds, by hand. Households reported catching up to 3 kg of frogs per trip. Frog catching is highly seasonal, and after the first month households reported that they did not catch frogs because the water in the fields and ponds became too high for them to easily catch them, and that further into the wet season frogs would often be full of worms and therefore inedible.
### Sources of Cash Income, and the Relative Importance of Wild Meat and Fish
During the PRA exercise to identify the relative importance of different household activities for household cash incomes, households allocated 62% (SE ± 2.8) of counters to arable farming on average. Salaried work or commerce was allocated 15% (SE + 1.9), forest products 12% (SE ± 1.7) [including 1.7% (SE ± 0.5) for wild meat], livestock 9% (SE ± 1.4), and fishing (including shrimp, crabs, and turtle) 1.8% (SE ± 0.7) of counters (**Figure 2**).
### Arable Farming
All three villages relied heavily on arable farming for cash incomes (mainly corn, mung bean, sesame, and some soy bean; **Figure 2**). Households sold almost all of their crop harvest, except for rice, which was grown for household consumption only. Crops were sold twice a year to town traders, with trade generally organized by the village as a collective. Households therefore had good knowledge of the quantity and price of each crop sold, allowing yearly gross agricultural incomes to be estimated from our one-off household interviews. However, without conducting a longer-term study of household incomes and outgoings, estimates of net profits will remain imprecise. With this in mind, average gross income/year/household were \$2,000/year (SE ± 260), ranging from \$6,500/year for the largest farm to only \$80/year for the smallest. Average gross income was highest for Village 3 (\$2,940/year/household, SE ± 600), followed by Village 1 (\$2,160/year/household, SE ± 260), and lowest for Village 2 (\$950/year/household, SE ± 260). Participants of the PRA exercises agreed that in an average year (one in which a normal harvest was achieved) ∼1/3 of farm incomes might be spent on herbicides, pesticide and additional labor. Net arable profits might therefore be in the region of \$1,900, \$1,400 and \$600/year/household for Villages 3, 1, and 2, respectively. Smaller farmers reported making overall losses, either due to existing debt, land rents or crop failures.
Village 2 had relatively low incomes from arable crops and households allocated the lowest number of counters to arable incomes in the PRA exercise [48% (±2.9) compared to 73% (SE ± 3.6) and 69% (SE ± 4.9) for Villages 1 and 3, respectively; **Supplementary Material S9**]. Households in
Village 2 also owned/used fewer hectares of arable fields than in the other two villages [an average of 3.8 ha (SE ± 0.76) for Village 1, 3.1 (SE ± 0.38) for Village 3, and only 1.5 (SE ± 0.35) for Village 2]. At the time of our study, Village 2 had the government land mine clearance organization camped in the village. Many of the fields were not yet cleared of land mines, were therefore not being used by their owners. This may explain the lower profits from agriculture in Village 2. Several interviewees in Village 2 highlighted the issue of land tenure rights for rural households. Interviewees reported that while they were obviously keen to seen land mines cleared from their fields, they were also concerned that after clearance, the land would be leased to multinational agricultural companies and that, paradoxically, the clearance of land mines would therefore lead to the loss of their lands.
Every household in the three villages had a market garden (a small area behind the house), containing vegetables, such as eggplant, chili, herbs, gourd, cabbage, pumpkin, cassava, and beans, and fruit, such as oranges, mango, banana, papaya, coconut, milk fruit, lemon, and jackfruit. Households generally did not sell much produce for their market gardens, however there was a strong barter economy in the villages and households reported that they would often give fruit and vegetables to their neighbors for free and know that they would get the same in return. In Village 2, market gardening scored as highly as many agricultural products in terms of cash incomes (11% of beans allocated; **Supplementary Material S9**), and this may again be due to the smaller size of agricultural fields in Village 2, leading to a diversification of livelihood activities.
### Employment, Commerce, and Labor
There were very few employment opportunities available in the three villages. Aside from several government jobs (village chief, deputies and one teacher, paid \$7–15 a month), nine households contained members with either army pensions or disability allowances (often due to war wounds, or landmine injuries). Few households engaged in commerce (two households baked cakes, making between \$1.25 and \$4 a day, one household owned a truck, and had set up a corn trading business between the village and corn traders in the nearby town, earning a reported \$500 a year. One household ran a village shop, and another had just set up a restaurant). Most employment came in the form of households with small farms selling their labor to households with larger farms in the planting and harvesting season. Of the 41 households interviewed, 28 sold their labor at some point during the year. One day's labor cost ∼\$3, and households reported making between \$10 and \$50 dollars from selling labor over the year. Although this does not represent a large income when compared to the money that can be made from arable farming, for poorer households with little land, incomes from selling labor can still be important. In the PRA exercise the land-poor Village 2, where field sizes are limited by the number of remaining land mines, incomes from selling labor scored more highly [17.2% (SE ± 3.1)] than Village 1 or 3 [5.1% (SE ± 1.1) and 6.3% (SE ± 1), **Supplementary Material S9**]. The scores that an individual household gave to "selling labor" as a cash-income were also negatively correlated with the gross agricultural incomes of
the household (Pearson's product moment correlation, n = 32, p = 0.019, R <sup>2</sup> = 0.14).
### Livestock
Household generally kept a few animals, with each household owning on average 10 chickens, and 1–2 cows or pigs. Livestock in the study villages is not often kept for household consumption, and instead is used in the same way as a savings account, providing source of emergency funds, or to pay for one-off purchases. 16 of the 41 households interviewed sold livestock over the last year, and household received, on average, \$430 (SE ± 134) from livestock sales, not accounting for rearing costs. Cows and pigs were the main species sold, and of the 10 households that sold cows in the previous year, 5 of these did so to buy a mini tractor ("koyun"), or to invest in building a new house.
### NTFPs
The two main sources of income from forest products, aside from wild meat and fish, were mushrooms and medicinal plants. Three main species of medical plants were sold to traders from town, known locally as "Tacao," "Krawine," and "Moi Roi Kun" (or "100 uses"). Medicinal plants were collected by 30 of the 41 surveyed households and sold to town traders by 21 of these households, with profits of ∼US \$100/year. The main collecting season was June–September, and families might spend a few weeks collecting plants in June and July. "Pok" Mushrooms were collected by 36 of the 41 surveyed households and sold by 11 of these households. The Pok mushroom season runs for a few months (June–August) during the wet season and is sold to town traders. Pok mushrooms sell for \$2.5–3/kg, and households reported making between 10 and 50 dollars each year.
In Eastern Cambodia, the most valuable NTFP is often resin, which is tapped from dipterocarp trees. Although resin trees exist in the study villages, they are of a different species and the resin produced is much less valuable (US \$0.5 per kilo). It is used to caulk boats, and was only collected by five of the 41 households, generally by younger boys. Unlike in Eastern Cambodia, there are no resin collectives in the village.
In Village 3, insects were included as a category in the PRA exercise. However, no counters were allocated to insects for cash incomes.
### Fishing and Hunting
While fishing and hunting was conducted by most households, they were only reported as primary income-generating activities by a few households. Households allocated a mean of 1.8% (SE ± 0.41) counters to fishing incomes and 1.7% (SE ± 0.54) of counters to wild meat (**Figure 2**), and 16 and 30% of households sold fish or wildmeat, respectively. Only 1 of the interviewed households allocated more than 10% of counters to wild meat, and only 2 households for fishing.
Interviewees explained that the amount of fish caught in a night was often only enough for 1 or 2 meals, leaving no excess to sell. Hunter interviews suggested that only a few men in each village were "commercial hunters" engaging in as a primary livelihood activity and source of income. These men would hunt almost every day, penetrate further into the forested hills around the village, and were more likely to catch larger-bodied and higher-value species. Interviewees gave four explanations for the low number of commercial hunters in each village.
Low PRA scores for hunting may also reflect an unwillingness to talk about hunting, rather than true low sales. This was probably partially true in Village 1 (where households allocated 0.18% (SE ± 0.18) of counters to hunting incomes on average, **Supplementary Material S9**), which was located close to a protected area ranger station and had been visited by conservation NGOs in the past. However, hunters in Villages 2 and 3 [which allocated 2.6% (SE ± 1.07) and 2.2% (SE ± 0.96) of counters to hunting, respectively], while being aware of the wildlife law, were quite happy to talk about catching protected species. Wild meat was sold in front of us, and we were offered wild meat (wild pig, porcupine, slow loris, and monitor lizard eggs) to eat.
Hunters reported that wild meat for consumption (rather than medicinal use) was sold almost exclusively within the village, and that demand for wild meat was high because of its perceived health benefits compared with meat from the market; there is therefore no need to sell to traders. In Village 1, the trade in meat seemed to be kept hidden; hunters reported that they would sell meat at their house to the neighbors that they trusted, as there was a spy in the village (whose identity was well-known to everyone), and he might tell the authorities. In Villages 2 and 3 trade in meat was out in the open, and women would come around with meat (generally wild pig) for sale while we were conducting interviews.
Species sold to traders outside the village were generally sold for their perceived medicinal properties (**Table 1**), rather than as a source of meat. Many species are used as a general health tonic, thought to cure a range of ailments. For example, one preferred species for medicinal use was the slow loris. Households said that slow loris were easy to find and could be used to treat "101 medical ailments" ("moi roy kun"), including joint pain. One of the households that we interviewed demonstrated its use for us. The slow loris had been smoked (charred) over a fire to preserve the carcass, and small amount of the charred flesh was then grated into rice wine and drunk as a tonic. Bones and antlers of animals were also used medically, and serow bone, porcupine jawbone, and sambar antler could all be grated into, or rubbed into, rice wine/rice water to create a health tonic. In addition to medical uses, deer antlers were traded for decoration, and we saw both sambar and Indian muntjac horns used in this way in small towns close to the study villages. Animal parts could also be used as good luck totems (for example, loris tattoo paint, pangolin scales, and bear claws).
Hunters universally identified the Sunda pangolin as the species that they most hope to catch, due to its high market value. Two pangolins were caught while we were in the study villages, both weighing 1.5 kg (with a value of ∼\$220 each), and were sold immediately to traders in Battambang, who were contacted by mobile phone. Hunters with a good pangolin dog might catch 1– 2 pangolins per month (2–3 if exceptional, but in some years, as few as 1–2 per year might be caught), and well-trained dogs can sell for \$1,000–2,000. In each village, only a few (1 or 2) hunters possessed trained pangolin-hunting dogs. Although animals like serow and black bear can fetch high prices (**Table 1**), both species were difficult and dangerous to catch. Only 2 of 31 households reported having caught a serow in the previous year; none of the interviewed households reported having caught a black bear.
Hunters in Village 2 reported that if they caught a high-value species, such as Sunda pangolin, serow, gaur, or bear, they would immediately hide the animal bones/skin/horns somewhere safe. They would then call/text a trader in Battambang, who would come and collect the animal. Less frequently a hunter might hold onto the wildlife items, such as bone, horn and skin) until his next trip to Battambang. In Village 1, households reported that there used to be a Battambang trader who would make frequent trips to the village to buy animals for traditional medicine trade, but the chance of being caught and fined by the rangers stopped him from making the trip, and now hunters transport the animals themselves to Battanbang. Hunters from Village 3 also reported that commercial hunting groups would come from Pursat and Battambang to hunt in the forest around the village.
Hunters were aware that the trade in wildlife was illegal and were taking precautions to keep the trade secret from the local authorities, although they also seemed happy to talk to us about the trade and their activity. Hunters only reported problems with buyer demand in the case of snakes (boa) and monkeys (pet trade); all other trades that we discussed with the hunters reported no problems with finding a buyer.
### Source of Non-cash Income, and the Relative Importance of Wild Meat and Fish
During the PRA exercise to identify the relative importance of different household activities for household non-cash incomes, households allocated 41% (SE ± 2.5) of counters to arable farming on average, of which 13% (SE ± 1.2) were for market garden products. Forest products were allocated 37% (SE ± 2.1) [including 4% (±0.6) for wild meat], fishing 16% (SE ± 1.4), and livestock 7% (SE + 0.8) of counters. While arable crops were the main source of cash-income for households, non-cash incomes were more diversely spread between different products from both farms and forest (**Figure 3**), with no one product represented by more than 13% of the counters on average.
### Arable Farming
All surveyed households in the villages, as is common in Cambodia, used rice as their staple carbohydrate and a main source of calories. However, in the Cardamom mountains, the hilly terrain does not provide a good growing environment for rice. In the three study villages, only 40% of households grew their own rice, and kept it for personal consumption. No households grew enough rice to support the family for the whole year (only one household could grow enough rice to feed the family for more than 6 months). Interviewees suggested that households would use the bulk of their arable incomes to buy rice. During this study, we were unable to determine what percentage of arable incomes were used to buy rice, but we suspect that households with low arable yields may have been in "rice debt," unable to afford the amount of rice needed to support their families. The level of "rice debt" in the cardamoms warrants further investigation. Market gardens also gained a high score for non-cash incomes, in comparison with their use as a source of cash incomes, suggesting that these gardens are an important source of food for local people.
### Livestock
Households ranked chicken as the only real non-cash input in terms of livestock, with pigs and cows reared for their value (and as a form of household savings) rather than household consumption.
### Hunting and Fishing
Both fishing and hunting were allocated a larger proportion of counters as a source of non-cash incomes (i.e., for food and medicine) than as a source of cash incomes. Wild meat is seasonally available to most households, with wild pig, muntjack, and civet crop raiding during the harvest seasons. As with produce from market gardens, wild meat is bartered between households; a household fortunate enough to catch a wild pig in their fields will not be able to eat or store the whole pig for personal use, and so instead will share with his neighbors, who will then return the favor when they catch an animal. As with cash incomes, non-cash income from wild meat was not correlated with agricultural activity or land ownership.
### Other NTFPS
In comparison with cash-income activities (where only pok mushrooms and tacao medical plants are sold in any quantity) households in all three villages use a range of forest products. Most households in used firewood for cooking—which explains
#### TABLE 1 | Uses and prices of wild meat species (from hunter and household interviews).
*m, meat; tm, traditional medicine, d, decorative. Prices are those given by hunters and households during the study in 2013.*
*<sup>a</sup>While the organ used for medicine was described as the "gallbladder," and was reported to be found in the front quarters (where the liver, lungs and heart are found), Muntjac do not have gallbladder, and so this must be a different organ.*
*<sup>b</sup>Tigers were not seen or caught during the study period, and were declared functionally extinct in Cambodia in 2016.*
its relatively high PRA ranking—as well as forest fruit and vegetables, mushroom, and medicinal plants. In Village 3, where insects were included as a PRA category, insects were allocated 4.9% (SE ± 1.5) of counters on average (**Supplementary Material S10**).
### Wild Meat and Fish as a Component of the Household Diet
Both household interviews (where we asked about the frequency of consumption of different meat proteins in terms of whether it was eaten daily/weekly etc.)
and PRA exercises (where we asked households to distribute counters to illustrate the importance of different meat proteins to the household diet) highlighted the importance of fish in the daily diet. In all three villages households allocated, on average, over 47% (SE ± 2.0) of counters to fish, and most households ate fish daily or weekly (**Figures 4**, **5**).
In comparison, wild meat was ranked second to last as a source of meat protein in the PRA exercise (**Figure 4**), and species, such as wild pig, muntjac, and porcupine, while eaten at some point in the year by most households, were generally eaten on a monthly or quarterly basis (**Figure 5**). Wild meat consumption is highly seasonal, occurring only for a month or so before harvest and we visited villages during the wet season harvest, when crops were ready to harvest. At this time, crop-raiding by these three species was frequent, and wild pig meat was observed hanging to dry outside village houses. Several households during interview expressed a preference for wild meat over meat bought from local markets, describing wild meat as being healthier, and bought meat as containing "chemicals." Several households said that they preferred wild meat because they could not tell where the bought meat had come from, or what had been done to it. In Village 3, where insects were included in the PRA exercises, we found that, while scoring lower [9% (SE ± 1.4)] than all other forms apart from beef, insects probably still represent an important component of freely available protein, which can be gathered from around the fields and village.
Both the PRA exercise and household interviews highlight the frequency of consumption of bought meat protein. Bought meat protein represented, on average, approximately half of the allocated counters in total in the PRA exercise. Pork, beef, and frozen fish were all bought from traders traveling from Battambang by motorbike each week, and the prices for different proteins are shown in **Table 2**. The cheapest bought protein was frozen fish (often sea fish), which interviewees suggested was imported from Vietnam and Thailand. All three villages allocated a similar proportion (23–27%) of counters to frozen fish in the PRA exercise (see **Supplementary Material S11** for score by village). Although frozen fish was consumed more than pork, several households reported a preference for pork, when given the choice. One woman, when asked why she ate frozen fish instead of pork said that "when you are poor, frozen fish tastes excellent. But when you are rich, pork tastes better." Very few households consumed beef, which was reported as being too expensive, not often provided by traders coming into the village, and bad for the health. Although the price of town chicken is less than pork, all but a few households ate chicken from their own stock rather than buying from traders. This may partly be due to availability, as traders brought pork every week, but did not seem to be trading chicken. The lack of trade in chicken may be because of the long journey time to the villages from Battambang, in which time chicken might spoil.
Village 1 allocated a higher proportion of counters to bought meat and fish [57% (SE ± 5.1)] than Village 2 [48% (SE ± 3.1)]
and Village 3 [39% (SE ± 5.6)]. This may be explained by market and river access. Village 1 was situated on a small stream, where large fish could not easily be found, and fishing was focused on shrimp, crab and small fry. In comparison Villages 2 and 3 were situated next to sizeable rivers, where catfish and trout could be fished. In addition, Villages 1 and 2 had road access to Battambang, whereas Village 3 was more remote, and during the wet season (when the study was conducted), access to the village, even by motorbike, can be difficult, and interviewees reported that trader visits to the village during months with heavy rain reduced from weekly visits to monthly visits.
### DISCUSSION
### What Are the Key Characteristics and Drivers of Current Wildlife Use in the Study Villages?
Our exploratory surveys in three Cardamom mountain villages suggest that most households hunt, and that most mammal species in the Cardamom mountains are harvested. At the same time, they also suggest that hunting incomes represent a low proportion of household incomes, and a low proportion of the household meat consumption. Household incomes are
*Prices were obtained in 2013, from local shops in the villages.*
predominantly agricultural, and both fishing and hunting only provide small additional incomes for most households. Owncaught fish, and even bought fish, pork, and chicken, are reported to be consumed much more than own-caught wildmeat.
The relative importance of fish over wildmeat in local diets has been observed in villages in Southern Cambodia, where a survey conducted in 2003 (Richardson, 2003) found that locally-made fermented fish paste (Prahoc) was the dominant protein source, with about half of the meals including some fish paste (but commonly only five grams or less per person), compared to wild/domestic meat protein used in 10 percent of meals, and no protein in 40 percent of meals. Most households reported experiencing protein shortages. In this study we did not quantify the amounts of meat and fish consumed by households, and it is therefore possible that households in the Cardamoms are similarly consuming less than the recommended level of protein or other micronutrients per day. While hunting incomes are lower that agricultural ones, and consumption of fish and bought proteins are higher than consumption of wild meat, the additional income and protein may therefore represent the difference between producing a deficit and breaking even, financially and nutritionally speaking. A preference for wildmeat over domestic meats was stated by several interviewees, and low levels of wild meat consumption may be due to low availability, rather than preference for bought meat. Nutritional surveys, to further investigate the amount of protein and other nutrients that are gained from different food sources, would help to better understand the role and importance of wild meat and fish in the diet.
It seems surprising that remote rural households would buy more meat than they hunt. However, with more context in terms of village characteristics and livelihood strategies, some plausible hypotheses for why this might be emerged. Village livelihoods in the Cardamoms have been shifting generally from subsistence use to a more trade-based system. Prior to the civil conflict, in most Por communities, produce from shifting cultivation, fishing and hunting would have been primarily for own-consumption (Sarou, 2009). However, there have been high levels of immigration of lowland Khmer peoples since the conflict, for whom rice is a culturally important staple food. Rice does not grow well or easily in the high-sloped Cardamom Mountains, and so households grow other arable crops more suited to the area, such as corn, mung bean, and sesame, and sell almost all of it. The profits from these crops are mainly used to buy rice.
In addition to this increased agricultural trade, and due to a combination of factors, it has become more cost effective, and less risky, to buy domestic meat rather than to hunt wildmeat. Where there are only a few adult men in a household, interviewees suggested that households will prioritize their available labor for agricultural work, which brings in stable household incomes, over hunting where incomes are unpredictable. Furthermore, there are still high densities of landmines in the forest surrounding these villages, and therefore the risk of hunting to life is significant. Wildlife population densities, following high levels of hunting pressure during the conflict, are also low, which reduces the potential return (CPUE) from hunting, compared with other livelihood activities which bear less risk. At the same time, improvements in road networks, and agricultural trade between the villages and nearby towns, has increased the availability of cheaper pork and sea fish.
While most households hunt, and a wide range of species are consumed over the year, the primary aim of hunting is often crop protection, with meat from hunting a welcome by-product. Some men also hunt during the pursuit of other livelihood activities, such as fishing. Only a few men in each village hunt commercially, to capture the potentially high incomes from species, such as pangolin, which are in high demand due to their perceived medicinal properties. Likewise, wildlife traders did not make scheduled trips to the study villages, maybe in part due to the low capture rate of wildlife and the remoteness of the area, and in part to hide their trade from wildlife authorities. However, interviews and past surveys (Wutty and Simms, 2005) suggest that commercial hunting gangs, unconnected to village communities in the Cardamoms, are highly active, and could be harvesting much higher numbers of large-bodied, target species which have naturally low population densities and reproductive rates and are therefore even more vulnerable to overhunting (Ripple et al., 2015). These results reflect those of Mckenney et al. (2004), who found that, in villages surrounding Preah Vihear and Kampong Thom (Northern and Central Cambodia, respectively), only 10% of households contained a skilled trapper or hunter, with other households generally only hunting and trapping around their agricultural fields. McKenney et al. also note the existence of militarytrained hunting groups external to the village, taking part in the wildlife trade.
Even low levels of hunting can have high impacts on biodiversity where target species populations are already depleted from past levels of hunting, and therefore natural levels of replenishment are low (Milner-gulland and Mace, 1998; Coad et al., 2018). Rarity can increase demand and in turn price, which means that even when species populations and catchper-unit-effort (CPUE) decline, hunting can remain financially worthwhile (Challender et al., 2015b; Shairp et al., 2016). Pangolins are a prized animal in China, thought to cure a range of ailments, which has result in swift declines in Chinese pangolin populations and increases in their value (Challender et al., 2015a). This has driven a voracious demand for pangolins from neighboring Asian countries, and as these populations decline, is now incentivizing international trade of pangolins from Africa to China (Mambeya et al., 2018). The price of a
live pangolin in Cambodia has increased dramatically since 2005, when Wutty and Simms (2005) recorded the price for a premium quality live pangolin at \$60; hunter interviews conducted for this study suggest that the price in 2013 was between \$150–200 per live animal. As pangolin populations decline due to high demand, the chance of catching pangolins becomes too much of a gamble for hunting to be relied on as a core income-generating activity. However, for households with two or three men—enough to cover agricultural labor requirement—allocating additional manhours to hunting, where returns are unreliable but potentially lucrative, can be worth the risk.
### Under What Circumstance Might Village Hunting Activity Increase?
All three villages studied here were highly dependent on agricultural incomes to buy staple foods—rice, fish, and domestic proteins. Households in each village collectively agree each year on which crops to grow, which are then harvested at the same time and sold to traders in town in bulk. However, the reliance on arable incomes and this strategy of bulk sales comes with risks, mainly due to the declining road network. To conduct our survey, it took us 1–2 days by motorbike from Battambang to reach each study village. For Village 2 we were unable to reach the village on the first attempt because a bridge had recently washed away. The possibility that the corn truck to take produce to market would not be able to get to Village 2 was causing great worry to the community. Due to the current state of the roads, only a few traders with good trucks can reach the villages, and this, combined with a lack of storage for crops, means that prices are set by traders and fluctuate with demand. If roads continue to decline, and traders can no longer reach these villages, households could easily lose most of their season's income. Households may then fall back on incomes from wildlife, which is easily transported by motorbike and can provide a high per kilo profit, as well replacing bought meats with wildmeat and fish. This was shown to be the case in Veal Veng in the Central Cardamoms, where during the wet season road access to villages can be completely cut off, limiting trade in agricultural products, and increasing local dependence on forest products for consumption, such as wild meat (Daltry and Momberg, 2000). There is mixed evidence from our results as to whether smaller agricultural incomes might result in a higher reliance on wildmeat. Village 2 gave the highest scores for wildmeat and had the lowest availability of agricultural land and agricultural incomes. However, analyzing our results by household, we found no correlation between agricultural incomes and wildmeat scores. Further investigation of how households respond to "shocks," such as large losses at harvest-time would help form a better understanding of the links between agricultural production and security, and wildmeat use.
## If These Findings Are Taken to Be Representative, Which Management Approaches for Sustainable Hunting Might Be Most Appropriate in These Villages?
Our results suggest that despite low animal abundance, and the relatively higher availability and consumption of domestic meats, hunting continues in these Cardamom mountain village due to a combination of factors:
In addition, it may be that only small amounts of fish and domestic meat are eaten by households, and in this case even small amounts of wildmeat could have an important impact on household nutrition. We were unable to measure amounts of fish and domestic proteins consumed by households, but this should be a priority for further surveys.
In these three case study villages we suggest that there are three main types of hunters, who may respond to different management approaches:
1. **"Farm" hunters**: for whom arable farming is their main livelihood activity. Most men will set snares to protect their crops and provide meat for the family; hunting increases in the harvest seasons and we hypothesis that is may also increase during times of low fish or crop production. Key species include wild pig, porcupine, muntjac, monitor lizard, turtle, mouse deer. The importance of arable farming (and possibly the relatively low densities of wildlife) mean that these farmers do not have the time to monitor large trap lines far into the mountains. We would suggest that a deeper understanding of how arable farming and forest use (including hunting) interact would be of great use in designing conservation strategies in this region. Where livelihoods and community concerns are focused predominantly on income from arable farming, "wildlife friendly farming" approaches (e.g., Clements et al., 2010) may have potential to help local communities sell their agricultural produce at a fair price, while also benefiting biodiversity. These projects provide benefits to farmers in the form of increased arable yields (providing technical help to farmers), increased access to arable markets (providing reliable transport of crops to market) and guaranteed sales and prices for arable products. In return farmers sign an agreement which may include an agreement not to expand arable fields further into the forest, not to hunt key conservation species, and/or not to use certain hunting methods. In addition, help should be provided to farmers to protect their fields from crop-raiding species that are not in the list of species that can be hunted. Wildlife-friendly farming projects often focus on target species for conservation or reducing habitat conversion (such as large-bodied species at risk from overexploitation, and protected species), as a total cessation of hunting for most rural communities is impractical (and often, in the case of small-bodied, fast reproducing, crop-raiding species, unwarranted). These agreements rely on a local level of project monitoring, to ensure that agreements are being adhered to.
Examples of successful "wildlife friendly farming" projects in Cambodia include the "Ibis Rice" model (Clements et al., 2010), where rice farmers in the Northern Plains are given a favorable price for their rice in return for not using certain herbicides and pesticides on their fields, and not clearing forests, with the aim of protecting habitat for the Giant Ibis and other bird species. Rice is then sold within Cambodia under the "Ibis Rice" Wildlife Friendly-certified brand, and profits reinvested into the project. To determine whether wildlife-friendly farming (for crops other than rice) is an option for communities in the Cardamoms, more detailed baseline studies over a larger number of study sites, including value-chain analyses, would be needed to better understand the farming practices and the importance of farming incomes for local communities in the Cardamoms, and how agricultural practices interact with hunting practices and habitat loss. Meetings with local communities to discuss their main livelihood concerns, and how shocks (such as loss of incomes) are mitigated for, would help to inform the development of conservation projects, as well as developing important lines of communication between conservation organizations and local communities.
Most conservation/development interventions aiming to work with local communities to support biodiversity-friendly land practices will require security of land tenure, whether this is customary or private (Robinson et al., 2018). Interviews in Village 2 suggest that local people feel little security of land tenure, to the extent where local people are concerned about the removal of land mines from agricultural fields, due to the potential of re-classification and redistribution of these areas for large-scale commercial agricultural plantations (e.g., rubber) once they are clear. In addition, under current wildlife laws most hunting is illegal, or legality is ambiguous, and therefore management approaches aiming to promote sustainable harvests may find themselves to be operating illegally. Land tenure insecurity and impractical wildlife laws are common issues facing sustainable wildlife practices in many tropical regions of the world (Coad et al., 2018).
2. **"Forest" hunters**: men belonging to households that have enough able-bodied men and women to cover the labor requirements of arable farming, so that at least one male family member has time to dedicate to more "high risk/high return" livelihood strategies, such as hunting. Hunting may focus more on high value species, such as pangolin, sambar, and bear species. In this case, where "Wildlife friendly farming" agreements are signed at a household or community level, and agreements are monitored, younger members of families engaging in hunting may be instructed by older members to adhere to agreements not to hunt key species. However, where monitoring and implementation of agreements is low, the trade in commercial species could easily continue. In the case of the commercial trade, enforcement may be better targeted at wildlife traders in Battambang and Pursat, or on reducing demand from consumers, rather than at village hunters.
3. **Commercial external hunters**: while we did not collect data on external hunters, Wutty and Simms (2005) suggest that groups of town and forest-based hunters (external to village hunters) target large-mammals for the commercial trade. Hunters in Village 3 reported meeting groups of hunters who were not from the village in the forests surrounding the village. These hunters are unlikely to be influenced by conservation approaches at the village level. Conservation approaches may include increased wildlife trade law enforcement within Cambodia and demand-reduction strategies in consumer countries.
### Thoughts on Future Research Priorities
Our preliminary study was based on four short field visits of 10 days each, in only three villages; these results and discussion of management scenarios should therefore be taken as first hypotheses, based on preliminary field visits. We would strongly encourage further research into village livelihoods in the Cardamom mountains and would especially prioritize studies into the nutritional importance of wild meat and fish, especially in times of economic stress. Household nutritional surveys, and more in-depth surveys of daily hunting returns where possible, would also help better gauge the number of hunted species; our species cards and questions focused on large and medium-bodied mammals and therefore the true number of species hunted (especially birds, reptiles and amphibians) is likely to be much higher.
The continued decline of wildlife populations in the Cardamoms is already flagged as a conservation priority for Cambodia but could also have significant impacts on local rural communities, especially in the context of declining road networks, where closed roads could result in large losses of income for rural farmers, and an increased reliance on own-caught wild meat and fish. In addition, it seems likely that international demand (especially from China) for medicinal wildlife products will continue to incentivize the creation and activity of specialized hunting groups in the Cardamom mountains. This external demand will likely have negative impacts on both biodiversity and in turn local food security, unless there are greater efforts at the international level to change consumer behavior and enforce national and international wildlife trade regulations.
### ETHICS STATEMENT
This study was carried out in accordance with the recommendations of Fauna and Flora International, and following PROFOR guidelines as outlined in their Poverty-Forests Linkages Toolkit, with verbal informed consent from all subjects. All subjects gave verbal informed consent in accordance with the Declaration of Helsinki. The protocol was approved by Fauna and Flora International. Verbal consent was taken rather than written consent. This was due to the low level of literacy in the communities that we were working with. In addition, written documents are often perceived to be very official, and asking for signatures could have caused concern and upset to community members. We therefore decided that verbal consent was more culturally appropriate in this circumstance. As detailed in the manuscript, we emphasized that participation was completely optional, and the identity of all villages and interviewees was kept anonymous. Names of interviewees or households were not recorded at any point.
### AUTHOR CONTRIBUTIONS
All authors designed the study, conducted the fieldwork, and collaborated on drafting the manuscript. LC conducted analyses of the field data.
### REFERENCES
### FUNDING
We would like to thank USAID for funding the field surveys and for further funding to analyse the data and write the manuscript (under award no. AID-BFS-IO-17-00005).
### ACKNOWLEDGMENTS
Many thanks to the three communities that we worked with during this study; we are so grateful for your hospitality, kindness and friendliness. Special thanks to the commune and village chiefs who gave their permission. Many thanks to the University of Phnom Penh for their support and advice. Thanks to Mathew Maltby who commissioned an oversaw the project and was a huge help with logistics.
### SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fevo. 2019.00296/full#supplementary-material
**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Copyright © 2019 Coad, Lim and Nuon. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
# Unsustainable vs. Sustainable Hunting for Food in Gabon: Modeling Short- and Long-Term Gains and Losses
#### David S. Wilkie<sup>1</sup> \*, Michelle Wieland<sup>2</sup> and John R. Poulsen<sup>3</sup>
*<sup>1</sup> Global Conservation Program, Wildlife Conservation Society, Bronx, NY, United States, <sup>2</sup> Africa Program, Wildlife Conservation Society, Bronx, NY, United States, <sup>3</sup> Nicholas School of the Environment, Duke University, Durham, NC, United States*
Today, rural people continue to consume wild animals (aquatic and terrestrial) because they are often cheaper and more available than farmed livestock and fish. In many places where the meat from wild animals is an important source of food and income for poor rural families, the capture, consumption or trade of wild animals is illegal and remains within the informal sector and outside of national accounting and regulatory systems. Few studies exist to help policy makers and wildlife managers develop and implement systems designed to halt unsustainable hunting, prevent species loss, and maintain, over the long term, flows of wildlife available to people as a source of food and income. This paper uses empirical data from a tropical forest area in Gabon within a heuristic simulation model to explore how hunter capture rates would need to change over time to halt unsustainable hunting and to maximize the nutritional and economic value of wildlife as a source of food and income over the long term. Results show that sustainable hunting of wildlife populations that are at or near 50% of carrying capacity (0.5 K) generates more biomass available for consumption and income generation over 25 years than either hunting to maintain current population densities or continuing to hunt unsustainably. Unsustainable hunting generates more biomass than sustainable hunting but only for the first 1 to 3 years after which offtake dwindles rapidly. Achieving sustainable hunting will require that hunters reduce their offtake for 3–13 years until depleted populations recover, which may be unlikely unless they have access to alternative sources of food and income.
#### Keywords: bushmeat, hunting, unsustainable, protein deficit, simulation
### INTRODUCTION
Across the planet many forests, grasslands, rivers, lakes, and coastal waters are empty or being emptied of their wildlife to meet growing human demand for animal-source foods. If demand exceeds the capacity of a wildlife population to replace harvested individuals the population will decline, potentially to local extinction.
The human population continues to grow by over 80 million people annually and is expected to reach 8.6 billion by 2030 (Desa, 2015) – nearly nine times the human population in 1,800. Half of the future increase in human population will be in Africa, and by 2030 one in every four people will be African (Desa, 2015). We are increasingly becoming an urban species, and global economic
#### Edited by:
*Krithi K. Karanth, Wildlife Conservation Society, India*
#### Reviewed by:
*Zuzana Burivalova, Princeton University, United States Nathalie Van Vliet, Center for International Forestry Research, Indonesia*
> \*Correspondence: *David S. Wilkie dwilkie@wcs.org*
#### Specialty section:
*This article was submitted to Conservation, a section of the journal Frontiers in Ecology and Evolution*
Received: *19 October 2018* Accepted: *09 September 2019* Published: *24 September 2019*
#### Citation:
*Wilkie DS, Wieland M and Poulsen JR (2019) Unsustainable vs. Sustainable Hunting for Food in Gabon: Modeling Short- and Long-Term Gains and Losses. Front. Ecol. Evol. 7:357. doi: 10.3389/fevo.2019.00357* development and poverty alleviation is making most of us wealthier, changing what we want and can afford to eat (Bodirsky et al., 2015). This combination of growth, urbanization, and wealth is driving up demand for animalsource foods (Alexandratos and Bruinsma, 2012). In many places, production of animal-source foods is not keeping pace with growing demand, and hunting and fishing of wild animals often increases to unsustainable levels to fill the gap (Wilkie et al., 2011; Ripple et al., 2016).
If this situation continues unchanged, we risk the wide-spread loss of aquatic and terrestrial biodiversity as hungry people eat species after species to extinction (Ripple et al., 2016). Millions of people, particularly poor people, will face a looming protein deficit that will make child malnutrition and "failure to thrive" an appalling norm (Golden et al., 2011). Loss of wildlife populations used as food will change plant and animal species composition and distribution within ecosystems (Poulsen et al., 2013; Trolliet et al., 2019). As a consequence, food webs will be disrupted and destabilized, decreasing ecosystem resilience to climate shocks and risking a cascade of species extinctions.
Estimates of the current volume and value of wild caught terrestrial and aquatic animals are encouraging many governments and development assistance organizations to promote policies that would legalize the trade in wildlife as food and bring this largely informal, weakly regulated economy into a more regulated marketplace. However, guiding policy reform solely on the current dietary and financial value of wildlife used as food fails to take into account that most wildlife populations captured for food are being over-exploited, and thus current capture rates are too high and cannot be sustained over the long-term.
Few studies exist to help policy makers and wildlife managers develop and implement systems designed to halt unsustainable hunting, prevent species loss, and maintain, over the long term, flows of wildlife as a source of food and income (Robinson and Bennett, 2000b; Bennett et al., 2007). Studies that explore the impact that different management decisions would likely have on wildlife population status and annual productivity are particularly lacking. Exploring different scenarios would help determine the approaches for optimizing both species conservation and maintenance of wildlife as a provisioning ecosystem service delivering food and income to families.
This paper uses empirical estimates of the abundance of hunted species in the forests of Gabon to simulate, over 25 years, the impact of three different hunting management scenarios, on wildlife populations, annual production of wildlife, and the benefits wildlife confer to people as a source of food and income. The three scenarios are: (1) reduction of hunting to sustainable levels at current wildlife population densities; (2) managing annual harvest levels (offtake) to move the hunted wildlife population levels toward 50% of carrying capacity (0.5 K) where annual production is greatest, enabling maximum sustainable offtake; and, (3) business-as-usual (i.e., continuation of unsustainable hunting).
We acknowledge that the model presented in this paper is, like all models, a simplification of reality. But as the statistician George Box so sagely noted "All models are wrong, but some are useful" (Box, 1979). Our model is offered as a heuristic device designed to explore the outcomes of different approaches to managing hunting of wildlife for food and income. Whether managers and policy makers: (a) should attempt to halt unsustainable hunting through spatial closures, individual or community quotas, exclusion of non-rights holders, or taxation and (b) how they should assess what is or is not sustainable offtake is beyond the scope of this paper.
### MATERIALS AND METHODS
To explore what happens to the supply of wild meat as hunting transitions from being unsustainable to sustainable, we will use an example based on data reported from recent research in 5,807 km<sup>2</sup> of tropical moist forest around the town of Makokou, Gabon (Koerner et al., 2017). The authors conducted surveys of terrestrial and arboreal wildlife within three zones of increasing distance from villages (0–6 km to 1,871 km<sup>2</sup> , 6–15 km to 2,126 km<sup>2</sup> , and >15 km to 1,830 km<sup>2</sup> ).
For this hunting simulation we focus on three monkey species (Cercopithecus mona pogonias—crested mona monkey, C. n. nictitans—great white-nosed monkey, and Lophocebus albigena—gray-cheeked mangabey), and small (Cephalophus monticola—blue duiker) and medium-size duikers (C. callipygus—peters duiker, C. leucogaster- white-bellied duiker, and C. dorsalis castaneus—bay duiker). These were chosen because Koerner et al. (2017) report estimates of their densities (**Table 1**) and they are the most commonly hunted, traded and consumed species when wildlife populations are not depleted from overhunting (Abernethy et al., 2013). For each species or species group (e.g., medium-sized duikers), we used density data within each zone to calculate total abundance. To calculate sustainable yield, H–the number of animals that can be taken from a population of any size over an indefinite period without depleting the stock, we used the standard Gordon-Schaefer equation under the assumption of logistic population growth.
### rSt(1 − St/K) = H<sup>t</sup>
The intrinsic rate of population growth (r) for each species was drawn from the literature (Fa et al., 1995) or computed using Cole's equation (Cole, 1954). Several studies suggest that most hunting in central Africa occurs within approximately 15 km of settlements (Abernethy et al., 2013; Coad et al., 2013; Beirne et al., 2019). Given this we assumed that the area further than 15 km from villages can serve as a largely unhunted reference point where populations are close to or at carrying capacity (K). S<sup>t</sup> is the stock (abundance) and H<sup>t</sup> the sustainable yield of a hunted species at time t. Because the annual population growth rate (dS/dt) when plotted against population size (S) is shaped like an inverted U (**Figure 1**), sustainable yield will be the same when a population is near zero (say 0.01 K) or near carrying capacity (say 0.99 K), and will be maximized around 0.5 K (i.e., maximum sustainable yield—MSY) where a population is growing at its fastest rate. We understand the risks of harvesting a population at MSY (Reynolds et al., 2001)
Density from Koerner et al. (2017)
and are not advocating this as a hunting management strategy. Rather, we are simply using the concept in the model to estimate the maximum sustainable offtake that is theoretically possible. In reality, a more precautionary approach where offtake is set less than MSY would reduce the risk of over-exploitation and local extinction.
We simulated offtake over a 25-year period for three scenarios: (1) sustainable hunting at current population densities; (2) maximum sustainable yield when population levels were at 0.5 K; and, (3) unsustainable business-as-usual hunting. Achieving MSY requires that the initial population at t<sup>0</sup> grow or shrink to 0.5 K. To allow a population to grow, offtake was set at 10, 25, or 50 percent below sustainable yield to leave a surplus to increase S in time t+1. If the initial population was above 0.5 K offtake was increased to 10, 25, or 50 percent above sustainable yield to deplete the population to 0.5 K. By manipulating offtake to 10, 25, or 50 percent below or above sustainable yield, the model is able to influence how rapidly the population level of each hunted species reaches 0.5 K—the higher the percentage the more quickly 0.5 K is obtained. For the business-as-usual scenario, we set offtake 10, 25, or 50 percent above sustainable yield. For all three scenarios, the abundance of each species within the near (<6 km) and intermediate (6–15 km) zones at t<sup>0</sup> was set using their empirically measured density (Koerner et al., 2017). Abundance within the far zone (>15 km) at t<sup>0</sup> was set at 90% of the empirical value (i.e., 0.9 K) so that the sustainable yield calculation would be >0. We converted offtake of individuals to biomass using average adult body weight from the literature (Kingdon et al., 2013).
From meat consumption studies in Gabon (Starkey, 2004; Wilkie et al., 2005; Foerster et al., 2012), we know that an average individual (reported as an Adult Male Equivalent) consumes 0.25 kg of wild meat per day in rural villages. This is approximately 100% of daily protein requirements as specified by the World Health Organization. Daily consumption of wild meat declines to 0.12 kg/AME/day in provincial towns (like Makokou) and 0.02 kg/AME/day in large cities (like Libreville). Using these figures, we calculated how many consumers in villages, towns or cities could be supplied, based on current consumption patterns, with wild meat over time from unsustainable or sustainable hunting within 6 and 15 km of settlements, and within the total study area of 5,807 km<sup>2</sup> .
### RESULTS
Assuming that wildlife populations in the largely unhunted area >15 km from settlements are close to or at carrying capacity, the Koerner et al. (2017) data suggest that the near zone wildlife populations within 6 km from settlements are already overexploited averaging 0.3 K. The intermediate zone populations are currently being hunted at 0.56 K (i.e., close to MSY). As expected based on hunters' preference for large bodied species, mid-size duikers appear to be more heavily depleted than smaller bodied species both in the near (0.16 vs. 0.38 K) and intermediate (0.36 vs. 0.64 K) zones (**Table 1**).
Within the near zone the abundance of all species populations at t<sup>0</sup> are below 0.5 K. In the intermediate zone only midsize duiker abundance is below 0.5 K at t0. Given this, even if hunters decided to hunt wildlife within the three zones to maintain, permanently, their current abundance (i.e., scenario 1–sustainable yield) this would not maximize offtake over a 25-year period, because wildlife densities are either above or below 0.5 K (**Figure 2**). For example, the blue duiker population is currently below 0.5 K at t<sup>0</sup> within the near zone (0.38 K) and above 0.5 K within the intermediate zone
(0.64 K). If hunting was maintained at, or reduced to, sustainable levels such that current population levels of blue duikers did not change (scenario 1), hunting would generate 22% less biomass over a 25-year period than would a management system that increased or decreased the population to 0.5 K and then allowed hunting at MSY. For all modeled species combined, if hunting was to be managed at t<sup>0</sup> population levels sustainable offtake would generate 26–30% less biomass available for consumption or income generation than when hunted at 0.5 K. So simply halting unsustainable hunting at t<sup>0</sup> population levels would not be rational if hunters maximize the rate of offtake following Charnov's (1976) marginal value theorem.
TABLE 2 | Estimated number of years that depleted wildlife populations within 6 km of villages in Gabon might take to recover to 0.5 K when offtake is reduced by 10, 25, or 50%.
The estimated time to population recovery to 0.5 K in the near zone when offtake is reduced by 10, 20, and 50% below sustainable yield takes on average 9, 4, and 3 years, respectively (**Table 2**), with mid-sized duikers taking the longest time (13 years) to recover to 0.5 K.
Reducing offtake below sustainable yield within 6 km from villages to allow depleted populations to recover to 0.5 K would require that hunters leave (i.e., not hunt) 12,679 to 15,340 animals in the forest (**Table 3**) during the recovery years (**Table 2**). As recovery is faster when offtake is reduced 50% below sustainable yield, the total number of animals not hunted is lower than for a 25 or 10% reduction. Reducing hunting to allow recovery of depleted wildlife populations would also reduce wild meat biomass available for consumption by 97,019 to 120,551 kg. Similarly income to all hunters combined (assuming they sell 50% of their catch) would be reduced by \$26,680 to \$33,151 based on an average sales price of \$0.55/kg (Gally and Jeanmart, 1996).
Though this simulation only includes a subset of all wildlife species hunted for food near Makokou, the simulated estimate of 103 kg/km<sup>2</sup> for MSY is within the lower range of estimates for maximum sustainable production in tropical forests (Robinson and Bennett, 2000a).
Hunting wildlife populations so that their abundance recovered or declined to 0.5 K, and then hunting them at MSY always generated a higher total biomass over 25 years than sustainable yield scenario 1, and the unsustainable businessas-usual scenario 3 (**Figure 2**). That said, for the first 2–3 years offtake in the business-as-usual scenario was higher than sustainable yield (scenario 1), but dropped to 50% of MSY after 19 years when offtake is 10% higher than is sustainable, 9 years when offtake is 25% higher than is sustainable, and 6 years when offtake is 50% higher than is sustainable. Unsustainable offtake falls to below 1% of MSY after 23 years when offtake is 25% higher than is sustainable and 13 years when offtake is 50% higher than is sustainable (**Figure 2**).
To calculate the Net Present Value of wildlife hunted for food (i.e., the current value relative to future cash returns over a given time period), we assigned a price of \$1 per kg and used a 20% discount rate, which is a realistic cost of capital in Gabon (i.e., the price lenders charge borrowers). NPV was higher for the business-as-usual scenario only during the first 5 years with a 10% unsustainable hunting rate. In all other TABLE 3 | Reduction in: (a) animals hunted; (b) biomass available for consumption or sale; and, (c) economic value to hunters when hunting of depleted wildlife within 6km of villages is reduced by 10, 25, or 50% below sustainable yield (*H*) to allow populations to recover to 0.5K.
timeframes and unsustainable hunting rates, NPV was higher for the MSY scenario.
Though sustainable hunting at current (t0) population levels generates less biomass available for consumption than hunting at MSY, it would produce enough wild meat for a village population of 0.75 people/km<sup>2</sup> which is close to the global estimate (Robinson and Bennett, 2000a; Peres and Nascimento, 2006) of the human carrying capacity of tropical forests in terms of protein supply (i.e., 1 person/km<sup>2</sup> ).
If the whole area is hunted at MSY (i.e., when all hunted populations are at 0.5 K), the three primate and four ungulate species could provide a sustainable supply of animal source foods over a 25-year period to an average of: a) 6,185 people in villages, covering 100% of daily protein requirements; or b) 13,402 people in provincial towns, meeting 46% of daily protein requirements; or c) 80,411 people in large cities, meeting 8% of daily protein requirements.
In contrast, when wildlife are being hunted unsustainably (business-as-usual), over 25 years the 10% depletion scenario supplied wild meat on average to only: (a) 3,755 people in villages; or (b) 8,137 people in towns, or (c) 48,820 people in cities, and the 50% depletion scenario supplied wild meat on average to: (a) 1,271 people in villages; or (b) 2,754 people in towns; or (c) 16,525 people in cities.
Unsustainable hunting (i.e., the 10, 25, and 50% business-asusual scenario) only increased the supply of wild meat to village, town and city dwellers for the first 1 or 2 years with supply plunging to <50% of the MSY scenario by years 6, 10 and 19 in the 50, 25, and 10% unsustainable hunting scenarios, respectively.
### DISCUSSION
In 1998, 40% of Central African forest was within 10 km of a road (Abernethy et al., 2013). By 2017 that had increased to 53% (Koerner et al., 2017). Of the 177 species that are hunted in Central Africa (Taylor et al., 2015) for food, 97 are being hunted at unsustainable levels according to the IUCN Red List. In this paper, we showed that populations of commonly hunted species (3 primates and 4 ungulates) are likely being hunted at unsustainable levels (i.e., population abundance has fallen to an average of 0.3 K) in a zone within 6 km of settlements, near Makokou in Gabon. Larger-bodied red duikers appear to be unsustainably hunted (0.36 K) up to 15 km away from settlements.
Establishing and enforcing rules to render hunting sustainable at current wildlife population levels across all zones would: (a) require hunters to substantially reduce their offtake for the initial years when unsustainable hunting generates higher offtake (**Figure 2**); and (b) generate approximately 20% less biomass than managing wildlife populations at 0.5 K (scenario 2). Similarly, allowing populations to reach 0.5 K so that they could be hunted at MSY would require hunters to reduce offtake by as much as 282,166 kg over 2 to 3 years or until wildlife populations recovered to 0.5 K. Persuading all hunters to reduce their offtake even for a few years is unlikely unless they are compensated for lost food and income. Legalizing hunting, only for hunters from villages with legitimate historical claims to nearby forest, may also not solve the problem if: (a) the majority of current hunters have legitimate claims to hunt; and, (b) traditional hunting zones do not extend beyond 15 km from villages, and thus are not under the jurisdiction of legitimate village hunters.
The government of Gabon has on more than one occasion voiced an interest in legalizing the trade in wildlife as food and using the tax revenue to finance wildlife conservation both within and outside of national parks and reserves. Results from this study show that for depleted populations to recover offtake would have to decrease substantially at least for a few years. As a result, tax revenues from a legalized trade would see a comparable decline during the recovery years, making it unlikely that taxes would even cover the costs of tax collection, let alone increase investment in wildlife law enforcement (Wilkie et al., 2006).
A shift from unsustainable to MSY hunting does initially impose costs on hunters. But the alternative business-as-usual scenario causes offtake to decline rapidly so that within 10 to 18 years, for the 50 and 25% unsustainable hunting scenarios, wildlife populations have been effectively wiped out (**Figure 2**).
The Koerner et al. (2017) data as interpreted in this paper show that wildlife populations in 32% of the study area (near zone <6 km) have already been depleted to 0.3 K, and in 37% of the area (intermediate zone 6–15 km) wildlife have been depleted to 0.56 K. The remaining 31% (>15 km from villages), we assume, based on the reported paucity of human sign, to be only rarely hunted and thus wildlife populations might be near carrying capacity (i.e., 0.9 to 1.0 K). Taken together, this means that wildlife populations within 68% (i.e., >6 km from settlements) of the study area near Makokou are still in relatively robust health (i.e., they are near or above 0.5 K). This is maybe not surprising as Gabon has a small human population (1.7 million–CIA World Factbook) for its geographic area. Moreover, 89% of Gabonese live in urban areas and 88% of the country is still covered in forest.
If the entire forest was depleted to the same level as the near zone (i.e., 0.3 K), the forest might be emptied of wildlife in 5 to 10 years under current, unsustainable, hunting levels. Setting the initial abundance of all 7 species at 0.1 K, leads to their extirpation within 3–4 years. In many areas of central Africa, the forest is already almost empty.
Even under the most optimistic scenario, where the near, intermediate and far zones are all hunted at MSY, the forest will only feed a small number of people. This means that legalization of hunting is not a solution for the poor rural families who depend on wildlife as their primary source of dietary protein. Even if hunting was legalized everywhere within 50 km from settlements (5,807 km<sup>2</sup> – an area almost two times the size of either the state of Rhode Island or the country of Luxembourg) the maximum possible sustainable offtake would feed 6,185 people. This number drops to 4,235 if hunting is limited to within 15 km of settlements, and to 1,930 if hunting is restricted to 6 km of settlements. With human population growth in Gabon estimated at 1.92% in 2017 (CIA World Factbook), in 20 years the population will have increased by more than 50%, with much of that growth being in urban areas. That said, in the future, all other things held constant, an even smaller percentage of the Gabonese village population can expect to get their protein supply solely from wildlife.
The situation in towns and cities is more complicated given that wild meat is a rival good, and the meat eaten in villages is no longer available to be consumed in towns and cities. Hunters can either use wild meat to feed their extended families or barter or sell some or all of it to purchase necessities or luxury items. If, implausibly, all animals from all zones hunted at MSY were traded to the nearest town and none were consumed in the hunters' villages, then 98% of the town of Makokou could get a minority (46%) of their daily protein requirements from wildlife. This is not only unrealistic, but the residents of Makokou would still need to find 54% of their daily protein requirements from other animal source foods.
The production of wildlife is limited by the availability of high-quality habitat. In Gabon and across central Africa, wildlife habitat is shrinking, not expanding, as forest lands are converted to farms, plantations, mines, roads and settlements (Austin et al., 2017; Kleinschroth et al., 2019). So, supply is either optimistically likely to remain static at MSY (assuming we can manage hunting sustainably) or, more realistically, will shrink as non-hunting factors like habitat destruction begin to drive down wildlife populations.
Rarely do Gabonese hunters trade more than 50% of the animals they hunt (Coad, 2007; Table 5.1), but this could change if consumers were willing to pay higher prices. If supply shrinks and/or demand increases, we might expect the price to rise, because there is evidence that demand for wild meat in Gabon is relatively price inelastic (Wilkie et al., 2005; Foerster et al., 2012) when substitutes are more expensive or unavailable. If price rises, then hunters may be motivated to sell a greater portion of the animals they capture. With less wild meat, levels of protein and micro-nutrient deficiency and "failure to thrive" amongst children will increase in wild meat-dependent villages (Golden et al., 2011).
As the human population of Gabon continues to grow, and as successful poverty alleviation efforts increase household income, we expect demand for animal-source foods to increase substantially (Wilkie et al., 2016) coupled with static or declining wild meat availability. To avoid this looming protein deficit and to prevent protein-hungry people from eating wild animals to extinction, conservation organizations must convince development organizations and donors to invest in increasing the supply of sustainably produced animal-source foods. Additionally, these investments should focus on feeding growing provincial towns close to still relatively abundant wildlife populations, and large metropolitan cities where per capita demand for wild meat is small but the aggregate demand of millions of consumers is huge.
Loss of wildlife from unsustainable hunting and fishing will have irrevocable, long-term impacts on forest species composition, distribution, productivity, and carbon content (Poulsen et al., 2013). But this conservation crisis cannot solely be solved with the classical conservation solution (i.e., establishing and managing wildlife populations within state protected areas and community reserves). Rather, to avoid this looming protein deficit these provincial towns and metropolitan cities must be able to develop profitable and sustainable enterprise that can supply animal-source foods in sufficient quantity to meet demand.
Thankfully these towns and cities are large enough to support profitable private-sector livestock, farmed fish, marketing, butchering, and veterinary-care enterprises. A focus on family-scale, back-yard production of new, more disease-resistant and productive breeds of poultry and other small livestock (guinea pigs and rabbits), makes sense for several reasons. Back yard production, minimizes capital costs, helps empower women as they are often the small livestock owners and producers, increases opportunities for unemployed and under-employed youth, avoids the need for cold chains (i.e., refrigerated supply chains), and is scalable as additional back-yard producers adopt observably successful innovations. Avoiding this looming protein deficit will not only help conserve wildlife hunted for food, it will increase household food and income security, reduce unemployment, and might also reduce motivation for youth and whole families to leave relatives and their homelands to seek a better life elsewhere.
This paper shows that current levels of hunting of wildlife for food and income risks fully depleting wildlife populations within 6 km of settlements. Gabon's growing human population will increase wild meat demand. Coupled with habitat loss and a reduction in wildlife production, our model suggests that current levels of hunting risk depleting wildlife across the landscape. To transition from unsustainable to sustainable hunting will require substantial reduction in hunting levels to allow depleted wildlife populations to recover. During this recovery period the supply of wildlife for food and income will be significantly lower than it is currently. Hunters are unlikely to willingly reduce the benefits they gain from hunting at current levels, even if they are not sustainable over the next 20 years. The government of Gabon and its conservation partners will need to find ways to offset the short-term losses of food and income until wildlife populations recover, otherwise hunters will have little interest in complying with sustainable hunting regulations and may take actions to undermine them.
### AUTHOR CONTRIBUTIONS
DW and MW contributed the initial concept. JP provided the wildlife abundance data from Gabon. DW developed the simulations. DW, MW, and JP contributed to the writing of the paper.
### REFERENCES
Cole, L. C. (1954). The population consequences of life history phenomena. Quart. Rev. Biol. 29, 103–137. doi: 10.1086/400074
Desa, U. (2015). World Population Prospects: The 2015 Revision, Key Findings and Advance Tables. Working Paper No. ESA/P/WP.241.
### FUNDING
The European Commission Directorate-General for International Cooperation and Development (DG DEVCO) and the United Nations Food and Agricultural Organization for financial support to DW and MW. Duke University provided funds to JP for field data collection.
### ACKNOWLEDGMENTS
The authors would like to thank the Government of Gabon, in particular the Center National de la Recherche Scientifique et Technologique (CENAREST) and the Agence Nationale des Parcs Nationaux (ANPN) for permission to collect field data, and would like to thank Dr. Timothy O'Brien and Dr. Samantha Strindberg for their comments on the concept and a technical review of the simulation. We would also like to thank the European Commission Directorate-General for International Cooperation and Development (DG DEVCO) and the United Nations Food and Agricultural Organization for financial support to DW and MW. Field data collection for this paper was supported by funds provided to JP from Duke University.
laws be enforced and wildlife survive in Central Africa: evidence from Gabon. J. Int. Wildlife Law Policy 9, 335–349. doi: 10.1080/138802906010 39287
**Conflict of Interest:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Copyright © 2019 Wilkie, Wieland and Poulsen. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
## Impact of Low-Intensity Hunting on Game Species in and Around the Kanuku Mountains Protected Area, Guyana
Matthew T. Hallett 1,2,3,4 \*, Anouska A. Kinahan<sup>5</sup> , Rayon McGregor <sup>6</sup> , Thadaigh Baggallay <sup>5</sup> , Timothy Babb<sup>6</sup> , Howard Barnabus <sup>7</sup> , Asaph Wilson<sup>8</sup> , Fernando M. Li <sup>7</sup> , Wesley W. Boone<sup>4</sup> and Brittany A. Bankovich<sup>9</sup>
*<sup>1</sup> School of Natural Resources and Environment, University of Florida, Gainesville, FL, United States, <sup>2</sup> Institute for the Environment and Sustainability, Miami University, Oxford, OH, United States, <sup>3</sup> Conservation Department, Jacksonville Zoo and Gardens, Jacksonville, FL, United States, <sup>4</sup> Department of Wildlife Ecology and Conservation, University of Florida, Gainesville, FL, United States, <sup>5</sup> Frankfurt Zoological Society, National Park, Georgetown, Guyana, <sup>6</sup> Protected Areas Commission, National Park, Georgetown, Guyana, <sup>7</sup> Caiman House Research Station, Yupukari Village, Guyana, <sup>8</sup> South Rupununi Conservation Society, Wichabai Ranch, Guyana, <sup>9</sup> Lovett E. Williams Jr. Wildlife Research Laboratory, Fish and Wildlife Research Institute, Florida Fish and Wildlife Conservation Commission, Gainesville, FL, United States*
#### Edited by:
*Robert Nasi, Center for International Forestry Research, Indonesia*
#### Reviewed by:
*Charlotte H. Chang, National Institute for Mathematical and Biological Synthesis (NSF), United States Fiona Maisels, Wildlife Conservation Society, United States*
> \*Correspondence: *Matthew T. Hallett mhallett2320@ufl.edu*
#### Specialty section:
*This article was submitted to Conservation, a section of the journal Frontiers in Ecology and Evolution*
Received: *05 January 2019* Accepted: *15 October 2019* Published: *08 November 2019*
#### Citation:
*Hallett MT, Kinahan AA, McGregor R, Baggallay T, Babb T, Barnabus H, Wilson A, Li FM, Boone WW and Bankovich BA (2019) Impact of Low-Intensity Hunting on Game Species in and Around the Kanuku Mountains Protected Area, Guyana. Front. Ecol. Evol. 7:412. doi: 10.3389/fevo.2019.00412* Unsustainable harvest is driving population declines in tropical forest species across the globe. Despite maintaining the second highest percent forest cover in the world (85%), concern is increasing in Guyana that unmanaged commercial and subsistence hunting activities could result in defaunation, and the cascading ecological effects of "empty forests." The Rupununi region of southwestern Guyana, home to the Kanuku Mountains Protected Area (KMPA), hosts one of the world's lowest human population densities (0.42 people/km<sup>2</sup> ), as well as large, intact tracts of both Neotropical savanna and forest habitats, making it one of the country's most biodiverse regions. Indigenous Makushi and Wapichan communities that reside there have maintained subsistence lifestyles mediated by traditional beliefs and management practices for millennia. However, as human populations and access to markets increase, there is a corresponding increase in the harvest of natural resources. Protected areas have long been recognized for their role in biodiversity conservation, while also serving as a reserve for subsistence hunters. The KMPA, one of Guyana's newest protected areas, allows for the continued sustainable use of its resources by indigenous communities. It is critical to understand the patterns, impacts, and levels of hunting that are sustainable in and around the protected area so that biodiversity can be managed and conserved effectively. Our study shows that the impact of current hunting intensity in and around the KMPA remains relatively low and supports the hypothesis that Neotropical forests can support hunting pressure of <1 person/km<sup>2</sup> . While our results show that current levels of hunting in the region can be considered sustainable, small shifts in activity patterns and distribution of preferred game species were observed in sites subject to higher hunting intensity, which in turn appears to have had cascading effects on non-hunted species. Our results serve as a caution for the Kanuku Mountains region and an indication of the truly low levels of harvest that some species can sustain before populations begin to show declines. Further, we suggest a system-level approach to monitoring that incorporates both preferred game and non-hunted species, as well as indigenous knowledge of patterns of use and trends in populations of game species. This approach to monitoring would serve as an effective early warning system, allowing communities, managers, and policy makers to intervene before animal populations are significantly impacted by overhunting.
Keywords: bushmeat, hunting intensity, protected areas, indigenous lands, Kanuku Mountains, Guyana, large mammals, Rupununi
### INTRODUCTION
Hunting intensity has reached unsustainable levels across much of the tropics, representing the most pressing threat to game mammal and bird populations after habitat loss (Redford, 1992; Fa and Peres, 2001; Nasi et al., 2011; Wilkie et al., 2011; Maxwell et al., 2016; Ripple et al., 2016; Young et al., 2016). Studies from across the global tropics assessing game mammal and bird species abundances under moderate and heavy hunting intensities have shown declines by an average of 83 and 58%, respectively (Wilkie et al., 2011). Hunting of long-lived, largebodied species is only considered sustainable under low intensity harvest regimes—when exploitation is ≤ 20% of production (Fa et al., 2002). For tropical forest species, establishing sustainable harvest regimes has become increasingly important to ensure the long-term survival of hunted species, while at the same time maintaining a reliable, low-cost source of protein for subsistencebased communities. Determining the levels of harvest that are sustainable for tropical forest species is also particularly important for policy makers who are responsible for setting rules and regulations related to hunting, as well as for managers of conservation areas that allow for the continued use of resources by indigenous communities, which is typically the case in the Neotropics.
Large-bodied species tend to carry a higher risk of local extinction from overhunting when compared to smaller-bodied animals, due to a combination of particular biological traits (i.e., low reproductive rates, and naturally low population densities), behavioral traits (i.e., diurnally active, high visibility, slow moving, repeated use of den/resting sites), and external environmental factors (i.e., limited geographic range) (Cardillo et al., 2005; Fa and Brown, 2009). Studies have shown an overall preference for large-bodied frugivorous and herbivorous mammals among hunters in the Neotropics (Redford and Robinson, 1987; Jerozolimski and Peres, 2003), which has resulted in documented local population declines of lowland tapir (Tapirus terrestris), red brocket (Mazama americana) and white-tailed deer (Odocoileus virginianus), white-lipped peccary (Tayassu pecari), and giant armadillo (Priodontes maximus) (Cullen Jr. et al., 2000; Peres, 2001; Weber and Gonzalez, 2003; Di Bitetti et al., 2008; Superina et al., 2014).
Little of the substantial plant biomass present in tropical forests is readily available as food for the large, terrestrial frugivores and herbivores (DeWalt and Chave, 2004) that are preferred by hunters (Peres, 2001; Ripple et al., 2016), as it tends to be either out of reach in the upper canopy or indigestible (Waterman and McKey, 1989; Fa and Peres, 2001; Fa and Brown, 2009). Large, tropical forest mammals, therefore, exist naturally at relatively low densities (Arita et al., 1990), making them particularly vulnerable to overhunting even at moderate or low levels of hunting intensity (Wilkie et al., 2011). Further, omnivorous species (i.e., peccaries) achieve higher reproductive rates and crude and metabolic biomass than their purely herbivorous counterparts (Bodmer, 1989), putting herbivorous species like lowland tapir, red brocket and white-tailed deer at a higher risk of overhunting. Declines in the populations of these key species can have significant negative effects on tropical forest ecosystems through decreased seed dispersal and seedling survival, changes in vegetation cover and composition, and functional compensation (Peres and Dolman, 2000; Terborgh et al., 2001; Peres and van Roosmalen, 2002; Stoner et al., 2007; Wright et al., 2007; Beck et al., 2013). It is estimated that for communities dependent exclusively on wildlife for protein, tropical forests can support ∼1 person/km<sup>2</sup> (Robinson and Bennett, 2000), and therefore, a negative impact on the populations of preferred game species can still occur even if only a few animals are hunted per square kilometer per year (Mena et al., 2000).
In Guyana, the demand for wild meat is steadily increasing in the country's growing urban centers. An estimated 625 tons of wild meat per year (0.2 tons/km<sup>2</sup> /year) are consumed in Guyana's capital (Puran et al., 2017)—a rate of consumption comparable to the Amazon Basin (0.23 tons/km<sup>2</sup> /year) (Rushton et al., 2005; Nasi et al., 2011), but much lower than the Congo Basin (1.98 tons/km<sup>2</sup> /year) (Fa and Purvis, 1997; Fa et al., 2002; Nasi et al., 2011). Lack of employment opportunities, coupled with increasing access to markets (Wilkie et al., 2000; Laurance et al., 2006; Puran et al., 2017), means that many indigenous hunters in Guyana's interior have shifted from hunting as a predominately subsistence activity to hunting that yields enough to both meet subsistence needs and supplement incomes.
Protected areas have long been recognized as important refuges for biodiversity. They can serve as a key buffer against local extinctions driven by overhunting (Le Saout et al., 2013), but are also recognized for their ability to function as a reserve that can meet the continued needs of subsistence hunters (Robinson and Bennett, 2000). The Kanuku Mountains Protected Area (KMPA) is one of Guyana's richest protected areas in terms of biodiversity (Montambault and Massa, 2002), and, like all of Guyana's protected areas, the KMPA is categorized as an IUCN category VI protected area, which allows for the continued sustainable use of resources by surrounding indigenous communities. The KMPA is bordered by 11 titled villages and 10 satellite communities, located in the adjacent Rupununi savannas. These largely indigenous Makushi and Wapichan communities are dependent on the KMPA's forest resources for subsistence. Approximately 20 and 55% of Rupununi households reported that wild meat and wild-caught fish are their primary source of protein, respectively (Luzar et al., 2012). More recently, a 2016 survey showed that all 21 communities have at least some residents that hunt within the KMPA and adjacent indigenous titled lands, albeit to varying extents, intensities, and for different purposes (Protected Areas Commission, unpublished data).
The extent to which the Kanuku Mountains region is supplying the growing demand for wild meat in Guyana's urban centers is unknown; however, the local trade in wild meat currently fetches the highest return on investment for any food product in the region (FAO, 2015). Although this area supports one of the lowest human population densities in the world (0.42 people/km<sup>2</sup> ), the indigenous population of the region has been increasing steadily (Bureau of Statistics (Guyana), 2016), while also transforming from primarily subsistence to increasingly commercial livelihoods. Demand for economic development, the opportunity presented by growing markets for wild meat, continued erosion of traditional beliefs and practices (Iwamura et al., 2016), and climate change models that predict this region will likely trend toward progressively hotter and drier conditions (Bovolo et al., 2012), creates the potential for significant negative effects on wild animal populations. The shift in hunting patterns raises particular concern among conservation managers and community leaders alike, with several communities already identifying overhunting as a driver of observed population declines, and even local extinctions, of some preferred game species (Protected Areas Commission, unpublished data).
Using a combination of household surveys, focus group discussions, community workshops, and camera-trap data, this study examines the type, level, and perceived and quantifiable impacts of low-intensity hunting on the occupancy, relative abundance, and activity patterns of important game species in the KMPA and surrounding indigenous titled lands.
### MATERIALS AND METHODS
### Study Area
The Kanuku Mountains Protected Area (KMPA) encompasses 611,000 hectares of largely intact tropical forest in south-western Guyana (**Figure 1**). The mountains are divided into eastern and western ranges by the north-south course of the Rupununi River, a tributary of Guyana's largest river, the Essequibo. The protected area is largely composed of gallery, lowland, and montane, deciduous and evergreen tropical forest, with 1% comprising the surrounding Rupununi savannas. The Kanuku Mountains highest peak reaches 1,067 m asl, with a number of minor peaks in its western range above 900 m asl. The surrounding Rupununi savannas lie between 120 and 150 m asl, are ecologically connected to Brazil's Rio Branco savanna system (Montambault and Massa, 2002), and are analogous to the cerrado savannas of eastern Brazil (Eden and McGregor, 1992). The region experiences two rainy seasons, one long (May-August) and one short (December), and a long dry season (September–April), with an average annual rainfall of 1,500– 2,000 mm.
The Kanuku Mountains and associated savannas are extremely rich in biological diversity, hosting approximately 70% of mammal, 53% of bird (including 17 of the 25 bird species endemic to the Guiana Shield), and 26% of plant species recorded in Guyana (Montambault and Massa, 2002). The KMPA also harbors healthy populations of many species that are listed on the IUCN Red List of Threatened Species, such as the harpy eagle (Harpia harpyja), giant river turtle (Podocnemis expansa), arapaima (Arapaima arapaima), and giant river otter (Pteronura brasiliensis).
The 21 communities adjacent to the KMPA are composed primarily of indigenous Makushi and Wapichan people who maintain traditional livelihoods, including subsistence fishing, farming, and hunting. The Kanuku Mountains region supports populations of game species preferred by hunters, including lowland tapir, white-lipped peccary, red brocket deer, lowland paca (Cuniculus paca), and black curassow (Crax alector).
### Assessing Hunting Patterns and Intensity
In 2015 and 2016 individual household surveys were carried out in all 21 communities surrounding the KMPA. A minimum of 50% of households were surveyed in each community, however in communities where the total number of households totaled <50, 100% of households were surveyed. In total, 815 households participated across the 11 villages (21 communities). The survey was designed to obtain a baseline of the Knowledge, Attitudes, and Practices (KAP) of participating communities as they relate to resource use in and around the protected area. Specific to this study, we calculated the percentage of households that engage in hunting activities in each community, as well as the species hunted. For each species identified as hunted, we also determined how frequently they were hunted and whether respondents perceived the abundance of each of the species as less, more, or the same as 10 years ago. Respondents were also asked their primary purpose for hunting (i.e., home use, selling, sharing, or multiple purposes).
Following these KAP surveys, in 2017, we carried out resource use mapping exercises in each of the 21 communities. Using participatory mapping and focus group discussions, each community worked with facilitators to create a sketch map of their resource use areas, and to identify hunting areas, hunting methods, the purpose of hunting, and hunting patterns using a seasonal calendar. Lists of hunted species developed from the KAP surveys were presented to focus groups and were verified or amended as necessary. Following the KAP surveys, communities were re-visited, and, at open village meetings, were asked similar questions to those in the KAP surveys, such as which species are hunted the most, which are getting less abundant, and which species villagers are most concerned about. A participatory voting approach was used in these meetings to verify trends in the individual responses from the KAP surveys. Once confirmed, maps of hunting areas were developed for each community.
boundaries of the KMPA (solid line) and indigenous community titled lands (dotted line) (Arino et al., 2012; Environmental Systems Research Institute (ESRI), 2015).
### Camera-Trap Surveys
Camera-trap photos were obtained as part of multi-species camera-trap studies of the Rupununi Region, following well-established methods for camera-trap research (Karanth and Nichols, 1998; Silver, 2004). Camera-traps (Bushnell Trophy Cam #119447C, #119734C, #119736C, and #119837C; Bushnell <sup>R</sup> , KS, USA) were set 2–3 km apart, with a single camera at each trap location, set 30–40 cm from the ground in proximity to observed animal sign. Cameras were active 24 h per day, with a 1 second delay between captures, recording the date and time with each 3-image sequence. Images of the species of interest that occurred at the same trap site within a period of 30 min were excluded to ensure that photo occasions were independent (Silver, 2004). In an effort to reduce wariness around cameras and avoid biased capture rates, no scents or lures were used, and all cameras employed were equipped with infrared flash.
Camera-trap sites were selected for inclusion in this study based on their proximity to hunting areas documented by KAP surveys and resource use mapping exercises. This included areas surveyed within the KMPA, as well as the adjacent titled lands of 14 indigenous communities. Camera-traps were set in clusters of 20–30 cameras and left in the field for 40–365 trap nights between May 2012 and October 2016. In order maximize the number of sampling locations across the study area, we integrated data from two different projects that employed identical methodologies, with the exception of the number of trap nights at each camera location. To standardize sampling effort and trap spacing, we selected the data from the first 40 trap nights at each camera location and removed overlapping trap locations with spacing <2 km.
In total, our sample includes 221 camera-trap locations (**Figure 1**) and 8,840 trap nights, which resulted in 51,036 photographs and 17,012 occasions. We detected a total of 102 species (46 mammals, 48 birds, 7 reptiles, 1 amphibian), of which 17 species were selected for inclusion in our analysis based on their status as either a game species targeted by local hunters, a terrestrial predator, or as a species of ecological importance/conservation interest.
### Hunting Intensity Analysis
The boundaries of each site (**Figure 1**) were determined by the combination of (a) the locations of hunting areas identified by community members, (b) the number of people and number of villages using each hunting area, (c) the habitat type associated with hunting areas identified, and (d) the availability of overlapping camera trap data. Of 21 communities surveyed, data from 14 communities (7 of the 11 titled villages) were selected for inclusion in this study based on their overlap with camera-trap surveys. In cases of overlapping hunting sites between multiple communities, each combination of communities using a given area was recorded. Hunting intensity for each site, defined as the number of households hunting in a particular area (HH/km<sup>2</sup> ), was determined by calculating the number of households from each community hunting in a given site, divided by the total area of that polygon.
We calculated the number of households by taking the percentage of respondents from the KAP surveys that hunted and extrapolating it to the total number of households known to occur in that village. For example, if the KAP surveys showed 10% of respondents said they hunted, and there was a total of 80 households in that village, we would calculate that eight households hunted. In this manner, the sum of all the households hunting would be combined to give an estimate of the total number of households hunting in a given site. Although we recognize that these are not an accurate reflection of the actual intensities for each site, calculating in a consistent and standardized manner across sites allows us to make inferences about their intensities relative to one another. Sites with the highest intensity value (HH/km<sup>2</sup> ) were considered as the most intensely hunted sites.
GPS locations of hunting areas were used in determining the boundaries of each site, but intellectual property agreements with partner communities prohibits specific locations from being shown here. Variation in the overall size of each polygon (due to some communities hunting across larger areas than others) was standardized by considering camera trap density (traps/km<sup>2</sup> ) at each site (**Table 1**). After excluding sites with insufficient cameratrap survey effort (<20 trap locations—Sites 3 and 5), seven sample sites were identified (**Figure 1**)—four in mixed lowland and upland tropical forest (largely within the protected area), and three within the savanna-forest mosaic (largely outside the protected area).
### Camera-Trap Data Analysis
Occupancy, relative abundance, and activity patterns were calculated for each species of interest at each camera location to account for the influence of hunting intensity (as calculated above) on spatial and temporal distribution. To account for additional variation in detection probability on occupancy, we implemented season (rainy or dry) and trail type (anthropogenic or natural) as covariates in our model. Forest cover was determined by placing a 1-km buffer around each camera-trap and calculating the percent forest cover (Hansen et al., 2013) within each buffer using the Tabulate Area function. Distance to village was determined by calculating the Euclidean distance (in meters) from each camera-trap to the nearest village. Season was determined by the date range during which each camera was active and trail type was determined visually at each camera location. Cameras were placed into 15 spatial groups using the Grouping Analysis tool to control for spatial autocorrelation. We chose the K-nearest neighbors method and used trial and error to determine the optimum number of nearest neighbors (K = 20). All spatial analyses were conducted in ArcMap 10.3.1 (Environmental Systems Research Institute (ESRI), 2015).
We used a Bayesian multi-species hierarchical occupancy modeling approach (Dorazio and Royle, 2005) to assess the relationship between occupancy and hunting pressure, distance to the closest village, and percent forest cover. This approach enabled us to estimate species-specific and aggregate community occupancy while accounting for imperfect detection (MacKenzie et al., 2002; Dorazio et al., 2011). We recorded a binary measure of detection (1 = observed, 0 = not observed) for each species at each camera location for each 24 h period from when the cameras were deployed (n = 40). The number of trap nights available for each camera site varied, but subsetting the data to include the first 40 trap nights from each site allowed us to maximize the number of camera-trap sites included in our sample, while managing reasonable computation time. We investigated variables that we believed a priori would influence species-specific occurrence or detection using a generalized linear mixed modeling (GLMM) approach (Dorazio and Royle, 2005; Russell et al., 2009). We modeled the relationship of season (0 = rainy, 1 = dry) and trail type (0 = anthropogenic, 1 = natural) with species-specific and community detection probability. We modeled the relationship of hunting pressure, expressed as our hunting intensity index ranging from 0 to 1, distance to the closest village, and the percent of forest cover within a 1 km buffer around each camera, with species-specific and community occurrence probability. We standardized variables accounting for variability in detection and occurrence using a z-transformation and modeled them as random effects with species-level variation drawn from a common distribution inclusive of an estimated mean and variance (i.e., hyperparameters). Additionally, we modeled discrete spatial clusters of cameras as a random effect to account for spatial autocorrelation. We fitted a single full model inclusive of all a priori variables (Zipkin et al., 2010).
We estimated the posterior distributions of each parameter using Markov chain Monte Carlo (MCMC) implemented in JAGS (version 4.2.0) using the R2Jags R package (Plummer, 2011). We used uniform (uninformative) priors (Gelman et al., 1995; Gilks et al., 1996) and generated 3 chains of 50,000 iterations with a burn-in of 10,000 iterations and a thinning rate of 50, yielding 3,000 samples. We then assessed convergence of MCMC chains with trace plots and the Gelman-Rubin diagnostic (Rhat), where values <1.1 indicated convergence (Gelman and Hill, 2007). We considered model covariates with 95% Bayesian credibility intervals (CRI) not inclusive of zero to be relevant predictors of occurrence or detection.
We used the package "overlap" in R to estimate the percent overlap in the activity patterns of species of interest across all sites, as well as within pair-wise comparisons between sites with the highest and lowest hunting intensities (Ridout and Linkie, 2009). Package "overlap" observes capture times as random samples from a continuous distribution, and the "coefficient of overlap" as a non-parametric measurement of the overlap between the probability distribution functions of these underlying distributions estimated by bootstrapping (Ridout and Linkie, 2009). Avoidance of heat stress is a known factor influencing circadian rhythms of mammals in open habitats (Terrien et al., 2011), thus we excluded sites with <90% canopy cover (all savanna forest sites) from analysis of activity patterns to isolate effects related to hunting activity. Lastly, we calculated relative abundance indices (RAI's) of each species by dividing the number of occurrences of each species by the number of nights at each camera and standardizing for 100 trap nights (O'Brien, 2011). R code for each analysis can be found in **Appendix 1**.
### RESULTS
### Communities' Hunting Activity and Patterns
All communities surveyed, but not all households in each community, indicated that they hunt. On average, 25% (range = 0.02–0.61) of households across all communities surveyed hunt. Of the 21 communities, all hunt on their village lands and 18 out the 21 also hunt inside the KMPA. Typically, hunting occurs throughout the year, with small groups of hunters going out for a few days at a time. Hunting activities increase in frequency and intensity, and hunting parties increase in size, around key celebrations, such as Amerindian heritage month (September), Easter, and Christmas. During these celebrations, certain largebodied species, such as lowland tapir, are highly sought after for holiday meals.
Outside annual celebrations, two key hunting seasons were commonly identified across communities. The first is during the peak fruiting seasons (August and December), when animals are more accessible as they feed on fruits that drop to the forest floor. The second is during the peak dry season between January and March, when normally elusive animals gather around drying ponds, creeks, and springs, making them easier prey for hunters. Village residents mainly hunt for home or celebration (subsistence) purposes, but on average 62% of respondents across the 21 communities reported hunting for both home use and to generate income. Although hunting methods are highly variable depending on species and season, the most common method for hunting year-round across all villages was pursuit with dogs, bow and arrow, and/or firearms, with the level of offtake generally being consistent across hunters and villages.
Both KAP surveys and focus group discussions showed similar results in species hunted. The most commonly and frequently hunted species across all communities surveyed were (from most common to least) lowland paca, red brocket deer, redrumped agouti (Dasyprocta leporina), collared peccary (Pecari tajacu), black curassow, lowland tapir, white-tailed deer, great long-nosed armadillo (Dasypus kappleri), nine-banded armadillo (Dasypus novemcinctus), yellow-footed tortoise (Chelonoidis denticulatus), red-footed tortoise (Chelonoidis carbonarius), capybara (Hydrochoerus hydrochoeris), and spectacled caiman (Caiman crocodilus). Lowland paca, red-rumped agouti, collared peccary, red brocket deer, and lowland tapir were the top five preferred mammal species across our survey sites (**Table 2**). Correlation analysis showed a very strong positive relationship between species most frequently hunted and those perceived to be less abundant than they were 10 years ago (r <sup>2</sup> = 0.89; **Figure 2**).
### Hunting Intensity
Hunting intensity varied across sites, ranging from very low and low hunting intensity (Site 2 and Site 8) to higher intensity (Site 4 and Site 7) (**Table 1**). We found no correlation in any of the 21 communities surveyed between the proportion of households hunting in a village and the distance to the nearest market town, the nearest road, the protected area, or the number of shops available in their community. However, hunting intensities tended to be slightly higher in savanna forest habitats, with all sites experiencing medium and high hunting intensity compared to forested habitats where hunting intensity ranged from very low to high (**Table 1**).
### Occurrence and Detection Probability
Our Bayesian multi-species hierarchical occupancy model indicated that although cumulative community detection probability was positively associated with dry season (β 0.07; CRI −0.17 to 0.33; Rhat 1.00) and natural (game) trails (β 0.05; CRI −0.08 to 0.18; Rhat 1.00), neither were a statistically significant TABLE 2 | The top three most hunted species at each site.
predictor of detection. Similarly, we also found that cumulative community occurrence was positively (but not significantly) associated (did not exhibit a clear directional association) with hunting intensity (β 0.05; CRI −0.30 to 0.41; Rhat 1.00), and distance to nearest village (β 0.06; CRI −0.35 to 0.47; Rhat 1.00). Percent forest cover was found to be a significant predictor of community occurrence (β 0.56; CRI 0.18 to 0.95; Rhat 1.00), as occupancy increased with forest cover for most species included in our analyses.
Gelman-Ruben diagnostic (Rhat) results showed that all individual species models converged (Rhat < 1), implying confidence in our estimates (**Table 3**). Trail type was not a good predictor of detection for any of the 17 species included in our analyses. Conversely, season proved a significant predictor of detection for several species, with lowland tapir (β 0.59; CRI 0.33–0.86), jaguar (Panthera onca) (β 0.90; CRI 0.40 to 1.44), and ocelot (Leopardus pardalis) (β 0.48; CRI 0.13 to 0.86) showing a positive correlation with detection in the dry season, while collared peccary (β −0.41; CRI −0.78 to −0.04), lowland paca (β −0.30; CRI −0.48 to −0.11), red acouchi (Myoprocta acouchy) (β −0.33; CRI −0.50 to −0.18), and gray-winged trumpeter (Psophia crepitans) (β −0.36; CRI −0.53 to −0.19) all had higher detection probability in the rainy season.
Forest cover had the most significant relationship with species occurrence, with 11 out of the 17 species showing a significant relationship with this variable (**Appendix II**). Increased forest cover was positively associated with occurrence of the red acouchi (β 2.16; CRI 1.30 to 3.21), black curassow (β 0.68; CRI 0.28 to 1.10), collared peccary (β 1.02; CRI 0.46 to 1.63), Amazonian brown brocket deer (Mazama nemorivaga) (β 1.70; CRI 0.82 to 2.75), gray-winged trumpeter (β 2.10; CRI 1.36 to 2.97), lowland paca (β 0.50; CRI 0.14 to 0.89), red brocket deer (β 0.72; CRI 0.31 to 1.16), and red-rumped agouti (β 0.71; CRI 0.33 to 1.09), and a significant negative association with the occurrence of crab-eating fox (Cerdocyon thous) (β −1.12; CRI −1.56 to −0.69), giant anteater (Myrmecophaga tridactlya) (β −0.77; CRI −1.65 to −0.21), and white-tailed deer (β −0.61; CRI −1.34 to −0.04). Lowland tapir (β 0.52; CRI 0.10–0.95) and both species of long-nosed armadillos (β 0.44; CRI 0.04 to 0.87) demonstrated a significant positive correlation with increased distance from the nearest village, while white-tailed deer (β −1.13; CRI −2.23 to −0.15) and crab-eating fox (β −1.15; CRI −2.17 to −0.17) exhibited significant negative relationships (**Appendix II**). Hunting intensity did not appear to have a significant impact on species occurrence for any of our 17 species (**Appendix II**), with the exception of the crab-eating fox (β 0.82; TABLE 3 | Species-specific summaries of covariate effects on occupancy (psi) and detection (p) for 18 species of interest.
*(Continued)*
β*, standard deviation (SD), and Bayesian credibility interval (CRI) are based on the model averaged posterior distribution. We considered model covariates to be relevant predictors of species occurrence or detection when 95% CRI's did not cross zero and assumed Rhat values of* <*1.1 indicated convergence.*
CRI 0.19 to 1.48) which showed a significant positive relationship with hunting intensity (**Table 3**).
### Relative Abundance Index
Given the influence of forest cover and season in predicting species occurrence, we compared the RAI of species between lower and higher hunting intensity sites post analysis, when habitat type and season were similar. In this manner, we compared Site 2 (very low intensity) with Site 4 (high intensity), both in tropical forest habitat sampled during the rainy season, as well as Site 6 (medium intensity) and Site 7 (high intensity), which represented savanna-forest mosaic habitat sampled in the dry season (**Table 1**). In forested habitats in the rainy season, and in savanna-forest mosaic in the dry season, nearly all species showed differences in their RAI's between low and high intensity sites (**Table 4**), however none of these differences proved statistically significant (Wilcoxon's test: forest: z = −0.37, n = 15, p = 0.71; savanna: z = −1.48, n = 15, p = 0.14). Lowland tapir, all three species of deer, red-rumped agouti, puma (Puma concolor), and ocelot consistently showed higher RAI's in sites that had lower hunting intensity levels regardless of habitat type, while lowland paca, black curassow, gray-winged trumpeter, and jaguar showed the same trend within forested habitats only. Conversely, collared peccary and red acouchi consistently showed higher RAI's in both higher hunting intensity sites, while lowland paca, black curassow, gray-winged trumpeter, crab-eating fox, and jaguar showed this trend in savanna-forest habitat only.
TABLE 4 | Comparison of relative abundance indices (RAI) of species of interest between sites with "low" and "high" hunting intensity, keeping habitat type and season constant.
\*\**, species preferred by hunters,* \**, species hunted, but not preferred, no asterisks, species not targeted by hunters.*
### Activity Patterns
Although we plotted the activity patterns of 13 of the species included in the analyses above (**Appendix III**), we applied more detailed analyses to the top five hunted species as indicated by our KAP surveys, as well as two non-hunted species—red acouchi and ocelot. Plots of general activity patterns showed that red-rumped agouti and collared peccary were primarily diurnal, lowland paca and both long-nosed armadillos exclusively nocturnal, lowland tapir were mostly nocturnal with some activity during the day, and red brocket deer were active both day and night (**Figure 3**). In non-hunted species, the red acouchi were crepuscular and ocelots were active both day and night with peak activities occurring at dawn, dusk, and midday.
In pair-wise comparisons of activity patterns between the lowest (Site 2) and highest intensity hunting sites (Site 4) where both habitat and season were constant, we observed shifts in activity patterns consistent with temporal avoidance of human (hunting) activity. Large-bodied game species that are targeted by hunters showed a decrease in diurnal activity at the more intensively hunted site (**Figure 3**), with activity patterns of lowland tapir, red brocket deer, and collared peccary shifting by 18.1, 19.8, and 20% respectively (**Table 5**). Medium-sized game animals showed more subtle shifts in activity patterns (**Figure 3**), with lowland paca (10.9%), red-rumped agouti (10%), and the long-nosed armadillos (10.1%) each shifting away from peak hunting times early in the morning (**Table 5**). Pair-wise comparisons of the overlap of activity patterns of medium-sized non-hunted species showed much larger shifts in activity, with red acouchi and ocelot shifting toward increased activity in the morning by 29 and 27%, respectively.
TABLE 5 | Percent overlap and overall trends in activity patterns of species of interest between Site 2 (very low hunting intensity) and Site 4 (high hunting intensity).
### DISCUSSION
As the demand for wild animal meat continues to increase, maintaining healthy populations of game mammals and birds represents a significant challenge facing conservation managers around the world. Understanding shifts in hunting preferences, patterns, and intensities, as well as the potential negative impacts of hunting, is critical for policy makers who are responsible for setting rules and regulations related to hunting. This is particularly true in regions that allow for the continued use of resources within conservation areas by indigenous communities, as is typically the case in the Neotropics. Boasting one of the world's lowest deforestation
rates and human population densities, Guyana's interior still hosts healthy populations of many globally and locally important species. However, indigenous communities surrounding the Kanuku Mountains Protected Area, one of the country's most biodiversity rich protected areas with a human population density in adjacent communities of only 0.42/km<sup>2</sup> , are reporting declines, and even local extinctions, of some preferred game species.
Hunting is typically considered sustainable when the number of individuals removed from the population is the same as the number added through normal population growth (Clark, 1990; Milner-Gulland et al., 2009). Our results indicate that current levels of hunting in and around the KMPA may currently be sustainable, however small shifts detected in the behavior and distribution of species preferred by hunters match the perceived changes observed by KMPA's indigenous communities, suggesting that species under elevated hunting pressure may be changing their behavior to evade predation by humans.
### Species Occurrence
Our results showed that hunting intensity, distance to village, forest cover, and season all serve as significant predictors of species occurrence and detectability, depending on the species. Interestingly, trail type did not have a significant relationship with detection for any of the species included in our model, despite previous research indicating that this was an important predictor for some species (Harmsen et al., 2010). We believe that this contradictory result can be attributed to the fact that none of the camera-traps in our sample were set along roads, and that the human-made trails that were utilized for camera-trapping were only lightly used.
Of the variables examined, forest cover was the best predictor of occurrence across species, showing a significant relationship with 11 out of the 17 species examined. Red and Amazonian brown brocket deer, collared peccary, lowland paca, red-rumped agouti, red acouchi, black curassow, and gray-winged trumpeter (includes four out of the top five species preferred by indigenous communities surrounding the KMPA), all showed a significant positive association with increased forest cover. On the other hand, white-tailed deer, giant anteater, and crab-eating fox showed a significant negative association with percent forest cover. These results are not surprising and reflect the general habitat preference typical for each species. White-tailed deer, giant anteater, and crab-eating fox are considered savannadwelling species (with giant anteater also inhabiting tropical forest but in lower densities), while the remaining species listed above are typically associated with dense tropical forests.
Forest cover was not a significant predictor of occurrence for lowland tapir and the long-nosed armadillos, however, suggesting that these species may use the savanna-forest mosaic more readily than other "forest-dwelling" game species. This result is not unexpected, as both species are known inhabitants of savanna, grassland, and shrubland habitats of the Cerrado, Chaco, Llanos, and Pantanal regions of South America (Loughry et al., 2014; Varela et al., 2019). Similarly, forest cover was not a good predictor of occurrence for jaguars, pumas, or ocelots, perhaps unsurprisingly as the savanna-forest mosaic hosts sufficient prey (including domestic animals) and the distribution and movement of these species is well-documented to be a function of prey availability (Weckel et al., 2006). Season was also a key variable in detection probability for many species, with lowland tapir and jaguar more detectable in the dry season, while probability of detection for collared peccary, lowland paca, red acouchi, and the long-nosed armadillos was higher in the rainy season. As with our interpretation of species relationships with forest cover, we suggest that these findings are more likely attributed to behavior and responses to the availability of water and prey/food, than for any other reason.
While not associated with forest cover, occurrence of both lowland tapir and the long-nosed armadillos showed a significant positive correlation with increased distance to villages. Studies show that distance to village is an important predictor for other elusive species, particularly in areas where hunting pressure is high (Phan et al., 2019). Lowland tapirs, in particular, are known to prefer less disturbed areas further away from human presence (Tobler, 2002; Licona et al., 2011). Generally, hunters use areas close to settlements more frequently than more distant areas (De Souza-Mazurek et al., 2000) and the communities in our study are no exception, with all 21 villages hunting more frequently in the savanna-forest mosaic close to villages (Protected Areas Commission unpublished data). KAP surveys and mapping exercises also showed that hunting intensities were higher in savanna-forest mosaic habitat near to villages, compared to forested habitats which are typically further from villages and closer to or within the KMPA (**Table 2**). So, while tapir and armadillos more readily utilize savanna forest habitats than other "forest-dwelling" species, distance from human settlement seems to be a more influential factor in determining the distribution of these species. Interestingly, lowland tapir and the long-nosed armadillos were ranked as the fifth and sixth most preferred forest game mammal species across the 21 communities surveyed. Both were also frequently cited as a species that communities were most worried about with regard to their availability in the future (Protected Areas Commission unpublished data), indicating that even relatively low hunting pressure may affect the distribution of species that are particularly sensitive to human activity.
Our study generally supports Robinson and Bennett (2000) hypothesis that hunting intensity <1 hunter/km<sup>2</sup> is sustainable. Though a number of hunted species showed a negative association, hunting intensity was not a significant predictor of occurrence for any of the species that we examined, with the exception of the crab-eating fox which showed a significant positive relationship with hunting intensity. The crab-eating fox is not hunted within our study area, and we suggest that this relationship is likely the result of its preference for scavenging, as well as the reduction in niche occupancy of competing species. Although omnivorous, studies have shown that fruit is the most prevalent item in the crab-eating fox's diet (Rocha et al., 2008), and thus there is likely significant dietary overlap with a number of the species that were identified as hunted by the KAP surveys. Occurrence of crab-eating fox also showed a significant negative relationship with increased distance to villages, which is not surprising as crab-eating fox frequently occur in humandominated areas where they are known to scavenge (Rocha et al., 2008) and likely feed on fruits from cultivated trees often found in association with homesteads (mango, cashew, citrus).
### Relative Abundance
Although not statistically significant, our results show that the RAI's of nearly all hunted species were higher in sites with low hunting intensity when compared to those with higher hunting intensity, with habitat (tropical forest) and season (rainy) kept constant. These trends are synonymous with what would be expected if hunting is having an impact on species, in so far as there are higher RAI's in the low hunting intensity site for species targeted by hunters, with the largest differences in those species that have a greater predisposition to overhunting—i.e., lowland tapir and deer (>40% increase in RAI in low hunting intensity sites) (Bodmer, 1989; Cardillo et al., 2005; Fa and Brown, 2009; De Thoisy et al., 2010). Smaller-bodied, more prolific breeding species such as lowland paca and red-rumped agouti, species known to be somewhat resistant to the effects of harvesting (Mayor et al., 2007), showed much smaller differences (∼25% increase in RAI in low hunting intensity sites) in RAI between sites.
For species whose movements and distribution are largely determined by prey availability, such as jaguar and puma (Weckel et al., 2006), trends in RAI mirrored those for most potential prey species. Conversely, we found that collared peccary and red acouchi showed higher RAI's in more intensely hunted sites. We suggest that this may be a result of these species occupying niches that have become available as competition for resources is reduced (Brown and Davidson, 1977). Previous studies (Fragoso et al., 2016) and anecdotal reports indicate that white-lipped peccaries were once relatively abundant across our study area. The lack of white-lipped peccary occurrences in our data suggests that the rapid, unexplained population declines that have transpired across the range of this species (Altrichter et al., 2012) may have also occurred in and around the KMPA. The disappearance of a superior competitor with which there is a high degree of niche overlap (Desbiez et al., 2009), in combination with their generalist nature and high reproductive rates, may explain increased RAI observed in collared peccaries. In rodents, this effect tends to be more prevalent in specialized species (Brown and Davidson, 1977). Red acouchi diets overlap with lowland paca and red-rumped agouti but comprises only about 50% of the diversity of their superior competitors filling the same niche (Dubost and Henry, 2006). As expected, in species such as ocelot, which is not hunted nor has much dietary overlap with other species included in our analysis, we saw little difference in RAI's between low and high intensity hunting sites.
In savanna-forest habitat in the dry season, while we see similar patterns in lowland tapir and jaguar, the potential impacts of hunting on RAI's are less pronounced. We suggest therefore, that additional confounding variables are likely at play. Firstly, the difference in hunting intensity between the two sites compared is relatively small (0.11 compared to >0.42 in forest habitats). Secondly, the site with the highest hunting intensity in our study, Site 7, is in close proximity to a major river, while Site 6, a site with medium hunting intensity, is located far from any major water source (**Figure 1**). We suspect that, particularly in the dry season, the presence of a permanent water source may have a greater influence in species' RAI's than the impacts of hunting. This hypothesis is further supported by the fact that we see a >500% increase in the RAI of lowland paca in Site 7—a species that is known to remain close to permanent water sources (Hutchins et al., 2003).
### Activity Patterns
As with RAI's, we compared the activity patterns of a variety of hunted and non-hunted species at sites with high and low hunting intensity, keeping habitat and season constant. We predicted that hunted species would shift their activity away from peak hunting times while unhunted species, such as the red acouchi, would shift its activities to avoid superior competitors (red-rumped agouti and lowland paca), and that ocelot would show no change in activity patterns. Ocelot are not hunted, there is little dietary overlap with large carnivores, and they have shown not to shift activity patterns even in areas of very high human disturbance (Kolowski and Alonso, 2010).
Activity patterns generally reflected those observed in other studies for large game species, as well as ocelot. Collared peccary were typically diurnal, lowland tapir primarily nocturnal, and red brocket deer (Tobler et al., 2009) and ocelot (Kolowski and Alonso, 2010) active day or night. Among medium-sized hunted species, red-rumped agouti were largely diurnal, following similar activity patterns to those found in previous studies (Dubost, 1988). However, we observed variation in the activity patterns of lowland paca and red acouchi when compared to those previously documented in the literature. Typically, lowland paca are largely crepuscular (Hutchins et al., 2003) and red acouchi predominantly diurnal (Dubost, 1988). However, our data showed lowland paca as exclusively nocturnal, a pattern typically observed in this species in heavily hunted areas (Hutchins et al., 2003). Additionally, our data showed that red acouchi were predominantly crepuscular, suggesting temporal avoidance of the diurnal red-rumped agouti and nocturnal lowland paca.
Shifts in circadian rhythms are considered important for species targeted by hunters, as they may alter reproductive success and survival (Gaynor et al., 2018). Activity pattern overlap analysis showed that large-bodied game species, like lowland tapir, red brocket deer, and collared peccary showed ∼20% shifts in activity patterns in the more intensely hunted site, with tapir and deer shifting toward increasingly nocturnal and collared peccary toward increasingly crepuscular activity. Although relatively small, these shifts away from peak hunting times are synonymous with temporal avoidance of human (hunting) activity (Kilgo et al., 1998). Red-rumped agouti and lowland paca also showed minor shifts away from peak hunting times (∼10%), with red-rumped agouti showing greater activity around dawn and lowland paca after dusk in hunted sites.
The red acouchi demonstrated a large shift in activity patterns (29%), with increased activity during the day in our most intensively hunted site. We infer that this result demonstrates a cascading shift in activity, with red acouchi shifting toward peak hunting times to avoid peak redrumped agouti activity around dawn. In this case, hunting activity may have reduced competition with red-rumped agouti and lowland paca, thus opening a temporal niche to this smaller-bodied, inferior competitor. Surprisingly, ocelots also showed a large shift in activity (30%) and appeared to reduce their activity in the middle of the day—similar to the shifts in activity observed in tapir and deer. Shift in ocelot activity may be a response to increased human activity, which would contradict previous studies that suggest ocelots are not impacted by human disturbance (Kolowski and Alonso, 2010). However, competition is also known to be an important influence over the structure of Neotropical carnivore communities (de Oliveira et al., 2010; de Oliveira and Pereira, 2014)—one that cannot be ruled out here and needs to be examined further.
### Perceptions vs. Camera-Trap Data
Data from the KAP surveys showed a strong positive correlation between the species most frequently hunted and those perceived by hunters to be less abundant than they were 10 years ago (**Figure 2**). Our camera-trap data supports these perceptions, with observed shifts in the behavior and distribution of species preferred by hunters that are consistent with what would be expected if species were demonstrating avoidance behavior in hunted areas. Indigenous knowledge (IK) is accumulated over a lifetime of traditional use of species (Gilchrist et al., 2005), and has long been recognized as an important and reliable source of information when used in modeling the abundance and distribution of species (Walters and Holling, 1990; Zabel et al., 2002; Anadón et al., 2009). We suggest, therefore, that a combination of indigenous knowledge and participation, as well as quantifiable data should be applied to informed decisionmaking on species management and the identification of levels of sustainable hunting for species of interest.
### CONCLUSION
Most previous studies on hunting in the tropics have focused on areas where levels of hunting are already unsustainable (Wilkie et al., 2011), therefore limiting our understanding of the level of harvest that is sustainable for a number of species of conservation concern. In lowland tapirs, it is well-documented that hunting is generally unsustainable even at very low levels as a result of their low population growth rate and density (Tobler et al., 2014). Although our study showed that current hunting levels are not having a statistically significant impact on the occurrence of species preferred by hunters, and this result supports previous studies that conclude that hunting intensity of <1 person/km<sup>2</sup> is sustainable, we treat this conclusion with caution. Our models did detect negative relationships with hunting intensity in a number of the preferred game species (lowland paca, collared peccary, lowland tapir, giant armadillo) in our study area. We view this result as an early warning for the Kanuku Mountains region and, considering the low overall hunting intensity at our site when compared to much of the rest of the Neotropics, an indication of the truly low levels of harvest that some species can sustain before populations begin to show declines. Once a significant relationship is found between species occurrence and hunting intensity, its likely hunting has already exceeded sustainable levels, and thus provides little insight for managers and policy makers tasked with preventing or reversing population decline.
Fa et al. (2002) suggests that harvest is no longer sustainable when species exploitation is ≥20% of its productivity. Sustainability therefore is largely attributed to balancing individual removal with population recovery (Clark, 1990) and only accounts for impacts on the populations of hunted species. Our study shows that even with relatively low intensity hunting, shifts in the distribution and behavior of hunted species can trigger cascading effects on non-hunted species, which in turn could have an impact on ecosystems as a whole. Our results support the notion, therefore, that levels of sustainable hunting should not only consider population trends in hunted species but should also include the structure and function of the community as a whole (Milner-Gulland et al., 2009).
We suggest that determining occupancy, relative abundance, and activity patterns of hunted and non-hunted species, as well as incorporating indigenous knowledge of trends in use and populations over time, would be a more effective early alarm system than monitoring the occurrence of hunted species alone. Further, as the human population in the Rupununi region and market demand for wild meat outside of the region grows, and infrastructure developments make the region more accessible to outsiders, our data can serve as a baseline that can be monitored over time. Monitoring can be done alongside surveys of changes in hunting practices, thus allowing for preventative measures to be adopted by managers and indigenous communities to ensure that the level of hunting in the region remains sustainable into the future.
### ETHICS STATEMENT
Research methodologies received approval from the IACUC and IRB offices at Miami University (#870, #E00547) and the University of Florida (#201408378, #2015-U-400). Methodologies were also reviewed and approved by the Frankfurt Zoological Society, the Protected Areas Commission, and the Ministry of Indigenous People's Affairs (MoIPA) prior to obtaining a research permit from the Environmental Protection Agency (#011315 BR 002) in Guyana. Permission to conduct research within indigenous titled lands was only granted by MoIPA following the submission of individual letters of support signed by the councils of each village who chose to participate in this project.
### AUTHOR CONTRIBUTIONS
MH determined study design, primary facilitator of cameratrap data collection, and co-primary author and editor of paper. AK determined study design, facilitated KAP surveys and community mapping, co-facilitated implementation of cameratrap data collection in some sites, and co-primary author and editor of paper. RM was responsible for KAP survey data management, facilitated analysis of KAP survey data, and contributed to writing and paper editing. TBag informed design and facilitated implementation of KAP surveys and community mapping, co-facilitated implementation of cameratrap data collection, and contributed to writing and paper editing. TBab facilitated spatial analysis of KAP survey data, GIS, and map making. HB contributed to design and implementation of camera-trap data collection, coordinated research teams, contributed to camera-trap data management, and reviewed and provided feedback to paper draft. AW contributed to design and implementation of camera-trap data collection, coordinated research teams, contributed to camera-trap data management, and reviewed and provided feedback to paper draft. FL was a official local collaborator, contributed to design and implementation of camera-trap data collection, coordinated research teams, contributed to camera-trap data management, and reviewed and provided feedback to paper draft. WB was responsible for writing code and running updated Bayesian models, wrote portions of the methods and results sections pertaining to these models, and worked with the team to interpret results. BB contributed to design and implementation of cameratrap data analysis, writing up of statistical methodology, R coding, production of figures, and contributed to writing and paper editing.
### FUNDING
Funding for this project provided through between the Protected Areas Commission, Frankfurt Zoological Society, and Jacksonville Zoo & Gardens, as well as the Ron Magill Conservation Award (2015), Phoenix Zoo Field Conservation and Science Grant (2013), and Field Research Grants from Project Dragonfly (2012), the Institute for the Environment and Sustainability (2012) and the Graduate School (2013) at Miami University, and the Tropical Conservation and Development Program (2014), School of Natural Resources and Environment (2016) and Department of Wildlife Ecology and Conservation (2016) at the University of Florida.
### ACKNOWLEDGMENTS
The authors are grateful to the whole team who made the collection of data possible for this study, particularly PAC staff,
### REFERENCES
Francisco Gomes, Jessica George, Octavious Hendricks, Jenkins Lawrence, Vercida Gomes, Norbert Atkinson, Sara Henry, and Tana Yussuff and the community-based team from the Rupununi Wildlife Research Unit—Ashley Holland, Ramlal Roberts, Brian Duncan, Jose George, Bernard Lawrence, Herman Phillips, Manuel Mandook, Anthony Roberts, Gerry Bell, Henry Clement, Nicholas Mandook, Magnus David, Flavian Thomas, Micah Davis, Martin Carter, Angus Jackman, Rhomayne Li, and Kyle Hallett. We are grateful to the Guyana Environmental Protection Agency, Ministry of Indigenous People's Affairs, Protected Areas Commission and the Kanuku Mountains Community Representative Group for issuing permission and permits to carry out this study. Lastly, we ever grateful to the leadership and residents of Katoka (Simonie), Maruranau, Moco Moco, Nappi (Parishara, Hiowa), Parikwaranau, Rupunau, Sand Creek, Shulinab (Meriwau, Quiko), Shea, St. Ignatius (Kumu, Quarrie), and Yupukari (Fly Hill, Quatatta, Kaicumbay) villages for their overwhelming kindness, hospitality, generosity, cooperation, support, and participation, without which none of this study could have been carried out.
### SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fevo. 2019.00412/full#supplementary-material
brocket deer (Mazama americana and M. nana). Biotropica 40, 636–645. doi: 10.1111/j.1744-7429.2008.00413.x
**Conflict of Interest:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Copyright © 2019 Hallett, Kinahan, McGregor, Baggallay, Babb, Barnabus, Wilson, Li, Boone and Bankovich. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. | doab | 2025-04-07T04:13:03.621753 | 18-11-2021 17:38 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffc27b79-9210-4e73-a820-6f98a7ec95e7",
"url": "https://www.frontiersin.org/research-topics/7120/shifting-the-paradigms-for-sustainable-wildmeat-use-in-tropical-and-sub-tropical-regions",
"author": "",
"title": "Shifting the Paradigms for Sustainable Wildmeat Use in Tropical and Sub-Tropical Regions",
"publisher": "Frontiers Media SA",
"isbn": "9782889662388",
"section_idx": 0
} |
ffc3321d-1b71-44dd-8de0-4fc5781eefa7 | ## Open Book Classics
Friedrich Hölderlin
*Hölderlin's only novel, Hyperion, is sti ll not widely enough known in the English-speaking world despite its unquesti onable importance. Gaskill's version is the fruit of long years of loving att enti on, and it catches much of the beauty and texture of the original — its rhythms, dicti on,*
*This is a superb version of Hölderlin's novel: accurate, elegant and inspiring. It at long last brings*
Friedrich Hölderlin's only novel, *Hyperion* (1797–99), is a fi c� onal epistolary autobiography that juxtaposes narra� on with cri� cal refl ec� on. Returning to Greece a� er German exile, following his part in the abor� ve uprising against the occupying Turks (1770), and his failure as both a lover and a revolu� onary, Hyperion assumes a hermi� c existence, during which he writes his le� ers. Confron� ng and commen� ng on his own past, with all its joy and grief, the narrator undergoes a transforma� on that culminates in the realisa� on of his true voca� on.
Though Hölderlin is now established as a great lyric poet, recogni� on of his novel as a supreme achievement of European Roman� cism has been belated in the Anglophone world. Incorpora� ng the aesthe� c evangelism that is a characteris� c feature of the age, *Hyperion* preaches a message of redemp� on through beauty. The resolu� on of the contradic� ons and an� nomies raised in the novel is found in the act of ar� cula� on itself. To a degree remarkable in a prose work of any length, what it means is inseparable from how it means. In this skilful transla� on, Gaskill conveys the beau� ful music and rhythms of Hölderlin's language to an
As with all Open Book publica� ons, this en� re book is available to read for free on the publisher's website. Printed and digital edi� ons, together with supplementary digital material,
*Cover image: Rudolf Lohbauer, 'Hyperions Fahrt nach Kalaurea' (1824). Cover design: Francesca Alabaster.*
*this major text of German literature within the reach of the Anglophone public.*
—Prof. Charlie Louth, University of Oxford
—Prof. Jeremy D. Adler, King's College London
*variety and shift s of style.*
Friedrich Hölderlin
**Hyperion,**
Translated by Howard Gaskill
or the Hermit in Greece
English-speaking reader.
e
**OPEN ACCESS**
book ebook and OA edi� ons also available
can also be found at www.openbookpublishers.com
**Hyperion**
**OBP**
# **Hyperion,** or the Hermit in Greece
FRIEDRICH HÖLDERLIN TRANSLATED BY HOWARD GASKILL
## To access digital resources including: blog posts videos online appendices
and to purchase copies of this book in: hardback paperback ebook editions
Go to: **https://www.openbookpublishers.com/product/941**
Open Book Publishers is a non-profit independent initiative. We rely on sales and donations to continue publishing high-quality academic works.
# HYPERION, OR THE HERMIT IN GREECE
# Hyperion, or the Hermit in Greece
*By Friedrich Hölderlin*
*Translated and with an Afterword by Howard Gaskill*
https://www.openbookpublishers.com
Translation and Afterword © 2019 Howard Gaskill
This work is licensed under a Creative Commons Attribution 4.0 International license (CC BY 4.0). This license allows you to share, copy, distribute and transmit the text; to adapt the text and to make commercial use of the text providing attribution is made to the authors (but not in any way that suggests that they endorse you or your use of the work). Attribution should include the following information:
Friedrich Hölderlin *Hyperion, or the Hermit in Greece*. Translated and with an Afterword by Howard Gaskill. Cambridge, UK: Open Book Publishers, 2019, https://doi.org/10.11647/ OBP.0160
In order to access detailed and updated information on the license, please visit, https:// www.openbookpublishers.com/product/941#copyright
Further details about CC BY licenses are available at, https://creativecommons.org/ licenses/by/4.0/
All external links were active at the time of publication unless otherwise stated and have been archived via the Internet Archive Wayback Machine at, https://archive.org/web
Updated digital material and resources associated with this volume are available at, https://www.openbookpublishers.com/product/941#resources
Every effort has been made to identify and contact copyright holders and any omission or error will be corrected if notification is made to the publisher.
Open Book Classics Series, vol. 10 | ISSN: 2054-216X (Print); 2054-2178 (Online)
ISBN Paperback: 978-1-78374-655-2 ISBN Hardback: 978-1-78374-656-9 ISBN Digital (PDF): 978-1-78374-657-6 ISBN Digital ebook (epub): 978-1-78374-658-3 ISBN Digital ebook (mobi): 978-1-78374-659-0 ISBN Digital (XML): 978-1-78374-667-5 DOI: 10.11647/OBP.0160
Cover image: Rudolf Lohbauer, 'Hyperions Fahrt nach Kalaurea' [Hyperion's voyage to Calauria] (1824), reproduced by kind permission of the owner. See Walter Ludwig, 'Rudolf Lohbauers Bild "Hyperions Fahrt nach Kalaurea"', *Hölderlin-Jahrbuch* 30 (1996– 1997), pp. 359–80.
Cover design: Francesca Alabaster.
All paper used by Open Book Publishers is SFI (Sustainable Forestry Initiative), PEFC (Programme for the Endorsement of Forest Certification Schemes) and Forest Stewardship Council® (FSC® certified).
## Contents
Map of Greece and Asia Minor, with locations mentioned in *Hyperion*, CC BY 4.0.
# *Hyperion*, *or the Hermit in Greece*
## Volume One
*Non coerceri maximo*, *contineri minimo*, *divinum est.*
[Not to be constrained by the greatest, to be contained by the smallest, is divine.]
I'd happily promise this book the love of the Germans. But I fear some will read it like a compendium and be overly concerned with the *fabula docet*, 1 whilst others will take it too lightly, and neither party will understand it.
Those who merely sniff my flower mistake its nature, and so do those who pluck it merely for instruction.
The resolution of the dissonances in a particular character is neither for mere reflection nor empty pleasure.
The setting for what follows is not new, and I confess that I was once childish enough to try changing the book in this respect. But I became convinced that this setting was the only appropriate one for Hyperion's elegiac character, and was ashamed at my weakness in having been unduly swayed by the probable verdict of the public.
I regret that for now it is not yet possible for everyone to judge the design of the work. But the second volume is to follow as soon as possible.
<sup>1 [</sup>Translator's note] what the fable teaches, the moral of the story.
## Hyperion to Bellarmin [I]
The beloved soil of my fatherland gives me joy and grief once more.
I'm up now every morning on the heights of the Isthmus of Corinth, and often, like the bee from flower to flower, my spirit flits back and forth between the seas to right and left that cool the feet of my glowing mountains.
One of those two gulfs would specially have delighted me, had I stood here some thousand years ago.
Then, like a conquering demi-god between the glorious wilderness of Helicon and Parnassus, where the rosy light of dawn plays on a hundred snow-covered peaks, and the paradisal plain of Sicyon, the shining gulf surged in towards the city of joy, youthful Corinth, pouring forth before its favourite the accumulated bounty from all corners of the earth.
But what is that to me? The howl of the jackal, singing its wild dirge amidst the rubble of antiquity, jolts me from my dreams.
Happy the man for whom a flourishing fatherland gladdens and fortifies the heart! Being reminded of mine is like being pitched into the mire, like having the coffin lid slammed shut over me, and whenever anyone calls me Greek, I always feel I'm being throttled with a dog collar.
And see, my Bellarmin! whenever I'd burst out with such remarks, as often as not with tears of anger in my eyes, along came the wise gentlemen who so delight in gibbering among you Germans, those wretches for whom a grieving disposition is such a welcome opportunity to unload their maxims; they were in their element, and made so bold to tell me: 'Don't moan, act!'
Oh, that I had never acted! how many hopes I'd now be richer by! —
Yes, just forget that men exist, starving, vexed and deeply harassed heart! and return whence you came, into the arms of nature, neverchanging, beautiful and tranquil.
#### Hyperion to Bellarmin [II]
I have nothing I might truly call my own.
Far away and dead are those I loved, and through no voice I hear from them, nothing ever more.
My business on earth is done. I set about my work with a will, bled over it, and made the world not a penny richer.
I return alone and unrenowned and wander through my fatherland, stretching about me like a vast graveyard, and it may be that what awaits me is the knife of the hunter who keeps us Greeks for sport like forest game.
But you still shine, sun of heaven! You still green, holy earth! Still the rivers rush into the sea, and shady trees whisper in the height of day. Spring's blissful song sings my mortal thoughts to sleep. The plenitude of the all-living world nourishes and fills with drunkenness my starving spirit.
O blissful nature! I can't tell what comes over me when I lift up my eyes before your beauty, but all the joy of heaven is in the tears I weep before you, the lover before the beloved.
My whole being stills and listens when the gentle ripple of the breeze plays about my breast. Often, lost in the immensity of blue, I look up into the aether and out into the hallowed sea, and it's as if a kindred spirit opened its arms to me, as if the pain of isolation were dissolved in the life of the godhead.
To be one with everything, that is the life of the godhead, that is the heaven of man.
To be one with everything that lives, to return in blissful self-oblivion into the all of nature, that is the summit of thoughts and joys, that is the holy mountain pinnacle, the place of eternal peace where noon loses its sultriness and the thunder its voice and the boiling sea becomes like a waving corn-field.
To be one with everything that lives! At these words virtue lays aside its wrathful harness, the mind of man its sceptre, and all thoughts melt away before the vision of the world's eternal oneness like the toiling artist's rules before his heavenly Urania, and iron fate renounces its dominion, and from the covenant of beings death disappears, and indivisibility and eternal youth blesses, makes beautiful the world.
On this height I often stand, my Bellarmin! But a moment of reflection casts me down. I begin to think, and find myself as I was before, alone, with all the pains of mortality, and my heart's sanctuary, the world's eternal oneness, is no more; nature's arms are closed, and I stand before her like a stranger and cannot comprehend her.
Oh! had I never gone into your schools. It's learning that lured me down into the pit, in my youthful folly I thought to find in it the proof of my pure joy, and it has ruined everything for me.
Amongst you I became so very rational, learnt to distinguish myself perfectly from what is around me, and now I'm set apart in the beautiful world, expelled from the garden of nature in which I grew and bloomed, and shrivel under the noonday sun.
Oh, man is a god when he dreams, a beggar when he thinks, and when inspiration's gone he's left standing there like a delinquent son, cast out of the house by his father, staring at the pitiful pennies given as alms to help him on his way.
#### Hyperion to Bellarmin [III]
I thank you for asking me to tell you about myself, for making me remember former times.
That's what really drove me back to Greece, wanting to live nearer to the playground of my youth.
As into quickening sleep the labourer, so my beleaguered being often sinks into the arms of the innocent past.
Peace of childhood! heavenly peace! how often do I stilly stand before you in loving contemplation, and think to grasp you! Yet we can only conceive of that which once was bad and has been made good again; of childhood, innocence we can have no conception.
When I was still a tranquil child, knowing nought of all that is around us, was I not then more than I am now, after all the heart's travail and all the mind's toiling and striving?
Yes! a divine being is the child as long as it's not been dipped in the chameleon colours of men.
It's wholly what it is and that's why it's so beautiful.
The force of law and fate can't touch it; in the child alone is freedom.
In the child is peace; it's not yet at variance with itself. Richness is in the child; it's still to know its heart, the penury of life. It is immortal, for it knows nothing of death.
But this men cannot bear. That which is divine must become like one of them, must learn that they too are there, and before nature expels the child from its paradise, men cajole and drag it out onto the ground of the curse, that it may, like them, grind away its life in the sweat of its face.
But the time of awakening is beautiful too, if only we're not woken out of season.
Oh, they are hallowed days in which our heart first tests its wings, when full of quick and fervent growth we stand there in the glorious world, like the young plant when it unfolds to the morning sun and stretches up its slender arms towards the endless heaven.
How I felt impelled to roam amongst the mountains and along the shore! oh, how I often sat with throbbing heart upon the heights of Tinos, and gazed after the falcons and the cranes, and the doughty sprightly ships as they shrank below the horizon! 'Down there!' I thought, 'down there you too one day will wander,' and I felt like one who, parched with heat, plunges into the cooling pool and splashes the spumy waters on his brow.
Sighing I'd then turn back towards my home. 'If only my school years were over,' I often thought.
Dear boy! They're far from over yet.
That in man's youth he thinks the goal so near! That is the most beautiful of all illusions with which nature helps our weakness.
And often when I lay amongst the flowers and basked in the soft spring sunlight, and looked up into the bright blue that embraced the warm earth, when I sat under the elms and willows, in the womb of the mountain, after a quickening shower, when the branches still quivered from the caresses of heaven and golden clouds moved above the dripping woods, or when full of peaceful spirit the evening star rose with those ancient youths, the other heroes of the heavens, and I watched as the life within them propelled itself through the aether in eternal effortless order, and the peace of the world enfolded and elated me, so that I roused and listened, not knowing what came over me — 'do you love me, good father in heaven!' I'd then silently ask, and felt so blissful and sure his answer in my heart.
O you whom I'd invoke as if you were above the stars, whom I called creator of heaven and earth, amiable idol of my childhood, you won't be angry I've forgotten you! — Why is the world not so wanting as to make one seek an entity outside it?2
Oh, if she's the daughter of a father, glorious nature, is the daughter's heart not his heart? her inmost self, is it not He? But do I possess it then? do I know it then?
It's as if I saw, but then again I take fright, as if it were my own image I'd seen, it's as if I felt him, the spirit of the world, like the warm hand of a friend, but I awake and think it's my own fingers I've been holding.
#### Hyperion to Bellarmin [IV]
Do you know how Plato and his Stella loved each other?
That's how I loved, how I was loved. Oh, I was a lucky lad!
It's a joy when like and like are joined, but when a great man raises lesser to his level, it's divine.
A kindly word from a brave man's heart, a smile that conceals the consuming glory of the spirit, is little and much, like a magical password hiding life and death in its innocent syllable, like living water welling up from deep inside the mountains and conveying to us in each crystal drop the secret energy of the earth.
But how I hate all the barbarians who think themselves wise because they no longer have a heart, all the vulgar brutes who find a thousand different ways to kill and destroy youth's beauty with their stupid petty principles of manhood!
Good God! This is the owl wanting to drive from the nest the young eagles, wanting to show them the way to the sun!
Forgive me, spirit of my Adamas! for recalling these people before you. That's what we gain from experience that we can imagine nothing excellent without its malformed opposite.
<sup>2 [</sup>Hölderlin's note] It should go without saying that such utterances, as mere manifestations of the human mind, ought properly to give no grounds for offence.
Oh, would that you were ever present to me, with all that is akin to you, grieving demi-god I cherish! Those you enfold with your tranquillity and strength, conqueror and warrior, those you confront with your love and wisdom, let them flee or become like you! What's ignoble and weak stands no chance beside you.
How often you were near to me when you were long since far away from me, you glorified me with your light, warmed me that my frigid heart began to stir again, like a frozen stream when it's touched by the ray of heaven! Then I felt like fleeing to the stars with my bliss, so that it not be debased by the world around me.
I'd grown up like an unpropped vine, and the wild tendrils spread aimlessly across the ground. As you well know, there's many a noble energy that perishes with us because it isn't used. I flitted about like a will-o'-the-wisp, grasped at everything, was gripped by everything, but then only for the moment, and my clumsy energies exhausted themselves to no purpose. Everywhere I felt wanting, and still couldn't find my goal. So he found me.
For long enough he'd practised patience and art on his material, the so-called cultivated world, but his material had been and stayed stone and wood; it might, when occasion demanded, outwardly assume the noble human form, but that's not what my Adamas was about; he wanted human beings, and to create them he'd found his art too poor. That those he sought had once existed, those his art was too poor to create, this he clearly saw. Where they'd existed, he also knew. That's where he wished to go, to probe beneath the rubble for their genius and with it while away his lonely days. He came to Greece. So I found him.
I still see him approaching me in smiling contemplation, I still hear his greeting and his questions.
Like a plant when its peace soothes the striving spirit and simple contentment returns to the soul — so he stood before me.
And I, was I not the echo of his quiet inspiration? did the melodies of his being not reverberate in me? What I saw I became, and it was divine what I saw.
How feeble is even the most honest human industry compared with the sheer power of unbroken inspiration.
This doesn't linger on the surface, doesn't merely touch us here or there, has no need of time or means; nor does it need command, compulsion and conviction; it takes hold of us in one moment on all sides and on all levels, low and high, and before we know it's there, before we can wonder what is coming over us, it turns us through and through into its blissfulness and beauty.
Happy the man whose path has thus been crossed in early youth by a noble spirit!
Oh, these are golden unforgettable days, full of the joys of love and sweet activity!
Adamas led me now into the world of Plutarch's heroes, now into the magical land of the Greek gods, now he used number and measure to bring to my youthful impetuousness order and composure, now he took me up into the mountains: by day to see the flowers of field and forest and the wild mosses of the rocks, by night to see the holy stars above us, and understand them after the manner of men.
There is a luscious feeling of well-being within us when our inner self can thus draw strength from its material, separating from it to bond with it more faithfully, and step by step the spirit becomes empowered.
But with threefold force I felt him and myself when, like shades from the past, in pride and joy, in anger and grief, we journeyed up as far as Athos and from there shipped eastwards to the Hellespont, then down to the shores of Rhodes and Taenarum's mountain chasms, through the silent islands all; when longing drove us inland from the coasts, into the sombre heart of ancient Peloponnese, to the lonely banks of the Eurotas, oh! the desolate valleys of Elis and Nemea and Olympia; when leaning against a pillar of one of forgotten Jupiter's temples, hugged by laurel roses and evergreens, we gazed into the wild riverbed, and the vibrance of spring and the ever youthful sun reminded us that man too had once been there and now is gone, that the glorious nature of humanity is barely there any more, like the fragment of a temple, or in memory as the image of one dead — then I sat playing sadly beside him, plucking the moss from a demi-god's pedestal, digging some hero's marble shoulder out of the rubble, and cutting away the brambles and the heather from the half-buried architraves, whilst my Adamas sketched the landscape that fondly held the ruins in its comforting embrace, the corn-covered hill, the olives, the herd of goats clinging to the mountain crag, the forest of elms sweeping down from the peaks to the valley; and the lizard frisked at our feet, and the flies buzzed about us in the stillness of noon — Dear Bellarmin! I'd love to give a point-by-point account in Nestor's manner; I range through the past like a gleaner through a field of stubble, when the lord of the land has reaped; one picks up every piece of straw. And the time I stood beside him on the heights of Delos, what a day it was that dawned for me as I climbed with him the ancient marble steps up Cynthus' granite face. Here once dwelt the sun-god, amidst the heavenly festivals where, like golden clouds, assembled Greece glowed all around him. It's here the youths of Greece immersed themselves in floods of joy and inspiration, like Achilles in the Styx, and like the demi-god emerged invincible. In the groves, in the temples their souls awoke, each sounding in the other, and all faithfully preserving the rapturous chords.
But why am I speaking of this? As if we still had an inkling of those days! Oh! not even a beautiful dream can thrive under the curse that weighs upon us. Like a howling north wind the present blasts the blossoms of our spirit and sears them in their bloom. And yet it was a golden day that enfolded me on Cynthus! We reached the summit with dawn still breaking. Now he rose in his eternal youth, the ancient sungod; serene and effortless as ever, the immortal titan with a thousand joys of his own soared upward, smiling down on his wasted land, on his temples, his pillars that fate had tossed before him like withered rose petals, mindlessly ripped from the bush by a passing child and scattered over the earth.
'Be like him!' Adamas cried out to me, grasped my hand and held it out towards the god, and I felt as if the morning winds were carrying us away, bringing us into the train of the holy being that now rose up to the summit of heaven, kindly and grand, and wonderfully infused the world and us with his spirit and his power.
Still my inmost self grieves and rejoices over every word Adamas spoke to me then, and I can't understand my privation when often I feel as he must have then. What is loss when man thus finds himself in a world which is his own? In us is everything. If a hair should fall from his head, what is it to him? Why does man so strive for bondage when he could be a god! 'You will be lonely, my dearest!' Adamas also said to me then, 'you will be like a crane, left behind by his kin in the harsh season whilst they seek out the spring in distant lands.'
And there you have it, dear friend! That's what makes us poor for all our wealth, that we cannot be alone, that as long as we live the love within us will not die. Give me back my Adamas, and come with all my kindred that amongst us may renew itself the ancient world of beauty, that together we may gather and commingle in the arms of our godhead, nature, and you will see, then I'll know nothing of need.
But let no one say it's fate that parts us! It's we, we ourselves who do it! we take our delight in plunging into the night of the unknown, into the cold alien terrain of some other world, and were it possible, we would quit the sun's realm and storm beyond the bounds of our wandering star. Alas! for man's wild breast there can be no home; and as the sun's ray sears the plants of earth it first unfolded, so man murders the sweet flowers that flourished at his breast, the joys of kinship and of love.
I might seem to bear a grudge against my Adamas for leaving me, but I bear him no grudge. Oh, he was going to return!
Hidden in the depths of Asia there's said to be a people of rare virtue; it's thither he was driven by his hopes.
I kept him company as far as Nios. Those were bitter days. I've learnt to suffer pain, but have no strength in me for such a parting.
With every moment that brought us closer to the final hour emerged more clearly how this man was woven into my being. As a dying man to his fleeing breath, so clove my soul to him.
At Homer's grave we passed a few more days, and Nios became for me the most hallowed of the islands.
Finally we tore ourselves away. My heart had worn itself weary. By the last moment I was calmer. I lay on my knees before him, clasped him for the last time in these arms: 'Give me a blessing, my father!' I softly cried up to him, and he smiled grandly, and his brow widened before the stars of morning, and his eye pierced the spaces of heaven — 'Preserve him for me,' he cried, 'you spirits of a better time! and raise him to your immortality, and all you friendly powers of heaven and earth, be with him!'
'There is a god in us,' he added more calmly, 'who steers our fate like rivers of water, and all things are his element. Be this god with you above all!'
So we parted. Farewell, my Bellarmin!
### Hyperion to Bellarmin [V]
Whither could I flee from myself, had I not the dear days of my youth?
Like a spirit that finds no rest by Acheron, I revisit the deserted scenes of my life. All things age and renew themselves. Why are we excluded from nature's beautiful cycle? Or does it hold for us too?
I'd gladly believe it, but for one thing in us that boils up from the depths of our being like the titan in Etna, the monstrous striving to be everything.
And yet who'd not rather feel it in himself like seething oil than own that he was born for the yoke and the whip? Which is nobler: a rampant battle-steed or a droopy-eared nag?
Dear friend! there was a time when my breast too basked in great hopes, when for me too the joy of immortality throbbed in every pulse, when I would wander amid grand designs as if in some vast sylvan night, when, like the fish of the ocean, I'd happily press on and ever onwards in my shoreless future.
How boldly, blissful nature! did the youth leap from your cradle! how he rejoiced in his untested arms! His bow was ready strung and his arrows rustled in the quiver, and the immortals, the sublime spirits of antiquity, were his leaders, and his Adamas was in their midst.
Wherever I was or went, these glorious forms kept me company; like flames heroic deeds from all the ages mingled in my mind, and just as those gigantic shapes, the clouds of heaven, merge together into one exultant storm, so merged in me the hundredfold triumphs of the Olympiads, became one single endless triumph.
Who can withstand it, who is not floored by the terrible glory of antiquity, like young woods flattened by a hurricane, when it seizes him as I was seized, and when, like me, he lacks the element in which to gain a firming sense of self?
Oh, like a storm the greatness of the ancients surely bowed my head, it blasted the bloom from my cheek, and often I would lie where no eye could see me, under a thousand tears, like a fallen fir when it lies by the stream and hides its withered crown beneath the waters. How gladly I'd have bought with blood a moment from a great man's life!
But what help to me was that? The fact is, no one wanted me.
Oh, it's pitiful to see oneself so crushed; and let him who finds this hard to understand not trouble himself further, and give his thanks to nature for having like the butterflies created him for joy, and go and in his life speak nevermore of misery and pain.
I loved my heroes as a fly loves the light; I'd seek their dangerous nearness and flee from it and seek it out again.
Like a bleeding hart into the stream, I would often plunge headlong into the whirlpool of joy, to cool my burning breast and wash away the glorious raving dreams of fame and greatness, but what help was that?
And often when at midnight my heated heart drove me down into the garden beneath the dewy trees, and the burn's lullaby and the balmy breeze and the moonlight soothed my senses, and so tranquil and free the silver clouds above me stirred, and the ebbing voice of the surging sea sounded faintly from afar, how fondly all its love's great phantoms played then with my heart!
'Farewell, you heavenly ones!' I often spoke in spirit, when over me the softly sounding melody of morning's light began, 'you glorious dead, farewell! Could I only follow you, shake off all this age has given me and set forth into the freer realm of the shades!'
But I languish parched in my chains and snatch with bitter joy the beggarly bowl that's offered for my thirst.
#### Hyperion to Bellarmin [VI]
—————
My island had grown too strait for me since Adamas's leaving. I'd been bored in Tinos for years. I wanted to get out into the world.
'Go first to Smyrna,' said my father, 'learn there the arts of seamanship and war, learn the speech of polished peoples, learn about their constitutions and opinions and manners and customs, prove all things and hold fast the best! — Then for my part go on as you will.'
'Learn a little patience too,' my mother added, and I accepted the advice and thanked her for it.
It's rapturous to take the first step beyond the bounds of youth, and when I look back to my parting from Tinos, it's like thinking of the day of my birth. There was a new sun above me, and I relished land and sea and air as if for the first time.
The living activity with which I now pursued my education in Smyrna, and my rapid progress, did much to soothe my heart. And from this time I can recall many a blissful evening, my labours done. How often would I stroll beneath the evergreen trees on the banks of the Meles, by the birthplace of my Homer, and pick sacrificial flowers and cast them into the hallowed stream! Then in my peaceful dreams I trod the path to the nearby grotto where, they say, the venerable poet sang his Iliad. I found him. Every sound in me stilled before his presence. I opened his divine poem and it was as if I'd never known it, so wholly new the way it came alive in me now.
I also have fond memories of roaming through the country around Smyrna. It's a glorious land, and a thousand times I've wished for wings to fly just once a year to Asia Minor.
From the plain of Sardis I made my way up through the crags of Tmolus.
I'd spent the night in a friendly hut at the mountain's foot, amid myrtles, amidst the fragrance of the rockrose bush, with the swans playing beside me in the golden waters of the Pactolus, and an ancient temple of Cybele, like a bashful ghost, glancing out from between the elms into the bright moonlight. Five lovely pillars stood mourning over the rubble, and a regal portal lay toppled at their feet.
Now through a thousand blooming bushes my path grew upwards. From the craggy slope whispering trees bowed, showering their gossamer flakes on my head. I'd set out in the morning. By noon I'd reached the crest of the range. I stood there, gladly gazing, relishing the purer airs of heaven. Those were blissful hours.
The land from which I'd climbed lay before me like a sea, youthful, full of living joy; it was with a heavenly unending play of colours that spring greeted my heart, and as the sun of heaven found itself again in the thousandfold changes of light given back to it by the earth, so my spirit knew itself in the fullness of life embracing it, besieging it from every side.
To the left the torrent, exulting like a giant, plunged down into the woods from the marble cliff that beetled above me, where the eagle played with its young and the snowy peaks glistened up into the blue aether; to the right storm clouds came rolling over the forests of Mount Sipylus; I didn't feel the gale that carried them, only the breath of a breeze in my hair, but I heard their thunder, as one hears the voice of the future, and I saw their flames like the distant light of dimly sensed divinity. I turned to the south and wandered on. There it lay open before me, the whole paradisal landscape through which the Cayster flows with many a charming wimple, as if it couldn't linger long enough amidst all the affluence and loveliness surrounding it. Like the zephyrs my spirit fluttered blissfully from beauty to beauty, from the nameless peaceful hamlet, huddled deep below at the foot of the mountain, out to where the range of Messogis looms.
I returned to Smyrna like a drunk man from a banquet. My heart was too full of delights for me not to endow mortality with something of its surplus. I'd too happily absorbed as bounty nature's beauty for me not to fill with it the gaps in human life. My shabby Smyrna was clothed in the colours of my inspiration and stood there like a bride. I found myself attracted by the sociable city-dwellers. The absurdity in their manners amused me like a child's buffoonery, and since I was by nature beyond all their established forms and customs, I toyed with them all, donning and doffing them like carnival costumes.
But what really added spice for me to the vapid fare of everyday communion were those good faces and figures that merciful nature here and there still sends like stars into our darkness.
What hearty joy I took in them! how full of faith I read those friendly hieroglyphics! But it was much like my erstwhile experience with the birches in spring. I'd heard about the sap of these trees and had extravagant expectations of the delicious drink their lovely stems must yield. But it lacked both strength and spirit.
Alas! and how utterly hopeless was everything else I heard and saw.
Sometimes, when I went about among these polished people, it really seemed to me that human nature had broken down into so many species of beast. Here, as everywhere, it was the men who were especially degenerate and corrupt.
There are animals that howl when they hear music. But when talk turned to beauty of the spirit and youthfulness of heart, my more mannerly humans laughed. Wolves slink off when someone strikes a light. When these people saw a spark of reason they'd turn their backs like thieves.
When I spoke the odd warm word of ancient Greece, they yawned and opined, it was this day and age we had to live in; and another added weightily, good taste was still alive and well.
And sure enough, this was soon in evidence. One of them japed like a deckhand, the other puffed out his cheeks and spouted gobbets of wisdom.
Or there'd be one who'd act enlightened, snap his fingers at heaven and proclaim: he'd never worried about the birds in the bush, give him the bird in the hand anytime! Yet mention death to him and he would at once clasp his hands, and in the course of conversation observe that it's a dangerous thing our priests now count for nothing.
The only ones I sometimes had a use for were the tellers of tales, the walking roll calls of foreign cities and countries, the talking peep boxes in which one can see spires and market places and potentates on chargers.
I finally grew weary of throwing myself away, looking for grapes in a desert and flowers on a field of ice.
I now lived more determinedly alone, and the gentle spirit of my youth had almost wholly vanished from my soul. The incurableness of the age had become plain to me from so much I tell and leave untold, and the beautiful consolation of finding my world in a single soul, embracing my kind in but one friendly figure, this too was denied me.
Dear friend! what were life without hope? A spark that leaps from the coal and perishes, and like a gust of wind heard in the dreary season, it soughs for a moment and then it's gone — should we think that's the way it is with us?
Even the swallow seeks friendlier climes in winter, the hart runs around in the heat of the day and its eyes seek streams of water. Who tells the infant its mother won't withhold her breast? And look! it seeks it still.
Nothing could live if it had not hope. My heart now locked its treasures away, but only to keep them safe for better times, for the one, the holy, the true that at some period of existence should surely meet my thirsting soul.
How blissfully I'd often cling to it in hours of forefeeling when it gently played like moonlight around my soothed brow? Already then I knew you, already then you looked upon me like a spirit from the clouds, you who'd one day rise for me, in the peacefulness of beauty, out of the dismal wave of the world! Now this heart would never more battle and burn.
As a lily sways in a silent breeze, so stirred my being in its element, in rapturous dreams of her.
#### Hyperion to Bellarmin [VII]
Smyrna was soured for me now. My heart had grown altogether weary over time. Now and then the wish might rear in me to wander round the world, or find some war to fight in, or else seek out my Adamas and burn away my rancour in his fire; but wish was all it stayed, and my futile sapless life refused to be refreshed.
Summer was now coming to an end; already I could sense the dreich dank days and the whistling of the winds and the brawling of the rain-swollen streams; and nature, which like a foaming fountain had swelled up in every flower and tree, stood already now before my darkened mind, dwindling and closed down and turned in upon itself, like me.
I wanted to take with me what I could of all this fleeting life; all I'd grown to love without I wanted to make safe within, for well I knew that the returning year would find me no longer amidst these trees and mountains; and so now, more than before, I walked and rode about throughout the region.
But what mainly drove me abroad was the secret longing to catch sight of one I'd encountered daily for some time, passing beneath the trees before the city gates.
Like a young titan the glorious stranger strode amongst this race of pygmies who with timid joy drank in his beauty, sizing up his strength and stature, feasting with furtive glance upon the glowing bronzed Roman face, as if upon forbidden fruit; and it was a glorious moment each time this man's eye, for whose gaze the vastness of the aether seemed too small a bound, so laid aside all pride and searched and strove until it felt itself in mine, and blushingly we passed each other by, looking all the while.
One day I'd ridden deep into the forests of Mount Mimas, and didn't turn home till late evening. I'd dismounted and was leading my horse down a steep, broken path, over tree-roots and stones, and as I wound my way down through the scrub into the hollow that now opened before me, I was suddenly set upon by Karaburun robbers, and at first I struggled to fend off their two drawn sabres, but they were already weary from other work, and in the end I coped. I calmly got back on my horse and rode on down.
At the foot of the mountain, amidst the woods and the piled-up crags, a small glade opened up before me. It grew light. The moon had just risen over the gloomy trees. From a distance I saw horses stretched out on the ground and men beside them in the grass.
'Who are you?' I called out.
'That's Hyperion!' cried a hero's voice, joyfully surprised. 'You know me,' the voice went on; 'I see you every day under the trees by the city gates.'
My horse flew to him like an arrow. The moonlight shone bright in his face. I knew him; I jumped down.
'Good evening!' cried the gentle strapping fellow, giving me a look both wild and tender, and his brawny fist squeezed mine and made me feel his meaning through and through.
Oh, now my bootless life was at an end!
Alabanda — this was the stranger's name — now told me that he and his servant had been set upon by robbers and the two that I'd encountered had been sent packing by him, that he'd lost his way out of the forest and hence had been forced to stay put till I came by. 'This has cost me a friend,' he added, pointing to his dead horse.
I gave mine to his servant and we proceeded on foot.
'It serves us right,' I began, as we made our way out of the forest arm in arm; 'why did we have to hold back for so long and pass each other by, so that it took this misadventure to bring us both together.'
'But I must tell you,' replied Alabanda, 'that you were more to blame, you were the colder one. I rode after you today.'
'My glorious friend!' I cried, 'just you wait and see! you shall never surpass me in love.'
Each moment brought us closer, more joyfully together.
As we approached the city we passed a well-built caravansary, resting amid murmuring fountains and fruit trees and sweet-scented meadows.
We decided to spend the night there. We sat up late together, with the windows open. A sublime spirit of stillness enfolded us. Earth and sea were sunk in blissful silence, like the stars that hung above us. Barely a breeze floated into the room from the sea and gently played with our light, and perhaps the more plangent tones of distant music drifted to us, while the thundercloud lay cradled in the aether's bed and sounded fitfully through the stillness from afar, like a slumbering giant when he sighs more deeply in his dreadful dreams.
Our souls were bound to be all the more powerfully drawn together, because they'd been unwillingly closed off from one another. We met like two streams that come rolling down the mountain and hurl aside the ballast of earth and stones and rotten wood and all the sluggish mess that holds them back, clearing their way to each other and bursting through to where, embracing and embraced with equal force, they mingle into one majestic river and begin their wandering course into the vastness of the sea.
He, by fate and men's barbarism driven from hearth and home and hunted hither and thither amongst strangers, from early youth made savage and embittered, and yet also with an inner heart full of love, full of longing to break out of the coarse husk and into a congenial element; I, so inwardly detached from everything already, so wholly alien and lonely amongst men, in the dearest melodies of my heart accompanied so grotesquely by the world's clanging cymbals; I, for whom the blind and lame were hated of my soul, yet myself too blind and lame, so utterly encumbered by all I had in common, be it ever so remotely, with the sophisticates and pedants, the barbarians and the wiseacres — and so full of hope, so full of single-hearted expectation of a more beautiful life —
How could these two youths not rush to embrace each other in temptestuous joy?
O you, my friend and brother-in-arms, my Alabanda, where are you? I could believe you've gone across into the unknown land, to peace, and have become again what once we were as children.
Sometimes, when a thunderstorm passes overhead, dispensing its divine energies on forests and seeded fields, or when the waves of the surging sea disport themselves, or a chorus of eagles soars around the mountain peaks where I wander, my heart can stir as if my Alabanda were not far away; but more visibly, more vividly, more palpably he lives within me, just as once he stood before me, a fierily stern and terrible accuser when he rehearsed the sins of the age. How then my spirit awakened in its depths, how the thundering words of remorseless justice rolled over my tongue! Like harbingers of Nemesis, our thoughts scoured the earth and cleansed it till not a trace of any curse remained.
We also summoned the past before our tribunal; proud Rome didn't daunt us with its glory, nor Athens seduce us with its youthful bloom.
Like storms when they, exulting and relentless, blast their way through woods and over mountains, so our souls forged ever onwards in their colossal projects; not that we, like milksops, brought forth our world as if by incantation, callowly expecting to encounter no resistance — Alabanda was too astute and too manful for that. But often even effortless enthusiasm can be soldierly and shrewd.
One day I recall with particular vividness.
We'd gone together out into the countryside, sat in intimate embrace in the dusky shade of the evergreen laurel, looking together in our Plato where he speaks so wonderfully sublimely of ageing and renewal, and now and then resting our eyes upon the muted leafless landscape where heaven, more beautifully than ever, with cloud and sunlight played about the trees in their autumnal sleep.
Then we spoke much of present-day Greece, both with bleeding hearts, for this same dishonoured soil was fatherland to Alabanda too.
And he was indeed uncommonly moved.
'When I see a child,' he cried, 'and consider how demeaning and corrupting is the yoke that it will bear, and that it will languish, as we do, that it will search for human beings, as we do, and ask, as we do, after the beautiful and the true, that it will fruitlessly waste away because it will be alone, as we are, that it — O Greeks, just take your sons from their cradles and cast them into the river, so that at least they'll be spared your shame!'
'Surely, Alabanda, things will surely change!' I said.
'How?' he retorted; 'our heroes have lost their fame, our sages their disciples. Great deeds, without a noble people to heed them, are no more than a hefty blow to a numb skull, and lofty words that find no lofty hearts in which to echo are like a dying leaf rustling down into the dung. What will you do about it?'
'I will,' I said, 'take my shovel and fling the dung into a pit. A people in whom greatness of thought and deed no longer inspires greatness of thought and deed has nothing more in common with others that still are human beings, it has forfeited all rights, and it's an empty farce, a superstition, to go on reverencing such will-less corpses as if there beat a Roman heart within them. Away with them! It shall not stand where it stands, the rotten withered tree, for it steals light and air from the young life ripening for a new world.'
Alabanda flew to me and hugged me, and his kisses pierced my soul. 'Comrade!' he cried, 'dear comrade! Oh, now I've got one hundred arms with which to fight!
'I hear my melody at last,' he went on, in a voice that stirred my heart like a battle cry, 'and that's enough for me! They're glorious words you've spoken, Hyperion! What? the god shall depend on the worm? The god in us, for whom the path to infinity beckons, shall stand and wait till the worm gets out of his way? No! no! We don't ask whether you're willing! You never are, you slaves and barbarians! Nor will we try to change you for the better, for that's a waste of effort! we just want to see to it you make way for the triumphant march of man. Oh! let someone light the torch for me that I may sear the weed from the heath! let the mine be made for me that I may blast the sluggish clods from the earth!'
'Where possible we'll push them gently aside,' I interjected.
Alabanda was silent for a while.
'I take my joy in the future,' he spoke again at last, ardently clasping both my hands. 'Thank God! I'll meet no common end. To be happy is to be torpid in the mouths of the slaves. To be happy! when I hear you slaves speak to me of being happy, it's like having pap and lukewarm water on the tongue. So fatuous and futile is everything for which you sacrifice your laurel crowns, your immortality.
'O holy light that wanders restlessly above us, working throughout its vast realm, and transmits its soul to me too in the rays that I drink, may your happiness be mine!
'The sons of the sun feed on their deeds; it's by victory they live; with their own spirit they rouse themselves, and their strength is their joy.' —
The spirit of this man would often seize you so, you might almost have been ashamed, feeling yourself so carried away light as a feather.
'O heaven and earth!' I cried, 'that's what I call joy! — Those are other times, that's not a voice from my childish age, that's not the soil where men's hearts heave under the slave driver's whip. — Yes! yes! by your glorious soul, man! You will save the fatherland with me.'
'That I will,' he cried, 'or perish.'
From this day on we grew ever holier and dearer to one another. There was now a deep and ineffable earnestness between us. But we were only the more blissfully together. Each of us lived only in the eternal ground tones of his being, and onward we strode without frills from one great harmony to another. Our common life was full of glorious rigour and boldness.
'How is it you've become so still?' Alabanda once asked me smilingly. 'In the torrid zones,' said I, 'nearer the sun, even the birds don't sing.'
But in this world all things rise and fall, and man with all his giant strength holds nothing fast. Once I saw a child stretch out its hand to catch the moonlight; but the light went calmly on its way. So too are we left standing there, wrestling to hold back wandering fate.
Oh, could we only ponder it as calmly as we watch the stars in their course!
The happier you are, the less it takes to bring you down, and such blissful days as Alabanda and I were living are like the peak of a sheer crag where it needs the merest touch from your companion to send you abruptly tumbling over the lacerating edges, down into the endless glooming depths.
We'd taken a glorious trip to Chios, and found a thousand joys in one another. Like breezes over the surface of the sea, the friendly enchantments of nature swayed over us. In joyful surprise we looked at one another without speaking a word, but our eyes were saying, 'I've never seen you like this before!' So glorified were we by the powers of earth and heaven.
During the trip we'd argued with bright ardour about a good few things; this time, as before, my heart had taken its joy in watching this spirit on his bold erratic course, as he so heedless of all rules, with such cheerful abandon and yet mostly so sure-footedly pursued his path.
As soon as we were back on land we made haste to be alone.
'You can't really convince anyone,' I now said most lovingly, 'you win people over, you beguile them before you begin; when you speak it's impossible to doubt, and he who doesn't doubt isn't convinced.'
'Proud flatterer,' he responded, 'you lie! but it's as well that you remind me! only too often you've robbed me of my reason! I wouldn't wish to be free from you, not for all the crowns in the world, but yet I often find it troubling that you should be so indispensable to me, that I'm so fettered to you; and look,' he went on, 'so that you may have me all, you shall know all about me too! up till now, for all the joy and glory, we've never thought to delve into the past.'
Then he told me of his fate; it was like watching a young Hercules struggle with Megaera.
'Will you now forgive me,' he concluded the tale of his adversities, 'will you now have more patience if often I'm coarse and rancid and hard to stomach!'
'Oh stop, stop!' I cried, deeply moved; 'but to think that you're still here, that you preserved yourself for me!'
'That's right! for you!' he cried, 'and I'm delighted you yet find me palatable fare. And though sometimes I might taste to you like a crab, just keep on pressing me until I'm fit to drink.'
'Leave me be! leave me be!' I cried; I resisted in vain; the man made a child of me; nor did I conceal it from him; he saw my tears, and woe to him if he shouldn't have seen them!
'We're carousing,' Alabanda now resumed, 'we're killing time in drunkenness.'
'We're having our bridegroom days together,' I blithely cried, 'so why shouldn't it still sound as if we were in Arcadia. — But to return to what we were discussing earlier!
'Say what you will, you grant too much power to the state. It may not demand what it cannot enforce. Yet what love gives, and the spirit, cannot be enforced. Let it leave that untouched, or we should take its precious law and nail it to the pillory! By heaven! he little knows how much he sins who would turn the state into a school of morals. It's this that's always made the state into a hell, that man has tried to make of it his heaven.
'The state is the coarse husk round life's kernel and nothing more. It's the wall around the garden of human fruits and flowers.
'But what help is the wall around the garden when the soil lies parched? Then all that helps is rainfall from heaven.
'O rainfall from heaven! O inspiration! You will bring again to us the springtime of the peoples. The state cannot command you hither. But let it not hinder you, then you will come, come you will, with your almighty joys, wreathe us in golden clouds and lift us high above mortality, and we shall marvel and wonder whether we're still who we were, we, the needy who'd ask the stars if a blossoming spring awaited us there — do you ask me when this shall come to pass? Then when the darling of time, the youngest, most beautiful daughter of time, the new church arises from these soiled and worn-out forms, when the wakened feeling for the divine brings back to man his divinity and beautiful youth to his breast, when — I can't proclaim it, for I barely sense it, but it shall surely, surely come. Death is a herald of life, and that we're now slumbering in our sickhouses, that's a sure sign there'll soon be a healthy awakening. Then, only then shall we truly be, then is the spirits' element found!'
Alabanda was silent, and looked at me astonished for a while. I was rapt with endless hopes; divine energies carried me away like a wisp of cloud —
'Come!' I cried, grasping Alabanda by his garment, 'come, who can stand it any longer in the benighted dungeon that surrounds us?'
'Whither, my drunken dreamer?' answered Alabanda dryly, and a shade of derision seemed to pass across his face.
I felt that I'd been flung down from the clouds. 'Go!' I said, 'you're a little man!'
At that moment several strangers entered the room, striking figures, mostly gaunt and pale, as far as I could tell in the moonlight, and calm, but there was something in their mien that pierced the soul like a sword, and it was like standing in the presence of omniscience; one might have doubted whether these were the aspects of needy natures, had not murdered feeling left behind its traces here and there.
One I found particularly striking. The stillness of his features was the stillness of a battlefield. Within this man wrath and love had raged, and intelligence beamed out above the wreckage of his heart like the eye of a hawk, perched on palace ruins. Deep contempt lay on his lips. One sensed it was no trivial purpose that engaged this man.
Another more likely owed his calm to natural hardness of heart. He displayed barely any trace of violence, either freely willed or inflicted by fate.
A third could rather have wrung from life his coldness by force of conviction, and often be at war still with himself, for there was a lurking contradiction in his being; to me he seemed to feel the need to keep himself in check. He had the least to say.
As soon as they came in Alabanda leapt up like sprung steel.
'We've been looking for you,' one of them cried.
'You'd find me,' he said laughingly, 'if I hid myself in the centre of the earth. They're my friends,' he added, turning towards me.
They seemed to eye me rather sharply.
'Here's another keen to see the world a better place,' cried Alabanda in a while, pointing to me.
'Are you serious about that?' asked one of the three.
'Bettering the world is no joking matter,' I said.
'You've said a great deal in few words!' cried one of them in his turn. 'You're our man!' another.
'You think that way too?' I asked.
'Ask us what we do!' was the response.
'And if I asked?'
'Then we'd tell you we're here to clean up the earth, that we clear the stones from the field, and break the hard clods with the mattock, and dig furrows with the plough, and seize the tares by the roots, cut them through at the roots, rip them out with the roots, that they may wither away in the searing sun.'
'Not that we expect to reap,' another interjected, 'for us the reward will come too late; we shall not live to see the harvest ripen.
'We're in the evening of our days. We erred often, hoped much and achieved little. We liked to take more risks than time to think. We were eager to be quickly done and trusted to good fortune. We spoke much of joy and pain, and loved and hated both. We played with fate and it did the same with us. From beggar's staff to crown it tossed us up and down. It swung us as one swings a glowing censer, and we glowed till the coals turned to ash. We've ceased to speak of fortune good or ill. We've grown beyond life's middle where all is green and warm. But that which outlives youth is not the worst. It's from hot metal the cold sword is forged. And they say that burnt-out dead volcanoes yield no mean young wine.'
'It's not for our sakes we say this,' cried another now more urgently, 'it's for your sakes we say it! We don't go begging for men's hearts. We don't need their hearts, their wills. For there's no way they can be against us, for everything is for us, and the foolish and the prudent and the simple and the wise and all vices and all virtues of crudity and culture are in our service without hire, and help us blindly on towards our goal — only we would wish that there be some to savour what we do, and so, from amongst the thousand blind helpers, we select the best that we may turn them into seeing helpers — but if no one wants to dwell where we have built, it's not our fault and not our loss. We did our part. If no one wants to gather where we ploughed, who will blame us for that? Who will curse the tree when its apple falls into the mire? I've often told myself, "you're offering to corruption," and still went on to finish my day's work.'
'These are impostors!' all the walls cried out to my quickened sense. I felt like one about to suffocate in smoke, who smashes doors and windows to get out, so did I thirst for air and freedom.
And soon they saw how ill at ease I was and broke off. Day was already dawning when I stepped out of the caravansary where we'd been together. I felt the breath of the morning breeze like balsam on a burning wound.
I was already too riled by Alabanda's derision for his mysterious connections not to make me lose all faith in him now.
'He's bad,' I cried, 'yes, he's bad. He feigns unbounded openness and consorts with men like these — and conceals it from you.'
I felt like a bride when she learns her beloved is secretly living with a whore.
Oh, this was not the kind of pain one likes to nurse, holding it to one's heart like a child, singing it to slumber with the tones of the nightingale!
Like a wrathful serpent when it climbs remorselessly up knees and loins, enwinding every limb, and sinks its venomed fangs now into breast, now into nape, so was my pain, so did it grip me in its terrible embrace. I summoned my highest heart for help and strove for noble thoughts so as to still hold back, and for a few moments too it worked, but now I'd also found the strength for rage, and now, like a malicious blaze, I killed each spark of love within me.
'He must,' I thought, ' — these are his people after all — he must be conspiring with them against you! What did he want with you anyway? What could he have looked for in you, the dreamer? Oh, had he but gone his way! But then they get their peculiar pleasure in taking on their opposite! having such a strange beast in their stable suits them just fine!' —
And yet I'd been ineffably happy with him, had so often sunk in his embrace and emerged with heart invincible, so often like steel been tempered and purged in his fire!
When once on a serene midnight I pointed out to him the Dioscuri, and Alabanda laid his hand on my heart and said: 'Those are mere stars, Hyperion, mere letters with which the names of the heroic brothers are written in the heavens; it's in us they are! living and true, with their courage and their divine love, and you, you are the son of the god and share your immortality with your mortal Castor!' —
When I roamed with him through the forests of Ida, and we descended down into the valley to ask the silent burial mounds about their dead, and I said to Alabanda that amongst the mounds one perhaps belonged to the spirit of Achilles and his beloved friend, and Alabanda confessed to me that he was often a child and imagined our one day falling in a single vale of battle and resting together beneath a single tree — who would have thought then that it would come to this?
I turned it over and over with all the strength of mind that I had left, I accused him, I defended him, and accused him the more bitterly again; I struggled against my mood and tried to cheer myself, only for all my efforts to end in utter gloom.
Alas! my eye had been left red and raw from many a hefty blow and was hardly even beginning to heal, how should it now see more soundly?
Alabanda visited me next day. My heart seethed when he came in, but I restrained myself, however much I was nettled and angered by his pride and his composure.
'The air is glorious,' he finally said, 'and it will be a very beautiful evening, let's walk up to the acropolis together.'
I agreed. For a long time no word was spoken. 'What is it you want?' I asked at last.
'You can ask me that?' replied this savage spirit with a sorrow that pierced my soul. I was taken aback, confounded.
'What am I to think of you?' I began again at last.
'Think me what I am!' he answered calmly.
'You need to make apology,' I said in a changed voice, looking at him proudly, 'absolve yourself! cleanse yourself!'
That was too much for him.
'How can it be,' he cried indignantly, 'that this fellow should bend me to his every whim? — It's true, they'd let me out of school too early, I'd dragged all the chains and torn them all asunder, there was just one I'd missed, only one still left to smash, I hadn't yet been chastised by a fantast — go on, murmur all you like! I've kept silent long enough!'
'O Alabanda! Alabanda!' I cried.
'Be still,' he replied, 'and don't use my name against me like a dagger!'
Now my own resentment exploded in its turn. We didn't let it rest till there was hardly any going back. We violently destroyed the garden of our love. We often stopped and stood there saying nothing, and would so gladly have fallen on each other's necks with thousand joys, but wretched pride stifled every sound of love that rose up from the heart.
'Farewell!' I cried at last and rushed away. Despite myself I had to look back, despite himself Alabanda had come after me.
'That's a strange beggar, isn't it, Alabanda?' I called out to him, 'his last farthing and he flings it in the mire!'
'And if that's the case may starve for all I care,' he cried and went his way.
Stunned I staggered on and presently was standing by the sea, gazing onto the waves — oh! down strove my heart, down into the deep, and my arms flew out towards the free flowing surge; but soon, as if from heaven, a gentler spirit came over me and with its peaceful wand brought order to the turmoil of my sorrowing soul; more tranquilly I pondered my fate, my faith in the world, my dismal experiences, I reflected on my fellow men, as from early youth I'd come to know and apprehend them, brought up in their multifarious ways, and everywhere found dull or strident discord; only in simple childlike limitation did I still find pure melodies — 'it's better,' I told myself, 'to become a bee and build one's house in innocence than to rule with the rulers of the world and howl with them as with wolves, than to lord it over whole peoples and sully one's hands on such polluted material.' I wanted to return to Tinos and live for my gardens and fields.
Smile all you like! I was very much in earnest. For the life of the world consists in alternate unfolding and closing, issuing forth and returning to self, why not too the heart of man?
It's true this new lesson was hard for me to accept, it's true I was loath to leave behind the proud errancy of my youth — for who gladly tears off his wings? — but it had to be!
I went through with it. I was now really embarked. A fresh mountain breeze carried me out of Smyrna's harbour. In wonderful tranquillity, just like a child knowing nothing of the next moment, I lay there on my ship and gazed on the trees and mosques of the city, my green walks along the shore, my footpath up to the acropolis, all this I gazed upon, and let it draw away, further and further away; but now, as I came out onto the high sea and everything was slowly sinking lower, like a coffin into the grave, suddenly then it was as if my heart were broken — 'O heaven!' I cried, and all the life in me awoke and struggled to hold back the fleeing present, but it was gone, gone!
Like a mist it lay before me, that heavenly land where, like a hind let loose on open pasture, far and wide I'd roamed over vales and heights and brought the echo of my heart to springs and streams, to breadths and depths of the earth.
Yonder onto Tmolus I'd walked in lonely innocence; down there to where Ephesus once stood in happy youth and Teos and Miletus, up there into the holy grieving Troad I'd wandered with Alabanda — with Alabanda — and like a god I'd ruled over him, and like a child, tender and trusting, I'd done his eye's bidding with joy in my soul, full of jubilant delight in his being, always happy, whether holding his horse's bridle for him, or when, exalted above myself, in glorious resolve, in daring thoughts, in fiery eloquence, I matched his soul!
And now it had come to this, now I was nothing, I'd been so hopelessly despoiled of everything, become the poorest among men and didn't even know how.
'O endless erring!' I thought to myself, 'when will man tear himself free from your chains?
'We speak of our hearts, our plans, as if they were ours, and yet it's a power outside ourselves that tosses us about and lays us in the grave just as it pleases, and we know neither whence it comes nor whither it goes.
'We'd like to grow high up here and spread our boughs and branches out there, and yet it's soil and weather that decide the way it goes with us, and if the lightning falls upon your crown and splits you down to the roots, poor tree! what has it to do with you?'
That's what I thought. Does it anger you, my Bellarmin? Just wait to hear what's still to come.
You see, the sad thing, my dear friend! is that the mind so readily takes on the errant heart's complexion, so readily holds fast to fleeting sorrow, that thought, which ought to heal the pains, becomes itself sick, that the gardener whose task it is to plant the rosebushes so often lacerates his hand on them, oh! it's this that's made many a man a fool before those whom, like another Orpheus, he'd otherwise have ruled, it's this that's so often made the noblest nature an object of derision before such men as can be found in any street, this is the rock on which heaven's favourites founder, that their love is powerful and tender as their mind, that often their heart's waves move more strongly and more swiftly than the trident with which the sea-god rules them, and therefore, my dear friend, let no man get above himself!
#### Hyperion to Bellarmin [VIII]
Can you bear to hear it, will you be able to grasp it when I tell you of my long sick sorrow?
Take me as I give myself and consider that it's better to die because one lived than to live because one never lived! Don't envy the griefless, the idols of wood who lack nothing because they're so poor in soul, who care nothing whether it's rain or shine because they've nothing that needs nurturing.
Yes! yes! it's oh so easy to be happy and at peace when you've a shallow heart and a narrow spirit. You're welcome to it; for who is bothered that the wooden target doesn't moan when the arrow strikes it, and that the empty pot makes so hollow a sound when flung against the wall?
Only you ought to learn your limits, dear people, ought to wonder in all silence if you cannot grasp that others aren't so happy, so selfcontented, ought to be wary of making your wisdom law, for if you were obeyed that would be the end of the world.
I was living now a very still and simple life on Tinos. And I really did let the world's appearances pass by like autumn mists, sometimes laughing moist-eyed at my heart when it flew out for a morsel, like a bird to painted grapes, and I stayed still and friendly all the while.
I now happily left every man his opinions, his foibles. I was converted, I no longer wanted to convert others, only it saddened me to see that people thought I left their farce unchallenged because I set such great store by it as they. It's not so much that I wished to submit to their silliness, but I tried to spare their feelings where I could. 'It gives them joy after all,' I'd think, 'it's what keeps them going!'
Often I even suffered myself to join in, and no matter how spiritless and impassive my presence, nobody noticed, nothing whatever was found wanting, and had I begged forgiveness for my demeanour, they'd have stood there in wonder and asked: 'What have you done to offend us then?' How very tolerant!
Often when I stood by my window of a morning and the bustling day approached me, I too could forget myself a moment, could look about me as if eager to engage in some activity in which my being still found its pleasure as before, but then I would chide myself, then I'd bethink myself, like a man who lets slip sounds from his mother tongue in a land where no one understands it — 'whither, my heart?' I'd sensibly tell myself and obey.
'What does it mean that man desires so much?' I often asked, 'why this infinity in his breast? infinity? where is it then? who has perceived it then? He wants more than he can have! that's surely true! Oh! that's something you've experienced often enough. And that's the way it must be too. That's what imparts the sweet intoxicating sense of energy that it doesn't flow forth as it will, that's what makes the beautiful dreams of immortality and all the lovely and colossal phantoms that enrapture man a thousandfold, that's what creates for man his Elysium and his gods that his life's line doesn't run straight, that his path isn't true as an arrow's and an alien force impedes him in his flight.
'The heart's wave wouldn't foam up so beautifully, becoming spirit, if fate, that ancient silent rock, didn't stand in its way.
'But yet the impulse dies within our breast, and with it our gods and their heaven.
'The fire leaps up in joyous shapes from out of the dark cradle where it slept, and its flame rises and falls and breaks apart and wraps around itself again in joy, until its fuel is spent, then it smokes and writhes and expires; what's left behind is ash.
'So it is with us. That's the essence of everything the sages tell us in mysteries both awful and alluring.
'And you? why ask yourself? That there are times when something rears up within you, and your heart, like the mouth of the dying man, in a single moment opens and closes so violently, that's the truly fatal sign.
'Just be still and let things take their course! no more futile industry! don't childishly try to add a cubit to your stature! — It's like wanting to create another sun, and spawn for it new satellites, an earth and a moon.'
So I dreamed on. Patiently, little by little I took my leave of everything. — O you who share with me this age! don't turn to your doctors, don't turn to the priests, when you find yourselves inwardly dwindling!
You've lost faith in all that is great; so you're bound, you're bound to waste away, unless this faith return like a comet from alien heavens.
#### Hyperion to Bellarmin [IX]
There's a forgetting of all existence, a silencing of our being, when we feel we've found everything.
There's a silencing, a forgetting of all existence, when we feel we've lost everything, a night of the soul, when no shimmer of a star, when not even a rotting log can give us light.
I'd now become calm. Now nothing roused me any more at midnight. Now I no longer singed myself in my own flame.
Lone and still I now looked straight before me, no roving with my eye in past and future. Now things far and near no longer jostled in my mind; unless men forced me to, I didn't see them.
Once this age would often lie before my mind like the eternally empty vessel of the Danaids, and with prodigal love my soul poured itself out to fill up the holes; now I saw no more holes, now life's tedium bore down on me no more.
Now never more did I say to the flower, 'you're my sister!' nor to the springs, 'we're of a kind!' Now, like an echo, I faithfully gave to every thing its name.
Like a stream past arid banks, where no willow-leaf mirrors in the water, the world untouched by beauty passed me by.
## Hyperion to Bellarmin [X]
Nothing can flourish and nothing so deeply dwindle as man. To the night of the abyss he often likens his suffering and to the aether his bliss, and how little does that tell us?
Yet there is nothing lovelier than when it begins to dawn in him once more, after a long death, and pain goes forth like a brother to meet distantly dawning joy.
Oh, it was with heavenly anticipation I now greeted the coming spring! As from afar on a silent breeze the sounds of the loved one's lyre when all are sleeping, so spring's soft melodies wafted round my breast, as if from Elysium I sensed its coming when the dead branches stirred and a gentle breath caressed my cheek.
Lovely heaven of Ionia! never had I so cleaved to you, but then never in its blithe and tender play had my heart been so akin to you as now. —
Who doesn't yearn for the joys of love and great deeds when in the eye of heaven and the bosom of the earth the spring returns?
I rose as from my sickbed, softly and slowly, but my breast trembled so blissfully with secret hopes that I forgot to ask myself what this might signify.
Now in sleep more beautiful dreams enfolded me, and when I woke they were in my heart like the trace of a kiss on the cheek of the beloved. Oh, the morning light and I, we now approached each other like friends freshly reconciled, when they still affect aloofness and yet already hold within their soul the nearing endless moment of embrace.
And indeed my eye opened once again, though no longer as before, steeled and full of its own power; it had become more beseeching, it begged for life, but deep within I still felt things could again be with me as before, and better.
I beheld again my fellow men, as if I too looked to play my part and find my joy amongst them. I truly joined in everywhere with a whole heart.
Heavens! what gloating glee there was that the proud maverick had ended up as one of them! what a joke: the stag of the forest driven by hunger to come running into their barnyard! —
Oh! it's my Adamas I was searching for, my Alabanda, but there appeared to me none.
Finally I wrote to Smyrna, and as I wrote, it was as if all the tenderness and might of man were compressed into a single moment; three times so I wrote, but no reply, I begged, threatened, minded him of all the hours of love and boldness, but no reply from him, the unforgettable, loved unto death — 'Alabanda!' I cried, 'O my Alabanda! you've broken the staff over me. You kept me upright, you were the last hope of my youth! There's now nothing left for me! that's now holy and certain!'
We pity the dead as if they felt death, and yet the dead have peace. But this, this is the pain beyond all pain, this is relentless feeling of total annihilation when our life so loses all meaning, when the heart says to itself, 'down you must go and nothing will remain of you; no flower have you planted, no hut have you built, only to be able to say: I leave behind a trace on earth.' Oh! and yet the soul can still be so full of yearning, for all its despondency!
I was always searching for something, but I didn't dare look up in human company. I had hours I could be frightened by the laughter of a child.
Yet mostly I was very still and patient, and often had a wondrous foolish faith in the healing powers of certain things; from a dove that I bought, from a trip in a boat, from a valley the mountains concealed from me, I could promise myself solace.
Enough! enough! had I grown up with Themistocles, had I lived amongst the Scipios, believe me, never would my soul have come to know itself from this side.
#### Hyperion to Bellarmin [XI]
Still there were times when mind and spirit stirred in me. But all to only ravaging effect!
'What is man?' I could begin; 'how can it be that such a thing is in the world that ferments like a chaos or moulders like a rotting tree, and never grows to ripeness. How can nature suffer this sour grape amid her sweet clusters?
'To the plants he says, "I too was once like you!" and to the pure stars, "I will become like you, in another world!" Meanwhile he breaks asunder and keeps practising his arts upon himself, as if he could, once it's come apart, put a living thing together again like a wall of bricks; but it doesn't even trouble him when nothing is improved by all his efforts; what he works will always stay mere artifice.
'Oh, you poor wights who feel this, you who too are loath to talk of man's vocation, who too are gripped so through and through by the Nothing that sways over us, so clearly see that we are born for Nothing, what we love is Nothing, what we believe in is Nothing, that we grind ourselves down for Nothing, only to slowly pass over into Nothing — can I help it that your knees give way when you take it all in? For more than once I too have been brought low in such thoughts and have cried out, "why do you lay the axe to my root, cruel spirit?", and still I'm here.
'Oh once, you dismal brethren! it was different. Then everything above us was so beautiful, so beautiful and blithe was all before us; these hearts too welled over before the distant blissful phantoms, and in bold exuberance our spirits too forged upwards and broke through the bounds, and when they looked around, oh woe, all was endless void.
'Oh! I can fling myself on my knees and wring my hands and beseech I don't know whom? for other thoughts. But I can't overcome it, the screaming truth. Have I not convinced myself twice over? When I look into life, what's at the end of it all? Nothing. When I soar in the spirit, what's at the summit of all? Nothing.
'But be still, my heart! It's your last strength you're wasting! your last strength? and you, you would storm heaven? where then are your hundred arms, Titan, where your Pelion and Ossa, your stairway to the fortress of the father of the gods, so that you may climb up and cast down the god and his table and all the immortal peaks of Olympus, and preach to mortal men: "Stay below, children of the moment! don't strive to scale these heights, for there's nothing up here."
'And you yourself can leave off trying to see what sways the lives of others. For you your new lesson applies. Above you and before you all is waste and void, and that's because within you all is waste and void.
'And indeed if you others are richer than I, you could at least help just a little.
'If your garden's so full of flowers, why does their breath not rejoice me too? — If you're so full of divinity, give me to drink of it. For no one starves at a feast, not even the poorest of men. But there's only one who holds his feasts amongst you; and that is death.
'Need and fear and night are your masters. They sunder you, they drive you together with blows. Hunger you call love, and where you can see nothing more, there dwell your gods. Gods and love?
'Oh, the poets are right, there's nought so small and trifling that it can't be made the stuff of inspiration.'
That's what I thought. How it came to be in me I still can't comprehend.
## Hyperion to Bellarmin [XII]
I'm living now on the island of Ajax, beloved Salamis.
I love this Greece all over. It bears the colour of my heart. Wherever you look a joy lies buried.
And yet around you is so much loveliness and grandeur too.
On the headland I've built me a hut of mastic twigs, and round about it planted moss and trees and thyme and all manner of shrubs.
There I spend my sweetest hours, there I sit whole evenings and gaze across to Attica, until at last my heart begins to race; then I take my tackle, go down to the bay and catch some fish.
Or up there on my promontory I read about the glorious ancient sea battle that once in wild and wisely managed turmoil spent its rage at Salamis, and take my pleasure in the spirit that could steer and tame the furious chaos of friend and foe, like a rider his steed, and feel deep shame at my own career as warrior.
Or I look out upon the sea and ponder my life, its rising and sinking, its bliss and its grief, and my past often sounds to me like a lyre when the master runs through all the tones, throwing together discord and harmony in hidden order.
It's thrice beautiful up here today. Two gentle days of rain have cooled the air and the life-weary earth.
The ground has grown greener, the field has unfolded. Endless stands the golden wheat, mingled with joyful cornflowers, and light and bright a thousand hopeful crowns rise from the depths of the grove. Gently and grandly each line in the distance weaves through the space; like steps the mountains climb up to the sun, one after another in ceaseless procession. All heaven is limpid. The white light only softly suffuses the aether, and like a silver cloudlet the bashful moon goes floating by in the brightness of day.
#### Hyperion to Bellarmin [XIII]
It's been a long time since I felt as I do now.
Like Jupiter's eagle to the song of the muses I hark to the wonderful endless harmony in me. Untroubled in mind and soul, strong and blithe, in smiling earnest I sport in spirit with fate and the three sisters, the holy Parcae. Full of godly youth my whole being exults in itself, in all things. Like the starry heavens I'm both still and moved.
I've waited long for such a festal spell as this to write to you once more. Now I'm strong enough; now let me tell you.
In the midst of my dark days an acquaintance from Calauria invited me across. I should come into his mountains, he wrote: life there was lived more freely than elsewhere, and besides, amongst the pinewoods and rushing waters, there bloomed lemon groves and palms and sweet herbs and myrtles and the holy vine. High in the mountains he'd built a garden and a house, shaded by lush trees at the rear, and in the searing summer days cooling breezes softly played about it; like a bird from the top of the cedar you could look down into the depths, to the villages and green hills and contented flocks of the island that lay all around the magnificent mountain like children and nourished themselves from its foaming streams.
This at least roused me a little. It was a bright blue April day when I sailed across. The sea was unusually beautiful and limpid, and light the air as in higher regions. In the gliding craft you left the earth behind you, like a luscious dish when holy wine is offered.
A gloomy mood is no match for the influence of sea and air. I gave myself up, fretted no more about myself and others, sought for nothing, pondered nothing, let myself be rocked into half-slumber by the boat, imagining I lay in Charon's bark. Oh, it's sweet to drink like this from the chalice of oblivion.
My cheerful boatman would have liked to chat with me, but I kept very taciturn.
He pointed with his finger to show me right and left this or that blue island, but I didn't look for long and next moment was back in my own dear dreams.
But when he pointed out to me the tranquil peaks in the distance and said that we should soon be in Calauria, I paid at last more heed, and my whole being opened itself to the wonderful power that all of a sudden sweetly and peacefully and inexplicably played with me. Wideeyed, marvelling and full of joy I gazed out into the mysteries of the distance, my heart trembled lightly and a hand escaped and clutched my boatman in friendly haste — 'Really?' I cried, 'that's Calauria?' And when he gave me an odd look, I didn't know what to make of myself either. I greeted my friend with wonderful tenderness. My whole being was full of sweet unrest.
That very same afternoon I wanted to roam through part of the island. The woods and secret valleys attracted me ineffably, and the clement day lured all out into the open.
It was so plain to see how everything that lives craves more than its daily bread, how even the birds and the beasts have their feast days.
It was a rapturous sight! As when the mother teasingly asks wherever her dearest might be, and all of the children rush into her lap, and even the smallest stretches its arms from the cradle, so flew and leapt and strove all life into the heavenly air, and beetles and swallows and doves and storks cavorted together in the heights and the depths in joyous confusion, and what the earth held fast found its steps turning to flight, the horse went hurtling over the ditches and over the hedges the deer, and from the ground of the sea the fishes rose up and hopped above the surface. The motherly air infused the hearts of all, uplifted them and drew them to itself.
And the people came out of their doors and felt wonderfully the spiritual breath as gently it stirred the wisps of hair on their brows, as it cooled the streaming sunlight, and benignly they loosened their clothing to take it to their bosom, breathed more sweetly, touched more tenderly the light and lucent soothing sea in which they lived and moved.
O sister of the spirit which lives and sways in us with fiery power, O holy air! how beautiful it is that, all-pervasive and immortal, you're there with me wherever I may wander!
It's with the children that the sublime element played most beautifully.
This one hummed happily to himself, a tuneless ditty slipped from the lips of that one, and from another came full-throated cries of joy; this one stretched, that one leapt high; another strolled round, lost in thought.
And all this was the language of one single sense of well-being, all one single response to the caresses of the rapturous breezes.
I was filled with ineffable longing and peace. A strange power swayed me. 'Friendly spirit,' I said to myself, 'whither do you call me? to Elysium or where?'
I went walking in a wood, up by rippling water where it trickled over rocks, where it glided harmless over pebbles, and step by step the valley narrowed, becoming an archway, and lonely the noonday light played in the silent gloom —
Here — would that I could speak, my Bellarmin! that I could write to you calmly!
Speak? oh, I'm a layman in joy, I will speak!
For stillness dwells in the land of the blissful, and above the stars the heart forgets its anguish and its language.
I've kept it sacred! like a palladium I've carried it within me, the divine that was revealed to me! and if fate henceforth should seize and plunge me down from abyss to abyss and drown in me all energy and all reason, yet shall this one and only outlive myself in me and shine in me and reign in eternal, indestructible glory! —
So you lay there lissomely, sweet life, so you looked up, rose and stood in slender fullness, divinely calm, your heavenly face still filled with the serene rapture I'd disturbed!
Oh, he who's gazed into the stillness of these eyes, he for whom these sweet lips have opened, of what more shall he speak?
Peace of beauty! divine peace! he who's once soothed in you his seething life and doubting spirit, how can ought else help him?
I cannot speak of her, but there are hours when the best and the most beautiful appears as in clouds, and the heaven of perfection opens before the intuition of love, then, Bellarmin! then think of her being, then go down on your knees with me, and think of my bliss! but don't forget that I possessed what you can only sense, that with these eyes I saw what only as in clouds appears to you.
That men will sometimes say that they feel joyful! Oh, believe you me, you've never had even an inkling of joy! Not the shadow of its shadow has appeared to you! Oh, go, you blindlings, and speak no more of the blue aether!
That we can become as children, that still the golden time of innocence returns, the time of peace and freedom, that there's yet one joy, one resting place on earth!
Is man not grown old and withered, is he not like a fallen leaf that can't find back to its stem and is whirled around with the winds until it's buried by the sand?
And yet his spring will come again!
Do not weep when the best decays! soon it will renew itself! Do not grieve when your heart's melody falls silent! soon there'll come a hand to make it sound again!
How then was it with me? was I not like a shattered lyre? I sounded still a little, but they were death tones. I'd sung myself a sombre swansong! I'd gladly have woven myself a funeral wreath, but I had only winter flowers.
And where then was it now, the deathly silence, the night and desolation of my life? all the beggarly mortality?
It's true that life's a poor and lonely thing. We dwell here in the deep like the diamond in the mine. We ask in vain how we came down so that we may find the way back up again.
We're like fire that slumbers in the tinder-dry branch or the flint; and all the while we writhe and seek the end of close confinement. But they come, they make up for eons of struggle, those moments of liberation, when the divine bursts open the dungeon, when the flame frees itself from the wood and surges victorious over the ash, ha! when we feel as if the unbound spirit, forgotten its suffering, its servile form, were returning in triumph to the halls of the sun.
## Hyperion to Bellarmin [XIV]
Once I was happy, Bellarmin! Am I not still? Would I not be even if that hallowed moment when I first saw her had been the last?
I have seen it once, the one and only that my soul was seeking, and the consummation we project beyond the stars, that we defer until the end of time, I've felt it in the here and now. It was here, the most high, in this cycle of human nature and of things, it was here!
I no longer ask where it is; it was in the world, it can return in the world, it's in the world now, only more hidden. I no longer ask what it is; I've seen it, I've known it.
O you who seek the highest and the best, be it in the depths of knowledge, in the turmoil of action, in the darkness of the past, in the labyrinth of the future, in the graves or above the stars! do you know its name? the name of what is one and all?
Its name is beauty.
Did you know what you were seeking? I don't yet know it, but I can sense it, the new kingdom of the new godhead, and rushing towards it lay hold of the others and carry them with me, like the river the rivers into the ocean.
And it was you, my love, who showed me the way! With you I began. They're not worthy of words, the days when I didn't yet know you —
O Diotima, Diotima, heavenly being!
#### Hyperion to Bellarmin [XV]
Let's forget that time exists and not count life's days!
What are centuries to the moment when two beings so sense and near one another?
I still see the evening when Notara took me for the first time to her house.
She lived only a few hundred steps from us at the foot of the mountain.
Her mother was a thoughtful tender being, a simple cheerful lad the brother, and both conceded happily in all their doings that Diotima was the household's queen.
Oh! everything was hallowed, made beautiful by her presence. Wherever I looked, whatever I touched, her carpet, her cushion, her little table, everything was in secret unison with her. And when she called me by name for the first time, when she even came so close to me that her innocent breath touched my listening being! —
We spoke very little together. You're ashamed of your language. You'd like to become pure tone and blend with the other in one celestial song.
And what was there to speak of? It's only each other we saw. To speak of ourselves we were shy.
We spoke at last of the life of the earth.
So fiery and childlike a hymn has never been sung to her.
It did us good to strew our hearts' abundance in the good mother's lap. We felt ourselves lightened like the trees when the summer wind shakes their fruitful branches and pours their sweet apples into the grass.
We called the earth one of the flowers of heaven, and heaven we called the infinite garden of life. As the roses gladden each other with golden dust, we said, so the heroic sunlight gladdens the earth with its rays; she is a glorious living being, we said, equally divine whether raging fire or mild clear water spring from her heart, ever happy whether she nourish herself on dew drops or storm clouds, prepared for her pleasure with heaven's help, she, the ever more truly loving half of the sun-god, in the beginning more inly one with him perhaps, but then sundered from him by an almighty fate, so that she might seek him, come close, draw away, and in joy and grief ripen to supreme beauty.
So we spoke. I give you the substance, the spirit of it. But what's that without the life?
It was growing dark and we had to leave. 'Good night, you angel eyes!' I thought in my heart, 'and do soon appear to me again, lovely divine spirit, with your peace and fullness!'
#### Hyperion to Bellarmin [XVI]
Some days later they came up to visit us. We walked around together in the garden. Diotima and I strayed ahead, immersed, and tears of joy would come to my eyes over the hallowed being who walked so unassumingly beside me.
Now we stood close by the edge of the mountain's summit and gazed out into the endless east.
Diotima's eye opened wide and softly as a bud unfolds, her lovely little face unfolded before the breezes of heaven, became pure speech and soul, and as if about to fly into the clouds, her whole figure stood there stretching lightly up in airy majesty and her feet barely touched the earth.
Oh, I could have clasped her under the arms, like the eagle his Ganymede, and with her flown away above the sea and its islands.
Now she stepped further forward and gazed down the sheer cliff face. She took her delight in plumbing the awesome depth and losing herself in the night of the forests that stretched up their bright crowns from broken rocks and foaming rain-swollen streams below.
The rail on which she leaned was rather low. And so she let me hold her a little, the delightful creature, as she bent forward. Oh! hot trembling bliss swept through my being and sent all my senses reeling and raging, and my hands burned like coals when I touched her.
And then my heart's joy at standing beside her in such closeness, and my gentle childish fear that she might fall, and my delight in the inspiration of the glorious maiden!
What is all that men have done and thought in whole millennia compared to a single moment of love? For is it not of all the things in nature the most consummate, the most divinely beautiful! it's thither all steps lead on the threshold of life. Thence we come and thither we go.
#### Hyperion to Bellarmin [XVII]
Only her song should I forget, only these sounds from the soul should never more return in my ceaseless dreams.
One doesn't know the proudly-sailing swan when it sits slumbering on the bank.
Only when she sang did one truly know this loving silent being who was so loath to speak.
Then, only then, did she, heavenly reticent, come forth in all her majesty and loveliness; then it breathed out from her tender blossoming lips, often so coaxingly and so caressingly, often like the commandment of a god. And how the heart stirred in that divine voice, how all greatness and humility, all the joy and all the grief of life seemed graced in the nobility of those tones!
As the swallow snatches bees in flight, so she always seized us all.
It wasn't pleasure, it wasn't admiration, it was the peace of heaven that came among us.
A thousand times I've told her and myself: the most beautiful is also the most hallowed. And so was everything about her. As her song, so too her life.
#### Hyperion to Bellarmin [XVIII]
Amongst the flowers her heart was at home, as if it were one of them. She called them all by name, made them new and more beautiful names out of love and unerringly knew the happiest season of each.
Like a sister when from every nook a loved one comes clamouring, each wanting to be greeted first, so this still being was busy with hand and eye, blissfully distracted, whenever we walked through wood or meadow.
And this wasn't at all learned or assumed, in her it had naturally grown.
For it's a truth eternally valid and everywhere manifest: the more innocent and beautiful a soul, the closer the bond with those other happy lives that we call soulless.
#### Hyperion to Bellarmin [XIX]
A thousand times I've laughed in the joy of my heart at those who imagine a sublime spirit cannot possibly know how to prepare a vegetable. If the occasion arose Diotima was quite capable of speaking with real passion of the hearth, and there's surely nothing nobler than a noble maiden who tends the wholesome flame and, akin to nature herself, prepares the food that gladdens the heart.
#### Hyperion to Bellarmin [XX]
What's all the artful knowledge in the world, what's the whole proud perfection of human thought, compared to the unstudied tones of this spirit that knew neither what it knew nor what it was?
Who wouldn't rather have their grape in a cluster, full and fresh, just as it sprang from the root, than the dried plucked berries that the merchant crams into the crate and sends out into the world? What's the wisdom of a book compared to the wisdom of an angel?
She seemed always to say so little, and said so much.
Once I walked her home in the gloaming; like dreams wisps of dewy mist came stealing over the meadow, like eavesdropping genii the blissful stars peeped through the boughs.
You seldom heard a 'how beautiful!' from her lips, even though this pious heart left no lisping leaf, no murmur of a burn unheeded.
But this time she did indeed speak it out to me — 'how beautiful!'
'Well, that will be for our sake!' I said, rather in the way that children say things, neither in jest nor in earnest.
'I can imagine what you mean,' she responded, 'I best like to picture the world to myself as the life of a couthie household where everyone fits in with the others without even having to think about it, and we live for the pleasure and joy of each other because that's how it comes from the heart.'
'Happy, sublime faith!' I cried.
She was silent a while.
'So we too are children of the house,' I finally resumed, 'we are and we shall be.'
'Shall for ever be,' she responded.
'Shall we?' I asked.
'In this I put my trust in nature,' she continued, 'as I do every day.'
Oh, to have been Diotima when she said this! But you don't know what she said, my Bellarmin! You didn't see it and didn't hear it.
'You're right,' I cried to her; 'eternal beauty, nature suffers no loss in herself, just as she suffers no addition. Her costume will be tomorrow different from today; but our best, us, us she cannot do without and you least of all. We believe we're eternal, for our soul feels the beauty of nature. She'll be mere patchwork, no longer divine, no longer complete, if ever you're missing in her. She doesn't deserve your heart if she has to blush before your hopes.'
#### Hyperion to Bellarmin [XXI]
I've never known anything so free from wants, so divinely self-sufficient.
As the wave of the ocean round the shores of blissful islands, so my restless heart swirled round the peace of the heavenly maiden.
I had nothing to give her but a spirit full of wild contradictions, full of bleeding memories, nothing to give but my unbounded love with its thousand cares, its thousand tumultuous hopes; but there she stood before me in changeless beauty, effortlessly, in smiling perfection, and all the yearning, all the dreaming of mortality, oh! everything that in golden hours of morning the genius foretells of higher spheres, was all fulfilled in this single tranquil soul.
They say the struggle abates above the stars, and it's only in the future, so they promise us, once our yeast has sunk, that fermenting life will turn into the noble wine of joy; otherwise the blessed's peace of heart is now nowhere to be sought upon this earth. I know different. I've come the shorter way. I stood before her and heard and saw the peace of heaven, and in the midst of groaning chaos there appeared to me Urania.
How often have I stilled my sorrowing before this image! how often have overweening life and striving spirit been assuaged when, sunk in blissful contemplation, I looked into her heart, as one looks into the source when it softly trembles from the touches of heaven, trickling down on it in silver droplets!
She was my Lethe, this soul, my holy Lethe, from which I drank oblivion of existence, so that I stood before her like an immortal, and joyfully scolded myself, and as after heavy dreams could not but smile at all the chains that had borne me down.
Oh, with her I should have become a happy man, a great man!
With her! but that has gone awry, and now I err around in what's before and in me and beyond, and don't know what to make of myself and other things.
My soul is like a fish cast out of its element upon the sand of the shore, and writhing and thrashing about till it shrivels up in the heat of the day.
Oh! if only there were still something in the world for me to do! if there were some work, a war for me, that should quicken me!
Babes torn from their mother's breasts and cast into the desert were once, so they say, suckled by a wolf.
My heart is not so lucky.
#### Hyperion to Bellarmin [XXII]
All that I can speak of her is scattered words. I must forget what she is whole if I'm to speak of her. I must make believe she lived in times of old, that I knew of her through tales, if her living image isn't so to seize me that I perish in rapture and pain, if I'm not to die of my joy in her and my grief for her.
#### Hyperion to Bellarmin [XXIII]
It's no use; I can't conceal it from myself. Whither I flee with my thoughts, up to the heavens and into the abyss, to the beginning and end of time, even to him who was my last refuge, who'd once consume every last care in me, who'd once sear all life's pleasure and all its pain in me with the flame of fire in which he revealed himself, even if I throw myself into his arms, the glorious secret spirit of the world, dive down into his depths as into the bottomless ocean, even there, even there shall the sweet dread find me out, the sweet bewildering lethal dread, that Diotima's grave is near.
Do you hear? do you hear? Diotima's grave!
Yet my heart had grown so still, and my love lay buried with the dead one I loved.
You know, my Bellarmin! for long I didn't write to you of her, and when I wrote, I wrote to you calmly, I think.
And now what?
I go down to the shore and gaze out towards Calauria where she rests, that's what.
Oh, let no one lend his boat to me, let no one take pity and offer me his oar and help me across to her!
Let the goodly waters not stay calm, lest I hew myself a piece of timber and swim across to her!
But I will fling myself into the raging sea and beg its wave to cast me up on Diotima's shore! —
Dear brother! I comfort my heart with all kinds of fancies, I grant myself many a sleeping draught; and it would no doubt be nobler to free oneself forever than to make do with palliatives; but isn't that the way it is with all of us? So I'm content it should be so.
Content? oh, that would be a fine thing! that would be help indeed where never a god can help.
Now! now! I've done what I could! I demand from fate my soul.
## Hyperion to Bellarmin [XXIV]
Was she not mine, you sisters of fate, was she not mine? The pure springs I call to witness and the innocent trees that overheard us, and the daylight and the aether! was she not mine? not one with me in all the chords of life?
Where is the being that knew her like mine? in what mirror were gathered the rays of that light as in me? was she not joyfully startled at her own glory when first she became aware of herself in my joy? Oh! where is the heart that, like mine, was everywhere close to her, that, like mine, filled her and was filled by her, that was there solely to embrace hers, as the eyelid is there for the eye.
We were but a single flower, and our souls lived in one another like the flower when it loves and conceals its tender joys in its closed chalice.
And yet, and yet was she not torn from me like a crown usurped, and laid into the dust?
## Hyperion to Bellarmin [XXV]
Before either of us knew it we belonged to one another.
When, with all the homage of my heart, blissfully overcome, I stood before her saying nothing, and my whole life gave itself up in the beams of the eye that saw only her, embraced only her, and she gazed at me in turn, tenderly doubting, not knowing where my thoughts were taking me, and when, buried in beauty and delight, I often discreetly watched her at some charming occupation, and my soul flitted and flew about her slightest movement as the bee about the swaying twigs, and when then in peaceful thought she turned towards me and, startled by my joy, had to conceal my joy from herself, and sought and found tranquillity again in her beloved task —
When, wondrously all-knowing, she revealed to me every harmony, every discord in the depths of my being at the moment it began, before I even noticed it myself, when she saw every shadow of a cloudlet on my brow, every shadow of sadness, of pride on my lips, every spark in my eye, when she caught the ebb and flow of my heart, and caringly sensed gloomy hours approaching as my spirit too intemperately and prodigally consumed itself in tumid talk, when the dear being, more faithfully than a mirror, betrayed to me every change in my cheek, and often reproved me in loving concern at my unstable being, and chided me like a cherished child —
Oh! when once, you innocent, you counted on your fingers the steps from our mountain down to your house, when you showed me your walks, the places where you used to sit and told me how the time had passed for you there, and finally said that now you felt as if I'd always been there too —
Had we not then long since belonged to one another?
## Hyperion to Bellarmin [XXVI]
I'm building a grave for my heart, that it may rest; I spin a cocoon around myself, because everywhere it's winter, in blissful memories I wrap myself against the storm.
Once we sat in Diotima's garden under blossoming almond trees, together with Notara — that's the friend with whom I was staying — and some others who also, like us, belonged to the mavericks in Calauria, and we spoke amongst other things of friendship.
I'd contributed little; for some time now I'd grown chary of making many words about things that first concern the heart, my Diotima had made me so taciturn —
'When Harmodius and Aristogeiton lived,' someone cried at last, 'there was still friendship in the world.' That delighted me too much for me to keep my peace.
'They should weave you a crown for those words!' I cried to him; 'do you really have an inkling, do you have a likeness for the friendship of Aristogeiton and Harmodius? Forgive me! But by the Aether! you'd have to be Aristogeiton to feel how Aristogeiton loved, and it didn't do to be afraid of lightning if a man would be loved with the love of Harmodius, for unless I'm much mistaken the terrible youth loved with the rigour of Minos. Few could withstand such a test, and it's no easier to be the friend of a demi-god than, like Tantalus, to sit at the table of the gods. But there's also nothing more glorious on earth than when as proud a pair as these are so in thrall to one another.
'And this is my hope, my joy in lonely hours, that in the symphony of the world's course tones as great and greater still must one day come again. Love brought forth millennia full of living men; friendship will bring them forth again. From childhood harmony the peoples once set out; with the harmony of spirits a new world history will begin. From the happiness of plants men began and grew up, grew until they ripened; from then on they fermented forth unceasingly, both inwardly and outwardly, till now the human race, in infinite dissolution, lies there like a chaos, and all those who still feel and see are gripped by dizziness; but beauty takes flight from the life of men, up into the spirit; what was nature becomes the ideal, and though the tree may be withered and weathered from below, a fresh crown has still sprung from it and greens in the sun's glory, as the stem once did in the days of youth; what was nature is the ideal. And it's by this ideal, this rejuvenated godhead, that the few shall know each other and they are one, for the oneness is in them, and from them, these few, shall begin the second age of the world — I've said enough to make clear what I think.'
You should have seen Diotima then as she leapt up and held out both her hands to me and cried: 'I've understood it, dearest, wholly understood, though it says so much.
'Love brought forth the world, friendship will bring it forth again.
'Oh then, you who are to come, you new dioscuri, then linger a little when you pass the place where Hyperion sleeps, linger mindfully over the ashes of the forgotten man, and say: "He'd be one of us were he here now."'
This I have heard, my Bellarmin! this I have known, and won't go willingly to my death?
Yes! yes! I've had my dues already, I have lived. More joy a god could bear, but not me.
## Hyperion to Bellarmin [XXVII]
Do you ask how I felt at this time? Like someone who's lost everything to gain everything.
It's true I often came from Diotima's trees like one drunk with triumph, often I'd have to rush away from her lest I betray any of my thoughts; so raged the joy in me, and the pride, the all-inspiring faith of being loved by Diotima.
Then I'd search out the highest mountains and their winds, and like an eagle whose bleeding wing has healed, my spirit stirred in the boundless air and spread itself out over the visible world as if this were its own; oh, wonder! I often felt as if the things of the earth were purified and fused together in my fire like gold, and a thing divine became of them and me, so raged in me the joy; and how I lifted up the children and pressed them to my throbbing heart, how I greeted the plants and the trees! I could have wished for a magic charm to gather the timid deer and all the wild birds of the forest like a family of little folk around my liberal hands, in such blissful folly did I love all things!
But not for long, then that was all extinguished in me like a light, and I'd sit there mute and mournful as a shade and sought my vanished life. I had no use for lamentation, nor did I want consolation. Hope I cast away, like a cripple whose crutch is hateful to him; I was ashamed to weep; I felt ashamed for existing at all. Yet in the end my pride broke out in tears, the pain I'd gladly have denied became dear to me, and I laid it like a child to my breast.
'No,' cried my heart, 'no, my Diotima! it doesn't hurt. Preserve your peace for yourself and let me go my way. Don't let your tranquillity be disturbed, you lovely star! when below you there's ferment and turbidness.
'Oh, don't let your rose fade, you blissful youth of the gods! Don't let your beauty age in the troubles of the earth. For this is my joy, sweet life! that you bear within you carefree heaven. You shall not suffer want, no, no! you shall not see within yourself the penury of love.'
And when I then went down to her again — I'd have liked to ask the breeze and read in the drift of the clouds how it would be with me in an hour! and how delighted I was when some friendly face met me on the way and, so long as it wasn't too dryly, called out to me his 'Good day!'
When a little girl came out of the woods and held out a bunch of strawberries for me to buy, with a look as though she'd rather make a gift of them, or when a peasant, sitting picking cherries in his tree, called down from the branches as I passed and asked if I'd not care to try a handful: these were good omens for my superstitious heart!
To top it all, if facing the path down which I came, one of Diotima's windows was standing open, how good that could make me feel!
Perhaps she'd looked out not long before.
And now I stood before her, breathless and unsteady, pressing my crossed arms against my heart so as not to feel its trembling, and like the swimmer from rushing waters, my spirit struggled and strove not to go under in endless love.
'What are we talking about just now?' I could cry, 'it's often so hard, you can't find the matter to hold fast your thoughts.'
'Are they taking off into the air again?' replied my Diotima. 'You'll have to bind lead to their wings, or I'll tie them to a string, like the boy his flying dragon, so they can't get away from us.'
The dear girl was trying with a jest to help herself and me, but it was to little avail.
'Yes, yes!' I cried, 'as you like, as you think best — shall I read to you? Your lute will still be tuned from yesterday — I don't have anything to read to you right now —'
'You've promised more than once,' she said, 'to tell me how you lived before we met, won't you tell me now?'
'That's true,' I replied; my heart gladly pounced on this, and so then I told her, as I've told you, of Adamas and my lonely days in Smyrna, of Alabanda and how I was parted from him, and of the baffling sickness of my being before I came over to Calauria — 'now you know it all,' I told her calmly when I'd finished, 'now you will find in me less cause for offence; now you will say,' I added with a smile, 'don't mock this Vulcan if he limps, for twice he's been hurled from heaven to earth by the gods.'
'Hush,' she cried in a choking voice and hid her tears in her handkerchief, 'oh hush, and don't make fun of your fate, of your heart! for I understand it, and better than you.
'Dear — dear Hyperion! there's no helping you, I fear.
'Do you really know,' she continued with heightened voice, 'do you really know what it is you're starving for, the one thing that you lack, what you're searching for like Alpheus his Arethusa, what you've been grieving for in all your grief? It didn't depart just years ago, one can't say precisely when it was there, when it went away, but it was, it is, in you it is! It's a better age you're seeking, a more beautiful world. It was only that world you embraced in your friends, with them you were that world.
'In Adamas it had dawned for you; it had also departed with him. In Alabanda its light appeared to you a second time, only more fierce and fervid, and that's why it was like midnight for your soul when he left your life.
'And do you see now why the smallest doubt about Alabanda was bound to become despair in you? why you only cast him off because he wasn't quite a god?
'You didn't want human beings, believe me, you wanted a world. The loss of all the golden centuries, as you felt them, compressed into a single happy moment, the spirit of all the spirits of better times, the energy of all the energies of the heroes, all this was to be made good for you by a single mere man! — Do you see now how poor, how rich you are? why you should be so proud and also so cast down? why for you joy and grief alternate so terribly?
'It's because you have everything and nothing, because the phantom of the golden days which are to come is yours and yet isn't there, because you're a citizen in the regions of justice and beauty, a god amongst gods in the beautiful dreams that steal upon you by day, and awake to find yourself standing on modern Greek soil.
'Twice, you said? oh, you're hurled from heaven to earth seventy times in a single day. Shall I say it? I fear for you, you'll scarce be able to bear the fate of these times. There are diverse things you'll still try, you'll —
'Oh God! and your final refuge will be a grave.'
'No, Diotima,' I cried, 'no, by heaven, no! As long as one melody still sounds for me, I'll not fear the deathly stillness of the wilderness beneath the stars; as long as the sun only shines and Diotima, there is for me no night.
'Let the death knell toll for all the virtues! it's you I hear, you, your heart's song, my love! and find immortal life while all else burns out and withers.'
'O Hyperion,' she cried, 'what are you saying?'
'I speak as I must. I can't, I can't any longer hide all the bliss and fear and care — Diotima! — Yes, you know it, must know it, have long since seen it, that I shall go under unless you give me your hand.'
She was taken aback, confounded.
'And to me,' she cried, 'to me Hyperion would cling? well then I wish, for the first time now I wish to be more than a mere mortal maiden. But I'll be for you what I can.'
'Oh then you're everything for me!' I cried.
'Everything? you wicked hypocrite! and what of mankind that's in the end your only love?'
'Mankind?' I said. 'I would that mankind made its watchword Diotima and painted your image on its banners, and proclaimed: "Today the divine shall triumph!" Angel of heaven! what a day that would be!'
'Go,' she cried, 'go and show your glory to heaven! it must not be so close to me.
'You will go, won't you, dear Hyperion?'
I obeyed. Who would not have done? I went. Never had I gone from her like this before. O Bellarmin! that was joy, stillness of life, peace of the gods, heavenly, wonderful, unknowable joy.
Words are useless here, and whoever seeks a likeness for it has never felt it. The only thing that could express such joy was Diotima's song, floating in golden mean between high and low.
O you willow banks of Lethe! you sunset paths in forests of Elysium! you lilies by the valley's streams! you rose-wreaths of the hill! I believe in you, in this friendly hour, and speak to my heart: 'there you shall find her again, and all the joy you have lost.'
#### Hyperion to Bellarmin [XXVIII]
I will tell you ever more of my bliss.
I will prove my breast on the joys of the past until it becomes like steel, I will practise on them until I'm invincible.
Ha! they may often fall like the stroke of a sword on my soul, but I'll play with the sword until I'm inured to it, I'll hold my hand in the fire until I can bear it like water.
I will not flinch; yes! I will be strong! I will conceal nothing from myself, I will conjure the most blissful of all my blisses from the grave.
It's beyond belief that man should fear what is most beautiful; yet it is so.
Oh, have I not fled a hundred times before these moments, this deadly delight of my memories, and averted my eyes like a child from lightning! and yet in the lush garden of the world there grows nothing lovelier than my joys, yet in heaven and on earth there flourishes nothing nobler than my joys.
But only to you, my Bellarmin, only to a pure free soul such as yours do I tell it. I will not be as prodigal as the sun with its rays, I will not cast my pearls before the foolish mob.
After that last soulful converse I knew myself less with each passing day. I felt there was a holy secret between me and Diotima.
I marvelled, dreamed. As if some blissful ghost had appeared to me at midnight and chosen me to walk with him, such was my state of mind.
Oh, it's a strange mixture of blissfulness and melancholy when it's thus revealed to us that we're forever removed from ordinary existence.
Since then I'd never managed to see Diotima alone. There'd always have to be a third person to disturb us, keeping us apart, and the world lay between her and me like an endless void. Six days of dread passed in this way without my knowing anything of Diotima. It was as if the others around us lamed my senses, as if they killed all my outer life, so that there was no way my locked-in soul could reach across to her.
If I would seek her with my eye, all went black before me, if I would address a word to her, it was throttled in my throat.
Oh! the holy nameless longing often almost tore my breast asunder, and mighty love often raged in me like a fettered titan. So deeply, so wholly implacably my spirit had never strained against the chains forged for it by fate, against the iron, inexorable law of being divided, not being a single soul with its desirable other half.
The star-bright night had now become my element. Then when all was still, as in the depths of the earth where gold in secret grows, it was then the more beautiful life of my love began.
Now my heart indulged its right to play the poet. It told me how Hyperion's spirit, before coming down to earth, had sported in pre-Elysium with his lovely Diotima, in divine childhood to the harmonies of the source, and beneath boughs such as we see the boughs of the earth when gilded they sparkle from golden streams.
And, like the past, the gates of the future opened in me.
Then we flew, Diotima and I, then we wandered like swallows from one springtime of the world to the next, through the vast realm of the sun and beyond to the other islands of heaven, to Sirius' golden shores, into the ghostly vales of Arcturus —
Oh, this is truly a thing to be desired, drinking the joy of the world from a single cup with the beloved!
Drunk from the blissful lullaby I sang myself, I fell asleep amidst the glorious phantoms.
But when the life of the earth took fire again from the ray of the morning light, I looked up and sought the dreams of the night. Like the beautiful stars they had vanished, and only the joy of grief bore witness to them in my soul.
I sorrowed; but I believe that amongst the blissful there's such sorrowing too. It was the harbinger of joy, this grief, it was the greying dawn from which would burgeon the countless roses of the blushing day. —
The glowing summer's day had now driven everything into dark shade. Around Diotima's house, too, all was still and empty, and at every window the envious curtains stood in my way.
I lived in thoughts of her. 'Where are you,' I thought, 'where shall my lonely spirit find you, sweet maiden? Do you gaze into space and muse? Have you put your work to one side, are you resting your elbow on your knee and your head on your little hand, yielding yourself to lovely thoughts?
'Let nothing disturb my tranquil one when she refreshes her heart with sweet fancies, let nothing handle this cluster of grapes and graze the quickening dew from its tender berries!'
So I dreamed. But whilst my thoughts sought her out between the walls of the house, my feet searched for her elsewhere, and before I was aware of it I was walking under the archways of the sacred wood behind Diotima's garden, where I'd seen her for the first time. What was this? Since then I'd so often mingled with these trees, become more intimate with them, more at peace beneath them; now a force seized me, as if I were stepping into Diana's shadows to die before the present deity.
Still I carried on. With every step I felt more wonderfully strange. I could have taken flight, so strongly did my heart propel me forwards; but it was as if there were lead in my shoes. My soul had hurried ahead, leaving my earthly limbs behind. I could no longer hear, and all shapes dimmed and shimmered before my eyes. My spirit was already with Diotima; the tree top played in the morning light, while the lower boughs still felt the chill of dawn.
'Ah! my Hyperion!' a voice now called out to me; I rushed towards it; 'my Diotima! O my Diotima!' beyond that I had no word and no breath, no consciousness.
Fade, fade, mortal life, beggarly business where the lonely spirit keeps inspecting the pennies it's gathered and counting them over! we are all called to the joy of the godhead!
Here there's a gap in my existence. I died, and awoke to find myself pressed to the heart of the heavenly maiden.
O life of love! how you had burst forth in her in full lovely bloom! as if sung into gentle slumber by blissful genii, the charming little head lay on my shoulder, smiled sweet peace, and opened up its aethereal eyes to me in joyful naive wonder, as if just now for the first time gazing out into the world.
Long we stood like this in lovely self-forgetful contemplation and neither knew what came over us, till at last too much joy heaped up in me, and amidst tears and sounds of rapture my lost speech began again, and roused my silently entranced Diotima fully back to life.
Then at last we looked again about us.
'O my old familiar trees!' cried Diotima, as if she hadn't seen them for a long time, and the remembrance of her former lonely days played about her joys, lovingly, like shadows about the virgin snow when it blushes and glows in the joyous glory of evening.
'Angel of heaven!' I cried, 'who can grasp you? who can say that he has fully comprehended you?'
'Do you wonder,' she replied, 'that I'm so very fond of you. Dear man! proud and modest man! Do you think I too am one of those who can't believe in you, have I not fathomed you, have I not found out the genius in his clouds? Veil yourself all you like and don't see yourself; I will conjure you forth, I will —
'But he's already here, he's burst forth like a star; he's broken through the husk and stands there like the spring; like a crystal stream from the gloomy grotto he's burst forth; this is not the dismal Hyperion, this is not the wild grief any more — O my own, my glorious boy!'
All this was like a dream to me. Could I believe in this miracle of love? could I? the joy would have killed me.
'Divine Diotima!' I cried, 'are your words meant for me? can you so deny yourself, you so blissfully self-sufficient! how can you take such joy in me? Oh, I see it now, I know now what I've often sensed, man is a garment that often a god wraps round himself, a chalice into which heaven pours its nectar so that its children may taste of the best.' —
'Yes, yes!' she broke in with an enthused smile, 'your namesake, the glorious Hyperion of heaven is in you.'
'Let me,' I cried, 'let me be yours, let me forget myself, let all life in me and all spirit fly only to you; to you alone, in blissful endless contemplation! O Diotima! so too I once stood before the shadowy divine effigy that my love fashioned for itself, before the idol of my lonely dreams; I nourished it lovingly; with my life I quickened it, with the hopes of my heart I refreshed it, warmed it, but it gave me nothing save what I'd given, and when I grew poor it left me poor, and now! now I have you in my arms and feel the breath of your breast, and feel your eyes in my eyes, this beautiful presence seeps into all of my senses, and I can sustain it, I hold what's most glorious and tremble no more — yes! I'm truly not who I used to be, Diotima! I'm become like you, and divine now plays with divine, as children play among themselves. —'
'But still you must grow a little calmer,' she said.
'You're right, you lovable creature!' I joyfully cried, 'else the graces won't appear to me; else I'll not see in the sea of beauty its gentle lovely ripples. Oh, I will yet learn to overlook nothing about you. Just give me time!'
'Flatterer!' she cried, 'but we're finished for today, dear flatterer! the golden cloud of evening is my reminder. Oh, don't be sad! Preserve for you and me the pure joy! Let it echo in you till the morrow, and don't kill it with ill humour! — the flowers of the heart want loving care. Their roots are everywhere, but they themselves only flourish in cheerful weather. Farewell, Hyperion!'
She tore herself loose. My whole being flamed up in me as she faded from view in her glowing beauty.
'O you!' — I cried and rushed headlong after her and gave my soul into her hand in endless kisses.
'God!' she cried, 'what will become of this!'
That hit me. 'Forgive me, heavenly one!' I said; 'I'm going. Good night, Diotima! think of me a little!'
'I will,' she cried, 'good night!'
And now not a word more, Bellarmin! It would be too much for my long-suffering heart. I'm shaken, I feel. But I will go out amongst the plants and trees and lay me down amongst them, and pray that nature bring me to such tranquillity.
#### Hyperion to Bellarmin [XXIX]
Our souls now lived ever more freely and beautifully together, and everything in and around us commingled into golden peace. It seemed as if the old world had died and a new one were beginning with us, so spiritual and strong and loving and light had everything become, and we and all beings floated in blissful union, like a chorus of a thousand inseparable tones, through the boundless aether.
Our converse glided away like a cerulean stream whence the goldsand now and then glitters, and our silence was like the silence of mountain-peaks where in glorious lonely height, far above the realm of the storms, only the divine breeze whispers still in the locks of the bold wanderer.
And the wonderful holy sorrow when the hour of parting rang into our rapture, when often I cried: 'now we are mortal again, Diotima!' and she said to me: 'Mortality is illusion, it's like the colours that quiver before our eye when it looks long into the sun!'
Oh! and all the sweet games of love! the pretty speeches, the petty worries, the touchiness, the strictness and indulgence.
And the omniscience with which we saw through one another, and the infinite faith with which we glorified one another!
Yes! man is a sun, all-seeing, all-illuming when he loves, and if he doesn't he's a dark dwelling in which there burns a reekie little lamp.
I ought to keep silent, ought to forget and keep silent.
But the alluring flame will tempt me till I plunge into it whole and perish like the fly.
In the midst of all this blissful wholehearted giving and taking, I felt one day that Diotima was becoming stiller and ever stiller.
I asked and I begged; but this only seemed to distance her the more; finally she begged that I no longer ask, that I should go, and speak of something else when I returned. This brought about a painful dumbness in me too, which I myself couldn't come to grips with.
I felt as if an unfathomable sudden fate had sworn death to our love, and all life was gone, from me and everything.
It's true I was ashamed of this; I knew for sure that Diotima's heart was not ruled by caprice. But always she remained a mystery to me, and my spoiled despondent disposition demanded that love be always palpable and present; locked-up treasures were treasures lost for me. Oh! in my happiness I'd unlearned hope, I was then still like impatient children who cry over the apple on the tree as though it weren't there at all unless it's kissing their mouths. I had no peace, again I begged, violently and meekly, tenderly and angrily, love armed me with all of its all-overpowering humble eloquence and now — O my Diotima! now I had it, the delightful confession, now I have and will hold it till the tide of love carries me too, with all that I am, back to the ancient home, into the womb of nature.
The innocent! she'd yet to know the powerful abundance of her heart, and sweetly dismayed at the wealth within her, she buried it deep in her breast — and when she now confessed, holy simplicity, when she confessed with tears that she loved too much, and when she took leave from all she'd cradled till then to her heart, oh, when she cried: 'I've fallen away from May and summer and autumn, and don't heed the day and the night as once I did, don't belong any longer to heaven and earth, belong just to one, only to one, but the blossom of May and the flame of summer and the ripeness of autumn, the lightness of day and the gravity of night, and earth and heaven, all is fused for me into this one! so do I love!' — and when she now looked at me, in the fullness of her heart's content, when she, in bold and holy joy, took me into her beautiful arms and kissed me on brow and mouth, oh! when that divine head, dying in bliss, slipped down my open neck, and those lovely lips rested on my beating breast and the sweet breath touched my soul — O Bellarmin! my senses fail and my mind runs adrift.
I see, I see how this must end. The rudder has fallen into the surge and the ship will be seized like a child by the feet and hurled against the rocks.
## Hyperion to Bellarmin [XXX]
There are in life great hours. We look up to them like colossal figures of the future and antiquity, we fight a glorious fight with them, and should we hold our own against them, they then become like sisters and never leave us.
Once we sat together on our mountain, on a stone of the ancient city of this island, and spoke of how here the lion Demosthenes had found his end, how he'd here with holy self-elected death helped himself to freedom from the Macedonian chains and daggers — 'That glorious spirit went from this world with a jest,' one of us cried. 'And why not?' I said, 'there was nothing more for him here; Athens had become Alexander's whore and the world, like a hart, hounded to death by the great huntsman.'
'O Athens!' cried Diotima; 'there's many a time I've grieved when I gazed across and out of the blue twilight the phantom of the Olympieion rose before me!'
'How far is it to get there?' I asked
'A day's journey, perhaps,' Diotima replied.
'A day's journey,' I cried, 'and I haven't been there yet? We must go across together straightaway.'
'Right then!' cried Diotima; 'we shall have a fair sea tomorrow, and all's now still greenness and ripeness.
'One needs the eternal sun and the life of the immortal earth for such a pilgrimage.'
'Tomorrow then!' I said, and our friends agreed.
Early, to the rooster's song, we set out from the roadstead. We and the world were shining in fresh brilliance. Golden peaceful youth was in our hearts. The life in us was like a newborn isle of the ocean in which the first spring is beginning.
For some time now, under Diotima's influence, more stability had come into my soul; today I felt it with threefold purity, and the scattered swarming energies were all gathered in one single golden mean.
We spoke together about the excellence of the ancient Athenians, how it arose, in what it consisted.
One said it was the climate that made it; another: art and philosophy; a third: religion and polity.
'Athenian art and religion, and philosophy and polity,' I said, 'are flowers and fruits of the tree, not the soil and the roots. You take the effects for the cause.
'And anyone who tells me that this was all shaped by the climate should consider that we too still live in it.
'Less disturbed in every respect, freer from violent influence than any other people on earth, so grew the people of the Athenians. No conqueror weakens them, no military fortune inebriates them, no alien cult dulls them, no hasty wisdom drives them to untimely ripeness. Left to itself, like the growing diamond, such is their childhood. Almost nothing is heard of them up to the times of Pisistratus and Hipparchus. They took but little part in the Trojan War which, as in a hothouse, prematurely kindled and quickened most of the Greek peoples. — It's not extraordinary fate that brings forth man. Great and colossal are the sons of such a mother, but beautiful beings, or, what is the same thing, men they never become, or not until late when the contrasts war with each other too harshly for them not to make peace in the end.
'In exuberant vigour Lacedaemon rushes ahead of the Athenians and for this very reason would also have scattered and dissolved sooner, had not Lycurgus come and with his discipline contained its intemperate nature. And so from then on everything about the Spartan was acquired, all excellence achieved and bought through labour and selfconscious effort, and in as much as one can, in a certain sense, speak of the simplicity of the Spartans, yet it follows naturally that true childlike simplicity was wholly lacking amongst them. The Lacedaemonians breached too soon the order of instinct, they grew out of kind too soon, and so then, with them, discipline too had to begin too soon; for all discipline and art begins too soon where man's nature hasn't yet come to ripeness. Perfected nature must dwell in the human child before it enters school, so that the image of childhood may show it the way back from school to perfected nature.
'The Spartans remained eternally a fragment; for whoever was never a finished child can hardly become a finished man. —
'True, heaven and earth also did their part for the Athenians, as for all the Greeks, dispensing neither dearth nor abundance. The rays of heaven didn't fall upon them like a rain of fire. The earth didn't pamper and ravish them with caresses and lavish gifts as here and there the foolish mother otherwise may do.
'And then there was the prodigious gesture of Theseus, the voluntary curtailment of his own monarchic power.
'Oh! such a seed, thrown into the hearts of the people, must beget an ocean of golden grains, and visibly continues even late to burgeon and bear fruit amongst the Athenians.
'Once again then! that the Athenians grew up so free from violent influence of any kind, on such moderate fare, this is what made them so excellent, and this alone could!
'Leave man undisturbed from the cradle on! don't drive him from the enclosing oneness of his being's bud, out of his childhood's tabernacle! don't do too little lest he manage without you and so distinguish you from himself, don't do too much lest he feel your power or his own and so distinguish you from himself, in short, don't let man know until late that there are men, that there is something outside himself, for only in this way will he become man. But man is a god as soon as he's man. And once he's a god, then he's beautiful.'
'Strange!' cried one of the friends.
'You've never yet spoken so deeply from my soul,' cried Diotima.
'I owe it to you,' I answered.
'That's how the Athenian was man,' I continued, 'that's how he had to become man. Beautiful he came from the hands of nature, beautiful in body and soul, as the saying goes.
'The first child of human, of divine beauty is art. In art divine man renews and repeats himself. He wants to feel himself, and so he confronts himself with his own beauty. Thus man gave himself his gods. For in the beginning man and his gods were one, when, unknown to itself, eternal beauty was. — I'm speaking mysteries, but they are. —
'The first child of divine beauty is art. So it was with the Athenians.
'Beauty's second daughter is religion. Religion is love of beauty. The wise man loves beauty itself, infinite and all-embracing; the people love its children, the gods, who appear to them in multifarious forms. So it was too with the Athenians. And without such love of beauty, without such religion any state is a shrivelled skeleton without life and spirit, and all thought and deed a tree without a top, a column whose crown has been lopped off.
'But that it was really the case with the Greeks, and especially the Athenians, that their art and their religion are the true children of eternal beauty — of perfected human nature — and could issue only from perfected human nature, that is clearly evident if one is only willing to consider with an unbiased eye the objects of their holy art and the religion with which they loved and honoured those objects.
'There are blunders and blemishes everywhere, and so here too. But one thing is certain withal: what one finds in the objects of their art is mostly ripened man. Here there's not the pettiness, not the monstrousness of the Egyptians and Goths, here there's human sense and human form. They sheer off less than others into the extremes of the supersensual and the sensual. Their gods stay more than others within the beautiful mean of man.
'And as the object, so also the love. Not too servile and not all too familiar! —
'From the Athenians' beauty of spirit there followed too the necessary feeling for freedom.
'The Egyptian bears without pain the despotism of arbitrary power, the son of the North without resentment the despotism of the law, legally enshrined injustice; for the Egyptian possesses from the womb an instinct for servility and idolatry; in the North they have too little faith in the pure free life of nature not to be superstitiously thralled to legality.
'The Athenian can't abide arbitrary power because his divine nature brooks no disturbance, he can't abide legality everywhere because he has no need of it everywhere. Draco isn't for him. He expects to be treated gently, and has every right to do so too.'
'Fair enough!' someone interjected, 'that I understand, but how this poetic religious people should also come to be a philosophical one, that I don't see.'
'Without poetry,' I said, 'they would never even have been a philosophical people!'
'What has philosophy,' he responded, 'what has the cold sublimity of that science to do with poetry?'
'Poetry,' I said, sure of my case, 'is the beginning and end of that science. Like Minerva from Jupiter's head it springs from the poetry of infinite divine being. And so too what is irreconcilable in it will finally flow together again in the mysterious wellspring of poetry.'
'He's a paradoxical fellow,' cried Diotima, 'yet I sense his meaning. You're both digressing though. It's Athens we're talking about.'
'The man,' I began again, 'who has not at least once in his life felt in himself full unsullied beauty, when the powers of his being played into one another like the colours of the rainbow, who has never come to know that only in hours of inspiration all is inmost concord, this man will not even become a philosophical sceptic, his spirit is not even fitted for tearing down, let alone for building up. For believe you me, the sceptic finds flaws and contradiction in all that is thought only because he knows the harmony of flawless beauty that never is thought. He disdains the dry crust that human reason proffers him well-meaningly only because he's secretly feasting at the table of the gods.'
'Dreamer!' cried Diotima, 'that's why you too were a sceptic. But the Athenians!'
'I was just coming to them,' I said. 'The mighty phrase of Heraclitus, εν διαϕερον εαυτω (the One differentiated in itself), could have been arrived at only by a Greek, for it is the essence of beauty and before that was found, there was no philosophy.
'Now one could define, the whole was there. The flower had ripened, one could now dissect.
'The moment of beauty had now been made known among men, it was there in life and spirit, infinite oneness was.
'One could take it apart, divide it in the mind, think the divided together afresh, could so come more and more to know the essence of the highest and the best and turn that knowledge into laws for the manifold realms of the mind.
'Do you see now why the Athenians in particular had also to be a philosophical people?
'The Egyptian couldn't. He who doesn't live loving, and loved by, heaven and earth in equal measure, who doesn't in this sense live at one with the element in which he moves, is by nature also not so at one with himself and cannot know eternal beauty, at least not so readily as a Greek.
'Like a sumptuous despot the oriental clime prostrates its inhabitants with its power and its splendour, and before man has even learned to walk he must kneel, before he has learned to talk he must pray; before his heart has found its balance it must bow, and before the spirit is strong enough to bear flower and fruit, fate and nature sap from it all strength with searing heat. The Egyptian's surrendered before he's a whole, and so he knows nothing of the whole, nothing of beauty, and what he calls the highest is a veiled power, a ghastly enigma; the dumb dismal Isis is his alpha and omega, an empty infinity, and nought meaningful ever came out of that. Even the sublimest nothing begets nothing.
'The North, by contrast, drives its nurslings into themselves too soon, and if the spirit of the fiery Egyptian, too wanderlusting, rushes out into the world, in the North the spirit prepares to return into itself before it's even ready to fare forth.
'In the North man must already have sense before even there's any ripe feeling within him, he imputes to himself the guilt of all things before innocence even has come to its beautiful end; he must become rational, a self-conscious spirit, before he's a man, a worldly-wise man before he's a child; the harmony of whole human being, beauty, may not blossom and ripen in him before he's developed and cultured. Mere understanding, mere reason are ever the kings of the North.
'But from mere understanding there never came anything sensible, from mere reason never came anything rational.
'Understanding without beauty of spirit is like a ministering menial who fashions the fence from rough timber, according to plans he's been given, and nails together the posts that he's hewn for the garden his master wants to make. The understanding's whole business is work of necessity. By laying down rules it guards us from folly and wrongness; but being safe from folly and wrongness is not the pinnacle of human excellence.
'Reason without beauty of spirit, beauty of heart, is like a slavedriver whom the master of the house has set over the servants; he knows as little as the servants what good is served by all the endless toil, and only shouts: 'Get a move on,' and is almost loath to see it happen, for in the end he'd have no more driving to do and his part would be played out.
'From mere understanding comes no philosophy, for philosophy is more than only the limited perception of what is.
'From mere reason comes no philosophy, for philosophy is more than the blind demand for never-ending progress in the unification and differentiation of whatever it can be applied to.
'But when the divine εν διαϕερον εαυτω, the ideal of beauty illumines striving reason, it doesn't demand blindly, it knows why it demands and for what.
'When, like a May morning into the artist's workshop, the sun of the beautiful shines for the understanding in its activity, it's true that it doesn't rush forth and abandon its work of necessity, yet still it thinks fondly forward to the festal day when it will wander in the quickening light of spring.'
This was as far as I'd got when we landed on the coast of Attica. Ancient Athens now was too much in our minds for us to feel much like orderly talk, and I wondered now myself at the nature of my utterances. 'How ever did I come,' I cried, 'to fetch up on those barren mountain peaks on which you saw me?'
'It's always like that,' replied Diotima, 'when we're feeling elated. Exuberant energy looks to find work. Young lambs butt their heads together when they've had their fill of mother's milk.'
We now made our way up Lycabettus, for all our haste stopping at times in thought and wonderful anticipation.
It's good that man finds it so hard to convince himself of the death of what he loves, and likely no one has ever gone to his friend's grave without some quiet hope of actually meeting the friend there. The beautiful phantom of ancient Athens seized me like the shape of a mother returning from the realm of dead.
'O Parthenon!' I cried, 'pride of the world! at your feet lies Neptune's realm like a vanquished lion, and like children the other temples are gathered around you, and the eloquent Agora and the grove of Academe — '
'Can you transport yourself so into the ancient times,' said Diotima.
'Don't remind me of those times!' I replied; 'there was divine life and man was the central point of nature. Spring, when it blossomed all around Athens, it was like a demure flower at the virgin's bosom; the sun rose red with shame over the glories of the earth.
'The marble rocks of Hymettus and Pentelikon leapt from their slumbering cradle like children from the mother's womb, and took on form and life under the delicate hands of the Athenians.
'Nature proffered honey, and the fairest violets and myrtles and olives.
'Nature was priestess and man her god, and all the life in her and every shape and every sound of hers but a single rapturous echo of the glorious being to whom she belonged.
'Him alone she celebrated, to him alone she sacrificed.
'And he was worthy of it, whether he sat lovingly in the holy workshop, embracing the knees of the divine image he'd made, or on the headland, on Sunium's green peak, ensconced amidst the listening disciples, whiling away the hours with lofty thoughts, or he might be running in the Stadium, or from the speaker's tribune dispensing, like the storm god, rain and sun and lightning bolts and golden clouds —'
'Oh look!' Diotima now suddenly called out to me.
I looked, and could have died at the almighty spectacle.
Like a vast shipwreck when the storms are silenced and the sailors flown, and the body of the shattered fleet lies unrecognizable upon the sandbank, so lay before us Athens, and the orphaned pillars stood before us like the naked tree trunks of a forest which greened still in the evening and then went up in fire in the night.
'Here,' said Diotima, 'you learn to be still about your own fate, be it good or ill.'
'Here you learn to be still about everything,' I continued. 'Had the reapers who mowed this cornfield enriched their barns with its haulms, nothing would have been lost and I should be content to stand here as a gleaner; but who gained by it?'
'All of Europe,' replied one of the friends.
'Oh, yes!' I cried, 'they've hauled off the pillars and statues and sold them to each other, prizing the noble figures not a little on account of their rarity, the way one prizes parrots and monkeys.'
'Don't say that!' replied the same man; 'and should they really lack the spirit of all the beauty, it would be because that couldn't be carried away, couldn't be bought.'
'Quite!' I cried. 'For that spirit had perished even before the destroyers fell upon Attica. Only when the houses and temples are desolate do the wild beasts venture into the gates and streets.'
'For anyone who does have that spirit,' said Diotima consolingly, 'Athens still stands like a fruit tree in bloom. The artist can easily for himself restore the torso to wholeness.'
The next day we went out early, saw the ruins of the Parthenon, the site of the ancient theatre of Bacchus, the temple of Theseus, the sixteen columns that still remain standing of the divine Olympieion; but what most moved me was the ancient gate through which in former times one came out of the old town into the new, and where once a thousand beautiful people must have greeted each other in a single day. Now you come neither into the old nor into the new city through this gate, and mute and waste it stands there like a dried-up fountain from whose pipes the clear fresh water once sprang forth with a friendly plash.
'Alas!' I said, as we went walking round, 'it's a splendid game fate plays, is it not, toppling down the temples here and leaving the shattered stones for children to throw about, making the mutilated gods into benches before the peasant's hut and the tombs here a restingplace for the grazing bull, and such prodigality is more regal than the wantonness of Cleopatra when she drank the molten pearls; but even so, the pity of all that grandeur and beauty!'
'Dear Hyperion!' cried Diotima, 'it's time for you to leave; you're pale and your eyes are weary, and it's useless trying to help yourself with fancies. Come away! out into the green! in amongst the colours of life! that will do you good.'
We went out into the nearby gardens.
The others had fallen into conversation on the way with two British scholars reaping their harvest among the antiquities of Athens, and they were not to be budged. I was happy to leave them to it.
My whole being was uplifted when once again I saw myself alone with Diotima; in a glorious struggle with the sacred chaos of Athens she had prevailed. Like the lyre of the heavenly muse above the warring elements, so Diotima's still thoughts held sway above the wreckage. Like the moon from a tender cloud, so her spirit rose up from sweet sorrow; the heavenly maiden stood there in her sadness like the flower that breathes out its loveliest fragrance in the night.
On and on we walked, and in the end our walk was not in vain.
O you groves of Angele where the olive tree and the cypress, whispering round one another, cool themselves with friendly shade, where the golden fruit of the citron tree gleams through the dark green, where the swelling grape grows wantonly over the hedge, and the ripe orange, like a smiling foundling, lies on the way! you fragrant hidden paths! you tranquil seats where the image of the myrtle bush smiles from the spring! you I shall never forget.
Diotima and I walked about for a while beneath the glorious trees until a large cheerful glade opened up before us.
Here we sat down. There was a blissful stillness between us. My spirit floated about the divine form of the maiden like the butterfly about a flower, and all my being lightened and became one in the joy of inspiring contemplation.
'Are you so soon comforted again, you lighthead?' said Diotima.
'Yes! yes! I am,' I answered. 'What I thought was lost I have, what I languished for, as if it had vanished from the world, is now before me. No, Diotima! the wellspring of eternal beauty hasn't yet run dry.
'I've told you once before, I no longer need gods and men. I know that heaven is desolate and void, and the earth, which once overflowed with beautiful human life, has almost become like an anthill. But there is still a place where the old heaven and the old earth smile for me. For all gods of heaven and all godly humans of earth I forget in you.
'What's the shipwreck of the world to me, I know of nothing but my blissful isle.'
'There's a time for love,' said Diotima in gentle earnest, 'as there's a time to live in the happy cradle. But life itself drives us out.
'Hyperion!' — here she ardently seized my hand, and her voice was raised with grandeur — 'Hyperion! it seems to me that you are born for higher things. Do not misjudge yourself! it's the lack of material that's held you back. Things didn't move quickly enough. That cast you down. Like the young fencers you'd lunge too soon, before your aim was yet sure and your hand skilled, and since, as is only natural, you took more hits than you gave, you became chary and doubted yourself and all things; for you're as sensitive as you're hotheaded. But thereby nothing is lost. Had your mind and your actions matured so early, your spirit would not be what it is; you would not be the thinking man, had you not been the suffering one, the man in ferment. Believe me, you would never so clearly have come to know the equilibrium of beautiful humanity, had you not so utterly lost it. Your heart has found peace at last. I'll believe it. I understand it. But do you really think you've now reached your end? Will you lock yourself up in the heaven of your love and leave a world which has need of you to wither and grow cold beneath you? Down you must like the ray of light, like the all-refreshing rain you must descend into the land of mortality, you must illumine like Apollo, shake and quicken like Jupiter, or you are not worthy of your heaven. I beg you, go back into Athens once again, and this time look at the people there walking about amongst the ruins, the raw Albanians and the other good artless Greeks, who console themselves with a merry dance and a pious fable for the vile oppression that weighs down on them — can you say: "I'm ashamed of this material"? I think it could still prove malleable after all. Can you turn away your heart from those in need? They're not evil, they've done you no harm!'
'What can I do for them?' I cried.
'Give them what you have within you,' Diotima replied, 'give — '
'Not a word, not a word more, great soul!' I cried, 'else you will sway me, else it will be as if you'd forced me to it —
'They won't be happier, but nobler, no! they'll be happier too. They must come forth, they must rise like young mountains from the surging sea, driven by their subterranean fire.
'It's true I stand alone and come unrenowned among them. Yet just one who is a man, can he not do more than hundreds who are only fragments of men?
'Holy nature! you are the same within me and without. It cannot be so hard to unite what is without with the divine that is within me. If the bee can build its thriving little realm, then why shouldn't I be able to plant and raise what's needed?
'What? the Arab merchant sowed his Koran and there grew up for him like an endless forest a people of disciples, and that field should not also flourish where the ancient truth shall come again in newly quickened youth?
'Let there be fundamental change! From humanity's roots let the new world spring! Let a new godhead rule over them, a new future brighten before them.
'In the workshop, in the home, in the assemblies, in the temples, everywhere let there be change!
'But first I must go forth to learn. I'm an artist, but I lack the craft. I can mould in my mind, but I haven't yet learned to direct my hand — '
'You shall go to Italy,' said Diotima, 'to Germany, France — how many years do you need? three — four — I should think three are enough; you're not of the slow sort, and will be seeking out only the greatest and most beautiful —'
'And then?'
'You will become the educator of our people, you'll be a great man, I hope. And when I then embrace you like this, then I'll dream as if I were a part of the glorious man, then I'll rejoice as if you had gifted me, like Pollux to Castor, half of your immortality, oh! I shall become a proud maiden, Hyperion!'
I fell silent for a while. I was filled with ineffable joy.
'Is there then contentment between the decision and the deed,' I finally resumed, 'is there a calm before the victory?'
'It's the calm of the hero,' said Diotima, 'there are decisions which, like the dicta of the gods, are command and fulfilment in one, and such is yours. —'
We walked back, as after our first embrace. Everything had become strange and new for us.
Now I stood over the ruins of Athens like the farmer on the fallow field. 'Just lie peaceful,' I thought, as we took ship again, 'just lie peaceful, slumbering land! Soon young life will green from you and wax towards the blessings of heaven. Soon the clouds will no more rain in vain, soon the sun will find again its ancient progeny.
'Do you ask after men, nature? Do you lament like a lyre which the brother of chance, the wind alone plays, since the artist who kept it in tune is dead? They will come, your men, nature! A rejuvenated people will rejuvenate you too, and you will become as its bride, and the ancient covenant of spirits will renew itself with you.
'There will be but one beauty; and humanity and nature will unite into one all-embracing godhead.'
## Volume Two
μη φυναι, τον απαντα νιχα λογον. το δ'επει ϕανη, βηναι χειδεν, οδεν περ ηχει, πολν δεντερου ως ταχιστα.
Sophocles
[Not to be born is, past all prizing, best. But when a man has seen the light of day, this is the next best by far, that with the utmost speed he should go back whence he came.]
## Hyperion to Bellarmin [XXXI]
We were living the last lovely moments of the year, on our return from Attica.
A brother of spring was the autumn for us, full of mellow fire, a feast time for commemorating sorrows and past joys of love. The withering leaves wore the red hues of dusk, only the spruce and the laurel stood in eternal green. In the bright breezes wandering birds lingered, others swarmed in vineyard and garden, merrily reaping what people had left. And the heavenly light streamed pure from the open sky, through all the branches smiled the holy sun, the goodly, that never I name without joy and thanks, that often in deep grief has healed me with a glance and purged my soul of discontent and cares.
We revisited all our dearest paths, Diotima and I, vanished hours of bliss would meet us everywhere.
We remembered the past May, we'd never before seen the earth as then, we said, it had been transformed, a silver cloud of blossom, a joyful flame of life, freed of all coarser stuff.
'Oh! all was so full of joy and hope,' cried Diotima, 'so full of ceaseless growth and yet so effortless too, so blissfully at peace, like a child lost in play with not a thought in the world.'
'It's by this that I know it, the soul of nature,' I cried, 'by this still fire, by this lingering in its mighty haste.'
'And it's so dear to the happy, this lingering,' cried Diotima; 'you remember? we stood together on the bridge one evening, after a violent storm, and the red mountain torrent shot away like an arrow beneath
us, but beside it the woods were greening in peace and the bright beech leaves barely stirred. How it warmed our hearts that the soulful green did not fly away from us too like the stream, and the beautiful spring held still for us like a tame bird, but now just the same it's over the hills and long gone.'
We smiled at the phrase, although we felt more like grieving.
So too our own bliss would pass, and this we foresaw.
O Bellarmin! who then may say he stands fast when even the beautiful ripens so towards its fate, when even the divine must humble itself and share mortality with all that is mortal!
#### Hyperion to Bellarmin [XXXII]
I'd lingered with the lovely maiden before her house until the lamp of the night shone into the tranquil twilight; now I came back to Notara's dwelling, thoughtful, full of overflowing heroic life, as always when I went from her embraces. A letter had come from Alabanda.
'Things are on the move, Hyperion,' he wrote, 'Russia has declared war on the Porte; they're coming with a fleet into the Archipelago;3 the Greeks shall be free if they join the uprising to drive the Sultan to the Euphrates. The Greeks will do their part, the Greeks will be free and I'm heartily glad that there's finally something to do again. I couldn't face the light of day before we got as far as this.
'If you're still the old Hyperion, then come! You'll find me in the village before Coron when you come by way of Mistra. I'm staying on the hill, in the white villa by the woods.
'I've broken with the people whom you met with me in Smyrna. You were right, with your finer sense, not to step into their sphere.
'I long to see us both together again in the new life. Until now the world was too wretched for you to make yourself known to it. Since you wouldn't do bondsman's work you did nothing, and doing nothing made you dreamy and morose.
'You wouldn't swim in the swamp. Come now, come, and let us bathe in open sea!
'That shall do us good, my sole beloved!'
<sup>3 [</sup>Hölderlin's note] In the year 1770.
So he wrote. At first I was taken aback. My face burned with shame, my heart seethed like hot springs, and I couldn't keep still, so did it pain me to be outflown by Alabanda, bested once and for all. Yet all the more eagerly did I now embrace the work ahead. —
'I've grown too idle,' I cried, 'too fond of the quiet life, too fey, too sluggish! — Alabanda looks out into the world like a noble pilot, Alabanda is busy and searches the waves for booty; and your hands are asleep in your lap? and you would make do with words and conjure the world with magic formulas? But your words are feckless as snowflakes and merely make the air murkier, and your incantations are for the pious, but the unbelievers don't hear you. — Yes! to be gentle in season, that may be a beautiful thing, but to be gentle out of season, that's ugly, because it's cowardly! — But Harmodius! it's your myrtle I will be like, your myrtle in which the sword was concealed. I will not have walked in idleness for nothing, and my sleep shall become like oil when it's touched by the flame. I will not look on when it counts, will not go around asking for news when Alabanda takes the laurel.'
#### Hyperion to Bellarmin [XXXIII]
Diotima's sudden pallor, when she read Alabanda's letter, pierced my soul. She then began, calmly and earnestly, to advise me against this step, and we spoke much back and forth. 'O you men of violence!' she cried at last, 'you who are so quick to go to the extreme, remember Nemesis!'
'For those who suffer extremes,' I said, 'for them extreme is right.'
'Even if it's right,' she said, 'it's not what you were born for.'
'So it would seem;' I said, 'I've loitered long enough. Oh, I'd like to load an Atlas on my shoulders to discharge the debts of my youth. Do I have a sense of myself? do I have inner stability? Oh, let me go, Diotima! It's here, in just such work as this, that I must gain it.'
'That's sheer hubris!' Diotima cried; 'you were lately more modest, lately when you said: "first I must go forth to learn".'
'Dear sophist!' I cried, 'our talk then was of something completely different. Leading my people to the Olympus of divine beauty, where from springs for ever young there flows the true with all things good, is something I'm not yet fit to do. But I've learnt to use a sword, and for now that's all that's needed. The new covenant of spirits can't live in the air, the holy theocracy of the beautiful must have a free state to live in, and that demands space on earth and this space we shall surely conquer.'
'You will conquer,' cried Diotima, 'and forget what for? at best you'll force your free state into being and then say: "for what have I built?" Oh! it will have been consumed, all the beautiful life that was to stir there will have been spent, even in you! The savage struggle will tear you apart, beautiful soul, you will grow old, blissful spirit! and lifeweary ask at the end, "where are you now, you ideals of my youth?"'
'That's cruel, Diotima,' I cried, 'grasping the heart in this way, holding me fast by my own fear of death, by my highest joy of life, but no! no! no! bondage kills, but a just war brings every soul alive. It's what gives to gold the colour of the sun, that it gets thrown into the fire! This and only this gives to man his whole youth, that he bursts his bonds asunder! It's this alone can save him, that he goes forth and crushes the viper underfoot, the crawling century that poisons in the bud all beautiful nature! — You say I'll grow old, Diotima! if I liberate Greece? grow old and wretched, a mean little man? Oh, then he must have been right sapless and empty and godforsaken, that Athenian youth, the victory messenger from Marathon, when he came over the peak of Pentelikon and looked down into the vales of Attica!'
'Dearest! dearest!' cried Diotima, 'be still! I won't say another word. You shall go, you shall go, proud man! Oh! when you're like this I have no power, no claim over you.'
She wept bitterly and I stood before her like a criminal. 'Forgive me, divine maiden!' I cried, sunk down before her, 'oh, forgive me that I must! I'm not making a choice, I'm not deliberating. There's a power within me and I cannot tell if it's my self that drives me to this step.' 'It's your full soul that commands you,' she answered. 'Not to follow it often leads to ruin, but then following it likely does too. It's best you go, for that is greater. You act; I will bear it.'
#### Hyperion to Bellarmin [XXXIV]
Diotima was from now on marvellously changed.
With joy I'd seen how since our love began the muted life had opened out in glances and lovely words, and her genial tranquillity had often met me as sparkling inspiration.
But how alien to us grows the beautiful soul when after its first blossoming, after the morning of its course, it must rise toward high noon! One barely knew the blissful child any more, so sublime and so sorrowful had she become.
Oh, how many times I lay before that grieving image of divinity and thought to weep away my soul in pain for her, and stood up in wonderment and myself full of all-powerful energies! A flame had risen to her eyes from her oppressed breast. Her bosom had grown for her too crowded with wishes and sorrows; that's why the maiden's thoughts were so glorious and bold. A new greatness, a visible power over all that could feel, held sway within her. She was a higher being. She belonged no more to mortal man.
O my Diotima, could I have imagined then how it would end?
#### Hyperion to Bellarmin [XXXV]
Even canny Notara was captivated by the new project, promised me strong backing, hoped soon to occupy the Corinthian Isthmus and grasp Greece here as by the helve. But fate willed otherwise and rendered his work futile before it reached its goal.
He advised me not to go to Tinos, but to head straight down the Peloponnese, and as unobtrusively as possible. I should write to my father on the way, he said; the cautious old man would more easily condone a step that had been taken than allow one that hadn't. I wasn't wholly at ease with that, but we so readily sacrifice our own feelings when a great goal stands before our eyes.
'I doubt,' continued Notara, 'whether you'll be able to count on your father's help in a case like this. So I shall give you what you're surely going to need along the way so as to live and do your work for a time, come what may. If you can someday, then pay it back; if not, then what was mine was also yours. Don't be ashamed to take the money,' he added with a smile; 'even the steeds of Phoebus don't live on air alone, so the poets tell us.'
#### Hyperion to Bellarmin [XXXVI]
Now came the day of parting.
Throughout the morning I'd remained up in Notara's garden, in the fresh winter air, beneath the evergreen cypresses and cedars. I was composed. The great energies of youth sustained me, and the suffering I sensed bore me aloft like a cloud.
Diotima's mother had asked that Notara and the other friends and I should spend the last day together at her home. The good people all had taken their joy in me and Diotima, and what was godly in our love had not been lost on them. They were now to bless my leaving too.
I went down. I found the dear maiden at the hearth. It seemed for her a holy priestly office to tend to the house on this day. She had prepared everything, made everything in the house beautiful, and no one was allowed to help. She had gathered all the flowers that still remained in the garden, and managed to put together roses and fresh grapes even in this late season.
She knew my footfall when I approached, and came softly towards me; her pale cheeks glowed from the flame of the hearth and her earnest swollen eyes were glistening from tears. She saw how overcome I was. 'Go in, my dear,' she said, 'mother's inside and I shall follow shortly.'
I went in. There the noble woman sat and held out her lovely hand to me — 'are you come,' she cried, 'are you come, my son! I should be angry with you, you have taken my child from me, have talked all reason and sense out of me, and you do what you list and make off; but forgive him, you heavenly powers! if what he intends is wrong, and if it is right, oh, then do not hold back with your help for the dear youth!' I wanted to speak, but just then Notara came in with the other friends and behind them Diotima.
We were silent for a while. We honoured the grieving love that was in us all, we were fearful to make free with it in speeches and lofty thoughts. Finally, after a few fleeting words, Diotima asked me to tell of Agis and Cleomenes: I'd often mentioned those great souls with fiery reverence and claimed that they were demi-gods as surely as Prometheus, and their struggle with the fate of Sparta more heroic than any in the lustrous myths. The genius of these men, I'd said, was the golden sunset of the Greek day, just as Homer and Theseus had been its aurora.
I told their story and at the end we all felt stronger and uplifted.
'Happy the man,' cried one of the friends, 'whose life alternates between heartfelt joy and fresh battles.'
'Yes!' cried another, 'that is eternal youth when there are always ample energies at play and we can keep ourselves wholly in pleasure and toil.'
'Oh, that I could go with you!' Diotima cried to me.
'But it's good you stay behind, Diotima!' I said. 'The priestess may not leave the temple. You shall be the keeper of the holy flame, in quiet seclusion the keeper of the beautiful, that I may find it again in you.'
'Yes, you're right, my dear, that's best,' she said, and her voice trembled and her ethereal eyes were hidden behind her handkerchief, so their tears, their confusion shouldn't be seen.
O Bellarmin! it almost tore my breast asunder to have made her so blush with shame. 'Friends!' I cried, 'keep this angel safe for me. I can be sure of nothing more if I can't be sure of her. O heaven! I dread to think what I could do if I were to lose her.'
'Be calm, Hyperion!' Notara interjected.
'Calm?' I cried; 'O you good people! you're ready enough to worry how the garden's going to bloom or what the harvest will be like, you're capable of praying for your vines, and I'm supposed to part without wishes from the only thing my soul serves?'
'No, good Hyperion!' cried Notara, moved, 'no! you shall not part from her without wishes! no, by the godly innocence of your love! of my blessing you both may be sure.'
'You remind me,' I quickly cried. 'She shall bless us, this dear mother, with all of you she shall be our witness — come Diotima! your mother shall hallow our union until the beautiful community, which is our hope, join us in marriage.'
So I fell upon one knee; with a wide gaze, blushing and solemnly smiling, she too sank down by my side.
'Long since, O nature!' I cried, 'has our life been one with you, and heavenly-youthful, like you and all your gods, is our own world become through love.'
'In your groves we wandered,' Diotima continued, 'and were like you, by your springs we sat, and were like you; over yon mountains we went, with your children the stars, like you.'
'When we were far from one another,' I cried, 'when, like a harp's whispering, our coming rapture barely began to thrum for us, when we found one another, when there was no more sleep and all the tones within us woke into the full chords of life, divine nature! then we were always like you, and now too when we part and joy dies, we are, like you, full of sorrow and yet good, therefore a pure mouth shall witness for us that our love is hallowed and eternal, like you.'
'I witness to it,' said the mother.
'We witness to it,' cried the others.
Now no words remained for us to speak. I felt my heart at its highest; I felt myself ripe for parting. 'Now I will forth, dear friends!' I said, and the life drained from all their faces. Diotima stood like a marble image and I felt her hand die in mine. I had killed everything around me, I was alone and reeled before the boundless silence in which my overflowing life no longer found a hold.
'Oh!' I cried, 'my heart burns hot within me, and you all stand there so cold, dear friends! and only the house-gods bend their ears? — Diotima! — you are silent, you don't see! — oh, it's well for you that you don't see!'
'Go then,' she sighed, 'it has to be; go then, you dear heart!'
'O sweet sound from these blissful lips!' I cried, and stood as in adoration before the lovely statue — 'sweet sound! one more time breathe upon me, dawn one more time, dear light of these eyes!'
'Don't speak that way, dearest!' she cried, 'speak in more earnest, speak with more greatness of heart to me!'
I wanted to hold myself back, but I was as in a dream.
'Woe!' I cried, 'this is a parting without return.'
'You'll kill her,' cried Notara. 'See how gentle she is, and you so beside yourself.'
I looked at her and tears burst from my burning eyes.
'Farewell then, Diotima!' I cried, 'heaven of my love, farewell! — Let us be strong, dear friends! dear mother! I gave you joy and grief. Farewell! farewell!'
I staggered forth. Diotima followed me alone.
Evening had come and the stars rose in the heavens. Below the house we stood in stillness. Eternity was in us, above us. Soft as the aether Diotima enfolded me. 'Foolish man, what is separation?' she whispered to me mysteriously, with the smile of an immortal.
'I do feel different now,' I said, 'and I cannot tell which of the two is a dream, my sorrow or my joyfulness.'
'They both are,' she answered, 'and both are good.'
'Perfect one!' I cried, 'I'll speak like you. By the starry heaven shall we know each other. Let it be the token between me and you for so long as our lips are mute.'
'Let it be!' she said with a lingering tone I'd never heard before — it was her last. Her image vanished away from me in the twilight, and I cannot tell if it was really she when I turned for the last time and the fading figure quivered one more moment before my eyes and then died into the night.
#### Hyperion to Bellarmin [XXXVII]
Why am I telling you, repeating my pain and stirring up again my restless youth in me? Isn't it enough to have gone through mortality once? why do I not remain still, in the peace of my spirit?
This is why, my Bellarmin! because every breath in life remains dear to the heart, because all changes in pure nature belong equally to her beauty. Our soul, when it casts off its mortal experiences and only lives alone in holy peace, is it not like a leafless tree? like a head without locks? Dear Bellarmin! I have rested a while; like a child I've been living among the quiet hills of Salamis, mindless of fate and the striving of men. Much has since changed in my eyes, and now I have peace enough within me to stay calm at every glance into human life. O friend! in the end the spirit reconciles us with all things. You won't believe it, at least not of me. But I think you should even see from my letters how daily my soul grows stiller and stiller. And I will go on speaking of it till you do believe it.
Here are letters from Diotima and me that we wrote to each other after my parting from Calauria. They are the most precious thing I confide to you. They are the warmest image from those days of my life. Of the clamour of war they tell you little. So much the more of my very own life, and that of course is what you want. And, oh, you must also see how loved I was. That I could never tell you, that Diotima alone can tell.
#### Hyperion to Diotima [XXXVIII]
I've awoken from the death of parting, my Diotima! fortified, as if from sleep, my spirit lifts itself up.
I'm writing to you from an Epidaurian mountain peak. There in the depths your island shimmers distant, Diotima! and out there my stadium where I must triumph or fall. O Peloponnese! O you springs of the Eurotas and Alpheus! That's where it will be decided! Out of the Spartan forests the ancient genius of the land shall swoop down like an eagle with our army, as on rushing wings.
My soul is full of lust for deeds and full of love, Diotima, and my eye gazes out into the Greek valleys as if it might magically command: arise again, you cities of the gods!
A god must be in me, for I hardly even feel our separation. Like the blissful shades of Lethe my soul now lives with yours in heavenly freedom and fate sways over our love no more.
#### Hyperion to Diotima [XXXIX]
I'm now in the middle of the Peloponnese. I'm spending tonight in the same hut where once I stayed when, barely more than a boy, I roamed through these parts with Adamas. How happily I sat then on the bench before the house and listened to the tinkling bells of the caravan, coming from afar, and the plash of the nearby fountain, pouring its silvery waters into the basin beneath blossoming acacias.
Now I'm happy again. I wander through this land as through Dodona's grove, where the oaks resounded with oracles prophesying glory. I see nothing but deeds, past ones, future ones, even when I wander from morning till evening under the open sky. Believe me, anyone travelling through this land who still can thole a yoke upon his neck, who doesn't become another Pelopidas, that man is empty of heart or lacking in sense.
So long this land has slept — so long has time dragged by, like the river of hell, drumbly and dumb, in desolate idleness?
And yet everything lies ready. The mountain folk hereabout are full of vengeful energies, couched there like a silent thundercloud that waits only to be driven by the storm-wind. Diotima! let me breathe the breath of God among them, let me speak to them a word from the heart, Diotima. Have no fear! They won't be so savage. I know raw nature. It scorns reason, but it's in league with inspiration. Whoever works with all his soul will never go astray. He has no need for overthought, for no force is against him.
#### Hyperion to Diotima [XL]
Tomorrow I shall be with Alabanda. It's a delight for me to ask the way to Coron, and I ask more often than is needed. I wish I could take the wings of the sun and fly to him, and yet I also like to linger so and ask myself: how will he be?
The kingly youth! why was I born later? why did I not spring with him from one cradle? I cannot suffer the difference that's between us. Oh, why did I live still like an idle shepherd-boy on Tinos and could only dream of the likes of him when already he was putting nature to the test in living labour, already battling sea and wind and all the elements? did I not also feel impelled towards the bliss of action?
But I will catch him up, I will be swift. By heaven! I'm overripe for work. My soul will only rage against itself if I don't soon free myself through living activity.
Noble maiden! how was I able to stand before you? How did you find it possible to love such a deedless being.
#### Hyperion to Diotima [XLI]
I have him, dearest Diotima!
Light is my breast and swift are my sinews, ha! and the future lures me, just as a clear watery deep lures us to leap in and cool our intemperate blood in the bracing bath. But that's just idle prattle. We're dearer to each other than ever, my Alabanda and I. We're freer together, and yet there's all fullness and deepness of life as before.
Oh, how right were those ancient tyrants to forbid friendships like ours! For then you're as strong as a demi-god and will thole no abuse in your sphere! —
It was evening when I entered his room. He had just put his work to one side, sat in a moonlit corner by the window, nursing his thoughts. I stood in the dark, he didn't recognize me, glanced perfunctorily in my direction. Heaven knows who he could have thought I was. 'So how's it going then?' he cried. 'Not too bad!' I said. But my dissembling was in vain. My voice was full of suppressed rejoicing. 'What's this?' he started up, 'is it you?' 'Of course it is, you blindling!' I cried, and flew into his arms. 'Oh now!' cried Alabanda at last, 'now there shall be change, Hyperion!'
'I think so too,' I said, and joyfully shook his hand.
'Do you still know me then,' Alabanda continued presently, 'do you still have your old pious faith in Alabanda? Generous soul! things meanwhile haven't gone as well for me as when I felt myself in the light of your love.'
'What?' I cried, 'Alabanda asks this? That was not proudly spoken, Alabanda. But it's the sign of the times that the old heroic nature goes begging for honour, and the living human heart, like an orphan, frets for a droplet of love.'
'Dear boy!' he cried; 'I've aged, that's all. This flabby life everywhere and the business with the old men whose disciple I tried to make you in Smyrna —'
'Oh, that's bitter,' I cried; 'even this man she's dared to assail, the goddess of death, the nameless one that they call fate.'
Light was brought and we gazed at each other anew with gentle loving scrutiny. My dearest's aspect was much changed since the days of hope. Like the mid-day sun from a blanched sky, his great ever-living eye gleamed at me from a faded face.
'My good young friend!' Alabanda cried, in friendly irritation at the gaze I gave him, 'enough of the doleful looks, there's a good fellow! I know full well I've gone downhill. O my Hyperion! I long so much for something great and true, and I hope to find it with you. You've grown beyond me, you're freer and stronger than you were and, believe it not! that makes me heartily glad. I'm the parched land and you come like a timely tempest — oh, it's glorious that you're here!'
'Hush!' I said, 'you rob me of my senses, and we shouldn't speak of ourselves at all until we're in the thick of life, amidst the deeds.'
'Quite right!' cried Alabanda joyously, 'it's only when the hunting horn sounds that the hunters feel fully what they are.'
'Will it soon begin then?' I said.
'It will,' cried Alabanda, 'and I tell you, dear heart! it should be quite a fire. Ha! may it reach right up to the tip of the tower and melt its vane, and rage and swirl about until the tower bursts and tumbles! — and don't let yourself be put off by our allies. I know full well the good Russians would like to use us as their guns. But don't you worry! once our stout Spartans have had the chance to find out who they are and what they can do, and when we've then conquered the Peloponnese, we'll laugh the North Pole in the face and build us a life of our own.'
'A life of our own,' I cried, 'a new, an honourable life. Were we then born of the swamp, like a will-o'-the-wisp, or do we stem from the victors of Salamis? How is it then now? how have you become a maidservant, free nature of the Greeks? how have you sunk so low, race of my fathers, of whom the divine effigies of Jupiter and Apollo were once merely the copy? — But hear me, heaven of Ionia! hear me, earth of the fatherland, you that half-naked, like a beggar-woman, wrap yourself in the rags of your ancient glory, I will thole it no longer!'
'O sun that reared us!' cried Alabanda, 'you shall be witness when under the toil our mettle grows, when under the blows of fate, like iron under the hammer, our project takes shape.'
We each of us kindled the other.
'And let not a blot stick,' I cried, 'none of the grotesques the age smears us with, like the rabble on walls.'
'Oh,' cried Alabanda, 'that's why war is so good —'
'Right, Alabanda,' I cried, 'like all great work where human energy and spirit alone helps and no crutch and no waxen wing. Then we shall cast off the slaves' livery on which fate has stamped its brand for us —'
'Then we'll have no more use for anything affected and forced,' cried Alabanda, 'then, just as in the Nemean races, we'll go unadorned, unfettered, naked to our goal.'
'To our goal,' I cried, 'where the young free state dawns and from Greek earth the pantheon of all that's beautiful rises.'
Alabanda was silent for a while. A fresh redness rose in his face and his form grew high, like a plant refreshed.
'O youth! youth!' he cried, 'then will I drink from your fount, then will I live and love. I'm full of joy, heaven of the night,' he went on as if drunk, walking to the window, 'you arch over me like a bower of vine leaves, and your stars hang down like grapes.'
#### Hyperion to Diotima [XLII]
It's as well for me I'm wholly busy with my work. I'd surely tumble from one folly to another, so full is my soul, so fuddled am I by this man, wonderful and proud, who loves nothing but me and heaps on me alone all the humility that's in him. O Diotima! this Alabanda has wept before me, has begged me like a child to forgive him what he did to me in Smyrna.
Who am I then, you dear ones, that I call you mine, that I may say, 'they are mine own,' that I stand between you like a conqueror and embrace you as my booty.
O Diotima! O Alabanda! noble beings, grand, serene! how must I achieve if I'm not to flee before my happiness, before you?
Just as I was writing I received your letter, dearest.
Do not grieve, lovely being, do not grieve! Save yourself, unscathed by sorrow, for the future festivals of the fatherland! Diotima! for the glowing festal day of nature save yourself, and for all the blithe days in honour of the gods!
Can you not see Greece already?
Oh, can you not see how the eternal stars, glad of their new neighbours, smile above our towns and groves, how the ancient ocean, when it sees our people stroll along the shore, is minded again of the beautiful Athenians and again brings us fortune, just as then for its darlings, on joyful waves?
Soulful maiden! you're already so beautiful now! how shall you blossom then, in ravishing glory, when the right climate nourishes you!
#### Diotima to Hyperion [XLIII]
I'd mostly shut myself in since you went away, dear Hyperion! Today I was once again outdoors.
In the clement February air I've gathered life and bring you what I've gathered. It's still done me good too, the fresh warming of heaven, still I've been able to share in it, the new bliss of the plant world, pure and ever the same, where everything grieves and joys again in its season.
Hyperion! O my Hyperion! why don't we also tread the quiet paths of life? They are holy names, winter and spring and summer and autumn! but we do not know them. Is it not sin to grieve in spring? why then do we do it?
Forgive me! the children of the earth live through the sun alone; I live through you, I have other joys, is it then wonder if I have other grief? and must I grieve? must I then?
Brave one! dear one! should I wither when you shine? should my heart wilt when the lust of victory wakes in all your sinews? Had I once heard that a Greek youth were setting forth to drag our good people from its shame, to return it to maternal beauty whence it sprang, how I should have roused from childhood's dream and thirsted for the image of the worthy one! and now that he's here, now that he's mine, can I still weep? O silly girl! is it not real? isn't he that glorious youth, and is he not mine? O you shades of blissful time, you my beloved memories!
For it seems to me barely yesterday, that magical evening when the holy stranger met me for the first time, when like a grieving genius he glanced into the sylvan shades where the carefree maiden sat in the dream of youth — in the breath of May he came, Ionia's magical breath of May, and this made him blossom the more for me, it waved his locks, it unfolded his lips like flowers, it dissolved into smiles his sorrow, and O you beams of heaven! how you did shine on me out of those eyes, those inspiriting springs where in the shade of sheltering arches eternal life shimmers and wells! —
Benevolent gods! how beautiful he became with his gaze upon me! how the whole youth, grown taller by a span, stood there in easy vigour, save that he let his loving arms hang humbly down as if they were nothing! and how he then looked up in rapture, as if I'd soared into heaven and were no longer there, oh! how he now smiled and blushed in all the grace of his heart when he became aware of me again and between the clouding tears his Phoebus eye shone through, to ask 'is it you? is it really you?'
And why did he meet me with such reverence, so full of loving superstitious awe? why had he first bowed his head, why was the divine youth so full of longing and sorrow? His genius was too blissful to remain alone, and too poor the world to comprehend it. Oh, that made for such a lovely image, woven of greatness and grief! But now it's different! the grief is gone! He's found his work to do, he's the sick man no more! —
I was full of sighing when I began to write to you, my love! Now I'm nought but joy. So does to speak of you make happy. And see! that's how it shall stay. Farewell!
#### Hyperion to Diotima [XLIV]
We've still found time to celebrate your feast day, lovely life! before the tumult begins. It was a heavenly day. Sweet springtime wafted and shone from the orient, it drew forth from us your name, just as it drew forth the blossoms from the trees, and all the blissful secrets of love poured out of me. A love like ours my friend had never known, and it was delightful how the proud man became all ear, how his eye and spirit glowed to grasp your image, your being.
'Oh,' he cried at last, 'then it's worth the effort fighting for our Greece if it still bears plants like these!'
'That's right, my Alabanda,' I said; 'then we'll go cheerful into battle, then heavenly fire will drive us on to deeds when our spirit is made young again by the image of such natures, and then we won't be chasing after petty goals, then we won't be bothering with this and that, tinkering with externals and ignoring the spirit, drinking the wine for the cup's sake; and we will not rest, Alabanda, till genial bliss is mystery no more, till all eyes are turned into triumphal arches from which the human spirit, so long absent, shines forth out of the erring and the suffering, and joyful in victory greets the paternal aether. — Ha! by the flag alone shall no one know our future people; all must be made new, there must be fundamental change; pleasure full of earnest and all labour full of cheer! nothing, not even the most paltry and commonplace, without the spirit and the gods! Love and hate and every sound from us must dumbfound the more vulgar world and never a single moment shall remind us of the humdrum past!'
#### Hyperion to Diotima [XLV]
The volcano's erupting. In Coron and Modon the Turks are under siege and we're pushing onwards with our mountain folk up the Peloponnese.
Now all melancholy's at an end, Diotima, and my spirit is firmer and quicker since I've been engaged in living labour, and see! I now even have a daily schedule.
I begin with the sun. Then I go out to where my warrior-folk lie in the forest shade, and greet the thousand bright eyes that now open up before me with wild affection. An awakening army! I know nothing like it, and all life in towns and villages is a swarm of bees by comparison.
Man cannot deny that once he was happy, like the deer of the forest, and after untold years there still glimmers in us a yearning for the days of the primal world when each roamed the earth like a god, before who knows what? made man tame, and instead of walls and dead wood the soul of the world, the all-pervasive holy air enveloped him still.
Diotima! I'm often filled with wonder when I walk amongst my carefree folk and, as if sprung from the earth, one after another rises and stretches towards the morning light, and among the troops of men the crackling flame leaps up where the mother sits with the shivering babe, where the quickening food cooks, whilst the horses, scenting the day, snort and whinny, and the forest resounds with all-shaking war music, and all around shimmers and rustles with weapons — but these are words, and the special delight of such a life is beyond telling.
Then my company gathers eagerly round me, and it's wonderful how even the oldest and thrawnest honour me in all my youth. We've come to confide in each other, and there's many a man will tell how he's been treated by life, and my heart often swells at this fate or that. Then I begin to talk of better days, and their eyes open wide and light up when they think of the covenant that shall join us and the proud image of the budding free state looms before them.
'All for each and each for all!' There's a joyful spirit in the words, and it always grips my men like a commandment from the gods. O Diotima! thus to see then how with hopes the hardened nature mellows and all its pulses beat higher and the grand designs smooth and clear the darkened brow, to stand thus in a sphere of human beings, surrounded by faith and joy, that's more than beholding earth and heaven and sea in all their glory.
Then I drill them in weapons and marching until noon. Their good cheer makes them willing pupils, and me it makes a master. Now they stand close-ranked in Macedonian phalanx and only move their arms, now they fly apart like rays of light to bolder strife in separate squads where supple force varies in every position and each man is his own commander, and then they assemble again at a point of safety — and always, wherever they move or stand in such a dance of weapons, there floats before their eyes and mine the image of the slaves of tyranny and the real battlefield.
Afterwards, when the sun shines hotter, council is held in the inner forest, and it's a joy to sway with quiet mind over our great future. We'll take away from chance its power, we shall master fate. We'll let resistance come on our terms, we'll goad the enemy into action for which we're well prepared. Or we'll look on, seemingly daunted, and let him come nearer till he offers his neck for the blow, we'll also confound him with our speed and that's my panacea. But more seasoned physicians think nothing of such cure-alls.
How good I then feel in the evenings with my Alabanda when for pleasure we roam the sun-red hills on sprightly steeds, and on the peaks where we linger the breeze plays in the manes of our mounts, and its friendly whispering mingles with our talk whilst we gaze out into the distances of Sparta that are to be our battle prize! and when we're returned and sit together in the lovely coolness of the night, and the wine cup wafts its fragrance, and the moonlight shines on our frugal meal, and in the midst of our smiling stillness the history of the ancients rises like a cloud from the sacred soil that bears us, how blissful is it then to grasp each other's hands in such a moment!
Then it may be that Alabanda speaks of many another tormented by the tedium of the age, of many a crooked path forged by life since its direct course has been blocked, and then my Adamas also comes to mind, with his journeyings, that longing of his which lured him into the depths of Asia — 'those are mere makeshifts, dear old man!' I'd like to call to him then, 'come! and build your world! with us! for our world is also yours.'
And your world too, Diotima, for it is a copy of you. O you, with your Elysian stillness, could we only create what you are!
#### Hyperion to Diotima [XLVI]
We've now had three successive victories in small skirmishes, but where the fighting men crisscrossed each other like lightning and everything was one consuming flame. Navarin is ours and now we stand before the fortress of Mistra, the remains of ancient Sparta. The flag I snatched from an Albanian horde I've planted on a ruin that lies before the city, I've flung for joy my Turkish turban into the Eurotas and have been wearing the Greek helmet ever since.
And now I'd like to see you, O maiden! I'd like to see you and take your hands and press them to my heart, for which the joy perhaps may soon be too great! soon! in a week perhaps it will be freed, the ancient, noble, holy Peloponnese.
O then, my dearest! teach me piety! then teach my overflowing heart a prayer! I should stay silent, for what have I done? and even had I done anything to speak of, yet how much still remains to do? But how can I help it if my thought is swifter than time? I'd much rather have it the other way round, with time and deed outflying thought, and winged victory overtaking hope itself.
My Alabanda blossoms like a bridegroom. From his every glance the coming world smiles out at me, and with that I can as yet just about still my impatience.
Diotima! I wouldn't swap this budding bliss for the finest heyday of ancient Greece, and the smallest of our victories means more to me than Marathon and Thermopylae and Plataea. Is it not true? Is not the convalescing life dearer to the heart than the untouched one that is yet to know sickness? Only when youth is gone do we love it, and only when lost youth comes again does it then gladden all the depths of the soul.
My tent stands by the Eurotas, and when I wake after midnight the ancient river-god murmurs reproachfully past me, and smiling I take the flowers of the bank and strew them on his glancing waves and say to him: 'Take this as a sign, you lonely one! Soon the life of old will blossom around you again.'
#### Diotima to Hyperion [XLVII]
I've received the letters, my Hyperion, that you wrote to me along the way. You seize me powerfully with all that you tell me, and amidst my love I often shudder to see the gentle youth who wept at my feet transformed into this warlike being.
Won't you then forget how to love?
But change as you will! I shall follow you. I think that if you could hate me I'd even then feel like you, I'd do my best to hate you too, and so our souls would remain akin and that's no idle vaunt, Hyperion.
I too am wholly changed from what I used to be. I'm missing my serene outlook on the world and my free delight in everything that lives. It's only the field of the stars that still attracts my eye. For that, I dwell the more fondly on the great spirits of ancient days and how they ended on earth, and those noble Spartan women have won my heart. Yet I don't forget the new warriors withal, the lusty ones whose hour is come; often I hear their triumphal clamour, nearer and nearer to me, roaring up through the Peloponnese, often I see them surging there like a cataract down through the Epidaurian woods, and their weapons glancing from afar in the sunlight that escorts them like a herald, O my Hyperion! and you hasten across to Calauria and greet the tranquil woods of our love, greet me and then fly back to your work; — and do you think I fear the outcome? Dearest! there are times I'm on the point of being overwhelmed, but my larger thoughts, like flames, hold the frost at bay. —
Farewell! carry it through as the spirit commands you! and don't let the war last too long, for the sake of the peace, Hyperion, for the sake of the beautiful, new, golden peace when, as you said, in our book of statutes will one day be inscribed the laws of nature, and when life itself, when divine nature, that can't be written into any book, will be in the heart of the community. Farewell.
## Hyperion to Diotima [XLVIII]
You should have soothed me, my Diotima! should have told me not to rush, told me to wrest from fate the victory bit by bit, as one does from tight debtors their dues. O maiden! standing still is worst of all. My blood dries up in my veins, so do I thirst to advance, and have to stand here idle, have to besiege and besiege day in and day out. Our people want to storm, but that would inflame their restive spirits beyond all temperance, and then woe to our hopes if the savage being seethes up and bursts the bonds of discipline and love.
I don't know, it can only be a matter of days before Mistra must surrender, but I wish we were further forward. Here in the camp it's like being in storm-laden air. I'm impatient, and I'm not happy with my men. There's a frightening wantonness amongst them.
But I'm silly to make so much of my mood. And it's surely worth enduring a worry or two for ancient Lacedaemon before it's won.
#### Hyperion to Diotima [XLIX]
It's over, Diotima! our people have plundered, murdered without distinction, even our brothers are slain, the Greeks in Mistra, the innocent, or they wander about helpless and the lifeless anguish on their faces cries out to heaven and earth for vengeance against the barbarians at whose head I stood.
Now I can go forth and preach my good cause. Oh, now all hearts will fly to me!
But I went about it cleverly, didn't I? I knew my people. Indeed! it was an extraordinary undertaking, to plant my Elysium with a band of brigands.
No! by holy Nemesis! I got what I deserved and I will bear it too, bear it till the pain tears my final consciousness asunder.
Do you think I'm raving? I have an honourable wound, dealt me by one of my faithful followers whilst I tried to avert the atrocity. If I were raving I'd rip the bandage from it, and then my blood would flow whither it belongs, into this grieving earth.
This grieving earth! this naked earth! that I sought to clothe with sacred groves, that I sought to adorn with all the flowers of Greek life!
Oh, it would have been beautiful, my Diotima.
Do you call me disheartened? Dear maiden! there's too much calamity. Demented hordes burst in on all sides; rapine rages in Morea like a plague and anyone who doesn't himself seize the sword is put to flight, butchered, and yet the maniacs claim they're fighting for our freedom. Others from amongst the raw people are in the pay of the Sultan and they behave just the same.
I've just heard that our inglorious army is now scattered. At Tripolitsa the cowards met an Albanian troop that numbered half as many. But since there was nothing to plunder, the wretches all ran away. The Russians who'd ventured on the campaign with us, forty brave men, were the only ones to stand their ground, and all found their deaths.
And so now I'm alone again with my Alabanda, as before. Since he saw me fall and bleed in Mistra this faithful friend has forgotten all else, his hopes, his lust of victory, his despair. The man of wrath who plunged amongst the plunderers like an avenging god, he now so gently guided me out of the turmoil and my clothing was wet with his tears. He's stayed with me too in the hut where I've since lain, and I'm now the more pleased about that. For had he gone on with the others, he'd now be lying in the dust at Tripolitsa.
How things will go from now on, I don't know. Fate thrusts me out into confusion and that's what I deserve; I'm banished from you by my own shame, and who knows for how long?
Oh! I promised you a Greece and now you get a dirge instead. Be your own comfort!
#### Hyperion to Diotima [L]
I can barely bring myself to speak.
We delight in speech, to be sure, we chatter away like the birds so long as the world wafts over us like a May breeze; but between noon and evening things can change, and what is lost in the end?
Believe me, and mind that I say this to you from my deepest soul: language is a great superfluity. The best is ever for itself and rests in its depths like the pearl at the bottom of the sea. — But what I really wanted to write to you, since the painting must have its frame and man his daily task, I intend to take service for a while with the Russian fleet; for I want nothing more to do with the Greeks.
O dear girl! it's grown very dark around me!
#### Hyperion to Diotima [LI]
I've lingered, I've struggled. But finally it has to be.
I see what needs to happen, and because I see it, it shall come to pass. Do not mistake my meaning! do not condemn me! I must advise you to leave me, my Diotima.
I can be for you nothing more, you lovely being! This heart has run dry, and my eyes see living life no more. Oh, my lips are withered; the sweet breath of love wells up in my bosom no more.
A single day has robbed me of all youth; by the Eurotas my life has wept itself weary, oh! by the Eurotas that in irredeemable shame wails with all its waves past Lacedaemon's rubble. It's there, there I was mown down by fate. — Am I to possess your love like alms? — I'm so utterly nothing, as unrenowned as the meanest menial. I'm shunned, cursed like a common rebel, and many a Greek in Morea will henceforth recount our heroic deeds to his children's children as a tale of thieves.
Oh! and one thing I've long kept from you. My father solemnly disowned me, banished me beyond recall from the house of my youth, never wants to see me again, either in this life or the next, as he puts it. This was his response to the letter in which I told him of my undertaking.
But don't ever let yourself be led astray by pity. Believe me, there's a joy that still remains for us throughout. True pain inspires. Whoever mounts his misery stands higher. And what's glorious is that only in suffering do we fully feel the freedom of the soul. Freedom! understand the word who can — it's a profound word, Diotima. I'm so deeply beleaguered, so incredibly hurt, I'm without hope, without a goal, I'm utterly bereft of honour, and yet there's a power within me, an indomitable something that sends sweet shudders through my bones whenever it stirs in me.
And I still have my Alabanda. He has as little to gain as I do. Him I can keep for myself without harm! Oh! the kingly youth would have deserved a better lot. He's become so gentle and so still. That often nearly rends my heart. But each of us sustains the other. We say nothing to each other; what should we say? But for all that there's a blessing in many a little labour of love we perform for one another.
There he sleeps and smiles contentedly, in the midst of our misfortune. The good soul! he doesn't know what I'm doing. He wouldn't suffer it. 'You must write to Diotima,' he charged me, 'and must tell her to make ready soon to flee with you to a more liveable land.' But he doesn't know that a heart that's learned despair, such as his and mine, can be nothing more for the beloved. No! no! you could never ever find peace with Hyperion, you would have to become unfaithful, and that I will spare you.
And so farewell then, you sweet girl! farewell! I'd like to tell you, 'go hither, go thither; there the springs of life purl forth.' I'd like to point you to a freer land, a land full of beauty and full of soul, and say: 'Escape thither!' But oh heaven! if I could do that, then I'd also be another and so there'd be no need for me to part from you — part? Alas! I know not what I do. I weened myself so calm, so cool. Now my head reels and my heart flings itself about like a restless patient. Woe on me! I destroy my last joy. But it has to be and nature's 'Alas!' avails nothing here. I owe this to you, and besides I was born to be homeless and have no resting place. O earth! O you stars! shall I find nowhere to dwell in the end?
Just one more time I should like to return to your bosom, no matter where! Aether eyes! one more time in you to meet myself! to hang upon your lips, you lovely! you ineffable one! to drink in your rapturous, holy-sweet life — but don't listen to that! I beseech you, pay no heed to that! I'd say I was a seducer, if you were to listen to it. You know me, you understand me. You know how deeply you heed me by not pitying me, by not listening to me.
I can, I may no more — how can the priest go on living if his god is no more? O genius of my people! O soul of Greece! down I must go, in the realm of the dead I must seek you.
#### Hyperion to Diotima [LII]
I've waited long, I'll confess to you, I've been hoping longingly for a word of farewell from your heart, but you remain silent. That too is the language of your beautiful soul, Diotima.
Yet it's true, is it not, this doesn't mean the holier chords must cease? it's true, is it not, Diotima, though the gentle moonlight of love may sink, the higher stars in its heaven still continue to shine? Oh, this is my final joy, that we remain inseparable, though no sound more return from you to me, no shadow of our lovely days of youth!
I look out into the sunset sea, I stretch out my arms towards the airt where you live, far away, and my soul warms once more with all the joys of love and youth.
O earth! my cradle! all bliss and all pain is in the leave we take of you.
You dear Ionian islands! and you, my Calauria, and you my Tinos, you're all before my eye, distant as you are, and my spirit flits with the breezes over the lively waters; and you that there shimmer dimly to one side, you shores of Teos and Ephesus where once I walked with Alabanda in the days of hope, you shine again for me as then, and I'd like to sail across to the land and kiss the soil and warm the soil at my bosom, and stammer all the sweet words of parting to the silent earth before I fly up into freedom.
Pity, pity, that things don't go better now amongst men, else I'd gladly remain on this good star. But I can do without this earthly orb, and that's worth more than all that it can give.
'Let us, O child! suffer slavery in the sunlight,' said the mother to Polyxena, and her love of life could not have spoken more beautifully. But the sunlight it is that deters me from slavery, it won't let me stay on this degraded earth and, like paths leading home, the hallowed rays draw me on.
For long the majesty of the fateless soul has been more present to me than all else; I've often lived within myself in glorious solitude; I've grown used to shaking off all outward things like flakes of snow; how should I then be feared to seek out so-called death? have I not freed myself a thousand times in thought, how should I then hold back from really doing it for once? Are we then like serfs, fettered to the soil we plough? are we like tame fowl that durstn't run out of the farmyard because that's where they're fed?
We're like the young eagles the father drives from the nest, that they may search for their prey in the high aether.
Tomorrow our fleet goes into combat and this one will be hot enough. I look on the battle as a bath to wash the dust off me; and I expect to find what I wish for; wishes like mine are readily granted on the spot. And so I would end after all by achieving something through my campaign, and see that amongst men no effort goes to waste.
'Pious soul!' I'd like to say, 'think of me when you come to my grave.' But they'll probably cast me into the surging sea, and I'm happy to see my remains sink down where the sources all and the streams that I loved gather together, and where the storm-cloud rises up and waters the hills and the valleys that I loved. And we? O Diotima! Diotima! when shall we see one another again?
It's impossible, and my innermost life rebels at the thought that we could lose each other. I would wander the stars for millennia, clothe myself in all forms, in all of life's languages just to meet you once more. But I believe that like will soon find like.
Great soul! you'll be able to accept this parting, and so let me wander! Greet your mother! Greet Notara and the other friends!
Also greet the trees where I first met you, and the merry rivulets where we walked, and the lovely gardens of Angele, and let my image meet you when you do, my love! Farewell.
## Hyperion to Bellarmin [LIII]
I was in a lovely dream when I copied out for you the letters that I once exchanged. Now I write to you again, my Bellarmin! and take you further down, down to the deepest depths of my sorrows, and then, last of my loved ones! come forth with me to the place where a new day shines upon us.
The battle I'd written about to Diotima began. The ships of the Turks had taken refuge in the channel between the island of Chios and the Asian coast, and stood along the mainland off Chesma. My admiral left the line with his ship, the one I was on, and began the prelude with the first ship of the Turks. At the very first attack the raging pair were heated to a frenzy, it was a terrible revenge-sodden turmoil. The ships were soon bound fast together with their cordage; the furious fighting grew ever tighter and tighter.
A deep feeling of life coursed through me once more. I felt warm and well in every limb. Like one tenderly departing, my spirit felt itself for the last time in all its senses. And now, full of hot resentment that I knew no better way than to let myself be slaughtered among a throng of barbarians, and with tears of rage in my eyes, I stormed to where my death was certain.
I found the foe close enough to hand and, of the Russians who fought at my side, in a matter of moments not even one remained. I stood there alone, full of pride, and cast my life like a beggar's penny before the barbarians, but they didn't want me. They looked on me as someone whom one fears to sin against, and fate seemed to reverence me in my despair.
In sheer self-defence one of them at last laid into me and struck me so that I fell. From then on I was aware of nothing more until I came to again on Paros, whither I'd been shipped.
From the servant who carried me out of the battle I afterwards heard that the two ships that began the combat had blown up only moments after he'd taken me off in a skiff with the surgeon. The Russians had hurled fire into the Turkish ship, and because their own was entangled with it, both went up together.
How this terrible battle ended you will know. 'So does one poison purge the other,' I cried, when I heard that the Russians had burned the entire Turkish fleet — 'so do tyrants wipe each other out.'
#### Hyperion to Bellarmin [LIV]
For six days after the battle I lay in a tortured death-like sleep. My life was like a night, punctuated by pain as by lancing lightning. The first thing I recognized was Alabanda. He had, as I heard, not left my side for an instant, he'd cared for me almost alone, with unbelievable solicitude, with a thousand tender homely attentions that would never in his life have occurred to him otherwise, and he'd been heard on his knees before my bed, crying: 'O live, dear friend! that I may live!'
It was a happy awakening, Bellarmin! when my eye now opened again to the light, and with tears of reunion this glorious man stood before me.
I reached out my hand to him and, proud as he was, he kissed it with all the rapture of love. 'He lives,' he cried, 'O nature! you kindly all-healing saviour! you at least don't forsake your wretched pair, these errants without a fatherland! Oh, I'll never forget, Hyperion! how your ship went up in fire before my eyes and thundering swept up the sailors with it into the raging flames, and amongst those few to be saved there was no Hyperion. I was out of my senses, and the furious clamour of battle did nothing to calm me. Yet I soon got word of you, and flew after you as soon as we were altogether finished with the foe.' —
And how he now watched over me! how with loving care he held me captive in the magic circle of his favours! how without a word he taught me through his great calmness to understand without rancour and like a man the free course of the world!
O you sons of the sun! you freer souls! much has been lost in this Alabanda. I've sought, and besought life in vain since he went away; such a truly Roman nature I never found again. Carefree, deeply wise, courageous, noble! Where is there a man if he was not one? And when he was mild and meek, it was as when the evening light plays in the gloom of the majestic oak and its leaves drip from the storm of the day.
#### Hyperion to Bellarmin [LV]
It was in the beautiful days of autumn when, half-recovered from my wound, I went to the window again for the first time. I returned to life with calmer senses and my soul had become more attentive. With its gentlest magic heaven breathed upon me, and mild as a rain of blossom the bright sunbeams streamed down. There was a great, still, tender spirit in the season, and the peace of fulfilment, the bliss of ripeness in the murmuring boughs, enfolded me like the renewed youth the ancients hoped for in their Elysium.
I'd long not enjoyed it with pure soul, this childlike life of the world; now my eyes opened up with all the joy of seeing anew, and blissful nature had remained changeless in her beauty. My tears flowed before her as an offering of atonement, and a fresh heart rose shuddering from my former discontent. 'O sacred plant world!' I cried, 'we strive and we scheme, and yet we have you! we struggle with our mortal powers to build the beautiful, and yet it grows up effortless beside us! it's true, is it not, Alabanda? man is made to care for bare necessity, everything else looks after itself. And yet — I cannot forget how much more I wanted.'
'Be happy, dear friend! that you are,' cried Alabanda, 'and no longer let your quiet working be hampered by grief.'
'And I will rest too,' I said. 'Oh, all the designs, all the demands I will rip up like letters of debt. I will keep myself pure as an artist does; you will I love, innocent life, life of the groves and the springs! you will I honour, O sunlight! on you will I still myself, beautiful aether, you that inspirit the stars and here too breathe around these trees and here stir us deep within our breast! O wilfulness of men! like a beggar I bowed my neck and the silent gods of nature gazed at me with all their gifts! — You're smiling, Alabanda? Oh, how often in our early days together did you smile that way when your boy prattled away before you in the drunken exuberance of youth, whilst you stood there like a silent temple pillar, amid the rubble of the world, and had to suffer the wild tendrils of my love to wind around you — look! it's like a blindfold falling from my eyes and the golden days of old come to life again.'
'Ah!' he cried, 'this earnestness in which we lived and this lust for life!'
'When we hunted in the forest,' I cried, 'when we bathed in the surging sea, when we sang and we drank where the sun and the wine and eyes and lips sparkled through the shade of the laurel — it was a life beyond compare, and our spirit illumed like a shining heaven our youthful happiness.' 'And that is why neither can forsake the other,' said Alabanda.
'Oh, I have a grave confession to make to you,' I said. 'Will you believe me when I say I wanted away? from you! that I violently sought my death! was that not heartless? madness? ah, and my Diotima! she should leave me, I wrote to her, and then another letter, on the eve of the battle —' 'and there you wrote to her,' he cried, 'that in the battle you hoped to find your end? O Hyperion! But she may well not yet have received your last letter. You just have to make haste and write to her that you're alive.'
'Dearest Alabanda!' I cried, 'that is solace! I'll write at once and dispatch my servant with the letter. Oh, I will offer him all I have to rush and reach Calauria in time. —'
'And the other letter, in which you spoke of renouncing her, the good soul will understand and readily forgive,' he added.
'She'll forgive?' I cried; 'O all you hopes! yes! if I could still be happy with the angel!'
'You will still be happy,' cried Alabanda; 'you still have left the most beautiful of life's seasons. The youth is a hero, the man a god if he can live to see it.'
There was a wondrous dawning in my soul as he spoke.
The tree tops softly shuddered; like flowers from the dark earth, stars sprang forth from the womb of the night and heaven's spring sparkled on me in holy joy.
#### Hyperion to Bellarmin [LVI]
A few moments later, just as I was about to write to Diotima, Alabanda came back gleefully into the room. 'A letter, Hyperion,' he cried; I startled and flew across.
'How long,' wrote Diotima, 'had I to live without a sign from you! You wrote to me about the fateful day at Mistra and I responded swiftly; yet it seems you never received my letter. Soon afterwards you wrote to me again, briefly and gloomily, and told me you were minded to join the Russian fleet; again I replied; but you didn't receive that letter either; then it was my turn to wait in vain, from May till now, the end of summer, until a few days ago the letter came, telling me I should renounce you, my love!
'You counted on me, had faith enough in me to know this letter couldn't offend me. That gave me heartfelt joy in the midst of my grief.
'Unhappy sublime spirit! I've understood you only too well. Oh, it's so wholly natural that you no longer want to love, since your greater wishes languish. For must you not spurn food when you're dying of thirst?
'I soon came to know that I couldn't be everything to you. Could I loosen the bonds of mortality for you? could I still the flame in your breast for which no fountain flows and no vine grows? could I hand you the joys of a world in a chalice?
'That's what you want. That's what you need, and you can do no other. The boundless impotence of your fellow-men has robbed you of your life.
'He who, like you, has suffered hurt in his whole soul can no longer rest in solitary joy, he who, like you, has suffered shallow nothingness quickens only in the highest spirit, he who has suffered death, like you, heals among the gods alone.
'Happy all those who do not understand you! He who understands you must share your greatness and your despair.
'I found you as you are. Life's first curiosity drove me to that wondrous being. Ineffably the tender soul drew me on, and childishly fearless I played about your dangerous flame. — The beautiful joys of our love appeased you; but only, wicked man! to make you more savage. They soothed, they solaced me too, they made me forget you were fundamentally beyond solace, and that I wasn't far from becoming so myself, ever since I'd looked into your beloved heart.
'In Athens, among the ruins of the Olympieion, it seized me anew. I might well still have thought in a carefree hour that the grief of the youth could not really be so earnest and unrelenting. It's so rare that with his first step into life a man has felt so all at once, so minutely, so quickly, so deeply the whole fate of his age, and that it inheres indelibly within him, this feeling, because he isn't rough enough to drive it out and not weak enough to weep it away; that, my dearest! is so rare that we deem it almost unnatural.
'Now, amidst the rubble of bright Athens, now it came home to me all too painfully how the tables are turned, that now it's the dead who walk above the earth and the quick, the divine men, who are below, now I also saw it too plainly and palpably written on your face, now once and for all I conceded you were right. But at the same time you appeared greater to me too. A being full of hidden power, full of deep, undeveloped meaning, a youth uniquely promising, so you seemed to me. He to whom fate speaks so loud is entitled to speak still louder back to fate, I told myself; the more unfathomably he suffers, the more unfathomably powerful he is. From you, from you alone I hoped for all healing. I saw you travelling. I saw you working. Oh, what a transformation! Founded by you, the grove of Academe greened again above the listening pupils, and the plane tree of Ilissus again heard holy discourse as of old.
'In your school the genius of our youth soon gained the earnestness of the ancients and its ephemeral games grew more immortal, for it became ashamed, holding butterfly flight to be imprisonment. —
'This one might have been content to steer a steed; now he's general. His whole pleasure another might have found in warbling an idle ditty; now he's an artist. For you'd laid bare before them in open combat the energies of the heroes, the energies of the world; you'd given them the riddles of your heart to solve; and so the youths learned to bring together what is great, learned to understand nature's play, full of soul, and put aside frivolity. — Hyperion! Hyperion! have you not made me, the mute maiden, into a muse? So it went with the others too.
'Oh! now these congenial men didn't leave one another so lightly; no longer did they drift aimlessly amongst one another like sand in the storm of the desert, nor did youth and age scorn each other, nor the stranger want a welcoming host, and fellow countrymen never more hived themselves off and lovers never more wearied of each other; from your springs, nature, they refreshed themselves, oh! with the holy joys that well up mysteriously from your depths and renew the spirit; and the gods quickened again the withering souls of men; every bond of friendship amongst them was guarded by the heart-sustaining gods. For you, Hyperion! had healed the eyes of your Greeks, that they could see living life; and the inspiration sleeping amongst them like fire in wood had been kindled by you, that they could feel the still and steady inspiration of nature and its pure children. Oh! now men no longer took the beautiful world like laymen the artist's poem when they praise the words and note the utility. A magical model you became for the Greeks, living nature! and kindled by the bliss of the ever-youthful gods, all man's doings were a festival as of old; and to their deeds the young heroes were escorted by Helios' light, more beautiful than any martial music.
'Hush! hush! It was my most beautiful dream, my first and my last. You're too proud to busy yourself further with this knavish race. And you're quite right in that. You led them to freedom and they thought of pillage. You lead them victorious into their ancient Lacedaemon and these monsters plunder, and by your own father you're cursed, great son! and no wilderness, no cave is safe enough for you on this Greek soil that you've reverenced like holy ground, that you've loved more than me.
'O my Hyperion! I'm the gentle maid no more since I've known all this. Outrage drives me upwards, that I can barely look to earth and my hurt heart won't stop trembling.
'We must part. You're right. Besides, I don't want any children; for I grudge them to the world of slaves, and in this drouth the poor plants would wither away before my eyes.
'Farewell! you dear youth! go where it seems to you worth the effort to offer up your soul. The world must have one battlefield, one place of sacrifice where you can find release. It would be a pity if so the good energies all passed away like a dream. But however you find your end, you'll return to the gods, return to the holy, free, youthful life of nature whence you came, and that is all you long for and I too.'
So she wrote to me. I was shaken to the marrow, full of fear and joy, yet I sought to compose myself and find words for a reply.
'You consent, Diotima?' I wrote, 'you think my renunciation right? you could comprehend it? — Faithful soul! you could reconcile yourself to that? Even to my darkest erring you could reconcile yourself, heavenly patience! and surrendered yourself, gloomed yourself for love, happy child of nature! and became like me and sanctified by sharing it my grief? Beautiful heroine! what crown have you not earned?
'But let there now be an end to grieving, my love! You've followed me into my night, now come! and let me follow you into your light, to your grace let us return, beautiful heart! oh, your tranquillity let me see again, blissful nature! and for ever let my hubris pass away before the image of your peace.
'It's true, isn't it, dearest! that it's not yet too late for my return, and you'll take me back and can love me again as before? it's true, isn't it, that the happiness of past days is not yet lost to us?
'I've taken it all to extremes. I've acted the ingrate towards my motherly earth, flinging away like a serf's pittance my blood and all the loving gifts she gave me, and oh! a thousand times more of an ingrate towards you, holy maiden! who once received me into her peace, me, a feral, fractured being, from whose deeply oppressed breast stole hardly a glimmer of youth, like the odd blade of grass on bare-trodden paths. Had you not called me into life? was I not yours? how could I then — oh, you don't yet know, I hope, haven't yet in your hands the unhappy letter I wrote to you before the last battle? I wanted to die then, Diotima, and imagined I was doing holy work. But how can that be holy which pulls lovers apart? how can that be holy which ruptures our life's pious happiness? — Diotima! life born of beauty! for that I've now become the more similar to you in what is uniquely yours, I've finally learned to reverence, I've learned to preserve what's good and essential on earth. Oh, even if I could land up there in the radiant islands of heaven, would I find more than I do with Diotima?
'Hear me now, beloved!
'In Greece I can remain no longer. That you know. When he cast me off my father sent me from his surplus funds enough for us to flee into some holy vale of the Alps or Pyrenees, and there to buy a couthie cottage and enough green earth to meet the needs of life's golden mean.
'If you're willing, then I'll come at once and lead you and your mother on my faithful arm, and we'll kiss Calauria's shores and dry away our tears, and hasten across the Isthmus to the Adriatic Sea whence a safe ship shall carry us onwards.
'Oh, come! in the depths of the mountain world the secret of our hearts shall rest like the precious stone in the mine, in the womb of the heaven-soaring forests we'll feel ourselves as amidst the pillars of the innermost temple where the godless dare not near, and we'll sit by the source, contemplating our world in its mirror, heaven and house and garden and us. Often on bright nights we'll stroll in the shade of our fruit trees and listen for the god in us, the loving one, while the plant from midday slumber raises its sunken head and the gentle life of your flowers quickens when they bathe their tender arms in the dew and the night breeze breathes coolingly around them, infusing them, and over us blossoms the meadow of heaven with all its sparkling flowers, and to the side behind westering clouds, as if out of love, shyly the moonlight mimics the sun-youth's setting — and then in the mornings when, like a river-bed, our valley fills with warm light and the golden flood runs tranquilly through our trees and flows around our house, enhancing the loveliness of the rooms, your creation, and you walk in the sunlit radiance and in your grace you sanctify the day for me, my love! when then, as we thus celebrate the bliss of morning, the bustling life of the earth kindles before our eyes like a sacrificial fire, and we go forth on our daily tasks to cast them too, as a part of us too, into the rising flame, will you not then say that we're happy, that we're become again like the ancient priests of nature, holy and joyful, those who were pious ere ever a temple stood.
'Have I said enough? decide now my fate, dear maiden, and soon! — It's as well that I'm still half an invalid from the last battle, and that I've yet to be released from service; else I couldn't stay here, I'd need to be off to ask you in person, and that wouldn't be good, that would be to besiege you. —
'O Diotima! foolish qualms strike me at the heart, and yet — I can't thole the thought that this hope too shall founder.
'Are you not grown too great to return to the happiness of earth? The fierce flame of the spirit, kindled by your sorrow, will it not consume everything mortal in you?
'I know full well that he who lightly falls out with the world more lightly makes peace with it too. But you, with your childlike stillness, you so happy once in your high humility, Diotima! who will appease you when fate arouses you?
'Dear life! is there no more healing power for you in me? of all the sounds from the heart is there none to call you back again to human life where you once so lovingly lingered in low flight? oh come, oh remain in this twilight! This shadowland is the element of love, and only here does the still dew of sadness run from the heaven of your eyes.
'And do you mind our golden days no more? those lovely, divinely melodious days? do they not whisper to you from all the groves of Calauria?
'And look! there's so much has perished in me, and I no longer have many hopes. Your image with its heavenly disposition I managed still to rescue like a house-god from the flames. Our life, ours, is still intact in me. Should I now go and bury this too? Shall I go forth, restless and without a goal, from one alien land to another? Is it for this I learned to love?
'Oh no! you first and you last! Mine you were and mine you will stay.'
### Hyperion to Bellarmin [LVII]
I was sitting with Alabanda on a nearby hill, in pleasantly warming sun, and around us played the wind with the fallen leaves. The land lay still; only here and there from the forest there came the sound of a toppling tree, felled by the landman, and beside us the ephemeral rainstream murmuring down to the tranquil sea.
I was as good as free of worry; I hoped now soon to see my Diotima, soon to live with her in quiet happiness. Alabanda had talked me out of all my doubts; so sure was he himself about it. He too was blithe; but in another sense. The future had no power over him any more. Oh, I didn't know it; he was at the end of his joys; with all his claims on the world, with all his conquering nature, he saw himself useless, ineffectual and isolated, and took it just as if he'd played some trivial game and lost.
A messenger now approached us. He brought us the discharge from war service that we'd both requested from the Russian fleet, since for us there was nothing more to do that seemed worth the effort. I could now leave Paros when I pleased. And I was now fit enough to travel. I didn't want to wait for Diotima's answer, I wanted away, to her; it was as if a god propelled me to Calauria. When Alabanda heard this his colour changed, and he looked at me in sadness. 'Is it so easy for my Hyperion,' he cried, 'to leave his Alabanda?'
'Leave,' I said, 'how so?'
'Oh, you dreamers!' he cried, 'can't you see that we must part?'
'How should I see it?' I replied; 'you never speak of it; and if now and then I noticed anything about you that might have hinted at a parting, I took it readily for whimsy, abundance of the heart —'
'Oh, I know it well,' he cried, 'this divine game of affluent love that creates want for itself so as to unburden itself of its own fullness, and I would it were so with me, good Hyperion! but this is meant in earnest!'
'In earnest?' I cried, 'and why then?'
'Because, my Hyperion,' he said gently, 'I wouldn't wish to disturb your future happiness, that's why; because I must fear Diotima's nearness. Believe me, it's a parlous thing to live around lovers, and a deedless heart, such as mine now is, will hardly be able to bear it.'
'O good Alabanda!' I said with a smile, 'how you mistake yourself! You're not made of wax, and your steadfast soul won't so lightly leap beyond its bounds. For the first time in your life you're being capricious. You played the sick-nurse here for me, and one sees how little you're made for the part. Sitting still has made you skittish —'
'You see,' he said, 'that's just the point. Shall I be any more active living with the two of you? and if it were any other woman! but this Diotima! can I do otherwise? can I feel her with half my soul? she who's through and through so wholly one, one single divinely undivided life? Believe me, it's childish folly to seek to see this being without love. You're staring at me as if you didn't know me? Well, I've become alien to myself these past few days, ever since her being came so alive in me.'
'Oh, why can't I give her to you?' I cried.
'Stop it!' he said. 'Don't try to comfort me, for here there can be no comfort. I'm alone, alone, and my life's running out like an hourglass.'
'Great soul!' I cried, 'must it come to this with you?'
'Put your mind at rest!' he said. 'I was already beginning to wither when we found each other in Smyrna. Yes! when I was still a ship's boy, and my spirit and all my limbs became quick and strong on rude fare and brave toil! when in the clear air after a stormy night I'd cling to the top of the mast, beneath the fluttering flag, and gaze out after the seabirds across the shining deep, when in battle our wrathful ships often churned up the sea like the boar's tusk the earth, and I'd stand bright-eyed by my captain's side — then I was alive, oh, then I was alive! And long afterwards, when the young Tiniot crossed my path on the Smyrna shore, with his earnestness, his love, and my hardened soul thawed again under the glances of the youth and learned to love and hold holy all that's too good to be governed, when with him I began a new life, and there burgeoned within me new and more soulful energies for enjoying and battling the world, then I hoped again — oh! and all that I hoped and had was chained to you; I pulled you to me, tried to drag you forcibly into my fate, lost you, found you again, our friendship alone was my world, my worth, my fame; now that too is over, for ever, and my whole existence is in vain.'
'Is that really true?' I replied with a sigh.
'True as the sun,' he cried, 'but never you mind! it's all taken care of.' 'How so, my Alabanda?' I said.
'Let me tell you,' he said. 'There's something I've never fully spoken to you about. And then — it will calm you and me a little too, if we speak about the past.
'I was once walking helpless by the harbour of Trieste. The privateer I used to serve on had foundered some years before, and together with a few others I'd barely been able to save myself, washing ashore at Seville. My captain had drowned, and my life and my sopping clothes were all I had left. I undressed and rested in the sun, drying my clothes on the bushes. Then I walked on along the road towards the town. Before I'd reached the gates I saw a merry company in the gardens, went in and sang a cheerful Greek ditty. I didn't know a sad one. It made me burn with shame and pain, displaying my misfortune in this way. I was an eighteen-year-old lad, wild and proud, and hated like death to be the object of people's attention. "Forgive me," I said, when I was done with my song; "I've just been shipwrecked and for the moment can't think of anything better to do for the world than sing." I'd said this in Spanish as best I could. A man with a distinguished face came up to me, gave me money, and said in our tongue with a smile: "Here! use that to buy yourself a whetstone and learn to sharpen knives, and make your wandering way on solid earth." The advice pleased me. "Sir! that's indeed what I'll do," I replied. Having also been generously rewarded by the others, I went and did as the man had advised, and for some time roved about Spain and France.
'What I experienced during this time, how on the thousandfold forms of bondage my love of liberty honed itself, and how from the many hardships my mettle and my shrewdness grew, I've often relished telling you.
'I pursued my innocent wandering work with pleasure, but in the end it became soured for me.
'People took it for a mask, because I can't have looked suitably vulgar; they imagined I was covertly engaged in some dangerous activity, and I was in fact arrested twice. That induced me then to give it up, and with what little money I'd earned I set out on the journey back to the home from which I'd once run away. I'd got as far as Trieste, and was about to move on, down through Dalmatia. Then the rigours of travel caused a sickness that laid me low, and with that my little wealth was all used up. And so, half recovered, I came to be walking sadly by the harbour of Trieste. Suddenly there stood before me the man who'd once helped me out when I was cast ashore at Seville. He was peculiarly glad to see me again, told me he often thought of me and asked me how I'd fared the while. I told him everything. "I see," he cried, "that it wasn't in vain to send you for a little to the school of fate. You've learned to endure, now if you wish you shall act."
'The words, his tone, the clasp of his hand, his mien, his glance—with the force of a god all this struck my being, made now more flammable than ever by much suffering, and I gave myself up to it.
'The man I'm speaking about, Hyperion, was one of those you saw with me in Smyrna. The very next night he inducted me into a solemn society. A shudder ran through me when I stepped into the hall and, as I entered, my companion pointed to the grave men and said: "this is the League of Nemesis." Drunk with the grand sphere of action that opened up before me, I solemnly made over my blood and my soul to these men. Soon afterwards the gathering was adjourned, to be resumed years later elsewhere, and each of us set out on the allotted path he was to take through the world. I was assigned to those you found with me in Smyrna some years after.
'The constraint under which I lived often tormented me, I saw precious little of the grand actions of the League, and my lust for deeds found scant nourishment. Yet none of this sufficed to make me fall away. It was my passion for you that finally seduced me. I've often told you, I was as if without air and sun when you were gone; and I had no other choice; I had to give up either you or my League. What I chose you see.
'But all of the works of men have in the end their punishment, and it's only gods and children whom Nemesis doesn't strike.
'I preferred the divine right of the heart. For the sake of my darling I broke my oath. Was that not right? must not the noblest longing be the freest? — My heart has taken me at my word; I gave it freedom and, as you see, it's putting it to use.
'Pay homage to the genius but once, and he'll heed no mortal hindrance any more and tear in two all life's bonds for you.
'Obligation I broke for the sake of the friend, friendship I would break for the sake of love. For Diotima's sake I would betray you, and in the end murder myself and Diotima because we should still not be one. But that's not the way it's going to be; if I'm to atone for what I did, then I will do it with freedom; I shall choose my own judges; those whom I've failed shall have me.'
'You mean your brothers of the League?' I cried; 'O my Alabanda! don't!'
'What can they take from me but my blood?' he replied. Then he gently grasped my hand. 'Hyperion!' he cried, 'my time is up, and all that's left for me is a noble end. Leave me be! Don't belittle me, put faith in my word! I know as well as you do that I could still fabricate an existence for myself, could, since the banquet of life is all consumed, still play with the crumbs, but that's not my way; nor is it yours. Need I say more? Do I not speak from your soul? I thirst for air, for a cooling breeze, Hyperion! My soul wells over of itself and will no longer be held in the old circle. Soon the beautiful winter days will come when the dark earth is nothing more than the foil to shining heaven; then would be a good time, then the islands of light more liberally sparkle besides! — you wonder at my words? Dearest! those taking leave of life all speak like drunken men and like to put on a festive manner. When the tree begins to wither, don't its leaves all wear the red hues of dawn?'
'Great soul,' I cried, 'must I bear pity for you?'
By his exaltation I could feel the depth of his suffering. Never in my life had I known such woe. And yet, O Bellarmin! yet I also felt the greatness of all joys, having such an image of divinity in my eyes and arms. 'Yes! die then,' I cried, 'die! your heart is glorious enough, your life is ripe as grapes on an autumn day. Go, you consummate one! I would go with you, were there no Diotima.'
'Do I have you now?' replied Alabanda, 'do you speak this way? How deep, how soulful all becomes, once my Hyperion grasps it!'
'He's flattering me,' I cried, 'trying to coax from me a second time my ill-considered word! good gods! and gain my permission for the trip to the blood-tribunal!'
'I'm not flattering you,' he replied in earnest, 'I've a right to do what you would hinder, and it's no mean one! honour that!'
There was a fire in his eyes that struck me down like a god's command, and I felt ashamed to say another word against him.
'They won't,' I kept telling myself, 'they can't. It's too senseless putting such a glorious life to the slaughter like some sacrificial beast,' and this conviction served to calm me.
It was a special bonus to hear him still the following night, after each of us had prepared for his own journey and we'd gone out again before daybreak to be alone together once more.
'Do you know,' he said, amongst other things, 'why I've never been awed by death? I feel in me a life that no god has created and no mortal man begotten. I believe we owe our being to ourselves and it's only of our own free desire that we're so closely bound into the all.'
'I've never heard you talk like this before,' I answered.
'And what,' he continued, 'what would this world be, were it not a unison of free beings? did not of their own joyous impulse the living join together from the very beginning in one full-voiced life, how wooden would it be, how cold? what a heartless piece of work would it be?'
'So here in the highest sense it were true,' I replied, 'without freedom everything is dead.'
'Yes indeed,' he cried, 'there's not a blade of grass grows up unless it has its own germ of life within! how much the more in me! and therefore, my dear friend! because I feel myself free in the highest sense, because I feel myself beginningless, therefore I believe that I'm endless, that I'm indestructible. If a potter's hand has made me, then let him smash his vessel as he pleases. But that which lives within must be unbegotten, must be divine of nature in its germ, sublime beyond all might and all art, and therefore invulnerable, eternal.
'We each have our mysteries, dear Hyperion! our more secret thoughts; these have been mine; ever since I could think.
'What lives is indestructible, stays free in its most deeply servile form, stays intact and though you split it to the ground, stays unscathed and though you shatter it to the marrow, and its essence flies away in triumph through your fingers. — But the morning wind stirs; our ships are awake. O my Hyperion! I have overcome; I've been able to bring myself to speak my heart's death sentence and separate you and me, darling of my life! be gentle with me now! spare me the parting pains! let's be quick! come!' —
A chill shot through all my bones as he uttered these words.
'Oh, by your troth, Alabanda!' I cried, prostrate before him, 'must it be, must it really be? You fuddled me unfairly, you carried me away and left me reeling. Brother! you took from me even the presence of mind to ask, "where are you going?"'
'I may not name the place, dear heart!' he replied; 'yet perhaps we shall meet again one day.'
'Meet again?' I replied; 'so that's one faith I'm the richer by! and so I'll become ever richer in faith and in the end there'll be nothing but faith for me.'
'Dearest!' he cried, 'where words are no help, let us be still! let us make a manful end! You ruin the last moments for yourself.'
Meanwhile we'd come closer to the harbour.
'One more thing!' he said, when we were at his ship. 'Greet your Diotima! Love each other! be happy, you beautiful souls!'
'O my Alabanda!' I cried, 'why can't I go in your stead?'
'Your calling is more beautiful,' he replied; 'stay with it! you belong to her, that lovely being is from now on your world — Oh! since there is no happiness without sacrifice, take me as sacrifice, O fate, and leave the loving their joy!' —
His heart began to overwhelm him, and he tore himself from me and leapt into the ship to shorten the parting for himself and me. I felt this moment like a thunderbolt that's followed by night and deathly stillness, but in the midst of the desolation my soul reared up to hold him, the dear departing, and my arms darted towards him of their own accord. 'Woe! Alabanda! Alabanda!' I cried, and heard from the ship a muffled 'Farewell'.
#### Hyperion to Bellarmin [LVIII]
It chanced that the vessel due to take me to Calauria delayed until the evening, Alabanda having already gone his way that morning.
I stayed on the shore, weary with the pains of parting, stilly gazing into the sea from one hour to another. My spirit told over the sorrowful days of slowly dwindling youth, and adrift like the beautiful dove, it hovered over the future. I wished to give myself strength, I took out my long-forgotten lute to sing myself a song of fate that I'd once, in blithely innocent youth, repeated after my Adamas.
> You wander high in the light On gentle ground, you blissful genii! Heavenly radiant airs Touching you soft, As the hand of the harpist her Hallowing strings.
Fatelessly, like the slumbering Suckling, breathe the celestials; Chaste preserved In the modest bud, Blossoms ever For them the spirit, And the blissful eyes Gaze in eternal Tranquil glory.
But to us it is given In no place to rest, They dwindle, they fall, The suffering mortals Blindling from one Hour to another, Like water from rock To rock hurled down, Year long into confusion below.
So I sang to the strings. I'd hardly finished when a boat came in and I immediately recognized my servant who brought me a letter from Diotima.
'So you're still on earth?' she wrote, 'and see the daylight still? I'd thought to find you elsewhere, my love! The letter you wrote before the battle of Chesma I received earlier than you afterwards wished, and so for a week I lived in the belief you'd thrown yourself into the arms of death, until your servant arrived with the joyful news that you're still alive. Besides, I'd heard a few days after the battle that the ship on which I knew you to be had blown up with all hands lost.
'But O sweet voice! once again I heard you, the language of the beloved touched me once more like a May breeze, and for a moment your beautiful hopeful joy, the lovely phantom of our future happiness deluded me too.
'Dear dreamer, why must I wake you? why can't I say, "come and make true the beautiful days you promised me!" But it's too late, Hyperion, it's too late. Your maiden has withered since you went away, a fire within has gently consumed me and there's only a small rest remaining. Don't be dismayed! All that is natural purges itself, and everywhere life's blossom winds itself freer and freer of coarser stuff.
'Dearest Hyperion! little did you think to hear this year my swansong.
#### Continued
'It began soon after you'd left, and even in the days before our parting. A force of spirit before which I took fright, an inner life before which the life of the earth paled and dwindled like lamps of the night in the glow of morning — shall I say it? I could have gone to Delphi and built the god of inspiration a temple beneath the cliffs of ancient Parnassus, and, a modern Pythia, kindled the flagging peoples with divine utterance, and my soul knows that my maidenly mouth would have opened the eyes of all the godforsaken and smoothed their dulled brows, so powerful was the spirit of life in me! Yet ever more weary grew my mortal limbs and dread weight pulled me relentlessly down. Oh! often in the stillness of my bower I have wept for the roses of youth! they withered and withered, and only from tears did your maiden's cheek redden. They were still the trees of old, it was the bower of old — there once stood your Diotima, your child, Hyperion, before your joyful eyes, a flower among flowers, and the powers of earth and heaven came peacefully together in her; now she walked a stranger among the buds of May, and her familiars, the lovely plants, nodded to her amiably, but she could only grieve; yet I passed none by, yet one after another I bade farewell to all the playmates of my youth, the groves and springs and whispering hills.
'Oh! as long as I still was able I would often make my sweet and weary way up to the heights where you'd stayed with Notara, and speak about you with our friend, as lightsomely as possible, so that he shouldn't write to you about me; but soon, when her heart grew too loud, the hypocrite stole out into the garden, and then there I was at the rail above the cliff where I once gazed down with you and out into open nature, oh! where I stood, held by your hands, caressed by your watchful eyes, in the first shuddering warming of love and wished to pour out my overflowing soul like sacrificial wine into life's abyss; there I now lurched around and cried my sorrow to the wind, and like a timid bird my glance flitted about and barely dared look at the beautiful earth from which I was to part.
#### Continued
'That's the way it's gone with your maiden, Hyperion. Don't ask how? don't explain this death to yourself! Whoever thinks to fathom such a fate will end by cursing himself and the world, and yet there's not a soul to blame.
'Shall I say it's grief for you that's killed me? oh no! oh no! it was welcome to me, this grief, it gave the death I bore within me form and grace; it's in honour of your darling that you're dying, I could now tell myself. —
'Or did my soul become too ripe for me in all the raptures of our love, and is that why, like a headstrong youth, it will be held no more in its humble home? tell me! was it my heart's luxuriance that estranged me from mortal life? is nature in me become through you, you glorious man! too proud to put up with it any longer on this mediocre star? But if you taught my soul to fly, why won't you also teach it to return to you? If you kindled the aether-loving fire, why wouldn't you tend it for me? — Hear me, my love! for the sake of your beautiful soul! don't you blame yourself for my death!
'Could you hold me back when your fate pointed you on the selfsame path? and if you had, in your heart's heroic struggle, preached to me — "be content, my child! and come to terms with the times" — would you not have been the vainest of all the vain?
#### Continued
'I will tell you plainly what I think. Your fire was alive in me, your spirit had passed into me; but that would scarce have harmed, and only your fate has made my new life deadly to me. Too mighty for me had my soul become through you, it would through you have also stilled again. You drew my life away from earth, you would have also had the power to fetter me to earth, you would have bound my soul as in a magic circle into your embracing arms; oh! just one of your heartfelt glances would have held me fast, just one love speech of yours would have turned me back into a happy wholesome child; yet when your own fate drove you into spiritual solitude, like floodwater onto a mountain peak, oh, only when I fully believed the storm of battle had burst open the dungeon and my Hyperion had flown up into ancient freedom, that was the turning point for me and soon it will be over.
'I've made many words, and yet the great Roman perished silently when her Brutus and the fatherland were writhing in their death throes. But what better could I do in the best of my last days of life? — And there are many things I still feel driven to say. Mute was my life; my death is eloquent. Enough!
#### Continued
'There's just one thing more I must say to you.
'You would have to go under, you'd have to despair, yet the spirit will save you. No laurel will comfort you, and no wreath of myrtle; Olympus will, the living and present, that eternally youthful blossoms around all your senses. The beautiful world is my Olympus; in this you will live, and with the holy beings of the world, the gods of nature, with them you will be joyful.
'Oh, be welcome, you good, you faithful! you deeply missed and misjudged! children and elders! sun and earth and aether with all the living souls that play about you, as you play about them, in eternal love! Oh, take all-venturing men, take back the fugitives into the family of gods, take them up into the home of nature from which they absconded! —
'You know this word, Hyperion! You began it in me. You will fulfil it in yourself, and only then rest.
'That's enough for me to die in joy as a Greek maiden.
'Those poor souls who know nothing but their own sorry handiwork, who are merely slaves of need and scorn genius, and who do not honour you, childlike life of nature! let them fear death. Their yoke is become their world; they know nothing better than their bondage; dread the freedom of the gods that death gives us?
'But I don't! I have gone beyond the patchwork made by hand of man, I have felt the life of nature that passes all thought — even if I became a plant, would the harm be so great? — I shall be. How should I be lost from the sphere of life where the love eternal, that is common to all, holds all natures together? How should I depart from the covenant that binds all beings? This is not so easily broken as the loose bonds of our age. It's not like a market day when folk flock together and make a great bustle and then disperse. No! by the spirit that unites us, by the god's spirit that is proper to each and common to all! no! no! in nature's covenant troth is no dream. We part to be only more inwardly one, more divinely at peace with all, with ourselves. We die in order to live.
'I shall be; I do not ask what I shall become. To be, to live, that is enough, that is the glory of the gods; and that is why all that is life is equal in the divine world, and there are there no masters and menials. Natures live together like lovers; they have all in common, spirit, joy and eternal youth.
'Permanence is the choice of the stars, in silent fullness of life they constantly wheel, knowing no age. We embody perfection in change, splitting the great chords of joy into wandering melodies. Like harpers around the thrones of the elders, we live, ourselves divine, around the silent gods of the world, with life's fleeting song we soften the blissful earnest of the sun-god and the others.
'Look up into the world! is it not like a wandering triumphal pageant where nature celebrates its eternal victory over all corruption? and does not life for its greater glory take with it death in golden chains, as once the general would take with him captive kings? and we, we're like the virgins and youths who with dance and song, in shifting shapes and sounds, accompany the majestic procession.
'Now let me be still. To say more would be too much. We shall likely meet again. —
'Grieving youth! soon, soon you will be happier. Your laurel failed to ripen and your myrtles faded, for priest you shall be of divine nature, and your poetic days burgeon already. Oh, could I only see you in your future beauty! Farewell.'
At the same time I received a letter from Notara in which he wrote:
'The day after she'd written to you for the last time she grew quite still, spoke only a few more words, and then said that she would rather part in fire from the earth than be buried, and we should gather her ashes in an urn and set it in the woods at the spot where you, my dear friend! had first met her. Soon after, when it began to grow dark, she said goodnight to us as if she wished to sleep, and wrapped her arms around her lovely head; until toward morning we heard her breathing. As it then became wholly still and I heard nothing more, I went over to her and listened.
'O Hyperion! what more shall I say? It was finished and our plaints awakened her no more.
'It is a terrible mystery that such a life should die, and I have to confess to you that I myself have neither sense nor faith since I witnessed this.
'Yet better always a beautiful death, Hyperion! than such a torpid life as ours now is.
'Warding off flies, that's our work in the future, and gnawing on the things of the world like children on dried orris root, that's our joy in the end. Growing old amongst youthful peoples seems to me a delight, but growing old where all is old seems to me worse than anything. —
'I'm tempted to advise you, my Hyperion! not to come here. I know you. It would rob you of your senses. Besides, you're not safe here. My dear friend! think of Diotima's mother, think of me and spare yourself!
'I have to confess, I shudder when I ponder your fate. But then I also believe, it's not the deeper springs that searing summer dries up, only the shallow rain-stream. I've seen you at moments, Hyperion! when you seemed to me a higher being. Now you're put to the test, and that's going to show who you are. Farewell.'
So Notara wrote; and you ask, my Bellarmin! how I feel now, telling you this?
My dearest friend! I'm at peace, for I want no better than the gods. Must not everything suffer? And the nobler it is, the more deeply. Does not holy nature suffer? O my godhead! that you could grieve as you are blissful is something I long failed to grasp. But the bliss that does not suffer is sleep, and without death there is no life. Should you be, like a child, eternal and slumber as nothingness? forego the victory? not pass through all the perfections? Yes! yes! worthy is pain to lie at men's hearts and be your familiar, O nature! For it alone leads from one bliss to the next, and there can be no other companion. —
Once I'd begun to revive again, I wrote to Notara from Sicily where I'd first been brought by ship from Paros:
'I've obeyed you, dear friend! I'm already far away from you all, and want to give you my news now; but words come hard to me; you won't mind my confessing that. The blessed, among whom Diotima is now, do not speak much; in my night, in the depths of the grieving, speech too is at an end.
'My Diotima died a beautiful death; you're right; and it's this that rouses me and gives me back my soul.
'But it's no longer the same world I return to. I'm a stranger, like the unburied when they come up from Acheron, and were I even on my home island, in the gardens of my youth that are barred to me by my father, oh! still, still I should be a stranger in the earth and no god can any longer link me to the past.
'Yes! it's all over. I just must tell myself that time and again, have the thought bind my soul so that it stay calm and not become inflamed through senseless childish efforts.
'It's all over; and even could I weep, beautiful divinity, as once you wept for Adonis, my Diotima will still not return to me and my heart's word has lost its power, for the winds alone hear me.
'Oh, God! and that I myself am nothing, and the meanest workman can say he's achieved more than I! that they may comfort themselves, the poor in spirit, and smile and chide me as a dreamer because my deeds failed to ripen, because my arms are not free, because my times are like the raging Procrustes who cast the men he captured into a child's cradle and, so that they'd fit into the little bed, would hack off their limbs.
'If only it weren't too utterly dreary to fling oneself into the foolish mob and be torn to pieces by it! or if only noble blood could feel no shame at mingling with the blood of slaves! oh, if there were a banner, gods! that my Alabanda might serve under, a Thermopylae where I could bleed it out with honour, all the lonely love I'll never more need! To be sure it would be better if I could live, live, and in the new temples, in the newly assembled agora of our people, still the great grief with great joy; but on that I shall be silent, for I'll merely weep my strength wholly away if I think upon it all.
'Oh, Notara! with me too it's over; my own soul has grown hateful to me because I must blame it for Diotima's being dead, and the ideas of my youth I deemed so grand mean nothing to me any more. They poisoned my Diotima, didn't they!
'And now tell me, where is there still a refuge? — Yesterday I was up on Etna. There came to my mind the great Sicilian who once, wearied with counting the hours and intimate with the soul of the world, plunged himself down in his bold lust for life, into the glorious flames; for the frigid poet had needed the fire to warm himself by, as someone mockingly said of him.
'Oh, how gladly would I have drawn such mockery upon myself! but one must think more highly of oneself than I do to fly thus uncalled into the heart of nature, or whatever else you like to call it, for truly! the way I am now, I have no name for things and all is confusion.
'Notara! and now tell me, where is there still a refuge?
'In the woods of Calauria? — Yes! in the verdant darkness there, where stand our trees, confidants of our love, where like the red glow of sunset their dying leaves fall upon Diotima's urn, and their lovely heads bend above Diotima's urn, gently growing old until they too sink down over the beloved ashes, — there, there I might dwell as I would wish!
'But you counsel me to stay away, think I'm not safe in Calauria, and it may be so.
'I know full well you'll point me to Alabanda. But just hear this! he's destroyed! even that firm and limber stem is blasted, and the boys will pick up the splinters to make themselves a jolly fire. He's gone; he's got certain good friends who will ease things for him, who have quite special skills in relieving anyone on whom life lies somewhat heavy; he's paying them a visit, and why? because there's nothing else for him to do, or, if you would know it all, because there's a passion gnawing at his heart, and do you know for whom? for Diotima; he thinks she's still alive, wedded to me and happy — poor Alabanda! and now she belongs to you and me both!
'He headed out to the east, and as for me, I'm taking ship northwestward, since chance will have it so. —
'And now farewell to you all! all you dear ones who've lain close to my heart, friends of my youth and you parents and all you dear Greeks, you who suffer!
'You winds that nourished me in tender childhood, and you dark laurel woods and you cliffs of the coast and you majestic waters that taught my spirit to sense great things — and oh! you monuments of mourning, where my melancholy began, you holy walls that girdled heroes' cities, and you ancient gates that many a beautiful wanderer passed through, you temple pillars and you rubble of the gods! and you, O Diotima! and you valleys of my love, and you streams that once witnessed the blissful figure, you trees where she would gladden, you springtimes where she the lovely lived, together with her flowers, don't, don't part from me! yet if it must be, you sweet memories! then you too fade away and leave me, for man can change nothing and the light of life comes and goes as it will.'
#### Hyperion to Bellarmin [LIX]
So I came among the Germans. I didn't ask for much and was prepared to find even less. Humbly I came like homeless blind Oedipus to the gates of Athens, where the grove of the gods welcomed him and he was met by beautiful souls —
How differently I fared!
Barbarians of old, made yet more barbarous by industry and learning and even religion, deeply incapable of any godly feeling, spoiled to the core for the joys of the holy graces, in every degree of profligacy and pettiness an offence to every decent soul, dull and full of discord like the shards of a discarded vessel — these, my Bellarmin! were my comforters.
They're harsh words, but I'll say them nonetheless, because it's the truth: I can imagine no people more fragmented than the Germans.
You see craftsmen, but no human beings, philosophers, but no human beings, priests, but no human beings, masters and menials, youths and elders, but no human beings — is this not like a battlefield where hands and arms and all the limbs lie about in pieces while the spilled life-blood seeps away into the sand?
To each his own trade, you will say, and so do I. Only he must pursue it with his whole soul, must not choke every energy within him that doesn't quite befit his title, must not with such niggardly timidity pretend to be to the letter only what he's called; with earnest, with love he must be what he is, then there lives a spirit in all he does, and if he's boxed into a specialism in which the spirit cannot thrive, then let him thrust it from him with contempt and learn to use a plough! But your Germans like to stick with what's most needful, and that's why there's so much bungled work amongst them, so little that is free and genuinely pleasing. But this one could put up with if only such beings didn't have to be feelingless for all beautiful life, if only the curse of godforsaken unnature didn't everywhere rest upon such a people. —
The virtues of the ancients were but glittering vices, some wicked tongue — I can't tell you which — once said; and yet even their vices are virtues, for then still lived a childlike, a beautiful spirit, and nothing they did was done without soul. But the virtues of the Germans are a glittering evil and nothing more; for they are only works of necessity, in craven fear wrung from the wasted heart with slavish effort, and leave without solace any pure soul that would nourish itself on the beautiful, oh! that, spoiled by the holy harmony in nobler natures, cannot thole the discord screeching in all the dead orderliness of such beings.
I tell you: there's nothing sacred that's not profaned, that's not debased to miserable makeshift with these people, and what is even amongst savages mostly kept divinely pure, these all-calculating barbarians approach as one might go about a trade, and they can do no other, for once a human being has been drilled, it serves its ends, it seeks its profit, it dreams no more, God forbid! it remains staid, and when it celebrates and when it loves and when it prays and even when the lovely festival of spring, when the world's season of reconciliation dissolves all care and conjures innocence into a guilty heart, when drunk with the sun's warm rays the slave joyfully forgets his chains, and softened by the god-inspired breath the foes of man are full of peace, like children — when even the caterpillar takes wing and the tipsy bee swarms, still will the German stick inside his box and not be bothered much about the weather!
But you shall judge, holy nature! For if only they were modest, these people, didn't make themselves the measure for the better amongst them! if only they'd not belittle what they're not, and yet, if belittle they must, if only they didn't mock the divine! —
Or isn't divine what you mock and call soulless? Isn't better than your babble the air that you drink? the rays of the sun, aren't they nobler than all of you in your cleverness? the earth's springs and the morning dew refresh your groves; can you do that? Oh! you can kill, but you can't bring to life if love doesn't do it, and that's not from you, you didn't invent it. You fret and you brood about how to flee fate, and are flummoxed when your childish arts prove bootless; meanwhile the stars wander blithely above. You debase it, you tear it apart, where it puts up with you, tolerant nature, and yet it endures in unending youth, its autumn and spring you can't drive away, its aether is something you won't defile.
Oh, it must be divine because you're suffered to destroy and yet it doesn't age and in your spite the beautiful stays beautiful! —
And it's heart-rending to see your poets, your artists, and all who still honour the genius, who love the beautiful and nurture it. The good souls! They live in the world like strangers in their own house, they are just like long-suffering Ulysses when he sat at his door in beggar's garb while the shameless suitors roistered in the hall and asked: 'Who has brought us this vagabond?'
Full of love and spirit and hope its youthful sons of the muses grow towards maturity among the German people; you see them seven years later, and they wander like shades, silent and cold, they're like soil the enemy has sown with salt so that it shall never more bring forth a blade of grass; and when they speak, woe unto him that understands them! who can see in the storming titanic energy, as in their protean skills, the despairing battle their troubled beautiful spirit must fight against the barbarians with which it has to deal.
'All things on earth are imperfect,' that's the eternal refrain of the Germans. If only someone would tell these godforsaken people that with them all things are only so imperfect because they leave nothing pure unsullied, nothing sacred unfumbled by their clumsy hands, that with them nothing thrives because they will not honour the root of all thriving, divine nature, that with them life is precisely so vapid and fretful and brimming with cold dumb discord because they scorn the genius that brings energy and nobility to human activity, serenity to suffering, and love and brotherhood to cities and homes.
And that too is the reason why they're so afraid of death, and for the sake of their oyster-like existence suffer all the shame, because they know of nothing higher than the handiwork they've cobbled for themselves.
O Bellarmin! where a people loves the beautiful, where it honours the genius in its artists, there blows like breath of life a universal spirit, the bashful mind unfolds, self-conceit melts away, and all hearts are pious and great, and inspiration brings forth heroes. The homeland of all men is with such a people and there the stranger loves to linger. But where divine nature and its artists are so abused, oh! there life's greatest joy is gone, and any other star is better than the earth. There men, though born beautiful, become ever more wild and wasted; a servile disposition grows and with it rude courage; intemperance grows with cares, and with opulence hunger and fear of famine; each year's blessing becomes a curse and all the gods take flight.
And woe to the stranger who wanders out of love and comes to such a people, and thrice woe to him who, driven like me by great pain, a beggar of my kind, comes to such a people! —
Enough! you know me, you won't take it amiss, Bellarmin! I spoke in your name too, I spoke for all who live in this land and suffer as I suffered there.
#### Hyperion to Bellarmin [LX]
I wanted to get away from Germany now. I looked for nothing more amongst these people, I'd been hurt enough by merciless affronts, and didn't want my soul to wholly bleed to death amongst such men. But the heavenly spring held me back; it was the only joy that remained to me, it was my last love, how could I still think of other things and leave the land where spring was too?
Bellarmin! I'd never so fully felt the old and firm and fateful saying that for the heart a new bliss rises when it holds out and endures through the midnight of grief, and that, like nightingale's singing in darkness, it's in deep suffering that first divinely sounds for us the lifesong of the world. For now I lived with the blossoming trees as with genii, and the limpid streams flowing beneath them whispered, like voices of gods, the sorrow from my breast. And so it fared with me everywhere, dear friend! — when I rested in the grass and tender life greened all about me, when I climbed the warm hill where the rose grew wild around the stony path, and when I boated along the breezy banks of the river and round all the islands it tenderly hugs.
And when on many a morning, like the sick to the healing spring, I'd climb to the mountain's summit through sleeping flowers, whilst to my side, sated with sweet slumber, the dear birds flew from the bush, in the half-light reeling and greedy for day, and the quickening breeze bore up already the prayers of the valleys, the voices of the herd, and the tones of the morning bells, and now the lofty light, divinely serene, came along its habitual path, enchanting the earth with immortal life, so that her heart warmed and all her children felt themselves once more — oh, like the moon, that still lingered in heaven to share in the joy of the day, I too then stood lonely above the plains, weeping tears of love down to the shores and the glancing waters, and long couldn't turn away my gaze.
Or of an evening when I'd wander far into the valley, to the cradle of the spring where around me rushed the dark crowns of the oaks, and nature entombed me in her peace like a dying saint; when the earth was now a shadow and viewless life whispered through the branches, through the treetops, and above the treetops the evening cloud stood still, a shining mountain range from which the rays of heaven flowed down to me like streams to quench the wanderer's thirst —
'O sun, O you breezes,' I'd then call out, 'with you alone my heart still lives, as amongst brothers!'
Thus I gave myself up more and more to blissful nature, and almost too endlessly. How gladly I'd have become a child to be closer to her! how gladly I'd have known less and become like the pure ray of light to be closer to her! Oh, to feel myself for a moment in her peace, her beauty, how much more that meant to me than years full of thinking, than all the ventures of all-venturing men! What I'd learned, what I'd done in my life melted away like ice, and all youth's designs died away in me; and O you dear ones, who are distant, you dead and you living, how wholly one we were!
Once I sat far away in a field, by a burn, in the shade of ivy-green rocks and overhanging flowering bushes. It was the most beautiful noon I've known. Sweet breezes wafted and in morning freshness the land still shone and serene in its native aether smiled the light. The people had gone away to rest at the homely table from their toil; my love was alone with the spring, and there was an unfathomable longing in me. 'Diotima,' I cried, 'where are you? oh, where are you?' And it seemed to me I heard Diotima's voice, the voice that used to gladden me in the days of joy —
'I'm with my kindred,' she called, 'with your kindred, those the erring human spirit does not know!'
A gentle terror seized me and all thought passed away in me.
'O dear words from holy mouth,' I cried, once again aroused, 'dear riddle, can I grasp you?'
And one last time I looked back into the cold night of men and shuddered and wept for joy that I was so blissful, and words I spoke, it seems to me, but they were like the fire's rush when up it flares and leaves behind the ashes —
"O you," so I thought, "with your gods, nature! I've dreamed it out, the dream of human things, and say that only you live, and all that peaceless men have wrought and thought, it melts away like pearls of wax before your flames!
"How long have they done without you? oh, how long have their rabble abused you, called you vulgar, you and your gods, the quick, the blissfully serene!
"Men fall from you like rotting fruit, oh, let them perish, for then they return to your root, and I, O tree of life, may I green again with you and breathe around your crown with all your budding branches! peacefully and deeply, for we've all of us grown high from that seed of golden grain!
"You springs of the earth! you flowers! and you forests and you eagles and you fraternal light! how old and new is our love! — We are free, we don't anxiously strive to be outwardly equal; how should not vary the modes of life? but we all love the aether, and deep in our innermost being we are all of us like.
"We too, we too are not parted, Diotima! and the tears for you haven't grasped it. We are living tones, we sound together in your harmony, nature! who can split that asunder? who will part lovers? —
"O soul! soul! beauty of the world! indestructible! rapturing! with your eternal youth! you are; what then is death and all the woe of men? — Oh! many empty words have been made by these strange brethren. Yet all things proceed from pleasure, and everything ends in peace.
"The dissonances of the world are like lovers' tiffs. There's reconciliation in the middle of strife, and all that's apart comes together again.
"The arteries part and return in the heart and one eternal glowing life is All."
So I thought. More anon.
On 3 August 1942, W. H. Auden wrote to Norman Holmes Pearson: 'My hostess [Caroline Newton] is translating Hölderlin's Hyperion, on which I hope one day to write an essay. It is the most perfect exposition of romanticism I have ever read.'1 It seems unlikely that Newton's translation itself came to much, the only surviving evidence being a gathering of some twenty-four leaves in a box in Princeton University Library.2 Nor does Auden's essay ever seem to have materialized. But the fact that he wanted to write it at all, together with his judgment of the novel, may be considered noteworthy, particularly given the relatively early date of the letter. For even in Germany it was not until the opening years of the twentieth century, and particularly with the publication of Hellingrath's edition on the eve of the First World War, that Friedrich Hölderlin (1770–1843) suddenly emerged from virtual obscurity as a poet of the first rank, at least amongst writers and intellectuals. But Auden's generation was not slow to pick up on the excitement of the discovery. It was the Orcadian Edwin Muir, himself a fine poet, who was really the first to mediate Hölderlin to the anglophone literary world in a series of brilliant essays in the 1920s and 1930s.3 Another significant figure who was reading the Swabian poet intensively in the late 1930s is Samuel Beckett, and indeed it is in *Hyperion* that most of the marginalia in his Hölderlin edition are to be found.4 However, it would
<sup>1</sup> *The Complete Works of W.H. Auden*: *Prose*: *Volume II*, *1939–1948*, ed. Edward Mendelson (Princeton: Princeton University Press, 2002), p. 506.
<sup>2</sup> My thanks to Stephen Ferguson, Curator of Rare Books, for this information.
<sup>3</sup> See Gaskill, 'Edwin Muir as a Critic of Hölderlin', *Forum for Modern Language Studies*, 14 (1978), 345–64.
<sup>4</sup> Dirk Van Hulle and Mark Dixon, *Samuel Beckett's Library* (Cambridge: Cambridge University Press, 2013), pp. 91–93.
generally be true to say that for most of those displaying an interest in Hölderlin — and he was rapidly becoming the poets' poet — it was the later 'hymns' and fragments, written in the years immediately preceding the collapse of his mind in 1806, and the fact of the madness itself, that proved to be the main source of fascination.5 The novel, which alone had just about kept Hölderlin's name alive in the nineteenth century, tended now to be relegated to the background. In the English-speaking world, the absence of a complete translation certainly did nothing to help.6
Even in specialist scholarship *Hyperion* was relatively late in receiving due critical attention. It would be fair to say that it was not until 1965, with the publication of Lawrence Ryan's monograph, that the modern reassessment of the status of the novel really began, and its significance within Hölderlin's oeuvre came to be properly recognized.7 Ryan's incisive examination of the narrative structure of *Hyperion* and the implications for interpretation proved to be an eye-opener for many, myself included. The voluminous secondary literature may since have variously modified, extensively revised, or even rejected his main argument, but, like it or not, we all remain in his debt.8 The distinction between Hyperion as narrator and Hyperion as a figure in his own narrative, between the hermit and the would-be hero/lover, cannot be ignored, whatever one chooses to make of it. Whether the development of the narrator, if such it is, may be seen to lead to any kind of resolution of the dissonances in his character, even a fragile or provisional one, is what now tends to split the critics.9 I freely admit that I belong to the
<sup>5</sup> As instanced by Muir's own poem 'Hölderlin's Journey' (1937); also David Gascoyne's collection *Hölderlin's Madness* (London: Dent & Sons, 1938).
<sup>6</sup> The first rendering of any part of *Hyperion* into English appears to have been made, possibly with help, by none other than Ralph Waldo Emerson. It is a translation of a substantial excerpt from the novel's penultimate letter (the scathing of the Germans) quoted in Theodor Mundt's *Geschichte der Literatur der Gegenwart* (1842), pp. 86–88 — see *The Dial*: *A Magazine for Literature*, *Philosophy*, *and Religion*, 4, 1843, 'A Letter', 262–70, pp. 265–66. For this discovery and (slightly puzzled) speculation as to Emerson's intentions, see Cyrus Hamlin, 'Transplanting German Idealism to American Culture', in *Translating Literatures Translating Cultures*: *New Vistas and Approaches in Literary Studies*, ed. by Kurt Mueller-Vollmer and Michael Irmscher (Berlin: Erich Schmidt, 1998), 107–24, pp. 108–11.
<sup>7</sup> Lawrence Ryan, *Hölderlins Hyperion*: *Exzentrische Bahn und Dichterberuf* (Stuttgart: Metzler, 1965).
<sup>8</sup> For an excellent history of critical reception, see Marco Castellari, *Friedrich Hölderlin*: *Hyperion nello specchio della critica* (Milan: C.U.E.M., 2002).
<sup>9</sup> For an emphatic rejection of any notion of a successfully embodied telos, see Hansjörg Bay, *Ohne Rückkehr*: *Utopische Intention und poetischer Prozess in Hölderlins Hyperion* (Munich: Wilhelm Fink Verlag, 2003).
'harmonizing' tendency. Seeing the novel primarily in terms of aporias and unreconciled contradictions goes against my aesthetic experience of the text as a thing of great brilliance and beauty. It is this experience that I am attempting to mediate. I would like to think that an anglophone readership might learn to love *Hyperion*.
#### A Novel in Letters10
*Hyperion* was published in two volumes, the first appearing in spring 1797, the second in autumn 1799. It is unlikely that Hölderlin, rather than Cotta, his publisher, was responsible for the two-and-a-half year gap. As may be seen from the conclusion of the Preface, the author seems none too happy that readers of the first volume will, for the time being at least, not be in any position to judge the novel's design. Similar misgivings were expressed to Schiller, who had mediated publication, in a letter of 20 June 1797. And it seems that they were justified, in that Cotta found difficulty in shifting copies of the second volume when it eventually did appear.
Hölderlin's one and only novel was a long time in gestation, the initial idea going back as far as 1792. The final version adopts the epistolary form, as does the only other version available to Hölderlin's contemporaries, the 'Fragment von Hyperion' which appeared in Schiller's *Thalia* in 1794.11 In the intervening period Hölderlin experimented with a variety of forms, including verse, as can be seen from the surviving manuscripts. Common to all versions, however, is that they confront us with a firstperson narrator (or alternatively an 'I' within the narrative), telling the story of his past life and commenting on it in the process. This indicates that the retrospective review was regarded as the most important element by Hölderlin, and that it must therefore have been instrumental in determining his final choice of narrative form. That choice remains, nevertheless, an unusual one.12
<sup>10</sup> In the following I make liberal use of my short monograph on the novel, *Hölderlin's Hyperion* (Durham: Durham Modern Language Studies, 1984). I am grateful to Michael Thomson and Sam Bootle, editor and former editor of the DMLS series, for permission to do so.
<sup>11</sup> *Neue Thalia*, 4 (1793), 181–221. The issue did not in fact appear until November 1794.
<sup>12</sup> The most informed and informative study of *Hyperion* as an epistolary novel is Gideon Stiening's *Epistolare Subjektivität*: *Das Erzählsystem in Friedrich Hölderlins Briefroman Hyperion oder der Eremit in Griechenland* (Tübingen: Niemeyer, 2005).
There were of course many precedents for the adoption of the epistolary convention in prose fiction (as well as non-fictional discourse). In the second half of the eighteenth century most novels with any pretension to artistic merit (and many without it) tended to be either epistolary or self-consciously comic. Common to both forms is a degree of preoccupation with the act of writing itself. It is straight third-person narrative, without irony or authorial intrusion, that was the exception. Thus it is not Hölderlin's choice of the epistolary form as such that may seem surprising, but the particular use he makes of it. In the great majority of epistolary novels there is correspondence between a number of characters, a medley of voices. It is not all one way. There were, it is true, examples of one-sided correspondence in works of philosophy or literary criticism. But in fiction — at least after Samuel Richardson's successful exploitation of polyperspectivism in *Clarissa* (1748) — such limitation seemed to have little to commend it. Goethe's enormously successful novel of 1774, *Die Leiden des jungen Werthers* (*The Sorrows of Young Werther*), in which we see only the protagonist's letters and can merely infer the content of any replies from his single correspondent, may thus be said to represent a departure from prevailing conventions. This also makes it an obvious point of reference for any German writer who subsequently adopts the epistolary form in an apparently similar manner. It seems pointless to deny that *Werther* exercised any significant influence on *Hyperion*. There are numerous verbal echoes and thematic correspondences, quite apart from the similarity of form. For *Hyperion*, too, consists of letters in one direction. (The letters from Diotima, Notara, and Alabanda are copied out by Hyperion and so form part of his correspondence with Bellarmin.) However, what certainly does distinguish the two novels from each other is that, whereas Werther's letters are written on a day-to-day basis, relating things largely from the immediacy of the present or near-present, Hyperion's are used to recount his past experiences, the related events all having taken place before the first letter is written.
Naturally, in neither case do the letters consist solely of narrative. What fascinates us in Goethe's novel is Werther's gradual disintegration, his slide towards mental and emotional catastrophe. His 'affliction' reveals itself in his reaction to events, but equally in his reflections on nature and the world around him. There are good grounds for seeing the attempted articulation of his problems as an important contributory factor in the hero's downfall. Writing itself affects Werther's frame of mind and intensifies his difficulties, since it encourages him to finger his emotional wounds both old and new and wallow in his distress. It is not therapeutic or cathartic; it leads him further into the labyrinth, instead of enabling him to transcend and transmute his experiences by lending them — through articulation — mediated, reflected coherence. Whether it is because he lacks the necessary detachment or innate artistic ability, or both, the effect is to accelerate the fragmentation of Werther's world. It seems to me that, seen in this light, *Hyperion* does indeed have much in common with *Werther*. In Hölderlin's novel the past experience is further in the past, and this has a bearing on the narrator's ability to confront it and articulate it. But the experience itself is just as harrowing as Werther's (a good deal more so in fact), and in attempting to relive it, give it meaning, and commit it to paper, Hyperion is exposing himself to real danger. The narrator's present is one in which he experiences extreme oscillations of mood and undergoes a series of violent emotional shocks. At times he is very near to madness, and it is important to see this and take the possibility seriously. The narrator may know the outcome of the events within the narrative, but he is very far from knowing the outcome and consequences of the narrating activity itself, with all its associated inner turmoil. His anxiety is expressed clearly enough in various key passages which we have little excuse for overlooking or misinterpreting. For instance, when Hyperion finally summons up the courage to begin the account of his meeting and blossoming relationship with Diotima, he tells Bellarmin that he has hitherto kept the image of his love sacred and carried it within him like a holy relic: 'and if fate henceforth should seize and plunge me down from abyss to abyss and drown in me all energy and all reason, yet shall this one and only outlive myself in me and shine in me and reign in eternal, indestructible glory!' (p. 44).13 (The language he uses here recalls and pre-echoes that of the 'Song of Fate'.) Before confronting this most crucial episode in his own past, the narrating Hyperion anticipates
<sup>13</sup> Letter XIII: 'und wenn hinfort mich das Schiksaal ergreift und von einem Abgrund in den andern mich wirft, und alle Kräfte ertränkt in mir und alle Gedanken, so soll diß Einzige doch mich selber überleben in mir, und leuchten in mir und herrschen, in ewiger, unzerstörbarer Klarheit!' (StA [= *Große Stuttgarter Ausgabe*] III, 51).
what might happen to him if he persists. He is badly scarred from his experiences, and realizes that reopening the old wounds, by evoking both the ecstasies and agonies, could well destroy him for good. He has been holding them back, and when at last the pressure is released, they virtually erupt into his narrative and threaten to run completely out of his tenuous control. Already at this early stage the implication would seem to be that his evolving monument to the memory of Diotima could cost him his sanity. This emerges more clearly in a later passage, which demonstrates the extreme difficulty the narrator finds in sufficiently detaching himself from his experiences to be able to write coherently about them:
All that I can speak of her is scattered words. I must forget what she is whole if I'm to speak of her. I must make believe she lived in times of old, that I knew of her through tales, if her living image isn't so to seize me that I perish in rapture and pain, if I'm not to die of my joy in her and my grief for her. (p. 51) 14
Having once conjured up memories of such intensity, Hyperion must attain rational control of them, if they are not to overwhelm him. This is not idle hyperbole. To be able to write here, certainly as far as his mental and emotional stability is concerned, is a matter of life and death. But the effort almost breaks him. The letter that begins with Hyperion's proclaimed intention of proving his breast on the joys of the past till it becomes like steel, inuring himself to their 'deadly delights', the same letter that climaxes in the embrace, ends with the narrator so severely shaken that he cannot continue.15 In the following letter an apparently calmer Hyperion resumes his account of the development of the relationship, but then any semblance of smooth narrative progression is suddenly fractured by a remark which demonstrates just how vulnerable and threatened he feels: 'I ought to keep silent, ought to forget and keep silent. // But the alluring flame will tempt me till I
<sup>14</sup> Letter XXII: 'Ich kann nur hie und da ein Wörtchen von ihr sprechen. Ich muß vergessen, was sie ganz ist, wenn ich von ihr sprechen soll. Ich muß mich täuschen, als hätte sie vor alten Zeiten gelebt, als wüßt' ich durch Erzählung einiges von ihr, wenn ihr lebendig Bild mich nicht ergreiffen soll, daß ich vergehe im Entzüken und im Schmerz, wenn ich den Tod der Freude über sie und den Tod der Trauer um sie nicht sterben soll' (StA III, 59).
<sup>15</sup> Letter XXVIII (pp. 59–63): StA III, 69–74.
plunge into it whole and perish like the fly' (p. 64).16 This image of death and disaster is clearly and unambiguously related to the activity of Hyperion as narrator, and the fate that awaits him if he continues. But he does go on, as he must, and what follows is his narration of Diotima's confession, her admission of the way in which her love for Hyperion has fractured the harmony with herself and the world, and brought about a distressing detachment from her earthly environment, from nature. This is in fact the beginning of a process which will eventually kill her, and the narrating Hyperion knows this. It is small wonder, then, that when he evokes the image of Diotima, having made her confession, embracing him and resting her head on his breast, it should prove too much for him: 'O Bellarmin! my senses fail and my mind runs adrift. // I see, I see how this must end. The rudder has fallen into the surge and the ship will be seized like a child by the feet and hurled against the rocks' (p. 65).17 What has happened here — and it strikes me as the only plausible interpretation — is that the *narrating* Hyperion feels he has gone too far. It is quite simply a crude (if not uncommon) misreading of the text to take him as referring in this passage to future events in his own past — as if he were momentarily projecting himself back into the perspective of his former self and expressing the presentiment that the love relationship is doomed and will end in tears. He has no need to tell himself this, or Bellarmin (or indeed the reader). Hyperion knows full well how it ended. All that he has been writing since making his home on Salamis — and since the twenty-third letter we know it too — has been written in the shadow of that 'sweet bewildering lethal dread' (p. 52), the knowledge that Diotima's grave is near.18 It is knowledge he has tried in vain to repress, the (relative) proximity of the island where she lived and died exercising a powerful and potentially fatal attraction, in the same way as he has resisted giving way to the deadly bliss of his memories. All to no avail, as we see. And if we look at the drastic image
<sup>16</sup> Letter XXIX: 'Ich sollte schweigen, sollte vergessen und schweigen. // Aber die reizende Flamme versucht mich, bis ich mich ganz in sie stürze, und, wie die Fliege, vergehe' (StA III, 75).
<sup>17</sup> End of Letter XXIX: 'o Bellarmin! die Sinne vergehn mir und der Geist entflieht. // Ich seh', ich sehe, wie das enden muß. Das Steuer ist in die Wooge gefallen und das Schiff wird, wie an den Füßen ein Kind, ergriffen und an die Felsen geschleudert' (StA III, 76).
<sup>18</sup> Letter XXIII: 'die süßen verwirrenden tödtenden Schreken, daß Diotima's Grab mir nah ist' (StA III, 60).
he employs at the end of letter XXIX, it is one of impending catastrophe brought about *by loss of control*: the vessel dashed against the rocks is surely Hyperion, the rudder his drowning reason. This passage marks the worst crisis in the narrator's development. He does eventually recover his control and finally, so it seems, come to terms with his experiences through the act of writing itself. But the issue is not, for the narrating Hyperion, predetermined, and he very nearly goes under in the process. We are meant to take this possibility seriously.
From the above it should be clear that we are dealing in *Hyperion* with a form of suspense that is by no means primarily tied to the events related in the narrative. Indeed, it is because the novel has too often been read as if it were the narrator's 'story' and nothing more, that it has been so badly underrated. However, once we have been alerted to the function and significance of the many comments and interpolations that punctuate the narrative, we begin to appreciate that any loss of conventional suspense — after all, Hyperion must have survived to be able to tell the tale of his (mis)adventures — is more than compensated for by the tension introduced between the narrator and his subject matter. It is a tension which in this degree would not be possible without the appreciable temporal distance between the narrated events and the narrator's present. Simplifying, one could say that in the first half of the novel we have a gloomy or even despairing narrator relating largely happy to blissful experiences from his own past; in the second half of the novel we have an increasingly calm, almost serene narrator telling of grief and disaster.
Contrary to the impression sometimes given by critics, the procedure adopted by Hölderlin in *Hyperion*, the use of the epistolary convention to narrate a life-story, is not unique to this novel. There are precedents, some of them respectable, others less so. Hölderlin could be seen here as resurrecting a form that was not in fact uncommon in the early stages of the development of the epistolary novel. Of these 'autobiographies' in letters perhaps one of the most celebrated is Marivaux' unfinished *La vie de Marianne*, *ou Les avantures de madame la comtesse de \*\*\** (1731–42). The most notorious is probably Cleland's *Memoirs of a Woman of Pleasure* of 1748, popularly known as *Fanny Hill*. One might well ask what reasons a writer might have for cutting up a past story into letters, instead of telling it as continuous narrative in the first person. One answer would be that the temporal detachment of the narrator from the main events of his or her narrative offers opportunity for subtle interplay between the perspectives of present and former self, accommodating side by side experience, reaction, feeling on two distinct time levels. And then there is the often extremely sensitive and intimate nature of the subject matter. It would tend to undermine the moral credibility of the narrator (one thinks here particularly of Fanny Hill) if he or she were seen to be voluntarily exposing all to a curious world. Even the built-in distancing comments and mature critical judgments might not be an adequate defence against the charge of shameless exhibitionism. Not only must the confessions ostensibly be mediated in a private manner, as secrets entrusted to one who is worthy to receive them; but also, the narrator's reluctance to speak at all must be credibly documented. This is a fiction which it would be difficult to maintain if the narrator were seen to have a vision of his or her life in hardback from the outset. It is the function of the correspondent (about whom we may know nothing except the name, and perhaps not even that) to press the central figure into writing about past experiences which may be painful and embarrassing. The narrator has good reason for being reluctant to rake over the coals, and will need to be coaxed.
We may know virtually nothing about Hyperion's Bellarmin except that he appears to be German, lives in Germany, and is presumably a youngish man of enlightened liberal persuasion, something that would put him very much in a minority amongst his compatriots. But the most important thing about him is that he is there at all, pestering his correspondent to tell of his past life: 'I thank you for asking me to tell you about myself, for making me remember former times' (p. 9).19 It is important to remember that Hyperion the narrator begins his tale as a hermit (hence the subtitle of the novel), his self-imposed isolation largely the result of disgust with the poverty of spirit of his fellows. For that reason alone any ostensibly public form of communication (one in which, if anyone is addressed, it must be the general reader), can be excluded as a possibility, at least initially. The narrator starts as a self-confessed elitist who would address himself, if at all, only to the privileged few. He will have no desire to cast his pearls before
<sup>19</sup> Letter III: 'Ich danke dir, daß du mich bittest, dir von mir zu erzählen, daß du die vorigen Zeiten mir in's Gedächtniß bringst' (StA III, 10).
swine: 'I will tell you ever more of my bliss … But only to you, my Bellarmin, only to a pure free soul such as yours do I tell it. I will not be as prodigal as the sun with its rays, I will not cast my pearls before the foolish mob' (p. 59).20 The sun imagery in this passage is significant, given the mythological associations of Hyperion's name and his mentor Adamas' injunction in the fourth letter that he should live up to it. For this would certainly imply detachment ('You will be lonely', p. 14), but also an active role in the development of the things of this world. The narrator's reluctance to be as liberal with his communication as his illustrious heavenly namesake with its rays suggests that he is not yet equal to his mission. This immature, misanthropic elitism is one reason why he is inclined to keep his past to himself (and why even Bellarmin has been made to wait). The other is, as mentioned above, that he is anxious about stirring up painful memories (painful, either because in themselves distressing, or because they evoke vanished happiness). In the beginning he is highly selective about what he chooses to communicate. Even so, the repression is not entirely successful. We are meant to sense the tension in the narrator caused by the awareness of what he has yet to reveal. In the penultimate version of the novel we are explicitly, and perhaps somewhat unsubtly, told that the narrator can be writing of one thing and thinking of another: 'It's true that I managed to stay silent about it long enough, could often restrain myself when amongst the other memories this one would seize me; just you look! you'll find frenzied tears on many a trivial page; they belong here; I dried them and wrote of other things.'21 Since it would be difficult for anyone but Bellarmin to verify this, it is perhaps just as well that the passage was omitted from the final version. But it does at least emphasize the strain that his narrating activity places on Hyperion. Given the threat it poses for his stability, he needs a convincing reason for attempting it at all. If the mastering of this threat constitutes a major theme of the novel,
<sup>20</sup> Letter XXVIII: 'Ich will dir immer mehr von meiner Seeligkeit erzählen … Aber nur dir, mein Bellarmin, nur einer reinen freien Seele, wie die deine ist, erzähl' ich's. So freigebig, wie die Sonne mit ihren Strahlen, will ich nicht seyn; meine Perlen will ich vor die alberne Menge nicht werfen' (StA III, 69).
<sup>21 &#</sup>x27;Zwar konnt' ich doch lange genug davon schweigen, konnte oft mich halten, wenn unter den andern Erinnerungen diese mich ergriff; siehe nur hin! du wirst tobende Thränen finden auf mancher unbedeutenden Seite; sie gehören hieher; ich troknete sie und schrieb von andern Dingen' (StA III, 250–51).
as I would contend, it is difficult to see what other form could be chosen. Complaints about the perilous recalcitrance of one's subject matter and one's own inadequacy in grappling with it would appear hollow and silly, unless written with ironic intent, which is here obviously not the case. Hyperion's communications may grow into a book, an address to the public. They must on no account be seen to have been conceived as one from the outset by the narrator.
When Hölderlin temporarily abandoned the epistolary convention in the intermediate versions of *Hyperion*, in order to experiment with alternative forms, he still retained as the basic element a life-story with built-in commentary. We are presented in these versions with an already mature figure who looks back on the days of his youth, and comments and judges from a solid base of acquired insight. The higher perspective is already fixed, and is not seen to change or develop in the course of narration. How could it? In abandoning the letter form Hölderlin has deprived himself of that option. In returning to it he opens up the possibility of a double development: that of the narrator as a figure within his own narrative, and that of the narrator as he confronts his past life and attempts to come to terms with it in his writing. This would represent something more than the oscillation between different time levels of narration that one would expect to find in any epistolary novel. Such a balance of emphasis between past action and developing present, if this is indeed what we have in *Hyperion*, would seem to be Hölderlin's innovation.
It is the openness towards the future, in terms of the perspective of the individual narrator, that must have attracted Hölderlin to the epistolary form. Throughout the novel he is at pains to show that Hyperion, the writer of the letters, has a developing present and, for long stretches, an extremely uncertain future. He does this in various ways. One involves the repeated reference, already mentioned, to the narrating activity itself and the toll it is taking. Nor is it simply a matter of reflections or despairing interjections, indicating a changing mental state. The narrator's present is more than a featureless vantage point. It is given dimension in time and space. Whilst it is true that his letters have neither address nor dates, we know where he is when he writes them, and are given an impression of the passage of time through periodic references to the changing seasons. Thus, the correspondence opens in spring, with Hyperion, having just returned from Germany, in the vicinity of Corinth. Late summer finds him on Salamis, directing his gaze towards Calauria, Diotima's island, and preparing to tell the story of his love, something he has delayed until he feels sufficient strength within himself to do so. There is a reference to winter at the beginning of the twenty-sixth letter, as the account of the relationship draws towards its climax and the narrator approaches his worst crisis.22 It is in the second volume that Hyperion indicates recovery, after his protracted stay on Salamis, and it would be tempting to assume that both narrative and narration conclude in the spring. There is admittedly no direct evidence for this. But we know that Hyperion begins writing in the spring, and that his letters end with an account of that same spring as experienced in Germany, prior to his return to Greece. And since I will argue that there is a deliberate fusion of temporal perspectives at the end of the novel, the assumption seems to me to be a reasonable one. The writing of the letters would then have taken a complete year.23
Hölderlin's novel is subtitled 'the hermit in Greece', itself a sufficient indication that its subject is the writer of the letters and what happens to him in his isolation. One would not expect very much to happen to a hermit. The essence of his 'activity' here is the reflective and imaginative engagement with his own past, the inner processing of the events that drove him into his retreat in the first place, and this we witness in its entirety. If changes are to come about through this activity, it must be clearly seen to have extension in time, and Hölderlin does enough to
<sup>22</sup> Letter XXVI: 'I'm building a grave for my heart, that it may rest; I spin a cocoon around myself, because everywhere it's winter, in blissful memories I wrap myself against the storm' (p. 54); 'Ich baue meinem Herzen ein Grab, damit es ruhen möge; ich spinne mich ein, weil überall es Winter ist; in seeligen Erinnerungen hüll' ich vor dem Sturme mich ein' (StA III, 62).
<sup>23</sup> Knaupp illustrates the narrative structure of the novel with a useful diagram in the shape of an ammonite — see Michael Knaupp, *Friedrich Hölderlin*: *Hyperion* (Stuttgart: Reclam, 1997), p. 77. This would suggest that the narrated time goes from pre-1768 (the year Hyperion moves to Smyrna) to spring 1771, the narrating time from spring 1771 to (spring) 1772. Amongst the external reference points: the Greek uprising in the Peloponnese (Orlov revolt) began in February 1770; we know that Hyperion is injured in the battle of Chesma, 5–7 July 1770, and that, after a period of convalescence on Paros, he leaves for Germany (via Sicily) in the late autumn or winter of that year. For a detailed chronology and a map of Hyperion's journeys, see *Friedrich Hölderlin*: *Sämtliche Werke und Briefe*, ed. by Michael Knaupp, III (Munich: Hanser, 1992), pp. 318–19, 324–25.
convey this. More would have been distracting and alien to his purpose. All we really need to know is that the writer of the letters is living alone on an island, and that months are passing whilst he writes them. Despite his naturally-motivated orientation towards the past, he must be seen to have a present within which any development may take place, and this present must be open-ended. The final words of the novel are 'So I thought. More anon.'24 What might happen next is a question that will concern us later. For the moment it is sufficient to observe that the narrating Hyperion has his story which is both distinct from, and integrally related to, the story he is telling.
#### The Foreword
One thing emerges clearly from the note prefacing the first volume of *Hyperion*: Hölderlin is well aware of the likelihood that he will not be properly understood. As with his other famous preface, the one that introduces his great hymn 'Friedensfeier' ('Celebration of Peace', 1801), the tone here seems to be a mixture of apology and provocation.25 He fears that we may not know what to make of his work, but then that may perhaps be our fault. One clearly legitimate reason for Hölderlin's misgivings is that his public will initially have only the one volume available to them. In a novel with two lines of development (the hero within, and also outside and above his narrative), both running parallel on the printed page but the one running into the other chronologically (the letter writing begins when the narrated events have ended), the separate publication of one half will present more than ordinary problems in this respect. But at least Hölderlin underlines the status of his novel as a structured work. It has a 'Plan', and he knows what it is, even if the reader is going to
<sup>24</sup> Letter LX: 'So dacht' ich. Nächstens mehr' (StA III, 160).
<sup>25 &#</sup>x27;Ich bitte dieses Blatt nur gutmüthig zu lesen. So wird es sicher nicht unfaßlich, noch weniger anstößig seyn. Sollten aber dennoch einige eine solche Sprache zuwenig konventionell finden, so muß ich ihnen gestehen: ich kann nicht anders' (StA III, 532). In Michael Hamburger's translation: 'All that I ask is that the reader be kindly disposed towards these pages. In that case he will certainly not find them incomprehensible, far less objectionable. But if, nonetheless, some should think such a language too unconventional, I must confess to them: I cannot help it.' — See *Friedrich Hölderlin*: *Hyperion and Selected Poems*, ed. Eric L. Santner (New York: Continuum, 1990), p. 229.
have to wait to judge. However, Hölderlin's apparently well-founded scepticism about the likely reception of his novel is really rooted in doubts about the qualities and capabilities of his potential readership. They will be Germans, those of his own day, the very same who are going to be unmercifully castigated in Hyperion's penultimate letter. 'I'd happily promise this book the love of the Germans,' the preface begins, and the conditional is used advisedly.26 'Love', communal awareness, the sense of belonging to the whole, openness to the spirit of life in all its manifestations, reverence for nature and beauty, sensitivity to its articulation in poetry — that is precisely what will not be found amongst the Germans, who are, according to Hyperion, not only politically, but also and above all spiritually fragmented.27 It seems that the Germans represent, albeit in extreme form, the diseases of the modern world (similar criticisms are levelled by Hyperion at his own Greek contemporaries). In their fragmentation and division, their egoism and slavish disposition, the Germans are the direct antithesis of the ancient Greeks, as seen by Hyperion. (It should be noted that the praise of the latter and the attack on the former occupy corresponding positions, near the end of the first and second volumes respectively. It seems unlikely that this structural correspondence is fortuitous.) In view of the severity of Hyperion's censure of the Germans, one might think it naïve of Hölderlin even to entertain the possibility that they might learn to love his book. But then his anger is really a measure of his deep emotional commitment to them, or rather to what they could and should be, and might indeed even become, were his seed to take root. *Hyperion* is in no small degree an exploration of the ways in which change might be brought about, and the answer seems to be the novel itself (Hölderlin was not a modest man). It is intended to be educative, though not in a preaching, narrowly didactic sense. The lesson is aesthetic. Beauty is redemptive. Were the Germans to be such that they could respond to the novel in the proper manner, were they capable of 'loving' it, then they would indeed be changed.28
<sup>26 &#</sup>x27;Ich verspräche gerne diesem Buche die Liebe der Deutschen' (StA III, 5).
<sup>27</sup> See Letter LIX, p. 131: StA III, 155: 'ich kann kein Volk mir denken, das zerrißner wäre, wie die Deutschen …'
<sup>28</sup> Cf. Diotima's (past) vision of her ideal community (Letter LVI): 'Oh! now men no longer took the beautiful world like laymen the artist's poem when they praise the words and note the utility' (p. 113); 'Ach! nun nahmen die Menschen die schöne
Hölderlin points to two ways in which he expects his novel to be misread:
But I fear some will read it like a compendium and be overly concerned with the *fabula docet*, whilst others will take it too lightly, and neither party will understand it.
Those who merely sniff my flower mistake its nature, and so do those who pluck it merely for instruction.
The resolution of the dissonances in a particular character is neither for mere reflection nor empty pleasure. (p. 5)29
One is used to writers defending their fictions on the grounds that they both instruct and entertain, mixing the pleasant with the useful.30 Here Hölderlin appears to be saying that neither of these things, in themselves, is his main concern. Or rather, that any arbitrary separation of the two functions will lead to a fundamental misreading. Plundering the novel for ideas, the philosophical message, is just as aberrant as taking it simply as a story to while away one's leisure hours. In one sense what Hölderlin is arguing here is the autonomy of the work of art, the rejection of its subservience to ends and purposes outside itself, whether these be enlightenment or entertainment. He was of course not alone in defending the integrity of art against external manipulation, the attribution of objective purposefulness to what of its nature must be free and self-determined. It is, however, necessary to tread carefully here, lest one ascribe to Hölderlin a form of poetic solipsism which was alien to him. His art is not hermetic. Like all properly creative forms of human activity, it is for him a service rendered to nature.31 It is the
Welt nicht mehr, wie Laien des Künstlers Gedicht, wenn sie die Worte loben und den Nuzen drin ersehn' (StA III, 131).
<sup>29 &#</sup>x27;Aber ich fürchte, die einen werden es lesen, wie ein Compendium, und um das *fabula docet* sich zu sehr bekümmern, indeß die andern gar zu leicht es nehmen, und beede Theile verstehen es nicht. // Wer blos.an meiner Pflanze riecht, der kennt sie nicht, und wer sie pflükt, blos, um daran zu lernen, kennt sie auch nicht. // Die Auflösung der Dissonanzen in einem gewissen Karakter ist weder für das bloße Nachdenken, noch für die leere Lust' (StA III, 5).
<sup>30</sup> Horace, *Ars Poetica*, line 343.
<sup>31</sup> See the letter to his half-brother of 4 June 1799: 'Du siehest, Lieber, daß ich Dir das Paradoxon aufgestellt habe, daß der Kunst- und Bildungstrieb mit allen seinen Modifikationen und Abarten ein eigentlicher Dienst sei, den die Menschen der Natur erweisen.' In Charlie Louth's translation: 'You can see, dear Karl, that I have presented you with the paradox that the artistic and creative impulse with all its modifications and varieties is actually a service human beings render unto
unity of all life, 'einiges, ewiges, glühendes Leben', that is celebrated in the final letter of the novel, and it is Life that remains for Hölderlin the highest value.32 It is in art, specifically in poetry, that it receives articulation, quintessential expression. The work of art is autonomous in the same way as a living organism. It is an indivisible whole, yet at the same time part of the greater whole, the most intense individual manifestation of the all-pervading spirit of Life itself. It is not gratuitous that Hölderlin refers to his novel as a plant.
But what, one may ask, is the correct response to a flower? In so far as the work of art is incarnate beauty, an articulated vision of wholeness, it requires that we respond as wholes, and its effect should be to make us whole. Any one-sided appropriation is deleterious, whether we use the ideas in the work to school our intellects or simply allow the 'plant's' fragrance to give our senses an intoxicating lift. In the case of *Hyperion* the warning is not idly given. For disregard of the novel's peculiar form and structure must lead to a severe misunderstanding of the content. In fact, it is difficult to think of many other novels where such neglect can lead so fundamentally astray as here. The embodied ideas seem to be relativized at every turn, either through the narrator's distancing comments, or by becoming stages in the evolving consciousness of the narrator himself. Even the beautiful hymn to nature at the end of the novel, where one might expect to find a definitive statement, seems (at least initially) to be accorded merely provisional status: it is placed in inverted commas and qualified by the concluding 'So I thought.' Context is everything. Yet if the novel defies us to extract the message or messages, it does in fact advance the solution to a problem which is usually stated in philosophical terms. That which is neither for intellectual instruction nor for superficial, undemanding enjoyment — 'neither for mere reflection nor empty pleasure' — is said to be the 'resolution of the dissonances in a particular character'. Hölderlin does not here spell out the precise nature
nature.' — in *Hölderlin*: *Essays and Letters*, edited and translated with an introduction by Jeremy Adler and Charlie Louth (London and New York: Penguin Classics, 2009), p. 136.
<sup>32</sup> Cf. M. H. Abrams, *Natural Supernaturalism*: *Tradition and Revolution in Romantic Literature* (London: Norton, 1971), pp. 431 ff.: 'Life' is the 'ground-concept', 'the generator of the controlling categories of Romantic thought … Life is the premise and paradigm for what is most innovative and distinctive in Romantic thinkers. Hence their vitalism: the celebration of that which lives, moves and evolves by an internal energy, over whatever is lifeless, inert, and unchanging.'
of those dissonances, but in a sense he has already alluded to them. In so far as they are concerned with the fragmentation of human faculties, the division between thought and feeling, the discursive and the intuitive, detachment and involvement, reflection and spontaneity, it could be said that they are already adumbrated in the mutually contradictory approaches which Hölderlin imputes to his potential readers: *mere* reflection or *empty* pleasure. This implies no denigration of reflection ('Nachdenken'), or indeed pleasure ('Lust') — the adjectives suggest only that both are impoverished when detached from one another. We shall be able to understand the resolution of the dissonances only if, at least for the duration of our reading, we are able to resolve them in ourselves.
#### 'Not to be constrained by the greatest …'
It might well be objected that the preface to the final version of *Hyperion* is too cryptic to be of much help to the bemused reader. The earlier prefaces are certainly more explicit.33 They also have the considerable drawback of encouraging us to see the work in terms of a detachable scheme or programme, with a heavy philosophical bias. They concentrate in fact on the *fabula docet*, and whilst this might have been in some measure appropriate for the particular versions they preface, Hölderlin is now clearly anxious to avoid the impression that he is merely offering an exemplification of preconceived truths. Unlike the paradigmatic states of harmony and perfection on which he dwells in the earlier prefaces, the unspecified 'resolution of the dissonances' is something that must be realized or enacted in the work itself. What Hölderlin does, however, retain from the first preface is the quotation he now uses as a motto to introduce the first volume: 'Non coerceri maximo, contineri a minimo, divinum est.'34 This time there is no attribution, no explanation given, and we are left to apply it as we may. It is in fact taken from the first section of an anonymous literary epitaph in praise of the founder of the Jesuits, Ignatius Loyola, first published in 1640.35 How Hölderlin came by it
<sup>33</sup> See StA III, 163, 235–37.
<sup>34 &#</sup>x27;Not to be constrained (confined, enclosed) by the greatest, to be contained (enclosed) by the smallest, is divine.'
<sup>35</sup> For details and a fuller context, see StA III, 437–38. In the original, and in the preface to the 'Fragment von Hyperion', it reads: 'contineri tamen a minimo'.
is not known for certain, but his most likely source would seem to be a virulently anti-Jesuit history of the order that began to appear in 1789.36 One is perhaps surprised by the peripheral intrusion of Jesuits into Hölderlin's novel, for whatever else might have motivated the choice of name for Hyperion's correspondent, as a trained theologian Hölderlin must have been aware of Cardinal Robert(o) Bellarmin(o) (1542–1621), as were no doubt many of his readers. But whether the latter would have been able to recognize the unattributed quotation or connect it with Jesuits is quite another matter. One who nowadays could is Francis, the first Jesuit Pope, and he is apparently fond of citing it. It also seems to appeal to his predecessor who even refers to *Hyperion* in this context.37
When Hölderlin decided to omit any guiding interpretation of the Loyola epitaph from the final version of *Hyperion*, he almost certainly did so because he realized that it would be unnecessarily limiting. Quite apart from the general considerations that motivated the excision of explicit theorizing from the preface, he would have recognized in the epitaph a peculiar aptness to all his major concerns, whether religious, political, philosophical or aesthetic. Even the addition of Loyola's name might have seemed in some measure prescriptive. Interesting is the addition of the 'divinum est'. This is present in the original epitaph, but was omitted from the preface to the 'Fragment' where we are confronted with the contrary drives — glossed as man wanting at the same time to be *in* everything and *above* everything38 — together with the implication that a notional reconciliation would represent the highest state achievable, but also that it is the very striving to achieve such balance which can be dangerous and lead to catastrophe: the attempt to reconcile the polar tensions could result in our being ripped apart by them.39 In the preface to the penultimate version of the novel, which dispenses
<sup>36</sup> Peter Philipp Wolf, *Allgemeine Geschichte der Jesuiten von dem Ursprunge ihres Ordens bis auf gegenw*ärtige *Zeiten*, 4 vols (Zurich: Orell, Geßner, Füssli, 1789–92). The epitaph is quoted in full in the first volume, pp. 214–17.
<sup>37</sup> Josef Cardinal Ratzinger, *Introduction to Christianity* (San Francisco: Ignatius Press, 2004), pp. 146–47.
<sup>38 &#</sup>x27;Der Mensch möchte gerne *in* allem und *über* allem seyn' (StA III, 163).
<sup>39</sup> Similar tensions give Goethe's Werther cause for reflection in his letter of 21 June: 'I've had all manner of thoughts about the desire [*Begier*] human beings have to extend themselves … to rove far and wide; and then about the impulse [*Trieb*] in them willingly to accept constraints'; translated by David Constantine, *The Sorrows of Young Werther* (Oxford: Oxford University Press, 2012), p. 24.
with Loyola and epitaph altogether, we — and Hyperion — are seen as torn between extremes of massive, solipsistic expansion on the one hand, and contraction to the point of obliteration on the other. Both are dangerous, because they are literally self-destructive. This dialectical tension between self and world is an inescapable part of human experience, and yet it is the drives that create it which also charge us to overcome it.40
The restoration of 'divinum est' suggests a religious dimension, if not in any narrow confessional sense. In its original context the statement is naturally closely associated with death, the narrow grave confining the mortal remains of one whose indomitable spirit knew no bounds:
Cuius Animus Vastissimo coerceri non potuit unius orbis ambitu, Ejus Corpus Humili hoc angustoque tumulo continetur. Qui magnum aut *Pompeium* aut *Caesarem* aut *Alexandrum* cogitas, Aperi oculos veritati, Majorem his omnibus leges *Ignatium*. Non coerceri maximo, contineri tamen a minimo, divinum est.41
The contrast is familiar to us from Shakespeare. One thinks here particularly of Prince Henry's lines on the death of Hotspur: 'When that this body did contain a spirit, / A kingdom for it was too small a bound, / But now two paces of the vilest earth / Is room enough.'42 These lines are echoed in a text that was of great, if still largely unacknowledged significance for Hölderlin, namely Macpherson's *Ossian* (1762–65): 'Narrow is thy dwelling now; dark the place of thine abode. With three steps I compass thy grave, O thou who wast so great before!'43 Mention
<sup>40</sup> See StA III, 236.
<sup>41</sup> Wolf, I, pp. 214–15. 'He whose spirit could not be contained by a whole globe's bounds, his body is enclosed in this low and narrow tumulus. You who deem great Pompey or Caesar or Alexander, open your eyes to the truth: as the greater than all these you will choose Ignatius …'
<sup>42</sup> *Henry IV* (Part I), Act V, Scene 4.
<sup>43</sup> James Macpherson, *The Poems of Ossian and Related Works*, ed. by Howard Gaskill (Edinburgh: Edinburgh University Press, 1996), pp. 24, 168. The passage is included in the 'Songs of Selma', translated by Goethe's Werther.
of Caesar and Alexander might also remind us of the Yorick scene in *Hamlet*:
Alexander died, Alexander was buried, Alexander returneth into dust, the dust is earth, of earth we make loam, and why of that loam, whereto he was converted, might they not stop a beer barrel?
Imperious Caesar, dead and turn'd to clay, Might stop a hole to keep the wind away. Oh, that that earth, which kept the world in awe, Should patch a wall t' expel the winter's flaw!44
But then Hamlet can express the great/small contrast in terms of a familiar paradox without any reference to death: were it not for bad dreams, 'I could be bounded in a nutshell, and count myself a king of infinite space.'45
By omitting any specific reference to Loyola or to the motto's being an epitaph, Hölderlin ensures that any primary association with death is absent, and thus widens the scope of the statement considerably without in any way falsifying it. In fact, it does not seem to me that it was the anonymous author's intention to restrict his meaning either; quite the reverse. He is not simply saying that it is divine to have a great spirit and a dead body, but rather seeing in Loyola's situation the illustration of a glorious universal truth, by associating his hero with Christ himself. For this gnomic statement ought to be recognized as an allusion to the miracle of Incarnation. It is a variation of a commonplace in medieval hymnology. Sometimes, admittedly, it can be used to refer to the death of Christ: 'Brevo sepulcro clauditur, qui caelo non capitur' ('He is enclosed in the narrow tomb, he whom the heavens cannot contain').46 But often the enclosing space is not the tomb but the Virgin's womb: 'Quem nequit totus cohibere mundus / Claudis in alvo' ('He whom the whole world cannot contain, / You enclose in your womb').47 Or: 'Intra te clauditur, / Qui claudit omnia' ('In you is enclosed / he who encloses everything').48 It is the paradox of incarnate divinity.
<sup>44</sup> *Hamlet*, Act V, Scene 1.
<sup>45</sup> *Hamlet*, Act II, Scene 2.
<sup>46</sup> Peter Damian, 'Pascalis rhythmus ad procedendum', in *Poésie latine chrétienne du Moyen âge*: *IIIe–XVe siècle*, Textes recueillis, traduits et commentés par Henry Spitzmuller (Paris: Desclée, De Brouwer, 1971), p. 404.
<sup>47</sup> Peter Damian, 'In assumptione hymnus', in Spitzmuller, p. 408; cf. also 'De virgine hymnus', pp. 408 ff.
<sup>48</sup> Alexander Neckham [?], 'De virgine cantio', in Spitzmuller, p. 747.
*Hyperion* was, for the most part, written at a time when Hölderlin was furthest removed from any kind of Christian orthodoxy. The sudden irruption of Christ into his poetry comes after 1800, although there are obvious intimations of his later concerns, notably in the shape of the martyr and redeemer Empedokles, the subject of Hölderlin's unfinished drama (1797–1800). In November 1798 he can ask his halfbrother: 'when will people recognize that the highest power is in its expression also the most modest and that the divine, when it makes itself manifest, can never be without a certain sadness and humility?'49 It seems to me that the notion of suffering incarnate divinity is already clearly implied here. It is this generalized insight that will open the way for the (attempted) reintegration of Christ into Hölderlin's world, albeit on the poet's own terms. And it is in the novel that the process really begins. For the mature insight which Hölderlin offers to his brother is one which is worked out in *Hyperion*, and by Hyperion. The absolute is available to us through beauty, embodied perfection, the individual manifestations of which, whether in the glories of ancient Greece or the figure of Diotima herself, are temporal and therefore transient. It takes Hyperion a long time to come to terms with this: 'O Bellarmin! who then may say he stands fast when even the beautiful ripens so towards its fate, when even the divine must humble itself and share mortality with all that is mortal!' (p. 82).50 If this represents the narrator's gradual and reluctant recognition of a necessary truth, marking already a measurable advance in understanding, it is transformed near the end into joyful affirmation:
… and you ask, my Bellarmin! how I feel now, telling you this?
My dearest friend! I'm at peace, for I want no better than the gods. Must not everything suffer? And the nobler it is, the more deeply. Does not holy nature suffer? O my godhead! that you could grieve as you are blissful is something I long failed to grasp. But the bliss that does not suffer is sleep, and without death there is no life. Should you be, like a child, eternal and slumber as nothingness? forego the victory? not pass
<sup>49 28</sup> November 1798: 'O, Lieber! wann wird man unter uns erkennen, daß die höchste Kraft in ihrer Äußerung zugleich auch die bescheidenste ist, und daß das Göttliche, wenn es hervorgeht, niemals ohne eine gewisse Trauer und Demuth seyn kann?' Translation by Charlie Louth, in *Hölderlin*: *Essays and Letters*, p. 111.
<sup>50</sup> Letter XXXI: 'O Bellarmin! wer darf denn sagen, er stehe vest, wenn auch das Schöne seinem Schiksaal so entgegenreift, wenn auch das Göttliche sich demüthigen muß, und die Sterblichkeit mit allem Sterblichen theilen!' (StA III, 94).
through all the perfections? Yes! yes! worthy is pain to lie at men's hearts and be your familiar, O nature! For it alone leads from one bliss to the next, and there can be no other companion. (p. 129)51
The radical dualism posited in the chronologically much earlier 'Song of Fate' (p. 123), between oblivious gods and suffering humanity, is here overcome. Hyperion finds his 'peace' (if not his resting place) in the revelation that it is the very essence of the gods' divinity to be, not only above everything, but also in everything, and as such, in their temporal aspect, subject to the same limitations as all living things, the same laws of change and decay. Without this they are nothing. For to be both is divine. The bliss of what is most intensely alive is inseparable from pain, and the featureless limbo of comatose non-feeling is no alternative. The 'condescension' of the gods, their submission to the shackles of time and space, is indeed redemptive, since through it man and all creation is raised, becoming a fitting and necessary element in the divine pageant. Everything that lives is holy.
The preface speaks of the 'resolution of the dissonances'. The Latin motto, as it stands, implies that it is divine to combine two apparently mutually contradictory qualities or modes of being. How then is this related to the conflicts sustained by Hyperion, and their eventual outcome? Each mode is in fact individually represented by a central character in the novel, and both exist in a state of initially unresolved tension in Hyperion himself. For all his nobility, Alabanda can be seen to represent the hubristic dangers inherent in idealistic radicalism. Not being willing to be constrained is allied with the impulse to dominate, to do violence to the freedom of others. It is an expression of the titanic 'monstrous striving to be everything' (p. 16) that, left unchecked,
<sup>51</sup> Letter LVIII: 'und du fragst, mein Bellarmin! wie jezt mir ist, indem ich diß erzähle? Bester! ich bin ruhig, denn ich will nichts bessers haben, als die Götter. Muß nicht alles leiden? Und je treflicher es ist, je tiefer! Leidet nicht die heilige Natur ? O meine Gottheit! daß du trauern könntest, wie du seelig bist, das konnt' ich lange nicht fassen. Aber die Wonne, die nicht leidet, ist Schlaf, und ohne Tod ist kein Leben. Solltest du ewig seyn, wie ein Kind und schlummern, dem Nichts gleich? den Sieg entbehren? nicht die Vollendungen alle durchlaufen? Ja! ja! werth ist der Schmerz, am Herzen der Menschen zu liegen, und dein Vertrauter zu seyn, o Natur! Denn er nur führt von einer Wonne zur andern, und es ist kein andrer Gefährte, denn er. — ' (StA III, 150). Beissner in his annotation of this passage (StA III, 486) insists that 'ewig' in 'ewig seyn' is used adverbially, so that the question would mean: 'should you be eternally like a child …?' This strikes me as a grotesque (if common) misreading.
inevitably leads to tyranny, alienation, and catastrophe.52 Alabanda will not suffer constraint and roams far and wide in a world he despises, eager to put a new one in its place, and violently if necessary. When his efforts are frustrated, he remains defiant and unrepentant, his fiery spirit undiminished, his awareness of a pure core of indestructible selfhood his only faith:
'So … without freedom everything is dead.'
'Yes indeed,' he cried, 'there's not a blade of grass grows up unless it has its own germ of life within! how much the more in me! and therefore, my dear friend! because I feel myself free in the highest sense, because I feel myself beginningless, therefore I believe that I'm endless, that I'm indestructible. If a potter's hand has made me, then let him smash his vessel as he pleases. But that which lives within must be unbegotten, must be divine of nature in its germ, sublime beyond all might and all art, and therefore invulnerable, eternal.' (p. 121) 53
It is a vision that presupposes dissonance and conflict. Here freedom is expressed in an image of violence, as it has to be, since liberty is understood as the assertion of self against the constraints of inhibiting forces which are experienced as external and alien to the self. It is a vision based on division, no doubt valid in its context, as the expression of an explosive phase in the life of individuals or communities periodically necessary and justified, but no more fitted to represent the norm than are floods, volcanoes, or earthquakes within the economy of nature. For Diotima by contrast, the model is one of unity and harmony. Within a narrowly circumscribed environment she lives an idyll of naïve, unreflecting fulfilment, like a beautiful flower embedded in a protected garden. She does not feel constricted, since all she knows is part of her and she of it. She can express herself only by expressing the world around her; she is a living hymn to nature, and her proper medium is not speech but song.
<sup>52</sup> Letter V: 'das ungeheure Streben, Alles zu seyn' (StA III, 18).
<sup>53</sup> Letter LVII: 'So [wär' es hier im höchsten Sinne wahr, erwiedert' ich, daß] ohne Freiheit alles todt ist. // Ja wohl, rief er, wächst doch kein Grashalm auf, wenn nicht ein eigner Lebenskeim in ihm ist! wie viel mehr in mir! und darum, Lieber! weil ich frei im höchsten Sinne, weil ich anfangslos mich fühle, darum glaub' ich, daß ich endlos, daß ich unzerstörbar bin. Hat mich eines Töpfers Hand gemacht, so mag er sein Gefäß zerschlagen, wie es ihm gefällt. Doch was da lebt, muß unerzeugt, muß göttlicher Natur in seinem Keime seyn, erhaben über alle Macht, und alle Kunst, und darum unverlezlich, ewig' (StA III, 141). The 'potter's hand' alludes to Jeremiah, chapters 18 and 19; also Isaiah, 29:16, underlining the willed impiety of Alabanda's words.
It is the disruptive influence of Hyperion that wrenches her from her idyll and makes her articulate. The vicarious experience of an out-ofjoint world through the man she loves uproots and ultimately destroys her (as an individual), since it makes impossible a return to her haven of innocence. It is, however, the shocking knowledge of dissonance that also enables her to give voice and differentiated expression to what she has known and lived, that peace and harmony which pass all human understanding. Her conception of freedom, the cornerstone of her faith, is the experienced oneness with the life of nature. It is something that is available to her now only through death, the final shedding of all that makes her distinct and different, painfully detached from that life. Her eloquent and lovely swansong displays a form of egalitarian radicalism that puts to shame Alabanda's elitist and egocentric liberalism:
Those poor souls who know nothing but their own sorry handiwork, who are merely slaves of need and scorn genius, and who do not honour you, childlike life of nature! let them fear death. Their yoke is become their world; they know nothing better than their bondage; dread the freedom of the gods that death gives us?
But I don't! I have gone beyond the patchwork made by hand of man, I have felt the life of nature that passes all thought — even if I became a plant, would the harm be so great? — I shall be. How should I be lost from the sphere of life where the love eternal, that is common to all, holds all natures together? How should I depart from the covenant that binds all beings? This is not so easily broken as the loose bonds of our age. It's not like a market day when folk flock together and make a great bustle and then disperse. No! by the spirit that unites us, by the god's spirit that is proper to each and common to all! no! no! in nature's covenant troth is no dream. We part to be only more inwardly one, more divinely at peace with all, with ourselves. We die in order to live.
I shall be; I do not ask what I shall become. To be, to live, that is enough, that is the glory of the gods; and that is why all that is life is equal in the divine world, and there are there no masters and menials. Natures live together like lovers; they have all in common, spirit, joy and eternal youth. (p. 127)54
<sup>54</sup> Letter LVIII: 'Die Armen, die nichts kennen, als ihr dürftig Machwerk, die der Noth nur dienen und den Genius verschmähn, und dich nicht ehren, kindlich Leben der Natur! die mögen vor dem Tod sich fürchten. Ihr Joch ist ihre Welt geworden; Besseres, als ihren Knechtsdienst, kennen sie nicht; scheun die Götterfreiheit, die der Tod uns giebt? // Ich aber nicht! ich habe mich des Stükwerks überhoben, das die Menschenhände gemacht, ich hab' es gefühlt, das Leben der Natur, das höher
The stress is on equality, but here, as in Hyperion's own equivalent concluding statement, it is a non- or even anti-Jacobin conception of equality that is voiced. Diotima delights in the diversity of all genuine manifestations of Life, insisting only that they are all equal in value and fundamentally related, emphatically not that they should be forced to conform to one another. In this celebration of Life, from which all hierarchical notions are absent, Diotima wishes for nothing better than to be 'in everything'. She would be content to be 'contained by the smallest': 'even if I became a plant, would the harm be so great?' Whatever Hölderlin may have meant in the preface to the 'Fragment', it is clear that this carries no negative connotations here. There is a definite rightness about everything Diotima is meant to say throughout the novel, and her statements are never relativized in any way. It is only that their ultimate truth is not immediately grasped by Hyperion himself. Given the undoubted authority of Diotima's words, it is thus all the more significant that she stresses that her way should not be his; not because the ultimate goal of unity with the life of nature is wrong, but because he is to realize it in a less private, self-effacing manner. It is Diotima who assigns Hyperion his mission. After addressing her trinity of sun, earth, and aether, and praying that errant man will return from his exile into the divine fold, into the home of nature, she continues: 'You know this word, Hyperion! You began it in me. You will fulfil it in yourself, and only then rest.'55 In what sense he is to realize the promise of her words arguably becomes finally clear to Hyperion only when he copies out for Bellarmin the letter in which they occur. To the reader
ist, denn alle Gedanken — wenn ich auch zur Pflanze würde, wäre denn der Schade so groß? — Ich werde seyn. Wie sollt' ich mich verlieren aus der Sphäre des Lebens, worinn die ewige Liebe, die allen gemein ist, die Naturen alle zusammenhält? Wie sollt' ich scheiden aus dem Bunde, der die Wesen alle verknüpft? Der bricht so leicht nicht, wie die losen Bande dieser Zeit. Der ist nicht, wie ein Markttag, wo das Volk zusammenläuft und lärmt und auseinandergeht. Nein! bei dem Geiste, der uns einiget, bei dem Gottesgeiste, der jedem eigen ist und allen gemein! nein! nein! im Bunde der Natur ist Treue kein Traum. Wir trennen uns nur, um inniger einig zu seyn, göttlicherfriedlich mit allem, mit uns. Wir sterben, um zu leben. // Ich werde seyn; ich frage nicht, was ich werde. Zu seyn, zu leben, das ist genug, das ist die Ehre der Götter; und darum ist sich alles gleich, was nur ein Leben ist, in der göttlichen Welt, und es giebt in ihr nicht Herren und Knechte. Es leben umeinander die Naturen, wie Liebende; sie haben alles gemein, Geist, Freude und ewige Jugend' (StA III, 147–48).
<sup>55 &#</sup>x27;Du kennst diß Wort, Hyperion! Du hast es angefangen in mir. Du wirsts vollenden in dir, und dann erst ruhn' (StA III, 147).
it should be obvious that she envisages for him a social and public function which, for the time being at least, would deny him the luxury of dying into nature. This function is unambiguously related to the bringing back of his fellow-men out of their lamentable alienated state to the harmony that is their birthright. It is not enough for Hyperion to experience unity; he must himself unify. Taking the three watchwords of the French Revolution, it could be said that Alabanda's 'liberty' and Diotima's 'equality' find their resolution in the 'fraternity' which Hyperion must strive to realize.
Hyperion actually uses the adjective 'brüderlich' (fraternal) in the vision of unity celebrated in the final letter of the novel:
"You springs of the earth! you flowers! and you forests and you eagles and you fraternal light! how old and new is our love! — We are free, we don't anxiously strive to be outwardly equal; how should not vary the modes of life? but we all love the aether, and deep in our innermost being we are all of us like." (p. 136)56
Here the word is applied to the light and any political meaning might seem oblique. But since it is immediately followed by a sentence featuring the adjective 'frei' (free) and the verb 'gleichen' (to be equal, like), it is difficult to believe that this is fortuitous and the language innocent of any revolutionary associations. On reading and being impressed by a substantial early fragment of the novel Hölderlin's mentor, Friedrich Stäudlin, told him to make sure to insert into the finished product 'cryptic passages about the spirit of the age'.57 That spirit was of course one of revolutionary turmoil and upheaval, and it is indeed reflected in the novel in a number of telling ways (though perhaps not so telling that any political import is immediately obvious). Like many of his generation — one might think here of Wordsworth, also born in 1770 — he was enthused by the promise of the French Revolution. Again, like many, he became appalled by the excesses of a revolutionary fanaticism which embodied the negation of the very ideals it purported to uphold. But unlike some, he continued to hold firm his faith in
<sup>56</sup> Letter LX: 'Ihr Quellen der Erd'! ihr Blumen! und ihr Wälder und ihr Adler und du brüderliches Licht! wie alt und neu ist unsere Liebe! — Frei sind wir, gleichen uns nicht ängstig von außen; wie sollte nicht wechseln die Weise des Lebens? wir lieben den Aether doch all' und innigst im Innersten gleichen wir uns' (StA III, 159).
<sup>57 &#</sup>x27;Unterlassen Sie doch nicht … versteckte Stellen über den Geist der Zeit in dieses Werk einzuschalten!!!' (4 September, 1793; StA III, 299).
the validity of the ideals themselves.58 Hölderlin situates his novel in Greece, the classical cradle of democracy; significantly, though, not in ancient Greece but in that of the near-present, the historical background being the unsuccessful Russian-backed insurrection against the occupying Turkish power, which took place in 1770.59 Hyperion, as one 'keen to see the world a better place' (p. 29), finds himself suddenly confronted, as he sees it, with the practical opportunity of making his dreams a reality. Disaster and disillusionment quickly follow. But we misinterpret Hyperion's failure as guerrilla and military leader, and indeed the failure of the insurrection itself (historical fact), if we take it to imply that revolutionary activism is wrong in any absolute sense. Surely Hölderlin would not be wishing to argue that the Greeks should not try to rid themselves of the Turkish yoke. The point is that Hyperion must not expect his ideal community to be brought into being by such methods. Diotima's misgivings about Hyperion's active participation in the revolt are well founded. Ideologically motivated violence has an inevitable tendency to lose touch with its original inspiration, and become an end in itself: 'You will conquer … and forget what for' (p. 84).60 And crucially, Hyperion himself is not destined to serve the cause in this way, though others may be: 'Even if it's right … it's not what you were born for' (p. 83).61
What then is Hyperion's true vocation? According to Diotima's vision for him at the end of the first volume, he is to become the educator of his people.62 It is difficult to see how he can do so by remaining a hermit. The subtitle of the novel is justified because what is documented is Hyperion's journey into isolation and his gradual emergence from it. The figure of the recluse thus stands in the middle of Hyperion's development, and yet at the same time at the beginning and end of the novel. It is the incipient hermit who writes the first letter, and it is the
<sup>58</sup> See my essays 'Hölderlin and Revolution', *Forum for Modern Language Studies*, 12 (1976), 118–36; and '"Diß ist die Zeit der Könige nicht mehr": Hölderlin and Republicanism', *Strathclyde Modern Language Studies*, *Special Issue*: *The French Revolution*: *German Responses*, 11 (1991), 39–56.
<sup>59</sup> For the so-called Orlov revolt, the nature of Hölderlin's sources, and the use he makes of them, see David Constantine, 'The Insurrection of 1770', in *In the Footsteps of the Gods*: *Travellers to Greece and the Quest for the Hellenic Ideal* (London: Taurus Parke Paperbacks, 2011), pp. 168–87.
<sup>60</sup> Letter XXXIII: 'Du wirst erobern … und vergessen, wofür' (StA III, 96).
<sup>61</sup> Letter XXXIII: 'Wenns auch recht ist … du bist dazu nicht geboren' (StA III, 96).
<sup>62</sup> Letter XXX: 'Du wirst Erzieher unsers Volks' (StA III, 89).
last of the letters that relates how he came to turn his back on humanity and throw himself into the arms of nature. But the Hyperion who writes at the end is no longer a hermit. He may still be detached from the world of men, but at least he is now looking towards and not away from it. For he is communicating with it, and in his last major direct statement, the scathing censure of the Germans, he is making himself the spokesman of those who suffer in it and would presumably wish to change it.63 There is of course a nice paradox in the last two letters of the novel, the Hebrew-prophet-like berating of the Germans being followed by the ecstatic hymn to nature. For it is the hermit, in his misanthropic isolation, who experiences the ineffable bliss of all-unity. And it is the mature Hyperion, no longer at odds with himself, and alive to his social responsibilities, who paints a most devastating picture of fragmentation and division. Implicit in both situations is a combination of involvement and detachment. But true synthesis is achieved only on the level of the narrator, for it is he who juxtaposes his philippic with the vision of all-unity which, though the product of the hermit's experience, is here articulated for the first time.
The precise status of the hymn to nature in the last letter has been the subject of much discussion. Formerly it could be widely assumed that the disillusioned and battle-scarred Hyperion retires to commune with nature, and that was that. If there were to have been anything more, it would have had to come in a third volume, which is indeed what some of Hölderlin's contemporaries seem to have expected. In the absence of that we are left with an uncompromising statement of radical pantheism, representing the sum of Hyperion's wisdom after his disasters in love and war. Recognition of the novel's sophisticated narrative structure, with the distinction between narrated and narrating time, has naturally led to a reappraisal, and even a devaluation of the significance of the last letter. For if one is going to argue that Hyperion develops in the course of his narrating, it seems difficult to admit the conclusive validity of an experience he has before he starts writing. And indeed, in so far as the vision of unity embodies coherent thoughts, Hyperion seems expressly to distance himself from them in the novel's final words: 'So I thought,'
<sup>63</sup> Letter LIX: 'I spoke in your name too, I spoke for all who live in this land and suffer as I suffered there' (p. 134); 'Ich sprach in deinem Namen auch, ich sprach für alle, die in diesem Lande sind und leiden, wie ich dort gelitten' (StA III, 156).
the implication being that he has moved on since then, though he has yet to tell us where. Yet by its very position at the end of the novel the hymn to nature is given authority. It is what is left ringing in our ears when we finish our reading. Moreover, attempts to differentiate the narrator's painfully acquired insights — 'My dearest friend! I'm at peace …' (p. 129) — from those expressed in the vision experienced in the German spring are forced to resort to casuistry. Though the accents might be different, it all seems to be there already: the acceptance of suffering (as another word for joy), its function within the dynamic harmony of nature, appreciation of essential unity in the midst of apparent chaos, the identification of life and love. It seems that we have come full circle and that the end result of the narrator's deliberations is to discover what he already knew. And in a sense this is true.
Perhaps the most important point to make about the vision of unity at the end of the novel is that it is not sustained. It is not simply that the ecstatic mood ebbs away, for it would not be humanly possible to prolong it beyond a certain point, but that the experience itself does not lead to lasting insight. Or rather the insights associated with it appear to have no sustaining power. That this is so is shown by the narrator's sorry plight in the opening letters, which of course postdate the beatific vision with which the novel closes. That vision, nevertheless, represents a turning-point in Hyperion's development and the axis of the novel itself. Though it apparently quickly fades, together with the confidence it inspires, it is — while it lasts — an experience of unique intensity and cosmic vastness, which takes him far beyond anything he has known before. His thoughts and feelings suffuse the whole of nature, no longer focused on the solitary mediatory figure of the loved one. It is the shock of reawakening and finding himself finite again that causes Hyperion most distress, and it is this jarring contrast which is treated at length in the second letter of the novel and constitutes perhaps the most obvious 'dissonance' of all. The experience of allunity may be transient, but it is repeatable. Once Hyperion has broken through the barriers and succeeded in universalizing his love for Diotima, extending it to the whole of nature, he is able to do so again and again. The first such experience, precisely because it represents a breakthrough, is the most important, and this in itself justifies its prominent position at the end of the novel. It marks the end of a phase
in Hyperion's life, and also the beginning of a new one with a new set of problems. In the second letter we find Hyperion struggling to come to terms with the provisional and fragile nature of a synthesis which had seemed decisive and absolute:
My whole being stills and listens when the gentle ripple of the breeze plays about my breast. Often, lost in the immensity of blue, I look up into the aether and out into the hallowed sea, and it's as if a kindred spirit opened its arms to me, as if the pain of isolation were dissolved in the life of the godhead.
To be one with everything, that is the life of the godhead, that is the heaven of man.
To be one with everything that lives, to return in blissful self-oblivion into the all of nature, that is the summit of thoughts and joys, that is the holy mountain pinnacle, the place of eternal peace where noon loses its sultriness and the thunder its voice and the boiling sea becomes like a waving corn-field.
To be one with everything that lives! At these words virtue lays aside its wrathful harness, the mind of man its sceptre, and all thoughts melt away before the vision of the world's eternal oneness like the toiling artist's rules before his heavenly Urania, and iron fate renounces its dominion, and from the covenant of beings death disappears, and indivisibility and eternal youth blesses, makes beautiful the world. (p. 8)64
Then, however, comes the inevitable fall:
On this height I often stand, my Bellarmin! But a moment of reflection casts me down. I begin to think, and find myself as I was before, alone, with all the pains of mortality, and my heart's sanctuary, the world's
<sup>64</sup> Letter II: 'Mein ganzes Wesen verstummt und lauscht, wenn die zarte Welle der Luft mir um die Brust spielt. Verloren in's weite Blau, blik' ich oft hinauf an den Aether und hinein in's heilige Meer, und mir ist, als öffnet' ein verwandter Geist mir die Arme, als löste der Schmerz der Einsamkeit sich auf in's Leben der Gottheit. // Eines zu seyn mit Allem, das ist Leben der Gottheit, das ist der Himmel des Menschen. // Eines zu seyn mit Allem, was lebt, in seeliger Selbstvergessenheit wiederzukehren in's All der Natur, das ist der Gipfel der Gedanken und Freuden, das ist die heilige Bergeshöhe, der Ort der ewigen Ruhe, wo der Mittag seine Schwüle und der Donner seine Stimme verliert und das kochende Meer der Wooge des Kornfelds gleicht. // Eines zu seyn mit Allem, was lebt! Mit diesem Worte legt die Tugend den zürnenden Harnisch, der Geist des Menschen den Zepter weg, und alle Gedanken schwinden vor dem Bilde der ewigeinigen Welt, wie die Regeln des ringenden Künstlers vor seiner Urania, und das eherne Schiksaal entsagt der Herrschaft, und aus dem Bunde der Wesen schwindet der Tod, und Unzertrennlichkeit und ewige Jugend beseeliget, verschönert die Welt' (StA III, 8–9).
eternal oneness, is no more; nature's arms are closed, and I stand before her like a stranger and cannot comprehend her. (p. 9)65
'On this height I often stand …' The feeling of unity with all that lives, first experienced with overwhelming intensity in the German spring, has since recurred and can be reproduced by Hyperion (in short bursts), as it seems, almost at will: 'just forget that men exist, starving, vexed and deeply harassed heart! and return whence you came, into the arms of nature, never-changing, beautiful and tranquil' (p. 8).66 Yet it appears that Hyperion has lost as much as he has gained. In the first place, the experience is achieved at the cost of isolation from the rest of humanity. The 'world's eternal oneness' provides a refuge, his 'heart's sanctuary', into which the hermit escapes, effectively denying his brotherhood with those he should be helping: 'And one last time I looked back into the cold night of men and shuddered and wept for joy that I was so blissful' (p. 136).67 And secondly, having once achieved this bliss, he has to face the fact that it too is ephemeral and cannot be held, but at best recaptured in isolated moments. Moreover, the very intensity of this periodical, and paradoxically exclusive experience of unity makes its loss or absence, which constitutes Hyperion's normal state, all the more bitter and difficult to bear. Thus what had seemed to offer an ultimate resolution of Hyperion's conflicts proves in fact to have exacerbated them by unbearably intensifying the contrast, in terms of his own experience, between unity and separation, fulfilment and despair. To have only one foot in Eden, as opposed to both or neither, proves to be the worst fate of all.
This is Hyperion's situation at the beginning of the novel, that is to say, when he begins to write. His writing proceeds from the experience in extreme form of both unity and alienation, and the dominant tone at the beginning is one of despair at the periodicity and ultimate
<sup>65 &#</sup>x27;Auf dieser Höhe steh' ich oft, mein Bellarmin! Aber ein Moment des Besinnens wirft mich herab. Ich denke nach und finde mich, wie ich zuvor war, allein, mit allen Schmerzen der Sterblichkeit, und meines Herzens Asyl, die ewigeinige Welt, ist hin; die Natur verschließt die Arme, und ich stehe, wie ein Fremdling, vor ihr, und verstehe sie nicht' (StA III, 9).
<sup>66</sup> Letter I: 'vergiß nur, daß es Menschen giebt, darbendes, angefochtenes, tausendfach geärgertes Herz! und kehre wieder dahin, wo du ausgiengst, in die Arme der Natur, der wandellosen, stillen und schönen' (StA III, 8).
<sup>67</sup> Letter LX: 'Und Einmal sah' ich noch in die kalte Nacht der Menschen zurük und schauert' und weinte vor Freuden, daß ich so seelig war' (StA III, 159).
ineffectualness of the moments of harmony. They cannot be preserved, carried over into his normal waking life and made fruitful there. They presuppose abandonment of self, the temporary eclipse of consciousness and individuation. As soon as Hyperion attempts to grasp what it is he is experiencing, the experience itself dissolves into nothing. And this is as it must be. For thinking is a divisive activity that presupposes a conscious, and therefore distinct, finite, and transient subject reflecting on, and therefore detached from a world which it can only know as an object, something external and alien to the self. And this reflective, analytical tendency is something which has been intensified beyond the ordinary in Hyperion by his stay in Germany and his introduction to the desert of philosophical speculation, divorced from its primal sources: 'Amongst you I became so very rational, learnt to distinguish myself perfectly from what is around me, and now I'm set apart in the beautiful world, expelled from the garden of nature in which I grew and bloomed, and shrivel under the noonday sun' (p. 9).68 Here we see again how crucial the German experience is for Hyperion. For it brings into sharp focus both the tendencies which he shares with, and also inherits from Alabanda and Diotima. Germany gives him nature in all its glorious beauty and vitality, inviting ecstatic communion and total immersion. It also aggravates his propensity towards elitist detachment, both by providing him with a soulless and horribly fragmented society to despise, and by schooling that faculty within him which automatically destroys feelings of oneness and unity, since it is dependent for its operation on the antithetical opposition of subject and object. The dilemma, which is implied at the end of the novel by the juxtaposition of the attack on the Germans and the hymn to nature, is here, at the beginning of the novel stated in explicit terms, and there would appear to be no way out. To be in everything and above everything at the same time, consciously to grasp the ineffable joys of self-abandonment; to retain and preserve them beyond the moment of the experience itself, so that they could inform and enhance our everyday lives, this would indeed seem to be a task beyond mere mortals. And yet it is the statement of the dilemma which itself incorporates the promise of its own solution.
<sup>68</sup> Letter II: 'Ich bin bei euch so recht vernünftig geworden, habe gründlich mich unterscheiden gelernt von dem, was mich umgiebt, bin nun vereinzelt in der schönen Welt, bin so ausgeworfen aus dem Garten der Natur, wo ich wuchs und blühte, und vertrokne an der Mittagssonne' (StA III, 9).
#### '… return whence he came'
It should again be stressed that the task undertaken by the narrator is both hazardous and necessary. It is Hyperion's self which is to be rescued, since it cannot survive in a state of constant oscillation between extremes of dissolution and alienation. Madness threatens, either through loss of rational control, or through autistic isolation. These states are characterized, in their extreme form, by the absence of relationship, since relationship presupposes both cohesion and distinction, and cannot exist where identity or division are absolute. The integrity of the self can be achieved and maintained, only if it can comprehend itself in relation to the whole of which it is part, in a state of distinction without division. For Hyperion this requires a mode of reflection which must be more than merely an articulation of the dilemma (although this is certainly a start). If the problem is that the ecstatic union with nature occurs outside the limits of consciousness and is not amenable to rational analysis, then the only way in which it can be comprehended, integrated into the temporal experience of the individual self, is as something absent, something lost. It must be supplied with a framework, a context in time. It must be given a history.
It is a characteristic feature of some of Hölderlin's greatest hymns that the glorious and overpowering epiphany of an unspecified deity is followed by a gesture of withdrawal on the part of the poet.69 The unmediated presence of the deity cannot be sustained, nor can his nature be grasped, except by means of a distancing procedure that can involve a survey of the god's mediated presence (incarnation) and subsequent absence in human history. What is experienced as immediately near cannot be grasped, and yet it must be grasped if its intensity is not to overwhelm. Undifferentiated divinity is consuming fire. For the narrating Hyperion, as we have seen, the 'living image' of Diotima assumes the same threatening quality as these numinous presences. And the reaction is similar. The detachment required if the mind is not to break would ideally be effected by transposition into temporal distance: 'I must make believe she lived in times of old, that I knew
<sup>69</sup> See my essay, 'Meaning in History: "Chiliasm" in Hölderlin's "Patmos"', *Colloquia Germanica*, 11 (1978), 19–52, pp. 41–42.
of her through tales' (p. 51).70 Whether he is confronting the 'terrible glory of antiquity' (p. 16),71 Diotima, or his own beatific experiences of all-unity, the necessary combination of involvement and detachment, being 'in' and 'above', can be realized only by the development of an integrating historical consciousness. Memories must be both evoked and interpreted as meaningful. The constant oscillations must be seen to form part of a progression. It is only in memory that the contradictions can be accommodated and reconciled. And Mnemosyne is, of course, the mother of the Muses.
Innocence is never known to itself, and awareness of it can only ever be as something lost, in the recollection of the consciousness that destroys it. Golden ages are always in the past, and this is true both of the individual and the race. Hölderlin's Christ departs in order that he should be preserved and understood in man's memory, and it is 'remembrance' of him that is celebrated in the Eucharist. The perfect mystic vision of Plato's divinely-possessed philosopher is a product of anamnesis, the sacred recollection of the Ideas: his soul regains its wings because it is 'always dwelling in memory as best it may upon those things which a god owes his divinity to dwelling upon.'72 It is in the contemplation of beauty that the recollection is awakened. For the Hölderlin who wrote 'Menons Klagen um Diotima' ('Menon's Lament for Diotima'), 'Andenken' ('Remembrance'), and 'Mnemosyne', it is certainly true that 'recollection had come to symbolize … the poet's link with the Absolute.'73 That Hölderlin is indeed a writer of elegiac temper is not to be denied, and it is clearly significant that he confers upon the protagonist of his novel a pronounced 'elegischer Karakter'. But it ought also to be clear that there is much more to this than mere wistful nostalgia for a vanished fantasy. The source of the sadness is intoxicating joy, grief and sorrow the medium through which this joy is articulated: 'Many have tried, but in vain, with joy to express the most joyful; / Here at last, in grave sadness, wholly I find it expressed.'74 Harmony requires
<sup>70</sup> See above, n. 14.
<sup>71</sup> Letter V: 'die schrökende Herrlichkeit des Altertums' (StA III, 18).
<sup>72</sup> Plato, *Phaedrus*, 249c (translation by Walter Hamilton).
<sup>73</sup> David H. Miles, 'The Past as Future: "Pfad" and "Bahn" as Images of Temporal Conflict in Hölderlin', *Germanic Review*, 46 (1971), 95–118, p. 115.
<sup>74 &#</sup>x27;Sophokles' (c. 1799) ('Viele versuchten umsonst, das Freudigste freudig zu sagen / Hier spricht endlich es mir, hier in der Trauer sich aus' (StA I, 305). Translation by Michael Hamburger, in *Hyperion and Selected Poems*, ed. Santner, p. 139. Note that
dissonance and discord in order that it may be expressed at all. In the letter to his half-brother, quoted above (p. 159, n. 49), Hölderlin writes 'that the highest power is in its expression also the most modest and that the divine, when it makes itself manifest, can never be without a certain sadness and humility?' It is this, it will be noted, *in its expression*. 'There's reconciliation in the middle of strife', Hyperion writes at the end.75 Unity is expressed through differentiation, the latter being the necessary image or echo of the former, the only way in which this is capable of being reproduced. The source of poetry is the experience of primal unity, its function to give it expression. But the act of articulation itself requires both the detachment of the poet and the prismatic division of that which is being expressed. In this sense it will not be surprising if the poet and his work display 'an elegiac character'.
Achim von Arnim called *Hyperion* an elegy,76 and clearly it does have some of the characteristics conventionally associated with the word. The fulfilment enjoyed in happier times weighs heavily on Hyperion, both in the form of his own past experience and also that of the longdeparted glory of his fatherland. Modern Greece indeed represents an appropriate choice of setting for a character such as Hyperion. (Given the potentially subversive nature of the subject matter, one cannot help feeling that Hölderlin would have been running something of a risk if he had set the novel anywhere near Germany.) The transience of a great culture, the ephemeral nature of human happiness, the fleeting quality of moments of fulfilment, these are things which Hyperion finds very difficult to come to terms with. The particular narrative technique adopted in the novel is uniquely appropriate to a character of this kind. It enables him to indulge a sense of loss and vast regret, mournfully to reflect on the passing of things of great beauty even as he conjures them before our eyes. Nothing lasts. To say that in itself takes time, and we are made to feel it passing whilst Hyperion writes. Transience is the real challenge Hyperion has to face, and it seems that the task laid upon him is to accept and affirm it, and also in doing so evolve a strategy for defeating it. The problem is a human one and its solution is poetic. There is therefore nothing arbitrary about the elegiac character of the
Hamburger is constrained by the metrical demands of the elegiac couplet, hence his 'grave sadness' for 'Trauer' ('sadness'/'grief').
<sup>75</sup> Letter LX: 'Versöhnung ist mitten im Streit' (StA III, 160).
<sup>76</sup> See StA III, 319.
novel and its central figure. It corresponds perfectly to Hölderlin's convictions about the nature of life and poetry. And lest we be inclined to dismiss these as merely eccentric, it must be pointed out that he by no means stands alone. In the peculiar significance that he attaches to recollection and memory Hölderlin could be said to articulate, arguably in its purest form, one of the most characteristic tendencies of his age. For his exact contemporary Wordsworth 'poetry takes its origin from emotion recollected in tranquillity', and it has been suggested that the formulation might owe something (via Coleridge) to Schiller.77
It is indeed to Wordsworth's *Prelude* that some of the most perceptive critics are inclined to compare *Hyperion*. One of the more obvious affinities is the circular structure of both works. As M. H. Abrams observes of *The Prelude*: 'Its structural end is its own beginning; and its temporal beginning … is Wordsworth's entrance on the stage of his life at which it ends.'78 In both works it is the evolving consciousness of the narrator that conditions the structure, and produces the combination of retrospection and progression. 'Wordsworth does not tell his life as a simple narrative in past time but as the present remembrance of things past in which form and sensation "throw back our life" and evoke the former self in a multiple awareness that Wordsworth calls "two consciousnesses".'79 Could it not also be said that *Hyperion*, too, constitutes the 'prelude' to itself, in that it embodies an account of its own genesis? For the purposes of the argument it is necessary to suppose that it is Hyperion's novel (rather than Hölderlin's), and that we take the concluding pages to be indeed the climax of the work. And this would seem to be the great virtue of the comparison, as presented by Cyrus Hamlin in a very fine, if neglected essay, that it rescues and reinforces the hymn to nature at the end of the novel. As Hamlin writes: 'For both poets the act of recollection or remembrance
<sup>77</sup> Wordsworth, Preface to the *Lyrical Ballads* (1800) — see *Schiller*: *On the Aesthetic Education of Man in a Series of Letters*, edited and translated by Elizabeth M. Wilkinson and L. A. Willoughby (Oxford: Clarendon Press, 1968), p. clxvii; also Michael John Kooy, *Coleridge*, *Schiller and Aesthetic Education* (London: Palgrave Macmillan, 2002), pp. 36–37. The relevant passage is from Schiller's review of Bürger's poems: 'aus der sanftern und fernenden Erinnerung mag er dichten', translated in Kooy as: 'from tranquil and distanced recollection is poetry made'.
<sup>78</sup> M. H. Abrams, *Natural Supernaturalism*: *Tradition and Revolution in Romantic Literature* (London: Oxford University Press, 1971), p. 79.
<sup>79</sup> Abrams, p. 75.
is identical with the poetic process. Poetry may thus be understood as a retrospect by the poetic self upon itself and as the recreation of moments of experience which are beyond understanding.'80 In the case of Hyperion, who is or becomes the poet here, the experience which is being re-created, that from which the writing proceeds and in which it ends, is the experience of all-unity in the German spring. Hamlin argues that:
the entire experience as presented to us in the final letter of the novel serves as a paradigm for the poetic process itself, as it applies both to the composition of *Hyperion* and to Hölderlin's subsequent practice as a poet. The origin of poetry is the experience of ecstasy or inspiration, here represented as a visitation by the spirit of Diotima, Hyperion's Muse. Only for such experience is the poetic vision, or intuition, authentic. And the motive for writing the poem is found in the immediate response to this experience, as indicated with regard to the novel by the fact that Hyperion apparently returned to Greece immediately after the experience of epiphany in order to start writing his letters to Bellarmin … The fact that the novel concludes with the narration of this event in Hyperion's life suggests that here, in comprehending this experience, the novel achieves fulfilment and completion.81
At first sight this might seem to be open to a number of objections. Firstly, as has been argued, the experience of all-unity here is not unique, except perhaps in intensity, for it proves subsequently to be repeatable. Moreover, as I have also argued, such experience serves, precisely because of its fleeting nature, rather to exacerbate Hyperion's problems than to solve them. And furthermore, how is one to account for the apparently relativizing inverted commas around the concluding passage, and the narrator's distancing 'So I thought'? Nevertheless, it seems to me that Hamlin's insights here are valid. The only qualification I would perhaps apply to his argument is that, as formulated here, it might appear to be suggesting that Hyperion returns home specifically to write his letters and recapture his experience. Now although this is in effect what happens, the conscious motivation is something that grows in the course of writing and is hardly present at all in the initial letters, where Hyperion has to be prodded into communication in the
<sup>80</sup> Cyrus Hamlin, 'The Poetics of Self-consciousness in European Romanticism: Hölderlin's *Hyperion* and Wordsworth's *Prelude'*, *Genre*, 6 (1973), 142–77.
<sup>81</sup> Hamlin, p. 152.
first place, and, once embarked on it, still has formidable obstacles to overcome. Insight into the true nature of his activity comes very late indeed.
But what of the above-mentioned objections? An obvious point to be made is that the first and decisive experience of all-unity is the last to be related, and therefore the first to be fully comprehended. The experiences of which Hyperion writes in the second letter are, for him, at the time of writing, elements of an unresolved and apparently unresolvable tension which is driving him to distraction. By the time he reaches that stage in his narrative when the first experience occurs, he knows what he is about, understands the experience and its meaning, and can present it in the form of a solution, not a problem. This is accomplished by a subtle fusing of perspectives that takes place in the last letter.
The experience related in that letter is, like all experience *within Hyperion's narrative*, recounted, recollected experience. It is something he had then: 'so I thought,' he says, not once but twice.82 It is not, of course, the first time he has recounted his past, as opposed to his present, thoughts; though nowhere else does he use quotation marks for anything other than direct speech, and even then only when there is no accompanying saying verb and to distinguish speakers from one another.83 Admittedly, he did speak at the time, or thinks he did, but has no recollection of what he might have said: 'and words I spoke, it seems to me, but they were like the fire's rush when up it flares and leaves behind the ashes' (p. 136).84 Curiously, this is then immediately followed by the passage in question, in which we appear to have an exact reconstruction of the thoughts behind the hypothetical words — set off from the rest of the text by speech marks at the beginning of each line.85 Yet the thoughts, as we have them, are not known to have been
<sup>82</sup> Apart from the last line of the novel, the phrase occurs at the beginning of the passage in quotation marks: 'O du, so dacht' ich, mit deinen Göttern, Natur! ich hab' ihn ausgeträumt, von Menschendingen den Traum' (StA III, 159); '"O you," so I thought, "with your gods, nature! I've dreamed it out, the dream of human things"' (p. 136).
<sup>83</sup> For more on Hölderlin's use of quotation marks, and the procedure adopted in the translation, see below p. 191.
<sup>84</sup> Letter LX: 'und Worte sprach ich, wie mir dünkt, aber sie waren wie des Feuers Rauschen, wenn es auffliegt und die Asche hinter sich läßt' (StA III, 159).
<sup>85</sup> Forty-five of them in the original Cotta edition, all exclusively on the left. There are no concluding quotation marks at the end of the 'speech', though Beissner supplies them: StA III, 160.
uttered at the time of the experience itself, nor can they even be said to have been consciously formulated in Hyperion's mind at that time. They would appear then to be the articulation of (re)creative reflection on the part of the narrator. Hamlin argues, plausibly, that what we are in fact presented with here is a purposeful confusion introduced by Hölderlin between the experience of the character Hyperion and his subsequent narration of it: 'The confusion may be regarded as an attempt to fuse the two perspectives. The speech is certainly the product of creative reflection … but it is also a fully accurate recreation of the hero's revery in response to his visitation …'86 In this way the circle is closed. The two perspectives come together as a paradigm of the poetic process itself, 'which originates in ecstasy and concludes in reflective self-understanding.'87
Yet we still have to confront the question as to why, having ingeniously closed the circle (the final image in the novel is one of circulation), Hölderlin then, disconcertingly, opens it again in the novel's last line. Let it once more be stressed that this is in no way intended to prejudice the authority of the vision itself. The glorious affirmation of Life is not to be dismissed as 'relatively immature' (compared to the reflections that follow the transcription of Notara's account of Diotima's death).88 That is to say, it may have been at the time of the initial experience, but is no longer so in its present formulation. The distinction is crucial. Whatever we make of the final passage, there can be no doubt that the last letter does voice a critical attitude towards the hermit's self-indulgent self-abandonment. 'Thus I gave myself up more and more to blissful nature,' writes the mature Hyperion, 'and almost too endlessly. How gladly I'd have become a child to be closer to her!'89 Here the (almost) total immersion in nature is clearly associated with an attitude which, from the point of view of the mature narrator, must
<sup>86</sup> Hamlin, p. 152.
<sup>87</sup> Hamlin, review of Friedbert Aspetsberger, *Welteinheit und epische Gestaltung*: *Studien zur Ichform von Hölderlins Roman 'Hyperion'* (Munich: Wilhelm Fink Verlag, 1971), in *Journal of English and Germanic Philology*, 73 (1974), 292–96, p. 295.
<sup>88 &#</sup>x27;My dearest friend! I'm at peace …' (p. 129); cf. also above, p. 159. For Ulrich Gaier, in 'Hölderlins "Hyperion": Compendium, Roman, Rede', *Hölderlin-Jahrbuch*, 21 (1978/79), 88–143, p. 108, n. 66, the concluding hymn to nature is 'vergleichsweise leichtsinnig'.
<sup>89 &#</sup>x27;So gab ich mehr und mehr der seeligen Natur mich hin und fast zu endlos. Wär' ich so gerne doch zum Kinde geworden, um ihr näher zu seyn!' (StA III, 158).
be regarded as regressive. For the naïve, un(self)conscious innocence of childhood, which is longed for by the hermit and which is still lovingly and nostalgically celebrated in the third letter, is not only recognized by the narrator as something which is not realistically available to him, but is actually explicitly rejected as inadequate, even as an ideal. After all, he has written only a few pages earlier (p. 129): 'But the bliss that does not suffer is sleep, and without death there is no life. Should you be, like a child, eternal and slumber as nothingness? forego the victory? not pass through all the perfections?'90 Yet in the final letter the narrator who proclaims this conviction is confronted with a former self who would like nothing better than to sink into oblivious communion with nature, abjuring his fellow man, himself and all he knows and has learnt, and in fact is on the point of doing just that, losing himself in ecstatic, spontaneous babbling, ephemeral froth without substance. This is not to deny the reality of the experience while it lasts. The point is that it cannot last. Because it is so totally spontaneous, and above or beyond or below rational comprehension, it leaves nothing behind but a sense of loss and despair. It is only in the interpreting recollection of the narrator that the experience acquires structure and content; and that only at the end of a long process of reflection which produces in Hyperion a comprehensive sense of himself and the meaning of his life. The thoughts with which Hyperion is credited in this concluding passage are valid only if taken in conjunction with all that precedes them. They are valid as the culmination of a series of experiences and reflections on those experiences, and indeed reflections on the reflections. They are of limited validity as the spontaneous outpouring of the hermit. The Hyperion who has the experience has turned his back on his fellows and relinquished self-control. He is not even in a position to preserve his ecstatic intuitions for himself, still less communicate them to others. He is 'in' his experience to such a degree that he can no more grasp it than the aether. Yet the experience will have lasting value for himself and others only if it can be given some form of articulation, and this the hermit is patently incapable of doing. Even if he is already at this stage in a position to sense the unity of all things, including the positive value
<sup>90</sup> Letter LVIII: 'Aber die Wonne, die nicht leidet, ist Schlaf, und ohne Tod ist kein Leben. Solltest du ewig seyn, wie ein Kind und schlummern, dem Nichts gleich? den Sieg entbehren? nicht die Vollendungen alle durchlaufen?' (StA III, 150).
of pain and suffering, the conviction as yet has no sustaining power; it belongs to a passing phase, another swing of the pendulum. The activity of the narrator is not pure cerebration, but ultimately practical and sustained demonstration of a truth which emerges 'in eternal, indestructible glory' only in the narrator's own creation, as a result of that creation.91 It's important therefore that the provisional and inchoate nature of the hermit's bliss be indicated, and the reader referred back (and forward) to the next stage in Hyperion's development, which begins with the writing of the letters. Hence the justification for the 'So I thought. More anon' which obliges us to ask what could follow in Hyperion's narrative, the answer being that he returned to Greece and started to write the letters we have just been reading.
At the end of the first letter Hyperion writes (addressing his own heart): 'just forget that men exist … and return whence you came, into the arms of nature' (p. 8).92 This is echoed by the motto from *Oedipus Coloneus* with which Hölderlin prefaces the second volume of the novel and which might be translated: 'Not to be born is, past all prizing, best. But when a man has seen the light of day, this is the next best by far, that with the utmost speed he should go back whence he came.' The echo makes it seem likely that Hölderlin uses the motto for his own purposes, so that it need not necessarily strike the pessimistic note which it has in its Sophoclean context (though that might be in some measure appropriate for the nature of the events to be narrated). Certainly, death and reunion with nature may be seen as equivalents, witness Diotima. But we would also do well to note the final sentence of Hyperion's hymnic vision: 'The arteries part and return in the heart and one eternal glowing life is All' (p. 137). It is hardly fortuitous that the novel ends with an image of circulation (of the bloodstream), nor that in the German it is a metrical line whose catalectic final foot is completed by being joined to the first.93 The motto from Sophocles
<sup>91</sup> Letter XIII: 'in ewiger unzerstörbarer Klarheit' (StA III, 51). For the rendering of 'Klarheit' as 'glory', see below p. 200.
<sup>92 &#</sup>x27;vergiß nur, daß es Menschen giebt … und kehre wieder dahin, wo du ausgiengst, in die Arme der Natur' (StA III, 8).
<sup>93 &#</sup>x27;Es scheiden und kehren im Herzen die Adern und einiges, ewiges, glühendes Leben ist Alles' (StA III, 160). As Hans Gerhard Steimer notes in his excellent 'Hölderlins Klauseln', *Hölderlin-Jahrbuch*, 31 (1998/99), 281–328, p. 282, after the unstressed opening syllable, there is a sequence of eight dactyls in a row followed at the end by a trochee which, if joined up to the first syllable of the sentence, yields a ninth
can be seen as a hint to the reader not to ignore the circular structure of the novel. The structure of *Hyperion* is circular and at the same time open. And if it were not open, it would not be obviously circular. The narrator's concluding comment ensures that it is both. For if the fusion of the two perspectives at the end were too perfect, we would be in danger of taking the end to be the end, and might well miss the second line of development altogether. The snake must be *seen* to have its tail in its mouth. Yet we are dealing here with a circularity that accommodates progression (which is why critics can talk of the novel as having a spiral structure). The progression is dependent on the circularity. As linear development it is incomplete, and there is a risk that the ending will be seen as an ultimate solution, a detachable end-product (the novel's message) which transforms all that precedes it into redundant scaffolding. Coleridge claims 'the common end of all *narrative*, nay of *all* Poems, is to convert a *series* into a *Whole'*. 94 And this is precisely what Hölderlin (or perhaps one should say, Hyperion) has done here. In interpretative recollection what is initially a disjointed series is transformed into a whole, and the truth is the whole. The circle and the line have to be combined. For what is purely cyclical lacks progression, and what is purely linear lacks wholeness. At the risk of appearing wilfully paradoxical, one could add that the linear sequence of experiences, the raw material of Hyperion's narrative, betrays a seasonal, cyclical rhythm (from joy to despair), and that it only achieves retrospective linearity, in the sense of meaningful progression, through the adoption of the circular narrative procedure whereby end is joined to beginning. At the same time the circular route to understanding issues in a definitive advance, so that the mature narrator ends on an ascending course, where the hero of the narrative is still trapped within the cycle of bliss and despair.
The way out and the way forward for Hyperion has been found in writing. Not that his writing is obviously therapeutic, at least initially. He begins by giving vent to his despair and then, at the instigation of Bellarmin, proceeds with justifiable reluctance to recall the harrowing events of his life, at the same time recording his present reaction to
dactyl. I cannot believe this is not intentional. (Nor have I found any satisfactory way of replicating it in English.)
<sup>94</sup> To Joseph Cottle, 1807 (Letter 135); Abrams, p. 271.
them. By the beginning of the second book of the first volume he begins to suspect that there may be a hidden pattern, and therefore meaning, to his experiences, though for the time being it remains hidden.95 At the same time there are indications that his state of mind is becoming more stable. He begins to employ oxymoronic expressions, suggesting an emergent ability to embrace polarities, rather than being thrown off balance by them.96 The second book of the first volume proves to be the most severe test of all for the narrating Hyperion, since it is here that he is forced to confront the most blissful experience of his past life, the blossoming of his love for Diotima, knowing all the time what the fate of the relationship is going to be. The strain drives the narrator to the brink. Once this ordeal has been successfully sustained, there seems to be a steady advance towards tranquillity and serenity of vision. The gap between Hyperion as narrator and as the central figure of his own narrative widens, almost to the point where his ability to empathize with his former self is challenged. The straightforward transcription of the lovers' correspondence, with very little in the way of comment, might indicate this. It seems to me questionable to speculate, as has been argued, that the division of Diotima's swansong into instalments should be seen as suggesting that Hyperion is periodically overcome as he copies it out. The 'continuations' are surely to be taken as those of the dying Diotima, indicating resumption of writing, and telling evidence of her physical weariness and exhaustion. But it is of course Hyperion who does the transcribing, so that even here he is rewriting his life. And by the time he copies out the swansong, he realizes what has been happening. It is not simply that he has grown calmer by learning to accept his fate. He has made sense of it in the only way open to him, by fulfilling the mission assigned to him by Diotima. 'You would have to go under,' she writes:
<sup>95</sup> See p. 41: 'Or I look out upon the sea and ponder my life, its rising and sinking, its bliss and its grief, and my past often sounds to me like a lyre when the master runs through all the tones, throwing together discord and harmony in hidden order.' Letter XII: 'Oder schau' ich auf's Meer hinaus und überdenke mein Leben, sein Steigen und Sinken, seine Seeligkeit und seine Trauer und meine Vergangenheit lautet mir oft, wie ein Saitenspiel, wo der Meister alle Töne durchläuft, und Streit und Einklang mit verborgener Ordnung untereinanderwirft' (StA III, 47).
<sup>96</sup> Cf. p. 42: 'in smiling earnest', 'like the starry heavens I'm both still and moved'. Letter XIII: 'mit lächelndem Ernste', 'wie der Sternenhimmel, bin ich still und bewegt' (StA III, 48).
you'd have to despair, yet the spirit will save you. No laurel will comfort you, and no wreath of myrtle; Olympus will, the living and present, that eternally youthful blossoms around all your senses. The beautiful world is my97 Olympus; in this you will live, and with the holy beings of the world, the gods of nature, with them you will be joyful. …
Grieving youth! soon, soon you will be happier. Your laurel failed to ripen and your myrtles faded, for priest you shall be of divine nature, and your poetic days burgeon already. Oh, could I only see you in your future beauty! Farewell. (pp. 126, 128)98
The double reference to myrtles and laurels should, as Ryan has persuasively argued, be taken as alluding to Hyperion's activity as a lover and a man of action.99 With the death of Diotima and the failure of the insurrection, fulfilment is denied to him in both these spheres. Diotima's consolation for him is that his true vocation lies elsewhere. And what she has to say about it must be seen in conjunction with her earlier vision of his mission at the end of the first volume, when the lovers contemplate the ruins of Athens. There Diotima exhorts him to become the 'educator' of his people. Hyperion appears excited by the prospect, and what on that occasion he envisages as the goal of his future efforts emerges clearly enough in his recorded thoughts, which conclude the volume and are offered without comment by the narrator, suggesting that the ideal they celebrate remains valid:
Do you ask after men, nature? Do you lament like a lyre which the brother of chance, the wind alone plays, since the artist who kept it in tune is dead? They will come, your men, nature! A rejuvenated people will rejuvenate you too, and you will become as its bride, and the ancient covenant of spirits will renew itself with you.
<sup>97</sup> Not wishing to follow some editors in assuming a misprint in the original edition here (*mein* for *dein*), I have opted for 'my' rather than 'your'. For the reasons, see below, p. 190.
<sup>98</sup> Letter LVIII: 'Du müßtest untergehn, verzweifeln müßtest du, doch wird der Geist dich retten. Dich wird kein Lorbeer trösten und kein Myrthenkranz; der Olymp wirds, der lebendige, gegenwärtige, der ewig jugendlich um alle Sinne dir blüht. Die schöne Welt ist dein Olymp; in diesem wirst du leben, und mit den heiligen Wesen der Welt, mit den Göttern der Natur, mit diesen wirst du freudig seyn. … Trauernder Jüngling! bald, bald wirst du glüklicher seyn. Dir ist dein Lorbeer nicht gereift und deine Myrthen verblühten, denn Priester sollst du seyn der göttlichen Natur, und die dichterischen Tage keimen dir schon. O könnt' ich dich sehn in deiner künftigen Schöne! Lebe wohl' (StA III, 147, 149).
<sup>99</sup> Ryan, *Exzentrische Bahn und Dichterberuf*, pp. 198–99.
There will be but one beauty; and humanity and nature will unite into one all-embracing godhead. (p. 77)100
In whatever way this rejuvenation of his people, this higher synthesis of mankind and nature is to be achieved, it will clearly be the result of public activity. It will not be brought about by a complete retreat into silent or inarticulate communion with nature, and this is not the way in which Diotima's final words should be construed. Becoming a priest of nature will entail something more than transient and subjective experiences of total absorption, though it may be conditional on an ability to have them. The experience of all-unity must be rendered capable of mediation. It must be communicated. As that other Diotima tells Socrates in *The Symposium*, the object of love is not simply beauty; its object is to 'procreate and bring forth in beauty'.101 Hyperion must be active and not merely passively receptive to the gods of nature. The 'spirit' which is to rescue him must, in one sense at least, be his own. In the words of the woman of Mantinea, Diotima could be said to hope for Hyperion that 'having his eyes fixed on beauty in the widest sense, he may,' with her death, 'no longer be the slave of … devotion to an individual example of beauty … but by gazing upon the vast ocean of beauty to which his attention is now turned, may bring forth in the abundance of his love of wisdom many beautiful and magnificent statements and ideas …'102 Hyperion will indeed find solace in nature, but only if the experience engages his understanding, and above all, issues in creative activity. Priests mediate. The days that Diotima prophesies for Hyperion will be 'poetic' ('dichterisch'), and I take the adjective to have both a subjective and an objective sense. That is to say, Hyperion will both experience and communicate poetically. He will have been made capable of appreciating universal beauty, Diotima's 'Olympus'. And he will also be able to articulate his perceptions, sharing
<sup>100</sup> Letter XXX: 'Du frägst nach Menschen, Natur? Du klagst, wie ein Saitenspiel, worauf des Zufalls Bruder, der Wind, nur spielt, weil der Künstler, der es ordnete, gestorben ist? Sie werden kommen, deine Menschen, Natur! Ein verjüngtes Volk wird dich auch wieder verjüngen, und du wirst werden, wie seine Braut und der alte Bund der Geister wird sich erneuen mit dir. // Es wird nur Eine Schönheit seyn; und Menschheit und Natur wird sich vereinen in Eine allumfassende Gottheit' (StA III, 90).
<sup>101</sup> Plato, *The Symposium*, 206e (translation by Walter Hamilton).
<sup>102</sup> *Symposium*, 210e.
and mediating his experience of a beautiful world, and developing a similar response in his fellows.
Clearly if the narrating Hyperion is to accept his vocation, he is obliged to emphasize his detachment from the incipient hermit at the end of the narrative. Priests are not hermits. They may stand apart, but they are not permanently turned away from the communities they serve. What they receive, they transmit. Their unworldliness, their apparent aloofness from human affairs is justified as a necessary concomitant of their mission, since it is only by remaining detached from the world that they are able to serve and influence it. Their detachment is thus an expression of their commitment to mankind, and only as such is it valid. As an end in itself it has no value, issuing merely in social sterility and impotence. The hermit's isolation is a denial of humanity, and in rejecting his fellowship with others, he is condemning himself. Literal perdition, loss of self and soul, awaits those who would abandon humanity; their speech is without echo, their gratification chimerical (for they lack a self to fulfil), 'and they go under in their solitude.'103 One might think here of the vision of Tree of Life in Hyperion's final speech, where he sees himself at the top and the rest of humanity down amongst the roots — but it is still the same tree. The distance from ordinary humanity implied in Hyperion's words still holds good (and indeed receives its justification, like everything else in the hymn to nature) when they cease to be mere intuitions and become articulated speech. The difference is that now they are communicated, so that the isolation is lifted and relationship established. The hermit addresses only nature. The poet-priest is addressing men.
The hermit's retreat from the world is an extreme reaction to the frustrations of excessive engagement with it. Either way, he would seem to have deprived himself of the possibility of 'proper effectivity'.104 As a hermit Hyperion has turned his back on his fellows, and his beatific experience of nature remains subjective, private and uncommunicated. And he pays the penalty for this. He loses himself in the experience, and is left with nothing. Because the experience is not articulated, it is not preserved. But then, on Hyperion's return to Greece, the recovery
<sup>103</sup> The last of 'Seven Maxims' — see Adler's translation in *Hölderlin*: *Essays and Letters*, p. 243; 'und Sie gehen in ihrer Einsamkeit unter' (StA IV (1), 247).
<sup>104 &#</sup>x27;rechte Wirksamkeit' — see above, n. 103.
begins. It is a process of recovery in that an anguished soul gains peace of mind; in the sense that the self comes to itself and establishes its identity through reflective recollection; and also in the related sense that lost experiences of beauty are recreated in such a way that they acquire permanence and objective validity. It is in the communication of such experience that Hyperion finds both his mission and his salvation. It is in and through his writing that he develops the embracing consciousness, of himself and the world, that constitutes the resolution of the dissonances. And it is only in creative writing ('Dichtung') that such consciousness could possibly be fostered and expressed. Who but a poet could be both in everything and above everything? And how else could he accomplish this but through the medium of his own work? The simultaneity of experience and reflection, involvement and detachment, is available only to the mediating consciousness of the poet. It is difficult to see how it could be achieved directly, in life, for we are not gods. But it can be created and sustained in the work itself, which is the product of poetic recollection. Discrete experiences can be reconstituted and held together in the interpretative, reflective consciousness of the poet, whose primal impulse is the sense of the unity and cohesion of all life. If the essence of beauty is 'εν διαϕερον εαυτω' (pp. 70, 71), the poetic work, reflecting this, will be a differentiated, an 'organized', articulated unity, a linguistic incarnation.105 And like God in relation to the world, the poet will be both in and above his creation. And this is nowhere better demonstrated than in the last letter of *Hyperion* where the narrator contrives to empathize with an ecstatic experience of his former self, conveying it to us from the inside, as it were, and at the same time makes it clear that he is presenting us with a retrospective articulation of something beyond words. And moreover, implied within the act of articulation itself is criticism of the self-abandonment which was a pre-condition of the initial experience. Intense involvement is combined with clear-headed detachment, and the latter is emphasized by the narrator's concluding comment, which brings us back to his own present, refers us to his past, and anticipates his future. Hyperion is in control. He is not wholly contained in the work he has just created, but is
<sup>105</sup> Letter XXX: 'the One differentiated in itself'; Heraclitus, rendered by Hyperion as 'das Eine in sich selber unterschiedne'. Cf. also Abrams, p. 260, where the allusion is to Blake.
poised to continue writing, presumably on the basis of the combination of experience and insight, and insight into that combination, which he has evidenced in his letters. Thus Hyperion's novel, no less than *The Prelude*, 'incorporates the discovery of its own *ars poetica*'.106 In its turn it promises to be the prelude to future works. 'More anon.'
If the arguments advanced above are valid, *Hyperion* deserves pride of place in any list of 'self-begetting'novels. The type is defined by Steven Kellman as follows:
Truly *samizdat* in the original sense of self-publishing, it is an account, usually first-person, of the development of a character to the point at which he is able to take up his pen and compose the novel we have just finished reading. Like an infinite recession of Chinese boxes, the selfbegetting novel begins again where it ends. Once we have concluded the central protagonist's story of his own sentimental education, we must return to page one to commence in a novel way the product of that process … We are at once confronted with both process and product, quest and goal, parent and child.107
Far-fetched as it may seem, Hölderlin might even be said to anticipate Proust by linking 'within a single book the story of the spiritual genesis of a vocation, the story of the genesis of a work, and the very embodiment of that work.'108 It should not, however, be necessary to argue Hölderlin's modernity here. For what he offers us in his novel is a remarkable practical realization of the Romantic aspirations of his own generation. If we think of Romanticism in terms of the poetics of self-consciousness, we are confronted here with a work which, as narrative, contains its own built-in commentary and critique; which is about the growth of comprehensive self-awareness in its narrator as he narrates; and which turns out to be writing about writing in that it culminates in the narrator's recognition of his poetic vocation, realized both practically and theoretically in the work he has just written. It is more surely a novel of poetic initiation than Novalis' *Heinrich von Ofterdingen* (1802). Hölderlin has his poet initiated by making him the
<sup>106</sup> Abrams, p. 78.
<sup>107</sup> Steven Kellman, *The Self-Begetting Novel* (London: Macmillan, 1980), p. 3. This stimulating study is based largely on French models, and gives no consideration to *Hyperion* and — as indicated by the absence of Hoffmann, Jean Paul, Raabe and others from the index — precious little to a rich German tradition of self-conscious fiction.
<sup>108</sup> Kellman, p. 27, quoting Germaine Brée on *À la recherche du temps perdu*.
writer of the work that constitutes his initiation. The writing is therefore self-contained and self-justifying. At the same time, it embodies a statement about the function of poetry ('Dichtung') and the vital role of the poet in society, exemplifying the aesthetic evangelism which is such a characteristic feature of the age. Poetry is redemptive — has it not just saved Hyperion? The poet is the teacher of his people, and Hyperion's first pedagogical act as an (almost) fully-fledged poet is to berate Bellarmin's, and of course Hölderlin's people, like a fiery Hebrew prophet. The *disjecta membra* of German society are confronted with a gospel of reunification through beauty. The roots of division are conveyed in terms of familiar polarities: subject and object, self and world, individuation and unity, thinking and feeling. And the resolution is seen to be accomplished in the aesthetic act itself. Bearing in mind the observations above on the linearity and circularity of the novel, it could be argued that in *Hyperion* we have perhaps the most perfectly achieved example of 'progressive universal poetry', to use the terms of Friedrich Schlegel's famous Romantic programme (1798).109 That the combination of involvement and detachment here does not yield irony in the conventional sense will be readily acknowledged, though it in no way vitiates the argument. The dialectical tension between the narrator and the subject of his narrative (the feelings, thoughts and actions of his former self) is expressed not only by the device of juxtaposing narrative and distancing comment, but also quite remarkably by fusing them and making the same words serve both perspectives, in the consummating vision at the end of the novel.
Hölderlin's obsession with his own art has, however, always to be seen in the context of his devotion to nature. His thorough-going pantheism survives the speculative acrobatics of Kantian and post-Kantian transcendentalism, and he is never really tempted to find beauty solely in the eye of the beholder or interpret nature as alienated ego. The problem is rather that man is alienated nature and has to be taught to recognize the primacy of the source of his own being. This is accomplished by accepting the alienation itself as a necessary and divinely-sanctioned differentiation of primal oneness, a prismatic division whose function is to articulate the unnamed and unknowable in quintessential form. The division itself cannot be reversed, except
<sup>109</sup> Friedrich Schlegel, 'Athenäums-Fragment', No. 116.
in death (or self-transcending peak experiences of short duration and variable degrees of legitimacy), and the inevitable pain and suffering of separation has to be borne. Indeed, it has to be affirmed as the source of ultimate bliss, since without it there is no feeling of any kind. But it can only be affirmed as the result of the experienced conviction of the unity of 'one eternal glowing life'. And it is this, above all, that poetry must communicate.
#### Englishing *Hyperion*
This is not the first English translation of Hölderlin's novel, nor will it, I hope, be the last. It is not intended as a critical response to the perceived shortcomings of other versions, two of which appeared after I had begun work on my own.110 On the contrary, I am grateful to, and have learnt from all of them. We are each of us contributing in our own different ways to the afterlife of a remarkable and beautiful work, one that deserves to be far better known than it is in the anglophone world.
Every translation presupposes an 'original', although the concept can be a rather slippery one, and a translator may often find her/himself in the role of editor as well.111 In the present instance this should not really be a problem. After all, in the post-war period Hölderlin has been the subject of extensive and meticulous editorial activity, issuing in two multi-volumed historical-critical editions (Appendix A). And whereas much of his major poetry remained unpublished in his lifetime — ensuring that the sorting and deciphering of the manuscripts he left behind, sometimes virtual palimpsests, has proved to be a task of hideous complexity — *Hyperion* actually appeared in print. Nor has the copy sent to the publisher survived, only various drafts in manuscript, including substantial fragments of one which is quite close to the finished version, as issued by Cotta in 1797 and 1799. On the face of it, therefore, the situation would not seem unduly complicated. However, the original edition inevitably contains misprints. Some of these were picked up post-publication by Hölderlin himself and corrections and
<sup>110</sup> Those of Ross Benjamin and India Russell — see Appendix B.
<sup>111</sup> On the instability of originals, and translators as editors, see Karen Emmerich's stimulating *Literary Translation and the Making of Originals* (London: Bloomsbury Academic, 2017).
additions inserted into the dedicatory copies for Susette Gontard (the model for Diotima) and his good friend Franz Wilhelm Jung. These corrections and modifications have been duly incorporated into the two historical-critical editions and their offspring. But Hölderlin was not the world's greatest proof-reader, either before or after publication, and there are numerous occasions when the modern editor will see the need to intervene (though not to the extent evinced by nineteenth-century editions, including the second Cotta of 1822, published of course without Hölderlin's imprimatur). In almost all cases these will be minor matters, affecting mainly consistency of spelling and punctuation. Sometimes, however, the most conscientious editor can make arbitrary decisions or even mistakes.
It might seem from the referencing conventions adopted in the 'Afterword' that I favour Beissner's *Große Stuttgarter Ausgabe* (StA) over Sattler's *Frankfurter Hölderlin Ausgabe* (FHA). This is not the case. The reason I quote and use page references from the former is simply that the Stuttgart edition is readily available online. Beissner's *Hyperion* does in fact contain blemishes. For instance, in the twenty-seventh letter when Diotima speaks of the implications of Hyperion's idealism, his search for a better world, she is made to say: 'one can't say precisely when it was there, when it went away' (p. 57). Here Beissner's text, and all subsequent editions based on it, including Schmidt's, has: 'man kann so genau nicht sagen, wenn es da war, wenn es weggieng'.112 In the critical apparatus we are told that, instead of the conjunction 'wenn', Cotta's second edition of 1822 has 'wann'.113 Given the wide variety of eighteenth-century practice in the use of conjunctions — for instance, the occurrence of 'wenn' as an interrogative — Hölderlin's usage here would not raise many eyebrows. Admittedly, judging by *Hyperion*, he seems more likely to use 'wann' where one would nowadays expect 'wenn', rather than vice versa.114 And in fact, if one actually consults
<sup>112</sup> StA III, 66–67.
<sup>113</sup> StA III, 362 (line 34).
<sup>114</sup> One example, but by no means the only one, in Letter LV: 'Du lächelst, Alabanda? o wie oft, in unsern ersten Zeiten, hast du so gelächelt, wann dein Knabe vor dir plauderte, im trunknen Jugendmuth' (StA III, 127); ['You're smiling, Alabanda? Oh, how often in our early days together did you smile that way when your boy prattled away before you in the drunken exuberance of youth' (pp. 109–110).] Here Hölderlin uses 'wann' as the equivalent of 'whenever', although elsewhere he
Cotta's first edition, it is clear that Beissner is quite simply wrong. There in the first volume we read: 'man kann so genau nicht sagen, wann es da war, wann es weggieng' (p. 118). It seems then that Beissner has mistakenly taken over the 'wenn' reading from the second Cotta edition of 1822. This is clearly what the editors of the FHA assume him to have done.115 The trouble is that it is not there either. In the same place on the same page in the same volume we read: 'man kann so genau nicht sagen, wann es da war, wann es weggieng'.116 But at least in this instance, even if the footnote is wrong, the Frankfurt edition gives us an accurate primary text, without editorial intervention. This cannot be said of the change of possessive pronoun, already referred to, in the final instalment of Diotima's swansong: 'The beautiful world is my Olympus' (p. 126). Here the FHA follows nineteenth-century editions in substituting 'dein' ('your') for 'mein' ('my').117 One could perhaps understand the reasoning behind the change, if any were offered. But as far as I can tell, it has never been properly argued. I see no problem in having Diotima say the equivalent of: 'for me the beautiful world is Olympus'. And whatever shortcomings Hölderlin might have had as a proof-reader of his own work, one would have thought that he would have given particular attention to Diotima's last letter, and therefore himself picked up and corrected any blatant error in the copy he dedicated to her.
For the purposes of the translation I have been extremely reluctant to assume any but the most obvious of misprints. There is just one notable exception, which I take to be sanctioned by Hölderlin himself. In the eighteenth letter we are told that Diotima's heart was at home amongst the flowers:
She called them all by name, made them new and more beautiful names out of love and unerringly knew the happiest season of each.
also tends to have 'wenn'. As the conjunction for non-recurrent events he usually has 'da', even 'wie', hardly ever 'als'.
<sup>115</sup> FHA 11, 662: '[line] 12 StA (mit der zweiten Auflage von 1822) *wenn es da war*, *wenn*.'
<sup>116</sup> It seems to me that the editors of the FHA have not actually consulted the Cotta second edition themselves, relying rather on Gustav Schlesier's later compilation of a list of corrections made in red ink in a copy of the original Cotta, in preparation for the 1822 edition. For this list, see Gustav Schlesier, *Hölderlin-Aufzeichnungen*, ed. by Hans Gerhard Steimer (Weimar: Verlag Hermann Böhlaus Nachfolger, 2002), pp. 165–68. The FHA editors must have assumed that the correction is one that Schlesier missed, for it is not in his list either.
<sup>117</sup> See above, p. 182; FHA 11, 766.
Like a sister when from every nook a loved one comes clamouring, each wanting to be greeted first, so the still being was busy with hand and eye, blissfully distracted, whenever we walked through wood or meadow. (p. 49) 118
The image of the sister being besieged by a multitude of boisterous siblings perhaps recalls Lotte in Goethe's *Werther*. However, any element of riotousness is absent from the original as published, where the verb I have rendered as 'comes clamouring' is simply the rather tame 'entgegenkömmt' ('comes towards'). Beissner himself points to the possibility that this might be down to compositor's error, since the surviving last draft has 'entgegentönt' (literally: 'sounds towards').119 Originally this was 'ruft und winkt' ('calls and waves'), which also suggests lively competing for attention. Why Hölderlin should later have amended 'entgegentönt', and substituted a bland nothing verb, is beyond me. I have therefore chosen to assume that he did not.
With regard to punctuation, this translation generally tries to be faithful to usage in the original so far as this is not merely a matter of the grammatical conventions of German, but characteristic of Hölderlin's (or Hyperion's) style and reflects the rhythms of the source. Thus, I generally preserve the position of the many question and exclamation marks if possible, even if this might occasionally look odd in English. However, I have deviated in one significant respect, and that concerns the use of inverted commas. As mentioned above, the convention adopted in the original — and also in the published 'Fragment' and Hölderlin's manuscripts — seems to be that they are only used for direct speech when there is no saying verb that makes it clear who is speaking. For instance, "Werden wir das?" in the draft becomes in the final version: Werden wir das? fragt' ich.120 The principle is fairly straightforward, or would be if editors did not occasionally decide to remove quotation
<sup>118 &#</sup>x27;Unter den Blumen war ihr Herz zu Hause, als wär' es eine von ihnen. // Sie nannte sie alle mit Nahmen, schuff ihnen aus Liebe neue, schönere, und wußte genau die fröhlichste Lebenszeit von jeder. // Wie eine Schwester, wenn aus jeder Eke ein Geliebtes ihr entgegenkömmt, und jedes gerne zuerst gegrüßt seyn möchte, so war das stille Wesen mit Aug und Hand beschäftigt, seelig zerstreut, wenn auf der Wiese wir giengen, oder im Walde' (StA III, 56).
<sup>119</sup> StA III, 259, 525. A diplomatic transcription and facsimile of the MS may be found in FHA X, 308–309.
<sup>120</sup> Letter XX: "'Shall we?' I asked" (p. 50); StA III, 260, 57.
marks for no apparent reason.121 Nevertheless, it will not be a convention familiar to many modern anglophone readers, who may be inclined to ask what is so special about the sporadic instances where it is employed (and may well also find the plethora of saying, crying, asking verbs a little repetitive, but that cannot be helped). I have therefore made the decision to use quotation marks for direct speech throughout, whether or not it is made explicit by a verb. Like Trask (Appendix A) I also use them for direct thought. There are numerous occasions throughout the narrative where Hyperion records the thinking of his former self, and if one is going to argue that it is important to be able to distinguish between the perspective of the hero within the narrative and that of the maturing narrator, then it seems to me that visual assistance of this kind is no bad thing. I have chosen to use single quotation marks, except for Hyperion's final 'speech'. As already mentioned, Hölderlin emphasizes its unique significance by having quotation marks at the beginning of each of the original's forty-five lines. Rather than trying to reproduce this, I have here chosen to distinguish the passage by using double quotation marks.
The modern sanitized texts of German 'classics' tend to obscure (if not conceal) the fact that these were produced when even the written language had yet to be thoroughly standardized. Nor is it only a matter of orthography, which can nowadays look decidedly antiquated, although that is perhaps the most striking difference.122 If the English of my translation comes across to the reader as a strange mixture of the poetically highly charged with the colloquial and regional, my justification is the nature of Hölderlin's original and the impression it makes on me. There are numerous occasions where he deviates from currently accepted norms. There follow some examples. I have already mentioned the use of temporal conjunctions ('wenn', 'wann'), although any unorthodoxy will not be reflected in an English translation, where just the one will do, namely 'when'. Similarly, though still common
<sup>121</sup> Beissner arbitrarily deletes the first quotation marks in the novel: ["]Und wenn ich fragte?["] (StA III, 33). And both StA and FHA delete the second set: ["]sind es und werden es seyn["] (StA III, 57; FHA XI, 650; Cotta I, 102), though they are clearly there in the draft.
<sup>122</sup> And can lead translators astray. For instance, Hölderlin spells 'glimmen' ('glimmer') as 'klimmen', though it has nothing to do with 'climbing'; Letter XLV: 'klimmt noch in uns ein Sehnen nach den Tagen der Urwelt' (StA III, 112).
enough in spoken German, using the conjunction 'wie' in comparisons of inequality would nowadays attract the red pen and a firm underlining in student work.123 Similarly, the use of 'so' as a relative pronoun.124 What would doubtless elicit a double underlining with an exclamation mark in the margin, are vestiges of the Swabian strong adjectival ending following the definite article in the plural. These are a feature of Hölderlin's language in his letters, and not every instance is picked up in the Cotta edition of *Hyperion*. 125 There are instances of double negatives for emphasis.126 I wish I dared to replicate these in the English, but since the construction has been frowned on too much for too long, I fear it would jar unacceptably. There are sundry deviations from conventional word order, particularly the position of the auxiliary in subordinate clauses, although in an eighteenth-century context they are perhaps not quite so striking.127 Some apparent solecisms appear to be Swabianisms, for instance, 'nimmt' instead of 'nehmt', as the secondperson plural familiar form of 'nehmen'.128 This is 'corrected' in the 1822 edition, though not the anacoluthon in 'wir fürchteten uns, sich ihrer [der Liebe] zu überheben in Reden und stolzen Gedanken', where in the
<sup>123</sup> Cf. Letter XLVIII: 'stille zu stehn, ist schlimmer, wie alles' (StA III,116). Elsewhere he can be more orthodox — cf. Letter 27: 'ich versteh' es und besser, als du' (StA III, 66); Letter 58: 'alt zu werden, da wo alles alt ist, scheint mir schlimmer, denn alles' (StA III, 149). But that such 'correctness' perhaps goes against the grain, is suggested by the following examples from the drafts: 'ich war besser, wie sie!' (StA III, 242); 'Warum bin ich später geboren, wie er' (StA III, 273).
infinitive clause a third-person reflexive pronoun ('sich') is used instead of the first-person plural ('uns').129
One does not have to read very far in the German *Hyperion* before encountering examples of elision more commonly associated with the spoken language. One might think here of the frequent omission of the second 'e' in 'gehen', here clearly outnumbered by 'gehn'. Many involve the contraction of 'es' ('it') to 's': with pronouns — 'ich's', 'mirs', 'dus', 'dirs', 'ihr's', 'man's'; with conjunctions — 'wie's', 'wenn's', 'weils'; with verbs — 'braucht's', 'geht's', 'giebt's', 'hab's', 'ists', 'sind's', 'war's', 'sahen's'. And there are literally hundreds of instances of the elision of the final 'e' of a verb before a word starting with a vowel. On the first page alone we have 'wär' ich', 'würd' ich', 'schnürt' er'.130 Bearing all this in mind, I had no compunction about using English contractions in the translation. If they might seem unsuitably colloquial for a literary text of such high seriousness, then so be it. But one ought also to consider that the novel is epistolary, consisting of letters addressed to a familiar, so that the contractions are perhaps not out of place. There is, however, a further argument that is relevant here. It was the young Nietzsche who claimed that the prose of his favourite poet's novel is music.131 Like many of the deviations from 'normal' word order, the elisions of Hölderlin's language are often there to serve euphony and rhythm. He abhors vowel sounds in hiatus and generally does his best to avoid them. Whether hiatus can ever be pleasing on the ear in either language is perhaps moot. In any case, in the translation I have taken my cue from the original by, for instance, using contractions such as 'I'm', 'you're', 'we're', 'they're', for 'I am', 'you/we/they are', and as far as possible trying to circumvent hiatus elsewhere.
As Hölderlin indicates in the Preface, the meaning of his novel is the whole. What it means is inseparable from how it means. In order to render that meaning one must at least try to approximate to its linguistic beauty, which is not an incidental bonus, but an integral part of the message. In one of her letters Susette Gontard expresses surprise
<sup>129</sup> Letter XXXVI: StA III, 100, 471–72. Beissner argues that this also corresponds to Swabian usage, quite apart from here avoiding an ugly 'uns, uns' sequence.
<sup>130</sup> StA III, 7. There would appear to be no fewer than 327 examples of verbs ending with an apostrophe followed by 'ich'.
<sup>131</sup> Friedrich Nietzsche: 'diese Prosa ist Musik', 'Brief an meinen Freund, in dem ich ihm meinen Lieblingsdichter zum Lesen empfehle' (19 October, 1861).
that Hölderlin should call his dear *Hipperion* a novel. For her it is more like a 'beautiful poem'.132 And indeed it confronts the translator with demands that are normally associated with lyric poetry, rather than prose. I confess that I find myself unequal to the task, in the sense that I can only hope to hint at the rhythms of the novel's language, but there is no way in which I can closely replicate them in English. The reader should still at least be able to recognize that the original is written in poetic prose, though not perhaps that some of the lines resolve into hexameters.133 At the most basic level I can follow Hölderlin in as far as possible avoiding sequences of more than two unaccented syllables, and also steering clear of stress clashes.134 This often involves rejecting polysyllabic Latinate vocabulary. In the rendering of the 'Song of Fate' (p. 123) I have tried to be more faithful to the stress patterns and rhythms, but with limited success, since accented sounds may still be of varying length — there cannot be many four-syllable words that take less time to say than 'Götterlüfte'. In the third stanza I opted for the archaic northern 'blindling' for 'blindlings', not only because it looks almost identical to the German, but because I hear it as longer than 'blindly'.135
If being faithful to Hölderlin's text means mediating its beauty, there are bound to be occasions when the literary will have to be privileged over the literal. But I have tried to keep these to a minimum. I would rather take liberties with the conventions of English than with Hölderlin, subject only to the results being consistent with my notions of what is poetically effective. It is a besetting sin of literary translations to be linguistically conservative. This is especially inappropriate when we are dealing
<sup>132</sup> Letter of 19 March, 1799 (StA VII (1), 75).
<sup>133</sup> Cf. Letter V: 'wie ein Geist, der keine Ruhe am Acheron findet' (StA III, 17). I am grateful to Wolf Schmidt for this example.
<sup>134</sup> See Steimer, 'Hölderlins Klauseln', p. 287: 'Für die Diktion des Romans ist nämlich charakteristisch, daß sie den Hebungsprall ebenso meidet wie Folgen von mehr als zwei gleichmäßig tonlosen Senkungen.'
<sup>135</sup> Rightly or wrongly, I read the German 'blindlings' here as a near-spondee. I was tempted to follow Edwin Muir and have the variant spelling 'blindlings' in English, but then the word would likely be taken as a noun ('blind person'). I would rather it suggest a neologism: 'blindly (s)tumbling' (?). For Muir's fine part-translation, see 'A Note on Friedrich Hölderlin', *The Freeman*, 7 (1 August 1923), p. 489. Together with two stanzas from 'Patmos' in the same article, this represents Muir's earliest translation from Hölderlin, and the only one he published of the 'Song of Fate', much though it meant to him.
with the work of a writer who was developing into one of the greatest poets of world literature, and who himself used translation to push the boundaries of literary German.136 As David Constantine observes: 'There is in [translations], very often, a poorer deployment of the host language's lexical and grammatical possibilities, altogether less variety of utterance … Translations die fast because there is, on the whole, less adventure, less risk, less departure from the norm in them than in the originals.' And, of particular relevance to the translation of an eighteenth-century source: 'You have to write in a language neither antiquarian nor up-tothe-minute modern; which is to say a language which is, in relation to the text equivalently poetic.'137 There are excellent lessons to be learnt here, but applying them successfully is of course a tall order.
One of the things I have tried to do in the translation is to use, wherever feasible, English words that are cognate with the German. Occasionally this involves slight shifts of meaning, a subtle approaching of the source to the target language. For instance, 'Strom' normally suggests something rather more imposing than 'stream'; 'wirken' would not normally be rendered by 'work'; and 'wandern' is given here throughout as 'wander', even though there is an element of aimlessness in the English verb which is not quite so present in the German. In these and other cases, I have allowed myself to be swayed by my own feeling for the euphony of Hölderlin's language and a desire to approximate to it, if necessary at the cost of strict accuracy. For instance, at the beginning of the second volume:
A brother of spring was the autumn for us, full of mellow fire, a feast time for commemorating sorrows and past joys of love. The withering leaves wore the red hues of dusk, only the spruce and the laurel stood in eternal green. In the bright breezes wandering birds lingered, others swarmed in vineyard and garden, merrily reaping what people had left. (p. 81)138
<sup>136</sup> Particularly in his translations from Pindar and Sophocles. See StA V: Übersetzungen.
<sup>137</sup> David Constantine, *Poetry* (Oxford: Oxford University Press, 2014), pp. 46–47.
<sup>138</sup> Letter XXXI: 'Ein Bruder des Frühlings war uns der Herbst, voll milden Feuers, eine Festzeit für die Erinnerung an Leiden und vergangne Freuden der Liebe. Die welkenden Blätter trugen die Farbe des Abendroths, nur die Fichte und der Lorbeer stand in ewigem Grün. In den heitern Lüften zögerten wandernde Vögel, andere schwärmten im Weinberg, und im Garten und erndteten fröhlich, was die Menschen übrig gelassen' (StA III, 93).
In the first clause of the last sentence the loveliness of the season is causing the birds to delay their migration. But rightly or wrongly, there is no way in which I could countenance such a jarring word as 'migrating' in this context.
As far as 'Mädchen' is concerned, like other translators I have felt obliged to opt in the great majority of instances for the cognate 'maiden' (and this definitely does not imply an answer to a meaningless question). Given the pedestal he places her on, it seemed to me to be utterly incongruous to have Hyperion referring to Diotima as a 'heavenly girl'.139 On the other hand, she would be hardly likely to address herself as a 'silly maiden'.140 One decision in favour of a cognate word might seem more contentious.
'What are we talking about just now?' I could cry, 'it's often so hard, you can't find the matter to hold fast your thoughts.'
'Are they taking off into the air again?' replied my Diotima. 'You'll have to bind lead to their wings, or I'll tie them to a string, like the boy his flying dragon, so they can't get away from us.'141
In modern German 'Drachen' means 'kite', and is distinguished, at least in the nominative, from 'Drache', the mythical fire-breathing monster. In Hölderlin's day there appears to have been no distinction in the word, at least to judge by a contemporary bilingual dictionary, which gives as one of four definitions for 'der Drache': 'the Dragon, a Machine of Paper &c. which the Boys let fly into the Air.'142 According to the Oxford English Dictionary 'dragon' as 'paper kite' seems no longer to be current, except in Scots usage. Nevertheless, I have opted for it here. It seems to me that there is an ironic contrast between the formidable lofty thoughts and the callow youth who has them, and this is brought
<sup>139</sup> Letter XXI: 'des himmlischen Mädchens' (StA III, 58); Letter XXX: 'das himmlische Mädchen' (StA III, 86). Other epithets preceding 'Mädchen' include: 'herrliches' ('glorious'/'magnificent'), 'edles' ('noble'), 'göttliches' ('divine'), 'hohes' (literally 'high'), and 'heiliges' ('holy'/'hallowed').
<sup>140</sup> Letter XLIII: 'o des albernen Mädchens!' (StA III, 109).
<sup>141 &#</sup>x27;Wovon sprechen wir doch geschwind? konnt' ich rufen, man hat oft seine Mühe, man kann den Stoff nicht finden, die Gedanken daran festzuhalten. // Reißen sie wieder aus in die Luft? erwiederte meine Diotima. Du mußt ihnen Blei an die Flügel binden, oder ich will sie an einen Faden knüpfen, wie der Knabe den fliegenden Drachen, daß sie uns nicht entgehn' (StA III, 66).
<sup>142</sup> Joh(an)n Ebers, *The New and Complete Dictionary of the German and English Languages* (Leipzig: Breitkopf and Haertel, 1796), I, p. 608.
out better by the use of 'dragon' here. Certainly, Hyperion's slightly peeved reaction suggests he is sensitive to the implications of the simile.
In a very modest way I have tried to make room for a fuller range of literary English than may be usual in a translation of this kind. This involves what I hope is judicious use of northern or Scots words, not for their own sake, but when they seem to me to be the best choice. Some, such as 'burn', 'dreich', 'couthie', 'reekie', 'fey', 'thrawn', 'airt', 'drouth' ought to present no problem. Others, such as 'drum(b)ly' or 'wimple' should be self-explanatory from the context, I hope. If not, they can usually be found in good dictionaries. And they are words that appeal to me. It is not that I hear Hyperion speaking English with a Scottish accent. But then neither do I hear him using bland Received Pronunciation. When reading Hölderlin for myself in German, I try, as best I can, to incorporate a Swabian lilt — and make 'Geist' as frothy as possible.
Finally, something ought to be said here about intertextuality. It is difficult for anyone reading the original to avoid picking up the odd echo from *Werther*, and Hölderlin clearly wrote *Hyperion* in the expectation that his readership would know Goethe's novel. However, there have been at least twenty different translations of the latter into English, including four in the last fifteen years, so that it is clearly extremely unlikely that any precise verbal echoes would be registered by an anglophone — not even the 'einiges, ewiges, glühendes Leben' at the end of Hölderlin's novel, which surely recalls Werther's 'das innere glühende, heilige Leben der Natur'.143 *Hyperion* also contains a possible allusion to Mignon's song in Goethe's *Wilhelm Meisters Lehrjahre* (1795): 'Kennst du das Land, wo die Zitronen blühn, / Im dunkeln Laub die Gold-Orangen glühn …'. It is difficult to avoid thinking of this when we read Hyperion's evocation of the groves of Angele: 'wo die goldne Frucht des Zitronenbaums aus dunklem Laube blinkt'.144 Goethe's lines are themselves a reminiscence from James Thomson's *Seasons* (1744 edition): 'Bear me, *Pomona!* to thy Citron-Groves; / To where the Lemon and the piercing Lime, / With the deep Orange, glowing thro' the Green,
<sup>143</sup> See Werther's letter of 18 August, in Constantine's translation: 'the holy fires of the inner life of Nature' (p. 45).
<sup>144</sup> Letter XXIX: 'where the golden fruit of the citron tree gleams through the dark green' (p. 74)
/ Their lighter Glories blend.'145 In the earlier unpublished version of Goethe's novel, the foliage from which the lemons glow had been 'green' rather than 'dark'.146 Since the relevant passage in Thomson was added too late to be included in Brockes' translation (1745), and Goethe learned to read English early with his father, the only source for the echo in Mignon's song must be the original. As far as we know, Hölderlin never developed a reading knowledge of English, the only book in the language he is known to have fleetingly had in his possession being 'Monk' Lewis's translation of Schiller's *Kabale und Liebe*. Goethe's novel obviously made an impression on Hölderlin, and so I tend to assume an echo in *Hyperion*, though it is unlikely to resonate in the translation. The same will be true of the probable allusion to Augustine's *Confessions* at the beginning of the fifth letter, 'Whither could I flee from myself …?', which surely bears more than a passing resemblance to 'quo a me ipso fugerem?' from the fourth book of the *Confessiones*. 147 The Bible is another matter.
It is no surprise that someone who grew up in an orthodox Protestant (if perhaps pietistically inclined) environment in provincial Württemberg in the last third of the eighteenth century, who spent five years at the theological seminary ('Stift') in Tübingen (1788–93), and whose whole education was in fact predicated on the assumption that he would become a minister of religion, should be thoroughly conversant with the Bible, particularly in Martin Luther's translation (1545). And there are of course numerous turns of phrase in *Hyperion* that are obviously biblical. One example may serve for many. When Hyperion's father offers the Pauline advice: 'prüfe alles und wähle das Beste!', I have translated with something close to the equivalent in the King James Bible (1611): 'prove all things and hold fast the best!' (p. 17).148 That may perhaps be seen as making the biblical allusion too
<sup>145</sup> James Thomson, 'Summer', *The Seasons* (London: A. Millar, 1744), p. 82.
<sup>146</sup> Curiously enough, Thomson's friend and fellow Scot, John Armstrong, has a very similar line in his *Art of Preserving Health* (1745) 'Thro' the green shade the golden Orange glows' (Bk 2, line 331) — See Adam Budd, *John Armstrong's The Art of Preserving Health*: *Eighteenth-Century Sensibility in Practice* (Farnham: Ashgate, 2011), p. 76.
<sup>147 &#</sup>x27;Wohin könnt' ich mir entfliehen [, hätt' ich nicht die lieben Tage meiner Jugend]?' (StA III, 17); Augustine, *Confessiones*, Book IV, ch. 7 (12).
<sup>148</sup> Letter VI: StA III, 20. See 1 Thessalonians 5:22/21: 'Prove all things: hold fast that which is good'; 'Prüfet aber alles / und das Gute behaltet.'
explicit, but on the other hand it enables me to use the cognate 'prove' for 'prüfe' without appearing to indulge in antiquarianism. Whenever I think I have picked up a biblical reference, image, or idiom, I have tried if possible to reflect it in the English. Occasionally this might result in a translation that appears to deviate wilfully from the obvious. A case in point would be the final lines of the second stanza of the 'Song of Fate': 'Und die seeligen Augen / Bliken in stiller / Ewiger Klarheit', where I have 'And the blissful eyes / Gaze in eternal / Tranquil glory'.149 As mentioned above (p. 143), the imagery of the Song is anticipated in the thirteenth letter:
Ich hab' es heilig bewahrt! wie ein Palladium, hab' ich es in mir getragen, das Göttliche, das mir erschien! und wenn hinfort mich das Schiksaal ergreift und von einem Abgrund in den andern mich wirft, und alle Kräfte ertränkt in mir und alle Gedanken, so soll diß Einzige doch mich selber überleben in mir, und leuchten in mir und herrschen, in ewiger, unzerstörbarer Klarheit!150
The language in the last clause recalls the shining light of the glorious gospel of Christ, the image of God.151 I cannot be the only translator to be struck by the inadequacy of 'clarity' as a rendering for 'Klarheit', either here or in the 'Song of Fate'. Quite apart from the prosaic flatness of the word, it lacks all religious resonance by comparison with the German. For 'Klarheit' occurs fifteen times in the Luther Bible (always in the New Testament), but 'clarity' not at all in King James. On every occasion bar one the word used here is 'glory', and even the exception features the adjective 'glorious'. Examples: 'And, lo, the angel of the Lord came upon them, and the glory of the Lord shone round about them: and they were sore afraid' (Luke 2:9); 'And, behold, there talked with him two men, which were Moses and Elias: Who appeared in glory … But Peter and they that were with him were heavy with sleep: and when they were awake, they saw his glory' (Luke 9:30–32); or 'There is one glory of the
<sup>149</sup> Letter LVIII (p. 123): StA III, 143. I have reversed the order of the epithets for reasons of rhythm.
<sup>150</sup> StA III, 51; 'I've kept it sacred! like a palladium I've carried it within me, the divine that was revealed to me! and if fate henceforth should seize and plunge me down from abyss to abyss and drown in me all energy and all reason, yet shall this one and only outlive myself in me and shine in me and reign in eternal, indestructible glory!' (p. 44).
<sup>151</sup> Cf. 2 Corinthians 4:4.
sun, another of the moon, and another glory of the stars: for one star differeth from another star in glory' (1 Corinthians 15:41).152 If we look at the definitions of 'glory' in the OED, we find amongst them: 'the majesty and splendour attendant upon a manifestation of God'; 'resplendent beauty or magnificence … effulgence of light such as is associated with conceptions of heaven'; 'the splendour and bliss of heaven'. It seems to me that, in the circumstances, 'glory' represents in almost every respect a better choice than 'clarity', particularly in the 'Song of Fate'. It does not, admittedly, contrast as well with 'confusion' in the last line of the 'Song'. But one cannot have everything.
One respect in which *Hyperion* might seem to have no similarity with *Werther* is in the significance for the latter of Macpherson's *Ossian*. No less than seven percent of Goethe's novel consists of translation from two of the Ossianic poems, the recital of which by the protagonist precipitates the tragic outcome.153 And throughout the novel there are sundry echoes of his favourite reading in the letters he writes. There would appear to be nothing comparable in *Hyperion*. And yet, a twentieth-century critic can claim that the novel is 'unthinkable' without *Ossian*. 154 He does not go into specifics. But then neither do those contemporaries of Hölderlin who are reminded of *Ossian* when they read *Hyperion*. The earliest reviewer (1799) states explicitly that whoever does not like *Ossian* will cast the book aside after the first letter.155 Reactions that suggest an Ossianic influence on the novel, whether approving or not, continue well into the nineteenth century.156 The reason they have not been much in evidence in more recent years is not far to seek. Macpherson's work is assumed, wrongly, to be totally fraudulent; it is assumed, equally wrongly, to be aesthetically worthless. Consequently, it is left unread.
<sup>152</sup> It should be noted that in the Luther Bible, at least since 1984, 'Klarheit' has been replaced here by 'Glanz', and the total instances of the word have been reduced to two.
<sup>153</sup> For the role of Ossian in *Werther*, also the way it is tackled by anglophone translators, see Howard Gaskill, '"Arise, O magnificent effulgence of Ossian's soul!": Werther the Translator in English Translation', in *Translation and Literature*, 22 (2013), 302–21.
<sup>154</sup> Herbert Schöffler, 'Ossian: Hergang und Sinn eines großen Betrugs', in Schöffler, *Deutscher Geist im 18.Jahrhundert. Essays zur Geistes- und Religionsgeschichte*, 2nd edition (Göttingen 1967), pp. 135–54, pp. 149–50. The essay first appeared in 1941.
<sup>155</sup> See StA III, 323 for this benevolent anonymous review of the first volume.
<sup>156</sup> One of the last comes from Karl Rosenkranz in his Hegel biography of 1843 — see Howard Gaskill, 'Hölderlin und Ossian', *Hölderlin-Jahrbuch*, 27 (1990–91), 100–30 , pp. 114–15.
Embarrassment combines with ignorance to repress the uncomfortable truth, that historically *Ossian* is one of the most influential works in world literature.157 Whether we like it or not, Hölderlin knew the Ossianic poetry extremely well. He can be shown to have devoured it as an adolescent, declaring his intention to read it until he has it (half) off by heart.158 Nor was his admiration confined to a youthful phase. Probably the last work he prepared for publication, the stunningly beautiful *Pindar-Fragmente* (1804?), features Ossian in the final sentence of the final Fragment.159 Nor will *Ossian* have been too far from his mind when he was working on *Hyperion*. Franz Wilhelm Jung, whose acquaintance Hölderlin first made on his arrival in Frankfurt in January 1796, was himself producing a translation of *Ossian* in free rhythms and immediately sought to gain the poet's interest in it. Successfully, as it seems, for when Jung was negotiating with the publisher Cotta, late in 1797, the manuscript was with Hölderlin, whose judgement is expected to carry authority. Almost a year later Jung claims in a letter to Fichte that Hölderlin is pressing him to publish the translation. And in 1799, when Hölderlin was attempting to set up his abortive literary periodical *Iduna*, it was his intention to include commented excerpts from Jung's *Ossian*. Hölderlin's affection for Jung is demonstrated by the fact that, together with Susette Gontard, he has a corrected copy of *Hyperion* (first volume) dedicated to him. Under the circumstances, it would not be surprising to find a plenitude of Ossianic resonances in the novel, if one knew what to look for. I have attempted to translate in such a way that anglophones familiar with *Ossian* — and they do exist — might also be reminded of it when they read *Hyperion*.
James Macpherson wrote his translations/adaptations/fabrications of ancient Gaelic verse in what he calls 'measured prose', whilst continually
<sup>157</sup> For evidence of the ubiquity of *Ossian*, including the impact on art and music, see *Ossian in Europe*, ed. by Howard Gaskill (London: Thoemmes Continuum, 2004); also 'Versions of Ossian: Receptions, Responses, Translations', ed. by Howard Gaskill, in *Translation and Literature*, 22 (2013), 293–435. For Ossian in Germany, see the magisterial study by Wolf Gerhard Schmidt, *'Homer des Nordens' und 'Mutter der Romantik'*: *James Macphersons Ossian und seine Rezeption in der deutschen Literatur*, 4 vols (Berlin: de Gruyter, 2003–4); for Hölderlin, see especially II, pp. 901–26; for *Hyperion*, pp. 905 ff.
<sup>158</sup> To Immanuel Nast, March 1787: 'da leß ich ihn so lang, biß ich ihn halb auswendig kan' (StA VI, 16).
<sup>159</sup> StA V, 290; Adler, *Hölderlin*: *Letters and Essays*, p. 339.
underlining the lyrical beauties of his originals. Translators of the English text did not feel constrained to follow Macpherson in reproducing it in prose, poetic or not, but one who did is Schiller's friend Johann Wilhelm Petersen, whose *Die Gedichte Ossians neuverteutschet* appeared in 1782.160 This is, in my view, the translation that exerted the greatest influence on Hölderlin, the one he was going to read and re-read until he had it by heart.161 The first German translation, in fact the first complete translation of Macpherson's *Ossian* into any language, had been made into hexameters by Michael Denis, appearing in 1768–69.162 In 1784 he adapted his version to accord with Macpherson's revised edition, *The Poems of Ossian* of 1773, and took the opportunity to issue it together with three volumes of his own poetry, written under his bardic name Sined (Denis spelt backwards).163 It is known that Hölderlin must have had access to this edition, or at least to one of its volumes, since as an eighteen-year-old he uses as a motto lines taken from one of Sined's poems.164 Although the Denis was probably not so suitable for everyday use (I find it difficult to imagine Hölderlin wandering around with expensive bulky quarto volumes under his arm — Petersen's translation comes in a handy single octavo), it is tempting to think that he might have found time to peruse the German version of Hugh Blair's 'Critical Dissertation on the Poems of Ossian', with which Denis opens the third volume. Blair writes: 'The "joy of grief," is one of Ossian's remarkable expressions, several times repeated.'165 Denis translates: 'Die *Wonne der Wehmuth* ist einer von Ossians merkwürdigen Ausdrücken, den er zu verschiednen Malen wiederholt.'166 This felicitous phrase, the first ever occurrence, was coined by Denis (or suggested to him) too late for inclusion in the first two volumes of 1768, though he made sure to incorporate it throughout in the revised edition of 1784. And it appealed to others too, whatever they might have thought of Denis's hexameters. For it is taken over by all German translators of the complete *Ossian*
<sup>160</sup> Tübingen: Heerbrandt.
<sup>161</sup> For the reasons, see Gaskill, 'Hölderlin und Ossian', pp. 106–9.
<sup>162</sup> Denis, *Die Gedichte Ossians eines alten celtischen Dichters*, 3 vols (Vienna: Trattner, 1768–69).
<sup>163</sup> *Ossians und Sineds Lieder* (Vienna: Wappler, 1784).
<sup>164</sup> See StA VI (2), 508; the lines are from the fourth volume, p. 163.
<sup>165 &#</sup>x27;Critical Dissertation', in *Poems of Ossian* (ed. Gaskill), 343–408, p. 381.
<sup>166</sup> *Ossians und Sineds Lieder*, III, p. xcv (Denis's italics).
before 1800, including of course the three known to be known to Hölderlin (Petersen, Denis, Jung).167 When he has Hyperion write, then: 'Wie aber am Strahle des Morgenlichts das Leben der Erde sich wieder entzündete, sah ich empor und suchte die Träume der Nacht. Sie waren, wie die schönen Sterne, verschwunden, und nur die Wonne der Wehmuth zeugt' in meiner Seele von ihnen', he is deliberately evoking *Ossian*. <sup>168</sup> Amongst historians of eighteenth-century German literature there has been a widespread tendency to misattribute 'joy of grief' (usually to Edward Young), or to assume that 'Wonne der Wehmuth', if not a catchphrase spontaneously self-generated in an age of sensibility, derives from Goethe, whose short and joyfully weepy poem of that title was written in 1775, but did not appear in print until 1789. Certainly, if editors think it worth a comment, they refer only to Goethe, never to *Ossian*. 169 Yet I cannot believe that someone who knows the work as well as Hölderlin can possibly write 'Wonne der Wehmuth' in all innocence of its Ossianic associations. I have therefore translated the passage: 'But when the life of the earth took fire again from the ray of the morning light, I looked up and sought the dreams of the night. Like the beautiful stars they had vanished, and only the joy of grief bore witness to them in my soul' (p. 61).
Ironically, it is with some slight reluctance that I use 'joy of grief' here.170 Both this phrase and its German equivalent are too readily associated with mawkish sentimentality, tears without fears. Blair defines it as 'that gratification, which a virtuous heart often feels in the indulgence of a tender melancholy.'171 But there is more to it than that. Naturally, one can regard the 'joy of grief' as typical of the contemporary predilection for diluted mixed feeling and the pleasures of melancholy in which anything genuinely painful is kept at arm's length. But *Hyperion* is made of sterner stuff, and even *Ossian*'s pathos is not groundless. When we read: 'There is a joy in grief when peace dwells in the breast
<sup>167</sup> A successful, if unpoetic prose version was that of Edmund de Harold, *Die Gedichte Ossian's eines alten celtischen Helden und Barden*, 3 vols (Düsseldorf, 1775; also Mannheim, 1782; Münster, 1795).
<sup>168</sup> Letter XXVIII: StA III, 71.
<sup>169</sup> Cf. Beissner, StA III, 462.
<sup>170</sup> David Schwarz is the only other anglophone translator to do so, in his re-working of Trask (who has 'ecstasy of grief'). I presume that Schwarz recognizes the allusion — see *Hölderlin*: *Hyperion and Selected Poems*. ed. by Santner, p. 57.
<sup>171 &#</sup>x27;Critical Dissertation', *Poems of Ossian*, p. 381.
of the sad. But sorrow wastes the mournful,'172 may we not think of 'emotion recollected in tranquillity'? The narrating Hyperion's journey is one from joy and grief to the joy of grief, and embodies perhaps its finest celebration in world literature.
<sup>172</sup> *Poems of Ossian*, p. 381.
Over the years that this translation has been in gestation, many friends, colleagues, acquaintances, and scholars I have never even met, have given generously of their time, and I cannot recall an email sent in vain. They have of course been pestered in unequal measure. The first individual mention here must therefore go to the long-suffering Iain Galbraith with whom I appear to have exchanged 251 emails on *Hyperion* and how best to translate particular passages. Others who received and responded to multiple queries are Wolf Gerhard Schmidt, Gerald Bär, and David Constantine. Special thanks must go to my colleagues in the language sections of the University of Edinburgh, in particular Andrew Barker, Eleoma Bodammer, Peter Davies, Peter France, Peter Graves, and Bill Webster. I have also gratefully received help and advice from Rudolf Brandmeyer, Marco Castellari, Sheila Dickson, Stuart Gillespie, David Hill, Duncan Large, Gauti Kristmannsson, Anthony Krupp, Charlie Louth, Jürgen Link, Matthias Löwe, Luigi Reitani, Thomas Roberg, Ritchie Robertson, Mark Roche, Gerhard Sauder, Hans Gerhard Steimer, Gideon Stiening, Jochen Schmidt, and Seán Williams. In addition, I should mention the stimulating responses to queries addressed to lists such as German-Studies, C18-L, and non\_current\_German.
I am greatly indebted to Roger Paulin for his close reading of the finished translation, and his many helpful observations and suggestions. Any errors or oddities that remain may be put down to my stubbornness.
I have been immensely encouraged by the interest, going back many years, that Jim Devin has taken in my work. There cannot be anyone with a more extensive knowledge of Hölderlin in English, and he has frequently supplied me with copies of contributions of which I was completely ignorant.
Very special thanks are due to Jutta Gaskill. She has been an invaluable and patient sounding-board from the beginning, and anything about the translation that is good owes much to her inspiration. She and our daughter, Fiona Blair, also did sterling work as proof-readers of the English *Hyperion*, for which I am extremely grateful.
Finally, I should like to express my indebtedness to the other translators consulted: for confirming my intuitions, putting me right, or provoking me into my own choices.
For their patience and care my heartfelt thanks to Alessandra Tosi and the team at Open Book Publishing, without whom this *Hyperion* would never have seen the light of day.
### Editions consulted
## Translations
## English
#### Other translations consulted
*Hölderlin*: *Hypérion ou l'Hermite de Grèce*, *précédé du Fragment Thalia*, translated by Phillippe Jaccottet (Paris: Gallimard, 1973).
<sup>1</sup> I am grateful to Jim Devin for alerting me to this and providing me with photocopy.
<sup>2</sup> I am grateful to Professor Reitani for providing early access to this version.
## Select bibliography in English
Constantine, David, *Hölderlin* (Oxford: Clarendon Press, 1988).
*Arethusa* — see *Alpheus*.
*Aristogeiton* — see *Harmodius*.
*Athos* — mountain and peninsula in north-eastern Greece.
*Castor —* see *Dioscuri*.
*Chios* — island off the Anatolian coast.
*Lacedaemon* — another name for *Sparta*.
*Pentelikon* — mountain range in Attica, north-east of Athens.
# This book need not end here...
At Open Book Publishers, we are changing the nature of the traditional academic book. The title you have just read will not be left on a library shelf, but will be accessed online by hundreds of readers each month across the globe. OBP publishes only the best academic work: each title passes through a rigorous peer-review process. We make all our books free to read online so that students, researchers and members of the public who can't afford a printed edition will have access to the same ideas.
This book and additional content is available at: https://doi.org/10.11647/OBP.0160
## Customise
Personalise your copy of this book or design new books using OBP and thirdparty material. Take chapters or whole books from our published list and make a special edition, a new anthology or an illuminating coursepack. Each customised edition will be produced as a paperback and a downloadable PDF. Find out more at:
https://www.openbookpublishers.com/section/59/1
## Donate
If you enjoyed this book, and believe that research like this should be available to all readers, regardless of their income, please become a member of OBP and support our work with a monthly pledge — it only takes a couple of clicks! We do not operate for profit so your donation will contribute directly to the creation of new Open Access publications like this one.
https://www.openbookpublishers.com/supportus
# You may also be interested in:
#### **Don Carlos Infante of Spain: A Dramatic Poem**
*Friedrich Schiller. Translated by Flora Kimmich, with an Introduction by John Guthrie*
https://doi.org/10.11647/OBP.0134
#### **Wallenstein: A Dramatic Poem**
*Friedrich Schiller. Translated by Flora Kimmich, with an Introduction by Roger Paulin*
https://doi.org/10.11647/OBP.0101
#### **Fiesco's Conspiracy at Genoa**
*Friedrich Schiller. Translated by Flora Kimmich, with an Introduction by John Guthrie*
https://doi.org/10.11647/OBP.0058
## Friedrich Hölderlin Translated by Howard Gaskill **Hyperion,** or the Hermit in Greece
*Hölderlin's only novel, Hyperion, is sti ll not widely enough known in the English-speaking world despite its unquesti onable importance. Gaskill's version is the fruit of long years of loving att enti on, and it catches much of the beauty and texture of the original — its rhythms, dicti on, variety and shift s of style.*
—Prof. Charlie Louth, University of Oxford
Friedrich Hölderlin
**Hyperion**
**OBP**
FRIEDRICH HÖLDERLIN
**Hyperion,**
Open Book Classics
or the Hermit in Greece
TRANSLATED BY HOWARD GASKILL
*This is a superb version of Hölderlin's novel: accurate, elegant and inspiring. It at long last brings this major text of German literature within the reach of the Anglophone public.*
—Prof. Jeremy D. Adler, King's College London
Friedrich Hölderlin's only novel, *Hyperion* (1797–99), is a fi c� onal epistolary autobiography that juxtaposes narra� on with cri� cal refl ec� on. Returning to Greece a� er German exile, following his part in the abor� ve uprising against the occupying Turks (1770), and his failure as both a lover and a revolu� onary, Hyperion assumes a hermi� c existence, during which he writes his le� ers. Confron� ng and commen� ng on his own past, with all its joy and grief, the narrator undergoes a transforma� on that culminates in the realisa� on of his true voca� on.
Though Hölderlin is now established as a great lyric poet, recogni� on of his novel as a supreme achievement of European Roman� cism has been belated in the Anglophone world. Incorpora� ng the aesthe� c evangelism that is a characteris� c feature of the age, *Hyperion* preaches a message of redemp� on through beauty. The resolu� on of the contradic� ons and an� nomies raised in the novel is found in the act of ar� cula� on itself. To a degree remarkable in a prose work of any length, what it means is inseparable from how it means. In this skilful transla� on, Gaskill conveys the beau� ful music and rhythms of Hölderlin's language to an English-speaking reader.
As with all Open Book publica� ons, this en� re book is available to read for free on the publisher's website. Printed and digital edi� ons, together with supplementary digital material, can also be found at www.openbookpublishers.com
*Cover image: Rudolf Lohbauer, 'Hyperions Fahrt nach Kalaurea' (1824). Cover design: Francesca Alabaster.* | doab | 2025-04-07T04:13:03.670404 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffc3321d-1b71-44dd-8de0-4fc5781eefa7",
"url": "https://library.oapen.org/bitstream/20.500.12657/25803/1/9781783746576.pdf||https://library.oapen.org/bitstream/20.500.12657/25803/1/9781783746576.pdf||https://library.oapen.org/bitstream/20.500.12657/25803/1/9781783746576.pdf||https://library.oapen.org/bitstream/20.500.12657/25803/1/9781783746576.pdf",
"author": "",
"title": "Hyperion or the Hermit in Greece",
"publisher": "Open Book Publishers",
"isbn": "",
"section_idx": 0
} | |
ffc4003d-750c-413c-b63b-035e5dac13ff | **Methods in Molecular Biology 1910**
## Maria Anisimova *Editor*
# Evolutionary Genomics
Statistical and Computational Methods
*Second Edition*
#### M ETHODS IN M OLECULAR B IOLOGY
Series Editor John M. Walker School of Life and Medical Sciences University of Hertfordshire Hatfield, Hertfordshire, AL10 9AB, UK
For further volumes: http://www.springer.com/series/7651
## Evolutionary Genomics
### Statistical and Computational Methods
### Second Edition
Edited by
## Maria Anisimova
Institute of Applied Simulations, School of Life Sciences and Facility Management, Zurich University of Applied Sciences (ZHAW), W€adenswil, Switzerland Swiss Institute of Bioinformatics, Lausanne, Switzerland
Editor Maria Anisimova Institute of Applied Simulations School of Life Sciences and Facility Management Zurich University of Applied Sciences (ZHAW) W€adenswil, Switzerland
Swiss Institute of Bioinformatics Lausanne, Switzerland
ISSN 1064-3745 ISSN 1940-6029 (electronic) Methods in Molecular Biology ISBN 978-1-4939-9073-3 ISBN 978-1-4939-9074-0 (eBook) https://doi.org/10.1007/978-1-4939-9074-0
This book is an open access publication.
© The Editor(s) (if applicable) and The Author(s) 2012, 2019.
Open Access This book is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this book are included in the book's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the book's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This Humana imprint is published by the registered company Springer Science+Business Media, LLC, part of Springer Nature.
The registered company address is: 233 Spring Street, New York, NY 10013, U.S.A.
#### Preface
This volume is a thoroughly revised second edition of Evolutionary Genomics: Statistical and Computational Methods published in 2012. Like the first edition, the new volume includes comprehensive reviews of the most recent and fundamental developments in bioinformatics methods for evolutionary genomics and related challenges associated with increasing data size, heterogeneity, and its inherent complexity.
Throughout the volume, prominent authors address the challenge of analyzing and understanding the dynamics of complex biological systems, and elaborate on some promising strategies that would bring us closer to the ultimate "holy grail" of biology uncovering of the relationships between genotype and phenotype. Consequently, the presented collection of peer-reviewed articles also represents a synergy between theoretical and experimental scientists from a range of disciplines, working together towards a common goal. Once again, the revised volume reiterates the power of taking an evolutionary approach to study molecular data.
This book is intended for scientists looking for a compact overview of the cutting-edge statistical and computational methods in evolutionary genomics. The volume may serve as a comprehensive guide for both graduate and advanced undergraduate students planning to specialize in genomics and bioinformatics. Equally, the volume should be helpful for experienced researchers entering genomics from more fundamental disciplines, such as statistics, computer science, physics, and biology. In other words, the material presented here should suit both a novice in biology with strong statistics and computational skills and a molecular biologist with a good grasp of standard mathematical concepts. To cater to differences in reader backgrounds, Part I is composed of educational primers to help with fundamental concepts in genome biology (Chapter 1), probability and statistics (Chapter 2), and molecular evolution (Chapter 3). As these concepts reappear repeatedly throughout the book, the first three chapters will help the neophyte to stay "afloat". The exercises and questions offered at the end of each chapter serve to deepen the understanding of the material.
Part II of this volume focuses on sequence homology and alignment—from aligning whole genomes (Chapter 4) to disentangling orthologs, paralogs, and transposable elements (Chapters 5 and 6). Part III includes chapters on phylogenetic methods to study genome evolution. Chapter 7 presents multispecies coalescent methods for reconciling phylogenetic discord between gene and species trees. However, a mathematically convenient "binary tree" model does not always live up to scrutiny as numerous evolutionary processes act in reticulate (network-like) fashion, complicating the statistical description of evolutionary models and increasing computational complexity, often to prohibitive levels. One simplification is to assume that some molecular sequence units (genes, gene segments) still evolve in a treelike manner. If so, Chapter 8 describes one practical approach to meaningfully summarize the binary tree distributions for a set of genomes as a "forest of trees". Alternatively network-like phylogenetic relationships can be represented by graphs (Chapter 9). Dating methods for genome-scale data are discussed in Chapter 10, while Chapter 11 provides more examples of non-treelike processes in a comparative review of genome evolution in different breeding systems.
By disentangling different evolutionary forces acting on genomes, we hope to understand the origins of biological innovation, which is often thought to be coupled with natural selection. After all, how do we explain that, by the words of Darwin, "from so simple a beginning endless forms most beautiful and most wonderful have been, and are being, evolved"? This is the main topic of Part IV that discusses the methodology for evaluating selective pressures on genomic sequences (Chapters 12–14) and genomic evolution in light of protein domain architecture and transposable elements (Chapters 15 and 16). Part V of this book is dedicated to population genomics and other omics, with example applications to disease. Indeed, as evolution starts in populations, there is much interest in generating and studying population genome data for a wide range of species. Chapter 17 discusses models for genetic architectures of complex disease and genome-wide association studies for finding susceptibility variants. Chapter 18 reviews approaches to study ancestral population genomics. Chapters 19, 20 and 21 illustrate first principles of analyzing environmental sequences and applications to clinical trials and systems genetics. Finally, Part VI concludes the book by discussing current bottlenecks in handling and analyzing genomic data. Chapter 22 focuses on challenges and approaches for large and complex data representation and simultaneous querying of heterogeneous databases. Chapter 23 makes the case for using efficient high-performance computing strategies for computationally demanding phylogenetic analyses, in particular in the Bayesian framework. Solutions for scalable workflows and sharing programming resources are presented in Chapters 24 and 25.
On behalf of all authors, I hope that this book will become a source of inspiration and new ideas for our readers. Wishing you a pleasant reading!
Wadenswil, Switzerland € Lausanne, Switzerland
Maria Anisimova
#### Acknowledgements
This renewed edition of Evolutionary Genomics: Statistical and Computational Methods is a result of a dedicated effort by 94 co-authors of the book representing research institutions from nearly two dozen different countries. Special thanks go to almost 50 independent reviewers whose constructive and detailed comments have greatly contributed to improving the overall quality of the book chapters and the clarity of the presentation. As for the first edition of this book, the cover image was made by the author of Chapter 6 and a talented photography artist, Wojciech Makałowski, from the University of Mu¨nster, Germany.
By a mutual agreement between all authors of the book, all chapters are available Open Access. Swiss Institute of Bioinformatics (SIB) and Zurich University of Applied Sciences (ZHAW) have generously contributed to cover a part of the Open Access publication fees. Finally, I would like to thank my colleagues at the Institute of Applied Simulations and the School of Life Sciences and Facility Management of ZHAW (Zurich University of Applied Sciences) as well as my family for their support and encouragement.
#### Contents
#### x Contents
#### PART IV NATURAL SELECTION AND INNOVATION IN GENOMIC SEQUENCES
#### Contributors
AMIT PANDE Institute of Bioinformatics, University of Muenster, Muenster, Germany
NICOLAS RODRIGUE Department of Biology, Carleton University, Ottawa, ON, Canada; Institute of Biochemistry, Carleton University, Ottawa, ON, Canada; School of Mathematics and Statistics, Carleton University, Ottawa, ON, Canada
## Part I
#### Introduction: Bioinformatician's Primers
## Chapter 1
#### Introduction to Genome Biology and Diversity
#### Noor Youssef, Aidan Budd, and Joseph P. Bielawski
#### Abstract
Organisms display astonishing levels of cell and molecular diversity, including genome size, shape, and architecture. In this chapter, we review how the genome can be viewed as both a structural and an informational unit of biological diversity and explicitly define our intended meaning of genetic information. A brief overview of the characteristic features of bacterial, archaeal, and eukaryotic cell types and viruses sets the stage for a review of the differences in organization, size, and packaging strategies of their genomes. We include a detailed review of genetic elements found outside the primary chromosomal structures, as these provide insights into how genomes are sometimes viewed as incomplete informational entities. Lastly, we reassess the definition of the genome in light of recent advancements in our understanding of the diversity of genomic structures and the mechanisms by which genetic information is expressed within the cell. Collectively, these topics comprise a good introduction to genome biology for the newcomer to the field and provide a valuable reference for those developing new statistical or computation methods in genomics. This review also prepares the reader for anticipated transformations in thinking as the field of genome biology progresses.
Key words Organism diversity, Viruses, Prokaryotes, Eukaryotes, Organelles, DNA, RNA, Protein, Regulatory DNA, Epigenetics, Plasmids, Transcription, Translation, DNA replication, Chromatin, Gene structure
#### 1 Introduction
Following the introduction of the concept of the genome in 1920 [1], the field of genome science has grown to encompass a vast range of interconnected topics (e.g., nucleic acid chemistry, molecular structure, replication and expression biochemistry, mutational processes, evolutionary dynamics, and interactions with cellular processes). Although the notion of the genome as a fundamental biological unit has been with us for nearly a century, it is only within the last decade that genomics has emerged as a transformative discipline within biology and the health sciences [2]. Its rapid development was in large part due to advances in massively parallel next-generation sequencing [3], which yielded unprecedented levels of genomic data. Those data revealed extensive natural
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_1, © The Author(s) 2019
variation in the way that genomes are structured and processed. This led modern biologists to reevaluate the fundamental definition of the genome.
The typical definition of the genome is often dualistic, referencing both structural features and its function to store and transmit biological information [4]. For example, the US National Institutes of Health (NIH) uses the following definition: "A genome is an organism's complete set of DNA, including all of its genes. Each genome contains all of the information needed to build and maintain that organism. In humans, a copy of the entire genome—more than three billion DNA base pairs—is contained in all cells that have a nucleus." This conception, as with many others, is structural with regard to physical features (viz., genes and DNA base pairs) and informational with regard to its role in carrying out cellular functions (viz., to build and maintain the organism). Through increased knowledge of genome diversity, the field has come to realize that both conceptions of the genome are sometimes insufficient [4]. We now understand that the physical structures of the genome can be transient and that the expression of information contained within a genome is often conditioned on non-genomic factors. The science of genome biology is entering a new era based on a deeper understanding of the relationship between genotype and phenotype [5].
The purpose of this review is to provide a condensed overview of genome biology and to anticipate transformations in thinking that will occur as the field progresses. The remainder of this article is structured into four parts, with the next section providing a brief overview of the diversity of organismal cell types. The two subsequent sections introduce the structural and informational aspects of genomes, respectively. In the final section, we reassess the definition of the genome through selected biological examples and conclude with an updated perspective on the nature of the genome as an informational entity.
#### 2 Organism Diversity and Cell Types
Cells are the smallest living unit of an organism. All cells have three attributes in common: cell membrane, cytoplasm, and genome. Structurally, cells can be divided into two basic types: prokaryotic and eukaryotic cells. Eukaryotic cells tend to be more complex. They possess a nucleus and other membrane-bound organelles, which are specialized components in the cell that perform unique functions (e.g., nucleus, mitochondria, plastids). Conversely, prokaryotic cells lack membrane-bound organelles. Although similar in cell structure, prokaryotes include two fundamentally distinct domains: the eubacteria (true bacteria, often referred to simply as bacteria) and the archaea.
Cellular life is detected in almost every environment on Earth. As life has colonized and adapted to the vast number of niches, cells have evolved an incredible amount of diversity in regard to size [6], form [7], lifestyle [8, 9], and complexity [10]. Understanding the basis of such diversity remains one of the central aims of biology. Readers interested in the latest understanding of Earth's biodiversity, the unique characteristics of its organisms, and how both extant and extinct forms are related to each other are encouraged to explore the following resources: the University of California Museum of Paleontology "History of life through time" exhibit [11], the Tree of Life Web Project [12], the Encyclopedia of Life [13].
2.1 Viruses Viruses are infectious agents of living cells that are unable to reproduce in the absence of a host. Viruses are not considered cellular entities since they lack two of the essential attributes that define a cell; they possess neither a cell membrane nor cytoplasm. The discovery of virophages, viruses that parasitize other viruses, resurrected the debate on their classification as living organisms [14]. Some consider viruses to be living entities since they can be hosts to other viruses, with a virophage infection leading to the eventual death of the host virus, implying an initial "living" state [15]. The opposing view asserts that a virus' inability to reproduce outside of a cellular host makes them nonliving entities [16, 17]. Irrespective of their delineation as living or nonliving, viruses are relevant to this review as they possess genomes and are the most abundant biological replicators in the biosphere [18].
> Outside of their host, viruses exist as viral particles (virions) consisting of a protein capsule that protects and encloses their genome. Once a virion has entered a host cell, it "hijacks" the host's cellular structures and processes to carry out the metabolically active phase of the viral life cycle. At this stage, the virus exhibits physiological properties reminiscent of living cells; they metabolize, grow, and reproduce. There is a wide range of viral lifestyles, with corresponding diversity in viral forms, sizes, hosts, and genomes [16]. The largest known virus, the mimivirus, was originally identified as an infectious agent of an amoeba [19] and can itself become a host for virophages [14]. To put this in context, the virion of a mimivirus can be larger than some prokaryotic cells [16]. At the other end of the scale are viruses such as the circoviruses, some of which have small genomes made up of less than 2000 nucleotides [20]. A more detailed account of viral diversity can be found at the ViralZone website [21].
2.2 Bacteria The bacterial cell is prokaryotic, and it is relatively simple as compared to eukaryotic cells. It has no membrane-bound organelles, and the chromosome (usually one) is not separated from the other components of the cell. While predominantly unicellular, they often live in biofilms, a community of cells bound together by a secreted polymer matrix [22], displaying a range of cooperative behaviors [23]. They can also exhibit regulated differentiation into different cell types, where two cells with the same genome have different morphology and function [22, 24].
Only a very small fraction of bacterial diversity (less than 1%) can be cultured and grown in the laboratory [25]. The problem of uncultivable bacteria is a consequence of our limited knowledge of their physiological diversity and the interactions necessary for their growth [26]. To this end, efforts are being made to study bacteria in nature [27–29] but with limited progress given the immense metabolic diversity of bacteria. Even within the incomplete sampling of cultivable bacteria, there is considerable diversity in cell shape [30], mode of reproduction [9], and cell cycle regulation [31].
The bacterial cell cycle involves the coordination of genome replication and segregation of replicated copies into daughter cells, followed by cell division. In this way, the transmission of genetic material is "vertical" from one cell generation to the next. Under certain conditions, some bacteria, such as E. coli, can initiate a new round of genome replication prior to completion of cell division [32, 33], thereby resulting in an increase in the number of gene copies near the origin of replication as compared to loci replicated later [31]. Other bacteria, such as Caulobacter, maintain a tightly regulated cell cycle to ensure a single replication event per division [34]. Under optimal conditions, some species can complete their cell cycle every 20 min, implying that a single cell could produce more than a billion descendants in a mere 10 h. In addition to vertical transfer, genetic information can be transferred "horizontally" between unrelated cells via the processes of transformation, conjugation, or transduction [35]. An event that transfers gene(s) between different species (or cells) by any of these three processes is referred to as a horizontal gene transfer (HGT) event.
2.3 Archaea Archaea are single-celled organisms that appear strikingly similar to bacteria under light and electron microscopes. Like bacteria they often have a single circular chromosome and lack a nucleus, and for a long period of time the archaea were wrongly categorized as bacteria. The first indication that the archaea might be a separate domain of life was obtained from phylogenetic analyses of the 16S rRNA gene [36]. Advancements in genome sequencing and analysis yielded further evidence of the evolutionary distinction between the bacterial and archaeal domains [37]. Despite their superficial cellular similarity to bacteria, the archaea have many molecular-level similarities to eukaryotes, leading researchers to hypothesize that the ancestor of the eukaryotes arose within the archaea [38].
> Previously, archaea were assumed to be a minor group of organisms inhabiting extreme environments beyond the tolerance of bacteria (salt brines, hydrothermal vents, acidic and anoxic
conditions, etc.). Through culture-independent methods, archaea were discovered to be much more widespread and metabolically diverse. Archaea are now known to inhabit the human gut, and through mutualistic community relationships, they play a key role in human health and metabolism [39–41]. There is increasing evidence for archaea playing a significant role in global nutrient cycling [42]. They contribute major mechanisms for anaerobic methane oxidation [42], ammonia oxidation [43], and other parts of the nitrogen cycle including nitrogen fixation [44]. The archaea also appear to be ecologically competitive with bacteria, as they make significant contributions to the microbial communities of non-extreme soil, aquatic, and marine environments [43, 45]. Although they can be highly abundant in such environments, archaeal diversity is greatest in the more extreme habitats [45].
Archaea possess an array of bacteria-like, eukaryote-like, and archaea-specific features. The archaeal cell wall is chemically and structurally diverse, yet they systematically lack a cell wall peptidoglycan, murein, that is ubiquitous among the bacteria [46, 47]. Their membrane lipids are chemically different from those found in either bacteria or eukaryotes [48], and they possess many novel enzymes that are required for the biosynthesis of their unique membranes [49, 50]. Consequently, most archeoviruses are unique to archaea [51]. Even structural appendages that initially appeared to be homologous to bacterial appendages are often structurally distinct and have different genetic basis than the bacterial counterparts [52–54]. At the biochemical level, the archaea use many sources of energy and are metabolically diverse, probably more so than either bacteria or eukaryotes [55].
#### 2.4 Eukaryotes All complex multicellular organisms are eukaryotes (animals, plants, fungi, red algae, and brown algae), as are many unicellular organisms [56, 57]. Eukaryotic cells are found in a wide diversity of sizes and shapes [58, 59]. They are generally larger and have a more complex internal organization than the bacteria and archaea. A key characteristic of the eukaryotic intracellular organization is the use of lipid membranes to separate their contents into different compartments [60, 61]. The bulk of the eukaryotic genetic material is surrounded by a nuclear envelope and is thus maintained in a separate organelle, the nucleus. This provides a fundamental perspective on how eukaryotic cells differ from bacterial and archaeal cells and has important consequences on the expression of eukaryotic genetic information.
In addition to the nucleus, other organelles (mitochondria and plastids) contain small genomes that encode additional genes. Both mitochondria and plastids originated from ancient endosymbiosis events between ancestral eukaryotic cells and bacterial organisms. Following these events, the invading bacteria underwent a process of genome reduction in which they transitioned from autonomous organisms to cell-dependent organelles [62].
Despite our familiarity with plants, animals, and fungi, the vast majority of eukaryotic diversity lies outside of those groups and is largely microbial [63]. These "other" eukaryotes are collectively called protists. They do not form a monophyletic group, i.e., protists do not from a phylogenetic group that is comprised of a shared common ancestor and all of its descendants [57, 64]. The term protist is used largely for convenience to classify all eukaryotes that are not plants, animals, or fungi. Protists embody extensive ecological and structural diversity and include several important groups of unicellular eukaryotes involved in human diseases [65]. For example, the unicellular apicomplexan eukaryote Plasmodium is the causative agent of malaria, which affects around 10% of the world population [65]. More positively, protist species are important primary producers and are an essential link in the ocean's biogeochemical cycles [66].
#### 3 Genome Structure and Organization
The notion of the gene as the physical carrier of hereditary information existed years before its physical and chemical structures were known. In 1902, Sutton provided the first clear support for the chromosomal theory of inheritance, allocating genes to segments on chromosomes [67]. The modern view of the gene is more often focused on a particular chemical sequence of nucleic acids rather than a chromosomal locus, but the two are not independent. The genetic instructions encoded within an organism's nucleic acid molecules comprise the organism's genotype. The physical manifestation of such genetic information, which will depend on environmental interactions, comprises the organism's phenotype.
There are two types of nucleic acids: deoxyribonucleic acid (DNA) and ribonucleic acid (RNA). Both are polymers consisting of chains of nucleotides. Each nucleotide includes three components: a 5-carbon sugar, a phosphate group, and a nitrogenous base. A nitrogenous base together with the sugar (without the phosphate group) is called a nucleoside. The sugar component in RNA, ribose, is a normal sugar with one hydroxyl group (OH) attached to each carbon atom. Deoxyribose, the sugar present in DNA, differs only in the absence of one oxygen atom at the 2<sup>0</sup> carbon atom (H instead of OH). This chemical difference is crucial for enabling enzymes to distinguish between RNA and DNA polymers. The 50 sugar carbon carries a phosphate group and is referred to as the 5<sup>0</sup> end of the polynucleotide molecule (DNA or RNA). The 3<sup>0</sup> end has a free hydroxyl (OH) group that is available to form chemical bonds with other atoms. As a result, synthesis of DNA and RNA in the cell proceed through the addition of a nucleotide to a 30 terminal hydroxyl group. The polynucleotides, therefore, exhibit directionality, and synthesis occurs in a 5<sup>0</sup> to 3<sup>0</sup> direction.
All living cells employ the double helical structure of DNA as a chemical means to store information. Each of the two longitudinal strands is an alternating sequence of phosphate and a 5-carbon sugar. At each sugar, the two strands are bridged by two nitrogenous bases, one purine molecule (of type adenine [A] or guanine [G]) and the other a pyrimidine molecule (of type cytosine [C], thymine [T], or uracil [U]). The chemical bridges between purine and pyrimidine molecules (called base pairs) are held together by hydrogen bonds. Each purine can be complemented by only one pyrimidine: A forms two hydrogen bonds with T (or U in RNA) and C forms three hydrogen bonds with G. These are referred to as the canonical or Watson-Crick pairings. Given this pairing pattern, the sequences of the double-stranded DNA are said to be complementary, and the sequence of one strand can be deduced from the sequence of its complementary strand. The order of the nitrogenous bases in DNA (or RNA) is what confers the meaning of the information encoded in the genome.
A vital feature of genetic information is its ability to be replicated and passed on to daughter cells. The core mechanisms that copy DNA are conserved in all three domains of cellular life: bacteria, archaea, and eukaryotes [68]. Accurate DNA replication is essential to produce viable offspring—too many alterations in the DNA impede the production of functional proteins, thereby increasing the chances of nonviable progeny. Therefore, most DNA replicates with high fidelity. However, mistakes do occur. In humans, on average one error occurs in 30 million bases copied per cell division [69]. The cells produced from these altered genes are called mutants.
Although all living things carry DNA, the processes through which genetic information is physically transferred from DNA to RNA (called transcription) and then used to create a polypeptide molecule with a unique sequence of amino acids (called translation) differ between domains of life. The lack of membrane-bound nuclei in prokaryotes permits the simultaneous occurrence of transcription and translation [70]. In eukaryotes, those processes are separated by the nuclear membrane; DNA is first transcribed to RNA in the nucleus, and the RNA product is subsequently translated to an amino acid sequence in the cytoplasm, ultimately leading to the construction of a protein.
Organisms from all domains of life, and many of the viruses that parasitize them, have a very large genome compared with the size of the cell or compartment to which it is confined. For instance, the human nuclear DNA consists of approximately three billion base pairs; when stretched out, it amounts to about 2 m of total DNA per cell. The average human cell size is merely 10 μm. The impressive ability to store DNA within the cell is possible through a process of genome packaging. In eukaryotes and some archaea, the DNA wraps around histone proteins to form nucleosomes. In humans, this results in a two-million-fold decrease in size, allowing the DNA to compact into the nucleus [68]. Prokaryotic DNA compaction is achieved using a combination of supercoiling, macromolecular crowding, and association with DNA-binding proteins [71]. The degree of the supercoiling used in prokaryotes varies considerably between different species.
Prokaryotic cells tend to have efficient genomes, with most of their genetic material composed of protein-coding regions. Archaeal genomes are, on average, more compact than bacterial genomes [72]. An increase in prokaryotic genome size is therefore often accompanied by an increase in the number of genes encoded. This trend is not evident in eukaryotes, for which there is little association between genome size and the number of proteincoding genes [73]. Consider the E. coli genome, more than 90% of its DNA encodes proteins. This is in stark contrast with the modest 2% protein-coding regions present in human DNA [74]. Most eukaryotic genomes are riddled with non-protein-coding regions (see Subheading 4.2 for an evolutionary mechanism). This results in them having larger genome sizes on average than prokaryotic cells [74].
3.1 Viral Genomes Viruses use any combination of either RNA or DNA, either singleor double-stranded molecules, in either circular or linear forms, to encode their genetic instructions [75, 76]. The viral genetic material is typically referred to as segments rather than chromosomes. Viral genomes composed of multiple segments are referred to as segmented. When different strains of the same segmented viral species infect a cell, genomes from the different strains can mix to produce hybrids—a process known as reassortment. Hybrid flus such as the H1N1 swine influenza A virus originated in this way [77].
> Viral strains package their genomes in various ways. Most DNA and RNA viruses with small genomes (<20 kb) employ energyindependent packaging systems where capsid assembly and genome condensation are coupled. One example is the RNA genome of the HIV retrovirus that, in the mature virion, forms a RNA-protein complex with one of the cleavage products of the Gag polyprotein [78]. Other viruses, such as the lambda bacteriophage, require ATP to pump their genome directly into a preassembled capsid [79]. The latter type of machinery is ubiquitous in bacterial viruses. Alternatively, large viruses package their genome using histone-like proteins that are critical for eukaryotic genome packaging [80]. For a review on genome packaging in viruses, see ref. 81.
3.2 Bacterial Genomes
Despite not being confined within a membrane-bound compartment, the prokaryotic genome will be unevenly distributed throughout the cell. It often clusters in an irregularly shaped viscous region known as the nucleoid that makes up about a quarter of the intracellular volume [82]. The organization and distribution of the nucleoid are dynamic and dependent on the growth rate and presence of antibiotics [83].
It was previously thought that all bacterial cells possessed a single circular chromosome. In 1989, the first linear bacterial chromosome was discovered in the spirochaete Borrelia burgdorferi, the causative agent of Lyme disease [84, 85]. Additionally, recent advancements have revealed that many cells retain multiple circular or linear chromosomes [86]. These often consist of a primary chromosome, which is larger and harbors a higher density of essential genes compared to the secondary chromosome(s) [87].
The replication of bacterial DNA initiates at a well-defined sequence, called the origin of replication. The proteins involved in replication bind to the origin site and DNA synthesis proceeds in both directions. Circular chromosomes require a single origin, and replication is terminated by either a stop signal or when the two replication forks meet [88]. Linear bacterial chromosomes typically have a central origin, and replication proceeds bidirectionally much as in circular chromosomes. However, replication enzymes are unable to synthesize new DNA at the ends of a linear chromosome, and this results in the gradual shortening of DNA after each replication event [89]. Linear chromosomes, therefore, require terminal structures known as telomeres to protect against DNA degradation. Telomeres are characterized by the presence of multiple tandem repeats of short noncoding nucleotide sequences.
Linear prokaryotic chromosomes have evolved two different types of telomeres [90]. The first, best understood in the streptomycetes, uses a terminal protein complex covalently attached to the 5<sup>0</sup> end of the DNA molecules. During replication, DNA polymerase binds the first synthesized nucleotide directly to the terminal protein. This replication strategy allows for the complete duplication of the linear molecule with no loss of genetic information [91]. The second type, best studied in the spirochetes, involves the formation of closed hairpin structures at the termini [92]. Replication of the linear DNA proceeds as expected. Once duplication of each DNA strand is completed the newly synthesized DNA are temporarily still attached—forming a structure superficially resembling a circular chromosome. A specific enzyme is then recruited to separate the two linear strands and re-form the telomeres [93]. For an overview of telomeric structures, see ref. 94.
3.3 Archaeal Genomes Archaeal genomes share features with both bacteria and eukaryotes. Archaea typically possess circular chromosomes reminiscent of bacteria genomes; some have a single chromosome and a single origin of replication, while other species have multiple chromosomes and multiple origins on each [95, 96]. Given that archaea have the prokaryote cell type that lacks membrane-bound organelles (and hence nuclei), they are similar to bacteria in permitting the simultaneous occurrence of transcription and translation. Nonetheless, there are fundamental differences from the bacteria in the processing of genomic information. The initiation of amino acid synthesis in archaea more closely resembles that used in the eukaryotic transcription process. Additionally, the core archaeal transcription machineries are more closely related to eukaryotes [97, 98]. Archaeal and eukaryotic DNA replication and repair systems have also been shown to have many features in common [99].
Relatively little is known about the structure of archaeal genomes [100], but some are packaged into chromatin via histone proteins. Chromatin is a compact and organized chromosome structure that consists of DNA in close association with proteins. Interestingly, this form of chromatin is present in all eukaryotes and missing from bacteria [101]. Among the archaea that use histones (i.e., Thermoproteales and Euryarchaea), the geometry of their histone-mediated chromatin is the same as in eukaryotes [102]. However, archaeal histones are often shorter than the eukaryotic histones [101]. Groups of archaea that lack histones (e.g., Crenarchaea) encode other DNA-binding proteins associated with the architecture of bacterial chromatin [100]. Another family of DNA-binding proteins called Alba (acetylation lowers binding affinity) is ubiquitous among archaea. They are abundant small proteins that facilitate genome compaction, play a key role in determining the architecture of archaeal chromatin, and regulate gene expression on a genomic scale [101]. Alba proteins have been detected in both histone-lacking and histone-containing archaea [103].
3.4 Eukaryotic Genomes Eukaryotes sequester their linear chromosomes within a membrane-bound nucleus. Linear eukaryotic chromosomes have three essential structural elements: a centromere, a pair of telomeres, and origins of replication. The centromere is the attachment point for spindle microtubules—the filaments responsible for physically moving chromosomes during cell division. Telomeres are the protective ends of a linear chromosome. The origins of replication are the sites where DNA synthesis begins. Eukaryotes typically have multiple linear chromosomes, each with many origins of replication. The larger genome size and slower replication machinery in eukaryotes necessitate the need of multiple origins to speed up the replication process.
> In eukaryotic cells, nuclear DNA compaction involves the association of DNA with the protein products of a family of genes, the histones, whose sequence variants provide for a variety of different functions. The eukaryotic chromosome is organized at the lowest
level by wrapping the DNA around histones, forming nucleosomes. This structure constitutes the basic unit of the chromatin fiber, which is further organized into higher-order structures mediated by other proteins [104, 105]. Sequence variation in histones, in combination with posttranslational modification of the protein, affects the structural properties of chromosomal nucleosomes and gene expression.
Eukaryotic DNA consists of at least three types of sequences: unique-sequence DNA, moderately repetitive DNA, and highly repetitive DNA. Unique-sequence DNA are regions that are present only once or at most a few times in the genome. Most proteincoding regions fall within this category. Alternatively, more than half of the total DNA in all eukaryotic genomes is made up of repeated sequence motifs that are either moderately or highly repetitive [106]. Moderately repetitive DNA are sequences from 160 to 180 base pairs (bp) in length that are repeated thousands of times [106]. Some of these sequences perform important functions for the cell, such as coding for types of RNA [107]. Highly repetitive DNA are short sequences, less than 60 bp that are present in hundreds of thousands of copies repeated throughout the genome. Repeats that are 2–10 bp are known as microsatellites, whereas motifs that are 10–60 bp are termed minisatellites [108].
Most of the repetitive sequences arise through transposition (see Subheading 4.2). The repeated sequences can be found either in tandem arrays, i.e., appearing adjacent to each other, or interspersed throughout the genome. The evolution and maintenance of nonfunctional repeated sequences have spurred the interest of genome scientists, with some classifying these motifs as selfish-genes that reproduce to propagate themselves and provide no positive contribution to the organism's phenotype or fitness [106]. Repeats also represent technical challenges for bioinformaticians developing software for sequence alignment and genome assembly. From a computational perspective, repeats create ambiguities that are challenging to resolve. For a review on computational challenges and solutions, see ref. 108.
3.5 Auxiliary DNA Structures Both prokaryotes and eukaryotes have secondary chromosomal structures. For eukaryotes, this refers to any form of DNA found outside of a nucleus—although the discovery of microDNA extends this classification [109]. Eukaryotic auxiliary DNA often contains essential genes that are necessary for normal cell production. For example, the DNA chromosome located within the mitochondrial organelle encodes genes that are involved in oxidative phosphorylation and the creation of different types of RNA [110]. For prokaryotes, auxiliary DNA refers to any DNA that is not associated with the primary chromosome, and unlike eukaryotes, the genes encoded in such DNA are often dispensable. For example, small circular chromosomes, called plasmids, often contain genes that allow the bacterium to survive various environmental conditions; however they are not usually essential for normal cell function [110].
3.5.1 Mitochondrial DNA The mitochondrion is a double membrane-bound organelle that is ubiquitous in eukaryotic cells. There is only one known case of a eukaryotic cell able to survive without a mitochondrion [111]. Mitochondria are essential because they are the site of production for most of the cell's energy, which is produced as ATP by the oxidative phosphorylation metabolic pathway. Additionally, the mitochondrion is the site of iron-sulfur (Fe/S) cluster assembly. Fe/S clusters are protein cofactors that are essential for various extramitochondrial pathways [112]. The mitochondrialacking eukaryote, a species of Monocercomonoides, is unique in that it lives only within the intestine of the chinchilla and has evolved different strategies for Fe/S cluster formation and obtaining energy absorbed from its environment [111].
> Mitochondria are the derivatives of prokaryotic cells that were engulfed by a common ancestor of all eukaryotes. The DNA within these organelles are the remnants of the DNA genome of the ancestral prokaryotic endosymbiont. Thus, the mitochondrial DNA (mtDNA) more closely resembles a prokaryotic genome. For example, in most animals and fungi, mtDNA consists of a single circular chromosome. However, small linear mtDNA chromosomes with defined telomeres have been identified within various protists, animals, and fungi [113, 114]. Additionally, the architecture of mtDNA is not determined by histones but instead by a set of small DNA-binding proteins that induce structures analogous to the bacterial chromatin. Mitochondrial genomes have been categorized into six different types depending on shape, size, structure, and number (see ref. 115).
> In humans, the mitochondrial genome encodes 13 of the 80 proteins that are directly involved in oxidative phosphorylation. The remaining proteins are encoded in the nuclear chromosomes [110]. The exact contribution from mitochondrial and nuclear genomes varies across eukaryotes. Nonetheless, in the vast majority of known eukaryotic species, the mtDNA is essential to produce important proteins involved in energy production, demanding that all cells have faithfully inherited the mtDNA.
3.5.2 Plastid DNA Plastids are similarly derived from an endosymbiosis with a bacterium, with the organelle retaining remnants of that ancestral bacterial genome. Like the mitochondrion, the plastid is a double membrane-bound cytoplasmic organelle. Unlike the mitochondrion, plastids often contain pigment used in photosynthesis. Plastids are found in the cytoplasm of protists and all higher plants. Plastid DNA (ptDNA) is highly reduced relative to the genomes of extant photosynthetic bacteria. In part, the reduction in genome size is due to gene loss with some regions excised and incorporated into the host nuclear DNA [116]. The ptDNA encodes important proteins that are essential for cell viability [117]. Almost all plastids have circular DNA, with the alveolate Chromera velia being the single known case of linear ptDNA. The linear extrachromosomal ptDNA has a telomere arrangement resembling those of linear mtDNA [117, 118].
Genes encoded in ptDNA are involved in the synthesis and storage of various cellular components, including those necessary for photosynthesis. Plastids have diverged to carry out different functions with multiple types identified. For example, chloroplasts are specialized for carrying out photosynthesis; chromoplasts contain pigments that provide petal colors, whereas amyloplasts are used for bulk storage of starch [117].
3.5.3 Nucleomorph DNA A nucleomorph is a vestigial eukaryotic nucleus found in cryptomonads and chlorarachniophytes, which are both plastidcontaining algae. The nucleomorph is located in these organisms between the inner and outer membranes of the plastid and is believed to be derived from the nucleus of an endosymbiotic algal cell engulfed by a larger eukaryotic cell [119]. Thus, the plastid organelle in this case evolved from two endosymbiotic events: a prokaryote was engulfed by a eukaryote which thereby became photoautotrophic and that cell was then engulfed by another eukaryote. The nucleomorph genomes are extremely small compared to the typical nuclear genome, being comprised of mostly single-copy housekeeping genes and having no mobile elements. The nucleomorph genome of the cryptomonads suggests that it was derived from a red algal ancestor, whereas the nucleomorph genome of the chlorarachniophytes suggests a green algal ancestor [119].
3.5.4 Plasmid DNA Plasmids are present in bacteria, archaea, and eukaryotes [120]. Most plasmids are circular, although linear plasmids have been identified [121]. The genes carried on plasmids tend to be associated with functions that enable or enhance survival and growth under specific conditions. They can be horizontally transferred between prokaryotic cells and represent an important vehicle for sharing genetic information [122]. For example, a plasmid that has evolved an antibiotic resistance gene(s) can be transferred to neighboring bacteria promoting their rapid adaptation to various stresses associated with an antibiotic environment.
> The eubacteria E. coli is estimated to have more than 270 plasmids having different distributions among and within cells; some promote mating, while others contain genes that kill other bacteria. The number of plasmids known and sequenced is much higher in bacteria as compared to archaea, with the lowest number having
been identified in eukaryotes [122]. In recent years, plasmids have been used extensively in genetic engineering as a means of introducing and modifying target genes [122, 123].
3.5.5 MicroDNA In 2012, Shibata et al. discovered a new form of extrachromosomal DNA in eukaryotes, called microDNA [122]. In contrast with other auxiliary DNA, microDNA is derived from non-repetitive sequences that are often associated with functional genes. They are circular DNA between 200 and 400 bp and are found in the nuclei of mammalian cells [122]. microDNA is thought to be associated with the repair and maintenance processes of nuclear DNA. It is not yet clear if microDNA plays a functional role in these processes or if they are merely an unavoidable by-product. For the time being, detection of specific microDNA is being proposed as a screening measure to aid the successful eradication of tumors in humans and as a potential method for cancer diagnosis and prognosis [124].
#### 4 Genomic Storage and Processing of Information
It was not possible to understand how hereditary information was encoded and transmitted across generations without first having knowledge of the structure of DNA. Knowledge of DNA structure led to a structure-oriented conception of genomes as linear sequences of ordered nucleotides. Once protein synthesis was linked to gene sequences, the structural view of the genome began to be supplanted by the informational view [125]. Genetic information was initially viewed as a static property belonging to the specific sequence of ordered subunits. However, others have argued that the static view of information is not satisfactory (e.g., [4, 125]). Barbieri [125] contends that "it is only when a sequence provides a guideline to a copymaker that it becomes information for it. It is only an act of copying, in other words, that brings organic information into existence." Based on Barbieri's viewpoint, information is not always a property of a specific structure (e.g., DNA or RNA); rather his view is that such molecules are information relevant only when they are used to perform a biological function. A DNA sequence, for example, is said to have information if it is transcribed or interacts with a protein in a biologically relevant way. Similarly, an mRNA transcript also encodes information as it is translated into a protein. Also then, a protein could be viewed as an informational entity in the sense that it is necessary to carry out a biological function. Therefore, under this new conception, as well as the static view, it is clear that biological information can be manifest in different biological molecules; an observation that has complicated the notion of the genome as the fundamental unit of biological information [4].
We now understand that storage of the genetic information required to sustain life does not need to be restricted to biological molecules. This was vividly illustrated in the laboratory when a bacterial genome was chemically sequenced, its information stored within a computer (a completely different medium composed of binary states), then resynthesized in the form of a new DNA chromosome, and that synthetic DNA ultimately used as the sole means to maintain a living cell [126]. Although the information required for life can be stored independently of the chemical structure of the DNA, it cannot be expressed in a biologically useful form without various proteins and RNA molecules. Thus, expression of information encoded within a genome (bringing that information into existence) is contingent on its cellular context. In this section, we examine different ways in which information may be contained within a genome and mechanisms that result in biologically useful expression of that information.
4.1 Gene Expression Mere knowledge of the DNA sequence of a genome is often insufficient to predict phenotype. The amount and timing of gene expression play a key role. For example, human cells with a nucleus have copies of almost identical DNA sequences. Yet cells perform varying functions, and they organize to create the multiple organs that constitute the human body. Cells achieve this primarily by differentially regulating the rate of transcription and/or translation of genes.
> DNA transcription and protein translation comprise elementary levels of information transfer from genotype to phenotype. Maintaining control of these processes is fundamental for all organisms. Genetic elements involved in regulating gene expression are referred to as regulatory elements. They often represent sequences found on the DNA or RNA. In this way, regulatory information can be encoded directly within the nucleic acid sequence. Direct structural proximity if often not necessary, as regulatory elements may be found proximal or distal to the genes they affect. In humans, approximately 8% of nuclear DNA is composed of elements involved directly in regulation such as promoters, enhancers, silencers, and insulators (defined in Subheading 4.1.1; [127, 128]).
> If all genetic and regulatory information is encoded in the DNA sequence, why can't any cell with a complete genome be used to produce a viable organism? The specificity of cells suggests that additional regulatory markers also exist outside of the primary DNA sequence. This type of regulation is epigenetic (above the genes) and is essential for normal development. Epigenetic information is derived from chemical modifications of the chromosome (e.g., DNA methylation or histone modification) that do not change the primary sequence of chromosomal DNA and can be
passed from one generation to the next [129, 130]. It is only through the collective actions of all cellular processes that gene products contribute to biochemical pathways and participate in the network of regulatory interactions to produce a complex organism or phenotype.
4.1.1 Transcriptional Regulation DNA transcription is the chemical process through which information is transferred from DNA to RNA. The transcribed RNA may itself carry out some biological function or may be part of an intermediate information-carrying class of RNA known as messenger RNA (mRNA). mRNA along with other RNA molecules (tRNA and rRNA) are part of the machinery used to synthesize proteins. The flow of genetic information from DNA to RNA to protein is present in all forms of life. However, it is important to note that information transfer is not exclusively unidirectional. The enzyme reverse transcriptase can transfer genetic information from an RNA template into DNA.
> The basic model of transcriptional regulation requires that regulatory proteins called transcriptional factors (TFs) bind specific DNA sequences in regulatory modules (RMs). TFs are protein products that are themselves subjected to regulation of gene expression. RMs are defined according to both the primary DNA sequence to which TFs bind and their role in the process of regulating gene expression. One type of RMs are promoters. They are specific motifs on DNA that are necessary regulatory elements for RNA transcription in prokaryotes and eukaryotes. They bind the basal transcriptional machinery, RNA polymerase and general TFs. Enhancers are RMs that bind activator proteins and enhance the affinity of RNA polymerase to the promoter region. They, therefore, result in an upregulation of transcription of a gene or set of genes. Enhancers often act by stabilizing RNA polymerase binding through structural histone modifications [131]. Silencers are regulatory elements that when bound to repressor proteins function to prevent gene transcription. Silencers and enhancers are often distance-independent, meaning that they can act on gene(s) that are proximal or distal to their location [132]. Enhancers can be thought of as on-switches for gene expression, whereas silencers are the off-switches.
4.1.2 Translational Regulation The fate of all mRNAs, transcribed from protein-coding genes, is not the same. The mRNA is often subjected to translational regulation depending on cellular and environmental conditions. These regulatory mechanisms affect the rate of protein synthesis. In prokaryotes and eukaryotes, most translational regulation involves structural changes in the mRNA molecule that impact its accessibility [133, 134]. The mRNAs can be sequestered in stress granules or localized in specific regions of a cell's cytoplasm [135–137]. Another mechanism of translational regulation is RNA interference (RNAi). This regulation strategy is common in eukaryotes and involves short noncoding RNAs—microRNA (miRNA) or small interfering RNA (siRNA)—that bind with imperfect complementarity to their target mRNA transcripts. The binding of miRNA (or siRNA) to mRNA destabilizes (or degrades) the target mRNA, thereby inhibiting its translation. The imperfect pairing allows a single RNAi molecule to affect the expression of multiple genes. In the human genome, almost 50% of mRNA transcripts are regulated by one or more miRNAs [138].
In prokaryotes, transcription and translation are more tightly coupled than in eukaryotes, and this allows prokaryotes to regulate their gene expression primarily by controlling the amount of transcription. Nevertheless, prokaryotes can still conduct translational regulation. They can employ fundamentally different types of translational regulatory machinery: the recently discovered CRISPR-Cas system. Although the CRISPR loci were first identified in prokaryotes in 1987 [139], it was only recently described as a bacterial immune defense system [140]. The CRISP-Cas system is most commonly known to target external DNA (viral or plasmid) and degrade it before it can be transcribed or translated. Recent advancement suggests that some CRISPR-Cas systems are more general and have the capacity to target RNA molecules. This was first discovered in Pyrococcus furiosus [141]; similar RNA targeting was later found in Sufolobus solfataricus [142]. Throughout these advancements, CRISPR-Cas system was still strictly viewed as an immune response to target and degrade external nucleic acid molecules. It was only in 2016 that a CRISPR-Cas system was discovered that targets cellular mRNAs and thereby participates in translational regulation [143].
4.1.3 Epigenetics The term epigenetics was coined in 1942 by Waddington [144]. He defined it as changes in an organism's phenotype without an underlying alteration of its genome. It is now understood that epigenetic effects cause variation in phenotypes not associated with a change in the primary sequence but by chemical alterations of the DNA. Consider this analogy: throughout this review, whenever a word was being defined it was written in this format. If this chapter was rewritten with all bolds and italics removed, the informational content would be unaltered; however, the emphasis would be different. These "decorative" changes in font are akin to chemical epigenetic markers appended to the DNA. DNA methylation is a type of chemical decoration that is analogous to striking through a phrase. Specifically, it corresponds to the addition of a methyl group to parts of the DNA that results in gene silencing [145]. This additional information is not directly encoded within the primary DNA sequence but is manifested through chemical changes in nucleotides [145]. Thus, DNA methylation is one form of epigenetic control of gene expression. Epigenetic factors may also have an impact on regulation by changing protein-DNA binding. In eukaryotes, epigenetic factors may bind to consecutive histones moving them closer to each other. This results in local DNA compaction and prevents the expression of the gene(s) in this location.
Importantly, an organism's exposure to certain environmental conditions can impact the epigenetic markers on its genome. Because epigenetic mechanisms ultimately affect the physiological form of the chromosome, such environmental exposures can lead to heritable changes in gene expression with no change to the underlying DNA sequence. It was initially thought that these alterations are not heritable and that following fertilization all epigenetic markers are removed from the zygote genome. Accumulating evidence suggests that such erasure of epigenetic marks occurs for most but not all genes [129, 130].
4.2 Mobile Genetic Elements Also known as transposons or jumping genes, mobile genetic elements are sequences that can move around within a genome independently of the complex networks which otherwise regulate gene expression [146]. Through their movement, transposons often cause mutations either by inserting into a gene and disturbing its function or by promoting DNA rearrangement. If a transposon is inserted within a protein-coding region, then it will undoubtedly affect the expression of this gene by altering the final protein product. Transposons may also be inserted into regulatory regions resulting in over- or under-expression of certain gene(s). The capability of these DNA sequences to produce new copies of themselves elsewhere in a genome is called transposition. The two types of transposition are:
Transposable elements are found in all cell types. The kinds of transposable elements vary within and between prokaryotes and eukaryotes. They are often viewed as genetic parasites since they rely on a host cell for information processing systems (replication, transcription, and/or translation). In humans, about 44% of the genome is comprised of sequences that are related to transposable elements [148]. These mobile genetic elements had an important impact on eukaryotic evolution [149, 150]. For example, siRNA regulation is believed to have evolved to regain control of the expression of transposable elements [151]. For a review of the regulatory mechanisms of transposable elements, see ref. 152.
#### 5 The Role of the Genome as an Informational Entity in Biology
Although the information contained within a genome is necessary to maintain a living cell, it is not sufficient on its own. Expression of biologically useful information requires a complex network of cellular components for processing and regulation of the genome. This dependency on external cellular components permits considerable flexibility in how the information is stored. As we have seen, the information essential for eukaryotic life is partitioned between chromosomes located in nuclear and organelle compartments, with some nuclear-encoded proteins being transported to the organelle for assembly with other proteins synthesized within the organelle [110]. Thus, as long as the cellular mechanisms for expression and processing are in place, genomic information can be physically dispersed within the cell. The Cryptophytes have taken this to an extreme, having their genomic information distributed across four cellular compartments: the nucleus, nucleomorph, mitochondria, and plastids [153]. Clearly, the physical location of the genome is not a constraint to information storage and processing. Furthermore, the storage of that information need not remain in a particular physical location. In the case of temperate phages, genomic information is transferred, for a period of time, to the genome of its host where it is maintained by its host's replication processes [154]. These examples, and others (e.g., [126]), underscore the importance of viewing the genome foremost as an informational entity irrespective of its physical location.
In a well-argued critique of conventional notions of the genome, Goldman and Landweber [4] argue that viewing DNA as the sole source of information leads to additional difficulties. Recall that the NIH definition refers to the genome as containing all of the information needed to build and maintain that organism. We now understand that even the cell and its associated cytoplasm are not always sufficient for realization of all functional capabilities encoded within a genome. In other words, the genome, as conventionally defined, appears to be an incomplete informational entity [4]. Genome research has identified a variety of extracellular informational entities that can influence, and in some cases are even essential to, the creation and maintenance of an organism. Below we review selected examples of this phenomenon prior to reassessing the definition of the genome in light of modern genome science.
Marine cyanobacteria (Prochlorococcus and Synechococcus) are among the most abundant photosynthetic organisms in the world's oceans. The viruses that infect them (cyanophages) were discovered to possess copies of some of their hosts photosynthesis genes (e.g., PsbA and PsbD: [155, 156]). Through the process of HGT, the cyanophages acquired host genes, which they express after infection to optimize their own gene expression and broaden their host range [157]. As novel as this discovery was, it was completely unexpected that the cyanobacteria and their phages continued to exchange genetic variation through homologous recombination [157]. Through such exchanges, the PsbA and PsbD genes participate in gene pools that extend beyond the photosynthetic species boundaries [157]. Given that cyanobacteria contribute as much as 30% of carbon fixation worldwide, those findings suggest that viral gene pool dynamics have influenced the evolution of oceanic photosynthesis on a global scale. This case demonstrates that to fully understand the origin and distribution of photosynthetic diversity, one must be aware that relevant genetic information can reside outside of the genomes of the photosynthetic organisms.
The bacterial genus Listeria is comprised of ecologically divergent lineages that share gene pools through the process of homologous recombination [158, 159]. Listeria monocytogenes is a pathogen closely related to the nonpathogenic species L. innocua. L. monocytogenes evolved as a pathogen through the process of HGT [160] and then subsequently evolved into ecologically divergent lineages differing in population structure and ability to respond to environmental stress [161]. Among Listeria, recombination is frequent enough to permit natural selection to act independently of the variability present at unlinked loci, thereby promoting or impeding exchangeability of genes among species and ecotypes residing in different niches [159]. This is just one example of the "mosaic genome" model of prokaryotic genome evolution, where the combined effects of recombination, drift, and selection lead to genomes comprised of a mosaic of differentially extendible trans-species gene pools. A wide variety of bacterial species are now thought to have genome dynamics consistent with the mosaic genome model [159, 162–165]. In some cases, the process of genomic divergence can even become decoupled from the process of ecological divergence [159, 163]. Thus, the physical genomes of some species of prokaryotes are incomplete informational entities.
The single-celled stichotrichous ciliates Oxytricha and Stylonychia have two nuclei that store genomic information in very different forms [166]. One nucleus, called the macronucleus, contains information in the form required for growth and maintenance of a cell. Hence, the macronuclear DNA is often referred to as "active." The second nucleus, called the micronucleus, contains the same information in a "stored" form, which is used to produce the active form of the DNA in the next generation. However, information storage in the micronucleus is extremely complex. Protein-coding genes expressed by the macronucleus are partitioned into small segments, inverted, and scrambled among ~1 GB of other DNA sequences within the micronucleus. Furthermore, the production of a working macronucleus in the next generation cannot be accomplished without information contained within both small RNA molecules (piRNA) and long RNA templates (lncRNA), which are passed across generations via the cytoplasm of the maternal macronucleus [167, 168]. The piRNA are crucial to the elimination of DNA during the development of an active macronucleus, and the lncRNA mediate (1) unscrambling of the inactive micronuclear DNA, (2) regulation of gene dosage in the macronucleus, and (3) epigenetic transfer of somatic (macronuclear) alterations that are not found within the germ-line (micronuclear) DNA [167]. Thus, without those RNA molecules, the DNA genome of the stichotrichous ciliates is an incomplete informational entity [4]. Furthermore, emerging work on both Oxytricha and Stylonychia suggests that epigenetic modification of their DNA may play a role in the production of active macronuclear DNA [166, 169–171]
Complex microbial communities live in close association with the human body and have a strong impact on human health and disease. Host genetic variation is known to influence the composition of those communities [172], and, conversely, microbial variability is thought to influence various host disease states [173]. This association is so intimate that the microbiome has been referred to as an additional "human organ" [174], and substantial amounts of missing heritability associated with many complex human diseases are now being attributed, in part, to a failure to adequately account for microbial genetic variation [175]. Taking inflammatory bowel disease (IBD) as an example, host human genetic variation accounts for less than 50% of its estimated heritability [176]. This result implies that there exists undiscovered context dependence of human genetic variation for IBD. We have since come to understand that there is extensive inter-individual variation in the genetic composition of the gut microbiome and this metagenomic variation can influence healthy and dysregulated human immune responses [177] and is predictive of IBD patient outcomes [178]. Because the development of the IBD phenotype is related to gut microbiome variability, and because genetically similar human hosts can have different microbiomes, heritability estimates for human DNA variation will be impacted [175]. In other words, the expression of similar IBD phenotypes in humans is a function of both human and microbial genetics. Regardless of whether such interactions should be formally included within any future conception of the genome, this example illustrates how the human genome is also an incomplete informational entity with respect to prediction of healthy and disease states.
Goldman and Landweber [4] suggest that the notion of the genome should be reconceptualized in light of our modern, and deeper, understanding of genomic diversity and the mechanisms of information storage and processing. We agree and follow Goldman and Landweber [4] when they call for a "more expansive definition of the genome as an informational entity, often but not always manifest as DNA, encoding a broad set of functional capabilities that, together with other sources of information, produce and maintain the organism." At first glance, this appears to be consistent with the controversial idea that a collection of functionally integrated organisms, called a holobiont, is a fundamental unit of biological organization and their set of genomes, called a hologenome, is itself a unit subject to evolution by natural selection [179]. However, we cannot go this far. We expect that any hologenome composed of informational entities having even a little independence is analogous to intra-genomic epistasis with just a little recombination. In the latter case, adaptive coevolution is not very effective at moving the system on its fitness landscape via compensatory substitutions [180]. Further, when informational entities are largely independent, either through high recombination (as observed in Listeria) or through independent replication (as within the human gut microbiome), the process of genomic divergence can become decoupled from ecological dynamics. Thus, we cannot agree with the notion of the hologenome as a unit of selection. Rather, we view the genome as a potential mosaic of gene pools subject to different evolutionary dynamics, and we follow Goldman and Landweber [4] by considering it foremost as an informational entity, which may be incomplete and which does not have to manifest exclusively as the DNA within a species boundary.
#### References
in animals and humans. Crit Rev Microbiol 40 (2):97–116
metabolism in archaea. Curr Opin Microbiol 8(6):695–705
quiescence. J Biol Chem 289 (16):11342–11352
the human genome. Nature 489 (7414):57–74
interference-based immune system in prokaryotes: computational analysis of the predicted enzymatic machinery, functional analogies with eukaryotic RNAi, and hypothetical mechanisms of action. Biol Direct 1(1):7
Proc Natl Acad Sci U SA 104 (35):14092–14097
wide association studies. Inflamm Bowel Dis 21(6):1428
structure are associated with sustained remission after nutritional treatment of pediatric Crohn's disease. Inflamm Bowel Dis 22 (12):2853–2862
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Chapter 2
#### Probability, Statistics, and Computational Science
#### Niko Beerenwinkel and Juliane Siebourg
#### Abstract
In this chapter, we review basic concepts from probability theory and computational statistics that are fundamental to evolutionary genomics. We provide a very basic introduction to statistical modeling and discuss general principles, including maximum likelihood and Bayesian inference. Markov chains, hidden Markov models, and Bayesian network models are introduced in more detail as they occur frequently and in many variations in genomics applications. In particular, we discuss efficient inference algorithms and methods for learning these models from partially observed data. Several simple examples are given throughout the text, some of which provide the basis for models that are discussed in more detail in subsequent chapters.
Key words Bayesian inference, Bayesian networks, Dynamic programming, EM algorithm, Hidden Markov models, Markov chains, Maximum likelihood, Statistical models
#### 1 Statistical Models
Evolutionary genomics can only be approached with the help of statistical modeling. Stochastic fluctuations are inherent to many biological systems. Specifically, the evolutionary process itself is stochastic, with random mutations and random mating being major sources of variation. In general, stochastic effects play an increasingly important role if the number of molecules, or cells, or individuals of a population is small. Stochastic variation also arises from measurement errors. Biological data is often noisy due to experimental limitations, especially for high-throughput technologies, such as microarrays or next-generation sequencing [1, 2].
Statistical modeling addresses the following questions: What can be generalized from a finite sample obtained from an experiment to the population? What can be learned about the underlying biological mechanisms? How certain can we be about our model predictions?
In the frequentist view of statistics, the observed variability in the data is the result of a fixed true value being perturbed by
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_2, © The Author(s) 2019
random variation, such as, for example, measurement noise. Probabilities are thus interpreted as long-run expected relative frequencies. By contrast, from a Bayesian point of view, probabilities represent our uncertainty about the state of nature. There is no true value, but only the data is real. Our prior belief about an event is updated in light of the data.
Statistical models represent the observed variability or uncertainty by probability distributions [3, 4]. The observed data are regarded as realizations of random variables. The parameters of a statistical model are usually the quantities of interest because they describe the amount and nature of systematic variation in the data. Parameter estimation and model selection are discussed in more detail in the next section. In this section, we first consider discrete, and then continuous random variables and univariate (1-dimensional) before multivariate (n-dimensional) ones. We start by formulating the well-known Hardy–Weinberg principle [5, 6] as a statistical model.
Example 1 (Hardy–Weinberg Model): The Hardy–Weinberg model is a statistical model for the genotypes in a diploid population of infinite size. Let us assume that there are two alleles, denoted A and a, and hence three genotypes, denoted AA, Aa ¼ aA, and aa. Let <sup>X</sup> be the random variable with state space X¼f AA, Aa, aa} describing the genotype. We parametrize the probability distribution of X by the allele frequency p of A and the allele frequency <sup>q</sup> ¼ <sup>1</sup> p of a. The Hardy–Weinberg model is defined by:
$$P(X = \text{AA}) = p^2,\tag{1}$$
$$P(X = \text{Aa}) = \mathcal{Q}p(1-p),\tag{2}$$
$$P(X = \text{aa}) = \left(1 - p\right)^2. \tag{3}$$
The parameter space of the model is <sup>Θ</sup> ¼ fp<sup>∈</sup> j <sup>0</sup> <sup>p</sup> <sup>1</sup>g¼½0,1, the unit interval. We denote the Hardy–Weinberg model by HWðpÞ and write <sup>X</sup> HWðpÞ if X follows the distribution (Eqs. 1–3). □
The Hardy–Weinberg distribution P(X) is a discrete probability distribution (or probability mass function) with finite state space: We have 0 <sup>P</sup>(<sup>X</sup> ¼ <sup>x</sup>) 1 for all <sup>x</sup>∈X and <sup>P</sup> <sup>x</sup>∈X <sup>P</sup>ð<sup>X</sup> <sup>¼</sup> <sup>x</sup>Þ ¼ <sup>p</sup><sup>2</sup> þ <sup>2</sup>pð<sup>1</sup> <sup>p</sup>Þþð<sup>1</sup> <sup>p</sup>Þ 2 ¼ ½<sup>p</sup> þ ð<sup>1</sup> <sup>p</sup>Þ<sup>2</sup> ¼ 1. In general, any statistical model for a discrete random variable with n states defines a subset of the (<sup>n</sup> -1)-dimensional probability simplex:
$$\Delta\_{n-1} = \left\{ (\mathfrak{p}\_1, \dots, \mathfrak{p}\_n) \in [0, 1]^n \mid \mathfrak{p}\_1 + \dots + \mathfrak{p}\_n = 1 \right\}.\tag{4}$$
The probability simplex is the set of all possible probability distributions of X, and statistical models can be understood as specific subsets of the simplex [7].
The Hardy–Weinberg distribution is of interest because it arises under the assumption of random mating. A population with major allele frequency p has genotype probabilities given in Eqs. 1–3 after one round of random mating. We find that the new allele frequency:
$$p' = P(\text{AA}) + P(\text{Aa})/2 = p^2 + 2p(1-p)/2 = p,\tag{5}$$
is equal to the one in the previous generation. Thus, genetic variation is preserved under this simple model of sexual reproduction, and the population is at equilibrium after one generation. In other words, Eqs. 1–3 describe the set of all populations at Hardy–Weinberg equilibrium. The parametric representation:
$$\begin{aligned} \{ (\not p\_{\rm AA}, \not p\_{\rm AA}, \not p\_{\rm aa}) \in \Delta\_2 \mid \not p\_{\rm AA} = p^2, \not p\_{\rm Aa} = 2p(1-p), \\ \not p\_{\rm aa} = \left(1-p\right)^2 \}, \end{aligned} \tag{6}$$
of this set of distributions is equivalent to the implicit representation as the intersection of the Hardy–Weinberg curve:
$$\mathbf{4} \ p\_{\rm AA} \ p\_{\rm za} - p\_{\rm Aa}^2 = 0 \tag{7}$$
with the probability simplex Δ<sup>2</sup> (Fig. 1).
The simplest discrete random variable is a binary (or Bernoulli) random variable X. The textbook example of a Bernoulli trial is the flipping of a coin. The state space of this random experiment is the set that contains all possible outcomes, namely, whether the coin lands on heads (<sup>X</sup> ¼ 0) or tails (<sup>X</sup> ¼ 1). We write X¼f0,1g to denote this state space. The parameter space is the set that contains all possible values of the model parameters. In the coin tossing example, the only parameter is the probability of observing tails, p, and this parameter can take any value between 0 and 1, so we write <sup>Θ</sup> ¼ {<sup>p</sup> j <sup>0</sup> <sup>p</sup> 1} for the parameter space. In general, the event <sup>X</sup> ¼ 1 is often called a "success," and <sup>p</sup> ¼ <sup>P</sup>(<sup>X</sup> ¼ 1) the probability of success.
Fig. 1 De Finetti diagram showing the Hardy–Weinberg curve <sup>4</sup> <sup>p</sup>AA <sup>p</sup>aa p<sup>2</sup> Aa <sup>¼</sup> 0 inside the probability simplex <sup>Δ</sup><sup>2</sup> <sup>¼</sup> {( <sup>p</sup>AA, <sup>p</sup>Aa, <sup>p</sup>aa)<sup>j</sup> <sup>p</sup>AA <sup>+</sup> <sup>p</sup>Aa <sup>+</sup> <sup>p</sup>aa <sup>¼</sup> 1}. Each point in this space represents a population as described by its genotype frequencies. Points on the curve correspond to populations in Hardy–Weinberg equilibrium
Example 2 (Binomial Distribution): Consider n independent Bernoulli trials, each with success probability p. Let X be the random variable counting the number of successes k among the <sup>n</sup> trials. Then, <sup>X</sup> has state space X¼f0, ... , <sup>n</sup>g and
$$P(X=k) = \binom{n}{k} p^k (1-p)^{n-k}.\tag{8}$$
This is the binomial distribution, denoted Binomðn, <sup>p</sup>Þ. Its parameter space is <sup>Θ</sup> ¼ ½0,1. Examples of binomially distributed random variables are the number of "heads" in n successive coin tosses or the number of mutated genes in a group of species. □
Important characteristics of a probability distribution are its expectation (or expected value, or mean) and its variance. They are defined, respectively, as:
$$\operatorname{E}(X) = \sum\_{\varkappa \in \mathcal{X}} \varkappa \, P(X = \varkappa), \tag{9}$$
$$\operatorname{Var}(X) = \sum\_{\boldsymbol{\varkappa} \in \mathcal{X}} [\boldsymbol{\varkappa} - \operatorname{E}(X)]^2 \ P(X = \boldsymbol{\varkappa}).\tag{10}$$
The standard deviation is ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi VarðXÞ <sup>p</sup> . For the binomial distribution, <sup>X</sup> Binomðn, <sup>p</sup>Þ, we find E(X) ¼ np and VarðXÞ ¼ npð<sup>1</sup> <sup>p</sup>Þ.
Example 3 (Poisson Distribution): The Poisson distribution PoisðλÞ with parameter <sup>λ</sup> 0 is defined as:
$$P(X=k) \quad = \quad \frac{\lambda^k}{k!} \quad , \quad k \in \mathbb{N}.\tag{11}$$
It describes the number X of independent events occurring in a fixed period of time (or space) at average rate λ and independently of the time since (or distance to) the last event. The Poisson distribution has equal expectation and variance, <sup>E</sup>ðXÞ ¼ VarðXÞ ¼ <sup>λ</sup>. □
The Poisson distribution is used frequently as a model for the number of DNA mutations in a gene after a certain time period, where λ is the mutation rate. Both the binomial and the Poisson distribution describe counts of random events. In the limit of large n and fixed product np, the two distributions coincide, Binomðn, <sup>p</sup>Þ ! PoisðnpÞ, for <sup>n</sup> !1.
Example 4 (Shotgun Sequencing): Let us consider a simplified model of the shotgun approach to DNA sequencing. Suppose that n reads of length L have been obtained from a genome of size G. We assume that all reads have the same probability of being sequenced. Then, the probability of hitting a specific base with one read is <sup>p</sup> ¼ <sup>L</sup>/G, and the average coverage of the sequencing run is <sup>c</sup> ¼ np. Under this model, the number of times <sup>X</sup> a single base is sequenced is distributed as Binomðn, <sup>p</sup>Þ: For large <sup>n</sup>, we have
Fig. 2 Coverage distribution of a shotgun sequencing experiment with <sup>n</sup> ¼ <sup>108</sup> reads of length <sup>L</sup> ¼ 100 of the human genome of length <sup>G</sup> ¼ <sup>3</sup> <sup>109</sup> . The average coverage is <sup>c</sup> ¼ np ¼ 3.4, where <sup>p</sup> ¼ <sup>L</sup>/G. Dots show the binomial coverage distribution Binomðn, <sup>p</sup>Þ and the solid line its approximation by the Poisson distribution PoisðnpÞ. Note that the Poisson distribution is also discrete and just shown as a line to distinguish it from the binomial distribution
$$P(X=k) = \binom{n}{k} p^k (1-p)^{n-k} \approx \frac{c^k}{k!} \, . \tag{12}$$
For example, using next-generation sequencing technology, one might obtain <sup>n</sup> <sup>¼</sup> <sup>108</sup> reads of length <sup>L</sup> ¼ 100 bases in a single run. For the human genome of length <sup>G</sup> ¼ <sup>3</sup> <sup>10</sup><sup>9</sup> , we obtain a coverage of <sup>c</sup> ¼ 3.4. The distribution of the number of reads per base pair is shown in Fig. 2. In particular, the fraction of unsequenced positions is <sup>P</sup>(<sup>X</sup> ¼ 0) ¼ <sup>e</sup> c ¼ 3.57%. □
A continuous random variable <sup>X</sup> takes values in X ¼ and is defined by a nonnegative function f(x) such that:
$$P(X \in B) = \int\_{B} f(\varkappa) d\varkappa,\quad \text{for all subsets } B \subseteq \mathbb{R}.\tag{13}$$
The function f is called the probability density function of X. For an interval:
$$P(X \in [a, b]) = P(a \le X \le b) = \int\_{a}^{b} f(x) dx. \tag{14}$$
The cumulative distribution function is
$$F(b) = P(X \le b) = \int\_{-\infty}^{b} f(\varkappa) d\varkappa, \quad b \in \mathbb{R}.\tag{15}$$
Thus, the density is the derivative of the cumulative distribution function, <sup>d</sup> dx <sup>F</sup>ðxÞ ¼ <sup>f</sup> <sup>ð</sup>xÞ.
In analogy to the discrete case, expectation and variance of a continuous random variable are defined, respectively, as:
$$\operatorname{E}(X) \, := \, \int\_{-\infty}^{\infty} \mathfrak{x} \, f(\mathfrak{x}) \, d\mathfrak{x}, \tag{16}$$
$$\text{Var}(X) \quad = \int\_{-\infty}^{\infty} \left[ \varkappa - \text{E}(X) \right]^2 f(\varkappa) \, d\varkappa. \tag{17}$$
Example 5 (Normal Distribution): The normal (or Gaussian) distribution has the density function:
$$f(\varkappa) = (2\pi\sigma^2)^{-1/2} \exp\left[-\frac{(\varkappa-\mu)^2}{2\sigma^2}\right].\tag{18}$$
The parameter space is <sup>Θ</sup> ¼ fðμ, <sup>σ</sup>2Þ j <sup>μ</sup><sup>∈</sup> , <sup>σ</sup>2<sup>∈</sup> þg. A normal random variable <sup>X</sup> Normðμ, <sup>σ</sup>2Þ has mean E(X) ¼ <sup>μ</sup> and variance VarðXÞ ¼ <sup>σ</sup>2. Normð0,1Þ is called the standard normal distribution. □
The normal distribution is frequently used as a model for measurement noise. For example, <sup>X</sup> Normðμ, <sup>σ</sup>2<sup>Þ</sup> might describe the hybridization intensity of a sample to a probe on a microarray. Then, μ is the level of expression of the corresponding gene and σ <sup>2</sup> summarizes the experimental noise associated with the microarray experiment. The parameters can be estimated from a finite sample {x(1), ..., x(N) }, i.e., from N replicate experiments, as the empirical mean and variance, respectively:
$$\overline{\boldsymbol{\pi}} = \frac{1}{N} \sum\_{i=1}^{N} \boldsymbol{\pi}^{(i)},\tag{19}$$
$$s^2 = \frac{1}{N-1} \sum\_{i=1}^{N} \left(\varkappa^{(i)} - \overline{\varkappa}\right)^2. \tag{20}$$
The normal distribution plays a special role in statistics due to the central limit theorem. It asserts that the average X- <sup>N</sup> ¼ ðXð1<sup>Þ</sup> þþ <sup>X</sup>ðN<sup>Þ</sup> Þ=<sup>N</sup> of <sup>N</sup> independent (see below) and identically distributed (i.i.d.) random variables X(i) with equal mean μ and variance σ <sup>2</sup> converges in distribution to the standard normal distribution:
$$\sqrt{N}\left(\frac{\bar{X}\_N - \mu}{\sigma}\right) \xrightarrow{d} \text{Norm}(0, 1),\tag{21}$$
irrespective of the shape of their distribution. As a consequence, many test statistics and estimators are asymptotically normally distributed. For example, the Poisson distribution PoisðλÞ is approximately normal Normðλ, <sup>λ</sup>Þ for large values of <sup>λ</sup>.
We often measure multiple quantities at the same time, for example the expression of several genes, and are interested in correlations among the variables. Let X and Y be two random variables with expected values μ<sup>X</sup> and μ<sup>Y</sup> and variances σ<sup>2</sup> <sup>X</sup> and σ<sup>2</sup> Y , respectively. The covariance between X and Y is
$$\text{Cov}(X, \mathcal{T}) = \text{E}[(X - \mu\_X)(\mathcal{T} - \mu\_{\mathcal{T}})] = \text{E}[X\mathcal{T}] - \text{E}[X]\text{E}[\mathcal{T}] \tag{22}$$
and the correlation between <sup>X</sup> and <sup>Y</sup> is <sup>ρ</sup>X,<sup>Y</sup> <sup>¼</sup> CovðX,<sup>Y</sup> <sup>Þ</sup>=ðσ<sup>X</sup> <sup>σ</sup><sup>Y</sup> <sup>Þ</sup>. For observations (x(1), <sup>y</sup> (1)), ..., (x(N) , y (N) ), the sample correlation coefficient is
$$r\_{\mathbf{x},\mathbf{y}} = \frac{\sum\_{i=1}^{N} (\mathbf{x}^{(i)} - \overline{\mathbf{x}})(\mathbf{y}^{(i)} - \overline{\mathbf{y}})}{(N-1)\mathfrak{s}\_{X}\mathfrak{s}\_{\Upsilon}},\tag{23}$$
where sX and sY are the sample standard deviations of X and Y , respectively, defined in Eq. 20.
So far, we have worked with univariate distributions and we now turn to multivariate distributions, i.e., we consider random vectors <sup>X</sup> <sup>¼</sup> (X1, ..., Xn) such that each Xi is a random variable. For the case of discrete random variables Xi, we first generalize the binomial distribution to random experiments with a finite number of outcomes.
Example 6 (Multinomial Distribution): Let K be the number of possible outcomes of a random experiment and θ<sup>k</sup> the probability of outcome <sup>k</sup>. We consider the random vector <sup>X</sup> <sup>¼</sup> (X1, ..., XK) with values inX ¼ <sup>K</sup> , where Xk counts the number of outcomes of type <sup>k</sup>. The multinomial distribution Multðn, <sup>θ</sup>1, ... , <sup>θ</sup><sup>K</sup> <sup>Þ</sup> is defined as:
$$P(X = \varkappa) = \frac{n!}{\varkappa\_1! \cdots \varkappa\_K!} \theta\_1^{\varkappa\_1} \cdots \theta\_K^{\varkappa\_K} \tag{24}$$
if P<sup>K</sup> <sup>k</sup>¼<sup>1</sup> xk <sup>¼</sup> <sup>n</sup>, and 0 otherwise. The parameter space of the model is <sup>Θ</sup> <sup>¼</sup> <sup>Δ</sup><sup>K</sup>-1. For <sup>K</sup> ¼ 2, we recover the binomial distribution (Eq. 8). Each component Xk of a multinomial vector has expected value E(Xk) <sup>¼</sup> <sup>n</sup>θ<sup>k</sup> and VarðXkÞ ¼ <sup>n</sup>θ<sup>k</sup>ð<sup>1</sup> <sup>θ</sup><sup>k</sup>Þ. The covariance of two components is CovðXk, XlÞ¼<sup>n</sup>θkθ<sup>l</sup> , for <sup>k</sup> 6¼ <sup>l</sup>. □
In general, the covariance matrix Σ of a random vector X is defined by:
$$\Sigma\_{ij} = \text{Cov}(X\_i, X\_j) = \text{E}[(X\_i - \mu\_i)(X\_j - \mu\_j)],\tag{25}$$
where μ<sup>i</sup> is the expected value of Xi. The matrix Σ is also called the variance–covariance matrix because the diagonal terms are the variances <sup>Σ</sup>ii <sup>¼</sup> CovðXi, XiÞ ¼ VarðXiÞ.
A continuous multivariate random variable X takes values in X ¼ n. It is defined by its cumulative distribution function:
$$F(\varkappa) = P(X \le \varkappa), \quad \varkappa \in \mathbb{R}^n \tag{26}$$
or, equivalently, by the probability density function:
$$f(\boldsymbol{\omega}) = \frac{\mathfrak{d}^n}{\mathfrak{d}\boldsymbol{\omega}\_1 \cdots \mathfrak{d}\boldsymbol{\omega}\_n} F(\boldsymbol{\omega}\_1, \dots, \boldsymbol{\omega}\_n), \quad \boldsymbol{\omega} \in \mathbb{R}^n. \tag{27}$$
Example 7 (Multivariate Normal Distribution): For <sup>n</sup> 1 and x ∈n, the multivariate normal (or Gaussian) distribution has density:
$$f(\varkappa) = (2\pi)^{-n/2} \det(\Sigma)^{-1/2} \exp\left[ -\frac{1}{2} (\varkappa - \mu)^f \Sigma^{-1} (\varkappa - \mu) \right],\tag{28}$$
$$\zeta\_{\varkappa} = \frac{1}{\varkappa - \varkappa} \int\_{\varkappa}^{\varkappa} \frac{1}{\varkappa - \varkappa} d\varkappa$$
with parameter space <sup>Θ</sup> ¼ fðμ, <sup>Σ</sup>Þ j <sup>μ</sup> ¼ ðμ1, ... , <sup>μ</sup><sup>n</sup>Þ<sup>∈</sup> <sup>n</sup> and <sup>Σ</sup> ¼ <sup>σ</sup><sup>2</sup> i j ∈nng, where <sup>Σ</sup> is the symmetric, positive-definite covariance matrix and μ the expectation. We write <sup>X</sup> ¼ ðX1, ... , XnÞ Normðμ, <sup>Σ</sup>Þ for a random vector with such a distribution. □
We say that two random variables X and Y are independent if P (X, <sup>Y</sup> ) ¼ <sup>P</sup>(X)P(<sup>Y</sup> ) or, equivalently, if the conditional probability <sup>P</sup>(<sup>X</sup> j <sup>Y</sup> ) ¼ <sup>P</sup>(X, <sup>Y</sup> )/P(<sup>Y</sup> ) is equal to the unconditional probability P(X). If X and Y are independent, denoted X ⊥ Y , then E[XY ] ¼ E[X]E[<sup>Y</sup> ] and Varð<sup>X</sup> þ <sup>Y</sup> Þ ¼ VarðXÞ þ Varð<sup>Y</sup> Þ. It follows that independent random variables have covariance zero. However, the converse is only true in specific situations, for example if (X, Y ) is multivariate normal, but not in general because correlation captures only linear dependencies.
This limitation can be addressed by using statistical models which allow for a richer dependency structure. Subheading 7 is devoted to Bayesian networks, a family of probabilistic graphical models based on conditional independences. Let X, Y , and Z be three random vectors. Generalizing the notion of statistical independence, we say that X is conditionally independent of Y given <sup>Z</sup> and write <sup>X</sup> <sup>⊥</sup> <sup>Y</sup> j <sup>Z</sup> if <sup>P</sup>(X, <sup>Y</sup> j <sup>Z</sup>) ¼ <sup>P</sup>(<sup>X</sup> j <sup>Z</sup>)P(<sup>Y</sup> j <sup>Z</sup>). Bayes' theorem states that
$$P(\mathcal{T} \mid X) = \frac{P(X \mid \mathcal{T})P(\mathcal{T})}{P(X)},\tag{29}$$
where <sup>P</sup>(<sup>Y</sup> ) is called the prior probability and <sup>P</sup>(<sup>Y</sup> j <sup>X</sup>) the posterior probability. Intuitively, the prior P(Y ) encodes our a priori knowledge about <sup>Y</sup> (i.e., before observing <sup>X</sup>), and <sup>P</sup>(<sup>Y</sup> j <sup>X</sup>) is our updated knowledge about Y a posteriori (i.e., after observing X).
We have <sup>P</sup>(X) <sup>¼</sup>∑YP(X, <sup>Y</sup> ) if <sup>Y</sup> is discrete, and similarly <sup>P</sup>(X) ¼ R YP(X, Y )dY if Y is continuous. Here, P(X) is called the marginal and P(X, Y ) the joint probability. This summation or integration is known as marginalization (Fig. 3).
Since <sup>P</sup>(X) <sup>¼</sup>∑YP(X, <sup>Y</sup> ) ¼∑YP(<sup>X</sup> j <sup>Y</sup> )P(<sup>Y</sup> ), Bayes' theorem can also be rewritten as:
$$P(\mathcal{T} \mid X) = \frac{P(X \mid \mathcal{T})P(\mathcal{T})}{\sum\_{\mathcal{Y}' \in \mathcal{Y}} P(X \mid \mathcal{Y})P(\mathcal{Y})},\tag{30}$$
where P(y<sup>0</sup> ) ¼ <sup>P</sup>(<sup>Y</sup> ¼ <sup>y</sup><sup>0</sup> ) and Y is the state space of <sup>Y</sup> .
Fig. 3 Marginalization. Left: two-dimensional histogram of a discrete bivariate distribution with the two marginal histograms. Right: contour plot of a two-dimensional Gaussian density with the marginal distributions of each component
Example 8 (Diagnostic Test): We want to evaluate a diagnostic test for a rare genetic disease. The binary random variables D and <sup>T</sup> indicate disease status (<sup>D</sup> ¼ 1, diseased) and test result (<sup>T</sup> ¼ 1, positive), respectively. Let us assume that the prevalence of the disease is 0.5%, i.e., 0.5% of all people in the population are known to be affected. The test has a false positive rate (probability that somebody is tested positive who does not have the disease) of <sup>P</sup>(<sup>T</sup> ¼ <sup>1</sup> j <sup>D</sup> ¼ 0) ¼ 5% and a true positive rate (probability that somebody is tested positive who has the disease) of <sup>P</sup>(<sup>T</sup> ¼ <sup>1</sup> j <sup>D</sup> ¼ 1) ¼ 90%. Then, the posterior probability of a person having the disease given that he or she tested positive is
$$\begin{aligned} P(D=1 \mid T=1) &= \\ \hline P(T=1 \mid D=1)P(D=1) &= \\ \hline P(T=1 \mid D=0)P(D=0) + P(T=1 \mid D=1)P(D=1) &= 0.083, \\ &\quad (31) \end{aligned} \tag{31}$$
that is, only 8.3% of the positively tested individuals actually have the disease. Thus, our prior belief of the disease status, P(D), has been modified in light of the test result by multiplication with <sup>P</sup>(<sup>T</sup> j <sup>D</sup>) to obtain the updated belief <sup>P</sup>(<sup>D</sup> j <sup>T</sup>). □
Exercise 9 (Conditional Independence): Let X, Y , and Z be random variables. Using the laws of probability, show that X and Y are conditionally independent given <sup>Z</sup> (i.e., <sup>X</sup> <sup>⊥</sup> <sup>Y</sup> j <sup>Z</sup>) if and only if <sup>P</sup>(<sup>X</sup> j <sup>Y</sup>, <sup>Z</sup>) ¼ <sup>P</sup>(<sup>X</sup> j <sup>Z</sup>).
#### 2 Statistical Inference
Statistical models have parameters and a common task is to estimate the model parameters from observed data. The goal is to find the set of parameters with the best model fit. There are two major approaches to parameter estimation: maximum likelihood (ML) and Bayes.
The maximum likelihood approach is based on the likelihood function. Let us consider a fixed statistical model M with parameter space Θ and assume that we have observed realizations <sup>D</sup> ¼ fxð1<sup>Þ</sup> , ... , xð<sup>N</sup> <sup>Þ</sup> g of the discrete random variable <sup>X</sup> <sup>M</sup>ðθ<sup>0</sup>Þ for some unknown parameter θ<sup>0</sup> ∈ Θ. For the fixed data set D, the likelihood function of the model is
$$L(\theta) = P(\mathcal{O} \mid \theta),\tag{32}$$
where we write <sup>P</sup>ð<sup>D</sup> j <sup>θ</sup>Þ to emphasize that, here, the probability of the data depends on the model parameter θ. For continuous random variables, the likelihood function is defined similarly in terms of the density function, <sup>L</sup>ðθÞ ¼ <sup>f</sup> ð<sup>D</sup> j <sup>θ</sup>Þ. Maximum likelihood estimation seeks the parameter θ ∈ Θ for which L(θ) is maximal. Rather than L(θ), it is often more convenient to maximize <sup>ℓ</sup>ðθÞ ¼ logLðθÞ, the log-likelihood function. If the data are i.i.d., then:
$$\ell(\theta) = \sum\_{i=1}^{N} \log P(X = x^{(i)} \mid \theta). \tag{33}$$
Example 10 (Likelihood Function of the Binomial Model): Suppose we have observed <sup>k</sup> ¼ 7 successes in a total of <sup>N</sup> ¼ 10 Bernoulli trials. The likelihood function of the binomial model (Eq. 8) is
$$L(\mathfrak{p}) = \mathfrak{p}^k \left(1 - \mathfrak{p}\right)^{N-k},\tag{34}$$
where p is the success probability (Fig. 4). To maximize L, we consider the log-likelihood function:
$$\ell(\not p) = \log L(\not p) = k \log(\not p) + (N - k) \log(1 - \not p) \tag{35}$$
and the likelihood equation <sup>d</sup>ℓ/dp ¼ 0. The ML estimate (MLE) is the solution p ^ML <sup>¼</sup> <sup>k</sup>=<sup>N</sup> <sup>¼</sup> <sup>7</sup>=10. Thus, the MLE of the success probability is just the relative frequency of successes. □
Example 11 (Likelihood Function of the Hardy–Weinberg Model): If we genotype a finite random sample of a population of diploid individuals at a single locus, then the resulting data consists of the numbers of individuals nAA, nAa, and naa with the respective genotypes. Assuming Hardy–Weinberg equilibrium (Eqs. 1–3), we want to estimate the allele frequencies <sup>p</sup> and <sup>q</sup> ¼ <sup>1</sup> p of the
Fig. 4 Likelihood function of the binomial model. The underlying data set consists of <sup>k</sup> ¼ 7 successes out of <sup>N</sup> ¼ 10 Bernoulli trials. The likelihood <sup>L</sup> ( <sup>p</sup>) <sup>¼</sup> pk (1 p) N<sup>k</sup> is plotted as a function of the model parameter p, the probability of success (solid line). The MLE is the maximum of this function, p ^ML <sup>¼</sup> <sup>k</sup>=<sup>N</sup> <sup>¼</sup> <sup>7</sup>=10 (dashed line)
population. The likelihood function of the Hardy–Weinberg model is <sup>L</sup>ðpÞ ¼ <sup>P</sup>ðAAÞ <sup>n</sup>AA <sup>P</sup>ðAaÞ <sup>n</sup>AaPðaaÞ <sup>n</sup>aa and the log-likelihood is
$$\begin{split} \ell(p) &= n\_{\rm AA} \log p^2 + n\_{\rm AA} \log 2 \rho (1 - p) + n\_{\rm a} \log (1 - p)^2 \\ &\propto (2n\_{\rm AA} + n\_{\rm A}) \log p + (n\_{\rm Aa} + 2n\_{\rm a}) \log (1 - p), \end{split} \tag{36}$$
where we have dropped the constant nAalog2. The MLE of p ∈ [0, 1] can be found by maximizing ℓ. Solving the likelihood equation:
$$\frac{\partial \ell}{\partial p} = \frac{2n\_{\text{AA}} + n\_{\text{Aa}}}{p} - \frac{n\_{\text{Aa}} + 2n\_{\text{Aa}}}{1 - p} = 0 \tag{37}$$
yields the MLE p ^ML ¼ ð2nAA <sup>þ</sup> <sup>n</sup>AaÞ=ð2<sup>N</sup> <sup>Þ</sup>, where <sup>N</sup> <sup>¼</sup> <sup>n</sup>AA <sup>+</sup> nAa + naa is the total sample size. For example, if we sample <sup>N</sup> <sup>¼</sup> 100 genotypes with <sup>n</sup>AA <sup>¼</sup> 81, <sup>n</sup>Aa <sup>¼</sup> 18, and <sup>n</sup>aa <sup>¼</sup> 1, then we find p ^ML ¼ ð<sup>2</sup> ð<sup>81</sup> <sup>þ</sup> <sup>18</sup>ÞÞ=ð<sup>2</sup> <sup>100</sup>Þ ¼ <sup>0</sup>:9 for the frequency of the major allele. □
MLEs have many desirable properties. Asymptotically, as the sample size <sup>N</sup> !1, they are normally distributed, unbiased, and have minimal variance. The uncertainty in parameter estimation associated with the sampling variance of the finite data set can be quantified in confidence intervals. There are several ways to construct confidence intervals and statistical tests for MLEs based on the asymptotic behavior of the log-likelihood function <sup>ℓ</sup>ðθÞ ¼ logLðθÞ and its derivatives. For example, the asymptotic normal distribution of the MLE is
$$
\hat{\theta}\_{\text{ML}} \stackrel{\mathfrak{a}}{\sim} \text{Norm}\left(\theta, J(\theta)^{-1}\right), \tag{38}
$$
where <sup>I</sup>(θ) ¼ -∂2 <sup>ℓ</sup>/∂θ<sup>2</sup> is the Fisher information and <sup>J</sup>(θ) ¼ E[<sup>I</sup> (θ)] the expected Fisher information. This result gives rise to the Wald confidence intervals:
$$\left[\hat{\theta}\_{\text{ML}} \pm \,\, z\_{1-a/2} \,\, J(\theta)^{-1}\right],\tag{39}$$
where <sup>z</sup>1<sup>α</sup>=<sup>2</sup> <sup>¼</sup> inf <sup>f</sup><sup>x</sup> <sup>∈</sup> <sup>j</sup> <sup>1</sup> <sup>α</sup>=<sup>2</sup> <sup>F</sup>ðxÞg is the (1 α/2) quantile and F the cumulative distribution function of the standard normal distribution. Equation 38 still holds after replacing J(θ) with the standard error seð<sup>θ</sup> ^MLÞ¼½<sup>I</sup> ð<sup>θ</sup> ^MLÞ-1 <sup>2</sup> or ½Jð<sup>θ</sup> ^MLÞ-1 2, and it also generalizes to higher dimensions. Other common constructions of confidence intervals include those based on the asymptotic distribution of the score function <sup>S</sup>(θ) ¼ <sup>∂</sup>ℓ/∂<sup>θ</sup> and the log-likelihood ratio logðLð<sup>θ</sup> ^MLÞ=LðθÞÞ [8].
We now discuss another more generic approach to quantify parameter uncertainty, not restricted to ML estimation, which is applied frequently in practice due to its simple implementation. Bootstrapping [9] is a resampling method in which independent observations are resampled from the data with replacement. The resulting new data set consists of (some of) the original observations, and under i.i.d. assumptions, the bootstrap replicates have asymptotically the same distribution as the data. Intuitively, by sampling with replacement, one is pretending that the collection of replicates thus obtained is a good proxy for the distribution of data sets that one would have obtained, had we been able to actually replicate the experiment. In this way, the variability of an estimator (or more generally the distribution of any test statistic) can be approximated by evaluating the estimator (or the statistic) on a collection of bootstrap replicates. For example, the distribution of the ML estimator of a model parameter θ can be obtained from the bootstrap samples.
Example 12 (Bootstrap Confidence Interval for the ML Allele Frequency): We use bootstrapping to estimate the distribution of the ML estimator p ^ML of the Hardy–Weinberg model for the data set (nAA, <sup>n</sup>Aa, <sup>n</sup>aa) ¼ (81, 18, 1) of Example 11. For each bootstrap sample, we draw <sup>N</sup> ¼ 100 genotypes with replacement from the original data to obtain random integer vectors of length three summing to 100. The ML estimate is computed for each of a total of B bootstrap samples. The resulting distributions of p ^ML are shown in Fig. 5, for <sup>B</sup> ¼ 100, 1000, and 10,000. The means of these empirical distributions are 0.899, 0.9004, and 0.9001,
Fig. 5 Bootstrap analysis of the ML allele frequency. The bootstrap distribution of the maximum likelihood estimator p ^ML ¼ ð2nAA <sup>þ</sup> <sup>n</sup>AaÞ=ð2N<sup>Þ</sup> of the major allele frequency in the Hardy–Weinberg model is plotted for <sup>B</sup> <sup>¼</sup> 100 (left), <sup>B</sup> <sup>¼</sup> 1000 (center), and <sup>B</sup> <sup>¼</sup> 10, 000 (right) bootstrap samples, for the data set (nAA, <sup>n</sup>Aa, <sup>n</sup>aa) ¼ (81, 18, 1)
respectively, and 95% bootstrap confidence intervals can be derived from the 2.5 and 97.5% quantiles of the distributions. For <sup>B</sup> ¼ 100, 1000, and 10,000, we obtain, respectively, [0.8598, 0.9350], [0.860, 0.940], and [0.855, 0.940]. The basic bootstrap confidence intervals have several limitations, including bias of the bootstrap estimator and skewness of the bootstrap distribution. Other methods exist for constructing confidence intervals from the bootstrap distribution to address some of them [9]. □
The Bayesian approach takes a different point of view and regards the model parameters as random variables [10]. Inference is then concerned with estimating the joint distribution of the parameters θ given the observed data D. By Bayes' theorem (Eq. 30), we have
$$P(\theta \mid \mathcal{O}) = \frac{P(\mathcal{O} \mid \theta)P(\theta)}{P(\mathcal{O})} = \frac{P(\mathcal{O} \mid \theta)P(\theta)}{\int\_{\theta \in \Theta} P(\mathcal{O} \mid \theta)P(\theta) \, d\theta},\tag{40}$$
that is, the posterior probability of the parameters is proportional to the likelihood of the data times the prior probability of the parameters. It follows that, for a uniform prior, the mode of the posterior is equal to the MLE.
From the posterior, credible intervals of parameter estimates can be derived such that the parameter lies in the interval with a certain probability, say 95%. This is in contrast to a 95% confidence interval in the frequentist approach because, there, the parameter is fixed and the interval boundaries are random variables. The meaning of a confidence interval is that 95% of similar intervals would contain the true parameter, if intervals were constructed independently from additional identically distributed data.
The prior P(θ) encodes our a priori belief in θ before observing the data. It can be used to incorporate domain-specific knowledge into the model, but it may also be uninformative or objective, in which case all observations are equally likely, or nearly so, a priori. However, it can sometimes be difficult to find noninformative priors. In practice, conjugate priors are most often used. A conjugate prior is one that is invariant with respect to the distribution family under multiplication with the likelihood, i.e., the posterior belongs to the same family as the prior. Conjugate priors are mathematically convenient and computationally efficient because the posterior can be calculated analytically for a wide range of statistical models.
Example 13 (Dirichlet Prior): Let <sup>T</sup> <sup>¼</sup> (T1, ..., TK) be a continuous random variable with state space <sup>Δ</sup>K-1. The Dirichlet distribution DirðαÞ with parameters <sup>α</sup>∈<sup>K</sup> þ has probability density function:
$$f(\theta\_1, \ldots, \theta\_K) \quad = \frac{\Gamma\left(\sum\_{i=1}^K a\_i\right)}{\prod\_{\substack{i=1 \\ i \neq 1}}^K \prod\_{i=1}^K \theta\_i^{a\_i - 1}, \tag{41}$$
where Γ is the gamma function. The Dirichlet prior is conjugate to the multinomial likelihood: If <sup>T</sup> DirðαÞ and ð<sup>X</sup> j <sup>T</sup> ¼ <sup>θ</sup>Þ Multðn, <sup>θ</sup>1, ... , <sup>θ</sup><sup>K</sup> <sup>Þ</sup>, then <sup>ð</sup><sup>θ</sup> <sup>j</sup> <sup>X</sup> <sup>¼</sup> <sup>x</sup>Þ Dirð<sup>α</sup> <sup>þ</sup> <sup>x</sup>Þ. For <sup>K</sup> <sup>¼</sup> 2, this distribution is called the beta distribution. Hence, the beta distribution is the conjugate prior to the binomial likelihood. □
Example 14 (Posterior Probability of Genotype Frequencies): Let us consider the simple genetic system with two loci and two alleles each of Example 1, but without assuming the Hardy–Weinberg model. We regard the observed genotype frequencies (nAA, nAa, <sup>n</sup>aa) ¼ (81, 18, 1) as the result of a draw from a multinomial distribution Multðn, <sup>θ</sup>AA, <sup>θ</sup>Aa, <sup>θ</sup>aaÞ. Assuming a Dirichlet prior DirðαAA, <sup>α</sup>Aa, <sup>α</sup>aa), the posterior genotype probabilities follow the Dirichlet distribution DirðαAA <sup>þ</sup> <sup>n</sup>AA, <sup>α</sup>Aa <sup>+</sup> <sup>n</sup>Aa, <sup>α</sup>aa <sup>+</sup> <sup>n</sup>aa). In Fig. 6, the prior Dirð10, 10, 10Þ is shown on the left, the multinomial likelihood <sup>P</sup>((nAA, <sup>n</sup>Aa, <sup>n</sup>aa) ¼ (81, 18, 1) j <sup>θ</sup>AA, <sup>θ</sup>Aa, <sup>θ</sup>aa) in the center, and the resulting posterior Dirð<sup>10</sup> þ 81, 10 þ 18, 10 þ <sup>1</sup>Þ on the right. Note that the MLE is different from the mode of the posterior. As compared to the likelihood, the nonuniform prior has shifted the maximum of the posterior toward the center of the probability simplex. □
We often have two or more competing models and would like to assess which one describes best the given data. For example, we may have observed genotypes from the set {AA, Aa, aa} and want to test whether the Hardy–Weinberg model (Example 1) is a more appropriate description of the genotype data than the multinomial model of the previous Example 14. Intuitively, we might want to select the model that fits the data best, for example, by comparing
Fig. 6 Dirichlet prior for multinomial likelihood. The Dirichlet prior is conjugate to the multinomial likelihood. Shown are contour lines of the prior Dirð10, 10, 10<sup>Þ</sup> on the left, the multinomial likelihood <sup>P</sup>((nAA, <sup>n</sup>Aa, <sup>n</sup>aa) ¼ (81, 18, 1)jθAA, <sup>θ</sup>Aa, <sup>θ</sup>aa) in the center, and the resulting posterior Dirð91, 28, 11Þ on the right. The posterior is the product of prior and likelihood
their likelihoods. However, the Hardy–Weinberg model has only one parameter, namely the allele frequency p, whereas the multinomial model has three parameters subject to the constraint θAA + <sup>θ</sup>Aa <sup>+</sup> <sup>θ</sup>aa <sup>¼</sup> 1. Hence, the number of free parameters is one and two, respectively, for the two models. This difference in the complexity of the models makes a comparison based only on the goodness of fit invalid, because models with more parameters, i.e., higher complexity, can generally provide a better fit. Estimating model complexity and scoring models based on both model complexity and goodness of fit is therefore essential for model comparison and model selection.
The goal of model selection is to find the model that best generalizes to unseen data, rather than just fits the observed data, because we seek the model capable of the most accurate predictions. A model that fits well but generalizes poorly is said to overfit the data. Models that are too complex tend to overfit the data. Model selection can be regarded as finding the right level model complexity for the given data, such that the predictive performance is optimized. This involves defining a criterion of optimality and a procedure for finding the optimal model.
A common frequentist approach to model selection are likelihood ratios. For a data set D, we compare a null model, M0, to an alternative model, M1, at given point estimates using the ratio of their likelihoods:
$$\Lambda(\mathcal{O}) = \frac{L(\hat{\theta}\_0)}{L(\hat{\theta}\_1)}\tag{42}$$
If <sup>Λ</sup>ðDÞ <sup>&</sup>lt; <sup>c</sup>, for a defined threshold <sup>c</sup>, we reject the null model and favor the alternative model. The choice of c should be informed by the distribution of Λ under the null. If the two models are nested, i.e., if M0 can be obtained from M1 by specifying a subset of the parameters, then -2 logΛ is approximately χ 2 -distributed with degrees of freedom equal to the difference in the number of free parameters between M1 and M0.
In the Bayesian framework, it is natural to compare the posterior probabilities of the two models. By Bayes theorem, we have, for <sup>i</sup> ¼ 0, 1:
$$P(\mathbf{M}\_i \mid \mathcal{O}) = \frac{P(\mathcal{O} \mid \mathbf{M}\_i)P(\mathbf{M}\_i)}{P(\mathcal{O})} \tag{43}$$
where:
$$P(\mathcal{O} \mid \mathbf{M}\_i) = \int P(\mathcal{O} \mid \theta\_i, \mathbf{M}\_i) P(\theta\_i \mid \mathbf{M}\_i) \, d\theta\_i \tag{44}$$
is the marginal likelihood. The marginal likelihood accounts for model complexity and for uncertainty in parameter estimates, but is usually analytically intractable and costly to compute. Various approximations of the marginal likelihood exist that give rise to model selection scores, such as the Bayesian information criterion (BIC; see Subheading 7) and the Akaike information criterion (AIC) [11].
For Bayesian model comparison, we consider the posterior odds:
$$\frac{P(\mathbf{M\_0} \mid \mathcal{O})}{P(\mathbf{M\_1} \mid \mathcal{O})} = \frac{P(\mathcal{O} \mid \mathbf{M\_0})}{P(\mathcal{O} \mid \mathbf{M\_1})} \,\,\frac{P(\mathbf{M\_0})}{P(\mathbf{M\_1})}\tag{45}$$
The ratio of the marginal likelihoods, i.e., the first factor on the right-hand side of Eq. 45, is called the Bayes factor. With equal priors, a Bayes factor larger than 20 is often considered strong support for M0 over M1 [12].
Exercise 15 (Poisson Distribution): We wish to model the number of bacterial colonies in a Petri dish and assume that the count data of this experiment follows a Poisson distribution PoisðλÞ (Example 3). Derive the log-likelihood function of this model and calculate the MLE of the model parameter λ. Suppose now that the number of bacterial colonies on a Petri dish follows the Poisson distribution with mean <sup>λ</sup> ¼ 5. What is the probability of finding exactly three colonies?
#### 3 Hidden Data and the EM Algorithm
We often cannot observe all relevant random variables due to, for example, experimental limitations or study designs. In this case, a statistical model <sup>P</sup>(X, <sup>Z</sup> j <sup>θ</sup> <sup>∈</sup> <sup>Θ</sup>) consists of the observed random variable X and the hidden (or latent) random variable Z, both of which can be multivariate. In this section, we write <sup>X</sup> <sup>¼</sup> (X(1), ..., X(N) ) for the random variables describing the N observations and refer to X also as the observed data. The hidden data for this model is <sup>Z</sup> <sup>¼</sup> (Z(1), ..., <sup>Z</sup>(N) ) and the complete data is (X, Z). For convenience, we assume the parameter space Θ to be continuous and the state spaces X of <sup>X</sup> and <sup>Z</sup> of <sup>Z</sup> to be discrete.
In the Bayesian framework, one does not distinguish between unknown parameters and hidden data, and it is natural to assess the joint posterior <sup>P</sup>(θ, <sup>Z</sup> j <sup>X</sup>) / <sup>P</sup>(<sup>X</sup> j <sup>θ</sup>, <sup>Z</sup>)P(θ, <sup>Z</sup>), which is <sup>P</sup>(X, <sup>Z</sup> j <sup>θ</sup>)P(θ) if priors are independent, i.e., if <sup>P</sup>(θ, <sup>Z</sup>) ¼ <sup>P</sup>(θ)P(Z). Alternatively, if the distribution of the hidden data Z is not of interest, it can be marginalized out. Then, the posterior (Eq. 40) becomes
$$P(\theta \mid X) = \frac{\sum\_{Z} P(X, Z \mid \theta) P(\theta)}{\int\_{\theta \in \Theta} \sum\_{Z} P(X, Z \mid \theta) P(\theta) \, d\theta}. \tag{46}$$
In the likelihood framework, it can be more efficient to estimate the hidden data, rather than marginalizing over it. The hidden (or complete-data) log-likelihood is
$$\ell\_{\text{hid}}(\theta) = \log P(X, Z \mid \theta) = \sum\_{i=1}^{N} \log P(X^{(i)}, Z^{(i)} \mid \theta). \tag{47}$$
For ML parameter estimation, we need to consider the observed log-likelihood:
$$\begin{split} \ell\_{\text{obs}}(\theta) &= \log P(X \mid \theta) = \log \sum\_{Z} P(X, Z \mid \theta) \\ &= \log \sum\_{Z^{(l)} \in \mathcal{Z}} \dots \sum\_{Z^{(N)} \in \mathcal{Z}} \prod\_{i=1}^{N} P(X^{(i)}, Z^{(i)} \mid \theta) . \end{split} \tag{48}$$
This likelihood function is usually very difficult to maximize and one has to resort to numerical optimization techniques. Generic local methods, such as gradient descent or Newton's method, can be used, but there is also a more specific local optimization procedure, which avoids computing any derivatives of the likelihood function, called the expectation maximization (EM) algorithm [13].
In order to maximize the likelihood function (Eq. 48), we consider any distribution q(Z) of the hidden data Z and write
$$\ell\_{\rm obs}(\theta) = \log \sum\_{Z} q(Z) \frac{P(X, Z \mid \theta)}{q(Z)} = \log \mathrm{E}[P(X, Z \mid \theta) / q(Z)],\tag{49}$$
where the expected value is with respect to q(Z). Jensen's inequality applied to the concave log function asserts that log E½<sup>Y</sup> <sup>E</sup>½log<sup>Y</sup> . Hence, the observed log-likelihood is bounded from below by E log ½ ðPðX, <sup>Z</sup> j <sup>θ</sup>Þ=qðZÞÞ , or
$$\ell\_{\rm obs}(\theta) \ge \mathcal{E}[\ell\_{\rmhid}(\theta)] + H(q), \tag{50}$$
where <sup>H</sup>ðqÞ¼-E log ½ <sup>q</sup>ðZÞ is the entropy. The idea of the EM algorithm is to maximize this lower bound instead of ℓobs(θ) itself. Intuitively, this task is easier because the big sum over the hidden data in Eq. 48 disappears on the right-hand side of Eq. 50 upon taking expectations.
The EM algorithm is an iterative procedure alternating between an E step and an M step. In the E step, the lower bound (Eq. 50) is maximized with respect to the distribution q by setting <sup>q</sup>(Z) ¼ <sup>P</sup>(<sup>Z</sup> j <sup>X</sup>, <sup>θ</sup>(t) ), where θ(t) is the current estimate of θ, and computing the expected value of the hidden log-likelihood:
$$\mathcal{Q}(\theta \mid \theta^{(t)}) = \mathcal{E}\_{Z|X, \theta^{(t)}}[\ell\_{\text{hid}}(\theta)]. \tag{51}$$
In the M step, Q is maximized with respect to θ to obtain an improved estimate:
$$\theta^{(\iota+1)} = \arg\max\_{\theta} \underline{Q}(\theta \mid \theta^{(\iota)}).\tag{52}$$
The sequence θ(1), θ(2), θ(3), ... converges to a local maximum of the likelihood surface (Eq. 48). The global maximum and, hence, the MLE is generally not guaranteed to be found with this local optimization method. In practice, the EM algorithm is often run repeatedly with many different starting solutions θ(1), or with few very reasonable starting solutions obtained from other heuristics or educated guesses.
Example 16 (Naive Bayes): Let us assume that we observe realizations of a discrete random variable (X1, ..., XL) and we want to cluster observations into K distinct groups. For this purpose, we introduce a hidden random variable Z with state space <sup>Z</sup> ¼ ½K¼f1, ... , <sup>K</sup>g indicating class membership. The joint probability of (X1, ..., XL) and Z is
$$\begin{split}P(X\_1, \ldots, X\_L, Z) &= P(Z)P(X\_1, \ldots, X\_L \mid Z) \\ &= P(Z) \prod\_{n=1}^L P(X\_n \mid Z). \end{split} \tag{53}$$
The marginalization of this model with respect to the hidden data Z is the unsupervised naive Bayes model. The observed variables Xn are often called features and Z the latent class variable (Fig. 7).
The model parameters are the class prior P(Z), which we assume to be constant and will ignore, and the conditional probabilities <sup>θ</sup>n,kx <sup>¼</sup> <sup>P</sup>(Xn <sup>¼</sup> <sup>x</sup> <sup>j</sup> <sup>Z</sup> <sup>¼</sup> <sup>k</sup>). The complete-data likelihood of observed data <sup>X</sup> <sup>¼</sup> (X(1), ..., <sup>X</sup>(N) ) and hidden data <sup>Z</sup> ¼ (Z(1), ..., Z(N) ) is
$$P(X, Z \mid \theta) = \prod\_{i=1}^{N} P(X^{(i)}, Z^{(i)} \mid \theta) = \prod\_{i=1}^{N} P(Z^{(i)}) \prod\_{n=1}^{L} P(X\_n^{(i)} \mid Z^{(i)}) \tag{54}$$
Fig. 7 Graphical representation of the naive Bayes model. Observed features Xn are conditionally independent given the latent class variable Z
$$\propto \quad \prod\_{i=1}^{N} \prod\_{n=1}^{L} \theta\_{n, Z^{(i)} X\_n^{(i)}} = \prod\_{i=1}^{N} \prod\_{n=1}^{L} \prod\_{k \in [K]} \prod\_{x \in \mathcal{X}} \theta\_{n, kx}^{I\_{n, kx}(Z^{(i)})},\tag{55}$$
where In,kx(Z(i) ) is equal to one if and only if <sup>Z</sup>(i) ¼ <sup>k</sup> and <sup>X</sup>ði<sup>Þ</sup> <sup>n</sup> <sup>¼</sup> <sup>x</sup>, and zero otherwise.
To apply the EM algorithm for estimating θ without observing Z, we consider the hidden log-likelihood:
$$\ell\_{\text{hid}}(\theta) = \log P(X, Z \mid \theta) = \sum\_{i=1}^{N} \sum\_{n=1}^{L} \sum\_{k \in [K]} \sum\_{x \in X} I\_{n,k\varepsilon}(Z^{(i)}) \log \theta\_{n,k\varepsilon} \tag{56}$$
In the E step, we compute the expected values of Z(i) :
$$\begin{split} \mathcal{I}\_{n,\boldsymbol{k}\boldsymbol{x}}^{(i)} = \operatorname{E}\_{\boldsymbol{Z}\mid \boldsymbol{X}\_{\boldsymbol{n}}=\boldsymbol{\kappa},\boldsymbol{\theta}}\left[\boldsymbol{Z}^{(i)}\right] &= \frac{P(\boldsymbol{X}\_{\boldsymbol{n}}^{(i)}=\boldsymbol{\kappa}\mid \boldsymbol{Z}^{(i)}=\boldsymbol{k})}{\sum\_{\boldsymbol{k}'\in\boldsymbol{K}}P(\boldsymbol{X}\_{\boldsymbol{n}}^{(i)}=\boldsymbol{\kappa}\mid \boldsymbol{Z}^{(i)}=\boldsymbol{k}')} \\ &= \frac{\theta\_{n,\boldsymbol{k}\boldsymbol{x}}^{\prime}}{\sum\_{\boldsymbol{k}'\in\boldsymbol{K}}\theta\_{n,\boldsymbol{k}\boldsymbol{\ell}}^{\prime}}, \end{split} \tag{57}$$
where θ<sup>0</sup> is the current estimate of θ. The expected value γ ðiÞ <sup>n</sup>, kx is sometimes referred to as the responsibility of class k for observation XðiÞ <sup>n</sup> <sup>¼</sup> <sup>x</sup>. The expected hidden log-likelihood can be written in terms of the expected counts <sup>N</sup> <sup>n</sup>, kx <sup>¼</sup> <sup>P</sup><sup>N</sup> <sup>i</sup>¼<sup>1</sup> <sup>γ</sup> ðiÞ <sup>n</sup>, kx as:
$$\operatorname{E}\_{Z \mid X, \theta^\*} [\ell\_{\text{hid}}(\theta)] = \sum\_{n=1}^{L} \sum\_{k \in [K]} \sum\_{\mathbf{x} \in \mathcal{X}} N\_{n,k\mathbf{x}} \log \theta\_{n,k\mathbf{x}}.\tag{58}$$
In the M step, maximization of this sum yields θ ^n, kx <sup>¼</sup> <sup>N</sup> <sup>n</sup>, kx<sup>=</sup> P <sup>x</sup>0N <sup>n</sup>, kx0. □
#### 4 Markov Chains
A stochastic process fXt,t∈T g is a collection of random variables with common state spaceX. The index setT is usually interpreted as time and Xt is the state of the process at time t. A discrete-time stochastic process <sup>X</sup> ¼ (X1, <sup>X</sup>2, <sup>X</sup>3, ... ) is called a Markov chain [14], if Xn+1 <sup>⊥</sup> Xn-<sup>1</sup> <sup>j</sup> Xn for all <sup>n</sup> 2 or, equivalently, if each state depends only on its immediate predecessor:
$$P(X\_n \mid X\_{n-1}, \dots, X\_1) = P(X\_n \mid X\_{n-1}), \quad \text{for all } n \ge 2. \tag{59}$$
We consider here Markov chains with finite state space X¼½K¼f1, ... , <sup>K</sup>g that are homogeneous, i.e., with transition probabilities independent of time:
$$T\_{kl} = P(X\_{n+1} = l \mid X\_n = k), \quad \text{for all } k, l \in [K], \ n \ge 2. \qquad (60)$$
The finite-state homogeneous Markov chain is a statistical model denoted MCðΠ, <sup>T</sup> Þ and defined by the initial state distribution <sup>Π</sup> <sup>∈</sup> <sup>Δ</sup>K-1, where <sup>Π</sup><sup>k</sup> <sup>¼</sup> <sup>P</sup>(X<sup>1</sup> <sup>¼</sup> <sup>k</sup>), and the stochastic <sup>K</sup> <sup>K</sup> transition matrix <sup>T</sup> ¼ (Tkl).
We can generalize the one-step transition probabilities Tkl to:
$$T\_{kl}^{\
u} = P(X\_{\nu+j} = l \mid X\_j = k),\tag{61}$$
the probability of jumping from state k to state l in n time steps. Any (n + m)-step transition can be regarded as an n-step transition followed by an m-step transition. Because the intermediate state i is unknown, summing over all possible values yields the decomposition:
$$T\_{kl}^{n+m} = \sum\_{i=1}^{K} T\_{ki}^{n} T\_{il}^{m}, \quad \text{for all } n, m \ge 1, \ k, l \in [K], \tag{62}$$
known as the Chapman–Kolmogorov equations. In matrix notation, they can be written as <sup>T</sup>(n+m) ¼ <sup>T</sup>(n) T(m) . It follows that the n-step transition matrix is the n-th matrix power of the one-step transition matrix, <sup>T</sup>(n) ¼ <sup>T</sup><sup>n</sup> .
A state l of a Markov chain is accessible from state k if T <sup>n</sup> kl > 0. We say that <sup>k</sup> and <sup>l</sup> communicate with each other and write <sup>k</sup> <sup>l</sup> if they are accessible from one another. State communication is reflexive (<sup>k</sup> <sup>k</sup>), symmetric (<sup>k</sup> <sup>l</sup> ) <sup>l</sup> <sup>k</sup>), and, by the Chapman–Kolmogorov equations, transitive (<sup>j</sup> <sup>k</sup> <sup>l</sup> ) <sup>j</sup> <sup>l</sup>). Hence, it defines an equivalence relation on the state space. The Markov chain is irreducible if it has a single communication class, i.e., if any state is accessible from any other state.
A state is recurrent if the Markov chain will reenter it with probability one. Otherwise, the state is transient. In finite-state Markov chains, recurrent states are also positive recurrent, i.e., the expected time to return to the state is finite. A state is aperiodic if the process can return to it after any time <sup>n</sup> 1. Recurrence, positive recurrence, and aperiodicity are class properties: if they hold for a state k, then they also hold for all states communicating with k.
A Markov chain is ergodic if it is irreducible, aperiodic, and positive recurrent. An ergodic Markov chain has a unique stationary distribution π given by:
$$\pi\_l = \lim\_{n \to \infty} T\_{kl}^n = \sum\_{k=1}^K \pi\_k T\_{kl}, \qquad l \in [K], \quad \sum\_{l=1}^K \pi\_l = 1 \tag{63}$$
independent of the initial distribution Π. In matrix notation, π is the solution of <sup>π</sup><sup>t</sup> ¼ <sup>π</sup><sup>t</sup> T.
Example 17 (Two-State Markov Chain): Consider the Markov chain with state space {1, 2} and transition probabilities <sup>T</sup><sup>12</sup> <sup>¼</sup> <sup>α</sup> <sup>&</sup>gt; 0 and <sup>T</sup><sup>21</sup> <sup>¼</sup> <sup>β</sup> <sup>&</sup>gt; 0. Clearly, the chain is ergodic and its stationary distribution π is given by:
$$(\begin{pmatrix} \pi\_1 & \pi\_2 \end{pmatrix} = (\pi\_1 & \pi\_2) \begin{pmatrix} 1-a & a\\ \beta & 1-\beta \end{pmatrix} \tag{64}$$
or, equivalently, απ<sup>1</sup> <sup>¼</sup> βπ2. With <sup>π</sup><sup>1</sup> <sup>+</sup> <sup>π</sup><sup>2</sup> <sup>¼</sup> 1, we obtain <sup>π</sup><sup>t</sup> <sup>¼</sup> (<sup>α</sup> <sup>+</sup> β) -1 (α, β). □
In Example 17, if <sup>α</sup> ¼ 0, then state 1 is called an absorbing state because once entered it is never left. In evolutionary biology and population genetics, Markov chains are often used to model evolving populations, and the fixation probability of an allele can be computed as the absorption probability in such models.
Example 18 (Wright–Fisher Process): We consider two alleles, A and a, in a diploid population of size N. The total number of A alleles in generation n is described by a Markov chain Xn with state space {0, 1, 2, ..., 2N}. We assume that individuals mate randomly and that maternal and paternal alleles are chosen randomly such that <sup>ð</sup>Xnþ<sup>1</sup> <sup>j</sup> XnÞ Binomð2<sup>N</sup> , <sup>k</sup>=ð2<sup>N</sup> ÞÞ, where <sup>k</sup> is the number of A alleles in generation n. The Markov chain has transition probabilities:
$$T\_{kl} = \binom{2N}{l} \left(\frac{k}{2N}\right)^l \left(\frac{2N-k}{2N}\right)^{2N-l}.\tag{65}$$
If the initial number of A alleles is <sup>X</sup><sup>1</sup> <sup>¼</sup> <sup>k</sup>, then E(X1) <sup>¼</sup> <sup>k</sup>. After binomial sampling, E(X2) ¼ <sup>2</sup>N(k/(2N)) ¼ <sup>k</sup> and hence E(Xn) ¼ <sup>k</sup> for all <sup>n</sup> 0. The Markov chain has the two absorbing states 0 and 2N, which correspond, respectively, to extinction and fixation of the A allele. To compute the fixation probability hk of A given k initial copies of it:
$$h\_k = \lim\_{n \to \infty} P(X\_n = 2N \mid X\_1 = k),\tag{66}$$
we consider the expected value, which is equal to k, in the limit as <sup>n</sup> !1 to obtain
$$k = \lim\_{n \to \infty} \mathcal{E}(X\_n) = 0 \cdot (1 - h\_k) + 2 \,\mathrm{N} \cdot h\_k. \tag{67}$$
Thus, the fixation probability is just hk <sup>¼</sup> <sup>k</sup>/(2N), the initial relative frequency of the allele. The Wright–Fisher process [15, 16] is a basic stochastic model for random genetic drift, i.e., for the variation in allele frequencies only due to random sampling. □
If we observe data <sup>X</sup> ¼ (X(1), ..., <sup>X</sup>(N) ) from a finite Markov chain MCðΠ, <sup>T</sup> Þ of length <sup>L</sup>, then the likelihood is
$$\begin{split} L(\Pi, T) &= \prod\_{i=1}^{N} P(X^{(i)}) = \prod\_{i=1}^{N} P(X\_1^{(i)}) \prod\_{n=1}^{L-1} P(X\_{n+1}^{(i)} \mid X\_n^{(i)}) \\ &= \prod\_{i=1}^{N} \Pi\_{X\_1^{(i)}} \prod\_{n=1}^{L-1} T\_{X\_n^{(i)}, X\_{n+1}^{(i)}}, \end{split} \tag{68}$$
which can be rewritten as:
$$\begin{split} L(\Pi, T) &= \prod\_{i=1}^{N} \prod\_{k \in [K]} \Pi\_k^{N\_k(X^{(i)})} \prod\_{k \in [K]} \prod\_{l \in [K]} T\_{kl}^{N\_{kl}(X^{(i)})} \\ &= \prod\_{k \in [K]} \Pi\_k^{N\_k} \prod\_{k \in [K]} \prod\_{l \in [K]} T\_{kl}^{N\_{kl}}. \end{split} \tag{69}$$
with Nkl(X(i) ) the number of observed transitions from state k into state l in observation X(i) , and <sup>N</sup> kl <sup>¼</sup> <sup>P</sup><sup>N</sup> <sup>i</sup>¼<sup>1</sup> <sup>N</sup> klðXði<sup>Þ</sup> Þ the total number of k-to-l transitions in the data, and similarly Nk(X(i) ) and Nk the number of times the i-th chain, respectively all chains, started in state k.
Exercise 19 (Markov Chains): Let us consider a simple infectious disease model, where each individual is either healthy (H) or diseased (D). We assume the following two-state Markov chain to describe infection-related disease and recovery via clearance of the pathogen:
The probability of a healthy individual becoming sick due to infection is <sup>α</sup> ¼ 0.6, and the probability of a diseased individual to clear the infection and recover is <sup>β</sup> ¼ 0.9. The initial probabilities for health and disease are <sup>P</sup>(H) ¼ 0.7 and <sup>P</sup>(D) ¼ 0.3. Write down the transition matrix T of this Markov chain. What is the probability of observing the disease trajectories DDHHD and HDHDH? Calculate the stationary distribution of the Markov chain.
#### 5 Continuous-Time Markov Chains
A continuous-time stochastic process {X(t), <sup>t</sup> 0} with finite state space [K] is a continuous-time Markov chain if
$$\begin{aligned} P[X(t+\mathfrak{s}) = l \mid X(\mathfrak{s}) = k, X(\mathfrak{u}) = \mathfrak{x}(\mathfrak{u}), \mathbf{0} \le \mathfrak{u} < \mathfrak{s}] \\ = P[X(t+\mathfrak{s}) = l \mid X(\mathfrak{s}) = k] \end{aligned} \tag{70}$$
for all <sup>s</sup>, <sup>t</sup> 1, <sup>k</sup>, <sup>l</sup>, <sup>x</sup>(u) <sup>∈</sup> [K], 0 <sup>u</sup> <sup>&</sup>lt; <sup>s</sup>. The chain is homogeneous if Eq. 70 is independent of s. The transition probabilities are then denoted:
$$T\_{kl}(t) = P[X(t+\mathfrak{s}) = l \mid X(\mathfrak{s}) = k]. \tag{71}$$
It can be shown that the transition matrix T(t) is the matrix exponential of a constant rate matrix R times t:
$$T(t) = \exp(Rt) = \sum\_{j=0}^{\infty} \frac{1}{j!} (Rt)^j. \tag{72}$$
Example 20 (Jukes–Cantor Model): Consider a fixed position in a DNA sequence, and let Tkl(t) be the probability that, due to mutation, nucleotide k changes to nucleotide l after time t at this position (Fig. 8). The Jukes–Cantor model [17] is the simplest DNA substitution model. It assumes that the transition rates from any nucleotide to any other are equal:
$$R = \begin{pmatrix} -3a & a & a & a \\ a & -3a & a & a \\ a & a & -3a & a \\ a & a & a & -3a \end{pmatrix}. \tag{73}$$
$$\begin{pmatrix} \ddots \\ \x \\ \vdots \\ \x \end{pmatrix} \begin{pmatrix} \cdot \\ \cdot \\ \cdot \\ \cdot \end{pmatrix}$$
Fig. 8 Nucleotide substitution model. The state space and transitions of a general nucleotide substitution model are shown. For the Jukes–Cantor model (Example 20), all transitions from any nucleotide to any other nucleotide have the same probability <sup>1</sup> <sup>4</sup> <sup>ð</sup><sup>1</sup> e-4αt Þ
The resulting transition matrix <sup>T</sup> ðtÞ ¼ expðRtÞ is
$$T(t) = \frac{1}{4} \begin{pmatrix} 1 + 3e^{-4at} & 1 - e^{-4at} & 1 - e^{-4at} & 1 - e^{-4at} \\ 1 - e^{-4at} & 1 + 3e^{-4at} & 1 - e^{-4at} & 1 - e^{-4at} \\ 1 - e^{-4at} & 1 - e^{-4at} & 1 + 3e^{-4at} & 1 - e^{-4at} \\ 1 - e^{-4at} & 1 - e^{-4at} & 1 - e^{-4at} & 1 + 3e^{-4at} \end{pmatrix} \tag{74}$$
and the stationary distribution as <sup>t</sup> !1 is uniform, <sup>π</sup> ¼ (1/4, 1/4, 1/4, 1/4)<sup>t</sup> . □
Example 21 (The Poisson Process): A continuous-time Markov chain X(t) is a counting process, if X(t) represents the total number of events that occur by time t. It is a Poisson process, if in addition <sup>X</sup>(0) ¼ 0, the increments are independent, and in any interval of length t the number of events is Poisson distributed with rate λt:
$$P[X(t+\mathfrak{s}) - X(\mathfrak{s}) = k] = P[X(t) = k] = e^{-\lambda t} \frac{\left(\lambda t\right)^k}{k!}.\tag{75}$$
The Poisson process is used, for example, to count mutations in a gene. □
Example 22 (Exponential Distribution): The exponential distribution ExpðλÞ with parameter <sup>λ</sup> <sup>&</sup>gt; 0 is a common distribution for waiting times. It is defined by the density function:
$$f(\varkappa) = \lambda \varepsilon^{-\lambda \varepsilon}, \quad \text{for } \varkappa \ge 0. \tag{76}$$
If <sup>X</sup> ExpðλÞ, then <sup>X</sup> has expectation E(X) ¼ <sup>λ</sup> -<sup>1</sup> and variance VarðXÞ ¼ <sup>λ</sup>-2. The exponential distribution is memoryless, which means that <sup>P</sup>(<sup>X</sup> <sup>&</sup>gt; <sup>s</sup> <sup>+</sup> <sup>t</sup> j <sup>X</sup> <sup>&</sup>gt; <sup>t</sup>) ¼ <sup>P</sup>(<sup>X</sup> <sup>&</sup>gt; <sup>s</sup>), for all <sup>s</sup>, <sup>t</sup> <sup>&</sup>gt; 0. An important consequence of the memoryless property is that the waiting times between successive events are i.i.d. For example, the waiting times <sup>τ</sup><sup>n</sup> (<sup>n</sup> 1) between the events of a Poisson process, the sequence of interarrival times, are exponentially distributed, <sup>τ</sup><sup>n</sup> ExpðλÞ, for all <sup>n</sup> 1. □
Exercise 23 (Kimura Model): The Kimura two-parameter model is a DNA substitution model that distinguishes between transitions, i.e., purine-to-purine and pyrimidine-to-pyrimidine substitutions, from transversions, i.e., purine-to-pyrimidine and pyrimidine-topurine substitutions [18]. It is defined by the rate matrix:
$$R = \begin{pmatrix} -2\beta - a & \beta & a & \beta \\ \beta & -2\beta - a & \beta & a \\ a & \beta & -2\beta - a & \beta \\ \beta & a & \beta & -2\beta - a \end{pmatrix},$$
where <sup>α</sup>, <sup>β</sup>∈þ are the two substitution rates. Assuming that the Markov chain is ergodic, derive its stationary distribution.
#### 6 Hidden Markov Models
A hidden Markov model (HMM) is a statistical model for hidden random variables <sup>Z</sup> <sup>¼</sup> (Z1, ..., ZL), which form a homogeneous Markov chain, and observed random variables <sup>X</sup> <sup>¼</sup> (X1, ..., XL). Each observed symbol Xn depends on the hidden state Zn. The HMM is illustrated in Fig. 9. It encodes the following conditional independence statements:
$$Z\_{n+1} \perp Z\_{n-1} \mid Z\_n, \qquad 2 \le n \le L - 1 \qquad \text{(Markov property)}\tag{77}$$
$$X\_n \perp X\_m \mid Z\_n, \qquad 1 \le m, n \le L, \ m \ne n \tag{78}$$
The parameters of the HMM consist of the initial state probabilities <sup>Π</sup> <sup>¼</sup> <sup>P</sup>(Z1), the transition probabilities Tkl <sup>¼</sup> <sup>P</sup>(Zn <sup>¼</sup> <sup>l</sup> <sup>j</sup> Zn-<sup>1</sup> <sup>¼</sup> <sup>k</sup>) of the Markov chain, and the emission probabilities Ekx <sup>¼</sup> <sup>P</sup>(Xn <sup>¼</sup> <sup>x</sup> <sup>j</sup> Zn <sup>¼</sup> <sup>k</sup>) of symbols <sup>x</sup>∈X. The HMM is denoted HMMðΠ, <sup>T</sup> , <sup>E</sup>Þ. For simplicity, we restrict ourselves here to finite state spaces <sup>Z</sup> ¼ ½K of <sup>Z</sup> and X of <sup>X</sup>. The joint probability of (Z, <sup>X</sup>) factorizes as:
$$\begin{split}P(X,Z) &= P(Z\_1) \prod\_{n=1}^{L-1} P(X\_n \mid Z\_n) P(Z\_{n+1} \mid Z\_n) \\ &= \Pi\_{Z\_1} \prod\_{n=1}^{L-1} E\_{Z\_n, X\_n} T\_{Z\_n, Z\_{n+1}}.\end{split} \tag{79}$$
The HMM is typically used to model sequence data <sup>x</sup> ¼ (x1, <sup>x</sup>2, ..., xL) generated by different mechanisms zn which cannot be observed. Each observation x can be a time series or any other object with a linear dependency structure [19]. In computational biology, the HMM is frequently applied to DNA and protein sequence data, where it accounts for first-order spatial dependencies of nucleotides or amino acids [20].
Fig. 9 Hidden Markov model. Shaded nodes represent observed random variables (or symbols) Xn, and clear nodes represent hidden states (or the annotation). Directed edges indicate statistical dependencies which are given, respectively, by transition and emission probabilities among hidden states and between hidden states and observed symbols
Example 24 (CpG Islands): CpG islands are CG-enriched regions in a DNA sequence. They are typically a few hundreds to thousands of base pairs long. We want to use a simple HMM to detect CpG islands in genomic DNA. The hidden states Zn∈<sup>Z</sup> ¼ f-, þg indicate whether sequence position n belongs to a CpG island (+ ) or not (-). The observed sequence is given by the nucleotide at each position, Xn∈X¼fA, <sup>C</sup>, <sup>G</sup>, <sup>T</sup>g.
Suppose we observe the sequence <sup>x</sup> ¼ (C, <sup>A</sup>, <sup>C</sup>, <sup>G</sup>). Then, we can calculate the joint probability of x and any state path z by Eq. 79. For example, if <sup>z</sup> ¼ (+, -, -, +), then <sup>P</sup>(<sup>X</sup> <sup>¼</sup> <sup>x</sup>, <sup>Z</sup> <sup>¼</sup> <sup>z</sup>) <sup>¼</sup> <sup>Π</sup><sup>+</sup> <sup>E</sup>+, <sup>C</sup>T+,- <sup>E</sup>-,AT-,- <sup>E</sup>-,CT-,+ E+,G. □
Typically, one is interested in the hidden state path <sup>z</sup> ¼ (z1, <sup>z</sup>2, ..., zL) that gave rise to the observation x. For biological sequences, z is often called the annotation of x. In Example 24, the genomic sequence is annotated with CpG islands. For generic parameters, any state path can give rise to a given observed sequence, but with different probabilities. The decoding problem is to find the annotation z <sup>∗</sup> that maximizes the joint probability:
$$z^\* = \underset{z \in Z}{\text{argmax}} \ P(X = \varkappa, Z = z). \tag{80}$$
There are KL possible state paths such that already for sequences of moderate length, the optimization problem (Eq. 80) cannot be solved by enumerating all paths.
However, there is an efficient algorithm solving (Eq. 80) based on the following factorization along the Markov chain:
$$\max\_{Z} P(X, Z) = \max\_{Z\_1, \dots, Z\_L} P(Z\_1) \prod\_{n=1}^{L-1} P(X\_n \mid Z\_n) P(Z\_{n+1} \mid Z\_n)$$
$$= \max\_{Z\_L} P(Z\_L \mid Z\_{L-1}) P(X\_L \mid Z\_L)$$
$$\left[ \dots \left[ \max\_{Z\_2} P(Z\_3 \mid Z\_2) P(X\_2 \mid Z\_2) \right]$$
$$\left[ \max\_{Z\_1} P(Z\_2 \mid Z\_1) P(X\_1 \mid Z\_1) \cdot P(Z\_1) \right] \dots \right]. \tag{81}$$
Thus, the maximum over state paths (Z1, ..., ZL) can be obtained by recursively computing maxima over each Zn. Each of the L terms in parenthesis defines a probability distribution over K states by maximizing over K values. Hence, the time complexity of the algorithm is O(LK<sup>2</sup> ), despite the fact that the maximum is over KL paths. This procedure is known as dynamic programming and it is the workhorse of biological sequence analysis. For HMMs, it is known as the Viterbi algorithm [21].
In order to compute the marginal likelihood <sup>P</sup>(<sup>X</sup> ¼ <sup>x</sup>) of an observed sequence <sup>x</sup>, we need to sum the joint probability <sup>P</sup>(<sup>Z</sup> ¼ <sup>z</sup>, <sup>X</sup> ¼ <sup>x</sup>) over all hidden states <sup>z</sup>∈Z. The length of this sum is KL , but it can be computed efficiently by the same dynamic programming principle used for the Viterbi algorithm:
$$\begin{split} \sum\_{Z} P(X, Z) &= \sum\_{Z\_{1}, \dots, Z\_{L}} P(Z\_{1}) \prod\_{n=1}^{L-1} P(X\_{n} \mid Z\_{n}) P(Z\_{n+1} \mid Z\_{n}) \\ &= \sum\_{Z\_{L}} P(Z\_{L} \mid Z\_{L-1}) P(X\_{L} \mid Z\_{L}) \end{split} \tag{4.10}$$
$$\left[ \dots \left[ \sum\_{Z\_{2}} P(Z\_{3} \mid Z\_{2}) P(X\_{2} \mid Z\_{2}) \right]$$
$$\left[ \sum\_{Z\_{1}} P(Z\_{2} \mid Z\_{1}) P(X\_{1} \mid Z\_{1}) \cdot P(Z\_{1}) \right] \dots \right]. \tag{8.2}$$
Indeed, this factorization is the same as in Eq. 81 with maxima replaced by sums. The recursive algorithm implementing (Eq. 82) is known as the forward algorithm. In each step, it computes the partial solution <sup>f</sup>(n, Zn) ¼ <sup>P</sup>(X1, ..., Xn, Zn).
The factorization along the Markov chain can also be done in the other direction starting the recursion from ZL down to Z1. The resulting backward algorithm generates the partial solutions b(n, Zn) <sup>¼</sup> <sup>P</sup>(Xn+1, ..., XL <sup>j</sup> Zn). From the forward and backward quantities, one can also compute the position-wise posterior state probabilities:
$$P(Z\_n \mid X) = \frac{P(X, Z\_n)}{P(X)} = \frac{P(X\_1, \dots, X\_n, Z\_n)P(X\_{n+1}, \dots, X\_L \mid Z\_n)}{P(X)}$$
$$= \frac{f(n, Z\_n)b(n, Z\_n)}{P(X)}.\tag{83}$$
For example, in the CpG island HMM (Example 24), we can compute, for each nucleotide, the probability that it belongs to a CpG island given the entire observed DNA sequence. Selecting the state that maximizes this probability independently at each sequence position is known as posterior decoding. In general, the result will be different from Viterbi decoding.
Example 25 (Pairwise Sequence Alignment): The pair HMM is a statistical model for pairwise alignment of two observed sequences over a fixed alphabet A. For protein sequences, A is the set of 20 natural amino acids and for DNA sequences, A consists of the four nucleotides, plus the gap symbol ("-"). At each position of the alignment, a hidden variable Zn∈<sup>Z</sup> ¼ fM, <sup>X</sup>, <sup>Y</sup>} indicates whether there is a (mis-)match (M), an insertion (X), or a deletion (Y) in sequence y relative to sequence x. For example:
```
z ¼ MMMMMMMMMMMMMXXMMMMMMMMMMMMYMMMMYMMMMM
x ¼ CTRPNNNTRKSIRPQIGPGQAFYATGD-
IGDI-
RQAHC
y ¼ CGRPNNHRIKGLR-
-
IGPGRAFFAMGAIRGGEIRQAHC
```
The emitted symbols are pairs (Xn, Yn) of aligned sequence characters with state space ðA AÞ∖fð-, -Þg. Thus, a pairwise alignment is a probabilistically generated sequence of pairs of symbols.
The choice of transition and emission probabilities corresponds to fixing a scoring scheme in nonprobabilistic formulations of sequence alignment. For example, the emission probabilities <sup>P</sup>[(a, <sup>b</sup>) j <sup>M</sup>] from a match state encode pairwise amino acid preferences and can be modeled by substitution matrices, such as PAM and BLOSUM [20].
In the pair HMM, computing an optimal alignment between x and y means to find the most probable state path <sup>z</sup><sup>∗</sup> ¼ argmaxzPð<sup>X</sup> ¼ <sup>x</sup>, <sup>Y</sup> ¼ <sup>y</sup>, <sup>Z</sup> ¼ <sup>z</sup>), which can be solved using the Viterbi algorithm. Using the forward algorithm, we can also compute efficiently the marginal probability of two sequences being related independent of their alignment, <sup>P</sup>(X, <sup>Y</sup> ) <sup>¼</sup>∑ZP (X, Y, Z). In general, this probability is more informative than the posterior <sup>P</sup>(<sup>Z</sup> j <sup>X</sup>, <sup>Y</sup> ) of an optimal alignment <sup>z</sup> <sup>∗</sup> because many alignments tend to have the same or nearly the same probability such that <sup>P</sup>(<sup>Z</sup> ¼ <sup>z</sup> ∗ j <sup>X</sup>, <sup>Y</sup> ) can be very small. Finally, we can also compute the probability of two characters xn and ym being aligned by means of posterior decoding. □
Example 26 (Profile HMM): Profile hidden Markov models represent groups of related sequences, such as protein families. They are used for searching homologous sequences and for building multiple sequence alignments. They can be regarded as unrolled versions of the pair HMM. A profile HMM is a statistical model for observed sequences, which are regarded as i.i.d. realizations. It has sitespecific emission probabilities En(a) <sup>¼</sup> <sup>P</sup>(Xn <sup>¼</sup> <sup>a</sup>). In its simplest form allowing only gap-free alignments, the probability of an observation x is just
$$P(X = \boldsymbol{\pi}) = \prod\_{n=1}^{L} E\_n(\boldsymbol{\pi}\_i). \tag{84}$$
The matrix ð Þ Enða<sup>Þ</sup> <sup>1</sup>nL, <sup>a</sup>∈A is called a position-specific scoring matrix (PSSM).
Profile HMMs can also model indels. Figure 10 shows the hidden state space of such a model. It has match states Mn, which can emit symbols according to the probability tables En, insert states In, which usually emit symbols in an unspecific manner, and delete states Dn, which do not emit any symbols. The possible transitions between those states allow for modeling alignment gaps of any length.
Fig. 10 Profile hidden Markov model. The hidden state space and its transitions are shown for the profile HMM of length <sup>L</sup> <sup>¼</sup> 3. Match states are denoted Mn, insert states In, and delete states Dn. B and E denote silent begin and end states, respectively. With match and insert states, probability tables for the emissions of symbols (amino acids or nucleotides, and gaps) are associated
A given profile HMM for a protein family can be used to detect new sequences that belong to the same family. For a query sequence x, we can either consider the most probable alignment of the sequence to the HMM, <sup>P</sup>(<sup>X</sup> ¼ <sup>x</sup>, <sup>Z</sup> ¼ <sup>z</sup> <sup>∗</sup>), or the marginal probability independent of the alignment, <sup>P</sup>(<sup>X</sup> <sup>¼</sup> <sup>x</sup>) <sup>¼</sup>∑ZP(<sup>X</sup> ¼ <sup>x</sup>, <sup>Z</sup>), to decide about family membership. □
Parameter estimation in HMMs is complicated by the presence of hidden variables. In Subheading 2, the EM algorithm has been introduced for finding a local maximum of the likelihood surface. For HMMs, the EM algorithm is known as the Baum–Welch algorithm [22]. For simplicity, let us ignore the initial state probabilities Π and summarize the parameters of the HMM by <sup>θ</sup> ¼ (T, <sup>E</sup>). For ML estimation, we need to maximize the observed log-likelihood:
$$\begin{split} \ell\_{\text{obs}}(\theta) &= \log P(X \mid \theta) = \log \sum\_{Z} P(X, Z \mid \theta) \\ &= \log \sum\_{Z^{(1)}, \dots, Z^{(N)}} \prod\_{i=1}^{N} P(X^{(i)}, Z^{(i)} \mid \theta), \end{split} \tag{85}$$
where X(1), ..., X(N) are the i.i.d. observations. For each observation, we can rewrite the joint probability as:
$$P(X^{(i)}, Z^{(i)} \mid \theta) = \prod\_{k \in [K]} \prod\_{\mathbf{x} \in \mathcal{X}} E\_{k\mathbf{x}}^{N\_{k\mathbf{x}}(Z^{(i)})} \cdot \prod\_{k \in [K]} \prod\_{l \in [K]} T\_{kl}^{N\_{kl}(Z^{(i)})},\tag{86}$$
where Nkx(Z(i) ) is the number of x emissions when in state k and Nkl(Z(i) ) the number of k-to-l transitions in state path Z(i) (cf. Eq. 68).
In the E step, the expectation of Eq. 85 is computed with respect to <sup>P</sup>(<sup>Z</sup> j <sup>X</sup>, <sup>θ</sup><sup>0</sup> ), where θ<sup>0</sup> is the current best estimate of θ. We use Eq. 86 and denote by Nkx and Nkl the expected value of ∑iNkx(Z(i) ) and ∑iNkl(Z(i) ), respectively, to obtain
$$\begin{split} \mathbb{E}[\ell\_{hid}(\theta)] &= \sum\_{Z} P(Z \mid X, \theta') \log P(X, Z \mid \theta) \\ &= \sum\_{Z^{(1)}, \dots, Z^{(N)}} P(Z \mid X, \theta') \\ &= \left[ \sum\_{k, \, \, x} N\_{kx}(Z^{(i)}) \log E\_{kx} + \sum\_{k, \, \, l} N\_{kl}(Z^{(i)}) \log T\_{kl} \right] \\ &= \sum\_{k, \, \, x} N\_{kx} \log E\_{kx} + \sum\_{k, \, \, l} N\_{kl} \log T\_{kl} . \end{split} \tag{87}$$
The expected counts Nkx and Nkl are the sufficient statistics [11] of the HMM, i.e., with respect to the model, they contain all information about the parameters available from the data. The expected counts can be computed using the forward and backward algorithms. In the M step, this expression is maximized with respect to <sup>θ</sup> <sup>¼</sup> (T, <sup>E</sup>). We find the MLEs <sup>T</sup>^ kl <sup>¼</sup> <sup>N</sup> kl<sup>=</sup> P mN km and <sup>E</sup>^kx <sup>¼</sup> <sup>N</sup> kx<sup>=</sup> P <sup>y</sup>N ky.
#### 7 Bayesian Networks
Bayesian networks are a class of probabilistic graphical models which generalize Markov chains and HMMs. The basic idea is to use a graph for encoding conditional independences among random variables (Fig. 11). The graph representation provides not only an intuitive and simple visualization of the model structure, but it is also the basis for designing efficient algorithms for inference and learning in graphical models [23–25].
A Bayesian network (BN) for a set of random variables <sup>X</sup> <sup>¼</sup> (X1, ..., XL) consists of a directed acyclic graph (DAG) and local probability distributions (LPDs). The DAG <sup>G</sup> ¼ (V, <sup>E</sup>) has vertex set <sup>V</sup> ¼ [L] and edge set <sup>E</sup> <sup>V</sup> <sup>V</sup> . Each vertex <sup>n</sup> <sup>∈</sup> <sup>V</sup> is identified with the random variable Xn. If there is an edge Xm ! Xn in G, then Xm is a parent of Xn and Xn is a child of Xm. For each vertex <sup>n</sup> <sup>∈</sup> <sup>V</sup> , there is an LPD <sup>P</sup>ðXn <sup>j</sup> <sup>X</sup>paðnÞÞ, where paðn<sup>Þ</sup> is the set of parents of Xn in G. The Bayesian network model is defined as the family of distributions for which the joint probability of X factors into conditional probabilities as:
$$P(X\_1, \ldots, X\_L) = \prod\_{n=1}^{L} P(X\_n \mid X\_{\text{pa}(n)}).\tag{88}$$
Fig. 11 Example of a Bayesian network. Vertices correspond to random variables and edges represent conditional probabilities. The graph encodes conditional independence statements about the random variables U, V , W, X, Y , and Z. Their joint probability factors according to the graph as <sup>P</sup>(U, <sup>V</sup>, <sup>W</sup>, <sup>X</sup>, <sup>Y</sup> ) ¼ <sup>P</sup>(U) <sup>P</sup>(<sup>Y</sup> )P(<sup>V</sup> j <sup>U</sup>, <sup>Y</sup> )P(<sup>W</sup> j <sup>V</sup> )P(<sup>X</sup> j <sup>U</sup>)
In this case, we write <sup>X</sup> BNðG, <sup>θ</sup>Þ, where <sup>θ</sup> ¼ (θ1, ..., <sup>θ</sup>L) denotes the parameters of the LPDs.
For the Bayesian network shown in Fig. 11, we find P(U, V, W, <sup>X</sup>, <sup>Y</sup> ) ¼ <sup>P</sup>(U)P(<sup>Y</sup> )P(<sup>V</sup> j <sup>U</sup>, <sup>Y</sup> )P(<sup>W</sup> j <sup>V</sup> )P(<sup>X</sup> j <sup>U</sup>). The graph encodes several conditional independence statements about (U, <sup>V</sup>, <sup>W</sup>, <sup>X</sup>, <sup>Y</sup> ), including, for example, <sup>W</sup> <sup>⊥</sup>{U, <sup>X</sup>} j <sup>V</sup> .
Example 27 (Markov Chain): A finite Markov chain is a Bayesian network with the DAG <sup>X</sup><sup>1</sup> ! <sup>X</sup><sup>2</sup> ! ! XL, denoted <sup>C</sup>, and joint distribution:
$$P(X\_1, \ldots, X\_n) = P(X\_1)P(X\_2 \mid X\_1)P(X\_3 \mid X\_2) \cdots P(X\_L \mid X\_{L-1}).\tag{89}$$
If <sup>X</sup> MCðΠ, <sup>T</sup> <sup>Þ</sup> is homogeneous, then the LPDs are <sup>θ</sup><sup>1</sup> <sup>¼</sup> <sup>P</sup> (X1) <sup>¼</sup> <sup>Π</sup> and <sup>θ</sup>n+1 <sup>¼</sup> <sup>P</sup>(Xn+1 <sup>j</sup> Xn) <sup>¼</sup> <sup>T</sup> for all <sup>n</sup> <sup>∈</sup> [<sup>L</sup> - 1] such that MCðΠ, <sup>T</sup> Þ ¼ BNðC, <sup>θ</sup>Þ. Similarly, HMMs are Bayesian networks with hidden variables Z and factorized joint distribution given in Eq. 79. □
The meaning of the parameters θ of a Bayesian network depends on the family of distributions that has been chosen for the LPDs. In the general case of a discrete random variable with finite state space, θ<sup>n</sup> is a conditional probability table. If each vertex Xn has K possible states, then:
$$\theta\_n = \left( P(X\_n = a \mid X\_{\text{pa}(n)} = b) \right)\_{b \in [K]^{\text{pa}(n)}, a \in [K]} \tag{90}$$
has <sup>K</sup>paðn<sup>Þ</sup> ð<sup>K</sup> - <sup>1</sup><sup>Þ</sup> free parameters. If Xn depends on all other variables, then <sup>θ</sup><sup>n</sup> has the maximal number of KL - 1 parameters, which is exponential in the number of vertices. If, on the other hand, Xn is independent of all other variables, paðnÞ¼;, then <sup>θ</sup><sup>n</sup> has (<sup>K</sup> - 1) parameters, which is independent of L. For the chain (Example 27) where each vertex has exactly one outgoing and one incoming edge, we find a total of (<sup>K</sup> - 1) + (<sup>L</sup> - 1)K(<sup>K</sup> - 1) free parameters which is of order O(LK<sup>2</sup> ).
A popular model for continuous random variables Xn is the linear Gaussian model. Here, the LPDs are Gaussian distributions with mean a linear function of the parents:
$$P(X\_n \mid X\_{\text{pa}(n)}) = \text{Norm}(\nu\_n + \boldsymbol{w}\_n^t \cdot X\_{\text{pa}(n)}, \sigma\_n^2), \tag{91}$$
with parameters vn∈ and wi ∈paðn<sup>Þ</sup> specifying the mean and variance σ<sup>2</sup> <sup>n</sup>. The number of parameters increases linearly with the number of parents, but only linear relationships can be modeled. All marginal and conditional probabilities of (X1, ..., XL) are also Gaussians.
Learning a Bayesian network BNðG, <sup>θ</sup>Þfrom data<sup>D</sup> can be done in different ways following either the Bayesian or the maximum likelihood approach as introduced in Subheading 2. In general, it involves first finding the optimal network structure:
$$G^\* = \underset{G}{\text{argmax}} \ P(G \mid \mathcal{O}),\tag{92}$$
and then estimating the parameters:
$$\theta^\* = \underset{\theta}{\text{argmax}} \ P(\theta \mid G^\*, \,\,\mathcal{O}) \tag{93}$$
for the given optimal structure G∗. The first step is a model selection problem as introduced in Subheading 2.
Model selection for Bayesian networks is a particularly hard problem because the number of DAGs increases superexponentially with the number of vertices rendering exhaustive searches impractical, and the objective function in Eq. 92 is difficult to compute. Recall that the posterior <sup>P</sup>ð<sup>G</sup> j <sup>D</sup>Þ is proportional to the product <sup>P</sup>ð<sup>D</sup> j <sup>G</sup>ÞPðGÞ of marginal likelihood and network prior, and the marginal likelihood:
$$P(\mathcal{O} \mid G) = \int P(\mathcal{O} \mid \theta, \ G) P(\theta \mid G) \, d\theta \tag{94}$$
is usually analytically intractable. Here, <sup>P</sup>(<sup>θ</sup> j <sup>G</sup>) is the prior distribution of parameters given the network topology.
To address this limitation, the marginal likelihood (Eq. 94) can be approximated by a function that is easier to evaluate. A popular choice is the Bayesian information criterion (BIC) [26]:
$$\log P(\mathcal{O} \mid G) \approx \log P(\mathcal{O} \mid \hat{\theta}\_{\text{ML}}, G) - \frac{1}{2} \nu \log N,\qquad(95)$$
where ν is the number of free parameters of the model and N the size of the data. The BIC approximation can be derived under certain assumptions, including a unimodal likelihood. It replaces computation of the integral (Eq. 94) by evaluating the integrand at the MLE and adding the correction term ðνlog<sup>N</sup> Þ=2, which penalizes models of high complexity.
The model selection problem remains hard even with a tractable scoring function, such as BIC, because of the enormous search space. Local search methods, such as greedy hill climbing or simulated annealing, are often used in practice. They return a local maximum as a point estimate for the best network structure. Results can be improved by running several local searches from different starting topologies.
Often, data are sparse and we will find diffuse posterior distributions of network structures, which might not be represented very well by a single point estimate. In the fully Bayesian approach, we aim at estimating the full posterior <sup>P</sup>ð<sup>G</sup> j <sup>D</sup>Þ / <sup>P</sup>ð<sup>D</sup> j <sup>G</sup>ÞPðGÞ. One way to approximate this distribution is to draw a finite number of samples from it. Markov chain Monte Carlo (MCMC) methods generate such a sample by constructing a Markov chain that converges to the target distribution [27].
In the Metropolis–Hastings algorithm [28], we start with a random DAG G(0) and then iteratively generate a new DAG G(n) from the previous one G(n-1) by drawing it from a proposal distribution Q:
$$G^{(n)} \sim \mathcal{Q}(G^{(n)} \mid G^{(n-1)}).\tag{96}$$
The new DAG is accepted with acceptance probability:
$$\min\left\{\frac{P(\mathcal{O}\mid G^{(n)})P(G^{(n)})Q(G^{(n-1)}\mid G^{(n)})}{P(\mathcal{O}\mid G^{(n-1)})P(G^{(n-1)})Q(G^{(n)}\mid G^{(n-1)})},1\right\}\tag{97}$$
Otherwise, the model is left unchanged and the next sample is drawn. With this acceptance probability, it is guaranteed that the Markov chain is ergodic and converges to the desired distribution. After an initial burn-in phase, samples from the stationary phase of the chain are collected, say G(m) , ..., G(N) . Any feature f of the network (e.g., the presence of an edge or a subgraph) can be estimated as the expected value:
$$\mathcal{E}(f) = \sum\_{G} f(G)P(G \mid \mathcal{O}) \approx \frac{1}{N} \sum\_{n=n}^{N} f(G^{(n)}).\tag{98}$$
A critical point of the Metropolis–Hastings algorithm is the choice of the proposal distribution Q, which encodes the way the network space is explored. Because not all graphs, but only DAGs, are allowed, computing the transition probabilities <sup>Q</sup>(G(n) <sup>j</sup> <sup>G</sup>(n-1)) is usually the main computational bottleneck.
Parameter estimation, i.e., solving (Eq. 93), can be done along the lines described in Subheading 2 following either the ML or the Bayesian approach. If the model contains hidden random variables, then the EM algorithm (Subheading 3) can be used. However, this approach is feasible only if efficient inference algorithms are available. For hidden Markov models (Subheading 6), the forward and backward algorithms provided an efficient way to compute marginal probabilities and the expected hidden log-likelihood. These algorithms can be generalized to the sum–product algorithm for tree-like graphs and the junction tree algorithm for general DAGs. The computational complexity of the junction tree algorithm is exponential in the size of the largest clique of the so-called moralized graph, which is obtained by dropping edge directions and adding edges between any two vertices that have a common child in the original DAG [11].
Alternatively, if exact inference is computationally too expensive, then approximate inference can be used. For example, Gibbs sampling [29] is an MCMC technique for generating a sample from the joint distribution P(X1, ..., XL). The idea is to iteratively sample from the conditional probabilities of P(X1, ..., XL), starting with Xðnþ1<sup>Þ</sup> <sup>1</sup> <sup>P</sup>ðX<sup>1</sup> <sup>j</sup> <sup>X</sup>ðn<sup>Þ</sup> <sup>2</sup> , ... , <sup>X</sup>ðn<sup>Þ</sup> <sup>L</sup> <sup>Þ</sup> and cycling through all variables in turns:
$$X\_j^{(n+1)} \sim P(X\_j \mid X\_1^{(n+1)}, \dots, X\_{j-1}^{(n+1)}, X\_{j+1}^{(n)}, \dots, X\_L^{(n)}) \qquad (99)$$
$$\text{for all } j = 2, \dots, L.$$
Gibbs sampling can be regarded as a special case of the Metropolis–Hastings algorithm. It is particularly useful, if it is much easier to sample from the conditionals <sup>P</sup>(Xk <sup>j</sup> <sup>X</sup>∖k) than from the joint distribution P(X1, ..., XL), where X∖<sup>k</sup> denotes all variables Xn except Xk. For graphical models, the conditional probability of each vertex Xk depends only on its Markov blanket XMB(k), defined as the set of its parents, children, and co-parents (vertices with the same children), <sup>P</sup>(Xk <sup>j</sup> <sup>X</sup>∖k) <sup>¼</sup> <sup>P</sup>(Xk <sup>j</sup> <sup>X</sup>MB(k)).
Example 28 (Phylogenetic Tree Models): A phylogenetic tree model [30] for a set of aligned DNA sequences from different species is a Bayesian network model, where the graph is a tree in which the leaves represent the observed contemporary species and the interior vertices correspond to common extinct ancestors (Fig. 12). The topology (graph structure) S defines the branching order and the branch lengths correspond to (phylogenetic) time. The LPDs are defined by a nucleotide substitution model (Subheading 5).
Let X(i) ∈ {A, <sup>C</sup>, <sup>G</sup>, <sup>T</sup>} <sup>L</sup> denote the i-th column of a multiple sequence alignment of L observed species. We regard the alignment columns as independent observations of the evolutionary process. The character states of the hidden (extinct) ancestors are denoted Z(i) . The likelihood of the observed sequence data <sup>X</sup> ¼ (X(1), ..., X(N) ) given the tree topology S and the branch lengths t is
$$P(X \mid S, t) = \sum\_{Z} \prod\_{i=1}^{N} P(X^{(i)}, Z^{(i)} \mid S, t), \tag{100}$$
where P(X(i) , <sup>Z</sup>(i) j <sup>S</sup>, <sup>t</sup>) factors into conditional probabilities according to the tree structure. This marginal probability can be computed efficiently with an instance of the sum–product algorithm known as the peeling algorithm (or Felsenstein algorithm) [31].
Fig. 12 Phylogenetic tree model. The observed random variables Xi represent contemporary species and the hidden random variables Zi their unknown common ancestors
For example, in the tree displayed in Fig. 12, each observation X has probability:
$$P(X) = \sum\_{Z} P(X, Z) \tag{101}$$
$$\begin{split} \mathcal{Q} &= \sum\_{\mathcal{Z}} P(X\_1 \mid Z\_4) P(X\_2 \mid Z\_1) P(X\_3 \mid Z\_1) P(X\_4 \mid Z\_2) . \\ &\quad \cdot P(X\_5 \mid Z\_2) P(Z\_1 \mid Z\_3) P(Z\_2 \mid Z\_3) P(Z\_3 \mid Z\_4) P(Z\_4) \end{split} \tag{102}$$
$$\begin{split} &= \sum\_{\mathcal{Z}\_4} P(Z\_4) P(X\_1 \mid Z\_4) \left[ \sum\_{Z\_3} P(Z\_3 \mid Z\_4) \right. \\ &\quad \left[ \sum\_{Z\_2} P(Z\_2 \mid Z\_3) P(X\_4 \mid Z\_2) P(X\_5 \mid Z\_2) \right] \\ &\quad \cdot \left[ \sum\_{Z\_1} P(Z\_1 \mid Z\_3) P(X\_2 \mid Z\_1) P(X\_3 \mid Z\_1) \right] \Bigg], \end{split} \tag{103}$$
where we have omitted the dependency on the branch length t. Several software packages implement ML or Bayesian learning of phylogenetic tree models. □
In the simplest case, we suppose that the observed alignment columns are independent. However, it is more realistic to assume that nucleotide substitution rates vary across sites because of varying selective pressures. For example, there could be differences between coding and noncoding regions, among different regions of a protein (loops, and catalytic sites), or among the three bases of a triplet coding for an amino acid. More sophisticated models can account for this rate heterogeneity. Let us assume site-specific substitution rates ri such that the local probabilities become <sup>P</sup>(X(i) j ri, <sup>t</sup>, <sup>S</sup>). To model the distribution of the rates, often a gamma distribution is used.
Example 29 (Gamma Distribution): The gamma distribution Gammaðα, <sup>β</sup>Þ is parametrized by a shape parameter <sup>α</sup> and a rate parameter β. It is defined by the density function:
$$f(\varkappa) = \frac{\beta^{\alpha}}{\Gamma(\alpha)} \varkappa^{\alpha - 1} e^{-\beta \varkappa}, \quad \text{for } \varkappa \ge 0. \tag{104}$$
Its expectation is E(X) <sup>¼</sup> <sup>α</sup>/<sup>β</sup> and its variance VarðXÞ ¼ <sup>α</sup>=β2. The gamma distribution generalizes several other distributions, for example Gammað1, <sup>λ</sup>Þ ¼ ExpðλÞ (Example 22). □
Another approach to account for varying mutation rates are phylogenetic hidden Markov models (phylo-HMMs).
Example 30 (Phylo-HMM): Phylo-HMMs [32] combine HMMs and phylogenetic trees into a single Bayesian network model. The idea is to use an HMM along the linear chain of the genomic sequence and, at each position, to condition a phylogenetic tree model on the hidden state (Fig. 13). This architecture allows for modeling different evolutionary histories at different sites of the genome. In particular, the model can account for heterogeneity in the rate of evolution, for example, due to functionally conserved elements, but it also allows for a change in tree topology along the sequence, a situation that can result from recombination [23]. Phylo-HMMs are also used for gene finding. □
Exercise 31 (Inference in Bayesian Networks): Consider the gene network on five genes denoted A, B, C, D, E, with the graph structure displayed below. Gene expression profiles under different conditions C1–C9 have been observed and are summarized in the table below, where a zero indicates that the gene is not expressed and a one that it is expressed.
Fig. 13 Phylo-HMM. Shown are the first four positions of a Phylo-HMM. The hidden Markov chain has random variables Z. In the trees, Y denote the hidden common ancestors and X the observed species. Note that the tree topology changes between position 2 and 3
$$P(X\_i \mid X\_{\mathrm{pa}(i)}) \approx \frac{N(X\_i, X\_{\mathrm{pa}(i)})}{\sum\_k N(X\_i = k, \, X\_{\mathrm{pa}(i)})},$$
where <sup>N</sup> <sup>ð</sup>Xi, <sup>X</sup>pað<sup>i</sup>ÞÞis the number of joint observations of Xi and its parents.
#### References
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
#### A Not-So-Long Introduction to Computational Molecular Evolution
#### Ste´phane Aris-Brosou and Nicolas Rodrigue
#### Abstract
In this chapter, we give a not-so-long and self-contained introduction to computational molecular evolution. In particular, we present the emergence of the use of likelihood-based methods, review the standard DNA substitution models, and introduce how model choice operates. We also present recent developments in inferring absolute divergence times and rates on a phylogeny, before showing how state-of-the-art models take inspiration from diffusion theory to link population genetics, which traditionally focuses at a taxonomic level below that of the species, and molecular evolution. Although this is not a cookbook chapter, we try and point to popular programs and implementations along the way.
Key words Likelihood, Bayes, Model choice, Phylogenetics, Divergence times
#### 1 Introduction
Many books [1–7] and review papers [8–10] have been published in recent years on the topic of computational molecular evolution, so that updating our previous primer on the very same topic [11] may seem redundant. However, the field is continuously undergoing changes, as both models and algorithms become even more sophisticated, efficient, robust, and accurate. This increase in refinement has not been motivated by a desire to complicate existing models, but rather to make an old wish come true: that of having integrated methods that can take unaligned sequences as an input, and simultaneously output the alignment, the tree, and other estimates of interest, in a sound statistical framework justified by sound principles: those of population genetics.
The aim of this chapter is still to provide readers with the essentials of computational molecular evolution, offering a brief overview of recent progress, both in terms of modeling and algorithm development. Some of the details will be left out as they are dealt with by others in this volume. Likewise, the analysis of
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_3, © The Author(s) 2019
genomic-scale data is briefly touched upon, but the details are left to other chapters.
#### 2 Parsimony and Likelihood
2.1 A Brief Overview of Parsimony The simplest phylogenetic question pertains to the reconstruction of a rooted tree with three sequences (Fig. 1). The sequences can be made of DNA, RNA, amino acids, or codons, but for the sake of simplicity we focus on DNA throughout this chapter. In the basic example below, based on [12], DNA sequences are assumed to have been sampled from three different species that diverged a "long time ago." In this context, we assume that the data or gene sequences have been aligned (see Subheading 6), and that the DNA alignment is:
The objective is to estimate which of the three fully resolved topologies in Fig. 1 is supported by the data. In order to go further, we recode the data in terms of site patterns, which correspond to the patterns observed in each column of our alignment. This recoding implies that columns, or sites, in our alignment evolve according to an identically and independently distributed (iid) process. With this in mind, our alignment can be recoded in the following manner. When all the characters (nucleotides) in a column are identical, the same letter is assigned to each character, for example, x, irrespective of the actual character state. When a substitution occurs in one of the three sequences, we have three corresponding site patterns: xxy, xyx, and yxx, where the order within each site pattern respects the order of the sequences in the alignment, s1s2s3.
Fig. 1 The simplest phylogenetic problem. With three species, s<sup>1</sup> s2, and s3, four rooted trees are possible: T0, the star tree, and the three resolved topologies T1–T<sup>3</sup>
Table 1 The winning-site strategy
The data alignment is reduced to a frequency table of site patterns. In the case of three sequences, only the last three site patterns are informative
The first informative site pattern, xxy, implies that at this particular site, sequences s<sup>1</sup> and s<sup>2</sup> are more similar than any of these to s3, so that this site pattern supports topology T1, which groups sequences s<sup>1</sup> and s<sup>2</sup> together (Fig. 1). The most intuitive idea, called the winning-site strategy, is that the topology supported by the data corresponds to the fully resolved topology that has the largest number of site patterns in its favor. In the example shown above, topology T<sup>1</sup> is supported by three columns (with site pattern xxy), topology T<sup>2</sup> by two columns (xyx), and T<sup>3</sup> by one column (yxx; see Table 1). This is the intuition behind parsimony, which minimizes the amount of change along a topology. Strictly speaking, unordered parsimony cannot distinguish these three trees as they all require at least one single change. Yet, it can be argued that if tree T<sup>1</sup> is the true tree, site pattern xxy is more likely than any other patterns as xxy requires at least one change along a long branch (the one leading to sequence s3) while both xyx and yxx require a change along a short branch (see p. 28 sqq. in [13]; [12]).
A number of methodological variations exist. A very condensed overview can be found in the books by Durbin [14] or, with more details, Felsenstein [15]. Most computer programs that implement substitution models where sites are iid condense the alignment as an array of site patterns; some, like PAML [16], even output these site patterns.
Note that in obtaining this topology estimate, most of the site columns were discarded from our alignment (all the xxx site patterns, representing 89% of the site in our example above). Most of our data were phylogenetically uninformative (for parsimony). We also failed to take evolutionary time into account, or any process of basic molecular biology, such as the observation that transitions (substitution of a purine [A or G] by a purine, or a pyrimidine by a pyrimidine) are more frequent than transversions (substitution between a purine and a pyrimidine).
2.2 Assessing the Reliability of an Estimate: The Bootstrap As with any statistical exercise estimating a quantity of interest, we would like to have a confidence interval, taken at a particular level, so that we can gauge the reliability of our estimate. A standard approach to derive confidence intervals is the bootstrap [17], a computational technique that resamples data points with replacement to simulate the distribution of any test statistic under the null hypothesis that is tested. The bootstrap, particularly useful in complicated nonparametric problems where no asymptotic results can be obtained [18], was adapted by Felsenstein to the nonstandard phylogenetic problem [19]. Indeed, the problem is nonstandard in that the object for which we wish to assess accuracy is not a realvalued parameter, but a graph.
> The basic idea, clearly explained in [20], consists in resampling columns of the alignment, with replacement, to construct a "synthetic" alignment of the same size as the original alignment. This synthetic or bootstrap replicate is then subjected to the same treereconstruction algorithm used on the original data (Fig. 2). This exercise is repeated a large number of times (e.g., <sup>106</sup> ), and the proportion of each original bipartition (internal node) in the set of bootstrapped trees is recorded. In Fig. 2, for instance, the bipartition s1s2|s<sup>3</sup> is found in two bootstrap trees out of three, so the bootstrap support for this node is 66.7%. In this simple case with three sequences, the bootstrap support for topology T<sup>1</sup> is also 66.7%. This bootstrap proportion for topologies (or for trees when branch lengths are taken into account, in a maximum likelihood context, for instance—see below) can be computed very quickly by bootstrapping the sitewise log-likelihood values, instead of the columns of the alignment; this bootstrap is called RELL, for "resampling estimated log-likelihood" [21].
> However, this approach is no longer used or cited extensively since 2008 (source: ISI Thompson). One alternative that has gained momentum is the one based on the approximated likelihood ratio test (aLRT) [22], implemented, for instance, in phyml [23, 24]. Instead of resampling any quantity (sites or sitewise log-likelihood values), the aLRT tests the null hypothesis that an interior branch length is zero. In spite of being slightly conservative in simulations, the approach is extremely fast and hence highly practical [22].
> The meaning of the bootstrap has been a matter of debate for years. As noted before [8] (see also [22]), the bootstrap proportion P can be seen as assessing the correctness of an internal node, and
#### Original sequence alignment
Fig. 2 The (nonparametric) bootstrap. See text for details
failing to do so [25], or 1 <sup>P</sup> can be interpreted as a conservative probability of falsely supporting monophyly [26]. Since bootstrap proportions are either too liberal or too conservative depending on the actual interpretation of P [27], it is difficult to adjust the threshold below which monophyly can be confidently ruled out [28]. Alternatively, an intuitive geometric argument was proposed to explain the conservativeness of bootstrap probabilities [18] and was further developed into the approximately unbiased or AU test, implemented in CONSEL [29]. In spite of these difficulties, the bootstrap is still widely used—and mandatory in all publications featuring a phylogeny—to assess the confidence one can have in the tree estimated from the data under a particular scheme or model (see Subheading 2.9.3 below). Lastly, note that bootstrap support has often been abused [30], as a high value does not necessarily indicate high phylogenetic signal, and can be the result of systematic biases [31] due to the use of the wrong model of evolution, for instance, as detailed below.
*T*1
2.3 Parsimony and LBA Now that we have a means of evaluating the support for the different topologies, we can test some of the conditions under which parsimony estimates the correct tree topology. Ideally, a good method should return the correct answer with a probability of one when the number of sites increases to infinity. This desirable statistical property is called consistency. One serious criticism of parsimony is its sensitivity to long branch attraction, or LBA, even in the presence of an infinite amount of data (infinite alignment length) [31]. In other words, parsimony is not statistically consistent.
> Different types of model misspecification can lead to LBA, and new ones are continually identified. The topology originally used to demonstrate the artifact is represented in Fig. 3, where two long branches are separated by a shorter one. Felsenstein demonstrated that, under a simple evolutionary process, the artifact or LBA tree is reconstructed. Note that parsimony is not the only phylogenetic method affected by LBA, but because it posits a very simple model of evolution [32–34], parsimony is particularly sensitive to the artifact. In spite of this, one particular journal chose to enforce the use of parsimony, stating that authors should estimate their phylogenies by parsimony but also that, if estimated by some other method, they would need to defend their position "on philosophical grounds" [35]; there is of course no valid scientific justification for taking such a step—derided in the "Twittersphere" as "#parsimonygate."
> The LBA artifact has been shown to plague the analysis of numerous data sets, and a number of empirical approaches have been used to detect the artifact [36, 37]. Most recent papers based on multigene analyses (e.g., [38, 39]) now examine carefully the effect of across-site and across-lineage rate variation (in addition to the use of heterogeneous models). For both sites and lineages, the
Fig. 3 The long branch attraction artifact. The true tree topology has two long branches separated by a short one. The tree reconstructed under a simple model of evolution (a) is the artifact or LBA tree on the left. The tree reconstructed under the correct model of evolution (b) is the correct tree, on the right
Fig. 4 Saturation of DNA sequences. As time increases, the observed number of differences between pairs of sequences reaches a plateau, whereas the actual number of substitutions keeps increasing
procedure is the same and consists in successively removing either the sites that evolve the fastest or the taxa that show the longest root-to-tip branch lengths.
Consequently, we would like a tree-reconstruction method that accounts for multiple substitutions. We would also like a method that (1) takes into account less parsimonious as well as most parsimonious state reconstructions (intervals, tests), (2) weights changes differently if they occur on branches of different length (evolutionary time), and (3) weights different kinds of events (transitions, transversions) differently (biological realism). Likelihood methods include such considerations explicitly, as they require modeling the substitution process itself.
2.5 Modeling Molecular Evolution The basic model of DNA substitution (Fig. 5) is defined on the DNA state space, made of the four nucleotides thymine (T), cytosine (C), adenine (A), and guanine (G). Note that T and C are pyrimidines (biochemically, six-membered rings), while A and G are purines (fused five- and six-membered heterocyclic compounds). Depending on these two biochemical categories, two different types of substitutions can happen: transitions within a category, and transversions between categories. Their respective rates are denoted α and β in Fig. 5.
> The process we want to model should describe the substitution process of the different nucleotides of a DNA sequence. Again, we
Fig. 5 Molecular evolution 101. Specification of the basic model of DNA substitution
will make the simplifying assumption that sites evolve under a timehomogeneous Markov process and are iid, as above. We can therefore concentrate on one single site for now (e.g., [41]).
At a particular site, we want to describe the change in nucleotide frequency after a short amount of time dt, during which the nucleotide frequency of A, for instance, after dt will change from fA(t) to fA(t + dt). According to Fig. 5, fA(t + dt) will be equal to what we had at time t, fA(t), minus the quantity of A that "disappeared" by mutation during dt, plus the quantity of A that "appeared" by mutation during dt. Denoting the mutation rate as μ, the quantity of A that "disappeared" by mutation during dt is simply fA(t)μAdt. These mutations away from A generated quantities of T, C, and G, in which we are not interested at the moment since we only want to know what happens to A. There are three different ways to generate A: from either T, C, or G (Fig. 5). Coming from <sup>T</sup>, mutation will generate fT(t)μT!Adt of <sup>A</sup> during dt. Similar expressions exist for C and for G, so that in total, over the three non-A nucleotides, mutation will generate <sup>∑</sup>i6¼<sup>A</sup> fi(t)μi!Adt. Mathematically, we can express these ideas as:
$$f\_A(t+dt) = f\_A(t) - f\_A(t)\mu\_A dt + \sum\_{i \neq A} f\_i(t)\mu\_{iA} dt \tag{1}$$
Equation 1 describes the change of frequency of A during a short time interval dt. Similar equations can be written for T, C, and G, so that we actually have a system of four equations describing the change in nucleotide frequencies over a short time interval dt:
$$\begin{cases} f\_T(t+dt) = f\_T(t) - f\_T(t)\mu\_T dt + \sum\_{i \neq T} f\_i(t)\mu\_{iT} dt \\ f\_C(t+dt) = f\_C(t) - f\_C(t)\mu\_C dt + \sum\_{i \neq C} f\_i(t)\mu\_{iC} dt \\ f\_A(t+dt) = f\_A(t) - f\_A(t)\mu\_A dt + \sum\_{i \neq A} f\_i(t)\mu\_{iA} dt \\ f\_G(t+dt) = f\_G(t) - f\_G(t)\mu\_G dt + \sum\_{i \neq G} f\_i(t)\mu\_{iG} dt \end{cases} (2)$$
which, in matrix notation, can simply be rewritten as:
A Not-So-Long Introduction to Computational Molecular Evolution 79
$$F(t+dt) = F(t) + \mathcal{Q}F(t)dt\tag{3}$$
with an obvious notation for F, while the instantaneous rate matrix Q is
$$\mathcal{Q} = \begin{pmatrix} -\mu\_T & \mu\_{TC} & \mu\_{TA} & \mu\_{TG} \\ \mu\_{CT} & -\mu\_C & \mu\_{CA} & \mu\_{CG} \\ \mu\_{AT} & \mu\_{AC} & -\mu\_A & \mu\_{AG} \\ \mu\_{GT} & \mu\_{GC} & \mu\_{GA} & -\mu\_G \end{pmatrix} \tag{4}$$
In all the following matrices, we will use the same order for nucleotide: T, C, A, and G, which follows the order in which codon tables are usually written. Recall that μij is the mutation rate from nucleotide i to nucleotide j. Note also that the sum of each row is 0.
Let us rearrange the matrix notation from Eq. 3 as:
$$F(t+dt) - F(t) = \mathcal{Q}F(t)dt\tag{5}$$
and take the variation limit when dt ! 0:
$$\frac{dF(t)}{dt} = \mathcal{Q}F(t) \tag{6}$$
which is a first-order differential equation that can be integrated as:
$$F(t) = e^{\mathcal{Q}t} F(0) \tag{7}$$
Very often, this last equation <sup>7</sup> is written as <sup>F</sup>(t) ¼ <sup>P</sup>(t)F(0), where <sup>F</sup>(0) is conveniently taken to be the identity matrix and <sup>P</sup>(t) <sup>¼</sup> {Pi, <sup>j</sup>(t)} ¼ <sup>e</sup> Qt is the matrix of probabilities of going from state i to j during a finite time duration t. Note that the right-hand side of this equation is a matrix exponentiation, which is not the same as the exponential of all the elements (row and columns) of that matrix. The computation of the term e Qt demands that a spectral decomposition of the matrix Q be realized. This means finding a diagonal matrix D of eigenvalues and a matrix M of (right) eigenvectors so that:
$$P(t) = Me^{Dt}M^{-1} \tag{8}$$
The exponential of the diagonal matrix D is simply the exponential of the diagonal terms.
Except in the simplest models of evolution, finding analytical solutions for the eigenvalues and associated eigenvectors can be tedious. As a result, numerical procedures are employed to solve Eq. 8. Alternatively, a Taylor expansion can be used to approximate P(t).
If all entries in Q are positive, any state or nucleotide can be reached from any other in a finite number of steps (all states "communicate") and the base frequencies have a stationary distribution <sup>π</sup> <sup>¼</sup> (πT, <sup>π</sup>C, <sup>π</sup>A, <sup>π</sup>G). This is the steady state reached after an "infinite" amount of time, or long enough for the Markov process to forget its initial state, starting from "random" base frequencies.
Fig. 6 Likelihood computation on a small tree. See text for details
#### 2.6 Computation on a Tree
Now that we know how to determine the rate of change of nucleotide frequencies during a time interval dt, we can compute the probability of a particular nucleotide change on a tree. The simplest case, though somewhat artificial with only two sequences, is depicted in Fig. 6.
We are looking at a particular nucleotide position, denoted j, for two aligned sequences. The observed nucleotides at this position are T in sequence 1, and C in sequence 2. The branch separating T from C has a total length of t<sup>0</sup> + t1. For the sake of convenience, we set an arbitrary root along this path. The likelihood at site j is then given by the probability of going from the fictive root i to T in t0, and from i to C in t1. Any of the four nucleotides can be present at the fictive root. As we do not know which one was there, we sum these probabilities over all possible state, weighted by their prior probabilities, the equilibrium frequencies πi. In all, we have the likelihood ℓ<sup>j</sup> at site j:
$$\ell\_j = \sum\_{i=\left\{T, C, A, G\right\}} \pi\_i P\_{i, T}(t\_0) P\_{i, C}(t\_1) \tag{9}$$
which is equivalent to the Chapman–Kolmogorov equation [42]. As all the sites are assumed to be iid, the likelihood of an alignment is the product of the site likelihoods in Eq. 9. Because all these sitewise probabilities can be small, and that the product of small numbers can become smaller than what a computer can represent in memory (underflow), all computations are done on a logarithmic scale and may include some form of rescaling [43].
Note that this example is somewhat artificial: with only two sequences, we can compute the likelihood directly with πTPT, <sup>C</sup>(t<sup>0</sup> <sup>+</sup> <sup>t</sup>1) <sup>¼</sup> <sup>π</sup>CPC,T(t<sup>0</sup> <sup>+</sup> <sup>t</sup>1); the full summation over unknown states as in Eq. 9 is required with three sequences or more. When analyzing a multiple-sequence alignment of S sequences, there will be many nodes in the tree for which the character state is unknown, which means that the summation required will involve many terms. Specifically, the sum will be over 4S<sup>3</sup> terms. Fortunately, terms can be factored out of the summation, and a dynamic programing algorithm with a complexity of the order of Oð42SÞ, called the pruning algorithm [44], can be used (see [15] for details).
#### 2.7 Substitution Models and Instantaneous Rate Matrices Q
Now that we have almost all the elements to compute the likelihood of a set of parameters, including the tree (i.e., the set of branch lengths and the tree topology; see Subheading 2.10), the only missing element required to compute the likelihood at each site, as in Eq. 9, for instance, is the specification of the instantaneous rate matrix Q as in Eq. 4. Remember that the μi,<sup>j</sup> represent mutation rates from state (nucleotide) i to j. This matrix is generally rewritten as:
$$\mathcal{Q} = \mu \begin{pmatrix} - & r\_{TC} & r\_{TA} & r\_{TG} \\ \\ r\_{CT} & - & r\_{CA} & r\_{GG} \\ r\_{AT} & r\_{AC} & - & r\_{AG} \\ \\ r\_{GT} & r\_{GC} & r\_{GA} & - \end{pmatrix} \tag{10}$$
so that each entry rij is a rate of change from nucleotide i to nucleotide <sup>j</sup>. The diagonal entries are left out, indicated by a "," and are in fact calculated as the negative sum of the off-diagonal entries (as rows sum to 0).
The simplest specification of Q would be that all rates of change are identical, so that Q becomes (leaving out the mutation rate μ and indexing the matrix to indicate the difference):
$$\mathcal{Q}\_{\text{JC}} = \begin{pmatrix} - & 1 & 1 & 1 \\ 1 & - & 1 & 1 \\ & 1 & 1 & - & 1 \\ 1 & 1 & 1 & - \end{pmatrix} \tag{11}$$
which is the model proposed by Jukes and Cantor [45] and often noted "JC" or "JC69". Under the specification of Eq. 11, this model has no free parameter. The process is generally scaled such that the unit of branch lengths can be interpreted as an expected number of substitutions per site.
Of course, this model is extremely simplistic and neglects a fair amount of basic molecular biology. In particular, it overlooks two observations. First, base frequencies are not all equal in actual DNA sequences, but are rather skewed, and second, transitions are more frequent than transversions (see Subheading 2.5).
The way to account for this first "biological realism" is as follows. If DNA sequences were made exclusively of As, for instance, that would mean that all mutations are towards the observed base, in this case A, whose equilibrium or stationary frequency is πA. The same reasoning can be used for arbitrary equilibrium frequencies π, so that all relative rates of change in Q become proportional to the vector of equilibrium frequency π of the target nucleotide. In other words, the instantaneous rate matrix Q becomes:
$$\mathcal{Q}\_{\text{F81}} = \begin{pmatrix} - & \pi\_C & \pi\_A & \pi\_G \\\\ \pi\_T & - & \pi\_A & \pi\_G \\\\ \pi\_T & \pi\_C & - & \pi\_G \\\\ \pi\_T & \pi\_C & \pi\_A & - \end{pmatrix} \tag{12}$$
again with the requirement that rows sum to 0. This matrix represents the Felsenstein or F81 model [44]. This model has four parameters (the four base frequencies), but since base frequencies sum to 1, we only have three free parameters.
The second "biological realism," accounting for the different rates of transversions and transitions, can be described by saying that transitions occur κ times faster than transversions. From Fig. 5, recall that transitions are mutations from T to C (and vice versa) and from A to G (and vice versa). This translates into:
$$\mathcal{Q}\_{\rm K80} = \begin{pmatrix} - & \kappa & 1 & 1 \\ \kappa & - & 1 & 1 \\ 1 & 1 & - & \kappa \\ 1 & 1 & \kappa & - \end{pmatrix} \tag{13}$$
This model is called the Kimura two-parameter model or K80 (or K2P) [46]. The model is alternatively described with the two rates α and β (see Fig. 5). In the "κ version" of the model as in Eq. 13, there is only one free parameter.
Of course it is possible to account for both kinds of "biological realism," unequal equilibrium base frequencies and transition bias, all in the same model, whose generator Q becomes:
$$\mathcal{Q}\_{\rm HKY} = \begin{pmatrix} - & \pi\_C \kappa & \pi\_A & \pi\_G \\\\ \pi\_T \kappa & - & \pi\_A & \pi\_G \\\\ \pi\_T & \pi\_C & - & \pi\_G \kappa \\\\ \pi\_T & \pi\_C & \pi\_A \kappa & - \end{pmatrix} \tag{14}$$
which corresponds to the Hasegawa–Kishino–Yano or HKY (or HKY85) model [47]. This model has four free parameters: κ and three base frequencies.
The level of "sophistication" goes "up to" the general timereversible model [48], denoted GTR or REV, which has for generator:
$$\mathcal{Q}\_{\rm GTR} = \begin{pmatrix} - & a\pi\_C & b\pi\_A & c\pi\_G \\ a\pi\_T & - & d\pi\_A & c\pi\_G \\ b\pi\_T & d\pi\_C & - & \pi\_G \\ c\pi\_T & c\pi\_C & \pi\_A & - \end{pmatrix} \tag{15}$$
The number of free parameters is now eight (three base frequencies plus five nucleotide propensities). The name is derived from the time-reversibility constraint, which implies that the likelihood is independent of the actual orientation of time.
In fact, there exist only a few "named" additional substitution models [15], most of which are time-reversible models, while a total of 203 models can be derived from GTR [49]. We have focused solely on DNA models in this chapter, but the problem is similar with amino acid or codon models, except that the number of parameters increases quickly. We have also limited ourselves to time-reversible time-homogeneous models, but irreversible non-homogeneous models were developed some time ago [50] and are used, for instance, to root phylogenies [51] or to help alleviate the effects of LBA [39].
For a given substitution model, how should parameters be estimated, given the (potentially) high dimensionality of the model? Analytical solutions consist in determining when the first derivative of the likelihood function is equal to zero (with a change of sign in the second derivative). However, finding the root of the likelihood function analytically is only possible in the simple case of three sequences of binary characters under the assumption of the molecular clock (see Subheading 3.1) [12]. As a result, numerical solutions must be found to maximize the likelihood function.
A number of ideas have been combined to search efficiently for the parameter values that maximize the likelihood function. Most programs will start from a random starting point, for example, ðθ ð0Þ <sup>1</sup> , θ ð0Þ <sup>2</sup> <sup>Þ</sup>, denoted by an <sup>x</sup> in Fig. 7, where we limit ourselves to a two-parameter example. The optimization procedure can follow
Fig. 7 Two optimization strategies. The likelihood surface of a function with two parameters θ<sup>1</sup> and θ<sup>2</sup> (e.g., two branch lengths) is depicted as a contour plot, whose highest peak is at the + sign. (a) Optimization one parameter at a time. (b) Optimization of all parameters simultaneously. See text for details
#### 2.8 Some Computational Aspects
2.8.1 Optimization of the Likelihood Function
one of the two strategies. In the first one, parameters are optimized one at a time. In Fig. 7a, parameter θ<sup>1</sup> is first optimized to maximize the likelihood function with a line search, which defines a direction along which the other parameter (θ2) or parameters in the multidimensional case are kept constant. Once θ ð1Þ <sup>1</sup> is found, a new direction is defined to optimize θ2, and so on so forth until convergence to the maximum of the likelihood function. As shown in Fig. 7a, many iterations can be required, in particular when the parameters θ<sup>1</sup> and θ<sup>2</sup> are correlated. The alternative to optimizing one parameter at a time is to optimize all parameters simultaneously. In this case (Fig. 7b), an initial direction is defined at ðθ ð0Þ <sup>1</sup> , θ ð0Þ <sup>2</sup> <sup>Þ</sup> such that the slope at this point is maximized. The process is repeated until convergence. More technical details can be found in [52]. The simultaneous optimization procedure generally requires fewer steps than optimizing parameters one at a time, but not always. Since the computation of the likelihood function is the most expensive computation of these algorithms, the simultaneous optimization is much more efficient, at least in our toy example.
How general is this result? Simultaneously optimizing parameters of the substitution model, while optimizing branch lengths one at a time, was shown to be more effective on large data sets [43], potentially because of the correlation that exists between some of the parameters entering the Q matrix (see Subheading 2.7).
2.8.2 Convergence Convergence is usually reached either when the increment in the log-likelihood score becomes smaller than an ε value, usually set to a small number such as 10<sup>6</sup> (but yet a number larger than the machine ε: the smallest number that a given computer can represent), or when the log-likelihood score has not changed after a predetermined number of iterations. However, none of these stopping rules guarantees that the global maximum of the likelihood function has been found. Therefore, it is generally recommended to run the optimization procedure at least twice, starting from different initial values of the model parameters, and to check that the likelihood score after optimization is the same across the different runs (Fig. 8). If this is not the case, additional runs may be required, and the one with the largest likelihood is chosen for inference (e.g., [53]).
In many instances though, different substitution models will give different tree topologies, and therefore different biological conclusions. One difficulty is therefore to know which model should be used to analyze a particular data set.
2.9 Selection of the Appropriate Substitution Model One important issue in model selection is about the trade-off between bias and variance [55]: a simple model will fail to capture all the sophistication of the actual substitution process, and will
Fig. 8 Likelihood surfaces behaving badly. Schematic of the probability surface of the function p(X|θ) is plotted as a function of θ. Most line search strategies will converge (CV) to the MLE when the initial value is in the "CV" interval, and fail when it is in the "no CV" interval. Adapted with permission from [54]
therefore be highly biased even if all the parameters can be estimated with tight precision (little variance). Alternatively, a highly parameterized model will "spread" the information available from the data over a large number of parameters, hereby making their estimation difficult (flat likelihood surface; see Subheading 2.8.1), with a large variance, in spite of perhaps being a more realistic model with less bias. The objective of most model selection procedure is therefore to find not the best model in terms of likelihood score, but the most appropriate model, the one that strikes the right balance between bias and variance in terms of number of parameters. However, we argue that optimizing for this bias–variance trade-off works only for statistical procedures, be they, for instance, frequentist (LRT, likelihood ratio test) or Bayesian (BF, Bayes factor). On the other hand, information-theoretic criteria (e.g., AIC, Akaike information criterion) aim at selecting the model that is approximately closest to the "true" biological process.
The bias–variance trade-off mainly concerns the comparison of models that are based on the same underlying rationale, for instance, choosing among the 203 models that can be derived from GTR. We may also be interested in comparing models that are based on very different rationales. The likelihood ratio test is suited for assessing the bias–variance trade-off, while Bayesian and information-theoretic approaches, as well as cross-validation (CV), can be used for more general model comparisons. Here we review four approaches to model selection: LRT, BF, AIC, and CV.
2.9.1 The Likelihood Ratio Test The substitution models presented above have one key property: it is possible to reduce the most sophisticated time-reversible named model (GTR+Γ+I) to any simpler model by imposing some constraints on parameters. As a result, the models are said to be nested, and statistical theory (the Neyman–Pearson lemma) tells us that there is an optimal (most powerful) way of comparing two nested models (a simple null vs. a simple alternative hypothesis) based on the likelihood ratio test or LRT.
θ
The test statistic of the LRT is twice the log-likelihood difference between the most sophisticated model (which by definition is always the one with the highest likelihood—if this is not the case, there is a convergence issue; see Subheading 2.8.1) and the simpler model. This test statistic follows asymptotically a χ <sup>2</sup> distribution (under certain regularity conditions), and the degree of freedom of the test is equal to the difference in the number of free parameters between the two models.
The null hypothesis is that the two competing models explain the data equally well. The alternative is that the most sophisticated model explains the data better than the simpler model. If the null hypothesis cannot be rejected at a certain level (type-I error rate), then, based on the argument developed above, the simpler model should be used to analyze the data. Otherwise, if the null hypothesis can be rejected, the more sophisticated model should be used to analyze the data. Note that a test never leads to accepting a null hypothesis; the only outcomes of a test are either reject or fail to reject a null hypothesis.
Intuitively, we can see the null hypothesis H<sup>0</sup> as stating that a certain parameter θ is equal to θ0. The maximum likelihood estimate (MLE) is at θ ^, which is our alternative hypothesis H1, left unspecified. We note the log-likelihood as ln <sup>p</sup>ðXjθÞ ¼ <sup>ℓ</sup>ðθÞ, where <sup>X</sup> represents the data. Under <sup>H</sup>0, we have <sup>θ</sup> ¼ <sup>θ</sup>0, while under <sup>H</sup><sup>1</sup> we have <sup>θ</sup> <sup>¼</sup> <sup>θ</sup> ^. The log-likelihood ratio is therefore ln LR ¼ <sup>ℓ</sup>ð<sup>θ</sup> ^Þ <sup>ℓ</sup>ðθ0Þ. Under the null <sup>H</sup>0, <sup>ℓ</sup>ð<sup>θ</sup> ^Þ ¼ 0 (by definition). The log-likelihood ratio then reduces to ln LR ¼ ℓðθ0Þ. We can then take the Taylor expansion of the log-likelihood function ℓ around θ ^, which gives us <sup>ℓ</sup> <sup>1</sup> 2 ðθ ^ <sup>θ</sup>0Þ 2 d2ℓ <sup>d</sup>θ<sup>2</sup> (recall that <sup>ℓ</sup>ð<sup>θ</sup> ^Þ ¼ 0, so that the first terms of the series "disappear"). Therefore, log-likelihood ratio can be approximated by <sup>1</sup> 2 ðθ ^ <sup>θ</sup>0Þ 2 d2ℓ <sup>d</sup><sup>θ</sup>2. Recall that Fisher's information is negative reciprocal of the second derivative of the likelihood function, so that:
$$\ln LR \approx \frac{1}{2} \frac{\left(\hat{\theta} - \theta\_0\right)^2}{\nu a r(\theta)}\tag{16}$$
which follows asymptotically half a χ <sup>2</sup> distribution. Hence the usual approximation:
$$2\ln LR = 2 \times (\ell\_1 - \ell\_0) \sim \chi^2\_k \tag{17}$$
with k being the difference in the number of free parameters between the two models 0 and 1. The important points in this intuitive outline of the proof are that (1) the two hypotheses need to be nested and (2) taking the Taylor expansion around θ ^ requires that the likelihood function be continuous at that point, which implies that ℓ is differentiable left and right of θ ^. Therefore, testing points at the boundary of the parameter space cannot be done by approximating the distribution of the test statistic of the LRT by a regular χ <sup>2</sup> distribution, as noted many times in molecular evolution [56–64]. A solution still involves the LRT, but the asymptotic distribution becomes a mixture of χ <sup>2</sup> distributions [65].
An approach that has become popular under the widespread adoption of computer programs such as ModelTest [66] and jModelTest [67] is the hierarchical LRT (hLRT). This hierarchy goes from the simplest model (JC) to the set of most complex models (+Γ+I), traversing a tree of models. The issue is that there is more than one way to traverse this tree of models, and that depending on which way is adopted, the procedure may end up selecting different models [68, 69].
Information theory provides us with a number of solutions to circumvent the three limitations of the LRT (nestedness, continuity, and dependency on the order in which models are compared).
The core of the information-based approach is the Kullback–Leibler (KL) distance, or information [70], which measures the distance between an approximating model g and a "true" model f [55]. This distance is computed as:
$$d\_{\rm KL}(f, \mathcal{g}) = \int f(\boldsymbol{\kappa}) \ln \frac{f(\boldsymbol{\kappa})}{\mathcal{g}(\boldsymbol{\kappa}|\boldsymbol{\theta})} \, d\boldsymbol{\kappa} \tag{18}$$
where θ is a vector of parameters entering the approximating model g and x represents the data. Note that this distance is not symmetric, as typically <sup>d</sup>KL( <sup>f</sup>, <sup>g</sup>) 6¼ <sup>d</sup>KL( <sup>g</sup>, <sup>f</sup> ), and that the "true" model f is unknown. The idea is to rewrite dKL( f, g) in a slightly different form, to make it clear that Eq. 18 is actually a difference between two expectations, both taken with respect to the unknown "truth" f:
$$d\_{\rm KL}(f, \mathcal{g}) = E\_f[f(\boldsymbol{\omega}) \, \ln f(\boldsymbol{\omega})] - E\_f[f(\boldsymbol{\omega}) \, \ln \mathcal{g}(\boldsymbol{\omega}|\boldsymbol{\theta})] \tag{19}$$
Equation 19 therefore measures the loss of information incurred by fitting g when the data x actually come from f. As f is unknown, dKL( f, g) cannot be computed as such.
Two points are key to deriving the criterion proposed by Akaike (see [55]). First, we usually want to compare at least two approximating models, g<sup>0</sup> and g1. We can then measure which one is closest to the "true" process f by taking the difference between their respective Kullback–Leibler distances. In the process, the direct reference to the "true" process cancels out. As a result, the "best" model among g<sup>0</sup> and g<sup>1</sup> is the one that is closest to the "true" process f: it is the model that minimizes the distance to f. By setting model parameters to their MLEs, we now deal with estimated distances, but these are still with respect to the unknown f.
Second, in the context of a frequentist approach, we would repeat the experiment of sampling data an infinite number of times.
2.9.2 Information-Theoretic Approaches We would then compute the expected estimated KL distance, so that model selection can be done on the sole estimated log-likelihood value. Akaike, however, showed that this latter approximation is biased, and must be adjusted by a term that is approximately equal to the number of parameters K entering model g (see [55]). For "historical reasons" (similarity with asymptotic theory with the normal distribution), the selection criterion is multiplied by 2 to give the well-known definition of the Akaike information criterion or AIC:
$$\text{AIC} = -2\ln\ell(\hat{\theta}) + 2K \tag{20}$$
Unlike the case of the hLRT, where we were selecting the "most appropriate model" (with respect to the bias–variance trade-off), in the case of AIC we can select the best model. This best model is the one that is closest to the "true" unknown model ( f ), with the smallest relative estimated expected KL distance. The best AIC model therefore minimizes the criterion in Eq. 20.
A small-sample second-order version of AIC exists, where the penalty for extra parameters (2K in Eq. 20) is slightly modified to account for the trade-off between information content in the data and K (see [55]). In our experience, we find it advisable to use this small-sample correction irrespective of the actual size of the data, since this correction vanishes in large and informative samples, but corrects for proper model ranking when K becomes very large compared to the amount of information (e.g., in phylogenomics where models are partitioned with respect to hundreds of genes).
The AIC has been shown to tend to favor parameter-rich models [71–75], which has motivated the use and development of alternative approaches in computational molecular evolution. These include, the Bayesian information criterion [76], and the decision theory or DT approach, which is based on ΔAIC weighted by squared branch length differences [71]. Most of these approaches, including the hLRT, have recently been compared in a simulation study that suggests, in agreement with empirical studies [72, 77], that both BIC and DT have the highest accuracy and precision [75].
One particular drawback of these information-theoretic approaches is that they require that every single model of evolution, or at least the most "popular" models (the few named ones), be evaluated. This step can be time-consuming, especially if a full maximum likelihood optimization is performed under each model. A first set of heuristics consists in fixing the tree topology to a tree estimated with a quick distance-based method such as BioNJ [78], and then estimating just the branch lengths and the parameters of the substitution model, as implemented in jModelTest [67]. As the optimizations are independent of each other under each substitution model, these computations are typically forked to multiple cores or processors [79]. Further heuristics exist to avoid all these independent optimizations [79], as implemented in SMS (Smart Model Selection in PhyML), which is reported to be cutting runtimes in half without forfeiting accuracy [80].
Note finally that all these approaches are not limited to selecting the most appropriate or the best model of evolution. Disregarding the hLRT, which requires that models be nested (to be able to use the χ <sup>2</sup> approximation; otherwise, see [65]), AIC, BIC, etc. allow us to compare non-nested models and, in particular, phylogenetic trees (branch lengths plus topology).
2.9.3 The Bayesian Approach The Bayesian framework has permitted the development of two main approaches, which are actually two sides of the same coin: one based on finding the model that is the most probable a posteriori, and one based on ranking models and estimating a quantity called the Bayes factor.
> In a nutshell, the frequentist approaches developed in the previous sections are based on the likelihood, which is the probability of the data, given the parameters: p(X|θ). However, this approach may not be the most intuitive, since most practitioners are not interested in knowing the conditional probability of their data, as the data were collected to learn more about the processes that generated them. It can therefore be argued that the Bayesian approach, which considers the probability of the parameters given the data or p(θ|X), is more intuitive than the frequentist approach. Unlike likelihood, which relies on the function p(X|θ) and permits point estimation, Bayesian inference is based on the posterior distribution p(θ|X). This distribution is often summarized by a centrality measure such as its mode, mean, or median. Measures of uncertainty are based on credibility intervals, the Bayesian equivalent of confidence intervals. Typically, credibility intervals are taken at the 95% cutoff and are called highest posterior densities (HPDs).
> The connection between posterior probability and likelihood is made with Bayes' inversion formula, also called Bayes' theorem, by means of a quantity called the prior distribution p(θ):
$$\varrho(\theta|X) = \frac{\varrho(X|\theta)\ p(\theta)}{\varrho(X)}\tag{21}$$
The prior represents what we think about the process that generated the data, before analyzing the data, and is at the origin of all controversies surrounding Bayesian inference. In practice, priors are more typically chosen based on statistical convenience, and often have nothing to do with our genuine state of knowledge about parameters before observing the available data. We will see in Subheading 3.1 that priors can be used to distinguish between parameters that are confounded in a maximum likelihood analysis (model), so that we argue that the frequentist vs. Bayesian controversy is sterile, and we advocate a more pragmatic approach, that often results in the mixing of both approaches (in their concepts and techniques) [81, 82].
All models have parameters. Subheading 2.7 treats substitution models, which can have eight free parameters in the case of GTR + Γ. Most people are not really interested in these parameters θ or in their estimates θ ^, but have to use them in order to estimate a phylogenetic tree τ. These parameters θ are called nuisance parameters because they enter the model but are not the focus of inference. The likelihood solution consists in setting these parameters to their MLE, ignoring the uncertainty with which they can be estimated, while the Bayesian approach will integrate them out, directly accounting for their uncertainty:
$$\not{p}(X|\mathfrak{r}) = \int\_{\theta} \not{p}(X|\mathfrak{r}, \theta) \not{p}(\theta) \, d\theta \,\tag{22}$$
One difficulty in Bayesian inference is about the denominator in Eq. 21, as this denominator often has no analytical solution. In spite of being a normalizing constant, p(X) requires integrating out nuisance parameters by means of prior distributions as in Eq. 22. Thus, it is easy to see from Eq. 21 that the posterior distribution of the variable of interest (e.g., τ) can quickly become complicated:
$$\not{p}(\tau|X) = \int\_{\theta} \frac{\not{p}(X|\tau,\theta)\not{p}(\tau)\not{p}(\theta)}{\sum\_{T} \not{p}(X|\tau,\theta)\not{p}(\tau)\not{p}(\theta)}\ d\theta \tag{23}$$
where τ and θ are assumed to be independent and the discrete sum is taken over the set T of all possible topologies (see Subheading 2.10.1). However, the ratio of posteriors evaluated at two different points will simplify: as the denominator in Eq. 23 is a constant, it will cancel out from the ratio. This simple observation is at the origin of an integration technique for approximating the posterior distribution in Eq. 23: Markov chain Monte Carlo (MCMC) samplers. A very clear introduction can be found in [83].
Building on this, two approaches can be formulated to compare models in a Bayesian framework. The first is to treat the model as a "random variable," and compute its posterior probability. The best model is then the one that has the highest posterior probability. This approach is typically implemented in a reversible-jump MCMC (or rjMCMC) sampler (e.g., see [49]).
The alternative is to use the Bayesian equivalent of the LRT, the Bayes factor. Rather than comparing two likelihoods, the Bayes factor compares the probability of the data under two models, M<sup>0</sup> and M1:
$$\text{BF}\_{0,1} = \frac{p(X|M\_0)}{p(X|M\_1)}\tag{24}$$
More specifically, BF0,1 evaluates the weight of evidence in favor of model M<sup>0</sup> against model M1, with BF0,1 > 1 considered as evidence in favor of M0. Just as in a frequentist context, where a null hypothesis is significantly rejected at a certain threshold, 5%, 1%, or less depending on different costs or error types, Bayes factors can be evaluated on a specific scale [84]. However, because this scale is just as ad hoc as in a frequentist setting, it might be preferable to use the probability of the data under a particular model p(X|Mi) as a means of ranking models Mi.
The quantity p(X|M0), which is the denominator in Eq. 23 (where we did not include the dependence on the model in the notation), is called the marginal likelihood. Note that it is also an expectation with respect to a prior probability distribution:
$$\not{p}(X|M\_0) = \int\_{\theta} \not{p}(X|\theta, M\_0) \not{p}(\theta|M\_0) \, d\theta \,\tag{25}$$
A number of approximations to evaluate Eq. 25 exist and are reviewed in [85] (see also [86, 87]). The simplest one is based on the harmonic mean of the likelihood sampled from the posterior distribution [88], also known as the harmonic mean estimator (HME). The way this estimator is derived demands to understand how integrals can be approximated. Briefly, to compute <sup>I</sup> ¼ <sup>Ð</sup> <sup>g</sup>ðθ<sup>Þ</sup> <sup>p</sup>ðθ<sup>Þ</sup> <sup>d</sup>θ, generate a sample from a distribution <sup>p</sup>⋆(θ) and calculate the simulation-consistent estimator <sup>I</sup> <sup>¼</sup> <sup>P</sup>wi <sup>g</sup>ðθÞ<sup>=</sup> Pwi where wi is the importance function p(θ)/ <sup>p</sup>⋆(θ). Take <sup>g</sup> ¼ <sup>p</sup>(X|θ) and <sup>p</sup>⋆(θ) ¼ <sup>p</sup>(X|θ) <sup>p</sup>(θ)/p(X), then I ^ ¼ <sup>p</sup> ^ðXjM0Þ ¼ lim<sup>N</sup>!1ð <sup>1</sup> N P <sup>1</sup> <sup>p</sup>ðXjθiÞ Þ 1 with <sup>θ</sup> <sup>p</sup>(θ|X) (see supplementary information in [89]). As a result, a very simple way to estimate the marginal likelihood and Bayes factors is to take the output of an MCMC sampler and compute the harmonic mean of the likelihood values (not the log-likelihood values) sampled from the posterior distribution.
Because of its simplicity, this estimator is now implemented in most popular programs such as MrBayes [90] or BEAST [91]. However, it might be considered as the worst estimator possible, because its results are unstable [88, 92] and biased towards the selection of parameter-rich models [86]. An alternative and reliable estimator, based on thermodynamic integration (TI; [86]—also known as path sampling; [93, 94]), is much more demanding in terms of computation. Indeed, it requires running MCMC samplers morphing one model into the other (and vice versa), which can increase computation time by up to an order of magnitude [86]. Improvements of the TI estimator are however available. The stepping-stone (SS) approach builds on importance sampling and TI to speed up the computation while maintaining the accuracy of the standard TI estimator [87, 95].
Moving away from the estimation of marginal likelihoods, an analogue of AIC that can be obtained through the output of an MCMC sampler (AICM) was proposed [96]. In essence, it relies on the asymptotic convergence of the posterior distribution of the log-likelihood on a gamma distribution [97]. As such, it becomes possible to estimate the effective number of parameters as twice the sample variance of posterior distribution of the log-likelihood, which itself can be estimated by a resampling procedure [96]. This gives a very elegant means of estimating AIC, from the posterior simulations. However, although AICM seems to be a more stable measure of model ranking than HME, both TI and SS still seem to outperform this estimator, at least in the case of the comparison of demographic and relaxed molecular clock models [96] (see Subheading 3).
2.9.4 Cross-Validation Cross-validation is another model selection approach, which is extremely versatile in that it can be used to compare any set of models of interest. Besides, the approach is very intuitive. In its simplest form, cross-validation consists in dividing the available data into two sets, one used for "training" and the other one used for "validating." In the training step (TS), the model of interest is fitted to the training data in order to obtain a set of MLEs. These MLEs are then used to compute the likelihood using the validation data (validation step, VS). Because the validation data were not part of the training data, the likelihood values computed during VS can be directly used to compare models, without requiring any explicit correction for model dimensionality.
The robustness of the cross-validation scores can be explored in various ways, such as repeating the above procedure with a switched labeling of training and validation data (hence the expression crossvalidation). Of course, this simple 2-fold cross-validation could be extended to n-fold cross-validation, where the data are subdivided into <sup>n</sup> subsets, with <sup>n</sup> 1 subsets serving for training, and one for validation. Ideally, the procedure is repeated <sup>n</sup> 1 additional times.
We know of only two examples of its use in phylogenetics, one in the ML framework [98] and one with a Bayesian approach [99]. Given the increasing size of modern data sets, putting aside some of the data for validation is probably not going to dramatically affect the information content of the whole data set. As a result, model selection via cross-validation, which is statistically sound, could become a very popular approach.
#### 2.10 Finding the Best Tree Topology
2.10.1 Counting Trees
Now that we can select a model of evolution (Subheading 2.9) and estimate model parameters (Subheading 2.8) under a particular model (Subheading 2.5), how do we find the optimal tree? The basic example in Subheading 2.1 suggested that we score all possible tree topologies and choose for inference the one that has the highest score. However, a simple counting exercise shows that an exhaustive examination of all possible topologies is not realistic.
Figure 9 shows how to count tree topologies. Starting from the simplest possible unrooted tree, with three taxa, there are three positions where a fourth branch (leading to a fourth taxon) can be added. As a result, there are three possible topologies with four taxa. For each of these, there are four places on the tree where a fifth branch can be added, which leads to a total of 3 <sup>5</sup> ¼ 15 topologies with five taxa. A recursion appears immediately, and it can be shown that the total number of unrooted topologies with n taxa is equal to 1 <sup>3</sup> <sup>2</sup><sup>n</sup> 5 [100] (see [15] for the deeper history), which, as given in [101], is equal to:
$$N\_{\text{unrooted}}^{T(n)} = \frac{(2n-5)!}{2^{n-3}(n-3)!} = \frac{2^{n-2}\Gamma(n-\frac{3}{2})}{\sqrt{\pi}}\tag{26}$$
where the Γ function for any real number x is defined as <sup>Γ</sup>ðxÞ ¼ <sup>Ð</sup> <sup>1</sup> <sup>0</sup> t <sup>x</sup><sup>1</sup> e<sup>t</sup> dt. An approximation based on Stirling number is also given in [101].
The same exercise can be done for rooted trees (Fig. 10), where the number of possible rooted topologies with n taxa becomes <sup>1</sup> <sup>3</sup> <sup>2</sup><sup>n</sup> 3, which is
Fig. 9 Procedure to count the number of unrooted topologies. The top line shows the current number of taxa included in the tree below. Gray arrows indicate locations where an additional branch can be grafted to add one taxon. Black arrows show the resulting number of topologies after addition of a branch (taxon). Only one such possible topology is represented at the next step. The bottom line indicates the number of possibilities. These numbers multiply to obtain the total number of trees
Number of tree topologies are given for the unrooted and rooted cases
Fig. 10 Procedure to count the number of rooted topologies. See Fig. 9 for legend and text for details
$$N\_{\text{root}}^{T(n)} = \frac{(2n - 3)!}{2^{n - 2}(n - 2)!} = \frac{2^{n - 1} \Gamma\left(n - \frac{1}{2}\right)}{\sqrt{\pi}} \tag{27}$$
Note that N <sup>T</sup> <sup>ð</sup>n<sup>Þ</sup> unrooted <sup>¼</sup> <sup>N</sup> <sup>T</sup> <sup>ð</sup>n1<sup>Þ</sup> rooted , as Table 2 clearly suggests.
As a result, the number of possible topologies quickly becomes very large when the number n of sequences increases, even with a very modest n, so that heuristics become necessary to find the bestscoring tree.
2.10.2 Some Heuristics to Find the Best Tree
The simplest approach builds upon the idea presented in Figs. 9 and 10. Stepwise addition, for instance, starts with three sequences drawn at random among the n sequences to be analyzed, and adds sequences one at a time, keeping only the tree that has the highest score at each step (e.g., [52]). However, there is no guarantee that the final tree is the optimal tree [44]. The idea behind branch-andbound [102], refined in [103], is to have a look-ahead routine that prevents entrapment in suboptimal trees. This routine sets a bound on the trees selected at each round of additions, such that only the trees that have a score at least as good as that of the trees obtained in the next round are kept in the search algorithm. Solutions found by the branch-and-bound algorithm are optimal, but computing time becomes quickly prohibitive with more than 20 sequences.
As a result, most tree-search algorithms will start with a quickly obtained tree, often reconstructed with an algorithm based on pairwise distances such as neighbor-joining [104] or a related approach [78, 105], and then alter the tree randomly until no further improvement is obtained or after a certain number of unsuccessful attempts are reached. Examples of such algorithms include nearest neighbor interchange (NNI), subtree pruning and regrafting (SPR), or tree bisection and reconnection (TBR), see, for instance, [6] for a full description. While the details are of little importance here, the critical point is the extent of topological rearrangement in each case. With, e.g., NNI, each rearrangement can give rise to two topologies. The result is that exploring the topology space is slow, especially in problems with large n. On the other hand, TBR has, among the three methods cited above, the largest number of neighbors. As a result, the topology space is explored quickly, but the optimal tree can be "missed" simply because a dramatic change is attempted, so that the computational cost increases. Alternatively, the chance of finding the optimal tree ^τ when ^τ is very different from the current tree is higher when the algorithm can create some dramatic rearrangements. Some programs, such as PhyML ver. 3.0, now use a combination of NNI and SPR to address this issue [24]. MCMC samplers that search the tree space implement somewhat similar tree-perturbation algorithms that are either "global," and modify the topology dramatically, or "local" [106] (see also [107] for a correction of the original local moves). As a result, MCMC samplers are affected by the same issues as traditional likelihood methods. Much of the difficulty therefore comes from this kind of trade-off between larger rearrangements that are expected to improve accuracy and the computational burden associated with these extra computations [108].
2.10.3 Cutting Corners with ABC and AI As some of the above computations can become quite costly (high runtimes, heavy memory footprints, poor scalability with large data sets, etc.), computational workarounds have been and are being explored. One of these resorts to approximate Bayes computing (ABC), which is essentially a likelihood-free approach. First developed in the context of population genetics [109, 110], the driving idea is to bypass the optimization procedures and replace them with simulations in the context of a rejection sampler. In population genetics, the problem could be about a gene tree, which is usually appropriately described by a coalescence tree [111, 112], for which we want to estimate some model parameters. As we are able to simulate trees from such a process, it is possible to place prior distributions on these model parameters, and simulate trees by drawing parameters until the simulated trees "look like" the observed tree. The set of parameters thus drawn approximates the posterior distribution of the corresponding variables. This forms the basis of a naı¨ve rejection sampler, that is quite flexible as it does not even require that a probabilistic model be formulated, but one that can be inefficient, especially if the posterior distribution is far from the prior distribution—which is usually the case. As a result, a number of variations have been described, trying either to correlate sample draws as in MCMC samplers [113] or to resample sequentially from the past [114, 115]. In spite of recent reviews of the computational promises and deliveries of ABC samplers [116–118], the few applications in molecular evolution have been, to date, mostly limited to molecular epidemiology [119–122]. One of the major challenges to estimate a phylogenetic tree from a sequence alignment with ABC is the lack of a proper and efficient simulation strategy: it is possible to simulate trees under various processes (we saw the coalescent above), it is also possible to simulate an alignment from a given (possibly simulated tree), so that in theory one could imagine an ABC algorithm that would use this backward process to estimate phylogenetic trees by comparing a simulated alignment to an "actual" alignment. This, however, would most likely be a very inefficient sampler.
A second area that holds promises is the use of artificial intelligence (AI), and more specifically of machine learning (ML), in molecular evolution. Here again, attempts have been made to using standard ML approaches such as support vector machines [123] to guide the comparison of tree shapes, for instance, [124], which can then be used in epidemiology [121], but estimating a phylogenetic tree has proved more challenging. In one notable exception, an alignment-free distance-based tree-reconstruction method was proposed [125], but its main legacy seems to be in the development of k-mers, or unaligned sequences chopped into words of length k, to reconstruct phylogenetic trees—in particular in the context of phylogenomics (phylogenetics at a genomics scale) [126, 127]. To the best of our knowledge, nobody has ever tried, yet, to train a neural network or even a deep learning algorithm [128–130] on a database of phylogenetic trees with corresponding alignments such as TreeBASE [131] or PANDIT [132]. As applications of deep learning start emerging in genomics [133] and proteomics [134], it is likely that phylogenetics will come next.
#### 3 Uncovering Processes and Times
Fig. 11 The strict molecular clock. The tree is ultrametric. The node marked with a star indicates the presence of a fossil, dated in this example to 10 million years ago (MYA). This is the point that we will use to calibrate the clock, that is, to estimate the global rate of evolution. The number of substitutions that accumulated from the marked node to the tips (present) is indicated on the right weights in at 0.1 substitutions/site. The node that is the most recent common ancestor of S2 and S5 is the node of interest. The number of substitutions from this node to the tips is 0.02 substitutions/site
sub/site/my) and (2) date all the other nodes on the tree. For instance, the most recent common ancestor of S2 and S5 is separated from the tips by a branch length of 0.02 sub/site. Its divergence time is therefore 0.02/0.01 ¼ 2 MYA.
As with any hypothesis, the strict clock can be tested. Tests based on relative rates assess whether two species evolve at the same rate as a third one, used as an outgroup. Originally formulated in a distance-based context [143], likelihood versions have been described [44, 144]. However, because of their low power [145] their use is on the wane. The most powerful test is again the LRT (see Subheading 2.9.1). The test proceeds as usual, first calculating the test statistic 2Δℓ (twice the difference of log-likelihood values). The null hypothesis (strict clock) is nested within the alternative hypothesis (clock not enforced), so that 2Δℓ follows a χ <sup>2</sup> distribution. The degree of freedom is calculated following Fig. 12. With an alignment of <sup>n</sup> sequences, we can estimate <sup>n</sup> 1 divergence times under the null model (disregarding parameters of the substitution model) and we have 2<sup>n</sup> 3 branch lengths under the alternative model. The difference in number of free parameters is therefore <sup>n</sup> 2, which is our degree of freedom. This version of the test actually assesses whether all tips are at the same distance from the root of the tree [44]. For time-stamped data, serially sampled in time as in the case of viruses, the alternative model incorporates information on tip dates [146].
This linear regression model suggested by the molecular clock hypothesis has often been portrayed as a recipe [147], which gave rise in the late twentieth to early twenty-first century to a veritable cottage industry [148–151], culminating with a paper suggesting that the age of the tree of life might be older than the age of planet Earth [152]. This recipe was put down by two factors: (1) the
Fig. 12 Testing the strict molecular clock. The divergence times that can be estimated under the strict clock assumption are denoted ti . The branch lengths that can be estimated without the clock are denoted bi . In the case depicted, with <sup>n</sup> ¼ 7 sequences, we have <sup>n</sup> <sup>1</sup> ¼ 6 divergence times and 2<sup>n</sup> <sup>3</sup> ¼ 11 branch lengths
publication of a piece written in a rather unusual style for a scientific paper [153], and (2) new methodological developments. The main points made in [153] are that (1) most of the early dating studies relied on one analysis [149] that used a fossil-based calibration point for the divergence of birds at 310 MYA to estimate a number of molecular dates for vertebrates, and that (2) these molecular dates were then used in subsequent studies as a proxy for calibration points, disregarding their uncertainty. As a result, estimation errors were passed on and amplified from study to study, leading to the nonsensical results in [152].
Clocks Sanderson [159] (see also [160]), who considered that rates of evolution can change from branch to branch on a tree. By constraining rates of evolution to vary in an autocorrelated manner on a tree, it is possible to devise a method that minimizes the amount of rate change.
The idea of an autocorrelated process governing the evolution of the rates of evolution is attributed to [161] in [159], but could all the same be attributed to Darwin. Thorne et al. developed this idea further in a Bayesian framework [162]. Building upon the basic theory covered in Subheading 2.9.3, the idea is to place prior distributions on the quantities in the right-hand side of Eq. 28. The target distribution is p(t|X). It is proportional to p(X|t) p(t) according to Bayes' theorem, but all that we can estimate is
$$\not{p}(b|X) = \frac{\not{p}(X|b)\not{p}(b)}{\not{p}(X)} = \frac{\not{p}(X|r,t)\not{p}(r,t)}{\not{p}(X)}\tag{29}$$
One of the possible ways to expand the joint distribution of rates and times is p(r, t) is p(r|t) p(t), which posits a process where rate change depends on the length of time separating two divergences. The "art" is now in choosing prior distributions, conditional on the obvious constraint that rates and times should take positive values. A number of such prior distributions for rates have been proposed and assessed [163] and one of the best-performing model for rates is, in our experience, the log-normal model [162, 164]. The prior on times is either a pure-birth (Yule) model or a birth-and-death process possibly incorporating species sampling effects [165]. If sequences are sampled at the population level, a coalescent process is more appropriate (see [112] for an introduction). In this case, the past demography of the sampled sequences can be traced back taking inspiration from spline regression techniques [166, 167] or multiple change-point models [168].
Once these priors are specified, an MCMC sampler will draw from the target distribution in Eq. 29, and marginal distributions for times and rates can easily be obtained. The rationale behind the sampler is represented in Fig. 13. As per Eq. 28, the relationship
Fig. 13 The relaxed molecular clock. See text for details
between rates and times is the branch of a hyperbolic curve, where the priors on rates and on times define a region of higher posterior probability, symbolized here by a contour plot superimposed on the hyperbolic curve. On top of this, fossil information is incorporated into the analysis as constraints on times. A very influential piece stimulated a discussion about the shape of these prior distributions [153], which was taken up [169], and further developed in [170]. Briefly, fossil information is usually imprecise, as paleontologists can only provide minimum and maximum ages (Fig. 13). Of these two ages, the minimum age is often the most reliable. Under the assumption that the placement of the fossil on the tree is correct, the idea is to place on fossil dates a prior distribution that will be highly skewed towards older (maximum) ages. A "hard bound" can be placed on the minimum age, possibly by shifting this prior distribution by an offset equal to the minimum age, while the tails of the prior distribution will act as "soft bounds," because they do not impose on the tree a strict (or hard) constraint. Empirical studies agree, however, that both reliability and precision of fossil calibrations are critical to estimating divergence times [136, 171].
3.1.4 Uncorrelated Relaxed Clocks Because of the autocorrelation between the rate of each branch and that of its ancestral branch (except for the root, which obviously requires a special treatment), the tree topology is fixed under the autocorrelated models described above. By relaxing this assumption about rate autocorrelation, [172] were able to implement a model that also integrates over topological uncertainty. In spite of the somewhat counter-intuitive nature of the relaxation of the autocorrelated process, as implemented in BEAST [91, 173], empirical studies have found this approach to be one of the bestperforming (e.g., [157]).
> When first published, it was proposed that making use of an uncorrelated relaxed molecular clock could improve phylogenetic inference [172]. The idea was that calibration points and their placement on the tree could act as additional information. However, a simulation study suggests that relaxed molecular clocks might not improve phylogenetic accuracy [174], a result that might be due to the lack of calibration constraints in this particular simulation study.
3.1.5 Some Applications of Relaxed Clock Models Since the advent of relaxed molecular clocks, two very exciting developments have seen the light of day. The first concerns the inclusion of spatial statistics into dating models [175, 176]. Spatial statistics are not new in population genetics [177] and have been used with success in combination with analyses in computational molecular evolution (e.g., [178]). However, the originality in [176], for instance, is to combine in a single statistical framework molecular data with geographical and environmental information to infer the diffusion of sequences through both space and time. While these preliminary models seem to deal appropriately with natural barriers to gene flow such as coastlines, a more detailed set of constraints on gene flow may further enhance their current predictive power.
The second development coming from relaxed molecular clocks concerns the mapping of ancestral characters onto uncertain phylogenies. This is not a novel topic, as a Bayesian approach was first described in 2004 [179, 180]. The novelty is that we now have the tools to correlate morphological and molecular evolution in terms of their absolute rates and to allow both molecular and morphological rates of evolution to vary in time [181]. Further development will certainly integrate over topological uncertainty. While there has been a heated controversy about the existence of such a correlation in the past [182], all previous studies were using branch length as a proxy for rate of molecular evolution, which is clearly incorrect. We can therefore expect some more accurate results on this topic very soon. More details and examples can be found in recent and extensive reviews [183–185] that further discuss applications to biogeographic studies [186], or extensions to viral [187, 188], as well as other types of genomic [189] and morphological [190] data.
#### 4 Molecular Population Phylogenomics
Population genetics is rich in theory regarding the relative roles of mutation, drift, and selection. Much research in population genomics is now focusing on using this theory to develop statistical procedures to infer past processes based on population-level data, such as those of the 1000-genome project [191], the UK's 10,000 genome project [192], and always more ambitious projects [193]. One limitation of these inference procedures is that they all focus on a thin slice of evolutionary time by studying evolution at the level of populations. If we wish to study longer evolutionary time scales, for example, tens or hundreds of millions of years, we must resort to interspecific data. In such a context, which is becoming intrinsically phylogenetic, the most important event is a substitution, that is, a mutation that has been fixed. Yet substitution rates can be defined from several features. In particular, from a population genetics perspective, it is of interest to model both mutational features and selective effects, combining them multiplicatively to specify substitution rates. We review briefly how substitution models that invoke codons as the state space lend themselves naturally to these objectives in a first section below (Subheading 4.1), before explaining the origin (and a shortcoming) of all the approaches developed so far (Subheading 4.2).
#### 4.1 Bridging the Gap Between Population Genetics and Phylogenetics
Assuming a point-mutation process, such that events only change one nucleotide of a codon during a small time interval, Muse and Gaut proposed a codon substitution model with rates specified from the QGTR nucleotide-level matrix (see Subheading 2.7), along with one parameter that modulates synonymous events and another one that modulates nonsynonymous events [194]. In most subsequent formulations, the parameter associated with synonymous events is assumed to be fixed, such that the model only modulates nonsynonymous rates by means of a parameter denoted ω. This parameter has traditionally been interpreted as the nonsynonymous to synonymous rate ratio, and is generally associated with a different formulation of the codon model proposed by Goldman and Yang [195]. More details on codon models can be found in Chapter 4.1 [196]. There continues to be a debate regarding the interpretation of the ω parameter [197, 198]. Regardless of how this issue is settled, it is clear that ω is aimed at capturing the net overall effects of selection, irrespective of the exact nature of these effects.
With the intention to model selective effects themselves, Halpern and Bruno proposed a codon substitution model that combines a nucleotide-level layer, as described above, for controlling mutational features, along with a fixation factor that is proportional to the fixation probability of the mutational event [199]. The fixation factor is in turn specified from an account of amino acid or codon preferences. One objective of the model, then, consists in teasing apart mutation and selection. While [199] proposed their model with site-specific fixation factors, later work has explored simpler specifications, where all sites have the same fixation factor [200]. Other models that aimed at capturing across-site heterogeneities in fixation factors were proposed using nonparametric devices and empirical mixtures [201]. Another core idea behind these approaches is to construct a more appropriate null model against which to test for features of the evolutionary process. This idea has been put into practice for the detection of adaptive evolution in protein-coding genes [202, 203]. Recent developments include sequence-wide fixation factors [9, 197, 204, 205], and we predict that these models will play a role in bridging the gap between molecular evolution at the population and at the species levels.
4.2 Origin of Mutation–Selection Models: The Genic Selection Model
In order to understand a shortcoming of these models, we need to go back to the development of fixation probabilities that took place in the second half of the twentieth century. The basic unit or quantum of evolution is a change in allele frequency p. Allele frequencies can be affected by four processes: migration, mutation, selection, and drift. Because of the symmetry between migration and mutation [206], which only differ in their magnitude, these two processes can be treated as one. We are left with three forces: 4.2.1 Fixation Probabilities
mutation, selection, and drift. The question is then, what is the fate of an allele under the combined action of these processes? Our development here follows [207] (but see [208] for a very clear account).
Of the three processes affecting allele frequencies, mutation and selection can be seen as directional forces in that their action will shift the distribution of allele frequencies towards a particular point, be it an internal equilibrium, or fixation/loss of an allele. On the other hand, drift is a non-directional process that will increase the variance in allele frequencies across populations, and will therefore spread out the distribution of allele frequencies. This distribution is denoted Ψ( p, t). We also must assume that the magnitude of all three processes, mutation, selection, and drift, is small and of the order of <sup>1</sup> 2N <sup>e</sup> , where Ne is the effective population size. To derive the fate of an allele after a certain number of generations, we also need to define g( p, ε;dt), the probability that allele frequency changes from p to p + ε during a time interval dt.
In phylogenetics (and population genetics) we are generally interested in predicting the past. The tool making this possible is called the Kolmogorov backward equation, which predicts the frequency of an allele at some time t, given its frequency p<sup>0</sup> at time t0:
$$\Psi(p, t + dt | p\_0) = \int \Psi(p, t | p\_0 + \varepsilon) \, g(p\_0, \varepsilon; dt) \, d\varepsilon \tag{30}$$
We can take the Taylor expansion of Eq. 30 around p0, neglect all terms whose order is larger than two (oðp<sup>2</sup> <sup>0</sup>Þ) and since <sup>Ψ</sup> is not a function of ε, we obtain:
$$\Psi(\not p, t + dt \vert p\_0) = \Psi \int \not p \; d\varepsilon + \frac{\partial \Psi}{\partial \not p\_0} \int \not g \, d\varepsilon + \frac{\partial^2 \Psi}{\partial \not p\_0^2} \int \frac{\epsilon^2}{2} \mathcal{g} \, d\varepsilon \tag{31}$$
This formulation leads to the definition of two terms that represent the directional processes affecting allele frequencies (M) and the non-directional process, or drift (V ):
$$\begin{cases} M(\not p) \not dt &= \\ V(\not p) \not dt &= \end{cases} \begin{cases} \not \varrho \not e \not d \varepsilon \\ \not \varrho \not e^2 d \varepsilon \end{cases} \tag{32}$$
that we can substitute into Eq. 31. At equilibrium, <sup>∂</sup><sup>Ψ</sup> <sup>∂</sup><sup>t</sup> <sup>¼</sup> 0 and, after a bit of calculus, we obtain:
$$\frac{\partial \hat{\Psi}}{\partial p\_0} = C \, e^{-\int \frac{2M}{V} dp} \, dp \tag{33}$$
Table 3 The standard selection models
Models are represented for one locus with two alleles, A<sup>1</sup> and A2, which define three genotypes A1A1, A1A2, and A2A<sup>2</sup> of fitness w1, w2, and w3. The selection coefficient is s (positive in this table, but not necessarily so) and the dominance is governed by h (h ∈ [0, 1])
for which we need to specify boundary conditions and a model of selection. The boundary conditions are the two absorbing states of the system: (1) once fixed, an allele remains fixed (Ψ(1, 1; 1) ¼ 1) and (2) once lost, an allele remains lost (Ψ(1, 1; 0) ¼ 0). With these two requirements, the probability that the allele frequency is 1 given that it was p<sup>0</sup> in the distant past is the fixation probability:
$$\Psi(1,\infty;p\_0) = \frac{\int\_0^{p\_0} e^{-\int \frac{2M}{V} \, dp} \, dp}{\int\_0^1 e^{-\int \frac{2M}{V} \, dp} \, dp} \tag{34}$$
We therefore only need to compute M and V under a particular model of selection to fully specify the fixation probability of an allele in a mutation–selection-drift system. All that is required now to go further is a selection model.
4.2.2 The Case of Genic Selection
We are now ready to derive an explicit form to <sup>Ψ</sup>(1, 1; <sup>p</sup>0) in Eq. 34 in the case of the genic selection model (Table 3; [209]). We obtain:
$$
\overline{w} = 1 + \mathfrak{s}p^2 + 2pqhs = 1 + 2phs + \mathfrak{s}p^2(1 - 2h) \tag{35}
$$
which can be approximated by 1 + 2phs (the result is exact only when <sup>h</sup> <sup>¼</sup> 1/2). Therefore, dw=dp ¼ <sup>2</sup>hs, and we can calculate the M and V terms to obtain the popular result:
$$\Psi(1,\infty;p\_0) = \frac{\int\_0^{p\_0} e^{-\int \frac{2M}{V} dp} dp}{\int\_0^1 e^{-\int \frac{2M}{V} dp} dp} = \frac{e^{-4N\_el\nu p\_0} - 1}{e^{-4N\_el\nu} - 1} \tag{36}$$
Now, the initial frequency of a mutation in a diploid population of (census) size <sup>N</sup> is <sup>p</sup><sup>0</sup> <sup>¼</sup> 1/2<sup>N</sup> (following [208]; [207] considered that <sup>p</sup><sup>0</sup> <sup>¼</sup> 1/2Ne; this debate is beyond the scope of this chapter), which leads to:
$$\Psi\left(1,\infty;\frac{1}{2N}\right) = \frac{e^{-2N\_\epsilon \hbar \nu/N} - 1}{e^{-4N\_\epsilon \hbar \nu} - 1} \tag{37}$$
If Ne is of the order of N, the numerator of the right-hand side of Eq. 37 becomes approximately e 2hs 1, whose Taylor approximation around hs ¼ 0 is simply <sup>2</sup>hs. We then obtain the result used in [199], and in all the papers that implemented mutation–selection (-drift) models (e.g., [197, 199–201, 204]):
$$\Psi\left(1,\infty;\frac{1}{2N}\right) = \frac{2hs}{1 - e^{-4N\_cbs}}\tag{38}$$
Two critical points should be noted here. First, none of the recent codon models [197, 199–202, 204, 210, 211] ever investigated the role of dominance h, as they all consider that the allele under (positive) selection is fully dominant. Second, Table 3 shows that another class of selection models, those based on balancing selection, has never been considered so far. The impact of the selection model on the predictions made by the mutation–selection (-drift) models is currently unknown.
#### 5 High-Performance Computing for Phylogenetics
5.1 Parallelization Because of the dependency of the likelihood computations on the shape of a particular tree (see Subheading 2.6), most phylogenetic computations cannot be parallelized to take advantage of a multiprocessor (or multicore) environment. Nevertheless, two main directions have been explored to speed up computations: first, in computing the likelihood of substitution models that incorporate among-site rate variation and second, in distributing bootstrap replicates to several processors, as both types of computations can be done independently. A third route is explored in Chapter 7.4 [212].
> In the first case, among-site rate variation is usually modeled with a Γ distribution [213] that is discretized over a finite (and small) number of categories [214]. The likelihood then takes the form of a weighted sum of likelihood functions, one for each discrete rate category, so that each of these functions can be evaluated independently. The route most commonly used is the plain "embarrassingly parallel" solution, where completely independent computations are farmed out to different processors. Such is the case for bootstrap replicates, for which a version of PhyML [24] exists, or in a Bayesian context for independent MCMC samplers [215] (see Subheading 2.9.3). The PhyloBayes-MPI package implements distributed likelihood calculations across sites over several compute-cores, allowing for a genuinely parallelized MCMC run [216, 217].
5.2 HPC and Cloud Computing More recent work has focused on the development of heuristics that make large-scale phylogenetics amenable to high-performance computing (HPC) that are performed on computer clusters. Because of the algorithmic complexity of resolving phylogenetic trees, an approach based on "algorithmic engineering" was developed [218]. The underlying idea is akin to the training phase in supervised machine learning [123], except that here the target is not the performance of a classifier but that of search heuristics. All of these heuristics reuse parameter estimates, avoid the computation of the full likelihood function for all the bootstrap replicates, or seed the search algorithm for every n replicate on the results of previous replicates [218]. For instance, in the "lazy subtree rearrangement" [219], topologies are modified by SPR (see Subheading 2.10.2), but instead of recomputing the likelihood on the whole tree, only the branch lengths around the perturbation are re-optimized. This approximation is used to rank candidate topologies, and the actual likelihood is evaluated on the complete tree only for the best candidates. These heuristics now permit the analysis of thousands of sequences in a probabilistic framework [220], but the actual convergence of these algorithms remains difficult to evaluate, especially on very large data sets (e.g., >104 sequences).
In addition to the reduction of the memory footprint for sparse data matrices [221], an alternative direction to "tweaking likelihood algorithms" has been to take direct advantage of the computing architecture available. One particular effort aims at tapping directly into the computing power of graphics processing units or GPUs, taking advantage of their shared common memory, their highly parallelized architecture, and the comparatively negligible cost of spawning and destroying threads on them. As a result, it is possible to distribute some of the summation entering the pruning algorithm (see Subheading 2.6) to different GPUs [222]. The number of programs taking advantage of these developments is widening and includes popular options such as BEAST [91] and MrBayes [223].
All these fast algorithms can be installed on a local computer cluster, a solution adopted by many research groups since the late 1990s. However, installing a cluster can be demanding and costly because a dedicated room is required with appropriate cooling and power supply (not to mention securing the room, physically). Besides, redundancy requirements, both in terms of power supply and data storage, as well as basic software maintenance and user management, may demand hiring a system administrator. An alternative is to run analyses on a remote HPC server, in the "cloud." Canada, for instance, has a number of such facilities, thanks to national funding bodies (CAC at cac.queensu.ca, SHARCNET at www.sharcnet.ca, or Calcul Quebec at www.calculquebec.ca, just to name a few), and commercial solutions are just a few clicks away (e.g., Amazon Elastic Compute Cloud or EC2). Researchers can obtain access to these HPC solutions according to a number of business models (free, on demand, yearly subscription, etc.) that are associated with a wide spectrum of costs [224]. But in spite of the technical support offered in the price, users usually still have to install their preferred phylogenetic software manually or put a formal request to the team of system administrators managing the HPC facility, all of which is not always convenient.
To make the algorithmic and technological developments described above more accessible, the recent past has seen the emergence of cloud computing [225] dedicated to the phylogenetics community. Examples include the CIPRES Science Gateway (www. phylo.org), or Phylogeny.fr (www.phylogeny.fr, [226]). Many include web portals that do not require that users be well versed in Unix commands, while others may include an application programming interface to cater to the most computer-savvy users. One potential limitation of these services is the bandwidth necessary to transfer large files, and storage requirements—especially in the context of next generation sequencing data. The management of relatively large files will remain a potential issue, unless phylogenetics practitioners are ready to discard these files after analysis, the end product of which is a single tree file a few kilobytes in size, in the same way that people involved in genome projects delete the original image files produced by massively parallel sequencers. Data security or privacy might not be a problem in most applications, except in projects dealing with human subjects or viruses such as HIV that expose the sexual practices of subjects. However, once these various hurdles are out of the way, users could very well imagine running their phylogenetic analyses with millions of sequences from their smartphone while commuting.
#### 6 Conclusions
Although most of the initial applications of likelihood-based methods were motivated by the shortcomings of parsimony, they have now become well accepted as they constitute principled inference approaches that rely on probabilistic logic. Moreover, they allow biologists to evaluate more rigorously the relative importance of different aspects of evolution. The models presented in this chapter have the ability to disentangle rates from times (Subheading 3), or mutation from selection (Subheading 4), while in most cases accounting for the uncertainty about nuisance parameters. But the latest developments described above still make a number of restrictive assumptions (Subheading 4.2), and while many variations in model formulations can be envisaged, they still remain to be explored in practice.
Although some progress has been made in developing integrative approaches (e.g., [176, 181]), throughout this chapter we have assumed that a reliable alignment was available as a starting point. A number of methods exist to co-estimate an alignment and a phylogenetic tree (see Part I of this book), but the computational requirements and convergence of some of these approaches can be daunting, even on the smallest data sets by today's standards.
This brings us, finally, to the issue of tractability of most of these models in the face of very large data sets. The field of phylogenomics is developing quickly (see Part III), at a pace that is ever increasing given the output rate of whole genome sequencing projects. Environmental questions are drawing more and more attention, and metagenomes (see Part VI) will be analyzed in the context of what will soon be called metaphylogenomics. Exploring the numerous available and foreseeable substitution models in such contexts will require continued work in computational methodologies. As such, modeling efforts will continue to go hand-in-hand with, and maybe dependent on, algorithmic developments [227]. It is also not impossible that in the near future, the use of likelihoodfree approach such as ABC or machine learning algorithms in computational molecular evolution be more thoroughly explored.
#### Acknowledgements
We would like to thank Michelle Brazeau, Eric Chen, Ilya Hekimi, Benoıˆt Page´, and Wayne Sawtell for their critical reading of a draft of the original chapter, as well as Jonathan Dench and George S. Long for their careful reading of the most recent draft. This work was supported by the Natural Sciences Research Council of Canada (SAB, NR).
#### References
adaptive evolution in human influenza virus A. J Mol Evol 51:423–432
methods for detecting adaptive evolution in protein coding sequences and for identifying positively selected sites. Genetics 168:1041–1051
topology has little effect on the decisiontheoretic approach to model selection in phylogeny estimation. Mol Biol Evol 22:691–703
precambrian explosive diversification of the Metazoa. Mol Biol Evol 20:1947–1954
sequences. Philos Trans R Soc Lond B Biol Sci 363:3931–3939
Liu L, Suchard MA, Huelsenbeck JP (2012) MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Syst Biol 61:539–542
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Part II
#### Genomic Alignment and Homology Inference
## Chapter 4
## Whole-Genome Alignment
#### Colin N. Dewey
#### Abstract
Whole-genome alignment (WGA) is the prediction of evolutionary relationships at the nucleotide level between two or more genomes. It combines aspects of both colinear sequence alignment and gene orthology prediction and is typically more challenging to address than either of these tasks due to the size and complexity of whole genomes. Despite the difficulty of this problem, numerous methods have been developed for its solution because WGAs are valuable for genome-wide analyses such as phylogenetic inference, genome annotation, and function prediction. In this chapter, we discuss the meaning and significance of WGA and present an overview of the methods that address it. We also examine the problem of evaluating whole-genome aligners and offer a set of methodological challenges that need to be tackled in order to make most effective use of our rapidly growing databases of whole genomes.
Key words Sequence alignment, Whole-genome alignment, Homology map, Toporthology, Genome evolution, Comparative genomics
#### 1 Introduction
When the problem of biological sequence alignment was first described and addressed in the 1970s, sequencing technology was limited to obtaining the sequences of individual proteins or mRNAs or short genomic intervals. As such, classical sequence alignment (as described in Chapter 7 [1]) is typically focused on predicting homologous positions within two or more relatively short and colinear sequences, allowing for the edit events of substitution, insertion, and deletion. Although limited in its scope, this type of alignment remains extremely important today, with genesized alignments forming the basis of most evolutionary studies.
Starting in 1995 with the sequencing of the 1.8 Mb-sized genome of the bacterium H. influenzae [2], biologists have had access to a different scale of biological sequences, those of whole genomes. DNA sequencing technology has rapidly improved since that time, and as a result, we have seen an explosion in the availability of whole-genome sequences. As of the writing of this chapter, there are 9071 published complete genome sequences (8380
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_4, © The Author(s) 2019
bacterial, 281 archaeal, and 410 eukaryotic), according to the GOLD database [3]. Whole-genome sequencing remains popular, with over 140,000 sequencing projects that are either ongoing or completed.
Along with the ascertainment of these sequences, the problem of whole-genome alignment (WGA) has arisen. As each genome is sequenced, there is interest in aligning it against other available genomes in order to better understand its evolutionary history and, ultimately, the biology of its species. Like classical sequence alignment, WGA is about predicting evolutionarily related sequence positions. However, aligning whole genomes is made more complicated by the fact that genomes undergo large-scale structural changes, such as duplications and rearrangements. In addition, a set of genomes may contain pairs of sequence positions whose evolutionary relationships can be described by any of the three major subclasses of homology: orthology, paralogy, and xenology. As orthologous positions are typically of primary interest, WGA also involves the classification of homologous relationships.
In this chapter, we describe the problem of WGA and the methods that address it. We begin with a thorough definition of the problem and discuss the important downstream applications of WGAs. We then categorize the WGA methods that have been developed and describe the key computational techniques that are used within each category. In addition to describing whole-genome aligners, we also discuss the various approaches that have been used for evaluating the alignments they produce. Lastly, we lay out a number of current methodological challenges for WGA.
#### 2 The Definition and Significance of WGA
2.1 WGA as a Correspondence Between Genomes
In imprecise terms, a WGA is a "correspondence" between genomes. For each segment of a given genome, a WGA tells us where its "corresponding" segments are in other genomes. A segment may be one or more contiguous nucleotide positions within a genome. What does it mean for two genomic segments to "correspond" to each other? In most situations, we consider two segments to be "corresponding" if they are orthologous. Orthologous sequences are those that are evolutionarily related (homologous) and that diverged from their most recent common ancestor (MRCA) due to a speciation event [4]. In contrast, paralogous sequences are homologs that diverged from the MRCA due to a duplication event. Thus, by definition, orthologous sequences are the most closely related pieces of two genomes and, as is more thoroughly discussed later and in Chapter 9 [5], are of primary interest because they are useful for applications such as function prediction and species tree inference. As such, WGA is most commonly taken to be the prediction of orthology between the components of entire genome sequences. When a WGA also predicts paralogy, typically only paralogs whose MRCA is at least as recent as the MRCA of entire set of genomes are considered, as there is extensive ancient homology within extant genomes.
It is important to note that the orthologous relationships between two genomes do not create a one-to-one correspondence. Duplication events that have occurred since the time of the MRCA of the species can result in a genomic segment in one species having multiple orthologous segments in another. This is a particularly important issue when the genome of one lineage has undergone a whole-genome duplication event since the time of the MRCA. In this situation, few segments of the genome of the nonduplicated lineage have a single ortholog in the other genome.
2.2 Toporthology In many cases, WGAs do not aim to predict all orthologous sequences. Instead, they only predict toporthology (positional orthology), a distinguished subset of orthology [6, 7]. The concept of toporthology captures the notion that not all orthologous relationships are equivalent in terms of the evolutionary history of the genomic context of the orthologs. Figure 1 gives an example scenario in which toporthology helps to distinguish between two orthologous relationships.
> The definition of toporthology relies on a classification of duplication events. A duplication event is considered to be "symmetric" if the removal of either copy of the duplicated genomic material (immediately after the event) reverts the genome to its original (preduplication) state. Examples of symmetric duplications are tandem and whole-genome duplications. If only one specific copy can be removed to undo a duplication event, then the event is considered "asymmetric." In the asymmetric case, the removable copy is referred to as the "target," with the other copy referred to as the "source." Retrotransposition and segmental duplication both belong to the asymmetric class.
> With this classification of duplication events in hand, we can now define toporthology. Two genomic segments are toporthologous if they are orthologous and neither segment is derived from the target of an asymmetric duplication event since the time of the MRCA of the segments. Thus, two orthologous segments are toporthologous if their evolutionary history (since the MRCA) only involves symmetric duplication events or asymmetric duplications in which their ancestral segment was part of the source copy.
> The important property of toporthologs is that, in the absence of rearrangement events, they share the same ancestral genomic context. As the context of a gene or genomic segment has functional consequences, toporthologous sequences are generally expected to be more similar in their function than orthologous sequences that are not toporthologous (atoporthologs) [6]. However, there is no guarantee that toporthologs share a common
Fig. 1 A hypothetical evolutionary scenario in which the relation of toporthology distinguishes between two ortholog pairs. The bullet-like shapes indicate genomic segments. Both YB1 and YB2 are orthologous to YA. However, only YB1 is toporthologous to YA because YB2 was derived from the target of an asymmetric duplication since the time of the most recent common ancestor, Y, of YB2 and YA
function or that two genomic intervals that have the same function are toporthologs. Thus, a rigorous functional analysis of genomes should consider all classes of homology. Nevertheless, WGAs that focus on toporthology produce a good first approximation to a functional correspondence between genomes.
2.3 Definition and Representation To be more precise, a WGA is, in general, the prediction of homologous pairs of positions between two or more genome sequences. Often, as we have previously discussed, only orthologous or toporthologous relations are predicted in WGAs. And while alignment is typically focused on homologous relationships between sequences, whole-genome comparisons can also include alignments within genomes, which represent paralogous sequences.
Note that we define WGA as homology prediction at the level of nucleotides. Although the concept of homology is more commonly used with respect to entire genes or proteins, it is easily used and, in fact, more naturally defined at the level of single nucleotides. Homology of nucleotide positions is established through templatedriven nucleotide synthesis, and the definitions of orthology, paralogy, and xenology for nucleotides follow those for genes [7].
While a WGA can be defined as a prediction of homology statements, it is usually represented as a set of nucleotide-level alignment matrices or "blocks," each block made up by segments of the genomes that are both homologous and colinear. Homologous genomic segments are colinear if they have not been broken by a rearrangement event since the time of their MRCA. Since rearrangement events, such as inversions, are common at the scale of entire genomes, WGAs are typically made up of many blocks. In general, a block contains two or more genomic segments, and multiple segments in the same block may belong to the same genome (indicating paralogous sequence). One specific WGA representation, the "threaded blockset" [8], requires that every
Fig. 2 An example WGA of three genomes represented as a set of alignment blocks. (a) The positions of the genomic segments that are in the alignment blocks are shown as shaded bullet-like shapes (the direction of the bullet indicates the orientation of the segment). In this example, not all genomic segments belong to a block (note the unshaded intervals). (b) The alignment blocks of the WGA. Note that blocks do not need to contain a segment from all genomes (e.g., block Y) and that some blocks can contain multiple segments from the same genome (e.g., blocks X and Z). (c) A slice of alignment block Z, which is a nucleotide-level alignment
position belongs to a block and thus additionally allows a block to contain just a single segment, which would represent a unique genomic sequence. Figure 2 depicts a hypothetical example of a WGA, with some blocks containing both orthologous and paralogous sequences.
As more genomes are added to an alignment or the total evolutionary divergence between them is increased, the blocks in a WGA decrease in size and increase in number. One might imagine that in the limit of an infinite number of genomes or an infinite amount of time, all blocks might have length one (a single column), which makes the concept of an "alignment matrix" irrelevant. However, rearrangements in certain segments of the genome are likely to be highly deleterious to an organism and will thus never be observed. Such segments are referred to as genomic "atoms" [9] and prevent all blocks from becoming single alignment columns.
2.4 Comparison to Other Homology Prediction Tasks WGA is closely related to classical sequence alignment (the alignment of two or more relatively short and colinear sequences), and most whole-genome aligners rely on classical alignment techniques (e.g., the Needleman–Wunsch [10] and Smith–Waterman [11] pairwise alignment algorithms and heuristics used for multiple alignments) as subroutines. However, there are three key differences between these two classes of alignment. First, and most importantly, classical alignment requires sequences to be colinear, which is often not the case for genome sequences due to rearrangement events. Second, even when restricted to toporthologous relationships, the correspondences between genomes are not one to one, which is also a requirement of classical alignment. Due, in part, to the complications of these first two issues, it is difficult to formulate a useful objective function (such as the sum-of-pairs score for classical alignment) for WGA. Thus, most genome alignment methods are heuristic procedures that lack an explicit objective. A last difference between classical alignment and WGA is the scale of the problem. Classical alignment typically focuses on the alignment of single genes, which are usually on the order of thousands of nucleotides long. Whole genomes, in contrast, are millions to billions of nucleotides in length. The facts that genomes are large and are often neither colinear nor in one-to-one correspondence with other genomes are what make WGA challenging.
Since WGA is often focused on orthologous relationships, it is also related to the "orthology prediction" problem (see Chapter <sup>9</sup> [5]). The key difference between the two problems is that orthology prediction is traditionally cast at the level of genes, whereas WGA operates at the level of nucleotides. For most orthology prediction methods, a genome is treated as an unordered set of genes. Whole-genome aligners, on the other hand, consider a genome to be a set of DNA sequences (chromosomes) within which genes are embedded. Thus, a WGA provides orthology predictions for both genes and intergenic regions. Due in part to their treatment of genomes as long nucleotide sequences, current WGA methods rely exclusively on sequence similarity and the ordering of nucleotides in a genome to predict orthology. In contrast, orthology prediction methods often use phylogenetic analyses, which can be more powerful than genome order and sequence similarity information alone. Thus, while the problem of WGA is broader in scope than that of orthology prediction, it is restricted to the analysis of relatively closely related genomes, for which homology of nongenic nucleotides is detectable and gene order is at least partially conserved. Gene-level orthology prediction is more appropriate for distantly related genomes, which may only have detectable homology at the amino acid level and little colinearity.
2.5 Significance WGAs are powerful because they allow for the analysis of molecular evolution at both large and small scales. At the large scale, one can use such alignments to estimate the frequency and location of rearrangement and duplication events. For example, one might use a WGA between human and mouse to identify colinear orthologous blocks, which are then given to a rearrangement analysis method (e.g., [12]) to determine a most parsimonious set of rearrangement events explaining the current structures of the two genomes. At the small scale, WGAs can be used to examine the rates of substitutions and indels across the entire genome. For example, one might look at alignments of ancestral repeats to estimate the neutral rates of nucleotide evolution. Both smalland large-scale mutational events identified from WGAs can be used as data for species tree inference. In combination with carefully constructed models of genome evolution at both scales, WGAs also enable the task of ancestral genome reconstruction [13, 14].
Beyond purely evolutionary studies, WGAs are valuable for identifying functional elements within genomes. Each class of functional element within the genome tends to have a unique "evolutionary signature," which can be searched for within WGAs [15]. For example, coding sequences tend to have mutational patterns with a predominance of substitutions at the third positions of codons, which are unlikely to affect the amino acid sequence. This characteristic evolutionary signature of coding sequence has led to the development of comparative gene-finding methods, which often use WGAs (Chapter 6 [16]). Noncoding RNA sequences can also be identified from WGAs but have more complex signatures involving compensatory mutations that maintain base pairing within RNA secondary structures [17]. More generally, one can search for evolutionarily constrained regions within WGAs, which can contain functional elements from a variety of classes [18]. When combined with the knowledge of transcription factor-binding motifs, this approach can be used to identify transcription factor-binding sites with a technique called "phylogenetic footprinting" [19]. The easiest evolutionarily constrained regions to pick out are those of "ultraconserved elements," which maintain high levels of sequence identity across large evolutionary distances and are primarily noncoding components of the genome [20].
WGAs also allow for the transfer of functional information about specific elements from one species to another. As WGAs typically predict orthology and orthologous sequences are likely to have similar functions, WGAs are valuable for function prediction. By aligning at the nucleotide level across the genome, they can aid in function prediction for both genes and nongenic regions, such as those that contain regulatory elements. For example, if we are interested in a specific disease-associated interval in the human genome, we might use an alignment to identify where its mouse orthologs are located. Knowledge of the mouse orthologs would enable us to have a better understanding of the evolutionary history of this genomic region and could lead to genetic manipulation experiments that can only be performed in mice.
#### 3 Methods for WGA
3.1 A Simplistic Approach
It is easier to understand the existing methods for performing WGA by first appreciating the shortcomings of a simplistic approach for comparing whole-genome sequences. One simple approach would be to run BLAST [21], or another similar local alignment tool, between all pairs of genomes. The WGA would then be defined as the union of all significant pairwise local alignments discovered by BLAST. By using a local alignment tool, we avoid the issues of rearrangements and duplications, as sets of local alignments are not constrained to be colinear or in one-to-one correspondence.
While this approach would certainly yield a large set of homology predictions between all pairs of genomes, it has a number of shortcomings. First, by only using a BLAST significance threshold, it makes no distinction between orthology, paralogy, and other refinements of homology. Second, the pairwise alignments that it produces are not guaranteed to be consistent with each other, even though homology, by definition, is a transitive relation. Third, BLAST may miss some homologous sequences that have low similarity but are strongly supported in their relatedness by flanking homologous sequences. BLAST's significance statistics are proven for ungapped sequences and good in practice for sequences with short indels [22], but are not designed for whole-genome comparisons, which often feature large-scale insertions and deletions and heterogeneous substitution rates. Lastly, this approach is overly computationally intensive. For example, it does not take advantage of the fact that homology is a transitive relation, that relationships between sequences are reasonably modeled by a tree, and that homologous sequences between genomes are often found in long colinear segments.
3.2 The Two Major Approaches to WGA Existing WGA methods attempt to address one or more of the weaknesses of this simple approach. These methods can be loosely classified into two major strategies which we refer to as the "hierarchical" and "local" approaches. The main idea behind the hierarchical approach is to split the WGA problem into a set of global multiple alignment problems. To do this, it first identifies the colinear and homologous (typically orthologous) segments of the genomes. Each set of colinear segments is then given to a specialized genomic global alignment method to produce a nucleotide-level alignment. In contrast, the first step of the "local" approach is to produce a large set of nucleotide-level alignments. Later steps involve the filtering and merging of these alignments to produce sets of pairwise or multiple alignments of homologous (typically orthologous) sequences. Despite their differences, both strategies typically begin with a local alignment step that is similar to the simplistic all-vs.-all alignment of the BLAST approach. A summary of all of the WGA methods described in this chapter and the role they play within one or both approaches is given in Table 1.
> Both approaches have advantages and disadvantages. The primary advantage of the hierarchical approach is that it can often be faster and breaks a WGA into a number of independent subproblems that can be solved in parallel. It is faster because the
(continued)
#### Table 1 (continued)
For each method, the approach it uses or the role it plays within a larger WGA system is given in the "category" column. Each method is labeled as either "pairwise" or "multiple" depending on whether it can be applied to generate multiple alignments. In addition, the primary type of evolutionary relationship predicted by each method is given in the "relationships predicted" column
> identification of long colinear and orthologous segments in the genomes can be accurately computed without the need for sensitive nucleotide-level alignments. However, because hierarchical methods do not often use the most sensitive aligners for this step, they tend to miss small rearranged or diverged segments. Thus, the primary advantage of the local method is in its sensitivity to these regions, although "glocal" alignment methods [23], which allow for small rearrangements, can partially ameliorate this weakness of hierarchical methods. Hierarchical methods also run the risk of being overconfident of the colinearity of genomic segments and can thus produce more false-positive aligned positions within sequences predicted to be colinear.
3.3 Local Pairwise Genomic Alignment Methods for both WGA strategies generally start by finding local alignments between, and perhaps within, the genomes. The Smith–Waterman algorithm is the classical solution to the pairwise local alignment problem, but is generally not used for WGA because it runs in time quadratic in the size of the genomes, which can be large. Instead, most methods adopt a "seed-andextend" approach for discovering high-scoring local alignments, much like BLAST. This approach first identifies short ungapped matches between the sequences using one of a variety of data structures. It then extends the short matches from both ends using a variant of the Smith–Waterman algorithm, stopping the extension when the score of the alignment drops below a specified threshold. In some cases, nearby and consistent (in terms of order and orientation) local alignments are "chained" together to form larger alignments.
There are a number of techniques used for discovering seeds at the genomic scale for the "seed-and-extend" approach to local alignment. A first distinction between the techniques is whether they find exact or inexact matching seeds. Exact seed discovery is often faster and easier to implement, whereas inexact seeds offer better sensitivity. Seed techniques also vary in whether they use "consecutive" or "spaced" seeds [24]. Consecutive seeds consider matches and mismatches at all positions within a sequence interval, whereas spaced seeds only check for matches at a subset of positions within an interval. The specific subset of positions checked is known as the "seed pattern," and there has been significant work on determining optimal sets of multiple seed patterns (e.g., [25, 26]). It has been shown that carefully chosen spaced seed patterns are superior to consecutive seeds in terms of sensitivity [27]. Lastly, seeds differ in whether their lengths are fixed or adaptive (variable). For WGA, adaptive seeds have been shown to allow for faster local alignment at the same level of sensitivity as fixed seeds [28].
Seed-finding techniques can often be improved by taking advantage of DNA evolutionary models. A generalization of spaced seeds is "subset seeds" [29], which allow subsets of bases to be considered equivalent when determining if there is a match at a given position. Subset seeds are particularly useful for taking into account that transitions are often more common than transversions in genome comparisons. Further taking into account biologically informed substitution patterns is the "translated" seed, which is a match at the amino acid level after translating genomic sequences in all six possible reading frames. Translated seeds enable increased sensitivity in comparisons of more diverged genomes. Lastly, when aligning a genome to a set of genomes for which a multiple WGA has already been constructed, one can take into account the substitution patterns and ancestral sequences inferred from the WGA to devise more sensitive seeds [30, 31].
The choice of seed type is the major determinant of the data structures used for seed discovery. For example, BLAT [32] uses a simple index of all possible k-mers for exact and translated seeds but uses a heuristic of indexing only nonoverlapping k-mers for memory efficiency. STELLAR [33] also uses an index of k-mers but implements an exact algorithm based on filtration for finding all local alignments with an error rate below a given threshold. LASTZ (the successor to BLASTZ [34]), which uses a carefully chosen spaced seed pattern introduced by [24], instead uses a hash table to find both exact and inexact matches. Not to be confused with LASTZ is the more recently developed LAST aligner [28], which uses adaptive seeds with highly configurable patterns that are identified via a suffix array data structure. MUMmer uses a suffix tree to rapidly find all exact consecutive seeds with some minimum length [35]. CHAOS [36], which is a component of the LAGAN-suite of genome alignment tools [37], uses a related structure, a "threaded trie," to find exact and inexact consecutive seeds.
For computational efficiency reasons, the extension step of the seed-and-extend approach typically only allows for ungapped alignments or alignments with short indels. However, genome alignments often feature large indels that are not discovered by extension from a seed. Thus, many local genomic alignment tools use a "chaining" step to link nearby and consistent local alignments discovered by the seed-and-extend strategy. For example, MUMmer includes a module for chaining together nearby exact matches using a variation of the longest increasing subsequence (LIS) problem [38]. CHAOS also uses an LIS-derived algorithm for chaining the inexact consecutive seeds it discovers. Chaining is often followed by more sensitive alignment between chained local alignments. For example, MUMmer runs a variant of Smith–Waterman alignment in between chained matches and LASTZ recursively searches for alignments with more sensitive seeds in between nearby alignments discovered in previous steps.
3.4 The Hierarchical Approach The hierarchical approach to WGA consists of two steps. First, a high-level homology map between the genomes is constructed. Second, a nucleotide-level alignment is obtained by running a genomic global alignment tool on each homologous and colinear set of genomic segments identified by the homology map. Hierarchical WGA methods vary in the exact techniques used for each step.
The idea behind the hierarchical approach is to separate the problem of identifying rearrangements and duplications from that of obtaining a nucleotide-level alignment. In the absence of rearrangements and duplications, WGA simply reduces to classical sequence alignment although at a much larger scale. Thus, if a WGA problem can be broken into a set of subproblems that do not contain these large-scale events, the numerous methods that have been developed for classical global alignment can be utilized.
The first step of the hierarchical strategy is to construct a homology map between the genomes of interest. A homology map is a collection of sets of genomic intervals, where each set of intervals is required to be homologous and colinear (i.e., free of rearrangements and duplications). Each set represents the sequences that will ultimately form a block within a WGA. Homology maps generally have the property that each genomic position belongs to at most one set and has all of its homologs contained within that set. For WGA, homology maps are often restricted in the evolutionary relationships that are captured, as only a subset of homologous relationships may be of interest. Typically, only orthologous relationships are captured, forming an "orthology map." When orthology maps are restricted to predicting one-to-one relationships, they are more likely to be representative of toporthology.
The concept of a homology map is closely related to the concepts of "conserved segments" and "syntenic blocks," which generally refer to sets of genomic intervals containing multiple homologous markers (e.g., genes) and featuring conserved orientations and adjacencies of these markers [39, 40]. Unfortunately, these concepts have long been poorly defined, and, as a result, methods for syntenic block identification differ markedly in their output [41]. In addition, methods for identifying syntenic blocks (or closely related concepts) are often focused on identifying sets of genomic intervals that exhibit levels of conservation of marker content or colinearity that exceed what one would expect if markers were randomly shuffled between genomes (e.g., [42–44]). This is in contrast to homology maps, which are concerned with colinear homology, regardless of biological significance. And, in practice, homology maps are intermediate objects in the process of WGA, whereas syntenic block predictions are often of direct interest.
Homology maps are most commonly constructed from local alignments, such as those computed by methods discussed in the previous section. As only a high-level correspondence is desired, these methods are often run in faster but less sensitive configurations. For example, local alignments between just the coding intervals of the genomes can be computed quickly and used for the construction of homology maps that are at least accurate with respect to protein-coding genes.
Although numerous pairwise homology mapping methods exist, in this chapter, we restrict our attention to methods that scale to more than two genomes, as the problem is significantly more challenging in the multiple genome case. Examples of multiple genome homology map methods include GRIMM-Synteny [40], its successor DRIMM-Synteny [45], Mercator [46], Enredo [47], OSfinder [48], SuperMap [49], and Sibelia [50]. The WGA programs M-GCAT [51], progressiveMauve [52], MUGSY [53], and Cactus [54] are integrated hierarchical methods that contain a homology mapping stage.
Many of these methods use graph-based data structures to find a mapping between multiple genomes simultaneously. Kehr et al. [55] characterized the relationships between four commonly used types of graphs: alignment graphs [56], A-Bruijn graphs [57, 58], Enredo graphs [47], and Cactus graphs [59]. The most straightforward graph is the alignment graph, which is a mixed graph with vertices representing genomic segments, directed edges representing adjacent segments, and undirected edges representing homologous segments. In an A-Bruijn graph, vertices instead represent sets of homologous segments, and directed edges represent adjacencies between pairs of segments (one from each set represented by the connected vertices). Relative to alignment graphs, A-Bruijn graphs are more compact and readily reveal the content of each genome. An Enredo graph is very similar to an A-Bruijn graph, but has a pair of vertices instead of a single vertex for each set of homologous segments, which captures information regarding the directionality of each segment within a homologous set. Lastly, cactus graphs flip the representation of adjacencies, with vertices corresponding to sets of adjacencies and edges corresponding to sets of homologous segments. Cactus graphs have a natural decomposition that provides advantages for analysis and visualization of WGAs.
Graph-based homology mapping methods generally produce an initial WGA graph using one of the four representations we have discussed and then refine the graph via modifications. Of the homology mapping methods we have listed, GRIMM-Synteny, Mercator, and MUGSY use alignment graphs. DRIMM-Synteny and OSfinder use A-Bruijn graphs and Sibelia uses de Bruijn graphs, of which A-Bruijn graphs are a generalization. And, as their names suggest, Enredo and Cactus use Enredo and cactus graphs, respectively. These methods use a variety of techniques for graph refinement. For example, MUGSY is unique in its use of flow network algorithms to identify breaks in colinearity. OSfinder uses a novel probabilistic model to determine a maximum likelihood multiple genome orthology map. And Cactus uses a simulated annealing-style algorithm, the Cactus alignment filter, to refine an initial cactus graph representing a homology map.
Unlike the graph-based methods that build a map between all genomes simultaneously, the SuperMap and progressiveMauve methods build a multiple genome map by progressively building pairwise maps up a guide tree. The pairwise SuperMap algorithm is essentially a symmetric version of the chaining method used by Shuffle-LAGAN [23], which allows for rearrangements and duplications in its chains of orthologous segments. The progressive-Mauve mapping method instead uses a "breakpoint elimination" algorithm to find colinear segments and does not allow for duplications, thus producing output indicative of one-to-one toporthology. This algorithm greedily removes local alignments one by one with the goal of maximizing an objective function that takes into account both the number of breakpoints implied by an alignment and substitution scores.
Once a homology map has been created, any one of a number of genomic global alignment methods can be used to align the orthologous and colinear segments identified by the map. As for our discussion of homology mapping methods, we restrict our attention to global aligners that can handle multiple genomes. Examples of such methods are MAVID [60], MLAGAN [37], DIALIGN [36], SeqAn::T-Coffee [61], PECAN [47], FSA [62], and the base-level alignment refinement (BAR) algorithm of Cactus [54]. For colinear sequences, the genomic alignment problem is the same as that of classical global alignment but is made more difficult by the fact that the sequences are long (possibly millions of nucleotides in length). Thus, global genomic aligners employ heuristics to speed up the process. By far, the most common heuristic used is to first identify short local alignments, or anchors, between the sequences, identify a chain of these anchors, and then perform global alignment between the adjacent chained anchors. This technique is similar to the strategy for hierarchical WGA, but is simpler, due to the fact that rearrangements and duplications do not need to be taken into account. MLAGAN and DIALIGN use the CHAOS local aligner, PECAN and FSA use Exonerate [63], and MAVID and SeqAn::T-Coffee use suffix trees or arrays to find anchors.
In addition to the specific local alignment technique used to speed up the alignment process, global genomic aligners also vary with respect to how they combine local pairwise alignments to build a multiple global alignment. First, MAVID, MLAGAN, SeqAn::T-Coffee, and Pecan all belong to the class of progressive alignment methods, which use a phylogenetic tree to guide their algorithms (see Chapter <sup>7</sup> [1]). For the alignment of non-leaf sequences during progressive alignment, MAVID uses maximum likelihood ancestral sequence inference, while MLAGAN, SeqAn:: T-Coffee, and Pecan use a sum-of-pairs objective function. Both SeqAn::T-Coffee and Pecan use a "consistency" technique, which adjusts the score between pairs of positions (or segments) based on the consistency of triplets of pairwise alignments. The nonprogressive methods, DIALIGN, FSA, and BAR, instead put together a multiple alignment by greedily merging consistent local pairwise alignments. While differing in their use of a tree, the FSA, Pecan, and BAR methods take advantage of probabilistic models of sequence alignment and attempt to maximize statistically grounded objective functions, as opposed to the heuristic score-based functions used by the other methods. BAR is unique in its ability to predict breakpoints when aligning groups of sequences that may contain the boundaries of rearrangement events.
Although the hierarchical approach breaks the WGA problem into a large number of subproblems (one per colinear segment set) that can be computed in parallel, it is still a significant computational effort to produce a WGA with this approach, particularly for large eukaryotic genomes. Thus, a number of Web sites host precomputed hierarchical WGAs. Alignments produced by the combination of Pecan with either Enredo or Mercator are hosted at the Ensembl Web site [64]. Similarly, the VISTA Web site [65] hosts WGAs generated by SuperMap and the LAGAN-suite of genomic aligners. Both sites offer visualizations of the WGAs, which are useful for looking at levels of conservation across genomes.
3.5 The Local Approach The local approach to WGA bypasses the high-level homology map construction phase of the hierarchical approach and instead begins by identifying a comprehensive set of nucleotide-level pairwise local alignments. The second step of this approach is to combine the pairwise local alignments into a cohesive WGA by filtering out nonorthologous relationships and merging pairwise alignments into multiple alignments. Because there is typically no additional pairwise nucleotide-level alignment performed in the second step, the local alignments generated by the first step are obtained with a more sensitive aligner than that used by hierarchical methods for homology map building. The two primary examples of local WGA methods are MUMmer, a pairwise genome aligner, and MULTIZ/ TBA, a multiple genome aligner [8].
> MUMmer was one of the first pairwise WGA methods to be developed and was initially targeted at the alignment of prokaryotic-sized genomes. The WGA ability of MUMmer is achieved through a combination of smaller modules that is orchestrated by the NUCmer or PROmer scripts. The first module identifies maximum unique matches (MUMs) between a pair of genomes with a suffix tree data structure. Nearby matches are clustered together, and a high-scoring colinear chain of matches is identified within each cluster. Finally, the matches within the chains are extended with a variant of the Smith–Waterman algorithm, and the resulting extended chains are output as a WGA. The raw WGA output by MUMmer can, in general, include all classes of homologous relationships. However, the chains are typically filtered to leave only those that are highest scoring or that result in a reference position being overlapped by only a single chain. Thus, a filtered WGA from MUMmer is usually representative of orthology.
> MULTIZ/TBA, which was instead designed for large eukaryotic genomes, starts by using LASTZ to generate sensitive local pairwise alignments between all pairs of genomes or between a reference genome and all others. MULTIZ is then used to identify local alignment blocks of subsets of genomes that should be combined and to merge these blocks using a banded variant of the Smith–Waterman algorithm. TBA is the program that is used to coordinate this entire process when all pairs of genomes are compared. Thus far, it does not appear that TBA has been used at the whole-genome scale, although MULTIZ is regularly used for reference-based WGAs hosted by the UCSC Genome Browser [66]. For these reference-based WGAs, the ungapped segments of LASTZ alignments are first processed with a chaining program (AXTCHAIN) to establish large colinear alignments between the reference and another genome. In contrast to the output of
chaining methods discussed in Subheading 3.3, a chain produced by AXTCHAIN is an ordered set of pairwise local alignments rather than a single long alignment that explicitly aligns between the short local alignments that form the chain. AXTCHAIN chains are typically filtered by the CHAINNET program to retain only the highest-scoring alignment at each position within the reference genome [67]. The remaining alignments, which most likely reflect orthologous relationships, are then combined into multiple alignments with MULTIZ.
3.6 Refining WGAs Because of the computational complexity of multiple alignment, particularly at the whole-genome scale, methods of both approaches to WGA use heuristics and simplified models to make WGA feasible. For example, most of the methods described in this chapter do not distinguish between different classes of genomic sequence (e.g., genic and intergenic) while constructing nucleotide-level alignments. And many methods disregard small, marginally significant, local alignments for the sake of speed. As a result, at a local level, the results of current WGA methods often leave room for improvement.
> To remedy this situation, a number of methods have been developed that may be used to refine WGAs. These methods take as input either a WGA, a single WGA block, or the set of homologous and colinear sequences that make up a WGA block. They can be generally grouped into one of three categories. The first is composed of methods that refine the local structure of a WGA. That is, they redefine the boundaries, or "breakpoints," of the homologous and colinear blocks in the WGA. A secondary category of methods focuses on optimizing individual WGA blocks with respect to an objective function. The last category includes methods that perform alignment while taking into account the structure and evolutionary dynamics of certain classes of genomic elements.
> PicoInversionMiner [68] and Cassis [69, 70] are two methods for refining the local structure of a WGA. PicoInversionMiner identifies very small "inplace" inversions between two genomes that are left undetected by an initial WGA. Such inversions are represented by alignments that would typically not have statistically significant scores at the genome level but can be detected via probabilistic models of local sequence evolution. In contrast to PicoInversionMiner, which identifies novel rearrangement events, Cassis refines the coordinates of breakpoints. The refinements produced by Cassis are the result of identifying weak similarities between sequences adjacent to segments of an initial orthology map and extending the boundaries of segments based on these similarities. The BAR algorithm of Cactus, which we have previously discussed in the context of hierarchical WGA, is also an alignment refinement method that identifies breakpoints.
Other methods for refining WGAs focus on improving local colinear multiple alignments with respect to a given objective function. For example, GenAlignRefine [71] attempts to optimize WGA blocks according to the COFFEE objective function [72] using a genetic algorithm. The PSAR-Align method [73] instead realigns blocks to optimize an expected accuracy objective function [74] using pairwise alignment probabilities estimated by the PSAR tool [75] and the sequencing annealing algorithm of the FSA multiple alignment method [62]. Lastly, the Phylo project [76, 77] refines WGAs by "crowd sourcing" the task of optimizing colinear alignment blocks, according to one of a number of objective functions. Phylo casts the multiple alignment problem as a casual game that may be played by "citizen scientists" at the project's website (http://phylo.cs.mcgill.ca/).
Lastly, a number of methods have been developed that can improve the alignments of specific classes of genomic elements, such as gene structures. The primary goal of these methods is generally to improve prediction of genomic elements, but a more accurate alignment often results as a side product. Among the oldest of such methods are comparative gene finders that perform protein-coding gene prediction and pairwise alignment simultaneously. These include SLAM [78] and DOUBLESCAN [79], both of which use pair hidden Markov models [80]. A related method, CESAR [81], was specifically designed for realignment and targets individual coding exons rather than full gene structures. Other methods focus on improving the alignment of noncoding regulatory regions by modeling the evolution of sets of transcription factor-binding sites with known motifs (e.g., MORPH [82], EMMA [83], and MAFIA [84]). Like the comparative gene finders, these methods also use statistical alignment techniques but with models extended to take into the account the conservation of binding sites instead of gene structures. SAPF [85] is also a method aimed at alignment of noncoding regulatory regions but more generally models sequences that are mixtures of "slow" and "fast" evolving elements without knowledge of binding motifs. Lastly, REAPR [86] focuses on the realignment and detection of noncoding RNAs by using alignment models that take into account the conserved secondary structures of such RNAs.
#### 4 Evaluation of WGAs
Just as for small-scale alignment (Chapter 7, [1]), assessing the accuracy of WGAs is hard because we rarely know the true evolutionary history of a set of genome sequences. In fact, the evaluation of WGAs is even harder than that of protein alignments. While protein aligners can be evaluated with "gold standard" benchmarking databases where the truth is established through protein structural information, genome aligners have no benchmarks of real data. In addition, WGAs must be assessed not only for whether they align truly homologous sequences but also for whether they correctly predict orthologous (or toporthologous) relationships. Thus, the evaluation of WGAs is related to that of gene orthology prediction, which is discussed in Chapter 9 [5]. Despite these challenges, a number of creative approaches have been used for determining the accuracy of WGA methods. The approaches generally fall into four categories: (1) simulation, (2) analysis of alignments to annotated regions, (3) comparison with predictions from other methods, and (4) alignment statistics.
Simulated data are appealing for evaluation as we know the entire evolutionary history of the simulated sequences and can thus thoroughly evaluate the accuracy of an alignment. Many of the WGA methods described in this chapter have used simulations for assessing their accuracies [8, 47, 52, 54, 62]. The Alignathon [87], one of the most comprehensive evaluations of WGA methods to date, relied heavily on simulated data sets. This study called attention to one potential pitfall of simulation-based evaluation, which is that the performance of a WGA method may be overestimated when that method was developed or trained with respect to the same simulator used for the assessment.
Simulating the evolution of whole genomes is a challenging task, and it is unclear if the current models used for simulation are close to reality. Such models are highly complex, as they have to account for many different types of evolutionary events, at both the small and large scales. For example, they need to model the random mutations of both single-nucleotide substitutions and megabasesized inversions. In addition, they also need to model natural selection, which alters the probability of these random mutations becoming fixed within a population. For example, an inversion that cuts an essential gene in half might have a much lower probability of becoming fixed than an inversion with both end points in intergenic regions. Despite these challenging model details, a number of genomic evolution simulators have been developed. Currently, only three simulators model both small-scale events (e.g., substitutions and indels) and large-scale rearrangements and duplications [88–90]. Other simulators focus only on nonrearranging events [8, 91–98] and are thus good for evaluating colinear genomic aligners but not homology mapping methods.
A second class of approaches to evaluating WGAs leverages our knowledge of various classes of elements within the genome. For example, with our understanding that most coding regions are conserved across closely related genomes, the fraction of exons in a genome "covered" by an alignment is an indirect measure of the sensitivity of a WGA [37, 49, 60, 99]. Specificity can also be roughly assessed with coding regions, either by counting the number of coding bases that are aligned to noncoding bases in other genomes [36, 100] or by checking that alignments in coding regions exhibit periodicities in their substitution patterns [99]. A related approach that instead assesses the accuracy of eukaryotic orthology maps is to check if exons from the same gene are mapped in the same order and orientation to other genomes [47]. For the subset of protein-coding and noncoding RNA genes that have curated "gold standard" alignments, the accuracy of a WGA with respect to those genes may be assessed [101]. However, the fact that genic regions are often highly conserved is also a disadvantage of using them for evaluation; the most conserved regions are the easiest to align, and some aligners use exon annotation information or translated matches. Because of these issues, repeat sequences, which are believed to evolve more neutrally, have been used for alignment evaluation [47, 99]. For example, in [99], sensitivity was assessed by alignments of ancestral repetitive elements, and specificity was inferred from the number of alignments to lineagespecific repeat elements (in this study, primate-specific Alu repeats).
Another common evaluation technique is to compare wholegenome aligners against other related methods. For example, a WGA produced by one method can be used as the "truth" with which to evaluate the sensitivity and specificity of other WGAs [53]. This technique is useful for judging the similarity of different WGAs but, unfortunately, does not provide much information about accuracy. Another technique is to compare with the results from gene orthology prediction programs [48, 49]. The advantage of this approach is that it provides a more independent test of accuracy, since gene orthology prediction programs generally use different algorithms and information sources to infer orthology. The disadvantages of this approach are that it only provides a genelevel measure of accuracy and does not evaluate alignments of noncoding regions. In addition, since WGA and gene orthology prediction share similar goals, we might expect that future methods will blend techniques from both and thus that this evaluation approach will decrease in usefulness.
A last class of evaluation techniques involves the computation of statistics for WGAs. These statistics can be subdivided into simple descriptive statistics and measures computed via statistical or sampling techniques. One of the most straightforward descriptive statistics of a WGA is the "coverage" or the fraction of the genomes included in an alignment or orthology map block [45, 47, 49, 53, 87]. Generally, the higher the coverage, the more sensitive the WGA is believed to be, although one can easily create highcoverage WGAs with poor sensitivity. As a check of large-scale specificity in mammalian WGAs, the authors of [47] checked the fraction of the X chromosome that was covered by alignments to autosomal chromosomes in other genomes (the assumption being that translocations into and out of the X chromosome are rare in mammals). Some more detailed nucleotide-level statistics of WGAs include the total number of "core" positions [53], which are gap-free alignment columns containing all genomes, and the average level of sequence identity in aligned columns [61].
More sophisticated statistics related to WGA accuracy are computed through the use of statistical or sampling techniques. Just as they are used for BLAST, Karlin and Altschul statistics [102] may be used to assess the significance of local pairwise alignments between genomes. StatSigMA extends these statistics to multiple alignments [103], and StatSigMA-w further extends this technique to detect dubiously aligned regions in WGAs of multiple genomes [104]. Whereas a given local pairwise alignment may be highly significant, the flanks of that alignment may be spurious, and a p-value may be computed assessing the possible "over-alignment" of a flank [105]. Within a multiple alignment, a number of techniques have been developed for estimating the accuracy of the alignment of pairs of residues or entire columns, including simply computing an alignment of reversed sequences [106], computing alignments with bootstrapped guide trees [107], sampling suboptimal multiple alignments [75], and evaluating consistency within a library of alternative alignments [108].
#### 5 Future Challenges
Despite the substantial progress made in WGA methodology development, there are a number of challenges that remain unsolved. First, we are in need of WGA methods that can scale to hundreds or thousands of genomes. Along with ever-improving sequencing technology, we are accumulating whole-genome sequences at an increasing rate. Projects such as the Genome 10K Community of Scientists [109], which aims to collect and sequence the genomes of 10,000 vertebrate species, will further push the WGA problem to new scales. While most WGA algorithms have been made efficient for long genomes, very few are practical for large numbers of genomes. Encouragingly, we are beginning to see methods capable of scaling to thousands of genomes for the simpler task of "coregenome alignment" of highly similar microbial-sized genomes [110]. However, methods scaling to thousands of genomes for the full WGA task or for mammalian-sized genomes do not currently exist. In addition to algorithmic advances, we will also be in need of novel approaches for storing and representing WGAs of thousands of genomes.
Second, advances are needed in the parameterization of WGA methods. Current methods are littered with large numbers of parameters that are often heuristic in nature and not easily determined. In some cases, the default parameters for a WGA method may be markedly suboptimal [111]. One solution to this problem is to adopt probabilistic models, which offer principled approaches to parameter estimation, such as maximum likelihood. In fact, probabilistic models of sequence evolution have already been adopted for the alignment of colinear genomic segments and have been shown to offer improved accuracy [47, 62]. However, we have yet to see a method that integrates probabilistic models of both small- and large-scale changes that is capable of constructing an entire WGA, although the recently introduced "split-alignment" pairwise WGA method is a promising step in this direction [112]. In addition, most WGA alignments use models or scoring schemes that assume homogenous rates of evolution across the genome. This assumption is obviously violated in real data, and new methods will need to be developed that take this into account. Simulated noncoding genomic alignments that represent a heterogeneous mix of evolutionary rates have been developed and should be useful for the development of new WGA methodology [97].
Lastly, more attention must be paid to the fact that a WGA is typically just a single estimate of the evolutionary history of a set of genomes and portions of this estimate may be highly uncertain. Encouragingly, methods for colinear genomic alignment have brought light to this issue at the nucleotide level [62, 113]. However, the issue of uncertainty at the large-scale orthology map level has not been sufficiently studied, perhaps due to the lack of probabilistic models for that level of the WGA problem. In addition, most efforts to address uncertainty in alignments simply assign levels of confidence to the components of a single alignment. It may be more useful to be presented with a set of near-optimal alignments so that alternative evolutionary histories can be examined by downstream analyses [114]. The determination and representation of uncertainty for all scales of a WGA will likely remain a challenging problem as the number of genomes included in alignments increases.
#### 6 Exercises
1. Download the whole-genome aligner MUMmer (http://mum mer.sourceforge.net) and FASTA-formatted genome sequences for the species Helicobacter pylori J99 and Helicobacter pylori B38 from GenBank (http://www.ncbi.nlm.nih.gov/genbank/, accessions NC000921 and NC012973, respectively). Run the NUCmer or PROmer programs on the two genome sequences. Visualize the resulting alignment with the mummerplot program. How many colinear blocks are there in the alignment? How many inversion events are implied by the alignment?
2. Visit the UCSC Genome Browser (http://genome.ucsc. edu) and browse the human genome version GRCh38/hg38. Search for and view the CFTR gene, mutations in which cause the disease cystic fibrosis. Turn on the Net tracks for alignments to
Fig. 3 The evolutionary scenario to be considered for Exercise 3. Each bullet-like shape corresponds to a genomic segment, with the direction of the bullet indicating the orientation of the segment
genomes of non-primate placental mammals by clicking on the "Placental Chain/Net" link (in the "Comparative Genomics" section) and choosing the appropriate configuration. Examine the Mouse Net track in the visualization and note the color of the mouse net alignments. Using the "Chromosome Color Key" (located in between the browser visualization and the track configuration section), identify the chromosome on which the mouse ortholog of CFTR is located. Looking at the net alignments for all of the placental mammals, does it appear that CFTR has been conserved across this clade?
3. Consider the evolutionary scenario giving rise to the genomes of three species shown in Fig. 3. For each of the relations listed below, give the pairs of genomic segments with that relation.
#### References
v.6: data updates and feature enhancements. Nucleic Acids Res 45:D446–D456
new generation of protein database search programs. Nucleic Acids Res 25:3389–3402
alignment: the problem of spurious alignment flanks. Nucleic Acids Res 36:5863–5871
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Chapter 5
#### Inferring Orthology and Paralogy
#### Adrian M. Altenhoff, Natasha M. Glover, and Christophe Dessimoz
#### Abstract
The distinction between orthologs and paralogs, genes that started diverging by speciation versus duplication, is relevant in a wide range of contexts, most notably phylogenetic tree inference and protein function annotation. In this chapter, we provide an overview of the methods used to infer orthology and paralogy. We survey both graph-based approaches (and their various grouping strategies) and tree-based approaches, which solve the more general problem of gene/species tree reconciliation. We discuss conceptual differences among the various orthology inference methods and databases and examine the difficult issue of verifying and benchmarking orthology predictions. Finally, we review typical applications of orthologous genes, groups, and reconciled trees and conclude with thoughts on future methodological developments.
Key words Orthology, Paralogy, Tree reconciliation, Orthology benchmarking
#### 1 Introduction
The study of genetic material almost always starts with identifying, within or across species, homologous regions—regions of common ancestry. As we have seen in previous chapters, this can be done at the level of genome segments [1], genes [2], or even down to single residues, in sequence alignments [3]. Here, we focus on genes as evolutionary and functional units. The central premise of this chapter is that it is useful to distinguish between two classes of homologous genes: orthologs, which are pairs of genes that started diverging via evolutionary speciation, and paralogs, which are pairs of genes that started diverging via gene duplication [4] (Fig. 1, Box 1). Originally, the terms and their definition were proposed by Walter M. Fitch in the context of species phylogeny inference, i.e., the reconstruction of the tree of life. He stated "Phylogenies require orthologous, not paralogous, genes" [4]. Indeed, since orthologs arise by speciation, any set of genes in which every pair is orthologous has by definition the same evolutionary history as the
Adrian M. Altenhoff and Natasha M. Glover are the Joint first authors
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_5, © The Author(s) 2019
Fig. 1 (a) Simple evolutionary scenario of a gene family with two speciation events (S<sup>1</sup> and S2) and one duplication event (star). The type of events completely and unambiguously define all pairs of orthologs and paralogs: The frog gene is orthologous to all other genes (they coalesce at S1). The red and blue genes are orthologs between themselves (they coalesce at S2), but paralogs between each other (they coalesce at star). (b) The corresponding orthology graph. The genes are represented here by vertices and orthology relationships by edges. The frog gene forms one-to-many orthology with both the human and dog genes, because it is orthologous to more than one sequence in each of these organisms. In such cases, the bi-directional best-hit approach only recovers one of the relations (the highest scoring one). Note that in contrary to BBH, the nonsymmetric BeTs approach—simply taking the best genome-wide hit for each gene regardless of reciprocity—would in the situation of a lost blue human gene infer an incorrect orthologous relation between the blue dog and red human gene
underlying species. These days, however, the most frequent motivation for the orthology/paralogy distinction is to study and predict gene function: it is generally believed that orthologs—because they were the same gene in the last common ancestor of the species involved—are likely to have similar biological function. By contrast, paralogs—because they result from duplicated genes that have been retained, at least partly, over the course of evolution—are believed to often differ in function. Consequently, orthologs are of interest to infer function computationally, while paralogs are commonly used to study function innovation.
#### Box 1: Terminology
Homology is a relation between a pair of genes that share a common ancestor. All pairs of genes in the below figure are homologous to each other.
(continued)
Orthology is a relation defined over a pair of homologous genes, where the two genes have emerged through a speciation event [4]. Example pairs of orthologs are (x1, y1) or (x2, z1). Orthologs can be further subclassified into one-to-one, one-to-many, many-to-one, and many-to-many orthologs. The qualifiers one and many indicate for each of the two involved genes whether they underwent an additional duplication after the speciation between the two genomes. Hence, the gene pair (x1, y1) is an example of a one-to-one orthologous pair, whereas (x2, z1) is a many-to-one ortholog relation.
Paralogy is a relation defined over a pair of homologous genes that have emerged through a gene duplication, e.g., (x1, x2) or (x1, y2).
In-Paralogy is a relation defined over a triplet. It involves a pair of genes and a speciation event of reference. A gene pair is an in-paralog if they are paralogs and duplicated after the speciation event of reference [5]. The pair (x1, y2) are in-paralogs with respect to the speciation event S1.
Out-Paralogy is also a relation defined over a pair of genes and a speciation event of reference. This pair is out-paralogs if the duplication event through which they are related to each other predates the speciation event of reference. Hence, the pair (x1, y2) are out-paralogs with respect to the speciation event S2.
Co-orthology is a relation defined over three genes, where two of them are in-paralogs with respect to the speciation event associated to the third gene. The two in-paralogous genes are said to be co-orthologous to the third (out-group) gene. Thus, x<sup>1</sup> and y<sup>2</sup> are co-orthologs with respect to z1.
Homoeology is a specific type of homologous relation in a polyploid species, which thus contain multiple "sub-genomes." This relation describes pairs of genes that originated by speciation and were brought back together in the same genome by allopolyploidization (hybridization) [6]. Thus, in the absence of rearrangement, homoeologs can be thought of as orthologs between sub-genomes.
In this chapter, we first review the main methods used to infer orthology and paralogy, including recent techniques for scaling up algorithms to big data. We then discuss the problem of benchmarking orthology inference. In the last main section, we focus on various applications of orthology and paralogy.
#### 2 Inferring Orthology
Most orthology inference methods can be classified into two major types: graph-based methods and tree-based methods [7]. Methods of the first type rely on graphs with genes (or proteins) as nodes and evolutionary relationships as edges. They infer whether these edges represent orthology or paralogy and build clusters of genes on the basis of the graph. Methods of the second type are based on gene/ species tree reconciliation, which is the process of annotating all splits of a given gene tree as duplication or speciation, given the phylogeny of the relevant species. From the reconciled tree, it is trivial to derive all pairs of orthologous and paralogous genes. All pairs of genes which coalesce in a speciation node are orthologs and paralogs if they split at a duplication node. In this section, we present the concepts and methods associated with the two types and discuss the advantages, limitations, and challenges associated with them.
2.1 Graph-Based Methods Graph-based approaches were originally motivated by the availability of complete genome sequences and the need for efficient methods to detect orthology. They typically run in two phases: a graph construction phase, in which pairs of orthologous genes are inferred (implicitly or explicitly) and connected by edges, and a clustering phase, in which groups of orthologous genes are constructed based on the structure of the graph.
2.1.1 Graph Construction Phase: Orthology Inference In its most basic form, the graph construction phase identifies orthologous genes by considering pairs of genomes at a time. The main idea is that between any given two genomes, the orthologs tend to be the homologs that diverged least. Why? Because assuming that speciation and duplication are the only types of branching events, the orthologs branched by definition at the latest possible time point—the speciation between the two genomes in question. Therefore, using sequence similarity score as surrogate measure of closeness, the basic approach identifies the corresponding ortholog of each gene through its genome-wide best hit (BeT)—the highest scoring match in the other genome [8]. To make the inference symmetric (as orthology is a symmetric relation), it is usually required that BeTs be reciprocal, i.e., that orthology be inferred for a pair of genes g1 and g2 if and only if g2 is the BeT of g1 and g1 is the BeT of g2 [9]. This symmetric variant, referred to as bi-directional best hit (BBH), has also the merit of being more robust against a possible gene loss in one of the two lineages (Fig. 1).
> Inferring orthology from BBH is computationally efficient, because each genome pair can be processed independently and high-scoring alignments can be computed efficiently using dynamic programming [10] or heuristics such as BLAST [11]. Overall, the
time complexity scales quadratically in terms of the total number of genes (Box 2). Furthermore, the implementation of this kind of algorithm is simple.
#### Box 2: Computational Considerations for Scaling to Many Genomes
Time complexity—the amount of time for an algorithm to run as a function of the input—is an important consideration when dealing with big data. This is relevant for inferring orthologs and paralogs due to the massive amounts of sequence data. Thus, it is necessary to consider the time complexity of the inference algorithms, especially when scaling for large and multiple genomes. In computer science, this is commonly denoted in terms of "Big O" notation, which expresses the scaling behavior of the algorithm, up to a constant factor. Below are listed the common time complexities for aspects of some orthology inference algorithms, in order of most efficient to least efficient.
#### Linear time
<sup>l</sup> O(n): Optimal algorithm to reconcile rooted, fully resolved gene tree and species tree [12]; Hieranoid algorithm, which recursively merges genomes along the species tree to avoid all-against-all computation [13].
Quadratic time
<sup>l</sup> O(n<sup>2</sup> ): The all-against-all stage central to many orthology algorithms scales quadratically, where n is total number of genes.
#### Cubic time
<sup>l</sup> O(n<sup>3</sup> ): The COG database's graph-based clustering merge triplets of homologs which share a common face until no more can be added.
#### NP-complete
<sup>l</sup> "Nondeterministic polynomial time," a large class of algorithms for which no solution in polynomial time is known, (e.g. scaling exponentially with respect to the input size), and thus are impractical. NP-complete problems are typically solved approximately, using heuristics. For instance, maximum likelihood gene tree estimation is NP-complete [14].
However, orthology inference by BBH has several limitations, which motivated the development of various improvements (Table 1).
Overview of graph-based orthology inference methods and their main properties
Allowing for More Than One Ortholog Some genes can have more than one orthologous counterpart in a given genome. This happens whenever a gene undergoes duplication after the speciation of the two genomes in question. Since BBH only picks the best hit, it only captures part of the orthologous relations (Fig. 1). The existence of multiple orthologous counterparts is often referred to as one-to-many or many-to-many orthology, depending whether duplication took place in one or both lineages. To designate the copies resulting from such duplications occurring after a speciation of reference, Remm et al. coined the term in-paralogs and introduced a method called InParanoid that improves upon BBH by potentially identifying all pairs of many-to-many orthologs [5]. In brief, their algorithm identifies all paralogs within a species that are evolutionarily closer (more similar) to each other than to the BBH gene in the other genome. This results in two sets of in-paralogs—one for each species—where all pairwise combinations between the two sets are orthologous relations. Alternatively, it is possible to identify many-to-many orthology by relaxing the notion of "best hit" to "group of best hits." This can be implemented using a score tolerance threshold or a confidence interval around the BBH [23, 34].
Evolutionary Distances Instead of using sequence similarity as a surrogate for evolutionary distance to identify the closest gene(s), Wall et al. proposed to use direct and proper maximum likelihood estimates of the evolutionary distance between pairs of sequences [31]. This estimate of evolutionary distance is based on the number and type of amino acid substitutions between the two sequences. Indeed, previous studies have shown that the highest scoring alignment is often not the nearest phylogenetic neighbor [35]. Building upon this work, Roth et al. showed how statistical uncertainties in the distance estimation can be incorporated into the inference strategy [36].
Differential Gene Losses As discussed above, one of the advantages of BBH over BeT is that by virtue of the bi-directional requirement, the former is more robust to gene losses in one of the two lineages. But if gene losses occurred along both lineages, it can happen that a pair of genes mutually closest to one another is in fact paralogs, simply because both their corresponding orthologs were lost—a situation referred to as "differential gene losses." Dessimoz et al. [37] presented a way to detect some of these cases by looking for a third species in which the corresponding orthologs have not been lost and thus can act as witnesses of non-orthology.
2.1.2 Clustering Phase: From Pairs to Groups The graph construction phase yields orthologous relationships between pairs of genes. But this is often not sufficient. Conceptually, information obtained from multiple genes or organisms is often more powerful than that obtained from pairwise comparisons only. In particular, as the use of a third genome as potential witness of non-orthology suggests, a more global view can allow identification and correction of inconsistent/spurious predictions. Practically, it is more intuitive and convenient to work with groups of genes than with a list of gene pairs. Therefore, it is often desirable to cluster orthologous genes into groups.
Tatusov et al. [8] introduced the concept of clusters of orthologous groups (COGs). COGs are computed by using triangles (triplets of genes connected to each other) as seeds and then merging triangles which share a common face, until no more triangle can be added. This clustering can be computed relatively efficient in time O(n3 ), where n is the number of genomes analyzed [38]. The stated objective of this clustering procedure is to group genes that have diverged from a single gene in the last common ancestor of the species represented [8]. Practically, they have been found to be useful by many, most notably to categorize prokaryotic genes into broad functional categories.
A different clustering approach was adopted by OrthoMCL, another well-established graph-based orthology inference method [29]. There, groups of orthologs are identified by Markov Clustering [39]. In essence, the method consists in simulating a random walk on the orthology graph, where the edges are weighted according to similarity scores. The Markov Clustering process gives rise to probabilities that two genes belong to the same cluster. The graph is then partitioned according to these probabilities and members of each partition form an orthologous group. These groups contain orthologs and "recent" paralogous genes, where the recency of the paralogs can be somewhat controlled through the parameters of the clustering process.
A third grouping strategy consists in building groups by identifying fully connected subgraphs (called "cliques" in graph theory) [23]. This approach has the merits of straightforward interpretation (groups of genes which are all orthologous to one another) and high confidence in terms of orthology within the resulting groups, due to the high consistency required to form a fully connected subgraph. But it has the drawbacks of being hard to compute (clique finding belongs to the NP-complete class of problems, for which no polynomial time algorithm is known; see Box 2) and being excessively conservative for many applications.
As emerges from these various strategies, there is more than one way orthologous groups can be defined, each with different implications in terms of group properties and applications [40]. In fact, there is an inherent trade-off in partitioning the orthology graph into clusters of genes, because orthology is a non-transitive relation: if genes A and B are orthologs and genes B and C are orthologs, genes A and C are not necessarily orthologs, e.g., consider in Fig. 1 the blue human gene, the frog gene, and the red dog gene. Therefore, if groups are defined as sets of genes in which all pairs of genes are orthologs (as with OMA groups), it is not possible to partition A, B, and C into groups capturing all orthologous relations while leaving out all paralogous relations.
2.1.3 Hierarchical More inclusive grouping strategies necessarily lead to orthologs and paralogs within the same group. Nevertheless, it can be possible to control the nature of the paralogs included. For instance, as seen above, OrthoMCL attempts at including only "recent" paralogs in its groups. This idea can be specified more precisely by defining groups with respect to a particular speciation event of interest, e.g., the base of the mammals. Such hierarchical groups are expected to include orthologs and in-paralogs with respect to the reference speciation—in our example all copies that have descended from a single common ancestor gene in the last mammalian common ancestor. Conceptually, hierarchical orthologous groups can be defined as groups of genes that have descended from a single common ancestral gene within a taxonomic range of interest.
Clustering
Several resources provide hierarchical clustering of orthologous groups. EggNOG [15] and OrthoDB [25], for example, both implement this concept by applying a COG-like clustering method for various taxonomic ranges. Another example, Hieranoid, produces hierarchical groups by using a guide tree to perform pairwise orthology inferences at each node from the leaves to the root inferring ancestral genomes at each node in the tree [13, 18]. Similarly, OMA GETHOGs is an approach based on an orthology graph of pairwise orthologous gene relations, where hierarchical orthologous groups are formed starting with the most specific taxonomy and incrementally merges them toward the root [21, 22]. Another method, COCO-CL, identifies hierarchical orthologous groups recursively, using correlations of similarity scores among homologous genes [41] and, interestingly, without relying on a species tree. By capturing part of the gene tree structure in the group hierarchies, these methods try in some way to bridge the gap between graph-based and tree-based orthology inference approaches. We now turn our attention to the latter.
2.2 Tree-Based Methods At their core, tree-based methods infer orthologs on the basis of gene family trees whose internal nodes are labeled as speciation or duplication nodes. Indeed, once all nodes of the gene tree have been inferred as a speciation or duplication event, it is trivial to establish whether a pair of genes is orthologous or paralogous, based on the type of the branching where they coalesce. Such labeling is traditionally obtained by reconciling gene and species trees. In most cases, gene and species trees have different topologies, due to evolutionary events acting specifically on genes such as duplications, losses, lateral transfers, or incomplete lineage sorting [42]. Goodman et al. [43] pioneered research to resolve these
Fig. 2 Schematic example of the gene/species tree reconciliation. The gene tree and species tree are not compatible. Reconciliation methods resolve the incongruence between the two by inferring speciation, duplication, and losses events on the gene tree. The reconciled tree indicates the most parsimonious history of this gene, constrained to the species tree. The simple representation (bottom right) suggests that the human and frog genes are orthologs and that they are both paralogous to the dog gene
incongruences. They showed how the incongruences can be explained in terms of speciation, duplication, and loss events on the gene tree (Fig. 2) and provided an algorithm to infer such events.
Most tree reconciliation methods rely on a parsimony criterion: the most likely reconciliation is the one which requires the least number of gene duplications and losses. This makes it possible to compute reconciliation efficiently and is tenable as long as duplication and loss events are rare compared to speciation events. In their seminal article, Goodman et al. [43] had already devised their reconciliation algorithm under a parsimony strategy. In the subsequent years, the problem was formalized in terms of a map function between the gene and species trees [44], whose computational cost was conjectured [45], and later proved [12, 46] to coincide with the number of gene duplication and losses. These results yielded highly efficient algorithms, either in terms of asymptotic time complexity [12] or in terms of runtimes on typical problem sizes [47]. With these near-optimal solutions, one might think that the tree reconciliation problem has long been solved. As we shall see in the rest of this section, however, the original formulation of the tree reconciliation problem has several limitations in practice, which have stimulated the development of various refinements to overcome them (Table 2).
Overview of gene/species tree reconciliation methods and their main properties
Table 2 bApproach to root gene tree (n.a. indicates that the initial rooting is assumed to be correct)
cApproach taken to handle reconstruction uncertainties of the gene tree (bootstrap, reconcile every bootstrap sample; multifurcate, splits in the gene tree with low supportcollapsed)
dUsed optimization framework (MP, maximum parsimony)
are 2.2.1 Unresolved Species Tree
A first problem ignored by most early reconciliation algorithms lies in the uncertainty often associated with the species tree, which these methods assume as correct and heavily rely upon.
One way of dealing with the uncertainties is to treat unresolved parts of the species tree as multifurcating nodes (also known as soft polytomies). By doing so, the reconciliation algorithm is not forced to choose for a specific type of evolutionary event in ambiguous regions of the tree. This approach is, for instance, implemented in TreeBeST [52] and used in the Ensembl Compara project [53].
Alternatively, Heijden et al. [57] demonstrated that it is often possible to infer speciation and duplication events on a gene tree without knowledge of the species tree. Their approach, which they call species overlap, identifies for a given split the species represented in the two subtrees induced by the split. If at least one species has genes in both subtrees, a duplication event is inferred; else a speciation event is inferred. In fact, this approach is a special case of soft polytomies where all internal nodes have been collapsed. Thus, the only information needed for this approach is a rooted gene tree. Since then, this approach has been adopted in other projects, such as PhylomeDB [59].
2.2.2 Rooting The classical reconciliation formulation requires both gene and species trees to be rooted. But most models of sequence evolution are time reversible and thus do not allow to infer the rooting of the reconstructed gene tree. One sensible solution is to root a gene tree so that it minimizes the number of duplication events [62]. Thus, this method uses the parsimony principle for both rooting and reconciliation. For cases of multiple optimal rootings, ties can be broken by selecting the tree that minimizes the tree height [63] or by picking the rooting which minimizes the number of gene losses [61].
> Another approach is to place the root at the "center of the tree"—also known as "midpoint rooting" [58]. The idea of this method goes back to Farris [64] and is motivated by the concept of a molecular clock. But for most gene families, assuming a constant rate of evolution is inappropriate [65, 66], and thus this approach is not used widely. A newly introduced refinement based on minimizing average deviations among children nodes holds promise of being more robust [67] but still relies on a molecular clock assumption.
> For the species tree, the most common and reliable way of rooting trees is by identifying an outgroup species. PhylomeDB uses genes from outgroup species to root gene trees [59]. One main potential problem with this approach is that in many situations, it can be difficult to identify a suitable outgroup. For example, in analysis covering all kingdoms of life, an outgroup species may not be available, or the relevant genes might have been lost
[68]. A suitable out-group needs to be close enough to allow for reliable sequence alignment, yet it must have speciated clearly before any other species separated. Furthermore, ancient duplications can cause outgroup species to carry in-group genes. These difficulties make this approach more challenging for automated, large-scale analysis [69].
2.2.3 Gene Tree Uncertainty Another assumption made in the original tree reconciliation problem is the (topological) correctness of the gene tree. But it has been shown that this assumption is commonly violated, often due to finite sequence lengths, taxon sampling [70, 71], or gene evolution model violations [72]. On the other hand, techniques of expressing uncertainties in gene tree reconstruction via support measures, e.g., bootstrap values, have become well established. Storm and Sonnhammer [58] as well as Zmasek and Eddy [63] independently suggested to extend the bootstrap procedure to reconciliation, thereby reducing the dependency of the reconciliation procedure on any one gene tree while providing a measure of support of the inferred speciation/duplication events. The downsides of using the bootstrap are the high computational costs and interpretation difficulties associated with it [73].
> Similarly to how unresolved species tree can be handled, unresolved parts of the gene tree can also be collapsed into multifurcating nodes. For instance, HOGENOM [55] and Softparsmap [61] collapse branches with low bootstrap support values.
> A third way of tackling this problem consists in simultaneously solving both the gene tree reconstruction and reconciliation problems [74]. They use the parsimony criterion of minimizing the number of duplication events to improve on the gene tree itself. This is achieved by rearranging the local gene tree topology of regions with low bootstrap support such that the number of duplications and losses is further reduced.
All the approaches mentioned so far try to minimize the number of gene duplication events. This is generally justified by a parsimony argument, which assumes that gene duplications and losses are rare events. But what if this assumption is frequently violated? Little is known about duplication and loss rates in general [75], but there is strong evidence for historical periods with high gene duplication occurrence rates [76] or gene families specifically prone to massive duplications (e.g., olfactory receptor, opsins, serine/threonine kinases, etc.)
Motivated by this reasoning, Arvestad et al. introduced the idea of a probabilistic model for tree reconciliation [49]. They used a Bayesian approach to estimate the posterior probabilities of a reconciliation between a given gene and species tree using Markov chain Monte Carlo (MCMC) techniques. Arvestad et al. [49]
2.2.4 Parsimony vs. Likelihood
modeled gene duplication and loss events through a birth-death process [77]. In the subsequent years, they refined their method to also model sequence evolution and substitution rates in a unified framework called gene sequence evolution model with iid rates (GSR) [49, 50].
Perhaps the biggest problem with the probabilistic approach is that it is not clear how well the assumptions of their model (the birth-death process with fixed parameters) relate to the true process of gene duplication and gene loss. Doyon et al. [78] compared the maximum parsimony reconciliation trees from 1278 fungi gene families to the probabilistically reconciled trees using gene birth/ death rates fitted from the data. They found that in all but two cases, the maximum parsimony scenario corresponds to the most probable one. This remarkably high level of consistency indicates that in terms of the accuracy of the "best" reconciliation, there is little to gain from using a likelihood approach over the parsimony criterion of minimizing the number of duplication events. But how this result generalizes to other datasets has yet to be investigated.
2.3 Graph-Based vs. Tree-Based: Which Is Better? Given the two fundamentally different paradigms in orthology inference that we reviewed in this section, one can wonder which is better. Conceptually, tree reconciliation methods have several advantages. In terms of inference, by considering all sequences from all species at the same time, it can also be expected that they can extract more information from the sequences. This in turn should translate into higher statistical power. In terms of their output, reconciled gene trees provide the user more information than pairs or groups of orthologs. For example, the trees display the order of duplication and speciation events, as well as evolutionary distances between these events. In practice, however, these methods have the disadvantage of having much higher computational complexity than their graph-based counterparts. Furthermore, the two approaches are in practice often not that strictly separated. Tree-based methods often start with a graph-based clustering step to identify families of homologous genes. Conversely, several hierarchical grouping algorithms also rely on species trees in their inference.
Thus, it is difficult to make general statements about the relative performance of the two classes of inference methods. One solution that can leverage the unique abilities of both tree-based and graph-based methods is to combine several independent orthology inference methods into one. We discuss this technique in the next section.
#### 3 Meta-methods
In recent years a new class of orthology inference tools has emerged which attempts to make the most out of multiple orthology prediction algorithms—meta-methods. These are approaches which combine several individual and distinct methods in order to produce more robust orthology predictions. These meta-methods are able to take advantage of the standardized formats of output which has been a goal of the orthology community [79], as well as the many new and well-established methods out there.
Generally, meta-methods assign a confidence score to a given predicted orthologous relation. In its most basic form, more weight is given to orthologs predicted by the most methods. Some examples include methods which simply take the intersection of several methods, such as GET\_HOMOLOGUES [80], COMPARE [81], HCOP [82], and DIOPT [83]. These methods maintain a high level of precision, but since they are based on intersections, they necessarily have a lower recall.
Additionally, post-processing techniques can be used to build upon the base of orthologs found by several methods—thus assigning more sequences as orthologs and improving performance. For example, MOSAIC (Multiple Orthologous Sequence Analysis and Integration by Cluster optimization) [84] uses an iterative graphbased optimization approach that works on ortholog sets predicted by several independent methods. MOSAIC captures orthologs which are missed by some individual methods, producing a 1.6 fold increase in the number of orthologs detected. Another example is the MARIO software, which looks for the intersection of several different orthology methods as seed groups and then progressively adds unassigned proteins to the groups based on HMM profiles [85]. MetaPhOrs' approach integrates phylogenetic and homology information derived from different databases [86]. They demonstrate that the number of independent sources from which an orthology prediction is made, as well as the level of consistency across predictions, can be used as confidence scores.
So far the previously mentioned meta-methods combine independent orthology prediction algorithms and give a higher score based on the more algorithms which predict a given orthologous relation. However, another emerging approach is to use machine learning techniques to recognize patterns among several different orthology inference methods. With this, one can predict previously unknown high-confidence orthologs. WORMHOLE is a tool which uses the information from 17 different orthology prediction methods to train support vector machine classifiers for predicting least diverged orthologs [87]. WORMHOLE was able to strongly re-predict least diverged orthologs in the reference set and also predict previously unclassified orthologous genes.
The type of meta-approach and its associated stringency depends on what the user is going after. For example, if the goal is to get very-high-confidence groups, methods which only combine for the intersection without trying to add more orthologs may be preferable. Studies requiring both high precision and recall may be better suited to use the meta-methods which use postprocessing or machine learning to predict orthologs. And as with all methods, it is important to understand which clades the method has been benchmarked in and which orthology tools have been combined. For example, if several methods have the same bias, one will just propagate the bias and end up with a false sense of security because the methods are not independent.
#### 4 Scaling to Many Genomes
In terms of orthology inference, the abundance of genomes now available has resulted in an emphasis on driving down computational processing time via efficient algorithms. When inferring orthology for many genomes, the bottleneck is generally the allagainst-all computations—aligning the proteins in every genome against the proteins in every other genome. This is the first step of nearly all graph-based methods. The all-against-all computation has an O(n<sup>2</sup> ) runtime, meaning it scales quadratically with the number of genomes analyzed (Box 2).
So far, two main techniques for scaling orthology prediction to many genomes have emerged. The first approach is by making the all-against-all comparisons faster. Because comparisons are independent of each other, the most obvious way of doing this is by taking advantage of a high-performance computing cluster, as this is an embarrassingly parallel computing problem. Many methods have implemented this, such as Hieranoid [13], PorthoMCL [88], or OMA [22]. Another way to save time on the all-against-all comparisons is by using very fast algorithms for the homology search. For example, preliminary results of SonicParanoid showed 160–750 speedup of orthology inference compared to InParanoid [89]. Innovations in alignment algorithms with methods such as DIAMOND [90] or MMSeq2 [91] have the potential to greatly reduce the time to do the all-against-all comparisons.
A second approach to efficiently scale up orthology inference to many genomes is by simply avoiding doing the entire all-against-all comparisons. This makes sense, since a significant amount of time is spent comparing unrelated gene pairs. For example, it is possible to avoid aligning many unrelated pairs by exploiting the transitive property of homology. Wittwer et al. [92] did this by first building clusters of homologous sequences with one representative sequence per cluster and subsequently performing the all-againstall within each cluster. Hieranoid avoids unnecessary all-against-all comparisons by using a species tree as a guide, reducing the number of comparisons to N 1 for N genomes, scaling linearly rather than quadratically [18]. Another way to avoid all-by-all comparison is by using a mapping strategy, whereby new proteomes are mapped onto precomputed orthologous groups. This strategy has been successfully implemented with the eggNOG database—each sequence in a new proteome is mapped to a precomputed orthologous cluster based on hidden Markov models. Then, orthology relations and function are transferred to the new sequence from the best matching sequence in the database [93].
#### 5 Benchmarking Orthology
Assessing the quality of orthology predictions is important but difficult. The main challenge is that the precise evolutionary history of entire genomes is largely unknown and thus, predictions can only be validated indirectly, using surrogate measures. To be informative, such measures need to strongly correlate with orthology/ paralogy. At the same time, they should be independent from the methods used in the orthology inference process. Concretely, this means that the orthology inference is not based on the surrogate measure and the surrogate measure is not derived from orthology/ paralogy.
#### 5.1 Benchmarking Approaches Several ways of benchmarking orthology inference have been developed in the past years. In the next sections, we go over the main approaches, bringing attention to the advantages and limitations to each.
5.1.1 Functional Conservation The first surrogate measures proposed revolved around conservation of function [94]. This was motivated by the common belief that orthologs tend to have conserved function, while paralogs tend to have different functions. Indeed, orthologs tend to be more conserved than paralogs in terms of GO annotation similarity [95]. Thus, "for a given evolutionary distance, more accurate orthology inference is likely to be correlated with more functionally similar gene pairs." Hulsen et al. [94] assessed the quality of ortholog predictions in terms of conservation of co-expression levels, domain annotation, and protein-protein interaction partners. Additionally, Altenhoff et al. [96] used similarity of experimentally validated GO annotations as well as Enzyme Commission (EC) numbers as a functional benchmark. Functional benchmarks have an advantage in that many researchers are interested in orthology because they want to find functionally conserved genes, thus making functional tests important for assessing different inference methods. The main limitation of these measures is that it is not so clear how much they correlate with orthology/paralogy. Indeed, it has been argued that the difference in function conservation trends between orthologs and paralogs might be much smaller than commonly assumed and indeed many examples are known of orthologs that have dramatically different functions [97].
5.1.2 Gene Neighborhood Conservation The fraction of orthologs that have neighboring genes being orthologs themselves is an indicator of consistency and therefore to some extent also of quality of orthology predictions [94]. Although synteny has been used as part of the orthology inference for several algorithms, to date it has not been used as part of large-scale benchmarking efforts. One possible problem is that gene neighborhood can be conserved among paralogs, such as those resulting from whole-genome duplications. Furthermore, some methods use gene neighborhood conservation to help in their inference process, which can bias the assessment done on such measures (principle of independence stated above).
5.1.3 Species Tree Discordance Test The quality of ortholog predictions can also be assessed based on phylogeny. By definition, the tree relating a set of genes all orthologous to one another only contains speciation splits and has the same topology as the underlying species. We introduced a benchmarking protocol that quantifies how well the predictions from various orthology inference methods agree with undisputed species tree topologies [96, 98]. Thus, the species tree discordance test judges the accuracy of ortholog predictions based on the correctness of the species tree which can be constructed from them. The advantage of this measure is that by virtue of directly ensuing from the definition of orthology, it correlates strongly with it and thus satisfies the first principle. However, the second principle, independence from the inference process, is not satisfied with methods relying on the species tree—typically all reconciliation methods but also most graph-based methods producing hierarchical groups. In such cases, interpretation of the results must be done carefully.
5.1.4 Gold Standard Gene Tree Test High-quality reference gene trees can also be used to assess orthology inferences. For this, one compares the pairs of orthologs from a given method to pairs of orthologs derived from these expertly curated gene trees [40, 99]. One drawback of this benchmark is that it is limited by the ability to curate the phylogeny—if the evolutionary history of the gene family is ambiguous, the resulting reference tree will unavoidably have mistakes. Another limitation is the small size of most benchmarks of this type. This casts doubts on their generalizability and makes them prone to overfitting.
5.1.5 Subtree Consistency Test For inference methods based on reconciliation between gene and species trees, Vilella et al. [53] proposed a different phylogenybased assessment scheme. For any duplication node of the labeled gene tree, a consistency score is computed, which captures the balance of the species found in the two subtrees. Unbalanced nodes correspond to an evolutionary scenario involving extensive gene losses and therefore, under the principle of parsimony, are less likely to be correct. Given that studies to date tend to support the adequacy of the parsimony criterion in the context of gene family dynamics (Subheading 2.2.4), it can be expected that this metric correlates highly with correct orthology/paralogy assignments. However, since virtually all tree-based methods themselves incorporate this very criterion in their objective function (i.e., minimizing the number of gene duplications and losses), the principle of independence is violated, and thus the adequacy of this measure is questionable. 5.1.6 Latent Class Analysis Chen et al. [100] proposed a purely statistical benchmark based on latent class analysis (LCA). Given the absence of a definitive answer on whether two given genes are orthologs, the authors argue that by looking at the agreement and disagreement of predictions made by several inference methods on a common dataset, one can estimate the reliability of individual predictors. More precisely, LCA is a statistical technique that computes maximum likelihood estimates of sensitivity and specificity rates for each orthology inference methods, given their predictions and given an error model. This is attractive, because it does not depend on any surrogate measure. However, the results depend on the error model assumed. Thus, we are of the opinion that LCA merely shifts the problem of assessing orthology to the problem of assessing an error model of various orthology inference methods. 5.1.7 Simulated Genomes Finally, simulated data can be used in benchmarking. By this, the precise evolutionary history of a genome can be validated, in terms of gene duplication, insertion, deletion, and lateral gene transfer [101]. Knowing for certain all aspects of the simulated genomes gives an advantage over assessments based on empirical data, where the true evolutionary history is unknown. On the other hand, how well the simulated data reflect "real" data is debatable. 5.2 Orthology Benchmarking Service The orthology benchmarking service is a web-based platform for which users can upload their ortholog predictions and run them through a variety of benchmarks. The user must use quest for orthologs (QFO) reference proteome set, which is a set of 66 genomes that covers a diverse set of species across all domains [79], to
infer pairwise or groups of orthologs. Several phylogenetic and function-based benchmarks are automatically run on the uploaded data, and then summary statistics of the results of each benchmark
#### 6 Applications
As we have seen so far, there is a large diversity in the methods for orthology inference. The main reason is that, although the methods discussed here all infer orthology as part of their process, many of them have been developed for different reasons and have different ultimate goals. Unfortunately, this is often not mentioned explicitly and tends to be a source of confusion. In this section, we review some of these ultimate goals and discuss which methods and representation of orthology are better suited to address them and why.
As mentioned in the introduction, most interest for orthology is in the context of function prediction and is largely based on the belief that orthologs tend to have conserved function. A conservative approach consists in propagating function between one-to-one orthologs, i.e., pairs of orthologous genes that have not undergone gene duplication since they diverged from one another. Several orthology databases directly provide one-to-one orthology predictions. But even with those that do not, it might still be possible to obtain such predictions, for instance, by selecting hierarchical groups containing at most one sequence in each species or by extracting from reconciled trees' subtrees with no duplication. A more sophisticated approach consists in propagating gene function annotations across genomes on the basis of the full reconciled gene tree. Thomas et al. [102], for instance, proposed a way to assign gene function to uncharacterized proteins using a gene tree and a hidden Markov model (HMM) among gene families. Engelhardt et al.. [103] developed a Bayesian model of function change along reconciled gene trees and showed that their approach significantly improves upon several methods based on pairwise gene function propagation. Ensembl Compara [53] and Panther [102] are two major databases providing reconciled gene trees.
Since Darwin, one traditional question in biology has always been how species are related to each other. As we recall in the introduction of this chapter, Fitch's original motivation for defining orthology was phylogenetic inference. Indeed, the gene tree reconstructed from a set of genes which are all orthologous to each other should by definition be congruent to the species tree. OMA Groups (OMA) have this characteristic and, crucially, are constructed without help of a species tree.
Yet another application associated with orthology are general alignments between genomes, e.g., protein-protein interaction (PPI) network alignments or whole-genome alignments. Finding an optimal PPI network alignment between two genomes on the basis of the network topology alone is a computationally hard problem (i.e., it is an instance of the subgraph isomorphism problem which is NP-complete [104]). Orthology is often used as heuristic to constrain the mapping of the corresponding genes between the two networks and thus to reduce the problem complexity of aligning networks [105]. For whole-genome alignments, people most often use homologous regions and use orthologs as anchor points [106]. These types of application typically rely on ortholog predictions between pairs of genomes, as provided, e.g., by InParanoid [5] or OMA [23].
#### 7 Conclusions and Outlook
The distinction between orthologs and paralogs is at the heart of many comparative genomic studies and applications. The original and generally accepted definition of orthology is based on the evolutionary history of pairs of genes. By contrast, there is a considerable diversity in how groups of orthologs are defined. These differences largely stem from the fact that orthology is a non-transitive relation and therefore, dividing genes into orthologous groups will either miss or wrongly include orthologous relations. This makes it important and worthwhile to identify the type of orthologous group best suited for a given application.
Regarding inference methods, while most approaches can be ordered into two fundamental paradigms—graph-based and treebased—the difference between the two is shrinking, with graphbased methods increasingly striving to capture more of the evolutionary history. On the other hand, the rapid pace at which new genomes are sequenced limits the applicability of tree-based methods, computationally more demanding.
Benchmarking this large variety of methods remains a hard problem—from a conceptual point as described above but also because of very practical challenges such as heterogeneous data formats, genome versions, or gene identifiers. This has been recognized by the research community and has led to the development of the QFO consortium benchmarking service [96].
Looking forward, we see potential in extending the current model of gene evolution, which is limited to speciation, duplication, and loss events. Indeed, nature is often much more complicated. For instance, lateral gene transfer (LGT) is believed to be a major mode of evolution in prokaryotes. While there has been several attempts at extending tree reconciliation algorithms to detecting LGT [107, 108], the problem is largely unaddressed in typical orthology resources [109]. Another relevant evolutionary process omitted by most methods is whole-genome duplications (WGD). Even though WGD events act jointly on all gene families, with few exceptions [110, 111], most methods consider each gene family independently.
Overall, the orthology/paralogy dichotomy has proved to be useful but also inherently limited. Reducing the whole evolutionary history of homologous genes into binary pairwise relations is bound to be a simplification—and at times an oversimplification. The shift toward hierarchical orthologous groups is thus a promising step toward capturing more features of the evolutionary history of genes. Yet further development will still be needed, as we are nowhere close to grasp the formidable complexity of gene evolution across the full diversity of life.
#### 8 Exercises
Assume the following evolutionary scenario
where duplications are depicted as star and all other splits are speciations.
connected subgraphs (cliques) that can be found. Second, reconstruct the COGs. COGs are built by merging triangles of orthologs whenever they share a common face. Remember that in both methods, a gene can only belong to a one cluster.
#### Acknowledgments
We thank Stefan Zoller for helpful feedback on an earlier version of the manuscript. We gratefully acknowledge support by the Swiss National Science Foundation grant PP00P3\_150654 to CD. Adrian M. Altenhoff and Natasha M. Glover contributed equally to this work.
#### References
genomics for over 1800 genomes. Bioinformatics 28:715–716
Proceedings of the fourth annual international conference on computational molecular biology. ACM, New York, NY, pp 138–146
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
#### Transposable Elements: Classification, Identification, and Their Use As a Tool For Comparative Genomics
Wojciech Makałowski, Valer Gotea, Amit Pande, and Izabela Makałowska
#### Abstract
Most genomes are populated by hundreds of thousands of sequences originated from mobile elements. On the one hand, these sequences present a real challenge in the process of genome analysis and annotation. On the other hand, they are very interesting biological subjects involved in many cellular processes. Here we present an overview of transposable elements biodiversity, and we discuss different approaches to transposable elements detection and analyses.
Key words Transposable elements, Transposons, Mobile elements, Repetitive elements, Genome analysis, Genome evolution
#### 1 Introduction
Most eukaryotic genomes contain large numbers of repetitive sequences. This phenomenon was described by Waring and Britten a half century ago using reassociation studies [1, 2]. It turned out that most of these repetitive sequences originated in transposable elements (TEs) [3], though the repetitive fraction of a genome varies significantly between different organisms, from 12% in Caenorhabditis elegans [4] to 50% in mammals [3], and more than 80% in some plants [5]. With such large contributions to genome sequences, it is not surprising that TEs have a significant influence on the genome organization and evolution. Although much progress has been achieved in understanding the role TEs play in a host genome, we are still far from the comprehensive picture of the delicate evolutionary interplay between a host genome and the invaders. They also pose various challenges to the genomic community, including aspects related to their detection and classification, genome assembly and annotation, genome comparisons, and mapping of genomic variants. They also pose various challenges to the genomic community, including aspects related to their detection and classification, genome assembly and annotation, genome
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_6, © The Author(s) 2019
comparisons, and mapping of genomic variants. Here we present an overview of TE diversity and discuss major techniques used in their analyses.
#### 2 Discovery of Mobile Elements
Transposable elements were discovered by Barbara McClintock during experiments conducted in 1944 on maize. Since they appeared to influence phenotypic traits, she named them controlling elements. However, her discovery was met with less than enthusiastic reception by the genetic community. Her presentation at the 1951 Cold Spring Harbor Symposium was not understood and at least not very well received [6]. She had no better luck with her follow-up publications [7–9] and after several years of frustration decided not to publish on the subject for the next two decades. Not for the first time in the history of science, an unappreciated discovery was brought back to life after some other discovery has been made. In this case it was the discovery of insertion sequences (IS) in bacteria by Szybalski group in the early 1970s [10]. In the original paper they wrote: "Genetic elements were found in higher organisms which appear to be readily transposed from one to another site in the genome. Such elements, identifiable by their controlling functions, were described by McClintock in maize. It is possible that they might be somehow analogous to the presently studied IS insertions" [10]. The importance of McClintock's original work was eventually appreciated by the genetic community with numerous awards, including 14 honorary doctoral degrees and a Nobel Prize in 1983 "for her discovery of mobile genetic elements" (http://nobelprize.org/nobel\_ prizes/medicine/laureates/1983/).
Coincidently, at the same time as Szybalski "rediscovered" TEs, Susumu Ohno popularized the term junk DNA that influenced genomic field for decades [11], although the term itself was used already before [12, 13].1 Ohno referred to the so-called noncoding sequences or, to be more precise, to any piece of DNA that do not code for a protein, which included all genomic pieces originated in transposons. The unfavorable picture of transposable and transposed elements started to change in early 1990s when some researchers noticed evolutionary value of these elements [14, 15]. With the wheel of fortune turning full circle and advances of genome sciences, TE research is again focused on the role of mobile elements played in the evolution of gene regulation [16–23].
<sup>1</sup>The historical background of the "junk DNA" term was recently discussed by Dan Graur in his excellent blog http://judgestarling.tumblr.com/post/64504735261/the-origin-of-the-term-junk-dna-a-historical
#### 3 Transposons Classification
#### 3.1 Insertion Sequences and Other Bacterial Transposons
The bacterial genome is composed of a core genomic backbone decorated with a variety of multifarious functional elements. These include mobile genetic elements (MGEs) such as bacteriophages, conjugative transposons, integrons, unit transposons, composite transposons, and insertion sequences (IS). Here we elaborate upon the last class of these elements as they are most widely found and described [24].
The ISs were identified during studies of model genetic systems by virtue of their capacity to generate mutations as a result of their translocation [10]. In-depth studies in antibiotic resistance and transmissible plasmids revealed an important role for these mobile elements in formation of resistance genes and promoting gene capture. In particular, it was observed that several different elements were often clustered in "islands" within plasmid genomes and served to promote plasmid integration and excision.
Although these elements sometimes generate beneficial mutations, they may be considered genomic parasites as ISs code only for the enzyme required for their own transposition [24]. While an IS element occupies a chromosomal location, it is inherited along with its host's native genes, so its fitness is closely tied to that of its host. Consequently, ISs causing deleterious mutations that disrupt a genomic mode or function are quickly eliminated from the population. However, intergenically placed ISs have a higher chance to be fixed in the population as they are likely neutral regarding population's fitness [25].
ISs are generally compact (Fig. 1). They usually carry no other functions than those involved in their mobility. These elements contain recombinationally active sequences which define the boundary of the element, together with Tpase, an enzyme, which processes these ends and whose gene usually encompasses the entire length of the element [26]. Majority of ISs exhibit short terminal inverted-repeat sequences (IR) of length 10–40 bp. Several notable exceptions do exist, for example, the IS91, IS110, and IS200/605 families.
The IRs contain two functional domains [27]. One is involved in Tpase binding; the other cleaves and transfers strand-specific reactions resulting in transposition. IS promoters are often positioned partially within the IR sequence upstream of the Tpase gene. Binding sites for host-specific proteins are often located within proximity to the terminal IRs and play a role in modulating transposition activity or Tpase expression [28]. A general pattern for the functional organization of Tpases has emerged from the limited numbers analyzed. The N-terminal region contains sequencespecific DNA binding activities of the proteins while the catalytic domain is often localized toward the C-terminal end [28].
Fig. 1 Schematic representation of insertion sequences (IS). dr direct repeats, IR inverted repeats, ORF open reading frame
The LTR retrotransposons are characterized by the presence of long terminal repeats (LTRs) ranging from several hundred to several thousand base pairs. Both exogenous retroviruses and LTR retrotransposons contain a gag gene that encodes a viral
Table 1 Prokaryotic transposable elements as presented in the IS Finder database [31]
a Presence (Y) or absence (N) of terminal inverted repeats
#### Table 2
#### Classification of eukaryotic transposable elements as proposed by Wicker et al. [33]
Please note that SVAs and retrogenes are not included in that classification a Not included in the original Wicker classification
> particle coat and a pol gene that encodes a reverse transcriptase, ribonuclease H, and an integrase, which provide the enzymatic machinery for reverse transcription and integration into the host genome. Reverse transcription occurs within the viral or viral-like particle (GAG) in the cytoplasm, and it is a multistep process [34]. Unlike LTR retrotransposons, exogenous retroviruses contain an env gene, which encodes an envelope that facilitates their migration to other cells. Some LTR retrotransposons may contain remnants of an env gene, but their insertion capabilities are limited to the originating genome [35]. This would rather suggest that they originated in exogenous retroviruses by losing the env gene. However, there is evidence that suggests the contrary, given that
Fig. 2 Structures of eukaryotic mobile elements. See text for detailed discussion
LTR retrotransposons can acquire the env gene and become infectious entities [36]. Presently, most of the LTR sequences (85%) in the human genome are found only as isolated LTRs, with the internal sequence being lost most likely due to homologous recombination between flanking LTRs [37]. Interestingly, LTR retrotransposons target their reinsertion to specific genomic sites, often around genes, with putative important functional implications for a host gene [35]. Lander et al. estimated that 450,000 LTR copies make up about 8% of our genome [38]. LTR retrotransposons inhabiting large genomes, such as maize, wheat, or barley, can contain thousands of families. However, despite the diversity, very few families comprise most of the repetitive fraction in these large genomes. Notable examples are Angela (wheat) [39], BARE1 (barley) [40], Opie (maize) [41], and Retrosor6 (sorghum) [42].
The DIRS order clusters structurally diverged group of transposons that possess a tyrosine recombinase (YR) gene instead of an integrase (INT) and do not form target site duplications (TSDs). Their termini resemble either split direct repeats (SDR) or inverted repeats. Such features indicate a different integration mechanism than that of other class I mobile elements. DIRS were discovered in the slime mold (Dictyostelium discoideum) genome in the early 1980s [43], and they are present in all major phylogenetic lineages including vertebrates [44]. It has been showed that they are also common in hydrothermal vent organisms [45].
Another order, termed Penelope-like elements (PLE), has wide, though patchy distribution from amoebae and fungi to vertebrates with copy number up to thousands per genome [46]. Interestingly, no PLE sequences have been found in mammalian genomes, and apparently they were lost from the genome of C. elegans [47]. Although PLEs with an intact ORF have been found in several genomes, including Ciona and Danio, the only transcriptionally active representative, Penelope, is known from Drosophila virilis. It causes the hybrid dysgenesis syndrome characterized by simultaneous mobilization of several unrelated TE families in the progeny of dysgenic crosses. It seems that Penelope invaded D. virilis quite recently, and its invasive potential was demonstrated in D. melanogaster [46]. PLEs harbor a single ORF that codes for a protein containing reverse transcriptase (RT) and endonuclease (EN) domains. The PLE RT domain more closely resembles telomerase than the RT from LTRs or LINEs. The EN domain is related to GIY-YIG intron-encoded endonucleases. Some PLE members also have LTR-like sequences, which can be in a direct or an inverse orientation, and have a functional intron [46].
LINEs [48, 49] do not have LTRs; however, they have a poly-A tail at the 3<sup>0</sup> end and are flanked by the TSDs. They comprise about 21% of the human genome and among them L1 with about 850,000 copies is the most abundant and best described LINE family. L1 is the only LINE retroposon still active in the human genome [50]. In the human genome, there are two other LINElike repeats, L2 and L3, distantly related to L1. A contrasting situation has been noticed in the malaria mosquito Anopheles gambiae, where around 100 divergent LINE families compose only 3% of its genome [51]. LINEs in plants, e.g., Cin4 in maize and Ta11 in Arabidopsis thaliana, seem rare as compared with LTR retrotransposons. A full copy of mammalian L1 is about 6 kb long and contains a PolII promoter and two ORFs. The ORF1 codes for a non-sequence-specific RNA binding protein that contains zinc finger, leucine zipper, and coiled-coil motifs. The ORF1p functions as chaperone for the L1 mRNA [52, 53]. The second ORF encodes an endonuclease, which makes a single-stranded nick in the genomic DNA, and a reverse transcriptase, which uses the nicked DNA to prime reverse transcription of LINE RNA from the 3<sup>0</sup> end. Reverse transcription is often unfinished, leaving behind fragmented copies of LINE elements; hence most of the L1-derived repeats are short, with an average size of 900 bp. LINEs are part of the CR1 clade, which has members in various metazoan species, including fruit fly, mosquito, zebrafish, pufferfish, turtle, and chicken [54]. Because they encode their own retrotransposition machinery, LINE elements are regarded as autonomous retrotransposons.
SINEs [48, 49] evolved from RNA genes, such as 7SL and tRNA genes. By definition, they are short, up to 1000 base pair long. They do not encode their own retrotranscription machinery and are considered as nonautonomous elements and in most cases are mobilized by the L1 machinery [55]. The outstanding member of this class from the human genome is the Alu repeat, which contains a cleavage site for the AluI restriction enzyme that gave its name [56]. With over a million copies in the human genome, Alu is probably the most successful transposon in the history of life. Primate-specific Alu and its rodent relative B1 have limited phylogenetic distribution suggesting their relatively recent origins. The mammalian-wide interspersed repeats (MIRs), by contrast, spread before eutherian radiation, and their copies can be found in different mammalian groups including marsupials and monotremes [57]. SVA elements are unique primate elements due to their composite structure. They are named after their main components: SINE, VNTR (a variable number of tandem repeats), and Alu [58]. Usually, they contain the hallmarks of the retroposition, i.e., they are flanked by TSDs and terminated by a poly(A) tail. It seems that SVA elements are nonautonomous retrotransposons mobilized by L1 machinery, and they are thought to be transcribed by RNA polymerase II. SVAs are transpositionally active and are responsible for some human diseases [59]. They originated less than 25 million years ago, and they form the youngest retrotransposon family with about 3000 copies in the human genome [58].
Retro(pseudo)genes are a special group of retroposed sequences, which are products of reverse transcription of a spliced (mature) mRNA. Hence, their characteristic features are an absence of promoter sequence and introns, the presence of flanking direct repeats, and a 30 -end polyadenosine tract [60]. Processed pseudogenes, as sometimes retropseudogenes are called, have been generated in vitro at a low frequency in the human HeLa cells via mRNA from a reporter gene [60]. The source of the reverse transcription machinery in humans and other vertebrates seems to be active L1 elements [61]. However, not all retroposed messages have to end up as pseudogenes. About 20% of mammalian protein-encoding genes lack introns in their ORFs [62]. It is conceivable that many genes lacking introns arose by retroposition. Some genes are known to be retroposed more often than others. For instance, in the human genome there are over 2000 retropseudogenes of ribosomal proteins [63]. A genome-wide study showed that the human genome harbors about 20,000 pseudogenes, 72% of which most likely arose through retroposition [64]. Interestingly, the vast majority (92%) of them are quite recent transpositions that occurred after primate/rodent divergence [64]. Some of the retroposed genes may undergo quite complicated evolutionary paths. An example could be the RNF13B retrogene, which replaced its own parental gene in the mammalian genomes. This retrocopy was duplicated in primates, and the evolution of this primate-specific copy was accompanied by the exaptation of two TEs, Alu and L1, and intron gain via changing a part of coding sequence into an intron leading to the origin of a functional, primate-specific retrogene with two splicing variants [65].
3.2.2 Class II: Mobile Elements Class II elements move by a conservative cut-and-paste mechanism; the excision of the donor element is followed by its reinsertion elsewhere in the genome. DNA transposons are abundant in bacteria, where they are called insertion sequences (see Subheading 3.1), but are present in all phyla. Wicker et al. distinguished two subclasses of DNA transposons based on the number of DNA strands that are cut during transposition [33].
Classical "cut-and-paste" transposons belong to the subclass I, and they are classified as the TIR order. They are characterized by terminal inverted repeats (TIR) and encode a transposase that binds near the inverted repeats and mediates mobility. This process is not usually a replicative one, unless the gap caused by excision is repaired using the sister chromatid. When inserted at a new location, the transposon is flanked by small gaps, which, when filled by host enzymes, cause duplication of the sequence at the target site. The length of these TSDs is characteristic for particular transposons. Nine superfamilies belong to the TIR order, including Tc1- Mariner, Merlin, Mutator, and PiggyBac. The second order Crypton consists of a single superfamily of the same name. Originally thought to be limited to fungi [66], now it is clear that they have a wide distribution, including animals and heterokonts [67]. A heterogeneous, small, nonautonomous group of elements MITEs also belong to the TIR order [68], which in some genomes amplified to thousands of copies, e.g., Stowaway in the rice genome [69], Tourist in most bamboo genomes [70], or Galluhop in the chicken genome [71].
Subclass II includes two orders of TEs that, just as those from subclass I, do not form RNA intermediates. However, unlike "classical" DNA transposons, they replicate without double-strand cleavage. Helitrons replicate using a rolling-circle mechanism, and their insertion does not result in the target site duplication [72]. They encode tyrosine recombinase along with some other proteins. Helitrons were first described in plants, but they are also present in other phyla, including fungi and mammals [73, 74]. Mavericks are large transposons that have been found in different eukaryotic lineages excluding plants [75]. They encode various numbers of proteins that include DNA polymerase B and an integrase. Kapitonov and Jurka suggested that their life cycle includes a single-strand excision, followed by extrachromosomal replication and reintegration to a new location [76].
#### 4 Identification of Transposable Elements
With the ever-growing number of sequenced genomes from different branches of the tree of life, there are increasing TE research opportunities. There are several reasons why one would like to analyze TEs and their "offsprings" left in a genome. First of all, they are very interesting biological subjects to study genome structure, gene regulation, or genome evolution. In some cases, they also make genome assembly and annotation quite challenging, especially with the current NGS technology that generates reads shorter than TEs. Nevertheless, TEs should be and are worthy to study. However, it is not a simple task and requires different approaches depending on the level of analysis. We will walk through these different levels starting with raw genome sequences without any annotation and discuss different methods and software used for TE analyses. In principle, we can imagine two scenarios: in the first one, genomic or transcriptome sequences are coming from a species for which there is already some information about the transposon repertoire, for instance, a related genome has been previously characterized or TEs have been studied before. In the second scenario, we have to deal with a completely unknown genome or a genome for which little information exists with regard to TEs. In the former case, one can apply a range of techniques used in comparative genomics or try to search specific libraries of transposons using the "homology search" approach. In the latter, which is basically an approach to identify TEs de novo, first we need to find any repeats in a genome and then attempt characterization and classification of newly identified repetitive sequences. In this approach, we will find any repeats, not necessarily transposons. There are many algorithms, and even more software, that can be applied in both approaches.
4.1 De Novo Approaches to Finding Repetitive Elements
There are several steps involved in the de novo characterization of transposons. First, we need to find all the repeats in a genome, then build a consensus of each family of related sequences, and finally classify detected sequences. For the first step, three groups of algorithms exist: the k-mer approach, sequence self-comparison, and periodicity analysis.
In the k-mer approach, sequences are scanned for overrepresentation of strings of certain length. The idea is that repeats that belong to the same family are compositionally similar and share some oligomers. If the repeats occur many times in a genome, then those oligomers should be overrepresented. However, since repeats and transposons in particular are not perfect copies of a certain sequence, some mismatches must be allowed when oligo frequencies are calculated. The challenge is to determine optimal size of an oligo (k-mer) and number of mismatches allowed. Most likely, these parameters should be different for different types of transposons, i.e., low versus high copy number, old versus young transposons, and those from different classes and families. Several programs have been developed based on the k-mer idea using a suffix tree data structure including REPuter [77, 78], Vmatch (Kurtz, unpublished; http://www.vmatch.de/), and Repeat-match [79, 80]. Another approach is to use fixed length k-mers as seeds and extend those seeds to define repeat's family as it was implemented in ReAS [81], RepeatScout [82], and Tallymer [83]. Another interesting algorithm can be found in the FORRepeats software [84], which uses factor oracle data structure [85]. It starts with detection of exact oligomers in the analyzed sequences, followed by finding approximate repeats and their alignment.
The second group of programs developed for de novo detection of repeated sequences is using self-comparison approach. Repeat Pattern Toolkit [86], RECON [87], PILER [88, 89], and BLASTER [90] belong to this group. The idea is to use one of the fast sequence similarity tools, e.g., BLAST [91], followed by clustering search results. The programs differ in the search engine for the initial step, though most are using some of the BLAST algorithms, the clustering method, and heuristics of merging initial hits into a prototype element. For instance, RECON [87], which was developed for the repeat finding in unassembled sequence reads, starts with an all-to-all comparison using WU-BLAST engine. Then, single-linkage clustering is applied to alignment results that is followed by construction of an undirected graph with overlapping. The shortest sequence that contains connected images (aligned subsequences) creates a prototype element. However, this procedure might result in composite elements. To avoid this, all the images are aligned to the prototype element to detect potential illegitimate mergers and split those at every point with a significant number of image ends.
PILER [88, 89] is using a different approach to find initial clusters. Instead of BLAST, it uses PALS (pairwise alignment of long sequences) for the initial alignment. PALS records only hit points and uses banded search of the defined maximum distance to optimize its performance. To further improve performance of the system, PILER uses different heuristics for different types of repeats, i.e., satellites, pseudosatellites, terminal repeats, and interspersed repeats. Finally, a consensus sequence is generated from a multiple sequence alignment of the defined family members.
Dot matrix is a simple method to compare two biological sequences. The graphical output of such an analysis is called a dotplot. Dotplots can be used to detect conserved domains, sequence rearrangements, RNA secondary structure, or repeated sequences. It compares every residue in one sequence to every residue in the other sequence or to every residue of the same sequence in the selfcomparison mode. In the latter case, there will be a main diagonal line representing a perfect match and a number of short diagonal lines representing similar regions (red circles in Fig. 3). Interestingly, simple repeats appear as diamond shapes on a main diagonal line or short vertical and horizontal lines outside the main diagonal line (red squares in Fig. 3). The method was introduced to biological analyses almost a half century ago [92, 93]. However, the first easy-to-use software with a graphical interface, DOTTER, was developed much later [94]. The major problem of this approach is the time required for the dotplot calculation, which is of quadratic complexity. This proved to be prohibitive for comparison of the genome-size sequences. One of the solutions to this problem is using a word index for the fast identification of substrings. Gepard implements the suffix array data structure to improve the execution time [95]. It is written in Java, which makes it platform-independent. Gepard enables analyses of sequences at the mega-base level in the matter of seconds, and it takes about an hour to analyze the whole human chromosome I [95]. The example of the dotplot produced by the Gepard is presented in Fig. 3.
4.2 Transposable Elements Determination in NGS Data With constant improvement of sequencing technology associated with decreasing sequencing cost, the number of new sequenced genomes is exploding. As of January 2019, there are more than 7000 eukaryotic and almost 180,000 prokaryotic genomes publicly available (information retrieved on January 16, 2019, from https:// www.ncbi.nlm.nih.gov/genome/browse/). However, this comes with a price; most of the recently sequenced genomes, due to the short read sequencing technology, are available at various levels of
Fig. 3 Graphical output of the Gepard. A 30 kb fragment of mouse chromosome 12 was compared to itself. Similar sequences are represented by diagonal lines if both fragments are located on the same strains or by reverse diagonal lines if the fragments with significant similarity are located on opposite strands. Some of the examples are marked with the red circles. Simple repeats are represented by either diamond shapes on the main diagonal or horizontal and vertical lines. Some of the examples are marked with the red squares
"completeness" or assembly. For most non-model organisms, we are presented with draft assemblies of rather short contigs. Moreover, these genomes usually are not very well annotated, with TEs not being on the annotation priority list. Unfortunately, genome annotation pipelines do not include TE annotation, focusing on protein-coding and RNA-coding genes. To fill the gap, a number of methods have been developed to detect repeats from short reads. Two algorithms dominate in attempts to determine repeats in NGS raw reads: clustering and k-mer. Transposome [96] and RepeatExplorer [97] employ the former approach, while RepARK [98], REPdenovo [99], and dnaPipeTE [100] utilize the latter one. Since NGS results in the relatively short reads, assembly of selected sequences into longer contigs representing TEs is required after initial clustering of the raw reads.
4.3 Population-Level Analyses of Transposable Elements Recent advances in sequencing technology and the sharp decrease in sequencing costs allow genomic studies at population level. Although initially focused on human populations [101–103], recent population studies of other species have been initiated as well [104, 105]. One of the common questions in such studies is Actual genomic fragment with a TE insertion
Reference genome - a TE missing
Fig. 4 Detection of a TE insertion (polymorphic TE) from the NGS data. The upper panel shows real genomic sequence with a TE, which is not present in the reference genome (lower panel). Hypothetical discordant pairreads (a, b, d, f, g, i, j, k, l, o, q, s, and t) have only one the pairs mapped to the reference genome, while the other would map to a consensus sequence of a TE. The hypothetical split reads (c, e, h, m, p, and r) will have part of the sequence mapped to the reference genome and the other to a TE consensus sequence
how much structural variation (SV) exists in different populations. TE insertions are responsible for about 25% of structural variants in human genomes [106]. In general, any tool designed for detection of SV should work for TE insertion analysis, but specialized software can take advantage of specific expectations related to insertions of TEs. Most of the SV-detection algorithms rely on pairedend reads and are based on discordant read pair mapping and/or split reads mapping (Fig. 4). A discordant pair read is defined as one that is inconsistent with the expected insert size in the library used for sequencing. For example, if the insert size of the library used for sequencing is 300 nt but the reads map to a reference genome within much larger distance or to two different chromosomes, such a pair is considered to be discordant. If, additionally, one of the reads maps to a TE, it might be an indication of a polymorphic TE. Usually some filtering is used to reduce a chance of false positives. These include minimum read number in the cluster mapped to a unique position, quality score of the reads, or consistency in reads orientation. However, the discordant read mapping cannot detect exact insertion position. Therefore another step is required that may include local assembly and split-read mapping.
A split read is defined as a read for which part of it maps uniquely to one position in the genome and the other part to another position. This is, for example, a very common feature of the mapping of RNA-seq data to eukaryotic genomes when reads span two exons. Split reads are being also observed if structural variants exist. In a case of a TE insertion, a part of the read will be mapped to a unique location and the rest to a TE in some other location or may not be mapped at all (Fig. 4).
Different methods for structure variant detection return different results on the same data. Recently published benchmarking demonstrates that TE detection is not an exception [107, 108]. Ewing [107] compared TranspoSeq [109] with two other tools, Tea [110] and TraFIC [111], on the same data sets. Results were not very encouraging as in both comparisons there was a high fraction of insertions detected only by a single program [107]. Similar conclusion was drawn by Rishishwar et al. [108] in a benchmark of larger number of tools including MELT [106], Mobster [112], and RetroSeq [113]. It is clear that different software have different biases, and each one can produce a high number of false positives. It is recommended then to employ several programs for high confidence results. Exhaustive tests run on real and simulated human genome data showed superior performance of MELT [106, 108]. TIPseqHunter is another tool developed to identify transposon insertion sites based on the transpose insertion profiling using next-generation sequencing [114]. It employs machine learning algorithm to ensure high precision and reliability. It is worth to note that all these tools were designed for short read sequencing methods. However, with current development of single-molecule long reads, sequencing technologies such as Pac-Bio and Oxford Nanopore may make these methods irrelevant and obsolete. Long reads should be of superior performance and make TE insertion detection relatively easy with more traditional aligners, such as MegaBLAST [115], BLAT [116], or LAST [117].
4.4 Comparative Genomics of TE Insertions To understand the general pattern of TE insertions in different genomes and evolutionary dynamics of TE families, a comparative approach is necessary. Although precomputed alignments of different genomes are publicly available, for example, the UCSC Genome Browser includes Multiz alignments of 100 vertebrate genomes [118], not many tools are available for such analyses. One of them is GPAC (genome presence/absence compiler) that creates a table of presence and absence of certain elements based on the precomputed multiple genomes alignment [119] (http://bioin formatics.uni-muenster.de/tools/gpac/index.hbi). The tool is quite generic, but is well suited for the TE comparative analysis (see Fig. 5 for an example).
Fig. 5 The output table of the GPAC software. Several Alu elements were analyzed for presence/absence in 11 primate species. The human genome was used a reference, and "hit coordinates" refer to that genome along with the information on the annotated elements in the hit region and a type of the region. For each genome, the presence (+) or absence (-) of the hit is presented. x/ denotes that only part of the original insertion (less than 20%) is present in a given genome, and ¼¼ indicates that more than 80% of the expected sequence is not alignable in a given locus. The optional phylogenetic tree constructed based on the obtained data is shown in the lower right corner
#### 4.5 Classification of Transposable Elements
Once the consensus of a repetitive element has been constructed, it can be subjected to further analyses. There are two major categories of programs dealing with the issue of TE classification: library or similarity-based and signature-based. The latter approach is very often used in specialized software, i.e., tailored for specific type of TEs. However, some general tools also exist, e.g., TEclass [120].
The library approach is probably the most common approach for TE classification. It is also very efficient and quite reliable as long as good libraries of prototype sequences exist. In practice, it is the recommended approach when we analyze sequences from wellcharacterized genomes or from a genome relatively closely related to a well-studied one. For instance, since the human genome is one of the best studied, any primate sequences can be confidently analyzed using the library approach. Most likely, the first software using the similarity-based approach for repeat classification was Censor developed by Jerzy Jurka in the early 1990s [121]. It uses RepBase [122] as a reference collection and BLAST as a search engine [91]. However, the most popular TE detection software is RepeatMasker (RM) (http://www.repeatmasker.org). Interestingly, RM is also using RepBase as a reference collection and AB-BLAST, RM-BLAST, or cross-match as a search engine. In both cases, original search hits are processed by a series of Perl scripts to determine the structure of elements and classify them to one of known TE families. Both Censor and RM also employ userprovided libraries, including "third-party" lineage-specific libraries, e.g., TREP [123]. Over the years, RepeatMasker has become a standard tool for TE analyses, and often its output is used for more biologically oriented studies (see below). The aforementioned programs have one important drawback: since they are completely based on sequence similarity, they can detect only TEs that had been previously described. Nevertheless, similarity searches, like in many other bioinformatics tasks, should be the first approach for the analysis of repetitive elements.
Signature-based programs are searching for certain features that characterize specific TEs, for example, long terminal repeats (LTRs), target site duplications (TSDs), or primer-binding sites (PBSs). Since different types (families) of elements are structurally different, they require specific rules for their detection. Hence, many of the programs that use signature-based algorithms are specific for certain type of transposons. There are a number of programs specialized in detection of LTR transposons, which are based on a similar methodology. They take into account several structural features of LTR retroposons including size, distance between paired LTRs and their similarity, the presence of TSDs, and the presence of replication signals, i.e., the primer-binding site and the polypurine tract (PPTs). Some of the programs check also for ORFs coding for the gag, pol, and env proteins. LTR\_STRUC [124] was one of the first programs based on this principle. It uses seed-and-extend strategy to find repeats located within userdefined distance. The candidate regions are extended based on the pairwise alignment to determine cognate LTRs' boundaries. Putative full-length elements are scored based on the presence of TSD, PBS, PPT, and reverse transcriptase ORF. However, because of the heuristics described above, LTR\_STRUC is unable to find incomplete LTR transposons and in particular solo LTRs. Another limitation of this program is its Windows-only implementation that significantly prohibits automated large-scale analysis. Several other programs have been developed based on similar principles, e.g., LTR\_par [125], find\_LTR [126], LTR\_FINDER [127], and LTRharvest [128]. Lerat tested performance of these programs [129], and although sensitivity of the methods was acceptable (between 40% and 98%), it was at the expense of specificity, which was very poor. In several cases, the number of falsely assigned transposons exceeded the number of correctly detected ones.
Another group of transposons that have a relatively conserved structure are MITEs and Helitrons. Several specialized programs were developed that take advantage of their specific structure. FINDMITE [130] and MUST [131] are tailored for MITEs, while HelitronFinder [132] and HelSearch [133] were developed for Helitron detection.
A further interesting approach to transposon classification was implemented by Abrusan et al. [120] in the software package called TEclass, which classifies unknown TE consensus sequences into four categories, according to their mechanism of transposition: DNA transposons, LTRs, LINEs, and SINEs. The classification uses support vector machines, random forests, learning vector quantization, and predicts ORFs. Two complete sets of classifiers are built using tetramers and pentamers, which are used in two separate rounds of the classification. The software assumes that the analyzed sequence represents a TE and the classification process is binary, with the following steps: forward versus reverse sequence orientation > DNA versus retrotransposon > LTRs versus nonLTRs (for retroelements) > LINEs versus SINEs (for nonLTR repeats). If the different methods of classification lead to conflicting results, TEclass reports the repeat either as unknown or as the last category where the classification methods agree (http://bioinfor matics.uni-muenster.de/tools/teclass/index.hbi).
#### 4.6 Pipelines Recent years witnessed some attempt to create more complex, global analyses systems. One such a system is REPCLASS [134]. It consists of three classification modules: homology (HOM), structure (STR), and target site duplication (TSD). Each module can be run separately or in the pairwise manner, whereas the final step of the analysis involves integration of the results delivered by each module. There is one interesting novelty in the STR module, namely, implementation of tRNAscan-SE [135] to
detect tRNA-like secondary structure within the query sequence, one of the signatures of many SINE families. The REPPET is another pipeline for TE sequence analyses. It uses "classical" three-step approach for de novo TE identification: self-alignment, clustering, and consensus sequences generation. However, the pipeline is using a spectrum of different methods at each step, followed by a rigorous TE classification step based on recently proposed classification of TEs [136]. Unfortunately, a complex implementation that makes installation and running the system rather difficult limits usage of the pipeline. The classification step seems to be unreliable as it may annotate lineage-specific TEs in wrong taxonomical lineages (Kouzel and Makalowski, unpublished data).
There are other attempts to create comprehensive systems for "repeatome" analysis. One of them is dnaPipeTE developed for mosquito genomes' analyses [100]. Interestingly, dnaPipeTE works on the raw NGS data, which makes the pipeline well suited for genomes with lower sequencing depth. The raw reads are first subjected to k-mer count on the sampled data. The sampling of the data to size less than 0.25 of the genome is required to avoid clustering reads representing unique sequences. The determined repetitive reads are assembled into contigs using Trinity [137]. Although Trinity was originally developed for transcriptome assembly from RNA-seq data, it proves to be very useful for TEs assembly from short reads as it can efficiently determine consensus sequences of closely related transposons. In the next step, dnaPipeTE annotates repeats using RepeatMasker with either built-in or user-defined libraries. This is probably the weakest point of the pipeline as it will not annotate any novel TEs, which have no similar sequences present in the provided libraries. It would be useful to complement this step with model-based or machine learning approaches (see Subheading 4.5). After contigs' annotation, copy number of the TEs are estimated using BLAST algorithm [91]. Finally, sequence identity between an individual TE and its consensus sequence is used to determine the relative age of the TEs. The pipeline produces a number of output files including several graphs, i.e., pie chart with the relative proportion of the main repeat classes and graph with the number of base pairs aligned on each TE contig and TE age distribution. Overall, the dnaPipeTE is very efficient, outperforming, according to the authors, RepeatExplorer by severalfold [100].
4.7 Meta-analyses Most of the software developed are focused on the TE discovery and rarely offer more biological oriented analyses. Consequently, researchers interested in TE biology or using TE insertions as tools for another biological investigations need to utilize other resources. One of them is TinT (transposition in transposition), tool that applies maximum likelihood model of TE insertion probability to estimate relative age of TE families [138] (http://bioinformatics. uni-muenster.de/tools/tint/index.hbi). In the first steps, it takes RepeatMasker output to detect nested retroposons. Then, it generates a data matrix that is used by a probabilistic model to estimate chronology and activity period of analyzed families. The method was applied to resolve the evolutionary history of galliformes [139], marsupials [140], lagomorphs [141], squirrel monkey [142], or elephant shark [143].
Another interesting application that takes advantage of TEs is their use for detecting signatures of positive selection [144], a central goal in the field of evolutionary biology. A typical research scenario for this application would be investigating whether a specific TE fragment exapted into resident genomic features, such as proximal and distal enhancers or exons of spliced transcripts, has undergone accelerated evolution that could be indicative of gain of function events. In short, the test first requires the identification of all genomically interspersed TE fragments that are homolog to the TE segment of interest, which can be done through alignments with a family consensus sequence. Based on multi-species genome alignments, a second step involves identification of lineage-specific substitutions in every single homolog fragment, which are then consolidated into a distribution of lineage-specific substitutions that provides the expectation (null distribution) for a segment evolving largely without specific constraints (neutrally). A significantly higher number of lineage-specific substitutions observed in the TE fragment of interest compared to the null distribution could then be interpreted as a molecular signature of adaptive evolution. However, the possibility of confounding molecular mechanisms, such as GC-biased gene conversion [145–147], needs to be evaluated. We note that building the null distribution based only on data from intergenic regions, where transcription-coupled repair is absent, results in a more liberal estimate of the expected substitutions, which in turn leads to a more conservative estimate of the adaptive evolution. Additionally, building the null distribution requires the detection of many homolog fragments, which limits the applicability of the test to TE families with numerous members in a given genome. Prime examples would be human Alu or murine B1 SINEs. In theory, this test could also be used for detecting signatures of purifying selection by searching for fragments depleted of lineage-specific substitutions. However, the low level or complete lack of lineage-specific substitution is characteristic to many TE fragments, obscuring the effect of potential purifying forces.
#### 5 Concluding Remarks
Annoying junk for some, hidden treasure for others, TEs can hardly be ignored [148]. With their diversity and high copy number in most of the genomes, they are not the easiest biological entities to analyze. Nevertheless, recent years witnessed increased interest in TEs. On the one hand, we observe improvement in computational tools specialized in TE analyses. Table 3 lists some of such tools and
(continued)
the up-to-date list can be found at our web site: http://www. bioinformatics.uni-muenster.de/ScrapYard/. On the other hand, improved tools and new technologies enable biologists to explore new research avenues that might lead to novel, fascinating insights into the biology of mobile elements.
#### References
(2017) Long terminal repeats power evolution of genes and gene expression programs in mammalian oocytes and zygotes. Genome Res 27(8):1384–1394. https://doi.org/10. 1101/gr.216150.116
Res 12(5):795–807. https://doi.org/10. 1101/gr.226102. Article published online before print in April 2002
retrotransposon, LINE-1. RNA Biol 7 (6):706–711
complete genomes. Bioinformatics 15 (5):426–427
repeatome with dnaPipeTE from raw genomic reads and comparative analysis with the yellow fever mosquito (Aedes aegypti). Genome Biol Evol 7(4):1192–1205. https://doi.org/10.1093/gbe/evv050
sequencing. Genome Res 24(7):1053–1063. https://doi.org/10.1101/gr.163659.113
CM, Lee BT, Karolchik D, Hinrichs AS, Haeussler M, Guruvadoo L, Navarro Gonzalez J, Gibson D et al (2018) The UCSC Genome Browser database: 2018 update. Nucleic Acids Res 46(D1): D762–D769. https://doi.org/10.1093/ nar/gkx1020
for de novo detection of LTR retrotransposons. BMC Bioinform 9:18. https://doi.org/ 10.1186/1471-2105-9-18. 1471-2105-9- 18 [pii]
genome. Nat Biotechnol 29(7):644–U130. https://doi.org/10.1038/nbt.1883
e1003684. https://doi.org/10.1371/jour nal.pgen.1003684
147. Gotea V, Elnitski L (2014) Ascertaining regions affected by GC-biased gene conversion through weak-to-strong mutational hotspots. Genomics 103(5–6):349–356. https://doi.org/10.1016/j.ygeno.2014.04. 001
148. Makalowski W (2000) Genomic scrap yard: how genomes utilize all that junk. Gene 259 (1–2):61–67. https://doi.org/10.1016/ S0378-1119(00)00436-4
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Part III
#### Phylogenomics and Genome Evolution
## Chapter 7
#### Modern Phylogenomics: Building Phylogenetic Trees Using the Multispecies Coalescent Model
#### Liang Liu, Christian Anderson, Dennis Pearl, and Scott V. Edwards
#### Abstract
The multispecies coalescent (MSC) model provides a compelling framework for building phylogenetic trees from multilocus DNA sequence data. The pure MSC is best thought of as a special case of so-called "multispecies network coalescent" models, in which gene flow is allowed among branches of the tree, whereas MSC methods assume there is no gene flow between diverging species. Early implementations of the MSC, such as "parsimony" or "democratic vote" approaches to combining information from multiple gene trees, as well as concatenation, in which DNA sequences from multiple gene trees are combined into a single "supergene," were quickly shown to be inconsistent in some regions of tree space, in so far as they converged on the incorrect species tree as more gene trees and sequence data were accumulated. The anomaly zone, a region of tree space in which the most frequent gene tree is different from the species tree, is one such region where many so-called "coalescent" methods are inconsistent. Second-generation implementations of the MSC employed Bayesian or likelihood models; these are consistent in all regions of gene tree space, but Bayesian methods in particular are incapable of handling the large phylogenomic data sets currently available. Two-step methods, such as MP-EST and ASTRAL, in which gene trees are first estimated and then combined to estimate an overarching species tree, are currently popular in part because they can handle large phylogenomic data sets. These methods are consistent in the anomaly zone but can sometimes provide inappropriate measures of tree support or apportion error and signal in the data inappropriately. MP-EST in particular employs a likelihood model which can be conveniently manipulated to perform statistical tests of competing species trees, incorporating the likelihood of the collected gene trees on each species tree in a likelihood ratio test. Such tests provide a useful alternative to the multilocus bootstrap, which only indirectly tests the appropriateness of competing species trees. We illustrate these tests and implementations of the MSC with examples and suggest that MSC methods are a useful class of models effectively using information from multiple loci to build phylogenetic trees.
Key words Introgression, Hybridization, Coalescent, Recombination, Neutrality, Molecular evolution
#### 1 Introduction
The concept of a phylogeny or "species tree," a bifurcating dendrogram graphically depicting the relationships among a group species, is one of the oldest and most powerful icons in all of biology. After Charles Darwin sketched the first species tree
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_7, © The Author(s) 2019
(in Transmutation of Species, Notebook B, 1837), he remained fascinated by the image for 22 years, eventually including a species tree as the only figure in On the Origin of Species [1]. Though species trees reached their aesthetic apogee with Ernst Haeckel's Tree of Life in 1886, the pursuit of ever-more scientifically accurate trees has kept phylogenetics a vibrant discipline for the 150 years since.
Because the direct evolution of species in the past is not observable (not even in the fossil record), relationships among species are often inferred by shared characteristics among extant taxa. Until the 1970s, this effort took place almost exclusively by using morphological characters. Although this approach had many successes, the paucity of characters and the challenges of comparing species with no obvious morphological homologies were persistent problems [2, 3]. When molecular techniques were developed in the late 1960s, it soon became clear that the sheer volume of molecular data that could be collected would represent a vast improvement. When DNA sequences became widely available for a range of species [4], molecular comparisons quickly became de rigueur [5–8]. Nonetheless, it was recognized early on that molecular phylogenies had their own suite of problems; the concept that not all gene tree topologies would match the true species tree topology (i.e., would not be speciodendric sensu Rosenberg [9]) was implicit in early empirical allozyme and mitochondrial DNA studies [10, 11]. However, it was generally assumed that the idiosyncratic genealogical history of any one gene, as reconstructed from extant mutations, was an acceptable approximation for the true history of the species given the potentially overwhelming quantity and seductive utility of molecular data [12–15]. Indeed, this assumption is still prevalent in the thinking of those who favor concatenation or supermatrix approaches as a means of combining information from multiple genes that may still differ in their genealogy from each other and from the species tree [16, 17]. In the meantime, the term "phylogeny" frequently became conflated with "gene tree," the entity produced by many of the leading phylogenetic packages of the day. The term "species tree," in use since the late 1970s to emphasize the distinction between lineage histories and gene histories (reviewed in [11, 18]), was only gradually acknowledged, despite the fact that species trees are the rightful heirs to the term "phylogeny" and better encapsulate the true goals of molecular and morphological systematics [19].
1.1 Stopgap Approaches to Gene Tree Heterogeneity
By and large, the ensuing decades of molecular phylogenetics has fulfilled much of its potential, revolutionizing taxonomies and resolving conundrums previously considered intractable. However, as the amount of genetic data per species becomes ever-more voluminous, it has become clear that the conflicts between individual genes with each other and with the overarching species tree, both in topology and branch lengths, can have practical consequences for phylogenetic analysis if not dealt with properly [18–23]. At first, some researchers treated this phenomenon as though it were an information problem: when working with only a few mutations, you were bound to occasionally get unlucky and sequence a gene whose random signal of evolution did not match that of the taxa being studied. The reasoning was surely more and/or longer sequences would fix that problem and cause gene trees to converge [16]. However, as more genes were sequenced, and as the properties of gene lineages within populations were studied in detail [24, 25], the twin realities of gene tree heterogeneity and "incomplete lineage sorting" [11] (ILS) became clear (Figs. 1 and 2). The probability of an event such as incomplete lineage sorting, which if considered alone would lead to inferring the wrong species tree, was worked out theoretically for the four allele/two species case first [26], followed by the three allele/three species case [7, 13] and more general cases [12, 27]. Pamilo and Nei [12] were among those that proposed that the solution was to simply acquire more gene sequences, after which the central tendency of this gene set would point to the correct relationships, a "democratic vote" method, where each gene was allowed to propose its own tree, and the topology with the most "votes" was declared the winner and therefore the true phylogeny. Though generally true for three-species case, it can sometimes produce the wrong topology with four or more species [28]. In fact, we now know that with four or more species, there is an "anomaly zone" for species trees with short branch lengths as measured in coalescence units, in which the addition of more genes for sampled taxa is guaranteed to lead to the wrong species tree topology for the democratic vote method [29, 30]. (Coalescent time units, equivalent to t/Ne where t is the number of generations since divergence and Ne is the effective population size of the lineage, are a convenient unit for discussions of gene tree/species tree heterogeneity. For a clear explanation, see Box 2 of Degnan and Rosenberg [28].) Such anomaly zones may be rare empirically [31], but empirical examples are emerging [32, 33], and even the theoretical possibility remains disconcerting. In addition, because the number of possible tree topologies increases as the double factorial of the number of tips, for species trees with more than four tips, a very large number of genes are required to determine which gene tree is in fact the most frequent. Advanced consensus methods [34] can circumvent some of the problems of the democratic vote by using novel assembly methods, such as rooted triple consensus [35], greedy consensus [36], or supertree methods [37]. However, although such methods suffer from lack of a biological model motivating the method of consensus, approaches such as that proposed by Steel and Rodrigo [38] might help approximate the dynamics of biological models while allowing for faster and more flexible extensions and should be further developed.
Fig. 1 An example showing the utility of multiple gene trees in producing species tree topologies. (a) Nine unlinked loci are simulated (or inferred without error) from a species group with substantial amounts of incomplete lineage sorting. Note that no single gene recovers the correct relationship between clades. Furthermore, despite identical conditions for all nine simulations, no two genes agree on the correct topology, let alone the correct divergence times. (b) Superimposing the nine gene trees on top of each other clarifies the relationships. It can be (correctly) inferred that the true tree is perfectly ordered, with (ABC) diverging from D about 1500 generations ago, the (AB)-C split occurring at 800, and A diverging from B about 600 generations ago. Also, the amount of crossbreeding within the recently diverged taxa implies (correctly) that C has the effective smallest population size
The second empirical approach to the problem of conflicting gene trees was to bypass it altogether. Concatenation methods appended the sequence of one gene to that of the next, to create long alignments or supermatrices [39], a technique that in some situations was superior to standard consensus methods in resolving discordance or achieving statistical consistency [40]. But some researchers, including those who questioned the "total-evidence"
Fig. 2 The relationship between gene trees and species trees. Lines within the species trees indicate gene lineages. Simplified gene trees are shown below each species tree. Whereas gene trees on the left vary due to deep coalescence, gene trees on the right are topologically concordant but vary slightly in branch lengths due to the coalescent. Modified with permission from [19]
approach to systematics (e.g., [41]), advocated against concatenation when, for whatever reason, gene trees appeared to conflict with one another. One problem with the concatenation approach was that it assumed full linkage across the supermatrix, a situation that would obviously not be the case if genes were on different chromosomes. Even when the lineage lengths in a species tree are long in coalescent units, such that gene tree topologies are congruent, the branch lengths of trees of genes on different chromosomes will differ subtly from one another due to the stochasticity of the coalescent process. The early implementations of this method also assumed the same distribution of mutation rates across the sequence, which was clearly not the case if the matrix included coding and noncoding regions. Like democratic vote methods, concatenation of many genes was sometimes defended as sufficient to override the conflicting signal across genes [42, 43], despite widespread acknowledgment that gene tree heterogeneity is ubiquitous and that concatenation can sometimes give the wrong answer, especially although not exclusively in the anomaly zone [44, 45].
Concatenation as a method of combining phylogenomic data still remains popular by default [16, 46], particularly among phylogenetic studies of higher taxa where incomplete lineage sorting is assumed to be rare. However, this logic suffers from two flaws frequently seen in the literature. First, "deep" phylogenetic studies among higher taxa are no more immune to the problems of ILS than are studies among closely related species, because it is the length of a given branch, not its depth in the tree, that is relevant to probability of gene tree discordance [28]. Detecting such ILS and ruling out gene tree congruence will indeed be more challenging in deep phylogenomic studies, but it should not be assumed that ILS will be less prevalent at deep scales than at shallow scales. Second, current implementations of concatenation represent only one way of species tree construction in which each gene is forced to have the same topology. The real distinction between concatenation and coalescent models is not the presence or absence of ILS but rather the possibility of conditional independence of gene trees as mediated by recombination between genes [47]. Even if all gene trees in an analysis are topologically identical, physically connecting different genes in a single supermatrix does not capture variation in branch lengths that recombination will allow in nature. More effort should be devoted to "supermatrix-like" methods that constrain gene trees to the same topology but allow recombination between genes and conditional independence of branch lengths, since these qualities will influence how signal is accumulated as more genes are added [47]. A final problem with concatenation is that, in a strict sense, concatenation also does not generate species trees, in so far as the method treats all nucleotides as if they were part of a single non-recombining gene, and thus does not distinguish between gene and species trees [19]. In the end, concatenation is best thought of as a special case of more general models of phylogenetic inference that acknowledge gene tree heterogeneity and conditional independence of genes. One such model is the multispecies coalescent model [23, 28, 48]. It is this model that provides the basis for a recent flurry of promising methods that permit efficient and consistent estimation of species trees under a variety of conditions.
#### 2 The Multispecies Coalescent Model
A plausible probabilistic model for analyzing multilocus sequences should involve not only the phylogenetic relationship of species (species tree) but also the genealogical history of each gene (gene tree) and allow different genes to have different histories. Unlike concatenation, such a multispecies coalescent model (MSC) explains the evolutionary history of multilocus sequences through two levels of biological hierarchy, the gene tree and the species tree, rather than just one [23, 49]. Models acknowledging these two levels require an explicit description of how sequences evolve on gene trees, the traditional likelihood equation of Felsenstein [50] and others, as well as how gene trees evolve in the species tree, the likelihood for which was first described by Rannala and Yang [48]. With a few exceptions (described below), the genealogical relationship (gene tree) of neutral alleles can be simply depicted by a coalescence process in which lineages randomly coalesce with each other backward in time. The MSC is a simple application of the single population coalescent model to each branch in a species tree [28]. It holds the standard assumptions found in many neutral coalescent models: no natural selection or gene flow among populations, no recombination within loci but free recombination between loci, random mating and a Wright-Fisher model of inheritance down each branch of the species tree. Despite these seemingly oversimplified assumptions, the pure coalescent model is fundamental in explaining the gene tree-species tree relationship because it forms a baseline for incorporating additional evolutionary forces on top of random drift [28, 49]. More importantly, the pure coalescent model provides an analytic tool to detect the evolutionary forces responsible for the deviation of the observed data (molecular sequences) from those expected from the model.
The coalescent process works, in effect, by randomly choosing ancestors with replacement from the population backward through time for each sequence in the original sample. Eventually, two of these lineages will share a common ancestor, and the lineages are said to "coalesce." The process continues until all lineages coalesce at the most recent common ancestor (MRCA). Multispecies coalescence works the same way but places constraints on how recently the coalescences occur, corresponding to the species' divergence times. Translating this model into computer algorithms for inferring species trees has led to a plethora of models [51–55], some of which first build gene trees by traditional methods and then combine them into a species tree with the highest likelihood or other criteria ("two-step" methods, e.g., [56] or [57]), others of which, particularly Bayesian methods [58–60], simultaneously estimate gene trees and species tree. In general for likelihood or Bayesian approaches, a species tree has been proposed, and the likelihood of each gene tree is evaluated using the MSC, with or without various priors, to evaluate the likelihood of the data (DNA sequences in the case of Bayesian methods or gene trees in the case of likelihood methods like MP-EST [56]) given the species tree or the posterior probability of the species tree. In this way, traditional multispecies coalescent methods are the converse of consensus methods; rather than each locus proposing a potentially divergent species tree, a common species tree is assumed and evaluated, given the sometimes divergent patterns observed among multiple loci.
A number of implementations of this idea have been developed (reviewed by Edwards [19, 54]). Several "two-step" packages are available for moving from independently built gene trees to species trees, including minimization of deep coalescence [61], STEM [62], JIST [63], GLASS [64], STAR, STEAC [65], NJst [66], and ASTRAL [57, 67]. Three methods to date utilize "one-step" Bayesian methods to infer gene trees and the species tree, with the input data being DNA sequences: BEST [58, 68, 69], \*BEAST2 [59], and a new model (A00) in the Bayesian Phylogenetics and Phylogeography (bpp) package [70–72]. An additional "one-step" method, SVD Quartets [73], derives species trees directly from aligned, unlinked single-nucleotide polymorphisms using the method of invariants in a coalescent framework. Species tree methods exhibit a number of attractive advantages over concatenation methods in terms of performance. These advantages are not restricted to the anomaly zone, occur across broad regions of tree space, and include less susceptibility to long-branch attraction [74] and missing data [75]. Another attractive aspect of species tree methods and multispecies coalescent models is that they deliver more appropriate estimated levels of confidence that are more evenly spread across genes and appear to be less susceptible to the inflation of posterior probabilities that was early on attributed to Bayesian analyses (e.g., [76, 77]) but may also be due to model misspecification due to concatenation [53]. Bayesian methods are generally agreed to be the most efficient and accurate, capturing all details of the MSC model seamlessly [52]. However, one drawback is that the estimation of larger numbers of parameters (population sizes and divergence times in addition to topologies) can slow computation, may not be relevant in some situations [78], and is generally not possible with the large data sets that are routinely seen today in phylogenomics [59]. Thus far, two-step methods such as ASTRAL, STAR, NJst, and MP-EST have proven the most widely used for large-scale phylogenomic studies, such as the Avian Phylogenomics Project [79] and large-scale phylogenomics of fish [80] and plants [81].
2.1 Sources of Gene Tree/Species Tree Discordance and Violations of the Multispecies Coalescent Model 2.1.1 Population Processes The "standard" and most common reason why gene trees are not speciodendritic is incomplete lineage sorting, i.e., lineages have not yet been reproductively isolated for long enough for drift to cause complete genetic divergence in the form of reciprocal monophyly of gene trees ([82]; Figs. 1 and 2). This source of gene tree heterogeneity is guaranteed to be ubiquitous, if only because it arises from the finite size populations of all species that have ever come into existence. Almost all the techniques and software packages discussed above are designed to approximate uncertainties in species tree topology arising from this phenomenon.
Delimitation of Species and Diverging Lineages For recent divergences, the definition of "species" can become problematic for species tree methods [63], and the challenge of delimiting species has, if anything, increased now that the overly conservative strictures of gene tree monophyly as a delimiter of species have been mostly abandoned [82]. This fundamental issue in a phylogenetic study—whether the extent of divergence among lineages warrants species status—has not gone away in the genomic era. However, traditional species tree methods using the MSC need not use "good" species as OTUs; they will work perfectly well on lineages that have recently diverged, so long as they have ceased exchanging genes. The key issue is not whether the OTUs in species tree analyses are in fact species but rather whether they have ceased exchanging genes, which has been shown to compromise traditional MSC methods [83, 84] (see below).
The problem of species delimitation may ultimately be solved by data other than genetics, and today few species concepts use strictly genetic criteria [85]. Some have suggested that the line between a population-level difference and a species-level difference can be drawn empirically and with consistency in well-studied taxa such as birds, using morphological, environmental, and behavioral data simultaneously [86]. Thus, there is some hope that species delimitation can be performed rigorously a priori in many cases. Researchers who opt for delimiting species primarily with molecular data have a wide array of techniques and prior examples available to them, although not all without controversy [71, 87–93]. Recent progress in species delimitation is motivated by the conceptual transition from "biological/reproductive isolation species" to the "lineage species concept," which defines species not in terms of monophyly of gene lineages but as population lineage segments in the species tree [93]. Under that expanded concept, boundaries of species (i.e., lineages in the species tree) can be facilitated by collection and analysis of gene trees in the framework of the multispecies coalescent model [72]. The recent suggestion that coalescent species delimitation methods define only structure but not species [90] was, in our view, already well-established, with confusion stemming largely from the term "species delimitation," as opposed to "delimitation of populations between which gene flow has ceased."
Gene Flow There are a number of other situations in which the assumptions of the coalescent are violated. MSC models involve a series of isolation events unaccompanied by gene flow. In this regard, they are like the isolation-migration models of phylogeography [94, 95] but without the migration. The assumption of no gene flow naturally restricts their utility, but gene flow of course compromises other methods of phylogenetic inference, including concatenation methods, as well. Additionally, situations in which gene flow yields a prominent molecular signal often are detectable primarily among very closely related species in the realm of phylogeography [96]. If some substantial gene flow continues between species after divergence, then the multispecies coalescent can quickly destabilize, especially for a small number of loci and as the rate of genetic introgression increases (Fig. 6 in [87, 97–99]). We recommend model comparison algorithms like PHRAPL [87] for determining whether a given data set conforms to the assumptions of the MSC.
2.1.2 Molecular Processes
In addition to species delimitation and gene flow, there are at least three mechanisms that generate discordance on the molecular level (Fig. 3). These include horizontal gene transfer (HGT), which can pose a serious risk to phylogenetic analysis; gene duplication, whose risks can be avoided by certain models; and natural selection, which generally poses no direct threat but, depending on its mode of action and consequences for DNA and protein sequences, can be the most challenging of all.
Fig. 3 Three examples of gene histories that depart from the standard multispecies coalescent model. (a) A duplication event that precedes a speciation event can lead to incorrect inference of divergence times in the species tree if copy 1 is compared to copy 2. This can be particularly difficult if one of the gene copies has been lost or not sequenced by the researcher. (b) Convergent evolution can occur at the molecular level, for example, in certain genes under strong natural selection or highly biased mutational processes. These processes will tend to bring together distantly related taxa in the phylogenetic tree and are likely to be given additional false support by morphological data. (c) Horizontal gene transfer causes difficulties in some current species tree methods, because it establishes a spurious lower bound to divergence times. Though rare in eukaryotes, it is by no means unknown and is likely to become a more difficult problem in the future when species trees are based on tens of thousands of loci
Horizontal Gene Transfer HGT is now known to be so widespread across the Tree of Life, especially in prokaryotes, that some have suggested a web of life may be a more appropriate paradigm for phylogenetic change [100–102]. Growing evidence shows that even eukaryotic genomes contain substantial amounts of "uploaded" genetic material from bacteria, archaea, viruses, and even fellow eukaryotes [103–105]. Even though effective techniques are not yet widely available for detecting HGT in eukaryotes, enough individual cases have been "accidentally" discovered that researchers have given up trying to list them all [103].
> The implications of HGT for species tree construction vary depending on the method used. For example, following the standard assumption in coalescent theory that allelic divergences must occur earlier in time than the divergences of species harboring those alleles, some species tree techniques [48, 58], as well as classical approaches (e.g., [13]), assume that the gene tree exhibiting the most recent divergence between taxon A and taxon B establishes a hard upper limit on the divergence time of those species in the species tree. For small sets of genes in taxa where HGT is rare, a researcher would need to be quite unlucky to choose a horizontally transferred gene for analysis. However, as the genomic era advances, it becomes more likely that at least one of the thousands of genes studied will have been transferred horizontally and thus establish a spurious upper bound for clade divergence at the species level. When selective introgression of genes from one species to another is considered, this number of genes coalescing recently between species will increase [106]. Although HGT is clearly a problem for some current methodologies, if transferred genes can first be identified, then they could be extremely useful as genomic markers for monophyletic groups that have inherited such genes and would otherwise be difficult to resolve [107]. However, for other species tree methods that calculate averages of coalescence times, such as STAR [65], HGT events will have less of an impact. Liu et al. [56] examined the effect of HGT on the pseudolikelihood method MP-EST and predicted that, mathematically, species tree branch lengths may be biased by HGT but that topologies were fairly robust. Davidson et al. [108] found that quartetbased methods, such as ASTRAL-II, were fairly robust to HGT in the presence of ILS. Removal of genes suspected to be transferred via HGT prior to species tree analysis would be warranted; however, some methods to detect such events rely both on having the true species tree already in hand and also on the absence of other mechanisms causing gene tree discordance [109–112]. Recent work aims to incorporate HGT into other mechanisms of gene tree incongruence (reviewed in [113]); how much we need to invest in such synthetic methods will likely depend on the prevalence of HGT in particular taxonomic groups.
Gene Duplication Gene duplication presents another violation of the basic MSC model (Fig. 3); like HGT, its potential problems are worst when they go unrecognized [49]. Imagine a taxon where a gene of interest duplicated 10 Mya into copy α and copy β; the taxon then split 5 Mya into species 1 and 2. A researcher investigating the daughter species would therefore sequence four orthologous genes, with the potential to compare α1 to β2 and β1 to α2 and thus generate two gene trees where the estimated split time was 10 Mya, rather than 5 Mya. Such a situation will be easily recognized if copy α and β have diverged sufficiently by the time of their duplication, and a number of methods of coalescent analysis have incorporated gene duplication (e.g., [114, 115]; reviewed in [116]). Additionally, failure to recognize the situation may not have drastic consequences for phylogenetic analysis if the paralogs have coalesced very recently or are species-specific, in which case the estimated gene coalescence would be approximately correct no matter which comparison was made. However, if one of the copies has been lost and only one of the remaining copies is sequenced, then the chances of inferring an inappropriately long period of genetic isolation are larger and will increase as the size of the family of paralogs increases. Assessing paralogs in phylogenomic data is a major challenge, particularly in groups like plants and fish, and a growing number of dedicated methods ([117]; assessed in [118]) or filtering protocols [119] for doing so exist. This problem will tend to overestimate gene coalescence times, and some species tree methods depend on minimum isolation times among a large set of genes. These deep coalescences might spuriously increase inferred ancestral population sizes. A systematic search for biases incurred by species tree methods due to gene duplication is needed.
Natural Selection Natural selection causes yet another violation of the multispecies coalescent model. Selection can cause serious problems in some cases, although in other circumstances it is predicted not to cause problems of phylogenetic analysis [47, 120]. The usual stabilizing selection can be helpful to taxonomists working at high levels because it slows the substitution rate; likewise selective sweeps, directional selection, and genetic surfing [121] tend to clarify phylogenetic relationships by accelerating reciprocal monophyly for genes in rapidly diverging clades. However, challenges to phylogenetic inference are posed by any evolutionary force that may bias the reconstruction of gene trees, including convergent neutral mutations (homoplasy), balancing selection, and selection-driven convergent evolution (e.g., [122]). Balancing selection tends to preserve beneficial alleles at a gene for long periods of time and is probably the most insidious form of selection with respect to accurately reconstructing gene trees and species trees.
#### 2.2 More About Violations and Model Fit of the Multispecies Coalescent Model
Many of the instances of violations of the coalescent model will occur at individual genes and usually will not dominate the signal of the entire suite of genes sampled for phylogenetic analysis. Reid et al. [123] conducted one of the few tests of the fit of the MSC to multilocus phylogenetic data. Although the title of their article suggests that the MSC overall provides a "poor fit" to empirical data, we suggest that their results provide a more hopeful picture. The most important thing is that they investigated the fit of the MSC to individual loci in phylogenetic data sets and were able to identify loci that failed to fit the MSC. They were less successful at identifying the causes of departure from the MSC for individual loci.
More common but still rare are efforts to determine which models of phylogenetic inference, the MSC or concatenation, provide a better fit to empirical phylogenomic data. Edwards et al. [124] and Liu and Pearl [58] both used the Bayesian species tree method BEST [68] to ask using Bayes factors whether the MSC or concatenation fits empirical data sets better. Uniformly, they found that the MSC fit empirical data sets better than concatenation, often by a large margin. However, further work in this area is still needed. Most discussions in the literature have focused on the perceived failings or violations of the MSC by empirical data sets—such as evidence for recombination within loci—even when such failings or assumptions also apply to concatenation [47]. Given that all models are approximations of reality, a better focus would be to ask which model better fits empirical data sets better. The limited research that has been done suggests overwhelmingly that the MSC provides a better fit to empirical data sets than concatenation.
Are there better models for phylogenomics than the MSC? Depending on the data set, almost surely there are (Fig. 4). Several authors working with phylogenomic data sets have suggested that gene flow is detectable, even among lineages that diverged a long time ago (e.g., [129, 130]). The increasing number of reports of hybridization and introgression among phenotypically distinct species suggests that hybridization may be a typical component of speciation and that even phylogenetic models can be improved by incorporating such reticulation (e.g., [47, 106, 131]). The pure MSC is best thought of as a special case of so-called "multispecies network coalescent" models, or MSNC [127, 132–134] (Fig. 4), in which gene flow connects some branches of the species tree. In the end, empiricists will need to decide what level of model fit they are willing to tolerate and which software packages can accommodate the large data sets that are now routine in phylogenomics.
2.2.1 Phylogenetic Outlier Loci Genes whose phylogenetic signal differs significantly from that of the remainder of data set can be thought of as phylogenetic outliers. These loci are conceptually similar to outliers in population genetics, which have been the focus of many studies (reviewed in
Gene flow: M=m/µ Divergence time: =µ*t*
Fig. 4 Diversity of phylogeographic models. Species trees estimated by the multispecies coalescent are naturally related to previous phylogeographic models by their shared demographic parameters, usually measured in units of mutation rate or substitutions per site (μ), including genetic diversity or effective population size (4Nμ, where N <sup>¼</sup> effective population size; gene flow M/μ, where M <sup>¼</sup> the scaled migration rate; 4Nm, where m is the number of migrants per generation; and divergence time <sup>τ</sup> <sup>¼</sup> <sup>μ</sup>t, where t is the divergence time in generations). (a) Equilibrium migration models as envisioned by early versions of the software MIGRATE [125]. (b) Isolation-migration models envisioned by Hey and coworkers [48, 95, 126]. Subscript A indicates ancestral population size. (c) Species tree models estimated by the multispecies coalescent [28]. (d) Multispecies network coalescent models or phylogenetic network models including divergence and gene flow [127, 128]
[135–137]). However, there has been little work in detecting phylogenomic outliers. Much attention has been paid to particular sites in a data set that differ from the majority and therefore exhibit homoplasy or incongruence with the rest of the data set [76, 138]. The sources of such incongruence are many and can include mutational processes (e.g., gene duplication), HGT, as well homoplasy (e.g., [139, 140]). Incongruence of particular sites, or entire loci, may also be due to technical issues such as contamination, misassembly, mistaken paralogy, annotation mistakes, and alignment errors (e.g., [119]). Here, in an analogy with work in population genetics, we will focus primarily on entire loci that deviate from the expected distribution governed by neutral processes due to natural selection. Understanding the distribution of gene tree topologies expected under the neutral multispecies coalescent [25] is a good starting point for identifying loci that may be targets of natural selection.
2.2.2 Genomic Signals of Phylogenetic Outliers When faced with a surprising or nonconvergent species tree, one possibility is that an unusual gene tree is to blame. Though techniques for dealing with violations of the coalescent model are in their infancy, researchers do have a few options. Below we list several ideas, some borrowed from classical phylogenetics or from methods used in bioinformatics. It is likely that the several tests constructed to detect phylogenetic outliers in classical phylogenetics can be extended slightly to incorporate the additional variation among genes expected due to the coalescent process. Of course, with larger data sets, at least with some coalescent methods, single anomalous genes may have little effect on the resulting species tree, particularly in species tree methods utilizing summary statistics [65]. However, as pointed out above, species tree methods such as BEST that relies on "hard" boundaries for the species tree by individual genes could be derailed due to the anomalous behavior of even a single gene.
> Jackknifing: A straightforward approach to detecting phylogenetic outliers under the multispecies coalescent model is to rerun the analysis n times, where n is the number of loci in the study, leaving one locus out each time. An outlier can then be identified if the analysis that does not include that gene differs from the remaining analyses in which that gene is included. This approach has been applied successfully in fruit flies by Wong et al. [21], who considered their problem resolved when the elimination of one of the ten genes unambiguously resolved a polytomy. There may be other metrics of success that are more robust or sensitive or do not depend as strongly on a priori beliefs about the relationships among taxa. Because some duplications or horizontal transfers may affect only one taxon, whole-tree topology summary statistics are unlikely to be sensitive enough to detect recent events. However, the cophenetic distance of each taxon to its nearest neighbor in the complete species tree could be compared across jackknife results. This procedure will produce a distribution of "typical" distances, and significance can therefore be assigned to highly divergent results. The drawback to such an approach is the
2.2.3 Simulation Approaches to Detecting Phylogenetic Outliers
computational demand. Species tree analyses on their own can be extremely time consuming to run even once, so jackknifing may prove intractable for studies involving many species and loci (see ref. 141).
Simulating gene trees from a species tree is another method for identifying gene trees that differ from the majority of loci in the data set. Several species tree methods yield estimate of the phylogeny that include branch lengths in coalescent units [56, 57, 70], which are required to simulate gene trees from a species tree. Branch lengths in the estimated species tree can be decomposed into a number of substitutions per site and an estimate of θ ¼ 4Nμ that are compatible with the original branch length in coalescent units. For example, using any number of algorithms, including maximum likelihood or Bayesian methods, the length of species tree branch lengths in substitutions per site can be approximated by fitting the concatenated alignment of genes to the estimated species tree topology, yielding a tree with the same topology but branch lengths in substitutions per site (μt, where t is the time span of the branch in either generations or years). With these branch lengths in hand, estimates of θ can then be applied to each branch so that the original coalescent units t/2N μt/θ from the species tree are retained. Care needs to be taken to preserve the appropriate ploidy units when simulating gene trees from an estimated species tree. Packages such as MP-EST yield estimates of species tree branch lengths in coalescent units of 4N generations, appropriate for diploids, whereas packages such as Phybase [142] simulate gene trees from a species tree in estimates of 2N units, appropriate for haploids. Another issue that is important to be aware of is the distinction between gene coalescence times and species tree branch lengths [143, 144]. Whereas species tree branch lengths are estimates of lineage or population branch lengths in the species tree, the DNA sequence alignment that is fitted to the species tree will yield branch lengths reflecting the coalescence time of genes in ancestral species. This discrepancy occurs because gene coalescence times by necessity predate and record a more ancient event than do species divergence times. The discrepancy may represent a small fraction of the branch length if species divergence times are large, but Angelis and dos Reis [143] have suggested that the discrepancy can be quite large even in comparisons of distantly related species, such as exemplars of mammalian orders. There is a great need for methods of molecular dating and combining fossils and DNA data that distinguish between gene coalescence times and speciation times, the latter of which is usually of primary interest.
Once the branch lengths of the species tree are prepared for simulation, gene trees can be simulated using a number of packages (Phybase, [142]; TreeSim, [145]; CoMus, [146]). Even packages traditionally used in phylogeography can be used to simulated gene trees on species trees, given the close relationship between species trees and phylogeographic models like isolation migration [147, 148]. One can then compare the distribution of gene tree topologies and branch lengths observed in one's data set with those simulated under the neutral coalescent model. A common approach is to calculate the distribution of Robinson-Foulds [149] distances among simulated gene trees and compare these to those observed in the original data set. Such approaches have been used to determine if a data set is consistent with the MSC or the percent of the observed gene tree variation that is explained by the MSC. Other statistics, such as the similarity in number of minority gene tree triplets produced by a given species tree at each node, can also be compared to the observed distribution. Song et al. [150] used coalescent simulations using Phybase to propose that the MSC could explain a large (>75%) fraction of the observed gene tree variation in a mammalian data set. Such simulations assume that the gene tree variation observed is biological in origin and not due to errors in reconstruction. They also noted that the near equivalence in frequency of minority triplets in gene trees at various nodes in the mammal tree suggested broad applicability of the neutral coalescent without gene flow or other complicating factors. Still, many papers observe some level of departure of the patterns in the observed data set from those expected under simulation. Usually the source of this departure is unknown. Natural selection or any other force such as HGT or anomalous mutation might be culprits in these cases. Heled et al. [151] proposed a simulation regime that incorporates gene flow between species and thus can be used to test for the effects of migration on gene trees and species tree estimation.
To detect possible phylogenetic outliers, Edwards et al. [152] applied a recently proposed method of detecting gene tree outliers, KDEtrees [153], to a series of phylogenomic data sets. KDEtrees uses the kernel density distribution of gene tree distances to estimate the 95% confidence limits on gene tree topologies in a given data set. Surprisingly, using default parameters, Edwards et al. [152] could not detect a higher-than-expected number of gene tree outliers in any data set, despite the fact that the data sets in several cases contained hundreds of loci. No data set possessed more than the expected 5% of outliers given the test implemented in KDEtrees. Clearly further work is needed to understand the pros and cons of various tests of phylogenetic outliers. For the time being, we can note the robustness of various species tree methods to phylogenetic outliers. One attractive prospect of algorithms for species tree construction that use summary statistics, such as STAR and STEAC, is that these methods are powerful and fast, yet they appear less susceptible to error due to deviations of single genes from neutral expectations. These methods do not utilize all the information in the data and hence can be less efficient than Bayesian or likelihood methods [52], yet they perform well with moderate amounts of gene tree outliers due to processes like HGT.
#### 3 Hypothesis Testing Using the Multispecies Coalescent Model
Hypothesis testing is a cornerstone of phylogenetic analysis but has received little attention in the context of the MSC (see ref. 154). Bayesian species tree inference [58, 59, 68–70] provides perhaps the most seamless approach to hypothesis testing. One can relatively easily assess the fit of the collected data to alternative tree topologies and compare the fit using Bayes factors or other approaches. One can also assess the fit of various models of analysis to the collected data [155]. Liu and Pearl [58] and Edwards et al. [124] used Bayes factors to determine whether concatenation or the MSC was a more appropriate model for several data sets; in all cases tested thus far, the MSC provides a far better fit to multilocus data (BF > 10) than does concatenation, in which all gene trees among loci are identical. Further work is needed to apply Bayes factors and likelihood ratio tests to multilocus data.
The bootstrap, introduced to phylogenetics by Felsenstein [156], is the most common statistic applied to phylogenetic trees [157]. In the era of multilocus phylogenetics, the "multilocus bootstrap" of Seo [158] has been recommended as a more suitable approach to assessing confidence limits than the traditional bootstrap. In the traditional bootstrap, sites within a locus, or a series of concatenated loci, are resampled with replacement to create pseudomatrices, which are then subjected to phylogenetic analysis, after which a majority rule consensus tree is usually made. By contrast, in the multilocus bootstrap, sites within loci and the loci themselves are resampled with replacement. In the context of the MSC, resampled pseudomatrices of the same number of loci as the original data set, which may contain duplicates of specific loci due to the random nature of the bootstrap, are then made into gene trees, from which a species tree can be made. The bootstrap and various other measures of branch-specific support [159] have been proposed as a means of assessing confidence in species trees made using the multilocus coalescent. Care should be taken in the comparison of different studies using different measures of support, since not all measures can be directly compared to one another. For example, as pointed out by Liu et al. [160], the measure of posterior support for ASTRAL trees proposed by Sayyari and Mirarab [159] is not the same as traditional bootstrap supports, and we do not yet know how they will scale under different conditions compared to the bootstrap. Edwards [161] summarized knowledge about the use of phylogenomic subsampling, in which data sets of increasing size or signal are analyzed so as to understand the stability and speed of approach to certainty of phylogenetic estimates under the MSC and under concatenation. He found that MSC methods tended to approach phylogenomic certainty more smoothly and monotonically than do concatenation methods, which jump around erratically in their certainty for sometimes conflicting topologies, especially when sampling smaller numbers of genes. Although we cannot simply translate many conclusions from the gene tree era of phylogenetics to the MSC era—for example, contrary to gene tree conclusions, it is not clear for MSC models that more taxa are always better than more loci [74]—many of these discussions about hypothesis testing echo early comparisons of posterior probabilities and bootstrap proportions used in the gene tree era of phylogenetics.
The bootstrap has always provided a means of hypothesis testing that is very indirect with respect to comparing alternative phylogenetic hypotheses. Aside from the tests allowed by Bayesian approaches, there have been few discussions of testing of alternative phylogenetic trees in the era of the multispecies coalescent. In this regard, the pseudo-likelihood model provided by MP-EST [56] provides a convenient framework for hypothesis testing using species trees. This framework is not available in most other species tree methods, including ASTRAL, STAR, and STEAC, since these methods do not employ a likelihood model. MP-EST takes advantage of the likelihood model of Rannala and Yang [48] to assess the fit of a species tree to a collection of gene trees and can thus be used to compare alternative species tree topologies and branch lengths directly.
To conduct a direct comparison of species trees using the likelihood ratio test, we first compare the likelihoods of two trees to find the most probable species tree that can explain the empirical set of gene trees. The likelihood of a set of gene trees given a species tree with branch lengths can be ascertained using functions in Phybase [142]. Let Tree 1 be the null tree and Tree 2 be the alternative tree. The likelihood ratio test statistic is t ¼ 2 (LTree2 LTree1), in which LTree1 and LTree2 are the log-likelihoods of the null and alternative hypotheses. The log-likelihood of the null hypothesis can be obtained from the output of the program MP-EST by fitting the branch lengths and topology of Tree 1 to the set of empirical gene trees. Similarly, we can find the log-likelihood of the alternative tree Tree 2 using MP-EST. The null distribution of the test statistic t is approximated by a parametric bootstrap. Specifically, we generate 100 or more bootstrap samples of gene trees under the null tree Tree 1. For each sample of these bootstrapped trees, we calculate the log-likelihoods of the null and alternative trees using the procedure described above. The null distribution of the test statistic t is approximated by the test statistics of the bootstrap samples. If t for the null and alternative species trees is outside the expected distribution of the bootstrap sample statistics, then the result can be considered significant.
We applied this approach to assessing alternative phylogenetic hypotheses to an example from birds (fairy wrens; [162]; Fig. 5). This data set consists of 18 genes and 26 taxa, with loci coming
Fig. 5 Example of hypothesis testing of alternative phylogenetic trees under the multispecies coalescent model. Top: alternative phylogenetic hypotheses involving the rearrangement of major groups of Australo-Papuan fairy wrens based on Lee et al. [162]. The three alternative phylogenetic trees are colored to indicate the three major groups whose relationships are being tested. Bottom: results of the likelihood ratio test (LRT) and estimates of confidence limits on the test statistic t using parametric bootstrapping. The plots show the distributions of the test statistic t resulting from gene trees built from resampled, bootstrapped sequence data. Despite the use of sequence data to generate the bootstrap gene tree distributions, the LRT is only an indirect test of the signal in the sequence data and instead is best thought of as a test of the fit of the estimated gene tree distribution on alternative phylogenies. See main text for further details
from a variety of marker types (exons, introns, anonymous loci). Lee et al. [162] applied a number of MSC approaches to this data set but did not compare alternative trees directly, having only used bootstrap approaches. Here, we consider three-species trees generated from the rearrangement of the three major clades of wrens: the core fairy wrens (Malurus), emu-wrens (Stipitirus), and grasswrens (Amytornis; Fig. 5). Rearranging these major clades results in three alternative rooted species trees. Based on traditional taxonomy and because the gene trees in this data set were highly variable, even among the three major clades, we consider these three alternative hypotheses true alternatives and not "straw men." Rooted maximum likelihood gene trees were built from the alignments of each locus using RaxML [163] and then used as input data for the likelihood ratio test described above. The LRT was applied first to Tree 1 (null) versus Tree 2 and was also applied to Tree 1 versus Tree 3 and Tree 2 versus Tree 3. The results indicate that Tree 1 fits to the empirical gene trees significantly better than does Tree 2 or Tree 3 does ( p < 0.01), and there is no significant difference between Trees 2 and 3 in their fit to the empirical gene trees ( p ¼ 0.52). Thus, the LRTs strongly favor Tree 1 over both Tree 2 and Tree 3.
It is important to note that the LRT described above is not a direct test of the phylogenetic signal in the DNA sequence data. Rather, it is a test of the distribution of gene trees inferred from the sequence data and assumes that the gene trees provided as data are without error. It does indirectly test the signal in the sequence data, because if the DNA sequences provide strong and consistent support of the gene trees, then the bootstrapped set of gene trees will be highly similar to one another, and the confidence limits on t will be very tight. By contrast, if the DNA sequence data does not have a strong signal, then the confidence limits on t will be very wide, and it will be difficult to reject alternative species trees. The LRT described here does not involve nested models. If the gene trees are known without error, then the value of t itself can be used to assess significance, assuming a chi-square distribution with 2 degrees of freedom. Further research is needed on methods for comparing and testing alternative species trees in the context of the MSC.
#### 4 Future Directions
Species tree methods are likely to continue to gain ascendancy as the strongest evidence of taxonomic relationship in phylogenetic research. As with any form of evidence, the conclusions of a species tree analysis are fallible, with each method susceptible to biases in the input data. For example, Xi et al. [164] showed that Phyml [165] yields biased gene trees when there is little information in the DNA sequences and can therefore result in biased species trees. This issue is particularly problematic when using MP-EST v. 1.5, which, unlike ASTRAL or MP-EST v. 2.0, does not randomly resolve or appropriately accommodate gene trees with polytomies or 0 or near 0-length branches. This bias may have affected the performance of MP-EST in previous side-by-side comparisons with ASTRAL. In the future, further work should be devoted to discovering and quantifying additional biases in inference of species trees. With the size of phylogenomic data sets increasing, even small biases can be amplified and result in poorly estimated species trees.
Many in the field agree that the most appealing statistical models for species tree inference using the MSC include Bayesian and full-likelihood models [52]. But it is still clear, at least to empiricists, not only that "two-step" methods of species tree inference work quite well in general but also that the large phylogenomic data sets available today prohibit the use of full-likelihood methods. Regardless, we now know that both types of models clearly outperform concatenation across wide swaths of parameter space, especially if one also evaluates the reliability of the confidence limits on the estimate of phylogeny and not only the point estimate of the topology. The major directions for future research in the field of species tree inference therefore include increasing the scalability of computational inference of species trees, further development of frameworks for hypothesis testing using the MSC, developing additional models of divergence with gene flow and network coalescent models (Fig. 4), and improvement in the estimation of gene trees and species trees from SNP data [166]. Linking mutations in species trees and heterogeneous gene trees to diverse phenotypic and ecological data will be another important avenue for the future [167, 168]. We view the MSC, with its application of population genetic models to higher-level systematics, as a key component of the long-term goal of uniting microevolution and macroevolution. Even if it proves incomplete in the long term, the neutral MSC provides a powerful null model for the understanding of genetic diversity across time and space.
#### 5 Practice Problems
BEST [68] or BPP [70] and the nonparametric method in STAR [65] to estimate the species tree for the four species, using Taeniopygia as the out-group. Do you estimate the same topology with both methods? What about the support for the single internal branch? If the support is not the same, what could be causing the difference? Answer: The BEST or BPP tree should have higher support than the STAR tree, but they both should have the same topology. The STAR tree might have lower support because in the data set about half of the gene trees have a topology differing from the species tree; whereas the full Bayesian model accommodates this variation accurately, nonparametric "two-step" methods interpret this type of gene tree variation as discordance, in conflict with the majority of the gene trees and with the species tree.
3. For the above data set, make individual gene trees using RaXml [170], and use the likelihood functions and bootstrap capabilities of Phybase [142] to conduct a likelihood ratio test of the two alternative species tree topologies for the four grass finches. Alternatively, you could use the posterior distribution of gene trees generated in BEST to estimate the confidence limits on the test statistic t. Is the tree estimated in question 2 significantly better than alternative trees? Answer: The LRT indicates that the tree estimated in question 2 is significantly better than alternative trees.
#### References
DM, Moritz C, Mable BK (eds) Molecular systematics. Sinauer, Sunderland, MA
gene lineages on a species tree. Proc Natl Acad Sci U S A 113:8002–8009
incongruence in molecular phylogenies. Nature 425:798–804
groups from orthologous gene pairs. PLoS One 8:e53786
zombie lineages do not compromise the dating of placental diversification. Proc Natl Acad Sci U S A 114:E9433–E9434
maximum likelihood. Infect Genet Evol 9:384–385
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
#### Genome-Wide Comparative Analysis of Phylogenetic Trees: The Prokaryotic Forest of Life
#### Pere Puigbo` , Yuri I. Wolf, and Eugene V. Koonin
#### Abstract
Genome-wide comparison of phylogenetic trees is becoming an increasingly common approach in evolutionary genomics, and a variety of approaches for such comparison have been developed. In this article we present several methods for comparative analysis of large numbers of phylogenetic trees. To compare phylogenetic trees taking into account the bootstrap support for each internal branch, the boot-split distance (BSD) method is introduced as an extension of the previously developed split distance (SD) method for tree comparison. The BSD method implements the straightforward idea that comparison of phylogenetic trees can be made more robust by treating tree splits differentially depending on the bootstrap support. Approaches are also introduced for detecting treelike and netlike evolutionary trends in the phylogenetic Forest of Life (FOL), i.e., the entirety of the phylogenetic trees for conserved genes of prokaryotes. The principal method employed for this purpose includes mapping quartets of species onto trees to calculate the support of each quartet topology and so to quantify the tree and net contributions to the distances between species. We describe the applications methods used to analyze the FOL and the results obtained with these methods. These results support the concept of the Tree of Life (TOL) as a central evolutionary trend in the FOL as opposed to the traditional view of the TOL as a "species tree."
Key words Forest of Life, Tree of Life, Phylogenomic methods, Tree comparison, Map of quartets
#### Abbreviations
BSD Boot-split distance CMDS Classical multidimensional scaling COG Clusters of orthologous genes FOL Forest of Life HGT Horizontal gene transfer ND Nodal distance NUTs Nearly universal trees QT Quartet topology
Electronic supplementary material: The online version of this chapter (https://doi.org/10.1007/978-1-4939- 9074-0\_8) contains supplementary material, which is available to authorized users.
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_8, © The Author(s) 2019
#### 1 Introduction
With the advances of genomics, phylogenetics entered a new era that is noted by the availability of extensive collections of phylogenetic trees for thousands of individual genes. Examples of such tree collections are the phylomes that encompass trees for all sufficiently widespread genes in a given genome [1–4] or the "Forest of Life" (FOL) that consists of all trees for widespread genes in a representative set of organisms [5]. It has been known since the early days of phylogenetics that trees built on the same set of species often have different topologies, especially when the set includes distant species, most notably, in prokaryotes [6, 7]. The availability of "forests" consisting of numerous phylogenetic trees exacerbated the problem as an enormous diversity of tree topologies has been revealed. The inconsistency between trees has several major sources: (1) problems with ortholog identification caused primarily by cryptic paralogy; (2) various artifacts of phylogenetic analysis, such as long branch attraction (LBA); (3) horizontal gene transfer (HGT); and (4) other evolutionary processes distorting the vertical, treelike pattern such as incomplete lineage sorting and hybridization [1, 8–10]. In order to obtain robust results in genomelevel phylogenetic analysis, for instance, to classify phylogenetic trees into clusters with (partially) congruent topologies or to identify common trends among multiple trees, reliable methods for comparing trees are indispensable.
The number and diversity of tree comparison methods and software have substantially increased in the last few years. The tree comparison methods variously use tree bipartitions, such as partition or symmetric difference metrics [11] and split distance [12]; distance between nodes such as the path length metrics [13], nodal distance [12, 14], and nodal distance for rooted trees [15]; comparison of evolutionary units such as triplets and quartets [16]; subtransfer operations such as subtree transfer distance [17], nearest-neighbor interchanging [18], subtree prune and regraft (SPR) using a rooted reference tree [19], SPR for unrooted trees [20] and tree bisection and reconnection (TBR) [17], and matching pair (MP) distance [21]; (dis)agreement methods such as agreement subtrees [22], disagree [12], corresponding mapping [23], and congruence index [24]; tree reconciliation [25]; and topological and branch lengths methods such as K-tree score [26]. Several algorithms have been proposed to analyze with multi-family trees. For example, the From Multiple to Single (FMTS) algorithm systematically prunes each gene copy from a multi-family tree to obtain all possible single-gene trees [12] and an algorithm implemented in TreeKO prunes nodes from the input rooted trees in which duplication and speciation events are labeled [27]. Another algorithm employs a variant of the classical Robinson-Foulds method to compare phylogenetic networks [28]. However, to the best of our knowledge, none of the available metrics for tree comparison takes into account the robustness of the branches, a feature that appears important to minimize the impact of artifacts (unreliable parts of a tree) on the outcome of comparative tree analysis. Here, we present the boot-split distance (BSD) method that calculates distances between phylogenetic trees with weighting based on bootstrap values. This method is implemented in the program TOPD/FMTS [12]. In our recent research, we used the BSD method combined with classical multidimensional scaling (CMDS) analysis to explore the main trends in the phylogenetic FOL and to explore the "Tree of Life" (TOL) concept in light of comparative genomics [5, 29].
Since the time (ca 1838) when Darwin drew the famous sketch of an evolutionary tree in his notebook on transmutation of species, with the legend "I think...," the thinking on the "Tree of Life" (TOL) has evolved substantially. The first phylogenetic revolution, brought about by the pioneering work of Zuckerkandl and Pauling [30] and later Woese and coworkers [31], was the establishment of molecular sequences as the principal material for phylogenetic tree construction. The second revolution has been triggered by the advent of comparative genomics when it has been realized that HGT, at least among prokaryotes, was much more common than previously suspected. The first revolution was a triumph of the tree thinking, when a well-resolved TOL started to appear within reach. The second revolution undermines the very foundation of the TOL concept and threatens to destroy it altogether [32–34].
The current views of evolutionary biologists on the TOL span the entire range from acceptance to complete rejection, with a host of moderate positions. The following rough classification may be used to summarize these positions (a) acceptance of the TOL as the dominant trend in evolution: HGT is considered to be rare and overhyped, and most of the observed "transfers" are deemed to be artifacts [35–38]; (b) the TOL is the common history of the (nearly) nontransferable core of genes, surrounded by "vines" of HGT [39–50]; (c) each gene has its own evolutionary history blending HGT and vertical inheritance; a statistical trend might exist in the maze of gene histories, and it could even be treelike [5, 29, 51, 52]; and (d) ubiquity of HGT renders the TOL concept totally obsolete (prokaryotic species and higher taxa do not exist, and microbial "taxonomy" is created by a pattern of biased HGT) [32, 34, 53–58].
Fig. 1 A schematic of the methods and concepts involved in the FOL analysis
We found that, although different trends and patterns have to be invoked to describe the FOL in its entirety, the main, most robust trend is the "statistical TOL," i.e., the signal of coherent topology that is discernible in a large fraction of the trees in the FOL, in particular, among the nearly universal trees (NUTs) [59, 60].
We further explored the FOL by analysis of species quartets [61]. A quartet is a group of four species which is the minimum evolutionary unit in unrooted phylogenetic trees; each quartet can assume three unrooted tree topologies [16]. We described a quantitative measure of the tree and net signals in evolution that is derived from an analysis of all quartets of species in all trees of the FOL. The results of this analysis indicate that, although diverse routes of netlike evolution jointly dominate the FOL, the pattern of treelike evolution that recapitulates the consensus topology of the NUTs is the single most prominent, coherent trend. Here, we report an extended version of these methodologies introduced to analyze the FOL and its trends, as well as new concepts of prokaryotic evolution under the FOL perspective (Fig. 1).
#### 2 Materials
#### 2.1 The Forest of Life (FOL) and Nearly Universal Trees (NUTs)
We analyzed the set of 6901 phylogenetic trees from [5] that were obtained as follows. Clusters of orthologous genes were obtained from the COG [62] and EggNOG [63] databases from 100 prokaryotic species (59 bacteria and 41 archaea). The species were selected to represent the taxonomic diversity of Archaea and Bacteria (for the complete list of species, see Additional File 1). The BeTs algorithm [62] was used to identify the orthologs with the highest mean similarity to other members of the same cluster ("index orthologs"), so the final clusters contained 100 or fewer genes, with no more than one representative of each species. The sequences in each cluster were aligned using the Muscle program [64] with default parameters and refined using Gblocks [65]. The program Multiphyl [66], which selects the best of 88 amino acid substitution models, was used to reconstruct the maximum likelihood tree of each cluster. The nearly universal trees (NUTs) are defined as trees from COGs that are represented in more than 90% of the species included in the study.
#### 3 Methods
3.1 Boot-Split Distance: A Method to Compare Phylogenetic Trees Taking into Account Bootstrap Support
3.1.1 Boot-Split Distance (BSD)
The BSD method compares trees based on the original split distance (SD) [12] method. Both methods work by collecting all possible binary splits of the two compared trees and calculating the fraction of equal splits, i.e., those splits that are present in both trees (different splits refer to splits that are present in only one of the two trees). Instead of considering all branches as being equal as is the case in SD, the BSD method takes into account the bootstrap values to increase or decrease the SD value proportionally to the robustness of individual internal branches. The BSD value is the average of the BSD in the equal splits (eBSD) and the BSD in the different splits (Eq. 1). Equations 2 and 3 give the formulas to calculate the eBSD and dBSD values, respectively.
$$\text{BSD} = \frac{\text{cBSD} + \text{dBSD}}{2} \tag{1}$$
$$\text{cBSD} = 1 - \left[\frac{\varepsilon}{a} M\_{\varepsilon}\right] \tag{2}$$
$$\text{dBSD} = \frac{d}{\mathfrak{a}} \cdot M\_d \tag{3}$$
Here e is the sum of bootstrap values of equal splits, d is the sum of bootstrap value of different splits, a is the sum of all bootstrap values, Me is the mean bootstrap value of equal splits, and Md is the mean bootstrap value of different splits.
The BSD algorithm proceeds in four basic steps to compare pairs of trees (Fig. 2). The first step is to obtain all possible splits from both trees. This procedure implies a binary split of the tree at each internal branch, so that the tree is partitioned into two parts each of which contains at least two species. Then, the common set of leaves between the two trees is obtained, that is, the set of shared species. Only trees with a common leaf set of at least four species can be compared. The third step consists in pruning all splits to the common leaf set of species; at this step, species that are present in only one of the two compared trees are removed from the split list. After this procedure, in partially overlapping trees, the algorithm checks whether each of the splits remains a valid partition, that is, a partition that separates at least two species from the rest of the tree. If a split is not a valid partition, it is removed. Finally, the algorithm calculates the BSD using Eqs. 1–3.
Fig. 2 The main algorithm of the BSD method. The algorithm to calculate the BSD between two trees includes four basic steps: (1) split both trees in all possible partitions, (2) read the common set of species of both trees, (3) prune the splits according with the common leaf set, and (4) calculate the BSD
3.1.2 The BSD Algorithm There are three possible types of comparisons for trees that do not include paralogs, that is, include one and only one sequence from each of the constituent species (Fig. 3). In the first case, the two trees completely overlap, that is, consist of the same set of species (Fig. 3a). In this case, step 2, the pruning procedure, is not necessary, and the comparison involves only obtaining all possible splits and the calculation of the BSD. In the second case, one of the compared trees is a subset of the other tree (Fig. 3b). In this case, the splits are only pruned and occasionally removed from the bigger tree. In the third case, when the two trees partially overlap or when a tree is a subset of another tree, a pruning procedure is required. In the example shown in Fig. 4, after the pruning procedure (step 3), there is only one remaining split (split: AB|CD) that is repeated several times in both trees. The remaining AB|CD split in Tree 1 is separated by four nodes that have different bootstrap values. In this case, the bootstrap of the remaining split is calculated using Eq. 4, where n is the total number of nodes between the two sides of the split and BSi is the bootstrap value (adjusted to the 0–1 range) of the node i.
$$\text{Bootstrap} = \text{l} - \Pi\_{i=1}^{\ast} (\text{l} - \text{BS}\_i) \tag{4}$$
The bootstrap value associated with a particular branch of a binary tree is taken as a measure of the probability that the four
Fig. 3 Examples of the BSD algorithm in single family trees. (a) Two trees of the same size. (b) Tree 1 is a subtree of the Tree 2. Two trees that partially overlap. SD split distance, BSD boot-split distance, eBSD BSD of equal splits, dBSD BSD of different splits, p number of equal splits, q number of different splits, m total number of splits, a sum of bootstraps in all splits, e sum of bootstraps in equal splits, d sum of bootstraps in different splits, Ma mean bootstrap value, Me mean bootstrap value in equal splits, Md mean bootstrap value in different splits
subtrees on the opposite ends of this branch are partitioned correctly. To estimate the probability of the correct partitioning of an arbitrary set of four subtrees, the internal branch of the quartet tree is mapped onto each of the internal branches of the original tree. The quartet is considered to be resolved correctly if it is resolved correctly relative to any of these branches. Under the assumption that bootstrap probabilities on individual branches are independent, Eq. 4 is obtained as the estimate of the bootstrap probability for the internal branch of the quartet tree.
3.1.3 Using a Bootstrap Threshold: Pros and Cons The key question regarding the BSD method is as follows: what is the best approach to phylogenetic tree comparison—using all branches, reliable or not, with the appropriate weighting, or using only branches supported by high bootstrap values? The first option is illustrated in Fig. 3, whereas Fig. 5 shows an example of a tree comparison that employs a bootstrap threshold of 70, i.e., only branches supported by a higher bootstrap are taken into account in the comparison. The second procedure appears reasonable and can be recommended in some cases. However, it is not advisable as a general approach because, when two large trees with varying
Fig. 4 Calculation of BSD for trees with an unequal numbers of species. The larger tree (1) is pruned prior to the calculation of BSD. The bootstrap value for the only shared internal branch is calculated according to Eq. 4
Fig. 5 Example of the BSD algorithm using a bootstrap cutoff. The figure shows the comparison of two phylogenetic trees that takes into account only those branches with bootstrap support greater than 70. SD split distance, BSD bootsplit distance, eBSD BSD of equal splits, dBSD BSD of different splits, p number of equal splits, q number of different splits, m total number of splits, a sum of bootstraps in all splits, e sum of bootstraps in equal splits, d sum of bootstraps in different splits, Ma mean bootstrap value, Me mean bootstrap value in equal splits, Md mean bootstrap value in different splits
bootstrap values are compared, using a strict threshold restricts the comparison to a small subset of robust branches, resulting in an artificially low BSD value. In other words, this procedure artificially inflates the similarity between the two trees by depreciating a large fraction of the branches. In addition, before considering the use of only most supported branches, one should take into account that the BSD method already uses bootstrap values to adjust the distance between trees, so if two trees are topologically similar (low SD) but supported by low bootstrap, the distance value increases (higher BSD), which is one of the advantages of the BSD method (see Eqs. 2 and 3).
3.1.4 Testing the BSD Method The performance of the BSD method was compared with that of the original SD method implemented in the TOPD/FMTS program [12]. Figure 6 shows the correlation of SD and BSD for trees with a number of species from 4 to 15 (a) and from 16 to 100 (b) from a recent large-scale analysis of the FOL [5]. The three-way comparison of SD, BSD, and tree size (number of species) shows a positive correlation between SD and BSD for all tree sizes (R<sup>2</sup> <sup>¼</sup> 0.8613 for trees with 4–16 species and <sup>R</sup><sup>2</sup> <sup>¼</sup> 0.7055 for trees with 16–100 species) (Fig. 6c). However, the SD follows a discrete distribution, which obviously is most conspicuous in the comparisons of small trees (Fig. 6a), whereas, thanks to the use of the bootstrap values, the BSD distribution is continuous (Fig. 7).
Figure 7 shows an example of the comparison (all-against-all) of three trees with six species each that differ in one, two, and three splits, resulting in SD values of 0.33, 0.66, and 1, respectively (Fig. 7a). Also, each tree was compared to itself resulting in a SD of 0. Then, bootstrap values were assigned randomly to the trees in order to compare the trees using the BSD method, and this procedure was repeated 1000 times. The resulting plot (Fig. 7b) shows that, for the comparison of trees with SD of 0 and 1, the BSD values ranged from 0 to 0.5 and from 0.5 to 1, respectively, and in principle, could assume all intermediate values. In the case of the comparisons that differed in one split (SD ¼ 0.33), the BSD value was greater than 0.33 in 75% of the comparison, whereas for the comparisons that differed in two splits (SD ¼ 0.67), 25% of the BSD values were greater than 0.67. Thus, the BSD method for tree comparison offers a better resolution than the SD method, especially, for trees with a small number of species.
Figure 8a shows the results of analysis of six simulated alignments with an increasing level of noise (divergence respect to the initial alignment) in each alignment, i.e., from the alignment 0 (without noise and producing trees with bootstrap values of 100) to alignment 5 with the maximum level of noise. For each alignment, a tree was constructed using the UPGMA method from the web server DendroUPGMA (http://genomes.urv.cat/
Fig. 6 Correlation of BSD and SD from the all-against-all tree comparisons of 6901 phylogenetic trees. (a) Trees containing 4–15 species. (b) Trees containing 16–100 species. (c) SD, BSD, and tree size for trees containing between 16 and 100 species
UPGMA). Distances were calculated using the Jaccard coefficient, and bootstraps were generated from 100 replicates. The results of the tree comparison (Fig. 8b) using three different methods, namely, nodal distance (ND), SD, and BSD, show that the BSD method presents a continuous distribution resulting in a better resolution of the distances than the other two methods. Indeed, the SD and ND methods fail to discern the similarity between trees after six changes, whereas the BSD method still reports discernible similarity (Fig. 8b). In order to compare the three tree comparison
Fig. 7 Comparisons of trees with six taxa. Bootstrap values were assigned randomly in each comparison
methods, the distance reported by each method was normalized to the maximum value in each case, i.e., after 46 changes (maximum number of changes in the simulation), the distance to the initial tree is 1.41, 0.30, and 0.42 for ND, SD, and BSD, respectively. All three distance values indicate that the trees are similar far above the random expectation, supporting the robustness of all methods, but the BSD method presents a better resolution in the tree comparison.
3.1.5 Analysis of Random Trees and the Significance of BSD Results To assess the significance of the tree comparison by the BSD method, we performed several tree comparisons using random trees containing between 4 and 100 species (Fig. 9). Each test is an all-against-all comparison of 1000 random trees (for complete results see Additional File 2). The results from random tree
Fig. 8 Comparison of six trees constructed from alignments with increasing noise levels. (a) Comparison of trees from six simulated alignments. The UPGMA tree from each alignment was reconstructed with the web
comparison have to be used to determine whether the detected similarities or differences between trees are significantly different from chance [12]. Figure 9 shows that the distance between random trees monotonically increases with the tree size up to a value of approximately 0.75 for BSD and approximately 0.999 for SD. In other words, although BSD is an extension of the SD method, the results obtained by the two methods are not directly comparable. Therefore, to assess whether the similarity between two trees is better than chance, one must consider the method used for the tree comparison (e.g. SD or BSD) and the size of the tree. For example, consider two trees with 15 species each for which the SD method reports a distance of 0.75. This value is far below randomness (Fig. 9), so the conclusion would be that the two trees are nonrandomly similar. However, if the same distance value (0.75) is reported by the BSD method, the conclusion would be the
Fig. 9 Random BSD and SD depending on the tree size. Results of the tree comparison of random trees (with different sizes ranging from 4 to 100 species) show that the BSD and SD increase up to 0.75 and 0.999, respectively
Fig. 8 (continued) server DendroUPGMA (http:/genomes.urv.cat/UPGMA) using the Jaccard coefficient as the measure of distance and generating 100 bootstraps replicates. Alignment 0 corresponds to the initial alignment without noise that perfectly separates all branches, resulting in a tree with bootstrap values of 100 for all internal nodes. Alignments 1 to 5 correspond to the derivatives of the initial alignment with increasing noise levels at each step. (b) Results of the comparison of each tree [1 to 5] with the initial tree (0). The trees were compared using three methods: split distance (SD), nodal distance (ND), and boot-split distance (BSD). For the purpose of comparison, the results obtained with each of the three methods were normalized to the maximum value in each case
opposite, namely, that the two trees are no more similar than two random trees of 15 species.
Another and probably the most important problem of the comparison of phylogenetic trees is how to interpret the results from a biological perspective. To address this issue, we generated random trees containing from 4 to 100 species and performed 1 to 100 permutations (swap of a pair of branches) in each tree. The resulting tree was then compared with the source tree (Fig. 10a, b). The results show the number of permutations required to obtain a particular BSD value for different tree sizes (number of species). For instance, BSD ¼ 0.3 in the comparison of two trees with
Fig. 10 The number of permutations and the BSD. (a) BSD depending on the number of permutations and tree size. (b) Mean and standard deviation of the BSD for up to 100 permutations for trees with 20 species
20 species indicates that the two trees are separated by one permutation whereas BSD ¼ 0.6 indicates that the trees are separated by approximately 9 permutations (for the complete listing of equivalences between BSD, SD and the number of permutations, see Additional File 3). Considering that each permutation corresponds to an HGT event, the BSD may be construed as the measure of the extent of HGT contributing to the topological difference between the compared trees. Given the discrete distribution of SD values, this measure cannot be used to infer the number of permutations with the same precision as BSD.
A key characteristic of the FOL is the degree of the topological (in) consistency between the constituent trees. To quantify this trend, we introduced the inconsistency score (IS), which is the fraction of the times that the splits from a given tree are found in all N trees that comprise the FOL. Thus, the IS may be naturally taken as a measure of how representative of the entire FOL is the topology of the given tree. The IS is calculated using Eqs. 5–7, where N is the total number of trees, X is the number of splits in the given tree, and Y is the number of times the splits from the given tree are found in all trees of the FOL.
$$\text{IS} = \frac{\frac{1}{\mathcal{T}} - \text{IS}\_{\text{min}}}{\text{IS}\_{\text{max}}} \tag{5}$$
$$\text{IS}\_{\text{min}} = \frac{1}{X \cdot N} \tag{6}$$
$$\text{IS}\_{\text{max}} = \frac{1}{X} - \text{IS}\_{\text{min}} \tag{7}$$
In addition to the calculation of a single value of IS for a given tree by comparing its topology to the topologies of rest of trees in the FOL, IS can be calculated along the depth of the trees, namely, split depth and phylogenetic depth. The split depth was calculated for each unrooted tree according to the number of splits from the tips to the center of the tree. The value of split depth ranged from 1 to 49 ([100 species/2] - 1). The phylogenetic depth was obtained from the branch lengths of a rescaled ultrametric tree, rooted between archaeal and bacterial species, and ranged from 0 to 1. The topology of the ultrametric tree was obtained from the supertree of the 102 NUTs using the CLANN program [67]. The branch lengths from each of the 6901 trees were used to calculate the average distance between each pair of species. The obtained matrix was used to calculate the branch lengths of the supertree of the NUTs. This supertree with branch lengths was then used to construct an ultrametric tree using the program KITSCH from the Phylip package [68] and rescaled to the depth range from 0 to 1. The resulting ultrametric tree was used for the analysis of the dependence of tree inconsistency on phylogenetic depth.
#### 3.2 Analysis of Topological Trends in a Set of Phylogenetic Trees
3.2.1 Calculation of the Tree Inconsistency
#### 3.2.2 Classical Multidimensional Scaling Analysis
#### 3.3 Analysis of Quartets of Species
3.3.1 Definition of Quartets and Mapping Quartets onto Trees
The classical multidimensional scaling (CMDS), also known as principal coordinate analysis, is the multifactorial method best suited to analyze matrices obtained from tree comparison methods like BSD and identify the main trends in a large set of phylogenetic trees. The CMDS embeds <sup>n</sup> data points implied by a [<sup>n</sup> <sup>n</sup>] distance matrix into an m-dimensional space (m < n) such that, for any k ∈ [1, m], the embedding into the first k dimensions is the best in terms of preserving the original distances between the points [69, 70]. In our analysis, the data points are distances between trees obtained using the BSD method. The choice of the optimal number of clusters is made using the gap statistics algorithm [71]. The number of clusters for which the value of the gap function for cluster k + 1 is not significantly higher than that for cluster k (z-score below 1.96, corresponding to 0.05 significance level) is considered optimal. The CMDS analysis was performed using the K-means function of the R package that implements the K-means algorithm. The CMDS approach has been previously employed by Hillis et al. for phylogenetic tree comparison, with the distances between trees calculated using the Robinson-Foulds distance [72].
The minimum evolutionary unit in unrooted phylogenetic trees is defined by groups of four species (or quartets), and each quartet may be best represented by the three possible unrooted tree topologies (Fig. 11a). A quartet defined by the set of species A, B, C, and D has three possible unrooted topologies: (1) AB|CD, (2) AC|BD, and (3) AD|BC. To analyze which quartet topology (QT) best represents the relationships among the four species in a quartet, each quartet was compared against the entire set of phylogenetic trees from 100 species (the FOL).
For 100 species, there are 3,921,225 quartets and, accordingly, 11,763,675 topologies (Fig. 11b). A mapping of quartets onto trees is produced using the SD method [12]. A binary version of this method was employed to compare quartets and trees (a quartet is represented in a tree when SD ¼ 0 and not represented when SD > 0). Figure 12a shows an example of quartet mapping onto a set of ten trees. Here q<sup>1</sup> is a resolved quartet, with the topology q1t<sup>1</sup> supported by eight of the ten trees. By contrast, for q2, three quartet topologies are equally supported, i.e., the topology of this quartet remains unresolved.
To analyze which of the three possible topologies best represents the almost four million quartets in the FOL, each quartet topology was compared with the entire set of 6901 trees, resulting in a total number of 8.12 <sup>10</sup><sup>10</sup> tree comparisons (Fig. 11b), and the number of trees that support each quartet topology was counted for the entire FOL or for the set of 102 NUTs (Fig. 11b).
Fig. 11 Quartets and quartet topologies. (a) Each quartet (qi ) is defined by a set of four species (different colors denote species) and may be represented by three possible unrooted tree topologies (qi ti ). (b) Quartet topologies (QT). In 100 species, the total number of quartets (Q) is 3,921,225. Each quartet may be represented by three distinct QTs, resulting in a total of 11,763,735 QTs. Each QT was mapped onto the FOL, i.e., for each QT, it was determined which of the three topologies is represented in each phylogenetic tree in the FOL (8.12 1010 comparisons). Modified from ref. <sup>61</sup>
rescaling each matrix of quartet distances to a 0–1 scale between the
3.3.2 Distance Matrices and Heat Maps Using the quartet support values for each quartet, a 100 100 between-species distance matrix was calculated as dij <sup>¼</sup> <sup>1</sup> - Sij/Qij where dij is the distance between two species, Sij is the number of trees containing quartets in which the two species are neighbors, and Qij is the total number of quartets containing the given two species. Then, this distance matrix was used to construct different heat maps using the matrix2png web server ([73], Fig. 12b). In contrast to the BSD method, which is best suited for the analysis of the evolution of individual genes, the distance matrices derived from maps of quartets are used to analyze the evolution of species and to disambiguate treelike evolutionary relationships and "highways" (preferential routes) of HGT. 3.3.3 The Tree-Net Trend (TNT) The quartet-based between-species distances were used to calculate the Tree-Net Trend (TNT) score. The TNT score is calculated by
Fig. 12 Mapping quartets. (a) Mapping quartets onto a set of ten trees. (b) A schematic of the procedure used to reconstruct a species matrix from the map of quartets
supertree-derived matrix (which is taken to represent solely the treelike evolution signal, hence the distance of 0) and the matrix obtained from permuted trees, with distance values around the random expectation of 0.67 (Fig. 13). Two situations may occur in the calculation of the TNT score depending on the relationship between the distance in the supertree matrix (Ds) and the distance in the random matrix (Dr ¼ 0.67). When Ds > Dr (e.g., in comparisons of archaea versus bacteria), <sup>S</sup>TNT <sup>¼</sup> (<sup>d</sup> - Dr)/ (Ds - Dr), where STNT is the TNT score and d is the distance between the two compared species in the matrix. When Ds < Dr (in comparisons between closely related species), <sup>S</sup>TNT <sup>¼</sup> <sup>1</sup> - ((<sup>d</sup> - Ds)/(Dr -Ds)).
#### 4 Phylogenetic Concepts in Light of Pervasive Horizontal Gene Transfer
4.1 Patterns in the Phylogenetic Forest of Life
The reconstruction of the evolutionary trends in the FOL is based on the idea that prokaryotes, effectively, share a common gene pool. This gene pool consists of genes with widely different ranges of phyletic spread, from universal to rare ones only present in a few species [74]. Thus, genes, as the elements of this gene pool, have
Fig. 13 The Tree-Net Trend (TNT). The figure shows a schematic of the TNT calculation and the rescaling procedure. Modified from ref. 61
their distinct evolutionary histories blending HGT and vertical inheritance (Fig. 14). In principle, the Forest of Life (FOL) encompasses the complete set of phylogenetic trees for all genes from all genomes. However, a comprehensive analysis of the entire FOL is computationally prohibitive (with over 1000 archaeal and bacterial genomes now available and the computational resources accessible to the authors, estimation of the phylogenetic tree for each gene represented in all these genomes would take weeks of computer time) so a representative subset of the trees needs to be selected and analyzed. Previously [5], we defined such a subset by selecting 100 archaeal and bacterial genomes, which are representative of all major prokaryote groups, and building 6901 maximum likelihood (ML) trees for all genes with a sufficient number of homologs and sufficient level of sequence conservation in this set of genomes; for brevity, we refer to this set of trees as the FOL. In this set of almost 7000 trees, only a very small portion of the forest is represented by nearly universal trees (Fig. 14). Furthermore, bacterial and archaeal universal trees are rare as well, as reflected in Fig. 14 by the small peaks around 41 and 59 species, i.e., all archaea and all bacteria, respectively. The dominant pattern in the major part of the FOL is completely different: the FOL is best represented by
Fig. 14 The Forest of Life (FOL). The distribution of the trees in the FOL by the number of species. Modified from ref. 5
Fig. 15 Distribution of the gene functions among the NUTs. The functional classification of genes was from the COG database [62]
numerous small trees, with about 2/3 of the trees including <20 species (Fig. 14).
4.2 The Nearly Universal Trees (NUTs) We define the nearly universal trees (NUTs) as trees for those COGs that were represented in more than 90% of the included prokaryotes. This definition yielded 102 NUTs. Not surprisingly, the great majority of the NUTs are genes encoding proteins involved in translation and the core aspects of transcription (Fig. 15). Among the NUTs, only 14 corresponded to COGs that consist of strict 1:1 orthologs (all of them ribosomal proteins), whereas the rest of NUTs included paralogs in some organisms (only the most conserved paralogs were used for tree construction [5]). The 1:1 NUTs were similar to the rest of the NUTs in terms of the connectivity in tree similarity (1-BSD) networks and their positions in the single cluster of NUTs obtained using CMDS.
> The 102 NUTs were compared to trees produced by analysis of concatenations of universal proteins [49]. The results showed that
most of the NUTs were topologically similar to a tree obtained by the concatenation of 31 universal orthologous genes [5]—in other words, the "Universal Tree of Life" constructed by Ciccarelli et al. [49] was statistically indistinguishable from the NUTs and showed properties of a consensus topology. Not surprisingly, the 1:1 ribosomal protein NUTs were even more similar to the universal tree than the rest of the NUTs, in part because these proteins were used for the construction of the universal tree and, in part, presumably because of the low level of HGT among ribosomal proteins.
4.3 The Tree of Life (TOL) as a Central Trend in the FOL We analyzed the matrix of all-against-all tree comparisons of the NUTs by embedding them into a 30-dimensional tree space using the CMDS procedure [69, 70]. The gap statistics analysis [71] reveals a lack of significant clustering among the NUTs in the tree space. Thus, all the NUTs seem to belong to one unstructured cloud of points scattered around a single centroid. This organization of the tree space is best compatible with individual trees randomly deviating from a single, dominant topology (which may be denoted the TOL), apparently as a result of random HGT (but in part possibly due to random errors in the tree-construction procedure). Therefore, there is an unequivocal general trend among the NUTs. Although the topologies of the NUTs were, for the most part, not identical, so that the NUTs could be separated by their degree of inconsistency (a proxy for the amount of HGT), the overall high consistency level indicated that the NUTs are scattered in the close vicinity of a consensus tree, with HGT events distributed randomly [5].
> Thus, the NUTs present a unique and strong signal of unity that seems to reflect the TOL pattern of evolution. The inconsistency score (IS) among the NUTs ranged from 1.4% to 4.3%, whereas the mean IS value for an equivalent set (102) of randomly generated trees with the same number of species was approximately 80%, indicating that the topologies of the NUTs are highly consistent and nonrandom [5].
> To further assess the potential contribution of phylogenetic analysis artifacts to observed inconsistencies between the NUTs, we analyzed these trees with different bootstrap support thresholds (i.e., only splits supported by bootstrap values above the respective threshold value were compared). Particularly low IS levels were detected for splits with high bootstrap support, but the inconsistency was never eliminated completely, suggesting that HGT is a significant contributor to the observed inconsistency among the NUTs (IS ranges from 0.3% to 2.1% and 0.3% to 1.8% for splits with a bootstrap value higher than 70 and 90, respectively) [5].
> Analysis of the supernetwork built from the 102 NUTs [5] showed that the incongruence among these trees is mainly concentrated at the deepest levels, with a much greater congruence at shallow phylogenetic depths. The major exception is the
unambiguous archaeal-bacterial split that is observed despite the apparent substantial interdomain HGT. Evidence of probable HGT between archaea and bacteria was obtained for approximately 44% of the NUTs (13% from archaea to bacteria, 23% from bacteria to archaea, and 8% in both directions), with the implication that HGT is likely to be even more common between the major branches within the archaeal and bacterial domains [5]. These results are compatible with previous reports on the apparently random distribution of HGT events in the history of highly conserved genes, in particular those encoding proteins involved in translation [75, 76], and on the difficulty of resolving the phylogenetic relationships between the major branches of bacteria [77–79] and archaea [5, 80, 81]. More specifically, archaeal-bacterial HGT has been inferred for 83% of the genes encoding aminoacyl-tRNA synthetases (compared with the overall 44%), essential components of the translation machinery that are known for their horizontal mobility [42, 82]. In contrast, no HGT has been predicted for any of the ribosomal proteins, which belong to an elaborate molecular complex, the ribosome, and hence appear to be non-exchangeable between the two prokaryotic domains [42, 76]. In addition to the aminoacyl-tRNA synthetases, and in agreement with many previous observations ([83] and references therein), evidence of HGT between archaea and bacteria was seen also for the few metabolic enzymes that belonged to the NUTs, including undecaprenyl pyrophosphate synthase, glyceraldehyde-3-phosphate dehydrogenase, nucleoside diphosphate kinase, thymidylate kinase, and others.
4.4 The NUTs Topologies as the Central Trend and Detection Distinct Evolutionary Patterns in the FOL
Using the BSD method, we compared the topologies of the NUTs to those of the rest of the trees in the FOL. Notably, 2615 trees (~38% of the FOL) showed a greater than 50% similarity (P-value <0.05) to at least one of the NUTs, being the mean similarity of the trees to the NUTs approximately 50% (Fig. 16). For a set of 102 randomized trees of the same size as the NUTs, only about 10% of the trees in the FOL showed the same or greater similarity, indicating that the NUTs were strongly and nonrandomly connected to the rest of the FOL.
We then analyzed the structure of the FOL by embedding the 3789 COG trees into a 669-dimensional space using the CMDS procedure [69, 70]. A CMDS clustering of the entire set of 6901 trees in the FOL was beyond the capacity of the R software package used for this analysis; however, the set of COG trees included most of the trees with a large number of species for which the topology comparison is most informative. A gap statistics analysis [69, 70] of K-means clustering of these trees in the tree space revealed distinct clusters of trees in the forest. The FOL is optimally partitioned into seven clusters of trees (the smallest number of clusters for which the gap function did not significantly increase with the increase of the number of clusters) (Fig. 17). Clusters 1, 4, 5, and 6 were enriched
Fig. 16 Topological similarity between the NUTs and the rest of the FOL. Percentage of trees connected to the NUTs at a different percentage of similarity. Modified from ref. 5
Fig. 17 Clusters and patterns in the FOL. The seven clusters identified in the FOL using the CMDS method and the mean similarity values between the 102 NUTs and all trees from each of the seven clusters are shown. Modified from ref. 5
for bacterial-only trees, all archaeal-only trees belonged to clusters 2 and 3, and cluster 7 consisted entirely of mixed archaeal-bacterial clusters; notably, all the NUTs form a compact group inside cluster 6.
The results of the CMDS clustering (Fig. 17) support the existence of several distinct "attractors" in the FOL. However, we have to emphasize caution in the interpretation of this clustering because trivial separation of the trees by size could be an important contribution. The approaches to the delineation of distinct "groves" within the forest merit further investigation. The most salient observation for the purpose of the present study is that all the NUTs occupy a compact and contiguous region of the tree space and, unlike the complete set of the trees, are not partitioned into distinct clusters by the CMDS procedure. Taken together with the high mean topological similarity between the NUTs and the rest of the FOL, these findings indicate that the NUTs represent a valid central trend in the FOL.
#### 4.5 The Tree and Net Components of Prokaryote Evolution
The TNT map of the NUTs was dominated by the treelike signal (green in Fig. 18a): the mean TNT score for the NUTs was 0.63 (Fig. 19b), so the evolution of the nearly universal genes of
Fig. 18 The Tree-Net Trend (TNT) score heatmaps. (a) The 102 NUTs. (b) The FOL without the NUTs (6799 trees). The TNT increases from red (low score, close to random, an indication of netlike evolution) to green (high score, close to the supertree topology, an indication of treelike evolution). The species are ordered according to the topology of the supertree of the 102 NUTs. In (a), the major groups of archaea and bacteria are denoted. Modified from ref. 61
Fig. 19 The Tree-Net Trend in the FOL and in the NUTs. (a) A hypothetical equilibrium between the tree and net trends. (b) A schematic representation of the tree tendency in the NUTs. (c) A schematic representation of the net tendency in the FOL
prokaryotes appears to be almost "two-third treelike" (i.e., reflects the topology of the supertree). The rest of the FOL stood in a stark contrast to the NUTs, being dominated by the netlike evolution, with the mean TNT value of 0.39 (Fig. 19c) (about "60% netlike"). Remarkably, areas of treelike evolution were interspersed with areas of netlike evolution across different parts of the FOL (Fig. 18b). The major netlike areas observed among the NUTs were retained, but additional ones became apparent including Crenarchaeota that showed a pronounced signal of a non-treelike relationship with diverse bacteria as well as some Euryarchaeota (Fig. 18b). The distribution of the tree and net evolutionary signals among different groups of prokaryotes showed a striking split among the NUTs: among the archaea, the tree signal was heavily dominant (mean TNTNUTs\_Archaea ¼ 0.80 0.20), whereas among bacteria the contributions of the tree and net signals were nearly equal (mean TNTNUTs\_Bacteria ¼ 0.51 0.38). Among the rest of the trees in the FOL, archaea also showed a stronger tree signal than bacteria, but the difference was much less pronounced than it was among the NUTs (mean TNTFOL\_Archaea ¼ 0.47 0.11 and mean TNTFOL\_- Bacteria ¼ 0.34 0.08). The conclusions on the treelike and netlike components of evolution made here are based on the assumption that the supertree of the NUTs represents the treelike (vertical) signal. We did not perform direct tests of the robustness of these conclusions to the supertree topology. However, observations presented previously [5] suggest that the results are likely to be robust given the coherence of the NUTs topologies as well as the similarity of the supertree topology and the topologies of the individual NUTs to the "Tree of Life" obtained from concatenated sequences of universally conserved ribosomal proteins [49].
#### 5 Conclusions
The analysis of the phylogenetic FOL is a logical strategy for studying the evolution of prokaryotes because each set of orthologous genes presents its own evolutionary history and no single topology may represent the entire forest. Thus, the methods introduced in this article that compare trees without the use of a preconceived representative topology for the entire FOL may be of wide utility in phylogenomics.
We have shown that, although no single topology may represent the entire FOL and several distinct evolutionary trends are detectable, the NUTs contain a strong treelike signal. Although the treelike signal is quantitatively weaker than the sum total of the signals from HGT, it is the most pronounced single pattern in the entire FOL.
Under the FOL perspective, the traditional TOL concept (a single "true" tree topology) is invalidated and should be replaced by a statistical definition. In other words, the TOL only makes sense as a central trend in the phylogenetic forest.
#### 6 Exercises
X ¼ (((A,B),C),D,E)
Y ¼ (((A,B),C),D,E); (A,B,(E,D); (((A,C),B),D,E); (A,C,(B, D); (A,B,(C,D); (A,B,(C,E); (A,E,(B,D); (((A,C),D),E,F); (((A,B),D),E,C); (((E,F),A),B,C)
#### Acknowledgment
The authors' research is supported by the Department of Health and Human Services intramural program (NIH, National Library of Medicine).
#### References
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
#### The Methodology Behind Network Thinking: Graphs to Analyze Microbial Complexity and Evolution
#### Andrew K. Watson, Romain Lannes, Jananan S. Pathmanathan, Raphae¨l Me´heust, Slim Karkar, Philippe Colson, Eduardo Corel, Philippe Lopez, and Eric Bapteste
#### Abstract
In the post genomic era, large and complex molecular datasets from genome and metagenome sequencing projects expand the limits of what is possible for bioinformatic analyses. Network-based methods are increasingly used to complement phylogenetic analysis in studies in molecular evolution, including comparative genomics, classification, and ecological studies. Using network methods, the vertical and horizontal relationships between all genes or genomes, whether they are from cellular chromosomes or mobile genetic elements, can be explored in a single expandable graph. In recent years, development of new methods for the construction and analysis of networks has helped to broaden the availability of these approaches from programmers to a diversity of users. This chapter introduces the different kinds of networks based on sequence similarity that are already available to tackle a wide range of biological questions, including sequence similarity networks, gene-sharing networks and bipartite graphs, and a guide for their construction and analyses.
Key words Sequence similarity network, Evolution, Lateral gene transfer (LGT), Metagenomics, Gene remodeling, Ecology
#### 1 Introduction
An evolutionary biologist is interested in how processes governing evolution have produced the diversity of genes, genomes, organisms, species, and communities that are observed today. For example, a biologist interested in the eukaryotes may wonder what symbiotic partners have contributed to their origins and evolution. Eukaryotic nuclear genomes are chimeric in nature, encoding many genes acquired from their alphaproteobacterial endosymbiont [1–3]. However, in recent years, it has been proposed that the ongoing gain of genes by both microbial [4–6] and multicellular eukaryotes [7, 8] via lateral gene transfer (LGT) has continued to contribute to eukaryotic evolution, though to a lesser extent than
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_9, © The Author(s) 2019
prokaryotes [9]. A biologist interested in prokaryotes may wish to investigate lateral gene transfer to explore the numbers and kinds of genes transferred between bacteria, archaea, and their mobile genetic elements [10–14]. These transfers are important for understanding the accessory genomes of prokaryotes [15–17]. Further, studying gene transfers in real bacterial communities from different environments can help to test the effect of LGT on ecology and evolution of communities [18]. Given the prevalence of introgression [9–11, 19], one interesting question is whether gene transfer has led to the formation of novel fusion genes that combine parts of genes originating from separate domains of life [20]. An ecologist may wish to analyze the distribution of genes and species in the environment [21]. A metagenome analyst may need to overcome an additional challenge exploring the nature of the large proportion of sequences in metagenome datasets that have little or no detectable similarity to characterize sequences and to study the "microbial dark matter" [22].
High-throughput sequencing technologies present new opportunities to investigate these diverse kinds of questions with molecular data; however, they also present challenges in terms of the scale of the analyses. Consequently, a number of network-based methods have recently been developed to expand the toolkit available to molecular biologists [23], and these have already made major contributions to our understanding of molecular evolution. Networks have been used to shed light on the nature of the "microbial dark matter" [24] and used in ecological studies to explore the geographical distribution of organisms or genes [25, 26] or the evolution of different lifestyles [27]. Their suitability for investigating introgressive events has been used to enhance our understanding of the chimeric origin of genes in the eukaryotic proteome [28, 29], the flow of genes between prokaryotes and their mobile genetic elements [30–35], and gene sharing across mobile elements to study the transfer of resistance factors [14, 36]. Networks have also been used to classify highly mosaic viral genomes [37, 38] and identify gene families [39, 40]. These approaches are highly complementary to traditional phylogenetic approaches, highlighted by the development of hybrid approaches and phylogenetic and phylogenomic networks [34, 41–43]. These hybrid networks are beyond the scope of discussion in this chapter but are covered in Chapters 7 and 8.
While the generation and analysis of networks were previously limited to biologists with programming experience, tools have recently been developed to simplify the process and broaden the availability of network analyses of molecular sequence data. This chapter introduces the different kinds of networks that are already available to biologists and a guide to how these networks can be constructed and analyzed for a large range of applications in molecular evolution. More precisely, this chapter will focus on three kinds of network and the types of analyses that are possible using these networks: sequence similarity networks, gene-sharing networks, and multipartite graphs [23].
#### 2 Sequence Similarity Networks (SSNs)
Sequence similarity networks are the bread and butter of networkbased molecular sequence analyses, with a huge range of applications in molecular biology. The use of SSNs for molecular sequence analysis first came to the fore in the late 1990s and early 2000s, when SSNs were suggested as a way to analyze the rapid influx of new molecular sequence data due to advances in sequencing technology and reduced cost, as well as to predict gene functions and protein-protein interactions [39, 44–46]. One of the earliest formal and heuristic uses of SSNs was to define the COG groups of homologous families and facilitate prediction of the functions of large numbers of genes based on homology [39, 40]. The need for efficient computation and analyses for large biological databases still pervades; however, more recently SSNs have been increasingly appreciated as useful approaches to describe complex biological systems, including inferring the "social networks" of biological life forms [30], producing maps of genetic diversity [27], detecting distant homologues [47–49], and exploring gene and genome rearrangements [50, 51].
A SSN is a graph in which each node is a sequence and edges connect any two nodes that are similar at the sequence level above a certain threshold (e.g., coverage, percent identity, and E-value) as determined by their pairwise alignment (Box 1) (Fig. 1). While the principle behind SSN construction is simple, the expression of similarity data in this structure can enable the use of powerful
Fig. 1 Constructing a simple sequence similarity network. A set of sequences (protein or DNA) in fasta format (a) are aligned in pairs using alignment tools (such as BLAST). These alignments (b) are scored with metrics such as the percentage identity between two sequences (the number of identical nucleotides/amino acids displayed above) or the E-value of the alignment. In the resulting network (c), sequences are represented as nodes. Two sequence nodes are joined with an edge if they can be aligned above a define threshold, with the weight of the edge often based on percentage identity or E-value
algorithms for graph analyses to study complex biological phenomena. Construction of a SSN is also frequently the starting point in a diversity of further graph analyses. A SSN can be constructed directly from fasta formatted sequence files using pipelines, such as EGN [52], the updated and faster performing EGN2 (forthcoming), or PANADA [53]. Visualization of networks can be performed with programs such as Cytoscape [54] or Gephi [55], both of which also have a range of internal tools and external plugins for network analysis. While these programs are useful for the visualization and analysis of relatively small networks, it can be difficult to load large and complex networks with a lot of edges (e.g., 50,000 edges). In these cases the iGraph library offers an extremely powerful and well-supported implementation of a broad range of commonly used methods for both complex graph generation and analysis in R, Python, and C++ [56]. However, using iGraph requires knowledge of programming in at least one of these languages. An additional package for network analysis in Python is NetworkX [57]. It is our goal here to further generalize network approaches by explaining how evolutionary biologists with less programming knowledge could analyze their data. A list including many of the tools and programs available for SSN generation is available at https://omictools.com.
#### Box 1: How to Build Your Own Sequence Similarity Network
1. Dataset assembly: The first and most important step of SSN construction is the assembly of a dataset of sequences relevant to your biological question, usually in fasta format. This can be used as the initial input for wizards such as EGN or EGN2 [52], which can fully automate the process. The nature of the dataset is highly dependent on the research question, so here we focus on the practicalities of database assembly. To construct the similarity network, all sequences in the dataset are aligned against one another in a similarity search. This similarity search is often the time-limiting step in an analysis, and the total number of searches required is quadratic to the number of sequences in the dataset. For large datasets, it is useful to benchmark the alignment using a subset of the data to estimate the timescale for the alignment. Large datasets can generate huge outputs, not only due to the number of sequences but also the length of their identifier. One way to reduce the output size is to replace each sequence name in the fasta file with a unique integer. The use of integers will reduce disk space use and the memory consumption for any software used to analyze the sequence data.
(continued)
2. Similarity search: To generate a sequence similarity network, all sequences must be aligned against one another in an allversus-all search, in which the dataset of sequences is searched against a database including the same sequences. For gene networks, the alignment is usually done with a fast pairwise aligner such as BLAST [58, 59] as implemented in EGN [52]. Filters are often used to remove low-complexity sequences from the search, as these can cause artefactual hits (BLAST options --seg yes, -soft-masking true). The BLAST method of alignment will be the focus of future discussion in this chapter; however, alternatives are available including BLAT [60] (also implemented in EGN), SWORD [61], USEARCH [62], and DIAMOND [63]. These alternatives generally include an option to produce a "BLAST" style tabulated output, making them compatible with programs commonly used in network analyses.
Within alignment tools like BLAST, it is possible to assign thresholds, such as the maximum E-value of the alignment. It is not recommended to set minimal thresholds for some parameters (such as % sequence identity) unless required due to memory constraints so that you can generate networks from a single sequence alignment with different thresholds for comparison (e.g., comparison of a 30% similarity threshold to a 90% threshold, where edges will only be drawn between highly similar genes).
Note: It may be intuitive to use additional CPUs to speed up the alignment process; however, in BLAST it can be more efficient to split the query file and launch multiple searches on separate cores instead of using the BLAST multithreading option. The pairwise alignment step is generally the most time-limiting part of generating a SSN, so benchmarking should be used to establish the optimal settings for the pairwise and/or determine the feasibility of a project given the size of the dataset and the available computational resources.
3. Filtering similarity search results: In an all-versus-all similarity search, any given query sequence will have a self-hit in the corresponding database. For example, with sequences A and B, a self-hit is query sequence A matching to sequence A in the database, cases of which must be removed prior to network construction (Fig. 2). When query sequence A in a similarity search is aligned with sequence B in the database, often the reciprocal result is also identified (an alignment between query sequence B and sequence A in the database). These are called reciprocal hits; while the sequences involved
(continued)
Fig. 2 Filtering sequence similarity results for network construction. In the output of an all-against-all sequence similarity search, there are a number of features that are often filtered out prior to network construction. Self-hits (1/ and 2/), where like sequences are paired in a sequence alignment, are not informative to network construction and are removed (highlighted by the red box surrounding the alignments). In cases where there are reciprocal hits (3/ and 4/) between two sequences, then only the alignment with the highest E-value is retained (highlighted with a green box around the retained alignment) to ensure only one edge representing the best possible alignment connects any two nodes in the network. The same is true for cases where a sequence has multiple hits against another sequence, such as when it aligns to another sequence in multiple positions (5/ and 6/)
are identical, the alignments and scores are not. Retaining both hits would generate two different edges between the same two nodes in a SSN, so generally only the best results from reciprocal hits are retained, based on a score such as the E-value (Fig. 2). Finally, a single query sequence may be significantly aligned multiple times in different positions of the same sequence in the database; however, for SSN construction only the best BLAST hit is generally retained (Fig. 2). The selection of the best BLAST hit is again generally often based on the E-value. Removing multiple hits against the same sequence allows the generation of an undirected network where a single edge connects two nodes, representing the best possible alignment between these nodes.
4. Thresholding and network construction: Constructing a SSN from a BLAST output is conceptually simple; an edge is created between two sequences (nodes) that have been aligned in the sequence similarity search. It is common to apply thresholding criteria such as minimal % ID and/or coverage and/or maximal E-value to determine whether an edge is drawn between two sequences in the network (Fig. 1). There are different ways to calculate the % coverage of an alignment. This could be based on the coverage of a single sequence in the alignment, selecting either the query or the database sequence in each alignment or the longest or shortest sequence in each alignment. Alternatively both (mutual coverage) can be used, retaining an alignment
when both values are above a given threshold. Edges above the thresholding criteria can be assigned a weight based on these criteria, producing a weighted sequence similarity network that retains information of the properties of the alignment between two sequences (Fig. 1). It is often useful to construct and compare several SSNs with variable stringencies defining the edges between sequences, for example, to optimize gene family detection within the SSN (discussed below).
As with other computational approaches, the scale of network analysis is limited by the available computational resources. The limiting factor in terms of the size of network it is possible to construct is predominantly governed by the pairwise alignment. All sequences in the dataset need to be aligned against one another in a pairwise manner, meaning the number of alignments is quadratic to the size of the dataset. For example, computing an allagainst-all comparison of 1,000,000 sequences requires computation of 10<sup>12</sup> alignments. BLAST [64] is the standard tool for this step, with a relatively good speed and accuracy for sequence similarity searches; however, the use of BLAST can be a bottleneck for the analysis of large datasets. This is an especially important consideration given the growth in the number of gene and genome sequences available in public databases. Several rapid alignment tools such as BLAT [60], USEARCH [62], Rapsearch [65], and Diamond [63] have been proposed to overcome this issue. For example, Diamond benchmarks suggest that it is almost as accurate as BLAST but is at least three orders of magnitude faster.
A second point to consider from the perspective of scalability is the complexity and size of the graph and the complexity of the algorithms used in their analysis. Algorithms where the number of calculations is linear to the size of the graph can generally be run on huge graphs with sufficient computational resources, for example, finding connected components using the "deep search first" algorithm. Algorithms for community detection (e.g., PageRank [66], Louvain) are also linear and particularly suited for detecting groups of closely related sequences in huge graphs (discussed in Subheading 4). In contrast, computing graph statistics such as the betweenness centrality are not linear to the size of the graph, even using the relatively efficient Brande algorithm for calculation [67], and are therefore more difficult to calculate for huge graphs. This has led to the development of toolkits specifically designed for the analysis of huge graphs (e.g., NetworKit) [68]. A recent book summarizes the challenges of the analysis of huge networks and some of the algorithms that have been developed to face these challenges [69].
#### 2.1 Scalability of Sequence Similarity Network Analysis
#### 2.2 Exploiting Sequence Similarity Networks for Identification of Gene Families
A gene family is usually defined as a group of sequences that are similar at the sequence level, indicative of homology and potentially of shared functions; however, there is no uniform way to define this similarity [70, 71]. One of the early contributions of SSNs in molecular sequence analysis was the construction of the COG database of homologous protein sequences [39, 40]. This study attempted to define gene families based on similarity at the sequence level using the results of sequence similarity searches. Within the results of an all-versus-all BLAST search, groups of at least three proteins encoded by different genomes that were more similar to each other than they were to other proteins found in the same genomes were defined as a likely orthologous gene family. Orthologous gene families are group of genes in different genomes that show sequence similarity, likely as a result of their shared evolutionary history.
The idea of using graphs to identify gene families is now a core part of many graph-based analyses. Members of a gene family aggregate in a sub-network in a SSN. These sub-networks are called connected components (CCs) at these defined thresholds, i.e., clusters of nodes connected by edges either directly or indirectly (via intermediate nodes) (Fig. 3). The size (number of nodes and edges in a CC) and density (the proportion of potential connections between all nodes in a CC that are actually connected by edges in the graph) of CCs will depend on the thresholds used for
Fig. 3 Louvain community detection in a sequence similarity network. The network is assembled from the results of an all-versus-all alignment, as previously described. Edges can be weighted by E-value, percentage of identity, or bitscore. For the purpose of simplification, we consider strong or weak weights rather than actual values. (a) A giant connected component at relaxed threshold. (b) Three connected components at a more stringent threshold. (c) Three communities with Louvain clustering algorithm, taking into account edge weights
constructing the SSN as well as the relationships between sequences in the network. For example, for a given dataset at a given mutual coverage threshold, a threshold of 90% sequence identity will identify a large number of small connected components that only include highly similar genes, while at a threshold of 30% sequence identity, there will be fewer but larger connected components including genes with more variation in sequence similarity. Commonly used thresholds for detecting homologous gene families are an E-value e5, mutual coverage 80%, and a percentage of identity 30% [23].
CCs are often detected in a SSN using the Depth-First Search (DFS) algorithm; however, there are also other approaches for the detection of gene families based on the idea of detecting "communities" [72]. In some cases, a CC can be further separated into communities of sequences that share more similarity to one another than to other sequences in the CC and thus are more highly linked in the SSN (Fig. 3). Communities are commonly identified by using graph clustering algorithms such as Louvain [73], MCL [74], or OMA [75]; however, different clustering algorithms will result in different outputs. The Louvain weighted method is widely used because it is simple to implement and scales very well to large graphs (Figs. 3 and 4) [73]. MCL is a strong deterministic algorithm that has been implemented, for example, in tribeMCL [74] and orthoMCL [76]. A potential drawback of MCL is that it requires user specification of the "inflation index," a parameter which controls cluster granularity (or "tightness"). A high inflation
Fig. 4 Giant connected component before and after community detection. (a) A single giant connected component from a sequence similarity network. (b) The same giant connected component after application of a community detection algorithm. Node colors correspond to the newly assigned communities
index increases the tightness of clustering, producing a larger number of clusters that are smaller on average than those that would be obtained clustering the same dataset using a low inflation index. Selecting an appropriate inflation index is not trivial and requires optimization [74].
A number of the above approaches have been used to compile additional databases of orthology that can act as useful reference datasets. OMA is a program that uses graph-based algorithms and exact Smith-Waterman alignments to identify orthology between genes [77–80]. OMA is also available as a web browser [81] including a database of orthologues that, in 2015, included more than 2000 genomes and more than seven million proteins [75]. SILIX is a software package [82] that aims at building families of homologous sequences by using a transitive linkage algorithm, and HOGENOM [83] is a database that contains families inferred by SILIX for seven million proteins.
In addition to clustering genes into families, valuable information can be extracted from the connected components using network metrics. Highly conserved sequences tend to form CCs where most of the nodes are connected to each other by edges, while sequences from more divergent families will tend to form more sparsely interconnected CCs. This information can be easily assessed for each component using the clustering coefficient. Conserved families will have a clustering coefficient close to 1, even for stringent thresholds. Identifying such conserved families can be useful to produce multiple sequence alignments (MSA) needed for phylogenetic reconstruction, but SSNs have also been demonstrated to unravel relationships between distant homologues by linking distantly related sequences together [24, 29, 48]. In a SSN, two distant sequences A and C which do not share similarity according to BLAST can be linked together due to sequence B which shows similarity to both A and C.
The idea of distant homology has been particularly illuminating regarding chimeric organisms such as eukaryotes which carry homologous genes inherited from a bacterial ancestor and from an archaeal ancestor [29]. A common way to analyze sequence similarity networks is to identify certain "paths" of interest, for example, the shortest possible paths between two nodes. This notion describes the path between two nodes in a connected component that minimizes the sum of the edge weights. Alvarez-Ponce et al. used this approach to explore the topology of connected components in a SSN including the complete proteomes of 14 eukaryotes, 104 prokaryotes (including archaea and bacteria), 2389 viruses, and 1044 plasmids. Eight hundred and ninety-nine CCs contained sequences from all three domains, and of these 208 contained eukaryotic sequences that were not directly similar to one another but only linked to one another via a "eukaryotearchaea-bacteria-eukaryote" shortest path. These are putatively distant homologues in eukaryotes that were present in both the archaeal host of the mitochondrial endosymbiont and in the alphaproteobacterial endosymbiont, with both copies subsequently retained in eukaryotes and as such strong evidence for the chimeric origin of eukaryotes [29]. This demonstrates the utility of networks in the study of ancient evolutionary relationships including the origin of eukaryotes [28] or rooting the tree of life [84]. Simple path analysis for a network is possible using existing plug-ins within visualization tools such as Cytoscape [54] and Gephi [55].
#### 2.3 Exploiting SSNs to Identify Signatures of "Tinkering" and Gene Fusion
When discussing identification of gene families, we have focused on networks where edges are drawn between protein sequences that show a high enough similarity across their entire length, defined by a high mutual coverage threshold (e.g., 80%). Sequence similarity can also be partial, for example, following gene remodeling or "tinkering" [85] producing new combinations of gene domains via gene fusion and fission events, or through the de novo sequence synthesis of gene extensions, adding to existing sequences. The term "Rosetta Stone sequence" was coined to define the formation of a new fusion protein in a species as the result of the fusion of two proteins that are found separate in another species, with authors originally predicting that these fusions could occur between proteins that physically interact in a common structural complex [86]. One of the earliest applications of sequence similarity searches to identify fusion proteins was an attempt to predict pairs of proteins that may physically interact in an organism based on whether they could be identified as a single "composite" fusion protein in another organism [44]. Beyond predicting protein-protein interactions, this kind of gene remodeling and recycling of existing gene parts has the potential to contribute to the expansion of functional diversity in genomes, creating new and unique combinations of domains and functions [51, 85, 87–91]. Similarity search-based screens have been implemented to identify composite genes and genome rearrangements in a range of prokaryotes [92–94], eukaryotes [87, 95–97], and viruses [98].
Early attempts to identify composite genes were based on the output of sequence similarity searches, but without formalizing the results of search methods into a graph structure. The first attempt to formalize the problem of identifying "composite" genes in networks was the "Neighborhood Correlation" approach, aiming to distinguish genuine multi-domain proteins sharing common ancestry (homologues) from novel multi-domain proteins that share domains due to insertions [99]. The later development of the FusedTriplets and MosaicFinder tools attempted to unify existing graph-based methods for detection of "composite" gene detection [50]. FusedTriplets is a graph-based implementation of the traditional gene-centered method for composite gene identification, originally introduced by Enright et al. [44], with additional cross-
Fig. 5 Composite gene identification using "minimal clique separators." (a) A multiple sequence alignment of composite genes (yellow) with two components (blue and magenta). (b) The sequence similarity network corresponding to the multiple sequence alignment. The composite genes (yellow) are a minimal clique separator for the network. Their removal (shown in c) decomposes the network to the two separate component families
checks on the absence of similarity between the two component genes contributing to a composite gene based on varying thresholds [50, 100]. MosaicFinder is a gene family-centered approach which will only identify highly conserved composite gene families that form "minimal clique separators" (Fig. 5) [50]. This graph topology implies that MosaicFinder may fail to detect divergent (e.g., ancient or fast evolving) composite gene families which will tend to form "quasi-cliques" without perfect separation. CompositeSearch [101] (available at http://www.evol-net.fr/index.php/ en/downloads) is a new program designed to overcome this limitation by identifying both conserved and divergent composite gene families (Box 2).
#### Box 2: How to Identify Composite Genes Using CompositeSearch
tendency of BLAST to produce overhanging alignments [100]. The output includes a node, edge, and information file including information on number of nodes, edges, and family connectivity from family detection. Two outputs are included for composite gene detection, a "composites" file with detailed information on each predicted composite gene in fasta format and a "compositesinfo" file, summarizing the data. Similarly, two files provide detailed information on composite gene families and a summary of composite gene families.
3. Filtering results: By default, CompositeSearch outputs all possible composite genes in "fast" mode or composite gene families in the full mode. These are given alongside a number of different metrics designed to help to filter families for more confident predictions, including the gene family size, number of composites directly predicted within the gene family, the number of domains, the number of component families, the number of singleton component families (families including only one sequence), the connectivity of the family, and a score based on the overlap between different components mapped to the composite gene.
Recent studies have explored composite gene formation as a source of innovation by "tinkering" [85] during major evolutionary transitions. These can be especially interesting when exploring genome evolution following introgression, raising the possibility of formation of new composite genes using components with different evolutionary origins [20, 51, 102]. For example, the gain of a cyanobacterial endosymbiont at the origin of photosynthetic eukaryotes was accompanied by the transfer of whole cyanobacterial genes to its new host genome, with gene functions related to the role of the plastid [103–105]. Identification of composite genes related to the origin of photosynthetic eukaryotes unraveled novel symbiogenetic composite genes, and unique fusions of genes encoded in the nucleus of photosynthetic eukaryotes that included components derived from the plastid endosymbiont. As with whole genes transferred to the nucleus, several of these components had predicted functions related to the role of the plastid, including redox regulations and light response [51].
2.4 Exploiting SSNs for Ecological Studies Ecological studies increasingly involve the assembly, analysis, and comparison of large metagenome datasets. In addition to identification of functions and organisms associated with a particular environment, these studies enable the investigation of important hypotheses in microbial ecology at the level of organism or function, such as the often quoted hypothesis that "everything is everywhere, but the environment selects" from Bass Becking: the idea that microbial lineages are limitlessly dispersible in the environment, but the environmental conditions will select for certain lineages and control their distribution rather than any specific geographical separation [21].
Networks are useful for these kinds of ecological studies because existing graph algorithms can be used to investigate the structure of the network. When investigating gene (or gene-sharing networks), it is possible to distinguish nodes by labeling them based on their properties, such as categories for taxonomic or environmental origins (Fig. 6). A simple way to represent this visually is to color nodes based on these properties in Cytoscape or Gephi. A formal way to explore the relationships between node properties is to use network metrics such as conductance [106], modularity [73], and assortativity coefficient (normalized modularity) [107]. Assortativity and conductance are different metrics that attempt to answer the same type of question: do nodes labeled as belonging to a particular category, such as environmental origin, tend to be connected with other nodes labeled as belonging to the same category? More precisely, conductance quantifies whether a given category of nodes shares more edges between themselves than with nodes from different categories. A low conductance approaching zero indicates that nodes of a given category are highly connected to one another, with few connections to nodes from different categories. A higher conductance is indicative that nodes of this category tend to be more sparsely interconnected and share more connections with nodes from different categories. Assortativity is a measure of the preference for a category of nodes in a network to attach to other nodes
Fig. 6 Exploring distribution of annotations in sequence similarity networks. In this example, nodes within a single connected component are assigned two colors, blue and yellow, corresponding to their having a different categorical annotation (e.g., originating from a different environmental source). Using the example of environmental source, genes in cluster A would all have the same environmental source (blue), indicating an environment-specific cluster of genes. Genes in cluster B are found in two different environmental sources (blue and yellow); however, nodes of the same type are preferentially linked to each other in the network than to genes from different environmental sources. This would result in a positive assortativity coefficient approaching 1 for environment and a low conductance score, suggesting a strong environmental community structure. Genes in cluster C are also found in two different environmental sources; however, there is no clear pattern for the distribution of genes with regard to environment. This network would have an assortativity approaching 0 and a high conductance score
from the same category. Normalized assortativity values range between 1 and 1, where 0 indicates random distribution of categories within the network, 1 indicates that nodes from the same categories tend to be connected to one another in the network, and 1 indicates that nodes from different categories tend to be connected in the network. A detailed description of the algorithms used in these calculations can be found in [108].
2.4.1 Assortativity as a Tool to Study Geographical and Habitat Distributions of Microbes and Genes
2.4.2 Conductance in the Comparison of Lifestyles and Evolutionary Histories
Forster et al. used assortativity (among other network statistics, including the previously discussed shortest path analysis) to explore the geographical dispersion patterns of marine ciliates in a network generated from ciliate SSU-rDNA sequences [25]. Sequences were clustered into two different levels of gene family—CCs and Louvain communities (LCs) as previously described. Sequences were assigned categorical labels based on their geographical point of origin (eight locations) or habitat of origin (three habitats), and assortativity was calculated. If sequences, and thus species, are broadly distributed across geographical categories, then assortativity of SSU-rDNA sequences labeled with these geographical categories would be low because similar sequences would be found in different environments. Contrarily, if similar sequences tend to be from the same geographical category, indicative of endemism, then assortativity of sequence geographical origin will be high (Fig. 6). The majority of CCs and LCs showed a positive assortativity for geographical origin, higher than expected by chance, indicative of geographical community structure as opposed to global dispersal of ciliates. Similar approaches were used by Fondi et al. and applied to a collection of environmental metagenome samples to test the "everything is everywhere" hypothesis at the gene pool and functional level. Gene pools were more strongly associated with a particular ecological niche than with specific geographical location, supporting the idea that microbial genes are found everywhere but the environment selects for them [26].
Conductance is used to explore the clustering of pairs of different node categories in a connected component. In a study by Cheng et al., the proteomes of 84 prokaryote genomes were categorized into four broad redox groups based on their lifestyle, methanogens, obligate anaerobes, facultative anaerobes, and obligate aerobes [27]. For each CC in a pan-proteome sequence similarity network including all 84 genomes, the conductance was calculated for pairs of redox categories and compared to values obtained following random relabelling of the components. The distributions of conductance values for methanogens and for obligate anaerobes groups indicated that the sequences in these groups have features distinct from those in other groups, that anaerobes and aerobes tend to be dissimilar, and that their sequences are more isolated from one another in the SSN than expected by chance.
An additional example of the use of conductance is in exploring the propensity of a gene family to lateral gene transfer. Within a network of archaeal and bacterial genes, CCs showing a low conductance for both archaeal and bacterial sequences indicate that the bacterial and archaeal genes within the corresponding families are structured in two separate and conserved groups (Fig. 6). Structuring gene families into two groups would indicate that there was little or no evidence for lateral gene transfer between archaea and bacteria within this particular gene family. This kind of gene family is rare, with only 86 gene families from 40,584 (0.2%) meeting this criteria [24].
2.5 SSNs in Remote Homologue Identification: Shedding Light on the Microbial Dark Matter Up to 99% of microbial species are not cultivable and thus have not been studied in isolated culture. Analysis of high-throughput sequencing and metagenomics datasets has shed light on these uncultivable organisms, often referred to as the "microbial dark matter" [109], and in some cases enabled the reconstruction of draft genomes [110–114]. A considerable portion of most metagenome studies have predicted ORFs showing no detectable similarity to any known proteins, termed metaORFans [115]. These can represent 25–85% of the total ORFs identified in metagenomes [22]. Identifying distant homologues of ORFans may help to predict their functions and begin to unravel the microbial dark matter. Recent work by Lopez et al. in 2015 probed the microbial diversity of metagenome datasets from a range of environments including the human gut microbiome, identifying homologues of genes from 86 ancient gene families that are distributed across archaea and bacteria. The majority of these gene families included environmental homologues that were highly divergent from any of their cultured homologues, and many branched deeply with the phylogenetic tree of life, highlighting our limited understanding of diverse elements of the microbial world and hinting at the existence of yet unknown major divisions of life [24] (Fig. 7).
2.6 Exploiting SSNs to Analyze Classifications Metagenomic and genomic data are providing scientists with a tantalizing amount of sequence data, casting the analysis of the extent of biodiversity as a major research theme in biology [116–120]. In theory, existing organismal and viral classifications are invaluable tools to structure and analyze this biodiversity. However, the way taxonomical classifications are constructed raises questions about their naturalness and their actual application scope [38, 120–128], in particular regarding genetic diversity surveys. There are three major reasons for this. First, organismal and viral diversity is still largely undersampled, which means that existing classifications are incomplete [119, 120]. Therefore, taxonomically unassigned sequences cannot be readily used in classbased genetic diversity surveys, since this dark matter remains outside existing classes. Second, classifications are constructed
Fig. 7 Remote homologue detection to help characterize the microbial dark matter. (a) A hypothetical highly conserved cluster of genes from genomes present in sequence databases, where the average % of identity is high (60%). (b) The same cluster after addition of divergent environmental sequences to the network. Environmental sequences in gray are more similar to those already identified from genome surveys (60% max identity) so are connected directly to the conserved gene cluster in the network. More divergent sequences in pink have <60% maximum identity to their homologues in the database. Many of these are only identified as linked to the sequences from the conserved database via intermediate gray nodes. This is the notion of "transitive homology"
using different features (i.e., for viruses, a mix of phylogenetic, morphological, and structural criteria, such as replication properties in cell culture, virion morphology, serology, nucleic acid sequence, host range, pathogenicity, epidemiology, or epizootiology); therefore their classes do not necessarily offer immediate proxies for quantifying genetic diversity per se. Third, evolutionary processes responsible for both genetic and organismal diversity are diverse, and they operate at different tempos and modes in different lineages [49, 123, 129–141]. As a result, genetic diversity within classes and between classes can be heterogeneous, meaning that existing classifications may lack efficiency to discriminate, predict, or compare taxa on genetic bases, potentially hampering diversity studies, a profound practical issue at a time where the analysis of metagenomic sequences is becoming a priority in biology.
Addressing these challenges is notably crucial for viral studies. Recently, the executive committee of the ICTV [142] proposed that network analyses methods that create similarity metrics based on the detection of homologous genes and their genetic divergence constitute a valuable strategy to assist classification of viruses. Consistently, basic network properties and metrics (Table 1) can quantify (1) whether genetic diversity is consistent within and between the classes of existing classifications and (2) describe what classes are the most homogeneous and distinctive in terms of genetic diversity. Three criteria can be used to estimate intra-class genetic heterogeneity (Fig. 8a–c). First, the average edge weights (measured as % of identity, PID) between pairs of sequences from genomes of the
#### Table 1
#### Schematic properties of two extreme kinds of taxonomic classes with respect to their genetic diversity
The three top properties inform about genetic diversity within classes (intra-class genetic diversity). The last two properties inform about the genetic distinctiveness (core and signature genes) of the classes. Interclass genetic heterogeneity identifies when genetic diversity of a class is not comparable with genetic diversity of another class in the classification. CCC, average proportion of genetic conservation between sequences from the same cluster and from the same taxonomic class; PID, average edge weights (% identity) between two sequences from genomes of the same class
> same class provide a trivial measure of intra-class genetic diversity. Second, the average proportion of Conserved Canonical Connections between sequences from the same connected component and from the same taxonomic class can be exploited (CCC, i.e., in each connected component of the SSN, the total number of edges connecting sequences of a given class i (intra-group edges, denoted Eii) divided by the theoretical maximal number of possible edges between sequences of that class in the connected component (CCC (i) ¼ 2\*Eii/(N<sup>i</sup> (N<sup>i</sup> 1)) where N<sup>i</sup> is the number of sequences of class i present in the connected component). CCC ranges between 0 and 1. Within a connected component, if all pairs of sequences from the same class are directly connected, CCC equals 1, since all these sequences are more conserved than a given %ID threshold. By contrast, low CCC are observed when sequences from genomes from the same class lack cohesive evolution, for example, when some related sequences evolved so fast that they show less than the minimal similarity required to be directly connected to their homologues in the graph. Third, the genetic consistency of a class can be estimated by (1) identifying what cluster of sequences was present in the largest number of genomes of the class and then (2) by quantifying the proportion (in %) of the class members harboring that most ubiquitous cluster (maxCore%). When max-Core% of a class is <100%, it means that, for this dataset, there is no gene family shared by all members of that class (i.e., no core genes). The SSN structure can also serve to estimate the genetic distinctiveness of each class, i.e., whether sequences from a given class are
Fig. 8 Intra- and interclasses heterogeneity measurements in weighted similarity networks. Sequences are represented by nodes. Each node is colored to represent the taxonomic class to which its host belongs. Nodes with the same color belong to the same class. Edge weight is represented by edge size proportional to the weight. Subgraphs correspond to clusters of sequences. Direct neighbors have a greater similarity than the threshold set to allow such connections. PID, average edge weights (% identity) between two sequences from genomes of the same class; CCC, average proportion of genetic conservation between sequences from the same cluster and from the same taxonomic class; maxCore%, conductance; and %-exclusive components correspond to the estimates used to assess genetic consistency of classes
more similar to one another than they are to sequences from other classes (Fig. 8d, e). Such sequences could be used as classificatory features to assign members to the class. In a SSN, this property translates to a low ratio of interclass edges over intra-class edges and is measured by conductance (Fig. 8d). Likewise, the proportion of clusters comprised exclusively of sequences from one class, a diagnostic feature of the class, provides an estimate of the class genetic distinctiveness. Genetically highly distinct classes have a high % of such exclusive clusters. Based on these network measures, interclass genetic heterogeneity can simply be diagnosed by contrasting estimates of genetic consistency for all the above measures for each class. There is interclass heterogeneity within a classification when the mean PID, mean CCC, maxCore%, DRC, and % of exclusive components differ between classes.
Such network analyses show that virus classifications face a pragmatic issue: overall genetic distinctiveness allows relatively safe assignments of viral sequences to existing classes; however, genetic diversity of viral taxa of similar ranks differs among the tested classifications. Therefore, virus classifications (especially ICTV classification at the family level) should be used carefully to avoid inaccurate estimates in metagenomic diversity surveys. Classes with broader genetic diversity will tend to be more easily detected in the environment than classes with reduced genetic diversity, since the former will necessarily be associated with more OTUs than the latter. Some alpha- and beta-diversity analyses of environmental data, which rely on counts and on contrasts of the abundance of taxonomic classes in different samples, will thus also be biased. A similar approach could be applied on different types of classified lineages, i.e., to identify what groups of bacteria, archaea, or eukaryotes with comparable taxonomical ranks are the most genetically heterogeneous and what ranks of their classification are the least genetically consistent.
#### 3 Gene-Sharing Networks
Gene-sharing networks are often called "genome networks" as they are best suited for summarizing what genes are shared between different genomes, highlighting routes of gene sharing. The ability to explore gene sharing between all genomes in a network in a simple graph can have useful properties for reflecting microbial social life, inherently inclusive of gene sharing both as a consequence of vertical inheritance and lateral gene transfer (LGT). Bacteriophage and plasmid genomes are typically highly mosaic in nature due to a high level of horizontal gene transfer, making it difficult to classify their genomes [37, 143]. Lima-Mendez et al. proposed the use of gene-sharing networks as a new classification method that tackles this problem of mosaicism by classifying viruses based on their genome's content [37]. Constructing gene-sharing networks using subsets of genes from different functional categories of genes can also be useful in exploring what kinds of genes are being shared by different genomes.
In a gene-sharing network, each genome is represented by a node, and two nodes are connected by an edge when the two corresponding genomes share homologous genes or gene families (Fig. 9). These gene families can be identified from SSNs (of as CCs of LCs) or by alternative methods. In gene-sharing networks, edges can be weighted by the number of genes or gene families shared between the genomes. In this way, gene-sharing networks enable the study of microbial social life, quantitatively displaying the gene families shared between genomes both as a result of vertical transmission and lateral gene transfer.
Gene-sharing networks are useful tools for exploring overall patterns of gene sharing between genomes. Recently, Lord et al. developed BRIDES, a software package that specifically identifies different kinds of patterns in evolving gene-sharing networks after the addition of new genome nodes [144]. However, in genesharing networks the kind of gene families that are being shared is often overlooked. To explore how functions are shared between different genomes, gene-sharing networks can be built from genes
Fig. 9 Translating gene networks to gene-sharing networks. (a) Gene network for three gene families. Gene nodes are colored based on their genome of origin. The background color corresponds to the gene family color in part c. (b) The gene-sharing network corresponding to the gene network in a. Edges are weighted on the number of gene families shared by the genomes. (c) Multiplex gene-sharing network corresponding to the gene network in a. Genomes are connected by multiple edges with colors corresponding to different gene families. These edges are weighted based on the number of genes shared between two genomes for each family
using different subsets of functions (Fig. 10) [29]. An alternative form of the gene-sharing network is the multiplex network. In this network nodes can be linked by edges of different types, for example, each edge representing a different gene family or different functional groups of gene families, thus retaining additional information compared to a simpler gene-sharing network (Fig. 9) [23]. Multiplex networks can be useful for small-scale analyses; however, with large datasets they can rapidly become difficult to interpret and analyze. Importantly, multiplex networks are unimodal projections of bipartite graphs (discussed in the Subheading 14) which can provide greater clarity and have a number of attractive properties for the analysis of larger datasets.
3.1 Classification of Entities Using Gene-Sharing Networks The possibility of summarizing gene sharing between sets of entities with complex evolutionary histories means that gene-sharing networks can be useful for classifying organisms based on their gene content. Lima-Mendez et al. analyzed bacteriophage genomes to generate two different phage gene-sharing networks that reflect their reticulate evolutionary history [37]. In the first gene-sharing network, phage genomes (nodes) were connected by edges when
Fig. 10 Functional gene-sharing network reflecting the chimeric nature of eukaryotes. These gene-sharing networks describing how genes in different functional categories are shared between bacteria (green), archaea (yellow), eukaryotes (gray), plasmids (purple), and viruses (red) from a published dataset [29]. In both cases, a giant connected component is shown alongside examples of smaller connected components (a) Gene-sharing network for COG category D: cell division control. In this network, sequences of eukaryote origin (gray) cluster with bacterial sequences, reflecting their origin in the alphaproteobacterial endosymbiont that would become the mitochondrion. (b) Gene-sharing network for COG category K: transcription machinery. In this network, eukaryote sequence (gray) cluster with archaeal sequences, reflecting the origin of these genes in the archaeal host for the eukaryotic endosymbiont
they shared significant similarity at the sequence level. This genesharing network was clustered using the previously discussed MCL algorithm [145], identifying distinct groups of phages with sequence similarity. Following clustering, membership to a particular cluster was reassessed based on shared similarity with viruses in other clusters, reflecting their reticulate evolutionary history, allowing the generation of a matrix assigning a score describing the relative membership of any given viral genome to a particular classification group. In the second approach, Lima-Mendez et al. generated a "module"-based gene-sharing network, where edges are drawn between two phage genomes if they share a "module," in this case defined as a group of genes with similar phylogenetic profiles, enabling the exploration of what kinds of genes are shared between different groups of phages or are "signatures" for a particular group of phage genomes [37].
3.2 Exploring Routes of Gene Sharing in Gene-Sharing Networks Two network metrics, also useful in the analysis of gene networks, can be used to attempt to identify "hubs" of gene sharing in the context of gene-sharing networks: node "degree" and "betweenness." Both metrics aim to determine the centrality of a node in a network. The degree of a node is simply the number of edges that it is connected to. The betweenness of a node is the frequency at which it is found in all the possible shortest paths between any two nodes in the network. Halary et al. used gene-sharing networks based on DNA sequence similarity to explore gene sharing between prokaryotes and mobile genetic elements [30]. Plasmids were identified as hubs of gene sharing within this pool of genomes, suggesting that they are key vectors for genetic exchange between cellular genomes and a potential DNA reservoir shared by genomes. Phages were more peripheral in the network and mostly linked prokaryotes from the same lineage. Thus, gene-sharing networks provided insights on the evolutionary processes that shape the gene content of prokaryote genomes.
The importance of plasmids in genetic worlds was further highlighted by exploring plasmid gene-sharing networks without inclusion of prokaryote genomes [14, 36]. Connecting 2343 plasmid genomes based on shared gene content in a single graph demonstrated that plasmids tended to cluster based on the phylogenetic class of their corresponding host prokaryote rather than habitat but that more mobile plasmids tended to be more "central" in the graph, indicating that these were hubs of gene sharing. Specifically, routes of gene sharing for gene families including antibiotic resistance markers were identified between actinobacterial plasmids and gammaproteobacterial plasmids, suggesting that Actinobacteria may act as a reservoir for antibiotic resistance genes for Gammaproteobacteria [14].
The finding that plasmids are hubs of gene sharing for prokaryote genomes was supported by analysis of gene sharing in a proteobacterial phylogenomic network including 329 proteobacterial genomes [32]. A phylogenomic network is a type of phylogenetic network that has been constructed from fully sequenced genomes. In this example the phylogenomic network is an alternative to a gene-sharing network, in which genome nodes within a phylogeny are linked by edges if they share genes [34]. This study identified extensive evidence for lateral gene transfer among Proteobacteria, with at least one LGT event inferred in 75% of all gene families. Of these putative LGTs, more were related to plasmid-related genes than phage-related genes, suggesting plasmid conjugation was a more frequent source of gene transfer [32]. Directed graphs exploring directionality of LGT events between 657 prokaryote genomes allowed the polarization of 32,028 putative LGT events finding that frequency of recent events correlates with genome sequence similarity and most LGTs occurring between donor-recipient pairs with <5% difference in GC content, suggesting that there are some barriers to lateral gene transfer between prokaryotes but that these are not insurmountable [31]. Later reconstruction of transduction events linking phage donors and recipients in a phylogenomic network demonstrated that LGT by transduction was generally highest in similar genomes and between clusters of closely related species but that this constraint was occasionally broken, resulting in LGTs over long evolutionary distances [35].
#### 4 Bipartite Graphs
Bipartite graphs are excellent at summarizing what genes are shared between sets of genomes, and as such are ideal for comparative genomics, including for the comparison of genomes reconstructed in metagenomic analyses. The potential to extend this approach to multilevel graphs, adding additional layers of information such as the environment in ecological studies, could provide a powerful summary of gene sharing in relatively complex datasets.
A multilevel network is a network in which edges exclusively connect nodes of different types, i.e., representing different levels of biological organization. Thus, a bipartite graph is a graph with two types of nodes (top and bottom nodes), where edges exclusively connect nodes of different types (Fig. 11) [146]. The types of nodes used can vary widely depending on the biological question, from linking diseases (top nodes) to their associated genes (bottom nodes) in order to explore the association between related disease phenotypes and their genetic causes [147, 148], to exploring the concept of flavor pairings in food based on a graph of ingredients (top nodes) and the flavor compounds they contain (bottom nodes) [149]. For applications in molecular biology, a typical example of a bipartite graph may describe the relationships between genomes (top nodes) and gene families (bottom nodes), with edges between nodes indicating that a genome encodes at least one member of the corresponding gene family (Fig. 11) [23, 33, 38, 150]. This kind of genome to gene family graph is particularly suited for the comparative analysis of the gene content of genomes in microbial communities and for exploring patterns of gene sharing, for example, between distantly related cellular genomes [33] or between cellular genomes and their mobile genetic elements (Corel et al. forthcoming). It is possible to represent all genes shared between a given set of genomes, as a result of both vertical inheritance and horizontal gene transfer, in a single bipartite graph [23].
Fig. 11 A bipartite graph and its reduction to a quotient graph: (a) An example of a bipartite graph displaying how five gene families are shared between three genomes. (b) A reduced form of the bipartite graph in which gene families are combined to "twin" nodes if they share identical taxonomic distributions. A single "articulation point" connects all three genomes
This feature was utilized by Iranzo et al. to explore gene sharing among the entire dsDNA virosphere, a group of entities typified by high rates of molecular evolution and gene transfer [38]. In this case, bipartite modularity was identified in the graph to identify groups of related viral genomes and their shared genes, with the modularity of the graph optimized to Barber's bipartite modularity [151]. A number of additional methods have been developed for detection of module structures within a bipartite graph including for weighted graphs [152]. Two recently developed tools, AcCNET [150] and MultiTwin (forthcoming), have simplified the process of constructing and analyzing multilevel graphs without the need for custom programming (Boxes 3 and 4).
#### Box 3: Generating Gene-Sharing Networks and Bipartite Graphs
- (a) In a gene-sharing network, two genomes are connected by an edge when they encode genes belonging to the same gene family. Generating this kind of network can be automated from BLAST or fasta sequence data using EGN [52].
- (b) In a bipartite graph, there are two types of node, genome nodes and gene family nodes. An edge is drawn between a genome node and a gene family node if that genome encodes a member of the gene family. AcCNET [150] and MultiTwin (forthcoming) tools both include pipelines for generating bipartite graphs from sequence data. MultiTwin can also generate a bipartite graph from two files: a tab-delimited file mapping gene identifiers to their corresponding genome identifier and a tab-delimited file mapping gene identifiers to their corresponding gene family.
Two topological features of bipartite graphs can be used to facilitate studies of gene sharing by an exact decomposition of the bipartite graph: twins and articulation points [23, 153]. A bipartite graph can be reduced to a quotient graph, a reduced variant of the bipartite graph where nodes from the bipartite graph have been combined based on sharing similar properties without the loss of information. For twin nodes ("twins"), this reduction is based on the combination of bottom nodes that have identical neighbors into a single "twin" supernode in the quotient graph (Fig. 11). This is a useful way of reducing the size of large graphs without losing information, but twin nodes also have useful properties for graph interpretation. The genomes supporting a twin node (its neighbors) define a club of genomes that share genes, through common ancestry and/or horizontal transfer, and the number of gene families making up the twin gives a simple description of how many gene families are shared between this club. For example, in any given dataset, any "core" set of gene families encoded by all species in the analysis will be represented by a single twin node. The gene families combined in twin supernodes can be viewed as gene families that are likely to be transmitted together [23]. An articulation point is a node that, when removed, will split the graph into two or more connected components. Within a gene family-genome bipartite graph, articulation points are expected to help to identify "public genetic goods," gene families that are shared by distantly related entities that may confer an advantage independent of genealogy [23, 154], as well as selfish genetic elements such as transposases that also spread across multiple genomes.
#### Box 4: Considerations for the Construction and Analysis of Bipartite Graphs Using AcCNET and MultiTwin
The default workflow for both ACcNet and MultiTwin takes protein sequence data in fasta format as input and generates a bipartite graph alongside a number of graph summary statistics and outputs for visualization in standard tools (such as Gephi and Cytoscape) but with a number of important differences, including:
<sup>l</sup> Graph levels: Both AcCNET and MultiTwin can generate a bipartite graph using their default workflow; however, Multi-Twin can also be used to explore additional graph levels by adding additional node types (e.g., a tripartite graph). Multipartite graphs mean that gene family level annotations can be associated with additional levels of biological information. This may be particularly useful for the comparison of samples in metagenomics studies or time course experiments, allowing gene families to be associated directly with features such as environmental origin or time point.
(continued)
AcCNET is available at: https://sourceforge.net/pro jects/accnet
MultiTwin is available at: http://www.evol-net.fr/index. php/en/downloads
4.1 Using Bipartite Graphs to Explore Patterns of Gene Sharing Between Diverse Entities
The simplest application of a bipartite graph is the summary of all genes shared between genomes in a single parsable graph, and this feature has been used to explore gene sharing in the dsDNA virome [38], a range of Escherichia coli genomes to investigate the E. coli pangenome [150] and between a broad range of prokaryotes that include newly discovered organisms [33]. In their analysis of prokaryote genomes, Jaffe et al. used the notion of "twins" to explore patterns of gene sharing between prokaryotes, including Archaea and the recently discovered ultrasmall "Candidate Phyla Radiation" and TM6 bacteria with extremely unusual and reduced genomes. The group found evidence for lateral gene transfer between ultrasmall bacteria and other prokaryotes, consistent with the suggestion that the ultrasmall bacteria may be symbionts [33]. In their exploration of the dsDNA virome, Iranzo et al. used graph module detection, algorithms designed to identify groups of densely connected nodes in a graph, to identify sets of densely connected viral genes and genomes that included viruses with broad host ranges, as well as 14 hallmark viral genes that account for most of the gene sharing between all different viral modules [38].
#### 5 Conclusions
This chapter has offered a brief introduction to the generation of commonly used sequence similarity networks in molecular biology and a guide to how they can be generated and applied to a broad range of studies (Fig. 12). Networks provide a highly scalable framework for the study of an increasingly broad range of applications in molecular biology and evolution and have already contributed to a number of important discoveries in the field. These include exploring patterns of introgression and horizontal transfer across all domains of life and mobile elements, the origin of eukaryotes, the contribution of new genes including novel fusion genes to major evolutionary transitions, shedding light on the "microbial dark matter" in metagenome sequencing datasets and in testing ecological hypotheses about organism and gene distribution and environmental selection. New methods and tools for network analysis are becoming increasingly user-friendly and accessible to biologists without extensive programming experience and enabling network analysis to become a more common part of a biologist toolkit in the analysis of molecular sequence data.
Fig. 12 A workflow highlighting some of the available routes for generation and analysis of SSNs, genesharing networks, and bipartite graphs. This workflow highlights just some of the many tools and routes for network construction and analysis
#### 6 Exercises
The exercises use EGN [52] and require access to a local installation of BLAST+ [58] and Perl. The fasta sequence file "example.faa" provided with EGN includes a dataset of protein sequences from Archaea, Bacteria, Eukaryota, and mobile genetic elements, available at http://www.evol-net.fr/index.php/fr/downloads:
- (a) Make the blast database using the "makeblastdb."
- <sup>l</sup> Command: "makeblastdb -dbtype prot -in example.faa –out example"
- (b) Performing the BLAST search using "blastp," remembering to output data in a tabular format for easy processing.
- <sup>l</sup> Command: "blastp -query example.faa -db example -evalue 1e-5 -seg yes -soft\_masking true - max\_target\_ seqs 5000 -outfmt "6 qseqid sseqid evalue pident bitscore qstart qend qlen sstart send slen" -out protein.blastpout"
- (a) Run EGN from the terminal using "perl egn.1.0.plus.pl" from the programs home directory.
- (b) Follow on-screen prompts sequentially to generate an alignment, filter the output, and generate a gene network with outputs compatible with both Cytoscape and Gephii.
- (a) In Cytoscape: Import files named "cc.\*.txt" as a network to visualize that set of connected components.
- <sup>l</sup> To associate nodes with their annotations, import "cc\*. atr" as a table.
- (b) In Gephi: Open "cc\*.gxf" files to import individual connected components from the network into Gephi. Use the "layout" menu to explore different kinds of layouts for the network.
#### Glossary
Articulation point A node in a graph whose removal increases the number of connected components of the resulting graph.
#### References
Genet 25:107–110. https://doi.org/10. 1016/j.tig.2008.12.004
https://doi.org/10.1101/gr.229202. Article published online before March 2002
Res 30:1575–1584. https://doi.org/10. 1093/nar/30.7.1575
protein-protein interactions from genome sequences. Science 285:751–753
6:2195–2205. https://doi.org/10.1093/ gbe/evu168
57:369–394. https://doi.org/10.1146/ annurev.micro.57.030502.090759
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Chapter 10
## Bayesian Molecular Clock Dating Using Genome-Scale Datasets
#### Mario dos Reis and Ziheng Yang
#### Abstract
Bayesian methods for molecular clock dating of species divergences have been greatly developed during the past decade. Advantages of the methods include the use of relaxed-clock models to describe evolutionary rate variation in the branches of a phylogenetic tree and the use of flexible fossil calibration densities to describe the uncertainty in node ages. The advent of next-generation sequencing technologies has led to a flood of genome-scale datasets for organisms belonging to all domains in the tree of life. Thus, a new era has begun where dating the tree of life using genome-scale data is now within reach. In this protocol, we explain how to use the computer program MCMCTree to perform Bayesian inference of divergence times using genome-scale datasets. We use a ten-species primate phylogeny, with a molecular alignment of over three million base pairs, as an exemplar on how to carry out the analysis. We pay particular attention to how to set up the analysis and the priors and how to diagnose the MCMC algorithm used to obtain the posterior estimates of divergence times and evolutionary rates.
Key words Molecular clock, Bayesian analysis, MCMC, Fossil, Phylogeny, Primates, Genome
#### 1 Introduction
The molecular clock hypothesis, which states that the rate of molecular evolution is approximately constant with time, provides a powerful way to estimate the times of divergence of species in a phylogeny. Since its proposal over 50 years ago [1], the molecular clock hypothesis has been used countless times to calibrate molecular phylogenies to geological time, with the ultimate aim of dating the tree of life [2, 3]. Several statistical inference methodologies have been developed for molecular clock dating analyses; however, during the past decade, the Bayesian method has emerged as the method of choice [4, 5], and several Bayesian inference software packages now exist to carry out this type of analysis [6–10].
In this protocol, we will explain how to use the computer program MCMCTree to estimate times of species divergences using genome-scale datasets within the Bayesian inference
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_10, © The Author(s) 2019
framework. Bayesian inference is well suited for divergence time estimation because it allows the natural integration of information from the fossil record (in the form of prior statistical distributions describing the ages of nodes in a phylogeny) with information from molecular sequences to estimate node ages, or geological times of divergence, of a species phylogeny [6, 11]. Another advantage of the Bayesian clock dating method is that relaxed-clock models, which allow for violations of the molecular clock, can be easily implemented as the prior on the evolutionary rates for the branches in the phylogeny [6]. MCMCTree allows analyses to be carried out using two popular relaxed-clock models (the autocorrelated and independent log-normally distributed rates models [12, 13]), as well as under the strict molecular clock. Furthermore, MCMCTree allows the user to build flexible fossil calibrations based on various statistical distributions (such as the uniform, truncated-Cauchy, and skew-t, and skew-normal distributions [12, 14, 15]). But perhaps the main advantage of MCMCTree is the implementation of an approximate algorithm to calculate the likelihood [6, 16], which allows the computer analysis of genome-scale datasets to be completed in reasonable amounts of time. The disadvantage of the algorithm is that it only works on fixed tree topologies. Several software packages that perform co-estimation of times and tree topology, but which do not use the approximation, are available [8, 9, 17, 18].
In this protocol, we focus on how to carry out a clock dating analysis with MCMCTree, paying particular attention to diagnosing the MCMC algorithm (the workhorse algorithm within the Bayesian method). Theoretical details of the Bayesian clock dating methods implemented in the program MCMCTree are described in [12–16, 19]. For general introductions to Bayesian statistics and Bayesian molecular clock dating, the reader may consult [20, 21].
#### 2 Software and Data Files
To run the protocol, you will need the MCMCTree and BASEML programs, which are part of the PAML software package for phylogenetic analysis [22]. The source code and compiled versions of the code are freely available from bit.ly/ziheng-paml. All the data files necessary to run the protocol can be obtained from github.com/ mariodosreis/divtime. Please create a directory called divtime in your computer and download all the data files from the GitHub repository. This protocol was tested with PAML version 4.9e.
You are assumed to have basic knowledge of the command line in Unix or Windows (also known as command prompt, shell, or terminal). Simple tutorials for users of Windows, Mac OS, and Linux are posted at bit.ly/ziheng-software. Install MCMCTree and BASEML in your computer system, and make sure you have the mcmctree and baseml executables in your system's path (see bit.ly/ziheng-paml for details on how to do this). Finally, it is helpful (but not indispensable) to have knowledge of the R statistical environment (www.r-project.org). R is quite useful to analyze the output of the program, perform convergence diagnostics, and create nice-looking plots. File R/analysis.R contains some examples for this tutorial.
In this protocol, we will estimate the divergence times of nine primates and one scandentian (an out-group), using a very long alignment (over three million nucleotides long). This dataset was chosen because it can be analyzed very quickly with MCMCTree and it is thus suitable to illustrate the method. We also provide a dataset of 330 species (276 primates and 4 out-groups) with a shorter alignment, to illustrate time estimation in a taxon-rich dataset (see Sect. 5.5 for details).
2.1 Tree and Fossil Calibrations The phylogenetic tree of the ten species is shown in Fig. 1. The tree encompasses members of all the main primate lineages. The ten species were chosen because they have had their complete genomes sequenced. They are a subset of the 36 mammal species analyzed in [23]. File data/10s.tree contains the tree with fossil calibrations in Newick format, which is the format required by MCMCTree. The eight fossil calibrations are shown in Table 1. The calibrations are the same used to estimate primate divergence times in [24]. We discuss fossil calibrations in detail in the "Sampling from the Prior" section. The time unit in the analysis is 100 million years (My). Thus, the calibration B(0.075, 0.10) means the node age is constrained to be between 7.5 and 10 million years ago (Ma).
2.2 Molecular Sequence Data The molecular data are an alignment of 5614 protein-coding genes from the ten species. All ambiguous codon sites were removed, and thus the alignment contains no missing data. The alignment was separated into two partitions: A partition consisting of all the first and second codon positions (2,253,316 nucleotides long) and a partition of third codon positions (1,126,658 nucleotides long). The alignment is a subset of the larger 36-mammal-species alignment in [23]. See also ref. 24. File 10s.phys in the data directory contains the alignment. The alignment is compressed into site patterns (a site pattern is a unique combination of character states in an alignment column) to save disk space.
#### 3 Tutorial
We seek to obtain the posterior distribution (i.e., the estimates) of the divergence times (t) and the molecular evolutionary rates (r, μ, σ 2 ) for the species in the phylogeny of Fig. 1. Here t ¼ (t11, ..., t19) are the nine species divergence times; r ¼ (r1,12, ..., r1,19, r2,12, ...,
Fig. 1 The tree of ten species. Nodes with fossil calibrations are indicated with black dots (see Table 1 for calibration densities). Internal nodes are numbered from 11 to 19 according to the nomenclature used by MCMCTree
a Node numbers as in Fig. <sup>1</sup> <sup>b</sup>
B(a, b, pL, pU) means the calibration is a uniform distribution between a and b, with probabilities p<sup>L</sup> and p<sup>U</sup> that the true node age is outside the calibration bounds. ST(location, scale, shape, df ) means the calibration is a skew-t distribution. S2N( p, location1, scale1, shape1, location2, scale2, shape2) means the calibration is a p:1 p mixture of two skewnormal distributions. G(α, β) means the calibration is a gamma distribution with shape α and rate β. See MCMCTree's manual for the full details on fossil calibration formats. The calibrations are from the primate analysis in [24]
> r2,19) are the 2 - 8 ¼ 16 molecular rates, one per branch and partition (i.e., there are eight branches in the tree and two partitions in the molecular data); and μ ¼ (μ1, μ2) and σ <sup>2</sup> <sup>¼</sup> (σ<sup>2</sup> 1, σ<sup>2</sup> 2) are the mean rates and the log-variance of the rates, for each partition. The posterior distribution is
$$f\left(\mathbf{t}, \mathbf{r}, \mu, \sigma^2 | D\right) \propto f(\mathbf{t}) f\left(\mathbf{r} | \mathbf{t}, \mu, \sigma^2\right) f(\mu) f\left(\sigma^2\right) f(D|\mathbf{r}, \mathbf{t}),$$
where f(t) is the prior on times; f(r|t, μ, σ 2 )f(μ)f(σ 2 ) is the prior on the branch rates, mean rates, and variances of the log-rates; and f(D|t, r) is the molecular sequence likelihood. The prior on the times is constructed by combining the birth-death process with the fossil calibration densities (see ref. 13 for details). The prior on the rates is constructed under a model of rate evolution, assuming, in this tutorial, that the branch rates are independent draws from a log-normal distribution with mean μ<sup>i</sup> and log-variance σ<sup>2</sup> <sup>i</sup> [13].
Bayesian phylogenetic inference using MCMC is computationally expensive because of the repeated calculation of the likelihood on a sequence alignment. The time it takes to compute the likelihood is proportional to the number of site patterns in the alignment. Thus, longer alignments take longer to compute. For genome-scale alignments, the computation time is prohibitive.
MCMCTree implements an approximation to the likelihood that speeds computation time substantially, making analysis of genome-scale data feasible. The approximate likelihood method for clock dating was proposed by Thorne et al. [6] and extended within MCMCTree [16]. The method relies on approximating the log-likelihood surface on the branch lengths by its Taylor expansion. Write ℓ(bj) ¼ log f(D| bj) for the log-likelihood as a function of the branch lengths b<sup>j</sup> ¼ (bj,<sup>i</sup> ¼ rj,iti) for the alignment partition j. The Taylor approximation is
$$\ell(\mathbf{b}\_{j}) \approx \ell(\hat{\mathbf{b}}\_{j}) + (\mathbf{b}\_{j} - \hat{\mathbf{b}}\_{j})^{\mathrm{T}} \mathbf{g}\_{j} + \frac{1}{2} \left(\mathbf{b}\_{j} - \hat{\mathbf{b}}\_{j}\right)^{\mathrm{T}} \mathbf{H}\_{j} (\mathbf{b}\_{j} - \hat{\mathbf{b}}\_{j}),$$
where b^<sup>j</sup> are the maximum likelihood estimates (MLEs) of the branch lengths and g<sup>j</sup> and H<sup>j</sup> are the gradient (vector of first derivatives) and Hessian (matrix of second derivatives) of the log-likelihood surface evaluated at the MLEs for the partition. The approximation can be improved by applying transformations to the branch lengths (see ref. 16 for details).
To use the approximation, one first fixes the topology of the phylogeny, and then estimates the branch lengths for each alignment partition on the fixed tree by maximum likelihood. The gradient and Hessian of the log-likelihood are obtained for each partition at the same time as the MLEs of the branch lengths. Note that parameters of the substitution model—such as the transition/ transversion ratio, κ, in the HKY model or the α parameter in the discrete gamma model of rate variation among sites—are estimated at this step. Thus, different substitution models will generate different approximations, because they will have different MLEs for the branch lengths, gradient, and Hessian. Note that the time it takes to compute the approximate likelihood depends only on the number of species (which determines the size of b and H) and not on the alignment length, that is, once g and H have been calculated, MCMC sampling on the approximation takes the same time regardless of the length of the original alignment.
- 1. Approximate likelihood calculation: First, we will calculate the gradient (g) and Hessian (H) matrix of the branch lengths on the unrooted tree. For this step, we will need to use the MCMCTree and BASEML programs (BASEML will carry out the actual computation of g and H). The substitution model is chosen at this step.
- 2. MCMC sampling from the posterior: Once g and H have been calculated and we have decided on our priors, we can use MCMCTree to perform MCMC sampling from the posterior distribution of times and rates. We will then look at the summaries of the posterior (such as posterior mean times and rates and 95% credibility intervals).
- 3. Convergence diagnostics: The MCMC algorithm is a stochastic algorithm that visits regions of the parameter space in proportion to the posterior distribution. Due to its very nature, it is possible that sometimes the MCMC chain is terminated before it has had a chance to explore the parameter space appropriately. The way to guard against this is to run the analysis two or more times and compare the summary statistics from the two (or more) MCMC chains. If the results from different runs are very similar, then convergence to the posterior distribution can be reasonably assumed.
- 4. MCMC sampling from the prior: Finally, we will sample directly from the prior of times and rates. This is particularly important in Bayesian molecular clock dating because in most cases the prior on times may look quite different from the fossil calibration densities specified by the user. Thus, sampling from the prior allows the user to check the soundness of the prior actually used.
Note that in this protocol we assume the user has chosen a suitable sequence alignment and a phylogenetic tree to carry out the analysis. For genome-scale alignments, it is important that the genes chosen among the various species are orthologous and that the alignment has been checked for accuracy. Several chapters in this volume can guide the user in this purpose.
Go into the gH directory, and open the mcmctree-outBV.ctl file using your favorite text editor. This control file contains the set of parameters necessary for MCMCTree to carry out the calculations of the gradient and Hessian needed for the approximate likelihood method. Figure 2 shows the contents of the mcmctree-outBV. ctl file.
3.2 Calculation of the Gradient and Hessian to Approximate the Likelihood
```
seqfile = ../data/10s.phys
treefile = ../data/10s.tree
ndata = 2
seqtype = 0 * 0: nucleotides; 1:codons; 2:AAs
usedata = 3 * 0: no data (prior); 1:exact likelihood;
* 2: approximate likelihood; 3:out.BV (in.BV)
clock = 2 * 1: global clock; 2: independent rates; 3: correlated rates
model = 4 * 0:JC69, 1:K80, 2:F81, 3:F84, 4:HKY85
alpha = 0.5 * alpha for gamma rates at sites
ncatG = 5 * No. categories in discrete gamma
cleandata = 0 * remove sites with ambiguity data (1:yes, 0:no)?
```
Fig. 2 The gH/mcmctree-outBV.ctl file, with appropriate options to set up calculation of the gradient and Hessian matrix for the approximate likelihood method
Fig. 3 The gH/out.BV file produced by BASEML. The first line has the number of species (10), the second line has the tree topology with MLEs of branch lengths, and the MLEs of branch lengths are given again in the third line. The fourth line contains the gradient, g, followed by the Hessian, H, for partition 1. This file will be renamed in.BV and placed into the mcmc/ directory to carry out MCMC sampling using the approximate likelihood method
> The first two items, seqfile and treefile, indicate the alignment and tree files to be used. The third item, ndata, indicates the number of partitions in the sequence file, in this case, two partitions. The fifth item, usedata, is very important, as it tells MCMCTree the type of analysis being carried out. The options are
0, to sample from the prior; 1, to sample from the posterior using exact likelihood; 2, to sample from the posterior using approximate likelihood; and 3, to prepare the data for calculation of g and H. The last is the option we will be using in this step. The next three items, model, alpha, and ncatG, set up the nucleotide substitution model, in this case the HKY + Gamma model [25]. Finally, the cleandata option tells MCMCTree whether to remove ambiguous data. Our alignment has no ambiguous sites, so this option has no effect in this case.
Using a terminal, go to the gH directory and type
\$ mcmctree mcmctree-outBV.ctl
(Don't type in the \$ as this represents the command prompt!) This will start the MCMCTree program. MCMCTree will prepare several tmp????.\* files and will then call the BASEML program to estimate g and H. For this step to work correctly, the baseml executable must be in your system's path. Once BASEML and MCMCTree have finished, you will notice a file called out.BV has been created. Figure 3 shows part of the contents of this file. The first line indicates the number of species (10), followed by the tree with branch lengths estimated under maximum likelihood for the first partition (first and second codon sites). Next, we have the MLEs of the 17 branch lengths (these are the same as in the tree but printed in a different order). Then we have the gradient, g1, the vector of 17 first derivatives of the likelihood at the branch length MLEs for partition 1. For small datasets, the gradient is usually zero. For large datasets, the likelihood surface is too sharp (i.e., bends downward sharply and it is very narrow at the MLEs), and the gradient is not zero for numerical issues. But this is fine. Next, we have the 17 - 17 Hessian matrix, H1, the matrix of second derivatives of the likelihood at the branch length MLEs for partition 1. If you scroll down the file, you will find the second block, with the tree, branch length MLEs, g2, and H<sup>2</sup> for partition 2 (third codon positions).
#### 3.3 Calculation of the Posterior of Times and Rates
3.3.1 Control File and Priors
Now that we have calculated g and H, we can proceed to MCMC sampling of the posterior distribution using the approximate likelihood method. Copy the gH/out.BV file into the mcmc directory, and rename it as in.BV. Now go into the mcmc directory. There you will find mcmctree.ctl, the necessary MCMCTree control file to carry out MCMC sampling from the posterior. Figure 4 shows the contents of the file. The first item, seed, is the seed for the random number generator used by the MCMC algorithm. Here it is set to 1, which tells MCMCTree to use the system's clock time as the seed. This is useful, as running the program multiple times will generate different outputs.
```
seed = -1
seqfile = ../data/10s.phys
treefile = ../data/10s.tree
mcmcfile = mcmc.txt
outfile = out.txt
ndata = 2
seqtype = 0 * 0: nucleotides; 1:codons; 2:AAs
usedata = 2 * 0: no data (prior); 1:exact likelihood;
* 2:approximate likelihood; 3:out.BV (in.BV)
clock = 2 * 1: global clock; 2: independent rates; 3: correlated rates
RootAge = '<1.0' * safe constraint on root age, used if no fossil for root.
model = 4 * 0:JC69, 1:K80, 2:F81, 3:F84, 4:HKY85
alpha = 0.5 * alpha for gamma rates at sites
ncatG = 5 * No. categories in discrete gamma
cleandata = 0 * remove sites with ambiguity data (1:yes, 0:no)?
BDparas = 1 1 0 * birth, death, sampling
kappa_gamma = 6 2 * gamma prior for kappa
alpha_gamma = 1 1 * gamma prior for alpha
rgene_gamma = 2 40 1 * gammaDir prior for rate for genes
sigma2_gamma = 1 10 1 * gammaDir prior for sigma^2 (for clock=2 or 3)
print = 1 * 0: no mcmc sample; 1: everything except branch rates 2: everything
burnin = 20000
sampfreq = 100
nsample = 20000
```
Fig. 4 The mcmc/mcmctree.ctl file necessary to sample from the posterior distribution using the approximate likelihood method
> The mcmcfile option tells MCMCTree where to save the parameters sampled (divergence times and rates) during the MCMC iterations. Here we will save them to a file named mcmc. txt. Once the MCMC sampling has completed, MCMCTree will read the sample from the mcmc.txt file and generate a summary of the MCMC output. This summary will be saved to a file called out.txt (outfile option).
> The option usedata is set to 2 here, which tells MCMCTree to calculate the likelihood approximately by using the g and H values saved in the in.BV file. Option clock sets the clock model. Here we use clock ¼ 2, which assumes rates are identical, independent realizations from a log-normal distribution [7, 26]. Option RootAge sets the calibration on the root node of the phylogeny, if none are present in the tree file. In our case, we already have a calibration on the root, so this option has no effect. The next three options, model, alpha, and ncatG, have no effect as the substitution model was chosen during estimation of g and H.
> The following options are very important as they determine the prior used in the analysis. BDparams sets the prior on node ages for those nodes without fossil calibrations by using the birth-death process [12]. Here we use 1 1 0, which means node ages are
3.3.2 Running and Summarizing the MCMC uniformly distributed between present time and the age of the root. Options kappa\_gamma and alpha\_gamma set gamma priors for the κ and α parameters in the substitution model. These have no effect as we are using the likelihood approximation. Options rgene\_gamma and sigma2\_gamma set the gamma-Dirichlet prior on the mean substitution rate for partitions and for the rate variance parameter, σ <sup>2</sup> [19]. The prior on the mean rate is Gamma(2, 40), which has mean 0.05 substitutions per time 100 My. A symmetric Dirichlet distribution with concentration parameter equal to 1 is used to spread the rate prior across partitions (thus rgene\_gamma <sup>¼</sup> 2 40 1). See ref. 19 for details. The prior on σ <sup>2</sup> is Gamma(1, 10) which has mean 0.1. A Dirichlet is also used to spread the prior across partitions.
The final block of options, print, burnin, sampfreq, and nsample, control the length and sampling frequency of the MCMC. We will discard the first 20,000 iterations as the burn-in and then print parameter values to the mcmc.txt file every 100 iterations, to a maximum of 20,000 + 1 samples. Thus, our MCMC chain will run for a total of 20,000 + 20,000 - 100 ¼ 2,020,000 iterations.
Go into the mcmc directory and type
\$ mcmctree mcmctree.ctl
This will start the MCMC sampling. First, MCMCTree will iterate the chain for a set number of iterations, known as the burnin. During this period, the program will fine-tune the step sizes for proposing parameters in the chain. Once the burn-in is finished, sampling from the posterior will start. Figure 5 shows a screenshot of MCMCTree in action. The leftmost column indicates the progress of the sampling as a percentage of the total (5%, 10% of total iterations, and so on). The next numbers represent the acceptance proportions, which are close to 30% (this is the result of fine-tuning by the program). After the five acceptance proportions, the programs prints a few parameters to the screen and in the last columns the log-likelihood and the time taken.
The above analysis takes about 2 min and 30 s to complete on a 2.2 GHz Intel Core i7 Processor. Once the analysis has finished, you will see that MCMCTree has created several new files in the mcmc directory. Rename mcmc.txt to mcmc1.txt and out.txt to out1.txt. Now, on the command line, type again
\$ mcmctree mcmctree.ctl
This will run the analysis a second time. The results should be slightly different to the previous run due to the stochastic nature of the algorithm. Once the second run has finished, rename mcmc.
```
0% 0.26 0.39 0.23 0.39 0.28 1.285 1.243 0.588 1.158 0.541 0.321 - 0.192 0.197 -16.9 0:02
(nsteps = 50)
Current Pjump: 0.26200 0.39475 0.23175 0.38650 0.28000 0.27550 0.39200 0.43750
0.40100 0.29725 0.33725 0.27525 0.32275 0.23475 0.23150 0.29875 0.31600 0.27800
0.25300 0.29975 0.29650 0.32575 0.27500 0.61150 0.29850 0.31225 0.35400 0.23200
0.30800 0.28250 0.33050 0.21325 0.22700 0.25900 0.26725 0.26900 0.33150 0.23725
0.31000 0.20700 0.24225 0.61625 0.30675 0.30150 0.32000 0.21975 0.27650 0.22500
0.36650 0.00000
Current finetune: 0.00365 0.00166 0.00586 0.00182 0.00503 0.00697 0.00486 0.00500
0.00835 0.24230 0.21346 0.71942 0.65595 0.01093 0.01230 0.01256 0.00960 0.01492
0.02008 0.02466 0.03547 0.03942 0.04624 0.17077 0.02425 0.04971 0.01513 0.03626
0.03661 0.04475 0.08082 0.00867 0.00949 0.01146 0.00861 0.01133 0.01263 0.02252
0.02728 0.03996 0.03790 0.14736 0.02025 0.04584 0.01209 0.02975 0.02776 0.03389
0.05173 0.00000
New finetune: 0.00313 0.00232 0.00438 0.00248 0.00465 0.00632 0.00675 0.00806
0.01194 0.23972 0.24532 0.65158 0.71499 0.00829 0.00918 0.01250 0.01020 0.01367
0.01654 0.02463 0.03499 0.04345 0.04183 0.47928 0.02411 0.05210 0.01846 0.02714
0.03776 0.04175 0.09064 0.00592 0.00694 0.00969 0.00755 0.01000 0.01422 0.01728
0.02835 0.02644 0.02976 0.42023 0.02079 0.04611 0.01305 0.02100 0.02527 0.02454
0.06589 0.00000
5% 0.34 0.30 0.31 0.32 0.28 1.163 0.981 0.622 0.893 0.464 0.295 - 0.129 0.154 -17.0 0:08
10% 0.35 0.30 0.31 0.32 0.27 1.189 0.943 0.607 0.859 0.457 0.293 - 0.128 0.153 -17.0 0:15
15% 0.36 0.30 0.30 0.31 0.27 1.156 0.920 0.604 0.837 0.457 0.290 - 0.133 0.160 -17.0 0:22
20% 0.35 0.30 0.30 0.32 0.26 1.126 0.908 0.600 0.825 0.453 0.290 - 0.137 0.165 -17.0 0:29
25% 0.36 0.30 0.30 0.31 0.26 1.139 0.912 0.605 0.829 0.458 0.293 - 0.138 0.165 -17.0 0:37
30% 0.36 0.30 0.30 0.31 0.26 1.153 0.918 0.609 0.834 0.460 0.293 - 0.136 0.163 -17.0 0:43
```
Fig. 5 Screenshot of MCMCTree's output during MCMC sampling of the posterior. Different runs of the program will give slightly different output values
> txt to mcmc2.txt and out.txt to out2.txt. If you want to conduct two runs simultaneously, you can create two directories (say r1/ and r2/) and copy the necessary files into them. Then open two terminal windows to start the runs from within each directory.
> Using your favorite text editor, open file out1.txt, which contains the summary of the first MCMC run. Scroll to the end of the file (see screenshot, Fig. 6). You will see the time used by the program (in my case 2:32), the posterior means of the parameters sampled, and three phylogenetic trees in Newick format. The first tree simply has internal nodes labelled with a number. This is useful to compare the tree with the posterior means of times at the end of the file. The second tree is the tree with branch lengths in absolute time units. The third tree is like the second by including the 95% credibility intervals (CIs) of the node ages. At the bottom of the file, you have a table with all the divergence times (from t\_n11 to t\_n19), the mean substitution rates for the two partitions (mu1 and mu2), the rate variation coefficients (sigma2\_1 and sigma2\_2), and finally the log-likelihood (lnL). The table gives the posterior means, equal-tail CIs, and high-posterior-density CIs. For example, the posterior age of the root (node 11, Fig. 1) is 116.8 Ma (95% CI, 144.2–92.4 Ma) while for the divergence
```
ln Lmax (unconstrained) = -4636133.236961
Time used: 2:26
mean of parameters using all iterations
1.16785 0.91766 0.60797 0.83447 0.46464 0.29132 0.17725 0.10441 0.08519 0.
Species tree for FigTree. Branch lengths = posterior mean times; 95% CIs = labels
(1_Tree_shrew, ((2_Bushbaby, 3_Mouse_lemur) 13 , (4_Tarsier, (5_Marmoset, (6_Rhesus, (7_Orangut
(Tree_shrew: 1.167850, ((Bushbaby: 0.607966, Mouse_lemur: 0.607966): 0.309693, (Tarsier: 0.8344
(Tree_shrew: 1.167850, ((Bushbaby: 0.607966, Mouse_lemur: 0.607966) [&95%={0.50317, 0.735468}]:
Posterior mean (95% Equal-tail CI) (95% HPD CI) HPD-CI-width
t_n11 1.1679 (0.9235, 1.4423) (0.9021, 1.4056) 0.5035 (Jnode 18)
t_n12 0.9176 (0.8015, 1.0484) (0.7965, 1.0423) 0.2458 (Jnode 17)
t_n13 0.6080 (0.5032, 0.7355) (0.5019, 0.7337) 0.2318 (Jnode 16)
t_n14 0.8345 (0.7236, 0.9602) (0.7192, 0.9538) 0.2346 (Jnode 15)
t_n15 0.4646 (0.3966, 0.5340) (0.3964, 0.5335) 0.1371 (Jnode 14)
t_n16 0.2913 (0.2526, 0.3380) (0.2499, 0.3333) 0.0833 (Jnode 13)
t_n17 0.1773 (0.1466, 0.2174) (0.1439, 0.2132) 0.0692 (Jnode 12)
t_n18 0.1044 (0.0995, 0.1164) (0.0988, 0.1139) 0.0152 (Jnode 11)
t_n19 0.0852 (0.0758, 0.0981) (0.0746, 0.0958) 0.0212 (Jnode 10)
mu1 0.0269 (0.0221, 0.0334) (0.0217, 0.0328) 0.0111
mu2 0.1110 (0.0898, 0.1396) (0.0877, 0.1364) 0.0488
sigma2_1 0.1370 (0.0607, 0.2833) (0.0484, 0.2511) 0.2027
sigma2_2 0.1634 (0.0755, 0.3201) (0.0625, 0.2883) 0.2258
lnL -17.0026 (-25.9750, -9.8710) (-24.9110, -9.1170) 15.7940
```
Fig. 6 The end of the mcmc/out.txt file produced by MCMCTree at the end of the MCMC sampling of the posterior
> between human and chimp (node 19, Fig. 1) is 8.52 Ma (95% CI, 7.58–9.81 Ma).
> You will also notice that MCMCTree created a file called Fig-Tree.tre. This contains the posterior tree in Nexus format, suitable for plotting in the program FigTree (tree.bio.ed.ac.uk/ software/figtree/). Figure 7 shows the posterior tree plotted in FigTree, with the time unit set to 1 My.
3.4 Convergence Diagnostics of the MCMC Diagnosing convergence of the MCMC chains is extremely important. Several software tools have been written for this purpose. For example, the user-friendly Tracer program (beast.bio.ed.ac.uk/ tracer) can be used to read in the mcmc1.txt and mcmc2.txt files and calculate several convergence statistics. Here we will use R to perform basic convergence tests (check out file R/analysis.R).
> The first step to assess convergence is to compare the posterior means among the different runs. You can visually inspect the posterior means reported in the out1.txt and out2.txt files (Fig. 8), although this may be cumbersome. Figure 8a shows a plot, made with R, of posterior times for run 1 vs. those from run 2. You can see that the points fall almost perfectly on the y ¼ x line, indicating that both runs have converged to the same distribution (hopefully the posterior!).
Fig. 7 The dated primate phylogeny with error bars (representing 95% CIs of node ages), drawn with FigTree. The time unit is 1 My
Another useful statistic to be calculated is the effective sample size (ESS). This gives the user an idea about whether an MCMC chain has been run long enough. Tracer calculates ESS automatically for all parameters. Function coda::effectiveSize in R will do the same. Figure 9 shows the posterior mean, ESS, posterior variance, and standard error of posterior means calculated with R for run 1 of the MCMC. The longer the ESS, the better. As a rule of thumb, one should seek ESS larger than 1000, although this may not always be practical in phylogenetic analysis. Note in Fig. 9 that some estimates have very low ESSs, while others have substantially higher ESSs. For example, t\_n11 has ESS ¼ 76.1, while t\_n19 has ESS ¼ 1261. Running the analysis again and increasing the total number of iterations (e.g., by increasing samplefreq or nsample) will lead to higher ESS values for all parameters.
Let v be the posterior variance of a parameter. The standard error of the posterior mean of the parameter is S.E. ¼ √(v/ESS). This is why having large ESS is important: Large ESS leads to small S.E. and better estimates of the posterior mean. For example, for t\_n11, the posterior mean is 116.8 Ma, with standard error
Fig. 8 Convergence diagnostic plots of the MCMC drawn with R (see R/analysis.R)
Fig. 9 Calculations of posterior mean, ESS, posterior variance, and standard error of the posterior mean in R (see R/analysis.R)
1.53 My (Fig. 9). That is, we have estimated the mean accurately to within 2 - 1.53 My ¼ 3.06 My. To reduce the S.E. by half, you need to increase the ESS four times. Note that independent MCMC runs can be combined into a single run. Thus, you may save time by running several MCMC chains in parallel for computationally expensive analyses, although care must be taken to ensure each chain has run long enough to exit the burn-in phase and explore the posterior appropriately.
Trace plots and histograms are useful to spot problems and check convergence. Figure 8b, c shows trace plots for t\_n19 and t\_n11, respectively. The trace of t\_n19, which has high ESS, looks like a "hairy caterpillar." Compare it to the trace of t\_n11, which has low ESS. Visual inspection of a trace plot usually gives a sense of whether the parameter has an adequate ESS without calculating it. Note that both traces are trendless, that is, the traces oscillate around a mean value (the posterior mean). If you see a persistent trend in the trace (such as an increase or a decrease), that most likely means the MCMC did not converge to the posterior and needs a longer burn-in period.
Figure 8d shows the smoothed histograms (calculated using density in R) for t\_n11 for the two runs. Notice that the two histograms are slightly different. As the ESS becomes larger, histograms for different runs will converge in shape until becoming indistinguishable. If you see large discrepancies between histograms, that may indicate serious problems with the MCMC, such as lack of convergence due to short burn-in or the MCMC getting stuck in different modes of a multimodal posterior.
3.5 MCMC Sampling from the Prior Note that fossil calibrations (such as those of Table 1) are represented as statistical distributions of node ages. MCMCTree uses these distributions to construct the prior on times. However, the resulting time prior used by the program may be substantially different from the original fossil calibrations, because the program applies a truncation so that daughter nodes are younger than their ancestors [14, 27]. Thus, it is advisable to calculate the time prior explicitly by running the MCMC with no data so that it can be examined and compared with the fossil calibrations and the posterior.
Go to the prior directory and type
\$ mcmctree mcmctree-pr.ctl
This will start the MCMC sampling from the prior. File mcmctree-pr.ctl is identical to mcmc/mcmctree.ctl except that option usedata has been set to 0. Sampling from the prior is much quicker because the likelihood does not need to be calculated. It takes about 1 min on the Intel Core i7 for MCMCTree to complete the analysis. Rename files mcmc.txt and out.txt to
Fig. 10 Prior (gray) and posterior (black) density plots of node ages plotted with R (see R/analysis.R)
mcmc1.txt and out1.txt, and run the analysis again. Rename the new files as appropriate. Check for convergence by calculating the ESS and plotting the traces and histograms.
Figure 10 shows the prior densities of node ages obtained by MCMC sampling (shown in gray) vs. the posterior densities (shown in black). Notice that for four nodes t\_n19, t\_n18, t\_n17, and t\_n16, the posterior times "agree" with the prior, that is, the posterior density is contained within the prior density. For nodes t\_n15, t\_n13, and t\_n11, there is some conflict between the prior and posterior densities. However, for nodes t\_n14 and t\_n12, there is substantial conflict between the prior and the posterior. In both cases the molecular data (together with the clock model) suggest the node age is much older than that implied by the calibrations. This highlights the problems in construction of fossil calibrations.
Each fossil calibration represents the paleontologist's best guess about the age of a node. For example, the calibration for the human-chimp ancestor is B(0.075, 0.10, 0.01, 0.20); thus, the calibration is a uniform distribution between 7.5 and 10 million years ago (Ma). The bounds of the calibration are soft, that is, there is a set probability that the bound is violated. In this case the probabilities are 1% for the minimum bound and 20% for the maximum bound. The bound probabilities are asymmetrical because they reflect the nature of the fossil information. Minimum bounds are usually set with confidence because they are based on the age of the oldest fossil member of a clade. For example, the minimum of 7.5 Ma is based on the age of {Sahelanthropus tchadensis, recognized as the oldest fossil within the human lineage [28]. On the other hand, establishing maximum bounds is difficult, as absence of fossils for certain clades cannot be interpreted as evidence that the clade in question did not exist during a particular geological time [29]. Our maximum here of 10 Ma represents the paleontologist's informed guess about the likely oldest age of the clade; however, a large probability of 20% is given to allow for the fact that the node age could be older. The conflict between the prior and posterior seen in Fig. 10 evidences this.
Note that when constructing the time prior, the Bayesian dating software must respect the constraints whereby daughter nodes must be younger than their parents. This means that calibration densities are truncated to accommodate the constraint, with the result that the actual prior used on node ages can be substantially different to the calibration density used (see Sect. 5.4). Detailed analyses of the interactions between fossil calibrations and the time prior and the effect of truncation are given in [14, 27].
#### 4 General Recommendations for Bayesian Clock Dating
Extensive reviews of best practice in Bayesian clock dating are given elsewhere [4, 20, 21, 30, 31]. Here we give a few brief recommendations.
4.1 Taxon Sampling, Data Partitioning, and Estimation of Tree Topology
In this tutorial we used a small phylogeny to illustrate Bayesian time estimation using approximate likelihood calculation. In practical data analysis, it may be desirable to analyze much larger phylogenies (see Sect. 5.5). In large phylogenies, there may be uncertainties in the relationships of some groups. The approximate method discussed here can only be applied to a fixed (known) tree topology. If the uncertainties in the tree are few so that just a handful of tree topologies appear reasonable, the approximate method can be used by analyzing each topology separately [23, 32]. This involves estimating g and H for each topology and then running separate MCMC chains on each topology to estimate the times. Several methods to co-estimate divergence times and tree topology are available [8, 9, 17, 18], although they do not implement the approximate likelihood method and are thus unsuitable for the analysis of genome-scale datasets.
We note that partitioning of sites in genomic datasets may have important effects on divergence time estimation. The infinite-sites theory [13, 33] studies the asymptotic behavior of the posterior distribution of times when the amount of molecular data (measured by the number of partitions and the number of sites per partition) increases in a relaxed-clock dating analysis. This theory shows that increasing the number of sites per partition will have minimal effects on time estimation when the sequences per partition are moderately long (>1000 sites, say), but the precision improves when the number of partitions increases, eventually approximating a limit when the number of partitions is infinite. The theory also predicts that very different time estimates may be obtained if the same genomic sequence alignment is analyzed as one partition or as multiple partitions [34]. Furthermore, while more partitions tend to produce more precise time estimates, with narrow CIs, they may not necessarily be more reliable, depending on the correctness of the fossil calibrations and the appropriateness of the partitioning strategies. Unfortunately it is hard to decide on a good partitioning strategy given the genome-scale sequence data, despite efforts to design automatic partitioning strategies for phylogenetic analysis and divergence time estimation [34–36]. Commonly used approaches partition sites in the alignment by codon position or by protein-coding genes of different relative rates [23]. We recommend the use of the infinite-sites plot [14], in which uncertainty in divergence time estimates (measured as the CI width) is plotted against the posterior mean of times. If the scatter points fall on a straight line, information due to the molecular sequence data has reached saturation, and uncertainty in time estimate is predominantly due to uncertainties in fossil calibrations.
4.2 Selection of Fossil Calibrations Fossil calibrations are one of the most important pieces of information needed to perform divergence time estimation and thus should be chosen after careful consideration of the fossil record, although this may involve some subjectivity [29]. Parham et al. [30] discuss best practice for construction of fossil calibrations. For example, minimum bounds on node ages are normally set to be the age of the oldest fossil member of the crown group. A small probability (say 2.5%) should be set for the probability that the node age violates the minimum bound (e.g., to guard against misidentified or incorrectly dated fossils). Specifying maximum bounds is more difficult, as absence of fossils for a given geological period is not evidence that the clade in question was absent during the period [31]. Current practice is to set the maximum bound to a reasonable value according to the expertise of the paleontologist (see ref. 29 for examples), although a large probability (say 10% or even 20%) may be required to guard against badly specified maximum bounds. Calibration densities based on statistical modeling of species diversification, fossil preservation, and discovery are also possible [15]. In so-called tip-dating approaches, fossil species are included as taxa in the analysis (which may or may not include morphological information for the fossil and extant taxa) [37–39]. Thus, in tip-dating, explicit specification of a fossil calibration density for a node age is not necessary.
4.3 Construction of the Time Prior The birth-death process with species sampling was used here to construct the time prior for nodes in the phylogeny for which fossil calibrations are not available. Varying the birth (μ), death (λ), and sampling (ρ), parameters can result in substantially different time priors. For example, using μ ¼ λ ¼ 1 and ρ ¼ 0 leads to a uniform distribution prior on node ages. This diffuse prior appears appropriate for most analyses. Varying the values of μ, λ, and ρ is useful to assess whether the time estimates are robust to the time prior. Parameter configurations can be set up to generate time densities that result in young node ages or in very old node ages (see p. 381 in [20] for examples).
4.4 Selection of the Clock Model In analysis of closely related species (such as the apes), the clock assumption appears to be appropriate for time estimation. A likelihood ratio test can be used to determine whether the strict clock is appropriate for a given dataset [40]. If the clock is rejected, then Bayesian molecular clock dating should proceed using one of the various relaxed-clock models available [7, 13]. In this case, Bayesian model selection may be used to choose the most appropriate relaxed-clock model [41], although the method is computationally expensive and thus only applicable to small datasets. The use of different relaxed-clock models (such as the autocorrelated vs. the independent log-normally distributed rates) may result in substantially different time estimates (see ref. 32 for an example). In such cases, repeating the analysis under the different clock models may be desirable.
#### 5 Exercises
#### 5.1 Autocorrelated Rate Model
Modify file mcmc/mcmctree.ctl and set clock <sup>¼</sup> <sup>3</sup>. This activates the autocorrelated log-normal rates model, also known as the geometric Brownian motion rates model [6, 13]. Run the MCMC twice and check for convergence. Compare the posterior times obtained with those obtained under the independent log-normal model (clock <sup>¼</sup> <sup>2</sup>). Are there any systematic differences in node age estimates between the two analyses? Which clock model produces the most precise (i.e., narrower CIs) divergence time estimates?
#### 5.2 MCMC Sampling with Exact Likelihood Calculation Modify file mcmc/mcmctree.ctl and set clock <sup>¼</sup> <sup>2</sup> (independent rates), usedata <sup>¼</sup> <sup>1</sup> (exact likelihood), burnin <sup>¼</sup> <sup>200</sup>, sampfreq <sup>¼</sup> <sup>2</sup>, and nsample <sup>¼</sup> <sup>500</sup>. These last three options will lead to a much shorter MCMC chain, with a total of 1200 iterations. Run the MCMC sampling twice, and check for convergence using the ESS, histograms, and trace plots. How long does it take for the sampling to complete? Can you estimate how long it would take to run the analysis using 2,020,000 iterations, as long as for the approximate method of Sect. 3.3.2? Did the two chains converge despite the low number of iterations?
5.3 Change of Fossil Calibrations There is some controversy over whether {Sahelanthropus, used to set the minimum bound for the human-chimp divergence, is indeed part of the human lineage. The next (younger) fossil in the human lineage is {Orrorin which dates to around 6 Ma. Modify file data/10s.tree and change the calibration in the humanchimp node to B(0.057, 0.10, 0.01, 0.2). Also change the calibration on the root node to B(0.615, 1.315, 0.01, 0.05). Run the MCMC analysis with the approximate method and again sampling from the prior. Are there any substantial differences in the posterior distributions of times under the new fossil calibrations? Which nodes are affected? How bad is the truncation effect among the calibration densities and the prior?
5.4 Comparing Calibration Densities and Prior Densities This is a difficult exercise. Use R to plot the prior densities of times sampled using MCMC (the same as in Fig. 10). Now try to work out how to overlay the calibration densities onto the plots. For example, see Fig. 3 in [23] for an idea. First, write functions that calculate the calibration densities. The dunif function in R is useful to plot uniform calibrations. Functions sn::dsn and sn::dst (in the SN package) are useful to plot the skew-t (ST) and skewnormal (SN) distributions. Calibration type S2N (Table 1) is a mixture of two skew-normal distributions [15]. How do the sampled priors compare to the calibration densities? Are there any substantial truncation effects?
5.5 Time Estimation in a Supermatrix of 330 Species Good taxon sampling is critical to obtaining robust estimates of divergence times for clades. In the data/ directory, an alignment of the first and second codon positions from mitochondrial protein-coding genes from 330 species (326 primate and 4 out-group species) is provided, 330s.phys, with corresponding tree topology, 330s.tree. First, place the fossil calibrations of Table 1 on the appropriate nodes of the species tree. Then obtain the gradient and Hessian matrix for the 330-species alignment using the HKY + G model. Finally, estimate the divergence times on the 330-species phylogeny by using the approximate likelihood method. How does taxon sampling affect node age estimates when comparing the 10-species and 330-species trees? How does uncertainty in node ages in the large tree, which was estimated on a short alignment, compare with the estimates on the small tree, but with a large alignment?
#### References
species divergence times. Syst Biol 67 (1):61–77
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Genome Evolution in Outcrossing vs. Selfing vs. Asexual Species
#### Sylvain Gle´min, Cle´mentine M. Franc¸ois, and Nicolas Galtier
#### Abstract
A major current molecular evolution challenge is to link comparative genomic patterns to species' biology and ecology. Breeding systems are pivotal because they affect many population genetic processes and thus genome evolution. We review theoretical predictions and empirical evidence about molecular evolutionary processes under three distinct breeding systems—outcrossing, selfing, and asexuality. Breeding systems may have a profound impact on genome evolution, including molecular evolutionary rates, base composition, genomic conflict, and possibly genome size. We present and discuss the similarities and differences between the effects of selfing and clonality. In reverse, comparative and population genomic data and approaches help revisiting old questions on the long-term evolution of breeding systems.
Key words Breeding systems, GC-biased gene conversion, Genome evolution, Genomic conflicts, Selection, Transposable elements
#### 1 Introduction
In-depth investigations on genome organization and evolution are increasing and have revealed marked contrasts between species, e.g., evolutionary rates, nucleotide composition, and gene repertoires. However, little is still known on how to link this "genomic diversity" to the diversity of life history traits or ecological forms. Synthesizing previous works in a provocative and exciting book, M. Lynch asserts that variations in fundamental population genetic processes are essential for explaining the diversity of genome architectures while emphasizing the role of the effective population size (Ne) and nonadaptive processes [1]. Life history and ecological traits may influence population genetic parameters, including Ne, making it possible to link species' biology and their genomic organization and evolution (e.g., [2–7])
Among life history traits affecting population genetic processes, breeding systems are pivotal as they determine the way genes are transmitted to the next generation (Fig. 1). Outcrossing,
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_11, © The Author(s) 2019
**asexuality**
Fig. 1 Reproduction and genotype transmission in outcrossing, selfing, and asexual species. In outcrossers, parental and recombinant (dotted lines) gametes from distinct zygotes are shuffled at generation n + 1. In selfers, only gametes produced by a given zygote can mate, which quickly increases homozygosity and reduces the recombination efficacy. Asexuals do not undergo meiosis or syngamy. They reproduce clonally
sexual species (outcrossers) reproduce through the alternation of syngamy (from haploid to diploid) and meiosis (from diploid to haploid), with random mating of gametes from distinct individuals at each generation. Outcrossing is a common breeding system that is predominant in vertebrates, arthropods, and many plants, especially perennials, etc. [8, 9]. Selfing species (selfers) also undergo meiosis, but fertilization only occurs between gametes produced by the same hermaphrodite individual. Consequently, diploid individuals from selfing species are highly homozygous (FIS ~ 1; see, for instance, ref. 10)—heterozygosity is divided by two at each generation, and the two gene copies carried by an individual have a high probability of being identical by descent. Selfing is common in various plant families (e.g., Arabidopsis thaliana), mollusks, nematodes (e.g., Caenorhabditis elegans), and platyhelminthes, among others [8, 9]. Note that many sexual species have intermediate systems in which inbreeding and outbreeding coexist. In organisms with a prolonged haploid phase (such as mosses, ferns, or many algae and fungi), a more extreme form of selfing can occur by taking place during the haploid phase (haploid selfing or intragametophytic selfing), leading instantaneously to genome-wide homozygosity [11]. Clonal asexual species, finally, only reproduce via mitosis, so that daughters are genetically identical to mothers unless a mutation occurs. In diploid asexuals, homologous chromosomes associated in a given zygote do not segregate in distinct gametes—they are co-transmitted to the next generation in the absence of any haploid phase. In contrast to selfing species, individuals from asexual diploid species tend to be highly heterozygous (FIS ~ 1, [12]), since any new mutation will remain at the heterozygote stage forever, unless the same mutation occurs in the homologous chromosome. Clonality is documented in insects (e.g., aphids), crustaceans (e.g., daphnia), mollusks, vertebrates, and angiosperms, among others [13–16]. As for selfing, clonality can also be partial, with sexual reproduction occurring in addition or in alternation with asexual reproduction. In addition to this common form of asexuality, other forms such as automixis imply a modified meiosis in females where unfertilized diploid eggs produce offspring potentially diverse and distinct from their mother, leading to different levels of heterozygosity [13]. This diversity of reproductive systems should be kept in mind, but for clarity we will mainly compare outcrossing, diploid selfing, and clonality.
Through the occurrence, or not, of syngamy, recombination, and segregation, breeding systems affect population genetic parameters (effective population size, recombination rate, efficacy of natural selection; Fig. 2) and thus, potentially, genomic patterns. A large corpus of population genetic theory has been developed to study the causes and consequences of the evolution of breeding systems (Table 1). Thanks to the exponentially growing amount of genomic data, and especially data from closely related species with contrasted breeding systems, it is now possible to test these theoretical predictions. Conversely, genomic data may help in understanding the evolution of breeding systems. Genomes should record the footprints of transitions in breeding systems and help in testing the theory of breeding system evolution in the long run, e.g., the "dead-end hypothesis," which posits that selfers and asexuals are doomed to extinction because of their inefficient selection and low adaptive potential [17, 18]. Since the first edition of this book, several theoretical developments have clarified the population genetics consequences of the different breeding systems, and empirical evidences have been accumulating, partly changing our view of breeding system evolution and consequences, especially for asexual organisms. We first review and update the consequences of
#### Table 1
Summary of the major theoretical predictions regarding breeding systems and evolutionary genomic variables, with outcrossing being taken as reference
TE transposable element abundance, LD linkage disequilibrium
breeding systems on genome evolution and then discuss and re-evaluate how evolutionary genomics shed new light on the old question of breeding system evolution.
#### 2 Contrasted Genomic Consequences of Breeding Systems
2.1 Consequences of Breeding Systems on Population Genetics Parameters Sex involves an alternation of syngamy and meiosis. In outcrossing sexual species, random mating allows alleles to spread across populations, while segregation and recombination (here in the sense of crossing-over) associated with meiosis generate new genotypic and haplotypic combinations. This strongly contrasts with the case of selfing and asexual species. In such species, alleles cannot spread beyond the lineage they originated from because mating occurs within the same lineage (selfers) or because syngamy is suppressed (asexuals). Recombination, secondly, is not effective in non-outcrossers. In selfers, while physical recombination does occur (r0), effective recombination (re) is reduced because it mainly occurs between homozygous sites, and it completely vanishes under complete selfing: for tight linkage, r<sup>e</sup> ¼ r0(1 FIS), where FIS is the Wright's fixation index [19], whereas for looser linkage, effective recombination is more reduced than predicted by this simple expression [20–22]. In asexuals, physical recombination is suppressed (r<sup>0</sup> ¼ r<sup>e</sup> ¼ 0). High levels of linkage disequilibrium (nonrandom association of alleles between loci) could therefore be expected in selfers and asexuals. The observed data are mainly consistent with these predictions. In the selfing model species Arabidopsis thaliana, LD extends over a few hundreds of kb, while in maize, an outcrosser, LD quickly vanishes beyond a few kb [23]. In a meta-analysis, Gle´min et al. [24] also found higher LD levels in selfers than in outcrossers. Beyond pairwise LD, selfing also generates higher-order associations, such as identity disequilibria (the excess probability of being homozygote at several loci, [25]) that alter population genetics functioning compared to outcrossing populations (e.g., [26]).
Theory also predicts that the effective population size, Ne, depends on the breeding system (Fig. 2). First, compared to outcrossers, selfing is expected to directly lower N<sup>e</sup> by a factor 1 + FIS by reducing the number of independent gametes sampled for reproduction [27]. From a coalescent point of view, selfing reduces coalescent time (again by the same factor 1 + FIS). Under outcrossing, two gene copies gathered in a same individual either directly coalesce or move apart at the preceding generation. Selfing prolongs the time spent within an individual, hence the probability of coalescing [19, 28]. In diploid asexuals, the picture is less obvious. Since genotypes, not alleles, are sampled, Balloux et al. [12] distinguished between the genotypic and allelic effective size. The genotypic effective size equals N, not 2N, i.e., the actual population size, similarly to the expectation under complete selfing. On the contrary, the allelic effective size tends toward infinity under complete clonality because genetic diversity within individuals cannot be lost [12]. This corresponds to preventing coalescence as long as gene copies are transmitted clonally [29, 30]. However, very low level of sex (higher than 1/2N) is sufficient to retrieve standard outcrossing coalescent behavior [29, 30], and as far as natural selection is concerned (see below), the genotypic effective size is what matters [31]. The ecology of selfers and asexuals may also contribute to decreasing N<sup>e</sup> as they supposedly experience more severe bottlenecks than outcrossers [32, 33]. On the contrary, higher population subdivision in selfers could contribute to increasing N<sup>e</sup> at the species scale. However, Ingvarsson [34] showed that, under most conditions, the extinction/recolonization dynamics is predicted to decrease N<sup>e</sup> in selfers, at both the local and metapopulation scale. Finally, because of low or null effective recombination, hitchhiking effects—the indirect effects of selection at a locus on other linked loci—reduce N<sup>e</sup> further [35]. Under complete selfing or clonality, because of full genetic linkage, selection at a given locus affects the whole genome. Most forms of selection, and especially directional selection, reduce the number of gene copies contributing to the next generation by removing deleterious alleles to the benefit of advantageous ones. Because of linkage, such a reduction spreads over the rest of the genome, globally reducing the effective population size (sensu lato) in non-outcrossing species. Background selection, the reduction in N<sup>e</sup> due to the removal of deleterious mutations at linked loci, can be particularly severe in highly selfing and clonal population, potentially reducing N<sup>e</sup> by one order of magnitude or more [22, 36]. And this effect is expected to be stronger in asexuals than in selfers [36]. In the predominantly selfing nematode C. elegans, nucleotide diversity has been shown to be reduced genome wide by both background selection [37] and selective sweeps [38], and in a comparative analysis, the effect of linked selection has shown to be more pronounced in selfing than in outcrossing species [39].
As genetic diversity scales positively with Neμ, where μ is the mutation rate, selfers are expected to be less polymorphic than outcrossers. Asexuals should also exhibit lower genotypic diversity, but the prediction is not clear for allelic diversity (see above). However, because of the lack of recombination, haplotype diversity should be lower for both breeding systems. The effect of selfing on the polymorphism level is well documented, and empirical data mainly agree with the theoretical predictions. Selfing species tend to be more structured, less diverse, and straightforwardly more homozygotes than outcrossers [6, 24, 40, 41]. Much fewer data exist regarding diversity levels in asexuals, but the available datasets confirm that genotypic diversity, at least, is usually low in such species (see discussion in ref. 12). At the population level, a recent comparative analysis of sexual and asexual Aptinothrips rufus grass thrips confirmed the expected lower nuclear genetic diversity of asexual populations while also evidencing that some asexuals with extensive migration can feature very high mitochondrial genetic diversity [42].
These predictions concerning polymorphism patterns implicitly assumed that mutation rates are the same among species with contrasted breeding systems. However, modifications in breeding systems can also affect various aspects of the species life cycle potentially related to the mutation rate. In asexuals, for instance, loss of spermatogenesis can reduce mutation rates, while loss of the dormant sexual phase can increase them (reviewed in [43]). Mutation rates can also be decreased in non-outcrossers due to the loss of recombination, which can be mutagenic [44, 45]. In selfers, meiosis and physical recombination do occur. However, the specific mutagenic process during meiosis depends on the level of heterozygosity, such as indel-associated mutations (IDAM): heterozygote indels could increase the point mutation rate at nearby nucleotides because of errors during meiosis [46, 47]. Consistent with this prediction, the IDAM process more strongly affects the outcrossing wild rice, Oryza rufipogon, than the very recent selfer and weakly
Fig. 3 Substitution rates relative to the neutral case (dN/dS) in outcrossers (thin lines), selfers (bold line), and asexuals (dotted lines) for different mutation dominance levels. The fitness of the resident, heterozygote, and homozygote mutant genotypes are 1, 1 hs, and 1 s, respectively. For asexuals, it is necessary to consider two substitution rates corresponding to the initial fixation of heterozygotes and the ultimate fixation of complete homozygote mutants from an initially heterozygote population [31]. Population size: N ¼ 10,000. To highlight the difference between selfers and asexuals due to segregation, demographic and hitchhiking effects reducing N<sup>e</sup> in asexuals and selfers are not taken into account
heterozygous domesticated rice, O. sativa. A. thaliana, a more ancient and mostly homozygous selfer, is very weakly affected by IDAM [48]. Overall, these processes should globally contribute to lowering mutation rates, and thus polymorphism, in selfing and asexual species.
The effective population size strongly affects the outcome of natural selection. The probability of fixation of a new mutation is a function of the Nes product, where s is the selection coefficient ([49] and see Fig. 3). As N<sup>e</sup> is reduced, a higher proportion of mutations behave almost neutrally. Weakly deleterious alleles can thus be fixed, while weakly advantageous ones can be lost. Genetic associations among loci generated by selfing and clonality also induce selective interferences [26, 50]. Because of their reduced effective population size and recombination rate, selection is thus expected to be globally less effective in selfers and asexuals than in outcrossers, which should result in various footprints at the molecular level (Table 1). Assuming that most mutations are deleterious (with possible back compensatory mutations), both the ratio of non-synonymous to synonymous polymorphism, πN/ πS, and the ratio of non-synonymous to synonymous substitutions, dN/dS, are predicted to be higher in selfers and asexuals than in outcrossers. Codon usage should also be less optimized in selfers and asexuals than in outcrossers.
Contrary to polymorphism surveys, few studies have tested these predictions empirically (Table 2). In the few available
#### 2.2 Breeding Systems and Selection Efficacy
2.2.1 Drift and Recombination: Parallel Reduction in Selection Efficacy in Selfers and Asexuals?
Table 2 Summary of studies comparing patterns of molecular evolution between different breeding systems
data
+ and
aNo
bTerminal vs. internal branches not controlled
cPossible
confounding
effect of hybrid origin of one asexual lineage
comparative studies, contrasted patterns were observed between selfers and asexuals. Compared to sexual ancestors, recent asexual lineages show a marked increase in the dN/dS ratio in Daphnia ([51] but see below), Timema stick insects [52], gastropods Campeloma [53] and Potamopyrgus [54], and the plant Boechera [55], in agreement with theoretical predictions (Table 2). However, no significant effect of asexuality on dN/dS was found in four aphid species [56] and in the plant Ranunculus auricomus [57]. Bdelloid rotifers, long considered as ancient asexuals (see below), exhibit a higher πN/πS ratio but not a higher dN/dS ratio than comparable sexual groups, suggesting that mildly deleterious mutations can segregate at a higher frequency in asexuals but are eventually removed. A higher πN/πS ratio in asexual lineages than in sexual relatives was reported from transcriptome data in Oenothera primroses [58] and Lineus nemerteans [59]. Note however that in the latter case, the increased πN/πS is primarily explained by the hybrid nature of the asexual Lineus pseudolacteus (Table 2). The recent origin of asexuality through introgression also challenges the interpretation of elevated dN/dS ratio in the mitochondrial genome of asexual lineages of Daphnia pulex [51], as less than 1% of mutations on the branches leading to asexual lineages would have arisen after the transition to asexuality [60]. Here, rather than being the direct cause of genomic degradation, asexuality may have evolved in already-degraded lineages.
All predictions are not equally supported by data in selfers. Polymorphism-based measures mostly support reduction in selection efficiency in selfers in various plant species, and this was recently confirmed by a meta-analysis of genome-wide polymorphism data ([6] and see Table 2). On the contrary, as far as dN/dS or base composition are compared, most studies, in plants, fungi, and animals, did not find evidence of relaxed selection in selfers (Table 2). A recent origin of selfing is often invoked to explain that effect of selfing is rarely observed in species divergence (e.g., [61, 62–64]), whereas a recent transition to selfing can leave a clear signature of relaxed selection at the polymorphism level [65]. In contrast, in the freshwater snail Galba truncatula where selfing is supposed to be old and ancestral to a clade of several species, relaxed selection in the selfing lineage was also observed at the divergence level [66]. The same rationale should apply to asexual species. However, in Campeloma, Potamopyrgus, Timema, and Boechera, clonality is also recent, yet the expected patterns are observed at the divergence level. The reduction in N<sup>e</sup> could simply be less severe in selfers than in asexuals as predicted by background selection models [36]. Furthermore, complete selfing is hardly ever noted in natural populations; residual outcrossing typically occurs. Among hitchhiking effects, some are very sensitive to the recombination level, such as Muller's ratchet [67], weak Hill-Robertson interferences [50], or hitchhiking of deleterious mutations during selective sweeps [68, 69]. If such mechanisms are the main cause of reduction of N<sup>e</sup> in selfers, then even a low recombination rate could be enough to maintain the selection efficacy. This is suggested by genomic patterns across recombination gradients in outcrossing species. In primates, no effect of recombination on the selection efficacy has been detected [70]. In Drosophila, Haddrill et al. [71] found little evidence of reduced selection in low recombining regions, except when recombination was fully suppressed, as in Y chromosomes. Differences between selfers and asexuals could thus simply result from different degrees of residual outcrossing. However, as stated above, selfers and asexuals also fundamentally differ as far as segregation is concerned, as we now discuss in more detail.
2.2.2 Segregation: Dealing with Heterozygotes Selfing affects the selection efficacy by increasing homozygosity and thus exposing recessive alleles to selection. This effect can counteract the effect of reducing Ne. Considering the sole reduction in N<sup>e</sup> due to non-independent gamete sampling, selection is less efficient under partial selfing for dominant mutations but more efficient for recessive ones (Fig. 3, and see ref. 72). More precisely, Gle´min [73] determined the additional reduction in N<sup>e</sup> (due to hitchhiking and demographic effects) necessary to overcome the increased selection efficacy due to homozygosity. This additional reduction can be high for recessive mutations. On the contrary, the lack of segregation in asexuals reduces selection efficacy and increases the drift load, as heterozygotes can fix [31]. The effects of selfing and clonality on the fixation probability of codominant, recessive, or dominant mutations are summarized in Fig. 3. Note that segregation may also have indirect effects. When recombination is suppressed, Muller's ratchet is supposed to reduce N<sup>e</sup> and contribute to the fixation of weakly deleterious alleles [74]. In selfers, the purging of partially recessive deleterious alleles slows down the ratchet [67], which suggests that the fixation of deleterious alleles at linked loci would be lower in selfers than in asexuals. The same mechanism also contributes to weaker background selection in selfers than in asexuals (see above, [36]). In the extreme case of intra-gametophytic selfing, purging could be even more efficient at removing deleterious alleles [11], as it has been suggested for moss species [75]. Segregation at meiosis could thus partly explain the differences between selfers and asexuals, but more data are clearly needed to confirm this hypothesis.
The two opposite effects of drift and segregation in selfers should also affect adaptive evolution. In outcrossers, new beneficial mutations are more likely to be rapidly lost if recessive, as they are initially present in heterozygotes and masked to selection—a process known as Haldane's sieve [76]. By unmasking these mutations in homozygotes, selfing could help adaptive evolution from recessive mutations [72, 73]. However, this advantage of selfing disappears when adaptation proceeds from pre-existing variation because homozygotes can also be present in outcrossers [77]. Selective interference in selfers also reduces their advantage of not experiencing Haldane's sieve, especially for weakly beneficial mutations [21], and the effect of background should globally reduce the rate of adaptation [73, 77, 78]. Conversely, the lack of segregation in asexuals delays the complete fixation of an advantageous mutation. Once a new advantageous mutation gets fixed in the heterozygotic state, additional lag time until occurrence and fixation of a second mutation is necessary to ensure fixation [79]. Little is known about the dominance levels of new adaptive mutations, but a survey of QTL fixed during the domestication process in several plant species confirmed the absence of Haldane's sieve in selfers compared to outcrossers [80]. This mostly corresponds to strong selection on new mutations or mutations in low initial frequencies in the wild populations. More generally, the effect of selfing on adaptive evolution will depend on the distribution of dominance and selective effects of mutations and the magnitude of genetic drift and linkage.
Few studies have tested for difference in positive selection between selfers and outcrossers. In their survey of sequence polymorphism data in flowering plants, Gle´min et al. [24] found, on average, more genes with a signature of positive selection in outcrossers than in selfers assessed by the McDonald-Kreitman test [81]. An extension of this method—where non-synonymous vs. synonymous polymorphism data are used to calibrate the distribution of the deleterious effects of mutations and then attribute the excess non-synonymous divergence observed to positive selection [82]—was applied to one plant [83] and one freshwater snail dataset. In both studies, a large fraction of non-synonymous substitutions was estimated to be adaptive in the outcrossing species (~40% in the plant Capsella grandiflora and ~55% in the snail Physa acuta), whereas this proportion was not significantly different from zero in the selfer (Arabidopsis thaliana and Galba truncatula, respectively). Based on methods where the dN/dS ratio is allowed to vary both among branches and sites, a comparative analysis of two outcrossing and two selfing Triticeae species [84] suggested that adaptive substitutions may have specifically occurred in the outcrossing lineages. This would contribute to explaining why selfing lineages did not show a higher dN/dS ratio than outcrossing ones (see above and Table 2). So the data available so far support an increased rate of adaptation in outcrossing species, suggesting that the effects of drift and linkage overwhelm the advantage of avoiding Haldane's sieve. A similar approach was used in Oenothera species suggesting also reduced adaptive evolution in clonal compared to sexual lineages [85].
Finally, the classical assumption of a lack of segregation in asexuals must be modulated. First, in some form of asexuality, such as automixis, female meiosis is retained, and diploidy restoration occurs by fusion or duplication of female gametes. Depending on how meiosis is altered, automixis generates a mix of highly heterozygous and highly homozygous regions along chromosomes. The genomes of such species could thus exhibit a gradient of signatures of selfing and diploid clonal evolution [86]. Secondly, mitotic recombination and gene conversion in the germline of asexual lineages can also reduce heterozygosity at a local genomic scale. Mitotic recombination has been well documented in yeast (see review in ref. 87) and also occurs in the asexual trypanosome T. b. gambiense [88] and in asexual Daphnia lineages [60, 89, 90]. If its frequency is of the order or higher than mutation rates, as reported in yeast and Daphnia, asexuals would not suffer much from the lack of segregation at meiosis. Especially, during adaptation, the lag time between the appearance of a first beneficial mutation and the final fixation of a mutant homozygote could be strongly reduced [87]. However, such mechanisms of loss of heterozygosity also rapidly expose recessive deleterious alleles in heterozygotes and generate inbreeding-depression-like effects [60].
2.2.3 Selection on Genetic Systems So far, we have only considered the immediate, mechanistic effects of breeding systems on population genetic parameters. Breeding systems, however, can also affect the evolution of genetic systems themselves, which modulates previous predictions. Theoretical arguments suggested that selfing, even at small rates, greatly increases the parameter range under which recombination is selected for [91–93]. These predictions have been confirmed in a meta-analysis in angiosperms in which outcrossers exhibited lower chiasmata counts per bivalent than species with mixed or selfing mating systems [94]. Higher levels of physical recombination (r0) could thus help break down LD and reduce hitchhiking effects. This could contribute to explaining why little evidence of longterm genomic degradation has been observed in selfers, compared to asexuals.
> Breeding systems may also affect selection on mutation rates. Since the vast majority of mutations are deleterious, mutation rates should tend toward zero, up to physiological costs of further reducing mutation rates being too high (e.g., [95, 96]). Under complete linkage, a modifier remains associated with its "own" mutated genome. Selection should thus favor lower mutation rates in asexuals and selfers (e.g., [95, 96]). However, Lynch recently challenged this view and suggested a lower limit to DNA repair may be set by random drift, not physiological cost [97]. Such a limit should thus be higher in asexuals and selfers. Asexuality is often associated with very efficient DNA repair systems (reviewed in [43]), supporting the view that selection for efficient repair may overwhelm drift in asexual lineages. Alternatively, only groups
having high-fidelity repair mechanisms could maintain asexuality in the long run. More formal tests of mutation rate differences between breeding systems are still scarce. The phylogenetic approach revealed no difference in dS, as a proxy of the neutral mutation rate, between A. thaliana and A. lyrata [61], nor did a mutation accumulation experiment that compared the deleterious genomic mutation rate between Amsinckia species with contrasted mating systems [98]. A similar experiment in Caenorhabditis showed that the rate of mutational decay was, on average, fourfold greater in gonochoristic outcrossing taxa than in the selfer C. elegans [99]. Recent mutation accumulation experiments on Daphnia pulex suggested a slightly lower mutation rate in obligate than in facultative asexual genotypes, except for one mutator phenotype which evolved in an asexual subline [90]. Overall, these results do not support Lynch's hypothesis of mutation rates being limited by drift in asexual and selfing species. However, such experiments are still too scarce, and quantifying how mutation rates vary or not with breeding systems is a challenging issue that requires more genomic data.
2.3 Breeding Systems and Genomic Conflicts Outcrossing species undergo various sorts of genetic conflict. Sexual reproduction directly leads to conflicts within (e.g., for access to mating) and between sexes (e.g., for resource allocations between male and female functions or between offspring). In selfers and asexuals, such conflicts occur because mates are akin or because mating is absent [100, 101]. Outcrossers are also sensitive to epidemic selfish element proliferation and to meiotic drive, because alleles can easily spread over the population through random mating. In contrast, selfers and asexuals should be immune to such genomic conflicts because selection only occurs between selfing or asexual lineages so that selfish elements should be either lost or evolve into commensalists or mutualists [102].
2.3.1 Relaxation of Sexual Conflicts in Selfers and Asexuals Some genes involved in sexual reproduction are known to evolve rapidly because of recurrent positive selection [103]. Arm races for mating or for resource allocation to offspring are the most likely causes of this accelerated evolution. In selfers and asexuals, selection should be specifically relaxed on these genes, not only because of low recombination and effective size but mainly because the selection pressure per se should be suppressed. According to this prediction, in the outcrosser C. grandiflora, 6 out of the 20 genes that show the strongest departure from neutrality are reproductive genes and under positive selection. This contrasts with the selfer A. thaliana, for which no reproductive genes are under positive selection [83].
More specifically, two detailed analyses provided direct evidence of relaxed selection associated with sexual conflict reduction. In the predominantly selfer C. elegans, some males deposit a copulatory plug that prevents multiple matings. However, other males do not deposit this plug. A single gene (plg-1), which encodes a major structural component of this plug, is responsible for this dimorphic reproductive trait [104]. Loss of the copulatory plug is caused by the insertion of a retrotransposon into an exon of plg-1. This same allele is present in many populations worldwide, suggesting a single origin. The strong reduction in male-male competition following hermaphroditism and selfing evolution explains that no selective force opposes the spread of this loss-of-function allele [104, 105]. In A. thaliana, similar relaxed selection has been documented in the MEDEA gene, an imprinted gene directly involved in the male vs. female conflict. MEDEA is expressed before fertilization in the embryo sac and after fertilization in the embryo and the endosperm, a tissue involved in nutrient transfer to the embryo. In A. lyrata, an outcrossing relative to A. thaliana, MEDEA could be under positive [106] or balancing selection [107], in agreement with permanent conflicting pressures for resource acquisition into embryos between males and females. Conversely, this gene evolved under purifying selection in A. thaliana, where the level of conflict is reduced.
Male vs. female diverging interests are also reflected by cytonuclear conflicts. When cytoplasmic inheritance is uniparental, as in most species, cytoplasmic male sterility (CMS) alleles favoring transmission via females at the expense of males can spread in hermaphroditic outbreeding species, leaving room for coevolution with nuclear restorers. Maintenance of CMS/non-CMS polymorphism leads to stable gynodioecy [108]. In selfers, CMS mutants also reduce female fitness—because ovules cannot be fertilized and are thus selected against. In the genus Silene, the mitochondrial genome of gynodioecious species exhibits molecular signatures of adaptive and/or balancing selection. This is likely due to cytonuclear conflicts as this is not, or is less, observed in hermaphrodites and dioecious [109–111]. Although less studied, cyto-nuclear conflicts are also expected in purely hermaphroditic species. In a recent study in A. lyrata, Foxe and Wright [112] found evidence of diversifying selection on members of a nuclear gene family encoding transcriptional regulators of cytoplasmic genes. Some of them show sequence similarity with CMS restorers in rice. Given the putative function of these genes, such selection could be due to ongoing cyto-nuclear coevolution. Interestingly, in A. thaliana, these genes do not seem to evolve under similar diversifying selection, as expected in a selfing species where conflicts are reduced.
2.3.2 Biased Gene Conversion as a Meiotic Drive Process: Consequences for Nucleotide Landscape and Protein Evolution
GC-biased gene conversion (gBGC) is a kind of meiotic drive at the base pair scale that can also be strongly influenced by breeding systems. In many species, gene conversion occurring during double-strand break recombination repair is biased toward G and C alleles (reviewed in [113]). This process mimics selection and can rapidly increase the GC content, especially around recombination hotspots [114, 115], and, more broadly, can affect genome-wide nucleotide landscapes. For instance, it is thought to be the main force that shaped the isochore structure of mammals and birds [116]. gBGC has been mostly studied by comparing genomic regions with different rates of (crossing-over) recombination (reviewed in [116]). However, comparing species with contrasted breeding systems offers a broader and unique opportunity to study gBGC. gBGC cannot occur in asexuals because recombination is lacking. Selfing is also expected to reduce the gBGC efficacy because meiotic drive does not occur in homozygotes [117]. To our knowledge, GC content has never been compared between sexual and asexual taxa, but there have been comparisons between outcrossers and selfers.
As expected, no relationship was found between local recombination rates and GC-content in the highly selfing Arabidopsis thaliana [117], and Wright et al. [118] suggested that the (weak) differences observed with the outcrossing A. lyrata and Brassica oleracea could be due to gBGC. Much stronger evidence has been obtained in grasses. Grasses are known to exhibit unusual genomic base composition compared to other plants, being richer and more heterogeneous in GC-content [119], and direct and indirect evidences of gBGC have been accumulating [119, 120–122]. Accordingly, GC-content or equilibrium GC values were found to be higher in outcrossing than in selfing species [24, 84, 120]. Difference in gBGC between outcrossing and selfing lineages has also been found in the plant genus Collinsia [123] and in freshwater snails [66], although difference in selection on codon usage cannot be completely ruled out.
gBGC can also affect functional sequence evolution, leaving a spurious signature of positive selection and increasing the mutation load through the fixation of weakly deleterious AT!GC mutations: gBGC would represent a genomic Achilles' heel [124]. Once again, comparing outcrossing and selfing species is useful for detecting interference between gBGC and selection. gBGC is expected to counteract selection in outcrossing species only. The Achilles' heel hypothesis could explain why relaxed selection was not detected in four grass species belonging to the Triticeae tribe [84]. In outcrossing species, but not in selfing ones, dN/dS was found to be significantly higher for genes exhibiting high than low equilibrium GC-content, suggesting that selection efficacy could be reduced because of high substitution rates in favor of GC alleles in these outcrossing grasses. In outcrossing species, gBGC can maintain recessive deleterious mutations for a long time at intermediate frequency, in a similar way to overdominance [125]. This could generate high inbreeding depression in outcrossing species, preventing the transition to selfing. In reverse, recurrent selfing would reduce the load through both purging and the avoidance of gBGC, thus reducing the deleterious effects of inbreeding. Under this scenario, gBGC would reinforce disruptive selection on mating systems. In the long term, gBGC could be a new cost of outcrossing: because of gBGC, not drift, outcrossing species could also accumulate weakly deleterious mutations, to an extent which could be substantial given current estimates of gBGC and deleterious mutation parameters [125]. Whether this gBGCinduced load could be higher than the drift load experienced by selfing species remains highly speculative. Both theoretical works, to refine predictions, and empirical data, to quantify the strength of gBGC and its impact on functional genomic regions, are needed in the future. Grasses are clearly an ideal model for investigating these issues, but comparisons with groups having lower levels of gBGC would also be helpful.
2.3.3 Transposable Elements in Selfers and Asexuals: Purging or Accumulation?
Considering the role of sex in the spread of selfish elements, TEs should be less frequent in selfers and asexuals than in outcrossers because they cannot spread from one genomic background to another through syngamy. However, highly selfing and asexual species derive from sexual outcrossing ancestors, from which they inherit their load of TEs. TE distribution eventually depends on the balance between additional transposition within selfing/clonal lineages on one hand and selection or excision on the other. Following the abandonment of sex, large asexual populations are expected to purge their load of TEs, provided excision occurs, even at very low rates. However, purging can take a very long time, and, without excision, TEs should slowly accumulate, not decline [126]. In small populations, even with excision, a Muller's ratchet-like process drives TE accumulation throughout the genome [126]. Transition from outcrossing to selfing should also rapidly purge TEs, but as for asexuals, in small fully selfing populations, TEs can be retained [127]. Using yeast populations, it was experimentally confirmed that sex increases the spread of TEs [128, 129]. TE numbers were also found to be higher in cyclically sexual than in fully asexual populations of Daphnia pulex [130–132] (Table 3), contrary to what was described in the parasitoid wasp Leptopilina clavipes and in root knot nematode species (Table 3). It should be noted that several comparative studies on asexual arthropods, nematodes, primroses, and green algae did not evidence any significant effect of breeding system on TE content or evolution (Table 3). At larger evolutionary scales, the putatively ancient asexual bdelloid rotifers strikingly exemplify the fact that
Table 3
Summary of studies comparing transposable element distribution and dynamics between different breeding systems
(continued)
Table 3 (continued)
asexuals can purge their load of TEs. Unlike all sexual eukaryotes, they appear to be free of vertically transmitted retrotransposon, while their genome contains DNA transposons, probably acquired via horizontal transfers [133, 134]. Examples of TE accumulation in asexuals are less common, maybe because species are doomed to extinction under this evolutionary scenario [135]. However, the increase in genome size in some apomictic lineages of Hypericum species may result from this process [136].
In selfers, the distribution of TEs depends not only on the population size but also on the mode of selection against TEs [127, 137]. Under the "deleterious" model, TE insertions are selected against because they disrupt gene functions. According to the "ectopic exchange" model, TEs are selected against because they generate chromosomal rearrangements through unequal crossing-over between TE at nonhomologous insertion sites. Under the first of these two models, homozygosity resulting from selfing increases the selection efficacy against TEs, while under the second one, under-dominant chromosomal rearrangements are less selected against in selfing than in outcrossing populations [127, 137]. A survey of Ty1-copia-like elements in plants suggests that they are less abundant in self-fertilizing than in outcrossing plants, thus supporting the "deleterious" rather than the "ectopic" exchange model [127]. The distribution of retrotransposons in self-incompatible and self-compatible Solanum species also supports the "deleterious" model, even though most insertions are probably neutral [138] (Table 3). In the selfer Arabidopsis thaliana, selection efficacy against TEs seems to be reduced compared to its outcrossing sister species A. lyrata [139, 140], but comparison of the two complete genomes revealed a higher load of TE in A. lyrata and a recent decrease in TE in number in A. thaliana, in agreement with the date of transition to selfing [141]. In the Capsella genus, while the very recent selfer C. rubella possesses a slightly higher number of TEs than the outcrossing C. grandiflora, the oldest selfer C. orientalis exhibits a significantly reduced load of TE [142] (Table 3). Other selfish elements, such as B chromosomes, are also less frequent in selfers, in support of the view that inbreeding generally prevents selfish element transmission [102].
2.4 Breeding Systems, Ploidy, and Hybridization
Atypical breeding systems are often associated with polyploidy [143], and the reasons for this association are not entirely clear. Polyploid mutants might be more likely to establish as new lineages in selfers and asexuals than in obligate outcrossers if crosses between polyploids and diploids are unfertile or counterselected. This is because at low population frequency a polyploid mutant will experience the disadvantage of mostly mating with diploids—the minority cytotype exclusion principle [144, 145]—unless it reproduces asexually or via selfing. In addition, by doubling gene copy number, polyploidy might alleviate the fitness cost of recessive deleterious mutations being exposed at homozygous state in selfers [146]. Kreiner et al. [147] reported that in Brassicaceae the rate of production of unreduced gametes is higher in asexuals than in outcrossers, suggesting that mating systems can influence not only the establishment rate but also the mutation rate to polyploidy.
Recent genome-wide data analyses have revealed that a number of polyploid selfers or asexuals actually correspond to allopolyploids (e.g., [59, 148–151]), highlighting the possibility that hybridization plays a role in breeding system and ploidy evolution. Hybridization between facultative asexuals might cause immediate transition to obligate asexuality if the two progenitor genomes are so divergent that meiosis is impaired—e.g., due to chromosomal rearrangements, or in case of genetic incompatibilities affecting genes involved in sexual reproduction [16]. Numerous selfing or asexual lineages, either diploid or polyploid, are known to be of hybrid origin (e.g., [13, 152–157]). Hybridization would therefore appear as a potential cause, and polyploidy a potential consequence, of atypical breeding systems [16], but more genome-wide data are obviously needed to draw firm conclusions on these complex relationships.
#### 2.5 Breeding Systems and Genome Size Evolution
As argued above, breeding systems can affect many aspects of genome content and organization. They should also affect the whole genome size. Following Lynch's theory [1], genome size should be higher in selfers and asexuals because of their reduced effective population size, hence reduced ability to get rid of useless, slightly costly sequences. However, the picture is probably more complex. First, because of the recent origin of many selfing and (at least some) asexual lineages, relaxed selection may not have operated longly enough to impact genome size. Second, because of their immunity to selfish element transmission, selfers and asexuals should exhibit lower genome size, especially in groups where TEs are major determinants of genome size. Hence, it is not clear whether genetic drift or resistance to selfish elements (or other processes) is the most important in governing genome size evolution in various breeding systems.
Meta-analyses performed in plants provided equivocal answers. Analysis of the distribution of B chromosomes showed a strong and significant positive association between outcrossing, the occurrence of B chromosomes, and genome size [102, 158]. However, after phylogenetic control, only the association between breeding systems and B chromosomes remains. Whitney et al. [159] simultaneously tested the effect of breeding systems (using outcrossing rate estimates) and genetic drift (using polymorphism data) on genome size in seed plants. Raw data showed a significant effect of both breeding systems and genetic drift, according to theoretical predictions. However, no effect was observed after phylogenetic control, leading the authors to reconsider the hypothesis of a role of nonadaptive processes in genome size evolution. Similarly, phylogenetic comparative analysis of 30 primrose species (Oenothera) covering several transitions to asexuality showed no significant relationship between reproductive mode and genome size [160].
Because breeding systems can evolve quickly, more detailed analyses at a short phylogenetic scale are needed to get a clearer picture of their effects on genome size evolution. Moreover, breeding systems are often correlated with other life history traits, such as lifespan, which can make it hard to clarify the causes and consequences of the observed correlations. A detailed analysis of genome size in the Veronica genus suggests that selfing, not annuality, is associated with genome size reduction [161]. A comparison of 14 pairs of plant congeneric species with contrasted mating systems also suggested a genome size reduction in selfers [162]. However, this could partly have been due to the four polyploid selfing species of the dataset—polyploidy can lead to haploid genome size reduction because of the loss of redundant DNA following polyploidization. A better understanding can be gained from the comparative analysis of genome composition and organization, not only genome size. In Caenorhabditis nematodes, the observed reduction in genome size is not driven by reduction in TEs but by a global loss of all genomic compartments [163]. This pattern contradicts the hypothesis of relaxed selection in selfers against the accumulation of deleterious genomic elements. Alternatively, it could be explained by deletion bias and high genetic drift in selfers. However, in mutation accumulation lines, insertions predominate over deletion in the selfing C. elegans, and deletions occurred at the whole gene level instead of being at random among genomic compartments, as predicted under a general deletion bias (see discussion in ref. 163). In this genus, Lynch's hypothesis that evolution of genome size should be driven by changes in N<sup>e</sup> does not apply. Alternatively, the authors suggested that it is a more direct consequence or even an adaptation to the selfing lifestyle, although the underlying mechanisms still remain unclear.
#### 3 A Genomic View of Breeding System Evolution
Because breeding systems can strongly affect genome structure and evolution, conversely, genomic approaches offer new powerful tools to reconstruct breeding system evolution and to test evolutionary hypotheses, especially concerning long-term evolution.
#### 3.1 Genomic Approaches to Infer Breeding System Evolution
3.1.1 Genomic Characterization of Breeding Systems
Genetic markers have long been used to determine breeding systems and quantify selfing rates or degrees of asexuality. For instance, current selfing rates can be inferred using molecular markers through FIS estimates or preferably—although more time consuming—through progeny analyses [164–166]. Multilocusbased estimates that take identity disequilibrium into account greatly improve the simple FIS-based method that is sensitive to several artifacts such as null alleles ([167], see also refs. 168, 169). This method, implemented in the RMES software [167], has proven to give results very similar to progeny-based methods [170]. To take advantage of the information potentially available in sequence data, coalescence-based estimators have also been proposed to infer long-term selfing rates, and they have been implemented more recently in a Bayesian clustering approach in the INSTRUCT software package [171]. However, this approach mostly captures information from recent coalescence events so that such approaches still estimate recent selfing rates [28]. Much more information about long-term selfing rates can be derived from LD patterns [19], but this has not been fully exploited for selfing rate estimators (for instance, LD is not taken into account in INSTRUCT). Similarly, recombination can be inferred using genetic markers or sequence data, and more generally, various methods have been proposed to characterize the degree of clonality in natural populations (for review see ref. 172) and recently implemented in the R package RClone [173].
Initially, such methods were applied with few markers, from which only global descriptions of breeding systems were deducible. Thanks to the considerable increase in sequencing facilities, it has become possible to finely characterize temporal and spatial variations in breeding systems. In A. thaliana, an analysis of more than 1000 individuals in 77 local stands using more than 400 SNP markers revealed spatial heterogeneity in outcrossing rates. Local "hotspots" of recent outcrossing (up to 15%) were identified, while other stands exhibited complete homozygosity with no detectable outcrossing [174]. Interestingly, at this local scale (from 30 m to 40 km), outcrossing rates have been found to be twofold higher on average in rural than in urban stands; hence, selfing could be associated with higher disturbance in urban stands.
Genomic data may also help characterize breeding systems in species with unknown or ill-characterized life cycles. In yeasts Saccharomyces cerevisiae and S. paradoxus, the analyses of linkage disequilibrium patterns allowed to quantify the frequency of (rare) sexual reproduction events and the proportion of inbreeding and outcrossing during these events [175, 176]. For instance, in the pico-algae Ostreococcus, no sexual form or process has been detected in the lab. However, the occurrence of infrequent recombination (about 1 meiosis for 10 mitoses) inferred from a population genomics approach and the presence of meiosis genes in the genome support the existence of a sexual life cycle [177]. Moreover, a strong negative correlation between chromosome size and GC-content has been observed [178]. In mammals and birds (among others), such a pattern has been interpreted as a longterm effect of gBGC acting on chromosomes with different average recombination rates [116]—small chromosomes having higher recombination rates because of the constraint of at least one chiasmata per chromosome arm. A similar interpretation for Ostreococcus is thus appealing. Genomic data also allow to test whether the theoretical signatures of long-term asexuality are observed in putative asexuals. As an example, whole-genome analyses of the trypanosome T. b. gambiense demonstrated an independent evolution and divergence of alleles on each homologous chromosome (the "Meselson effect" [179, 180]), which is indicative of strict asexual evolution [88]. In contrast, genomic studies of the putatively ancient asexual bdelloids recently uncovered the occurrence of inter-individual genetic exchanges ([181, 182] see below Subheading 3.2.2).
#### 3.1.2 Inferring and Dating Breeding System Transitions
Genomic approaches are also useful for analyzing the dynamics of breeding system evolution. A simple way is to map breeding system evolution on phylogenies, which could provide a raw picture of the frequency and relative timing of breeding system transitions (e.g., [183]). However, these approaches, based on ancestral character reconstruction, are hampered by numerous uncertainties. For instance, in the case of two sister species with contrasting breeding systems, such as A. thaliana and A. lyrata, it is impossible to know whether A. thaliana evolved toward selfing just after divergence (about five million years ago) or only very recently. At a larger phylogenetic scale, inferring rates of transition between characters and ancestral states can be biased if diversification rates differ between characters—this is typically expected with breeding systems for which asexuals and selfers should exhibit higher extinction rates than outcrossers [184].
Thanks to the genomic signatures left by contrasted breeding systems, it is possible to trace back transitions in the past and to date them more precisely. In diploid asexual species, because of the arrest of recombination, the two copies of each gene have diverged independently since the origin of asexuality. After having calibrated the molecular clock, it is thus possible to date this origin from the level of sequence divergence between the two copies. This so-called Meselson effect was observed and quantified in the trypanosome T. b. gambiense, suggesting that this species evolved asexually about 10,000 years ago [88]. However, no Meselson effect has been observed in other presumably ancient asexual species such as oribatid mites [185] or darwinulid ostracods [186], while data refute the possibility of cryptic sex. In such cases, it is thus not possible to infer when recombination actually stopped, presumably because of homogenizing processes such as very efficient DNA repair or automixis. Mitotic recombination could also obscure the pattern predicted under this Meselson effect. Of note, when asexuality originates by hybridization (see above Subheading 2.4), the last common ancestor of the two copies of a gene dates back to the ancestor of the two parental lineages, which can be much older than the hybridization date, faulting the above-described rationale.
Past transitions from outcrossing to selfing have also been investigated, through either population genomics approaches or the evolutionary analysis of self-incompatibility (SI) genes, which are directly involved in the transition to selfing. Since the evolution of selfing requires the breakdown of SI systems, initially constrained S-locus genes are expected to evolve neutrally after a shift to selfing. In A. thaliana, Bechsgaard et al. [187] reasoned that the dN/dS ratio in the selfing lineage should be the average of the neutral dN/dS (i.e., 1) and the outcrossing dN/dS—inferred from sister lineages—weighted by the time spent in the selfing vs. the outcrossing state. They deduced that SRK, one of the major SI genes, became a pseudogene less than 400,000 years ago. SRK, however, is not the only gene involved in SI. Mutations in other genes may have previously disrupted the SI system, thus confusing SRK-based dating. Indeed, coalescence simulations showed that the observed genome-wide pattern of linkage disequilibrium is compatible with the transition to selfing one million years ago [188], suggesting a possible but debated two-step scenario in the evolution of selfing [189, 190]. The persistence of three distinct divergent SRK haplotypes among extant A. thaliana individuals also suggests multiple loss of SI [191], but the recent discovery of the co-occurrence of the three haplotypes in Moroccan populations makes possible the evolution of selfing in a single geographic region [192]. In another Brassicaceae, i.e., Capsella rubella, analyses of both S-locus and genome-wide genes coupled with coalescence simulations suggested that selfing evolved very recently from the outcrosser C. grandiflora, around 50,000 years ago [193, 194] from a potentially large number of founding individuals followed by a strong reduction in N<sup>e</sup> [195]. In the tetraploid selfer Arabidopsis suecica, which originated as a hybrid between A. thaliana and the outcrossing A. arenosa, the genomic analysis of the S-locus also revealed the origin of selfing, suggesting an instantaneous loss of SI due to the fixation of nonfunctional alleles from both parents around 16,000 years ago [150].
#### 3.2 Matching Breeding System Evolution Theories with Genomic Data
3.2.1 Testing the Dead-End Hypothesis: Comparison Between Selfing and Asexuality
The expected reduction in N<sup>e</sup> in selfers and asexuals may increase the drift load (accumulation of slightly deleterious mutations) and preclude adaptation. Selfing and clonality are thus supposed to be evolutionary dead ends [17, 18]. The twiggy phylogenetic distributions of asexuals [196] and selfers [183] or self-compatible species [197] suggest they are mostly derived recently from outcrossing ancestors (but see ref. 198). However, this observation may not be sufficient to support the dead-end hypothesis, and neutral models can also explain this pattern [199–201]. In a comprehensive and epochal phylogenetic study of several Solanaceae genera, Goldberg et al. [202] went further by testing the irreversibility of transitions. Using a phylogenetic method developed for estimating the character effect on speciation and extinction [203, 204], they showed that self-compatible species have both higher speciation and extinction rates—with the resulting net diversification rates being lower—than self-incompatible species. This was the first direct demonstration of the dead-end hypothesis, and additional results have been obtained in Primula species [205]. On the contrary, in the Oenothera genus, asexuality has been found associated with increased diversification but frequent reversion toward the sexual system, suggesting that the form of asexuality in this group is not an evolutionary dead end [206].
Genomic data also provide an opportunity to investigate the genetic causes of such long-term evolutionary failures. The increased dN/dS ratios reported in asexuals (see above) suggest that deleterious point mutations contribute to the load. However, in Daphnia rapid exposure of recessive deleterious alleles through mitotic recombination or gene conversion likely has a much stronger effect on clone persistence than their long-term accumulation under Muller's ratchet [60]. TE could also contribute to the load and to the extinction of asexuals [135], though more data are still needed to unambiguously support this hypothesis (but see ref. 136). The pattern in selfers is less clear. While theory globally predicts a reduction in selection efficacy in selfers, models also highlight conditions under which selection can be little affected or even enhanced in selfers [72, 73, 207], especially regarding TE accumulation [127, 137]. Empirical data on both protein and TE evolution have not revealed any strong evidence of long-term accumulation of deleterious mutation in selfers, as compared to outcrossers, whereas polymorphism data mainly support relaxation of selection in selfers (Table 2). This is in agreement with the recent origin of selfing but makes difficult further inference of the underlying causes of higher extinction in selfers as trait-dependent diversification processes alter the relationship between life history traits and rate of molecular evolution [208]. A reduced ability to respond to environmental changes through adaptive evolution could also contribute to long-term extinction in asexuals (but see ref. 209) and selfers, especially if standing variation is needed to rescue populations experiencing environmental challenges [77, 210]. Few studies, however, have compared the rate of adaptation in selfers and outcrossers (see Table 2). Theoretical predictions regarding this effect, moreover, critically depend on the dominance level of new favorable mutations [72, 73, 77, 210], which are poorly known (but see ref. 80).
While several issues remain open, current knowledge suggests that selfers are less prone to extinction than asexuals. The wider distribution of selfing than clonality in plants supports this view [211, 212]. Selfers could go toward extinction more slowly than asexuals, and the causes of their extinction could differ. Since deleterious mutations should accumulate at a slower rate in selfers than in asexuals, as suggested by theory and current data, this process would likely not be sufficient to drive them to extinction. The reduced adaptive potential could be the very cause of their ultimate extinction as initially proposed by Stebbins [18], which could generally occur before sufficient deleterious mutations have accumulated to be detected via molecular measures of divergence. On the contrary, in asexuals, the accumulation of deleterious mutations could be fast enough to leave a molecular signature and contribute to extinction. Alternatively, demographic characteristics associated with uniparental reproduction, such as recurrent bottlenecks, fragmented populations, and extinction/recolonization dynamics, could be sufficient to drive population extension simply because of higher sensitivity to demographic stochasticity (see also ref. 213). Genomic degradation would only be the witness of the evolution toward selfing and clonality without being the ultimate cause of their extinctions. These hypotheses need to be further investigated by building more realistic demo-genetic model and by better integrating genomic and ecological approaches.
The literature reviewed above focuses on intrinsic factors that may affect the extinction rate of selfing and asexual species, taken as established lineages, compared to their sexual relatives. Alternatively, Janko et al. [199] suggested that if asexual mutants are produced at a relatively high rate and compete with each other, this would imply a rapid turnover between clonal lineages and a young expected age for extant asexuals, without the need to invoke any fitness effect (see also refs. 200, 201). Of note, this model invokes competitive exclusion among clonal lineages, but not between clonal and sexual ones—the ancestral sexual gene pool is assumed to be immune from extinction.
3.2.2 Evading the "Dead End" The few putatively ancient asexuals known so far seem to escape the mutational load predicted by the dead-end hypothesis and avoid extinction over long evolutionary time scales. For example, fossil evidence and decades of microscopic observations indicate that bdelloid rotifers have apparently persisted for over 40 million years without meiosis, males, or conventional sexual reproduction [15, 214]. As a matter of fact, the first genome assembly published for these organisms confirmed that their genome structure is incompatible with conventional meiosis [215]. However, two independent studies recently demonstrated that bdelloids could experience genetic exchanges between individuals.
A first article by Debortoli et al. [182] evidenced frequent horizontal exchanges of genetic fragments between individuals of the species Adineta vaga (Adinetidae). Such horizontal transfers could be promoted by the peculiar ecology of these rotifers, which experience frequent desiccations damaging their cell and nucleus membranes and thus allowing for the entry of foreign DNA in the cells. In addition, desiccation induces multiple DNA double-strand breaks, facilitating the integration of foreign DNA during repair processes.
Another study by Signorovitch et al. [181] identified a pattern of allele sharing between individuals of the species Macrotrachela quadricornifera (Philodinidae) that was incompatible with strict asexual evolution. The authors suggested that bdelloids had evolved an atypical meiotic mechanism similar to what has been described in some species of primroses (Oenothera), in which chromosomes organize into a ring during meiosis without requiring homologous chromosome pairing [216]. They advocated that even rare events of such unconventional sex could be enough to generate the observed pattern of allele sharing.
In the absence of conventional meiosis and syngamy, bdelloid rotifers might thus have escaped extinction by maintaining some level of genetic exchanges between individuals, either through horizontal gene transfers or unconventional Oenothera-like meiosis. Regardless of the underlying molecular mechanisms, bdelloids should not be considered as "ancient asexual scandals" anymore. These recent results call for a reassessment of the reproductive mode of all supposedly ancient asexuals (see Subheading 3.1.1 above). The rise of genomic studies in recent years will greatly contribute to decipher whether putative asexuals evolve as strict asexuals or have developed new alternatives to sex.
#### 4 Conclusion and Prospects
There is a large body of theory on the effects of breeding systems on molecular evolution. However, some of them have not been clearly verified by empirical data, and numerous questions remain. Genomic data have also partly unveiled the complexity of breeding systems, especially in asexual or presumably asexual species. Promising prospects include (1) analysis of the rate and pattern of transition to selfing/asexuality using densely sampled phylogenies with appropriate breeding system distributions combined with genome-wide molecular data, (2) distinguishing between the different forms of selection with a better characterization of the fitness effect of mutations, (3) explicitly accounting for the possible association between breeding system shifts and non-equilibrium demographic dynamics (e.g., bottlenecks in selfers, clone turnover in asexuals). A large theoretical corpus has already been developed, and thanks to the increasing availability of genomic data, qualitative patterns are now rather well described and partly understood. Another challenge in the future is also to make our predictions and tests more quantitative.
#### 5 Questions
#### Acknowledgments
This work was supported by ARCAD, a flagship project of Agropolis Fondation, by an ERC grant (PopPhyl) to N.G. and by the CoGeBi program (grant number ANR-08-GENM-036-01) and a Swiss National Research Found SINERGIA grant.
#### References
among hermaphroditic animals. Evolution 60 (9):1816–1824
Proc R Soc Lond B Biol Sci 273 (1604):3011–3019
selfing evolution in Triticeae (poaceae). Evolution 64(10):2855–2872
duplication and deletion in Daphnia pulex. Genome Res 26(1):60–69
landscapes in seed plants. Plant Cell 24 (4):1379–1397
thaliana genomes. Mol Biol Evol 34 (8):1878–1889
Weissenbach J, Wincker P, Jaillon O, Van Doninck K (2013) Genomic evidence for ameiotic evolution in the bdelloid rotifer Adineta vaga. Nature 500(7463):453–457
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Part IV
#### Natural Selection and Innovation in Genomic Sequences
## Chapter 12
## Selection Acting on Genomes
#### Carolin Kosiol and Maria Anisimova
#### Abstract
Populations evolve as mutations arise in individual organisms and, through hereditary transmission, may become "fixed" (shared by all individuals) in the population. Most mutations are lethal or have negative fitness consequences for the organism. Others have essentially no effect on organismal fitness and can become fixed through the neutral stochastic process known as random drift. However, mutations may also produce a selective advantage that boosts their chances of reaching fixation. Regions of genomes where new mutations are beneficial, rather than neutral or deleterious, tend to evolve more rapidly due to positive selection. Genes involved in immunity and defense are a well-known example; rapid evolution in these genes presumably occurs because new mutations help organisms to prevail in evolutionary "arms races" with pathogens. In recent years genome-wide scans for selection have enlarged our understanding of the genome evolution of various species. In this chapter, we will focus on methods to detect selection on the genome. In particular, we will discuss probabilistic models and how they have changed with the advent of new genome-wide data now available.
Key words Conserved and accelerated regions, Positive selection scans, Codon models, Selectionmutation models, Polymorphism-aware phylogenetic models
#### 1 Introduction
In the past selection studies mainly focused on the analysis of particular loci such as genes, proteins, or regular elements of interest. With the availability of comparative genomic data, the emphasis has shifted from the study of individual proteins to genome-wide scans for selection.
The search for selection can be performed on different levels comparing homologous nucleotide sequences or protein-coding genes in one or multiple genomes. The evolutionary processes in all these levels can be described by probabilistic models, which set the basis for evaluating selective pressures and selection tests. This book chapter will give an introduction into fundamental properties of the probabilistic models used to detect selection in the Subheading 3 as well as examples of genome-wide scans.
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_12, © The Author(s) 2019
Fig. 1 A diagram illustrating the different data and levels to analyze genomic sequences and the relationship of the various approaches modeling selection
In Fig. 1, we summarize the different data levels and time scales of modeling selection on genomes.
#### 2 Comparative Genome Data
Several whole genome sequence data sets are now available for selection scans. Mammalian genomes are well represented (in particular primates), and insect genomes are becoming more numerous (in particular Drosophila). These data can be downloaded as orthologous alignments from the Ensembl [1] and UCSC [2] browsers.
In light of recent advances in DNA sequencing, with so-called next generation sequencing (NGS) technologies that have dramatically reduced the cost and time needed to sequence an organism's entire genome, large-scale (involving many organisms) sequencing projects have been and are currently being undertaken. Just to name a few, genome projects re-sequencing 1000 D. melanogaster [3] and 1001 Arabidopsis [4] were accomplished, and the 100,000 human genome project [5] is ongoing. These polymorphism data from multiple individuals from several species enable us to detect very recent selection.
Together with the progress in sequencing technologies, algorithmic advances now allow the de novo assembly of genomes from NGS data, including complex mammalian genomes (e.g., giant panda genome [6]). Therefore, not only international consortia but also small groups and individual labs can now envisage to sequence the organisms of their interest. As a consequence platforms for sharing this data have been established. For example, the Genome 10K project aims to assemble a genomic zoo—a collection of DNA sequences representing the genomes of 10,000 vertebrate species, approximately one for every vertebrate genus. All these genomes can be subject to scans for selection, for which we outline methods below.
#### 3 Methods
#### 3.1 Probabilistic Models for Genome Evolution
The statistical modeling of the evolutionary process is of great importance when performing selection studies. When comparing reasonably divergent sequences, counting the raw sequence identity (percentage of sites with observed changes) underestimates the amount of evolution that has occurred because, by chance alone, some sites will have incurred multiple substitutions. In this chapter we discuss maximum likelihood (ML) and Bayesian methods to detect selection based on probabilistic models of character evolution. Such substitution models provide more accurate evolutionary distance estimates by accounting for these unobserved changes and often explicitly model the selection pressures.
One of the primary assumptions made in defining probabilistic substitution models is that future evolution is only dependent on its current state and not on previous (ancestral) states. Statistical processes with this lack of memory are called Markov processes. The assumption itself is reasonable, because during the evolution mutation and natural selection can only act upon the molecules present in an organism and have no knowledge of what came previously. However, some large-scale mutational events, such as recombination [7], gene conversion (e.g., see [8, 9]), or horizontal transfer [10] might not satisfy this "memoryless" condition.
To reduce the complexity of evolutionary models, it is often further assumed that each site in a sequence evolves independently from all other sites. There is evidence that the independence of sites assumption is violated. In real proteins, chemical interactions between neighboring sites or the protein structure affects how other sites in the sequence change. Steps have been made toward context-dependent models, where the specific characters at neighboring sites affect the sites evolution (e.g., see [11, 12]).
The Markov model asserts that one sequence is derived from another by a series of independent substitutions, each changing one character in the first sequence to another character in the second during the evolution. Thereby we assume independence of evolution at different sites. A continuous-time Markov process is fully defined by its instantaneous rate matrix Q ¼ {qij}i,j¼<sup>1</sup> ... <sup>N</sup>.
The diagonal elements of Q are defined by a mathematical requirement that the rows sum up to zero. For multiple sequence alignments, the substitution process runs in continuous time over a tree representing phylogenetic relations between the sequences. The transition probability matrix P(t) ¼ {pij(t)} ¼ e Qt consists of transition probabilities from residue i to residue j over time t and is found as a solution of the differential equation dP(t)/dt ¼ P(t)Q with P(0) being the identity matrix. In order for tree branches to be measured by the expected number of substitutions per site, the Qmatrix is scaled so that the average substitution rate at equilibrium equals 1.
As a matter of mathematical and computational convenience rather than biological reality, several simplifying assumptions are usually made. Standard substitution models allow any state to change into any other. Such Markov process is called irreducible and has a unique stationary distribution corresponding to the equilibrium codon frequencies π ¼ {πi}. Time reversibility implies that the direction of the change between two states i and j is indistinguishable, so that πipij(t) ¼ πjpji(t). This assumption helps to reduce the number of model parameters and is convenient when calculating the matrix exponential (Q-matrix of a reversible process has only real eigenvectors and eigenvalues [13]). Fully unrestrained Q-matrix for N characters defines an irreversible model with N (N - 1) - 1 free parameters, while for a reversible process this number is N(N + 1)/2 -2.
By comparing how well substitution models explain sequence evolution, and by examining the parameters estimated from data, ML and Bayesian inference can be used to address many biologically important questions. In this section we focus on probabilistic models that are used to detect selection.
3.2 Detecting Regions of Accelerated Genome Evolution Understanding the forces shaping the evolution of specific lineages is one of the most exciting areas in evolutionary genomics. In particular, regions of accelerated evolution in mammalian and insect species have been studied (e.g., see [14]). To eliminate nonfunctional regions, one strategy is to begin with a search for regions that are conserved through the mammalian history or longer. A likelihood ratio test (LRT) may be used to detect acceleration of rates in a lineage of interest, for example, the human lineage. Such LRT compares the likelihood of the alignment data under two probabilistic models. The null model has a single scale parameter representing shortening (more conserved) and lengthening (less conserved) of all branches of the tree. The alternative model has an additional parameter for the human lineage, which is constraint to be 1. This extra parameter allows the human branch to be relatively longer (accelerated) than the branches in the rest of the tree.
> For example, this approach was used to identify genomic regions that are conserved in most vertebrates but have evolved rapidly in humans. Interestingly, the majority of the human accelerated regions (HARs) were noncoding, and many were located near protein-coding genes with protein functions related to the nervous system [14].
> In contrast, the majority of Drosophila melanogaster accelerated regions (DMARs) are found in protein-coding regions and
primarily result from rapid adaptive change at synonymous sites [15]. This could be because flies have much more compact genomes compared to humans; however, even after considering the genomic content, in Drosophila a significant excess of DMARs occur in protein-coding regions. Furthermore, Holloway and colleagues observed a mutational bias from G|C to A|T, and therefore the accelerated divergence in DMARs might be attributed to a shift in codon usage and a fixation of many suboptimal codons.
In a similar manner, amino acid based models search for site- or lineage-specific rate accelerations and residues subject to altered functional constraints. Such sites are likely to be contributing to the change in protein function over time. The advantage of amino acid-based models is that they might be suitable for the analysis of deep divergences of fast-evolving genes, where sequences rapidly saturate over time. Also amino acid methods are not influenced by the effects of codon bias, a topic that is discussed at the end of this chapter. The idea is that adaptive change on the level of amino acid sequences may not necessarily correspond to an adaptive change in protein function but rather to peaks in the protein adaptive landscape reflecting the optimization of the protein function in a particular species to long-term environmental changes. One class of methods for detecting functional divergence searches for a lineagespecific change in the shape parameter of the gamma distribution that is used to model rate heterogeneity (see [16–19]). Other methods search for evidence of clade-specific rate shifts at individual sites (see [20–26]). For example, Gu [21] proposed a simple stochastic model for estimating the degree of divergence between two pre-specified clusters. The statistical significance was tested using site-specific profiles based on a hidden Markov model, which was used to identify amino acids responsible for these functional differences between two gene clusters. More flexible evolutionary models were incorporated in the maximum likelihood approach applicable to the simultaneous analysis of several gene clusters [27]. This was extended [28] to evaluate site-specific shifts in amino acid properties, in comparison with site-specific rate shifts. Pupko and Galtier [24] used the LRT to compare ML estimates of the replacement rate at an amino acid site in distinct subtrees.
3.3 Codon Models: Site, Branch, and Branch-Site Specificity
3.3.1 Basic Codon Models
In protein-coding sequences, nucleotide sites at different codon positions usually evolve with highly heterogeneous patterns (e.g., [29]). Thus DNA substitution models fail to account for this heterogeneity unless the sequences are partitioned by codon positions for the analysis. But even then, DNA models do not model the structure of genetic code or selection at the protein level. Indeed, one advantage of studying protein-coding sequences at the codon level is the ability to distinguish between nonsynonymous (AA replacing) and synonymous (silent) codon changes. Based on this distinction, the selective pressure on the proteincoding level can be measured by the ratio ω ¼ dN/d<sup>S</sup> of the nonsynonymous to synonymous substitution rates. The nonsynonymous substitution rate may be higher than the synonymous rate, and thus ω > 1 due to fitness advantages associated with recurrent AA changes in the protein, i.e., positive selection on the protein. In contrast, purifying selection acts to preserve the protein sequence, so that the nonsynonymous substitution rate is lower than the synonymous rate, with ω < 1. Neutrally evolving sequences exhibit similar nonsynonymous and synonymous rates, with ω 1.
First methods that used the ω ratio as a criterion to detect positive selection were based on pairwise estimation of d<sup>N</sup> and d<sup>S</sup> rates with "counting" methods (e.g., see [30]). However, ML estimates of pairwise d<sup>N</sup> and d<sup>S</sup> based on a codon model were shown to outperform all other approaches [31]. Moreover, a Markov codon model is naturally extended to multiple sequence alignments, unlike the counting methods. This, together with the benefits of the probabilistic framework within which codon models are defined, made codon models very popular in studies of positive selection in protein-coding genes.
The first two codon models were proposed simultaneously in the same issue of Molecular Biology and Evolution [32, 33]. The model of Goldman and Yang [32] included the transition/transversion rate ratio κ, and modeled the selective effect indirectly using a multiplicative factor based on Grantham [34] distances, but was later simplified to estimate the selective pressure explicitly using the ω parameter [35]. The main distinction between the first codon models concerns the way to describe the instantaneous rates with respect to equilibrium frequencies: (1) proportional to the equilibrium frequency of a target codon (as in Goldman and Yang [32]) or (2) proportional to the frequency of a target nucleotide (as in Muse and Gaut [33]).
In 2006, empirical codon models have been estimated (see [36, 37]) that summarize substitution patterns from large quantities of protein-coding gene families. In contrast to the parametric codon models that estimate gene-specific parameters (e.g., transition-transversion κ, selective pressure ω, etc.), the empirical codon models do not explicitly consider distinct factors that shape protein evolution. Standard parametric models assume that protein evolution proceeds only by successive single-nucleotide substitutions. However, empirical codon models indicate that model accuracy is significantly improved by incorporating instantaneous doublet and triplet changes. Kosiol et al. [37] also found that the affiliations between codon, the amino acid it encodes, and the physicochemical properties of the amino acid are main driving factors of the process of codon evolution. Neither multiple nucleotide changes nor the strong influence of the genetic code nor amino acid properties form a part of the standard parametric models.
On the other hand, parametric models have been very successful in applications studying biological forces shaping protein evolution of individual genes. Thus combining the advantages of parametric and empirical approaches offers a promising direction. Kosiol, Holmes, and Goldman [37] explored a number of combined codon models that incorporated empirical AA exchangeabilities from ECM while using parameters to study selective pressure, transition/transversion biases, and codon frequencies. Similarly, AA exchangeabilities from (suitable) empirical AA matrices may be used to alter probabilities of nonsynonymous changes, together with traditional parameters ω, κ, and codon frequencies π<sup>j</sup> [38]. In 2013, De Maio et al. [39] extended the ECM approach to accommodate site-specific variation of selective pressure and lineagespecific variation. Simulations showed that ECMs allowing for double and triple mutations is more conservative: they reduce the number of false positives and have less power to detect positive selection [39].
3.3.2 Accounting for Variability of Selective Pressures First codon models assumed constant nonsynonymous and synonymous rates among sites and over time. Although most proteins evolve under purifying selection most of the time, positive selection may drive the evolution in some lineages. During episodes of adaptive evolution, only a small fraction of sites in the protein have the capacity to increase the fitness of the protein via AA replacements. Thus approaches assuming constant selective pressure over time and over sites lack power in detecting genes affected by positive selection. Consequently, various scenarios of variation in selective pressure were incorporated in codon models, making them more powerful at detecting positive selection, and short episodes of adaptive evolution in particular. Evidence of positive selection on a gene can be obtained by a LRT comparing two nested models: a model that does not allow positive selection (constraining ω 1 to represent the null hypothesis) and a model that allows positive selection (ω > 1 is allowed in the alternative hypothesis). Positive selection is detected if a model ω > 1 fits data significantly better compared to the model restricting ω 1 at all sites and lineages. However, the asymptotic null distribution may vary from the standard χ <sup>2</sup> due to boundary problems or if some parameters become not estimable (e.g., see [40, 41]).
3.3.3 Case Study: Application of a Genome-Wide Scan of Positive Selection on Six Mammalian Genomes
In 2006, six high-coverage genome assemblies became available for eutherian mammals. The increased phylogenetic depth of this data set permitted Kosiol and colleagues [42] to perform several new lineage- and clade-specific tests using branch-site codon models. Of ~16,500 human genes with high-confidence orthologs in at least two other species, 544 genes showed significant evidence of positive selection using branch-site codon models and standard LRTs.
Interestingly, several pathways were found to be strongly enriched in genes with positive selection, suggesting possible coevolution of interacting genes. A striking example is the complement immunity system, a biochemical cascade responsible for the elimination of pathogens. This system consists of several small proteins found in the blood that cooperate to kill target cells by disrupting their plasma membranes. Of 78 genes associated with this pathway in KEGG (see http://www.genome.jp/kegg-bin/ show\_pathway?map04610 for the complement cascades), nine were under positive selection (FDR < 0.05), and five others had nominal P < 0.05. Most of genes under positive selection are inhibitors (DAF, CFH, CFI) and receptors (C5AR1, CR2), but some are part of the membrane attack complex (C7, C9, C8B), which punctures cell membranes to initiate cell lysis. Here we focus on the analysis of these proteins of the membrane attack complex.
First we calculate gene averaged ω value using the basic M0 model [32]. The ML estimates of ω < 1 (ω ¼ 0.31 for C7, ω ¼ 0.25 for C8B, and ω ¼ 0.44 for C9) indicate that most sites in these genes are under purifying selection. However, selection pressure could be variable at different locations of the membrane proteins, and we therefore continue our analysis by applying models that allow for variation in selective pressure across sites.
3.3.4 Selective Variability Among Codons: Site Models The simplest site models use the general discrete distribution with a pre-specified number of site classes. Each site class i has an independent parameter ω<sup>i</sup> estimated by ML together with proportions of sites pi in each class. Since a large number of site categories require many parameters, three categories are usually used (requiring five independent parameters). To test for positive selection, several pairs of nested site models were defined to represent the null and alternative hypotheses in LRTs. For example, model M1a includes two site classes, one with ω<sup>0</sup> < 1 and another with ω<sup>1</sup> ¼ 1, representing the neutral model of evolution (the null hypothesis). The alternative model M2a extends M1a by adding an extra site class with ω<sup>2</sup> 1 to accommodate sites evolving under positive selection. Significance of the LRT is tested using the χ<sup>2</sup> 2-distribution for the M1 vs. M2 comparison. We test the C7 gene for positive selection by the LRT comparing nested models M1a and M2a (Table 1).
> Model M2a has two additional parameters compared to model M1a. The resulting LRT statistic is 2(log L2 log L1) ¼ 2 (-6377.35 - (-6369.67)) ¼ 2 7.68 ¼ 15.36. This is much greater than the critical value of the chi-square distribution χ <sup>2</sup> (d<sup>f</sup> <sup>¼</sup> 2, at 5%) <sup>¼</sup> 5.99, and we calculate a <sup>p</sup>-value of P ¼ 5.0e-04. However, the M1a vs. M2a comparison for genes C8B and C9 is not significant.
#### Table 1 Parameter estimates and log-likelihoods for a LRT of positive selection for the complement immunity component C7
The model M2a is the alternative model with a class of sites with ω<sup>2</sup> 1. The null hypothesis M1a is the same model but with ω<sup>2</sup> ¼ 1 fixed
> Another LRT can be performed on the basis of the modified model M8 with two site classes: one with sites where the ω ratio is drawn from the beta distribution (with 0 ω 1 describing the neutral scenario) and the second, discrete class, with ω 1. Constraining ω ¼ 1 for this second class provides a sufficiently flexible null hypothesis, whereby all evolution can be explained by sites with ω from the beta distribution or from a discrete site class with ω ¼ 1. Significance of the LRT is tested the mixture <sup>1</sup> 2 χ2 0 þ 1 2 χ2 <sup>1</sup> for the M8 (ω ¼ 1) vs. M8 comparison. If the LRT for positive selection is found to be significant, specific sites under positive selection may be predicted based on the values of posterior probabilities (PP) to belong to the site class under positive selection (usually PP > 0.95, but see [43, 44]). Such posterior probabilities are estimated using the naı¨ve empirical Bayesian approach (NEB, [45]), full hierarchical Bayesian approach ([46]; BEB [44]), or a mid-way approach the Bayes empirical Bayes (BEB [44]). For a discussion on these approaches, see Scheffler and Seoighe [47] and Aris-Brosou [48]. Alternatively, Massingham and Goldman [49] proposed a site-wise likelihood ratio estimation to detect sites under purifying or positive selection.
> For the C7 gene, using BEB we identified several amino acids sites to be putatively under selection: residue R at position 223 (PP ¼ 0.94), H at position 239 (PP ¼ 0.93), and N at position 331 (PP ¼ 0.93). Unfortunately, the crystal structures of C7 (as well as C8B and C9) are not known, and we cannot relate the location of amino acids in the protein sequence to relevant 3D data, such as sites of protein-protein interaction or binding sites of the
protein. If such structural information were known, it would also be possible to use this biological knowledge in a model that is aware of the position of the different structural elements.
Site models that do not use a priori partitioning of codons (as those described above) are known as random-effect (RE) models. In contrast, fixed-effect (FE) models categorize sites based on a prior knowledge, e.g., according to tertiary structure for single genes, or by gene category for multigene data. Site partitions for FE models can be defined also based on inferred recombination breakpoints, useful for inferences of positive selection from recombining sequences (see [50, 51]); although the uncertainty of breakpoint inference is ignored in this way. FE models with each site being a partition should be avoided, as they lead to the "infinitely many parameter trap" (e.g., see [52]). Given a biologically meaningful a priori partitioning, FE models are useful to study heterogeneity among partitions. However, a priori information is not always available.
3.3.5 Selective Variability over Time: Branch Models A simple way to include the variation of the selective pressure over time is by using separate parameters ω for each branch of a phylogeny (known as free-ratio model; [35]). Compared with the oneratio model (which assumes constant selection over time), the freeratio model requires additional 2T - 4 ω parameters for T species. Figure 2 shows the estimates of the free-ratio model for the C8B gene. Although the ML estimates of ω values on the rodent lineages are visibly higher than on the primate lineages, none of the branches has ω > 1.
> Other branch models can be defined by constraining different sets of branches of a tree to have an individual ω. LRTs are used to decide (1) whether selective pressure is significantly different on a pre-specified set of branches and (2) whether these branches are under positive selection.
> However, branch models have relatively poor power to detect selection [53] in comparison to branch-site models that are discussed in the next section. Also note that testing of multiple hypotheses on the same data requires a correction, so the overall false-positive rate is kept at the required level (most often 5%). Correction for multiple testing further reduces the power of the method, especially when many hypotheses are tested simultaneously (see Subheading 4 later).
3.3.6 Temporal and Spatial Variation of Selective Pressure Several solutions were proposed to simultaneously account for differences in selective constraints among codons and the episodic nature of molecular evolution at individual sites. One of the first models—model MA [45]—assumes four site classes. Two classes contain sites evolving constantly over time: one under purifying selection with ω<sup>0</sup> < 1; another with ω<sup>1</sup> ¼ 1. The other two site
Fig. 2 An estimate of ω for each branch of a six-species phylogeny. Shown is the maximum likelihood estimate for the gene 8B. Each branch is labeled with the corresponding estimate of ω
classes allow selective pressure at a site to change over time on a pre-specified set of branches, known as the foreground. The two variable classes are derived from the constant classes so that sites typically evolving with ω<sup>0</sup> < 1 or ω<sup>1</sup> ¼ 1 are allowed to be under positive selection with ω<sup>2</sup> 1 on the foreground. Testing for positive selection on the rodent clade involves a LRT comparing a constrained version of MA (with ω<sup>2</sup> ¼ 1) vs. an unconstrained MA model. Compared to branch models, the branch-site formulation improves the chance of detecting short spills of adaptive pressure in the past even if these occurred at a small fraction of sites.
Returning to our example of gene C8B of the complement pathway, we perform a branch-site LRT for positive selection using the M1a vs. M2a comparison. Thereby we take mouse and the rat lineage, respectively, as foreground branches and all other branches as background branches. Significance of the LRT is tested the mixture <sup>1</sup> 2 χ2 0 þ 1 2 χ2 <sup>1</sup> with critical values to be 2.71 at 5%. For the C8B gene, we calculate 2(log L2 log L1) ¼ 2 2.23 ¼ 4.46 for the mouse lineage and 11.2 for the rate lineage, respectively.
A major drawback of described branch-site models is their reliance on a biologically viable a priori hypothesis. In context of detecting sites and lineages affected by positive selection, one possible solution is to perform multiple branch-site LRTs, each setting a different branch at the foreground [54]. In the example of six species (Fig. 2), a total of nine tests (for an unrooted tree) are necessary in the absence of an a priori hypothesis. Multiple test correction has to be applied to control excessive false inferences. This strategy tends to be conservative but can be sufficiently powerful in detecting episodic instances of adaptation. As with all model-based techniques, precautions are necessary for data with unusual heterogeneity patterns, which may cause deviations from the asymptotic null distribution and thus result in an elevated falsepositive rate.
In the case of episodic selection where any combination of branches of a phylogeny can be affected, a Bayesian approach in lieu of the standard LRTs and multiple testing have been suggested. The multiple LRT approach is most concerned with controlling the false-positive rate of selection inference and is less suited to infer the best-fitting selection history. In the hypothetical example (Fig. 2), a total of 2<sup>9</sup> - 1 ¼ 511 selection histories (excluding the history without selection on any branch) need to be considered. The Bayesian analysis allows a probability distribution over possible selection histories to be computed and therefore permits estimates of prevalence of positive selection on individual branches and clades. Such approach evaluates uncertainty in selection histories using their posterior probabilities and allows robust inference of interesting parameters such as the switching probabilities for gains and losses of positive selection [42].
Other models (e.g., with d<sup>S</sup> variation among sites [55]) may be extended to allow changes of selective regimes on different branches. This is achieved by adding further parameters, one per branch, describing the deviation of selective pressure on a branch from the average level on the whole tree under the site model. Such model is parameter-rich and can be used for exploratory purposes on data with long sequences but does not provide a robust way of testing whether ω > 1 on a branch is due to positive selection on a lineage or due to inaccuracy of the ML estimation.
Kosakovsky Pond and Frost [55] suggested detecting lineagespecific variation in selective pressure using the genetic algorithm (GA)—a computational analogue of evolution by natural selection. The GA approach was successfully applied to phylogenetic reconstruction. In the context of detecting lineage-specific positive selection, GA does not require an a priori hypothesis. Instead the algorithm samples regions of the whole hypotheses space according to their "fitness" measured by AICC. The branch-model selection with GA may also be adapted to incorporate d<sup>N</sup> and d<sup>S</sup> among site variation, although this imposes a much heavier computational burden.
In branch and branch-site models, change in selection regime is always associated with nodes of a tree, but the selective pressure remains constant over the length of each branch. Guindon et al. [56] proposed a Markov-modulated model where switches of selection regimes may occur at any site and any time on the phylogeny. In a covarion-like manner, this codon model combines two Markov processes: one governs the codon substitution, while the other specifies rates of switches between selective regimes. These models can be used to study the patterns of the changes in selective pressures over time and across sites, by estimating the relative rates of changes between different selective regimes (purifying, neutral, and positive).
Polymorphism-aware phylogenetic models (POMOs, [57, 58]) use polymorphism and divergence data simultaneously to estimate relative mutation rates and scaled selection coefficients. Similar to DNA substitution models, the PoMo approach is based on a continuous-time Markov process to model evolution of hereditary sequences along a species tree. However, not only evolution of a single reference site but rather evolution of a population is considered.
3.3.7 Polymorphism-Aware Phylogenetic
Models
PoMo includes polymorphisms as states of the Markov chain, in addition to the four nucleotide states of classical nucleotide models. Sequence evolution is modeled as a gradual process made by small allele frequency changes. PoMo accounts for ancestral polymorphisms and in particular for ancestral shared polymorphisms and incomplete lineage sorting (when two speciation events are separated by a lapse of time not sufficient for polymorphisms to reach fixation, see Maddison and Knowles [59]). The parameters in PoMo do not merely describe substitution rate but are also informative of mutation rates, fixation biases, root nucleotide frequencies, and branch lengths. All these parameters are estimated within a ML framework. De Maio et al. [57] performed a comprehensive study of evolutionary patterns of fourfold-degenerate sites in great apes populations. They show evidence in favor of variation in mutation and fixation rates between genomic regions with different base composition, contributing to the long-standing debate regarding the origin and maintenance of GC content variation (e.g., see Eyre-Walker and Hurst [60]). They found that both mutation rates and biased gene conversion vary with GC content. They also found lineage-specific differences, with weaker fixation biases in orangutan species, suggesting a reduced historical effective population size. As PoMo can distinguish between the contributions of mutation and fixation biases, it might also contribute to addressing the problem of disentangling signatures of selection and biased gene conversion (see Subheading 4.2).
3.4 Software The software PHAST (PHylogenetic Analysis with Space/Time models) includes several phylo-HMM-based programs. Two programs in PHAST are particularly interesting in the context of selection studies: PhastCons is a program for conservation scoring and identification of conserved elements (Siepel et al. [61]). PhyloP is designed to compute p-values for conservation or acceleration, either lineage-specific or across all branches (Pollard et al. [62]). Recently, the software can also be run through a webportal at http://compgen.cshl.edu/phastweb/.
A variety of codon models to detect selection, including branch-site models and the recent selection-mutation model, are implemented in the CODEML program of PAML [63]. HYPHY is another implementation that includes a large variety of codon models [64]. PoMo has been implemented as part of the IQ-TREE software package (http://www.iqtree.org/) by Schrempf et al. [65].
These programs are primarily developed for maximum likelihood inference on a fixed tree. ML inference of phylogeny under codon models is possible with CodonPhyML, which allows to explicitly account for selection on the protein level [66].
#### 4 Notes/Discussion
With the wider use of codon models to detect selection, some questioned the statistical basis of testing based on branch-site models. In 2004, Zhang found that the original branch-site test [67] produced excessive false positives when its assumptions were not met. The modified branch-site test was shown to be more robust to model violations (see [43, 68]) and is now commonly used in genome-wide selection scans (e.g., see [69]). Recently, however, another simulation study by Nozawa et al. [70] suggested that this modification also showed an excess of false positives. Yang and Dos Reis [52] defended the branch-site test by examining the null distribution and showing that Nozawa and colleagues [70] misinterpreted their simulation results. However, it is clear that even tests with good statistical properties will be affected by data quality and the extent of models violations. Below we list factors that can affect the test and so should be taken into account when analyzing genome-wide data.
#### 4.1 Quality of Multiple Alignments
The impact of the quality of sequence and the alignment is a major concern when performing positive selection scans. For example, in their analysis of 12 genomes Markova-Raina and Petrov [71] found that the results were highly sensitive to the choice of an alignment method. Furthermore, visual analysis indicated that most sites inferred as positively selected are in fact misaligned at the codon level. The rate of false positives ranged ~50% and more depending on the aligner used. Some of these results can be ascribed to the high divergence level of the 12 Drosophila species and could be addressed by better filtering of the data. Nevertheless, even in mammals where alignment is easier, problems have been observed.
Bakewell et al. [72] used the branch-site test to analyze ~14,000 genes from the human, chimpanzee, and macaque and detected more genes to be under positive selection on the chimpanzee lineage than on the human lineage (233 vs. 154). The same pattern was also observed by Arbiza et al. [73] and Gibbs et al. [74]. Mallick et al. [75] re-examined 59 genes detected to be under positive selection on the chimpanzee lineage by Bakewell et al. [72], using more stringent filters to remove less reliable nucleotides and using synteny information to remove misassembled and misaligned regions. They found that with improved data quality, the signal of positive selection disappeared in most of the cases when the branch-site test was applied. It now appears that, as suggested by Mallick et al. [75], the earlier discovery of more frequent positive selection on the chimpanzee lineage than on the human lineage is an artifact of the poorer quality of the chimpanzee genomic sequence. This interpretation is also consistent with a few recent studies analyzing both real and simulated data, which suggest that sequence and alignment errors may cause excessive false positives (see [76, 77]). Indeed, most commonly used alignment programs tend to place nonhomologous codons or amino acids into the same column (see [78, 79]), generating the wrong impression that multiple nonsynonymous substitutions occurred at the same site and misleading the codon models into detecting positive selection [77]. In 2012, Jordan and Goldman [80] investigated the effect of various multiple alignment and filtering programs on the identification of positive selection. They found that alignment software PRANK [79] and the filter Guidance [81] performed best in simulations. However, it remains very challenging to develop a pipeline to detect positive selection that is robust to errors in the sequences or alignments. Instead we advise to carefully check the alignments of genes that are putatively under selection by any method described here.
4.2 Biased Gene Conversion and Recombination
Mutation rate variation can also cause genomic regions to have different substitution rates without any change in fixation rate. Recent studies of guanine and cytosine (GC)-isochores in the mammalian genome have suggested the importance of another selectively neutral evolutionary process that affects nucleotide evolution. As described in the work of Laurent Duret and others (see [82, 83]), biased gene conversion (BGC) is a mechanism caused by the mutagenic effects of recombination combined with the preference in recombination-associated DNA repair toward strong (GC) versus weak (adenine and thymine [AT]) nucleotide pairs at non-Watson-Crick heterozygous sites in heteroduplex DNA during crossover in meiosis. Thus, beginning with random mutations, BGC results in an increased probability of fixation of G and C alleles. In particular, methods looking for accelerated regions in coding DNA but also codon models cannot distinguish positive selection from BGC (see [84, 85]). Therefore, the putatively selected genes should be checked for GC content and closeness to recombination hotspots and telomeres.
Most codon models assume a single phylogeny and a constant synonymous rate among sites, implying that rate variation among codons is solely due to the variation of the nonsynonymous rate. Recent studies question whether such assumptions are generally realistic (e.g., see [86]) suggesting that failure to account for synonymous rate variation may be one of the reasons why LRTs for positive selection are vulnerable on data with high recombination rates. Some selection scans try to control this problem by checking putatively selected genes for recombination either manually or automated with traditional detection software (e.g., RDP [87]). Also Drummond and Suchard [88] have recently developed a Bayesian approach to detect recombination within a gene.
Another approach is to explicitly consider recombination. For example, Scheffler, Martin, and Seoighe [89] extended codon models with both d<sup>N</sup> and d<sup>S</sup> site variation and allowed changes of topology at the detected recombination breakpoints. Certainly, fast-evolving pathogens (such as viruses) undergo frequent recombination which often changes either the whole shape of the underlying tree, or only the apparent branch lengths. While the efficiency of the approach depends on the success of inferring recombination breakpoints, the study demonstrated that taking into account alternative topologies achieves a substantial decrease of false-positive inferences of selection while maintaining reasonable power. In principle the correlation structure of a collection of orthologous sequences can be fully described by a network known as an ancestral recombination graph (ARG). However, methods for ARG inferences have not been fast enough for practical use, and for applications on large-scale genomic data, approximations are necessary (Rassmussen et al. [90]).
4.3 Selection on Synonymous Sites Most selection studies to date focused on detecting selection on the protein, since synonymous changes are often presumed neutral and so unaffected by selective pressures. However, selection on synonymous sites has been documented more than a decade ago. Codon usage bias is known to affect the majority of genes and species. In his seminal work, Akashi [91] demonstrated purifying selection on genes of Drosophila melanogaster, where strong codon bias favoring certain (optimal) codons serves to increase the translational accuracy. Pressure to optimize for translational efficiency, robustness, and kinetics leads to synonymous codon bias, which was shown to widely affect mammalian genes [92], as well as genes of fastevolving pathogens like viruses [93]. The standard approach to study selection on codon usage computes various codon adaptation
indexes on full-length protein-coding genes (see [94] for review). More recently, methods to study selection on synonymous changes adopted more sophisticated approaches, mainly the following strategies: (1) account for synonymous rate variation within sequences; (2) include codon fitness parameters within a modeling framework that connects population and intraspecific parameters; and (3) allow for selection on synonymous substitutions by introducing the dependency on the rate of protein production and nonsense error rates. Below we elaborate on these approaches.
In the past decade, evidence has accumulated to suggest that codon bias may vary not only between genomes and genes of the same genome but also within genes. Rather than just measuring codon biases in single sequences, a more powerful approach is to model evolution and selection across a set of homologous sequences. Taking the evolutionary perspective into account, Resch et al. [95] conducted a large-scale study of selection on synonymous sites in mammalian genes. They measured selection by comparing the average rate of synonymous substitutions (dS) to the average substitution rate in the corresponding introns (dI). While purifying selection was found to affect 28% of genes (dS/ d<sup>I</sup> < 1), 12% of genes were found to have been affected by positive selection on synonymous sites (dS/d<sup>I</sup> > 1). The signal of positive selection correlated with lower predicted mRNA stability compared to genes with negative selection on synonymous sites, suggesting that mRNA destabilization (affecting mRNA levels and translation) could be driving positive selection on synonymous sites.
An increasing number of experimental studies exemplify different scenarios explaining how synonymous mutation may be affected by positive or negative selection. Codon bias to match skews of tRNA abundances may influence translation [96]. Changes at silent sites can disrupt splicing control elements and create new "cryptic" splice sites, as well as mRNA and transcript stability can be affected through preference or avoidance of certain sequence motifs (see [92, 97]). Silent changes may affect gene regulation via constraints for efficient binding of miRNA to sense mRNA (e.g., [92, 98]). Selection may act on the choice of synonymous codons near miRNA targets, improving the binding site accessibility, binding efficiency and consequently the function of miRNA itself [99]. Programmed ribosomal frameshifting may be another reason for selection to act on specific codon positions [100]. Speeddependent protein folding also has been proposed to be a result of selective pressure [101]. According to the co-translational protein folding hypothesis, slower production could cause the protein to take an altered final form (as has been shown in multidrug resistance-1, [102]). Finally, synonymous changes may act to modulate expression by altering mRNA secondary structure, affecting protein abundance [103].
Models of codon evolution currently provide the most powerful approach for studying selection on silent sites. Models with variable synonymous rates (see [64, 104]) have been used to evaluate the extent of variability of synonymous rates in a gene and to predict specific sites with most extreme—low or high—synonymous rates (for example see [93]). A large-scale study of synonymous rate variation [105] described some intriguing general patterns and showed that the phenomenon is widespread in protein-coding genes. Genes displaying significantly varying synonymous rates increased association with several genetic diseases (especially cancers and diabetes) and were enriched for metabolic pathways. Other studies specifically focusing on human oncogenes revealed that a significant proportion of all cancer driver mutations were synonymous [106]. This suggests that synonymous rates cannot be automatically assumed fitness-neutral. Note that ω ¼ dN/dS, an accepted measure of selection on the protein, is not designed to detect selection on synonymous codons, particularly when d<sup>S</sup> is assumed constant. Yet, some cautioned that low synonymous rates preserved by purifying selection might erroneously lead to the detection of positive selection on the protein (e.g., Rubinstein et al. [107]). However, the usage of the ω ratio does not rely on the assumption that synonymous sites are neutral (pages 58–59 of Yang [108]; and Section 6.3 of Anisimova and Liberles [109]); rather, it is defined as a ratio of two ratios, comparing the proportions of nonsynonymous and synonymous sites after and before selection has operated on the protein (ω ¼ 1). In general we can assume that the evolutionary forces apply equally to synonymous and nonsynonymous sites. Forces that act differentially on synonymous and nonsynonymous sites should be rare in real data, but they can affect the validity of the ω measure. The only known example of such a natural force is probably synonymous phasing, considered by Xing and Lee [110]. But even in this case, and with a worst case scenario, the estimated effect is very weak. More crucially, an adequate description of mutational processes at the DNA level allows to circumvent biases in the estimation of the ω ratio [106].
Further testing, however, is necessary to decide whether any specific site has been affected by selection on synonymous codon usage. For example, Zhou, Gu, and Wilke [111] suggested distinguishing two types of synonymous substitution rates: the rate of conserving synonymous changes dSC (between "preferred" codons or between "rare" codons) and the rate of non-conserving synonymous changes dSN (between codons from the two different groups "rare" and "preferred"). Silent sites with dSN/dSC > 1 may be considered to be under positive selection, and significance can be tested based on a likelihood ratio test. Alternatively, synonymous rates at sites may be compared to the mean substitution rate in the corresponding intron, which can be implemented in a joint codon and DNA model, similar to the approach proposed by Wong and Nielsen [112].
Mutation-selection models include selective and mutational effects separately and allow estimating the fitness of various codon changes (see [113–115]). The relative rate of substitution for selected mutations to neutral mutations is given by ω ¼ 2γ/ (1 e -2γ ), where γ ¼ 2Ns is the scaled selection coefficient (see Exercise 3 for a derivation). Nielsen et al. [114] assumed that all changes between preferred and rare codons have the same fitness (and so the same selection coefficient). They used one selection coefficient for optimal codon usage for each branch of a phylogeny and estimated these jointly with the ω ratio by ML. Using this approach to study ancestral codon usage bias, Nielsen et al. [114] confirmed the reduction in selection for optimal codon usage in D. melanogaster. In contrast, Yang and Nielsen [115] estimated individual codon fitness parameters and used them to estimate optimal codon frequencies for a gene across multiple species. LRT is used to test whether the codon bias is due to the mutational bias alone. Nevertheless, one remarkable contribution of the mutationselection models is the connection they make between the interspecific and population parameters. Exploiting this further should provide insights to how changing demographic factors influence observed intraspecific patterns. Mutation-selection models also allow a new perspective on understanding codon models in the context of fitness landscapes with statistical implications as discussed in Subheading 4.2 of Chapter 13 by Jones, Susko, and Bielawski.
Finally, it is also possible to study selection on synonymous changes by introducing a parametric relationship between fitness and protein production cost. The idea was first described by Gilchrist [116], who assumed that, in addition to mutation and drift, the codon bias evolved under selection to reduce the cost of nonsense errors. Protein production cost can be computed as a ratio of the expected cost to the expected benefit [117]. Kubatko and colleagues [118] have extended a standard codon model to include the difference in protein production due to the usage of different codons (and therefore different elongation probabilities). However, such a model requires position-specific instantaneous rate matrices, and consequently also the probability transition matrices, making the approach computationally very intensive. To circumvent this, a GPU-based implementation was developed and used for phylogeny inference from 104 gene data set from Saccharomyces cerevisiae. Based on the standard model selection measure AIC, the new model outperformed the simplest model M0 as well as the mutation-selection model FMutSel of Yang and Nielsen.
#### 5 Exercises
Q1. Amino Acid and Codon Substitution Models
How many parameters need to be estimated in the instantaneous rate matrix Q defining a reversible empirical AA model? How many such parameters are necessary to estimate for a reversible empirical codon model? How many parameters are to be estimated in both cases if a model is nonreversible?
#### Q2. Positive Selection Scans
#### Q3. Selection-Mutation Models
Selection-mutation rely on a theoretical relationship between the nonsynonymous-synonymous rate ratio ω and the scaled selection coefficient γ ¼ 2Ns. The probability that a new mutation eventually becomes fixed is
$$\Pr(\text{fixation}) = \left(1 - e^{-2s}\right) / \left(1 - e^{-4Ns}\right) = 2s / \left(1 - e^{-4Ns}\right).$$
if we assume that the selection coefficient s is small and N is large and represents the effective population size, which is constant in time (Kimura and Ohta [119]). Furthermore, assume that synonymous substitutions are neutral and nonsynonymous have equal (and small) selection coefficients. Derive the relationship:
$$a = 4s/(1 - e^{-4Ns}) = 2\gamma/(1 - e^{-2\gamma})$$
that combines phylogenetic with population genetic quantities and is crucial for mutation-selection models.
#### Acknowledgments
C. K. is supported by a grant of the Vienna Science and Technology Fund (WWTF—MA016-061). M. A. receives funding from the Swiss National Science Foundation (grant 31003A\_176316).
#### References
The sequence and de novo assembly of the giant panda genome. Nature 463:311–317
model of molecular evolution. J Mol Evol 53:711–753
biomedical insights from the macaque genome. Science 316:222–234
misfolding: new research. Nova Science Publisher Inc, New York, NY
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Chapter 13
#### Looking for Darwin in Genomic Sequences: Validity and Success Depends on the Relationship Between Model and Data
#### Christopher T. Jones, Edward Susko, and Joseph P. Bielawski
#### Abstract
Codon substitution models (CSMs) are commonly used to infer the history of natural section for a set of protein-coding sequences, often with the explicit goal of detecting the signature of positive Darwinian selection. However, the validity and success of CSMs used in conjunction with the maximum likelihood (ML) framework is sometimes challenged with claims that the approach might too often support false conclusions. In this chapter, we use a case study approach to identify four legitimate statistical difficulties associated with inference of evolutionary events using CSMs. These include: (1) model misspecification, (2) low information content, (3) the confounding of processes, and (4) phenomenological load, or PL. While past criticisms of CSMs can be connected to these issues, the historical critiques were often misdirected, or overstated, because they failed to recognize that the success of any model-based approach depends on the relationship between model and data. Here, we explore this relationship and provide a candid assessment of the limitations of CSMs to extract historical information from extant sequences. To aid in this assessment, we provide a brief overview of: (1) a more realistic way of thinking about the process of codon evolution framed in terms of population genetic parameters, and (2) a novel presentation of the ML statistical framework. We then divide the development of CSMs into two broad phases of scientific activity and show that the latter phase is characterized by increases in model complexity that can sometimes negatively impact inference of evolutionary mechanisms. Such problems are not yet widely appreciated by the users of CSMs. These problems can be avoided by using a model that is appropriate for the data; but, understanding the relationship between the data and a fitted model is a difficult task. We argue that the only way to properly understand that relationship is to perform in silico experiments using a generating process that can mimic the data as closely as possible. The mutation-selection modeling framework (MutSel) is presented as the basis of such a generating process. We contend that if complex CSMs continue to be developed for testing explicit mechanistic hypotheses, then additional analyses such as those described in here (e.g., penalized LRTs and estimation of PL) will need to be applied alongside the more traditional inferential methods.
Key words Codon substitution model, dN/dS, False positives, Maximum likelihood, Mechanistic model, Model misspecification, Mutation-selection model, Parameter confounding, Phenomenological load, Phenomenological model, Positive selection, Reliability, Statistical inference, Site-specific fitness landscape
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_13, © The Author(s) 2019
#### 1 Introduction
Codon substitution models (CSMs) fitted to an alignment of homologous protein-coding genes are commonly used to make inferences about evolutionary processes at the molecular level (see Chapter 10 for examples of different applications of CSMs). Such processes (e.g., mutation and selection) are represented by a vector of parameters θ that can be estimated using maximum likelihood (ML) or Bayesian statistical methods. Here, we focus on ML and for convenience use CSM to indicate a model that is used in conjunction with the ML approach (see [21], for an example of the Bayesian approach). Considerable apprehension was expressed about the statistical validity of CSMs during their initial phase of development. In particular were concerns over the risk of falsely inferring that a sequence or codon site evolved by adaptive evolution [11, 22, 23, 46, 60–63, 85]. Many of the studies employed in the critique of CSMs were later shown to be flawed due to statistical errors or incorrect interpretation of results [70, 72, 77, 84]. In their reanalysis of the iconic MHC dataset [24], for example, Suzuki and Nei [61] based their criticism of the ML approach on results that were incorrect due to computational issues [70]. And in simulation studies by Suzuki [60] and Nozawa et al. [46], the branch-site model of Yang and Nielsen [79] was criticized as being too liberal because it falsely inferred positive selection at 32 out of 14,000 simulated sites, despite that this rate (0.0023) was well below the level of significance of the test (<sup>α</sup> ¼ 0.05) [77]. Concerns about the ML approach were eventually mollified by numerous simulation studies showing that the false positive rate is no greater than the specified level of significance of the LRT under a wide range of evolutionary scenarios [2, 3, 29, 31, 37, 70, 77, 82, 85, 86]. The validity and success of the approach is now well established [84], and this has led to the formulation of CSMs of ever-increasing sophistication [31, 41, 48–50, 55, 64, 65].
The most common use of a CSM is to infer whether a given process, such as adaptive evolution somewhere in the gene, the fixation of double and triple mutations, or variations in the synonymous substitution rate, actually occurred when the alignment was generated. Several factors can potentially undermine the reliability of such inferences. These include:
These same factors can impact any model-based effort to make inferences from data generated by complex biological processes, not only to the CSMs described here. The possibility of false inference due to any combination of these factors does not imply that the CSM approach is unreliable in principle. As has been demonstrated by numerous successful applications, CSMs generally extract accurate and useful information provided that the model is well suited for the data at hand [1, 71, 76]. We maintain that the validity of inferences is not a function of the model in and of itself, but is a consequence of the relationship between the model and the data.
Here, we explore this relationship via case studies taken from the historical development of CSMs. Our objective is to be candid about the limitations of CSMs to reliably extract information from an alignment. But, we emphasize that the impact of these limitations (i.e., false positives and confounding) is a consequence of a mismatch between the parameters included in the model and the often limited information contained in the alignment. The case studies are divided into two parts, each corresponding to a distinct phase in the development of CSMs. Phase I is characterized by pioneering efforts to formulate CSMs to account for the most prominent components of variation in an alignment [16, 42]. These include the M-series models that were among the first CSMs to account for variations in selection effects across sites [81], and the branch-site model of Yang and Nielsen [79] (hereafter, YN-BSM) formulated to account for variations in selection effects across both sites and branches. The first pair of case studies exemplifies concerns about the impact of low information content (Case Study A) and model misspecification (Case Study B) on the probability of falsely detecting positive selection in a gene or at a particular codon site. We also include a description of methods recently developed to mitigate the problem of false inference.
Phase II in the historical development is characterized by the general increase in the complexity of CSMs aimed to account for more subtle components of variation in an alignment.<sup>1</sup> Models used to detect temporal changes in site-specific selection effects
<sup>1</sup>The original CSM proposed by Goldman and Yang [16] was in fact quite complex in that it adjusted substitution rates between nonsynonymous codons to account for differences in physicochemical properties using the Grantham matrix [17]. This approach was later abandoned in favor of the simpler formulation now known as M0 [44], e.g., the first M-series model [81].
(e.g., [18, 31, 55]) or "heterotachy" [36] are representative. The movement toward complex parameter-rich models has resulted in a new set of concerns that are not yet widely appreciated. Principal among these is an increase in the possibility of confounding. Two components of the alignment-generating process are confounded if they can produce the same or similar patterns in the data. Such components can be impossible to disentangle without the input of further biological information, and their existence can lead to a statistical pathology that we call phenomenological load (PL). The second pair of case studies illustrates the possibility of false inference due to confounding (Case Study C) and PL (Case Study D). An essential feature of these studies is the use of a much more realistic generating model to produce alignments for the purpose of model evaluation.
Recent discoveries made using the mutation-selection (MutSel; [80]) framework of Halpern and Bruno [19], which is based on a realistic approximation of population dynamics at individual codon sites, have challenged the way we think about the relationship between parameters of traditional CSMs and components of the process of molecular evolution they are meant to summarize (e.g., [25, 26, 56, 57]). Previously, there has been a tendency to think about alignment-generating processes as if they occur in the same way they are modeled by a CSM. This way of thinking can be misleading because mechanisms of protein evolution can differ in important and substantial ways from traditional CSMs. To redress this issue, we begin this chapter with a brief overview of the conceptual foundations of MutSel as a more realistic way of thinking about the actual process of molecular evolution. This material is followed by a novel presentation of the ML statistical framework intended to illustrate potential limitations in what can reasonably be inferred when a CSM is fitted to data.
#### 2 Conceptual Foundations
#### 2.1 How Should We Think About the Alignment-Generating Process?
A codon substitution model represents an attempt to explain the way a target protein-coding gene changed over time by a combination of mutation, selection (purifying as well as adaptive), and drift. Adaptive evolution occurs at each site within a protein in response to a hierarchy of effects, including, but not limited to, changes in the network of the protein's interactions, changes in the functional properties of that network, and changes in both the cellular and organismal environment over time. The result of the complex interplay between these effects is typically viewed through the narrow lens of an alignment of homologous sequences X obtained from extant species, possibly accompanied by a tree topology τ (for our purposes, it is always assumed that τ is known). The information contained in X is evidently insufficient to resolve all of the effects of the true generating process, which would in any case be difficult or even impossible to parameterize with any accuracy. It is therefore necessary to base the formulation of a CSM on a number of simplifying assumptions. The usual assumptions include that:
The elements qij of a substitution rate matrix Q are typically defined for codons <sup>i</sup> 6¼ <sup>j</sup> as follows [44]:
ð1Þ
where κ is the transition bias and πi is the stationary frequency of the ith codon, both assumed to be the same for all codon sites. The ratio <sup>ω</sup> ¼ dN/dS of the nonsynonymous substitution rate dN to the synonymous substitution rate dS (both adjusted for "opportunity"<sup>2</sup> ) quantifies the stringency of selection at the site, with values closer to zero corresponding to sites that are more strongly conserved. We follow standard notation and use ω^ to represent the maximum likelihood estimate (MLE) of ω obtained by fitting Eq. 1 to an alignment.
Equation 1 provides the building block for most CSMs, yet it is unsuitable as a means to think about the substitution process at a site. For instance, the rate ratio in Eq. 1 is assumed to be the same for all nonsynonymous pairs of codons. If interpreted mechanistically, this is tantamount to the assumption that the amino acid occupying a site has fitness f and all other amino acids have fitness f + df, and that, with each substitution, the newly fixed amino acid changes its fitness to f and the previous occupant changes it fitness
<sup>2</sup> Single-nucleotide (SN) mutations that are nonsynonymous occur more frequently than those that are synonymous due to idiosyncrasies in the genetic code. This is accounted for in the formulation of dN and dS, so that dN can be interpreted as the proportion of nonsynonymous SN mutations that are fixed. Likewise, dS is the proportion of synonymous SN mutations that are fixed. See Jones et al. [25] for a discussion of various interpretations of dN/dS.
to f + df. Such a narrow view of the substitution process, akin to frequency-dependent selection [6, 25], is conceptually misleading for the majority of proteins. To be clear, CSMs are undoubtedly a valuable tool to make inferences about the evolution of a protein (e.g., [8, 52, 71, 76]); our point is that they do not necessarily provide the best way to think about the process.
The way we think about the substitution process should not be limited to unrealistic assumptions used to formulate a tractable CSM. It is more informative to conceptualize evolution at a codon site using the traditional metaphor of a fitness landscape upon which greater height represents greater fitness as depicted in Fig. 1. If sites are assumed to evolve independently, a site-specific fitness landscape can be defined for the hth site by a vector of fitness coefficients f <sup>h</sup> and its implied vector of equilibrium codon frequencies <sup>π</sup><sup>h</sup> . Combined with a model for the mutation process, <sup>π</sup><sup>h</sup> determines the evolutionary dynamics at the site, or the way it "moves" over its landscape (more formally, the way mutation and fixation events occur at a codon site in a population over time). This provides a way to think about evolution at a codon site in terms of three possible dynamic regimes: shifting balance, under which the site moves episodically away from the peak of its fitness landscape (i.e., the fittest amino acid) via drift and back again by positive selection (Fig. 1a); adaptive evolution, under which a change in the landscape is followed by movement of the site toward its new fitness peak (Fig. 1b); and neutral or nearly neutral evolution, under which drift dominates and the site is free to move over a relatively flat landscape limited primarily by biases in the mutation process. This way of thinking about the alignment-generating process is encapsulated by the MutSel framework [6, 7, 25]. The precise relationship between the MutSel framework and the three dynamic regimes will be presented in Case Study C.
#### 2.2 What Is the Objective of Model Building?
CSMs have become increasingly complex with the addition of more free parameters since the introduction of the M-series models in Yang et al. [81]. The prima facie objective of this trend is to produce models that provide better mechanistic explanations of the data. The assumption is that this will lead to more accurate inferences about evolutionary processes, particularly as the volume of genetic data increases [35]. However, the significance of a new model parameter is assessed by a comparison of site-pattern distributions without reference to mechanism. Combined with the possibility of confounding, this feature of the ML framework means that the objective of improving model fit does not necessarily coincide with the objective of providing a better representation of the mechanisms of the true generating process.
Given any CSM with parameters θM, it is possible to compute a vector <sup>P</sup> that assigns a probability to each of the 61<sup>N</sup> possible site patterns for an N-taxon alignment (i.e., a multinomial distribution
Fig. 1 It can be useful to think of the substitution process at a site as movement on a site-specific fitness landscape. The horizontal axis in each figure shows the amino acids at a hypothetical site in order of their stationary frequencies indicated by the height of the bars. Frequency is a function of mutation and selection, but can be construed as a proxy for fitness. The site-specific dN/dS ratio [25] is a function of the amino acid that occupies the site, and can be <1 (left of the red dashed line) or >1 (right of the dashed red line). (a) Suppose phenylalanine (F, TTT) is the fittest amino acid. The site-specific dN/dS ratio is much less than one when occupied by F because any nonsynonymous mutation will always be to an amino acid that is less fit. Nevertheless, it is possible for an amino acid such as valine (V, GTT) to be fixed on occasion, provided that selection is not too stringent. When this happens, dN/dS at the site is temporarily elevated to a value greater than one as positive selection moves the site back to F by a series of replacement substitutions, e.g., V (GTT) ! <sup>G</sup> (GGT) ! C (TGT) ! F (TTT). We call the episodic recurrence of this process shifting balance on a static fitness landscape. Shifting balance on a landscape for which all frequencies are approximately equal corresponds to nearly neutral evolution (not depicted), when dN/dS is always -1. (b) Now, consider what happens following a change in one or more external factors that impact the functional significance of the site. The relative fitnesses of the amino acids might change from that depicted in a to that in b for instance, where glutamine (Q) is fittest. If at the time of the change the site is occupied by F (as is most likely), then dN/dS would be temporarily elevated as positive selection moves the site toward its new peak at Q, e.g., F (TTT) ! Y (TAT) ! H (CAT) ! Q (CAA). This process of adaptive evolution is followed by a return to shifting balance once the site is occupied by Q
for 61<sup>N</sup> categories). We refer to <sup>P</sup> ¼ <sup>P</sup>M(θM) as the site-pattern distribution for that model. Figure 2 depicts the space of all possible site-pattern distributions for an N-taxon alignment. Each ellipse represents the family of distributions {PM(θM)jθ<sup>M</sup> <sup>∈</sup> <sup>Ω</sup>M}, where Ω<sup>M</sup> is the vector space of all possible values of θM. For example, {PM0(θM0)jθM0 <sup>∈</sup> <sup>Ω</sup>M0} is the family of distributions that can be
Fig. 2 The (61<sup>N</sup> 1)-dimensional simplex containing all possible site-pattern distributions for an N-taxon alignment is depicted. The innermost ellipse represents the subspace {PM0(θM0)jθM0 <sup>∈</sup> <sup>Ω</sup>M0} that is the family of distributions that can be specified using M0, the simplest of CSMs. This is nested in the family of distributions that can be specified using M1 (blue ellipse), a hypothetical model that has the same parameters as M0 plus some extra parameters. Similarly, M1 is nested in M2 (red ellipse). Whereas models are represented by subspaces of distributions, the true generating process is represented by a single point PGP, the location of which is unknown. The empirical site-pattern distribution <sup>P</sup> <sup>S</sup>ð<sup>θ</sup> ^<sup>S</sup>Þ corresponds to the saturated model fitted to the alignment; with large samples, <sup>P</sup> <sup>S</sup>ð<sup>θ</sup> ^<sup>S</sup>Þ - P GP. For any other model M, the member <sup>P</sup> <sup>M</sup>ð<sup>θ</sup> ^<sup>M</sup>Þ∈f<sup>P</sup> <sup>M</sup>ðθMÞ j <sup>θ</sup>M<sup>∈</sup> <sup>Ω</sup>Mg most consistent with <sup>X</sup> is the one that minimizes deviance, which is twice the difference between the maximum log-likelihood of the data under the saturated model and the maximum log-likelihood of the data under M
specified using M0, the simplest CSM that assumes a common substitution rate matrix Q for all sites and branches. This is nested inside {PM1(θM1)jθM1 <sup>∈</sup> <sup>Ω</sup>M1}, where M1 is a hypothetical model that is the same as M0 but for a few extra parameters. Likewise, M1 is nested in M2. The location of the site-pattern distribution for the true generating process is represented by PPG. Its location is fixed but unknown. It is therefore not possible to assess the distance between it and any other distribution. Instead, comparisons are made using the site-pattern distribution inferred under the saturated model.
Whereas a CSM {PM(θM)jθ<sup>M</sup> <sup>∈</sup> <sup>Ω</sup>M} can be thought of as a family of multinomial distributions for the 61<sup>N</sup> possible site patterns, the fitted saturated model <sup>P</sup>Sð<sup>θ</sup> ^SÞ is the unique distribution defined by the MLE θ ^<sup>S</sup> ¼ ðy1=n, ... ,ym=n<sup>Þ</sup> <sup>T</sup> , where yi <sup>&</sup>gt; 0 is the observed frequency of the ith site pattern, m is the number of unique site patterns, and n is the number of codon sites. In other words, the fitted saturated model is the empirical site-pattern distribution for a given alignment. Because it takes none of the mechanisms of mutation or selection into account, ignores the phylogenetic relationships between sequences, and excludes the possibility of site patterns that were not actually observed (i.e., yi/<sup>n</sup> ¼ 0 for site patterns <sup>i</sup> not observed in <sup>X</sup>), <sup>P</sup>Sð<sup>θ</sup> ^SÞ can be construed as the maximally phenomenological explanation of the observed alignment. An alignment is always more likely under the saturated model than it is under any other CSM. <sup>P</sup>Sð<sup>θ</sup> ^SÞ therefore provides a natural benchmark for model improvement.
For any alignment, the MLE over the family of distributions {PM(θM)jθ<sup>M</sup> <sup>∈</sup> <sup>Ω</sup>M} is represented by a fixed point <sup>P</sup>Mð<sup>θ</sup> ^MÞ in Fig. 2. <sup>P</sup>Mð<sup>θ</sup> ^MÞ is the distribution that minimizes the statistical deviance between <sup>P</sup>M(θM) and <sup>P</sup>Sð<sup>θ</sup> ^SÞ. Deviance is defined as twice the difference between the maximum log-likelihood (LL) of the data under the saturated model and the maximum log-likelihood of the data under M:
$$D(\hat{\theta}\_{\text{M}}, \hat{\theta}\_{\text{S}}) = \mathcal{Z}\{\ell(\hat{\theta}\_{\text{S}} \mid X) - \ell(\hat{\theta}\_{\text{M}} \mid X)\}\tag{2}$$
A key feature of deviance is that it always decreases as more parameters are added to the model, corresponding to an increase in the probability of the data under that model. For example, suppose {PM2(θM2)jθM2 <sup>∈</sup> <sup>Ω</sup>M2} is the same family of distributions as {PM1(θM1)jθM1 <sup>∈</sup> <sup>Ω</sup>M1} but for the inclusion of one additional parameter <sup>ψ</sup>, so that <sup>θ</sup>M2 <sup>¼</sup> (θM1, <sup>ψ</sup>). The improvement in the probability of the data under <sup>P</sup>M2ð<sup>θ</sup> ^M2Þ over its probability under <sup>P</sup>M1ð<sup>θ</sup> ^M1Þ is assessed by the size of the reduction in deviance induced by ψ:
$$\begin{split} \Delta D(\hat{\theta}\_{\text{M1}}, \hat{\theta}\_{\text{M2}}) &= D(\hat{\theta}\_{\text{M1}}, \hat{\theta}\_{\text{S}}) - D(\hat{\theta}\_{\text{M2}}, \hat{\theta}\_{\text{S}}) \\ &= \mathcal{Q}\{\ell(\hat{\theta}\_{\text{M2}} \mid X) - \ell(\hat{\theta}\_{\text{M1}} \mid X)\} \end{split} \tag{3}$$
Equation 3 is just the familiar log-likelihood ratio (LLR) used to compare nested models under the maximum likelihood framework.
Given this measure of model improvement, the de facto objective of model building is not to provide a mechanistic explanation of the data that more accurately represents the true generating process, but only to move closer to the site-pattern distribution of the fitted saturated model. Real alignments are limited in size, so there will always be some distance between <sup>P</sup>Sð<sup>θ</sup> ^S<sup>Þ</sup> and <sup>P</sup>GP due to sampling error (as represented in Fig. 2). But even with an infinite number of codon sites, when <sup>P</sup>Sð<sup>θ</sup> ^SÞ converges to <sup>P</sup>GP, the criterion of minimizing deviance does not inevitably lead to a better explanation of the data because of the possibility of confounding. Two processes are said to be confounded if they can produce similar patterns in the data. Hence, if ψ represents a process E that did not actually occur when the data was generated, and if E is confounded with another process that did occur, the LLR in Eq. 3 can still be significant. Under this scenario, the addition of ψ to M1 would engender movement toward <sup>P</sup>Sð<sup>θ</sup> ^SÞ and <sup>P</sup>GP, but the new model M2 would also provide a worse mechanistic explanation of the data because it would falsely indicate that E occurred. The possibility of confounding and its impact on inference is demonstrated in Case Study D.
#### 3 Phase I: Pioneering CSMs
The first effort to detect positive selection at the molecular level [24] relied on heuristic counting methods [43]. Phase I of CSM development followed with the introduction of formal statistical approaches based on ML [16, 42]. The first CSMs were used to infer whether the estimate ω^ of a single nonsynonymous to synonymous substitution rate ratio averaged over all sites and branches was significantly greater than one. Such CSMs were found to have low power due to the pervasiveness of synonymous substitutions at most sites within a typical gene [76]. An early attempt to increase the statistical power to infer positive selection was the CSM designed to detect ω^ > 1 on specific branches [78]. Models accounting for variations in ω across sites were subsequently developed, the most prominent of which are the M-series models [78, 81]. These were accompanied by methods to identify individual sites under positive selection. The quest for power culminated in the development of models that account for variations in the rate ratio across both sites and branches. The appearance of various branch-site models (e.g., [4, 10, 79, 86]) marks the end of Phase I of CSM development.
Two case studies are employed in this section to illustrate some of the inferential challenges associated with Phase I models. We use Case Study A to examine the impact of low information content on the inference of positive selection at individual codon sites. The subject of this study is the M1a vs M2a model contrast applied to the tax gene of the human T-cell lymphotropic virus type I (HTLV-I; [63, 82]). We use Case Study B to illustrate how model misspecification (i.e., differences between the fitted model and the generating process) can lead to false inferences. The subject of this study is the Yang–Nielsen branch-site model (YN-BSM; [79]) applied to simulated data.
3.1 Case Study A: Low Information Content
To study the impact of low information content on inference, we use a pair of nested M-series models known as M1a and M2a [70, 82]. Under M1a, sites are partitioned into two rate-ratio categories, 0 <sup>&</sup>lt; <sup>ω</sup><sup>0</sup> <sup>&</sup>lt; 1 and <sup>ω</sup><sup>1</sup> <sup>¼</sup> 1 in proportions <sup>p</sup><sup>0</sup> and <sup>p</sup><sup>1</sup> <sup>¼</sup> <sup>1</sup> <sup>p</sup>0. M2a includes an additional category for the proportion of sites <sup>p</sup><sup>2</sup> <sup>¼</sup> <sup>1</sup> <sup>p</sup><sup>0</sup> <sup>p</sup><sup>2</sup> that evolved under positive selection with ω<sup>2</sup> > 1. The use of multiple categories permits two levels of inference. The first is an omnibus likelihood ratio test (LRT) for evidence of positive selection somewhere in the gene, which is conducted by contrasting a pair of nested models. For example, the contrast of M1a vs M2a is made by computing the distance LLR ¼ <sup>Δ</sup>Dð<sup>θ</sup> ^M1a, θ ^M2aÞ between the two models and comparing the result to the limiting distribution of the LLR under the null model. In this case, the limiting distribution of LLR is often taken to be χ<sup>2</sup> <sup>2</sup> [75], which would be correct under regular likelihood theory because the models differ by two parameters. The second level of inference is used to identify individual sites that underwent positive selection. This is conducted only if positive selection is inferred by the omnibus test (e.g., if LLR > 5.99 for the M1a vs M2a contrast at the 5% level of significance). Let c0, c1, and c<sup>2</sup> represent the event that a given site pattern x falls into the stringent ( 0 <sup>&</sup>lt; <sup>ω</sup>^ <sup>0</sup> <sup>&</sup>lt; 1 ), neutral (ω^ <sup>1</sup> <sup>¼</sup> 1 ), or positive (ω^ <sup>2</sup> <sup>&</sup>gt; 1 ) selection category, respectively. Applying Bayes' rule:
$$\Pr(\boldsymbol{\varepsilon}\_{2} \mid \boldsymbol{\varkappa}, \hat{\boldsymbol{\theta}}\_{\text{M2a}}) = \frac{\Pr(\boldsymbol{\varkappa} \mid \boldsymbol{\varepsilon}\_{2}, \hat{\boldsymbol{\theta}}\_{\text{M2a}}) \hat{p}\_{2}}{\sum\_{k=0}^{2} \Pr(\boldsymbol{\varkappa} \mid \boldsymbol{\varepsilon}\_{k}, \hat{\boldsymbol{\theta}}\_{\text{M2a}}) \hat{p}\_{k}} \tag{4}$$
Sites with a sufficiently high posterior probability (e.g., Prðc<sup>2</sup> <sup>j</sup> <sup>x</sup>,<sup>θ</sup> ^M2aÞ <sup>&</sup>gt; <sup>0</sup>:95 ) are inferred to have undergone positive selection. Equation 4 is representative of the naive empirical Bayes (NEB) approach under which MLEs (θ ^M2a ) are used to compute posterior probabilities.
The NEB approach ignores potential errors in parameter estimates that can lead to false inference of positive selection at a site (i.e., a false positive). The resulting false positive rate can be especially high for alignments with low information content. An example setting with low information content arises when there are a substantial number of invariant sites, since these provide little information about the substitution process. The issue of low information content is well illustrated by the extreme case of the tax gene, HTLV-I [63]. The alignment consists of 20 sequences with 181 codon sites, 158 of which are invariant. The 23 variable sites have only one substitution each: 2 are synonymous and 21 are nonsynonymous. The high ratio of nonsynonymous-to-synonymous substitutions suggests that the gene underwent positive selection. This hypothesis was supported by analytic results: the LLR for the M1a vs M2a contrast was 6.96 corresponding to a pvalue of approximately 0.03 [82]. The omnibus test therefore supported the conclusion that the gene underwent positive selection. However, the MLE for p<sup>2</sup> under M2a was p ^<sup>2</sup> <sup>¼</sup> 1. Using this value in Eq. <sup>4</sup> gives Prðc<sup>2</sup> <sup>j</sup> <sup>x</sup>,<sup>θ</sup> ^M2aÞ ¼ 1 for all sites, including the 158 invariable sites. Such an unreasonable result can occur under NEB because, despite the possibility of large sampling errors in MLEs due to low information, θ ^M2a is treated as a known value in Eq. 4.
Bayes empirical Bayes (BEB; [82]), a partial Bayesian approach under which rate ratios and their corresponding proportions are assigned discrete prior distributions (cf. [21]), was proposed as an alternative to NEB. Numerical integration over the assumed priors tends to provide better estimates of posterior probabilities, particularly in cases where information content is low. Using BEB in the analysis of the tax gene, for example, the posterior probability was <sup>0</sup>:<sup>91</sup> <sup>&</sup>lt; Prðc<sup>2</sup> <sup>j</sup> <sup>x</sup>,<sup>θ</sup> ^M2aÞ <sup>&</sup>lt; <sup>0</sup>:93 for the 21 sites with a single nonsynonymous change and 0:<sup>55</sup> <sup>&</sup>lt; Prðc<sup>2</sup> <sup>j</sup> <sup>x</sup>,<sup>θ</sup> ^M2aÞ <sup>&</sup>lt; <sup>0</sup>:61 for the remaining sites [82]. Hence, the BEB approach mitigated the problem of low information content, as the posterior probability of positive selection at invariant sites was reduced. An alternative to BEB is called smoothed bootstrap aggregation (SBA) [38]. SBA entails drawing site patterns from X with replacement (i.e., bootstrap) to generate a set of alignments {X1, ..., Xm} with similar information content as <sup>X</sup>. The MLEs f<sup>θ</sup> ^ig m <sup>i</sup>¼<sup>1</sup> for the vector of model parameters θ are then estimated by fitting the CSM to each Xi ∈{X1, ..., Xm}. A kernel smoother is applied to these values to reduce sampling errors. The mean value of the resulting smoothed fθ ^ig m <sup>i</sup>¼<sup>1</sup> is then used in Eq. <sup>4</sup> in place of the MLE for <sup>θ</sup> obtained from the original alignment to estimate posterior probabilities. This approach was shown to balance power and accuracy at least as well as BEB. But, SBA has the advantage that it can accommodate the uncertainty of all parameter estimates (not just those of the ω distribution, as in BEB) and is much easier to implement. When SBA was applied to the tax gene, the posterior probabilities for positive selection were further reduced: 0:<sup>87</sup> <sup>&</sup>lt; Prðc<sup>2</sup> <sup>j</sup> <sup>x</sup>,<sup>θ</sup> ^M2aÞ <sup>&</sup>lt; <sup>0</sup>:<sup>89</sup> for the 21 sites with a single nonsynonymous change, and 0:55 < Prðc<sup>2</sup> <sup>j</sup> <sup>x</sup>,<sup>θ</sup> ^M2aÞ <sup>&</sup>lt; <sup>0</sup>:60 for the remaining sites [38].
The problem of low information content was fairly obvious in the case of the tax gene, as 158 of the 181 codon sites within that dataset were invariant. However, it can sometimes be unclear whether there is enough variation in an alignment to ensure reliable inferences. It would be useful to have a method to determine whether a given data set might be problematic. An MLE θ ^ will always converge to a normal distribution centered at the true parameter value θ with variance proportional to 1/n as the sample size n (a proxy for information content) gets larger, provided that the CSM satisfies certain "regularity" conditions (a set of technical conditions that must hold to guarantee that MLEs will converge in distribution to a normal, and that the LLR for any pair of nested models will converge to its expected chi-squared distribution). This expectation makes it possible to assess whether an alignment is sufficiently informative to obtain the benefits of regularity. The first step is to generate a set of bootstrap alignments {X1, ..., Xm}. The CSM can then be fitted to these to produce a sample distribution f<sup>θ</sup> ^ig m <sup>i</sup>¼<sup>1</sup> for the MLE of any model parameter <sup>θ</sup>. If the alignment is sufficiently informative with respect θ, then a histogram of fθ ^ig m <sup>i</sup>¼<sup>1</sup> should be approximately normal in distribution. Serious departures from normality (e.g., a bimodal distribution) indicate unstable MLEs, which are a sign of insufficient information or an irregular modeling scenario. Mingrone et al. [38] recommend using this technique with real data as a means of gaining insight into potential difficulties of parameter estimation using a given CSM.
3.1.1 Irregularity and Penalized Likelihood Issues associated with low information content can be made worse by violations of certain regularity conditions. For example, M2a is the same as M1a but for two extra parameters, p<sup>2</sup> and ω2. Usual likelihood theory would therefore predict that the limiting distribution of the LLR is χ<sup>2</sup> 2. However, this result is valid only if the regularity conditions hold. Among these conditions is that the null model is not obtained by placing parameters of the alternate model on the boundary of parameter space. Since M1a is the same as M2a but with <sup>p</sup><sup>2</sup> <sup>¼</sup> 0, this condition is violated. The same can be said for many nested pairs of Phase I CSMs, such as M7 vs M8 [81] or M1 vs branch-site Model A [79]. Although the theoretical limiting distribution of the LLR under some irregular conditions has been determined by Self and Liang [54], those results do not include cases where one of the model parameters is unidentifiable under the null [2]. Since M1a is M2a with <sup>p</sup><sup>2</sup> <sup>¼</sup> 0, the likelihood under M1a is the same for any value of ω2. This makes ω<sup>2</sup> unidentifiable under the null. The limiting distribution for the M1a vs M2a contrast is therefore unknown [74].
> A penalized likelihood ratio test (PLRT; [39]) has been proposed to mitigate problems associated with unidentifiable parameters. Under this method, the likelihood function for the alternate model (e.g., M2a) is modified so that values of p<sup>2</sup> closer to zero are penalized. This has the effect of drawing the MLE for p<sup>2</sup> away from the boundary, and can interpreted as a way to "regularize" the model. PLRT seems to be more useful in cases where the analysis of a real alignment produces a small value of p ^<sup>2</sup> accompanied by an unrealistically large value of ω^ 2. This can happen because ω^ <sup>2</sup> is influenced by fewer and fewer site patterns as p ^2 approaches zero, and is therefore subject to larger and larger sampling errors. In addition,ω^ <sup>2</sup> and p ^<sup>2</sup> tend to be negatively correlated, which further contributes to the large sampling errors. For example, Mingrone et al. [39] found that M2a fitted to a 5-taxon alignment with 198 codon sites without penalization gave ðp ^2,ω^ <sup>2</sup>Þ¼ð0:01,34:70Þ. These MLEs, if taken at face value, suggest that a small number of sites in the gene underwent positive
selection. However, such a large rate ratio is difficult to believe given that its estimate is consistent with only approximately 2 codon sites (e.g., an estimated 1% of the 198 sites or -2 sites). Using the PLRT, the MLEs were ð<sup>p</sup> ^2,ω^ <sup>2</sup>Þ¼ð0:09,1:00Þ. These suggest that selection pressure was nearly neutral at a significant proportion of sites in the gene. In this case, the rate ratio is consistent with 9% of the 198 sites or -18 sites and is therefore less likely to be an artifact of sampling error. We expect this approach to be useful in a wide variety of evolutionary applications that rely on mixture models to make inferences (e.g., [13, 34, 47, 66]).
Other approaches for dealing with low information content in the data for an individual gene include the empirical Bayes approach of Kosiol et al. [33] and the parametric bootstrapping methods of Gibbs [14]. Both methods exploit the additional information content available from other genes. Kosiol et al. [33] adopted an empirical Bayes approach, where ω values varied over edges and genes according to a distribution. Because empirical posterior distributions are used, the approach is more akin to detecting sites under positive selection (e.g., using NEB) than formal testing. By contrast, Gibbs [14] adopted a test-based approach and utilized parametric bootstrapping [15] to approximate the distribution of the likelihood ratio statistic using data from other genes to obtain parameter sets to use in the bootstrap. Whereas this approach can attenuate issues associated with low information content, it can also be computationally expensive, especially when applied to large alignments.
3.2 Case Study B: Model Misspecification The mechanisms that give rise to the diversity of site patterns in a set of homologous genes are highly complex and not fully understood. CSMs are therefore necessarily simplified representations of the true generating process, and are in this sense misspecified. The extent to which misspecification might cause an omnibus LRT to falsely detect positive selection was of primary concern during Phase I of model development. We use a particular form of the YN-BSM called Model A [79] to illustrate this issue. In its original form, the omnibus LRT assumes a null under which a proportion <sup>p</sup><sup>0</sup> of sites evolved under stringent selection with <sup>ω</sup><sup>0</sup> <sup>¼</sup> 0 and the remaining sites evolved under a neutral regime with <sup>ω</sup><sup>1</sup> <sup>¼</sup> 1 on all branches of the tree (i.e., model M1 in [44]). This is contrasted with Model A, which is the same as M1 except that it assumes that some stringent sites and some neutral sites evolved under positive selection with ω<sup>2</sup> > 1 on a prespecified branch called the foreground branch. The omnibus test contrasting M1 with Model A was therefore designed to detect a subset of sites that evolved adaptively on the same branch of the tree.
> During this period of model development, the standard method to test the impact of misspecification on the reliability of
Table 1 Rate ratios (ω) for regimes X and Z taken from Zhang [85]
an omnibus LRT was to generate alignments in silico using a more complex version of the CSM to be tested as the generating model. This usually involved adding more variability in ω across sites and/or branches than assumed by the fitted CSM while leaving all other aspects of the generating model the same. In Zhang [85], for example, alignments were generated using site-specific rate matrices, as in Eq. 1, with rate ratios ω specified by predetermined selection regimes, two of which are shown in Table 1. In one simulation, 200 alignments were generated using regime Z on a single foreground branch and regime X on all of the remaining branches of a 10 or 16 taxon tree. The gene therefore underwent a mixture of stringent selection and neutral evolution over most of the tree (regime X), but with complete relaxation of selection pressure on the foreground branch (regime Z). Positive selection did not occur at any of the sites. Nevertheless, the M1 vs Model A contrast inferred positive selection in 20–55% of the alignments, depending on the location of the foreground branch. Such a high rate of false positives was attributed to the mismatch between the process used to generate the data compared to the process assumed by the null model M1 [85].
The branch-site model was subsequently modified to allow <sup>0</sup> <sup>&</sup>lt; <sup>ω</sup><sup>0</sup> <sup>&</sup>lt; 1 instead of <sup>ω</sup><sup>0</sup> <sup>¼</sup> 0 (Modified Model A in [86]). Furthermore, the new null model is specified under the assumption that some proportion p<sup>0</sup> of sites (the stringent sites) evolved under stringent selection with 0 < ω<sup>0</sup> < 1 everywhere in the tree except on the foreground branch, where those same sites evolved neutrally with <sup>ω</sup><sup>2</sup> <sup>¼</sup> 1. All other sites in the alignment (the neutral sites) are assumed to have evolved neutrally with <sup>ω</sup><sup>1</sup> <sup>¼</sup> 1 everywhere in the tree. This is contrasted with the Modified Model A, which assumes that some of the stringent sites and some of the neutral sites evolved under positive selection with ω<sup>2</sup> > 1 on the foreground. Hence, unlike the original omnibus test that contrasts M1 with Model A, the new test contrasts Modified Model A with <sup>ω</sup><sup>2</sup> <sup>¼</sup> 1 against Modified Model A with ω<sup>2</sup> > 1. These changes to the YN-BSM were shown to mitigate the problem of false inference. For example, using the same generating model with regimes X and Z, the modified omnibus test falsely inferred positive selection in only 1–7.5% of the alignments, consistent with the 5% level of significance of the test [86].
This case study demonstrates how problems associated with model misspecification were traditionally identified, and how they could be completely corrected through relatively minor changes to the model. However, the generating methods employed by studies such as Zhang [85] and Zhang et al. [86], although sophisticated for their time, produced alignments that were highly unrealistic compared to real data. For example, it was recently shown that a substantial proportion of variation in many real alignments might be due to selection effects associated with shifting balance over static site-specific fitness landscapes [25, 26]. This process results in random changes in site-specific rate ratios, or heterotachy, that cannot be replicated using traditional CSMs as the generating model. While the mitigation of statistical pathologies due to low information content (e.g., using BEB or SBA) or model misspecification (e.g., by altering the null and alternative hypotheses or the use of penalized likelihood) were critical advancements during Phase I of CSM development, other statistical pathologies went unrecognized due to reliance on unrealistic simulation methods. This issue is taken up in the next section.
#### 4 Phase II: Advanced CSMs
A typical protein-coding gene evolves adaptively only episodically [59]. The evidence of adaptive evolution of this type can be very difficult to detect. For example, it is assumed under the YN-BSM that a random subset of sites switched from a stringent or neutral selection regime to positive selection together on the same set of foreground branches. The power to detect a signal of this kind can be very low when the proportion of sites that switched together is small [77]. Perhaps encouraged by the reliability of Phase I models demonstrated by extensive simulation studies [2, 3, 29, 31, 37, 70, 77, 82, 85, 86], combined with experimental validation of results obtained from their application to real data [1, 71, 76], investigators began to formulate increasingly complex and parameter-rich CSMs [31, 41, 48, 50, 55, 64, 65]. The hope was that carefully selected increases in model complexity would yield greater power to detect subtle signatures of positive selection overlooked by Phase I models. The introduction of such CSMs marks the beginning of Phase II of their historical development.
Phase II models fall into three broad categories:
1. The first consists of Phase I CSMs modified to account for more variability in selection effects across sites and branches than previously assumed, with the aim of increasing the power to detect subtle signatures of positive selection (e.g., the branch-site random effects likelihood model, BSREL; [31]).
An example of the first category of models is BSREL, which accounts for variations in selection effects across sites and over branches by assuming a different rate-ratio distribution fðωb <sup>i</sup> ,pb <sup>i</sup><sup>Þ</sup> : <sup>i</sup> <sup>¼</sup> 1, ... , kb Þg for each branch <sup>b</sup> of a tree [31]. BSREL was later found to be more complex than necessary, so an adaptive version was formulated to allow the number of components kb on a given branch to adjust to the apparent complexity of selection effects on that branch (aBSREL; [55]). A further reduction in model complexity led to the formulation of the test known as BUSTED (for branch-site unrestricted statistical test for episodic diversification; [41]), which we use to illustrate the problem of confounding in Case Study C. An example of the second category of models is the addition of parameters for the rate of double and triple mutations to traditional CSMs, the most sophisticated version of which is RaMoSSwDT (for Random Mixture of Static and Switching sites with fixation of Double and Triple mutations; [26]). This model is used in Case Study D to illustrate the problem of phenomenological load.
Models in the third category are the most ambitious CSMs currently in use, and are far more challenging to fit to real alignments than traditional models. One of the most impressive examples of their application is the site-wise mutation-selection model (swMutSel; [64, 65]) fitted to a concatenated alignment of 12 mitochondrial genes (3598 codon sites) from 244 mammalian species. Based on the mutation-selection framework of Halpern and Bruno [19], swMutSel estimates a vector of selection coefficients for each site in an alignment. This and similar models (e.g., [48–50]) appear to be reliable [58], but require a very large number of taxa (e.g., hundreds). Phase II models of this category are therefore impractical for the majority of empirical datasets. Here, we utilize MutSel as an effective means to generate realistic alignments with plausible levels of variation in selection effects across sites and over time rather than as a tool of inference.
#### 4.1 Case Study C: Confounding By expressing the codon substitution process in terms of explicit population genetic parameters, the MutSel framework facilitates the investigation of complex evolutionary dynamics, such as shifting balance on a fixed fitness landscape or adaptation to a change in selective constraints (i.e., a peak shift; [6, 25]) that are missing from alignments generated using traditional methods. Specifically, by assigning a different vector of fitness coefficients for the 20 amino acids to each site, MutSel can generate more variation in rate ratio across sites and over time than has been realized in the past simulation studies (e.g., Table 1). In this way, MutSel provides the basis of a generating model that can be adjusted to produce alignments that closely mimic real data [26]. MutSel therefore serves to connect demonstrably plausible evolutionary dynamics to the pathology we refer to as confounding.
Under MutSel, the dynamic regime at the hth codon site (e.g., shifting balance, neutral, nearly neutral, or adaptive evolution) is uniquely specified by a vector of fitness coefficients <sup>f</sup> <sup>h</sup> ¼ <sup>f</sup> <sup>h</sup> 1, ... , <sup>f</sup> <sup>h</sup> <sup>m</sup>. It is generally assumed that mutation to any of the three stop codons is lethal, so <sup>m</sup> ¼ 61 for nuclear genes and <sup>m</sup> ¼ 60 for mitochondrial genes. And, although it is not a requirement, it is typical to assume that the f <sup>h</sup> <sup>j</sup> are constant across synonymous codons [25, 57]. Given f h , the elements of a site-specific instantaneous rate matrix Ah can be defined as follows for all <sup>i</sup> 6¼ <sup>j</sup> (cf. Eq. 1):
$$A\_{ij}^b \propto \begin{cases} \mu\_{ij} & \text{if } s\_{ij}^b = 0\\ \mu\_{ij} \frac{s\_{ij}^b}{1 - \exp\left(-s\_{ij}^b\right)} & \text{otherwise} \end{cases} \tag{5}$$
where μij is the rate at which codon i mutates to codon j and sh ij <sup>¼</sup> <sup>2</sup><sup>N</sup> <sup>e</sup>ð<sup>f</sup> <sup>h</sup> <sup>j</sup> <sup>f</sup> <sup>h</sup> <sup>i</sup>Þ is the scaled selection coefficient for a population of haploids with effective population size Ne. The probability that the new mutant j is fixed is approximated by s h ij=f<sup>1</sup> expð<sup>s</sup> h ijÞg [9, 28].
The rate matrix <sup>A</sup><sup>h</sup> defines the dynamic regime for the site as illustrated in Fig. 3. The bar plot shows codon frequencies <sup>π</sup><sup>h</sup> ¼ <sup>π</sup><sup>h</sup> 1, ... , <sup>π</sup><sup>h</sup> <sup>m</sup> sorted in descending order. A site spends most of its time occupied by codons to the left or near the "peak" of its landscape. The codon-specific rate ratio for the site (dN <sup>h</sup> <sup>i</sup> =dS<sup>h</sup> <sup>i</sup> for codon i) is low near the peak (red line plot in Fig. 3) since mutations away from the peak are seldom fixed. However, if selection is not too stringent, the site will occasionally drift to the right into the "tail" of its landscape. When this occurs, the codon-specific rate
Fig. 3 Fitness coefficients for the 20 amino acids were drawn from a normal distribution centered at zero and with standard deviation <sup>σ</sup> ¼ 0.001. Bars show the resulting stationary frequencies (a proxy for fitness) sorted from largest to smallest. They compose a metaphorical site-specific landscape over which the site is imagined to move. The solid red line shows the codon-specific rate ratio dN<sup>h</sup> <sup>i</sup> =d S<sup>h</sup> <sup>i</sup> for the sorted codons. This varies depending on the codon currently occupying the site, and can be greater than one following a chance substitution into the tail (to the right) of the landscape. In this case, the codon-specific rate ratio for the site ranged from 0.21 to 4.94 with a temporally averaged sitespecific rate ratio of dN<sup>h</sup> /dS<sup>h</sup> ¼ 0.52
ratio will be elevated for a time until a combination of drift and positive selection moves the site back to its peak. This dynamic between selection and drift is reminiscent of Wright's shifting balance. It implies that, when a population is evolving on a fixed fitness landscape (i.e., with no adaptive evolution), its gene sequences can nevertheless contain signatures of temporal changes in site-specific rate ratios (heterotachy), and that these might include evidence of transient elevation to values greater than one (i.e., positive selection). Such signatures of positive selection due to shifting balance can be detected by Phase II CSMs [25].
For example, BUSTED [41] was developed as an omnibus test for episodic adaptive evolution. The underlying CSM was formulated to account for variations in the intensity of selection over both sites and time modeled as a random effect. This is in contrast to the YN-BSM, which treats temporal changes in rate ratio as a fixed effect that occurs on a prespecified foreground branch (although the sites under positive selection are still a random effect). We therefore refer to the CSM underlying BUSTED as the random effects branch-site model (RE-BSM) to serve as a reminder of this important distinction. Under RE-BSM, the rate ratio at each site and branch combination is assumed to be an independent draw from the distribution <sup>ð</sup>ω0, <sup>p</sup>0Þ,ðω1, <sup>p</sup>1Þ,ðω2, <sup>p</sup>2Þ . In this way, the model accounts for variations in selection effects both across sites and over time. BUSTED contrasts the null hypothesis that <sup>ω</sup><sup>0</sup> <sup>ω</sup><sup>1</sup> <sup>ω</sup><sup>2</sup> <sup>¼</sup> 1 with the alternative that <sup>ω</sup><sup>0</sup> <sup>ω</sup><sup>1</sup> <sup>1</sup> <sup>ω</sup>2. When applied to real data, rejection of the null is interpreted as evidence of episodic adaptive evolution.
Unlike the YN-BSM that aims to detect a subset of sites that underwent adaptive evolution together on the same foreground branches (i.e., coherently), BUSTED was designed to detect heterotachy similar to the type predicted by the mutation-selection framework: shifting balance on a static fitness landscape. Jones et al. [25] recently demonstrated that plausible levels of shifting balance can produce signatures of episodic positive selection that can be detected. BUSTED inferred episodic positive selection in as many as 40% of alignments generated using the MutSel framework. Significantly, BUSTED was correct to identify episodic positive selection in these trials. Even though the generating process assumed fixed site-specific landscapes (so there was no episodic adaptive evolution), and the long-run average rate ratio at each site was necessarily less than one [57], positive selection nevertheless did sometimes occur by shifting balance. This illustrates the general problem of confounding. Two processes are said to be confounded if they can produce the same or similar patterns in the data. In this case, episodic adaptive evolution (i.e., the evolutionary response to changes in site-specific landscapes) and shifting balance (i.e., evolution on a static fitness landscape) are confounded because they can both produce rate-ratio distributions that indicate episodic positive selection. The possibility of confounding underlines the fact that there are limitations in what can be inferred about evolutionary processes based on an alignment alone.
4.2 Case Study D: Phenomenological Load Phenomenological load (PL) is a statistical pathology related to both model misspecification (Case Study B) and confounding (Case Study C) that was not recognized during Phase I of CSM development. When a model parameter that represents a process that played no role in the generation of an alignment (i.e., a misspecified process) nevertheless absorbs a significant amount of variation, its MLE is said to carry PL [26]. This is more likely to occur when the misspecified process is confounded with one or more other processes that did play a role in the generation of the data, and when a substantial proportion of the total variation in the data is unaccommodated by the null model [26]. PL increases the probability that a hypothesis test designed to detect the misspecified process will be statistically significant (as indicated by a large LLR) and can therefore lead to the incorrect conclusion that the misspecified process occurred. Critically, Jones et al. [26] showed that PL was only detected when model contrasts were fitted to data generated with realistic evolutionary dynamics using the MutSel model framework.
> To illustrate the impact of PL, we consider the case of CSMs modified to detect the fixation of codons following simultaneous double and triple (DT) nucleotide mutations. The majority of
CSMs currently in use assume that codons evolve by a series of single-nucleotide substitutions, with the probability for DT changes set to zero. However, recent model-based analyses have uncovered evidence for DT mutations [32, 68, 83]. Early estimates of the percentage of fixed mutations that are DT were perhaps unrealistically high. Kosiol et al. [32], for example, estimated a value close to 25% in an analysis of over 7000 protein families from the Pandit database [69]. Alternatively, when estimates were derived from a more realistic site-wise mutation-selection model, DT changes comprised less than 1% of all fixed mutations [64]. More recent studies suggest modest rates of between 1% and 3% [5, 20, 27, 53]. Whatever the true rate, several authors have argued that it would be beneficial to introduce a few extra parameters into a standard CSM to account for DT mutations (e.g., [40, 83]). The problem with this suggestion is that episodic fixation of DT mutations can produce signatures of heterotachy consistent with shifting balance.
Recall the comparison of M1, a CSM containing parameters represented by the vector θ1, and M2, the same model but for the inclusion of one additional parameter <sup>ψ</sup>, so that <sup>θ</sup><sup>2</sup> <sup>¼</sup> (θ1, <sup>ψ</sup>). The parameter ψ will reduce the deviance of M2 compared to M1 by some proportion of the baseline deviance between the simplest CSM (M0) and the saturated modelPSð<sup>θ</sup> ^SÞ. We call this the percent reduction in deviance (PRD) attributed to ψ^:
$$\text{PRD}(\hat{\boldsymbol{\psi}}) = \frac{\Delta D(\hat{\theta}\_{\text{M1}}, \hat{\theta}\_{\text{M2}})}{\Delta D(\hat{\theta}\_{\text{M0}}, \hat{\theta}\_{\text{S}})} \tag{6}$$
Suppose M1 and M2 were fitted to an alignment and that the LLR ¼ <sup>Δ</sup>Dð<sup>θ</sup> ^M1, θ ^M2Þ was found to be statistically significant. This would lead an analyst to attribute the PRDðψ^Þ to real signal for the process ψ was meant to represent, possibly combined with some PL and noise. Now, consider the case in which the process represented by ψ did not actually occur (i.e., it was not a component of the true generating process). Under this scenario, PRDðψ^Þ would contain no signal, but would be entirely due to PL plus noise. When this is known to be the case, we set PRDðψ^Þ ¼ PLðψ^Þ. As illustrated below, PLðψ^Þ can be large enough to result in rejection of the null, and therefore lead to a false conclusion about the data generating process.
We illustrate PL by contrasting the model RaMoSS with a companion model RaMoSSwDT that accounts for the fixation of DT mutations via two rate parameters, α (the double mutation rate) and β (the triple mutation rate) [26]. RaMoSS combines the standard M-series model M3 with the covarion-like model CLM3 (cf., [12, 18]). Specifically, RaMoSS mixes (with proportion pM3) one model with two rate-ratio categories ω<sup>0</sup> < ω<sup>1</sup> that are constant over the entire tree with a second model (with proportion
Fig. 4 The box plot depicts the distribution of the phenomenological load (PL) carried by ðα^ , <sup>β</sup>^ Þ produced by fitting the RaMoSS vs RaMoSSwDT contrast to 50 alignments generated under MutSel-mmtDNA: the circles represent outliers of this distribution. The diamond is the percent reduction in deviance for the same parameters estimated by fitting RaMoSS vs RaMoSSwDT to the real mtDNA alignment
<sup>p</sup>CLM3 <sup>¼</sup> <sup>1</sup> <sup>p</sup>M3) under which sites switch randomly in time between ω<sup>0</sup> <sup>0</sup> < ω<sup>0</sup> <sup>1</sup> at an average rate of δ switches per unit branch length. Fifty alignments were simulated to mimic a real alignment of 12 concatenated H-strand mitochondrial DNA sequences (3331 codon sites) from 20 mammalian species as distributed in the PAML package [73]. The generating model, MutSel-mmtDNA [26], was based on the mutation-selection framework and produced alignments with single-nucleotide mutations only. Since DT mutations are not fixed under MutSel-mmtDNA, the PRD carried by ðα^, <sup>β</sup>^Þ in each trial can be equated to PL (plus noise). The resulting distribution of PLðα^, <sup>β</sup>^Þ is shown as a boxplot in Fig. 4.
Although DT mutations were not fixed when the data was generated, shifting balance on a static landscape can produce similar site patterns as a process that includes rare fixation of DT mutations (site patterns exhibiting both synonymous and nonsynonymous substitutions; [26]).<sup>3</sup> DT and shifting balance are therefore confounded. And since shifting balance tends to occur at a substantial proportion (approximately 20%) of sites when an alignment is generated under MutSel-mmtDNA, DT mutations were falsely inferred by the LRT in 48 of 50 trials at the 5% level of significance (assuming LLR χ<sup>2</sup> <sup>2</sup> for the two extra parameters α and β in RaMoSSwDT compared to RaMoSS). The PRD ðα^, <sup>β</sup>^Þ when RaMoSS vs RaMoSSwDT was fitted to the real mmtDNA is
<sup>3</sup> It has previously been noted that the rapid fixation of compensatory mutations following substitution to an unstable base pair (e.g., AT!GT!GC) can also produce site patterns that suggest fixation of DT mutations [74, p. 46].
shown as a diamond in the same plot. Although ðα^, <sup>β</sup>^Þ estimated from the real mmtDNA were found to be highly significant (LLR ¼ 84, <sup>p</sup>-value << 0.001), the PRDðα^, <sup>β</sup>^Þ was found to be just under the 95th percentile of PLðα^, <sup>β</sup>^Þ (PRD ¼ 0.060% compared to the 95th percentile of PL ¼ 0.061). The evidence for DT mutations in the real data is therefore only marginal, and it is reasonable to suspect that its PRDðα^, <sup>β</sup>^Þ, if not entirely the result of PL, is at least partially caused by PL.
#### 5 Discussion
CSMs have been subjected to a certain degree of censure, particularly during Phase I of their development [11, 22, 23, 46, 60–63, 85]. We maintain that it is not the model in and of itself, or the maximum likelihood framework it is based on, that gives rise to statistical pathologies, but the relationship between model and data. This principle was illustrated by our analysis of the history of CSM development, which we divided into two phases. Phase I was characterized by the formulation of models to account for differences in selection effects across sites and over time that comprise the major component of variation in an alignment. Starting with M0, such models represent large steps toward the fitted saturated model in Fig. 2, and also provide a better representation of the true generating process. The main criticism of Phase I models was the possibility of falsely inferring positive selection in a gene or at an individual codon site [62, 63, 85]. But, the most compelling empirical case of false positives was shown to be the result of inappropriate application of a complex model to a sparse alignment [63]. Methods for identifying (bootstrap) and dealing with (BEB, SBA, and PLRT) low information content were illustrated in Case Study A.
The other big concern that arose during Phase I development was the possibility of pathologies associated with model misspecification. The method used to identify such problems was to fit a model to alignments generated under a scenario contrived to be challenging, as illustrated in Case Study B. There, the omnibus test based on Model A of the YN-BSM was shown to result in an excess of false positives when fitted to alignments simulated using the implausible but difficult "XZ" generating scenario (e.g., with complete relaxation of selection pressure at all sites on one branch of the tree; Table 1). Subsequent modifications to the test reduced the false positive rate to acceptable values. Hence, Case Study B underlines the importance of the model–data relationship. However, it is not clear whether a model adjusted to suit an unrealistic datagenerating process is necessarily more reliable when fitted to a real alignment. This difficulty highlights the need to find ways, for the purpose of model testing and adjustment, to generate alignments that mimic real data as closely as possible.
Confidence in the CSM approach, combined with the exponential increase in the volume of genetic data and the growth of computational power, spurred the formulation CSMs of everincreasing complexity during Phase II. The main issue with these models, which has not been widely appreciated, is confounding. Two processes are confounded if they can produce the same or similar patterns in the data. It is not possible to identify such processes when viewed through the narrow lens of an alignment (i.e., site patterns) alone. This was illustrated by Case Study C, where shifting balance on a static landscape was shown to be confounded with episodic adaptive evolution [7, 25]. Confounding can lead to what we call phenomenological load, as demonstrated in Case Study D. In that analysis, the parameters (α, β) were assigned a specific mechanistic interpretation, the rate at which double and triple mutations arise. It was shown that (α, β) can absorb variations in the data caused by shifting balance; hence, the MLEs ðα^, <sup>β</sup>^Þ resulted in a significant reduction in deviance in 48/50 trials (Fig. 4), and therefore improved the fit of the model to the data. However, the absence of DT mutations in the generating process invalidated the intended interpretation of ðα^, <sup>β</sup>^Þ. This result underlines that a better fit does not imply a better mechanistic representation of the true generating process.
It is natural to assume that a better mechanistic representation of the true generating process can be achieved by adding parameters to our models to account for more of the processes believed to occur. The problem with this assumption is that the metric of model improvement under ML (reduction in deviance) is independent of mechanism. A parameter assigned a specific mechanistic interpretation is consequently vulnerable to confounding with other processes that can produce the same distribution of site patterns. As CSMs become more complex, it seems likely that the opportunity for confounding will only increase. It would therefore be desirable to assess each new model parameter for this possibility using something like the method shown in Fig. 4 whenever possible. The idea is to generate alignments using MutSel or some other plausible generating process in such a way as to mimic the real data as closely as possible, but with the new parameter set to its null value. To provide a second example, consider the test for changes in selection intensity in one clade compared to the remainder of the tree known as RELAX [67]. Under this model, it is assumed that each site evolved under a rate ratio randomly drawn from <sup>ω</sup><sup>R</sup> <sup>¼</sup> { ω1, ..., ωk} on a set of prespecified reference branches, and from a modified set of rate ratios <sup>ω</sup><sup>T</sup> ¼ fω<sup>m</sup> <sup>1</sup> , ... ,ω<sup>m</sup> <sup>k</sup> <sup>g</sup> on test branches, where m is an exponent. A value 0 < m < 1 moves the rate ratios in ωT closer to one compared to their corresponding values in ωR, consistent with relaxation of selection pressure at all sites on the test branches. Relaxation is indicated when the contrast of the null hypothesis that <sup>m</sup> ¼ 1 versus the alternative that <sup>m</sup> <sup>&</sup>lt; 1 is statistically significant. The distribution of PL(m^ ) can be estimated from alignments generated with <sup>m</sup> ¼ 1. The PRD(m^ ) estimated from the real data can then be compared to this to assess the impact of PL (cf. Fig. 4). This approach is predicated on the existence of a generating model that could have plausibly produced the site patterns in the real data. Jones et al. [26] present a variety of methods for assessing the realism of a simulated alignment, although further development of such methods is warranted. Software based on MutSel is currently available for generating data that mimic large alignments of 100-plus taxa (Pyvolve; [56]). Other methods have been developed to mimic smaller alignments of certain types of genes (e.g., MutSel-mmtDNA; [25]). It is only by the use of these or other realistic simulation methods that the relationship between a given model and an alignment can be properly understood.
#### References
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
#### Evolution of Viral Genomes: Interplay Between Selection, Recombination, and Other Forces
#### Stephanie J. Spielman, Steven Weaver, Stephen D. Shank, Brittany Rife Magalis, Michael Li, and Sergei L. Kosakovsky Pond
#### Abstract
Natural selection is a fundamental force shaping organismal evolution, as it both maintains function and enables adaptation and innovation. Viruses, with their typically short and largely coding genomes, experience strong and diverse selective forces, sometimes acting on timescales that can be directly measured. These selection pressures emerge from an antagonistic interplay between rapidly changing fitness requirements (immune and antiviral responses from hosts, transmission between hosts, or colonization of new host species) and functional imperatives (the ability to infect hosts or host cells and replicate within hosts). Indeed, computational methods to quantify these evolutionary forces using molecular sequence data were initially, dating back to the 1980s, applied to the study of viral pathogens. This preference largely emerged because the strong selective forces are easiest to detect in viruses, and, of course, viruses have clear biomedical relevance. Recent commoditization of affordable high-throughput sequencing has made it possible to generate truly massive genomic data sets, on which powerful and accurate methods can yield a very detailed depiction of when, where, and (sometimes) how viral pathogens respond to various selective forces.
Here, we present recent statistical developments and state-of-the-art methods to identify and characterize these selection pressures from protein-coding sequence alignments and phylogenies. Methods described here can reveal critical information about various evolutionary regimes, including whole-gene selection, lineage-specific selection, and site-specific selection acting upon viral genomes, while accounting for confounding biological processes, such as recombination and variation in mutation rates.
Key words Virus evolution, Molecular evolution, Recombination, Positive selection, Relaxed selection, Phylogenetics, Codon models
#### 1 Introduction
Natural selection is a powerful evolutionary force that shapes genomes of all living organisms. A variety of computational approaches have been developed to measure the strength and direction of selection directly from genomic data. Given an alignment of homologous gene sequences, the strength of natural selection acting on a given gene or genes can be measured in a phylogenetic
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_14, © The Author(s) 2019
context using codon models [1, 3]. A typical analysis on viral genomes, for example, might be performed for a single gene represented by isolates from different individuals (e.g., sequences from many HIV-1-infected hosts) or from different hosts (e.g., primate lentiviruses).
In the context of codon models, selection is typically measured using dN/dS (also referred to as ω, or Ka/Ks), which represents the ratio of the non-synonymous evolutionary rate (dN) to the synonymous evolutionary rate (dS). The synonymous evolutionary rate is used to provide a baseline rate of neutral evolution because the average selective effect of a synonymous substitution is assumed to be negligible compared to the effect of a non-synonymous substitution. <sup>1</sup> The selective regime can be deduced by establishing, with a degree of statistical confidence, that dN/dS differs from unity, i.e., the neutral expectation where dN/dS ¼ 1. Diversifying, balancing, or (sometimes) directional selection yields dN/dS > 1, whereas purifying selection effects dN/dS < 1. Comparative methods for selection detection estimate dN/dS, or dS and dN separately, at sites and/or branches and perform a statistical test to establish on which side of the neutral expectation the inferences fall. As with any statistical procedure applied to finite data, each inference can be a false positive or a false negative, although methods typically take care to control the rates of both.
While the question "Is this gene under selection?" is an obvious one, the nearly universally applicable answer to this question is "yes.'. That is because a functional gene is (or has been) subject to some form of selection, e.g., negative selection to maintain essential features. On the other extreme is the question that has an immediate biological significance: "Is changing a leucine to an arginine at position 209 in gene X along a specific branch in the phylogeny adaptive?". Without additional information, such as a carefully experimentally measured fitness impact of introducing said substitution, current comparative sequence approaches cannot answer this question. Indeed, such a scenario presents a sample size of one, which cannot be statistically meaningful.
In this chapter, we present a collection of statistical methods, each of which is designed to carefully address a biological question somewhere on the spectrum between the two extremes: sufficiently specific to be interesting, yet general enough to be answerable based only on the evolutionary history of homologous sequences. We will not discuss the technical details of codon substitution methods here (for details, please see one of the excellent available reviews Anisimova and Kosiol [1], Delport et al. [3], Yang [44], or the primary methods papers including Goldman and Yang [8],
<sup>1</sup> We note that there are a variety of well-documented situations where synonymous substitutions can have strong effects on fitness [11, 30].
Kosakovsky Pond and Muse [13], Muse and Gaut [27], Nielsen and Yang [29]). Instead, we present each method operationally ("How and when does one use this method?"), by addressing the following points:
We conclude by discussing how inferences can accommodate potentially confounding biological processes, including intragenic recombination and mutation rate variation. It is critical to model these processes, both in their own right and because ignoring their effects could bias selection inference tools and yield misleading results.
#### 2 Materials
The data used throughout the following tutorials and exercises are available at https://github.com/veg/evogenomics\_hyphy. A "README" file in the top directory of this repository provides a detailed description of all contents. Importantly, all datasets used here reside in the datasets directory. Please refer to http://www.hyphy.org for instructions on downloading and installing HyPhy to your system. All exercises have been validated using version 2.3.4. Throughout, we will use the hyphymp executable (MP ¼ multiprocessor). For all analyses, you will need the following information:
All methods will produce a final file of results in JSON (Java-Script Object Notation) format, a highly extensible format that is simple, relatively compact, and both machine- and humanreadable. JSON output files can be visually and interactively examined within our new web application, hyphy-vision, accessible at vision.hyphy.org.
All methods employ the general time reversible (GTR) nucleotide model for initial branch length optimization and correcting nucleotide substitution biases, followed by fitting a Muse–Gaut model (with general time reversible nucleotide biases) to obtain preliminary dN/dS estimates (see Kosakovsky Pond and Frost [12] for a detailed model description) for selection inference. Codon frequencies are estimated using the CF3x4 procedure [15]. In our view, the historical rationale for using simpler evolutionary models (e.g., K80, F81, or HKY85), namely, computational cost, to fit nucleotide data is no longer relevant.
Finally, we recommend different P-value thresholds depending on the given analysis method. As site-level methods (FEL, SLAC, and MEME) tend to be conservative on biological data, we recommend significance as P - 0.1 (or posterior probability 0.9 for FUBAR). By contrast, we recommend significance as P - 0.05 for alignment-wide methods—BUSTED, RELAX, and aBSREL.
#### 3 Methods
#### 3.1 How to Run a Selection Analysis
There is a uniform workflow to run any of the described methods, either locally (on one's own computer and/or a high-performance computing environment) in HyPhy or using the Datamonkey web-service, available at www.datamonkey.org. The version of HyPhy that supports all of the analyses is a command-line program, i.e., it must be run from a terminal prompt (similar to most other bioinformatics packages) in Linux or Mac OS X. It is also possible to run the program in Windows, with an appropriate POSIX emulation environment (e.g., MinGW) installed.
To execute a selection analysis locally, the following steps will need to be taken.
- <sup>l</sup> Two separate files containing the alignment and phylogeny, respectively. In this circumstance, HyPhy issues two successive prompts: the first for the file containing the alignment, and the second for the file containing the tree.
- <sup>l</sup> A single file containing an alignment in one of the formats supported by HyPhy (FASTA, MEGA, and PHYLIP), with a Newick-formatted phylogeny included at the bottom of this file. In this circumstance, HyPhy issues two successive prompts: the first for the file containing the alignment, and the second asking whether to accept the tree found in the file (provide the affirmative response, e.g., "y," to accept it).
- <sup>l</sup> A NEXUS file containing both the alignment and phylogeny. In this circumstance, HyPhy automatically accepts the provided phylogeny and therefore only issues a single prompt for the file containing the alignment. This is also the format that can be used to specify partitioned data, which is necessary to account for recombination.
Each method will provide live on-the-screen progress updates and, when finished, a text summary of the analysis. The output is generated in Markdown,2 which can either be read directly as text or formatted using one of many Markdown viewers.
When an analysis is finished, HyPhy will write a JSON file with numerous details about the analysis to disk. By convention, this file will be placed in the same directory as the input alignment file, with the added <method>.json extension, e.g., flu\_ha.nex. BUSTED.json for an input alignment named flu\_ha.nex analyzed by the method BUSTED. All results contained in this JSON file can be explored visually within a web browser using a web application from the hyphy-vision suite of tools, accessible at vision.hyphy.org. Since JSON files can be easily accessed by
<sup>2</sup> With the exception of GARD.
scripting and data-analysis languages, these are also well-suited for incorporation into pipelines.<sup>3</sup>
When run through www.datamonkey.org, this entire workflow is automated: one simply uploads an alignment, selects options for the analysis, and waits for the job to finish. Once the job has completed, the results will be displayed in an interactive application within the web browser. Note that Datamonkey will automatically remove duplicate sequences before executing any analysis.
#### 3.2 BUSTED
#### What Biological Question Is the Method Designed to Answer?:
Is there evidence that some sites in the alignment have been subject to positive diversifying selection, either pervasive (throughout the evolutionary tree) or episodic (only on some lineages)? In other words, BUSTED asks whether a given gene has been subject to positive, diversifying selection at any site, at any time [26]. If a priori information about lineages of interest is available (e.g., due to migration, change in the environment, etc.), then BUSTED can be restricted to test for selection only on a subset of tree lineages, potentially boosting power.
#### Recommended Applications
#### Statistical Test Procedure:
Each (branch, site) pair evolves with ω<sup>1</sup> ω<sup>2</sup> - 1, or ω<sup>3</sup> 1, with the ratio chosen independently of other (branch, site) pairs with probability p1, p2, p<sup>3</sup> (normalized to sum to 1). The threerate ω distribution is estimated jointly from the entire alignment, i.e., rates are shared by all (branch,site) combinations. Therefore, BUSTED is technically a "branch-site" model [16], although it is not intended to detect individual sites which drive signal of selection.
<sup>3</sup> Note that the method GARD does not provide markdown output or a JSON, and output is in a different format. This may be updated in a future HyPhy release.
The test for episodic diversifying selection is performed by comparing the full model versus the nested null model, where ω<sup>3</sup> is constrained to 1. Statistical significance is obtained by the likelihood ratio test, assuming the χ<sup>2</sup> <sup>2</sup> asymptotic distribution of the likelihood ratio statistic under the null model.
When only some of the branches are chosen for testing, and the remainder are designated as the background, two independent three-rate ω distributions are fitted: one for the test branches, and one for the background branches. Testing for selection is carried out by constraining the distribution on the test branches as described above.
#### Example Analysis:
To begin, we will perform a BUSTED analysis using a dataset of primate-specific KSR2, kinase suppressor of RAS2, genes from Enard et al. [5]. This gene has been implicated as a so-called 'virus-interacting protein,' and previous work has suggested it has experienced adaptation in mammalian lineages due to selective pressures exerted by viruses [5]. We will test all lineages for positive selection (rather than specifying a subset of "test" branches), thereby asking the question: "Has KSR2 been subject to diversifying selection at some time during evolution in primates?"
To run BUSTED, open a terminal session and enter HYPHYMP from the command line to launch the HyPhy analysis menu. Enter 1 (Selection Analyses) and then 5 to reach the BUSTED analysis menu, and supply values for the following prompts:
BUSTED will now run to completion, printing status indicators to screen while it runs. For an example of how this output will look when rendered into HTML (or similarly, PDF), see this link: http://bit.ly/2vsRZrh.
#### Listing 1 Partial BUSTED screen output:
### Branches to test for selection in the BUSTED analysis
\* Selected 15 branches to test in the BUSTED analysis: 'HUM, PAN, Node6, GOR, Node5, PON, Node4, GIB, Node3, MAC, BAB, Node12, Node2, MAR, BUS' ### Obtaining branch lengths and nucleotide substitution biases under the nucleotide GTR model \* Log ( L ) = -5768.01, AIC - c = 11582.06 (23 estimated parameters) ### Obtaining the global omega estimate based on relative GTR branch lengths and nucleotide substitution biases \* Log ( L ) = -5342.48, AIC - c = 10745.17 (30 estimated parameters) \* non - synonymous / synonymous rate ratio for \*test\* = 0.0342 ### Improving branch lengths , nucleotide substitution biases, and global dN/dS ratios under a full codon model \* Log ( L ) = -5333.46, AIC - c = 10727.13 (30 estimated parameters) \* non - synonymous / synonymous rate ratio for \*test\* = 0.0307 ### Performing the full ( dN / dS > 1 allowed) branch-site model fit \* Log ( L ) = -5319.67, AIC - c = 10707.62 (34 estimated parameters) \* For \* test \* branches , the following rate distribution for branch-site combinations was inferred | Selection mode | dN/dS |Proportion, %| Notes | |---------------------------------- |-------------- |-------------|----------------– | | Negative selection | 0.024 | 99.151 | | | Negative selection | 0.085 | 0.812 | | | Diversifying selection | 118.143 | 0.037 | | ### Performing the constrained (dN/dS > 1 not allowed) model fit \* Log ( L ) = -5326.18, AIC - c = 10718.63 (33 estimated parameters) \* For \* test \* branches under the null (no dN/dS > 1 model), the following rate distribution for branch-site combinations was inferred | Selection mode | dN/dS |Proportion, %| Notes | |-----------------------------|--------------|-------------|---------------------––| | Negative selection | 0.000 | 10.598 | | | Negative selection | 0.000 | 86.086 | Collapsed rate class | | Neutral evolution | 1.000 | 3.316 | | ---- ## Branch - site unrestricted statistical test of episodic diversification [BUSTED]
Likelihood ratio test for episodic diversifying positive selection, \*\*p = 0.0015\*\*.
#### Interpreting Results:
The results printed to the terminal indicate a highly significant result (P ¼ 0.0015) in the test for whole-gene selection. Analysis with BUSTED therefore provides robust evidence that KSR2 experienced episodic positive selection in the primates. Because we performed the original BUSTED analysis on the entire tree (i.e., without a specified set of test branches), we do not know from this result along which lineages KSR2 was subject to positive selection. We can conclude only that a non-zero proportion of sites on some lineage(s) in the primate tree experienced diversifying selection pressure.
The output additionally provided information about the specific BUSTED model fits to the test data, including the inferred ω distributions and corresponding weights. The BUSTED alternative model (shown under the output header Performing the full ( dN/dS >1 allowed) branch-site model fit) found that a very small proportion (only 0.037%) of sites evolved under a very large ω of over 100 (118.143 ). Importantly, neither of these estimates is precise because they were derived from a small subset of the data. As such, all the BUSTED tests establish the fact that the proportion of sites along test lineages (here, the entire phylogeny) with ω > 1 is non-zero. For example, if BUSTED had inferred a rate category of ω ¼ 10 on a different gene, it would not be correct to claim that this gene evolves under weaker selection than does KSR2. A formal statistical test would have to be carried out to establish such a claim.
Conversely, had the result not been statistically significant, we would not be able to reject the null hypothesis that no positive selection had occurred in KSR2. Importantly, however, a negative finding would not unequivocally rule out the presence of positive selection. This outcome could be due to a lack of statistical power wherein the provided data did not contain a sufficiently strong selection.
BUSTED's fixed a priori assumption of model complexity (a three-rate ω distribution) may lead to over-parameterized (or under-parameterized) models. For example, in the constrained model for KSR2, two of the three rate classes have the same value of ω(0.0), implying that one of them is unnecessary. HyPhy will report this to the screen as a diagnostic message Collapsed rate class, but there is no corrective action that needs to be taken. These messages simply point to low-complexity data.
We will additionally take this opportunity to showcase the visual power of our accompanying web browser, HyPhy-Vision. Figure 1 displays the rendering of the output ksr2.fna. BUSTED.json as it appears in HyPhy-Vision. On this site, users can interactively view and explore inference results, view figures and charts, and perform other tasks.
Fig. 1 Example analysis visualization in HyPhy-Vision of BUSTED results. (a) The summary section provides a brief overview of the analysis performed, including information about the inputted data (which can be downloaded via the linked file name) and primary results from the hypothesis test performed. (b) The model statistics section provides information about models fitted to the data. In BUSTED, this section additionally includes an interactive display of site evidence ratios, which can be interpreted as a descriptive measure for which sites may have contributed to the selection signal. (c) The tree section displays the phylogeny as fitted under all inferred models and data partitions, if specified. Tree views can be toggled under the Options drop-down menu. (d) Graphical views of each model's inferred ω distribution can be viewed when clicking on a given row's plot icon in the Model fits table seen in (b)
Rules of Thumb for BUSTED Use
#### 3.3 RELAX
#### What Biological Question Is the Method Designed to Answer?:
Is there evidence that the strength of selection has been relaxed (or conversely intensified) on a specified group of lineages (Test) relative to a set of reference lineages (Reference)? We note that the RELAX framework can perform this specific hypothesis test as well as fit a suite of descriptive models which address, for example, overall rate differences between test and reference branches or lineage-specific inferences of selection relaxation. We focus our attention here on RELAX's hypothesis testing abilities. More information about descriptive analyses is available on hyphy.org as well as in RELAX's primary publication [43]. Importantly, RELAX is not designed to detect diversifying selection specifically.
#### Recommended Applications
#### Statistical Test Procedure:
Given a tree with at least two sets of branches, one of which is designated as Test, and the other as Reference, the core version of RELAX compares two nested models, which follow the same general framework as BUSTED. Each (branch, site) combination is drawn independently from a 3-rate ω distribution. The evolutionary rates for Test branches are functions of those for Reference branches. Specifically, <sup>ω</sup>Test <sup>¼</sup> <sup>ω</sup><sup>K</sup> Reference, where K is the relaxation or intensification parameter. The alternative model infers K from the data, and the null model sets K ¼ 1. Statistical significance is obtained by the likelihood ratio test, assuming the χ<sup>2</sup> <sup>1</sup> asymptotic distribution under the null model. A significant result of K > 1 indicates that selection strength has been intensified along the test branches, and a significant result of K < 1 indicates that selection strength has been relaxed along the test branches. In other words, for K < 1 the Test ω values shrink toward neutrality (ω ¼ 1) relative to Reference, and for K > 1 they move away from neutrality.
If some branches in the tree belong to neither the Test or the Reference set, they are allocated to a group with its own (Unclassified) distribution of ω, which is uncoupled from the testing procedure.
#### Example Analysis:
We will perform a RELAX analysis using a dataset of Influenza A PB2 subunit sequences from Tamuri et al. [41]. The PB2 subunit, which is part of influenza's RNA polymerase complex, has emerged as a critical determinant of influenza infectivity and, as a consequence, host range [9, 18]. The dataset we examine here contains sequences from both avian host and human host strains.<sup>4</sup> Previous studies have shown that this host switch is correlated with significant shifts in selection pressures and preferred amino acids at key sites in PB2 [36, 40, 41]. We now re-analyze this dataset using RELAX to ask a different but related question: "Was the shift from avian to human hosts associated with a relaxation of selection pressures in Influenza A PB2?"
RELAX requires an a priori specification of test and reference lineages, although not all lineages in a tree need to be classified. As such, you must label your test (and reference, if desired) branches in the input phylogeny. We provide an online widget to assist with tree labeling at http://phylotree.hyphy.org. The dataset we have provided for this analysis already has a labeled phylogeny, with the human host lineages labeled as "test."
To run RELAX, open a terminal session and enter HYPHYMP from the command line to launch the HyPhy analysis menu. Enter 1 (Selection Analyses) and then 7 to reach the RELAX analysis menu, and supply values for the following prompts:
<sup>4</sup>The original dataset in Tamuri et al. [41] contained 401 sequences. For the purposes of this chapter, we analyze a subset of this alignment with only 35 sequences (20 from avian and 15 from human hosts), thereby achieving a tractable runtime on a personal machine.
Enter 1 to select the branches labeled as "test" as the test set in RELAX analysis. Note that when multiple labels are present in your tree, HyPhy will issue an additional prompt to choose the set of Reference branches, in the event that some branches should remain Unclassified.
5. Analysis type. This option asks you to specify the scope of RELAX analysis. Selecting "Minimal" will run the RELAX hypothesis test, and selecting "All" will run hypothesis testing and fit two additional descriptive models, described earlier. Here, we will perform only hypothesis testing to determine whether the data shows evidence for a relaxation or intensification of selection intensity between the test and reference lineages. Enter the option 2 to run the "Minimal" analysis.
RELAX will now run to completion, printing status indicators to screen while it runs.
#### Listing 2 Partial RELAX screen output:
```
### Obtaining branch lengths and nucleotide substitution biases under the
nucleotide GTR model
* Log ( L ) = -16755.26, AIC - c = 33660.66 (75 estimated parameters)
### Obtaining the global omega estimate based on relative GTR branch lengths
and nucleotide substitution biases
* Log ( L ) = -14410.97, AIC - c = 28988.46 (83 estimated parameters)
* non - synonymous / synonymous rate ratio for *Reference* = 0.0401
* non - synonymous / synonymous rate ratio for *Test* = 0.0604
### Improving branch lengths , nucleotide substitution biases, and global dN/dS
ratios under a full codon model
* Log ( L ) = -14354.67, AIC - c = 28875.86 (83 estimated parameters)
* non - synonymous / synonymous rate ratio for *Reference* = 0.0358
* non - synonymous / synonymous rate ratio for *Test* = 0.0609
### Fitting the alternative model to test K != 1
* Log ( L ) = -14337.22, AIC - c = 28849.02 (87 estimated parameters)
* Relaxation / intensification parameter (K) = 0.73
* The following rate distribution was inferred for **test** branches
```
440 Stephanie J. Spielman et al.
\* The following rate distribution was inferred for \*\*reference\*\* branches
### Fitting the null ( K := 1) model
\* Log ( L ) = -14342.33, AIC - c = 28857.22 (86 estimated parameters) \* The following rate distribution for test/reference branches was inferred
## Test for relaxation ( or intensification) of selection [RELAX] Likelihood ratio test \*\* p = 0.0014\*\*. > Evidence for \* relaxation of selection\* among \*\*test\*\* branches \_relative\_ to the \*\*reference\*\* branches at P<=0.05 ----
#### Interpreting Results:
On this data, RELAX has inferred a relaxation parameter K ¼ 0.73 with a highly significant P ¼ 0.0014. Therefore, there is evidence to reject the null hypothesis that selection pressure has not been shifted in the test (here, human host) lineages. We instead have strong evidence that selection has been relaxed (because the inferred K < 1) in the human host lineages. In other words, selection in the test branches has generally moved towards neutrality (ω ¼ 1) compared to the reference branches. This finding is consistent with the evolutionary changes that typically occur during a virus host-switching event, wherein selection stringency will be reduced to facilitate viral adaptation.
Keep in mind that RELAX defines relaxation (or intensification) in a fairly restrictive fashion. In other words, all selective regimes (i.e., all ω rates), both negative and positive, must weaken or strengthen. Therefore, certain relaxation scenarios, for example, when only positive selection is relaxed but negative selection is maintained, may result in a non-significant RELAX test even though selection has changed.
Rules of Thumb for RELAX Use
Since then, we have developed a greatly improved branch-site model called aBSREL ("adaptive branch-site random effects likelihood"). Rather than assuming that each branch should be fit with three rate classes, aBSREL infers, using small-sample Akaike Information Criterion correction (AICc), the optimal number of rate categories per branch. In this manner, computational complexity and the number of parameters are greatly reduced, leading to a tractable runtime for larger datasets that could not otherwise be studied with earlier branch-site models.
#### What Biological Question Is the Method Designed to Answer?:
Like classical branch-site models, aBSREL asks whether some proportion of sites is subject to positive selection along specific branches or lineages of a phylogeny.
Recommended Applications
1. Exploratory testing for evidence of lineage-specific positive diversifying selection in small- to medium-sized alignments (up to 100 sequences).
2. Targeted testing of branches selected a priori for positive diversifying selection. This includes alignments with prohibitive runtimes under older branch-site models (up to 1000 sequences) [37].
#### Statistical Test Procedure:
aBSREL uses the information-theoretic criterion AIC<sup>c</sup> to automatically determine the complexity of the evolutionary process at every branch [37]. As a heuristic optimization, aBSREL will always examine branches in order from longest to shortest, because longer branches tend to be the ones requiring more complex models. In this adaptive model, one rate class is allowed to assume any value of ω > 1, whereas for any other inferred rate class is constrained as ω - 1. In the null model, all ω categories are constrained as ω - 1. For any branch inferred to have sufficient rate variation (i.e., more than one rate category) where one rate category is described by ω > 1, aBSREL will proceed to fit a null model to this branch. In other words, if the maximum-inferred ω - 1 on a branch, the null model will have the same exact fit as the alternative model, and the resulting P-value is 1. The test for lineage-specific diversifying selection is performed by comparing the full model versus the nested null model, and statistical significance is obtained by the likelihood ratio test. Significance is evaluated using a mixture of 50%χ<sup>2</sup> <sup>0</sup>, 20%χ<sup>2</sup> <sup>1</sup>, and 30%χ<sup>2</sup> <sup>2</sup> distributions (proportions determined via simulations Smith et al. [37]). Finally, aBSREL will correct all P-values obtained from individual tests for multiple comparisons using the Bonferroni–Holm procedure to control family-wise false-positive rates (i.e., the probability of generating one or more false positives, when all null hypotheses are correct).
One can either select a specific set of branches in order to test a specific a priori hypothesis or one can perform an exploratory analysis across the entire phylogeny by testing all branches for selection. The former approach may have substantially more power to detect selection, especially if only a few branches in a large tree are chosen, due to the decreased volume of multiple testing. However, the approach does carry the risk of failing to identify branches subject to positive selection that have not been included in the test set.
#### Example Analysis:
Here, we will demonstrate aBSREL use and interpretation using a dataset of HIV-1 env sequences collected from an epidemiologically linked donor–recipient transmission pair [7]. This dataset can be found in the provided file hiv1\_transmission. fna.
To run aBSREL, open a terminal session and enter HYPHYMP from the command line to launch the HyPhy analysis menu. Enter 1 (Selection Analyses) and then 6 to reach the aBSREL analysis menu, and supply values for the following prompts:
aBSREL will now run to completion, printing status indicators to screen while it runs (some output abbreviated).
#### Listing 3 Partial aBSREL screen output:
### Obtaining branch lengths and nucleotide substitution biases under the nucleotide GTR model \* Log ( L ) = -5524.50, AIC - c = 11153.08 (52 estimated parameters) ### Fitting the baseline model with a single dN/dS class per branch, and no site-to-site variation.
\* Log ( L ) = -5402.40, AIC - c = 11009.72 (102 estimated parameters)
\* Branch - level non - synonymous / synonymous rate ratio distribution has median 0.66, and 95% of the weight in 0.00--5.41
### Determining the optimal number of rate classes per branch using a step up procedure
### Adaptive branch site random effects likelihood test Likelihood ratio test for episodic diversifying positive selection at Holm-Bonferroni corrected \_p = 0.0500\_ found \*\*3\*\* branches under selection among \*\*44\*\* tested.
#### Interpreting Results:
The first printed markdown table ("Determining the optimal number of rate classes per branch using a step up procedure") summarizes the model selection process. For example, when two ω rates were assigned to branch Separator, this improved the AICc score of the fit (compared to the single-rate model) from 11, 009.72 to 11, 004.02. However, allocating three ω rates to the same branch worsens the score to 11, 008.06. Therefore the aBSREL model will use two ω rates at the branch.
The second printed markdown table ("Testing selected branches for selection") shows the results of tests for episodic selection on individual branches. At branch 0564\_4, for example, the tested model includes two ω rates, with the positive selection class taking on value 28.89 (2.15% proportion of the mixture). Constraining this rate to range between 0 and 1 yields the likelihood ratio test statistic of 4.81, which maps to a P-value (before multiple test correction) of 0.03281.
Finally, aBSREL reports three branches under episodic diversifying selection pressure. Further examination of results using HyPhy-Vision shows that these branches are found (a) along the transmission event from donor to recipient, and (b) within a highly diverged clade in the donor (Fig. 2). The first finding is consistent with an expected increase in evolutionary rate when a virus infects a new host and encounters novel host immunity, and the second finding is consistent with intrahost adaptive dynamics of the donor's long-term HIV infection. Importantly, a close examination of the markdown-output table under the header "Testing selected branches for selection" reveals several nodes with uncorrected P-values whose significance was lost upon applying the Bonferroni–Holm correction, e.g., 0564\_4 whose uncorrected P ¼ 0.03281. This result illustrates the potential loss of power incurred by this aBSREL exploratory analysis.
Rules of Thumb for aBSREL Use
Fig. 2 HyPhy-Vision tree viewer depicting the fitted aBSREL Adaptive model to HIV-1 data. Branches are colored by their inferred ω distribution, as indicated in the legend. Lineages identified as positive selection at P < 0.05 after correction for multiple testing are shown with thick branches, with color distributions representing the relative values and proportions of inferred ω categories. Note that taxon labels beginning with "0554" represent HIV-1 sequences derived from the donor patient, and labels beginning with "0557" represent HIV-1 sequences derived from the recipient patient
amount of statistical signal does not increase for larger datasets. One option is to thin out large phylogenies (before performing any testing), retaining major clades and lineages of interest.
#### 3.5 Site-Level Selection: MEME, FEL,
#### SLAC, and FUBAR What Biological Question Is the Method Designed to Answer?:
The methods FEL, SLAC, and FUBAR address the question: Which site(s) in a gene are subject to pervasive, i.e., consistently across the entire phylogeny, diversifying selection? MEME addresses a more general question: Which site(s) in a gene are subject to pervasive or episodic, i.e., only on a single lineage or subset of lineages, diversifying selection?
Recommended Applications
- <sup>l</sup> FEL is our recommended method for analyzing small-tomedium size datasets when one wishes only to study pervasive selection at individual sites.
- <sup>l</sup> FUBAR is our recommended method for detecting pervasive selection at individual sites on large (> 500 sequences) datasets for which other methods have prohibitive runtimes, unless you have access to a computer cluster.
- <sup>l</sup> SLAC provides legacy functionality as a counting-based method adapted for phylogenetic applications. In general, this method will be the least statistically robust.
#### Statistical Test Procedure:
Each method presented here employs a distinct algorithmic approach to inferring selection. FEL uses maximum likelihood to fit a codon model to each site, thereby estimating a value for dN and dS at each site. FEL tests for selection with the likelihood ratio test using the χ<sup>2</sup> <sup>1</sup> distribution, asking whether the dN estimate is significantly greater than the inferred dS estimate.
SLAC represents the most basic inference method and is an extension of the Suzuki–Gojobori counting-based method [39] for phylogenetically related sequences (as opposed to sequence pairs). SLAC uses maximum likelihood to infer ancestral characters for each site across the phylogeny and then directly counts the number of synonymous and non-synonymous changes which have occurred at each site over evolutionary time. SLAC then tests for selection by testing whether or not there are too many or too few non-synonymous changes compared to what is expected under neutrality. The neutral expectation is derived based on the phylogeny-wide estimated numbers of synonymous and non-synonymous nucleotide sites at a given codon. The statistical test employs the binomial distribution to compute significance, e.g., how likely is it to observe 13 non-synonymous and 1 synonymous substitutions at a site, if the expected synonymous to non-synonymous substitution count ratio under neutrality is 1:4?
MEME employs a mixed-effects maximum likelihood approach. For each site, MEME infers two ω rate classes and corresponding weights representing the probability that the site evolves under each rate class at a given branch. To this end, MEME infers a single α (dS) parameter and two separate β (dN) parameters, β and β+. The ω rates per site, therefore, consist of β+/α and β/α. MEME uses this framework to fit a null and alternative model each, both models enforcing the constraint β α. The null model disallows positive selection by enforcing the constraint β<sup>+</sup> α, whereas the alternative model places no constraint on β+. MEME uses the likelihood ratio test to compare between null and alternative model fits, with significance assessed using the mixture of 33%χ<sup>2</sup> <sup>0</sup>, 30%χ<sup>2</sup> <sup>1</sup>, and 37%χ<sup>2</sup> 2.
FUBAR takes a Bayesian approach to selection inference and is a particular case of statistical models developed in the context of document classification (latent Dirichlet allocation). The key innovation to FUBAR's approach is its use of an a priori specified grid of dN and dS values (typically 20 20), spanning the range of negative, neutral, and positive selection regimes, whose likelihoods can be pre-computed and used throughout analysis (rather than having to re-compute likelihoods during optimization as traditional random-effects approaches do [12, 29]). This approach, combined with other algorithmic advances, speeds computation time by at least an order of magnitude compared to FEL, while yielding comparable statistical performance. FUBAR estimates every model parameter except the proportion of sites allocated to each grid point using simple (and fast) nucleotide models. The proportions are estimated using an MCMC procedure, and non-neutral evolution at each site is inferred using a straightforward naive
(continued)
empirical Bayes approach [29]. Sites are called positively or negatively selected if the corresponding posterior probabilities are sufficiently high.
Note that FEL and SLAC report both positively and negatively selected sites, but MEME and FUBAR report only sites under positive selection.
#### Example Analysis:
We will demonstrate the use and interpretation of site-level methods using data from influenza strain H3N2 (the "Hong Kong flu"), the primary circulating strain of seasonal influenza since the late 1960s. We specifically will assess selection on the H3 hemagglutinin, the influenza surface protein which is responsible for host cell binding. Hemagglutinin experiences rapid evolution triggered by host immune escape, and previous studies have identified numerous signatures of positive diversifying selection in H3 sequences with a particular concentration around the host-binding domain [28].
We base analyses here on an alignment from Meyer and Wilke [22] of H3 sequences sampled over time since the 1991–1992 influenza season. We removed all partial and strongly outlying sequences (i.e., those with excessive divergence) from the original dataset before proceeding, yielding 2555 sequences to comprise our "full" H3 dataset. We further subsetted this alignment to two smaller alignments with comparable numbers of taxa but spanning different evolutionary time frames: The first smaller alignment ("trunk") contains 163 sequences sampled along the influenza H3 trunk, whereas the second smaller alignment ("shallow") contains 121 sequences sampled from a single clade (Fig. 3). Therefore, while these two smaller datasets contain a comparable number of sequences, the trunk dataset spans a much longer time frame and contains substantially more sequence divergence relative to the shallow dataset. Indeed, the trunk dataset has a total tree length (sum of branch lengths, in units substitutions/site/unit time) of 0.43, whereas the shallow dataset had a total tree length of 0.12, meaning that the trunk dataset contains nearly four times the amount of sequence divergence seen in the shallow dataset. We have compiled results for all three datasets analyzed with all four methods (Table 1). We now describe, using the trunk dataset as an example, how to run each of these analyses in HyPhy.
Fig. 3 Phylogeny of H3 hemagglutinin sequences analyzed here. Tip colors indicate those selected for each dataset
#### Table 1 Sites identified as positively selected across the H3 datasets analyzed here
Bold sites are those identified by multiple methods for a given dataset. Bold italicized sites are those identified in more than one dataset, generally by more than one method. Numbers in parentheses give the total number of positively selected sites identified with the given method and dataset
<sup>∗</sup> For FUBAR, significance is assessed as posterior probability 0.9
FEL: Launch HyPhy from the command line, and enter options 1 (Selection Analyses) and then 2 to reach the FEL analysis menu, and supply values for the following prompts:
FEL will now run to completion and print status indicators to the screen, including results for any site found to be under selection (either positive or negative). Abbreviated results are shown below.
#### Listing 4 Partial FEL screen output:
```
### Obtaining branch lengths and nucleotide rates under the GTR model
* Log ( L ) = -7506.06
### Obtaining the global omega estimate based on relative GTR branch lengths
and nucleotide substitution biases
* Log ( L ) = -7302.10
* non - synonymous / synonymous rate ratio for *test* = 0.2923
### Improving branch lengths , nucleotide substitution biases, and global dN/dS
ratios under a full codon model
* Log ( L ) = -7289.65
* non - synonymous / synonymous rate ratio = 0.2598
### For partition 1 these sites are significant at p <=0.1
...
```
### \*\* Found \_3\_ sites under pervasive positive diversifying and \_115\_ sites under negative selection at p <= 0.1\*\*
> Inference details for codons with significant likelihood ratio tests for positive or negative selection are reported to the screen.
Note that the "Codon" and "Partition" columns are common to all site-specific analyses.
MEME and SLAC: SLAC and MEME follow identical menu prompts as FEL, with the exception that only FEL will prompt for synonymous rate variation. Instead, SLAC has a different prompt for Step 5: Select the number of samples used to assess ancestral reconstruction uncertainty. If this number is positive, then HyPhy will draw samples from the distribution of ancestral states and use them to measure whether or not inference is sensitive to ancestral inference uncertainty. When you encounter this option, provide the default value of 100 (or 0 to forego sampling). MEME does not emit any additional prompts.
#### Listing 5 Partial SLAC screen output:
```
### For partition 1 these sites are significant at p <=0.1
...
...
### Ancestor sampling analysis
> Generating 100 ancestral sequence samples to obtain confidence intervals
Resampling results for partition 1
...
```
...
... SLAC reports several key quantities for codons with significant P-values for positive or negative selection to the screen.
If the user elected to perform ancestral resampling, another table is reported, showing how much these quantities are affected by ancestral state reconstruction uncertainty. For example, at codon 177, some ancestral reconstructions yielded 3 synonymous substitutions, whereas others yielded 4; however, this was not sufficient to move the P-value on different sides of the threshold.
#### Listing 6 Partial MEME screen output:
FUBAR: To run FUBAR, launch HyPhy from the command line, and enter options 1 (Selection Analyses) and then 4 to reach the FUBAR analysis menu, and supply values for the following prompts<sup>5</sup> :
<sup>5</sup> Note that for all prompts with default values, simply pressing enter will choose this default.
#### Listing 7 Partial FUBAR screen output:
... ### Tabulating site - level results | Codon |Partition| alpha | beta | N.eff |Posterior prob for positive selection| |:-------:|:------–: |:----:|:---–: |:------–:|:----------------------------------–:| | 61 | 1 | 0.753 | 4.365| 64.549 | Pos. posterior = 0.9262 | | 64 | 1 | 0.753 | 3.920| 77.106 | Pos. posterior = 0.9095 | | 69 | 1 | 0.730 | 4.447| 64.182 | Pos. posterior = 0.9325 | | 154 | 1 | 0.637 | 6.595| 53.312 | Pos. posterior = 0.9826 | 456 Stephanie J. Spielman et al.
> Like other site analyses, FUBAR will print a number of inferences about each individual site detected to be under pervasive positive selection
#### Interpreting Results:
Sites identified as positively selected by each method, across all three datasets, are given in Table 1. In general, we expect MEME to be the most comprehensive and robust of all site-level methods because it uniquely considers both pervasive and episodic selection [24]. In addition, power studies have shown that FUBAR is expected to outperform FEL and SLAC under most circumstances [25]. Finally, we expect that SLAC will be the least robust method due to its reliance on a relatively naive counting-based approach [12].
These expectations are generally borne out in the results obtained here in our brief study of H3 selection. For the full H3 dataset of 2555 sequences, MEME identified 16 sites, and FEL identified 15 sites under positive selection. All sites were identical except for the following: MEME uniquely identified sites 151 and 208, and FEL uniquely identified with 237. Interestingly, site 208 was additionally identified as positively selected by all methods on the trunk H3 dataset. Combined, these results demonstrate MEME's ability to identify sites subject to both pervasive and episodic selection, as site 208 appears to be under pervasive selection only along the H3 trunk. Because FEL uses a less stringent test statistic distribution ( χ<sup>2</sup> <sup>1</sup> ) to call significance, occasionally sites subject to pervasive selection near the significance thresholds may be detected by FEL but missed by MEME (e.g., site 237, with FEL reporting P ¼ 0.08 and MEME reporting P ¼ 0.105).
FUBAR identified two fewer selected sites in the full H3 alignment compared to FEL (which is a directly comparable test), missing sites 19 (posterior 0.83), 277 (posterior 0.59), and 292 (posterior 0.89) relative to FEL, but adding site 160 (FEL P ¼ 0.8).
In addition to differences across methods, we expect to see some important differences for sites inferred across the full, shallow, and trunk H3 datasets. Because the trunk and full H3 datasets span similar time frames, we expect sites returned for these two datasets to have the most overlap. In addition, sites found to be under selection in the shallow lineage may not be detected across the full H3 phylogeny, as selection may have been fleeting, weak, or constrained to the specific shallow clade examined here. For example, site 49 was specifically selected in the shallow H3 lineage alone, as indicated by three of the four methods. In contrast, sites 19 and 241 were found to be selected in both the shallow and the full H3 datasets, but this signal was not apparent when the trunk lineage was examined independently, perhaps because these sites experience only transient changes that do not propagate along the trunk.
What are some potential reasons for seeing discrepancies in inferences across H3 datasets? The site 154, for example, is positively selected in both the full H3 phylogeny and the trunk H3 lineage, but not the shallow H3 lineage. This result suggests that site 154 may have experienced pervasive selection throughout H3 evolution, but its signal in the shallow clade alone was either too weak to detect or selection was attenuated in the shallow clade. In addition, sites which appeared only in the shallow clade analyses may have experienced lineage-specific selection where the signal was too weak to detect when the entire phylogeny was considered.
Furthermore, while MEME, FEL, and FUBAR were able to detect selected sites in the shallow H3 lineage, SLAC did not identify any such sites. This is because SLAC requires a large number of substitutions, which are unlikely to have occurred in the shallow sample, to achieve significance. Overall, we emphasize that in many cases different site-level methods will not identify exactly the same set of sites under selection, although, as the H3 example shows, the agreement between is typically good.
#### Rules of Thumb for Site-Level Detection of Selection
corrections will nearly always yield no significant results on small to moderate sized datasets. Thirdly, some key assumptions of methods for correcting false discovery rates are not applicable for site-level testing. For example, a typical collection of results from site-level testing will contain very few, if any, true sites with P-values supporting neutrality (dN/ dS ¼ 1).
3.6 Screening Sequences for Recombination A critical aspect of sequence analysis we have not yet covered is the detection of and correction for intragenic recombination in an alignment of homologous sequences. Because recombination is such a key biological process in many viral pathogens, we strongly advocate screening an alignment for recombination before proceeding with additional analyses, unless there is a sound biological reason to discount (i.e., intragenic recombination Influenza A is negligibly rare). Indeed, because recombination causes different regions of an alignment to be related by different phylogenies, its presence can heavily influence selection detection and other downstream applications.
> There are many computational approaches to finding evidence of recombination in a sequence alignment [32], however at their core, many such methods look for evidence of phylogenetic incongruence. Here, we demonstrate one such method, GARD (genetic algorithms for recombination detection) that we have found to perform very well among a wide range of approaches on simulated data [14]. Note that at this time, GARD will not produce a JSON file as output but instead several text files containing inference information, as well as a final partitioned alignment for downstream use if recombination was detected.
#### 3.7 GARD
#### What Biological Question Is the Method Designed to Answer?:
Have sequences in the given alignment undergone recombination, and if so what are the recombination breakpoints and segment-specific phylogenies?
Recommended Applications:
GARD is geared towards mapping the breakpoints and detecting segments of the alignment which can be adequately described by a single tree topology. Therefore, alignments, particularly alignments of viral sequences, should be screened for the presence of recombination before performing any selection inference. The NEXUS output from GARD can be directly used as input for most downstream selection detection analyses.
#### Statistical Test Procedure:
GARD employs a genetic algorithm to find a solution to a complex optimization problem by mimicking processes of biological evolution (mutation, recombination, and selection) in a population of competing solutions. In this application of genetic algorithms, we are evolving a population of "chromosomes" that specify different numbers and locations of recombination breakpoints in the alignment with the objective of detecting topological incongruence, i.e., support for different phylogenies by separate regions of the alignment. The "fitness" of each chromosome is determined by using maximum likelihood methods to evaluate a separate phylogeny for each non-recombinant fragment defined by the breakpoints (e.g., to the left and to the right of a breakpoint in Fig. 4), and computing a goodness of fit (AICc) for each such model. The genetic algorithm searches for the number and placement of breakpoints yielding the best AIC<sup>c</sup> and also reports confidence values for inferred breakpoint locations based on the contribution of each considered model weighted by how well the model fit the data. For computational expedience, the current implementation of GARD infers topologies for each segment using neighbor joining [37] based on the TN93 pairwise distance estimator [41] and then fits a user-specified nucleotide evolutionary model using maximum likelihood to obtain AIC<sup>c</sup> scores.
Fig. 4 Phylogenetic incongruence caused by the presence of a recombinant sequence in an alignment. Sequence R is a product of homologous recombination between sequences A and B. Phylogenies reconstructed from sequences A, B, R and an outgroup sequence (O) will differ based on which part of the alignment is being considered. To the left of the breakpoint, R clusters with A, whereas to the right of the breakpoint R clusters with B
Example Analysis 1: We will demonstrate the use of GARD, as well as its benefits for downstream analysis, using a dataset consisting of 13 glycoprotein sequences from Cache Valley Fever virus (cvf.fna). We will first use GARD to detect recombination in this dataset, and then we will process both the GARD-informed data and the original alignment (with no recombination assumed) with FEL to see how the presence of recombination may confound selection inference.
Importantly, GARD specifically requires the use of HyPhy's MPIenabled executable, HYPHYMPI. To run GARD from the command line, you will need an operating system with a MPI headers and libraries installed so that this executable can be compiled. Here, we will describe how to use GARD from the command line, but we emphasize that GARD is fully implemented and available on www. datamonkey.org and takes the same input options described here.
To run GARD, open a terminal session and start HYPHYMPI in the appropriate MPI environment (e.g., MPIRUN in OpenMPI) from the command line to launch the HyPhy analysis menu. Enter 12 (Recombination) and then 1 to reach the GARD analysis menu, and supply values for the following prompts:
5. Save results to. For this option, provide a full path to the output file to which you would like GARD to write results. The supplied file name will ultimately contain an HTML-formatted summary of the analysis. HyPhy will generate several other files with names obtained by appending suffixes (as in <file name>\_suffix) to the main result file. In particular, the \_finalout file stores the original alignment in NEXUS format with inferred non-recombinant sections of the alignment saved in the ASSUMPTIONS block and trees inferred for each partition in the TREES block. This NEXUS file can be input into many recombination-aware analyses in HyPhy and other programs that can read NEXUS. The \_ga\_details file contains two lines of information about each model examined by the genetic algorithm: its AICc score and the location of breakpoints in the model. Finally, the \_ga\_splits file stores information about the location of breakpoints and trees inferred for each alignment region under the best model found by the GA.
GARD will now run to completion, printing status indicators to screen while it runs:
#### Listing 8 Partial GARD output:
```
Fitting a baseline nucleotide model...
Done with single partition analysis. Log(L) = -5921.9511901113, c-AIC = 11914.85153276497
Starting the GA ...
GENERATION 2 with 1 breakpoints (~0% converged)
Breakpoints c - AIC Delta c - AIC [BP 1]
0 11914.85
1 11804.56 110.291 1393
GA has considered 92/ 328 (92 over all runs) unique models
Total run time 0 hrs 0 mins 2 seconds
Throughput 46.00 models/second
Allocated time remaining 999 hrs 59 mins 58 seconds (approx. 165599908 more models.)
...
GENERATION 52 with 4 breakpoints (~100% converged)
Breakpoints c - AIC Delta c - AIC [BP 1] [BP 2] [BP 3] [BP 4]
0 11914.85
1 11804.56 110.291 1445
2 11783.92 20.638 617 1490
3 11778.94 4.978 587 962 1475
4 11778.94 0.000 587 962 1475
GA has considered 268/ 473490550 (1356 over all runs) unique models
Total run time 0 hrs 4 mins 2 seconds
Throughput 5.60 models/second
Allocated time remaining 999 hrs 55 mins 58 seconds (approx. 20170544.82644628 more models.)
```
Performing the final optimization...
#### Interpreting Results:
GARD found evidence of recombination in this dataset with three breakpoints, yielding a 135.9 point AIC<sup>c</sup> improvement over the model without recombination. Among all models with three breakpoints in the Cache Valley Virus glycoprotein alignment, the best model places them at nucleotides 587, 962, and 1475. Importantly, if GARD had reported that the best model had 0 breakpoints, we could conclude that no evidence of recombination had been found. Note that because genetic algorithms are stochastic, there is no guarantee that replicate runs will converge to exactly the same quantitative results. When there is a strong signal of recombination breakpoints in the data, however, the qualitative results (number and general location of breakpoints) should be fairly robust.
Example Analysis 2: The NEXUS file that GARD produced is a partitioned dataset, wherein different groups of sites are described by different trees. Most HyPhy selection analyses discussed here,6 including MEME, FUBAR, FEL, SLAC, and BUSTED, are able to analyze partitioned data. To demonstrate the importance of screening for recombination, we will now compare results for a FEL analysis performed on the original alignment of 13 Cache Valley Virus glycoproteins, as well as on the GARD-inferred partitioned alignment. All steps here were carried out as described earlier in this chapter.
#### Interpreting Results:
FEL inference on the GARD-processed partitioned Cache Valley Virus data does not detect sites under selection at P - 0.1. By contrast, FEL inference on the unpartitioned Cache Valley Virus data (i.e., not pre-screened for recombination) detects three positively selected sites at P - 0.1 (212, 516, and 558 at P = 0.08, P = 0.03, and P = 0.09, respectively). From these results, we can clearly tell that not screening or recombination has the potential for adverse consequence including an increased false positive rate as seen here. As such, we strongly encourage users to screen alignments for recombination if such processes are suspected before proceeding to selection detection.
#### 3.8 Accounting for Synonymous Rate Variation
A critical genomic process that one must consider when detecting selection is the phenomenon of synonymous rate variation, wherein the rate of synonymous codon evolution (represented by dS in the
<sup>6</sup> Note that neither aBSREL nor RELAX accepts partitioned data because they require a consistent phylogeny to define branch sets.
context of codon models and representing mutation rate) varies across species, genes, and even intragenic positions. In particular, intragenic synonymous rate variation has been identified across domains of life [11, 20, 30] and can arise from a variety of evolutionary processes, including selection on mRNA secondary structure [2], gene expression [4], GC-biased gene conversion [10], and other neutral mutation processes. For example, even the genomic context of a given nucleotide can influence its mutation rate; indeed, experimental work has shown that GC-neighboring sites can feature up to a 75-fold increase in mutation rate [20, 38]. In addition, the synonymous rate at certain sites may be elevated due to the mutational vulnerability of the non-template DNA strand during transcription [20]. These processes must be accounted for in order to ensure an appropriate baseline dS is used when testing for selection.
We demonstrate the importance of considering synonymous rate variation for selection inference using a dataset of 10 mammalian CD2 genes, which code for a specific T-cell surface adhesion molecule [21]. We use FEL to detect selection in this dataset under two specifications: with synonymous rate variation ("yes" in prompt 4 in the FEL analysis menu), and without synonymous rate variation ("no" in prompt 4 in the FEL analysis menu).
#### Interpreting Results:
At P - 0.1, analysis of CD2 with synonymous rate variation revealed a total of 14 sites under positive selection. By contrast, CD2 analysis with FEL without dS variation only detected four sites under positive selection (Fig. 5). Similarly, analysis with dS variation revealed 27 sites under purifying selection, but analysis without dS variation revealed only 15 sites under purifying selection. Most importantly, all sites detected when dS was fixed to 1 were a subset of the sites identified by the model with dS variation (Fig. 5). Together, these results demonstrate that ignoring dS variation can induce both an increased false negative rate regarding positive selection detection and an overall decrease in power to detect any selective regime. We acknowledge that it is possible that the opposite conclusion might be true, namely, that additional sites identified by FEL with dS variation might instead be false positives. However, in our experience, this is much less frequently the case [12].
Fig. 5 Sites identified as positively (red) and negatively (blue) selected in CD2 at <sup>P</sup> - 0.1 by FEL run with (above the line) and without dS variation (below the line). Sites with arrows represent those identified as selected by FEL with dS variation that were not identified by FEL when dS variation was ignored
#### 4 Tips
Here we provide some helpful notes on HyPhy usage.
#### 5 Exercises
- <sup>l</sup> Is there evidence (compare model fits using the small sample AIC) that test branches have a different selective regime than the rest of the tree?
- <sup>l</sup> The entire dataset should provide evidence for episodic diversification, but the recipient only analysis should return a negative result. What does this mean biologically, i.e., where does the selection signal come from?
- <sup>l</sup> Which analysis generated more positive results?
- <sup>l</sup> Do you think these results are true or false positives? How does this compare to the FEL analysis we described in the text?
- <sup>l</sup> Compare site-wise estimates of substitution rates (e.g., α) between the two analyses. Is there a discernible bias introduced by not accounting for recombination?
#### References
and applications to the HIV-1 envelope gene. Genetics 148:929–936
revealed through mutation–accumulation experiments. Mol Biol Evol 32(7):1672–1683
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Chapter 15
#### Evolution of Protein Domain Architectures
#### Sofia K. Forslund, Mateusz Kaduk, and Erik L. L. Sonnhammer
#### Abstract
This chapter reviews current research on how protein domain architectures evolve. We begin by summarizing work on the phylogenetic distribution of proteins, as this will directly impact which domain architectures can be formed in different species. Studies relating domain family size to occurrence have shown that they generally follow power law distributions, both within genomes and larger evolutionary groups. These findings were subsequently extended to multi-domain architectures. Genome evolution models that have been suggested to explain the shape of these distributions are reviewed, as well as evidence for selective pressure to expand certain domain families more than others. Each domain has an intrinsic combinatorial propensity, and the effects of this have been studied using measures of domain versatility or promiscuity. Next, we study the principles of protein domain architecture evolution and how these have been inferred from distributions of extant domain arrangements. Following this, we review inferences of ancestral domain architecture and the conclusions concerning domain architecture evolution mechanisms that can be drawn from these. Finally, we examine whether all known cases of a given domain architecture can be assumed to have a single common origin (monophyly) or have evolved convergently (polyphyly). We end by a discussion of some available tools for computational analysis or exploitation of protein domain architectures and their evolution.
Key words Protein domain, Protein domain architecture, Superfamily, Monophyly, Polyphyly, Convergent evolution, Domain evolution, Kingdoms of life, Domain co-occurrence network, Node degree distribution, Power law, Parsimony
#### 1 Introduction
1.1 Overview By studying the domain architectures of proteins, we can understand their evolution as a modular phenomenon, with high-level events enabling significant changes to take place in a time span much shorter than required by point mutations only. This research field has become possible only now in the -omics era of science, as both identifying many domain families in the first place and acquiring enough data to chart their evolutionary distribution require access to many completely sequenced genomes. Likewise, the conclusions drawn generally consider properties averaged for entire
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_15, © The Author(s) 2019
species or organism groups or entire classes of proteins, rather than properties of single genes.
We will begin by introducing the basic concepts of domains and domain architectures, as well as the biological mechanisms by which these architectures can change. The remainder of the chapter is an attempt at answering, from the recent literature, the question of which forces shape domain architecture evolution and in what direction. The underlying issue concerns whether it is fundamentally a random process or whether it is primarily a consequence of selective constraints. We end by outlining some available software tools and resources for analysis of domain architectures and their evolution.
1.2 Protein Domains Protein domains are high-level parts of proteins that either occur alone or together with partner domains on the same protein chain. Most domains correspond to tertiary structure elements and are able to fold independently. All domains exhibit evolutionary conservation, and many either perform specific functions or contribute in a specific way to the function of their proteins. The word domain strictly refers to a distinct region of a specific protein, an instance of a domain family. However, domain and domain family are often used interchangeably in the literature.
1.3 Domain Databases By identifying recurring elements in experimentally determined protein 3D structures, the various domain families in structural domain databases such as SCOP [1] and CATH [2] were gathered. New 3D structures allow assignment to these classes from semiautomated inspection. The SUPERFAMILY [3] database assigns SCOP domains to all protein sequences by matching them to hidden Markov models (HMMs) that were derived from SCOP superfamilies, i.e., proteins whose evolutionary relationship is evidenced structurally. The Gene3D [4] database is similarly constructed but based on domain families from CATH.
This approach resembles the methodology used in pure sequence-based domain databases such as Pfam [5]. In these databases, conserved regions are identified from sequence analysis and background knowledge, to make multiple sequence alignments. From these, HMMs are built that are used to search new sequences for the presence of the domain represented by each HMM. All such instances are stored in the database. The HMM framework ensures stability across releases and high quality of alignments and domain family memberships. The stability allows annotation to be stored along with the HMMs and alignments. The InterPro database [6] is a meta-database of domains combining the assignments from several different source databases, including Pfam. The Conserved Domain Database (CDD) is a similar meta-database that also contains additional domains curated by the NCBI [7]. SMART [8] is a manually curated resource focusing primarily on signaling and extracellular domains. ProDom [9] is a comprehensive domain database automatically generated from sequences in UniProt [10]. Likewise, ADDA [11] is automatically generated by clustering subsequences of proteins from the major sequence databases, though it has not been updated for some time. Genome3D [12] is a recent consensus database which brings together several domain prediction tools as well as the SCOP and CATH databases for describing representative domain arrangements in a series of trusted, well-annotated genomes.
Since the domain definitions from different databases only partially overlap, results from analyses often cannot be directly compared. In practice, however, choice of database appears to have little effect on the main trends reported by the studies described here.
1.4 Domain Architectures The terms "domain architecture" or "domain arrangement" generally refer to the domains in a protein and their order, reported in N- to C-terminal direction along the amino acid chain. Another recurring term is domain combinations. This refers to pairs of domains co-occurring in proteins, either anywhere in the protein (the "bag-of-domains" model) or specifically pairs of domains being adjacent on an amino acid chain, in a specific N- to C-terminal order [13]. The latter concept is expanded to triplets of domains, which are subsequences of three consecutive domains, with the N- and C-termini used as "dummy" domains. A domain X occurring on its own in a protein thus produces the triplet N-X-C [14].
#### 1.5 Mechanisms for Domain Architecture Change
Most mutations are point mutations: substitutions, insertions, or deletions of single nucleotides. While conceivably enough of these might create a new domain from an old one or noncoding sequence or remove a domain from a protein, in practice we are interested in mechanisms whereby the domain architecture of a protein changes instantly or nearly so (but see below for an overview of recent work on the origin of new domains). Figure 1 shows some examples of ways in which domain architectures may mutate. In general, adding or removing domains requires genetic recombination events. These can occur either through errors made by systems for repairing DNA damage such as homologous [16, 17] or nonhomologous (illegitimate) [18, 19] recombination or through the action of mobile genetic elements such as DNA transposons [20] or retrotransposons [21, 22]. Recombination can cause loss or duplication of parts of genes, entire genes or much longer chromosomal regions.
In organisms that have introns, exon shuffling [23, 24] refers to the integration of an exon from one gene into another, for instance, through chromosomal crossover, gene conversion, or mobile genetic elements. Exons could also be moved around by being
Fig. 1 Examples of mutations that can change domain architectures. Adapted from Buljan et al. [25]. (a) Gene fusion by a mobile element. LINE refers to a Long Interspersed Nuclear repeat Element, a retrotransposon. The reverse transcriptase encoded within the LINE causes its mRNA to be reverse-transcribed into DNA and integrated into the genome, making the domain-encoding blue exon from the donor gene integrate along with it in the acceptor gene. (b) Gene fusion by loss of a stop signal or deletion of much of the intergenic region. Genes 1 and 2 are joined together into a single, longer gene. (c) Domain insertion through recombination. The blue domain from the donor gene is inserted within the acceptor gene by either homologous or illegitimate recombination. (d) Right: Gene fission by introduction of transcription stop (the letter Ω) and start (the letter A). Left: Domain loss by introduction of a stop codon (exclamation mark) with subsequent degeneration of the now untranslated domain
brought along by mobile genetic elements such as retrotransposons [24, 25].
Two adjacent genes can be fused into one if the first one loses its transcription stop signals. Point mutations can cause a gene to lose a terminal domain by introducing a new stop codon, after which the "lost" domain slowly degrades through point mutations as it is no longer under selective pressure [26]. Alternatively, a multi-domain gene might be split into two genes if both a start and a stop signal are introduced between the domains. Novel domains could arise, for instance, through exonization, whereby an intronic or intergenic region becomes an exon, after which subsequent mutations would fine-tune its folding and functional properties [25, 27].
Recent literature (see, e.g., [28]) has discussed the possibility of de novo domain creation through a variety of mutational mechanisms, with some support for this occurring more often than previously thought [29, 30]. The majority of such new domains arise as novel genes from noncoding sequence but may subsequently recombine to join with older domains. Furthermore, young domains in vertebrates tend more often to occur at the N-terminal of a protein and tend to experience higher relative rates of non-synonymous substitution than older domains, which may reflect the nature of the mechanisms through which novel domains arise. Moore, Bornberg-Bauer et al. explore the relative prevalence of domain loss, duplication, and de novo origination in arthropods [31] and plants [32], suggesting such novel domains most frequently are associated with environmental adaptations.
#### 2 Distribution of the Sizes of Domain Families
Domain architectures are fundamentally the realizations of how domains combine to form multi-domain proteins with complex functions. Understanding how these combinations come to be requires first that we understand how common the constituent domains of those architectures are and whether there are selective pressures determining their abundances. Because of this, the body of work concerning the sizes and species distributions of domain families becomes important to us.
Comprehensive studies of the distributions and evolution of protein domains and domain architectures are possible as genome sequencing technologies have made many entire proteomes available for bioinformatic analysis. Initial work [33–35] focused on the number of copies that a protein family, either single domain or multi-domain, has in a species. Most conclusions from these early studies appear to hold true for domains, for supra-domains (see below) and for domain architectures [36–38]. In particular, these all exhibit a dominance of the population by a selected few [35], i.e., a small number of domain families are present in a majority of the proteins in a genome, whereas most domain families are found only in a small number of proteins.
Looking at the frequency N of families of size X (defined as the number of members in the genome), in the earliest studies, this frequency was modeled as the power law
$$N = \iota X^{-a}$$
where a is an exponent parameter. The power law is a special case of the generalized Pareto distribution (GPD) [39]:
$$N = \mathfrak{a}(i+X)^{-a}$$
Power law distributions arise in a vast variety of contexts: from human income distributions, connectivity of internet routers, word usage in languages, and many other situations ([34, 35, 40, 41], see also [42], for a conflicting view). Luscombe et al. [35] described a number of other genomic properties that also follow power law distributions, such as the occurrence of DNA "words," pseudogenes, and levels of gene expression. These distributions fit much better than the alternative they usually are contrasted against, an exponential decay distribution. The most important difference between exponential and power law distributions in this context concerns the fact that the latter has a "fat tail," that is, while most domain families occur only a few times in each proteome, most domains in the proteome still belong to one of a small number of families.
Later work ([39, 43], see also [44]) demonstrated that proteome-wide domain occurrence data fit the general GPD better than the power law but that it also asymptotically fits a power law as X i. The deviation from strict power law behavior depends on proteome size in a kingdom-dependent manner [43]. Regardless, it is mostly appropriate to treat the domain family size distribution as approximately (and asymptotically) power law-like, and later studies typically assume this.
The power law, but not the GPD, is scale-free in the sense of fulfilling the condition
$$f(a\mathfrak{x}) = \mathfrak{g}(a)f(\mathfrak{x})$$
where f(x) and g(x) are some functions of a variable x and where a is a scaling parameter, that is, studying the data at a different scale will not change the shape of function. This property has been extensively studied in the literature and is connected to other attributes, notably when it occurs in network degree distributions (i.e., frequency distributions of edges per node). Here it has been associated with properties such as the presence of a few central and critical hubs (nodes with many edges to other nodes), the similarity between parts and the whole (as in a fractal), and the growth process called preferential attachment, under which nodes are more likely to gain new links the more links they already have. However, the same power law distribution may be generated from many different network topologies with different patterns of connectivity. In particular, they may differ in the extent that hubs are connected to each other [42]. It is possible to extend the analysis by taking into account the distribution of degree pairs along network edges, but this is normally not done.
What kind of evolutionary mechanisms give rise to this kind of distribution of gene or domain family sizes within genomes? In one model by Huynen and van Nimwegen [33], every gene within a gene family will be more or less likely to duplicate, depending on the utility of the function of that gene family within the particular lineage of organisms studied, and they showed that such a model matches the observed power laws. While they claimed that any model that explains the data must take into account family-specific probabilities of duplication fixation, Yanai and coworkers [45] proposed a simpler model using uniform duplication probability for all genes in the genome and also reported a good fit with data.
Later, more complex birth-death [43] and birth-death-andinnovation (BDIM) [29, 34, 39, 46] models were introduced to explain the observed distributions, and from investigating which model parameter ranges allow this fit, the authors were able to draw several far-ranging conclusions. First, the asymptotic power law behavior requires that the rates of domain gain and loss are asymptotically equal. Karev et al. [39] interpreted this as support for a punctuated equilibrium-type model of genome evolution, where domain family size distributions remain relatively stable for long periods of time but may go through stages of rapid evolution, representing a shift between different BDIM evolutionary models and significant changes in genome complexity. Like Huynen and van Nimwegen [33], they concluded that the likelihood of fixated domain duplications or losses in a genome directly depend on family size. The family will however only grow as long as new copies can find new functional niches and contribute to a net benefit for survival, i.e., as long as selection favors it.
Aside from Huynen and van Nimwegen's, none of the models discussed depend very strongly on family-specific selection to explain the abundances of individual gene families, nor do they exclude such selection. Some domains may be highly useful to their host organism's lifestyle, such as cell-cell connectivity domains to an organism beginning to develop multicellularity. Expansion of these domain families might therefore become more likely in some lineages than in others. To what extent these factors actually affect the size of domain families remains to be fully explored. Karev et al. [39] suggested that the rates of domain-level change events themselves—domain duplication and loss rates, as well as the rate of influx of novel domains from other species or de novo creation must be evolutionarily adapted, as only some such parameters allow the observed distributions to be stable. Van Nimwegen [47] investigated how the number of genes increases in specific functional categories as total genome size increases. He found that the relationship matches a power law, with different coefficients for each functional class remaining valid over many bacterial lineages. Ranea et al. found similar results. Also, Ranea et al. [48] showed that, for domain superfamilies inferred to be present in the last universal common ancestor (LUCA), domains associated with metabolism have significantly higher abundance than those associated with translation, further supporting a connection between the function of a domain family and how likely it is to expand.
Extending the analysis to multi-domain architectures, Apic et al. [37] showed that the frequency distribution of multi-domain family sizes follows a power law curve similar to that reported for individual domain families. It therefore seems likely that the basic underlying mechanisms should be similar in both cases, i.e., that duplication of genes, and thus their domain architectures, is the most important type of event affecting the evolution of domain architectures.
Have the trends described above stood the test of time as more genomes have been sequenced and more domain families have been identified? We considered the 1943 UniProt proteomes covered by version 30.0 of Pfam, plotted the frequency Y of domain families that have precisely X members as a function of X, and fit a power law curve to this. Figure 2a shows the resulting plots for three representative species, one complex eukaryote (Homo sapiens),
Fig. 2 (a) Distribution of domain family sizes in three selected species. Power law distributions were fitted to these curves such that for frequency <sup>f</sup> of families of size <sup>X</sup>, <sup>f</sup> <sup>¼</sup> cXa . For S. cerevisiae, a ¼ -1.9, for E. coli, a ¼ -1.7, and for H. sapiens, a ¼ -1.5. (b) Distribution of domain family sizes across the three kingdoms. Power law distributions were fitted to these curves such that for frequency <sup>f</sup> of families of size <sup>X</sup>, <sup>f</sup> <sup>¼</sup> cX<sup>a</sup> . For bacteria, a ¼ -0.9, for archaea, a ¼ -1.1, for eukaryotes, a ¼ -0.8, and for viruses, a ¼ -1.9
one simple eukaryote (Saccharomyces cerevisiae), and one prokaryote (Escherichia coli). Figure 2b shows the corresponding plots for all domains in all complete eukaryotic, bacterial, and archaeal proteomes. The power law curve fits decently well, with slopes becoming less steep for the more complex organisms, whose distributions have relatively more large families. The power law-like behavior suggests that complex organisms with large proteomes were formed by heavily duplicating domains from relatively few families. Figures 3a, b show equivalent plots, not for single domains but for entire multi-domain architectures. The curve shapes and the relationship between both species and organism groups are similar, indicating that the evolution of these distributions have been similar.
Fig. 3 (a) Distribution of multi-domain (architecture) family sizes in three selected species. Power law distributions were fitted to these curves such that for frequency <sup>f</sup> of families of size <sup>X</sup>, <sup>f</sup> <sup>¼</sup> cXa . For S. cerevisiae, a ¼ -2.0, for E. coli, a ¼ -1.8, and for H. sapiens, a ¼ -1.5. (b) Distribution of multidomain (architecture) family sizes across the three kingdoms. Power law distributions were fitted to these curves such that for frequency <sup>f</sup> of families of size <sup>X</sup>, <sup>f</sup> <sup>¼</sup> cX<sup>a</sup> . For bacteria, a ¼ -1.0, for archaea, a ¼ -1.1, for eukaryotes, a ¼ -1.1, and for viruses, a ¼ -2.0
#### 3 Kingdom and Age Distribution of Domain Families and Architectures
How old are specific domain families or domain architectures? With knowledge of which organism groups they are found in, it is possible to draw conclusions about their age and whether lineagespecific selective pressures have determined their kingdom-specific abundances. Domain families and their combinations have arisen throughout evolutionary history, presumably by new combinations of pre-existing elements that may have diverged beyond recognition or by processes such as exonization. We can estimate the age of a domain family by finding the largest clade of organisms within which it is found, excluding organisms with only xenologs, i.e., horizontally transferred genes [14]. The age of this lineage's root is the likely age of the family. The same holds true for domain combinations and entire domain architectures. This methodology allows us to determine how changing conditions at different points in evolutionary history, or in different lineages, have affected the evolution of domain architectures.
Apic et al. [36] analyzed the distribution of SCOP domains across 40 genomes from archaea, bacteria, and eukaryotes. They found that a majority of domain families are common to all three kingdoms of life and thus likely to be ancient. Kuznetsov et al. [43] performed a similar analysis using InterPro domains and found that only about one fourth of all such domains were present in all three kingdoms, but a majority was present in more than one of them. Lateral gene transfer or annotation errors can cause a domain family to be found in one or a few species in a kingdom without actually belonging to that kingdom. To counteract this, one can require that a family must be present in at least a reasonable fraction of the species within a kingdom for it to be considered anciently present there. For instance, using Gene3D assignments of CATH domains to 114 complete genomes, mainly bacterial, Ranea et al. [48] isolated protein superfamily domains that were present in at least 90% of all the genomes and at least 70% of the archaeal and eukaryotic genomes, respectively. Under these stringent cutoffs for considering a domain to be present in a kingdom, 140 domains, 15% of the CATH families found in at least one prokaryote genome, were inferred to be ancient. Chothia and Gough [49] performed a similar study on 663 SCOP superfamily domains evaluated at many different thresholds and found that while 516 (78%) superfamilies were common to all three kingdoms at a threshold of 10% of species in each kingdom, only 156 (24%) superfamilies were common to all three kingdoms at a threshold of 90%. They also showed that for prokaryotes, a majority of domain instances (i.e., not domain families but actual domain copies) belong to common superfamilies at all thresholds below 90%.
Extending to domain combinations, Apic et al. [36] reported that a majority of SCOP domain pairs are unique to each kingdom but also that more kingdom-specific domain combinations than expected were composed only of domain families shared between all three kingdoms. This would imply a scenario where the independent evolution of the three kingdoms mainly involved creating novel combinations of domains that existed already in their common ancestor.
Several studies have reported interesting findings on domain architecture evolution in lineages closer to ourselves: in metazoa and vertebrates. Ekman et al. [50] claimed that new metazoaspecific domains and multi-domain architectures have arisen roughly once every 0.1–1 million years in this lineage. According to their results, most metazoa-specific multi-domain architectures are a combination of ancient and metazoa-specific domains. The latter category are however mostly found as novel single-domain proteins. Much of the novel metazoan multi-domain architectures involve domains that are versatile (see below) and exon-bordering (allowing for their insertion through exon shuffling). The novel domain combinations in metazoa are enriched for proteins associated with functions required for multicellularity—regulation, signaling, and functions involved in newer biological systems such as immune response or development of the nervous system, as previously noted by Patthy [23]. They also showed support for exon shuffling as an important mechanism in the evolution of metazoan domain architectures. Itoh et al. [51] added that animal evolution differs significantly from other eukaryotic groups in that lineagespecific domains played a greater part in creating new domain combinations. Nasir et al. [52] analyzed the age and taxonomic distribution of domains drawing on species phylogenies reconstructed from domain repertoires, concluding among other things that most widespread domains are relatively old and suggesting high numbers of both domain gain and loss in the evolution of the three organismal superkingdoms. Bacterial and archaeal genes have tended to gain or lose domains encoding aspects of metabolic capacity, whereas those of eukaryotes—including multicellular ones—have gained domains enabling more elaborate extracellular processes such as immunity and regulatory capacities.
In the most recent datasets, what is the distribution of domains and domain combinations across the three kingdoms of life? Looking at the set of UniProt proteomes represented in version 30.0 of Pfam, the distribution of domains across the three kingdoms are as displayed in the Venn diagram of Fig. 4a. Figure 4b, c show the equivalent distributions of immediate neighbors and triplets of domains, respectively, and Fig. 4d the distribution of multi-domain architectures across kingdoms. The numbers are somewhat biased toward bacteria as 56% of the UniProt proteomes are from this kingdom. However, with this high coverage of all kingdoms
Fig. 4 (a) Kingdom distribution of unique domains. Values are given as percentages of the total, 10,330 domains. (b) Kingdom distribution of unique domain pairs. Values are given as percentages of the total, 31,287 domain pairs. (c) Kingdom distribution of unique domain triplets. Values are given as percentages of the total, 33,662 domain triplets. (d) Kingdom distribution of unique multi-domain architectures. Values are given as percentages of the total, 23,238 multi-domain architectures
(506 eukaryotic, 94 archaeal, and 1090 bacterial proteomes, as well as 253 viral entities), the results should be robust in this respect. Compared to most previous reports, we see a striking difference in that a much smaller portion of domains are shared between all kingdoms. There are some potential artifacts which could affect this analysis. If lateral gene transfer is very widespread, we may overestimate the number of families present in all three kingdoms. Moreover, there are cases where separate Pfam families are actually distant homologs of each other, which could lead to underestimation of the number of ancient families. To counteract this, we make use of Pfam clans, considering domains in the same clan to be equivalent. While not all distant homologies have yet been registered in the clan system, performing the analysis on the clan level reduces the risk of such underestimation.
Our finding that 10% of all Pfam-A domains are present in all three main kingdoms is strikingly lower than in the earlier works and is even lower than reported by Ranea et al. [48], who used very stringent cutoffs. However, a direct comparison of statistics for Pfam domains/clans and CATH superfamilies is difficult. The decrease in ancient families that we observe may be a consequence of the massive increase in sequenced genomes and/or that the recent growth of Pfam has added relatively more kingdom-specific domains. We further found that only 1.5% of all domains or domain combinations are unique to archaea, suggesting that known representatives of this lineage have undergone very little independent evolution and/or that most archaeal gene families have been horizontally transferred to other kingdoms. The trend when going from domain via domain combinations to whole architectures is clear—the more complex patterns are less shared between the kingdoms. In other words, each kingdom has used a common core of domains to construct its own unique combinations of multidomain architectures.
#### 4 Domain Co-occurrence Networks
A multi-domain architecture connects individual domains with each other. There are several ways to derive these connections and quantify the level of co-occurrence. The simplest method is to consider all domains on the same amino acid chain to be connected, but we can also limit the set of co-occurrences we consider to, e.g., immediate neighbor pairs or triplets. Regardless of which method is used, the result is a domain co-occurrence network, where nodes represent domains and where edges represent the existence of proteins in which members of these families co-occur. Figure 5 shows an example of such a network and the set of domain architectures which defines it. This type of explicit network representation is explored in several studies, notably by Itoh et al. [51], Przytycka et al. [53], and Kummerfeld and Teichmann [13]. It is advantageous as it allows the introduction of powerful analysis tools developed within the engineering sciences for use with artificial network structures such as the World Wide Web. The patterns of co-occurrences that we observe should be a direct consequence of the constraints and conditions under which domain architectures evolve, and because of this, the study of these patterns becomes relevant for understanding such factors.
The frequency distribution of node degrees in the domain co-occurrence network has been fitted to a power law [36] and a more general GPD as well [40]. The closer this approximation holds, the more the network will have the scale-free property.
Fig. 5 Example of protein domain co-occurrence network, adapted from Kummerfeld and Teichmann [13]. (a) Sample set of domain architectures. The lines represent proteins and the boxes their domains in N- to C-terminal order. (b) Resulting domain co-occurrence (neighbor) network. Nodes correspond to domains and are linked by an edge if at least one domain exists where the two domains are found adjacent to each other along the amino acid chain
This property can be thought of as a hierarchy in the network, where the more centrally connected nodes link to more peripheral nodes with the same relative frequency at each level. In the context of domains, this means that a small number of domains co-occur with a high number of other domains, whereas most domains only have a few neighbors—usually some of the highly connected hubs. The most highly connected domains are referred to as promiscuous [54], mobile, or versatile [14, 55, 56]. Many such hub domains are involved in intracellular or extracellular signaling, protein-protein interactions and catalysis, and transcription regulation. In general, these are domains that encode a generic function, e.g., phosphorylation, which is reused in many contexts by additional domains that Table 1
The 20 most densely connected hubs with regard to immediate domain neighbors, according to Pfam 30.0
confer substrate specificity or localization. Table 1 shows the domains (or clans) with the highest numbers of immediate neighbors in Pfam 30.0.
One way of evolving a domain co-occurrence network that follows a power law is by "preferential attachment" [53, 57]. This means that new edges (corresponding to proteins where two domains co-occur) are added with a probability that is higher the more edges these nodes (domains) already have, resulting in a power law distribution.
Apic et al. [37] considered a null model for random domain combination, in which a proteome contains domain combinations with a probability based on the relative abundances of the domains only. They showed that this model does not hold and that far fewer domain combinations than expected under it are actually seen. If most domain duplication events are gene duplication events that do not change domain architecture—or at the very least do not disrupt domain pairs—then this finding is not unexpected, nor does it require or exclude any particular selective pressure to keep these domains together in proteins. There is growing support for the idea that separate instances of a given domain architecture in general descend from a single ancestor with that architecture [58], with polyphyletic evolution of domain architectures occurring only in a small fraction of cases [53, 59, 60].
Itoh et al. [51] performed reconstruction of ancestral domain architectures using maximum parsimony, as described in the next section. This allowed them to study the properties of the ancestral domain co-occurrence network and thus explore how network connectivity has altered over evolutionary time. Among other things, they found increased connectivity in animals, particularly of animal-specific domains, and suggest that this phenomenon explains the high connectivity for eukaryotes reported by Wuchty [40]. For non-animal eukaryotes, they reported a correlation between connectivity and age, such that older domains had relatively higher connectivity, with domains preceding the divergence of eukaryotes and prokaryotes being the most highly connected, followed by early eukaryotic domains. In other words, early eukaryotic evolution saw the emergence of some key hub proteins, while the most prominent eukaryotic hubs emerged in the animal lineage. Parikesit et al. [61] studied the functional annotation of co-occurring domains in eukaryotes, concluding that while these may have different associated functional descriptors, these descriptors usually tend to fall within the same overall category within the gene ontology. Co-occurring domains thus tend to contribute to the same overall process type rather than have very widely divergent functional annotations. Hsu et al. [62] constructed a network linking domain architectures (i.e., each node is a multi-domain architecture, as opposed to in a regular domain co-occurrence network) where parsimonious reconstruction suggests evolution of one from the other, identifying "highly evolvable" architectures as hubs in this network. Proteins with such architectures were reported to be more widespread, less often essential, more often duplicated, and more often associated with gene functions involved in specific adaptation of organisms.
What is the degree distribution of current domain co-occurrence networks? We again used the domain architectures from all complete proteomes in version 30.0 of Pfam and considered the network of immediate neighbor relationships, i.e., nodes (domains) have an edge between them if there is a protein where they are adjacent. Each domain was assigned a degree as its number
Fig. 6 (a) Distribution of domain co-occurrence network node degrees in three selected species. Power law distributions were fitted to these curves such that for frequency <sup>f</sup> of families of size <sup>X</sup>, <sup>f</sup> <sup>¼</sup> cXa . For S. cerevisiae, a ¼ -2.2, for E. coli, a ¼ -2.0, and for H. sapiens, a ¼ -1.9. (b) Distribution of domain co-occurrence network node degrees across the three kingdoms. This corresponds to a network where two domains are connected if any species within the kingdom has a protein where these domains are immediately adjacent. Power law distributions were fitted to these curves such that for frequency f of families of size X, <sup>f</sup> <sup>¼</sup> cXa . For bacteria, a ¼ -1.6, for archaea, a ¼ -1.7, for eukaryotes, a ¼ -1.5, and for viruses a ¼ -2.0
of links to other domains. We then counted the frequency with which each degree occurs in the co-occurrence network. Figure 6a shows this relationship for the set of domain architectures found in the same species as for Figs. 2a, and 6b shows the equivalent plots for the three kingdoms as found among the complete proteomes in Pfam. Regressions to a power law have been added to the plots. The presence of a power law-like behavior of this type implies that few domains have very many immediate neighbors, while most domains have few immediate neighbors. Note that the observed degrees in our dataset were strongly reduced by removing all sequences with a stretch longer than 50 amino acids lacking domain annotation.
#### 5 Supra-domains and Conserved Domain Order
As we have seen, whole multi-domain architectures or shorter stretches of adjacent domains are often repeated in many proteins. These only cover a small fraction of all possible domain combinations. Are the observed combinations somehow special? We would expect selective pressure to retain some domain combinations but not others, since only some domains have functions that would synergize together in one protein. Often, co-occurring domains require each other structurally or functionally, for instance, in transcription factors where the DNA-binding domain provides substrate specificity, whereas the trans-activating domain recruits other components of the transcriptional machinery [63]. Vogel et al. [38] identified series of domains co-occurring as a fixed unit with conserved N- to C-terminal order but flanked by different domain architectures and termed them supra-domains. By investigating their statistical overrepresentation relative to the frequency of the individual domains in the set of nonredundant domain architectures (where "nonredundant" is crucial, as otherwise, e.g., whole-gene duplication would bias the results), they identified a number of such supra-domains. Many ancient domain combinations (shared by all three kingdoms) appear to be such selectively preserved supra-domains.
How conserved is the order of domains in multi-domain architectures? In a recent study, Kummerfeld and Teichmann [13] built a domain co-occurrence network with directed edges, allowing it to represent the order in which two domains are found in proteins. As in other studies, the distribution of node degrees fits a power law well. Most domain pairs were only found in one orientation. This does not seem required for functional reasons, as flexible linker regions should allow the necessary interface to form also in the reversed case [58], but may rather be an indication that most domain combinations are monophyletic. Weiner and Bornberg-Bauer [64] analyzed the evolutionary mechanisms underlying a number of reversed domain order cases and concluded that independent fusion/fission is the most frequent scenario. Although domain reversals occur in only a few proteins, it actually happens more often than was expected from randomizing a co-occurrence network [13]. That study also observed that the domain co-occurrence network is more clustered than expected by a random model and that these clusters are also functionally more coherent than would be expected by chance.
#### 6 Domain Mobility, Promiscuity, or Versatility
While some protein domains co-occur with a variety of other domains, some are always seen alone or in a single architecture in all proteomes where they are found. A natural explanation is that some domains are more likely to end up in a variety of architectural contexts than others due to some intrinsic property they possess. Is such domain versatility or promiscuity a persistent feature of a given domain, and does it correlate with certain functional or biological properties of the domain?
Several ways of measuring domain versatility have been suggested. One measure, NCO [40], counts the number of other domains found in any architectures where the domain of interest is found. Another measure, NN [37], instead counts the number of distinct other domains that a domain is found adjacent to. Yet another measure, NTRP [65], counts the number of distinct triplets of consecutive domains where the domain of interest is found in the middle. All of these measures can be expected to be higher for common domains than for rare domains, i.e., variations in domain abundance (the number of proteins a domain is found in) can hide the intrinsic versatility of domains. Therefore, three different studies [14, 55, 66] formulated relative domain versatility indices that aim to measure versatility independently of abundance. It is worth noting that most studies have considered only immediately adjacent domain neighbors in these analyses, a restriction based on the assumption that those are more likely to interact functionally than domains far apart on a common amino acid chain. More recent work [67] introduced a network versatility metric which can classify domains as being central or peripheral with regard to the large-scale structure of their bigram network (i.e., the network-linking domains found adjacent in proteins), observing how peripheral such domains exhibit relatively higher primary sequence conservation suggestive of adaptation to more specific functions, whereas the core domains may be more multifunctional.
The first relative versatility study was presented by Vogel et al. [66], who used as their domain dataset the SUPERFAMILY database applied to 14 eukaryotic, 14 bacterial, and 14 archaeal proteomes. They modeled the number of unique immediate neighbor domains as a power law function of domain abundance, performed a regression on this data, and used the resulting power law exponent as a relative versatility measure. Basu et al. [55] used Pfam and SMART [8] domains and measured relative domain versatility for 28 eukaryotes as the immediate neighbor pair frequency normalized by domain frequency. They then defined promiscuous domains as a class according to a bimodality in the distribution of the raw numbers of unique domain immediate neighbor pairs. Weiner et al. [14] used Pfam domains for 10,746 species in all kingdoms and took as their relative versatility measure the logarithmic regression coefficient for each domain family across genomes, meaning that it is not defined within single proteomes.
To what extent is high versatility an intrinsic property of a certain domain? Vogel et al. [66] only examined large groups of domains together and therefore did not address this question for single domains. Basu et al. [55] and Weiner et al. [14] instead analyzed each domain separately and concluded that there are strong variations in relative versatility at this level. Their results are very different in detail, however, reflected by the fact that only one domain family (PF00004, AAA ATPase family) is shared between the ten most versatile domains reported in the two studies. As they used fairly similar domain datasets, it would appear that the results strongly depend on the definition of relative versatility. Another potential reason for the different results is that Basu's list was based on eukaryotes only, while Weiner's analysis was heavily biased toward prokaryotes. Furthermore, the top ten list in Basu et al. [55] and their follow-up paper [56] only overlap by four domains, yet the main difference is that in the latter study all 28 eukaryotes were considered, while the former study was limited to the subset of 20 animal, plant, and fungal species. The choice of species thus seems pivotal for the results when using this method. They also used different methods for calculating the average value of relative versatility across many species, which may influence the results.
Does domain versatility vary between different functional classes of domains? Vogel et al. [66] found no difference in relative versatility between broad functional or process categories or between SCOP structural classes. In contrast to this, Basu et al. [55] reported that high versatility was associated with certain functional categories in eukaryotes. However, no test for the statistical significance of these results was performed. Weiner et al. [14] also noted some general trends but found no significant enrichment of gene ontology terms in versatile domains. This does not necessarily mean that no such correlation exists, but more research is required to convincingly demonstrate its strength and its nature. More recently, Cromar et al. [68] analyzed domain architectures in eukaryotic extracellular matrix proteomes, noting that these structures are organized around a set of versatile domains under the weighted bigram metric of Basu et al. [55].
Another important question is to what extent domain versatility varies across evolutionary lineages. Vogel et al. [66] reported no large differences in average versatility for domains in different kingdoms. The versatility measure of Basu et al. [55] can be applied within individual genomes, which means that according to this measure domains may be versatile in one organism group but not in another, as well as gain or lose versatility across evolutionary time. They found that more domains were highly versatile in animals than in other eukaryotes. Modeling versatility as a binary property defined for domains in extant species, they further used a maximum parsimony approach to study the persistence of versatility for each domain across evolutionary time and concluded that both gain and loss of versatility are common during evolution. Inferring ancestral domain architectures, Cohen-Gihon et al. [69] report an increase in versatility in many domains during eukaryotic evolution, in particular around the divergence of Bilateria. Weiner at al. [14] divided domains into age categories based on distribution across the tree of life and reported that the versatility index is not dependent on age, i.e., domains have equal chances of becoming versatile at different times in evolution. This is consistent with the observation by Basu et al. [55] that versatility is a fast-evolving and varying property. When measuring versatility as a regression within different organism groups, Weiner et al. [14] found slightly lower versatility in eukaryotes, which is in conflict with the findings of Basu et al. [55]. Again, this underscores the strong dependence of the method and dataset on the results.
Further properties reported to correlate with domain versatility include sequence length, where Weiner et al. [14] found that longer domains are significantly more versatile within the framework of their study, while at the same time, shorter domains are more abundant and hence may have more domain neighbors in absolute numbers. Basu et al. [55] further reported that more versatile domains have more structural interactions than other domains. To determine which of these reported correlations that genuinely reflect universal biological trends, further comprehensive studies are needed using more data and uniform procedures. This would hopefully allow the results from the studies described here to be validated and any conflicts between them to be resolved.
Basu et al. [55] further analyzed the phylogenetic spread of all immediate domain neighbor pairs ("bigrams") containing domains classified as promiscuous. The main observation this yielded was that although most such combinations occurred in only a few species, most promiscuous domains are part of at least one combination that is found in a majority of species. They interpreted this as implying the existence of a reservoir of evolutionarily stable domain combinations from which lineage-specific recombination may draw promiscuous domains to form unique architectures. Later work by Hsu et al. [70] analyzed the domain co-occurrence networks centered on each domain family, classifying such subnetworks as being either mostly starlike, taillike, or tetragon-like, with promiscuous domains forming cores of starlike architecture networks in this representation.
#### 7 Principles of Domain Architecture Evolution
What mutation events can generate new domain architectures, and what is their relative predominance? The question can be approached by comparing protein domain architectures of extant proteins. This is based on the likely realistic assumption that most current domain architectures evolved from ancestral domain architectures that can still be found unchanged in other proteins. Because of this, in pairs of most similar extant domain architectures, one can assume that one of them is ancestral. This agrees well with results indicating that most groups of proteins with identical domain architectures are monophyletic. By comparing the most similar proteins, several studies have attempted to chart the relative frequencies of different architecture-changing mutations.
Bjo¨rklund et al. [71] used this particular approach and came to several conclusions. First, changes to domain architecture are much more common by the N- and C-termini than internally in the architecture. This is consistent with several mechanisms for architecture changes such as introduction of new start or stop codons or mergers with adjacent genes, and similar results have been found in several other studies [15, 25, 26]. Furthermore, insertions or deletions of domains ("indels") are more common than substitutions of domains, and the events in question mostly concern just single domains, except in cases with repeats expanding with many domains in a row [72]. In a later study, the same group made use of phylogenetic information as well, allowing them to infer directionality of domain indels [50]. They then found that domain insertions are significantly more common than domain deletions.
Weiner et al. [26] performed a similar analysis on domain loss and found compatible results—most changes occur at the termini (see also discussion in [28]). Moreover, they demonstrated that terminal domain loss seldom involves losing only part of a domain, or rather, that such partial losses quickly progress into loss of the entire domain. However, it is important to ensure such observations are not confounded by cases where errors in gene boundary recognition make domain detection less accurate [73].
There is some support [23, 74, 75] for exon shuffling to have played an important part in domain evolution, and there are a number of domains that match intron borders well, for example, structural domains in extracellular matrix proteins. While it may not be a universal mechanism, exon shuffling is suggested to have been particularly important for vertebrate evolution [23].
Recognizing the potential role of gene duplications in domain architecture evolution, Grassi et al. [76] analyzed domain architecture shifts following either whole-genome duplication (WGD) or smaller-scale gene duplication events in yeast. Surviving WGD duplicates had retained ancestral architecture in ca 95% of cases, with approximately the same chance of architecture change in WGD as under local duplication. Genes retained over time from either type of duplication were enriched for a core of commonly occurring domains but with a subset of rarer domains additionally enriched in retained WGD duplicates compared to locally duplicated genes. The former category more often was associated with housekeeping-type gene functions, whereas the latter more often involved adaptive functions. Functional change was generally larger than architectural change following duplication. Zhang et al. [77] similarly studied domain architecture evolution in plants, noting that lineage-specific architecture expansions largely can be explained from differential retention of genes following successive whole-genome duplications. Another form of domain duplication particularly relevant in plants is amplification of the numbers of domain repeats in proteins, discussed, e.g., by Sharma and Pandey [78].
#### 8 Inferring Ancestral Domain Architectures
The above analyses, based on pairwise comparison of extant protein domain architectures, cannot tally ancestral evolutionarily events nearer the root of the tree of life. With ancestral architectures, one can directly determine which domain architecture changes have taken place during evolution and precisely chart how mechanisms of domain architecture evolution operate, as well as gauge their relative frequency. A drawback is that since we can only infer ancestral domain architectures from extant proteins, the result will depend somewhat on our assumptions about evolutionary mechanisms. On the upside, it should be possible to test how well different assumptions fit the observed modern-day protein domain architecture patterns.
Attempts at such reconstructions have been made using parsimony. Given a gene tree and the domain architectures at the leaves, dynamic programming can be used in order to find the assignment of architectures to internal nodes that require the smallest number of domain-level mutation events. This simple model can be elaborated by weighting loss and gain differently or by requiring that a domain or an architecture can only be gained at most once in a tree (Dollo parsimony) [79].
An early study of Snel et al. [80] considered 252 gene trees across 17 fully sequenced species and used parsimony to minimize the number of gene fission and fusion events occurring along the species tree. Their main conclusion, that gene fusions are more common than gene fissions, was subsequently supported by a larger study by Kummerfeld and Teichmann [81], where fusions were found to be about four times as common as fissions in a most parsimonious reconstruction. Fong et al. [82] followed a similar procedure on yet more data and concluded that fusion was 5.6 times as likely as fission.
Buljan and Bateman [15] performed a similar maximum parsimony reconstruction of ancestral domain architectures. They too observed that domain architecture changes primarily take place at the protein termini, and the authors suggested that this might largely occur because terminal changes to the architecture are less likely to disturb overall protein structure. Moreover, they concluded from reconciliation of gene and species trees that domain architecture changes were more common following gene duplications than following speciation but that these cases did not differ with respect to the relative likelihood of domain losses or gains.
Recently, Buljan et al. [25] presented a new ancestral domain architecture reconstruction study which assumed that gain of a domain should take place only once in each gene tree, i.e., Dollo parsimony [79]. Their results also support gene fusion as a major mechanism for domain architecture change. The fusion is generally preceded by a duplication of either of the fused genes. Intronic recombination and insertion of exons are observed but relatively rarely. They also found support for de novo creation of disordered segments by exonization of previously noncoding regions. More recently still a method for domain architecture history reconstruction using a network construct called a plexus was described [83]. Yang and Bourne [84] further described another parsimony-based reconstruction approach, as did Wu et al. [85], reporting that histories of signaling and development proteins are enriched for gene fusion/fission events. Stolzer et al. [86] present another method for domain architecture history inference, made available through the Notung software.
#### 9 Polyphyletic Domain Architecture Evolution
There appears to be a "grammar" for how protein domains are allowed to be combined. If nature continuously explores all possible domain combinations, one would expect that the allowed combinations would be created multiple times throughout evolution. Such independent creation of the same domain architecture can be called convergent or polyphyletic evolution, whereas a single original creation event for all extant examples on an architecture would be called divergent or monophyletic evolution. This is relevant for several reasons, not least because it determines whether or not we can expect two proteins with identical domain architectures to have the same history along their entire length.
A graph theoretical approach to answer this question was taken by Przytycka et al. [53], who analyzed the set of all proteins containing a given superfamily domain. The domain architectures of these proteins define a domain co-occurrence network, where edges connect two domains both found in a protein, regardless of sequential arrangement. The proteins of such a set can also be placed in an evolutionary tree, and the evolution of all multidomain architectures containing the reference domain can be expressed in terms of insertions and deletions of other domains along this tree to form the extant domain architectures. The question, then, is whether or not all leaf nodes sharing some domain arrangement (up to and including an entire architecture) stem from a single ancestral node possessing this combination of domains. For monophyly to be true for all architectures containing the reference domain, the same companion domain cannot have been inserted in more than one place along the tree describing the evolution of the reference domain. By application of graph theory and Dollo parsimony [79], they showed that monophyly is only possible if the domain co-occurrence network defined by all proteins containing the reference domain is chordal, i.e., it contains no cycles longer than three edges.
Przytycka et al. [53] then evaluated this criterion for all superfamily domains in a large-scale dataset. For domains where the co-occurrence network contained fewer than 20 nodes (domains), the chordal property and hence the possibility of complete monophyly of all domain combinations and domain architectures containing that domain held. By comparing actual domain co-occurrence networks with a preferential attachment null model, they showed that far more architectures are potentially monophyletic than would be expected under a pure preferential attachment process. This finding is analogous to the observation by Apic et al. [37] that most domain combinations are duplicated more frequently (or reshuffled less) than expected by chance. In other words, gene duplication is much more frequent than domain recombination [66]. However, for many domains that co-occurred with more than 20 other different domains, particularly for domains previously reported as promiscuous, the chordal property was violated, meaning that multiple independent insertions of the same domain, relative to the reference domain phylogeny, must be assumed.
A more direct approach is to do complete ancestral domain architecture reconstruction of protein lineages and to search for concrete cases that agree with polyphyletic architecture evolution. There are two conceptually different methodologies for this type of analysis. Either one only considers architecture changes between nodes of a species tree, or one considers any node in a reconstructed gene tree. The advantage of using a species tree is that one avoids the inherent uncertainty of gene trees, but on the other hand, only events that take place between examined species can be observed.
Gough [59] applied the former species-tree-based methodology to SUPERFAMILY domain architectures and concluded that polyphyletic evolution is rare, occurring in 0.4–4% of architectures. The value depends on methodological details, with the lower bound considered more reliable.
The latter gene-tree-based methodology was applied by Forslund et al. [60] to the Pfam database. Ancestral domain architectures were reconstructed through maximum parsimony of singledomain phylogenies which were overlaid for multi-domain proteins. This strategy yielded a higher figure, ranging between 6% and 12% of architectures depending on dataset and whether or not incompletely annotated proteins were removed. The two different approaches thus give very different results. The detection of polyphyletic evolution is in both frameworks dependent on the data that is used—its quality, coverage, filtering procedures, etc. The studies used different datasets which makes it hard to compare. However, given that their domain annotations are more or less comparable, the major difference ought to be the ability of the gene-tree method to detect polyphyly at any point during evolution, even within a single species. It should be noted that domain annotation is by no means complete—only a little less than half of all residues are assigned to a domain [5]—and this is clearly a limiting factor for detecting architecture polyphyly. The numbers may thus be adjusted considerably upwards when domain annotation reaches higher coverage. A later study by Zmasek and Godzik [87] reports much higher rates (25–75%) still of polyphyletic evolution of eukaryotic multi-domain architectures, arguing that previous datasets were too small to have the power to reveal this.
Future work will be required to provide more reliable estimates of how common polyphyletic evolution of domain architectures is. Any estimate will depend on the studied protein lineage, the versatility of the domains, and methodological factors. A comprehensive and systematic study using more complex phylogenetic methods than the fairly ad hoc parsimony approach, as well as effective ways to avoid overestimating the frequency of polyphyletic evolution due to incorrect domain assignments or hidden homology between different domain families, may be the way to go. At this point all that can be said is that polyphyletic evolution of domain architectures definitely does happen, but relatively rarely, and that it is more frequent for complex architectures and versatile domains. A detailed case study was made recently of netrin domaincontaining proteins, where polyphyletic evolution in metazoa seems well-supported [88]; these authors further suggest the term merology for such polyphyletic evolution. A series of papers by Nagy and Patthy et al. [73, 89, 90] further elaborates on challenges faced within this line of research; they report strong confounding influence of gene prediction errors. They further propose the term epaktology for gene similarity resulting from the independent acquisition of two proteins by the same additional domain. The authors suggest such cases inflate both estimates of terminal domain changes and estimates of gene fusion-driven changes in domain architecture. Beyond such changes, whether correctly inferred or not, the authors describe internal domain shuffling as an important mechanism for how domain architecture evolution has occurred.
#### 10 Conclusions
As access to genomic data and to increasing amounts of compute power has grown during the last decade-and-a-half, so has our knowledge of the overall patterns of domain architecture evolution. Still, no study is better than its underlying assumptions, and differences in the representation of data and hypotheses mean that results often cannot be directly compared. Overall, however, the current state of the field appears to support some broad conclusions.
Domain and multi-domain family sizes, as well as numbers of co-occurring domains, all approximately follow power laws, which implies a scale-free hierarchy. This property is associated with many biological systems in a variety of ways. In this context, it appears to reflect how a relatively small number of highly versatile components have been reused again and again in novel combinations to create a large part of the domain and domain architecture repertoire of organisms. Gene duplication is the most important factor to generate multi-domain architectures, and as it outweighs domain recombination, only a small fraction of all possible domain combinations is actually observed. This is probably further modulated by family-specific selective pressure, though more work is required to demonstrate to what extent. Most of the time, all proteins with the same architecture or domain combination stem from a single ancestor where it first arose, but there remains a fraction of cases, particularly with domains that have very many combination partners, where this does not hold.
Most changes to domain architectures occur following a gene duplication and involve the addition of a single domain to either protein terminus. The main exceptions to this occur in repeat regions. Exon shuffling played an important part in animals by introducing a great variety of novel multi-domain architectures, reusing ancient domains as well as domains introduced in the animal lineage.
In this chapter, we have reexamined with the most up-to-date datasets many of the analyses done previously on less data and found that the earlier conclusions still hold true. Even though we are at the brink of amassing enormously much more genome and proteome data thanks to the new generation of sequencing technology, there is no reason to believe that this will alter the fundamental observations we can make today on domain architecture evolution. However, it will permit a more fine-grained analysis, and also there will be a greater chance to find rare events, such as independent creation of domain architectures. Furthermore, careful application of more complex models of evolution with and without selection pressure may allow us to determine more closely to what extent the process of domain architecture evolution was shaped by selective constraints.
#### 11 Materials and Methods
Updated statistics were generated from the data in Pfam 30.0. All UniProt proteins in the SwissPfam set for Pfam 30.0 were included. These span 1090 bacteria, 506 eukaryotes, and 94 archaea. All Pfam-A domains regardless of type were included. However, as stretches of repeat domains are highly variable, consecutive subsequences of the same domain were collapsed into a single pseudodomain, if it was classified as type Motif or Repeat, as in several previous works [50, 60, 66, 82].
Domains were ordered within each protein based on their sequence start position. In the few cases of domains being inserted within other domains, this was represented as the outer domain followed by the nested domain, resulting in a linear sequence of domain identifiers. As long regions without domain assignments are likely to represent the presence of as-yet uncharacterized domains, we excluded any protein with unassigned regions longer than 50 amino acids (more than 95% of Pfam-A domains are longer than this). This approach is similar to that taken in previous works [59, 60, 71]. Other studies [50, 72] have instead performed additional, more sensitive domain assignment steps, such as clustering the unassigned regions to identify unknown domains within them.
Pfam domains are sometimes organized in clans, where clanmates are considered homologous. A transition from a domain to another of the same clan is thus less likely to be a result of domain swapping of any kind and more likely to be a result of sequence divergence from the same ancestor. Because of this, we replaced all Pfam domains that are clan members with the corresponding clan.
The statistics and plots were generated using a set of Perl and R scripts, which are available upon request. Power law regressions were done using the R nls function. For reasons of scale, the regression for a power law relation such as
$$N = cX^{-a}$$
was performed on the equivalent relationship
$$
\log(X) = (1/a)(\log(c) - \log(N))
$$
for the parameters a and c, with the exception of the data for Fig. 6, where instead the relationship
$$
\log(N) = \log(\iota) - \mathfrak{a}\log(X)
$$
was used. Moreover, because species or organism group datasets were of very different size, raw counts of domains were converted to frequencies before the regression was performed.
#### 12 Online Domain Database Resources
For further studies or research into this field, the first and most important stop will be the domain databases. Table 2 presents a selection of domain databases in current use.
#### Table 2 A selection of protein domain databases
#### 13 Domain Architecture Analysis Software
Several software tools have been described and made available that allow for analysis and visualization of domain architectures and their evolution. A selection of such tools is shown in Table 3.
A few of these tools allow domain architecture evolution analysis by visualizing each protein's domain architecture along a protein sequence tree. An example is the web tool TreeDom [96] which, given a protein domain family and an anchor sequence, fetches the family from Pfam and builds a tree with the nearest neighbors of the anchor sequence. An example output from TreeDom is shown in Fig. 7, in which a nonredundant set of representative proteomes were queried. Here one can see that while the NUDIX domain of the anchor sequence tends to co-occur with two other domains (zf-NADH-PPase and NUDIX-like), it also has recombined with many other domains over the course of evolution.
Other tools allow different types of analyses, for instance, searching for similar domain architectures or showing taxonomic distributions. Some of the protein domain databases listed in Table 2 include variants of such analyses, while external tools typically offer more specialized functionality. For example, the Pfam website allows searching for domain content, while the java tool PfamAlyzer allows searching Pfam for particular domain architecture patterns specified with a given domain order and spacing [94].
The RAMPAGE/RADS tools [95] make use of domain assignments for rapid homology searching. DoMosaics [92] is a software
Fig. 7 TreeDom output using as query the NUDIX domain (PF00293), the human NUDT12 (Q9BQG2) protein, 30 closest sequences, and RP15 (representative proteomes at 15% co-membership). The domains are green, NUDIX; blue, NUDIX-like (PF09296); yellow, zf-NADH-PPase (PF09297); red, Ocnus (PF05005); cyan, Ank\_2 (PF12796); black, Ank\_5 (PF13857); orange, Prefoldin (PF02996); and pink, Fibrinogen\_C (PF00147)
tool that can act as a wrapper for domain annotation tools, allowing detailed visualization and analysis of domain architectures, as does DomArch [97]. The DAAC algorithm [98] explicitly transfers functional annotation to query sequences based on domain architectural similarity to annotated homologs, as does FACT [93]. In the same vein, similarity measures between architectures are available using the WDAC [99] tool and in ADASS [100]. Domain architecture similarity is used for orthology detection in the porthoDom software [68]. The DOGMA tool makes use of domain content data to assess completeness of a proteome or transcriptome [101].
#### 14 Exercises/Questions
#### References
InterPro in 2017—beyond protein family and domain annotations. Nucleic Acids Res 45 (D1):D190–D199
understand their sequences. Nucleic Acids Res 43(D1):D382–D386
definition, properties, and implications. Internet Math 2(4):431–523
inferred ancestral domain architectures. Mol Biosyst 7(3):784–792
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
#### New Insights on the Evolution of Genome Content: Population Dynamics of Transposable Elements in Flies and Humans
#### Lain Guio and Josefa Gonza´ lez
#### Abstract
Understanding the abundance, diversity, and distribution of TEs in genomes is crucial to understand genome structure, function, and evolution. Advances in whole-genome sequencing techniques, as well as in bioinformatics tools, have increased our ability to detect and analyze the transposable element content in genomes. In addition to reference genomes, we now have access to population datasets in which multiple individuals within a species are sequenced. In this chapter, we highlight the recent advances in the study of TE population dynamics focusing on fruit flies and humans, which represent two extremes in terms of TE abundance, diversity, and activity. We review the most recent methodological approaches applied to the study of TE dynamics as well as the new knowledge on host factors involved in the regulation of TE activity. In addition to transposition rates, we also focus on TE deletion rates and on the selective forces that affect the dynamics of TEs in genomes.
Key words Long-read sequencing, Transposition rates, Self-regulation, Effective population size, Adaptation, Horizontal transfer
#### 1 Transposable Elements Are Abundant and Active Genome Denizens
Transposable elements (TEs) are short DNA sequences, typically from a few hundred bp to ~10 kb long, which have the ability to move around in the genome by generating new copies of themselves. In addition to active autonomous elements, genomes also contained nonautonomous elements that can be mobilized by the enzymatic machinery of active TEs from the same family. Additionally, genomes contain TEs that cannot be mobilized anymore due to accumulation of mutations in their sequences [1]. TEs are an ancient, extremely diverse, and exceptionally active component of genomes. TEs have been found in virtually all organisms studied so far including bacteria, archaea, fungi, protists, plants, and animals [2–5]. The main TE groups, class I and class II, are present in all kingdoms, revealing their persistence over evolutionary time
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_16, © The Author(s) 2019
[2]. These two classes of TEs differ in their transposition intermediates: while class I TEs transpose through RNA intermediates, class II TEs transpose directly as DNA. TEs within each class are further classified into (1) different orders, based on their insertion mechanism, structure, and encoded proteins; (2) different superfamilies, based on their replication strategy and on presence and size of target site duplications; and (3) different families, based on sequence conservation [2, 3]. Piegu et al. [1] criticized the current classification system, which accounts for sequence homology, structural features, and target site duplications, because it does not always take into account the evolutionary origins of the TEs [1–3]. As a consequence, phylogenetically unrelated classes or subclasses of TEs are grouped [1]. Piegu et al. [1] also suggested that a more inclusive classification that includes prokaryotic and eukaryotic TE classes should be considered. Recently, Arkhipova [6] proposed a TE classification system based on the replicative, integrative, and structural components of TEs, which integrates different aspects of all the existing classification systems [6].
TEs constitute a substantial albeit variable (from ~1% to almost 90%) proportion of genomes [7, 8] (Fig. 1). The identification methods, as well as the sequencing and assembly methods, have an important effect in the TE content estimation [4, 9–11]. In some cases, the TE-generated fraction of genomes is likely to be underestimated because methods for detecting TEs in genomic sequences are necessarily biased toward younger and more easily recognizable TEs. Indeed, new tools developed in recent years are able to identify TEs that remained hidden until now [4, 11]. As an example, when the human genome was first sequenced, ~40–45% of the genome was identifiable TEs, 5% was genes and other functional sequences (functional RNAs or regulatory regions), and the remaining ~50% of the genome had no identifiable origin [12]. de Koning et al. [13] using a highly sensitive new strategy named P-cloud found that at least 66–69% of the human genome is identifiable as repetitive sequences, most of them derived from TEs [13]. In Drosophila melanogaster, third-generation sequencing techniques (3GS) have allowed the detection of 37% more TE insertions in chromosome 2L compared to previously available short-read sequencing estimates (see below) [14]. In other Drosophila species such as D. buzzatii, the TE content has also been updated from 6% to 11%, thanks to the recent availability of wholegenome sequences [15].
As mentioned above, TEs are extremely active genomic denizens that are able to generate mutations of a great diversity of types [16–21]. TE-induced mutations range from subtle regulatory mutations to gross genomic rearrangements and often have phenotypic effects of a complexity that is not achievable by point mutations (Fig. 2). Among others, TEs can affect the expression of nearby genes by adding new splice sites, adenylation signals,
Fig. 1 TE content in the genome of different organisms expressed as percentage of the genome: Homo sapiens (~45% [12], >66% [13]) Mus musculus [143], Saccharomyces cerevisiae [144], Arabidopsis thaliana [145], Pyrococcus furiosus [146], Clostridium difficile [147], Danio rerio [133], Kryptolebias marmoratus [148], Bombyx mori [149], Hypothenemus hampei [150], Drosophila melanogaster (11%, [68], ~20% [69]), Pseudozyma antarctica, and Laccaria bicolor [151]. Zea mays [152] and Fritillaria imperialis [8]. All estimates were obtained with homology-based methods except [13] that uses P-cloud and [69] that uses de novo approaches
promoters, or transcription factor binding sites [22–24]. TEs can also be targets of epigenetic histone modifications that spread into adjacent genes affecting their expression [25, 26]. In addition to transcriptional changes, TEs have been shown to affect translation regulation when they are transcribed within a mRNA [27–29], to contribute to protein-coding regions both at the transcript and at the protein level [30–35], and TE-encoded proteins have been domesticated and are part of host genes [17, 36–40]. TE excision can lead to DNA deletions [41], and TE insertion can result in adding DNA through 3<sup>0</sup> and, less frequently, through 5<sup>0</sup> transduction [42, 43]. Finally, ectopic recombination between TEs causes
Fig. 2 Effects of TEs on the host genome. (a) TEs can affect the expression and/or structure of genes. Exons are represented as blue boxes and TEs as green boxes. (1) A TE inserted in the upstream region of a gene can add insulator sequences, transcription factor-binding site (TFBS), or can disrupt an existing promoter gene; (2) A TE inserted in an intron can truncate the mRNA or induce alternative splicing; (3) A TE inserted in the downstream region of a gene can add microRNA binding sites or alter the polyadenylation site; (4) A TE
deletions, duplications, and sequence rearrangements. Two recent studies in the human genome identified 516 chromosome rearrangements potentially generated by LINE-LINE nonallelic homologous recombination and 78 HERV-mediated rearrangements [44, 45]. Both studies used the annotations of LINEs and HERVs in the reference genome and look for evidence of rearrangements induced by these TEs using clinical databases of copy number variants containing information from thousands of patients. In addition to being associated with diseases [24, 46–49], the number of TE-induced mutations associated with positive effects on fitness-related traits also continues to increase both in humans and in Drosophila [50–63].
Overall, recent advances in sequencing technologies and in TE detection methods showed that, as expected, the TE content is higher than previously estimated. These new data also provided further evidence for the impact of TEs in genome function and genome structure. Thus, it is still indisputable that a thorough understanding of TE population dynamics is essential for the understanding of the eukaryotic genome structure, function, and evolution.
#### 2 Drosophila and Humans: Two Extremes in TE Diversity and Population Dynamics
Much of the detailed information on TE evolution still comes from two species with the best-studied genomes: fruit flies (D. melanogaster) and humans. Fortunately, these two genomes represent two extremes in terms of TE diversity and population dynamics and thus give a reasonably diverse picture of the TE evolution and dynamics. For the rest of this chapter, we focus
Fig. 2 (continued) inserted in the exon of a gene can lead to exonization of the TE or to transcript truncation; (5) the whole domain of a TE protein could insert in the coding region of a gene generating a chimeric gene with host and TE domains [5, 21]. In addition to these changes that depend on where the TE is inserted and on the sequences that the TE is adding, TEs can also alter the posttranslational modifications of histones. (b) TEs could also induce translation repression by generating secondary structure in the 3<sup>0</sup> UTR of genes that leads to changes in the localization of the mRNA. This secondary structure could bind to one of the protein components of paraspeckle (P54nrb) and translocate to paraspeckle, a group of subnuclear bodies, avoiding moving out of the nucleus. However, the same secondary structure could bind to the dsRNA-binding protein Staufen 1 (STAU1) and in this case translocate to cytoplasm. Once in the cytoplasm, the secondary structure could bind to STAU1 again allowing translation, but under some situations mRNA could bind to the ds-RNA- dependent protein kinase (PKR) repressing translation [23]. (c) Ectopic recombination between TE copies (green boxes with yellow arrows) in the same orientation can lead to deletions when recombination takes place between copies located on the same chromatid (1) or deletions and duplications when recombination takes place between copies in different chromosomes (2) (recombination between two nonhomologous chromosomes should lead to a translocation). Ectopic recombination between TE copies in opposite orientation leads to inversion of the DNA between the two TEs (3)
primarily on these two genomes and will highlight the similarities and differences observed between them.
As mentioned above, the human reference genome has millions of TE copies, with 66–69% of the genome mostly derived from TE sequences [13]. Two human retrotransposable element (class I) families, LINE1 (L1, long interspersed nuclear element 1) and Alu, account for 60% of all interspersed repeat sequences. The vast majority of the TEs in the human genome are fixed, and most families are inactive. However, some elements of the main families of human endogenous retrovirus (HERV-K) and LINE1 elements show autonomous transposition. Meanwhile, elements of Alu and the hybrid SVA elements formed by SINEs (short interspersed nuclear elements), VNTRs (variable number tandem repeat), and Alus show nonautonomous activity [64–66].
In contrast, the fruit fly D. melanogaster reference genome contains only thousands of individual TE copies (5416 TE copies in FlyBase R6.04) accounting for only ~5.5% of the euchromatin [67]. If the missing percentage of TEs detected in chromosome 2L is similar in other chromosomes, the euchromatin TE content might be higher (~ 8.7%) [14]. If heterochromatin is also included, TEs account for 11–20% of the D. melanogaster genome [68, 69]. D. melanogaster TEs belong to approximately 100 diverse families of both class I and class II elements [69, 70]. Each family consists of 1–304 copies with no dominant family corresponding to the majority of TEs. The only exception is INE-1 family that contains ~2000 copies and has been inactive for the past ~3–- 4.6 million years [71–73]. The majority of TE families are considered to be active in Drosophila: individual TE copies are generally polymorphic in the population and show a high sequence similarity [69, 70, 74, 75]. Indeed, there is experimental evidence showing that Gypsy and ZAM elements are active [76, 77]. Besides, there is indirect evidence for the activity of 24 D. melanogaster superfamilies based on a whole-genome sequencing experiment of mutation accumulation lines [75] (Table 1).
Why do these two genomes differ so profoundly in content, diversity, and activity of TEs? The answer must lie in different aspects of TE population dynamics within genomes and forces that lead to varying rates of TE family birth and extinction. In the rest of this review, we focus on the state of knowledge of different aspects of TE population dynamics and discuss aspects of TE family evolution. Specifically, we focus on rates of TE transposition, fixation, or loss in human and D. melanogaster populations due to stochastic forces and natural selection for or against TE insertions and forces that affect coexistence of multiple TE families and the standing diversity of TE types (Fig. 3).
#### Table 1 Summary of recent TE population dynamic studies
(continued)
#### 3 Methodology Used to Study TE Population Dynamics
TE dynamics continues to be studied using three main approaches: mathematical modeling, computer simulations, and the analysis of empirical data. Often a combination of these approaches is used to better understand TE abundance, diversity, and distribution (Table 2). Le Rouzic et al. [78] applied the statistical framework originally developed to infer speciation and extinction dynamics in species phylogenies to reconstruct the evolutionary history of TEs [78]. The model allows to estimate and to interpret the pattern of transposition activity that results in different TE copy number distributions [78]. The authors also performed computer simulations to provide reference dynamics that aid in the interpretation of the results obtained (Table 2).
Traditionally, mathematical models considered the relationship between the host and a homogenous group of active TEs. However, the TE content of any genome is a mixed of autonomous and nonautonomous insertions. Xue and Goldenfeld [79] proposed a mathematical model that considers the relationship between nonautonomous and autonomous TEs as a predator-prey dynamic. Unlike previous models that also use the analogy to ecological models, Xue and Goldenfeld model takes into account the
Fig. 3 Factors that influence the population and evolutionary dynamics of TEs. Our understanding of TE population and evolutionary dynamics is still incomplete. The different factors that affect TE population and evolutionary dynamics are interrelated, new factors have been identified in recent years, and future research is still likely to reveal existence of additional factors
molecular level interactions between transposable elements and the small copy number of the active transposons. The model predicts oscillations in the number of TEs in a time scale much longer than the cell replication time, suggesting that the genome stores the predator-prey state during successive generations [79].
TE dynamics have also been analyzed in variable environments [80, 81] (Table 2). Gogolesky et al. [81] proposed a stochastic computational model to analyze the dynamics of active TEs in genomes of sexual diploid organisms under environmental stress. They based their model in the Fisher geometrical model of fitness landscapes. Overall, the authors conclude that the presence of inactive copies of TEs is necessary for the transposition-selection equilibrium of autonomous copies and that the mutator capacity of TEs might be important when host populations face rapid environmental changes [81].
Other recently developed methods analyzed the influence of the mating system in TE dynamics, different modes of selection, or applied branching models for studying the propagation of particular TE classes [82–84] (Table 2).
In addition to mathematical modeling and simulations, multiple computational tools have been developed to analyze TEs in
#### Table 2
#### Summary of recent mathematical models and computer simulations applied to the study of TE dynamics
(continued)
#### Table 2 (continued)
sequenced genomes in the last 5 years. While some of these tools aimed at assessing the global abundance and diversity of TEs in the genome, such as dnaPipeTE, or to annotate TEs in assembled genomes, such as REPET, most of them are focused on discovering and/or genotyping individual copies of TEs in the genome using next-generation sequencing (NGS) data [11, 64, 85–90]. The diversity of methods available makes it difficult to choose the most appropriate one for the analyses of a given genome. To try to overcome this limitation, Nelson et al. [91] developed an integrated pipeline named McClintock that incorporates six complementary TE detection methods. McClintock generates standardized output for the different TE detection methods, thus facilitating the comparison of the results obtained with the different pipelines, as well as facilitating their installation and use [91]. This and other studies that compared the performance of several tools arrived to the same conclusion: several computational tools should be combined to increase the accuracy of TE analysis [64, 86, 91].
The availability of third-generation sequencing techniques (3GS) should help improve the detection and genotyping of TE insertions. Although 3GS was developed before 2010 [92], it has only been in the last few years when this technique has started to be used [14, 93]. Chakraborty et al. [14] reported the assembly of a D. melanogaster genome from a Zimbabwe strain using long-read single molecule real-time sequencing with 147X coverage. Among several novel structural variants described, they identified 37% additional TE insertions in the 2L chromosome compared with a previous study that used 70X coverage of short reads [14, 94]. 3GS technologies have also been applied to the sequencing of human genomes, although a detailed analysis of TE content based on longread data has not been performed yet [95–97].
Recently, Disdero and File´e [98] introduced the first tool that uses long-read sequences to identify TE insertions in the D. melanogaster genome: LoRTE [98]. The authors argue that available software based on short reads fail to correctly identify TEs that are present in highly repetitive regions of the genome, while long-read technologies should allow us to identify all TEs in a given genome. LoRTE, developed in Python, verifies presence and/or absence of previously annotated TEs and can also detect new insertions not previously annotated in the reference genome. LoRTE is able to work with low-coverage sequences (<10X) providing an efficient accurate TE annotation in a cost-effective manner [98].
#### 4 Rates of Transposition
4.1 Empirical Estimates of the Rates of Transposition in Drosophila and Humans
Transposition rates in D. melanogaster have been traditionally estimated empirically by in situ hybridization and by using PCR approaches. The activation of TEs following intra- and interspecific hybridization has been studied in different Drosophila species [99–101]. For example, Vela et al. [100] estimated transpositions rates in D. buzzatii-D. koepferae interspecific hybrid flies by in situ hybridization [100]. They found that hybrids showed at least one order of magnitude higher transposition rates than parental lines for at least three TE families [100]. Robillard et al. [102] estimated transposition rates by qPCR in an experimental evolution study in which a TE insertion was introduced in a strain lacking insertions from that particular family [102]. In the first generations after the introduction of the TE insertion, the transposition rate was 0.33–0.45 per copy per generation, while in the following generations, transposition rates were reduced at least one order of magnitude per copy per generation. These values represent the first steps in the invasion of a TE in a genome that is faster than the rate of transposition when measured in natural populations [102].
In the first edition of this chapter [103], we anticipated that NGS would allow studying transposition rates in a deeper and more accurate way. Indeed, recent studies have taken advantage of NGS data to estimate transposition rates in D. melanogaster. Rahman et al. [89] estimated using NGS data the transposition rate in the reference strain by comparing two available genomes that were sequenced with ~15 years difference. The average transposition rate for TEs belonging to different families was 7 - 10<sup>5</sup> , which is on the same order of magnitude as the previously reported rates (~10<sup>4</sup> –10<sup>5</sup> ). Furthermore, they confirmed the prediction of increased transposition rate in inbred lines: they estimated a higher average number of TE insertions in lab strains inbred for more generations compared with strains inbred for a smaller number of generations [89]. Adrion et al. [75] estimated spontaneous insertion and deletion rates in D. melanogaster mutation accumulation lines [75]. The authors identified 24 active superfamilies and estimated genome-wide insertion rates to be higher than deletion rates: 2.11 - <sup>10</sup><sup>9</sup> vs. 1.37 - 10<sup>10</sup> per site per generation, respectively. Superfamily-specific rates of insertion varied from 0 to 5.13 - 10<sup>3</sup> insertions per copy per generation and were within the range of previously estimated rates [75] (Table 1).
In humans, previous studies estimated the transposition rate as in 1 in 95 to 1 in 250 births for L1, 1 in 20 births for Alu insertions, and 1 in 916 births for SVA retrotransposons [104–107]. Although there are several recent studies that estimate transposition rate in humans using NGS data, they all focused on somatic transposition in the brain or in tumor samples [47, 48, 90].
4.2 Transposition Control Mechanisms Understanding the mechanisms controlling the transposition of TEs is central to our understanding of TE dynamics. Many different mechanisms of TE regulation have been described [43, 108, 109]. In this section, we will highlight recent advances in both TE self-regulation and regulation by host factors.
4.2.1 TE Self-Regulation Self-regulation of transposition was first described in prokaryotes and soon after in TEs involved in hybrid dysgenesis in Drosophila [110]. Recent studies have cast some doubt on one of the selfregulation mechanisms described: transposase overproduction inhibition. The transposase overproduction inhibition mechanism regulates the transposition of IS630-Tc1-mariner piggyBac and hobo-AC-Tam (hAT) superfamilies [111, 112]. However, several studies reported contradictory results suggesting that transposase inhibition by overproduction does not always happen [113]. Bire et al. [113] suggested that some works failed to detect transposase inhibition because cellular cofactors are necessary to execute this regulation system, and as such it can only be detected in in vivo experiments [113]. However, Woodard et al. [114] showed that aggregation of transposase proteins produces filamentous structures (rodlets) in the nucleus in a host independent manner [114]. The authors further showed that a decline in transposition occurs after transposase concentrations are high enough for filamentous structures to be visible [114]. Thus, it is still not clear why some in vitro experiments failed to detect transposase overproduction inhibition [114].
4.2.2 Regulation by Host Factors Small RNAs, such as small-interfering RNAs (siRNAs) and piwiinteracting RNAs (piRNAs), are well-known to play an essential role in silencing TEs and preventing transposition. Several recent reviews highlight the monumental progress in this field [115–119]. In addition to posttranscriptional regulation of TEs, small RNAs are involved in transcriptional regulation as well. In mouse, piRNAs are required for de novo methylation and silencing of TEs [120]. In Drosophila, Piwi proteins repress transcription and correlate with an increase in repressive chromatin marks at loci targeted by piRNAs [121].
> While the role of siRNAs and piRNAs has been established for several years, a role of micro RNAs (miRs) in suppressing the mobility of retrotransposons was only recently described [122]. The authors showed that mir-128 binds to L1 RNA and represses its integration in humans [122].
> New studies have also provided evidence for the role in TE repression of proteins previously known for their roles in other cellular processes such as interferon-stimulated proteins, the tumor suppressor p53, and the longevity regulating protein SIRT6. Several interferon-stimulated genes, such as the Moloney leukemia virus 10 (MOV10), the zinc-finger antiviral protein (ZAP), and the 3<sup>0</sup> repair exonuclease 1 (TREX1), which are associated with virus response, have been recently involved in the inhibition of L1 activity [66, 123]. Recently, it has also been shown that the p53 transcription factor, which is involved in stress response networks and acts to restrict oncogenesis, also restricts retrotransposon activity in zebra fish, flies, and humans [124]. The authors showed that p53 interacts with components of the piwiinteracting RNA to suppress retrotransposition [124]. Finally, the longevity regulating protein SIRT6 is also involved in retrotransposon repression by coordinating their packaging into transcriptionally repressive heterochromatin. SIRT6 binds to the 5<sup>0</sup> UTR region
of retrotransposons and mono-ADP ribosylates the Kru¨ppel-associated protein 1 (KAP1) facilitating the interaction of KAP1 with the heterochromatin protein 1α (HP1α) leading to chromatin compaction [125].
#### 5 Rate of Fixation and Frequency Distribution
#### 5.1 Natural Selection Against TE Insertions
Natural selection and stochastic processes influence both the rate of fixation and the frequency distribution of TEs in populations. The efficiency of selection depends on the effective population size, which largely differs between Drosophila and humans: >10<sup>8</sup> and ~104 , respectively [126, 127]. Thus, while in Drosophila the high efficiency of selection should led to the removal of slightly deleterious TE insertions, in humans, these insertions may accumulate in the genome. Indeed most of the TE sequences in the human genome are remnants of ancient insertions [12].
A review by Barro´n et al. [128] explored the latest insights on the nature of selection acting against the deleterious effects of TEs in D. melanogaster populations [128]. More recently, Kofler et al. [129] analyzed intraspecific TE dynamics between D. melanogaster and D. simulans populations to shed light on the long-term evolution of TEs [129]. They confirmed that most of the TEs are present at low frequencies in D. melanogaster and showed that the same pattern is present in D. simulans. Based on computer simulations showing that 50% of the TE families have temporally heterogeneous transposition rates, and on the differences in TE composition between populations of the same species, the authors suggested that TE activity has recently increased in the two species. They proposed that the demographic history of both species, with a recent colonization of different environments, could be the cause of the high TE activity detected [129].
In humans, a recent study took advantage of the 1000 Genome Project data that reports 16,192 polymorphic TEs to perform the most complete TE dynamics analysis to date [130]. Most of the polymorphic TEs were found to be present at very low frequencies: >93% of TEs showed <5% allele frequency in 26 human populations. These results confirm that overall polymorphic TE insertions are deleterious in humans as was previously suggested with smaller family-specific datasets [131].
5.2 TE-Induced Adaptations Several recent reviews have compiled results that showcase the adaptive role of TEs [19, 24, 50, 59, 128]. We would like to highlight the recent discovery of a TE in a fish-like marine chordate that encodes RAG-like proteins with endonuclease-transposase activity [39]. This discovery provides evidence that supports the TE origin hypothesis for the adaptive immune system in jawed vertebrates [39]. Two other recent publications provide experimental evidence for a role of TEs as providers of functional transcription factor binding sites (TFBS) involved in immune response and in cell pluripotency [50, 132]. A recent study linked ERV elements in humans with the interferon response pathway [50]. The authors showed that ERVs carrying enhancers have been co-opted to activate different genes involve in inflammatory response activated by interferon. This example shows how the exaptation of one family of TEs could shape a transcriptional network to activate different genes with one trigger system [50]. Sundaram et al. [132] reported mouse-specific TEs that contain multiple transcription factor binding sites for pluripotency transcription factors. The majority of the TEs were experimentally shown to exhibit enhancer activity in mouse embryonic stem cells including an in silico reconstructed ancestral TE. This latter result suggests that ancestral TEs already had transcriptional regulatory sites [132].
In Drosophila, the adaptive role of several TEs has also been identified. Most of the TEs characterized so far are involved in stress response: viral infection and xenobiotics (Doc1420, [60, 61]), oxidative stress (FBti0018880, [53]), xenobiotic stress (Accord, [62, 63], and FBti0019627, [52]), cold stress (FBti0019985, [55]), and heavy metal stress (FBti0019170, [56]), while FBti0019386 insertion was associated with faster developmental time [54]. Some of these adaptive insertions have been shown to affect gene expression through different molecular mechanisms, such as affecting the polyadenylation site choice [52], and adding TFBS [53], while others have been associated with gene duplication [60, 62].
#### 6 Rate of Loss
A recent study estimated genome-wide and superfamily-specific TE deletion rates in D. melanogaster inbred lines [75]. The authors found that most of the deletions involved retrotransposon elements suggesting that the deletions were due to ectopic recombination instead of excision. Deletion rates were smaller than insertion rates estimated in the same inbred lines [75].
In vertebrates, lineage-specific differences in TE deletion rates have been reported [133]. A possible explanation for this observation is that the success of some families results in a competition for the genome resources leading to the elimination of other TE families [133].
In addition to TE deletion rates, DNA loss rates should also be considered. In the human linage, estimates of DNA loss are smaller than estimates of DNA gain, 650 Mb vs. 815 Mb [134], while in D. melanogaster, the rate of DNA loss is higher than the rate of DNA gain [135–137].
#### 7 Horizontal Transfer of TE Insertions
In addition to parent to offspring transmission, TEs can also be horizontally transferred [138–141]. By combining simulation and analytical approaches, Groth and Blumenstiel [142] suggested that exposure rate to new TE families through horizontal transfer can be an important determinant of TE genomic content when the effects of drift in a population are weak [142]. Thus, larger populations are expected to carry a higher TE content if population exposure rate is proportional to population size [142]. So far, most of the evidence for TE horizontal transfer comes from closely related and geographically close species [140]. There are several examples of horizontal transfer of TEs in Drosophila species, while so far horizontal transfer of TEs has not been described in humans [138].
#### 8 Conclusion
Recent years have seen an increase in the number of reference genome sequences available as well as of population genome datasets. The availability of all these genome sequences and the development of new bioinformatics tools have allowed us to update our previous estimates of genomic TE content that have increased both in humans and in D. melanogaster. These data has also allowed us to gather more evidence for the functional impact, both detrimental and beneficial, of TE insertions. Thus, it is still indisputable that understanding TE population dynamics is essential to understand genome structure, genome function, and genome evolution.
New methods developed to analyze the dynamics of TEs in populations have shed light on the interplay between autonomous and nonautonomous TE copies, TE invasion dynamics, and how the mating system influences the dynamics of TEs in genomes. We have also considerably advanced our knowledge on the host factors that regulate TE activity as well as in the genome features that influence TE dynamics (Fig. 3). Finally, differences in effective population sizes that affect the efficiency of selection against new TE insertions and differences in the rates of TE loss between humans and D. melanogaster can still be considered two important factors that contribute to the different abundance, diversity, and activity of TEs in this two species [103].
#### 9 Questions
How differences in the rate of DNA loss can affect the evolutionary dynamics of TEs?
Why host regulation of transposition is relevant for TE dynamics?
#### Acknowledgment
We thank the reviewers for providing constructive comments on a previous version of this manuscript. This work has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation program (H2020- ERC-2014-CoG-647900) and from the Spanish Ministry of Economy and Competitiveness/FEDER (BFU2014-57779-P).
#### References
Macas J, Lysak MA (2011) Diverse retrotransposon families and an AT-rich satellite DNA revealed in giant genomes of Fritillaria lilies. Ann Bot 107(2):255–268
in Drosophila genomes. BMC Genomics 17:344
mammalian SINE transcripts. Nucleic Acids Res 23(10):1758–1765
Drosophila melanogaster. Mol Biol Evol 32 (7):1800–1814
Korlach J, Wilson RK, Eichler EE (2017) Discovery and genotyping of structural variation from long-read haploid genome sequence data. Genome Res 27(5):677–685
repeats are major structural variants in the human genome. Cell 141(7):1171–1182
Udawatta M, Ngo D, Chen Y, Paguntalan A, Ray T, Hughes A, Cohen BA, Wang T (2017) Functional cis-regulatory modules encoded by mouse-specific endogenous retrovirus. Nat Commun 8:14550
survey of retrotransposons revealed by the complete Saccharomyces cerevisiae genome sequence. Genome Res 8(5):464–478
mapping of transposable element diversity reveals links to gene regulation and epigenomic variation. Elife 5:e20777
156. Yan L, Gu YH, Tao X, Lai XJ, Zhang YZ, Tan XM, Wang H (2014) Scanning of transposable elements and analyzing expression of transposase genes of sweet potato Ipomoea batatas. PLoS One 9(3):e90895
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Part V
#### Population Genomics and Omics in Light of Disease and Evolution
## Chapter 17
#### Association Mapping and Disease: Evolutionary Perspectives
#### Søren Besenbacher, Thomas Mailund, Bjarni J. Vilhja´ lmsson, and Mikkel H. Schierup
#### Abstract
In this chapter, we give a short introduction to the genetics of complex diseases emphasizing evolutionary models for disease genes and the effect of different models on the genetic architecture, and we give a survey of the state-of-the-art of genome-wide association studies (GWASs).
Key words Complex diseases, Association mapping, Genome-wide association studies, Common disease/common variant
#### 1 Introduction
A combination of genes and environment determines our phenotype. The degree to which genotype or environment influences our phenotype—the balance of nature versus nurture—varies from trait to trait, with some traits independent of genotype and determined by the environment alone and others determined by the genotype alone and independent of the environment.
A measure quantifying the importance of genotype compared to the environment is the so-called heritability. It is the fraction of the total phenotypic variation in the population explained by variation in the genotype within the population [1]. A trait of interest, say a common disease, which exhibits a nontrivial heritability, tells us that genes are important for understanding this trait and that it is worthwhile to identify the specific genetic polymorphisms influencing the trait. The first step toward this is association mapping: searching for genetic polymorphisms that, statistically, associate with the trait. Polymorphisms associated with a given phenotype need not influence that phenotype directly, but it is among those associated genetic polymorphisms that we will find the causal ones.
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_17, © The Author(s) 2019
Genetic variants are correlated, a phenomenon called linkage disequilibrium (LD), so by examining the trait association of a few variants, we learn about the association of many others. Examining the association between a phenotypic trait and a few hundred thousand to a million genetic variants suffices to capture how most of the common variation in the entire genome associates with the trait [2–4]. When we find a genetic variant associated with the trait, we have not necessarily located a variant that has any functional effect on the trait, but we have located a genomic region containing genetic variation that does. LD is predominantly a local phenomenon, so correlated genetic variants tend to be physically near each other on the genome. If we observe an association between the phenotype and a variant, and the variant is not causally affecting the trait but is merely in LD with a causal variant, the causal variant is likely nearby. Further examination of the region might reveal which variants affect the trait, and how, but that often involves functional characterization and is beyond association mapping. With association mapping, we merely seek to identify genetic variation that associates with a trait.
#### 2 The Allelic Architecture of Genetic Determinants for Disease
Many complex diseases show a high heritability, typically ranging between 20% and 80%. Each genetic variant that increases the risk of disease contributes to the measured heritability of the disease and thus explains some fraction of the estimated total heritability of the trait. For most diseases investigated, many variants contribute, and the fraction of the heritability explained for each is therefore low. The number of contributing variants, their individual effects on the disease probability, their selection coefficient, and their dominance relations can be collectively termed the genetic architecture of a common disease. Insights into this architecture are slowly emerging and reveal differences between diseases [5].
Below we first consider two proposed genetic architectures based on theoretical arguments: the common disease common variant (CDCV) architecture and the common disease rare variant (CDRV) architecture. CDCV states that most of the heritability can be explained by a few high-frequency variants with moderate effects, while CDRV states that most of the heritability can be explained by moderate- or low-frequency variants with large effects. We present population genetic arguments for the two architectures and the consequences of the two architectures for association mapping. Later, in Subheading 5.1, we present empirical knowledge we have obtained about the genetic architectures of common diseases.
#### 2.1 Theoretical Models for the Allelic Architecture of Common Diseases
Understanding the distribution of the number and frequency of genetic variants in a population is the purview of population genetics. Using diffusion approximations we can derive the expected frequency distribution of independent mutations under mutationdrift-selection balance in a stable population (see, e.g., Wright [6]). Central parameters are the mutation rate, u, and the selection for or against an allele, measured by s, scaled with the effective population size, N. Mutations enter a population with a rate determined by Nu, and subsequently, their frequencies change in a stochastic manner. If a mutant allele is not subject to natural selection, for example, if it does not lead to any change in function, it is selectively neutral. Its frequency then rises and falls with equal probability. If the allele is under selection, it has a higher likelihood of increasing in frequency than decreasing if it is under positive selection (s > 0) and conversely for negative selection (s < 0).
At very high or very low frequencies, selection has an insignificant effect on the change in frequency, and the system evolves essentially entirely stochastic (genetic drift). At moderate frequencies, however, the effect of selection is more pronounced, and given sufficiently strong selection (of an order Ns 1), the direction of changes in the allele frequency is almost deterministically determined by the direction of selection. An allele subject to a sufficiently strong selection that happens to reach moderate frequencies either halts its increase if selection works against it, and drifts back to a low frequency, or if selection favors it, it rapidly rises to high frequencies, where eventually the stochastic effects again dominate (see Fig. 1).
The range of frequencies, where drift dominates, or selection dominates, is determined by the strength of selection (Ns) and the genotypic characteristics of selection, as, e.g., dominance relations between alleles. For strong selection or in large populations, the process is predominantly deterministic for most frequencies, while for weak selection or a small population, the process is highly stochastic for most frequencies. The time an allele can spend at moderate frequencies is also determined by Ns and selection characteristics.
Pritchard and Cox [7, 8] used diffusion arguments to show that common diseases are expected to be caused by a large number of distinct mutations. This implies that genes commonly involved in susceptibility exert their effect through multiple independent mutations rather than a single mutation identical by descent in all carriers (see Fig. 2). Each mutation, if under weak purifying selection, is unlikely to reach moderate frequencies, and since the population will only have few carriers of each disease allele, each can only explain little of the heritability. The accumulated frequency of several alleles, each kept to low frequency by selection, can, however, reach moderate frequencies. So the heritability can be
Fig. 1 Mutation, drift, and selection. New mutations enter a population at stochastic intervals, determined by the mutation rate, u, and the effective population size, N. For low or high frequencies, where the range of such frequencies is determined by the selection factor, s, and the effective population size, the frequency of a mutant allele changes stochastically. At medium frequencies, on the other hand, the frequency of the allele changes up or down, depending on s, in a practically deterministic fashion. If a positively selected allele reaches moderate frequency, it will quickly be brought to high frequency, at a speed also determined by s and N
Fig. 2 Accumulation of several rare frequencies. If selection works against a set of alleles, each will be kept at a low frequency. Their accumulated frequency, however, can be high in the population
explained either by many recurrent mutations or many independent loci affecting the disease: the CDRV architecture.
Implicitly, this model assumes a population in mutationselection equilibrium, and this does not necessarily match human populations. Humans have recently expanded considerably in numbers, and changes in our lifestyle, e.g., from hunter-gatherers to farmers might have changed the adaptive landscape driving selection of our genes.
The frequency range where drift, rather than deterministic selection, dominates is larger with a smaller population than with a larger population. We can think of the drift process as a birth– death process operating on individual copies of genes, which is highly stochastic. Only when we consider a large number of these processes do we get an almost deterministic process. At low allele frequencies, the process is stochastic because we only have a few copies of the allele to consider. At higher frequencies, we have many copies, so we get the deterministic behavior. The same number of copies, however, constitutes a higher frequency of a small population than of a larger population. Consequently, selection is effective at much lower frequencies in a large population than it is in a small population; the absolute number of copies of a deleterious allele might be the same in a small and a large population, but they constitute a smaller fraction of the large population. In large populations, we expect to see deleterious mutations to be found at small frequencies unless, as is the case for most human populations, the large population size is a consequence of recent dramatic growth [9]. This effect is illustrated as the "transient period" in Fig. 3, where common genetic variants may contribute much more to disease than under stable demographic conditions. Following expansion, alleles that would otherwise be held at low frequency by selection may be at moderate frequencies and thus contribute a larger part of the heritability: the CDCV architecture.
Similarly, a recent change in the adaptive landscape of a population might cause an allele that was previously held at low frequency to be under positive selection and now rise in frequency [10]. In this transition period, an allele may be at a moderate frequency and therefore contributes significantly to the heritability of disease susceptibility (see Fig. 4).
Depending on which architecture underlies a given disease, different strategies are needed to discover the genetic variants
Fig. 3 A population out of equilibrium following an expansion. In a transition period following a population expansion, the allele frequency patterns are different from the patterns in a stable population
Fig. 4 A population out of equilibrium following changes in the selective landscape. If the selection of an allele changes direction, so the positively selected allele becomes negatively selected and vice versa, it will eventually move through moderate frequencies. Following a change in the selective landscape, it is thus possible to find alleles at moderate frequencies that would not otherwise be found
> involved. When genome-wide association mapping was proposed as a strategy for discovering disease variants, the proposal was based on the hypothesis that, at least for some common diseases, the CDCV architecture underlies them. GWAS relies on the CDCV hypothesis for two practical reasons. The first is that the LD patterns across the genome greatly restrict examination to only a small fraction of the total possible variation. It is feasible to probe the common variants of a genome from a small selection of representative variants, but the association with rare variants is far less detectable. Second, statistical analysis of the association between polymorphism and disease is rather straightforward for moderatefrequency alleles but has far less power to detect association with low-frequency alleles.
> While the GWAS approach is only practical as an approach for variant discovery for common alleles, it was necessary to hypothesize that the CDCV architecture would be underlying diseases of interest. The actual genetic architecture behind common diseases was unknown, but there were no alternative methods aimed at CDRV, so GWAS was the only show in town.
2.2 The Allelic Frequency Spectrum in Humans The vast majority of human nucleotide variation is very rare because of our history of population bottlenecks followed by rapid growth. For instance, in the 2500 individuals of the 1000 genomes study, 64 million SNVs have frequency <0.5%, and 20 million SNVs have frequency >0.5% [11]. Nevertheless the majority of heterozygous variants observed within a single individual are not rare [11]. The rare variants are most often very recent and therefore specific to populations, and they are also more often deleterious because selection has not yet acted on them [12]. This is particularly clear for loss-of-function variants and other protein-coding variants. A study of 2636 Icelanders found that the fraction of variants with a minor allele frequency (MAF) below 0.1% was 62% for proteintruncating variants, 46% for missense variants, and 38% for synonymous variants [13].
The strong recent population expansions have also allowed variants to increase in frequency by surfing on the population expansion wave front even if they would be selected against in a population with stable size. Thus, rare variants with large effects on disease may exist. The GWAS studies so far have been successful in identifying a large set of common variants associated with disease, so common variants contributing to disease do exist. It is likely that rare variants with large phenotypic effects also contribute to the heritability of many common diseases, but the extend is likely to be disease specific.
#### 3 The Basic GWAS
The first GWASs were published around 2006 [14, 15] when Illumina and Affymetrix first introduced genotyping chips that made it possible to test hundreds of thousands of SNPs quickly and inexpensively. The GWASs' approach to find susceptibility variants for diseases boils down to testing approximately 0.3–- 2 million SNPs (depending on chip type) for differences in allele frequencies between cases and controls, adjusting for the high number of multiple tests. This approach is a wonderfully simple procedure that requires no complicated statistics or algorithms but only well-known statistical tests and a minimum of computing power. Despite the simplicity, some issues remain, such as faulty genotype data and confounding factors that can result in erroneous findings if not handled properly. The most important aspects of any GWAS are, therefore, thorough quality control of the data used and measures to avoid and reduce the effect of confounding factors.
#### 3.1 Statistical Tests The primary analysis in an association study is usually testing each variant separately under the assumption of an additive or multiplicative model. One way of doing that is by creating a 2 2 allelic contingency table as shown in Table 1 by summing the number of A and B alleles seen in all case individuals and all control individuals. Be aware that we are counting alleles and not individuals in this contingency table, so Ncases will be equal to two times the number of case individuals because each individual carries two copies of each variant unless we are looking at non-autosomal DNA. If there is no association between the variant and the disease in question, we
#### Table 2 Expected allele counts in case/control data
would expect the fraction of cases that have a particular allele to match the fraction of controls that have that allele. In that case, the expected allele count (EN) would be as shown in Table 2. To test whether the difference between the observed allele counts (in Table 1) and the expected allele counts (in Table 2) is significant, a Pearson χ <sup>2</sup> statistic can be calculated:
$$X^2 = \Sigma\_{\text{Phncopypc}} \Sigma\_{\text{Allc}} \left( N\_{\text{Phncorypc}, \text{Allc}} - \text{EN}\_{\text{Phncorypc}, \text{Allc}} \right)^2 / \text{EN}\_{\text{Phncorypc}, \text{Allc}}$$
This statistic approximates a χ <sup>2</sup> distribution with 1 degree of freedom, but if the expected allele counts are very low (<10), the approximation breaks down. This means that if the MAF is very low or if the total sample size, N, is small, an exact test, such as the Fisher's exact test, should be applied. An alternative to the tests that use the 2 2 allelic contingency table and thereby assumes a multiplicative model is the Cochran–Armitage trend test that assumes an additive risk model [16]. This test is preferred by some since it does not require an assumption of Hardy–Weinberg equilibrium in cases and controls combined [17].
While a 1 degree of freedom test that assumes an additive or multiplicative model is usually the first analysis, some studies also perform a test that would be better at picking up associations following a dominant or recessive pattern, for instance, by performing a 2 degrees of freedom test of the null hypothesis of no association between rows and columns in the 2 3 contingency table that counts genotypes instead of alleles.
3.2 Effect Estimates A commonly used way of measuring the effect size of an association is the allelic odds ratio (OR), which is the ratio of the odds of being a case given that you carry n copies of alleles A to the odds of being a case if you carry n 1 copies of allele A. Assuming a multiplicative model, this can be calculated as:
$$\begin{array}{l} \text{OR} = \left(N\_{\text{casc},\text{A}}/N\_{\text{control},\text{A}}\right) / \left(N\_{\text{casc},\text{B}}/N\_{\text{control},\text{B}}\right) \\ = N\_{\text{casc},\text{A}}N\_{\text{control},\text{B}}/N\_{\text{casc},\text{B}}N\_{\text{control},\text{A}} \end{array}$$
Another measure of effect size that is perhaps more intuitive is the relative risk (RR), which is the disease risk in carriers divided by the disease risk in noncarriers. This measure, however, suffers from the weakness that it is harder to estimate. If our cases and controls were sampled from the population in an unbiased way, the allelic RR could be calculated as:
$$\text{RR} = \left( N\_{\text{case},\text{A}} / N\_{\text{A}} \right) / \left( N\_{\text{case},\text{B}} / N\_{\text{B}} \right)$$
but it is very rare to have an unbiased population sample in association studies because the studies are generally designed to deliberately oversample the cases to increase the power. This oversampling affects the RR as calculated by the formula above but not the OR which is one of the reasons why the OR is usually reported in association studies instead of the RR.
3.3 Quality Control Data quality problems can be either variant specific or individual specific, and inspection usually results in the removal of both problematic individuals and problematic variants from the data set.
> Individual-specific problems can be caused by low DNA quality or contamination by foreign DNA. A sample of low DNA quality results in a high rate of missing data, where particular variants cannot be called, and there is a higher risk of miscalling variants. It is, therefore, recommended that individuals lacking calls in more than 2–3% of the variants are removed from the analysis. Excess heterozygosity is an indicator of sample contamination, and individuals displaying that should also be disregarded. Sex checks and other kinds of phenotype tests might also be applied to remove individuals, where the genotype information does not match the phenotype information due to a sample mix-up [18].
> For a given variant, the data from an individual can be suspicious in two ways: it can fail to be called by the genotype-calling program or it can be miscalled. Typically, a conservative cutoff value is used in the calling process securing that most problems show up as missing data rather than miscalls. Most problematic variants, therefore, reveal a high fraction of missing data, and variants missing calls above a given threshold (typically, 1–5%) are removed. Miscalls typically occur when the homozygotes are hard to distinguish from the heterozygotes, and some of the heterozygotes are being misclassified as homozygotes or vice versa. Both biases
manifest as deviation from Hardy–Weinberg equilibrium, and SNPs that show large deviations from Hardy–Weinberg equilibrium within the controls should be removed [19].
3.4 Confounding Factors Confounding in GWAS can arise if there are genotyping batch effects or if there is population or family structure in the sample. For example, if cases and controls in GWAS are predominantly collected from geographically distinct areas, association signals could arise due to genetic differences caused by geographic variation, and most of such genetic signals are unlikely to be causal. Such confounding due to population structure typically occurs when samples have different genetic ancestry, e.g., if the sample contains individuals of both European and Asian ancestry. Population structure confounding can also happen when the population structure is more subtle, especially for large sample sizes. Methods for inferring population substructure, such as principal components analysis, are useful for detecting outliers we can remove from the data [20]. However, this approach is not suitable when dealing with subtle structure, as a small bias can become significant in a large enough sample of individuals of similar genetic ancestry.
> Confounding in GWAS can be detected as inflation of the test statistics, beyond what is expected due to truly causal variants. A useful way of visualizing such inflation of test statistics is the so-called quantile–quantile (QQ) plot. In this plot, ranked values of the test statistic are plotted against their expected distribution under the null hypothesis. In the case of no true positives and no inflation of the test statistic due to population structure or cryptic relatedness, the points of the plot lie on the x ¼ y line (see Fig. 5a). True positives show an increase in values above the line in the right tail of the distribution but do not affect the rest of the points since
Fig. 5 QQ plots from a χ <sup>2</sup> distribution. (a) A QQ plot, where the observation follows the expected distribution. (b) A QQ plot, where the majority of observations follow the expected distribution, but where some have unexpectedly high values, i.e., are statistically significant. (c) A QQ plot, where the observations all seem to be higher than expected, which is an indication that the observations are not following the expected distribution
only a small fraction of the SNPs is expected to be true positives (Fig. 5b). Cryptic relatedness and population stratification lead to a deviation from the null distribution across the whole distribution and can, thus, be seen in the QQ plot as a line with a slope larger than 1 (Fig. 5c).
Several approaches accounting for population structure in GWAS have been proposed. Devlin and Roeder [21, 22] proposed genomic control, i.e., to shrink the observed χ <sup>2</sup> test statistic to make the median coincide with the expected value under the null model. However, studies by Yang et al. [23] and Bulik-Sullivan et al. [24] pointed out that the median and mean χ <sup>2</sup> statistic is expected to be inflated for polygenic traits, even when there is no population structure confounding. With that in mind, we recommend adjusting for the confounders in the statistical model instead of performing genomic control. One such approach is to include covariates that capture the relevant structure in the model. Price et al. [25] proposed including the largest principal components as covariates in the model to adjust for population structure. This approach has proved to be effective in most cases. However, if the sample includes related individuals or if it is very large, controlling for the top PCs may not be able to capture subtle structure. An alternative approach is to use mixed models [26, 27], where the expected genetic relatedness between the individuals is included in the model. Advances in computational efficiency of mixed models [28] now enable analysis of very large and complex data sets, such as the UK biobank data set [29].
Besides population structure, family structure or cryptic relatedness can also confound the analyses. Here one can identify closely related individuals by calculating a genetic relatedness matrix and prune the data so that it does not contain any close relatives. Lastly, sequencing batch effects due to incomplete randomizations can lead to structure, unrelated to genetics, which confounds the analysis. A study on polygenic prediction of longevity by Sebastiani et al. [30] serves as a warning. The researchers applied two different kinds of chips and failed to remove several SNPs that exhibited bad quality on only one of the chips [31]. If the fraction of the two different kinds of chips had been the same in both cases and controls that would probably not have resulted in false signals, unfortunately, the chip with the bad SNPs was used in twice as many cases as controls. When this genotyping batch effect was discovered, the authors had to retract their publication from Science. Type and frequency of errors that may happen during sample preparation and SNP calling are likely to vary through time and space, so case and control samples should be completely randomized as early as possible in the procedure of genotypic typing. Failure to carefully plan this aspect of an investigation introduces errors in the data that are hard, if not impossible, to disclose, and they may reduce interesting findings to mere artifacts.
3.5 Meta-analysis of GWAS The statistical power to detect association depends directly on the sample size used, all other things being equal. This fact has driven researchers to collaborate across institutions and countries in GWAS consortia, where they combine multiple cohorts in one large analysis. However, for logistic and legal reasons, it may not be possible to share individual-level genotypes, which are required for all of the GWAS approaches covered so far. Meta-analyses of GWASs performed in each cohort are a solution to this problem. These require coordination between the researchers, where they share GWAS summary statistics instead of individual-level genotypes. These summary statistics are then meta-analyzed using statistical approaches that either assume a constant effect across cohorts or not. In recent years many large-scale GWAS metaanalyses have been published, and the resulting summary statistics of these are often made public, providing a treasure trove for understanding genetics of common diseases and traits [32].
3.6 Replication The best way to make sure that a finding is real is to replicate it. If the same signal is found in an independent set of cases and controls, it means that the association is unlikely to be the result of a confounding factor specific to the original data. Likewise, if the association persists after typing the markers using another genotyping method, it means that it is not a false positive due to some artifact of the genotyping method used.
> When trying to replicate a finding, the best strategy is to try to replicate it in a population of similar ancestry. A marker that correlates with a true causal variant in one population might not be correlated with the same variant in a population of different ethnicity, where the LD structure can be different. This is especially problematic when trying to replicate an association found in a non-African population in an African population [33]. A marker might easily have 20 completely correlated markers in a European population, but no good correlates in an African population. To replicate a finding in the European population of one of these variants, it does not suffice to test one of the variants in an African population; all 20 variants must be tested. This, however, also offers a way to fine map the signal and possibly find the causative variant [34].
> Before spending time and effort to replicate an association signal in a foreign cohort, it is a good idea to search for the existing partial replication of the marker within the data. Usually, a marker is surrounded by several correlated markers on the genotyping chip, and if one marker shows a significant association, then the correlated markers should show an association too. If a marker is significantly associated with a disease, but no other marker in the region is, then it should be viewed as suspicious.
#### 4 Imputation: Squeezing More Information Out of Your Data
The current generation of SNP chip types includes only 0.3–- 2 million of the nine to ten million common SNPs in the human (i.e., SNPs with a MAF of more than 5%). Because of the correlation between SNPs in LD, however, the SNP chips can still claim to assay most of the common variants in the genome (in European populations at least). Although the Illumina HumanHap300 chip only directly tests about 3% of the ten million common SNPs, it still covers 77% of the SNPs in HapMap with a squared correlation coefficient (r 2 ) of at least 0.8 in a population of European ancestry [35]. The corresponding fraction in a population of African ancestry is only 33%, however.
These numbers expose two limitations of the basic GWAS strategy. First, there is a substantial fraction of the common SNPs that are not well covered by the SNP chips even in European populations (23% in the case of the HumanHap300 chip). Second, we rely on tagging to test a large fraction of the common SNPs, and this diluted signal from correlated SNPs inevitably causes us to overlook true associations in many instances. An efficient way of alleviating these limitations is genotype imputation, where genotypes that are not directly assayed are predicted using information from a reference data set that contains data from a large number of variants. Such imputation improves the GWAS in multiple ways: It boosts the power to detect associations, gives a more precise location of an association, and makes it possible to do meta-analyses between studies that used different SNP chips [36].
#### 4.1 Selection of Reference Data Set
The two important choices when performing imputation are the reference data set to use and the software to use. Usually, a publicly available reference data set, such as the 1000 Genomes Project [11] or the large Haplotype Reference Consortium [37], is used. Alternatively, researchers sequence a part of their study cohort and thus create their own reference data set. The latter strategy has the advantage that one can be certain that the ancestry of the reference data matches the ancestry of the study cohort. It is important that the reference data be from a population that is similar to the study population. If the reference population is too distantly related to the study population, the reliability of the imputed data will be reduced. The quality and nature of the reference data also limit the quality of the imputed data in other ways. A reference data set consisting of only a small number of individuals is not able to reliably estimate the frequency of rare variants and that in turn means that the imputation of rare variants lacks in accuracy. This means that there is a natural limit to how low a frequency a variant can have and still be reliably imputed.
The largest publicly available reference data set is the Haplotype Reference Consortium (HRC) that combines whole-genome sequence data from 20 studies of predominantly European ancestry [37]. The first release of this reference panel has data from 32,611 samples at 39,235,157 SNPs. The large sample size means that variants with minor allele frequencies as low as 0.1% can correctly be imputed using this data set.
The use of imputation methods does not only offer the possibility of increased SNP coverage, but, given the right reference data, also eases the analysis of common non-SNP variation, such as indels and copy number variations (CNVs). So far some reference panels have, however, only include SNVs and disregarded indels and structural variants. The increasing quality of wholegenome sequencing and software for calling structural variants means that better data sets that include structural variants should soon become available. Imputation will then make it possible to use the SNP chips to test many indels and structural variants that are not being (routinely) tested today [38].
4.2 Imputation Software The commonly applied genotype imputation methods, such as IMPUTE2 [39], BIMBAM [40], MaCH-Admix [41], and minimac3 [42], are all based on hidden Markov models (HMMs). Comparisons of these software packages have shown that they produce data of broadly similar quality but that they are superior to imputation software based on other methodological approaches [36, 43]. The basic HMMs used in these programs are similar to earlier HMMs developed to model LD patterns and estimate recombination rates.
When the sample size is large, imputation using these HMM-based methods imposes a high computational burden. One possible way of decreasing this burden is to pre-phase the samples so that resolved haplotypes are used as input for the imputation software instead of genotypes [44]. But even with pre-phasing, the computational task is far from trivial, and wholegenome imputation is not a task that can be performed on a single computer. This computational problem can be solved by using one of the two free imputation services that have recently been launched (https://imputationserver.sph.umich.edu, https://imputation. sanger.ac.uk). These services allow users to upload their data through a web interface and choose between a set of reference panels. The data set will then be imputed on a High Performance Computing Cluster, and the user will receive an email when the imputed data is ready for download.
4.3 Testing Imputed Variants Since imputation is based on probabilistic models, the output is merely a probability for each genotype for the unknown variants in a given individual. That is, instead of reporting the genotype of an individual as AG, say, the program reports that the probability of the genotype being AA is 5%, that of being AG is 93%, and that of being GG is 2%. This nature of the output data challenges the GWAS. The simplest way of analyzing the imputed data is to use the "best guess" genotype, i.e., assume the genotype with the highest probability and ignore the others. In the example above, the individual would be given the genotype AG at the SNP in question, and usually, an individual's genotype would be considered as missing if none of the genotypes have a probability larger than a certain threshold (e.g., 90%). The use of "best guess" genotype is problematic since it does not take the uncertainty of the imputed genotypes into account, may introduce a systematic bias, and lead to false positives and false negatives. A better way is to report a logistic regression on the expected allele count—in the example above, the expected allele count for allele A would be 1.03 (2pAA + pAG). This method has proved to be surprisingly robust at least when the effect of the risk allele is small [45], which is the case for most of the variants found through GWAS. An even better solution is to use methods that fully account for the uncertainty of the imputed genotypes [45–47].
#### 5 Current Status
After the first GWAS saw publication in 2005, it was followed by many more studies, and today almost 4000 such studies of human diseases or traits have been published (Fig. 6a). The first GWASs had moderate sample sizes with hundreds of samples, but over the years the sample sizes and thereby the power of the studies have gradually been increasing (Fig. 6b). Imputation and later also nextgeneration sequencing have resulted in a rapid increase in the
Fig. 6 GWAS statistics from the NHGRI-EBI GWAS Catalog [63] (accessed June 2017). (a) The cumulative number of GWASs published since 2005. (b) The initial sample sizes of the GWASs. For dichotomous traits the combined number of cases and controls is shown. Replication samples are not counted. (c) The number of tested variants in each study
number of variants that are tested in a GWAS (Fig. 6c). All these GWASs published in the last decade have increased our knowledge about the genetic architecture of common diseases a lot. In this section, we will go through some of the insights that have been revealed by these studies.
5.1 Polygenic Architecture of Common Diseases GWASs have consistently shown that most complex traits and diseases have very polygenic architectures with a large number of causal variants with small effects. The small effect sizes mean that enormous sample sizes are needed to detect the associated variants and that each variant only explains a small fraction of the heritability. Even though large sample sizes have led to the discovery of many loci affecting common diseases, the aggregated effect of all these loci still only explains a small fraction of the heritability.
> A good example is type 2 diabetes where researchers by 2012 had identified 63 associated loci that collectively only explained 5.7% of the liability-scale variance [48]. Such results led to much discussion about the possible source of the remaining "missing heritability" [49, 50]. A significant contribution to this debate was when researchers in 2010 started using mixed linear models to estimate the heritability explained by all common variants not only those that surpass a conservative significance threshold. These studies showed that a significant fraction of the so-called missing heritability was not truly missing from the GWAS data sets but only hidden due to small effect sizes. This was first illustrated in height where 180 statistically significant SNPs could only explain 10% of the heritability, but this fraction increased to 45% when all genotyped variants were considered [51].
> For common diseases, such analyses have typically shown that around half of the heritability can be explained by considering all common variants. Given the small individual contribution of each of the discovered variants and that the individual contribution of the yet to be found variants will be even smaller, it is likely that the actual number of causal variants will be much more than a thousand for many common diseases. Recent data shows that in many diseases these causal variants are relatively uniformly distributed along the genome. It has, for instance, been estimated that 71–100% of 1 MB windows in the genome contribute to the heritability of schizophrenia [52]. Another article recently estimated that most 100 kB windows contribute to the variation of height and that more than 100,000 markers have an independent effect on height. This strikingly large number leads the authors to propose a new "omnigenic" model in which most genes expressed in a cell type that is relevant for a given disease have a nonzero contribution to the heritability of that disease [53].
5.2 Pleiotropy The variants that have been discovered by GWASs so far reveal numerous examples where one genetic locus affects multiple often seemingly unrelated traits [54, 55]. One explanation for such a shared association between a pair of traits is mediation where the shared locus affects the risk of one of the traits, and that trait is causal for the other. Another possible explanation is pleiotropy where the shared locus is independently causal for both traits. It is possible to distinguish between mediation and true pleiotropy by adjusting or stratifying for one trait while testing the other. In the case of mediation, it is also possible to determine the direction of the causation. In general, it is difficult to make such causal inference from observational data, but Mendelian randomization, which uses significantly associated variants as instrumental variables, can in some circumstances be used to assess a causal relationship between a potential risk factor and a disease. For instance, Voight and colleagues used SNPs associated with lipoprotein levels to assess whether the correlation between different forms of lipoprotein and myocardial infarction risk was causal [56]. They found that while low-density lipoprotein (LDL) had a causal effect on disease risk, high-density lipoprotein (HDL) did not.
The fact that pleiotropy is widespread has several implications. One is that variants that have already been found to affect one trait can be prioritized in other studies since they are more likely also to affect another trait than a random variant is. Another implication is that we cannot always examine the effect of selection by studying one trait in isolation. There are multiple examples of antagonistic pleiotropy where a variant increases the risk of one disease while decreasing the risk of another.
5.3 Differences Between Diseases Because of differences in age of onset and severity, we do not expect identical allelic architectures in all common diseases. Using the currently available GWAS data sets, we can now start to identify these differences in the allelic architectures, but because of the significant differences in samples sizes and the number of tested variants, this is not an easy task.
The data available to date show that the degree of polygenicity differs between diseases with schizophrenia, for example, having more predicted loci than immune disorders [57] and hypertension [52]. Results also show that rare variants play a larger role in some diseases compared to others. Rare variants, for example, have a greater role in amyotrophic lateral sclerosis than in schizophrenia [58] and are even less important in lifestyle-dependent diseases such as type 2 diabetes [59].
#### 6 Perspectives
The price of whole-genome sequencing is still declining, and it is not unreasonable to expect that at some point in the future, a majority of people will get their genomes sequenced. At that point the availability of genetic data will no longer be a limiting factor in studies of common human diseases. In order to make the most of such huge data sets, the genetic information needs to be combined with high-quality phenotypic and environmental information. If that is achieved, we will be able to explain most—if not all—of the additive genetic variance for the common human diseases. Having large population data sets where genetic data is combined with extensive phenotypic data including information about lifestyle, diet and other environmental risk factors will also enable much better studies of pleiotropy and gene–environment interactions. A few large population data sets are already available now with the UK Biobank [29]—a prospective study of 500,000 individuals—being the best example.
While GWASs have found a lot of loci that are associated with common diseases, the actual causal variant and the functional mechanism driving the causation are still unknown for a large fraction of the loci. In order to understand the functional mechanism of a specific locus, it is necessary to combine sequence data with other types of data. This includes gene expression data (from the correct tissue) and epigenetic data such as methylation. Such data sets are fortunately also becoming cheaper to produce and thus more abundant as a result of falling sequencing costs. Furthermore large consortium data sets such as GTEx [60], ENCODE [61], and Roadmap Epigenomics [62] mean that each lab studying these mechanisms will not have to produce all the data themselves but can in part rely on these public data sets. It is thus likely that we in the future not only will find many more GWAS loci for each common disease but we will also have a much better understanding of how each of these loci affects the disease.
#### 7 Questions
#### References
stratification in genome-wide association studies. Nat Rev Genet 11:459–463. https://doi. org/10.1038/nrg2813
330:641–646. https://doi.org/10.1126/sci ence.1197005
Genet 45:1150–1159. https://doi.org/10. 1038/ng.2742
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Chapter 18
#### Ancestral Population Genomics
#### Julien Y. Dutheil and Asger Hobolth
#### Abstract
Borrowing both from population genetics and phylogenetics, the field of population genomics emerged as full genomes of several closely related species were available. Providing we can properly model sequence evolution within populations undergoing speciation events, this resource enables us to estimate key population genetics parameters such as ancestral population sizes and split times. Furthermore we can enhance our understanding of the recombination process and investigate various selective forces. With the advent of resequencing technologies, genome-wide patterns of diversity in extant populations have now come to complement this picture, offering an increasing power to study more recent genetic history.
We discuss the basic models of genomes in populations, including speciation models for closely related species. A major point in our discussion is that only a few complete genomes contain much information about the whole population. The reason being that recombination unlinks genomic regions, and therefore a few genomes contain many segments with distinct histories. The challenge of population genomics is to decode this mosaic of histories in order to infer scenarios of demography and selection. We survey modeling strategies for understanding genetic variation in ancestral populations and species. The underlying models build on the coalescent with recombination process and introduce further assumptions to scale the analyses to genomic data sets.
Key words Ancestral population, Coalescence, Demography, Divergence, Markov model, Migration, Recombination, Selection, Speciation
#### 1 Introduction
We are in the population genomics era where data sets from the 1000 human genomes project [1], the great apes project [2], and the 1001 arabidopsis genomes project [3] are available. The underlying data sets contain genotypic information for thousands of individuals in one or several species, in the form of de novo sequenced genomes or variation compared to an available "reference" genome (a.k.a. resequencing). By comparing genomes from several individuals of the same species or closely related species, we can obtain information about split times, population sizes, recombination events, and selection in contemporary and ancestral
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_18, © The Author(s) 2019
Fig. 1 Left: Isolation model of two species. Right: The coalescent process along the genomes of the two species. By comparing the two genomes we obtain information about the split time of the species and the ancestral population size. Furthermore the breakpoints along the genomes correspond to recombination events, so we also have information about the recombination process
species (see Fig. 1). In this chapter we discuss various models for obtaining this information.
Comparing homologous sequences available for a given locus to infer their degree of relatedness enables the discovery of the parental relationships of the sequences, depicted as a tree thereby named genealogy. When one sequence sampled from one individual of one species is compared with sequences from other species, the resulting genealogy contains information about the history of species, the so-called phylogeny. The phylogeny summarizes the relationship and the divergence times between the species.
Conversely, when sequences from several individuals within a species are sampled, we have access to the genetic variation in contemporary populations. The evolutionary forces that shape genetic variation within a species are genetic drift, mutation, recombination, and selection and are the subject of population genetics. The key modeling tool in population genetics is coalescent theory. Classical coalescent theory describes the genetic ancestry of a sample of homologous DNA sequences from the same species. This genealogical description includes times to common ancestry, which is measured back into the past.
Molecular phylogenetics and population genetics have accumulated 50 years of methodological developments. The convergence of these two fields and their key mathematical and statistical tools is needed in order to fully understand genomic sequence alignments, because comparing genealogies and phylogenies is at the heart of the study of the speciation process [4].
We describe the interplay between population genetics and phylogenetics by reviewing the methods and models that have been developed to understand evolutionary history from genomic data (see Table 1 for a comparative summary of all methods).
(continued)
This table summarizes and compares existing ancestral population genomics methods. Parameters correspond to the one in Fig. 4. RAS: Rate across site model, assuming an a priori distribution of evolutionary rate (usually a discretized gamma distribution) over alignment positions
#### 2 Coalescent Theory and Speciation
We start by describing the standard coalescent model within one population. The coalescent model describes the shape of the genealogy of several sequences sampled from a single population. For more information on the coalescent, we refer to [21, 22] and [23]. This section describes the coalescent process as a chronological process. In the next section, we will see how it can be modeled as a spatial process along the genome. In subsequent sections we extend the standard model to include two or more populations. In the cases where multiple populations are present we describe both the isolation model and the isolation-with-migration model.
2.1 The Standard Coalescent Model The standard coalescent model is a continuous-time approximation of the neutral Wright–Fisher model. In the Wright–Fisher model the number of chromosomes 2N (we consider diploid organisms) is fixed in each non-overlapping generation. Each chromosome in a new generation chooses its ancestor uniformly at random from the previous generation.
> Consider two chromosomes. The probability of the two chromosomes choosing the same ancestor is 1/(2N) and the probability of the two chromosomes not finding a common ancestor is 1 - 1/(2N). Let R<sup>2</sup> denote the number of generations back in time when the two individuals find a most recent common ancestor (MRCA). By repeating the argument above, the probability of the two chromosomes not finding a common ancestor r generations back in time is
$$P(R\_2 > r) = \left(1 - \frac{1}{2N}\right)^r.$$
$$\mathbb{P}\_{\bullet \bullet \bullet \bullet \bullet \bullet}$$
If we scale time <sup>t</sup> in units of 2N, i.e., set <sup>r</sup> ¼ <sup>2</sup>Nt, we get
$$P(R\_2 > r) = \left(1 - \frac{1}{2N}\right)^r = \left(1 - \frac{1}{2N}\right)^{2Nt} \approx \varepsilon^{-t},$$
where the approximation is valid for large N. In coalescent time units the waiting time <sup>T</sup><sup>2</sup> <sup>¼</sup> <sup>R</sup>2/(2N) before coalescence of two individuals is therefore exponentially distributed with mean one.
These considerations can be extended to multiple individuals. In general the time Tn before two of n individuals coalesce is exponentially distributed with rate <sup>n</sup> 2 .
The waiting time Wn for a sample of n individuals to find the most recent common ancestor (MRCA) is given by
$$W\_n = T\_n + T\_{n-1} + \dots + T\_2,$$
where Tk are independent exponential random variables with parameter <sup>k</sup> 2 ; see Fig. 2 for an illustration. It follows that the mean of Wn is
Fig. 2 Illustration of the coalescent process. The waiting time before two out of n individuals coalesce is Tn and the time before a sample of n individuals find common ancestry is Wn
$$\begin{aligned} E[W\_n] &= \sum\_{k=2}^n E[T\_k] = \sum\_{k=2}^n \frac{2}{k(k-1)} = 2\sum\_{k=2}^n \left(\frac{1}{k-1} - \frac{1}{k}\right), \\ &= 2\left(1 - \frac{1}{n}\right). \end{aligned}$$
$$\begin{aligned} \dots \end{aligned}$$
Note that limn!1E[Wn] <sup>¼</sup> 2. The variance of Wn is
$$\begin{aligned} \operatorname{Var}[W\_n] &= \sum\_{k=2}^n \operatorname{Var}[T\_k] = \sum\_{k=2}^n \binom{k}{2}^{-2} \\ &= 8 \sum\_{k=1}^{n-1} \frac{1}{k^2} - 4 \left(1 - \frac{1}{n}\right) \left(3 + \frac{1}{n}\right). \end{aligned}$$
Note that limn!1Var½<sup>W</sup> <sup>n</sup>¼ð8π<sup>2</sup> <sup>6</sup> -<sup>12</sup>Þ ¼ <sup>1</sup>:16.
The consequences of these calculations are that when we only sample within a population we are limited to relatively recent events. The expected time for a large sample to find their MRCA is approximately 2 (2N) ¼ <sup>4</sup><sup>N</sup> generations with standard deviation ffiffiffiffiffiffiffiffiffiffi <sup>1</sup>:<sup>16</sup> <sup>p</sup> ð2<sup>N</sup> Þ ¼ <sup>2</sup>:15<sup>N</sup> generations. As a consequence, a neutral sample within a population contains little information beyond 6N generations.
Humans have a generation time of approximately 20 years and an effective population size of approximately <sup>N</sup> ¼ 10, 000 (see [21, p. 251]), and therefore 6N generations correspond to approximately 1.2 million years (My) for humans. Therefore human diversity at neutral loci contains little demographic information beyond 1.2 My.
Now suppose mutations occur at a rate u per locus per generation. In a lineage of r generations, we then expect ru mutations or in the coalescent time units with <sup>r</sup> ¼ <sup>2</sup>Nt we expect 2Ntu mutations. We let <sup>θ</sup> ¼ <sup>4</sup>Nu be the mutation rate parameter. Since <sup>u</sup> is small we can make a Poisson approximation of the binomial number of mutations in a lineage of r generations
$$\text{Bin}(r,\mathfrak{u}) = \text{Bin}(2Nt, \theta/(2\cdot 2N)) \approx \text{Pois}(t\theta/2).$$
We have thus arrived at the following two-step process for simulating samples under the coalescent: (a) simulate the genealogy by merging lineages uniformly at random and with waiting times exponentially distributed with rate <sup>n</sup> 2 when n lineages are present; (b) on each lineage in the tree add mutations according to a Poisson process with rate θ/2.
Another possibility is to scale the coalescent process such that one mutation is expected in one time unit. In this case the exponentially distributed waiting times in (a) have rate <sup>n</sup> 2 ð2=θÞ, and in (b) the mutations are added with unit rate. We use the latter version of the coalescent-with-mutations process below.
For species where recombination occurs, different parts of the genome come from distinct ancestors, and therefore have a distinct history. Figure 3 exemplifies this phenomenon for two species. It displays the genealogical relationships for two sequences which underwent a single recombination event. In the presence of recombination, each position of a genome alignment therefore has a specific genealogy, and close positions are more likely to share the same one (recall Fig. 1). The genome alignment can therefore be described as an ordered series of genealogies, spanning a variable amount of sites, and then changing because of a recombination event [4]. The genealogy is therefore depicted as a complex graph with nodes representing both coalescence and recombination events, the ancestral recombination graph (ARG, Fig. 3c). A single genome thus contains different samples from the distribution of the age of the MRCA, and the distribution contains information about the ancestral population size and speciation time. The coalescent with recombination serves as a basis for modeling genome-wide genealogy, a point that we will further develop in Subheading 4.
#### 3 Adding Genetic Barriers and Gene Flow to the Picture: The Structured Coalescent
In this section we extend the standard coalescent model. We consider coalescent models with multiple species and introduce population splits or speciation events. The models that we describe are
2.3 Taking Recombination into Account
2.2 Adding Mutations to the Standard Coalescent
Model
Fig. 3 Ancestral recombination graph for two species. (a) Genealogy of four sampled sequences from two species. The bold line shows the divergence of two sequences of interest. (b) A single recombination event happened between the lineages of sequences 3 and 4 (horizontal line), so that in a part of the sequences, the genealogy is as depicted by the bold line and therefore displays an older divergence. (c) The corresponding ancestral recombination graph (in black) with the trees of each side of the recombination break point superimposed (red: left tree; blue: right tree). When going backward in time, a split corresponds to a recombination event and a merger to a coalescence event
shown in Fig. 4 (see also Table 1) and include: (a) The two species isolation model; (b) The two species isolation-with-migration models; (c) The three species isolation model (and incomplete lineage sorting); and (d) The three species isolation-with-migration model. We also discuss the general multiple species isolation-withmigration model. The two species isolation model was introduced in [24] and the isolation-with-migration model was introduced in [25].
3.1 Isolation Model with Two Species If the sequences are sampled from two distinct species that have diverged a time T ago (see Fig. 4a), then the distribution of the age of the MRCA is shifted to the right with the amount T, resulting in the distribution
Fig. 4 Speciation models and associated parameters. In all exemplified models effective population size is constant between speciation events, represented by dash lines. The timing of the speciation events, noted T are parameters of the models, together with ancestral effective population sizes, noted NA. In some cases, contemporary population sizes can also be estimated, and are noted Ni , where i is the index of the population. Models with post-divergence genetic exchanges have additional migration parameters labeled <sup>m</sup>from!to. The number of putative migration rates increases with the number of contemporary populations under study, and some models might consider some of them to be equal or eventually null to reduce complexity. (a) Isolation model with two species. (b) Isolation-migration model with two species. (c) Isolation model with three species. (d) Isolation-Migration model with three species
$$f\_{T\_2}(t) = \begin{cases} 0 & \text{if } t < T \\ \frac{2}{\theta\_A} e^{-2(t-T)/\theta\_A} & \text{if } t > T \end{cases}$$
where <sup>θ</sup><sup>A</sup> <sup>¼</sup> <sup>4</sup>NA <sup>u</sup> is the ancestral mutation rate. The mean time to coalescent is <sup>E</sup>[T2] ¼ <sup>T</sup> <sup>+</sup> <sup>θ</sup>A/2 and the average divergence time between two sequences is twice this quantity, that is, 2T + θA. Since <sup>θ</sup><sup>A</sup> <sup>¼</sup> <sup>4</sup>NAu it follows that the larger the size of the ancestral population, the bigger the difference between the speciation time and the divergence time.
The variance of the divergence time is Var½<sup>T</sup> <sup>2</sup> ¼ <sup>θ</sup><sup>2</sup> <sup>A</sup>=4. With access to the distribution of divergence times, we could estimate the speciation time and population size from the mean and variance of the distribution. Unfortunately we do not know the complete distribution of divergence times and it is not immediately available to us, because long regions are needed for precise divergence estimation but have experienced one or more recombination events.
Now consider the isolation model with three species depicted in Fig. 4c. Such a model is often used for the human–chimpanzee– gorilla (HCG) triplet (e.g., [10–12]).
The density function for the time to coalescence between sample 1 and sample 2 is given by
$$f\_{T\_2}(t) = \begin{cases} 0 & \text{if } t < T\_1 \\ \frac{2}{\theta\_{A1}} e^{-2(t - T\_1)/\theta\_{A1}} & \text{if } T\_1 < t < T\_{12} \\ \\ P\_{12} \frac{2}{\theta\_{A2}} e^{-2(t - T\_{12})/\theta\_{A2}} & \text{if } t > T\_{12}, \end{cases} \tag{1}$$
where
$$T\_{12} = T\_1 + T\_2 \quad \text{and} \quad P\_{12} = \varepsilon^{-2(T\_{12} - T\_1)/\theta\_{A1}}$$
is the probability of the two samples not coalescing in the ancestral population of sample 1 and sample 2. In the upper right corner of Fig. 5 we plot the density (Eq. 1) with parameters that resemble the HCG triplet.
If sample 1 and sample 2 do not coalesce in the ancestral population of sample 1 and sample 2, then the three trees ((1,2),3), ((1,3),2), and ((2,3),1) are equally likely. The probability of the gene tree being different from the species tree is thus
$$\Pr(\text{incongruenc}) = \frac{2}{3} P\_{12} = \frac{2}{3} e^{-2(T\_{12} - T\_1)/\theta\_{A1}}.\tag{2}$$
The event that the gene tree is different from the species tree is called incomplete lineage sorting (ILS). ILS is important because species tree incongruence often manifests itself as a relatively clear signal in a sequence alignment and thereby allows for accurate estimation of population parameters. In Fig. 6 we show the (in) congruence probability Eq. 2. We also refer to Exercise 1 (see Subheading 8.1) and Exercise 2 (see Subheading 8.2) for more discussion of ILS.
3.2 Isolation Model with Three or More Species and Incomplete Lineage Sorting
Fig. 5 Illustration of the density for coalescent in various models and data layout. The curves are the probability density functions. In the most simple case with two species, a constant ancestral population size and a punctual speciation (top left panel), more genomic regions find a common ancestor close to the species split (the vertical line), while a few regions have a more ancient common ancestor, distributed in an exponential manner (see Eq. 1). If speciation is not punctual and migration occurred after isolation of the species, then some sequences have a common ancestor which is more recent than the species split and the distribution in the ancestor becomes more complex (bottom left panel, see Eqs. 4 and 6). When a third species is added (right panel), then another discontinuity appears and all distributions depend on additional parameters, particularly when migration is allowed. We use <sup>θ</sup>A<sup>1</sup> <sup>¼</sup> 0.0062, <sup>θ</sup>A<sup>2</sup> <sup>¼</sup> 0.0033 and <sup>τ</sup><sup>1</sup> <sup>¼</sup> 0.0038 (the first vertical line), <sup>τ</sup><sup>2</sup> <sup>¼</sup> 0.0062 (the second vertical line) corresponding to the HCG triplet. Ancestral population sizes are taken from the simulation study in Table 6 in Wang and Hey [8]: <sup>θ</sup><sup>1</sup> <sup>¼</sup> 0.005 and <sup>θ</sup><sup>2</sup> <sup>¼</sup> 0.003. Migration parameters are all set to 50
In the three species isolation model the mean coalescent time for a sample from population 1 and a sample from population 2 is given by
$$E[T\_2] = T\_1 + (1 - P\_{12})\frac{\theta\_{A1}}{2} + P\_{12}\frac{\theta\_{A2}}{2}.\tag{3}$$
Fig. 6 Probability (Eq. 2) of gene tree and species tree being incongruent. In case of the HCG triplet we obtain (T<sup>12</sup> - <sup>T</sup>1)/θA<sup>1</sup> <sup>¼</sup> (0.0062 - 0.0038)/ 0.0062 ¼ 0.39 which corresponds to an incongruence probability of 30%
Burgess and Yang [9] describe the speciation process for human, chimpanzee, gorilla, orangutan (O), and macaques (M) using an isolation model with five species. The HCGOM model contains four ancestral parameters θHC, θHCG, θHCGO, and θHCGOM. In this case (Eq. 3) extends to
$$\begin{split} E[T\_2] &= T\_{\rm HC} + (1 - P\_{\rm HC}) \frac{\theta\_{\rm HC}}{2} P\_{\rm HC} (1 - P\_{\rm HCG}) \frac{\theta\_{\rm HCG}}{2} \\ &+ P\_{\rm HC} P\_{\rm HCG} (1 - P\_{\rm HCGO}) \frac{\theta\_{\rm HCGO}}{2} \\ &+ P\_{\rm HC} P\_{\rm HCG} P\_{\rm HCGO} (1 - P\_{\rm HCGM}) \frac{\theta\_{\rm HCGM}}{2} .\end{split}$$
3.3 Isolation-with-Migration Model with Two Species and Two Samples
The isolation-with-migration (IM) model with two species is shown in Fig. 4b. The IM-model has six parameters: The mutation rates θ1, θ2, and θA, the migration rates m<sup>1</sup> and m2, and the speciation time <sup>T</sup>. We let <sup>Θ</sup> ¼ (θ1, <sup>θ</sup>2, <sup>θ</sup>A, <sup>m</sup>1, <sup>m</sup>2, <sup>T</sup>) be the vector of parameters.
Wang and Hey [8] consider a situation with two genes. Before time T the system is in one of the following five states:
The instantaneous rate matrix Q is given by
Starting in state a, the density for coalescent in population 1 at time t < T is given by [26]
$$f\_1(t) = (\mathcal{e}^{\mathcal{Q}f})\_{aS\_{11}} (\mathcal{Q}/\theta\_1),\tag{4}$$
the density for coalescent in population 2 at time t < T is
$$f\_{\,2}(t) = (\mathfrak{e}^{\mathcal{Q}\_{\,f}})\_{a\mathcal{S}\_{22}}(\mathfrak{2}/\theta\_2),\tag{5}$$
and the total density for a coalescent at time t < T is
$$f(t) = f\_1(t) + f\_2(t). \tag{6}$$
Here <sup>e</sup><sup>A</sup> ¼ <sup>P</sup><sup>1</sup> <sup>i</sup>¼<sup>0</sup> <sup>A</sup><sup>i</sup> <sup>=</sup>ði!Þ is the matrix exponential of the matrix A and (e <sup>A</sup>)ij is entry (i, j) in the matrix exponential.
After time T the system only has two states: SAA corresponding to two genes in the ancestral population and SA corresponding to one single gene in the ancestral population. The rate of going from state SAA to state SA is 2/θA. The density for coalescent in the ancestral population at time t > T is therefore
$$f(t) = \left[ (\varepsilon^{\mathcal{Q}T})\_{aS\_{11}} + (\varepsilon^{\mathcal{Q}T})\_{aS\_{12}} + (\varepsilon^{\mathcal{Q}T})\_{aS\_{22}} \right] \frac{2}{\theta\_A} \varepsilon^{-(2/\theta\_A)(t-T)}.\tag{7}$$
In Fig. 5 we illustrate the coalescent density in the two species isolation-with-migration model.
The likelihood for a pair of homologous sequences X is given by
$$P(X|\theta) = L(\theta|X) = \int\_0^\infty P(X|t)f(t|\theta)dt\tag{8}$$
where <sup>f</sup>(t) ¼ <sup>f</sup> (t|Θ) given by Eqs. <sup>6</sup> and <sup>7</sup> is the density of the two sequences finding a MRCA at time t and P(X|t) is the probability of the two sequences given that they find a MRCA at time t. The latter term is calculated using a distance-based method. One possibility is to use the infinite sites model where it is assumed that substitutions happen at unique sites, i.e., there are no recurrent substitutions. In this case the number of differences between the two sequences follows a Poisson distribution with rate 1.
For an application of the isolation-with-migration model with two sequences, we refer to [8]; a discussion of their approach can be found in [27].
3.4 Isolation-with-Migration Model with Three or More Species and Three or More Samples
Hey [28] considered the multipopulation isolation-with-migration (IM) model. Recall from Fig. 4b that the two-population IM model has six parameters: two present population sizes, one ancestral population size, one speciation time, and two migration rates. The three-population IM model in Fig. 4d has fifteen parameters: three present population sizes, two ancestral population sizes, two speciation times, and eight migration rates. In general a k-population IM model has 3<sup>k</sup> - 2 + 2(<sup>k</sup> -1)<sup>2</sup> parameters:
See Fig. 5 for an example of divergence distribution with three species and migration and Exercise 3 (see Subheading 8.3) for a derivation of the number of migration rates in the general k-population model. For <sup>k</sup> ¼ 5, 6, and 7 we obtain 45, 66, and 91 parameters. Because the number of parameters becomes very large even for small k, Hey [28] suggests adding constraints to the migration rates, e.g., setting some rates to zero or introducing symmetry conditions where rates between populations are the same.
#### 4 Approximating the Coalescent with Recombination Along Genomes
Before the genomic era, multilocus population genetics models were addressing a small fraction of the complete ancestral recombination graph (ARG) by considering independent loci. As sequencing technologies evolved and allowed access to larger samples of genomic diversity, this independence assumption had to be relaxed and more explicit modeling of the ARG was required. Yet the complexity of the coalescent with recombination process makes its application to genome-scale data sets very challenging. Two directions of analysis methods have emerged: simulation-based or spatial approximations along the genome. In this chapter we focus on the latter and refer to Kelleher et al. [29] and Staab et al. [30] for the former. Simonsen and Churchill [31] described the first model of the joint distribution of genealogies at two loci for two genomes. Wiuf and Hein [32] extended this approach and described the coalescent as a spatial process along the genome. McVean and Cardin [33] further approximated the description with a Markov process. In this section we describe and discuss these types of approximations.
4.1 The Independent Loci Approach: Free Recombination Between, No Recombination Within
The simplest way to handle issues relating to the ancestral recombination graph is to divide the data into presumably independent loci. Such analyses are therefore restricted to candidate regions that are not too large (to avoid including a recombination point) and not too close (to ensure several recombination events happened between loci). Each region can then be described by a single underlying tree, reducing the analytical and computational load.
Using 15,000 loci distant from 10 kb totaling 7.4 Mb and the isolation model introduced above, Burgess and Yang [9] (Table 2, model (b) sequencing errors) find the following ancestral population sizes and speciation times estimates for human (H), chimpanzee (C), gorilla (G), orangutan (O), and macaque (M) ancestors: <sup>θ</sup>HC <sup>¼</sup> 0.0062, <sup>θ</sup>HCG <sup>¼</sup> 0.0033, <sup>θ</sup>HCGO <sup>¼</sup> 0.0061, <sup>θ</sup>HCGOM <sup>¼</sup> 0.0118 and <sup>T</sup>HC <sup>¼</sup> 0.0038, <sup>T</sup>HCG <sup>¼</sup> 0.0062, <sup>T</sup>HCGO <sup>¼</sup> 0.0137, <sup>T</sup>HCGOM <sup>¼</sup> 0.0260. Converting these estimates into time units requires an estimate of the substitution rate, either absolute or deduced from a scaling point. Using <sup>u</sup> ¼ <sup>10</sup>-<sup>9</sup> as an estimate for substitutions per year, this leads to an estimate of 3.8 My for the human–chimpanzee speciation, a very recent estimate. Using the same data, Yang [10] showed that the isolation-withmigration model was preferred. Yang finds a more ancient speciation time <sup>T</sup>HC <sup>¼</sup> 0.0053 (5.3 My with <sup>u</sup> <sup>¼</sup> <sup>10</sup>-9 ) when migration is accounted for.
4.2 State-Space Model: Simonsen–Churchill Framework
The coalescent with recombination for two loci and two sequences is originally described in Simonsen and Churchill [31] as a continuous-time Markov chain backward in time with eight states as shown in Fig. 7. This Markov chain is given a careful treatment in the textbooks by Durrett [34, Section 3.1.1] and Wakeley [21, Section 7.2.4], and we therefore only briefly explain the basic properties of the model here.
A single sequence is either linked ( , , , or meaning that it contains material ancestral to the sample at both loci, or it is unlinked ( , , , or ) when it contains material ancestral to the sample at only one locus. The coalescent rate is one for any two sequences, and the recombination rate is ρ/2 for any linked sequence. The chain begins at time zero in state 1 with two linked sequences. After an exponential waiting time with rate 1 + ρ the chain enters state 8 with probability 1/(1 + ρ) or state 2 with probability ρ/(1 + ρ). The transition from state 1 to state 8 is a coalescent event, and the left and right tree heights are identical. The transition from state 1 to state 2 is a recombination event that breaks apart one of the two sequences. All other transitions have similar interpretations. Common ancestry for a locus is marked
Fig. 7 State transition diagram for two loci and two sequences described as a continuous-time Markov chain backward in time. The figure is adapted from Figure 7.7 in Wakeley [21]. A line with a bullet or a cross at both ends is a linked sequence (ancestral material to the sample at both loci), whereas a line with a bullet or a cross at one end only is a sequence with ancestral material at one locus only. A cross denotes common ancestry. s and t denote the heights of the left and right trees, respectively
with a , so the transition from, e.g., state 1 to state 8 is a transition to the state .
The height S of the left tree is the first time at which the process enters one of the states 5, 7, or 8 (states with a left ), and the height T of the right tree is the first time at which one of the states 4, 6, or 8 is entered (states with a right ). When state 8 is entered from state 1 the two tree heights are identical. State 8 is absorbing because only the tree heights are of interest.
The two key ingredients for the state-space model are the conditional probability for staying in a state <sup>P</sup>(<sup>T</sup> ¼ <sup>s</sup>|<sup>S</sup> ¼ <sup>s</sup>) and the conditional density q(t|s) of a new tree height t conditional on a change and a previous tree height s. Hobolth and Jensen [35] show that the conditional probability of no change from the left to the right tree is
$$P(T=\mathfrak{s}|S=\mathfrak{s}) = \mathfrak{e}^{\mathfrak{s}}[\mathfrak{e}^{\Lambda\mathfrak{s}}]\_{11},\tag{9}$$
and the conditional density <sup>q</sup>(t|s) of <sup>T</sup> given <sup>S</sup> ¼ <sup>s</sup> and given <sup>T</sup> 6¼ <sup>S</sup> is
$$q(t|s) = \begin{cases} \left. e^{-(s-t)} \frac{[\varepsilon^{\Lambda t}]\_{12} + [\varepsilon^{\Lambda t}]\_{13}}{e^{-s} - [\varepsilon^{\Lambda t}]\_{11}} \right| & t < s, \\\\ \left. e^{-(t-s)} \frac{[\varepsilon^{\Lambda t}]\_{12} + [\varepsilon^{\Lambda t}]\_{13}}{e^{-s} - [\varepsilon^{\Lambda t}]\_{11}} & t > s, \end{cases} \tag{10}$$
where <sup>Λ</sup> denotes the 8 8 rate matrix from Fig. 7.
Wakeley [21, Section 7.2.4] noted that the transitions between state 4 and 6 and the transitions between state 5 and 7 can be removed from the chain if we are only interested in the tree heights. Actually, even more transitions can be removed from the chain. Note from Eqs. 9 and 10 that we only need the entries (1, 1), (1, 2), and (1, 3) in e <sup>Λ</sup><sup>t</sup> for calculating the probability of the same tree height in the next position and the transition density conditional on a change. These entries can be found from a reduced rate matrix where states 4, 5, 6, and 7 are removed and the rate from states 2 and 3 to a new absorbing state equals 2. In other words, define the reduced rate matrix
$$
\tilde{A}\_{\pm} = \begin{pmatrix}
1 & -(3+\rho/2) & \rho/2 & 2 \\
0 & 4 & -6 & 2 \\
0 & 0 & 0 & 0
\end{pmatrix},
$$
where states are numbered 1, 2, 3, and 4. The holding time and transition density for the model are now given by Eqs. 9 and 10 with Λ substituted by Λ~.
In the left plot in Fig. 8 we illustrate the probability (Eq. 9) of the same tree height in the left and right loci conditional on the tree height in the left locus and different recombination rates.
Fig. 8 (a) Probability of same tree height. (b) Density for right tree height conditional on the left tree height being equal to s and a recombination rate equal to ρ
As expected the probability for identical tree heights decreases with the height of the left tree and with the recombination rate.
In the right plot in Fig. 8 we illustrate the density (Eq. 10) of the right tree height conditional on the left tree height and a change in tree height. When the recombination rate increases, the density for the right tree height moves toward smaller tree heights. The reason is that at least one recombination is needed for having a change in tree height. We also observe that the density is continuous but not differentiable in the position of the left tree height.
#### 4.3 Time Discretization: Setting Up the Finite State HMM
Li and Durbin [14] and Mailund et al. [13] analyze pairs of sequences using a hidden Markov model (HMM). The hidden states are tree heights (times to the most recent common ancestor), and the tree height is discretized to obtain a finite hidden state space. The observed states of the HMM are alignment columns, with probabilities corresponding to a substitution process on the tree (see Fig. 9). In the Li and Durbin model, an infinite site model is assumed and observed states are converted to binary data, telling whether the site is heterozygous (one mutation) or homozygous (no mutation).
We now describe how we discretize time for the case of two sequences considered in the previous section. The discrete version of the Markov process is used to build a finite Markov chain along the two sequences. When the finite Markov chain is combined with a substitution process, we obtain an HMM as in Li and Durbin [14].
Let the discrete time points (backward in time) of the Markov chain be <sup>d</sup><sup>0</sup> <sup>¼</sup> <sup>0</sup> <sup>&</sup>lt; <sup>d</sup><sup>1</sup> <sup>&</sup>lt; <sup>d</sup><sup>2</sup> <sup>&</sup>lt; <sup>&</sup>lt; dM-<sup>1</sup> <sup>&</sup>lt; dM ¼ 1 and denote the corresponding states by 1, 2, ..., M. State m (m ∈{1, ..., M}) then corresponds to a tree height in the interval between dm-<sup>1</sup> and dm. The continuous stationary distribution is <sup>π</sup>ðtÞ ¼ expð<sup>t</sup>Þ, and therefore the discrete times are chosen such that
Fig. 9 (a) Graphical structure of the hidden Markov Model. (b) Simulation from the hidden Markov model
1 expðdmÞ ¼ <sup>m</sup>=M, or dm ¼ logð<sup>1</sup> <sup>m</sup>=MÞ, where we define logð0Þ ¼ -1.
We now get for 1 <sup>ℓ</sup>, <sup>r</sup> <sup>M</sup> the joint density
$$\begin{split} P(L=\ell, R=r) \\ \begin{cases} \sum\_{k \in \{5,7\}} \sum\_{j \in \{5,7\}} \sum\_{i \in \{1,2,3\}} [\epsilon^{\lambda d\_{\ell-1}}]\_{1i} \\ [\epsilon^{\lambda(d\_{\ell}-d\_{\ell-1})}]\_{ij} [\epsilon^{\lambda(d\_{r-1}-d\_{\ell})}]\_{jk} [\epsilon^{\lambda(d\_r-d\_{r-1})}]\_{k8} & \text{if } \ell < r \\ \sum\_{i \in \{1,2,3\}} [\epsilon^{\lambda d\_{\ell-1}}]\_{0i} [\epsilon^{\lambda(d\_{\ell}-d\_{\ell-1})}]\_{i8} & \text{if } \ell = r \\ P(L=r, R=\ell) & \text{if } \ell > r. \end{cases} \end{split}$$
ð11Þ The reason for the first case is that in order for the left tree height to be in state <sup>ℓ</sup> <sup>&</sup>lt; <sup>r</sup>, it must be in state 1, 2, or 3 at time <sup>d</sup>ℓ-<sup>1</sup> and in state 5 or 7 at time d<sup>ℓ</sup> (i.e., there have been no coalescent events before time <sup>d</sup>ℓ-<sup>1</sup> and a left coalescent event between time <sup>d</sup>ℓ-<sup>1</sup> and <sup>d</sup>ℓ), and similarly it must still be in state 5 or 7 at time dr-<sup>1</sup> and in state 8 at time dr (i.e., there have been no coalescent events between time <sup>d</sup><sup>ℓ</sup> and time dr-<sup>1</sup> and a right coalescent event between time dr-<sup>1</sup> and time dr). The next case corresponds to no coalescent events before time <sup>d</sup>ℓ-<sup>1</sup> and both a left and a right coalescent event between time <sup>d</sup>ℓ-<sup>1</sup> and dℓ. The last case is due to symmetry of the chain.
From the joint tree states (ℓ, r) we easily get the conditional tree states
$$P\_{(\ell,r)} = P(r|\ell) = P(R=r|L=\ell) = \frac{P(L=\ell, R=r)}{P(L=\ell)},$$
where <sup>P</sup>(<sup>L</sup> ¼ <sup>ℓ</sup>) ¼∑rP(<sup>R</sup> <sup>¼</sup> <sup>r</sup>, <sup>L</sup> <sup>¼</sup> <sup>ℓ</sup>). These probabilities are used in the HMM.
4.4 Careful Treatment of Mutation Process
A careful treatment of the mutation process allows for a more coarse binning procedure and is needed to avoid biasing the results. In continuous time the probability for a mutation given a tree height <sup>t</sup> is given by <sup>μ</sup>ðtÞ ¼ <sup>1</sup> expð<sup>θ</sup>tÞ, and the stationary tree height distribution is <sup>π</sup>ðtÞ ¼ expð<sup>t</sup>Þ. The probability of a mutation conditionally on the hidden state m becomes
$$\begin{split} \mu\_{m} &= p(\boldsymbol{\gamma}\_{i} = 1 | \boldsymbol{\chi}\_{i} = m) \\ &= p(\boldsymbol{\gamma}\_{i} = 1 | t \in (\boldsymbol{d}\_{m-1}, \boldsymbol{d}\_{m})) = \frac{p(\boldsymbol{\gamma}\_{i} = 1 | t \in (\boldsymbol{d}\_{m-1}, \boldsymbol{d}\_{m}))}{p(t \in (\boldsymbol{d}\_{m-1}, \boldsymbol{d}\_{m}))} \\ &= \frac{\int\_{\boldsymbol{d}\_{m-1}}^{\boldsymbol{d}\_{m}} p(\boldsymbol{\gamma}\_{i} = 1 | t) \pi(t) dt}{\int\_{\boldsymbol{d}\_{m-1}}^{\boldsymbol{d}\_{m}} \pi(t) dt} = \frac{\int\_{\boldsymbol{d}\_{m-1}}^{\boldsymbol{d}\_{m}} (1 - \varepsilon^{-\theta t}) \varepsilon^{-t} dt}{\int\_{\boldsymbol{d}\_{m-1}}^{\boldsymbol{d}\_{m}} \varepsilon^{-t} dt} \\ &= 1 - \varepsilon^{-\theta \boldsymbol{d}\_{m-1}} \frac{\left(1 - \varepsilon^{-(1+\theta)(\boldsymbol{d}\_{m} - \boldsymbol{d}\_{m-1})}\right)}{(1+\theta)\left(1 - \varepsilon^{-(\boldsymbol{d}\_{m} - \boldsymbol{d}\_{m-1})}\right)}. \end{split} \tag{12}$$
Note that with a fine discretization we have that the interval dm - dm-<sup>1</sup> is small and the first-order Taylor expansion expðazÞ <sup>1</sup> az for z small gives
$$p(\mathbf{y}\_i = \mathbf{l} | \mathbf{x}\_i = m) \approx \mathbf{l} - \boldsymbol{\varepsilon}^{-\theta d\_{m-1}},$$
as perhaps expected. We are, however, discretizing the interval [0, 1[, so it is not possible to avoid one or more large bins. Generally we have found that a careful treatment of the mutation process is crucial for accurate inference [36].
4.5 Statistical Inference of Population Parameters from Sequences Here we choose to focus on three inference methods for estimating the recombination rate. The first method is based on the full likelihood obtained from the classical forward (or backward) algorithm for HMMs. The second is based on the distribution of the distance between segregating sites. This summary statistics was used in Harris and Nielsen [37] for demographic inference. It is sometimes also described as the distribution of the distance between heterozygote sites, runs of homozygosity, or the nearestneighbor distribution. The third summary statistics is the probability that two sites at certain distance apart are both heterozygote sites. This probability is closely related to the pair correlation function from spatial statistics [36] and to the zygosity correlation introduced in [38].
4.5.1 Summary Statistics: Runs of Homozygosity and Pair Correlation
Recall that in continuous time the probability for a mutation given a tree height <sup>t</sup> is given by <sup>μ</sup>ðtÞ ¼ <sup>1</sup> expð<sup>θ</sup>tÞ, and the stationary tree height distribution is <sup>π</sup>ðtÞ ¼ expð<sup>t</sup>Þ. The marginal probability for a mutation is therefore given by
$$\int\_0^\infty \mu(t)\pi(t)dt = \theta/(1+\theta). \tag{13}$$
Fig. 10 (a) Stationary distribution of tree height conditional on a mutation. (b) Probability of a mutation at various distances away from a mutation. (c) Probability of the first mutation at various distances away from a mutation
We also get the stationary distribution
$$\phi(t) = \frac{\mu(t)\pi(t)}{\int\_0^\infty \mu(t)\pi(t)dt} = \frac{1+\theta}{\theta}e^{-t}\left(1 - e^{-\theta t}\right),$$
for a tree height t conditional on a mutation. Figure 10a shows ϕ(t) for different values of θ. Note that small mutation rates imply a higher tree height when we condition on a mutation. In discrete time the probability for a mutation given a tree height m was given by Eq. 12. Let <sup>μ</sup> <sup>¼</sup> (μ1, ..., <sup>μ</sup>M) be the vector of mutation probabilities. The stationary distribution <sup>ϕ</sup> ¼ (ϕ1, ..., <sup>ϕ</sup>M) for a state m conditional on a mutation is given by
$$\phi\_{\ell} = \frac{\mu\_{\ell}\pi\_{\ell}}{\sum\_{m=1}^{M} \mu\_{m}\pi\_{m}},$$
where <sup>π</sup><sup>m</sup> <sup>¼</sup> 1/<sup>M</sup> because this is how the time discretization was chosen.
The probability for a mutation at a distance r from a typical mutation is then given by
$$\kappa(r) = \mathfrak{gl}' P^r \mu,$$
where <sup>0</sup> denotes vector transpose. In Fig. 10b we show κ(r) as a function of ρ and θ. Note that the curves converge to θ/(1 + θ) and that the behavior for small r is determined by the recombination rate.
The distribution of runs of homozygosity is given by
$$\nu(r) = \Phi'[P\text{diag}(e-\mu)]^{r-1}P\mu.$$
Here <sup>e</sup> ¼ (1, ..., 1) is the vector of length <sup>M</sup> with 1 in every entry and diag(<sup>e</sup> <sup>μ</sup>) is the diagonal matrix with <sup>e</sup> μ on the diagonal. In Fig. 10c we show ν(r) as a function of ρ and θ.
We estimate the mutation rate using an estimating equation based on the marginal probability for a mutation (Eq. 13). If the observed frequency of a mutation is p ^, then the mutation rate is θ ^ ¼ <sup>p</sup> ^=ð<sup>1</sup> p ^Þ (see left plot in Fig. 11). The recombination rate is estimated using maximum likelihood for the HMM and goodness of fit for the pair correlation (see middle plot in Fig. 11) and runs of homozygosity (see right plot in Fig. 11).
We simulated 50 sequences of length 20,000 base pairs and with mutation rate <sup>θ</sup> ¼ 0.1 and recombination rate <sup>ρ</sup> ¼ 0.1. We estimated the mutation rate using the estimating equation and the recombination rate using maximum likelihood and the HMM, and goodness of fit for the pair correlation and nearest neighbor (Fig. 12) [35]. As expected the HMM procedure shows the best results because here we are using all the available information. It seems, however, that we are not losing too much power when applying the pair correlation function. This is in contrast to the nearest-neighbor summary statistics that perform much worse than the other two methods.
We have provided a detailed treatment of the main components involved in an analysis of pair of DNA sequences based on an HMM derived from coalescent theory. Pairwise sequentially Markov coalescent (PSMC) models have been extensively applied to various organisms, see, for instance [39–43].
4.5.2 Parameter Estimation
Fig. 11 Parameter estimation for summary statistics. (a) The mutation rate θ is estimated from the observed number of mutations and length of the region. (b) The recombination rate ρ is estimated using the empirical distribution of a mutation at various distances from a mutation. (c) The recombination rate is estimated using the empirical distribution of the first mutation from a mutation
#### 5 Extending the Pairwise Sequentially Markov Coalescent
Extending the SMC to more than two genomes has proved to be challenging. The number of hidden states becomes prohibitive, as several divergence times have to be modeled and combined with distinct possible topologies. Further simplifications are therefore needed to account for an increasing number of genomes.
Fig. 12 Results of parameter estimation for simulation study. The pair correlation summary performs rather well compared to the full HMM data analysis. Nearest neighbor is a poor summary statistics
#### 5.1 From 2 to n Genomes
5.1.1 The Multiple Sequentially Markov Coalescent (MSMC)
Schiffels and Durbin [15] proposed to extend the PSMC model [14] to more than two haploid genomes by modeling the most recent coalescence event in the sample. In this framework, the hidden states of the model are a combination of divergence times, taken from a discretized distribution, and identity of the corresponding haplotypes involved. The rationale for such simplification was that the PSMC showed poor resolution in the recent past [14], and considering more genomes would bring additional signal. The drawback of this implementation is that the more genomes are considered, the more "shifted" toward the present is the timeframe where population parameters can be inferred. As a result, the authors reported that with more than 8 diploid individuals (16 haploid genomes), parameters can virtually not be estimated (see also [44] for an illustration of this effect with simulations). Another consequence of this approach is that the recombination rate parameter cannot be reliably estimated [15]. The MSMC was used to infer the recent history of human population. In particular, the authors introduced the possibility to label individuals and look at cross-coalescence rate between groups, a way to get a fine-tuned view of population divergence [15, 45].
5.1.2 The Demographic Inference with Composite Approximate Likelihood (diCal)
An alternative approach was introduced by Song and colleagues [16–18]. The demographic inference with composite approximate likelihood (diCal) approach is based on the conditional sampling distribution, which computes the likelihood of one genome conditioned on the observation of others. Using the so-called composite likelihood formula, it is therefore possible to compute the likelihood of the data for n genomes as the product of the likelihood of one genome given the <sup>n</sup> - 1 other ones and the likelihood the remaining <sup>n</sup> -1 genomes:
$$P(D\_{1\ldots n}|\Theta) = \Pr(D\_1|D\_{2\ldots n}, \Theta) \times P(D\_{2\ldots n|\Theta}),$$
where Θ is the set of model parameters and D1...<sup>n</sup> denotes the data set with n genomes. By further noting that
$$P(D\_{2\ldots n}|\Theta) = P(D\_2|D\_{3\ldots n}, \Theta) \times P(D\_{3\ldots n-1}|\Theta)$$
the likelihood of the full data set can be computed by recursion. The terms P(Di|Di+1...n) form the conditional sampling distribution (CSD). Paul et al. [16] proposed a way to compute the CSD at the cost of introducing several additional hypotheses: (a) the haplotypes upon which the sample is conditioned are considered independent, that is, no coalescence events involving these haplotypes are allowed and (b) mutations can only occur once in any lineage (infinite site hypothesis). The likelihood resulting from this approximated CSD is therefore not exact. This approach was introduced by Li and Stephens [46] and is referred to as the product of approximate conditionals (PAC) model. Under the PAC model, the likelihood depends on the order by which the data is conditioned, which can be circumvented with permutation procedures. While the CSD-based SMC does not have the same drawbacks as the MSMC of Schiffels and Durbin [15], its computational efficiency decreases as the number of haplotypes considered increases and becomes impractical for more than 10 genomes [19]. An elegant feature of the diCal approach is that it can be extended to more complex demographic models, including population structure and gene flow [18, 45]. Such extension is of interest as the SMC approximation has been shown to be sensitive to strong population structure [47].
5.1.3 Extending the SMC with Conditional Site Frequency Spectra (CSFS) In order to use the large amount of data available in "1000 genomes" projects, Terhorst et al. [19] extended the PSMC in a different direction. Instead of modeling the genealogy of the complete sample, the authors proposed to model the divergence of two haplotypes (the PSMC model) as hidden states, yet considering the full set of genomes as observed states. In this approach, the transition probabilities of the coalescent HMM are similar to the PSMC (or to be more precise, similar to the MSMC with two haplotypes, as the original PSMC uses the SMC of McVean and Cardin [33] and not the SMC' of Marjoram and Wall [48]), but the emission probabilities are extended to account for the full site frequency spectrum of hundreds of genomes. This conditional site frequency spectrum (CSFS) is computed using coalescence theory, offering a generalization of the Poisson random field (PRF) model introduced by Sawyer and Hartl [49]. Just like the original PRF, however, the CSFS ignores linkage of observed states, only linkage between the two conditioned haplotypes is modeled via the SMC. Additional data reduction steps are therefore required to ensure that the independence condition of sampled sites is met.
While the ARG contains all historical information about a sample of genomes, genomes themselves contain very little information regarding the underlying ARG. As a result, in most statistical inference methods is the ARG treated as a variable accounted for, but not directly inferred. In the SMC models presented above, this is taken care of by the hidden Markov methodology, which computes a likelihood for a given sample by summing over all possible ARG (via the so-called forward algorithm). The Viterbi algorithm and the posterior decoding procedure are HMM algorithms that allow to reconstruct a posteriori the most likely ARG for a sample, such procedures are notably used for the inference of patterns of incomplete lineage sorting along genomes [11, 12, 50, 51]. Yet the variance in such estimation is typically very large [12].
Rasmussen et al. [20] proposed a different approach: they developed a Bayesian sampler of ARGs conditioned on a set of genome sequences. Similar in principle to the PAC and CSD approaches, the authors proposed to generate the ARG of <sup>n</sup> genomes conditioned on the ARG of <sup>n</sup> - 1 genomes, a procedure they refer to as threading. The generated ARGs can then be used to infer evolutionary processes of interest. Palacios et al. [52] developed a non-parametric method that allows to estimate the variation in time of the effective population size based on such reconstructed ARG. Rasmussen et al. further showed that while the model used for inference is purely neutral, the a posteriori inferred ARG contains signature of selection, visible for instance as a decrease of the time of the most common ancestor of two samples in the data close to coding sequences. Such approaches offer promising avenues for the development of new statistical methods to detect genomic regions with unusual history.
5.2 The Case of Multiple Species Hobolth et al. [11] developed a hidden Markov model (HMM) to infer the ancestral recombination graph between three closely related species. Because this model only contains one haploid genome per species, it only allows to infer population parameters in the ancestral species. Dutheil et al. [12] reparametrized this model in the context of the sequentially Markov coalescent. In contrast to the previous approaches, only four hidden states were
5.1.4 Explicit Reconstruction of the Ancestral Recombination Graph
Fig. 13 The coalescent process along genomes of three closely related species. (a) Four archetypes of coalescence scenarios with three species, exemplified with human, chimpanzee, and gorilla. In the first scenario, human and chimpanzee coalesce within the human–chimpanzee common ancestor. In the three other scenarios, all sequences coalesce within the common ancestor of all species, with probability 1/3 depending on which two sequences coalesce first. (b) Example of genealogical changes along a piece of an alignment. The alignment was simulated using the true coalescent process and parameters corresponding to the human–chimpanzee–orangutan history. The blue line depicts the variation along the genome of the human–chimpanzee divergence. The background colors depict the change in topology, red and yellow corresponding to incomplete lineage sorting. Each change in color or break of the blue line is the result of a recombination event
considered, corresponding to four alternative scenarios of lineage segregation (Fig. 13). In states 1 and 2, the genealogy is consistent with the phylogeny and lineages segregate in the same order as the species. In states 2, 3, and 4, allele divergence predates the first speciation event and ancestral polymorphism persists between the two speciation events, leading to incomplete lineage sorting. The scenarios depicted by states 2, 3 and 4 are equally likely, and in the case of states 3 and 4, the resulting topology is inconsistent with the phylogenetic tree. This model therefore does not rely directly on divergence variation along the genome alignment but uses patterns of topology variation instead to compute the speciation times and ancestral population sizes.
Using this approach, Hobolth et al. estimated a speciation time between human and chimpanzee around 4.1 My and a large ancestral effective population size of 60,000 for the human–chimpanzee ancestor. Dutheil et al. [12] found similar estimates with the same data set while accounting for substitution rate variation across sites and estimated an average recombination rate of 1.7 cM/Mb. With sequencing of more great ape genomes, this approach allowed to estimate population size in several ape ancestors ( [27, 50, 53], reviewed in [54]). As ILS is a proxy for ancestral effective population size, a major result of these studies is that the distribution of ILS is not uniform along the genome. For instance, it is reduced in proximity of genes, a pattern that can be explained by background selection [27, 50]. Large regions of the X chromosome were also found to be devoid of ILS, a pattern resulting from recurrent selective sweeps along the chromosomes [55].
#### 6 Specific Issues Faced When Dealing with Genomic Data
In previous sections we discussed population genetic models and methods for parameter estimation. We now describe several challenges encountered when analyzing whole-genome data sets, at the intra- and interspecific levels.
6.1 Sequencing Errors and Rate Variation Sequencing errors are a well-described source of bias in population genetics analyses, resulting in an excess of singletons [56]. At both the intra- and interspecific/populational level, such error therefore leads to incorrect estimates of local divergence, in particular for recent times. When more divergent sequences are compared, for instance, from distinct species, the issue becomes more complex as the error rate differs between and within sequences due to coverage variation, but also properties of the genome (base composition, repeated elements, etc.). Such errors result in a departure from the molecular clock hypothesis, thus potentially leading to biases in parameter estimates, such as asymmetries in genealogy frequencies [57, 58]. In this respect, data preprocessing becomes a crucial step in any genomic analysis. Methods would also benefit in many cases of inclusion of a proper modeling of such errors. Burgess and Yang noticed that sequencing errors can be seen as a contemporary acceleration in external branches, resulting in an extra branch length [9]. Such an extra length can be easily accommodated in many models. It has to be noted that only a differential in error rates between lineages results in a departure from molecular clock, and in such approaches, one still has to consider that at least one sequence is error-free. In addition, as noted by the authors, assuming a constant error rate over all genomic positions may also turn out to be inappropriate, and better models should allow this rate to vary across the sequence. Such approaches still have to be explored. Moreover, sequencing errors are not distinguishable from lineagespecific acceleration (or deceleration in another species). In that respect, sequence quality scores can be a valuable source of information. They are currently used to preprocess the data by removing doubtful regions, but can ultimately be used in the modeling framework.
The substitution rate also varies along the genome, which potentially affects the reconstruction of sequence genealogy, a phenomenon well known by phylogeneticists. In such case the tools developed for phylogenetic analysis can be applied with a reasonable cost. This generally consists in assuming a prior distribution of the site-specific rate and integrating the likelihood over all possible rates [8, 9, 12]. Alternatively, one can also use one or more outgroup sequences to calibrate the rate, as in [6, 7].
6.2 Diploid Data and Phasing While sequencing of diploid individuals allows to infer the two alleles present at heterozygous positions, establishing how these alleles are combined on each homologous chromosome requires an additional, error-prone step calling phasing. Analyses based on the comparison of individuals from distinct species do not require such information, as the coalescence time of two alleles from the same species is expected to have happened much after the speciation time of the compared species. In such case alleles at each heterozygous position can be sampled randomly [13] in order to build a composite haploid genome. The same rationale applies with respect to the use of the human reference genome, a composite genome obtained from multiple individuals. Conversely, inferences at the population level typically rely on the modeling of haploid genomes and therefore require phased data. A notable exception is the PSMC [14], as well as its extension SMC++ [19], which, when applied to one diploid individual, only requires the knowledge of the position of heterozygous positions. 6.3 Structural Genome data are intrinsically fragmented, firstly because of chro-
Variation and Genome Alignment
mosomal organization, but also because of rearrangements that prevent molecule-to-molecule alignment from one species to another. A genome data set is therefore a set of distinct alignments, one per synteny block. Synteny information can only be extracted when individual genomes are available, which is typically not the case for most "re-sequencing" data sets. At the population level, however, such large-scale variation is considered negligible (but see, for instance, [59] for an exception), while it becomes more prominent when genomes from distinct species are compared. In such cases, a genome alignment is constructed with potential errors ultimately leading to the comparison of nonhomologous regions. So far, the only way to deal with such errors is to restrict the analysis on regions where orthology can be unambiguous resolved, mostly by removing short synteny blocks and regions that contain a high proportion of repeated elements, gaps, and duplications.
#### 7 Discussion
Studying the speciation process with genome data implies new modeling challenges, as the basic configuration of a population genetics data set is drastically changed: instead of having a few loci sequenced in several individuals, we have an (almost) exhaustive set of loci sequenced in several individuals for multiple closely related species. The change involves the spatial dimension, but also time, as the process under study occurred much further back in time than the ones that are commonly studied with a "standard" population genetics data set. The use of the spatial signal has a major consequence, namely, that recombination has to be taken into account, even if it is not directly modeled.
Apart from these considerations, ancestral population genomics, as population genetics, heavily relies on the study of sequence genealogy, its shape, but also its variation. The underlying models build on existing intraspecies population modeling, as they only need to add the species divergence process, that is, a moment in time where two populations stop exchanging genetic material and evolve fully independently. The simplest isolation model assumes that the speciation is instantaneous, while the isolation-with-migration model assumes that the two neo-species can still exchange some material, at least for a certain time after the split. Such a model is not different from a pure isolation model where the ancestral population is structured into two subpopulations: in the first case the speciation time is defined as the time of the split, while in the second case it is the time of the last genetic exchange. Recent work on primates [10] suggests that the speciation of human and chimpanzee was not instantaneous. If the average divergence of the human and chimpanzee is a bit more than 6 My (using widely accepted mutation rate), then the split of the two species initiated around 5.5 My ago, and the last genetic exchange can be dated around 4 My.
The fact that we sample a large number of positions in the genome thus appears to have the power to counterbalance the reduced sampling of individuals within population, allowing the estimation of demographic parameters in the ancestor. Nonetheless, complexity limits are rapidly reached, when considering, for example, three closely related species that can exchange migrants. More complex demographic scenarios, incorporating, for instance, variation in population sizes, will also add additional parameters that might not all be identifiable.
If the ancient speciation processes have left signatures in the contemporary genomes, we do not know yet how far back in time this is true. Intuitively, the signal is maximal when the variation in divergence due to polymorphism is large enough compared to the total divergence. The divergence due to polymorphism is proportional to the ancestral population size, while the divergence of species is only dependent on the time when it happened. So the further back in time we are looking at, the bigger the population sizes need to be so that the ancient polymorphism leaves a signature in the total divergence time. In addition to this, one has to take into consideration sequence saturation due to the too large number of substitutions that accumulated since ancient splits, and the fact that demographic scenarios complexity increases with time. For instance, when considering the evolution of a species over several millions of generations, the probability that a bottleneck, resetting the signal from past events, occurred once is not negligible.
We are in the population genomics era. Data sets are available that allow us to understand the evolutionary processes that are associated with the formation and evolution of species. Analyzing such data sets with the current methodologies however offers major challenges: (1) developing the appropriate computational tools able to handle such data sets with current machines (both in terms of processor speed and memory usage) and (2) design realistic models with enough complexity to capture the most important historical events while remaining computationally tractable.
#### 8 Exercises
8.2 Estimating Ancestral Population Size from the Observed Amount of ILS Given that 30% of incomplete lineage sorting is observed between human, chimpanzee, and gorilla and assuming a generation time of 20 years and a that 2.5 My separate the splits between human/ chimpanzee and human—chimpanzee/gorilla, what is the effective ancestral population size compatible with this observed amount? Using Burgess and Yang's method [9], a researcher finds a higher estimate of Ne than expected. What could explain this discrepancy?
8.3 Number of Migration Rates in the General k-Population IM Model
In this exercise we show that a k-population IM model has 2(<sup>k</sup> -1)2 migration rates.
1. Starting at the bottom of the k-population IM model argue that the number of migration rates at the level of k populations is <sup>k</sup>(<sup>k</sup> -1).
#### Acknowledgements
We thank an anonymous reviewer for constructive comments and detailed suggestions on the manuscript.
#### References
Balanovska E, Fedorova S, Litvinov S, Malyarchuk B, Derenko M, Mosher MJ, Archer D, Cybulski J, Petzelt B, Mitchell J, Worl R, Norman PJ, Parham P, Kemp BM, Kivisild T, Tyler-Smith C, Sandhu MS, Crawford M, Villems R, Smith DG, Waters MR, Goebel T, Johnson JR, Malhi RS, Jakobsson M, Meltzer DJ, Manica A, Durbin R, Bustamante CD, Song YS, Nielsen R, Willerslev E (2015) Population genetics. Genomic evidence for the Pleistocene and recent population history of native Americans. Science 349(6250):aab3884
on the X chromosome in the humanchimpanzee ancestor explain its low divergence. PLoS Genet 11(8):e1005451
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Chapter 19
#### Introduction to the Analysis of Environmental Sequences: Metagenomics with MEGAN
#### Caner Bag˘ cı, Sina Beier, Anna Go´rska, and Daniel H. Huson
#### Abstract
Metagenomics has become a part of the standard toolkit for scientists interested in studying microbes in the environment. Compared to 16S rDNA sequencing, which allows coarse taxonomic profiling of samples, shotgun metagenomic sequencing provides a more detailed analysis of the taxonomic and functional content of samples. Long read technologies, such as developed by Pacific Biosciences or Oxford Nanopore, produce much longer stretches of informative sequence, greatly simplifying the difficult and timeconsuming process of metagenomic assembly. MEGAN6 provides a wide range of analysis and visualization methods for the analysis of short and long read metagenomic data. A simple and efficient analysis pipeline for metagenomic analysis consists of the DIAMOND alignment tool on short reads, or the LAST alignment tool on long reads, followed by MEGAN. This approach performs taxonomic and functional abundance analysis, supports comparative analysis of large-scale experiments, and allows one to involve experimental metadata in the analysis.
Key words Metagenomics, Software, MEGAN, Taxonomic analysis, Functional analysis, Long reads
#### 1 Introduction
Metagenomics is the study of microbiome samples, such as obtained from ocean water, soil, plant matter, or feces, say, using high-throughput DNA sequencing [1]. Metagenomic sequencing allows the study of microorganisms found in environmental samples without relying on culturing methods or prior knowledge of the composition of the community. With metagenomics, one can determine the taxonomic and functional content of samples.
While most metagenomic projects to date have used short read sequencing (next-generation sequencing), there is increasing interest in using long read sequencing technologies in this area. Long read technologies have been considered too expensive, difficult, or error-prone for application in metagenomics. However, this is changing and computational analysis methods designed for processing short reads now need to be modified to work well on long
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_19, © The Author(s) 2019
reads, so as to make good use of the ability of long reads to cover multiple genes.
A major computational challenge in metagenomics is the alignment of sequencing reads against a comprehensive reference database. Billions of reads can be aligned against a large protein reference database in reasonable time using high-throughput alignment tools such as DIAMOND [2]. Long reads require frame-shift aware alignment tools, such as LAST [3, 4], because insertions or deletions due to sequencing errors impact long reads, as discussed in Subheading 2.
In the following, we will first discuss how to perform basic alignment and analysis of short reads in Subheading 2.1 and long reads in Subheading 2.2. We will then show, in Subheading 3, how to compare large numbers of samples in MEGAN6 [5] and perform basic statistical analysis of the samples and their metadata. In Subheading 4 we briefly discuss the challenges we will have to face to further improve the analysis of data from environmental samples. Finally, in Subheading 4.1 we describe some additional resources available for using MEGAN 6.
#### 2 Workflows for Metagenomic Analysis with MEGAN
The basic workflow for using MEGAN consists of two main steps: read alignment against a reference database and then import an analysis of the alignments in MEGAN. The aim of pipeline is to perform taxonomic and functional binning of the input reads.
The alignment can be performed using a number of different tools depending on the type of sequencing data and on the chosen database, its sequence type, size, and available computer power. For smaller databases more sensitive tools can be chosen such as MALT [6] or even BLAST [7]. These tools generally offer higher sensitivity at the cost of a longer runtime. For large datasets and databases, it is more suitable to choose an alignment tool such as DIAMOND or LAST. We use the NCBI NR database [8] with both of the latter tools, because it is the largest and most comprehensive protein database available today. NCBI NR contains 144.5 million protein sequences (August 2017).
Before running the pipeline, one can optionally perform preprocessing, that is, quality control, trimming, and filtering, of the raw reads. However, these steps usually have little impact on the results of the alignment-based analysis described in this document.
Fig. 1 Basic pipeline for short read analysis
#### 2.1.1 Read Alignment with DIAMOND
DIAMOND uses double indexed alignment, which means both the reference database and the query are indexed for comparison. This leads to a large speedup especially for large queries and databases. Like BLASTX, DIAMOND uses the "seed and extend" method to find all matches between a query and the database. To further increase speed, DIAMOND utilizes spaced seeds, which are long seeds where only some positions are used for matching the seed. This leads to another increase of speed without decreasing sensitivity.
DIAMOND can be run either in fast or sensitive mode. Fast mode will run around 20,000 times faster than BLASTX on short reads and will be able to find 75–90% of all relevant matches that one would find with BLASTX, while sensitive mode provides a speedup of 2500 while recovering up to 94% of significant matches.
#### 2.1.2 Taxonomic and Functional Classification with MEGAN6 DIAMOND can save alignments in a compressed format called DAA (DIAMOND alignment archive) format. DAA files can be imported into MEGAN6 in multiple ways. A small number of small DAA files can easily be imported interactively using menu items provided in MEGAN. For larger datasets and or many files, one should use the command-line tools provided with MEGAN. These include daa2rma, which will generate a RMA file as used by MEGAN from one or two (for paired reads) DIAMOND files and daa-meganizer, which analyzes a DAA file and then appends the result to the end of the file. Such "meganized" DAA files can then be opened directly in MEGAN. The latter approach is much faster and is more space efficient. However, to use paired reads all alignments have to be in the same file.
One can use the program blast2rma to process the output of a range of different alignment programs, such as BLAST.
During the processing of alignments for MEGAN, the reads will be assigned to nodes in the NCBI taxonomy and any functional classifications that have been configured in the import dialog or on the command-line. Taxonomic binning of each read is done separately, by assigning it to the lowest common ancestor (LCA) of its significant matches. Matches can be filtered by multiple parameters, for example, e-value and bit-score, as well as sequence identity. Only matches passing those filters will be used to determine the LCA. It is also important to choose the minimum support (or minimum support percentage), the number or percentage of reads that must be assigned to a single taxon before it will be part of the final result. Reads assigned to a taxon that does not pass the minimum support filter will be pushed up the taxonomy until a taxon is found that passes the filter.
Functional binning is performed by mapping the NCBI database accessions for the matches of a read to identifiers of the selected functional classification. Mapping files are currently available for InterPro2GO [9, 10] (InterPro families embedded in a GO-based hierarchy), eggNOG [11], KEGG [12], and SEED [13].
2.1.3 Investigation of the Results The resulting files can be opened and interactively investigated using the MEGAN6 graphical user interface. The first view when opening a file is always a hierarchical representation of the taxonomic composition of the sample. Selecting different nodes of this tree, the user can uncover further information on the reads mapped to the represented taxon. Selecting Inspect Reads on a node will open the Inspector Window, which displays the reads assigned to that node, as well as their alignments. This functionality can be used both in the Taxonomy Viewer, where nodes represent taxa, and in any of the Functional Viewers. Figure 2a shows an example of the Inspector Window.
> Instead of just viewing a listing of the matches and alignments, it is also possible to select Show Alignments. This will open the Alignment Viewer (Fig. 2b), where for each of the database references with matches from the reads assigned to the selected node it is
Fig. 2 (a) The Inspector Viewer showing some reads that have been assigned to Alistipes ihumii. (b) The Alignment Viewer showing reads aligned to a reference sequence
Fig. 3 (a) Bar chart of taxonomic assignments on family level, sorted by abundance. (b) Radial chart of functional assignments to KEGG for the same sample from [14]
possible to show the alignment of all of those reads on the reference. This can be useful, say, to determine how much of a reference gene is covered by reads.
Apart from being able to investigating taxonomic diversity, the advantage of using metagenomic sequencing to study an environmental sample is the ability to study the functional potential of the community. MEGAN currently provides four different functional classification systems for this purpose: InterPro & GO, eggNOG, KEGG, and SEED.
Each functional classification is displayed as a tree. The nodes of the tree can be investigated very much like the nodes of the taxonomic tree. Abundances can be visualized using different visualization options from simple bar charts over box plots and heat maps to radial tree charts drawn based on the abundances of the selected nodes. Two examples show charts that are shown in Fig. 3.
Alignments or reads matching a selected function can be exported to a text file or extracted to a new MEGAN document. This makes it possible to study only a part of a microbial community that is of particular interest. For example, if you select nodes associated with antibiotic resistance genes, you can determine which taxonomic assignment the reads assigned to antibiotic resistance genes have. An example of this is shown in Fig. 4.
If you want to study the full gene sequence of proteins found in your samples and be able to compare variants of those genes, it can be helpful to use gene-centric assembly [15]. Gene-centric assembly uses the alignments to reference proteins to assemble the matching reads. One can thus obtain the gene sequences from different organisms found in a sample for further analysis steps.
We will introduce more possibilities for studying the taxonomic and functional diversity of multiple samples in comparison in Subheading 3.
Fig. 4 Taxonomic assignment of reads from the day 0 sample for "Alice" from the ASARI [14] dataset which have been assigned to "resistance of fluoroquinolones" in the SEED hierarchy
#### 2.2 Long Read Pipeline As presented in the previous section, using metagenomic short reads, one can assembly gene sequences and obtain variants of a single gene using a gene-centric assembly, or of course use other assembly techniques. However, using short read data, it is very difficult to establish whether different genes are present in the same organism. We can connect the genes if they are found on a single DNA molecule with long sequencing reads, provided by third generation sequencing technologies such as PacBio [16] or Oxford Nanopore [17]. The PacBio and Nanopore devices can produce reads that are hundreds of thousands of bases long, with error rates of around 10%, say [17]. In contrast to short reads, which each can be safely assumed to overlap with only a single gene, long read will usually overlap or contain multiple genes. Hence, many popular short read alignment and analysis algorithms may require modification so as to take into account that a given read can align to multiple genes.
As described in the following, for long reads alignment is performed using LAST, processing of the alignments requires an additional step and MEGAN provides some modified algorithms for processing and visualizing long reads.
Fig. 5 Basic pipeline for long read analysis
```
Score = 86 bits (159), Expect = 7e-13
Identities = 34/37 (92%), Positives = 34/37 (92%), Gaps = 2/37 (5%)
Frame = -1
Query: 1080 EAVMVLSLDAEA\LVGYRE/KFPAWMDADRFEIKPRK 976
EAVMVLSLDAEA LV YRE KFPAWMDADRFEIKPRK
Sbjct: 232 EAVMVLSLDAEA-LVRYRE-KFPAWMDADRFEIKPRK 266
```
Alternatively, the user can use DAA\_Converter (available at http://github.com/BenjaminAlbrecht84/DAA\_Converter), which converts a given MAF file to a DAA file. This has several advantages, including space compression and faster processing. Additionally, the output of LAST can directly be piped into DAA\_Converter which will then convert the output into a DAA file as LAST continues to operate. The trade-off when using DAA\_Converter currently is that the alignments are filtered out with the default settings in MEGAN6 and resulting DAA file only has the alignments that would pass the filter, making it impossible to change filtration parameters without running LAST again once the conversion is done.
Similar to short reads, these long read MAF and DAA can then be imported into MEGAN and each read will get assigned to a taxon and/or functional class(es) of any provided functional hierarchy. The filtration based on bit-score of alignments work differently for long reads. In case of short reads, the alignments are filtered globally—only those that are within top 10% (by default) of the best-scoring alignment are taken into account. For long reads, this filtration is applied to each "gene" separately, as one long read can contain many different genes along its length. The alignments that overlap significantly (>90% by default) are grouped into segments, denoting different genes, and each interval is then processed individually in the filtering step.
The LCA algorithm to assign reads to taxonomic classes is also modified for long reads. As there are multiple genes on a single long read, and each of them may be conserved in different clades of the taxonomic tree, the naı¨ve LCA is usually uninformative. Instead long reads are assigned to the most specific taxon that covers more than a fixed percentage (>80% by default) of every base pair that has an alignment. This algorithm assigns reads specifically to lower levels of taxonomy as long as they cover a gene which has low level conservation, other taxa gets lower percentages of coverage. Functional classification of long reads does not necessarily assign each read into one functional class, instead reads are assigned to the functional class of best-scoring alignment in each segment, thus each segment is assigned to one function and one read can be assigned to multiple different functional classes.
2.2.4 Investigation of the Results The first view the user gets when a long read dataset is loaded in to MEGAN6 is identical to that of a short read dataset; however, there are some underlying differences and several investigation modes designed specifically for long reads.
> Due to a large variability of read length of long reads [18], it is impractical for MEGAN to report number of reads assigned to class as a mean of abundance. Using the raw read length is also not feasible for Nanopore technology as reads tend to have "head"
Fig. 7 Long Read Inspector in MEGAN6. The read is drawn as a line in the middle and the protein alignments are drawn as arrows on their corresponding positions and strands on the read
and "tail" regions composed of random bases [19] (Fig. 7 shows a read whose tail region has no significant alignment to any protein in the database). Thus, the default mean of reporting the abundance for a particular taxon or functional class in long read pipeline is the number of aligned bases.
The number of alignments on a long read can easily exceed hundreds and complicates the Alignment Viewer and the Inspector features of MEGAN6. In order to simplify the investigation of alignments on the reads, MEGAN6 offers a Long Read Inspector window (Fig. 7), accessible via right-click on any of the nodes in the main view. This inspector draws reads as horizontal lines and alignments as arrows on their corresponding positions. The names of taxa or functional classes are also linked to these alignment arrows.
The Inspector Window helps particularly in the case of suspicious assignments. Figure 8a shows the inspector view for a read that was assigned to Trichuris trichiura, a human parasitic whipworm, in a sample of known mixture of microorganisms [20]. A closer inspection to Fig. 8a lets us see that, although the read is spanned by several alignments from Escherichia coli, it is assigned to T. trichiura because the total length of alignments to T. trichiura is longer than 80% whereas it is below that for E. coli and all other competing taxa.
For further analysis of such suspicious assignments, MEGAN6 offers a remote BLAST function, in which selected reads are aligned against a selected database (such as the nucleotide collection— NCBI nt) on the NCBI website and the resulting assignments are captured, processed, and presented in a new MEGAN document. In Fig. 8b, we see that our "suspicious" read is assigned to E. coli, which was in the known mixture of microorganisms, based on remote NCBI-BLAST against NCBI nt.
Similar to exporting alignments and reads as explained in the previous section, these can also be exported in general feature format (GFF) for downstream analysis. This provides a simple way of obtaining the annotation, especially for long reads and contigs. The annotations exported to the GFF files contain the accessions of
Fig. 8 MEGAN6 offers a remote BLAST functionality, namely "BLAST on NCBI," which can be used for suspicious assignments. (a) Long Read Inspector view for a read assigned to Trichuris trichiura, based on protein alignments against NCBI nr. (b) Long Read Inspector view for the same read as in (a), assigned to Escherichia coli, after searching it against nucleotide collection of NCBI using the remote BLAST functionality of MEGAN6
> references and their corresponding taxonomic and/or functional mappings depending on which mapping files were used during importing the dataset into MEGAN.
#### 3 Comparison of Multiple Samples
Most modern metagenomics experiments include the collection and analysis of multiple samples to compare different groups with controls or study the dynamic changes of a microbial community over time. Hence, a very important feature of MEGAN is the ability to load multiple datasets into a single "comparison document" (megan file). This is a light-weight file that does not contain the original reads and alignments, but allows one to compare the taxonomic and functional diversity of multiple samples.
To be able to easily compare groups of samples and relate findings to features attached to samples, it is helpful to import metadata. Metadata should be provided in tabular format and connect the sample IDs to attributes whose values can be text, numeric, or boolean values. Using this information you can group samples in different visualizations. For example, this allows easier interpretation of the principal component analysis (PCoA) plots in MEGAN. Principal components can be calculated using different distance measures including Bray–Curtis or simple Euclidean distances. MEGAN can include bi-plots and tri-plot vectors into the PCoA plot, which represent the top taxonomic or functional classes and metadata features, respectively, that correlate most with the differences between samples. Figure 9 shows multiple examples of PCoA plots including bi-plot and tri-plot vectors.
MEGAN can also calculate and visualize co-occurrence and correlation plots. For correlation there are two options. The first is useful for time series analysis, because it calculates correlations between different taxa. This can be used to determine how changes in abundance of one taxon influence changes in another, which makes it possible to detect potential interactions between taxa. To distinguish the effect of interactions between taxa from it being caused by an external influence, it is useful to check out the other attribute correlation plot, which calculates correlations between taxa and metadata. So, if, for example, two taxa are correlated to each other and correlated to the same external influence from the metadata, then they might be less likely to be influencing each other, but are perhaps both influenced by the same attribute of the metadata. An example of an attribute correlation plot is shown in Fig. 10.
Fig. 9 PCoA analysis of 12 samples associated with "Alice" (round shapes) and "Bob" (square shapes), from [14]. Time points of antibiotic intake are colored light blue, time points before and after antibiotic intake dark red. (a) A PCoA plot based on Bray–Curtis distances as calculated by MEGAN using the taxonomic abundances for the samples. The green vectors represent the bi-plot vectors. The samples are grouped by individual, showing the convex hulls of the groups as well as ellipses. (b) is based on the same data but using the abundances of GO terms in the InterPro2GO hierarchy and only showing the convex hulls of the group. Here the orange vectors are the tri-plot vectors, showing the relation of metadata values to the principal components
Fig. 10 Attribute correlation plot for the data from [14] for two healthy individuals taking antibiotics for 6 days (day 1–6). Correlation is shown as a heat map with red marking positive correlation between the attribute and the taxon and blue marking negative correlation. Correlations are shown for antibiotics intake (boolean) and time (day 0, 1, 3, 6, 8, and 34)
#### 4 Outlook
It goes without saying that the quality and quantity of the input sequencing data limits the reliability of the output analysis. More directly, quality of the MEGAN hierarchy assignments is determined by the quality of the read alignment, which, in turn, depends on the chosen database and alignment tool. On the one hand, the database needs to be well annotated and comprehensive, as it is only possible to analyze the organisms or entities present in it. On the other hand, the alignment tool needs to be sensitive in order to identify the matching sequence. It is especially difficult to deal with sets of very similar sequences. Currently, for the human gut microbiome sequencing data analyzed with the basic short read pipeline, as much as 30% of reads are not assigned to any node in the course of the taxonomic analysis.
In order to avoid the bias introduced by the database one can also use one of the database-free strategies, e.g., k-mer counting. They are good for tracking the global changes in the data, but it is difficult to correct for possible contaminations. Although MEGAN does not support this type of analysis, it enables global comparisons with PCoA based on the profiles computed for each of the samples.
Another approach is assembly based analysis. In brief, the reads are assembled and then the scaffolds or contigs are annotated and investigated. This approach provides some information on gene co-localization at a cost of data loss in the form of unassembled reads and short contigs. Full metagenomic read assembly [21] is a very complex and computationally expensive task that MEGAN does not address.
Application of the long read sequencing technologies opens new perspective for metagenomics analysis. Long reads provide information on gene co-location on a single DNA molecule, and make assembly much easier. But, long reads also pose new algorithmic challenges in aspects of the protein alignment, hierarchy assignment, and abundance computation. As long read technologies continue to evolve, so, too, must the corresponding analysis algorithms.
MEGAN is a powerful visual analytics tool that provides a wide range of the algorithms for analysis of metagenomics sequencing data. MEGAN can run on hundreds of samples along with hundreds of metadata columns. It is the main workhorse of the Tubiom project where metagenomics profiles of 10,000 volunteers are collected and mined for correlations with the vast metadata (www.tuebiom.de).
MEGAN Community software is freely available on the website: ab. inf.uni-tuebingen.de/data/software/megan6, together with the current mapping files for taxonomic and functional analysis.
Short read datasets presented in this chapter and used for visualizations are publicly accessible in MEGAN via MeganServer. The dataset used in the Long Read Pipeline section was downloaded from the supplementary material of Brown et al. [20]. Instructions for use of MEGAN and user support can be found on the MEGAN community website (megan.informatik. uni-tuebingen.de).
#### References
4.1 MEGAN Resources
determination through sequence-based metagenomics. Antimicrob Agents Chemother 59 (12):7335–7345
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Chapter 20
#### Multiple Data Analyses and Statistical Approaches for Analyzing Data from Metagenomic Studies and Clinical Trials
#### Suparna Mitra
#### Abstract
Metagenomics, also known as environmental genomics, is the study of the genomic content of a sample of organisms (microbes) obtained from a common habitat. Metagenomics and other "omics" disciplines have captured the attention of researchers for several decades. The effect of microbes in our body is a relevant concern for health studies. There are plenty of studies using metagenomics which examine microorganisms that inhabit niches in the human body, sometimes causing disease, and are often correlated with multiple treatment conditions. No matter from which environment it comes, the analyses are often aimed at determining either the presence or absence of specific species of interest in a given metagenome or comparing the biological diversity and the functional activity of a wider range of microorganisms within their communities. The importance increases for comparison within different environments such as multiple patients with different conditions, multiple drugs, and multiple time points of same treatment or same patient. Thus, no matter how many hypotheses we have, we need a good understanding of genomics, bioinformatics, and statistics to work together to analyze and interpret these datasets in a meaningful way. This chapter provides an overview of different data analyses and statistical approaches (with example scenarios) to analyze metagenomics samples from different medical projects or clinical trials.
Key words Metagenomics, Metatranscriptomics, Microbiome, Clinical trials, Comparative metagenomics
#### 1 Introduction
The diversity of species on earth is high, and most of them are microorganisms. Their ubiquitous presence makes it extremely difficult to identify and classify all microbes in a laboratory environment. Standard genomics tries to enrich pure cultures and study them: for example, the taxonomy, the genome, the genes, and the pathways. However, only a miniscule fraction of all microbes can be cultured because of their complex symbiosis and nutrient requirements in other organisms. The scientific community is now equipped with the development of new sequencing techniques
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_20, © The Author(s) 2019
and high-throughput analysis. The study of the genomic content of a sample of microorganisms obtained from a common habitat is made possible with the field of metagenomics, also known as environmental genomics [1]. Instead of taking the DNA for sequencing from isolated cultures it is obtained directly from the environment. Therefore, the analysis of microbes that are deemed unculturable (which means current laboratory culturing techniques are unable to grow them) with standard laboratory techniques becomes possible. Two main approaches commonly used in metagenomic studies: marker gene-based metagenomics (e.g., 16S amplicon sequencing) and metagenomic shotgun sequencing. In the first approach, DNA is used as the template for PCR to amplify a segment of the conserved 16S ribosomal RNA (rRNA) gene sequence. Universal primers complementary to conserved regions are used so that the region can be amplified from any bacteria. After purification of PCR products, sequencing of the 16S rRNA gene is performed [2]. In the second approach, shotgun sequencing, DNA is broken up randomly into multiple small segments, which are sequenced using the chain termination method to obtain reads. Multiple overlapping reads for the target DNA are obtained by performing several rounds of this fragmentation and sequencing. Computer programs then use the overlapping ends of different reads to assemble them into a continuous sequence [3].There are several publications discussing the differences in microbial biodiversity discovery between 16S amplicon and shotgun sequencing, for example see [4]. In a recent study using water samples from Brazil's major river floodplain systems, authors showed shotgun sequencing outdone by amplicon [5]. Here, the authors ascribed the poor performance of shotgun sequencing mainly to the weakness of the database used in the study, as compared to databases for the 16S rRNA gene. This study can be used as a caution for people working with rare environments (See article by Catherine Offord in The Scientist<sup>1</sup> ). Comparisons of the two methods in well-studied systems such as the gut microbiome have generally found that shotgun sequencing identifies more microbial diversity [6].
Further recent advancement of culturomics approach is shedding light on multiple high-throughput culture conditions [7, 8]. As the samples used in metagenomics do not contain the genome of just one but many different microorganisms, the possibility of analyzing their functional and metabolic interplay arises. Next-generation sequencing technology (NGS) has effectively transformed infectious disease research throughout the last decade, fuelling the growth in genetic data and providing huge number of DNA reads at an affordable cost. Many studies use these
<sup>1</sup> https://www.the-scientist.com/?articles.view/articleNo/50044/title/Shotgun-Sequencing-Outdone-by-Amplicon/.
techniques, which examine microorganisms that inhabit niches in the human body, sometimes causing disease, and researchers often try to correlate these microorganisms and their change with multiple treatment conditions (e.g., see [9]). Gene annotations in these studies support the association of specific genes or metabolic pathways with health and with specific diseases. In a recent article authors discussed how host gene–microbial interactions are major determinants for the development of multifactorial chronic disorders and thus for the relationship between genotype and phenotype [10]. There are many other reports based on the application of metagenomics in understanding oral health and disease [11–13]. As recently described by Forbes et al., metagenomics and other "omics" disciplines could provide the solution to a cultureless future in clinical microbiology, food safety, and public health [14].
No matter from which environment it comes, the analysis of datasets from such studies are similar to some extent. Most projects aim at determining either the presence or absence of specific species of interest, or to obtain an overview of the taxa represented in a given metagenome and comparing the biological diversity and the functional activity of a wider range of microorganisms within their communities. The importance increases for comparison of different datasets, as researchers will need to determine and understand the similarities and dissimilarities within the metagenomes of different environments. These environments can be multiple patients with different conditions, multiple drugs, or multiple time points of same treatment or same patient. Further, sometimes researchers also may compare different environments for example to study antibiotic resistance genes (ARG) and understand which environments are more prone to such ARGs. Thus, no matter how many hypotheses we have, we need a good understanding of genomics, bioinformatics, and statistics to work together to analyze and interpret these datasets in a meaningful way.
This chapter provides an overview of different data analyses and statistical approaches to analyze metagenomics samples from a number of clinically derived datasets. The methodological description of this chapter will be guided by three main scenarios. The first one is a published data set from human atherosclerotic plaque samples (Scenario 1) [15]; the second one is a clinical trial example comparing the effects of two omega-3 polyunsaturated fatty acids (PUFAs) supplements on healthy volunteers (Scenario 2) [16]; and the third one is another clinical trial example comparing the efficacy of two drugs for an infectious disease (Scenario 3).
The Scenarios 3 came from an ongoing unpublished project; therefore, the real datasets are not provided. This chapter is mainly focused on multiple data analyses/annotation and statistical approaches that can be used in similar situations, but any biological finding of the example scenarios is not explained here. Although all of these scenarios are derived from medical projects, the analyses approach can be adapted to environmental samples as well. On this occasion, I must emphasize the importance to have good metadata, that is, a detailed description of each parameter like health status or sampling site or age or any similar information relating to specific samples that may be important for the analyses. Good metadata are key to good analyses and noise reduction in data analysis processes.
#### 2 Description of Example Studies
2.1 Scenario 1: Metagenomic Analyses of Human Atherosclerotic Plaque Samples
To investigate microbiome diversity within human atherosclerotic tissue samples high-throughput metagenomic analysis was employed on (1) atherosclerotic plaques obtained from a group of patients who underwent endarterectomy due to recent transient cerebral ischemia or stroke and (2) presumed stabile atherosclerotic plaques obtained from autopsy from a control group of patients who all died from causes not related to cardiovascular disease. Our data provides evidence that suggest a wide range of microbial agents in atherosclerotic plaques, and an intriguing new observation that shows this microbiota displayed differences between symptomatic and asymptomatic plaques, as judged from the taxonomic profiles in these two groups of patients. Additionally, functional annotations reveal significant differences in basic metabolic and disease pathway signatures between these groups.
In this project, we demonstrate the feasibility of novel highresolution techniques aimed at identification and characterization of microbial genomes in human atherosclerotic tissue samples. Our analysis suggests that distinct groups of microbial agents might play different roles during the development of atherosclerotic plaques. These findings may serve as a reference point for future studies in this area of research. The workflow in Fig. 1 provides a brief description of the sample processing and analyses pipeline for the study described in Scenario 1. If readers want to know more details of the methodology, please refer to (15). This scenario is an example of analyzing host-associated metagenome samples.
2.1.1 Methodology Details For this study, we used atherosclerotic tissue samples from a group of 15 patients that underwent elective carotid endarterectomy following repeated transient ischemic attacks or minor strokes (samples from symptomatic atherosclerotic plaques as cases).2 Further, we have asymptomatic atherosclerotic plaques from seven
<sup>2</sup>All methods and experimental manuals were approved by The National Committee on Health Research Ethics (Danish) and was granted by the Ethical Committee of the region of Copenhagen (H-3-2011-013).
Fig. 1 Analysis pipeline for the study of human atherosclerotic plaque samples. Interested readers may refer to the full study here [15]
persons who died from causes not related to atherosclerotic disease (samples from stable plaques as controls).<sup>3</sup>
All 22 arterial plaque samples resulted in 2,610,268,774 shotgun sequencing reads. After mapping these reads against Hg19 using bowtie 2 [17] with "very-sensitive" parameters to filter all human-like sequences from our samples. The average amount of non-Hg19 reads is 884,727,044 (average 33.89% per sample, Table 1). These non-Hg19 reads were extracted and aligned against nonredundant (nr) protein database (version 30.07.2012) [18] using BLASTX (ncbi-blast-2.2.25+; Max e-value 10e3) [19]. After performing the BLASTX alignment, all output files of paired read sequences were imported and analyzed using the paired-end protocol of MEGAN5 [20]. For all non-Hg19 annotated reads, 2–16% (mean 4.6%) were assigned as bacteria in different samples. The rest of reads were assigned to Eukaryota. Table 1 provides details of sequencing read statistics and assignments of reads after different stages of data processing. R statistical
<sup>3</sup>These samples originated from the tissue bank at the Department of Forensic Medicine (Approval No. 1501230).
Table 1 Sample statistics and read assignments programming language [21] was used for multivariate statistics. Later in Subheading 3, we will describe few of the analysis approaches revisiting this study.
In this study our data provided evidence that suggest a wide range of microbial agents (some pathogens) in atherosclerotic plaques, and these microbes displayed differences between symptomatic and asymptomatic plaques as judged from the taxonomic profiles in these two groups of patients. Further, fluorescence in situ hybridization (FISH) was performed to validate the presence of biofilm-like structures of few pathogens (which have been previously predicted from taxonomic analyses) in the symptomatic atherosclerotic plague samples. FISH staining demonstrates the presence of live bacteria; thus, this is a very good approach for cross-validation of any computational finding in the lab.
There are also potentials of using this data for not only taxonomic annotation but also to reveal the functional profiles through partial assembly of specific members and their functional annotations. Functional annotations reveal significant differences in basic metabolic and disease pathway signatures between these groups. Here, we will not provide details of the whole study, but interested readers may refer to [15].
On this occasion, it is necessary to mention that in any similar project in future, for alignment purpose, we would have used DIAMOND [22] which uses improved algorithms and additional heuristics and works much faster compared to available other aligners. Scenario 1 is an example of analyzing shotgun sequence datasets obtained from tissue samples or host-associated metagenome. In case readers have shotgun sequence datasets from environmental samples or from fecal samples, they do not need to perform alignment step to get rid of the host-associated sequences, unless there is any doubt of contamination. Normally we suggest to have control or blank samples in two wells per 96-well plate to address any issue with contaminations.
2.2 Scenario 2: The Effect of Omega-3 Polyunsaturated Fatty Acid Supplements on the Human Intestinal Microbiota
A randomized, open-label, crossover trial of 8 weeks' treatment with 4 g mixed eicosapentaenoic acid (EPA)/docosahexaenoic acid (DHA) in two formulations (soft-gel capsules and drinks) with a 12-week "washout" period [16] is chosen. Healthy volunteers aged greater than 50 years of both genders were included in this study. Participants were randomized to take two types of EPA and DHA compositions (Fig. 2):
- 2. Four soft-gel capsules (each containing 250 mg EPA and 250 mg DHA as the ethyl ester) twice daily with meals (providing 2000 mg EPA and 2000 mg DHA per day), both for 8 weeks.
2.2.1 Study Design
Fig. 2 Schedule of visits for the study to understand the effect of omega-3 polyunsaturated fatty acid supplements on the human intestinal microbiota
After a 12-week "washout" period, participants took the second intervention for 8 weeks. We also included a final study visit after a second 12-week "washout" period (V5; Fig. 2). Fecal samples were collected at five time-points for microbiome analysis by 16S rRNA PCR and Illumina MiSeq sequencing. Parallel red blood cell (RBC) fatty acid analysis was performed by liquid chromatography–tandem mass spectrometry.
analyses were imported in MEGAN for detailed group-specific analyses, annotations, and plots [31]. R statistical programming language [21] was used for multivariate statistics and other plots.
This dataset and method pipeline are purely described as an example for similar analyses; thus, we will not explain the results here, but interested, readers may see [16]. Scenario 2 is a typical example of analyzing 16S sequence data. In Subheading 3, we will describe few of the analysis approaches using data from this study.
In a given situation suppose we need to compare treatment effect of two drugs (e.g., X and Y) or more, where we have time series data, that is, patient samples from multiple time points of the treatment course for both drugs. This time series data can be either collected every day of the treatment period or in intervals. Furthermore, for practical reasons we might not be able to obtain data at a desired day but 1/2 days. It is important to select an error threshold and be consistent with that throughout the project. For example, we need to have a similar depth of sequencing reads or need to follow subsample comparison as detailed later, and, also, we need to discard samples with very low number of reads. Further during alignment to reference database and during mapping to taxonomy similar scores and thresholds should be used for all samples (please check best parameter selections in individual websites while using specific tools). Additionally, there can be multiple fundamental factors in patient samples such as age, gender, and geography that may not contribute in a similar manner to resiliency. Figure 3 shows a schematic of the metadata structure, which may help to understand the complexity of a typical clinical trial.
Fig. 3 Schematic diagram of multiple factors in a clinical study
2.3 Scenario 3: Comparing Effects of Two Drug Treatments for an Infectious Disease
2.3.1 Sample Preparation and Sequencing and Data Analyses
In a clinically relevant setting this type of study wants to know which drug works better for a similar group of patients. Patients are randomized between drug arms to control any selection bias. Usually in this type of projects as we want to compare several factors, we need many samples to start with. Readers are advised to seek statistics help to do power calculation to obtain the preferred sample size. In general, as we end up having hundreds of samples, we usually go for 16S sequencing as a cost-effective solution. However, some projects can also use shotgun sequencing. Similar to previous examples, we assume that we have sequenced (either 16S or shotgun sequencing) our samples and performed further analysis process as outlined earlier to obtain taxonomic profile (following data analyses methods as described in previous scenarios) for each patient at each time point. Besides analyzing time series of each individual separately, we have also grouped them in certain time points such as baseline, mid-treatment, end of treatment, and follow-up. Besides treatment groups, patients are also compared based on multiple factors such as age, gender, and geography.
#### 3 General Methods for Annotation and Statistical Analyses
Broadening our focus beyond these studies, additional analysis techniques are explained below which are used in these studies and also can be used in similar projects.
3.1 Taxonomic and Functional Annotation Taxonomic annotation addresses the question, 'Who is out there?' or in other words tries to obtain information regarding the species composition of a given metagenome. On the other hand, functional annotation attempts to answer the question, 'What are they doing?' There are different approaches for metagenome analyses, among which one type of approach is to use phylogenetic markers to distinguish between different species in a sample. The most widely used marker is the small subunit ribosomal ribonucleic acid (SSU rRNA) gene (16S or 18S) and a second type of method is based on analyzing the nucleotide composition of reads. In a supervised approach the nucleotide composition of a collection of reference genomes is used to train a classifier, which is then used to place a given set of reads into taxonomic bins. In an unsupervised approach, reads are clustered by composition similarity and then the resulting clusters are analyzed in an attempt to place the reads. Subheading 4 of this chapter provide details of multiple approaches and available different tools which readers can use according to their preferences.
> In general, for annotating 16S rRNA sequences we use QIIME [27] and for shotgun sequencing we use MEGAN [31] which can also be used for 16S. MEGAN is a highly efficient program for
interactive analysis and comparison of microbiome data, allowing one to explore hundreds of samples and billions of reads. While taxonomic profiling is performed based on the NCBI taxonomy, MEGAN also provides a number of different functional profiling approaches. MEGAN Community Edition also supports the use of metadata in the context of principal coordinate analysis and clustering analysis [31]. In all the three scenarios explained in this chapter, MEGAN is used as primary tool for annotations. For more details on MEGAN tool, see Chapter 23.
If we have shotgun sequencing then we have good option for functional annotation, but with 16S sequences we can only perform taxonomic analyses with confidence although there are few tools which might predict metagenome functional content from marker genes [32, 33]. Most shotgun annotation pipelines (such as MEGAN [31], MG-RAST [34], IMG/MER [35], EBI Metagenomics [36]) support functional annotations and they often use databases such as KEGG [37], SEED [38], eggNOG [39], and COG/KOG [40], as well as protein domain databases such as TIGRFAM [41] and PFAM [42].
3.2 Metagenome Assembly Similar in nature to the genomic assembly, which is the reconstruction of genomes from the sequenced DNA segments (or reads), metagenome assembly is more complex. The main goal is to stitch together the fragments of the reads that could be from the same genome. Here the reads consist of mixture of DNA from different organisms and also may have widely different levels of abundance. Few recent reviews discussed new challenges and opportunities as well as assessed the most common and freely available metagenome assembly tools with respect to their output statistics, their sensitivity for low-abundance community members and variability in resulting community profiles as well as their ease of use. Interested readers please refer to reviews [43, 44].
3.3 Rarefaction Curves Rarefaction curves represent a powerful method for comparing species richness among habitats on an equal-effort basis based on the construction of the so-called rarefaction curves [45]. This is a very useful tool for statistical data analyses that helps us to Correct for bias in species number due to unequal sample sizes by standardization to the number of species expected in a sample if it had the same total size as the smallest sample. As an example, we have two sample groups, first having 50 individuals and second 30 individuals with multiple number of species obtained from their taxonomic analyses. Rarefaction helps us to compare the situation, if we would have same number of individuals in two sample groups. Rarefaction curves are used differently in case of 16S and shotgun metagenomics. Ni and colleagues have described methods for estimating a reasonable and practical amount for SSU rRNA gene sequencing and explained how much metagenomic sequencing is enough to
Fig. 4 Rarefaction. Rarefaction plot using annotated species profile for all 22 (unstable and stable) atherosclerotic plaque samples. These curves show the number of nodes that would be present if based on 10%, 20%, and up to 90% of the reads
achieve a given goal [46]. In metagenomic shotgun sequencing, the fraction of the metagenome represented in the data set is termed coverage, which can be assessed through rarefaction curve. Interested readers may refer to a recent publication which has advocated for the estimation of the average coverage obtained in metagenomic studies, and briefly presented the advantages of different approaches [47].
In Scenario 1, for comparing case and control groups from human atherosclerotic plaque samples, we computed rarefaction curves from the normalized profile of 22 samples using the bacterial reads, showing the number of nodes that would be present in the analysis if based from 10% to 90% of the reads (Fig. 4). From sequence statistics (Table 1) and the rarefaction curve (Fig. 4), it is apparent that 2 (sample 233 and 238) of the 22 samples had much higher sequencing depth than the other samples. Later in the study we therefore omitted these two samples from merged case vs. control analyses.
Similarly, in Scenario 2 also, rarefaction was performed at various levels to compare diversity for different sample groupings. All groups were rarefied to the lowest read number, and the diversity calculated using weighted and unweighted UniFrac as well as the non-phylogenetic Bray–Curtis dissimilarity measure.
Fig. 5 Genus level taxonomic comparison of patients' microbiome (median of each time point group) in two drug treatment groups for certain time points such as baseline, mid-treatment, end of treatment and follow up. Here different colors indicate different genera and the size of each color in the pie reflects the percentage of those genus in median microbiome for each time point group and for each drug
Fig. 6 Tree view at "family" level taxonomy comparing merged data from cases and control samples using data from Scenario 1
Form this figure we can easily see that the microbiome pattern in drug X over treatment period is more consistent (or more stable over the time) than in drug Y. Here with visual comparison we are not making any conclusion, but with these types of plots we can start to see if there is any trend in our data, which can later be investigated with appropriate statistical tests.
Further as metagenomic data are often hierarchical in nature, besides doing basic plots which can be done only at certain taxonomic levels (e.g., family/genus), often it is helpful to display the whole data as comparative tree view. For example in Scenario 1, samples from cases and controls have grouped closely (as can be seen later in Subheading 3.9), we can explore their broad differences by comparing total biome from cases and controls using comparative tree view (Fig. 6). This kind of tree view also help us to assess multiple time point samples from single patient or grouped data comparison for multiple factors (e.g., in Scenario 3).
3.6 Diversity Analyses Diversity analyses is one of the prominent statistical analysis approaches that address some of the downstream analysis steps associated with metagenomic studies. Species abundance estimates in the community are used to make inference about diversity on the whole community. The terms alpha, beta, and gamma diversity were all introduced by R. H. Whittaker to describe the spatial component of biodiversity [48]. Alpha diversity is just the diversity of each site (samples in each group). Beta diversity represents the differences in species composition among sites. Gamma diversity is the diversity of the entire landscape of different sites (all species pool from multiple samples). A diversity index measures how many different types (such as species) are there in a dataset (a community) and simultaneously takes into account how evenly the basic entities (such as individuals) are distributed among these types. Three commonly used measures of diversity, Simpson's index, Shannon's entropy, and the total number of species, are related to Renyi's definition of a generalized entropy, and are well explained and compared by Hill [49]. Interested readers may also refer to [50] for consistent terminology for quantifying species diversity. Many other publications also explain this topic very well.
3.7 Comparison Using Distance Matrices Another common technique to compare metagenomic datasets is using distance matrices. First, a taxonomic profile is computed for each data set. Second, a matrix of pairwise distances is determined using one of several possible ecological indices. Finally, the distances are represented using an appropriate visualization technique. Mitra et al. [51] explained multiple distance matrices (such as Bray–Curtis, Kulczynski, χ<sup>2</sup> , Hellinger, and Goodall) in the context of multiple metagenome comparison. In addition to these UniFrac is another distance metric used for comparing biological communities. It differs from dissimilarity measures such as Bray–Curtis by incorporating information on the relative relatedness of community members by incorporating phylogenetic distances between observed organisms in the computation [52–54]. Both weighted (quantitative) and unweighted (qualitative) variants of UniFrac are often used in microbial ecology, where the former accounts for abundance of observed organisms, while the latter only considers their presence or absence.
3.8 Boxplots In descriptive statistics, "boxplot" or alternatively called "box and whisker plot," is an important and one of the most informative tools that is used for graphically depicting groups of numerical data through their quartiles [55]. The boxplot is a quick way of examining multiple groups of data graphically, which easily provides information regarding quartiles, range, variation, and even outliers and enables us to compare within and between group samples. For example, Fig. 7 shows distribution of samples in multiple time point for both drugs (example data in Scenario 3). From this plot we can clearly gather the idea that diversity with drug X is consistently higher than that with drug Y. Further in Fig. 5 we have already seen that microbiome pattern in drug X showed less disruption, thus from these two figures we can hypothesize that drug Y
Fig. 7 Boxplot showing Simpson diversity indices for samples from each time point and for both the drugs X and Y
being more disruptive to the microbiome. Such hypotheses can help us in further statistical analyses.
3.9 Hierarchical Clustering Cluster analysis, especially hierarchical clustering [56, 57], is an important tool for the exploratory and unsupervised analysis (where we do not need a training dataset to feed the programme) of high dimensional datasets and often used in genomics and other fields for their ability to simultaneously uncover multiple layers of clustering structure. In our example, Fig. 8 depicts a hierarchical clustering result of family level taxonomic comparison data for all 22 samples. Interestingly, samples 238 and P0613 were mostly different, and among the other samples, all unstable plaques clustered together, apart from all stable plaque controls that clustered separately.
> Interestingly, the asymptomatic atherosclerotic plaques have more abundance of host microbiome-associated microbial families such as Porphyromonadaceae, Bacteroidaceae, Micrococcaceae, and Streptococcaceae than the symptomatic atherosclerotic plaques. In contrast, the symptomatic atherosclerotic plaques have more abundance of pathogenic microbial families such as Helicobacteraceae, Neisseriaceae, and sulfur-consuming families such as sulfur-
Fig. 8 Taxonomic comparison of all DNA samples. Hierarchical clustering result of "family" level taxonomic comparisons of data from Scenario 1: unstable atherosclerotic plaques from 15 patients with symptomatic atherosclerotic disease (unstable plaques) and stable plaques from a control group of seven patients that died from other causes than atherosclerosis (controls). Red indicates downregulation, green indicates upregulation, and black indicates no change in read abundance level comparing to all samples. Hierarchical clustering was computed with average linkage, whereas Pearson correlation was used for clustering the families (rows) and Spearman correlation was used for clustering the datasets (columns), respectively
oxidizing symbionts and Thiotrichaceae than the asymptomatic atherosclerotic plaques (Fig. 8). For P0613, the species profile appeared very different from all other samples. Thus, this sample also treated as an outlier in further analyses (see [15] if interested in actual study).
3.10 Principal Component Analysis (PCA) and Principal Coordinates Analysis (PCoA)
PCA and PCoA are tools for multivariate analysis. PCA uses an orthogonal transformation to convert a set of observations of possibly correlated variables into a set of values of linearly uncorrelated variables called principal components [58]. This is often used for quantitative variables, so the axes in graphic have a quantitative weight, and the positions of the samples are in relation with those weight. On the other hand, PCoA or multidimensional scaling (MDS) is a means of visualizing the level of similarity of individual cases of a dataset [59]. PCoA is similar to Polar ordination (PO; [60]) arranges samples between endpoints or 'poles' according to the distance matrix maximizing the linear correlation between the distances in the distance matrix. If further interested in these methods please see [61].
For multiple sample comparison we often use PCoA and PCA, these are among the best tools available for multivariate analysis. These can give us powerful information of similarities and dissimilarities within samples. When coupled with phenotypic data or metadata (using colors and symbols etc.), these can be very helpful tools to understand within group variations. As an example, we have used PCoA on 22 plaque samples from Scenario 1 (Fig. 9). Here we can see that sample 238 and 238 being very different possibly due to high sequence depth (as also seen in Fig. 4).
Biplots: In addition to PCA or PCoA, variables can also be plotted on the same diagram (this is called a biplot). The biplot provides a useful tool of data analysis and allows the visual appraisal of the structure of large data matrices [62]. In our examples, where taxa are variables, biplot can show important taxa which helps in determining relatedness represented as arrows. For example, in Scenario 2, β diversity was compared using principal coordinate analysis (PCoA) on all samples from all visits, where biplots are displayed with green arrows (Fig. 10). From this PCoA with biplot, we interpret that samples from volunteers 8, 13, and 16 are different than the other volunteers and that they have higher abundance of Succinivibrionaceae, Gammaproteobacteria, Aeromonadales, etc.
3.11 Canonical-Correlation Analysis (CCA) and Canonical-Correspondence Analysis (CCA)
CCA (correlation) seeks to find the linear combination of the Xi and Yj that have the greatest correlation with each other where X ¼ (X1, ..., Xn) and Y ¼ (Y1, ..., Ym) of random variables thus it is often used as a dimension–reduction method. The method was first introduced by Harold Hotelling [63]. On the other hand, CCA (correspondence) is a multivariate method to elucidate the relationships between biological assemblages of species and their
Fig. 9 principal coordinate analyses (PCoA) of "family" level taxonomic comparisons of data from Scenario 1: unstable atherosclerotic plaques from 15 patients with symptomatic atherosclerotic disease (cases: cyan) and stable plaques from a control group of seven patients that died from other causes than atherosclerosis (controls: magenta)
environment. This method by Cajo J. F. ter Braak involves a canonical correlation analysis and a direct gradient analysis [64]. By environment we mean any kind of metadata, such as some physicochemical parameters obtained from same group where the species data is obtained. The idea is to relate the prevalence of a set of species to a collection of environmental variables. Biplots are often used in CCA (correspondence) for visualization purpose. For example, in our Scenario 2, a typical illustration of correlation and correspondence analyses between the microbiome and RBC fatty acid data is displayed in Fig. 11.
In this occasion it is important to note that CCA does not perform variable selection. Further, when the number of variables exceeds the number of observations (or sample size), CCA cannot be applied directly due to singularity of the covariance matrix. In a recent study [65] the authors have discussed this problem and a few existing solutions. Additionally, they developed a method for structure-constrained sparse canonical correlation analysis
Fig. 10 principal coordinate analyses (PCoA) of level taxonomic comparisons of data from Scenario 2: all samples (V1–V5) for all participants, where biplots are displayed with green arrows. Each visit is denoted by a different color
(ssCCA) in a high-dimensional setting. ssCCA takes into account the phylogenetic relationships among bacteria, which provides important prior knowledge on evolutionary relationships among bacterial taxa (see [65] if interested).
3.12 Multivariate Analyses Multivariate data analysis refers to any statistical approach used to analyze data with more than one variable. For example, as described in Scenario 3 we have multiple factors. The key to identifying important microbial taxa associated with two treatments is that the large datasets from each patient are compared within groups, and then the metadata from the patients' groups are compared against each other. Analysis of multivariate data in response to factors, groups, or treatments in an experimental design needs sophisticated methods.
> To achieve this, we can use PERMANOVA (permutational multivariate analysis of variance) [66] to test the homogeneity of multivariate dispersions within groups, on the basis of any resemblance measure. PERMANOVA is a better approach than ANOVA (Analysis of variance)/MANOVA (Multivariate analysis of variance) for our study as PERMANOVA works with any distance measure that is appropriate to the data, and uses permutations to make it
Fig. 11 (a) Pearson correlation between genus level microbiome and RBC fatty acid data. (b) Canonical correspondence analysis of microbiome (genus level taxonomy) distribution in relation to blood parameters (biplot: represented by blue arrows). Red crosses represent taxa and black circles represents individual samples
distribution free, unlike assuming normal distributions. Finally, in addition to the above multiple comparisons, we can examine if there is consistency of microbiota changes and patterns across the geographical locales of treatment subjects; as our samples are from different countries. We are not showing the details of multivariate analyses, but there are multiple available packages for such analyses with good tutorials. Interested readers may visit these packages and websites as detailed below.
The Primer-E package [67] is commonly used by microbial ecologists and allows for multiple multivariate statistical analyses. We often use R statistical programming language [21] for multivariate statistics. Moreover R is used for several types of graphical representations. Particular packages provide in-built functions and libraries (within R environment) specially for metagenomic datasets such as Bioconductor [68], vegan [69], and phyloseq [70].
#### 4 Tools and Packages Commonly Used in Metagenomic Studies
A list of multiple tools is provided below for analyzing metagenomic data from raw sequence reads to final comparisons and statistical analyses. Discussion of all these tools are beyond the scope of this chapter, but interested readers can see recent review articles [71–74] and it must be noted that there can be other tools as well outside this list.
- (a) FastQC (https://www.bioinformatics.babraham.ac.uk/ projects/fastqc/).
- (b) Fastx\_toolkit (http://hannonlab.cshl.edu/fastx\_toolkit/).
- (c) Cut-adapt (both adapter trimming and quality trim) [25].
- (d) BBTools (http://jgi.doe.gov/data-and-tools/bbtools/).
- (e) Condetri (Read trimmer for Illumina data) [75].
- (f) Trimmomatic (allows multiple threads) [76].
- (g) SolexaQA [77].
- (h) PRINSEQ [78].
- (a) BLAST [18].
- (b) USEARCH [28].
- (c) DIAMOND [22].
- (d) Rapsearch [79].
- (e) PyNAST [29].
- (a) QIIME [27].
- (b) USEARCH [28].
- (c) RDP classifier [30].
- (d) SILVA (for 16S + 18S) [80].
- (e) Mothur [81].
- (f) SILVAngs (https://www.arb-silva.de/documentation/ silvangs/).
- (g) MEGAN [31].
- (h) AmpliconNoise [82].
- (i) Open reading frame (ORF) prediction, for example, with MG-DOTUR [83].
- (a) Reference-based assembly.
- <sup>l</sup> MIRA 4 [84].
- <sup>l</sup> MetaAMOS (https://www.cbcb.umd.edu/software/ metamos).
- (b) De novo assembly.
- <sup>l</sup> Newbler (Roche).
- <sup>l</sup> iAssembler [85].
- <sup>l</sup> EULER [86].
- <sup>l</sup> MetaVelvet-SL [90].
- <sup>l</sup> Meta-IDBA [91].
- <sup>l</sup> InteMAP [92].
- <sup>l</sup> SAT-Assembler [93].
- <sup>l</sup> IDBA-UD [94].
- (a) Bowtie 2 [17].
- (a) Composition-based binning algorithms.
- <sup>l</sup> S-GSOM [95].
- <sup>l</sup> PhylopythiaS [96].
- <sup>l</sup> TACAO [97].
- <sup>l</sup> PCAHIER [98].
- <sup>l</sup> ESOM [95].
- <sup>l</sup> ClaMS [99].
- (b) Similarity-based binning software include tools.
- <sup>l</sup> MEGAN [31].
- <sup>l</sup> IMG/MER 4 [35].
- <sup>l</sup> MG-RAST [34].
- <sup>l</sup> CARMA [100].
- <sup>l</sup> MetaPhyler [101].
- (c) Unsupervised binning.
- <sup>l</sup> PhylopythiaS+ [102].
- <sup>l</sup> PhymmBL [103].
- <sup>l</sup> ESOMs [104].
- <sup>l</sup> VizBin [105].
- <sup>l</sup> IFCM (fuzzy c-means method) [106].
- (a) ICoVeR [107].
- (b) MyCC [108].
- (c) MetaBAT [109].
- (d) GroopM [110].
- (a) MetaGeneMark [113].
- (b) Prodigal [114].
- (c) Orphelia [115].
- (d) FragGeneScan [116].
- (a) CRT [117].
- (b) PILER-CR [118].
- (c) IMG/MER [35].
- (a) MEGAN [31].
- (b) QIIME for 16S projects [27].
- (c) Galaxy platform.
- (d) MG-RAST [34].
- (e) IMG/MER [35].
- (f) Primer-E package [67].
- (g) Several packages built within R [21].
- <sup>l</sup> Vegan [69].
- <sup>l</sup> Phyloseq [70].
- <sup>l</sup> Bioconductor [68].
- (a) PICRUSt [33].
- (b) Tax4Fun [32].
- (a) R [21].
- (b) Many other tools can be used for statistical analyses.
- (a) The EBI Metagenomics service [36].
- (b) European Nucleotide Archive (ENA).
- (c) MG-RAST [34].
- (d) METAGENassist [119].
- (e) BusyBee Web [120].
- (f) Meta4 [121].
#### 5 Concluding Remarks
This chapter has illustrated multiple data analyses and annotation techniques in metagenomic studies with three case studies. This is not a chapter about any new method development but a description of optimized pipelines using various available tools. With these example scenarios, the use of multiple pipelines has been demonstrated to analyze and interpret the data starting from very raw sequence to the final statistical outputs. Example scenarios describe some of the tools that we have used for analyzing the projects selected for demonstration, but besides these there are plenty of other available tools for metagenomics, most of which are listed in Subheading 4. This chapter does not provide the details of the tools or describe their pros and cons but this can be a good starting point for the readers to explore available options to analyze and interpret their datasets. From this chapter readers shall get an idea of current research projects in medical studies and multiple approaches used to analyze the data originating from these projects, although readers should keep in mind that this is not an exclusive list of possible pipelines for analyzing metagenomic samples. There might be other approaches as well. While step-by-step instructions of all the tools is beyond the scope of this chapter, the methods outline here might be useful to researchers to plan, analyze, and interpret their research projects successfully.
#### References
Characterization of the gut microbiome using 16S or shotgun metagenomics. Front Microbiol 7:459
data". https://github.com/ ExpressionAnalysis/ea-utils
using phylogenetic networks based on ecological indices. ISME J 4(10):1236–1242
independent, community-supported software for describing and comparing microbial communities. Appl Environ Microbiol 75 (23):7537–7541
sequencing data. PLoS Comput Biol 10(8): e1003737
augmented binning of metagenomic data. Microbiome 3:1
palindromic repeats. BMC Bioinformatics 8:209
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Chapter 21
#### Systems Genetics for Evolutionary Studies
#### Pjotr Prins, Geert Smant, Danny Arends, Megan K. Mulligan, Rob W. Williams, and Ritsert C. Jansen
#### Abstract
Systems genetics combines high-throughput genomic data with genetic analysis. In this chapter, we review and discuss application of systems genetics in the context of evolutionary studies, in which high-throughput molecular technologies are being combined with quantitative trait locus (QTL) analysis in segregating populations.
The recent explosion of high-throughput data—measuring thousands of RNAs, proteins, and metabolites, using deep sequencing, mass spectrometry, chromatin, methyl-DNA immunoprecipitation, etc. allows the dissection of causes of genetic variation underlying quantitative phenotypes of all types. To deal with the sheer amount of data, powerful statistical tools are needed to analyze multidimensional relationships and to extract valuable information and new modes and mechanisms of changes both within and between species. In the context of evolutionary computational biology, a well-designed experiment and the right population can help dissect complex traits likely to be under selection using proven statistical methods for associating phenotypic variation with chromosomal locations.
Recent evolutionary expression QTL (eQTL) studies focus on gene expression adaptations, mapping the gene expression landscape, and, tentatively, define networks of transcripts and proteins that are jointly modulated sets of eQTL networks. Here, we discuss the possibility of introducing an evolutionary "prior" in the form of gene families displaying evidence of positive selection, and using that prior in the context of an eQTL experiment for elucidating host-pathogen protein-protein interactions.
Here we review one exemplar evolutionairy eQTL experiment and discuss experimental design, choice of platforms, analysis methods, scope, and interpretation of results. In brief we highlight how eQTL are defined; how they are used to assemble interacting and causally connected networks of RNAs, proteins, and metabolites; and how some QTLs can be efficiently converted to reasonably well-defined sequence variants.
Key words Systems genetics, Genetical genomics, QTL, eQTL, xQTL, R-genes, Evolution, R/qtl, LMM, GEMMA, NGS, Genomics, Metabolomics, Network inference, GeneNetwork
#### 1 Introduction
Genetics concerns the study of heritably quantitative or complex traits. Many agricultural traits of interest, such as milk production in cattle and response to fertilizer in crops and most human, animal, and plant diseases, are complex traits. Associating, or linking,
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_21, © The Author(s) 2019
complex traits with certain positions on the genome is achieved through the mapping of the so-called quantitative trait loci (QTL).
Mapping QTL in experimental populations is possible when linkage and/or association information is available. When we have a population of individuals with known genotypes, it may be possible to link a phenotype with a certain genotype. To genotype individuals, first marker maps are created. A marker is a known genomic location, where the genotype of an individual can be determined. In the early days, the genotype was determined by visible chromosome features, later with restriction fragment length polymorphism (RFLP) and amplified fragment length polymorphism (AFLP, see also [1–3]), and, increasingly, with SNP/haplotype data [4]. When all individuals with genotype A at a marker location somewhere on the genome are susceptible to a disease and all other individuals with genotype B are not, there is linkage/association or a QTL. If it is clear cut, i.e., single QTL explains all phenotype variance, it is likely to be a single gene effect. Often it is not clear cut, and we need statistics to determine the strength of association between phenotype and genotype.
It is also possible to use linkage disequilibrium (LD) to map QTL in outbred and natural populations. LD occurs when certain stretches of the genome (haplotypes) show nonrandom behavior based on allele frequencies and recombination. Associating haplotype frequencies with phenotypes potentially renders QTL. Kim et al. describe the genome-wide pattern of LD in a sample of 19 Arabidopsis thaliana accessions using SNP microarrays [5]. LD is tested, for example, by Dixon et al., to globally map the effect of polymorphism on gene expression in 400 children from families recruited through a proband with asthma [6].
The use of terms "association" and "linkage" can be confusing, even in literature. In this text we use association with haplotypes in natural populations of unrelated individuals and linkage with markers in families and groups of families, often termed experimental populations. Note some genetic studies are hybrids of both methods, such as Dixon et al. [6], and individuals are related, i.e., some within-family linkage information is available for 400 children from 206 families which should be accounted for in the analysis.
Statistical power can be increased by using experimental crosses instead of natural populations. For example, each individual line in a set of recombinant inbred lines (RILs) is homozygous across the genome, doubling the genetic variance, simplifying genetic models, and increasing statistical power. For model organisms, such as A. thaliana, Caenorhabditis elegans, Drosophila melanogaster, and Mus musculus, genotyped and even fully sequenced experimental crosses are available; i.e., for these species it is not necessary to generate a new cross, and for these crosses comprehensive SNP and sequence data may be available. One of the features of inbred model organisms is that they are "immortal" which means that experiments conducted more than 10, even 30, years ago can still be compared with those today. Databases, such as GeneNetwork [7, 8], contain thousands of studies conducted on the same individual mouse strains.
Systems genetics combines genetics with high-throughput molecular technologies. Combining gene expression, as measured by microarray probes or RNA sequencing, with linkage leads to gene expression QTL (eQTL). Such eQTL studies elucidate how genotypic variation underlies, for example, morphological phenotypes, by using gene expression levels as intermediate molecular phenotypes. In other words, the expression level, as measured by a microarray probe or probe set, is treated as a phenotype, i.e., a gene expression trait. This phenotype is associated with the genome in the form of one or more eQTL. With microarrays, the genomic location of the probe is usually known. Therefore, expression phenotype and probe connect two types of genomic information: eQTL location(s) and gene location. It is usually assumed that eQTL loci represent cis- or trans-transcription regulators of the target gene [9]. If the eQTL is located close to the gene on the genome, the eQTL may point to a cis-regulator. If the eQTL is located far from the gene on the genome, the eQTL may point to a trans-regulator of a single gene or even eQTL trans-bands that regulate multiple genes (see Fig. 1a and [10, 11]).
In a similar fashion, proteins and metabolites can be measured to map protein QTL ( pQTL) and metabolite QTL (mQTL). A remarkable study published in 1994 used two-dimensional protein electrophoresis and a restriction fragment length polymorphism map (RFLP) [12]. Deep sequencing, chromatin, and methyl-DNA immunoprecipitation are just a few of the latest technologies that add to the arsenal of tools available for the study of the genetic variation underlying quantitative phenotypes. Together, eQTL, mQTL, and pQTL are referred to as xQTL. Different xQTL appear to confirm each other, for example, with the A. thaliana glucosinolate pathway where eQTL, mQTL, and pQTL were mapped together and used to infer the underlying pathways [13]. Such causal inference can lead to dissecting pathways and gene networks which is an active field of research, e.g., [14–16] (see also Fig. 1).
1.1 Evolutionary xQTL Studies From the perspective of evolutionary biology, systems genetics has been applied to elucidate evolutionary adaptations of transcript regulation. For example, Fraser et al. introduced a test for lineage-specific selection on gene expression and analyzed the directionality of microarray eQTL for 112 haploid segregants of a genetic cross between two strains of the budding yeast Saccharomyces cerevisiae, reanalyzing the two-color cDNA microarray data of Brem and Kruglyak [17]. They found that hundreds of gene expression levels have been subjected to lineage-specific selection. Comparing these findings with independent population genetic
Fig. 1 In this hypothetical and schematic example related to mapped locations on a chromosome, prior information is combined with multiple phenotype-genotype QTL mappings to zoom in on genomic areas and to reason about causal relations between different layers of information. (a) The prior (red area on the chromosome) points out that certain sections are of interest; these sections consist of related genes with high homology showing evidence of positive selection, as discussed in the main text. The blue double arrow points out the confidence interval for each QTL, above the significance threshold (red dotted line). The accumulated evidence (light-blue areas) leads to a narrowed down section on the genome, where in this case the prior information is the most specific. In addition, expression phenotypes A and B point to exact gene locations (dotted line, based on exact probe information). (b) To infer causal relationships, network inference is possible. On the left (vertical I), traits A, B, and D map to one hot spot, where A may be a regulator of B because one QTL is shared. B causes metabolite phenotype C, again a shared QTL. Phenotype D matches A and B, and phenotype E matches A, B, and C. These causal relationships are drawn by arrows. The figure suggests that, even if individual QTL are not very informative, the accumulated evidence starts to paint a picture
evidence of selective sweeps suggests that this lineage-specific selection has resulted in recent sweeps at over a hundred genes, most of which led to increased transcript levels. Fraser et al. suggest that adaptive evolution of gene expression is common in yeast, that regulatory adaptation can occur at the level of entire pathways, and that similar genome-wide scans may be possible in other species, including human [18].
In another S. cerevisiae study, Zou et al., by reanalyzing the same two-color cDNA microarray data, uncovered genetic regulatory network divergence between duplicate genes. They found evidence that the regulation of the ancestral gene diverged due to gene duplication [19].
Li et al. studied plasticity of gene expression in C. elegans, using a set of 80 RILs generated from a cross of N2 (Bristol) and CB4856 (Hawaii), representing two genetic and ecological extremes of C. elegans. While the overall level of polymorphism among wild isolates of C. elegans is relatively low, the genetic distance between N2 and CB4856 is high, representing millions of years of genetic drift. Differential expression induced in a RIL population by temperatures of 16 C and 24 C has a strong genetic component. With a group of transgenes, there was prominent evidence for a common master regulator: an eQTL trans-band of 66 coregulated genes appeared at 24 C. The results suggest widespread genetic variation of differential expression responses to environmental impacts and demonstrate the potential of systems genetics for mapping the molecular determinants of phenotypic plasticity [11], leading to a more generalized systems genetics, where value is added from environmental perturbation [20].
Hager et al. determined that genetic architecture supports mosaic brain evolution and independent brain-body size regulation by a quantitative genetic approach involving over 10,000 BXD mouse RILs. The BXD family consists of over 100 lines derived from parental strains that differ at five million single nucleotide polymorphisms (SNPs), indels, transposons, and copy-number variants. This model system harbors naturally occurring genetic variation at a level approximating that of human populations. The study utilizes a high-density linkage analysis to map loci modulating phenotypic variation in overall brain size, body size, and the size of seven major brain parts: neocortex, cerebellum, striatum, olfactory bulb, hippocampus, lateral geniculate nucleus, and basolateral complex of the amygdala. Under the mosaic evolutionary hypothesis, the size of different systems evolves independently due to differential selective pressures associated with different tasks. They identified independent loci for size variation in seven key parts of the brain and observe that brain parts show low or no phenotypic correlation, as is predicted by a mosaic scenario. They also demonstrate that variation in brain size is independently regulated from body size [21].
Kliebenstein et al. detected significant gene network variation in 148 RILs originating from a cross between two A. thaliana accessions, Bay-0 and Shahdara. They were able to identify eQTL controlling network responses for 18 out of 20 a priori defined gene networks, representing 239 genes [22].
According to Gilad, eQTL studies show that (1) variation in gene expression levels is both widespread and highly heritable; (2) gene expression levels are highly amenable to genetic mapping; and (3) most strong eQTL are found near the target gene, suggesting that variation in cis-regulatory elements underlies much of the observed variation in gene expression levels [23]. Meanwhile, Alberts et al. suggest that sequence polymorphisms influencing the binding of microarray probes may cause many false cis eQTL, which should be accounted for [24].
1.2 Adding a Prior QTL mapping links complex traits with one or more locations on the genome (see Fig. 1). Such a location is a wide measure because a QTL is a statistical estimate and rarely a precise indicator. On the genome, a single QTL may represent tens, hundreds, and even thousands of real genes. Combining the QTL with highthroughput technologies, such as microarrays, can add information. To zoom in on the genes underlying QTL, information from other sources has to be utilized. Such a priori knowledge (prior) could consist of results from traditional linkage studies or association studies of, for example, human disease. That way one can assign a specific regulatory role to polymorphic sites in a genomic region known to be associated with disease [23]. Other useful priors can be derived from existing information on gene ontology terms, metabolic pathways, and protein-protein interactions, which can be used to identify genes and pathways [25], provided these databases are sufficiently informative.
Zou et al., for example, used gene ontology as a prior and concluded that trans-acting eQTL divergence between duplicate pairs of genes is related to a fitness defect under treatment conditions, but not with fitness under normal condition [19].
Chen et al. identified strong candidate genes for resistance to leaf rust in barley and on the general pathogen response pathway using a custom barley microarray on 144 doubled haploid lines of the St/Mx population [26]. Fifteen thousand six hundred and eighty-five eQTL were mapped from 9557 genes. Correlation analysis identified 128 genes that were correlated with resistance, of which 89 had eQTL colocating with the phenotypic QTL (phQTL) or classic QTL. Transcript abundance in the parents and conservation of synteny with rice prioritized six genes as candidates for Rphq11, the phQTL of largest effect [26].
In this chapter we discuss the steps needed to design an xQTL experiment to make use of systems genetics in evolutionary studies more concrete. As the prior we add information on plant host genes showing evidence of positive selection.
#### 2 Designing an Evolutionary xQTL Experiment
An experimental design based on systems genetics can highlight sections of the genome showing correlation with an evolutionary trait. One such evolutionary trait of interest is plant resistance against pathogens. Plants have developed mechanisms to defend themselves against pests. When a pathogen, such as potato blight Phytophthora infestans, or a nematode, such as Meloidogyne hapla, infects a plant, it uses a battery of so-called effectors to help invade the plant. Some of these effector molecules act to dissolve cellulose [27]. Intriguingly, other molecules are involved in actively reprogramming plant cells. Such plant-pathogen effectors have been shown to mimic plant transcription factors [28] and switch on genes that help the pathogen [29]. A susceptible plant allows the pathogen to suppress defense mechanisms and to change cell configuration. For example, the nematodes M. hapla and Globodera rostochiensis transform plant cells, so they become elaborate feeding structures. The genetics of this plant-pathogen interaction is potentially even relevant for human medicine, as an increased understanding of host-pathogen relationships may help understand the workings of the innate immune system and nematode immunomodulation [30, 31]. The innate immune system, through plant resistance genes (R-genes, see Box 1), influences susceptibility to infections in all multicellular organisms and is a much older evolutionary mechanism than the advanced adaptive immune system found in higher organisms.
#### Box 1: Adaptive evolution in R-genes
Plant resistance genes (R-genes) are a homologous family of genes, formed by gene duplication events and hypothesized to be involved in an evolutionary arms race with pathogen effectors. R-genes are involved in recognizing specific pathogens with cognate avirulence genes and initiating defense signaling that results in disease resistance [32]. R-genes are characterized by a molecular gene-for-gene interaction [33] in which a specific allele of a disease resistance gene recognizes an avirulence protein or pathogen allele. This specificity is often encoded, at least in part, in a relatively fast-evolving leucine-rich repeat (LRR) region [34], which consists of a varying number of LRR modules. Activation of at least some
(continued)
of these proteins is regulated in trans, as has been shown for RPM1 and RPS2 [35].
A single A. thaliana plant has about 150 R-genes, representing a subset of R-genes in the overall population. The protein products of R-genes are involved in molecular interactions. They generally have a recognition site which can dock against, i.e., recognize, one or more specific molecule(s). The proteins encoded by the largest class of R-genes carry a nucleotide-binding site LRR domain (NB-LRR, also referred to as NB-ARC-LRR and NBS-LRR). NB-LRR R-genes can be further subdivided based on their N-terminal structural features into TIR-NB-LRR, which have homology to the Drosophila Toll and mammalian interleukin-1 receptors and CC-NB-LRR, which contain a putative coiled-coil motif [36]. The LRR domain appears to mediate specificity in pathogen recognition, while the N-terminal TIR, or coiled-coil motif, is likely to play a role in downstream signaling [34]. When a molecule is docked, the R-protein is able to activate pathways in the cell, resulting in, for example, a hypersensitive response causing apoptosis and preventing spread of infection.
Meanwhile, one single R-protein only recognizes one type of invading molecule. Therefore, through its R-genes, one individual plant only recognizes a limited number of strains of invading pathogens, as the individual pathogens have variation in effectors too. When a pathogen evolves to use nonrecognized effectors, the plant becomes susceptible. The success of plant defense is determined by both evolution and the variation of specificity in a population. Unlike the evolved mammal immune system, which can change in a living organism and learn about invasions "on the fly" [37], plant R-genes depend on the variation inside a gene pool to provide the resistance against a pathogen; see, for example, Holub et al. [38]. Even so, many genes involved in pathogen recognition undergo rapid adaptive evolution [39], and studies have found that A. thaliana R-genes show evidence of positive selection, e.g., [40–42].
In this chapter we do not limit ourselves to (known) R-genes. Plants have evolved a complex array of chemical and enzymatic defenses, both constitutive and inducible, that are not involved in pathogen detection but whose effectiveness influences pathogenesis and disease resistance. The genes underlying these defenses comprise a substantial portion of the host genome. Based on genomic sequencing, it is estimated that some 14% of the 21,000 genes in A. thaliana are related to defense against pathogens [43]. Most of these genes are not involved in direct pathogen detection, but their protein products interact directly with pathogen proteins or protein products at the molecular level. Among these proteins, for example, are chitinases and endoglucanases that attack and degrade the cell walls of pathogens and which pathogens counterattack with inhibitors. Such systems of antagonistically interacting proteins provide the opportunity for molecular coevolution of individual systems of attack and resistance [39].
In this chapter we design an experiment to look for all gene families showing evidence of positive selection. This evidence of positive selection is the prior for eQTL analysis: combining known genomic locations of gene families with eQTL locations derived from gene expression variation in a host-pathogen interaction experiment, which hopefully results in zooming in on gene families involved in plant resistance. The prior adds statistical power in locating putative gene families involved in host-pathogen coevolution (Fig. 1). Note that, in this chapter, the term "interaction" is used in two ways. The first is for QTL interaction, where two QTL on the genome interact statistically. The second is for hostpathogen gene-for-gene interaction, where gene products from different species interact physically.
2.1 Create a Prior with PAML To create the prior, we use Ziheng Yang's codeml implementation of phylogenetic analysis by maximum likelihood (PAML) [44]. PAML can find amino acid sites which show evidence of positive selection using dN/d<sup>S</sup> ratios, which is the ratio of non-synonymous over synonymous substitution (ω, see [44]). The calculation of maximum likelihood for multiple evolutionary models is computationally expensive, and executing PAML over an alignment of a hundred sequences may take hours, sometimes days, on a PC. The software for generating the prior is prepackaged and makes up the workflow in Chap. 25, which includes BLAST [45], Clustal Omega [46], pal2nal [47], PAML [44], and BioRuby [48].
> It is possible to find nonoverlapping large gene families by using BLASTCLUST, a tool that is part of the BLAST tool set [45]. After fetching the A. thaliana cDNA sequences from the Arabidopsis Information Resource (TAIR) [49], convert the sequences to a protein BLAST database format. Based on a homology criterion, the identity score and genes are clustered into putative gene families by running BLASTCLUST with 70% amino acid sequence identity. Note that the percentage identity may not render all families and will leave out a number of genes. It is used here for demonstration purposes only. For A. thaliana such a genome-wide search finds at least 60 gene families, including some R-gene families.
After aligning all family sequences, use PAML's codeml to find evidence of positive selection in the gene families. Clustal Omega is used to align the amino acid sequences and create a phylogenetic tree. Next, pal2nal creates codon alignments, which can be used by PAML. Finally run PAML's codeml M0-M3 (one ratio vs. nearly neutral) tests and M7-M8 (beta vs. beta + ω) tests in a computing cluster environment as shown in Chap. 25.
An M0-M3 χ2 test finds that 43 gene families (out of 60) show significant evidence of positive selection. M7-M8, meanwhile, finds 35 gene families. Therefore, based on the described procedure, approximately half the families show significant evidence of positive selection and can be considered candidate gene families involved in host-pathogen interactions. Note that this number contains false positives because the evolutionary model may be too simplistic; see also [50]. Nevertheless, these candidate gene families can be used as an effective filter for further research.
When a gene family displays evidence of positive selection, the genome locations can be used as a prior for systems genetics (see Fig. 1). With the full genome sequence of A. thaliana available, the location of gene families showing evidence of positive selection is known. For example, in the Columbia (Col-0) ecotype, the majority of the 149 R-genes are combined in clusters spreading 2–9 loci; the remaining 40 are isolated. Clusters are organized in so-called superclusters [36, 51]. Phylogenetic analysis shows that such clusters are the result of both old segmental duplications and recent chromosome rearrangements [36, 52].
#### 2.2 Select a Suitable Experimental Population To select a suitable experimental population, the choice of parents is key. Because we want a descriptive evolutionary prior based on gene families with known genome locations, we also need a sequenced genome, from one parent and ideally from both of the parental strains. The choice of parents for QTL analysis is normally based on large (classical) phenotypic differences. For testing pathogen resistance, the choice would ideally be one susceptible parent and one resistant (nonsusceptible) parent. For eQTL, phylogenetic distance can be used, when there is no obvious phenotype. In general, it is a good idea to choose one or both parents from common library strains based on, for example, Columbia (Col-0), Landsberg erecta (Ler-0), Wassilewskija (Ws-0), or Kashmir (Kas-1). This is because a great number of experimental resources and online information will be available. In addition, a reference genetic background is provided in this way, which allows the comparison of the effects of QTL and mutant alleles [53]. A number of RIL populations can be found through TAIR, a model organism database providing a centralized, curated gateway to Arabidopsis biology, research materials, and community [49].
#### 2.3 Select an xQTL Technology
A large part of published xQTL studies is based on gene expression eQTL partly because gene expression probe provides a direct genomic link. When it comes to selecting single-color or two-color arrays, one consideration may be that two-color arrays have higher efficiency when using a distant pair design [54].
Deep sequencing technology (RNA-seq, [55]) is affordable for eQTL studies. The main advantage over microarrays is improved signal-to-noise ratios and possibly improved coverage depending on the reference genome. Microarrays are noisy partly due to cross hybridization, e.g., [56], and have limited signal on low-abundance transcripts or expressors; both facts are detrimental to significance. Deep sequencing is no panacea, however, since it accentuates the high expressors. High expressors are expressed thousands of times higher than low expressors. Low expressors may lack significance for differential expression. Worse because deep sequencing is stochastic, many low expressors may even be absent. Another point to consider is that currently at least 1 in 1000 nucleotide base pairs is misread, which makes it harder to disentangle error from genetic variation. Only when a sequence polymorphism is measured many times (say 20), it can be considered to represent genetic variation.
Also a choice for a certain eQTL technology should take into account that, when looking at differential gene expression analysis, different microarray platforms agree with each other, but overlap between microarray and deep sequencing is much lower, suggesting a technical bias [57].
For an example of a metabolite mQTL study, see Keurentjes et al. [58] and Fu et al. [59]. For a study integrating eQTL, pQTL, mQTL, and classical phenotypic QTL, see Fu et al. [60] and Jansen et al. [13].
The size of the experimental population should be large enough to give informative results. For classical QTL analysis, the sizing may be assisted using estimates of total environmental variance and the total genetic variance derived from the accessions, selected as parents. Roughly, population sizes of 200 RILs, without replications, will allow detection of large-effect QTL with an explained variance of 10% in confidence intervals of 10–20 cM. Detection of smalleffect QTL or mapping accuracy below 5% requires increasing the population size to at least 300 RILs [53]. It is important to note that QTL mapping accuracy is a function of marker density and population size. The number of strains to use differs between inbred lines. The promise of extreme dense marker maps, such as delivered by SNPs, does not automatically translate to higher accuracy. It is the number of recombination events in the population for a particular genomic region that limits QTL interval size. In fact, current marker maps, in the order of thousands of (evenly spread) markers per genome, suite population sizes of a few hundred RILs.
#### 2.4 Sizing the Experimental Population
It is a fallacy, for example, to expect higher mapping power when combining an ultradense SNP map with just 20 individuals.
For high-throughput xQTL, the experimental population should be sized against an acceptable false discovery rate (FDR), minimizing for type I and type II errors. This can be achieved using a permutation strategy to assess statistical significance, maintaining the correlation of the expression traits while destroying any genetic linkages or associations in natural populations: marker data is permuted while keeping the correlation structure in the trait data, such as presented by Breitling et al. [61]. Unfortunately, this information differs for every experiment and is only available afterward. Analyzing a similar experiment, using the same tissue and data acquisition technology, may give an indication [60], but when no such material is available, a crude estimate may be had by taking the thresholds of a (classic) single-trait QTL experiment and adjusting that for multiple testing by the Bonferroni correction (minimize type I errors) or Benjamini-Hochberg correction (minimize type II errors). Note that Bonferroni results in a very conservative estimate.
#### 2.5 Analyzing the xQTL Experiment with R/qtl
R/qtl is extensible, interactive free software for the mapping of xQTL in experimental crosses. It is implemented as an add-on package for the widely used statistical language/software R. Since its introduction, R/qtl has become a reference implementation with an extensive guide on QTL mapping [62].
R/qtl includes multiple QTL mapping (MQM), as described in [10], an automated procedure, which combines the strengths of generalized linear model regression with those of interval mapping. MQM can handle missing data by analyzing probable genotypes. MQM selects important marker cofactors by multiple regression and backward elimination. QTL are moved along the chromosomes using these preselected markers as cofactors. QTL are interval mapped using the most informative model through maximum likelihood. MQM for R/qtl brings the following advantages to QTL mapping: (1) higher power, as long as the QTL explain a reasonable amount of variation; (2) protection against overfitting, because MQM fixes the residual variance from the full model; (3) prevention of ghost QTL detection (between two QTL in coupling phase); and (4) detection of negating QTL (QTL in repulsion phase) [10].
MQM for R/qtl brings additional advantages to systems genetics data sets with hundreds to millions of traits: (5) a pragmatic permutation strategy for control of the FDR and prevention of locating false QTL hot spots, as discussed above; (6) highperformance computing by scaling on multi-CPU computers, as well as clustered computers, by calculating phenotypes in parallel, through the message passing interface (MPI) of the parallel package for R; and (7) visualizations for exploring interactions in a genomic circle plot and cis- and trans-regulation. MQM comes with a 40-page tutorial for MQM and is part of the software distribution of R/qtl [10, 63].
2.6 Matching the Prior After detecting eQTL, we have a map of gene regulation in the form of a cis-trans map. When taking a priori information into account, i.e., genomic locations derived through other methods, we can potentially match the genomic locations of genes and gene families with the eQTL cis-trans map. Until now, there has been no combined QTL and evolutionary study, involving PAML, for hostpathogen relationships in plants, though they have been conducted separately.
2.7 Combining xQTL Results: Causality and Network Inference
In addition to identifying eQTL or xQTL, it is possible to think in terms of grouping related traits by correlations. Molecular and phenotypic traits can be informative for inferring underlying molecular networks. When two independent non-correlated traits share multiple QTL, inference of a functional relationship is possible (Fig. 1b). Thus, distinguishing trait causality, reactivity, or independence can be based upon logic involving underlying QTL. This was the basic idea in Jansen and Nap 2001 [64]. Later, people started to use biological variation as an extra source for reasoning because if A affects B, biological variation in trait A is propagated to B and not vice versa. This assumes there is no hidden trait C affecting both A and B; see also Li et al. [15].
Mapping QTL for thousands of molecular phenotypes is the first step in attempting to reconstruct gene networks. Not only can network reconstruction be used within a particular layer, say within eQTL analysis, i.e., transcript data only, but also across layers. Such interlevel (system) analysis integrates transcript eQTL, protein pQTL, metabolite mQTL, and classical QTL [13].
The examination of pairwise correlation between traits can lead to the hypothesis of a functional relationship when that correlation is high. Beyond the detected QTL, the correlation between residuals among traits, after accounting for QTL effects, or correlations between traits conditional on other traits is further evidence for a network connection. To infer directional effects, it is necessary to analyze the correlations among pairs of traits in detail. If trait A maps to a subset of the QTL of trait B, then the common QTL can be taken as evidence for their network connection, while the distinct QTL can be used to infer the direction (Fig. 1b), unless all the common QTL have widespread pleiotropic effects, which is when a single gene influences multiple traits. If traits A and B have common QTL, without QTL that are distinct, then the inference is more complicated, and further analysis is needed to discriminate pleiotropy from any of the possible orderings among traits [13, 15].
Li et al. [15] point out that, despite the exciting possibilities of correlation analysis, extreme caution is advised, especially in intralevel analyses, owing to the potential impact of correlated measurement error (leading to false-positive connections). By introducing a prior, however, causal inference becomes feasible for realistic population sizes [15]. The outcome of a causal inference on two traits sharing a common QTL may be either that one is causal for the other or that they are independent. In the first case, QTL-induced variation is propagated from one trait to the other, while in the latter case, the two traits may even be regulated by different genes or polymorphisms within the QTL region, and their apparent relationship (correlation) is explained by linkage disequilibrium and not by a shared biological pathway [15].
#### 3 Discussion
A QTL is a statistical property connecting genotype with phenotype. In this chapter, we reviewed studies which, with various degrees of success, combine some type of prior information with xQTL. We propose that a search for genome-wide evidence of positive selection can produce a valid and interesting prior for xQTL analysis. This is achieved by combining information of genomic locations of putative gene families, possibly involved in plantpathogen interactions, with QTL locations derived from a systems genetics experiment. Both the eQTL example and the search for genome-wide evidence of positive selection pressure are essentially exploratory and result in a list of putative genes, or gene families, with known genomic locations. The combined information yields candidate genes and pathways that are under positive selection pressure and, potentially, involved in host-pathogen interactions. We explain that it is possible to design an eQTL experiment using existing experimental populations, e.g., using an A. thaliana RIL population, and analyze results with existing free and open-source software, such as the R/qtl tool set.
Systems genetics bridges the study of quantitative traits with molecular biology and gives new momentum to QTL population studies. Genetic variation at multiple loci in combination with environmental factors can induce molecular or phenotypic variation. Variation may manifest itself as linear patterns among traits at different levels that can be deconstructed. Correlations can be attributed to detectable QTL and a logical framework based on common and distinct QTL and propagation of biological variation, which can be used to infer network causality, reactivity, or independence [15]. Unexplained biological variation can be used to infer direction between traits that share a common QTL and have no distinct QTL, though it may be difficult to separate biological from technical variation. Prior knowledge and complementary experiments, such as deletion mapping followed by independent gene expression studies between parental lines, may validate or disprove implicated network connections [65].
Evolutionary systems genetics can help dissect the underlying genetics of pathogen susceptibility in plants. Where "evolutionary genetics" describes how evolutionary forces shape biodiversity, as observed in nature, "evolutionary systems genetics" describes how phenotype variation in a population is formed by genotype variation between, for example, host and pathogen involved in an evolutionary arms race.
For purpose of online analysis we created GeneNetwork.org (GN) [7], a free and open-source (FOSS) framework for web-based genetics that can be deployed anywhere. GN allows biologists to upload high-throughput experimental data, such as expression data from microarrays and RNA-seq, and also classical phenotypes, such as disease phenotypes. These phenotypes can be mapped interactively against genotypes using embedded tools, such as R/QTL [10] for model organisms and FaST-LMM [66] and GEMMA [67] which are suitable for human populations and outbred crosses, such as the mouse diversity outcross. Interactive D3 graphics are included from R/qtl charts, and presentation-ready figures can be generated. Recently we have added functionality for phenotype correlation [68], correlation trait loci [16], and network analysis [14]. For examples on using GeneNetwork, see also Mulligan et al. [8].
If you want to know more about eQTL, we suggest the review by Gilad et al. [23], which also discusses eQTL in genome-wide association studies (GWAS), useful in situations where experimental crosses are not available (such as with many pathogens and humans). For further reading on R-gene evolution, we recommend Bakker et al. [34]. For R/qtl analysis, we recommend the R/qtl guide [62] and our MQM tutorial online [63]. For integrating different xQTL methods and causal inference, we recommend Li et al. [15] and Jansen et al. [13].
#### 4 Questions
#### References
bacterial pathogens. Cell Host Microbe 4 (2):96–99
analysis shows major advances in robustness, resolution and inter-lab portability over five microarray platforms. Nucleic Acids Res 36 (21):e141
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Part VI
#### Handling Genomic Data: Resources and Computation
## Chapter 22
#### Semantic Integration and Enrichment of Heterogeneous Biological Databases
#### Ana Claudia Sima, Kurt Stockinger, Tarcisio Mendes de Farias, and Manuel Gil
#### Abstract
Biological databases are growing at an exponential rate, currently being among the major producers of Big Data, almost on par with commercial generators, such as YouTube or Twitter. While traditionally biological databases evolved as independent silos, each purposely built by a different research group in order to answer specific research questions; more recently significant efforts have been made toward integrating these heterogeneous sources into unified data access systems or interoperable systems using the FAIR principles of data sharing. Semantic Web technologies have been key enablers in this process, opening the path for new insights into the unified data, which were not visible at the level of each independent database. In this chapter, we first provide an introduction into two of the most used database models for biological data: relational databases and RDF stores. Next, we discuss ontology-based data integration, which serves to unify and enrich heterogeneous data sources. We present an extensive timeline of milestones in data integration based on Semantic Web technologies in the field of life sciences. Finally, we discuss some of the remaining challenges in making ontology-based data access (OBDA) systems easily accessible to a larger audience. In particular, we introduce natural language search interfaces, which alleviate the need for database users to be familiar with technical query languages. We illustrate the main theoretical concepts of data integration through concrete examples, using two well-known biological databases: a gene expression database, Bgee, and an orthology database, OMA.
Key words Data integration, Ontology-based data access, Knowledge representation, Query processing, Keyword search, Relational databases, RDF stores
#### Abbreviations
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_22, © The Author(s) 2019
#### 1 Introduction
Biological databases have grown exponentially in recent decades, both in number and in size, owing primarily to modern highthroughput sequencing techniques [1]. Today, the field of genomics is almost on par with the major commercial generators of Big Data, such as YouTube or Twitter, with the total amount of genome data doubling approximately every 7 months [2]. While most biological databases have initially evolved as independent silos, each purposely built by a different research group in order to collect data and respond to a specific research question, more recently significant efforts have been made toward integrating the different data sources, with the aim of enabling more powerful insights from the aggregated data, which would not be visible at the level of individual databases.
Let us consider the following example. An evolutionary biologist might want to answer the question "What are the human-rat orthologs, expressed in the liver, that are associated with leukemia?". Getting an answer for this type of question usually requires information from at least three different sources: an orthology database (e.g., OMA [3], OrthoDB [4], or EggNog [5]); a gene expression database, such as Bgee [6]; and a proteomics database containing disease associations (e.g., UniProt [7]). In the lack of a unified access to the three data sources, obtaining this information is a largely manual and time-consuming process. First, the biologist needs to know which databases to search through. Second, depending on the interface provided by these databases, he or she might need to be familiar with a technical query language, such as SQL or SPARQL (note: a list of acronyms is provided at the beginning of this chapter). At the very least, the biologist is required to know the specific identifiers (IDs) and names used by the research group that created the database, in order to search for relevant entries. An integrated view, however, would allow the user to obtain this information automatically, without knowing any of the details regarding the structure of the underlying data sources—nor the type of storage these databases use—and eventually not even specific IDs (such as protein or gene names).
Biological databases are generally characterized by a large heterogeneity, not only in the type of information they store but also in the model of the underlying data store they use—examples include relational databases, file-based stores, graph based, etc. Examples of databases considered fundamental to research in the life sciences can be found in the ELIXIR Europe's Core Data Resources, available online at https://www.elixir-europe.org/platforms/data. In this chapter we will mainly discuss two types of database models: the relational model (i.e., relational databases) and a graph-based data model, RDF (the Resource Description Framework).
Database systems have been around since arguably the same time as computers themselves, serving initially as "digitized" copies of tabular paper forms, for example, in the financial sector, or for managing airline reservations. Relational databases, as well as the mathematical formalism underlying them, namely, the relational algebra, were formalized in the 1970s by E.F. Codd, in a foundational paper that now has surpassed 10,000 citations [8]. The relational model is designed to structure data into so-called tuples, according to a predefined schema. Tuples are stored as rows in tables (also called "relations"). Each table usually defines an entity, such as an object, a class, or a concept, whose instances (the tuples) share the same attributes. Examples of relations are "Gene", "Protein", "Species", etc. The attributes of the relation will represent the columns of the table, for example, "gene name." Furthermore, each row has a unique identifier. The column (or combination of columns) that stores the unique identifier is called a primary key and can be used not only to uniquely identify rows within a table but also to connect data between multiple tables, through a Primary Key-Foreign key relationship. Doing such a connection is called a join. In fact, a join is only one of the operations defined by relational algebra. Other common operations include projection, selection, and others. The operands of relational algebra are the database tables, as well as their attributes, while the operations are expressed through the Structured Query Language (SQL). For a more in-depth discussion on relational algebra, we refer the reader to the original paper by E.F. Codd [8].
This chapter is structured as follows. In Sect. 2, we give a brief introduction to relational databases, through the concrete example of the Bgee gene expression database. We introduce the basics of Semantic Web technologies in Sect. 3. Readers who are already familiar with the Semantic Web stack might skip Sect. 3 and jump directly to Sect. 4, which presents an applied use case of Semantic Web technologies in the life sciences: modeling the Bgee and OMA databases. Section 5 represents the core of this chapter. Here, we present ontology-based data integration (Sect. 5.1) and illustrate it through the concrete example of a unified ontology for Bgee and OMA (Sect. 5.2), as well as the mechanisms required to further extend the integrated system with other heterogeneous sources such as the UniProt protein knowledge base (Sect. 5.3). We introduce natural language interfaces, which enable easy data access even for nontechnical users, in Sect. 5.4. We present an extensive timeline of milestones in data integration based on Semantic Web technologies in the field of life sciences in Sect. 6. Finally, we conclude in Sect. 7.
#### 2 Modeling a Biological Database with Relational Database Technology
In this section we will demonstrate how to model a biological database with relational database technology.
Figure 1 illustrates the data model of a sample extracted from the Bgee database. The sample contains five tables and their relationships, shown as arrows, where the direction of the arrow is oriented from the foreign key of one table to the primary key of a related one. For example, the Primary Key (PK) of the Species table is the SpeciesID. Following the relationships highlighted in bold, we see that the SpeciesID also appears in the two tables connected to Species: GlobalCond and Gene. In these tables, the attribute plays the
Fig. 1 Sample relational database (extracted from the gene expression database Bgee)
role of a Foreign Key (FK). The PK-FK relationships allow combining or aggregating data from related tables. For example, by joining Species and Gene, through the SpeciesID, we can find to which species a gene belongs. Concretely, let's assume we want to find the species where the gene "HBB" can be found. Given that this information is stored in the SpeciesCommonName attribute, we can retrieve it through the following SQL query:
#### SELECT SpeciesCommonName from Species JOIN Gene WHERE Gene.GeneName = 'HBB' and Species.SpeciesID = Gene. SpeciesID
This query enables retrieving (via the "SELECT" keyword) the attribute corresponding to the species name (SpeciesCommon-Name) by joining the Species and Gene tables, based on their primary key-foreign key relationship, namely, via the SpeciesID, on the condition that the GeneName exactly matches "HBB." For a more detailed introduction to the syntax and usage of SQL, we refer the reader to an online introductory tutorial [9], as well as the more comprehensive textbooks [10, 11].
Taking this a step further, we can imagine the case where a second relational database also stores information about genes, but perhaps with some additional data, such as associations with diseases. Can we still combine information across these distinct databases? Indeed, as long as there is a common point between the tables in the two databases, such as the GeneID or the SpeciesID, it is usually possible to combine them into a single, federated database and use SQL to query it through federated joins. An example of using federated databases for biomedical data is presented in [12].
So far, we have seen that relational databases are a mature, highly optimized technology for storing and querying structured data. Also, combined with a powerful and expressive query language, SQL, they allow users to federate (join) data even from different databases.
However, there are certain relationships that are not natural for relational databases. Let us consider the relationship "hasOrtholog". Both the domain and the range of this relationship, as defined in the Orthology Ontology [13], are the same—a gene. For example, the hemoglobin (HBB) gene in human has the Hbb-bt orthologous gene in the mouse (expressed via the relation hasOrtholog). In the relational database world, this translates into a so-called selfjoin. As the name suggests, this requires joining one table—in this case, Gene—with itself, in order to retrieve the answer. These types of "self-join" relations, while frequent in the real world (e.g., a manager of an employee is also an employee, a friend of a person is also a person, etc.), are inefficient in the context of relational databases. While there are sometimes ways to avoid self-joins, these
2.1 Limitations of Relational Databases and Emerging Solutions for Data Integration
require even more advanced SQL fluency on the part of the programmer [14].
Moreover, relational databases are typically not well-suited for applications that require frequent schema changes. Hence, NoSQL stores have gained widespread popularity as an alternative to traditional relational database management systems [15–17]. These systems do not impose a strict schema on the data and are therefore more flexible than relational databases in the cases where the structure of the data is likely to change over time. In particular, graph databases, such as Virtuoso [18], are very well suited for data integration, as they allow easily combining multiple data sources into a single graph. We discuss this in more detail in Sect. 3.
These and other considerations have led to the vision of the Semantic Web, formalized in 2001 by Tim Berners Lee et al. [19]. At a high-level, the Semantic Web allows representing the semantics of data in a structured, easy to interlink, machinereadable way, typically by use of the Resource Description Framework (RDF)—a graph-based data model. The gradual adoption of RDF stores, although widespread in the Web context and in the life sciences in particular, did not replace relational databases altogether, which lead to a new challenge: how will these heterogeneous data sources now be integrated?
Initial integration approaches in the field of biological databases have been largely manual: first, many of them (either relational or graph-based) have included cross-references to other sources. For example, UniProt contains links to more than 160 other databases. However, this raises a question for the user: which of the provided links should be followed in order to find relevant connections? While a user can be assumed to know the contents of a few related databases, we can hardly expect anyone to be familiar with more than 160 of them! To avoid this problem, other databases have chosen an orthogonal approach: instead of referencing links to other sources, simply copy the relevant data from those sources into the database. This approach also has a few drawbacks. First, it generates redundant data (which might result in significant storage space consumption), and, most importantly, it might lead to the use of stale, outdated results. Moreover, this approach is contradictory to best practices of data warehousing used widely across various domains in industry. For a discussion on this, we refer the reader to [20].
Databases such as UniProt are highly comprehensive, with new results being added to each release, results that may sometimes even contradict previous results. Duplication of this data into another database can quickly lead to missing out the most recent information or to high maintenance efforts required to keep up with the new changes. In the following sections, we discuss an alternative approach: integrating heterogeneous data sources through the use of a unifying data integration layer, namely, an integrative ontology, that aligns, but also enriches the existing data, with the purpose of facilitating knowledge discovery.
Throughout the remainder of this chapter, we will combine theoretical aspects of data integration with concrete examples, based on our SODA project [21], as well as from our ongoing research project, Bio-SODA [22], where we are currently building an integrated data access system for biological databases (starting with OMA and Bgee), using a natural language search interface. In the context of this project, Semantic Web technologies, such as RDF, are used to enhance interoperability among heterogeneous databases at the semantic level (e.g., RDF graphs with predefined semantics). Moreover, currently, several life science and biomedical databases such as OMA [3], UniProt [7], neXtProt [22], the European Bioinformatics Institute (EMBL-EBI) RDF data [24], and the WorldWide Protein Data Bank [25] already provide RDF data access, which also justifies an RDF-based approach to enable further integration efforts to include these databases. A recent initiative for (biological) data sharing is based on the FAIR principles [26], aiming to make data findable, accessible, interoperable, and re-usable.
#### 3 Semantic Web Technologies
The Semantic Web, as its name shows, emerged mainly as a means to attach semantics (meaning) to data on the Web [19]. In contrast to relational databases, Semantic Web technologies rely on a graph data model, in order to enable interlinking data from disparate sources available on the Web. Although the vision of the Semantic Web still remains an ideal, many large datasets are currently published based on the Linked Data principles [27] using Semantic Web technologies (e.g., RDF). The Linked Open Data Cloud illustrates a collection of a large number of different resources including DBPedia, UniProt, and many others.
In this section, we will describe the Semantic Web (SW) stack, focusing on the technologies that enhance data integration and enrichment. For a more complete description of the SW stack, we refer the reader to the comprehensive introductions in [28–30].
The Semantic Web stack is presented in Fig. 2. We will focus on the following standards or layers of the stack: URI, the syntax layer (e.g., Turtle (TTL), an RDF serialization format), RDF, OWL, RDFS, and SPARQL. These layers are highlighted in gray in Fig. 2.
3.1 Unique Resource Identifier (URI) A Uniform Resource Identifier (URI) is a character sequence that identifies an abstract or physical resource. A URI is classified as a locator, a name, or both. The Uniform Resource Locators (URLs) are a subset of URIs that, in addition to identifying a resource, provide a means of locating the resource by describing its primary
Fig. 2 The Semantic Web stack modified from [31]
Fig. 3 An example of a UniProt URI with a fragment
access or network "location." For example, https://bgee.org is a URI that identifies a resource (i.e., the Bgee gene expression website), and it implies solely a representation of this resource (i.e., an HTML Web page). This resource is accessible through the HTTPS protocol.
The Uniform Resource Name (URN) is also a URI that refers to both the "urn" scheme [32], which are URIs required to remain globally unique and persistent even when the resource does not exist anymore or becomes unavailable, and to any other URI with the properties of a name. For example, the URN urn:isbn:978-1- 61779-581-7 is a URI that refers to a previous edition of this book by using the International Standard Book Number (ISBN). However, no information about the location and how to get this resource (book) is provided.
The URI syntax consists of a hierarchical sequence of components referred to as the scheme, authority, path, query, and fragment [33]. Figure 3 describes a UniProt URI that includes these components.
An individual scheme does not have to be classified as being just one of "name" or "locator." Instances of URIs from any given scheme may have the characteristics of names (URN) or locators
Fig. 4 An RDF graph with two nodes (subject and object) and an edge connecting them (predicate)
(URL) or both (URN + URL). Further examples of URIs with variations in their syntax components are:
#### 3.2 Resource Description Framework (RDF)
The Resource Description Framework (RDF) is a framework for describing information about resources in the World Wide Web, which are identified with URIs. In the previous section, we have seen that data in relational databases is organized into tables, according to some predefined schema. In contrast, in RDF stores, data is mainly organized into triples, namely, <subject, predicate, object>, similarly to how sentences in natural language are structured. An informal example would be: <Bob, isFriendOf, Alice>. A primer on triples and the RDF data model, using this simple example, is available online [34]. Figure 4 illustrates the RDF triple: the subject represents the resource being described, the predicate is a property of that resource, and finally the object is the value of the property (i.e., an attribute of the subject).
Triples can be defined using the RDF. The data store for RDF data is also called a "triple store." Moreover, in analogy to the data model (or the schema) of a relational database, the high-level structure of data in a triple store can be described using an ontology. According to Studer et al. [35], an ontology is a formal, explicit specification of a shared conceptualization. "Formal" refers to the fact that the expressions must be machine readable: hence, natural language is excluded. In this context, we can mention description logic (DL)-based languages [36], such as OWL 2 DL (see Sect. 3.3 for further details) to define ontologies. A DL ontology is the equivalent of a knowledge base (KB). A KB is mainly composed of two components that describe different statements in ontologies: the terminological box (TBox, i.e., the schema) and the assertional box (ABox, i.e., the data). Therefore, the conceptual statements form the set of TBox axioms, whereas the instance level statements form the set of ABox assertions. To exemplify this, we can mention the following DL axioms: Man -Human u Male (a TBox axiom that states a man is a human and male) and john: Man (an ABox assertion that states john is an instance of man).
Given that one of the goals of the Semantic Web is to assign unambiguous names to resources (URIs), an ontology should be more than a simple description of data in a particular triple store. Rather, it should more generally serve as a description of a domain, for instance, genomics (see Gene Ontology [37]) or orthology (see Orth Ontology [13]). Different instantiations of this domain, for example, by different research groups, should reuse and extend this ontology. Therefore, constructing good ontologies requires careful consideration and agreement between domain specialists, with the goal of formally representing knowledge in their field. As a consequence, ontologies are usually defined in the scope of consortiums—such as the Gene Ontology Consortium [38] or the Quest for Orthologs Consortium [39]. A notable collaborative effort is the Open Biological and Biomedical Ontology (OBO) Foundry [40]. It established principles for ontology development and evolution, with the aim of maximizing cross-ontology coordination and interoperability, and provides a repository of life science ontologies, currently, including about 140 ontologies.
To give an example of RDF data in a concrete life sciences use case, let us consider the following RDF triples, which illustrate a few of the assertions used in the OMA orthology database to describe the human hemoglobin protein ("HBB"), using the first version of the ORTH ontology [13]:
oma:PROTEIN\_HUMAN04027 rdf:type orth:Protein. oma:PROTEIN\_HUMAN04027 oma:geneName "HBB". oma:PROTEIN\_HUMAN04027 biositemap:description "Hemoglobin subunit beta". oma:PROTEIN\_HUMAN04027 obo:RO\_0002162 <http://www.uniprot. org/taxonomy/9606>.
This simple example already illustrates most of the basics of RDF. The instance that is being defined—the HBB protein in human—has the following URI in the OMA RDF store: http:// omabrowser.org/ontology/oma#PROTEIN\_HUMAN04027
The URI is composed of the OMA prefix, http://omabrowser. org/ontology/oma# (abbreviated here as "oma:"), and a fragment identifier, PROTEIN\_HUMAN04027. The first triple describes the type of this resource—namely, an orth:Protein—based on the Orthology Ontology, prefixed here as "orth:," http://purl.org/ net/orth#. As mentioned previously, this is a higher-level ontology, which OMA reuses and instantiates. It is important to note that other ontologies are used as well in the remaining assertions: for example, the last triple references the UniProt taxonomy ID 9606. This is based on the National Center for Biotechnology Information (NCBI) organismal taxonomy [41]. If we follow the link in a Web browser, we see that it identifies the "Homo sapiens" species, while the property obo:RO\_0002162 (i.e., http://purl.obolibrary. org/obo/RO\_0002162) simply denotes "in taxon" in OBO [40]. Lastly, the concept also has a human-readable description, "Hemoglobin subunit beta."
3.3 RDF Schema (RDFS) RDF Schema (RDFS) provides a vocabulary for modeling RDF data and is a semantic extension of RDF. It provides mechanisms for describing groups (i.e., classes) of related resources and the relationships between these resources. The RDFS is defined in RDF. The RDFS terms are used to define attributes of other resources such as the domains (rdfs:domain) and ranges (rdfs: range) of properties. Moreover, the RDFS core vocabulary is defined in a namespace informally called rdfs here, and it is conventionally associated with the prefix rdfs:. That namespace is identified by the URI http://www.w3.org/2000/01/rdf-schema#.
> In this section, we will mostly focus on the RDF and RDFS terms used in this chapter. Further information about RDF/RDFS terms is available in [42].
- rdfs:Resource—all things described by RDF are called resources, which are instances of the class rdfs:Resource (i.e., rdfs:Resource is an instance of rdfs:Class).
- rdfs:Class is the class of resources that are RDF classes. Resources that have properties (attributes) in common may be divided into classes. The members of a class are instances.
- rdf:Property is a relation between subject and object resources, i.e., a predicate. It is the class of RDF properties.
- rdfs:Literal is the class of literal values such as textual strings and integers. rdfs:Literal is a subclass of rdfs:Resource.
- rdfs:range is an instance of rdf:Property. It is used to state that the values of a property are instances of one or more classes. For example, orth:hasHomolog rdfs:range orth:SequenceUnit (see Fig. 5a). This statement means that the values of orth:hasHomolog property can only be instances of orth: SequenceUnit class.
- rdfs:domain is an instance of rdf:Property. It is used to state that any resource that has a given property is an instance of one or more classes. For example, orth:hasHomolog rdfs: domain orth:SequenceUnit (see Fig. 5b). This statement means that resources that assert the orth:hasHomolog property must be instances of orth:SequenceUnit class.
- rdf:type is an rdf:Property that is used to state that a resource is an instance of a class.
Fig. 5 Examples of RDF/RDFS statements
3.4 Web Ontology Language (OWL) The first level above RDF/RDFS in the Semantic Web stack (see Fig. 2) is an ontology language that can formally describe the meaning of resources. If machines are expected to perform useful reasoning tasks on RDF data, the language must go beyond the basic semantics of RDF Schema [43]. Because of this, OWL and OWL 2 (i.e., Web Ontology languages) include more terms for describing properties and classes, such as relations between classes (e.g., disjointness, owl:disjointWith), cardinality (e.g., "exactly 2," owl:cardinality), equality (i.e., owl:equivalentClass), richer typing of properties, characteristics of properties (e.g., symmetry, owl: SymmetricProperty), and enumerated classes (i.e., owl:oneOf). The owl: prefix replaces the following URI namespace: http://www.w3. org/2002/07/owl#.
As a full description of OWL and OWL 2 is beyond the scope of this chapter, we refer the interested reader to [44, 45]. In the following, we focus solely on some essential modeling features that the OWL languages offer in addition to RDF/RDFS vocabularies.
<sup>l</sup> owl:Class is a subclass of rdfs:Class. Like rdfs:Class, an owl:Class groups instances that share common properties. However, this new OWL term is defined due to the restrictions on DL-based
Fig. 6 Examples of instances of orth:SequenceUnit and orth:Gene and object and datatype property assertions
Further information about OWL languages are available as World Wide Web Consortium (W3C) recommendations in [46] and [47].
3.5 RDF Serialization Formats RDF is a graph-based data model which provides a grammar for its syntax. Using this grammar, RDF syntax can be written in various concrete formats which are called RDF serialization formats. For example, we can mention the following formats: Turtle [48], RDF/XML (an XML syntax for RDF) [49], and JSON-LD (a JSON syntax for RDF) [50]. In this section, we will solely focus on the Turtle format.
> Turtle language (TTL) allows for writing an RDF graph in a compact textual form. To exemplify this serialization format, let us consider the following turtle document that defines the homologous and orthologous relations:
```
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix owl: <http://www.w3.org/2002/07/owl#> .
@prefix orth: <http://purl.org/net/orth#> .
# http://purl.org/net/orth#SequenceUnit
orth:SequenceUnit rdf:type owl:Class .
orth:hasHomolog rdf:type owl:ObjectProperty ;
rdf:type owl:SymmetricProperty ;
rdfs:domain orth:SequenceUnit ;
rdfs:range orth:SequenceUnit .
orth:hasOrtholog rdf:type owl:ObjectProperty ;
rdfs:subPropertyOf orth:hasHomolog .
```
This example introduces many of features of the Turtle language: @prefix and prefixed names (e.g., @prefix rdfs: http://www.w3.org/2000/01/rdf-schema#), predicate lists separated by ";" (e.g., orth:hasOrtholog rdf:type owl: ObjectProperty; rdfs:subPropertyOf orth:hasHomolog.), comments prefixed with "#" (e.g., # http://purl.org/net/ orth#SequenceUnit), and a simple triple where the subject, predicate, and object are separated by white spaces and ended with a "." (e.g., orth:SequenceUnit rdf:type owl:Class).
Further details about TTL serialization are available as a W3C recommendation in [48]
#### 3.6 Querying the Semantic Web with SPARQL
Once we have defined the knowledge base (TBox and ABox), how can we use it to retrieve relevant data? Similar to SQL for relational databases, data in RDF stores can be accessed by using a query language. One of the main RDF query languages, especially used in the field of life sciences, is SPARQL [51]. A SPARQL query essentially consists of a graph pattern, namely, conjunctive RDF triples, where the values that should be retrieved (the unknowns either subjects, predicates, or objects) are replaced by variable names, prefixed by "?". Looking again at the previous example, if we want to get the description of the "HBB" protein from OMA, we would simply use a graph pattern, where the value of the "description" the one we want to retrieve—is replaced by a variable as follows:
```
SELECT ?description WHERE {
?protein oma:geneName "HBB".
?protein biositemap:description ?description.
}
```
The choice of variable name itself is not important (we could have used "?x", "?var", etc., albeit with a loss of readability). Essentially, we are interested in the description of a protein about which we only know a name—"HBB."
In order to get a sense of how large bioinformatics databases currently are, but also to get a hands-on introduction into how they can be queried using SPARQL, we propose to retrieve the total number of proteins in UniProt in Exercise A at the end of this chapter. Furthermore, Exercise C will allow trying out and refining the OMA query introduced above, but also writing a new one, using the OMA SPARQL endpoint.
#### 4 Modeling Biological Databases with Semantic Web Technologies
In this section we show a concrete example of how we can use Semantic Web technologies to model the two biology databases Bgee and OMA.
Figure 7 illustrates a fragment of a candidate ontology describing the relational database sample from Bgee (see Fig. 1). The ellipses illustrate classes of the ontology, either specific to the Bgee ontology, such as AnatomicEntity (the equivalent of the anatEntity table in the relational view), or classes from imported ontologies, such as the Taxon class (the prefix "up:" denoting the UniProt ontology, http://purl.uniprot.org/core/). The advantage of using external (i.e., imported) classes is that integration with other databases which also instantiate these classes will be much simpler. For example, we will see that the class Gene serves as the "join point" between OMA and Bgee. Arrows define properties of the ontology: either datatype properties (similar to attributes of a table in the relational world), such as the speciesName or the stage-Name, or object properties, which are similar to primary key-foreign key relationships, given that they link instances of one class to those of another. If we compare Fig. 7 (the ontology view) against Fig. 1 (the relational view), we notice that the object properties isExpressedIn and isAbsentIn only appear explicitly in the ontology. This is because the values of these properties will actually be calculated on-the-fly, from multiple attributes in the relational database. Given that Bgee is mainly used to query gene expressions, these properties are exposed as new semantic properties in the domain ontology, namely, expression or absence of expression of a gene in a particular anatomic entity. This is one of the means through which the semantic layer can not only describe but also enrich the data available in the underlying layers (in this case, in the relational database). The domain of both the isExpressedIn and isAbsentIn properties is in this case a gene, while the range is an anatomic entity, such that triples that instantiate this relationship will have the structure: <Gene, isExpressedIn, AnatomicEntity>.
Given that the OMA ontology is significantly larger than the one for Bgee, we only show here the class hierarchy in Fig. 8. The
Fig. 7 A portion of the ontology defined over the relational database sample from Bgee. For readability purposes, we omitted the namespace ("bgee:") for the ontology properties
Fig. 8 The class hierarchy of the OMA ontology. Ellipses indicate class labels, while arrows indicate the "rdfs: subClassOf " property. Further details are available in [13]
most important concepts in the ontology are shown in the top right corner, namely, the cluster of orthologs and the cluster of paralogs, which store information about gene orthology (or paralogy) in a hierarchical tree structure (the gene-tree node). Similarly to the Bgee ontology, the Gene class in OMA is external. Arrows indicate the "rdfs:subClassOf" relationship—for example, both the "Cluster of Orthologs" and the "Cluster of Paralogs" classes—are subclasses of the "Cluster of Homologs" class. For a description of the ontology, as well as a discussion regarding its design within the Quest for Orthologs Consortium, we point the reader to [13]. Furthermore, the ontology can be explored or visualized in Web-VOWL [52] using the Web page of the OMA SPARQL endpoint [53] available online at https://sparql.omabrowser.org/sparql.
Until here we have explored a few relatively simple examples in order to get familiar with the basics of Semantic Web technologies (URIs, RDF triples, and SPARQL). However, we can now introduce a more complex query that will better illustrate the expressivity of the SPARQL query language for accessing RDF stores—that is, for integrating and joining data across different databases.
Since all RDF stores structure data using the same standard model for data interchange, the main requirements in order to efficiently join multiple sources are:
This is the reason why already today we can jointly query, for example, OMA and UniProt—essentially, integrating the two databases by means of executing a federated SPARQL query.
To illustrate this, let us consider the following example: what are the human genes available in the OMA database that have a known association with leukemia? OMA does not contain any information related to diseases, however, UniProt does. In this case, since OMA already cross-references UniProt with the oma: xrefUniprot property, we can write the following federated SPARQL query, which will be running at the OMA SPARQL endpoint:
```
select distinct ?proteinOMA ?proteinUniProt
where {
service <http://sparql.uniprot.org/sparql> {
?proteinUniProt a up:Protein .
?proteinUniProt up:organism taxon:9606 . # Homo Sapiens
?proteinUniProt up:annotation ?annotation . # annotations of this protein
entry
?annotation rdfs:comment ?text
filter( regex(str(?text), "leukemia") ) # only those containing the
text "leukemia"
}
?proteinOMA a orth:Protein.
?proteinOMA oma:xrefUniprot ?proteinUniProt.
}
```
We skip the details regarding the prefixes used in the example and focus on the new elements in the query. The main part to point out is the "service <http://sparql.uniprot.org/sparql>" block, delimited between the inner brackets. This enables using the SPARQL endpoint of UniProt remotely, as a service. Through this mechanism, the query will first fetch from UniProt all instances of proteins that are annotated with a text that contains "leukemia" (this is achieved by the filter keyword in the service block). Then, using the cross-reference oma:xrefUniprot property, the query will return all the equivalent entries from OMA. From here, the user can explore, either in the OMA browser or by further refining the SPARQL query, other properties of these proteins: for example, their orthologs in a given species available in the database. In Exercise D at the end of this chapter, we encourage the reader to try this out in the OMA SPARQL endpoint. Note that the same results can be obtained by writing this query in the UniProt SPARQL endpoint and referencing the OMA one as a service. For an overview of federation techniques for RDF data, we refer the reader to the survey [54].
The mechanisms illustrated so far, while indeed powerful for federating distinct databases, have a major drawback: they require the user to know the schema of the databases (otherwise, how would we know which properties to query in the previous examples?), and, more importantly, they require all users to be familiar with a technical query language, such as SPARQL. While very expressive, formulating such queries can quickly become overwhelming for non-programmer users. In the following, we will look at techniques that aim to overcome these limitations.
#### 5 Ontology-Based Integration of Heterogeneous Data Stores
So far we have seen some of the alternatives available for storing biological data—relational databases and triple stores. In this section, we look at how these heterogeneous sources can be integrated and accessed in a unified, user-friendly manner that does not require knowledge of the location or structure of the underlying data nor of the technical language (SQL or SPARQL) used to retrieve the data. The architecture we present is inspired by work presented in [21], which focused strictly on keyword search in relational databases.
The main four layers of an integrated data access system, as shown in Fig. 9, are:
Fig. 9 Integrated data access system
which identifiers from one data model correspond to which ones from the others. In the case of biological databases, this is usually done by taking into account cross-references, which already exist between most databases, as we have seen in the SPARQL query in Sect. 5.
While the data model layer can be seen as a local ontology, the integration layer will serve as a global ontology. The integration layer can be queried using, for example, SPARQL. However, in order to get the results from the underlying sources, the SPARQL query needs be translated in the native query languages of the underlying sources (e.g., SQL for relational databases). This is achieved by using the mappings defined in the global ontology. For example, the keyword "expressed in" does not have a direct correspondence in Bgee, but it can be translated into an SQL procedure (in technical terms, it represents an SQL view of the data). Without going into details, at a high level, the property "gene A expressed in anatomic entity B" will be computed by looking at the number of experiments stored in the database, showing the expression of A in B. It is conceivable that in another database, which could also form part of the integrated system, this information is available explicitly. In this case the mapping would simply be a 1-to-1 correspondence to the property value stored in the database. The role of the integration layer is to capture all the occurrences where a certain concept (entity or property) can be found, along with a mapping for each of the occurrences, defining how information about this concept can be computed from the base data.
To summarize, the integration layer abstracts away the location and structure of data in the underlying sources, providing users a unified access through a global ontology. One of the drawbacks of this approach is that, in the lack of a presentation layer, such as a user-friendly query interface (e.g., a visual query builder or a keyword-based search interface), the data represented in the global ontology is accessible mainly through a technical query language, such as SPARQL. Therefore, in order to be able to access the data, users are required to become fluent in the respective query language.
It is worth at this point mentioning that most data integration systems available at the time of this writing only offer the three layers presented so far. Examples of such systems, generically denoted as ontology-based data access (OBDA) systems, are Ontop [55], Ultrawrap [56], or D2RQ [57].
5.1.4 Presentation Layer The three layers presented so far already achieve data integration, but with a significant drawback, which is that the user is required to know a technical query language, such as SPARQL. The role of the presentation layer is to expose data from all integrated resources in an easy to access, user-friendly manner. The presentation layer abstracts away the structure of the integration layer and exposes data through a search interface that users (including non-programmers) are familiar with, such as keyword search [21, 58] or even full natural language search [59, 60].
The challenges in building the presentation layer are manyfold: first, human language is inherently ambiguous. As an example, let us assume a user asks: "Is the HBB gene expressed in the blood?" What does the user mean? The hemoglobin gene (HBB) in general? Or just in the human? The system should be proactive in helping the user clarify the semantics or intents of the question, before trying to compute the underlying SPARQL query. Second, the presentation layer should provide not only raw results but also an explanation—for example, what sources were queried, how many items from each source have been processed in order to generate the response, etc. This enables the user to validate the generated results or to otherwise continue refining the question. Third, the presentation layer must also rank the results according to some relevance metric, similarly to how search results are scored in Web search engines. Given that the number of results retrieved from the underlying sources can easily become overwhelming (e.g., searching for "HBB" in Bgee returns over 200 results), it is important that the most relevant ones are shown first.
From a technical point of view, the presentation layer maintains an index (i.e., the vocabulary) of all keywords stored in the lower layers, both data and metadata (descriptions, labels, etc.), such that each keyword in a user query can be mapped to existing data in the lower layers. An important observation is that the presentation layer highly relies on the quality of the annotations available in the lower layers. In the lack of human-readable labels and descriptions in the global ontology, the vocabulary collected by the presentation layer will miss useful terms that the user might search for. One way to detect and fix this problem is to always log user queries and improve the quality of the annotations "on demand," whenever the queries cannot be solved due to missing items in the vocabulary. For a more extended discussion on the topic of labels and their role in the Semantic Web, refer to [61].
Finally, it is worth noting that none of these layers need to be centralized—indeed, even in the case of the integration layer, although its role is to build a common view of all data in the physical storage, it can be distributed across multiple machines, just as long as the presentation layer knows which machine holds which part of the unified view.
5.2 A Concrete Example: A Global Ontology to Unify OMA and Bgee
So far we have seen an abstract view of a system for data integration across heterogeneous databases. It is time to look at how this translates into a real-world example, using the Bgee relational database and the OMA RDF database.
Fig. 10 A sample global ontology for integrating OMA and Bgee and an example assertion
The top part of Fig. 10, the terminological box, illustrates part of the global ontology (layer 3, integration layer) for the two databases, with most of the terms being part of OMA, except for Anatomic Entity, which is specific to Bgee. As mentioned previously, OMA extends the ORTH ontology, which is why the corresponding terms in the ontology are prefixed with "orth:." The Gene concept can actually be found in both Bgee and OMA; therefore the global ontology will define mappings to both sources. As we can see in the ontology, the Gene is the common point that joins together OMA and Bgee. The gene IDs used in both databases are Ensembl IDs [62], stored in the ensemblGeneId string property. For example, the human hemoglobin gene, "HBB," which we previously showed as an example entry in OMA, corresponds to the ENSG00000244734 Ensemble ID and can also be found in Bgee.
The lower part of Fig. 10, the assertional box, illustrates an example assertion—in this case, that the protein HUMAN22168 in OMA is orthologous to the protein HORSE13872 and that, furthermore, this protein is encoded by the gene with the Ensemble ID ENSG000001639936. Moreover, this gene is expressed in the brain (the Uberon ID for this being "UBERON:0000955"). The human-readable description is stored in the String literal label—as, for example, the name of the anatomic entity, "brain," shown in the bottom-right corner in the figure. Without labels, much of the available data would not be easily searchable by a human user nor by an information retrieval system.
Note that with this sample ontology, we can already answer questions related to orthology and gene expression jointly, such as the first part of our introductory query: "What are the human-rat orthologs, expressed in the liver...?". This question essentially refers to pairs of orthologous Genes (those in human and rat) and their expression in a given Anatomic Entity (the liver). Apart from the Species class, which is not explicitly shown, all of the information is already captured by the ontology in Fig. 10. A similar mechanism can be used to further extend this to UniProt (for instance, based again on gene IDs as the "join point," or by using existing crossreferences, as we have shown in the previous section), therefore enabling users to ask even more complex queries.
One of the main challenges in implementing technologies for the Semantic Web was recognized from early on (see the study published in 2001 by Calvanese et al. [63]) to be the problem of integrating heterogeneous sources. In particular, one of the observations made was that integrating legacy data will not be feasible through a simple 1-to-1 mapping of the underlying sources into an integrative ontology (e.g., mapping all attributes of tables in relational databases to properties of classes in an ontology), but rather through more complex transformations, that map views of the data into elements of the global ontology [63].
> To illustrate this with a concrete example, let us consider again the unified ontology for OMA and Bgee that we introduced in the previous section. Although Figure 10 shows properties such as "gene isExpressedIn" or "gene hasOrtholog," this data is actually not explicitly stored in the underlying databases but rather needs to be computed on-the-fly based on the available data. For example, the "isExpressedIn" property can be computed based on the number of experiments which show the expression of a gene in a certain anatomic entity in Bgee. Deciding the exact threshold for when a gene is considered as "expressed" according to the data available is not straightforward and needs to be agreed upon by domain specialists. Therefore, the integration layer will also serve to enrich the data available in the underlying layers, by defining new concepts
#### 5.3 How to Link a Database with an Ontology?
based on this data (e.g., the presence or absence of gene expression in an anatomic entity).
At this point it is worth clarifying an important question: why are mappings necessary? Why is it not enough to replicate the data in the different underlying formats into a single, uniform way (e.g., translate all RDB data into RDF)? The answer is that not only would such a translation require a lot of engineering effort, but more importantly, it would transform the data from a format that is highly optimized for data access, into a format that is optimized for different purposes (data integration and reasoning). Querying relational databases still is, today, the most efficient means of accessing very large quantities of structured data. Transforming all of it into RDF would in many cases mean downgrading the overall performance of the system. In some cases storing RDF data in the relational format was proven to be more efficient [64].
So how are mappings then created? One of the main mechanisms to achieve this is currently the W3C standard R2RML, available as a W3C recommendation online [65]. R2RML enables mapping relational data to the RDF model, as chosen by the programmer. For a concrete example of how mappings can be defined and what are the advantages of this approach, we refer the reader to [66]. A mapping essentially defines a view of the data, which is a query (in this case, an SQL query) that allows retrieving a relevant portion of the underlying data, in order to answer a higherlevel question (e.g., what is "expressed in"?). The materialization of this query (the answer) will be returned in RDF format, on demand, according to the mapping. This avoids duplicating or translating data in advance from the underlying relational database into RDF until it is really needed, in order to answer a user query.
For a discussion regarding the limitations of R2RML and alternative approaches to define mappings from relational data to RDF, we refer the reader to the survey [67].
5.4 Putting Things Together So far we have seen how individual sources can be represented into a single, unified ontology, and we had a high-level view of a data access system that enables users to ask queries and get responses in a unified way, without knowledge of where data is located or how it is structured. In this section we finally look at how all of these components can work together in answering natural language queries on biological databases. Although there are multiple alternatives to natural language interfaces, including visual query interfaces or keyword-based search interfaces, it has been shown that natural language interfaces are the most appropriate means to query Semantic Web data for non-technical end-users [68]. As a consequence, natural language querying, based on Semantic Web technologies, is currently one of the active areas of research, examples of recent systems implementing an ontology-based natural language interface including the Athena [59] and TRDiscover [60] systems.
First, recall the user question we formulated in the beginning of this chapter: "What are the human-rat orthologs, expressed in the liver, that are associated with leukemia?" Let us assume the resources at hand to answer this question are the biological databases OMA, Bgee, and UniProt. The four main steps required to translate the natural language question into the underlying query languages of OMA, Bgee, and UniProt will be:
(a) Identify entities in the query
This is the natural language processing step that extracts the main concepts the user is interested in, based on the keywords of the input query: orthologs, human, rat, expressed, liver, associated, and leukemia.
(b) Identify matches of the entities in the integrative ontology
The extracted keywords will be searched for in the vocabulary of the presentation layer, resulting in one or multiple URIs, given that a keyword can match multiple concepts. For example, the keyword "orthologs" can match either the entity "OrthologCluster" or the property "hasOrtholog" of a gene in OMA. The index of the presentation layer will also return the location the URI originates from (OMA or Bgee or UniProt).
(c) Construct subqueries for each of the matches
The extracted URIs will be used to construct subqueries on each of the underlying data sources. This step requires translating the original query into the native language of each underlying database, with specific mechanisms for each type of database (relational or triple store). At a high level, the translation process involves finding the minimal sub-schema (or subgraph in the case of RDF data) that covers all the keywords matched from the input query. Taking the example previously shown in Fig. 10, the minimal subgraph that contains "orthologs" and "expressed" will essentially contain only two nodes of the entire graph: Gene (which is both the domain and the range of the "hasOrtholog" property in the Orthology Ontology) and AnatomicEntity (which is the range of the "isExpressedIn" property in the Bgee ontology). All the unknowns of the query (e.g., which ortholog genes) are replaced by variables. The final subqueries for OMA and Bgee might therefore (informally) look like this:
```
OMA: select ?gene1 ?gene2 where {
?protein1 a Protein.
?protein1 inTaxon "Homo sapiens".
?protein1 isEncodedBy ?gene1.
?protein1 hasOrtholog ?protein2.
?protein2 inTaxon "Rattus norvegicus".
?protein2 isEncodedBy ?gene2.
}
```
Note that we have simplified the actual query for readability purposes (using the literals "Homo sapiens" and "Rattus norvegicus" instead of their corresponding URIs). This subquery will cover the keywords: ortholog, human, and rat. Notice that the query should return genes, not proteins, because the join point between OMA and Bgee is the Gene class.
```
Bgee: select ?gene where {
?gene a Gene.
?gene isExpressedIn ?anatomicEntity.
?anatomicEntity rdfs:label "liver".
}
```
This subquery will therefore cover the expressed and liver keywords. The final step will be then to get the similar subquery for UniProt (which we omit here for brevity) and to compute the joint result, namely, the intersection between all the sets returned by the subqueries.
(d) Join the results from each of the subqueries
This final step is essential in keeping the performance of the system to an acceptable level. Joining (federating) the results of several subqueries into a unified result is not an easy task and requires a careful ordering of the operations from all subqueries. To understand this problem, let us consider again our example and try to see how many results each of the subqueries will return. First, if we take a look at the OMA browser and try to find all orthologs between human and rat, this will amount to more than 21,000 results. However, is the user really interested in all of them? Certainly not, as the input query shows—the user is only interested in a small fraction of the orthologs, namely, those that are expressed in the liver and have an association with leukemia (according to the data stored in Bgee and UniProt). How many are these? If we now refer to UniProt and look for the disease leukemia, we will find that there are only 20 entries which illustrate the association with this disease. Clearly, getting only the orthologs of these 20 entries will be much more efficient than retrieving all 21,000 pairs from OMA first and then removing most of them to only keep relevant ones.
However, note that in this case, we only know this information because we constructed the queries and tried them out by hand first. How should the system estimate the number of results (i.e., the cardinality of each subquery) in advance? This question has been an active area of research for a long time. Some of the methods used to tackle this problem are either to precompute statistics regarding the number of results available in different tables of the underlying sources [69] or to use statistics regarding previously asked queries to optimize the new ones, for example, via statistical machine learning [70]. In the first case, we would, for instance, store the individual counts of different orthologous pairs while also keeping statistics about diseases if we expect these types of questions to be asked frequently, whereas in the second case, we would simply look at the number of results similar subqueries generated in the past, to optimize which results to fetch first. For a recent study of optimization methods for federated SPARQL queries, see [71].
(e) Present the user the final results
Finally, the joined results are returned to the user, along with an explanation regarding the constructed query and the entities that were matched in order to construct it. In this way, the user has the opportunity to validate the correctness of the answer or otherwise to further refine the question.
For a more in-depth discussion regarding natural language query interfaces in ontology-based data access systems, we refer the reader to Athena [59] and TRDiscover [60].
#### 6 Timeline of Semantic Web Technologies and Ontology-Based Data Integration in Life Sciences
The field of life sciences has been an early adopter of Semantic Web technologies, due to the need of interoperability and integration of biological data spread across different databases. In this section, we provide a brief timeline (see Fig. 11), including the example ontologies introduced in this chapter.
Fig. 11 A selective timeline of data integration efforts in life sciences
comprehensive resource of computable knowledge regarding gene functions and products.
published in RDF, and since 2013 a SPARQL endpoint is provided [78].
ontology). Updates continue to be provided for increased interoperability among bioinformatics databases [87, 88].
#### 7 Conclusions and Outlook
Data integration is arguably one of the most important enablers of new scientific discoveries, given that research data is currently growing at an unprecedented rate. This is especially true in the case of biological databases. While data integration poses many challenges, the emergence of standards, integrative ontologies, as well as the availability of cross-references between many of the biological databases make the problem easier to tackle. This chapter has provided a brief introduction to the methods that can be used to integrate heterogeneous databases using Semantic Web technologies while also providing a concrete example of achieving this goal for three well-known existing biological databases: OMA, Bgee, and UniProt.
Although there would be many more aspects to cover and much of the work for achieving wide-scale data integration still remains to be done, we would like to end this chapter by reinforcing the following conclusion, extracted from a study of Biological Ontologies for Biodiversity Knowledge Discovery [91]:
We hope that current work will spur interest and feedback from scientists and bioinformaticians who see data integration, interoperability, and reuse as the solution to bringing the past 300 years of biological exploration of the planet into currency for science and society.
#### 8 Exercises
A. Querying UniProt with SPARQL
The goal of this warm-up exercise is to get familiar with a SPARQL endpoint and to write your first SPARQL query. For this purpose, open the link to the UniProt SPARQL endpoint, http://sparql. uniprot.org/ in a Web browser. How many entries do you think are available in UniProt? To find out, simply check the bottom-left corner of the Web page—you will notice that the total number of triples is always kept up to date there. How many of these entries describe proteins? To find out, try running the following SPARQL query that counts all instances of the database that belong to the protein class. What is the result?
```
PREFIX up:<http://purl.uniprot.org/core/>
SELECT (count(?protein) as ?count)
WHERE
{ ?protein a up:Protein. }
```
Notice that the UniProt SPARQL web page includes many examples on the right-hand side—in order to get more familiar with UniProt and SPARQL, try further some of the sample queries provided there.
B. Exploring Biological Ontologies Through Keyword Search in the Ontology Lookup Service
We have seen in Sect. 3.6 an example assertion about the "HBB" gene in the human, including the following triple:
#### oma:PROTEIN\_HUMAN04027 obo:RO\_0002162 <http://www.uniprot. org/taxonomy/9606> .
This triple essentially asserts that the gene is located in the Homo sapiens taxon. However, as a regular user, how could you know what the URIs for "in taxon" and Homo sapiens are? One of the possible ways to get these identifiers is by searching for the keywords of interest in the Ontology Lookup Service (OLS). To do this, go to the Web page of the service https://www.ebi.ac.uk/ols/ index, and try to enter first "in taxon". What is the result? Try also Homo sapiens. What about "human"?
C. Querying OMA with SPARQL
Recall from Sect. 3.6 the sample query we presented for retrieving the description of the human hemoglobin gene from OMA. We provide it in a more explicit form here:
```
SELECT ?description WHERE {
?protein oma:geneName "HBB".
?protein <http://bioontology.org/ontologies/biositemap.owl#description> ?de-
scription.
}
```
First try to think about possible information that is missing from this query. For example, is this query guaranteed to return a single result (remember we are using an orthology database)?
Try to look again at how the human "HBB" protein is defined in Sect. 3. Then, try to run the SPARQL query as-is in the OMA SPARQL endpoint: https://sparql.omabrowser.org/sparql. What do you get? What is the reason? Try to print out more information about the protein, not just its description. For example, add another triple pattern to capture the oma:hasOMAId property value as well (don't forget to add it to the selected variables in the first line!), perhaps also the taxon ID in UniProt. What can you deduce? Can you correct the query so that it only gets the description we were originally interested in?
D. Federated Queries Using SPARQL (OMA and UniProt)
In Sect. 4 we presented an example Federated Query using the SPARQL endpoint of OMA and the remote SPARQL endpoint of UniProt, as a service. We recall the query here:
```
prefix up:<http://purl.uniprot.org/core/>
prefix taxon:<http://purl.uniprot.org/taxonomy/>
select distinct ?proteinOMA ?proteinUniProt
where {
service <http://sparql.uniprot.org/sparql> {
?proteinUniProt a up:Protein .
?proteinUniProt up:organism taxon:9606 . # Homo Sapiens
?proteinUniProt up:annotation ?annotation . # annotations of this
protein entry
?annotation rdfs:comment ?text
filter( regex(str(?text), "leukemia") ) # only those containing
the text "leukemia"
}
?proteinOMA a orth:Protein.
?proteinOMA oma:xrefUniprot ?proteinUniProt.
```
}
Try running this query in the OMA SPARQL endpoint, https://sparql.omabrowser.org/sparql. You might need to wait a couple of minutes to get the remote results. Next, try to look at the examples provided in the right side of the page to see how to get more properties of the proteinOMA variable—for example, try getting the description or the OMA ID. Next, try modifying this query so that it can run in the UniProt SPARQL endpoint, invoking the OMA one as a service. Remember to get the relevant prefixes and define them in the header of the query first ("oma," "orth"). You can get these by looking at "Namespace prefixes" in the OMA SPARQL Web page. Finally, test your modifications using UniProt, http://sparql.uniprot.org/.
#### Acknowledgments
We would like to thank all the members of the Bio-SODA SNF project at ZHAW Zurich University of Applied Sciences, at the University of Lausanne (UNIL) and at the SIB Swiss Institute of Bioinformatics for their valuable feedback: Adrian Altenhoff (UNIL, SIB), Maria Anisimova (ZHAW, SIB), Christophe Dessimoz (UNIL, SIB), Frederic Bastian (UNIL, SIB), Heinz Stockinger (SIB), Marc Robinson-Rechavi (UNIL, SIB), and Erich Zbinden (ZHAW, SIB). This work was supported by the Swiss National Science Foundation (SNSF) funded under grant NRP 75 Big Data.
#### References
natural language querying over relational data stores. Proc VLDB Endow 9(12):1209–1220. https://doi.org/10.14778/2994509. 2994536
toothbrush! Brief Bioinform 9(3):220–231. https://doi.org/10.1093/bib/bbn003
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
#### High-Performance Computing in Bayesian Phylogenetics and Phylodynamics Using BEAGLE
#### Guy Baele, Daniel L. Ayres, Andrew Rambaut, Marc A. Suchard, and Philippe Lemey
#### Abstract
In this chapter, we focus on the computational challenges associated with statistical phylogenomics and how use of the broad-platform evolutionary analysis general likelihood evaluator (BEAGLE), a highperformance library for likelihood computation, can help to substantially reduce computation time in phylogenomic and phylodynamic analyses. We discuss computational improvements brought about by the BEAGLE library on a variety of state-of-the-art multicore hardware, and for a range of commonly used evolutionary models. For data sets of varying dimensions, we specifically focus on comparing performance in the Bayesian evolutionary analysis by sampling trees (BEAST) software between multicore central processing units (CPUs) and a wide range of graphics processing cards (GPUs). We put special emphasis on computational benchmarks from the field of phylodynamics, which combines the challenges of phylogenomics with those of modelling trait data associated with the observed sequence data. In conclusion, we show that for increasingly large molecular sequence data sets, GPUs can offer tremendous computational advancements through the use of the BEAGLE library, which is available for software packages for both Bayesian inference and maximum-likelihood frameworks.
Key words Adaptive Markov chain Monte Carlo, Multipartite data, Generalized linear model, Highperformance computing, BEAGLE, BEAST, Pathogen phylodynamics, Data integration, Bayesian phylogenetics, Phylogenomics
#### 1 Introduction
Phylogenomics, a term coined by Eisen and Fraser [13], explores the intersection of evolutionary studies and genomic analyses. Accurate phylogenetic reconstruction using genomic data has important repercussions for answering particular questions in genome analysis, as phylogenomic analyses often involve estimating the underlying evolutionary history of sequences either as an intermediate goal or as an end point. The availability of more and more complete genomes can help to correct for phylogenetic reconstruction artifacts and contradictory results that often appeared in
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_23, © The Author(s) 2019
molecular phylogenies based on a single or few orthologous genes [21]. Expanding the number of characters that can be used in phylogenetic reconstruction from a few thousand to tens of thousands, these large quantities of data lead to reduced estimation errors associated with site sampling, to very high power in the rejection of simple evolutionary hypotheses and to high confidence in estimated phylogenetic patterns [4].
Among the phylogenetic reconstruction approaches that have attained widespread recognition, Bayesian inference has become increasingly popular, in large part due to the availability of opensource software packages such as the Bayesian evolutionary analysis by sampling trees (BEAST) software [11] and MrBayes [29]. Bayesian phylogenetic inference is based on a quantity called the posterior distribution of trees, which involves a summation over all trees and, for each tree, integration over all possible combinations of branch length and substitution model parameter values [20]. Analytical evaluation of this distribution is practically infeasible, and hence needs to be approximated using a numerical method, the most common being Markov chain Monte Carlo (MCMC). The basic idea is to construct a Markov chain that has as its state space the parameters of the statistical model and a stationary distribution that is the posterior distribution of the parameters (including the tree) [20]. While MCMC integration has revolutionized the field of phylogenetics [34], the continuously increasing size of data sets is pushing the field of statistical phylogenetics to its limits.
While promising approaches to improve MCMC efficiency have emerged recently from the field of computational statistics, such as sequential Monte Carlo (SMC; see, e.g., Doucet [10]) and Hamiltonian Monte Carlo (HMC; see, e.g., Neal [27]), these approaches do not yet find widespread use in phylogenetics. The primary difficulty in this adoption centers around the tree that encompasses both continuous and discrete random variables. Instead, considerable attention is being meted on techniques for parallelization [32] to improve phylogenetic software run-times. Obtaining sufficient samples from a Markov chain may take many iterations, due to the large number of trees that may describe the relationships of a group of species and high autocorrelation between the samples. It is therefore of critical importance to perform each iteration in a computationally efficient manner, making optimal use of the available hardware. High-performance computational libraries, such as the broad-platform evolutionary analysis general likelihood evaluator (BEAGLE) [3], can be useful tools to enable efficient use of multicore computer hardware (or even special-purpose hardware), while at the same time requiring minimal knowledge from the software user(s).
In this chapter, we first introduce the BEAGLE software library and its primary purpose, characteristics, and typical usages in Subheading 2, along with the hardware specifications of the devices used for benchmarks in this chapter. In Subheading 3, we present computational benchmarks on the different hardware devices for a collection of data sets that are typically analyzed with models of varying complexity. Subheading 4 presents a brief overview of studies for which GPU computing capabilities were critical to analyze the data in a timely fashion. Given the increasing capabilities over hardware devices, we present an interesting avenue for further research in Subheading 5, in the form of adaptive MCMC.
#### 2 The BEAGLE Library
BEAGLE [3] is a high-performance likelihood-calculation platform for phylogenetic applications. BEAGLE defines a uniform application programming interface (API) and includes a collection of efficient implementations for evaluating likelihoods under a wide range of evolutionary models, on graphics processing units (GPUs) as well as on multicore central processing units (CPUs). The BEA-GLE library can be installed as a shared resource, to be used by any software aimed at phylogenetic reconstruction that supports the library. This approach allows developers of phylogenetic software to share any optimizations of the core calculations, and any program that uses BEAGLE will automatically benefit from the improvements to the library. For researchers, this centralization provides a single installation to take advantage of new hardware and parallelization techniques.
The BEAGLE project has been very successful in bringing hardware acceleration to phylogenetics. The library has been integrated into popular phylogenetics software including BEAST [11], MrBayes [29], PhyML [19], and GARLI [35] and has been widely used across a diverse range of evolutionary studies. The BEAGLE library is free, open-source software licensed under the Lesser GPL and available at https://beagle-dev.github.io.
#### 2.1 Principles
2.1.1 Computing Observed Data Likelihoods The most effective methods for phylogenetic inference involve computing the probability of observed character data for a set of taxa given an evolutionary model and phylogenetic tree, which is often referred to as the (observed data) likelihood of that tree. Felsenstein demonstrated an algorithm to calculate this probability [16], and his algorithm recursively computes partial likelihoods via simple sums and products. These partial likelihoods track the probability of the observed data descended from an internal node conditional on a particular state at that internal node.
The partial likelihood calculations apply to a subtree comprising a parent node, two child nodes, and connecting branches. It is repeated for each unique site pattern in the data (in the form of a multiple sequence alignment), for each possible character of the state space (e.g., nucleotide, amino acid, or codon), and for each internal node in the proposed tree. The computational complexity of the likelihood calculation for a given tree is <sup>O</sup>( <sup>p</sup> <sup>s</sup> <sup>2</sup> <sup>n</sup>), where p is the number of unique site patterns in the sequence (typically on the order of 102 –10<sup>6</sup> ), s is the number of states each character in the sequence can assume (typically 4 for a nucleotide model, 20 for an amino-acid model, or 61 for a codon model), and n is the number of operational taxonomic units (e.g., species and alleles).
Additionally, the tree space is very large; the number of unrooted topologies possible for n operational taxonomic units is given by the double factorial function (2<sup>n</sup> 5)!! [15]. Thus, to explore even a fraction of the tree space, a very large number of topologies need to be evaluated, and hence a very great number of likelihood calculations have to be performed. This leads to analyses that can take days, weeks, or even months to run. Further compounding the issue, rapid advances in the collection of DNA sequence data have made the limitation for biological understanding of these data an increasingly computational problem. For phylogenetic inferences, the computation bottleneck is most often the calculation of the likelihoods on a tree. Hence, speeding up the calculation of the likelihood function is key to increasing the performance of these analyses.
Advances in computer hardware, specifically in parallel architectures, such as many-core GPUs, multicore CPUs, and CPU intrinsics (e.g., SSE and AVX), have created opportunities for new approaches to computationally intensive methods. The structure of the likelihood calculation, involving large numbers of positions and multiple states, as well as other characteristics, makes it a very appealing computational fit to these modern parallel processors, especially to GPUs.
BEAGLE exploits GPUs via fine-grained parallelization of functions necessary for computing the likelihood on a (phylogenetic) tree. Phylogenetic inference programs typically explore tree space in a sequential manner (Fig. 1, tree space) or with only a small number of sampling chains, offering limited opportunity for tasklevel parallelization. In contrast, the crucial computation of partial likelihood arrays at each node of a proposed tree presents an excellent opportunity for fine-grained data parallelism, which GPUs are especially suited for. The use of many lightweight execution threads incurs very low overhead on GPUs, enabling efficient parallelism at this level.
In order to calculate the overall likelihood of a proposed tree, phylogenetic inference programs perform a post-order traversal, evaluating a partial likelihood array at each node. When using BEAGLE, the evaluation of these multidimensional arrays is offloaded to the library. While each partial likelihood array is still
2.1.2 Parallel Computation
Fig. 1 Diagrammatic example of the tree sampling process and fine-grained parallel computation of phylogenetic partial likelihoods using BEAGLE for a nucleotide model problem with five taxa, nine site patterns, and four evolutionary rate categories. Each entry in a partial likelihood array L is assigned to a separate GPU thread t. In this example, 144 GPU threads are created to enable parallel evaluation of each entry of the partial likelihood array L(x0)
evaluated in sequence, BEAGLE assigns the calculation of the array entries to separate GPU threads, for computation in parallel (Fig. 1, partial likelihood). Further, BEAGLE uses GPUs to parallelize other functions necessary for computing the overall tree likelihood, thus minimizing data transfers between the CPU and GPU. These additional functions include those necessary for computing branch transition probabilities, for integrating root and edge likelihoods, and for summing site likelihoods.
Multicore CPU parallelization through BEAGLE can only be done via multiple instances of the library, such that each instance computes a different data partition. Multiple CPU threads can be used (e.g., one for each partition) if the application program (BEAST, for the remainder of this chapter) creates the BEAGLE instances in separate computation threads, which will be the case when using BEAST. This approach suits the trend of increasingly large molecular sequence data sets, which are often heavily partitioned in order to better model the underlying evolutionary processes. BEAGLE itself does not employ any kind of load balancing nor are the site columns computed in individual threads. Each BEAGLE instance only parallelizes computation on CPUs via SSE vectorization.
BEAGLE can also use GPUs to perform partitioned analyses, however for problem sizes that are insufficiently large to saturate the capacity of one device, efficient computation requires multiple GPUs. Recent progress has been made in parallelizing the computation of multiple data subsets on one GPU [1], and future releases of BEAGLE will include this capability.
Fig. 2 Layer diagram depicting BEAGLE library organization, and illustration of API use. Arrows indicate direction and relative size of data transfers between the client program and library
#### 2.2 Design
2.2.1 Library
The general structure of the BEAGLE library can be conceptualized as layers (Fig. 2, library), the upper most of which is the application programming interface. Underlying this API is an implementation management layer, which loads the available implementations, makes them available to the client program, and passes API commands to the selected implementation.
The design of BEAGLE allows for new implementations to be developed without the need to alter the core library code or how client programs interface with the library. This architecture also includes a plugin system, which allows implementation-specific code (via shared libraries) to be loaded at runtime when the required dependencies are present. Consequently, new frameworks and hardware platforms can more easily be made available to programs that use the library, and ultimately to users performing phylogenetic analyses.
Currently, the implementations in BEAGLE derive from two general models. One is a serial CPU implementation model, which does not directly use external frameworks. Under this model, there is a standard CPU implementation, and one with added SSE intrinsics, which uses vector processing extensions present in many CPUs to parallelize computation across character state values. The other implementation model involves an explicit parallel accelerator programming model, which uses the CUDA external computing framework to exploit NVIDIA GPUs. It implements fine-grained parallelism for evaluating likelihoods under arbitrary molecular evolutionary models, and thus harnessing the large number of processing cores to efficiently perform calculations [3, 32].
Recent progress has been made in developing new implementations for BEAGLE, beyond those described here, thus expanding the range of hardware that can be used. Upcoming releases of the library will include additional support for CPU parallelism via a multi-threaded implementation and will support the OpenCL standard, enabling the use of AMD GPUs [2].
2.2.2 Application Programming Interface The BEAGLE API was designed to increase performance via finescale parallelization while reducing data transfer and memory copy overhead to an external hardware accelerator device (e.g., GPU). Client programs, such as BEAST [11], use the API to offload the evaluation of tree likelihoods to the BEAGLE library (Fig. 2, API). API functions can be subdivided into two categories: those which are only executed once per inference run and those which are repeatedly called as part of the iterative sampling process. As part of the one-time initialization process, client programs use the API to indicate analysis parameters such as tree size and sequence length, as well as specifying the type of evolutionary model and hardware resource(s) to be used. This allows BEAGLE to allocate the appropriate number and size of data buffers on device memory. Additionally at this initialization stage, the sequence data is specified and transferred to device memory. This costly memory operation is only performed once, thus minimizing its impact.
> During the iterative tree sampling procedure, client programs use the API to specify changes to the evolutionary model and instruct a series of partial likelihood operations that traverse the proposed tree in order to find its overall likelihood. BEAGLE efficiently computes these operations and makes the overall tree likelihood as well as per-site likelihoods available via another API call.
2.3 Performance Peak performance with BEAGLE is achieved when using a highend GPU; however, the relative gain over using a CPU depends on model type and problem size as more demanding analyses allow for better utilization of GPU cores. Figure 3 shows speedups relative to serial CPU code when using BEAGLE with an NVIDIA P100 GPU for the critical partial likelihood function, with increasing unique site pattern counts and for two model types. Computing these likelihoods typically accounts for over 90% of the total execution time for phylogenetic inference programs and the relationship between speedups and problem size observed here primarily matches what would be observed for a full analysis.
> Figure 3 includes performance results for computing partial likelihoods under both nucleotide and codon models. The vertical axis labels show the speedup relative to the average performance of a baseline serial, single threaded and non-vectorized, CPU implementation. This nonparallel CPU implementation provides a consistent performance level across different problem sizes and
Fig. 3 Plots showing BEAGLE partial likelihood computation performance on the GPU relative to serial CPU code, under nucleotide and codon models and for an increasing number of unique site patterns. Speedup factors are on a log-scale
provides a relevant point of comparison as most phylogenetic inference software packages use serial code as their standard.
Using a nucleotide model, relative GPU performance over the CPU strongly scales with the number of site patterns. For very small numbers of patterns, the GPU exhibits poor performance due to greater execution overhead relative to overall problem size. GPU performance improves quickly as the number of unique site patterns is increased and by 10,000 patterns it is closer to a saturation point, continuing to increase but more slowly. At 100,000 nucleotide patterns, the GPU is approximately 64 times faster than the serial CPU implementation.
For codon-based models, GPU performance is less sensitive to the number of unique site patterns. This is due to the better parallelization opportunity afforded by the 61 biologically meaningful states that can be encoded by a codon. The higher state count of codon data compared to nucleotide data increases the ratio of computation to data transfer, resulting in increased GPU performance for codon-based analyses. For a problem size with 10,000 codon patterns, the GPU is over 256 times faster than the serial CPU implementation.
#### 2.4 Memory Usage When assessing the suitability of a phylogenetic analysis for GPU acceleration via BEAGLE, it is also important to consider if the GPU has sufficient on-board memory for the analysis to be performed. GPUs typically have less memory than what is available to CPUs, and the high transfer cost of moving data from CPU to GPU memory prevents direct use of CPU memory for GPU acceleration.
Figure 4 shows how much memory is required for problems of different sizes when running nucleotide and codon-model analyses in BEAST with BEAGLE GPU acceleration. Note that when multiple GPUs are available, BEAST can partition a data set into
Fig. 4 Contour plots depicting BEAGLE memory usage on GPUs for BEAST nucleotide and codon-model analyses with 4 rate categories in double-precision floating-point format, for a range of problem sizes with different numbers of taxa and of unique site patterns. Memory requirements shown here assume an unpartitioned dataset. Partitioned analyses and more sophisticated models that use multiple BEAGLE instances incur memory overhead per additional library instance. Black dots indicate memory usage requirements for the unpartitioned version of three data sets subsequently described in this chapter
separate BEAGLE instances, one for each GPU. Thus, each GPU will only require as much memory as necessary for the data subset assigned to it. Typical PC-gaming GPUs have 8 GB of memory or less, while GPUs dedicated to high-performance computing, such as the NVIDIA Tesla series, may have as much as 24 GB of memory.
2.5 Hardware Highly parallel computing technologies such as GPUs have overtaken traditional CPUs in peak performance potential and continue to advance at a faster pace. Additionally, the memory bandwidth available to the processor is especially relevant to data-intensive computations, such as the evaluation of nucleotide model likelihoods. In this measure as well, high-end GPUs significantly outperform equivalently positioned CPUs.
> BEAGLE was designed to take advantage of this trend of increasingly advanced GPUs and uses runtime compilation methods to optimize code for whichever generation of hardware is being used. Table 1 lists hardware specifications for the processors used in this chapter. We note that further advancements in the GPU market for scientific computing are on its way, with NVIDIA preparing the launch (at the time of writing) of the Tesla V100 in Q3 of 2017. The new NVIDIA Tesla V100 features a total of 5120 CUDA cores and comes equipped with 32 GB of on-board memory with 900 GB/s of bandwidth. As such, it seems to have the potential to reach 7.5 TFLOPs in double-precision peak performance (DP PP), a roughly 50% increase over their current flagship, the Tesla P100.
Table 1 Hardware specifications for the Intel CPUs and NVIDIA GPUs used in this chapter
Estimated performance in double-precision peak performance (DP PP) taken from the manufacturer's website. Note that the Quadro P5000 GPU only lists performance in single precision and we hence list its double-precision performance as not available (NA)
#### 3 Results
In this section, we compare the performance of various typical Bayesian phylogenetic, phylogenomic, and phylodynamic analyses on different multicore architectures. In Subheading 3.1, we analyze a data set of mitochondrial genomes [32] using a high-dimensional model of codon substitution which, albeit low in number of parameters, is particularly challenging in phylogenetic analyses specifically because of the high-dimensional state space. In Subheading 3.2, we analyze the largest Ebola virus data set at the time of publication [12] using a collection of nucleotide substitution models, i.e., one per codon position and an extra one for analyzing the intergenic regions, where the large number of taxa and unique site patterns offer an interesting test case for the comparison between CPU and GPU performance. Finally, Subheading 3.3 reports on the performance of analyzing data sets that complement sequence data with discrete trait data (typically host data or geographic data), for which transition rates between (a potential large number of) discrete trait states are parameterized as a generalized linear model (GLM). All performance evaluations in this results section were run for 100,000 iterations (which is usually insufficient to achieve convergence) in BEAST v1.8.4 [11], using double precision (both on CPU and GPU) and in conjunction with BEAGLE v2.1.2 [3]. By default, BEAST—through BEAGLE—uses SSE2 (Streaming SIMD Extensions 2), an SIMD instruction set extension to the x86 architecture, when performing calculations on CPU.
#### 3.1 Carnivores Selection is a key evolutionary process in shaping genetic diversity and a major focus of phylogenomics investigations [23]. Researchers frequently evaluate the strength of selection operating on genes or even individual codons in the entire phylogeny or in a subset of branches using statistical methods. Codon substitution models have been particularly useful for this purpose because they allow estimating the ratio of non-synonymous and synonymous substitution rates (dN/dS) in a phylogenetic framework. Goldman and Yang [18] and Muse and Gaut [26] developed the first codonbased evolutionary models (GY and MG, respectively), i.e., models that have codons as their states, incorporating biologically meaningful parameters such as transition/transversion bias, variability of a gene, and amino acid differences.
Full codon substitution models are computationally expensive compared to standard nucleotide substitution models due to their large state space. Compared to nucleotide models (4 states) and amino acid models (20 states), a full vertebrate mitochondrial codon model has 60 states (ignoring the four nonsense or stop codons). We restrict ourselves to the standard GY codon substitution model implementation in BEAST [11], employ the standard assumption that mutations occur independently at the three codon positions and therefore only consider substitutions that involve a single-nucleotide substitution, and assume that codons evolve independently from one another. Additionally, we allow for substitution rate heterogeneity among codons using a discrete gamma distribution (i.e., each codon is allowed to evolve at a different substitution rate) [33], which increases the computational demands of such an analysis fourfold (given that we allow for the standard assumption of four discrete rate categories).
As a first application of using state-of-the-art hardware in statistical phylogenetics, we reevaluate the performance of a full codon model on a set of mitochondrial genomes from extant carnivores and a pangolin outgroup [4, 32]. This genomic sequence alignment contains 10,869 nt columns that code for 12 mitochondrial proteins and when translated into a single 60-state vertebrate mitochondrial codon model, yields a total of 3623 alignment columns, of which 3601 site patterns are unique [32]. Figure 5 shows a comparison of the computational throughput between various CPU and GPU computing platforms. To this end, we make use of an option in BEAST [11] to split an alignment into two or more pieces of equal length, with each resulting alignment being evaluated on a separate processor core or computing device for optimal performance. Figure 5 shows that the analysis scales remarkably well on CPU, where the use of each additional processor core results in a performance increase. This can be attributed to the use of full codon models, which invokes a higher workload when evaluating each likelihood and hence more concurrent evaluation compared to thread communication. As the evaluation of the total **CPU performance**
Fig. 5 Performance evaluation of different CPU and GPU configurations when estimating a single full codon substitution model on a full genome mitochondrial data set. Left: performance comparison on multicore CPUs (with the dashed line indicating optimal GPU performance). Right: performance comparison on GPUs. The numbers below the bars indicate the number of partitions/threads used in each analysis. The GTX 590, a GPU released in 2011 for PC gaming, performs equally well as a 24-core CPU configuration. However, GPUs from the Tesla generation, aimed at scientific computing, drastically outperform the CPU system, with the recently released P100 showcasing the impressive improvements in the GPU market, running 166 times faster than a single-core CPU setup
amount of unique site patterns is split over more processor cores, the workload per core decreases and the communication overhead increases, resulting in smaller relative performance increases.
**GPU performance**
The performance of a 24-core CPU setup is easily matched by a single GPU (the GTX 590) that was originally aimed at the market of PC gaming. However, subsequent improvements in GPU cards for scientific computing have yielded impressive performance gains, with a single Tesla K20 GPU outperforming 2 GTX 590 GPUs. Whereas the advent of the Tesla K40 offered further performance increases, it was mainly welcomed for having twice the amount of on-board memory, allowing for much larger data sets to be analyzed on GPU. The recent introduction of the Tesla P100 GPU promised and delivered astonishing results, as shown in Fig. 5, with a single Tesla P100 GPU delivering six times the performance of a Tesla K40 GPU on these high-dimensional full codon models. We conclude that the use of a high-performance computational library such as BEAGLE, in combination with a powerful GPU, has significantly facilitated the evaluation of and phylogenetic inference with full codon models.
3.2 Ebola Virus The original developments within the BEAGLE library offered considerable computational speedup when evaluating codon models—up to a 52-fold increase when employing three GPU cards and nucleotide models—up to a 15-fold increase when using three GPU cards—in double precision [32]. This may have resulted in the perception that GPU cards are mainly useful when evaluating codon models, but that the benefit for fitting models was not sufficiently substantial to warrant GPUs. To offer an objective assessment of the usefulness of GPUs in such cases, we analyze the use of various CPU and GPU configurations on a full genome Ebola virus data set, consisting of 1610 publicly available genomes sampled over the course of the 2013–2016 Ebola virus disease epidemic in West Africa [12] (we discuss this study in more detail in Subheading 4). This data set encompassing 18,992 nt columns is modelled with four partitions: one for each codon position and one additional partition for the intergenic region (which consists of several noncoding regions interspersed in the genome). The three codon partitions contain, respectively, 2366, 2328, and 2731 unique site patterns, while the intergenic partition contains 2785 unique site patterns. We model among-site rate variation [33] in each partition independently, which confronts us with a computationally demanding analysis for this large number of taxa and unique site patterns.
> Figure 6 shows how the performance of such a large nucleotide data set scales with the available CPU and GPU resources. Contrary to the carnivores data set analysis in Subheading 3.1, this analysis does not scale particularly well with the number of CPU cores available, as the main benefit lies with splitting each partition into two subpartitions and only limited performance gains can be observed when using additional partitions or threads. Popular single GPU cards for scientific computing—such as the Tesla K40—match the optimal performance brought about by using 16 CPU cores, and may provide a useful alternative to multicore CPU systems. However, the decreasing cost for increasingly parallel multicore CPU systems makes this a difficult matchup for slightly older GPUs. More recently introduced GPU cards, such as the Tesla P100, are able to deliver a substantial performance improvement over a multicore CPU setup, with two Tesla P100 GPUs running in parallel offering over twice the performance of a 16-core CPU setup. We note that the GTX 590 cards, as well as a single Tesla K20 card, do not contain sufficient on-board memory to hold the full data set and as such, these benchmarks could not be run on those resources.
3.3 Phylogeography As shown in the results in Subheadings 3.1 and 3.2, different partitions of the aligned sequence data can contain a large number of unique site patterns, rendering phylogenomic inference challenging. However, other data types are also included more
Fig. 6 Performance evaluation of different CPU and GPU configurations when estimating a four-partition nucleotide substitution model on a full genome Ebola virus data set. Left: performance comparison on multicore CPUs (with the dashed line indicating optimal GPU performance). Right: performance comparison on GPUs. The numbers below the bars indicate the number of partitions/threads used in each analysis. Because of the lower dimensionality compared to full codon models, nucleotide substitution models currently take less advantage of the large amounts of cores present on GPUs. The fastest GPU configuration consisting of two Tesla P100 GPUs—outperforms by 107% the fastest CPU configuration that employs 16 threads on a 24-core Haswell CPU
frequently, such as trait data to be analyzed alongside the sequence data and hence potentially influencing the outcome of such an analysis (for an overview, see, e.g., Baele et al. [6]).
Arguably, the most frequently considered traits in phylodynamics, and molecular evolution in general, are spatial locations. The interest in spatial dispersal has developed into its own research field referred to as phylogeography, with Bayesian inference of discrete phylogenetic diffusion processes being adopted in the field of biogeography [30]. Jointly estimating the phylogeny and the trait evolutionary process, Lemey et al. [23] implemented a similar Bayesian full probabilistic connection between sequences and traits in BEAST [11], with applications focusing on spatiotemporal reconstructions of viral spread. These approaches offer extensive modelling flexibility at the expense of a quadratic growth in number of instantaneous rate parameters in the continuous-time Markov chain (CTMC) model as a function of the state dimensionality of the trait. This can be seen in Fig. 7, which shows two maps with different numbers of (discrete) locations and the corresponding CTMC models that describe the instantaneous rates of transition between these locations.
Fig. 7 Graphical depiction of the Benelux countries (top) and the provinces of Belgium (bottom). When these countries and provinces are used as discrete location states in a discrete trait model, this yields, respectively, a 3 3 and a 10 10 CTMC model with instantaneous rates of transition between each pair of locations. Such models are subject to the same restrictions as those used in popular substitution models, i.e., the rows sum to 0. As such, these CTMC models consist of, respectively, 6 and 90 free parameters to be estimated
Many phylodynamic hypotheses can be addressed through the combination of genetic and trait data, but additional data in the form of covariates can help explain the evolutionary or epidemiological process. Such covariates can be used in a GLM formulation on a matrix of transition rate parameters between locations defining a CTMC process. Lemey et al. [25] developed an approach to simultaneously reconstruct spatiotemporal history and identify which combination of covariates associates with the pattern of spatial spread. This approach involves parameterizing each rate of among-location movement, typically denoted as the ijth elements (λij) of the CTMC transition rate matrix, in the phylogeographic model as a log linear function of various potential covariates:
$$
\lambda \log \mathfrak{i}\_{ij} = \beta\_1 \delta\_1 \mathfrak{x}\_{i,j,1} + \beta\_2 \delta\_2 \mathfrak{x}\_{i,j,2} + \dots + \beta\_N \delta\_N \mathfrak{x}\_{i,j,N}, \tag{1}
$$
where β<sup>i</sup> is the estimated effect size of covariate xi, δ<sup>i</sup> is a binary indicator that tracks the posterior probability of the inclusion of covariate xi in the model, and N equals the number of covariates; further, in the case of Fig. 7: <sup>i</sup>, <sup>j</sup> <sup>∈</sup> {A, <sup>B</sup>, <sup>C</sup>} with <sup>N</sup> <sup>¼</sup> 3 (top), and <sup>i</sup>, <sup>j</sup> <sup>∈</sup>{A, <sup>B</sup>, <sup>C</sup>, <sup>D</sup>, <sup>E</sup>, <sup>F</sup>, <sup>G</sup>, <sup>H</sup>, <sup>I</sup>, <sup>J</sup>} with <sup>N</sup> <sup>¼</sup> 10 (bottom). Priors and posteriors for the inclusion probabilities (δ) can be used to express the support for each predictor in terms of Bayes factors (for more information, see Baele et al. [6], Lemey et al. [25]). We discuss examples of such possible predictors in a phylogeographic setting in Subheading 4 but focus here on performance benchmarks for such generalized linear models.
3.3.1 Bat Rabies We here assess the performance of a phylodynamic setup aimed at reconstructing the spatial dispersal and cross-species dynamics of rabies virus (RABV) in North American bat populations based on a set of 372 nucleoprotein gene sequences (nucleotide positions: 594–1353). The data set comprises a total of 17 bat species sampled between 1997 and 2006 across 14 states in the USA [31]. Two additional species that had been excluded from the original analysis owing to a limited amount of available sequences, Myotis austroriparius (Ma) and Parastrellus hesperus (Ph), are also included here [14]. We also include a viral sequence with an unknown sampling date (accession no. TX5275, sampled in Texas from Lasiurus borealis), which will be adequately accommodated in our inference. This leads to a total of 548 unique site patterns. Following Faria et al. [14], we employ two GLM-diffusion models for this analysis, one on the discrete set of 17 bat species and another on the discrete set of 14 location states.
> Figure 8 shows the performance of various multicore platforms on the bat rabies Bayesian phylodynamic analysis. In contrast to previous examples, the low number of sites (and hence unique site patterns) in the alignment does not offer many options for splitting the observed data likelihood over additional threads. While four CPU cores offer the optimal performance across our CPU platforms, using more threads for the analysis causes serious communication overhead, slowing down the analysis. Comparing the CPU results with the GPU results shows that, across all multicore platforms tested, a 4-core CPU offers the best performance.
> Nonetheless, this scenario provides a very interesting use case for employing multiple graphics cards for scientific computing. Even though the (relatively small) dimensions of this particular example do not allow for a performance increase, it will be beneficial for higher-dimensional cases to compute each diffusion model on a separate GPU. When assuming independent diffusion processes that only depend on the underlying phylogeny, each of the trait diffusion models can be computed on a different GPU, whereas the data alignment can be split into two subpartitions of equal complexity (i.e., with an equal number of unique site patterns) and hence also be computed in parallel over the two GPUs. However, the limited sequence data size and the relatively restricted
Fig. 8 Performance evaluation of different CPU and GPU configurations when estimating a single-partition nucleotide substitution model on a rabies virus data set, along with two GLMs. Left: performance comparison on multicore CPUs (with the dashed line indicating optimal GPU performance). Right: performance comparison on GPUs. Given the short length of the alignment, communication overhead becomes quite severe when adding a large number of CPU processor cores. Even two state-of-the-art Tesla P100 GPUs fail to outperform four CPU cores, given the low number of sequences and the low number of states for the GLMs
number of discrete locations make this data set less suited for illustrating performance increases using GPUs.
3.3.2 Ebola Virus We here assess the performance of a similar setup as in the previous section, but using the data from the 2013–2016 West African epidemic caused by the Ebola virus [12]. We hence use the nucleotide data from the previous Ebola example (see Subheading 3.2) and augment it with location states. Using a phylogeographic GLM that integrates covariates of spatial spread, we have examined which features influenced the spread of EBOV among administrative regions at the district (Sierra Leone), prefecture (Guinea), and country (Liberia) levels. This resulted in a GLM parameterizing transition rates among 56 discrete location states according to 25 potential covariates (see Dudas et al. [12] for more information), resulting in a computationally challenging analysis.
> As shown in Fig. 9, we have evaluated the performance of this challenging data set on our different multicore platforms. By comparing these benchmarks with those in Fig. 6, it's clear that the addition of a high-dimensional discrete trait model is much harder to process for any multicore CPU configuration. Adding more CPU cores to the analysis does not improve performance by much, indicative of the discrete trait model being the main
Fig. 9 Performance evaluation of different CPU and GPU configurations when estimating a four-partition nucleotide substitution model on a full genome Ebola virus data set along with a GLM-diffusion model for the location traits. Left: performance comparison on multicore CPUs (with the dashed line indicating optimal GPU performance). Right: performance comparison on GPUs. While a previous comparison of solely the nucleotide data resulted in a performance increase of over 100% for GPUs over CPUs, we observe a far more pronounced benefit for GPU in this case with two Tesla P100 GPUs outperforming a 12-core CPU setup by 568%
bottleneck in this analysis. This can be attributed to the high dimension of the discrete phylogeographic model [23] and the fact that this model describes a single column of characters, which cannot be split into multiple partitions. Relative to the computational complexity of modelling the location states, splitting the observed sequence data over multiple partitions/threads yields relatively small performance improvements.
Some of the (older) GPUs cannot fit the full data set in memory (such as the GTX 590 and a single Tesla K20), but those that are able to vastly outperform any CPU setup. Further, as these GPUs are better equipped to handle high-dimensional models, splitting the observed sequence data over multiple physical cards still yields noticeable performance gains. In contrast to Fig. 8, where two discrete phylogeographic models were used that could each be computed on different GPUs, the fact that this example only considers a single trait observation explains why less performance gains can be obtained by adding an additional GPU to the analysis.
#### 4 Examples
In this section, we highlight examples of large sequence data sets that are augmented with trait data in the form of discrete geographic locations, for which BEAGLE [3] offers impressive computational benefits, specifically when running these analyses on powerful graphics cards for scientific computing. Further, discrete phylogeographic models can be equipped with generalized linear models to identify predictors of pathogen spread. Both inclusion probabilities and conditional effect sizes for these predictors are estimated in order to determine support for such explanatory variables of (pathogen) spread.
4.1 Human Influenza H3N2 A potentially powerful predictor for the behavior of influenza and other infectious diseases comes in the form of information on global human movement patterns, of which the worldwide air transportation network is by far the best studied system of global mobility in the context of human infectious diseases [8].
> Lemey et al. [25] use a discrete phylogeographic model equipped with a GLM to show that the global dynamics of influenza H3N2 are driven by air passenger flows, whereas at more local scales spread is also determined by processes that correlate with geographic distance. For a data set that encompasses 1441 timestamped hemagglutinin sequences (sampled between 2002 and 2007) and up to 26 locations to be used in a discrete phylogeographic model equipped with a GLM, BEAGLE can offer substantial performance gains. A snapshot of a visual reconstruction through geographic space is presented in Fig. 10, which includes a summary of the support for the collection of covariates in the
Fig. 10 Snapshot of the geographic spread of human influenza subtype H3N2, based on 1441 hemagglutinin sequences sampled between 2002 and 2007 [25]. A discrete phylogeographic approach was used, allocating the sequence data into a discrete number of locations and employing a generalized linear model on the parameters that model geographic spread. Inclusion probabilities and Bayes factor support are shown for the most prominent predictors of H3N2 geographic spread. D3 visualization is made using SpreaD3 [7], with circular polygon areas proportional to the number of tree lineages maintaining that location at that time
GLM that offer the strongest contribution to spatial spread among those tested. As illustrated in Fig. 10 (but see Lemey et al. [25] for additional data), there is strong evidence that air passenger flow is among the most dominant drivers of the global dissemination of H3N2 influenza viruses. Further, geographic spread is found to be inversely associated (data not shown; but see Lemey et al. [25]) with geographical distance between locations and with origin and destination population densities, which may seem counterintuitive. As the authors state, this negative association of population density with viral movement may suggest that commuting is less likely, per capita, to occur out of, or into, dense subpopulations.
4.2 Ebola Virus During the two and a half years Ebola virus (EBOV) circulated in West Africa, it caused at least 28,646 cases and 11,323 deaths. As mentioned in Subheading 3.3.2, Dudas et al. [12] used 1610 genome sequences collected throughout the epidemic, representing over 5% of recorded Ebola virus disease (EVD) cases to reconstruct a detailed phylogenetic history of the movement of EBOV within and between the three most affected countries. This study considers a massive time-stamped data set that allows to uncover regional patterns and drivers of the epidemic across its entire duration, whereas individual studies had previously focused on either limited geographical areas or time periods. The authors use the phylogeographic GLM to test which features were important in shaping the spatial dynamics of EVD during the West African epidemic (see Fig. 11).
> The phylogeographic GLM allowed Dudas et al. [12] to determine the factors that influenced the spread of EBOV among administrative regions at the district (Sierra Leone), prefecture (Guinea), and country (Liberia) levels. The authors find that EBOV tends to disperse between geographically close regions, with great circle distances having among the strongest Bayes factor support for inclusion in the GLM among all covariates tested (along with four other predictors). Additionally, both origin and destination population sizes are equally strongly and positively correlated with viral dissemination (see Fig. 11). Dudas et al. [12] conclude that the combination of the positive effect of population sizes with the inverse effect of geographic distance implies that the epidemic's spread followed a classic gravity-model dynamic, with intense dispersal between larger and closer populations. Finally, the authors found a significant propensity for virus dispersal to occur within each country, relative to internationally, suggesting that country borders may have provided a barrier for the geographic spread of EBOV.
Fig. 11 Snapshot of the geographic spread of the Ebola virus during the 2013–2016 West African epidemic, based on 1610 whole genome sequences [12]. A discrete phylogeographic approach was used, allocating the sequence data into a discrete number of locations and employing a generalized linear model on the parameters that model geographic spread. Inclusion probabilities and Bayes factor support are shown for the most prominent predictors of Ebola virus geographic spread. D3 visualization is made using SpreaD3 [7], with circular polygon areas proportional to the number of tree lineages maintaining that location at that time
#### 5 Adaptive MCMC
The various data sets described in this chapter so far have shown the use and computational performance of a wide range of models in phylogenetic, phylogenomic, and phylodynamic research. Whether employing full codon models (e.g., the carnivores data set), codon partition models (e.g., the Ebola data set), or discrete phylogeographic models, the number of parameters of a typical Bayesian phylogenetic analysis has increased drastically over the years. This is exacerbated by the use of partitioning strategies, resulting also in a potentially large array of likelihoods that need to be evaluated simultaneously, increasing run times for most phylogenetic analyses. In a similar fashion, computational resources available to researchers have also markedly increased, both in the form of multicore CPU technology and increasingly powerful graphics cards targeted towards scientific computing. The ubiquitous availability of multiprocessor and multicore computers practically has motivated the design of novel parallel algorithms to make efficient use of these machines [22, 32].
Many Bayesian phylogenetics software packages, such as BEAST [11] and MrBayes [29], do not fully exploit the inherent parallelism of such multicore systems when analyzing partitioned data because they typically update one single parameter at a time (a practice called single-component Metropolis–Hastings; Gilks et al. [17]). Such a single parameter often belongs to an evolutionary model for a single data partition, leading to only one of the potentially large collection of (observed) data likelihoods to be modified at any one time. Such a strategy does not use the computational power of modern-day multiprocessor and multicore systems to its full advantage. Updating all the models' parameters at once however leads to multiple data likelihoods being modified simultaneously, thereby making better use of the resources offered by these multicore systems (see Fig. 12).
Fig. 12 Conceptual visualization on the potential benefits of an adaptive MCMC algorithm over singlecomponent Metropolis–Hastings (green bars indicate that a processor is computing a specific likelihood). In Bayesian phylogenetics, the common practice of updating a single parameter (i.e., either a, b, c, or d) at a time leaves many CPU cores idle, underusing the computational performance of such architecture. Adaptive MCMC allows to update a collection of continuous parameters simultaneously (i.e., a, b, c, and d), putting many cores (in this case: 4) to work in a parallel fashion. Quad-Core AMD Opteron processor silicon die is shown, courtesy of Advanced Micro Devices, Inc. (AMD), obtained from Wikimedia Commons
In recent work, Baele et al. [5] propose to use multivariate components to update blocks of parameters, leading to acceptance or rejection for all of those parameters simultaneously, rather than updating all the parameters one by one in a sequential fashion using low-dimensional or scalar components [17]. To this end, the authors developed an adaptable variance multivariate normal (AVMVN) transition kernel for use in Bayesian phylogenetics, based on the work of Roberts and Rosenthal [28], to simultaneously estimate a large number of partition-specific parameters. Baele et al. [5] implemented this adaptive MCMC approach in the popular open-source BEAST software package [11], which enables this transition kernel to exploit the computational routines within the BEAGLE library [3]. The authors applied this transition kernel to a collection of clock model parameters, speciation model parameters, coalescent model parameters, and partition-specific evolutionary model parameters (which include substitution model parameters, varying rates across sites parameters, and relative rate parameters), although this kernel may find its use on parameters in many additional models.
Baele et al. [5] show that such an AVMVN transition kernel tremendously increases estimation performance over a standard set of single-parameter transition kernels. Importantly, the use of an AVMVN transition kernel requires a paradigm shift in assessing performance of transition kernels in MCMC. It is common to judge the performance of Bayesian phylogenetic software packages strictly by the time they take to evaluate proposed parameter values, often expressed in time per number of states or iterations. However, comparing transition kernels that only require a single processor core to evaluate a proposed value against transition kernels that require a collection of processor cores to evaluate a collection of proposed values simultaneously is unfair, as the latter will logically take up more time as this involves more (computational) work (on multiple processor cores). Hence, a fair comparison involves calculating the effective sample size (ESS) per time unit, as this takes into account differences in execution speed while still reporting a main statistic of interest.
We note that the approach of Baele et al. [5] has been shown to yield performance increases on CPU, but that it still needs to be tested on GPU. This is due to the specific design of the BEAGLE library [3], which evaluates a collection of BEAGLE likelihoods/ instances sequentially on GPU. Current work is underway to simplify the run time process of the BEAGLE library on GPU, allowing for simultaneous evaluation of such a collection of instances.
#### 6 Conclusions
In this chapter, we have focused on the computational challenges associated with typical analyses in the fields of phylogenetics, phylogenomics, and phylodynamics. We have provided a detailed description of how the BEAGLE library can employ multicore hardware to perform efficient likelihood evaluations and have focused on its interaction with the BEAST software package. Using benchmarks collected on a range of multicore hardware, both from the CPU and GPU market, we have shown that employing the BEAGLE high-performance computational library can considerably decrease computation time on these different systems and this for data sets with different characteristics in terms of size and complexity. The BEAGLE library allows to simultaneously compute the likelihoods for different data partitions on different CPU cores or even on different hardware devices, such as multiple GPU cards. In addition, existing data partitions can be split into multiple subpartitions to be computed in parallel across multicore hardware, yielding potentially drastic performance increases as shown in the benchmarks discussed in this chapter.
Having employed the BEAGLE library on state-of-the-art multicore hardware for a range of commonly used evolutionary models, we conclude that the combination of using BEAGLE and running analyses on powerful graphics cards aimed at the scientific computing market allows for massive performance gains for many challenging data sets. Given that sequence data sets keep growing in size and are being complemented with associated trait data, we have paid particular attention to a popular discrete trait model that parameterizes the transition rates between its states as a GLM, to allow for the inclusion of covariates to help explain transitions in the trait data. Graphics cards can be particularly useful when dealing with such models, as shown in the benchmarks presented, and we have hence presented a number of examples from the literature in which such a setup was used to perform the analyses.
Discrete phylogeography approaches (or discrete trait analyses), as the ones presented in this chapter, treat the sampling locations of the sequences as informative data, rather than uninformative auxiliary variables [9, 23, 25]. As such, the posterior distribution of the parameters given the data contains not only the likelihood of the sequences given the genealogy and substitution model but also the likelihood of the sampling locations given the genealogy and migration matrix, calculated by integrating over all possible discrete state transition histories using Felsenstein's pruning algorithm [16]. What makes this computationally demanding is a potential large number (equal to the number of branches in the phylogeny) of potentially high-dimensional (depending on the number of sampling locations) matrices, which can be parallelized across a large number of computing cores such as those found on a GPU.
Similarly, structured coalescent approaches also contain the likelihood of the sequences given the genealogy and substitution model in their posterior distribution of the parameters given the data. The use of BEAGLE will yield equal benefits to both approaches when it comes to the computation of the likelihood of the sequences given the genealogy and substitution model. However, rather than the likelihood of the sampling locations, structured coalescent approaches require computation of the probability density of the genealogy and migration history under the structured coalescent given the migration matrix and effective population sizes. To compute this density, a product of exponentials one for each of the time intervals between successive events (coalescence, sampling, or migration)—needs to be calculated. If the number of demes is sufficiently large, a GPU implementation of the probability density of the genealogy and migration history under the structured coalescent may be able to compute the contribution to this density for each of those time intervals in a highly parallel manner.
Approximations to the structured coalescent include, for example, BASTA [9], which aims to compute the probability density of the genealogy under the structured coalescent, integrated over migration histories. The computational bottleneck of this approach lies with calculating and updating the probability distribution of lineages among demes, over all lineages and over all coalescent events. This involves computing the matrix exponential of the product of each time interval duration with the backwards-intime migration rate matrix, of which the diagonal elements are defined such that the rows sum to zero. BEAGLE is equipped with a parallel thread block design for computing such finite-time transition probabilities, and to construct the finite-time transition probabilities in parallel across all lineages, and therefore has the potential to provide performance increases for structured coalescent approximations such as BASTA. However, the application software that calls upon BEAGLE needs to be implemented to rely on BEAGLE's API in order to achieve the corresponding performance increases.
Graphics cards aimed at the scientific computing market have traditionally offered roughly three times the single-precision performance compared to their double-precision performance (Tesla K40, K20X and K20). Previous generation cards, such as the Tesla K10, offered poor double-precision performance and focused solely on single-precision performance (up to 24 times their double-precision performance). The latest generation of GPUs, specifically the Tesla P100, offers tremendous double-precision performance, while single-precision performance is still twice as high. We therefore expect a doubling in performance for the computations described in this chapter if we would run them in single precision on GPU, provided that the decrease in accuracy would not lead to rescaling issues which would slow down the evaluations.
In theory, single-precision likelihood evaluations will be twice as fast as double-precision likelihood evaluations on CPU as well, but with more rescaling issues hampering performance. However, the influence of switching to single precision is more difficult to assess for CPUs, as there are a number of other factors to consider. Single-precision floating points are half the size compared to double-precision floating points and hence they may fit into a lower level of cache with a lower latency, which potentially frees up cache space to cache more (or other) data. Additionally, they require half the memory bandwidth, which frees up that bandwidth for other operations to be performed. Nevertheless, the total overall bandwidth will still be limited compared to that of powerful graphics cards and this will not suffice to bridge the performance differences between CPUs and GPUs in phylogenetics.
Finally, we have presented an interesting avenue for further increasing computational performance on multicore hardware, in the form of a new adaptive MCMC transition kernel. Traditional MCMC transition kernels generally update single parameters in a serial fashion triggering sequential likelihood evaluations on single cores. The adaptive transition kernel however updates a collection of continuous parameters simultaneously, triggering multiple likelihood evaluations in parallel on multiple cores and hence allowing for potentially large improvements in computational efficiency. Further research into this area is needed to continuously advance MCMC kernels and keep computation time manageable for a wide range of models in Bayesian phylogenetics.
#### 7 Notes
1. We have showcased the potentially impressive performance gains brought about by using BEAGLE in conjunction with powerful graphics cards. However, users sometimes complain about the poor performance gains they experience when using a GPU for their analyses, which may have to do with their GPU being not particularly suited for scientific computing. We urge readers to be cautious as to which GPU they invest in, as there is an important distinction between graphics cards aimed at the gaming market and those aimed at the scientific computing market. Computer gaming cards mainly offer tremendous single-precision performance, but typically weak doubleprecision performance. We hence advise to invest in GPUs aimed at scientific computing, offering increased accuracy and performance in double precision. As a rule, computer gaming cards have a much reduced cost compared to scientific computing cards, but we advise readers to check the technical specifications of the card before purchase.
#### 8 Exercises
1. An important aspect to getting computations—such as those discussed in this chapter—up and running, is defining which hardware is available on your computer or server. This can easily be checked using BEAST once BEAGLE has been installed. To check this when using the BEAST graphical user interface (GUI), simply check the box that says "Show list of available BEAGLE resources and Quit"; alternatively, when using the command-line interface using a BEAST Java Archive (or JAR) file—which can usually be found in the lib directory within the BEAST folder—you can simply type:
java -jar beast.jar -beagle\_info
If the path to the BEAGLE library hasn't been set up automatically, be sure to add its location to the command by adding:
java -Djava.library.path=/usr/local/lib ...
On a typical desktop system equipped with a GPU fit for scientific computing, this will yield the following output to screen:
Using BEAGLE library v2.1.2 for accelerated, parallel likelihood evaluation. 2009-2013, BEAGLE Working Group. Citation: Ayres et al (2012) Systematic Biology 61: 170-173 BEAGLE resources available: 0 : CPU Flags: PRECISION\_SINGLE PRECISION\_DOUBLE ... 1 : Intel(R) HD Graphics 530 Global memory (MB): 1536 Clock speed (Ghz): 1.05 Number of compute units: 24 Flags: PRECISION\_SINGLE COMPUTATION\_SYNCH ... 2 : Tesla K40c Global memory (MB): 11520 Clock speed (Ghz): 0.74 Number of cores: 2880 Flags: PRECISION\_SINGLE PRECISION\_DOUBLE ...
In order to determine which resource to use for your computations, it's important to look into the specifications of the GPU as listed by the hardware vendor. For example, certain GPUs will be equipped with a large number of cores and yet they're aimed at the computer gaming market, which will result in poor double-precision performance. As we have shown in Table 1, the Tesla brand is typically well suited for GPU computing, but other cards may be appropriate as well if they deliver adequate double-precision peak performance. In the output printed above, it's quite obvious that we'll be interested in running our analyses on resource 2, i.e., a GPU equipped with thousands of computing cores (resource 1 is an integrated graphics unit, mainly fit for delivering graphics output to screen).
2. Once you have located a GPU fit for scientific computing on your desktop computer or server, try to perform your analysis both on the system's CPU and GPU to compare performance. Using BEAST's GUI, the default option is to run on CPU; if you'd like to run your analysis on a suitable GPU, use BEAST's GUI to select "GPU" where it says "Prefer use of:." However, most desktop computers don't come equipped with powerful graphics cards, and most often servers aimed at highperformance computing (HPC) will be used for performing these types of computations. As such servers are typically instructed using a command-line interface, BEAST offers the possibility to assign computations to one or more specific GPUs. Using the system described here, it would hence make sense to run your analysis on resource 2, which can be done as follows:
java -jar beast.jar -beagle\_gpu -beagle\_order 2 data.xml
Note that not specifying -beagle\_order will result in the analysis being run on the system's CPU, i.e., resource 0. Additionally, when employing a GPU for your analyses, adding the -beagle\_gpu argument is highly advised. Many different combinations of using resources arise when your data set is partitioned into multiple subsets, for example if your data is partitioned according to gene and/or codon position. In such cases, it may be beneficial to split those partitions onto multiple resources by using the -beagle\_order command-line option. For example, the Ebola virus data set (without trait data) has four partitions; it may be useful (although this depends on the actual hardware and needs to be tested) to compute the likelihood of one partition on the CPU (i.e., resource 0) and the other three likelihoods on the GPU (i.e., resource 2). This can be done as follows:
java -jar beast.jar -beagle\_gpu -beagle\_order 0,2,2,2 ebola.xml
3. In some cases, such as for example the carnivores data set analyzed in this chapter, only one (sequence) data partition is available. On CPU, drastic performance improvements can still be achieved by using a BEAGLE feature that allows to split up a data partition into multiple subsets, as can be seen in Fig. 5. This approach will lead to performance increases on most CPU systems, as many laptops now come equipped with 4-core processors; this can hence easily be tested on the system you're currently using. To split a (sequence) data partition into two subsets, you can use the following command:
java -jar beast.jar -beagle\_instances 2 carnivores.xml
To generate the results in Fig. 5, we have used this approach to split the data set into 2, 4, 8, 12, 16, 20, and 24 subpartitions, increasing performance every step of the way. Note that on GPU, this approach will only lead to an increased overhead and hence worse performance compared to keeping the single data partition, as all the likelihood calculations end up on the same (GPU) device. With multiple GPUs in a system however, this can also lead to drastic performance improvements, as shown throughout this chapter.
#### Acknowledgements
The authors acknowledge support from the European 2020 Union Seventh Framework Programme for research, technological development, and demonstration under Grant Agreement no. 278433- PREDEMICS. The VIROGENESIS project receives funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No. 634650. The COMPARE project receives funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No. 643476. This work is supported by National Science Foundation (NSF) [DMS 1264153 to M.A.S.; DBI 0755048 and DBI 1356562 to Michael P. Cummings who supported D.L.A. and is also acknowledged here] and National Institutes of Health [R01 AI117011 and R01 HG006139 to M.A.S.]. G.B. acknowledges support from the Interne Fondsen KU Leuven/Internal Funds KU Leuven and of a Research Grant of the Research Foundation— Flanders (FWO; Fonds Wetenschappelijk Onderzoek—Vlaanderen). We acknowledge funding of the "Bijzonder Onderzoeksfonds," KU Leuven (BOF, No. OT/14/115), and the Research Foundation—Flanders (FWO, G0D5117N and G0B9317N). We gratefully acknowledge support from the NVIDIA Corporation with the donation of parallel computing resources used for this research. The computational resources and services used in this work were provided by the VSC (Flemish Supercomputer Center), funded by the Research Foundation—Flanders (FWO) and the Flemish Government—Department EWI.
#### References
nonsynonymous nucleotide substitution rates, with application to the chloroplast genome. Mol Biol Evol 11(5):715–724
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Chapter 24
#### Scalable Workflows and Reproducible Data Analysis for Genomics
#### Francesco Strozzi, Roel Janssen, Ricardo Wurmus, Michael R. Crusoe, George Githinji, Paolo Di Tommaso, Dominique Belhachemi, Steffen Mo¨ ller, Geert Smant, Joep de Ligt, and Pjotr Prins
#### Abstract
Biological, clinical, and pharmacological research now often involves analyses of genomes, transcriptomes, proteomes, and interactomes, within and between individuals and across species. Due to large volumes, the analysis and integration of data generated by such high-throughput technologies have become computationally intensive, and analysis can no longer happen on a typical desktop computer.
In this chapter we show how to describe and execute the same analysis using a number of workflow systems and how these follow different approaches to tackle execution and reproducibility issues. We show how any researcher can create a reusable and reproducible bioinformatics pipeline that can be deployed and run anywhere. We show how to create a scalable, reusable, and shareable workflow using four different workflow engines: the Common Workflow Language (CWL), Guix Workflow Language (GWL), Snakemake, and Nextflow. Each of which can be run in parallel.
We show how to bundle a number of tools used in evolutionary biology by using Debian, GNU Guix, and Bioconda software distributions, along with the use of container systems, such as Docker, GNU Guix, and Singularity. Together these distributions represent the overall majority of software packages relevant for biology, including PAML, Muscle, MAFFT, MrBayes, and BLAST. By bundling software in lightweight containers, they can be deployed on a desktop, in the cloud, and, increasingly, on compute clusters.
By bundling software through these public software distributions, and by creating reproducible and shareable pipelines using these workflow engines, not only do bioinformaticians have to spend less time reinventing the wheel but also do we get closer to the ideal of making science reproducible. The examples in this chapter allow a quick comparison of different solutions.
Key words Bioinformatics, Evolutionary biology, Big data, Parallelization, MPI, Cloud computing, Cluster computing, Virtual machine, MrBayes, Debian Linux, GNU Guix, Bioconda, CWL, Common Workflow Language, Guix Workflow Language, Snakemake, Nextflow
Availability: All included software, scripts, and Docker images are based on free and open-source software and can be found at https://github.com/EvolutionaryGenomics/scalability-reproducibility-chapter.
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_24, © The Author(s) 2019
#### 1 Introduction
1.1 Overview In this chapter, we show how to create a bioinformatics pipeline using four workflow systems: CWL, GWL, Snakemake, and Nextflow. We show how to put them together, so you can adapt it for your own purposes while discussing in the process the different approaches. All scripts and source code can be found on GitHub. The online material allows a direct comparison of how such workflows are assembled with their syntax.
> Due to large volumes, the analysis and integration of data generated by high-throughput technologies have become computationally intensive, and analysis can no longer happen on a typical desktop computer. Researchers therefore are faced with the need to scale analyses efficiently by using high-performance compute clusters or cloud platforms. At the same time, they have to make sure that these analyses run in a reproducible manner. And in a clinical setting, time becomes an additional constraint, with motivation to generate actionable results within hours.
> In the case of evolutionary genomics, lengthy computations are often multidimensional. Examples of such expensive calculations are Bayesian analyses, inference based on hidden Markov models, and maximum likelihood analysis, implemented, for example, by MrBayes [1], HMMER [2], and phylogenetic analysis by maximum likelihood (PAML) [3]. Genome-sized data, or Big Data [4, 5], such as produced by high-throughput sequencers, as well as growing sample size, such as from UK Biobank, the Million Veterans Program, and the other large genome-phenome projects, are exacerbating the computational challenges, e.g., [6].
> In addition to being computationally expensive, many implementations of major algorithms and tools in bioinformatics do not scale well. One example of legacy software requiring lengthy computation is Ziheng Yang's CodeML implementation of PAML [3]. PAML finds amino acid sites that show evidence of positive selection using dN/d<sup>S</sup> ratios, i.e., the ratio of nonsynonymous and synonymous substitution rate. For further discussion see also Chapter. 12. Executing PAML over an alignment of 100 sequences may take hours, sometimes days, even on a fast computer. PAML (version 4.x) is designed as a single-threaded process and can only exploit one central processing unit (CPU) to complete a calculation. To test hundreds of alignments, e.g., different gene families, PAML is invoked hundreds of times in a serial fashion, possibly taking days on a single computer. Here, we use PAML as an example, but the idea holds for any software program that is CPU bound, i.e., the CPU speed determines program execution time. A CPU bound program will be at (close to) 100% CPU usage. Many legacy programs are CPU bound and do not scale by themselves.
Most bioinformatics (legacy) programs today do not make effective use of multi-core computers
The reason most bioinformatics software today does not make full use of multicore computers or GPUs is because writing such software is difficult. (See also the text box below for a further treatment of this topic; see Box 1.)
A common parallelization strategy in bioinformatics is to start with an existing nonparallel application and run it by dividing data into independent units of work or jobs which run in parallel and do not communicate with each other. This is also known as an "embar rassingly parallel" solution, and we will pursue this below.
1.2 Parallelization in the Cloud Cloud computing allows the use of "on-demand" CPUs accessible via the Internet and is playing an increasingly important role in bioinformatics. Bioinformaticians and system administrators previously had to physically install and maintain large compute clusters to scale up computations, but now cloud computing makes it possible to rent and access CPUs, GPUs, and storage, thereby enabling a more flexible concept of on-demand computing [7]. The cloud scales and commoditizes cluster infrastructure and management and, in addition, allows users to run their own operating system, usually not true with existing cluster and GRID infrastructure (a GRID is a heterogeneous network of computers that act together). A so-called hypervisor sits between the host operating system and the guest operating system, and it makes sure they are clearly separated while virtualizing host hardware. This means many guests can share the same machine that appears to the users as a single machine on the network. This allows providers to efficiently allocate resources. Containers are another form of light virtualization that is now supported by all the main cloud providers, such as Google, Microsoft, Rackspace OpenStack, and Amazon (AWS). Note that only OpenStack is available as free and open-source software.
> An interesting development is that of portable batch systems (PBS) in the cloud. PBS-like systems are ubiquitous in highperformance computing (HPC). Both Amazon EC2 and Microsoft Cloud offer batch computing services with powerful configuration options to run thousands of jobs in the cloud while transparently automating the creation and management of virtual machines and containers for the user. As an alternative, Arvados is an open-source product specifically aimed at bioinformatics applications that makes the cloud behave as if it is a local cluster of computers, e.g., [8].
> At an even higher level, MapReduce is a framework for distributed processing of huge datasets, and it is well suited for problems using large number of computers [9]. The map step takes a dataset and splits it into parts and distributes them to worker nodes. Worker nodes can further split and distribute data. At the
reduce step, data is combined into a result, i.e., it is an evolved scatter and gather approach. An API is provided that allows programmers to access functionality. The Apache Hadoop project includes a MapReduce implementation and a distributed file system [10] that can be used with multiple cloud providers and also on private computer clusters. Another similar example is the Apache Spark project based on resilient distributed datasets (RDD)—a fault-tolerant collection of elements that can be accessed and operated on in parallel. The advantage of such higher-level systems is that they go well beyond hardware virtualization: not only the hardware infrastructure but also the operating system, the job scheduler, and resource orchestration are abstracted away. This simplifies data processing, parallelization, and the deployment of virtual servers and/or containers. The downside is that users have less control over the full software stack and often needs to program and interact with an application programmers interface (API). Overall, in the last decade, both commercial and noncommercial software providers have made cloud computing possible. Bioinformaticians can exploit these services. 1.3 A Pipeline for the Cloud To create a bioinformatics pipeline, it is possible to combine remote cloud instances with a local setup. Prepare virtual machines or containers using similar technologies on a local network, such as a few office computers or servers, and then use these for calculations in the cloud when an analysis takes too long. The cloud computing resources may, for instance, support a service at peak usage, while regular loads are met with local infrastructure (i.e., burst compute). New ideas can be developed and pre-evaluated using modest in-house setups and then scaled to match the most demanding work. Cloud services can be used for burst computing – enabling local clusters to be much smaller – as small as a single computer In the following sections, we will provide instructions to deploy applications, and we will show how the use of workflow systems and reproducible environments can greatly simplify running scalable workflows on different environments, including the cloud. 1.4 Parallelization of Applications Using a Workflow In case of embarrassingly parallel applications, programs are run independently as separate processes which do not communicate with each other. This is also a scatter and gather approach, i.e., inputs split into several jobs are fed into each process by the user. Job outputs are collected and collated. In bioinformatics, such tasks are often combined into computational pipelines. With the PAML example, each single job can be based on one alignment, potentially giving linear speed improvements by distributing jobs across multiple CPUs and computers. In other words, the PAML software, by itself, does not allow calculations in parallel, but it is possible to parallelize multiple runs of PAML by splitting the dataset. The downside of this approach is the deployment and configuration of pipeline software, as well as the management and complexity of splitting inputs and the collecting and collating of outputs. Also, pipelines are potentially fragile, because there is no real interprocess communication. For example, it is hard to predict the consequences of a storage or network error in the middle of a week- or month-long calculation.
Even for multithreaded applications that make use of multiple CPUs, such as BLAST and MrBayes, it is possible to scale up calculations by using a workflow. For example, MrBayes-MPI version 3.1.2 does not provide between-machine parallelization and is therefore machine bound, i.e., the machine's performance determines the total run time. Still, if one needs to calculate thousands of phylogenetic trees, discrete jobs can be distributed across multiple machines. A similar approach is often used for large-scale BLAST analyses over hundreds of thousands of sequences.
A pipeline typically consists of linear components, where one software tool feeds into another, combined with a scattering of jobs across nodes and a gathering and collation of results.
In existing compute clusters, to distribute work across nodes, portable batch system (PBS) schedulers are used, such as Slurm [11]. Many pipelines in bioinformatics are created in the form of Bash, Perl, or Python scripts that submit jobs to these schedulers. Such scripted pipelines have the advantage that they are easy to write and adaptable to different needs. The downside is that they are hard to maintain and not very portable, since the description of the environment and the software packages are not part of these scripts, reducing or completing preventing the reproducibility of a certain analysis in a different context. This has led to the current state of affairs in bioinformatics that it is surprisingly hard to share pipelines and workflows. As a result much effort is spent reinventing the wheel.
Most existing bioinformatics pipelines cannot easily be shared and reproduced
In recent years, a number of efforts have started to address the problem of sharing workflows and making analyses reproducible. One example is the Common Workflow Language (CWL), a specification for describing analysis workflows and tools in a way that makes them portable and scalable across a variety of environments—from workstations to cluster, cloud, and HPC environments. CWL is a large bioinformatics community effort. Different platforms support CWL, including Arvados, Galaxy, and Seven Bridges [8].
A second workflow language is the Guix Workflow Language (GWL) built on top of the GNU Guix software deployment system. GWL aims to provide a deterministic and bit-reproducible analysis environment.
A third workflow language and orchestrator, Nextflow, allows scalable and reproducible scientific workflows to run seamlessly across multiple platforms from local computers to HPC clusters and the cloud, offering a concise and expressive DSL to describe complex workflows. Nextflow is routinely used in organizations and institutes, such as the Roche Sequencing, the Wellcome Trust Sanger Institute, and the Center for Genomic Regulation (CRG) Nextflow workshop.
Forth there is Snakemake, another widely used workflow manager system, written in Python and inspired by GNU Make. It allows for the composition of workflows based on a graph of rules whose execution is triggered by the presence, absence, or modification of expected files and directories.
It is interesting to note that all these workflow languages and systems originated in bioinformatics. It suggests that in this rapidly growing field, the increasing computational needs and moreover the diverse demands made more formal solutions a necessity. It also suggests that existing workflow engines used in astronomy and physics, for example, have different requirements.
#### Box 1: Understanding Parallelization
Parallel computing is related to concurrent computing. In parallelized computing, a computational task is typically broken down in several, often many, very similar subtasks that can be processed independently and whose results are combined afterward, upon completion, i.e., a simple scatter and gather approach. In contrast, in distributed computing, the various processes often do not address related tasks; or when they do, the separate tasks may have a varied nature and often require some interprocess communication during execution. The latter is also a hallmark of supercomputing where compute nodes have high-speed connections.
In the bioinformatics space, we usually discuss embarrassingly parallel computing which means similar tasks are distributed across multiple CPUs without interprocess communication. This can be among multiple cores within a single processor, a multiprocessor system, or a network of computers, a so-called compute cluster.
Even so, parallel multicore programming easily becomes complex. Typically, parallel programming has to deal with extra data and control flow; it has to deal with deadlocks,
where depending tasks wait for each other forever and, with race conditions, where tasks try to modify a shared resource (e.g., a file) at the same time resulting in a loss of data or an undetermined condition. This introduces additional complexity in software development, bug hunting, and code maintenance. Typically it takes more time to debug such code than to write it.
Writing programs that fully utilize multi-core architectures is hard
Not only is parallel programming intrinsically complicated; programmers also have to deal with communication overheads between parallel threads. MrBayes, for example, a program for calculating phylogenetic trees based on Bayesian analysis, comes with MPI support. MPI is a message-based abstraction of parallelization, in the form of a binary communication protocol implemented in a C programming library [12]. In some cases the parallelized version is slower than the single CPU version. For example, the MPI version calculates each Markov chain in parallel, and the chains need to be synchronized with each other, in a "scatter and gather" pattern. The chains spend time waiting for each other in addition to the communication overheads introduced by MPI itself. Later MrBayes adopted a hybrid use of coarsegrained OpenMPI and fine-grained use of pthreads or OpenMP leading to improved scalability, e.g., [13].
Another example of communication overhead is with the statistical programming language R [14], which does not have native threading support built into the language. One possible option is to use an MPI-based library which only allows coarse-grained parallelization from R, as each parallelized R thread starts up an R instance, potentially introducing large overheads, both in communication time and memory footprint. For a parallelized program to be faster than its single-threaded counterpart, these communication overheads have to be dealt with.
Parallelization in R is coarse-grained with large overhead
The need for scaling up calculations on multi-CPU computers has increased the interest in a number of functional programming languages, such as Erlang [15], Haskell [16], Scala [17], and Julia [18]. These languages promise to ease writing parallel software by introducing a higher level of abstraction of parallelization, combined with immutable data, automatic garbage collection, and good debugging support [5, 19]. For example, Erlang and Scala rely on Actors as an abstraction of parallelization and make reasoning about fine-grained parallelization easier and therefore less error prone.
Actors were introduced and explored by Erlang, a computer language originally designed for highly parallelized telecommunications computing. To the human programmer, each Actor appears as a linear piece of programming and is parallelized without the complexity of locks, mutexes, and semaphores. Actors allow for parallelization in a manageable way, where lightweight threads are guaranteed to be independent and each has a message queue, similar to MPI. Actors, however, are much faster, more intuitive, and, therefore, probably, safer than MPI. Immutable data, when used on a single multi-CPU computer, allows fast passing of data by reference between Actors. When a computer language supports the concept of immutability, it guarantees data is not changed between parallel threads, again making programming less error prone and easier to structure. Actors with support for immutable data are implemented as an integral part of the programming language in Erlang, Haskell, Scala, Elixir, and D [20].
Another abstraction of parallelized programming is the introduction of goroutines, part of the Go programming language [21]. Where MPI and Actors are related to a concept of message passing and mail boxes, goroutines are more closely related to Unix named pipes. Goroutines also aim to make reasoning about parallelization easier, by providing a pipe where data goes in and results come out, and this processing happens concurrently without use of mutexes, making it easier to reason about linear code. Goroutines are related to communicating sequential processes (CSP), the original paper by Tony Hoare in 1978 [22]. Meanwhile, recent practical implementations are driven by the ubiquity of cheap multicore computers and the need for scaling up. A Java implementation of CSP exists, named JCSP [23], and a Scala alternative named CSO [24]. Go made goroutines intuitive and a central part of the strongly typed compiled language.
Erlang, Elixir, Haskell, Scala, Julia, Go and D are languages offering useful abstractions and tools for multi-core programming
It is important to note that the problems, ideas, and concepts of parallel programming are not recent. They have been an important part of computer science theory for decennia. We invite the reader interested in parallel programming to read up on the languages that have solid built-in support high-level parallelization abstractions, in particular, Scala [17], Go [21], and D [20].
1.4.1 GPU Programming Another recent development is the introduction of GPU computing or "heterogeneous computing" for offloading computations. Most GPUs consist of an array of thousands of cores that can execute similar instructions at the same time. Having a few thousand GPU cores can speed up processing significantly. Programming GPUs, however, is a speciality requiring specialized compilers and communication protocols, and there are many considerations, not least the I/O bottleneck between the main memory and the GPU's dedicated RAM [5]. Even so, it is interesting to explore the use of GPUs in bioinformatics since they come with almost every computer today and clusters of GPU can increasingly be found in HPC infrastructure and in the cloud, alike. With the advent of "deep neural networks" and the general adoption of machine learning techniques for Big Data, GPUs have become a mainstream technology in data mining.
#### 2 Package Software in a Container
Container technologies, such as Docker and Singularity, have gained popularity because they have less overhead than full virtual machines (VMs) and are smaller in size [24]. Containers are fully supported by the major cloud computing providers and play an important role for portability across different platforms.
Adoption of container solutions on HPC has been problematic, mostly because of security concerns. Singularity [26] offers a decentralized environment encapsulation that works in user space and that can be deployed in a simpler way since no root privileges are required to execute tools provided with Singularity. That is, Singularity containers can be created on a system with root privileges but run on a system without root privileges—though it requires some special kernel support. Docker containers can be imported directly in Singularity, so when we present how to build Docker container images in the following sections, the reader should be aware that the same images can also be used with Singularity. Singularity is slowly being introduced in HPC setups [27].
GNU Guix also has support for creating and running Linux containers. One interesting benefit is that, because the software packaging system is read-only and provides perfect isolation, containers automatically can share specific software running on the underlying system, making running containers even lighter and extremely fast.
In this section we discuss three popular software distribution systems for Linux: Debian GNU/Linux (Debian), GNU Guix, and Conda can be used together on a single system allowing access to most bioinformatics software packages in use today. In this section we bundle tools that can be deployed in a Docker image, which can run on a single multicore desktop computer and a compute cluster and in the cloud.
software and even different variants or combinations of software, e.g., Apache web server with SSL and without SSL compiled on a single system.
As of November 2017, GNU Guix provides over 6500 soft ware packages, including a wide range of dedicated scientific software for bioinformatics, statistics, and machine learning
#### 2.2.1 Create a Docker Image with GNU Guix GNU Guix has native support for creating Docker images. Creating a Docker image with GNU Guix is a one liner:
guix pack -f docker -S /bin=bin paml clustal-omega which creates a reproducible Docker image containing PAML and Clustal Omega [33], including all of their runtime dependencies. Guix makes it very easy to write new package definitions using the Guile language (a LISP). If you want to include the definition of your own packages (that are not in Guix main line), you can include them dynamically. This is how we add pal2nal [34] in below GWL workflow example (see Subheading 3.3 below).
```
FROM conda/miniconda3
RUN conda config --add channels conda-forge
RUN conda install -y perl=5.22.0
RUN conda install -y -c bioconda paml=4.9 clustalo=1.2.4
wget=1.19.1
ADD pal2nal.pl /usr/local/bin/pal2nal.pl
RUN chmod +x /usr/local/bin/pal2nal.pl
```
Note that we provide the version numbering of the packages. If you want to build this container, you can use the Dockerfile provided in the GitHub repository [30] and then run:
docker build -t scalability .
We also added a pre-built container image on Docker Hub [31].
Conda can also be used outside any container system to install the software directly on a local computer or cluster. To do that first install the Miniconda package https://conda.io/miniconda.html, and then you can create a separate environment with the necessary software to run the workflows. Following is an example to set up a working environment:
conda create -n scalability source activate scalability conda config --add channels conda-forge conda install -y perl=5.22.0 conda install -y -c bioconda paml=4.9 clustalo=1.2.4 wget=1.19.1 wget http://www.bork.embl.de/pal2nal/distribution/pal2nal. v14.tar.gz tar xzvf pal2nal.v14.tar.gz sudo cp pal2nal.v14/pal2nal.pl /usr/local/bin sudo chmod +x /usr/local/bin/pal2nal.pl
Note that we use Miniconda here to bootstrap Bioconda. Bioconda can be bootstrapped in other ways. One of them is GNU Guix which contains a Conda package.
2.4 A Note on Software Licenses All above packaging systems use free and open-source software (FOSS) released under a permissible license, i.e., a license permitting the use, modification, and distribution of the source code for any purpose. This is important because it allows software distributions to distribute all included software freely. Software that is made available under more restrictive licenses, such as for "academic nonprofit use only," cannot be distributed in this way. An example is PAML that used to have such a license. Only when it was changed PAML got included into Debian, etc. Also, for this book chapter, we asked the author of pal2nal to add a proper license. After adding the GPLv2, it became part of the Debian distribution; see also https://tracker.debian.org/pkg/pal2nal. This means that above Docker scripts can be updated to install the pal2nal Debian package.
> When you use scientific software, always check the type of license under which it is provided, to understand what you can or cannot do with it. When you publish software, add a license along with your code, so others can use it and distribute it.
> Typical licenses used in bioinformatics are MIT (Expat) and BSD, which are considered very permissive, and also GPL and the Apache License, which are designed to grant additional protections with regard to derivative works and patentability. Whenever possible, free software licenses such as mentioned above are encouraged for scientific software. Check the guidelines of your employer and funding agencies.
#### 3 Create a Scalable and Reusable Workflow
3.1 Example Workflow We have created a number of examples to test a scalable and reproducible workflow, the full code, and examples that are available on GitHub [30]. In this case putative gene families of the oomycete Phytophthora infestans are tested for evidence of positive selection. P. infestans is a single-cell pathogen, which causes late blight of potato and tomato. Gene families under positive selection pressure may be involved in protein–protein interactions and are potentially of interest for fighting late blight disease.
As an example the P. infestans genome data [36] was fetched from http://www.broadinstitute.org/annotation/genome/phyto phthora\_infestans/MultiDownloads.html, and predicted genes were grouped by \name{blastclust} using 70% identity (see also Chapter. 21). This resulted in 72 putative gene families listed on the online repository on GitHub [30].
The example workflow aligns amino acid sequences using Clustal Omega, creates a neighbor joining tree, and runs CodeML from the PAML suite. The following is one example to look for evidence of positive selection in a specific group of alignments:
```
clustalo -i data/clusterXXXXX/aa.fa --guidetree-out=data/
clusterXXXXXX/aa.ph > data/clusterXXXXXX/aa.aln
pal2nal.pl -output paml data/clusterXXXXX/aa.aln data/clus-
terXXXXX/nt.fa > data/clusterXXXXX/alignment.phy
cd data/clusterXXXXX
Codeml ../paml0-3.ctl
```
First we align amino acid with Clustal Omega, followed by translation to a nucleotide alignment with pal2nal. Next we test for evidence of positive selection using PAML's \name{Codeml} with models M0–M3. Note that the tools and settings used here are merely chosen for educational purposes. The approach itself here may result in false positives, as explained by Schneider et al. [37]. Also, PAML is not the only software that can test for evidence of positive selection, for example, the HyPhy molecular evolution and statistical sequence analysis software package contains similar functionality and uses MPI to parallelize calculations [38]. PAML is used here because it is a reference implementation and is suitable as an example how a legacy single-threaded bioinformatics application can be parallelized in a workflow.
In the next section, different workflow systems are presented that can be used to run the described analysis: in a scalable and reproducible manner, locally on a desktop, on a computer cluster, or in the cloud. All the code and data to run these examples is available on GitHub [30]. To load the code on your desktop, clone the git repository locally. The examples can be executed from the repository tree:
```
git clone https://github.com/EvolutionaryGenomics/scalabil-
ity-reproducibility-chapter.git
```
#### 3.2 Common Workflow Language
Common workflow language (CWL, http://www.commonwl. org/) is a standard for describing workflows that are portable across a variety of computing platforms [39]. CWL is a specification and not a software in itself though it comes with a reference implementation which can be run with Docker containers. CWL promotes an ecosystem of implementations and supporting systems to execute the workflows across multiple platforms. The promise is that when you write a workflow for, e.g., Arvados, it should also run on another implementation, e.g., Galaxy.
Given that CWL takes inspiration from previously developed tools and GNU Make in particular [40], the order of execution in a CWL workflow is based on dependencies between the required tasks. However unlike GNU Make, CWL tasks are defined to be isolated, and you must be explicit about inputs and outputs. The benefits of explicitness and isolation are flexibility, portability, and scalability: tools and workflows described with CWL can transparently leverage software deployment technologies, such as Docker, and can be used with CWL implementations from different vendors, and the language itself can be applied to describe large-scale workflows that run in HPC clusters, or the cloud, where tasks are scheduled in parallel across many nodes.
CWL workflows are written in JSON or YAML formats. A workflow consists of blocks of steps, where each step in turn is made up of a task description that includes the inputs and outputs of the task itself. The order of execution of the tasks is determined automatically by the implementation engine. In the GitHub repository, we show an example of a CWL workflow to describe the analysis over the protein alignments. To test the workflow, you will need the CWL reference runner implementation:
pip install cwlref-runner and then to run the example from the repository tree:
CWL/workflow.cwl --clusters data
To run the CWL workflow on a grid or cloud multi-node system, we can install another CWL implementation, this one built upon the toil platform [41]:
pip install toil[cwl] toil-cwl-runner CWL/workflow.cwl --clusters data
Fig. 1 Workflow automatically generated from the CWL schema displays how PAML's Codeml receives inputs from two sources and outputs the dN/d<sup>S</sup> information. A workflow engine figures out that it has to run clustal first, followed by pal2nal and Codeml as a linear sequence. For each input, the job can be executed in parallel
CWL comes with extra tooling, such as visualization of CWL workflows (Fig. 1). See view.commonwl.org for more examples.
The Guix Workflow Language (GWL) extends the functional package manager GNU Guix [32] with workflow management capabilities. GNU Guix provides an embedded domain-specific language (EDSL) for packages and package composition. GWL extends this EDSL with processes and process composition.
In GWL, a process describes the computation, for example, running the clustalo program. A workflow in the GWL describes how processes relate to each other. For example, the Codeml program can only run after both clustalo and pal2nal finished successfully.
The tight coupling of GWL and GNU Guix ascertains that not only the workflow is described rigorously but also the deployment of the programs on which the workflow depends.
To run the GWL example, you need to install GNU Guix (https://www.gnu.org/software/guix/manual/html\_node/ Binary-Installation.html) and the GWL installed on your computer. Once GNU Guix is available, installing GWL can be done using:
guix package -i gwl
#### 3.3 Guix Workflow Language
The example can be run using:
```
cd scalability-reproducibility-chapter/GWL
guix workflow -r example-workflow
```
GWL also implements execution engines to offload computation on compute clusters, allowing it to scale. The process engines can use the package composition capabilities of GNU Guix to create the desirable form of software deployment—be it installing programs on the local computer or creating an application bundle, a Docker image, or a virtual machine image.
Running our example on a cluster that has Grid Engine:
guix workflow -r example-workflow -e grid-engine
GNU Guix + GWL can ensure full reproducibility of an analysis, including all software dependencies—all the way down to glibc. GNU Guix computes a unique string, a hash, on the complete set of inputs and the build procedure of a package. It can guarantee that a package is built with the same source code, dependency graph, and the same build procedure, and produces identical output. In GWL for each process and workflow, a hash is computed of the packages, the procedure, and the execution engine. By comparing hashes it is not only possible to compare whether the workflow is running using the exact same underlying software packages, and using the same procedures, but also the full graph of dependencies can be visualized. To obtain such an execution plot:
```
guix package -i graphviz
guix workflow -g example-workflow | dot -Tpdf > example-
workflow.pdf
```
Note that, unlike the other workflow solutions discussed here, GWL does not use the time stamps of output files. The full dependency graph is set before running the tools, and it only needs to check whether a process returns an error state. This means that there are no issues around time stamps and output files do not have to be visible to the GWL engine.
#### 3.4 Snakemake Snakemake [42] is a workflow management system that takes inspiration from GNU Make [40], a tool to coordinate the compilation of large programs consisting of interdependent source files (https://snakemake.readthedocs.io/en/stable/).
Snakemake provides a DSL that allows the user to specify generator rules. A rule describes the steps that need to be performed to produce one or more output files, such as running a shell script. These output files may be used as inputs to other rules. The workflow is described as a graph in which the nodes are files (provided input files, generated intermediate files, or the desired output files) and the edges are inferred from the input/output interdependencies of connected rules.
When a user requests a certain file to be generated, Snakemake matches the file name against concrete or wildcard rules, traverses the graph from the target file upward, and begins processing the steps for every rule for which no new output file is available. Whether or not an output file is considered new depends on its time stamp relative to the time stamp of prerequisite input files. In doing so, Snakemake only performs work that has not yet been done or for which the results are out of date, just like GNU Make. Snakemake can be configured to distribute jobs to batch systems or to run jobs on the local system in parallel. The degree of parallelization depends on the dependencies between rules.
Snakemake is written in Python and allows users to import Python modules and use them in the definition of rules, for example. It also has special support for executing R scripts in rules, by exposing rule parameters (such as inputs, outputs, concrete values for wildcards, etc.) as an S4 object that can be referenced in the R script.
Snakemake provides native support for the Conda package manager. A rule may specify a Conda [35] environment file describing a software environment that should be active when the rule is executed. Snakemake will then invoke Conda to download the required packages as specified in the environment file. Alternatively, Snakemake can interface with an installation of the Singularity container system [26] and execute a rule within the context of a named application bundle, such as a Docker image.
To run the Snakemake workflow, you need to install Snakemake (example showed with Conda):
conda install -y -c bioconda snakemake=4.2.0
And then to run the example from the repository tree:
cd Snakemake snakemake
3.5 Nextflow Nextflow [43] is a framework and an orchestration tool that enables scalable and reproducible scientific workflows using software containers (https://www.nextflow.io/). It is written in the Groovy JVM programming language [44] and provides a domain-specific language (DSL) that simplifies writing and deploying complex workflows across different execution platforms in a portable manner.
> A Nextflow pipeline is described as a series of processes, where each process can be written in any language that can be executed or
interpreted on Unix-like operating systems (e.g., Bash, Perl, Ruby, Python, etc.). A key component of Nextflow is the dataflow programming model, which is a message-based abstraction for parallel programming similar to the CSP paradigm (see [23]). The main difference between CSP and dataflow is that in the former, processes communicate via synchronous messages, while in the latter, the messages are sent in an asynchronous manner. This approach is useful when deploying large distributed workloads because it has latency tolerance and error resilience. In practical term the dataflow paradigm uses a push model in which a process in the workflow sends its outputs over to the downstream processes that waits for the data to arrive before starting their computation. The communication between processes is performed through channels, which define inputs and outputs for each process. Branches in the workflow are also entirely possible and can be defined using conditions that specify if a certain process must be executed or not depending on the input data or on user defined parameters.
The dataflow paradigm is the closest representation of a pipeline idea where, after having opened the valve at the beginning, the flow progresses through the pipes. But Nextflow can handle this data flow in a parallel and asynchronous manner, so a process can operate on multiple inputs and emit multiple outputs at the same time. In a simple workflow where, for instance, there are 100 nucleotide sequences to be aligned with the NCBI NT database using BLAST, a first process can compute the alignment of the 100 sequences independently and in parallel, while a second process will wait to receive and collect each of the outputs from the 100 alignments to create a final results file. To allow workflow portability, Nextflow supports multiple container technologies such as Docker and Singularity and integrates natively with Git and popular code sharing platforms, such as GitHub. This makes it possible to precisely prototype self-contained computational workflows, tracking also all the modifications over time and ensuring the reproducibility of any former configuration. Nextflow allows executing workflows across different computing platforms by supporting several cluster schedulers (e.g., SLURM, PBS, LSF and SGE) and allowing direct execution on the Amazon cloud (AWS), using services, such as AWS Batch or automating the creation of a compute cluster in the cloud for the user.
To run the Nextflow example, you need to have Java 8 and a Docker engine (1.10 or higher) installed. Next install Nextflow with:
curl -s https://get.nextflow.io | bash
Run the example from the repository tree:
./nextflow run Nextflow/workflow.nf -with-docker evolutionarygenomics/scalability
To save the graph of the executed workflow, it is sufficient to add the option "-with-dag workflow.pdf." The same example can also be run without Docker if the required packages have been installed locally following the Bioconda or Guix examples. In this case you can omit the "-with-docker" instruction. To run the example on a compute cluster or in the cloud, it is sufficient to specify a different executor (e.g., sge or awsbatch) in the Nextflow configuration file and ensure that those environments are configured to properly work with the Docker container.
#### 4 Discussion
In this chapter we show how to describe and execute the same analysis using a number of workflow systems and how these follow different approaches to tackle execution and reproducibility issues. It is important to assess underlying design choices of these solutions and also to look at the examples we provide online. Even though it may look attractive to opt for the simplest choices, it may be that the associated maintenance burden may be cause for regret later.
The workflow tools introduced in this chapter offer direct integration of software packages. The overall advantage of the bundling software approach is that when software deployment and execution environment are controlled, the logic of the analysis pipeline can be developed separately using descriptive workflows. This separation allows communities to build best practice shareable pipelines without worrying too much about individual system architectures and the underlying environments. An example is the effort by the Global Alliance for Genomics and Health (GA4GH, https://www.ga4gh.org) to develop and share best practice analysis workflows with accompanying container images [45].
In this chapter we also discussed the scaling up of computations through parallelization. In bioinformatics, the common parallelization strategy is to take an existing nonparallel application and divide data into discrete units of work, or jobs, across multiple CPUs and clustered computers. Ideally, running jobs in parallel on a single multicore machine shows linear performance increase for every CPU added, but in reality it is less than linear [46]. Resource contention on the machine, e.g., disk or network I/O, may have processes wait for each other. Also, the last, and perhaps longest, running job causes total timing to show less than linear performance, as the already finished CPUs are idle. In addition to the resource contention on a single machine, the network introduces latencies when data is moved around.
Running the example workflow in the cloud has similar performance and scalability compared to running it on a local infrastructure, after adjusting for differences in hardware and network speeds. Cloud computing is an attractive proposition for scaling up calculation jobs and storing data. Cloud prices for virtual servers and data storage have decreased dramatically, and the possibility of using spot or preemptible instances (i.e., virtual servers that can be priced down to 70% or 80% the normal price but that can be shut down in any moment by the cloud provider) is making cloud computing solutions competitive for high-performance and scientific computing. Cloud essentially outsources hardware and related plumbing and maintenance. Sophisticated tooling allows any researcher to run software in the cloud. We predict an increasing number of groups and institutes will move from large-scale HPC clusters toward tight HPC cluster solutions that can handle continuous throughput with burst compute in the cloud.
Reproducibility is a prime concern in science. Today several solutions are available to address reproducibility concerns. Systems such as Docker and Singularity are built around bundling binary applications and executing them in a container context. Advanced package managers such as Conda or Guix allow the user to create separate software environments where different application versions can be deployed without collisions while ensuring control and traceability over changes and dependencies. All these solutions represent a different approach to address the reproducibility challenge while also offering a different user experience and requiring different setups to work properly. For instance, container-based systems such as Docker and Singularity are not always a viable option in HPC environments since they may require updates to the existing computing infrastructure. Also, HPC operating system installations may include kernel versions that do not allow for the so-called user namespaces, a fundamental component among the many kernel features that together allow an application to run in an isolated container environment. Another downside of containers is that it is hard to assess what is in them—they act like a black box. When creating containers with above Docker scripts, it depends on the time they are assembled what goes in. A Debian or Conda update between creating containers, for example, may include a different software version therefore a different dependency graph. Only GNU Guix containers provide a clear view on what is contained.
Containers provide isolation from the underlying operating system. On HPC environments it may be required to run software outside a container. While applications built with Guix or Conda can be run in isolation when container support is available, they do not require these features at runtime. As a package manager Conda, neither depends on container features nor on root privileges, but it pays for this convenience with a lack of both process isolation and bit-reproducibility [47]. GNU Guix, meanwhile, provides the most rigorous path to reproducible software deployment. In order to guarantee that packages are built in a bit-reproducible fashion and share binary packages, Guix requires to store packages in the directory /gnu/store. There are several work-arounds for this; one of them is by using containers, and another is by mounting /gnu/ store from a host that has built privileges for that directory. A third option is to build packages targeted at a different directory, but this loses the bit-reproducibility and the convenience of binary installs. A fourth option is to provide relocatable binary installation packages that can be installed in a user available directory, similar to what Bioconda does. Such packages exist for sambamba, gemma, and the D-compiler.
Finally, each combination of these packaging and workflow solutions occupies a slightly different region in the solution space for the scalability and reproducibility challenge. Fortunately, the packaging tools can be used next to each other without interference, thereby providing a wealth of software packages for bioinformatics. Today, there is hardly ever a good reason to build software from source.
#### 5 Questions
#### References
performance of genomic pipelines. PeerJ 3: e1273
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
## Sharing Programming Resources Between Bio\* Projects
#### Raoul J. P. Bonnal, Andrew Yates, Naohisa Goto, Laurent Gautier, Scooter Willis, Christopher Fields, Toshiaki Katayama, and Pjotr Prins
#### Abstract
Open-source software encourages computer programmers to reuse software components written by others. In evolutionary bioinformatics, open-source software comes in a broad range of programming languages, including C/C++, Perl, Python, Ruby, Java, and R. To avoid writing the same functionality multiple times for different languages, it is possible to share components by bridging computer languages and Bio\* projects, such as BioPerl, Biopython, BioRuby, BioJava, and R/Bioconductor.
In this chapter, we compare the three principal approaches for sharing software between different programming languages: by remote procedure call (RPC), by sharing a local "call stack," and by calling program to programs. RPC provides a language-independent protocol over a network interface; examples are SOAP and Rserve. The local call stack provides a between-language mapping, not over the network interface but directly in computer memory; examples are R bindings, RPy, and languages sharing the Java virtual machine stack. This functionality provides strategies for sharing of software between Bio\* projects, which can be exploited more often.
Here, we present cross-language examples for sequence translation and measure throughput of the different options. We compare calling into R through native R, RSOAP, Rserve, and RPy interfaces, with the performance of native BioPerl, Biopython, BioJava, and BioRuby implementations and with call stack bindings to BioJava and the European Molecular Biology Open Software Suite (EMBOSS).
In general, call stack approaches outperform native Bio\* implementations, and these, in turn, outperform "RPC"-based approaches. To test and compare strategies, we provide a downloadable Docker container with all examples, tools, and libraries included.
Key words Bioinformatics, R, Python, Ruby, Perl, Java, Web services, RPC, EMBOSS, PAML
#### 1 Introduction
Bioinformatics has created its tower of Babel. The full set of functionality for bioinformatics, including statistical and computational methods for evolutionary biology, is implemented in a wide range of computer languages, e.g., Java, C/C++, Perl, Python, Ruby, and R. This comes as no surprise, as computer language design is the
Download: https://github.com/EvolutionaryGenomics/Cross-language-interfacing
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0\_25, © The Author(s) 2019
result of multiple trade-offs, for example, in strictness, convenience, and performance. In this chapter we discuss strategies for combining solutions from different languages and look at performance implications of combining cross-language functionality. In the process we also highlight implications of such strategic choices.
Computer languages used in bioinformatics today typically fall into two groups: those compiled and those interpreted. Java, C++, and D, for example, are statically typed compiled languages, while R, Perl, Ruby, and Python are dynamically typed interpreted languages. In principle, a compiled language is converted into machine code once by a language compiler, and an interpreted language is compiled every time at runtime, the moment it is run by an interpreter. Static typing allows a compiler to optimize machine code for speed. Dynamic typing requires an interpreter and resolves variable and function types at runtime. Such design decisions cause Java, C++, and D to have stronger compile-time type checking and faster execution speed than R, Perl, Ruby, and Python. When comparing runtime performance of these languages, compiled statically typed languages, such as C++, D, and Java, generally outperform interpreted dynamically typed languages, such as Python, Perl, and R. For speed comparison between languages, see, for example, the benchmarks game.
Statically typed compiled languages tend to produce faster code at runtime
Runtime performance, however, is not the only criterion for selecting a computer language. R, Perl, Ruby, and Python offer sophisticated interactive analysis of data in an interpreted shell which is not directly possible with C++, D, or Java. Another important criterium may be conciseness. Interpreted languages generally allow functionality to be written in less lines of code. The number of lines matter, as it is often easier to grasp something expressed in a short and concise fashion, if done competently, leading to easier coding and maintenance of software and resulting in increased programmer productivity. In general, with R, Perl, Ruby, and Python, it takes less lines of code to write software than with C+ +, D, or Java; this is also visible from the examples in the benchmarks game.
Interpreted languages allow for concise code that is easier to read and results in increased programmer productivity
Based on the conciseness criterium, computer languages fall into these two groups. This suggests a trade-off between execution speed and conciseness/programmer productivity. Even so, strong typing may help later when refactoring code, perhaps regaining some of that lost productivity. The authors also note that in their experience, the more programming languages one masters, the easier it becomes mastering new languages (with the exception, perhaps, of Haskell). Learning new programming languages is important when writing software.
Logically, to fully utilize the potential of existing and future bioinformatics functionality, it is necessary to bridge between computer languages. Bioinformaticians cannot be expected to master every language, and it is inefficient to write the same functionality for every language. For example, R/Bioconductor contains unique and exhaustive functionalities for statistical methods, such as for gene expression analysis [1]. The singular implementation of this functionality in R has caused researchers to invest in learning the R language. Others, meanwhile, have worked on building bridges between languages. For example, RPy and Rserve allow accessing R functionality from Python [2], and JRI and Rserve allow accessing R functionality from Java [3, 4]. Other languages have similar bindings, such as RSRuby that allows accessing R from Ruby.
Discussing other important criteria for selecting a programming language, such as ease of understanding, productivity, portability, and the size and dynamics of the supporting Bio\* project developer communities, is beyond the scope of this chapter. The authors, who have different individual preferences, wish to emphasize that every language has characteristics driven by language design and there is no single perfect all-purpose computer language. In practice, the choice of a computer language depends mainly on the individuals involved in a project, partly due to the investment it takes to master a language. Researchers and programmers have prior investments and personal preferences, which have resulted in a wide range of computer languages used in the bioinformatics community.
Contrasting with singular implementations, every mainstream Bio\* project, such as BioPerl [5], Biopython [6], BioRuby [7], R/Bioconductor [1], BioJava [8], the European Molecular Biology Open Software Suite (EMBOSS) [9], and Bio++ [10], contains duplication of functionality. Every Bio\* project consists of a group of volunteers collaborating at providing functionality for bioinformatics, genomics, and life science research under an open-source software (OSS) license. The BioPerl project does that for Perl, BioJava for Java, etc. Next to the language used, the total coverage of functionality, and perhaps quality of implementation, differs between projects. Not only is there duplication of effort, both in writing and testing code, but also there are differences in implementation, completeness, correctness, and performance. For example, implementations between projects differ even for something as straightforward as codon translation, e.g., in number of types of encoding and support for the translating of ambiguous nucleotides. EMBOSS, uniquely, attempts to predict the final amino acid in a sequence, even when there are only two nucleotides available for the last codon.
Whereas Chapter 25 discusses Internet data resources and how to share them, in this chapter, we discuss how to share functional resources by interfacing and bridging functionality between different computer languages. This is highly relevant to evolutionary biology as most classic phylogenetic resources were written in C, while nowadays phylogenetic routines are written in Java, Perl, Python, Ruby, and R. Especially for communities with relatively few software developers, we argue here that it is important to bridge these functional resources from multiple languages. For bridging, strategies are here discussed to invoke one program from another, use some form of remote procedure calls (RPC), or use a local call stack.
The most simple way of interfacing software is by invoking one program from another. This strategy is often used in Bio\* projects, for example, for invoking external programs. A regular subset would be PAML [11], HMMER [12], ClustalW [13], MAFFT [14], Muscle [15], BLAST [16], and MrBayes [17]. The Bio\* projects typically contain modules which invoke the external program and parse the results. The advantage of this approach is that it mimics running a program on the command line, so invocation is straightforward. Another advantage, in a web service context, is that if the called program crashes, it does not have to take the whole service down. There are also some downsides, however. Loading a new instance of a program every time incurs extra overhead. More importantly, nonstandard input and output makes the interface fragile, i.e., what happens when input or output differs between two versions of a program? A further downside is that external programs do not have fine-grained function access and have no support for advanced error handling and exceptions. What happens, for example, when the invoked program runs out of process memory? How to handle that gracefully? A final complication is that such a program is an external software deployment dependency, which may be hard to resolve for an end user.
Procedure Call In contrast to calling one program from another, true crosslanguage interfacing allows one language to access functions and/or objects in another language, as if they are native function calls. To achieve transparent function calls between different computer languages, there are two principal approaches. The first approach is for one language to call directly into another language's function or method over a network interface, the so-called remote procedure call (RPC). The second approach is to call into another language over a local "call stack."
> In bioinformatics, cross-language RPC comes in the form of web services and binary network protocols. A web service application programming interface (API) is exposed, and a function call gets translated with its parameters into a language-independent
#### 1.1 Bridging Functional Resources Calling from Program to Program
1.2 Remote
format, a procedure called "marshalling." After calling the function on a server, the result is returned in, for example, XML and translated back through "unmarshalling." Examples of cross-language XML protocols are SOAP [18] and XML/RPC [19].
More techniques exist for web service-type cross-language RPC. For example, representational state transfer (REST), or ReSTful [20], is a straightforward HTTP protocol, often preferred over SOAP because of its simplicity. Another XML-based protocol is Resource Description Framework (RDF), as part of the semantic web specification. Both REST and RDF can be used for RPC solutions.
In addition, binary alternatives exist because XML-based protocols are not very efficient. XML is verbose, increasing the data load, and requires parsing at both marshalling and unmarshalling steps. In contrast, binary protocols are designed to reduce the data transfer load and increase speed. Examples of binary protocols are Rserve [3], which is specifically designed for R, and Google protocol buffers [21]. Another software framework based on a binary protocol is Thrift, by the Apache software foundation, designed for scalable cross-language service development [22]. Finally, also worth considering are very fast interoperable messaging-based paradigms, such as ZeroMQ [23], and high-level message-level optimizers, such as GraphQL.
1.3 Local Call Stack The alternative to RPC is to create native local bindings from one language to another using a shared native call stack, essentially linking into code of a different computer language. With the call stack, function calls do not run over the network but over a stack implementation in shared computer memory. In a single virtual machine, such as the JVM and Erlang Beam, compiled code can share the same call stack, which can make cross-language calling efficient. For example, the languages Java, Jython, JRuby, Clojure, Groovy, and Scala can transparently call into each other when running on the same virtual machine using native speeds.
Native call stack sharing is also supported at the lowest level by the computer operating system through compiled shared libraries. These shared libraries have an extension .so on Linux, .dylib on OSX, and .dll on Windows. The shared libraries are designed so that they contain code and data that provide services to independent programs, which allows the sharing and changing of code and data in a modular fashion. Shared library interfaces are well defined at the operating system level, and languages have a way of binding them. Specialized interface bindings to shared libraries exist for every language, for example, R's C modules, the Java Native Interface (JNI) for the JVM, Foreign Function Interfaces (FFI) for Python and Ruby, the Parrot native compiler interface PerlXS for Perl.
With (dynamic) shared libraries, certain algorithms can be written in a low-level, high-performance compiled computer language, such as C/C++, D, or FORTRAN. And high-level languages, such as Perl, Python, Ruby, R, and even Java, can access these algorithms. This way, languages can be mixed to optimize solutions. Creating these shared library interfaces, however, can be a tedious exercise, which often calls for code generators. One such generator is the Simplified Wrapper and Interface Generator (SWIG) [24], which consists of a macro-language, a C header file parser, and the tools to bind low-level shared libraries to a wide range of languages. For C/C++, SWIG can parse the header files and generate the bindings for other languages, which, in turn, call into these shared libraries. The Boost project has similar facilities for mapping calls to SWIG. C FFI's that come with programming languages, such as Python's CFFI and Ruby's FFI, tend to be the easiest to work with.
Even though this extensive functionality for interfacing is available, the full potential of creating cross-language adapters is not fully exploited in bioinformatics. Rather than bridge two languages, researchers often opt to duplicate functionality. This is possibly due to a lack of information on the effort involved and the added complexity of creating a language bridge. Also, the impact on performance may be an unknown quantity. A further complication is the need to understand, to some degree, both sides of the equation, i.e., to provide an R function to Python requires some understanding of both R and Python, at least to the level of reading the documentation of the shared module and creating a working binding. Likewise, binding Python to C using a call stack approach requires some understanding of both Python and C. Sometimes, binding of complex functions can be daunting, and deployment may be a concern, e.g., when creating shared library bindings on Linux, they may not easily work on Windows or macOS.
1.4 Comparing Approaches Here, we compare bridging code from one language to another using the RPC approach and the call stack approach. As a comparison we also provide a program-to-program approach and show how dependencies can be fixated. The comparison is done in the form of short experiments (scripts) which can be executed by the reader. To measure performance between different approaches, we use codon translation as an example of shared functionality between Bio\* projects. Codon translation is a straightforward algorithm with table lookups. Such sequence translation is representative of many bioinformatics tasks that deal with genome-sized data and require many function calls with small-sized parameters.
In this chapter we first focus on comparing R and Python bindings. We include native Bio\* implementations, i.e., Biopython, BioRuby, BioPerl, BioJava, and EMBOSS (C) for an absolute speed comparison. Next we try bindings on the JVM.
Examples and tests can in principle be experimented with a computer running Linux, macOS, or Windows. To ease trials, we have defined GNU Guix packages that contain the tools and their dependencies. From this we have created a downloadable Docker image that supports all interfaces and performance examples (GNU Guix and Docker are discussed in Chapter 25).
#### 2 Results
2.1 Calling into R R is a free and open-source environment for statistical computing and graphics [25]. R comes with a wide range of functionality, including modules for bioinformatics, such as bundled in R/Bioconductor [1]. R is treated as a special citizen in this chapter because the language is widely used and comes with statistical algorithms for evolutionary biology, such as Ape [26] and SeqinR [27], both available through the comprehensive R archive network (CRAN).
> R defines a clear interface between the high-level language R and low-level highly optimized C and FORTRAN libraries, some of which have been around for a long time, such as the libraries for linear regression and linear algebra. In addition, the R environment successfully handles cross-platform packaging of C, C++, FOR-TRAN, and R code. The combination of features has resulted in R becoming the open-source language of choice in a number of communities, including statistics and some disciplines in biology. R/Bioconductor has gene expression analysis [1] and R/qtl [28] and R/qtlbim [29] for QTL mapping (see also QTL mapping in Chapter 21). Not all is lost, however, for those not comfortable with the R language itself. R can act as an intermediate between functionality and high-level languages. A number of libraries have been created that interface to R from other languages, either providing a form of RPC, through RSOAP or Rserve, or a call stack interface calling into the R-shared library and executing R commands, for example, RPy for Python, RSPerl for Perl, RSRuby for Ruby, and JRI for Java. Of the last call stack approaches, RPy currently has the most complete implementation; see also [2].
> In this chapter, we compare different approaches for invoking full R functionality from another language. To test cross-language calling, we elected to demonstrate codon translation. Codon-toprotein amino acid translation is representative for a relatively simple computation that potentially happens thousands of times with genome-sized data. Every Bio\* project includes such a translation function, so it is a fair way to test for language interoperability and performance. For data, we use a WormBase [30] C. elegans cDNA FASTA file (33 Mb), containing 24,652 nucleotide sequences, predicted to translate to protein (Fig. 1).
Fig. 1 Throughput of mRNA to protein translation using combinations of cross-language calling with a range of programming resources. WormBase C. elegans predicted protein coding DNA that was parsed in FASTA format and translated into amino acids. Tests were executed inside a container. Different file sizes were used containing 500, 1000, 5000, 15,000, and 25,000 sequences (X-axis) and the number of sequences processed per seconds (Y-axis log10 scale). Measurements were taken on an AMD Opteron(TM) 6128 8 cores at 2.0 GHz, 4 sockets 8 cores, with 512 GB RAM DDR3 ECC, and an HDD SATA of 2 TB. Broadly the figure shows that sustained throughput is reached quickly and flattens out. R-Biostrings performs poorly at 285 Seq/s, while R-GeneR and Rserve (Python+Rserve+GeneR) perform at the level of native Bio\* libraries, respectively, 658 Seq/s and 660 Seq/s. The cross-language Ruby-FFI at 6256 Seq/s calls EMBOSS C translation and outperforms all others
2.1.1 Using GeneR with Plain R "The R/Bioconductor GeneR package [31] supports fast codon translation with the strTranslate function implemented in C." GeneR supports the eukaryotic code and other major encoding standards. R usage is:
```
library(GeneR)
strTranslate("atgtcaatggtaagaaatgtatcaaatcagagcgaaaaattg-
gaaattttgt")
[1] "MSMVRNVSNQSEKLEIL"
```
The \name{R+GeneR} script (also available here) reads:
```
fasta = 'dna.fa'
library(GeneR)
idx = indexFasta(fasta)
```
```
lines <-readLines( paste(fasta,'.ix',sep='') )
index <-read.table(paste(fasta,'.ix',sep='') )[,1]
n=0
for (i in 1:times) {
for (name in index) {
readFasta (file=fasta, name = name)
ntseq = getSeq(0)
aaseq = strTranslate(ntseq)
cat(">",name," (",n,")\n",aaseq,"\n",sep="")
n = n+1
}
}
```
and parses the nucleotide FASTA input and outputs amino acid FASTA. Run the script:
```
docker run --rm -v 'pwd'/tmp:/tmp -v 'pwd'/scripts:/scripts -e \
BATCH_VARS=/tmp/test-dna-${i}.fa -t bionode bash -c "source
/etc/profile
cd /book-evolutionary-genomics
./scripts/create_test_files.rb
R -q --no-save --no-restore --no-readline --slave < src/R/
DNAtranslate_GeneR.R" > /dev/null
```
Used directly from R, the throughput of the GeneR module is about 658 sequences per second (Seq/s) on the test system, an AMD Opteron(TM) 6128 CPU at 2.00 GHz (see also Fig. 1). When checking the implementation by reading the source code, in the first edition, we found that the GeneR FASTA parser was a huge bottleneck. The FASTA parser implementation created an index on disk and reloaded the full index file from disk for each individual sequence, thereby incurring a large overhead for every single sequence.
To see if we could improve throughput, we replaced the slow FASTA parser with \name{R+Biostrings} which reads FASTA once into RAM using the R/Bioconductor BioStrings module and still uses GeneR to translate. At the time, this implementation was 1.6 times faster than GeneR. At this time GeneR is 3.2 on average faster than reading with Biostrings which had a throughput of 284.83 Seqs/s proving some work was done by the authors to improve GeneR. The second script can be found here.
2.1.2 Calling into R from Other Languages with RPC One strategy for bridging between languages is to use R as a network server and invoke remote procedure calls (RPC) over the network.
1. SOAP
SOAP allows processes to communicate using XML over HTTP in a client/server setup. SOAP is an operating system
2.1.3 Calling into R from Other Languages with the Call Stack Approach
and computer language "agnostic," so it can be used to bridge between languages. In the previous edition of this chapter {Ref to Previous Edition, same chapter}, we wrote a R/SOAP [32] adapter for codon translation and invoked it from Python (a Python to R bridge). That client script can be found here. The SOAP bridge was dropped from this chapter because the SOAP packages are not maintained and it was by far the slowest method of cross-language interfacing we tried! The marshalling and unmarshalling of simple string objects using XML over a local network interface takes a lot of computational resources. We do not recommend using SOAP.
2. Rserve
Rserve [3] is a custom binary network protocol, more efficient than XML-based protocols [3]. R data types are converted into Rserve binary data types. Rserve was originally written for Java, but nowadays connectors exist for other languages. With Rserve, Python and R do not have to run on the same server. Furthermore, all data structures will automatically be converted from native R to native Python and numpy types and back.
With RServe fired up a Python example is:
```
import pyRserve
conn = pyRserve.connect()
conn.eval('library(GeneR)')
conn.eval('strTranslate("atgtcaatggtaagaaatgtatcaaatcagagc-
gaaaaattggaaattttgt")')
'MSMVRNVSNQSEKLEIL'
```
where Rserve+GeneR uses the GeneR translate function. In our test Biopython [6] is used for parsing FASTA, and at 797 Seq/s, even with this network bridge, Python+Rserve's speed is on par with that of R. The script can be found here.
Another strategy for bridging language is to use a native call stack, i.e., data does not get transferred over the network. RPy2 executes R code from within Python over a local call stack [2]. Invoking the same GeneR functions from Python:
```
import rpy2.robjects as robjects
from rpy2.robjects.packages import importr
importr('GeneR')
strTranslate=robjects.r['strTranslate']
strTranslate("atgtcaatggtaagaaatgtatcaaatcagagcgaaaaattg-
gaaattttgt")[0]
'MSMVRNVSNQSEKLEIL'
```
This example uses Biopython for parsing FASTA and invokes GeneR translation over a call stack handled by RPy2. At 2049 Seq/ s, throughput is the highest of our calling into R examples. The Python implementation outperforms the other FASTA parsers, and GeneR is fast too when only the translation function is called (GeneR's strTranslate is actually written in C, not in R). Still, there are some overheads for bridging and transforming string objects from Python into R and back. The RPy2 call stack approach is efficient for passing data back and forth. The script can be found here.
2.2 Native Bio\* Implementations When dealing with cross-language transport comparisons, it is interesting to compare results with native language implementations. For example, Biopython [6] would be:
```
from Bio.Seq import Seq
from Bio.Alphabet import generic_dna
coding_dna = Seq("atgtcaatggtaagaaatgtatcaaatcagagcgaaaaattg-
gaaattttgt", generic_dna)
coding_dna.translate()
Seq('MSMVRNVSNQSEKLEIL', ExtendedIUPACProtein())
```
which runs at 797 Seq/s which is slower than the Python3+RPy2 +GeneR version. This is because the translate function is written in Python and not in C. It is, however, still faster than R+GeneR. Ruby+BioRuby runs faster at 1481 Seq/s. Perl+BioPerl is in the middle with 1165 Seq/s. We can assume the Biopython, BioPerl, and BioRuby implementations are reasonably optimized for performance. Therefore, throughput reflects the performance of these interpreted languages (see Fig. 1).
Java is a statically typed compiled language. Java+BioJava [8] outperforms the interpreters and runs at 2266 Seq/s.
The source code for all examples can be found here in the {Biopython}, {BioRuby}, {BioPerl}, and {BioJava} subdirectories.
2.3 Using the JVM for Cross-Language Support The Java virtual machine (JVM) is a "bytecode" standard that represents a form of computer intermediate language. This language conceptually represents the instruction set of a stackoriented capability architecture. This intermediate language, or "bytecode," is not tied to Java specifically, and in the last 10 years, a number of languages have appeared which target the JVM, including JRuby (Ruby on the JVM), Jython (Python on the JVM), Groovy [33], Clojure [34], and Scala [35]. These languages also compile into bytecode and share the same JVM stack. The shared JVM stack allows transparent function calling between different languages.
An example of calling BioJava translation from a Scala program:
import org.biojava.nbio.core.sequence.transcription.TranscriptionEngine import org.biojava.nbio.core.sequence.\_ val transcriber = TranscriptionEngine.getDefault() val dna = new DNASequence("atgtcaatggtaagaaatgtatcaaatcagagcgaaaaattggaaattttgt") val rna = dna.getRNASequence(transcriber) rna.getProteinSequence(transcriber) 'MSMVRNVSNQSEKLEIL' which uses the BioJava libraries. A native Java function, such as getProteinSequence, is directly invoked from the other language without overheads (the passed-in transcriber object is passed by reference, just like in Java). In fact, Scala compiles to bytecode, which maps one to one to Java, including the class definitions. The produced bytecode is a native Java bytecode; therefore, the performance of calling BioJava from Scala or Java is exactly the same. This also holds for other languages on the JVM, such as Clojure and Groovy. We have also included a JRuby example that calls into BioJava4 on the JVM and runs at 1413 Seq/s. JRuby is an interpreter on the JVM that still needs some translation calling into JVM functions. It is therefore slower than native calls. 2.4 Shared C Library Cross-Calling Using EMBOSS Codon Translation EMBOSS is a free and OSS analysis package specially developed for the needs of the molecular biology user community, mostly written in C [9]. 2.4.1 FFI Using Foreign Function Interface (FFI), it is possible to load dynamic libraries at runtime, define classes to map composite data types, and bind functions for a later use inside your host programming language. We used FFI to bind the EMBOSS translation function to Python and Ruby. The Python example: from ctypes import \* import os emboss = cdll.LoadLibrary(os.path.join(os.path.dirname(os.
```
path.abspath(__file__)),"emboss.so"))
```
```
trnTable = emboss.ajTrnNewI(1)
```
```
ajpseq = emboss.ajSeqNewNameC(b"atgtcaatggtaagaaatgtatcaaat-
```
cagagcgaaaaattggaaattttgt", b"Test sequence")
```
ajpseqt = emboss.ajTrnSeqOrig(trnTable,ajpseq,1)
```
```
seq = emboss.ajSeqGetSeqCopyC(ajpseqt)
seq = str(c_char_p(seq).value,'utf-8')
```
```
print(seq)
MSMVRNVSNQSEKLEILX
```
The Ruby example:
```
require 'ffi'
module Emboss
extend FFI::Library
ffi_lib "./emboss.so"
attach_function :ajTrnNewI, [:int], :pointer
attach_function :ajSeqNewNameC, [:pointer, :pointer], :
pointer
attach_function :ajTrnSeqOrig, [:pointer, :pointer, :int], :
pointer
attach_function :ajSeqGetSeqCopyC, [:pointer], :string
end
trnTable = Emboss.ajTrnNewI(1)
ajpseq = Emboss.ajSeqNewNameC("atgtcaatggtaagaaatgtatcaaatca-
gagcgaaaaattggaaattttgt", "Test sequence")
ajpseqt = Emboss.ajTrnSeqOrig(trnTable,ajpseq,1)
aa = Emboss.ajSeqGetSeqCopyC(ajpseqt)
print aa,"\n"
MSMVRNVSNQSEKLEILX
```
In both cases the advantage of FFI is that it does not require to compile any source code, just loading the shared library and binding what is needed. Python has a native library called ctypes, and more sophisticated libraries are available to help the programmer bind complex data structures and functions. Ruby has a dedicated gem called [ruby-ffi].
The Ruby and Python FFI outperforms all above methods at 6257 Seq/s and 4787 Seq/s, respectively (see Fig. 1). Plotting the time in seconds spent to translate the sequences, Ruby and Python FFI are the lowest (quickest) in the whole comparison (see Fig. 2). The high speed points out that (1) the invoked Biopython and BioRuby functions are reasonably efficient at parsing FASTA, (2) the FFI-generated call stack is efficient for moving data over the local call stack, and (3) the EMBOSS transeq DNA to protein translation is optimal C code.
```
2.5 Calling Program
to Program
Calling program to program is far more common than you may
think because even when you run a program in a shell, such as Bash,
you are calling program to program. You can invoke EMBOSS
from the command line:
```
Fig. 2 Number of seconds needed for processing mRNA to protein translation using cross-language calling with a range of programming resources. See Fig. <sup>1</sup> for the setup. The figure shows that for all the implementations, the time increases linearly with the number of sequences in input. R-Biostrings performs poorly with an upstart of 6.50 s and the highest slope. The cross-language Ruby-FFI, Python FFI, and Python-EMBOSS with an upstart slightly higher than Java have a very minimal slope; Ruby-FFI has a nearly constant time
transeq test-dna.fa test.pep
transeq is written in C and runs at a very fast 23,478 Seq/s. Invoking above EMBOSS' transeq in Python looks like this:
```
os.system("transeq "+fn+" out.pep")
for seq_record in SeqIO.parse("out.pep", "fasta"):
print(">",seq_record.id)
seq = str(seq_record.seq)
print(seq)
```
and this combination runs at 4768 Seq/s. That is close to Python FFI and a third of the speed of transeq on its own because of Python parsing the output. Every parsing step has a cost attached.
2.6 Web Services A discussion on bridging languages would not be complete if we did not include web services, particularly using REST API's. Service like TogoWS and EBI web services which include EMBOSS transeq (SOAP) offer functionality over http(s) and can be used from any programming language. Here a Ruby example of using TogoWS:
```
## Invoke irb by loading BioRuby
% irb -r bio
## Create a TogoWS object
>> togows = Bio::TogoWS::REST.new
=> #<Bio::TogoWS::REST:0x007f840faab9d8 @pathbase="/",
@http=#<Net::HTTP togows.dbcls.jp:80 open=false>,
@header={"User-Agent"=>"BioRuby/1.5.1"}, @debug=false>
## Search for UniProt entries by keywords
>> togows.search('uniprot', 'lung cancer')
=> "KKLC1_MACFA\nKKLC1_HUMAN\nDLEC1_HUMAN\n .....
## Retrieve one UniProt entry (or multiple entries if you like)
>> entry = togows.entry('uniprot', 'KKLC1_MACFA')
## See the entry content
>> puts entry
ID KKLC1_MACFA Reviewed; 114 AA.
AC Q4R717;
:
## Convert the retrieved UniProt entry into FASTA format
>> puts togows.convert(entry, 'uniprot', 'fasta')
>KKLC1_MACFA RecName: Full=Kita-kyushu lung cancer antigen
1 homolog;
MNVYLLLASGILCALMTVFWKYRRFQRNTGEMSSNSTALALVRPSSTGLINSNTDNNLSV
YDLSRDILNNFPHSIAMQKRILVNLTTVENKLVELEHILVSKGFRSASAHRKST
```
Web services can harness a lot of power because they use large databases and access up-to-date information. As an example, let's generate RDF from above entry:
```
## Retrieve PubMed entry and convert it into RDF/Turtle
(or JSON or XML if you like)
>> puts togows.entry('pubmed', '16381885', 'ttl')
@prefix dc: <http://purl.org/dc/elements/1.1/> .
@prefix dcterms: <http://purl.org/dc/terms/> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix prism: <http://prismstandard.org/namespaces/2.0/ba-
sic/> .
@prefix medline: <http://purl.jp/bio/10/pubmed/> .
```
```
<http://rdf.ncbi.nlm.nih.gov/pubmed/16381885> medline:pmid
"16381885" ;
```
```
rdfs:label "pmid:16381885" ;
dc:identifier "16381885" ;
medline:own "NLM" ;
```
Unfortunately, data centric web services can be slow, i.e., sending and retrieving data over the internet incurs large latency and throughput penalties. Sometimes they use powerful back ends, and it is possible to submit large batch jobs which compete with locally installed solutions. Examples are the BLAST service [16] and GeneNetwork [36].
#### 3 Discussion
The half-life of bioinformatics software is 2 years—Pjotr Prins
In this chapter we show that there are many ways of bridging between computer languages. Cross-language interfacing is a topic of importance to evolutionary genomics (and beyond) because computational biologists need to provide tools that are capable of complex analysis and cope with the amount of biological data generated by the latest technologies. Cross-language interfacing allows sharing of code. This means computer software can be written in the computer language of choice for a particular purpose. Flexibility in choice of computer programming language allows optimizing of computational resources and, perhaps even more important, software developer resources, in bioinformatics.
When some functionality is needed that exists in a different computer language than the one used for a project, a developer has the following options: either rewrite the code in the preferred language, essentially a duplication of effort, or bridge from one language to the other. For bridging, there are essentially two technical methods that allow full programmatic access to functionality: through RPC or a local call stack. A third option may be available when functionality can be reached through the command line, as shown above with transeq.
RPC function invocation, over a network interface, has the advantage of being language agnostic and even machine independent. A function can run on a different machine or even over the Internet, which is the basis of web services and may be attractive even for running services locally. RPC XML-based technologies, however, are slow because of expensive parsing and high data load. Our metrics suggest that it may be worth experimenting with binary protocols, such as Rserve and Apache Thrift.
When performance is critical, e.g., when much data needs to be processed, or functions are invoked millions of times, a native call stack approach may be preferred over RPC. Metrics suggest that the EMBOSS C implementation performs well and that binding to the native C libraries with FFI is efficient (see Fig. 2). Alternatively, it is possible to use R as an intermediate to C libraries. Interestingly, calling R libraries, many of which are written in C, may give higher performance than calling into native Bio\* implementations. For example, Python+RPy2+GeneR is faster that Biopython pure Python implementation of sequence translation, and it is also faster than R calling into GeneR directly—confirming a common complaint that R can be slow.
Even though RPC may perform less well than local stack-based approaches, RPC has some real advantages. For example, if you have a choice of calling a local BLAST library or call into a remote and ready NCBI RPC interface, the latter lacks the deployment complexity. Also the public resource may be more up to date than a copied server running locally. This holds for many curated services that involve large databases, such as PDB [37], Pfam [38], KEGG [39], and UniProt [40]. Chapter 25 gives a deeper treatment of these Internet resources.
From the examples given in this chapter, it may be clear that actual invocation of functions through the different technologies is similar, i.e., all listed Python scripts look similar, provided the underlying dependencies on tools and libraries have been resolved. The main difference between implementations is with deployment of software, rather than invocation of functionality. The JVM approach is of interest, because it makes bridging between supported languages transparent and deployment straightforward. Not only can languages be mixed, but also the advanced Java tool chain is available, including debuggers, profilers, load distributors, and build tools. Other shared virtual machines, such as .NET and Parrot, potentially offer similar advantages but are less used in bioinformatics.
In the first edition, we wrote that when striving for reliable and correct software solutions, the alternative strategy of calling computer programs as external units via the command line should be discouraged: not only is it less efficient that a program gets started every time a function gets called, but also a potential deployment nightmare is introduced. What happens when the program is not installed, or the interface changed between versions, or when there is some other error? With the full programmatic interfaces, discussed in this chapter, incompatibilities between functions get caught much earlier. In this edition of the chapter, we add that efficiency considerations still hold, and error handling can be problematic. When it comes to deployment, however, there now exist solutions that fixate versions of software and give control of the dependency graph, i.e., a tool like transeq can be coupled with its exact version against your software. To ascertain coupling: first there are containers, such as offered by Docker, that allow for bundling software binaries. Second, some recent software distributions allow for formal deployment solutions with reproducible dependency graphs. If you want to know more, check the GNU Guix and NixOS projects. It is possible to combine these deployment technologies. In fact, with this chapter, we provide tools and scripts defined as GNU Guix packages and hosted in a Docker container. These solutions are discussed in Chapter 25.
Choosing a computer language should not be based on runtime performance considerations alone. The maturity of the language and accompanying libraries, tools, and documentation should count heavily, as well as the activity of the community involved. The time saved by using a known language versus learning a new language should be factored in. The main point we are trying to make here is that it is possible to mix languages using different interfacing strategies. This allows leveraging existing functionality, as written by others, using a language of choice. Depending on one's needs, it is advisable to test possible alternatives for performance, as the different tests show that performance varies.
Whichever language and bridging technology is preferred, we think it important to test the performance of different ways of interfacing languages, as there is (1) a need for combining languages in bioinformatics and (2) it is not always clear what impact a choice of cross-language interface may have on performance. By testing different bridging technologies and functional implementations, the best solution should emerge for a specific scenario.
So far, we have focused on the performance of cross-language calling. In Chapter 25, scalability of computation is discussed by programming for multiple processors and machines.
#### 4 Questions
#### Acknowledgments
We thank all open-source software developers for creating such great tools and libraries for the scientific community.
#### References
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
### INDEX
#### A
#### B
Maria Anisimova (ed.), Evolutionary Genomics: Statistical and Computational Methods, Methods in Molecular Biology, vol. 1910, https://doi.org/10.1007/978-1-4939-9074-0, © The Author(s) 2019
#### C
#### 284, 296, 299
#### D
#### E
#### F
#### G
#### H
#### I
#### J
#### K
#### L
#### M
#### N
#### O
#### P
Population (cont.) genetics........................................................ 53, 71, 95, 97, 101–104, 223, 225, 331, 333–337, 343, 354, 356, 360, 392, 415, 416, 522, 534, 535, 555–586, 637 genomics.................................................356, 555–586 simulator.................................................................. 139 size ........................................................ 104, 213, 214, 222, 224, 331, 333, 335–337, 351, 352, 392, 416, 512, 520, 522, 536, 537, 555, 556, 560, 561, 563–565, 568, 580–582, 584, 585, 645, 648, 710 Positive selection, see Selection, positive Posterior decoding .....................................................59, 60, 580 probability ................................................... 41, 45, 46, 89, 90, 409, 410, 430, 450, 456, 705 probability distribution.................................... 91, 715 Power law distribution...............474, 476, 477, 483, 485 Preferential attachment.......................474, 483, 493, 500 Primary chromosome...................................................... 13 Primates ...................................................... 143, 182, 185, 186, 193, 194, 311, 312, 321, 328, 341, 374, 382, 392, 428, 433, 435, 557, 584, 585 Primer ....................................................71, 606, 611, 663 Principal components analysis (PCA)...................542, 621–622 coordinates analysis (PCoA).........601, 602, 621–622 Prior conjugate .............................................................46, 47 distribution .................................................. 89, 90, 96, 100, 101, 327, 410, 455, 583 Profile HMM.............................................................60, 61 Progressive alignment...................................................135 Prokaryotes........................................ 4, 9, 10, 12, 13, 15, 18–20, 22, 170, 221, 242, 243, 258–260, 264–266, 272, 280, 281, 285, 293, 297, 478, 484, 488, 518 Prokaryotic cell...............................................5, 10, 14, 15 Promoter .........................................................17, 18, 179, 185, 186, 507, 508 Protein architecture..................................................... 469–500 combination .................................................. 471, 472, 478, 479, 481, 483–484, 486, 489, 493 complex.................................................. 475, 477, 494 databases ADDA....................................................... 471, 497 CATH .............................470, 471, 478, 481, 497 Conserved Domain Database........................... 486 Gene3D family .................................470, 478, 497 InterPro ............................................470, 478, 497 Pfam......................................................... 470, 476, 479–481, 483–485, 487, 488, 494, 496–498
#### Q
#### R
#### S
#### T
#### U
URL, see Uniform resource locator UTR, see Untranslated regions
#### V
#### W
#### X
#### Y
```
Yeast ............................................343, 347, 354, 490, 639
```
#### Z
| doab | 2025-04-07T04:13:03.724911 | 10-2-2021 15:59 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/4.0/",
"book_id": "ffc4003d-750c-413c-b63b-035e5dac13ff",
"url": "https://library.oapen.org/bitstream/20.500.12657/23338/1/1006817.pdf||https://library.oapen.org/bitstream/20.500.12657/23338/1/1006817.pdf",
"author": "",
"title": "Evolutionary Genomics",
"publisher": "Springer Nature",
"isbn": "",
"section_idx": 0
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.0 | *Edited by Takafumi Nakamura*
This book provides the application of praxises in the field of engineering safety by learning from previous system failures. And it addresses the most recent developments in the theoretical and practical aspects of these important fields, which, due to their special nature, bring together in a systematic way, many disciplines of engineering, from the traditional to the most technologically advanced. The authors of these chapters are involved in using the system thinking and system engineering approaches at the scale of increased complexity and advanced computational solutions to such systems. The chapters cover the areas such as failure assessment in aeronautical engineering, seismic resistance of offshore pipeline engineering, electrical engineering, critical infrastructure failure, and system of system theory.
Published in London, UK © 2018 IntechOpen © flyparade / iStock
ISBN 978-1-78923-046-8
System of System Failures
| doab | 2025-04-07T04:13:03.838452 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 0
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.1 | System of System Failures
*Edited by Takafumi Nakamura*
**SYSTEM OF SYSTEM**
Edited by **Takafumi Nakamura**
**FAILURES**
| doab | 2025-04-07T04:13:03.838595 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 1
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.2 | **SYSTEM OF SYSTEM FAILURES**
Edited by **Takafumi Nakamura**
#### **System of System Failures**
http://dx.doi.org/10.5772/intechopen.68578 Edited by Takafumi Nakamura
#### **Contributors**
Liumila Muravieva, Steven Voldman, Martin Hromada, David Rehak, Jozef Zurek, Ryszard Kaleta, Datu Buyung Agusdinata, Takafumi Nakamura
#### **© The Editor(s) and the Author(s) 2018**
The rights of the editor(s) and the author(s) have been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights to the book as a whole are reserved by INTECHOPEN LIMITED. The book as a whole (compilation) cannot be reproduced, distributed or used for commercial or non-commercial purposes without INTECHOPEN LIMITED's written permission. Enquiries concerning the use of the book should be directed to INTECHOPEN LIMITED rights and permissions department (permissions@intechopen.com). Violations are liable to prosecution under the governing Copyright Law.
Individual chapters of this publication are distributed under the terms of the Creative Commons Attribution 3.0 Unported License which permits commercial use, distribution and reproduction of the individual chapters, provided the original author(s) and source publication are appropriately acknowledged. If so indicated, certain images may not be included under the Creative Commons license. In such cases users will need to obtain permission from the license holder to reproduce the material. More details and guidelines concerning content reuse and adaptation can be foundat http://www.intechopen.com/copyright-policy.html.
#### **Notice**
Statements and opinions expressed in the chapters are these of the individual contributors and not necessarily those of the editors or publisher. No responsibility is accepted for the accuracy of information contained in the published chapters. The publisher assumes no responsibility for any damage or injury to persons or property arising out of the use of any materials, instructions, methods or ideas contained in the book.
First published in London, United Kingdom, 2018 by IntechOpen eBook (PDF) Published by IntechOpen, 2019 IntechOpen is the global imprint of INTECHOPEN LIMITED, registered in England and Wales, registration number: 11086078, The Shard, 25th floor, 32 London Bridge Street London, SE19SG – United Kingdom Printed in Croatia
British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library
Additional hard and PDF copies can be obtained from orders@intechopen.com
System of System Failures Edited by Takafumi Nakamura p. cm. Print ISBN 978-1-78923-046-8 Online ISBN 978-1-78923-047-5 eBook (PDF) ISBN 978-1-83881-329-1
| doab | 2025-04-07T04:13:03.838903 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 2
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.3 | We are IntechOpen, the first native scientific publisher of Open Access books
3,400+ Open access books available
109,000+ International authors and editors
115M+ Downloads
151 Countries delivered to Our authors are among the
12.2%
Contributors from top 500 universities
Selection of our books indexed in the Book Citation Index in Web of Science™ Core Collection (BKCI)
| doab | 2025-04-07T04:13:03.839078 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 3
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.4 | Interested in publishing with us? Contact book.department@intechopen.com
Numbers displayed above are based on latest data collected. For more information visit www.intechopen.com
| doab | 2025-04-07T04:13:03.839132 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 4
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.5 | **Meet the editor**
Takafumi Nakamura has over 30 years of experience as an ICT infrastructure architect in the Japanese market and 5 years of experience in the Australian market. He is a principal IT architect at Fujitsu. His research interests include system management, software development, technical support for ICT systems, strategic maintenance planning, development of support technologies, and
risk management. He holds a leadership role as an ICT architect, system analyst, and system thinker. He is currently working for the Fujitsu Fsas Inc. He is a member of the board and the president in System Support Promotion Unit.
Contents
**Preface VII**
Takafumi Nakamura
Steven H. Voldman
**Technology 37**
**Section 3 Exploring Purposes 73**
**Section 2 Improving Goal Seeking and Viability 7**
**Electrostatic Discharge 9**
Chapter 3 **Vibration Strength of Pipelines 23** Muravieva Liudmila Victorovna
Józef Żurek and Ryszard Kaleta
Chapter 5 **Failures in a Critical Infrastructure System 75** David Rehak and Martin Hromada
Datu Buyung Agusdinata
Chapter 1 **Introductory Chapter: System of System Failures 3**
Chapter 2 **System and Component Failure from Electrical Overstress and**
Chapter 4 **Probabilistic Methods for Damage Assessment in Aviation**
Chapter 6 **Dealing with Complexities and Uncertainties in a System-of-Systems: Case Studies on Urban Systems 95**
**Section 1 Introduction 1**
| doab | 2025-04-07T04:13:03.839165 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 5
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.6 | Contents
#### **Preface XI**
Preface
tem of system theory.
tic contributions.
The purpose of this book is to foster discussion on the current applications in the field of engineering safety by learning from previous system failures. This book contains various application examples to promote a holistic view to manage and therefore mitigate system failures. The predominant worldview in the current engineering arena is that system fail‐ ures can be prevented at the design phase. This worldview is obvious if we examine the mainstream, current methodologies for managing the system failures. These methodologies use a reductionist approach. And it is often pointed out that most of such methodologies have difficulty coping with emergent properties in a proactive manner and preventing the introduction of various side effects from quick (i.e., temporary) fixes, which lead to repeat‐ ing failures of similar type. This book intends to provide managers with a comprehensive overview of the current state-of-the-art praxises in the field of engineering safety by holisti‐ cally examining the system failures to prevent further occurrence. Also, it provides manag‐ ers with a bird's-eye view learning from various approaches through utilizing system of system methodologies (SOSMs). The brief introduction and core concept of SOSMs are pro‐ vided in the introductory chapter. On the basis of SOSM framework, various approaches are developed for risk management and engineering system failure arena, that is, system of sys‐ tem failures (SOSFs). And SOSFs provide managers with a practical reflection to be able to bring to bear, on the complex, diverse, and rapidly changing problem situations they con‐
The chapters address the most recent developments in the theoretical and practical aspects of these important fields, which, due to their special nature, bring together in a systematic way, many disciplines of engineering, from the traditional to the most technologically ad‐ vanced. The authors of these chapters are various practitioners and theory developers in‐ volved in using the system thinking and system engineering approaches at the scale of increased complexity and advanced computational solutions to such systems. The chapters cover the areas such as failure assessment in aeronautical engineering, seismic resistance of offshore pipeline engineering, electrical engineering, critical infrastructure failure, and sys‐
The editor is grateful to all the authors and publishing process manager for their enthusias‐
**Dr. Takafumi Nakamura** Fujitsu Fsas Inc., Japan
front, holistic approaches based on the variety of possible perspectives.
| doab | 2025-04-07T04:13:03.839388 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 6
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.7 | Preface
The purpose of this book is to foster discussion on the current applications in the field of engineering safety by learning from previous system failures. This book contains various application examples to promote a holistic view to manage and therefore mitigate system failures. The predominant worldview in the current engineering arena is that system fail‐ ures can be prevented at the design phase. This worldview is obvious if we examine the mainstream, current methodologies for managing the system failures. These methodologies use a reductionist approach. And it is often pointed out that most of such methodologies have difficulty coping with emergent properties in a proactive manner and preventing the introduction of various side effects from quick (i.e., temporary) fixes, which lead to repeat‐ ing failures of similar type. This book intends to provide managers with a comprehensive overview of the current state-of-the-art praxises in the field of engineering safety by holisti‐ cally examining the system failures to prevent further occurrence. Also, it provides manag‐ ers with a bird's-eye view learning from various approaches through utilizing system of system methodologies (SOSMs). The brief introduction and core concept of SOSMs are pro‐ vided in the introductory chapter. On the basis of SOSM framework, various approaches are developed for risk management and engineering system failure arena, that is, system of sys‐ tem failures (SOSFs). And SOSFs provide managers with a practical reflection to be able to bring to bear, on the complex, diverse, and rapidly changing problem situations they con‐ front, holistic approaches based on the variety of possible perspectives.
The chapters address the most recent developments in the theoretical and practical aspects of these important fields, which, due to their special nature, bring together in a systematic way, many disciplines of engineering, from the traditional to the most technologically ad‐ vanced. The authors of these chapters are various practitioners and theory developers in‐ volved in using the system thinking and system engineering approaches at the scale of increased complexity and advanced computational solutions to such systems. The chapters cover the areas such as failure assessment in aeronautical engineering, seismic resistance of offshore pipeline engineering, electrical engineering, critical infrastructure failure, and sys‐ tem of system theory.
The editor is grateful to all the authors and publishing process manager for their enthusias‐ tic contributions.
> **Dr. Takafumi Nakamura** Fujitsu Fsas Inc., Japan
**Section 1**
**Introduction**
**Section 1**
| doab | 2025-04-07T04:13:03.839535 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 7
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.8 | **Introduction**
**Chapter 1**
**Provisional chapter**
**Introductory Chapter: System of System Failures**
**Introductory Chapter: System of System Failures**
DOI: 10.5772/intechopen.75793
Managers are expected to cope with ever-changing complexity and diversity. They are asked to tackle a much greater diversity of problems learning from previous failures. This book provides managers with a bird's eye view learning from various approaches through utilizing system of system methodologies (SOSM). In order to promote holistic view and promote creativity, Jackson [1, 2] introduced SOSM. SOSM classifies the world of objects into two dimensions: systems and participants. The system dimension has two domains: simple and complex. The participant dimension has three domains: unitary, plural, and coercive. On this
basis, holistic approaches can be classified into four types (**Figure 1**):
© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
© 2018 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use,
distribution, and reproduction in any medium, provided the original work is properly cited.
**Figure 1.** Systems approaches related to problem context in the system of system methodologies (SOSM).
Takafumi Nakamura
Takafumi Nakamura
**1. Introduction**
Additional information is available at the end of the chapter
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.75793
#### **Introductory Chapter: System of System Failures Introductory Chapter: System of System Failures**
DOI: 10.5772/intechopen.75793
#### Takafumi Nakamura Takafumi Nakamura
Additional information is available at the end of the chapter Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.75793
## **1. Introduction**
Managers are expected to cope with ever-changing complexity and diversity. They are asked to tackle a much greater diversity of problems learning from previous failures. This book provides managers with a bird's eye view learning from various approaches through utilizing system of system methodologies (SOSM). In order to promote holistic view and promote creativity, Jackson [1, 2] introduced SOSM. SOSM classifies the world of objects into two dimensions: systems and participants. The system dimension has two domains: simple and complex. The participant dimension has three domains: unitary, plural, and coercive. On this basis, holistic approaches can be classified into four types (**Figure 1**):
**Figure 1.** Systems approaches related to problem context in the system of system methodologies (SOSM).
© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2018 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
| doab | 2025-04-07T04:13:03.839589 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 8
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.9 | **2. System of system failures (SOSF)**
Based upon the SOSM framework, various approaches are developed for risk management and engineering system failure arena, that is, system of system failures (SOSF) [3–5].
**References**
John Wiley & Sons; 2003
[1] Jackson MC. Systems Thinking: Creative Holism for Managers. London and New York:
Introductory Chapter: System of System Failures http://dx.doi.org/10.5772/intechopen.75793 5
[2] Jackson MC. Creative holism: A critical systems approach to complex problem situations. Systems Research and Behavioral Science. September/October 2006;**23**(5):647-657
[3] Nakamura T, Kijima K. A methodology for learning from system failures and its applica-
[4] Nakamura T, Kijima K. System of system failures: Meta methodology for IT engineering safety. Systems Research and Behavioral Science. January/February 2009:**26**(1):29-47
[5] Nakamura T, Kijima K. Method for quantifying risk factors of system failures and its application to ICT. Risk Management. 2015;**16**:231-271. DOI: 10.1057/rm.2015.1
tion to PC server maintenance. Risk Management. 2008;**10**(1):1-31
In the Preface, the editor noted that this book intends to provide the reader with a comprehensive overview of the current state-of-the-art in engineering safety by holistically examining system failures for the purpose of preventing further occurrence of system failures. This provides managers a practical reflection to be able to bring to bear, on the complex, diverse and rapidly changing problem situations they confront, holistic approaches based on the variety of possible perspectives.
#### **3. The structure of the book**
A short conclusion closes the argument. In this introductory chapter, the editor sought to make clear the structure of the book and the logic underlying that structure. The book structure is summarized by SOSM in **Table 1**.
**Table 1.** Structure of the book.
### **Author details**
Takafumi Nakamura
Address all correspondence to: tk\_moek@nifty.com
Fujitsu Fsas Inc., Tokyo, Japan
| doab | 2025-04-07T04:13:03.839706 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 9
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.11 | **System and Component Failure from Electrical Overstress and Electrostatic Discharge Overstress and Electrostatic Discharge**
**System and Component Failure from Electrical**
DOI: 10.5772/intechopen.72677
Steven H. Voldman Steven H. Voldman Additional information is available at the end of the chapter
Additional information is available at the end of the chapter
http://dx.doi.org/10.5772/intechopen.72677
#### **Abstract**
Electrical overstress (EOS) and electrostatic discharge (ESD) have been an issue in devices, circuit and systems for electronics for many decades, as early as the 1970s, and continued to be an issue to today. In this chapter, the issue of EOS and ESD will be discussed. The sources of both EOS and ESD failure history will be discussed. EOS and ESD physical models, failure mechanisms, testing methods and solutions will be shown. The chapter will close with discussion on how to provide both EOS and ESD robust devices, circuits, and systems, design practices, and procedures, as well as EOS and ESD factory control programs. EOS sources also occur from design characteristics of devices, circuits, and systems.
**Keywords:** electrical overstress, electrostatic discharge, latchup, system failure, component failure
#### **1. Introduction**
Electrostatic discharge (ESD) and Electrical overstress (EOS) have been an issue with the coming of the electrical age, when electricity and electrical product were first introduced into the mainstream of society [1–5]. With the scaling of semiconductor components, electrostatic discharge (ESD) has been a growing issue [2]. With the introduction of electrical power systems, the telephone, and electronics, inventions such as circuit breakers, and fuses became the first type of electrical overstress protection concepts to avoid over-load of electronic systems [1, 16–25]. Electrostatic discharge (ESD) and electrical overstress (EOS) will be discussed in the following sections.
In electronic design, a plethora of electrical events can occur. **Figure 1** illustrates the type of topics including ESD, EOS, latchup as well as electromagnetic interference (EMI), and electromagnetic compatibility (EMC).
Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2018 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
© 2016 The Author(s). Licensee InTech. This chapter is distributed under the terms of the Creative Commons
The human body model is regarded as an electrostatic discharge (ESD) event, not an electrical overstress (EOS) event [1–11]. HBM represents the interaction the electrical discharge from a human being and component. The model assumes that the human being is the initial condition. The human body model (HBM) became of interest in early days in the mining industry in the 1950s. In the Bureau of Mines, investigation reports discussed the issue of electrostatic in the mining industry. A first publication was published by P.G. Guest, V.W. Sikora, and B.L. Lewis as the *Bureau of Mines, Report of Investigation 4833*, U.S. Department of Interior, January 1952 [7]. A second article of interest was published by D. Bulgin, referred as D. Bulgin. Static
System and Component Failure from Electrical Overstress and Electrostatic Discharge
http://dx.doi.org/10.5772/intechopen.72677
11
An early investigator of issues with the human body model standard was T. M. Madzy and L.A. Price II of IBM in 1979 discussed a test system titled "Module Electrostatic Discharge Simulator" [4]. In this article, it was discussed that the ESD simulator was used within IBM since 1974. In 1980, H. Calvin, H. Hyatt, H. Mellberg, and D. Pellinen proposed values for the resistance and capacitance for the human ESD event for the finger tip and field enhanced discharges in "Measurement of Fast Transients and Application to Human ESD," published in the 1980 Proceedings of the EOS/ESD Symposium [4, 10–11]. The proposed resistance for the finger tip was averaged 1920 Ω, and capacitance of 110 pF, whereas the field enhanced discharge was a resistance of 550 Ω, and 120 pF. In 1981, H. Hyatt, H. Calvin, and H. Mellberg investigated the human ESD event, published in the 1981 Proceedings of the EOS/ESD
HBM failure mechanisms are associated with permanent damage on the peripheral circuitry of a semiconductor chip [3]. Additionally, HBM failures can occur power rails and ESD power clamps between the power rails. HBM failures can occur in both passive and active semiconductor devices. The failure signature is typically isolated to a single device, or a few elements. ESD circuits are designed to be "tuned" to be responsive to specific pulse widths; this is an issue for EOS events since they are not "tuned" for EOS events. For example, the RC-triggered
HBM ESD failures are also distinct from EOS events [1, 4]. HBM events will not typically cause failures in the package, printed circuit board (PCB), or single component devices mounted on
Human body model (HBM) failures can occur in diode and MOSFET structures. Integrated circuit diode structures fail at the contact interface, silicon surface, or junction region. Human body model failure occurs in a metal oxide semiconductor field effect transistor (MOSFET) structure. Integrated circuit MOSFET structures failure occurs from MOSFET source-to-drain, or at the MOSFET gate. From HBM failures, typically, the failure is MOSFET source-to-drain failures [2, 3]. An example of an ESD protection network is known as a dual-diode network [3]. The dualdiode ESD network is a commonly used network for complimentary metal oxide semiconductor (CMOS) technology. A first p-n diode element is formed in an n-well region where the p-anode is the p-diffusion implant of the p-channel MOSFET device and the n-cathode is the n-well region connected to the power supply VDD. This is sometimes referred to as the "up diode." A second p-n diode element is formed in an p-well or p-substrate region
Electrofication*. British Journal of Applied Physics*, Supplemental 2, 1953 [8].
Symposium, titled "A closer look at the human ESD event" [4, 10–11].
ESD power clamp is tuned to the HBM pulse, not EOS events.
a printed circuit board.
**Figure 1.** ESD, EOS, Latchup, EMI and EMC.
| doab | 2025-04-07T04:13:03.839827 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 11
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.12 | **2. Electrostatic discharge (ESD)**
Electrostatic discharge (ESD) is a common form of component level failure from manufacturing, shipping, and handling. Today, the ESD models and performed for qualification and shipping of semiconductor components are as follows [4]:
Additional models that are still performed, but not used for qualification of components include [4]:
#### **2.1. Human body model (HBM)**
ESD pulse models have been established to quantify the interaction of semiconductor chips and human beings. An important model is the human body model (HBM). Today, HBM is the most widely established standard for the reliability and quality in the semiconductor industry [2–4, 6]. The HBM test is integrated into the qualification and release process of the quality and reliability teams for components in corporations, and foundries [6].
The human body model is regarded as an electrostatic discharge (ESD) event, not an electrical overstress (EOS) event [1–11]. HBM represents the interaction the electrical discharge from a human being and component. The model assumes that the human being is the initial condition.
The human body model (HBM) became of interest in early days in the mining industry in the 1950s. In the Bureau of Mines, investigation reports discussed the issue of electrostatic in the mining industry. A first publication was published by P.G. Guest, V.W. Sikora, and B.L. Lewis as the *Bureau of Mines, Report of Investigation 4833*, U.S. Department of Interior, January 1952 [7]. A second article of interest was published by D. Bulgin, referred as D. Bulgin. Static Electrofication*. British Journal of Applied Physics*, Supplemental 2, 1953 [8].
An early investigator of issues with the human body model standard was T. M. Madzy and L.A. Price II of IBM in 1979 discussed a test system titled "Module Electrostatic Discharge Simulator" [4]. In this article, it was discussed that the ESD simulator was used within IBM since 1974. In 1980, H. Calvin, H. Hyatt, H. Mellberg, and D. Pellinen proposed values for the resistance and capacitance for the human ESD event for the finger tip and field enhanced discharges in "Measurement of Fast Transients and Application to Human ESD," published in the 1980 Proceedings of the EOS/ESD Symposium [4, 10–11]. The proposed resistance for the finger tip was averaged 1920 Ω, and capacitance of 110 pF, whereas the field enhanced discharge was a resistance of 550 Ω, and 120 pF. In 1981, H. Hyatt, H. Calvin, and H. Mellberg investigated the human ESD event, published in the 1981 Proceedings of the EOS/ESD Symposium, titled "A closer look at the human ESD event" [4, 10–11].
**2. Electrostatic discharge (ESD)**
**Figure 1.** ESD, EOS, Latchup, EMI and EMC.
10 System of System Failures
• Human Body Model (HBM).
• Machine Model (MM).
• Transient Latchup.
include [4]:
• Latchup.
• Charged Device Model (CDM).
• Transmission Line Pulse (TLP).
**2.1. Human body model (HBM)**
• Very-Fast Transmission Line Pulse (VF-TLP).
shipping of semiconductor components are as follows [4]:
Electrostatic discharge (ESD) is a common form of component level failure from manufacturing, shipping, and handling. Today, the ESD models and performed for qualification and
Additional models that are still performed, but not used for qualification of components
ESD pulse models have been established to quantify the interaction of semiconductor chips and human beings. An important model is the human body model (HBM). Today, HBM is the most widely established standard for the reliability and quality in the semiconductor industry [2–4, 6]. The HBM test is integrated into the qualification and release process of the quality
and reliability teams for components in corporations, and foundries [6].
HBM failure mechanisms are associated with permanent damage on the peripheral circuitry of a semiconductor chip [3]. Additionally, HBM failures can occur power rails and ESD power clamps between the power rails. HBM failures can occur in both passive and active semiconductor devices. The failure signature is typically isolated to a single device, or a few elements. ESD circuits are designed to be "tuned" to be responsive to specific pulse widths; this is an issue for EOS events since they are not "tuned" for EOS events. For example, the RC-triggered ESD power clamp is tuned to the HBM pulse, not EOS events.
HBM ESD failures are also distinct from EOS events [1, 4]. HBM events will not typically cause failures in the package, printed circuit board (PCB), or single component devices mounted on a printed circuit board.
Human body model (HBM) failures can occur in diode and MOSFET structures. Integrated circuit diode structures fail at the contact interface, silicon surface, or junction region. Human body model failure occurs in a metal oxide semiconductor field effect transistor (MOSFET) structure. Integrated circuit MOSFET structures failure occurs from MOSFET source-to-drain, or at the MOSFET gate. From HBM failures, typically, the failure is MOSFET source-to-drain failures [2, 3].
An example of an ESD protection network is known as a dual-diode network [3]. The dualdiode ESD network is a commonly used network for complimentary metal oxide semiconductor (CMOS) technology. A first p-n diode element is formed in an n-well region where the p-anode is the p-diffusion implant of the p-channel MOSFET device and the n-cathode is the n-well region connected to the power supply VDD. This is sometimes referred to as the "up diode." A second p-n diode element is formed in an p-well or p-substrate region where the n-cathode is the n-diffusion implant of the n-channel MOSFET device, or the n+/n-well implant and the p-anode is the p-well region or p-substrate region connected to the power supply VSS. This is sometimes referred to as the "down diode." This circuit provides a "forward bias" ESD protection solution for positive and negative ESD pulse events to the two power rails VDD and VSS. An advantage of the dual-diode ESD network is that it is easily to migrate from technology generation to technology generation. In shallow trench isolation (STI) technology, this structure is scalable. A second advantage is that it has a low turn-on voltage of 0.7 V. A third advantage is that it can be designed with low capacitance, making it suitable for CMOS, advanced CMOS, and RF technologies. A fourth advantage is that it does not contain MOSFET gate dielectric failure mechanisms.
In the calibration and verification procedure, the JEDEC standard requires a 1 GHz oscilloscope, whereas the ESDA standard requires 3 GHz [13]. Both standards today are bandwidth limited signal since the CDM waveform is faster that 1 GHz. These oscilloscopes were chosen based on availability at the time. It is well know that the energy spectrum of the CDM pulse
System and Component Failure from Electrical Overstress and Electrostatic Discharge
http://dx.doi.org/10.5772/intechopen.72677
13
CDM event damage occurs in the semiconductor chip through the substrate. It can also occur through the power supply. Charge is stored on the package, and the substrate.; then the power supply rapidly discharges through the grounded pin. The CDM failure mechanism can be small "pin-hole" in a MOSFET gate structure; this can occur in receiver networks, as
The current path for charged device model (CDM) in components is significantly different from other electrostatic discharge (ESD) events. In the case of the charged device model (CDM), the package and/or chip substrate is charged through a power or ground rail. The component itself is charged slowly to a desired voltage state. As a result, the current flows from the component itself to the grounded pin during ESD testing. This is significantly from other ESD tests that ground a reference, and then apply an ESD event to a signal or power pin. As a result, the current path that a CDM event follows is from inside the component to pin that is grounded during test. To avoid CDM failures of the MOSFET gate structure, an additional charged device model (CDM) ESD network is used [3] The ESD network comprises of a first stage dual-diode network placed adjacent or in proximity of the signal pad. A second set of diodes (e.g. second stage network) are placed adjacent to the receiver circuit. A resistor is placed between the first and second stage. Three paths are possible for the CDM current from a charged ground rail (e.g. p-substrate) to the grounded receiver pin. For a positive charging of the substrate, the current flows from the substrate to any possible path that will reach the grounded signal pad node. A first path is through the n-channel MOSFET receiver circuit gate and to the second stage ESD diode. A second path is through the substrate to the second stage diode network. A
In the case of the first stage ESD protection circuit is far from the signal pad, the substrate resistance can be significant. For the third path, the total resistance from the grounded location to the grounded signal path is the sum of the substrate resistance and the ESD diode series resistance. In the case that the receiver network is adjacent to the second stage ESD network, the current will prefer to follow the second path instead of first path. When the impedance of the n-channel MOSFET receiver (e.g. Path A) is higher than resistance through the second path, the receiver gate structure can avoid rupturing of the MOSFET gate dielectric. To insure that the current flows through the second path through the second stage CDM ESD network, the circuit must be physically close to the receiver, and a low series resistance diode element.
Electrical overstress (EOS) has been an issue in devices, circuit and systems for electronics for many decades, as early as the 1970s, and continues to be an issue today [1]. EOS failures are occurring at the device manufacturer, supplier, assembly and the field. In the electronic
waveform can extend into the 5 GHz frequency.
third path is through the substrate to the first stage ESD network.
**3. Electrical overstress (EOS)**
well as metal interconnects.
*An example of a signal pin ESD network consisting of a grounded gate n-channel MOSFET device [3]. The grounded gate NMOS (also referred to as GGNMOS) ESD network is a commonly used network for complimentary metal oxide semiconductor (CMOS) technology. Typically, it is a n-channel MOS-FET whose MOSFET drain is connected to the signal pin, and whose MOSFET source and gate are connected to the ground power rail. This circuit remains "off" in normal operation. When the signal pin exceeds the MOSFET snapback voltage, this circuit discharges to the VSS power rail. When the signal pin is below the ground potential, the MOSFET drain forward biases to the p-well or p-substrate region. An advantage of the GGNMOS ESD network is that it is a natural scalable solution. As the technology scales, the MOSFET snapback voltage reduces, leading to an earlier turn-on of the MOSFET.*
#### **2.2. Charged device model (CDM)**
The charged device model is an electrostatic discharge (ESD) test method that is part of the qualification of semiconductor components [4]. The charged device model (CDM) standard is supported by ESD Association as ANSI/ESD ESD-STM5.3.1-1999 [12]. Presently, there are four CDM test standard (ESDA S5.3.1, JEDEC JESD22-C101, AEC-Q100-011 Rev. C, and JEITA ED-4701-300). Each require different test platform, testing, waveform, and calibration requirements [4]. The charged device model (CDM) event is associated with the charging of the semiconductor component substrate and package. The charging of the package occurs through direct contact charging, or field-induced charging process (e.g. the field-induced charge device model (FICDM)).
There is presently an effort to align the CDM standards between the ESD Association and the JEDEC organization, by establishing a joint ESDA/JEDEC standard. The ESDA/JEDEC joint standard (JS-002 2014) will replace existing CDM ESD standards JEDS22-C101 and ANSI/ESD S5.3.1. The new joint standard will preserve test systems in the field, and improve the waveform measurement process.
The charged device model (CDM) pulse is regarded as the fastest event of all the ESD events [4, 12–15]. Note that the CDM pulse waveform is influenced by the test platform and measurement metrology. The test platform is influenced by the field plate, field plate dielectric thickness and material type, and the probe assembly (e.g. test head, and ground plane). The metrology is influenced by the oscilloscope and verification module specifications.
First, the event is oscillatory. The CDM current pulse rise time is on the order of 250 ps, and with peak currents in the range of 10 A. The energy spectrum of the CDM pulse event extends to 5 GHz frequency. The CDM pulse waveform has a fast current pulse. The time scale of the CDM event is significantly lower than the thermal diffusion time; hence CDM events are in the "adiabatic regime" of a Wunsch-Bell power-to-failure curve [4].
In the calibration and verification procedure, the JEDEC standard requires a 1 GHz oscilloscope, whereas the ESDA standard requires 3 GHz [13]. Both standards today are bandwidth limited signal since the CDM waveform is faster that 1 GHz. These oscilloscopes were chosen based on availability at the time. It is well know that the energy spectrum of the CDM pulse waveform can extend into the 5 GHz frequency.
CDM event damage occurs in the semiconductor chip through the substrate. It can also occur through the power supply. Charge is stored on the package, and the substrate.; then the power supply rapidly discharges through the grounded pin. The CDM failure mechanism can be small "pin-hole" in a MOSFET gate structure; this can occur in receiver networks, as well as metal interconnects.
The current path for charged device model (CDM) in components is significantly different from other electrostatic discharge (ESD) events. In the case of the charged device model (CDM), the package and/or chip substrate is charged through a power or ground rail. The component itself is charged slowly to a desired voltage state. As a result, the current flows from the component itself to the grounded pin during ESD testing. This is significantly from other ESD tests that ground a reference, and then apply an ESD event to a signal or power pin. As a result, the current path that a CDM event follows is from inside the component to pin that is grounded during test.
To avoid CDM failures of the MOSFET gate structure, an additional charged device model (CDM) ESD network is used [3] The ESD network comprises of a first stage dual-diode network placed adjacent or in proximity of the signal pad. A second set of diodes (e.g. second stage network) are placed adjacent to the receiver circuit. A resistor is placed between the first and second stage. Three paths are possible for the CDM current from a charged ground rail (e.g. p-substrate) to the grounded receiver pin. For a positive charging of the substrate, the current flows from the substrate to any possible path that will reach the grounded signal pad node. A first path is through the n-channel MOSFET receiver circuit gate and to the second stage ESD diode. A second path is through the substrate to the second stage diode network. A third path is through the substrate to the first stage ESD network.
In the case of the first stage ESD protection circuit is far from the signal pad, the substrate resistance can be significant. For the third path, the total resistance from the grounded location to the grounded signal path is the sum of the substrate resistance and the ESD diode series resistance. In the case that the receiver network is adjacent to the second stage ESD network, the current will prefer to follow the second path instead of first path. When the impedance of the n-channel MOSFET receiver (e.g. Path A) is higher than resistance through the second path, the receiver gate structure can avoid rupturing of the MOSFET gate dielectric. To insure that the current flows through the second path through the second stage CDM ESD network, the circuit must be physically close to the receiver, and a low series resistance diode element.
| doab | 2025-04-07T04:13:03.840194 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 12
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.13 | **3. Electrical overstress (EOS)**
where the n-cathode is the n-diffusion implant of the n-channel MOSFET device, or the n+/n-well implant and the p-anode is the p-well region or p-substrate region connected to the power supply VSS. This is sometimes referred to as the "down diode." This circuit provides a "forward bias" ESD protection solution for positive and negative ESD pulse events to the two power rails VDD and VSS. An advantage of the dual-diode ESD network is that it is easily to migrate from technology generation to technology generation. In shallow trench isolation (STI) technology, this structure is scalable. A second advantage is that it has a low turn-on voltage of 0.7 V. A third advantage is that it can be designed with low capacitance, making it suitable for CMOS, advanced CMOS, and RF technologies. A fourth
advantage is that it does not contain MOSFET gate dielectric failure mechanisms.
**2.2. Charged device model (CDM)**
12 System of System Failures
form measurement process.
*An example of a signal pin ESD network consisting of a grounded gate n-channel MOSFET device [3]. The grounded gate NMOS (also referred to as GGNMOS) ESD network is a commonly used network for complimentary metal oxide semiconductor (CMOS) technology. Typically, it is a n-channel MOS-FET whose MOSFET drain is connected to the signal pin, and whose MOSFET source and gate are connected to the ground power rail. This circuit remains "off" in normal operation. When the signal pin exceeds the MOSFET snapback voltage, this circuit discharges to the VSS power rail. When the signal pin is below the ground potential, the MOSFET drain forward biases to the p-well or p-substrate region. An advantage of the GGNMOS ESD network is that it is a natural scalable solution. As the technology scales, the MOSFET snapback voltage reduces, leading to an earlier turn-on of the MOSFET.*
The charged device model is an electrostatic discharge (ESD) test method that is part of the qualification of semiconductor components [4]. The charged device model (CDM) standard is supported by ESD Association as ANSI/ESD ESD-STM5.3.1-1999 [12]. Presently, there are four CDM test standard (ESDA S5.3.1, JEDEC JESD22-C101, AEC-Q100-011 Rev. C, and JEITA ED-4701-300). Each require different test platform, testing, waveform, and calibration requirements [4]. The charged device model (CDM) event is associated with the charging of the semiconductor component substrate and package. The charging of the package occurs through direct contact charging,
or field-induced charging process (e.g. the field-induced charge device model (FICDM)).
There is presently an effort to align the CDM standards between the ESD Association and the JEDEC organization, by establishing a joint ESDA/JEDEC standard. The ESDA/JEDEC joint standard (JS-002 2014) will replace existing CDM ESD standards JEDS22-C101 and ANSI/ESD S5.3.1. The new joint standard will preserve test systems in the field, and improve the wave-
The charged device model (CDM) pulse is regarded as the fastest event of all the ESD events [4, 12–15]. Note that the CDM pulse waveform is influenced by the test platform and measurement metrology. The test platform is influenced by the field plate, field plate dielectric thickness and material type, and the probe assembly (e.g. test head, and ground plane). The
First, the event is oscillatory. The CDM current pulse rise time is on the order of 250 ps, and with peak currents in the range of 10 A. The energy spectrum of the CDM pulse event extends to 5 GHz frequency. The CDM pulse waveform has a fast current pulse. The time scale of the CDM event is significantly lower than the thermal diffusion time; hence CDM events are in
metrology is influenced by the oscilloscope and verification module specifications.
the "adiabatic regime" of a Wunsch-Bell power-to-failure curve [4].
Electrical overstress (EOS) has been an issue in devices, circuit and systems for electronics for many decades, as early as the 1970s, and continues to be an issue today [1]. EOS failures are occurring at the device manufacturer, supplier, assembly and the field. In the electronic industry, many products and applications are returned from the field due to "EOS" failure. To make progress in addressing the electrical overstress (EOS) issue, it is important to provide a framework for evaluation and analysis of EOS phenomena.
Electrical overstress (EOS) sources exist from natural phenomena, and power distribution [1, 14–25]. Switches, cables, and other power electronics that can be a source of electrical overstress. EOS sources exist in devices, circuits and systems. In the following sections, these issues will be discussed [1].
## **3.1. EOS design issues**
Many of the electrical overstress (EOS) issues can occur from the design of the semiconductor component, the system and its integration. Examples of EOS source design issues are as follows [1]:
> Field returns occur in all electronic components independent of the technology generation and period of time of evaluation. One of the key difficulties in the semiconductor industry is
System and Component Failure from Electrical Overstress and Electrostatic Discharge
http://dx.doi.org/10.5772/intechopen.72677
15
EOS events do not have a characteristic time response. EOS events are typically slower, and distinguishable from ESD events by having longer characteristic times. The time constant for
Electrical over-voltage (EOV), electric over-current (EOC), and electrical over-power (EOP) can lead to failure mechanisms; these can lead to melted packages, blown single component capacitors and resistors, ruptured packages, blown bond wires, cracked dielectrics, fused and
Visual external or internal inspection can be applied to evaluate EOS failure mechnanisms.
the ability to track, record and maintain a database of these field failures.
EOS events range from sub-microseconds to seconds.
The failure analysis process can comprise of the following steps:
melted metal layers, and molten silicon.
• Failure site identification and localization.
Visual damage signatures can include the following:
• Information gathering. • Failure verification.
**Figure 2.** Safe operating area (SOA).
• Root cause determination. • Feedback of root cause.
• Documentation reports.
**3.2. EOS failure mechanisms**
• Corrective action.
**Figure 2** illustrates the safe operating area (SOA) of a semiconductor device. There is a current limit, and a voltage limit on the borders of the SOA. At the corner of the SOA, the limitation is a thermal limit, and a second breakdown limit. Thermal limit has to do with the thermal limit of a device. The second breakdown limit has to do with second breakdown or thermal breakdown limit.
Testing and test simulation of devices, components and systems are an important part of the evaluation to electrical overstress (EOS) [4]. EOS test simulation is valuable part of understand EOS failures. EOS testing provides [1, 4]:
System and Component Failure from Electrical Overstress and Electrostatic Discharge http://dx.doi.org/10.5772/intechopen.72677 15
**Figure 2.** Safe operating area (SOA).
industry, many products and applications are returned from the field due to "EOS" failure. To make progress in addressing the electrical overstress (EOS) issue, it is important to provide
Electrical overstress (EOS) sources exist from natural phenomena, and power distribution [1, 14–25]. Switches, cables, and other power electronics that can be a source of electrical overstress. EOS sources exist in devices, circuits and systems. In the following sections, these
Many of the electrical overstress (EOS) issues can occur from the design of the semiconductor component, the system and its integration. Examples of EOS source design issues are as
**Figure 2** illustrates the safe operating area (SOA) of a semiconductor device. There is a current limit, and a voltage limit on the borders of the SOA. At the corner of the SOA, the limitation is a thermal limit, and a second breakdown limit. Thermal limit has to do with the thermal limit of a device. The second breakdown limit has to do with second breakdown or thermal
Testing and test simulation of devices, components and systems are an important part of the evaluation to electrical overstress (EOS) [4]. EOS test simulation is valuable part of under-
a framework for evaluation and analysis of EOS phenomena.
• Semiconductor process - application mismatch.
• Safe operating area (SOA) power rating violation. • Safe operating area (SOA) voltage rating violation. • Safe operating area (SOA) current rating violation.
• Transient safe operating area - di/dt and dv/dt.
stand EOS failures. EOS testing provides [1, 4]:
• Printed circuit board (PCB) inductance. • Printed circuit board (PCB) resistance.
issues will be discussed [1].
**3.1. EOS design issues**
14 System of System Failures
• Latchup sensitivity [5].
breakdown limit.
• Root cause analysis.
• Replication of failure signature.
• Technology benchmarking.
• System qualification.
• Technology EOS hardness evaluation.
• Component reliability qualification.
follows [1]:
Field returns occur in all electronic components independent of the technology generation and period of time of evaluation. One of the key difficulties in the semiconductor industry is the ability to track, record and maintain a database of these field failures.
EOS events do not have a characteristic time response. EOS events are typically slower, and distinguishable from ESD events by having longer characteristic times. The time constant for EOS events range from sub-microseconds to seconds.
Electrical over-voltage (EOV), electric over-current (EOC), and electrical over-power (EOP) can lead to failure mechanisms; these can lead to melted packages, blown single component capacitors and resistors, ruptured packages, blown bond wires, cracked dielectrics, fused and melted metal layers, and molten silicon.
The failure analysis process can comprise of the following steps:
#### **3.2. EOS failure mechanisms**
Visual external or internal inspection can be applied to evaluate EOS failure mechnanisms. Visual damage signatures can include the following:
Visual damage can also be evaluated from internal inspection. For internal inspection, the following visual damage signatures are:
these devices can be referred to as "voltage clamp" devices where dI/dV remains positive for all states; for the second group, there exists a region where dI/dV is negative. The first group can be classified as "voltage clamp devices" whereas the second group can be referred to as an "S-type I-V characteristic device", or as a "snapback device." In the classification of voltage suppression devices, the second classification can be associated with the directionality; a voltage suppres-
System and Component Failure from Electrical Overstress and Electrostatic Discharge
http://dx.doi.org/10.5772/intechopen.72677
17
The choice of electrical overstress (EOS) device to use in an application is dependent on the electrical characteristics, cost, and size. The electrical characteristics that are of interest are the
The types of voltage suppression devices used electrical overstress (EOS) are Transient Voltage Suppression (TVS) Diodes [22], Thyristor devices, Varistor devices [21], Polymer
Current-limiting devices can be used in a series configuration for electrical overstress (EOS)
The choice of the current-limiting EOS protection device is a function of the cost, size, rated
Diodes are uni-directional type EOS structure, but can be utilized in a forward or reverse breakdown mode of operation for a voltage limiting EOS solution [1]. Schottky diodes are also commonly used uni-directional electrical overstress (EOS) protection device [1]. Schottky diodes have a forward conduction state, and reverse blocking state. Schottky diodes have a lower forward turn-on (e.g. 0.35 V) compared to standard silicon p-n junction (e.g. 0.7 V). For electrical overstress, Schottky diodes are mounted on printed circuit board (PCB) by soldering in the leads through vias, or surface mount. Schottky diodes are not as commonly used within components to provide electrostatic discharge (ESD) protection due to lack of availability. Schottky diodes are uni-directional type EOS structure, but can be utilized in a forward or reverse breakdown mode of operation for a voltage limiting EOS solution. Zener diodes are
Zener diodes are used for electrostatic discharge (ESD) protection for high voltage and power applications. Zener diodes are not used for ESD protection for low voltage CMOS applications. For electrical overstress (EOS) single component Zener diodes are mounted on printed circuit board (PCB) through vias, or surface mount. Zener diodes are uni-directional type EOS
also used as a uni-directional electrical overstress (EOS) protection device [1].
t value, rated voltage, voltage drops, and application requirements.
Voltage Suppression (PVS) devices, and Gas Discharge Tube (GDT) devices [23].
protection. EOS current-limiting devices can be as follows [16–24]:
sion device can be "uni-directional" or "bidirectional."
breakdown voltage, and the forward conduction [1].
• Positive temperature coefficient (PTC) devices.
• Resistors.
• eFUSE.
• Resetting fuses.
• Circuit breakers.
current, time response, I2
• Non-resetting fuses.
There are certain categories of failures that electrostatic discharge (ESD) does not typically cause, and EOS events do cause. Failures that typically are caused by EOS phenomena but not ESD are as follows [1]:
Today, electrical overstress (EOS) is still an issue in today's electronic systems. To address electrical overstress in systems, electrical overstress (EOS) protection device are added to printed circuit boards (PCB), cards, and systems. The integration of EOS protection devices into systems.
#### **3.3. EOS protection devices**
Electrical overstress (EOS) protection devices are supported by a large variety of technologies. Although material and operation may differ between the EOS protection devices, their electrical characteristics can be classified into a few fundamental groups [16–25].
EOS protection networks can be identified as a voltage suppression device, or as a current-limiting device. The voltage suppression device limits the voltage observed on the signal pins or power rails of a component, preventing electrical over-voltage (EOV). The current-limiting device prevents a high current from reaching sensitive nodes, avoiding electrical over-current (EOC) [1].
Voltage suppression devices can also be sub-divided into two major classifications [1]. Voltage suppression devices can be segmented into devices that remain with a positive differential resistance, and those that undergo a negative resistance region. For positive differential resistance, these devices can be referred to as "voltage clamp" devices where dI/dV remains positive for all states; for the second group, there exists a region where dI/dV is negative. The first group can be classified as "voltage clamp devices" whereas the second group can be referred to as an "S-type I-V characteristic device", or as a "snapback device." In the classification of voltage suppression devices, the second classification can be associated with the directionality; a voltage suppression device can be "uni-directional" or "bidirectional."
The choice of electrical overstress (EOS) device to use in an application is dependent on the electrical characteristics, cost, and size. The electrical characteristics that are of interest are the breakdown voltage, and the forward conduction [1].
The types of voltage suppression devices used electrical overstress (EOS) are Transient Voltage Suppression (TVS) Diodes [22], Thyristor devices, Varistor devices [21], Polymer Voltage Suppression (PVS) devices, and Gas Discharge Tube (GDT) devices [23].
Current-limiting devices can be used in a series configuration for electrical overstress (EOS) protection. EOS current-limiting devices can be as follows [16–24]:
• Resistors.
• Package lead damage.
• Package discoloration.
• Melted metallurgy.
not ESD are as follows [1]:
• Package molding damage.
**3.3. EOS protection devices**
• Package pin damage. • Wire bond damage.
into systems.
• Molten silicon.
lowing visual damage signatures are:
• Printed circuit board (PCB) damage.
• Cracked inter-level dielectrics.
Visual damage can also be evaluated from internal inspection. For internal inspection, the fol-
There are certain categories of failures that electrostatic discharge (ESD) does not typically cause, and EOS events do cause. Failures that typically are caused by EOS phenomena but
Today, electrical overstress (EOS) is still an issue in today's electronic systems. To address electrical overstress in systems, electrical overstress (EOS) protection device are added to printed circuit boards (PCB), cards, and systems. The integration of EOS protection devices
Electrical overstress (EOS) protection devices are supported by a large variety of technologies. Although material and operation may differ between the EOS protection devices, their electri-
EOS protection networks can be identified as a voltage suppression device, or as a current-limiting device. The voltage suppression device limits the voltage observed on the signal pins or power rails of a component, preventing electrical over-voltage (EOV). The current-limiting device prevents a high current from reaching sensitive nodes, avoiding electrical over-current (EOC) [1]. Voltage suppression devices can also be sub-divided into two major classifications [1]. Voltage suppression devices can be segmented into devices that remain with a positive differential resistance, and those that undergo a negative resistance region. For positive differential resistance,
cal characteristics can be classified into a few fundamental groups [16–25].
• Foreign material.
16 System of System Failures
• Cracks.
• Corrosion.
The choice of the current-limiting EOS protection device is a function of the cost, size, rated current, time response, I2 t value, rated voltage, voltage drops, and application requirements.
Diodes are uni-directional type EOS structure, but can be utilized in a forward or reverse breakdown mode of operation for a voltage limiting EOS solution [1]. Schottky diodes are also commonly used uni-directional electrical overstress (EOS) protection device [1]. Schottky diodes have a forward conduction state, and reverse blocking state. Schottky diodes have a lower forward turn-on (e.g. 0.35 V) compared to standard silicon p-n junction (e.g. 0.7 V). For electrical overstress, Schottky diodes are mounted on printed circuit board (PCB) by soldering in the leads through vias, or surface mount. Schottky diodes are not as commonly used within components to provide electrostatic discharge (ESD) protection due to lack of availability. Schottky diodes are uni-directional type EOS structure, but can be utilized in a forward or reverse breakdown mode of operation for a voltage limiting EOS solution. Zener diodes are also used as a uni-directional electrical overstress (EOS) protection device [1].
Zener diodes are used for electrostatic discharge (ESD) protection for high voltage and power applications. Zener diodes are not used for ESD protection for low voltage CMOS applications. For electrical overstress (EOS) single component Zener diodes are mounted on printed circuit board (PCB) through vias, or surface mount. Zener diodes are uni-directional type EOS structure, but can be utilized primarily in reverse breakdown mode of operation for a voltage limiting EOS solution.
One of the disadvantages of the GDT devices is the slow turn-on times typically in the micro-seconds. An example of some of the electrical characteristics can exhibit d.c. breakdown from 75 to 600 V, with a single surge response of 40 kA in 10–20 s, or multiple surges of magnitude of 20 kA. The electrical circuit breaker is used in industrial, commercial, and residential electrical systems for high currents. Electrical circuit breakers have issues of physical size, weight, cost, and time response. Circuit breakers can be used to protect household appliances, and large scale switchgear high voltage circuits. The circuit breaker is an electrical switch designed for the purpose of electrical over-current events, short circuits, or fault detection. Circuit breakers are typically "tripped" by the high current event, and can be manually reset. The concept of
System and Component Failure from Electrical Overstress and Electrostatic Discharge
http://dx.doi.org/10.5772/intechopen.72677
19
A class of circuit breakers is the thermal-magnetic circuit breaker [1]. Thermal-magnetic circuit breakers are used to avoid "short-circuit" currents. Thermal-magnetic circuit breakers are sensitive to temperature. Thermal-magnetic circuit breakers contain a bi-metal switch and an electromagnet. The bi-metal switch provides over-current protection. During current over-load, the bi-metal switch heats up, leading to bending of the element. The electromagnet
Power controllers are used for low power and low voltage applications; power controllers typically are low voltage high efficiency products that can carry amperes of current per channel. Buck-converters use over-current protection logic and networks; over-current functions protect the switching converter from an output short by monitoring current flow in the application. Hence, in power applications, it is possible to integrate electrical over-voltage (EOV) and electrical over-current (EOC) within a component design. Many analog and power applications also contain thermal protection networks as well to avoid thermal runaway and EOS damage.
Future challenges exist in improve reliability and safety in components and systems due to electrostatic discharge (ESD) and electrical overstress (EOS). Challenges include the following:
• Maintaining chip and system level performance objectives without lowering of ESD and
In conclusion, ESD and EOS failures occur in devices, components and systems in electronics in the past, and in the future with the introduction of both single component to VLSI technology.
• Achieving EOS and ESD standards protection levels in future technology generations.
• Electronic system failure from CMOS latchup in scaled future technology.
• Electronic system failure from overheating in handheld and portable devices.
the circuit breaker was invented by Charles Grafton Page, in 1836 [17].
responds to short-circuit currents [1].
**4. Challenges in the future**
EOS protection levels.
**5. Conclusions**
Zener diodes are used uni-directional electrical overstress (EOS) protection device. Zener diodes are typically used as a voltage clamping EOS protection device, and typically used in the breakdown state. Schottky and Zener diodes can both be integrated into a given application.
An EOS protection device used for high voltages is the varistor. A varistor is also known as a voltage dependent resistor (VDR). The varistor element behaves like a diode, forming a nonlinear current-voltage (I-V characteristic).
Another EOS protection device is the metal oxide varistor (MOV) device; this is the most common varistor composition [1, 21]. Zinc oxide, combined with other metal oxides are integrated between two metal electrodes. Metal oxide varistors can also include bismuth, cobalt, and manganese. The operation of the MOV device is based on conduction through ZnO grains; current flows "diodelike" through the grain structures creating a low current flow at low voltages. At higher voltages, the current flow is dominated by a combination of thermionic emissions and tunneling. This diode-like behavior forms the diode-like characteristic provides the high resistance/low voltage state, and the low resistance/high voltage state. An advantage of the MOV structure is it has a high trigger voltage, making it suitable for EOS protection in power electronics (e.g. 120–700 V applications) [1, 21]. The disadvantage of these elements is that it has high capacitance, high on-resistance, high trigger voltage, and variability of the device response (e.g. on-resistance and clamping voltage) in the MOV device characteristics. Key device parameters of varistor are the energy rating, operating voltage, response time, maximum current and breakdown voltages.
Gas discharge tubes (GDT) devices can be used to avoid electrical overstress (EOS) in systems [1, 23]. Gas discharge tubes (GDT) are bidirectional, allowing for protection for both positive and negative EOS events. GDT elements are suitable from surge protection. GDT devices have high trigger voltages (unless used as a first stage followed by other low voltage secondary EOS solutions) [1].
Gas-filled tubes (GDT) utilize electrical discharge in gases. An applied voltage initiates the device by ionizing the electrical gas, followed by electrical glow discharge, and an electrical arc. With creation of an electrical arc, the GDT device becomes a low resistance shunt for EOS protection. These gas-filled tubes can contain hydrogen, deuterium, and noble gases (e.g. helium, neon, argon, krypton, and xenon). GDT devices can vary their electrical characteristics by choices of the gas type, pressure, electrode design, and spacings.
GDT devices undergo three states: (1) electrical breakdown, (2) glow discharge, and (3) electrical arc [1, 23]. The electrical breakdown is a high voltage low current state prior to triggering of the GDT device. A glow discharge region forms a second state which incorporates a low current high voltage state. Lastly, after full ionization of the gas, a low voltage high current state occurs with a low "on-resistance."
GDT devices have high trigger voltages suitable for LDMOS power electronic applications to HV LDMOS (e.g. 120 V), and UHV LDMOS applications (e.g. 600–700 V) [1, 23]. These devices are used in a number of high voltage switch devices, such as ignitrons, krytons, and thyratrons. One of the disadvantages of the GDT devices is the slow turn-on times typically in the micro-seconds. An example of some of the electrical characteristics can exhibit d.c. breakdown from 75 to 600 V, with a single surge response of 40 kA in 10–20 s, or multiple surges of magnitude of 20 kA.
The electrical circuit breaker is used in industrial, commercial, and residential electrical systems for high currents. Electrical circuit breakers have issues of physical size, weight, cost, and time response. Circuit breakers can be used to protect household appliances, and large scale switchgear high voltage circuits. The circuit breaker is an electrical switch designed for the purpose of electrical over-current events, short circuits, or fault detection. Circuit breakers are typically "tripped" by the high current event, and can be manually reset. The concept of the circuit breaker was invented by Charles Grafton Page, in 1836 [17].
A class of circuit breakers is the thermal-magnetic circuit breaker [1]. Thermal-magnetic circuit breakers are used to avoid "short-circuit" currents. Thermal-magnetic circuit breakers are sensitive to temperature. Thermal-magnetic circuit breakers contain a bi-metal switch and an electromagnet. The bi-metal switch provides over-current protection. During current over-load, the bi-metal switch heats up, leading to bending of the element. The electromagnet responds to short-circuit currents [1].
Power controllers are used for low power and low voltage applications; power controllers typically are low voltage high efficiency products that can carry amperes of current per channel. Buck-converters use over-current protection logic and networks; over-current functions protect the switching converter from an output short by monitoring current flow in the application. Hence, in power applications, it is possible to integrate electrical over-voltage (EOV) and electrical over-current (EOC) within a component design. Many analog and power applications also contain thermal protection networks as well to avoid thermal runaway and EOS damage.
| doab | 2025-04-07T04:13:03.840942 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 13
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.14 | **4. Challenges in the future**
structure, but can be utilized primarily in reverse breakdown mode of operation for a voltage
Zener diodes are used uni-directional electrical overstress (EOS) protection device. Zener diodes are typically used as a voltage clamping EOS protection device, and typically used in the breakdown state. Schottky and Zener diodes can both be integrated into a given
An EOS protection device used for high voltages is the varistor. A varistor is also known as a voltage dependent resistor (VDR). The varistor element behaves like a diode, forming a non-
Another EOS protection device is the metal oxide varistor (MOV) device; this is the most common varistor composition [1, 21]. Zinc oxide, combined with other metal oxides are integrated between two metal electrodes. Metal oxide varistors can also include bismuth, cobalt, and manganese. The operation of the MOV device is based on conduction through ZnO grains; current flows "diodelike" through the grain structures creating a low current flow at low voltages. At higher voltages, the current flow is dominated by a combination of thermionic emissions and tunneling. This diode-like behavior forms the diode-like characteristic provides the high resistance/low voltage state, and the low resistance/high voltage state. An advantage of the MOV structure is it has a high trigger voltage, making it suitable for EOS protection in power electronics (e.g. 120–700 V applications) [1, 21]. The disadvantage of these elements is that it has high capacitance, high on-resistance, high trigger voltage, and variability of the device response (e.g. on-resistance and clamping voltage) in the MOV device characteristics. Key device parameters of varistor are the energy rating,
Gas discharge tubes (GDT) devices can be used to avoid electrical overstress (EOS) in systems [1, 23]. Gas discharge tubes (GDT) are bidirectional, allowing for protection for both positive and negative EOS events. GDT elements are suitable from surge protection. GDT devices have high trigger voltages (unless used as a first stage followed by other low voltage second-
Gas-filled tubes (GDT) utilize electrical discharge in gases. An applied voltage initiates the device by ionizing the electrical gas, followed by electrical glow discharge, and an electrical arc. With creation of an electrical arc, the GDT device becomes a low resistance shunt for EOS protection. These gas-filled tubes can contain hydrogen, deuterium, and noble gases (e.g. helium, neon, argon, krypton, and xenon). GDT devices can vary their electrical characteris-
GDT devices undergo three states: (1) electrical breakdown, (2) glow discharge, and (3) electrical arc [1, 23]. The electrical breakdown is a high voltage low current state prior to triggering of the GDT device. A glow discharge region forms a second state which incorporates a low current high voltage state. Lastly, after full ionization of the gas, a low voltage high current
GDT devices have high trigger voltages suitable for LDMOS power electronic applications to HV LDMOS (e.g. 120 V), and UHV LDMOS applications (e.g. 600–700 V) [1, 23]. These devices are used in a number of high voltage switch devices, such as ignitrons, krytons, and thyratrons.
operating voltage, response time, maximum current and breakdown voltages.
tics by choices of the gas type, pressure, electrode design, and spacings.
limiting EOS solution.
18 System of System Failures
ary EOS solutions) [1].
state occurs with a low "on-resistance."
linear current-voltage (I-V characteristic).
application.
Future challenges exist in improve reliability and safety in components and systems due to electrostatic discharge (ESD) and electrical overstress (EOS). Challenges include the following:
| doab | 2025-04-07T04:13:03.842200 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 14
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.15 | **5. Conclusions**
In conclusion, ESD and EOS failures occur in devices, components and systems in electronics in the past, and in the future with the introduction of both single component to VLSI technology. Significant advancements have been made in the understanding of failure mechanisms, as well as solutions to address them have been applied in semiconductor electronics.
[13] JEDEC. JESD22-C101-A. A Field-Induced Charged Device Model Test Method for Electrostatic Discharge-Withstand Thresholds of Microelectronic Components; 2000 [14] Renninger R, Jon M, Lin D, Diep T, Welsher T. A field-induced charged device model simulator. In: Proceedings of the Electrical Overstress/Electrostatic Discharge (EOS/ESD)
System and Component Failure from Electrical Overstress and Electrostatic Discharge
http://dx.doi.org/10.5772/intechopen.72677
21
[15] Renninger RG. Mechanisms of charged device electrostatic discharges. In: Proceedings of the Electrical Overstress/Electrostatic Discharge (EOS/ESD) Symposium; 1991. p. 127-143
[17] Page CG. Improvement in induction-coil apparatus and in circuit breakers. US Patent
[18] Wright A, Newbury PG. Electric Fuses. 3rd ed. Institute of Electrical Engineers;
[19] Kraz V. Origins of EOS in manufacturing environment. In: Proceedings of the Electrical
[20] Kaschani KT, Gaertner R. The impact of electrical overstress on the design, handling and application of integrated circuits. In: Proceedings of the Electrical Overstress/ Elec-
[21] Philipp HR, Levinson LM. Transient Protection with ZnO Varistors: Technical Considerations. In: Proceedings of the Electrical Overstress / Electrostatic Discharge (EOS/ESD)
[22] Hopkins DC. Protective level comparisons for voltage transient suppressors. In: Proceedings of the Electrical Overstress/ Electrostatic Discharge (EOS/ESD) Symposium;
[23] Bazarian A. Gas tube surge arresters for control of transient voltages. In: Proceedings of the Electrical Overstress/Electrostatic Discharge (EOS/ESD) Symposium; 1980. p. 44-53
[24] Horgan EL. Analytical assessment of electrical overstress effects on electronic systems. In: Proceedings of the Electrical Overstress/Electrostatic Discharge (EOS/ESD) Sym-
[25] Durgin DL. An overview of the sources and effects of electrical overstress, In: Proceedings of the Electrical Overstress/ Electrostatic Discharge (EOS/ESD) Symposium; 1980.
Overstress/Electrostatic Discharge (EOS/ESD) Symposium; 2009. p. 44-48
[16] Edison TA. Fuse block. US Patent No. 438,305, October 14th; 1890
trostatic Discharge (EOS/ESD) Symposium; 2011. p. 220-229
Symposium; 1989. p. 59-71
No. 76,654; April 14th; 1868
Symposium; 1980. p. 26-34
posium; 1980. p. 140-148
1980. p. 35-43
p. 154-160
Piscataway, New Jersey, USA; 2004. p. 2-10
| doab | 2025-04-07T04:13:03.842269 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 15
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.16 | **Author details**
Steven H. Voldman
Address all correspondence to: voldman@ieee.org
Institute of Electrical Engineering, IEEE fellow, United States of America
| doab | 2025-04-07T04:13:03.842444 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 16
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.18 | 2. Vibration structure
It is used as a reliability factor of the structure under vibrations [1, 3]. Under harmonic vibrations, the vibration velocity can serve as a reliability criterion for the pipeline.
The harmonic vibrations are characterized by two parameters: frequency of vibrations and displacement amplitude:
$$y = Y\_0 \sin \omega t \tag{1}$$
M zð Þ¼ EIy00ð Þ¼� <sup>z</sup> EI <sup>i</sup>
<sup>σ</sup>makc <sup>¼</sup> EI <sup>i</sup>
<sup>ω</sup><sup>i</sup> <sup>¼</sup> <sup>i</sup> 2 π2 L2
2 π2
At initial approximation, certain typical ideal forms of vibrations are used for vibration analysis. Then, the natural vibration frequency of the pipeline is expressed by the following formula:
According to [3, 5], the maximum vibration velocity is determined from the following equa-
where vbase is a vibration velocity of the base plate; η is a dynamic magnification factor; k is a
Expansion bends are regarded as vibration damping elements for the pipelines to ensure their vibration resistance. The expansion bends prevent the transfer of vibrations along the pipeline.
Stresses across the cross sections of the pipeline under natural vibrations can be determined
<sup>0</sup> ydz= Ð 1 <sup>0</sup> <sup>y</sup><sup>2</sup> ffiffiffiffiffiffiffi EIg rF
s
where I is a moment of inertia.
Maximum stresses in the pipeline:
then W is a moment of resistance in the pipeline.
where F is a square section of the pipeline.
4. Analysis of pipeline vibrations
form engagement factor: <sup>k</sup> <sup>¼</sup> <sup>Ð</sup> <sup>1</sup>
tions and current coordinate.
from the following equation:
tion:
2 π2
Figure 1. Force excitation of vibrations. (а) Excitation through an elastic support; (b) kinematic excitation of vibrations.
<sup>L</sup><sup>2</sup> <sup>Y</sup><sup>0</sup> sin <sup>i</sup>π<sup>z</sup>
<sup>L</sup> (5)
Vibration Strength of Pipelines
25
http://dx.doi.org/10.5772/intechopen.72794
(7)
WL<sup>2</sup> <sup>Y</sup>0, (6)
Vmax ¼ vbaseKη (8)
dz, y, z are dimensionless forms of pipeline vibra-
Vibration velocity and vibration acceleration are expressed as follows:
$$
\omega v\_0 = \omega Y\_0; \quad \mathcal{g}\_0 = \omega^2 Y\_0 \tag{2}
$$
| doab | 2025-04-07T04:13:03.842483 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 18
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.19 | 3. Bending vibrations of the pipelines
Under bending vibrations of the pipelines, when distribution of stresses and vibration velocities is significantly different for various fixing conditions, factor с shall be determined individually for each case [3].
$$\frac{d^2y}{dx^2} + \frac{m\omega^2}{T}y = 0\tag{3}$$
Form of the elastic curve of the pipeline is expressed by a sine wave.
$$y(z) = Y\_0 \sin \frac{i\pi z}{L}, y(z) = Y\_0 \overline{y}(z) \tag{4}$$
where Y0 is an amplitude of vibrations.
Bending moment in the random location on the pipeline is equal to:
Figure 1. Force excitation of vibrations. (а) Excitation through an elastic support; (b) kinematic excitation of vibrations.
$$M(z) = EIy''(z) = -EI\frac{\text{i}^2 \pi^2}{L^2} Y\_0 \sin\frac{i\pi z}{L} \tag{5}$$
where I is a moment of inertia.
When calculations are made in regard of hazard assessment of the pipeline vibrations, criteria shall be used to determine the vibration behavior of the pipelines. Vibrations of the pipelines caused by external effects such as impact and earthquake can be described by a general
For example, in case of vibrations of the base plate to which a pipeline support is fixed (Figure 1а and b), the latter has a dynamic impact by which degree is determined by the
It is highly important to have an opportunity to determine stressed condition of the pipelines subject to vibrations in the form of elastic curves, which occur during vibrations caused by
It is used as a reliability factor of the structure under vibrations [1, 3]. Under harmonic
The harmonic vibrations are characterized by two parameters: frequency of vibrations and
<sup>v</sup><sup>0</sup> <sup>¼</sup> <sup>ω</sup>Y0; g<sup>0</sup> <sup>¼</sup> <sup>ω</sup><sup>2</sup>
Under bending vibrations of the pipelines, when distribution of stresses and vibration velocities is significantly different for various fixing conditions, factor с shall be determined individ-
mω<sup>2</sup>
d2 y dx<sup>2</sup> <sup>þ</sup>
y zð Þ¼ <sup>Y</sup><sup>0</sup> sin <sup>i</sup>π<sup>z</sup>
Form of the elastic curve of the pipeline is expressed by a sine wave.
Bending moment in the random location on the pipeline is equal to:
y ¼ Y<sup>0</sup> sin ωt (1)
Y<sup>0</sup> (2)
<sup>T</sup> <sup>y</sup> <sup>¼</sup> <sup>0</sup> (3)
<sup>L</sup> ,y zð Þ¼ <sup>Y</sup>0y zð Þ (4)
vibrations, the vibration velocity can serve as a reliability criterion for the pipeline.
Vibration velocity and vibration acceleration are expressed as follows:
3. Bending vibrations of the pipelines
where Y0 is an amplitude of vibrations.
integral of the forced vibration equation.
Stresses in the pipeline wall can be regarded as criteria.
support yield δA, Р = Y0δ<sup>A</sup> cos ωt.
external excitations.
24 System of System Failures
2. Vibration structure
displacement amplitude:
ually for each case [3].
Maximum stresses in the pipeline:
$$
\sigma\_{\text{makc}} = EI \frac{\text{i}^2 \pi^2}{\text{W} \text{L}^2} \text{Y}\_{0\text{v}} \tag{6}
$$
then W is a moment of resistance in the pipeline.
At initial approximation, certain typical ideal forms of vibrations are used for vibration analysis. Then, the natural vibration frequency of the pipeline is expressed by the following formula:
$$
\omega\_i = \frac{\dot{\mathbf{r}}^2 \pi^2}{L^2} \sqrt{\frac{\text{Elg}}{\rho F}} \tag{7}
$$
where F is a square section of the pipeline.
According to [3, 5], the maximum vibration velocity is determined from the following equation:
$$\mathbf{V}\_{\text{max}} = \mathbf{v}\_{\text{base}} \mathbf{K} \mathbf{\eta} \tag{8}$$
where vbase is a vibration velocity of the base plate; η is a dynamic magnification factor; k is a form engagement factor: <sup>k</sup> <sup>¼</sup> <sup>Ð</sup> <sup>1</sup> <sup>0</sup> ydz= Ð 1 <sup>0</sup> <sup>y</sup><sup>2</sup> dz, y, z are dimensionless forms of pipeline vibrations and current coordinate.
Expansion bends are regarded as vibration damping elements for the pipelines to ensure their vibration resistance. The expansion bends prevent the transfer of vibrations along the pipeline.
#### 4. Analysis of pipeline vibrations
Stresses across the cross sections of the pipeline under natural vibrations can be determined from the following equation:
$$
\sigma\_k = \frac{ED}{2} \mathcal{C}\_k(\omega) \frac{\partial^2 y\_k(z)}{\partial z^2} . \tag{9}
$$
p1 ¼ þ <sup>δ</sup> � <sup>4</sup>Et
p1 <sup>¼</sup> <sup>4</sup><sup>E</sup>
Combined stress in the axial direction due to pressure pulsations and vibration:
Or
5. Spectral transforms
STð Þ¼ f
are interrelated by a complex Fourier series.
At small values of <sup>ω</sup>0: <sup>S</sup>ð Þ <sup>ω</sup> <sup>≈</sup> <sup>1</sup>
ð∞ �∞
¼ 1 T ∙ X∞ m¼�∞
coefficients of the function (15) into a Fourier series as follows:
F tðÞ¼ sin ω0t at 0 < t <
ω.
Then amplitude spectrum is determined by the following expression:
sTð Þt e
<sup>S</sup> <sup>m</sup> T � �∙<sup>δ</sup> <sup>f</sup> � <sup>m</sup>
The function containing n vibrations is described by the following equations,
the time function
D2 ω2
<sup>D</sup><sup>2</sup> � <sup>ω</sup><sup>2</sup>
Spectral density corresponds to the spectral form of the internal pressure sinusoidal vibrations u(t) = sin ω0t. Periodic function spectrum ST(f) is determined by a direct Fourier transform of
> 1 T X∞ m¼�∞
> > m¼�∞
and F tðÞ¼ 0 at 0 > t >
�j2πftdt <sup>¼</sup> S fð Þ<sup>∙</sup>
T � � <sup>¼</sup> <sup>X</sup><sup>∞</sup>
> <sup>T</sup> <sup>∙</sup><sup>S</sup> <sup>m</sup> T � �
2ω<sup>0</sup> <sup>n</sup>πω ω0
� � � �
ω2 <sup>0</sup> � ω<sup>2</sup>
� � � �
Complex weights of S-functions at frequencies multiple of 1/Т. They represent expansion
Time function sТ(t) and spectral function С[т] of the impact are the main characteristics. They
2πn ω0
Sð Þ¼ ω
C m½ �¼ <sup>1</sup>
� �ta (12)
Vibration Strength of Pipelines
27
http://dx.doi.org/10.5772/intechopen.72794
� �td (13)
σ<sup>t</sup> ¼ σ<sup>Δ</sup>pt þ σ<sup>V</sup> (14)
<sup>δ</sup> <sup>f</sup> � <sup>m</sup> T � �
C m½ �∙<sup>δ</sup> <sup>f</sup> � <sup>m</sup>
T � �
> 2πn ω0 :
(15)
(16)
Stresses acting on the pipeline can be expressed as follows:
$$
\sigma = Y\_0 \frac{EI\overline{y''}}{W(z)'} ,
$$
where EI is bending rigidity of pipe, N�m<sup>2</sup> .
Allowable amplitude of vibration equals to:
$$[Y\_0] = \frac{[\sigma] \mathcal{W}(z)}{E I \overline{y''}},\tag{10}$$
where [σ] is a permissible stress in the pipe metal.
The analysis of pipeline vibrations is performed using root mean square of instantaneous vibration parameters over a period determined by the following formula [8]:
$$\overline{y} = \frac{1}{T} \int\_0^T y^2 dt. \tag{11}$$
Measurement results of real pipeline vibrations show that such vibrations are of complex, and in some cases, they are of random nature. For the determination of stresses in the pipeline walls, the process loads Рр due to operating pressure of the product being transported, the effect of hydrostatic water head pressure shall be taken into consideration as well as timevariable loading on the pipeline such as pressure pulsations and seismic forces. It is effective to use spectral method during the analysis of the pipeline random vibrations [1, 3].
The internal pressure of the gas line generates random vibrations.
Elastic stress in the pipe walls can be expressed as follows:
$$\delta = \frac{F}{A} = \frac{\overline{p}\_2 D}{2t}, \text{ when } \overline{p}\_2 = 2Et\frac{\Delta D}{D^2} = \frac{4Et}{D^2}d,\tag{11a}$$
where δ is a dynamic stres, N/m<sup>2</sup> . Pressure changes in this manner p sin ωt ¼ p2 þ p3, in this case the balance between the elastic and internal force in the pipe wall shall be equal to:
$$
\overline{\mathbf{p}}\_1 = \overline{\mathbf{p}}\_2 + \overline{\mathbf{p}}\_3 = \frac{4Et}{D^2}\mathbf{d} + \mathbf{t}\delta\mathbf{a}
$$
As displacement and acceleration, at a given frequency, are related by �ω<sup>2</sup> , the expression can be changed to:
#### Vibration Strength of Pipelines http://dx.doi.org/10.5772/intechopen.72794 27
$$\overline{\mathbf{p}}\_1 = + \left[ \delta - \frac{4Et}{D^2 \alpha^2} \right] \mathbf{t} \tag{12}$$
Or
<sup>σ</sup><sup>k</sup> <sup>¼</sup> ED
Stresses acting on the pipeline can be expressed as follows:
where EI is bending rigidity of pipe, N�m<sup>2</sup>
26 System of System Failures
Allowable amplitude of vibration equals to:
where [σ] is a permissible stress in the pipe metal.
<sup>2</sup> Ckð Þ <sup>ω</sup>
σ ¼ Y<sup>0</sup>
.
½ �¼ Y<sup>0</sup>
<sup>y</sup> <sup>¼</sup> <sup>1</sup> T ðT 0 y2
use spectral method during the analysis of the pipeline random vibrations [1, 3].
The internal pressure of the gas line generates random vibrations.
<sup>A</sup> <sup>¼</sup> <sup>p</sup>2<sup>D</sup>
As displacement and acceleration, at a given frequency, are related by �ω<sup>2</sup>
Elastic stress in the pipe walls can be expressed as follows:
<sup>δ</sup> <sup>¼</sup> <sup>F</sup>
where δ is a dynamic stres, N/m<sup>2</sup>
be changed to:
vibration parameters over a period determined by the following formula [8]:
∂2 ykð Þz
EIy} W zð Þ,
½ � σ W zð Þ
The analysis of pipeline vibrations is performed using root mean square of instantaneous
Measurement results of real pipeline vibrations show that such vibrations are of complex, and in some cases, they are of random nature. For the determination of stresses in the pipeline walls, the process loads Рр due to operating pressure of the product being transported, the effect of hydrostatic water head pressure shall be taken into consideration as well as timevariable loading on the pipeline such as pressure pulsations and seismic forces. It is effective to
<sup>2</sup><sup>t</sup> , when <sup>p</sup><sup>2</sup> <sup>¼</sup> <sup>2</sup>EtΔ<sup>D</sup>
case the balance between the elastic and internal force in the pipe wall shall be equal to:
p1 <sup>¼</sup> p2 <sup>þ</sup> p3 <sup>¼</sup> <sup>4</sup>E<sup>t</sup>
<sup>D</sup><sup>2</sup> <sup>¼</sup> <sup>4</sup>Et
<sup>D</sup><sup>2</sup> <sup>d</sup> <sup>þ</sup> <sup>t</sup>δ<sup>a</sup>
. Pressure changes in this manner p sin ωt ¼ p2 þ p3, in this
<sup>∂</sup>z<sup>2</sup> : (9)
EIy} , (10)
dt: (11)
<sup>D</sup><sup>2</sup> d, (11a)
, the expression can
$$
\overline{\mathbf{p}}\_1 = \left[\frac{4E}{D^2} - \omega^2\right]td\tag{13}
$$
Combined stress in the axial direction due to pressure pulsations and vibration:
$$
\sigma\_t = \sigma\_{\Delta pt} + \sigma\_V \tag{14}
$$
#### 5. Spectral transforms
Spectral density corresponds to the spectral form of the internal pressure sinusoidal vibrations u(t) = sin ω0t. Periodic function spectrum ST(f) is determined by a direct Fourier transform of the time function
$$\begin{split} S\_T(f) &= \int\_{-\infty}^{\infty} s\_T(t) e^{-j2\pi \theta} dt = S(f) \cdot \frac{1}{T} \sum\_{m = -\infty}^{\infty} \delta \left( f - \frac{m}{T} \right) \\ &= \frac{1}{T} \cdot \sum\_{m = -\infty}^{\infty} S\left( \frac{m}{T} \right) \cdot \delta \left( f - \frac{m}{T} \right) = \sum\_{m = -\infty}^{\infty} \mathbb{C}[m] \cdot \delta \left( f - \frac{m}{T} \right) \end{split} \tag{15}$$
Complex weights of S-functions at frequencies multiple of 1/Т. They represent expansion coefficients of the function (15) into a Fourier series as follows:
$$\mathbf{C}[m] = \frac{1}{T} \cdot \mathbf{S}\left(\frac{m}{T}\right).$$
Time function sТ(t) and spectral function С[т] of the impact are the main characteristics. They are interrelated by a complex Fourier series.
The function containing n vibrations is described by the following equations,
$$F(t) = \sin\omega\_0 t \text{ at } 0 < t < \frac{2\pi n}{\omega\_0} \text{ and } F(t) = 0 \text{ at } 0 > t > \frac{2\pi n}{\omega\_0}.$$
Then amplitude spectrum is determined by the following expression:
$$S(\omega) = \left| \frac{2\omega\_0 \frac{n\pi\omega}{a\_0}}{\omega\_0^2 - \omega^2} \right| \tag{16}$$
At small values of <sup>ω</sup>0: <sup>S</sup>ð Þ <sup>ω</sup> <sup>≈</sup> <sup>1</sup> ω. Upon the completion of the vibration analysis according to the scheme of the single-degree-offreedom system (which includes the reduced weight of the pipeline and its components, and elastic support action), stresses and deformations in the support elements shall be calculated.
Transfer function of the maximum stress relative to acceleration of the pipeline supports can be written as
$$|H\_{\sigma}(\omega, z)| = \frac{\mathbb{C}\_{k}(\omega)\frac{\mathrm{ED}}{2}y\_{k}''(z)}{\sqrt{\left(\omega\_{0}^{2} - \omega^{2}\right)^{2} + \left(2\beta\alpha\omega\_{0}\right)^{2}}}.\tag{17}$$
duration of the time period shall be long enough to consider all potential delays. Pipe laying
<sup>σ</sup>�<sup>1</sup> <sup>¼</sup> <sup>1</sup>:75σB=N<sup>0</sup>:<sup>12</sup>
The requirements [design documentation] state that "the maximum allowable amplitude of
Offshore pipeline specifications do not provide either limitations for the pressure pulsations or
Low-frequency vibrations of the pipelines under principal modes, when such vibrations are close to be harmonic, can be easily evaluated on the basis of the amplitude of vibration displacement since in this case they are proportional to the stresses induced in the pipelines
We get the following expression for k-form of the vibrations using formula [5], for the root
ω2
In the event of random vibrations, the combined stress in the pipeline is the following:
<sup>σ</sup><sup>2</sup> ð Þ¼ <sup>z</sup> <sup>X</sup> N
k¼1 σ2 <sup>k</sup> ð Þz :
However, in the regulatory documents for offshore pipelines, there are not only restrictions on pressure pulsations but also restrictions on vibrations. Low-frequency oscillations of pipelines along lower forms, when these oscillations are close to harmonic, can be conveniently estimated from the amplitude values of the vibrational displacement because in this case they are proportional to the stresses arising in the pipelines and are indicators of the strength of the pipelines.
<sup>Φ</sup>QQð Þ <sup>ω</sup> <sup>d</sup><sup>ω</sup> ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Vibration Strength of Pipelines
29
http://dx.doi.org/10.5772/intechopen.72794
<sup>0</sup> � <sup>ω</sup><sup>2</sup> � �<sup>2</sup> <sup>þ</sup> <sup>2</sup>βωω<sup>0</sup> � �<sup>2</sup> <sup>q</sup>
vibrations of the process pipelines is 0.2 mm at vibration frequency of max 40 Hz" [2].
σ�<sup>1</sup> value can be defined either using reference data or Manson formula [4]:
Pipeline vibration limiting regulations can be divided into the following categories:
• for pipeline vibration resistance under exposure to external vibrations.
and can be regarded as a strength factor of the pipelines.
ffiffiffiffiffiffiffiffiffiffiffi σ2 <sup>k</sup> ð Þz q
¼
ð∞ 0 Ckð Þ ω
2 6 4
period shall not exceed this time interval.
here N is a number of loading cycles.
6. Pipeline vibration limiting
• for pipeline soundness and quality
vibration limitations.
mean square value of vibrations:
Spectral density of the pipeline response to random excitations will be equal to
$$\boldsymbol{\Phi}\_{\rm YY}(\omega) = \frac{\boldsymbol{\Phi}\_{\rm QQ}(\omega)}{\left(\omega\_0^2 - \omega^2\right)^2 + \left(2\beta\omega\omega\_0\right)^4} \tag{18}$$
Considering the tensile strength and fatigue strength, the allowable amplitudes of stresses in the pipeline wall can be calculated from the following formulas:
$$\sigma\_{\Lambda m} = \frac{\sigma\_{-1}\beta k}{n\_m \left(1 + \frac{\sigma\_{-1}}{\sigma\_\beta} \frac{1 + q\_r}{1 - q\_r}\right)}, \quad \sigma\_{\Lambda t} = \frac{\sigma\_{-1}\beta k}{n\_t \left(1 + \frac{\sigma\_{-1}}{\sigma\_\beta} \frac{1 + q\_t}{1 - q\_r}\right)}\tag{19}$$
where σ<sup>В</sup> is a tensile strength; σ�<sup>1</sup> is a fatigue strength under symmetrical loading cycle; β is a coefficient that takes into consideration the effect of the pipeline surface finish on the fatigue strength: for new pipelines β = 0.80–0.85, for corrosion susceptible pipelines, this coefficient is reduced to β = 0.5; k is a stress concentration factor [1, 3, 4].
Stress ratio is shown as follows:
$$q\_r = \frac{P\_p - \Delta P}{P\_p + \Delta P}, \quad q\_t = \frac{P\_p D / (2\delta) - \sigma\_t}{P\_p D / (2\delta) + \sigma\_t}.$$
During the installation of the pipeline system, the rated natural load shall be assumed as maximum\* at the most probable sea condition for the time period under review, which is determined using (Нз, Тр), and applicable stream and wind conditions. Rated load is assumed as maximum at the most probable parameters of the natural environment (in other words, waves, stream, and wind- LE), and equals to.
$$\mathcal{R}(L\_E) = 1 - \frac{1}{N}$$
where R(LЕ) is a probability distribution function LЕ.
N is a number of loading cycles of minimum 3 h in length at a certain sea condition.
Note that the specific sea condition for the time period under review can be interpreted as a sea condition for an applicable location and period of pipe laying. Ordinary requirement is that the duration of the time period shall be long enough to consider all potential delays. Pipe laying period shall not exceed this time interval.
σ�<sup>1</sup> value can be defined either using reference data or Manson formula [4]:
$$\sigma\_{-1} = 1.75 \sigma\_{\rm B} / N^{0.12}$$
here N is a number of loading cycles.
Upon the completion of the vibration analysis according to the scheme of the single-degree-offreedom system (which includes the reduced weight of the pipeline and its components, and elastic support action), stresses and deformations in the support elements shall be calculated. Transfer function of the maximum stress relative to acceleration of the pipeline supports can be
> <sup>2</sup> y<sup>00</sup> <sup>k</sup> ð Þ<sup>z</sup> ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
� �<sup>2</sup> <sup>q</sup> : (17)
� �<sup>4</sup> (18)
� � (19)
<sup>0</sup> � <sup>ω</sup><sup>2</sup> � �<sup>2</sup> <sup>þ</sup> <sup>2</sup>βωω<sup>0</sup>
ΦQQð Þ ω
<sup>0</sup> � <sup>ω</sup><sup>2</sup> � �<sup>2</sup> <sup>þ</sup> <sup>2</sup>βωω<sup>0</sup>
j j <sup>H</sup>σð Þ <sup>ω</sup>; <sup>z</sup> <sup>¼</sup> Ckð Þ <sup>ω</sup> ED
Spectral density of the pipeline response to random excitations will be equal to
ΦYYð Þ¼ ω
the pipeline wall can be calculated from the following formulas:
σΛ<sup>m</sup> <sup>¼</sup> <sup>σ</sup>�<sup>1</sup>β<sup>k</sup> nm <sup>1</sup> <sup>þ</sup> <sup>σ</sup>�<sup>1</sup> σB 1þqr 1�qr
reduced to β = 0.5; k is a stress concentration factor [1, 3, 4].
qr <sup>¼</sup> Pp � <sup>Δ</sup><sup>P</sup>
Stress ratio is shown as follows:
waves, stream, and wind- LE), and equals to.
where R(LЕ) is a probability distribution function LЕ.
ω2
ω<sup>2</sup>
Considering the tensile strength and fatigue strength, the allowable amplitudes of stresses in
where σ<sup>В</sup> is a tensile strength; σ�<sup>1</sup> is a fatigue strength under symmetrical loading cycle; β is a coefficient that takes into consideration the effect of the pipeline surface finish on the fatigue strength: for new pipelines β = 0.80–0.85, for corrosion susceptible pipelines, this coefficient is
Pp <sup>þ</sup> <sup>Δ</sup><sup>P</sup> , qt <sup>¼</sup> PpD=ð Þ� <sup>2</sup><sup>δ</sup> <sup>σ</sup><sup>t</sup>
During the installation of the pipeline system, the rated natural load shall be assumed as maximum\* at the most probable sea condition for the time period under review, which is determined using (Нз, Тр), and applicable stream and wind conditions. Rated load is assumed as maximum at the most probable parameters of the natural environment (in other words,
R Lð Þ¼ <sup>E</sup> <sup>1</sup> � <sup>1</sup>
Note that the specific sea condition for the time period under review can be interpreted as a sea condition for an applicable location and period of pipe laying. Ordinary requirement is that the
N is a number of loading cycles of minimum 3 h in length at a certain sea condition.
N
PpD=ð Þþ 2δ σ<sup>t</sup>
,
� � , <sup>σ</sup><sup>Δ</sup><sup>t</sup> <sup>¼</sup> <sup>σ</sup>�<sup>1</sup>β<sup>k</sup>
nt <sup>1</sup> <sup>þ</sup> <sup>σ</sup>�<sup>1</sup> σB 1þqt 1�qrt
written as
28 System of System Failures
| doab | 2025-04-07T04:13:03.842668 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 19
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.20 | 6. Pipeline vibration limiting
Pipeline vibration limiting regulations can be divided into the following categories:
The requirements [design documentation] state that "the maximum allowable amplitude of vibrations of the process pipelines is 0.2 mm at vibration frequency of max 40 Hz" [2].
Offshore pipeline specifications do not provide either limitations for the pressure pulsations or vibration limitations.
Low-frequency vibrations of the pipelines under principal modes, when such vibrations are close to be harmonic, can be easily evaluated on the basis of the amplitude of vibration displacement since in this case they are proportional to the stresses induced in the pipelines and can be regarded as a strength factor of the pipelines.
We get the following expression for k-form of the vibrations using formula [5], for the root mean square value of vibrations:
$$\sqrt{\sigma\_k^2(z)} = \left[ \int\_0^\infty \mathbb{C}\_k(\omega) \frac{\mathfrak{d}\_{\mathbb{Q}\mathbb{Q}}(\omega) d\omega}{\sqrt{\left(\omega\_0^2 - \omega^2\right)^2 + \left(2\beta\alpha\omega\_0\right)^2}} \right] \frac{ED}{2}$$
In the event of random vibrations, the combined stress in the pipeline is the following:
$$
\overline{\sigma^2}(z) = \sum\_{k=1}^N \overline{\sigma\_k^2}(z).
$$
However, in the regulatory documents for offshore pipelines, there are not only restrictions on pressure pulsations but also restrictions on vibrations. Low-frequency oscillations of pipelines along lower forms, when these oscillations are close to harmonic, can be conveniently estimated from the amplitude values of the vibrational displacement because in this case they are proportional to the stresses arising in the pipelines and are indicators of the strength of the pipelines.
Vibration velocity can be written as:
$$V\_l = \int\_0^T a(t)dt = \int\_0^T B\_0 \sin\omega t dt = B\_0 n / 2\pi f\_K \tag{20}$$
Development of the standardized vibration limits is complicated due to a wide variety of the
Vibration Strength of Pipelines
31
http://dx.doi.org/10.5772/intechopen.72794
However, for determination of the pipeline system reliability exposed to vibrations, it is required to consider the effect of vibrations on failures and malfunctions on the basis of certain assumptions regarding damage accumulation in the structures and failure occurrence. If stresses in the elastic elements of pipelines (supports) are considered failure criteria, the vibration strength of the supports shall be evaluated using calculation models of support structures. The private oilextracting companies apply burying of pipelines. It is just concealment of a problem. In the
During the analysis of the first fundamental forms of vibrations which lie within 20 Hz,
The following stages are included into the obligatory vibration tests of the pipeline: study of the operating conditions of the system and analysis of the dynamic loads acting on the pipeline; determination of the potential failure patterns; and selection of failure occurrence criteria due
According to the regulations of the Ministry of Gas Industry, emergency vibration level [2] is measured using vibration velocity Ve = 18 mm/s, and the warning vibration level is estimated
During the pipeline vibration analysis, it is necessary to know rigidity characteristics of the system components. The rigidity of pipe of permanent round section is characterized by the
Up to the present moment, frequency ranges have not been yet defined for the offshore pipelines where one or another vibration parameter shall be used for vibration limitation purposes.
Vmax ¼ vbasekη Vmax:
Amplitude of vibration stresses at different frequencies is determined during calculation of the
The main criterion of the pipeline vibration strength is detuning of natural frequencies fj from
Let us determine safe vibration velocity for the pipeline kinematically excited on the movable base plate. Amplitudes of vibration voltage at different frequencies are determined by the results of calculation of forced oscillations of the pipeline. The way to ensure the vibration
Vibrations of marine pipelines with a protective coating to decrease the interaction of the pipeline with the coating result in cracking in the coating of corrosion and damage to the coating.
discrete frequencies of the excitatory loads fip, defined as described in 2.2 [6].
resistance of the pipeline is to detune the natural frequencies fj of the structure [7].
.
requirements to the characteristics of the vibration capacity of various equipments.
seismic phenomena, vibrations arise in the thickness of the Earth.
vibration displacement is frequently used as a test parameter.
to vibrations.
by the exceedance Ve = 41 mm/s [9, 10].
forced vibrations of the pipeline [8].
following parameters: EI is bending rigidity, N�m2
The allowable amplitude of vibrations is defined as follows:
where V<sup>l</sup> is vibration velocity (mm/s); a is vibration acceleration (mm/s<sup>2</sup> ) with amplitude В0; fk is frequency of multiplicity factor k.
During the evaluation of the vibration strength, the maximum amplitude of equivalent vibration stresses shall be determined for each representative section of the pipeline. This amplitude is obtained as a result of various modal superpositions. Dimension of the transfer function depends on the type of disturbance and response against which the transfer function is determined.
When exposed to random vibrations, root-mean-square value of the maximum stress in the pipeline can be found using the transfer function for the maximum stress with respect to acceleration of the pipeline supports:
$$|H\_{\sigma}(\omega, z)| = \frac{\mathsf{C}\_{k}(\omega)\frac{\mathsf{ED}}{2}y\_{k}''(z)}{\sqrt{\left(\omega\_{0}^{2} - \omega^{2}\right)^{2} + \left(2\beta\omega\omega\_{0}\right)^{2}}}\tag{21}$$
where D is outside diameter of pipe.
Pipeline response to broadband random vibrations can be defined as a combined effect of several narrowband random vibrations. The narrowband vibrations of the pipelines occur as a response to the broadband excitation under low damping. The mean frequency of the narrowband vibrations can be calculated from the Rice's formula:
$$
\omega\_0^2 = \frac{\int\_{-\infty}^{\infty} \omega^2 \Phi\_{YY}(\omega) d\omega}{\int\_{-\infty}^{\infty} \Phi\_{YY}(\omega) d\omega} = \frac{R\_y^{\cdot}(0)}{R\_y(0)} = \frac{\sigma\_{\dot{y}}^2}{\sigma\_y^2} \tag{22}
$$
Root mean square value of the pipeline movement to be subject to vibrations can be calculated from the following formula:
$$
\sigma = \left[ \int\_{f\_1}^{f\_2} \eta\_f^2 \Phi(f) df \right]^{1/2}, \quad \text{here } \sigma = \sqrt{\sum\_{i=1}^N \eta\_{f\_i}^2 \Phi(f\_i) \Delta f\_i}. \tag{23}
$$
where η<sup>f</sup> is a dynamic response factor-relation between displacement amplitude of the anchor points for the pipeline and relative displacement amplitude of the pipeline sections at specified frequency; Ф(f) is a spectral density of the disturbed random vibration in the frequency band f<sup>1</sup> and f2; Δfi is an interval of the frequency band segmentation f1, f2; N is a number of intervals of the frequency band segmentation.
Development of the standardized vibration limits is complicated due to a wide variety of the requirements to the characteristics of the vibration capacity of various equipments.
However, for determination of the pipeline system reliability exposed to vibrations, it is required to consider the effect of vibrations on failures and malfunctions on the basis of certain assumptions regarding damage accumulation in the structures and failure occurrence. If stresses in the elastic elements of pipelines (supports) are considered failure criteria, the vibration strength of the supports shall be evaluated using calculation models of support structures. The private oilextracting companies apply burying of pipelines. It is just concealment of a problem. In the seismic phenomena, vibrations arise in the thickness of the Earth.
During the analysis of the first fundamental forms of vibrations which lie within 20 Hz, vibration displacement is frequently used as a test parameter.
The following stages are included into the obligatory vibration tests of the pipeline: study of the operating conditions of the system and analysis of the dynamic loads acting on the pipeline; determination of the potential failure patterns; and selection of failure occurrence criteria due to vibrations.
According to the regulations of the Ministry of Gas Industry, emergency vibration level [2] is measured using vibration velocity Ve = 18 mm/s, and the warning vibration level is estimated by the exceedance Ve = 41 mm/s [9, 10].
During the pipeline vibration analysis, it is necessary to know rigidity characteristics of the system components. The rigidity of pipe of permanent round section is characterized by the following parameters: EI is bending rigidity, N�m2 .
Up to the present moment, frequency ranges have not been yet defined for the offshore pipelines where one or another vibration parameter shall be used for vibration limitation purposes.
The allowable amplitude of vibrations is defined as follows:
Vibration velocity can be written as:
is frequency of multiplicity factor k.
acceleration of the pipeline supports:
where D is outside diameter of pipe.
from the following formula:
σ ¼
of the frequency band segmentation.
ð<sup>f</sup> <sup>2</sup> f 1 η2 <sup>f</sup> Φð Þf df " #<sup>1</sup>=<sup>2</sup>
band vibrations can be calculated from the Rice's formula:
ω2 <sup>0</sup> ¼ Ð ∞
determined.
30 System of System Failures
Vl ¼ ðT 0
a tð Þdt ¼
where V<sup>l</sup> is vibration velocity (mm/s); a is vibration acceleration (mm/s<sup>2</sup>
ðT 0
During the evaluation of the vibration strength, the maximum amplitude of equivalent vibration stresses shall be determined for each representative section of the pipeline. This amplitude is obtained as a result of various modal superpositions. Dimension of the transfer function depends on the type of disturbance and response against which the transfer function is
When exposed to random vibrations, root-mean-square value of the maximum stress in the pipeline can be found using the transfer function for the maximum stress with respect to
> <sup>2</sup> y<sup>00</sup> <sup>k</sup> ð Þ<sup>z</sup> ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
> > <sup>y</sup>ð Þ0 Ryð Þ<sup>0</sup> <sup>¼</sup>
> > > X N
i¼1 η2 f i Φ f <sup>i</sup> � �Δ<sup>f</sup> <sup>i</sup> : vuut (23)
σ2 y\_ σ2 y
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
<sup>0</sup> � <sup>ω</sup><sup>2</sup> � �<sup>2</sup> <sup>þ</sup> <sup>2</sup>βωω<sup>0</sup>
j j <sup>H</sup>σð Þ <sup>ω</sup>; <sup>z</sup> <sup>¼</sup> Ckð Þ <sup>ω</sup> ED
ω2
Pipeline response to broadband random vibrations can be defined as a combined effect of several narrowband random vibrations. The narrowband vibrations of the pipelines occur as a response to the broadband excitation under low damping. The mean frequency of the narrow-
�<sup>∞</sup> <sup>ω</sup><sup>2</sup>ΦYYð Þ <sup>ω</sup> <sup>d</sup><sup>ω</sup>
�<sup>∞</sup> <sup>Φ</sup>YYð Þ <sup>ω</sup> <sup>d</sup><sup>ω</sup> <sup>¼</sup> <sup>R</sup>}
Root mean square value of the pipeline movement to be subject to vibrations can be calculated
, here σ ¼
where η<sup>f</sup> is a dynamic response factor-relation between displacement amplitude of the anchor points for the pipeline and relative displacement amplitude of the pipeline sections at specified frequency; Ф(f) is a spectral density of the disturbed random vibration in the frequency band f<sup>1</sup> and f2; Δfi is an interval of the frequency band segmentation f1, f2; N is a number of intervals
Ð ∞
B<sup>0</sup> sin ωtdt ¼ B0n=2πf <sup>K</sup> (20)
� �<sup>2</sup> <sup>q</sup> (21)
) with amplitude В0; fk
(22)
$$\mathbf{V}\_{\text{max}} = \upsilon\_{\text{base}} k \eta \text{ V}\_{\text{max}}$$
Amplitude of vibration stresses at different frequencies is determined during calculation of the forced vibrations of the pipeline [8].
The main criterion of the pipeline vibration strength is detuning of natural frequencies fj from discrete frequencies of the excitatory loads fip, defined as described in 2.2 [6].
Let us determine safe vibration velocity for the pipeline kinematically excited on the movable base plate. Amplitudes of vibration voltage at different frequencies are determined by the results of calculation of forced oscillations of the pipeline. The way to ensure the vibration resistance of the pipeline is to detune the natural frequencies fj of the structure [7].
Vibrations of marine pipelines with a protective coating to decrease the interaction of the pipeline with the coating result in cracking in the coating of corrosion and damage to the coating.
#### Example.
#### Vibration strength of pipelines.
Low-frequency vibrations of the pipelines under principal modes, when such vibrations are close to be harmonic, can be easily evaluated on the basis of the amplitude of vibration displacement because in this case they are proportional to the stresses induced in the pipelines and can be regarded as a strength factor of the pipelines.
σ�<sup>1</sup> value can be defined either using reference data or Manson formula [4]:
<sup>σ</sup>�<sup>1</sup> <sup>¼</sup> <sup>1</sup>, <sup>75</sup>σB=N<sup>0</sup>:12, here <sup>N</sup> is a number of loading cycles.
Let us consider the following example of calculating the allowable stress amplitudes in the pipeline wall.
Researches were made for vibration speed (response characteristic: root mean square value of Vmax = 0.0103 cm/s) and pressure pulsation amplitude of ΔP = 0.5 MPa for the landfall section of the offshore pipeline with rated pressure of 17.5 MPa. Outside diameter of the pipeline is D = 406 mm, wall thickness is 17.5 mm. Material grade is X52 (σ<sup>в</sup> = 455 MPa, σ�<sup>1</sup> ≈ 916 MPa, E = 0.20457�106MPa, <sup>σ</sup><sup>e</sup> = 358 MPa). Pipe section modulus is W = 0.00199 m3 .
strength of the pipeline steel (see Table 1), stress amplitude σ<sup>Δ</sup><sup>t</sup> ¼ 38.322 MPa in the pipeline
Let us review the seismogram of seismic intensity 6 to MSK-64 scale for the vibration strength analysis of the landfall section of the offshore pipeline. The seismic impact is characterized by the following parameters: maximum acceleration amax = 0.94485 cm/s<sup>2</sup> at t = 0.12 s (Figure 2).
> <sup>¼</sup> <sup>2</sup><sup>π</sup> N∙Δt
Periodogram results can be interpreted as dispersed data at frequencies given in Table 2 and
Ser.no. Frequency Period Periodogram Density Spectral density in hamming window
ð Þ rad=s
Vibration Strength of Pipelines
33
http://dx.doi.org/10.5772/intechopen.72794
wall. Fatigue strength factor of the pipeline shall be at least п = 2.0 [3].
Figure 2. Accelerogram recording used during calculations.
We get the line spectrum of the signal consisting of the harmonics:
0.00000 0.0 1488.55 0.035714 0.02174 46.0 2836.21 1828.18 0.241071 0.04348 23.0 1695.297 1839.68 0. 446429 0.06522 15.3333 1456.8 1528.67 0.241071 0.08696 11.5 1341.92 1355.41 0.035714
0.1087 9.2 1255.43 1258.18 0.13044 7.66667 1174.63 1174.32 0.15217 6.67143 1092.56 1091.35 0.17391 5.75 1007.15 1005.84 0.19585 5.11111 918.245 917.192 0.21739 4.6 826.556 825. 941
where Тmin is a minimum period.
Figure 3.
<sup>Δ</sup><sup>ω</sup> <sup>¼</sup> <sup>2</sup><sup>π</sup> Tmin
Kt≔1
σBB�β�k=806.08
$$\text{norm1} \cdot \left(1 + \frac{\sigma \text{BB}}{\sigma \text{B}} \cdot \frac{1 + \text{rq}}{1 - \text{rq}}\right) = 64.214$$
If length l of the pipeline section is 30 m, displacement y0 is determined by the formula (7) and is equal to 0.026 m, considering the formula (9) vmax = 0.035 m
If pressure pulsation ΔР is 0.5 MPa and work pressure is 15.7 MPa (β = 0.8), stress variation on the pipeline wall is 6.386 MPa. Stress ratio (17) qr is 0.938. Stress amplitude in the pipeline wall is equal to σ<sup>Δ</sup><sup>q</sup> ¼ 12.553 MPa considering steel tensile strength and fatigue strength (refer to Table 1).
Temperature fluctuations Δt = 5� result in stress variation of 12.36 MPa in the pipeline wall and are equal to rt = 0.817 as shown in the formula (19). Allowing for tensile strength and fatigue
Table 1. The main physical characteristics of steel pipe [2].
Figure 2. Accelerogram recording used during calculations.
Example.
32 System of System Failures
pipeline wall.
σBB�β�k=806.08
Table 1).
Vibration strength of pipelines.
and can be regarded as a strength factor of the pipelines.
<sup>σ</sup>�<sup>1</sup> <sup>¼</sup> <sup>1</sup>, <sup>75</sup>σB=N<sup>0</sup>:12, here <sup>N</sup> is a number of loading cycles.
σ�<sup>1</sup> value can be defined either using reference data or Manson formula [4]:
0.20457�106MPa, <sup>σ</sup><sup>e</sup> = 358 MPa). Pipe section modulus is W = 0.00199 m3
nm1 � 1 þ
is equal to 0.026 m, considering the formula (9) vmax = 0.035 m
Name of the characteristics Steel
Table 1. The main physical characteristics of steel pipe [2].
Low-frequency vibrations of the pipelines under principal modes, when such vibrations are close to be harmonic, can be easily evaluated on the basis of the amplitude of vibration displacement because in this case they are proportional to the stresses induced in the pipelines
Let us consider the following example of calculating the allowable stress amplitudes in the
Researches were made for vibration speed (response characteristic: root mean square value of Vmax = 0.0103 cm/s) and pressure pulsation amplitude of ΔP = 0.5 MPa for the landfall section of the offshore pipeline with rated pressure of 17.5 MPa. Outside diameter of the pipeline is D = 406 mm, wall thickness is 17.5 mm. Material grade is X52 (σ<sup>в</sup> = 455 MPa, σ�<sup>1</sup> ≈ 916 MPa, E =
Kt≔1
1 þ rq 1 � rq
¼ 64:214
σBB σB �
If length l of the pipeline section is 30 m, displacement y0 is determined by the formula (7) and
If pressure pulsation ΔР is 0.5 MPa and work pressure is 15.7 MPa (β = 0.8), stress variation on the pipeline wall is 6.386 MPa. Stress ratio (17) qr is 0.938. Stress amplitude in the pipeline wall is equal to σ<sup>Δ</sup><sup>q</sup> ¼ 12.553 MPa considering steel tensile strength and fatigue strength (refer to
Temperature fluctuations Δt = 5� result in stress variation of 12.36 MPa in the pipeline wall and are equal to rt = 0.817 as shown in the formula (19). Allowing for tensile strength and fatigue
Yield strength, МPа 358 448 Strength limit Rm, МPа 455 530 Permissible stresses in metal pipes σ, МPа 255.6 292.8
.
Х52∅406 Х65∅711
strength of the pipeline steel (see Table 1), stress amplitude σ<sup>Δ</sup><sup>t</sup> ¼ 38.322 MPa in the pipeline wall. Fatigue strength factor of the pipeline shall be at least п = 2.0 [3].
Let us review the seismogram of seismic intensity 6 to MSK-64 scale for the vibration strength analysis of the landfall section of the offshore pipeline. The seismic impact is characterized by the following parameters: maximum acceleration amax = 0.94485 cm/s<sup>2</sup> at t = 0.12 s (Figure 2).
We get the line spectrum of the signal consisting of the harmonics:
$$
\Delta w = \frac{2\pi}{T\_{min}} = \frac{2\pi}{N \cdot \Delta t} \left(\text{rad/s}\right),
$$
where Тmin is a minimum period.
Periodogram results can be interpreted as dispersed data at frequencies given in Table 2 and Figure 3.
Mean mathematic amplitude Вк over a period of observation Т shall be written as:
<sup>B</sup><sup>0</sup> <sup>¼</sup> <sup>1</sup> T ðT 0
where BKð Þt is a behavior of k-х components of vibration spectrum.
a tð Þdt ¼
where V<sup>l</sup> is vibration velocity (mm/s); a is vibration acceleration (mm/s<sup>2</sup>
ðT 0
According to the standards of the Ministry of Gas Industry [6], an emergency vibration level is equal to the vibration speed Ve = 18 mm/s, and an alarm vibration level exceeds Ve = 41 mm/s. For pipeline sections more than 0.5 m, the vibration displacement span is restricted to 0.5 mm (refer to the standards of the National Compressor Engineering Association "Souzcompre-
Based on the allowable strength from [2], the allowable amplitude of vibrations is calculated for the pipeline of 408 mm diameter and 17.5 mm wall thickness, and allowable stress is 255.6 MPa. The effective span length of the pipeline section under review is 30 m. Amplitude of forced vibrations Y0 is 0.061 m. Safe vibration speed of the pipeline, which is excited kine-
Offshore pipelines can be regarded as new constructions in Russia. Industry-Specific Construc-
None of the private oil companies is interested in studying vibration strength of the pipelines; they make references to the existing standards. Offshore pipeline vibration analysis is very important and requires government involvement. Such researches should be performed by the
Private companies use a pipe burial method. The pipelines can be buried; however, it is not a
Vibrations of marine pipelines with a protective coating to decrease the interaction of the
way to solve the problem. Seismic impacts come from under the ground.
Vl ¼ ðT 0
Vibration velocity can be written as:
is frequency of multiplicity factor k.
matically on the moving platform, is 0.035 m/s.
tion Standard specifies dynamic vibrations of the pipeline.
pipeline with the coating result in cracking in the coating.
ssmach") [2].
7. Conclusion
large scientific institutes.
BKð Þt dt, (24)
B0sinωtdt ¼ B0n=2πf <sup>K</sup> (25)
) with amplitude В0; fk
Vibration Strength of Pipelines
35
http://dx.doi.org/10.5772/intechopen.72794
Table 2. Vibration impact periodogram.
Figure 3. Vibration impact periodogram.
Mean mathematic amplitude Вк over a period of observation Т shall be written as:
$$B\_0 = \frac{1}{T} \int\_0^T B\_K(t)dt,\tag{24}$$
where BKð Þt is a behavior of k-х components of vibration spectrum.
Vibration velocity can be written as:
Ser.no. Frequency Period Periodogram Density Spectral density in hamming window
0.23913 4.18182 733. 242 733.16 0.26087 3.83333 639.703 640.204 0.28261 3.53846 547.464 548.566 0.30435 3. 28571 458.096 459.796 0.32609 3.06667 373.16 375.439 0.34783 2.875 294.168 296.99 0.36957 2.70588 222.536 225.855 0.3913 2.55556 159.563 163.32 0.41304 2.42105 106.391 110.52 0.43478 2.3 63.992 68.417 0.45652 2.19048 33.139 37.781 0.47826 2.09091 14.398 19.172
Table 2. Vibration impact periodogram.
34 System of System Failures
Figure 3. Vibration impact periodogram.
$$V\_{l} = \int\_{0}^{T} a(t)dt = \int\_{0}^{T} B\_{0} \sin \omega t dt = B\_{0} n / 2\pi f\_{K} \tag{25}$$
where V<sup>l</sup> is vibration velocity (mm/s); a is vibration acceleration (mm/s<sup>2</sup> ) with amplitude В0; fk is frequency of multiplicity factor k.
According to the standards of the Ministry of Gas Industry [6], an emergency vibration level is equal to the vibration speed Ve = 18 mm/s, and an alarm vibration level exceeds Ve = 41 mm/s. For pipeline sections more than 0.5 m, the vibration displacement span is restricted to 0.5 mm (refer to the standards of the National Compressor Engineering Association "Souzcompressmach") [2].
Based on the allowable strength from [2], the allowable amplitude of vibrations is calculated for the pipeline of 408 mm diameter and 17.5 mm wall thickness, and allowable stress is 255.6 MPa. The effective span length of the pipeline section under review is 30 m. Amplitude of forced vibrations Y0 is 0.061 m. Safe vibration speed of the pipeline, which is excited kinematically on the moving platform, is 0.035 m/s.
### 7. Conclusion
Offshore pipelines can be regarded as new constructions in Russia. Industry-Specific Construction Standard specifies dynamic vibrations of the pipeline.
None of the private oil companies is interested in studying vibration strength of the pipelines; they make references to the existing standards. Offshore pipeline vibration analysis is very important and requires government involvement. Such researches should be performed by the large scientific institutes.
Private companies use a pipe burial method. The pipelines can be buried; however, it is not a way to solve the problem. Seismic impacts come from under the ground.
Vibrations of marine pipelines with a protective coating to decrease the interaction of the pipeline with the coating result in cracking in the coating.
| doab | 2025-04-07T04:13:03.843836 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 20
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.21 | Author details
Muravieva Liudmila Victorovna
Address all correspondence to: rfludmia@yandex.ru
Saint Petersburg State Polytechnic University, St. Petersburg, Russian Federation
| doab | 2025-04-07T04:13:03.844973 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 21
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.23 | 2. Safety hazard due to sudden and developing damage in transport systems
Despite many efforts in the processes of designing and implementation of new aircraft constructions, the occurrence of not signalled (sudden) damage, the effects of which are serious, took place. They may occur in the process of operation for many reasons, and mostly due to the lack of complete recognition of many processes that take place during the aviation equipment operation. So far, there has been also a lack of methodology for optimal shaping of construction elements based on destruction models, which would take into account all possible types of their loads and assessment of the environmental impact effects in the long-term operation.
where R(t)—probability of failure to the catastrophic damage occurrence.
E[K]min
Figure 1. Diagram for determination of the acceptable range of the risk of losses.
Obszar dopuszczalnego ryzyka
Area of the acceptable risk
E[K]—expected value of the total cost E[K] = E[O] + E[S].
3.1. Classification of construction systems and wear processes
result of the occurrence of catastrophic damage in the operation process.
able' risk policy based on the principle 'as low the risk, as it is reasonably achievable'.
3. Types of damage causing threats and models for assessing the
order to eliminate catastrophic damage).
during operation resulting in catastrophic damage.
probability of their occurrence
be presented.
Koszty
Costs
E[S]
E[K]
E[O]
E[O]—expected value of the cost of developing the construction and reliability test (in
Poziom bezpieczeństwa
Safety level
Probabilistic Methods for Damage Assessment in Aviation Technology
http://dx.doi.org/10.5772/intechopen.72317
39
Reliability
1,0 R(t)
E[S]—expected value of the risk resulting from the frequency of the failure and losses as a
Figure 1 shows the possibility of moving from the zero-risk policy of a threat to the 'accept-
The basis for estimating the risk of threat is a forecast of the occurrence of negative phenomena
In this chapter, the selected models, which can be used for estimating the reliability and risk of the occurrence of catastrophic damage in the accepted time period of the device operation, will
By assumption, the aviation technology has high reliability requirements, which, in practice, are implemented through special inspection procedures and appropriate design solutions involving the introduction of excesses of structure, strength, power, information, etc. The structural excess is characterised by elements or functional systems, basic and reserveprotective ones. After damage to the basic system, the protection systems start functioning. It ensures a high safety level of aircraft flights, which is one of the most important issues in the
For example, it is possible to distinguish some causes of sudden (catastrophic) damage. They are as follows [1]:
All the causes leading to catastrophic damage are the subject of numerous experimental and theoretical considerations.
The occurrence of catastrophic damage to the aircraft equipment usually results in serious failures or catastrophes and large losses. Therefore, it is necessary to estimate the risk of incurring these losses.
The term of risk is understood as the probability of the occurrence of critical damage or an adverse event in case of the occurrence of losses. In our case, a negative phenomenon will include the occurrence of catastrophic damage to aircraft. The level of the accepted risk of negative events is determined by the frequency of their occurrence and costs. In Figure 1, a method for determination of the accepted risk range is presented.
Figure 1. Diagram for determination of the acceptable range of the risk of losses.
The second way is to use an external central computerised system of monitoring and analysing the operation process of the entire set of objects, including a system of recording operational events and an information processing system. The system of recording information obtained during diagnostic checks and detected damage divided by types, symptoms, causes and effects is of particular importance. The recorded data are used for determination of the reliability and safety characteristics, and provide the opportunity to estimate a functioning resource.
2. Safety hazard due to sudden and developing damage in transport
loads and assessment of the environmental impact effects in the long-term operation.
fatigue crack, which exceeded the critical value.
foreign matters between the cooperating elements.
circumstances conductive to uncontrolled explosions.
method for determination of the accepted risk range is presented.
• The random shortcuts in the electronic circuits.
Despite many efforts in the processes of designing and implementation of new aircraft constructions, the occurrence of not signalled (sudden) damage, the effects of which are serious, took place. They may occur in the process of operation for many reasons, and mostly due to the lack of complete recognition of many processes that take place during the aviation equipment operation. So far, there has been also a lack of methodology for optimal shaping of construction elements based on destruction models, which would take into account all possible types of their
For example, it is possible to distinguish some causes of sudden (catastrophic) damage. They
• The loss of the volume strength of the element, which can be damaged as a result of the occurrence of excessive permanent deformations, the occurrence of an ad hoc crack, or a
• The loss in material properties in the construction elements as a result of functioning of
• The loss of the element's usability as a result of the surface wear or the inclusions of
• The random increase in the concentration of vapours of chemicals and the occurrence of
All the causes leading to catastrophic damage are the subject of numerous experimental and
The occurrence of catastrophic damage to the aircraft equipment usually results in serious failures or catastrophes and large losses. Therefore, it is necessary to estimate the risk of
The term of risk is understood as the probability of the occurrence of critical damage or an adverse event in case of the occurrence of losses. In our case, a negative phenomenon will include the occurrence of catastrophic damage to aircraft. The level of the accepted risk of negative events is determined by the frequency of their occurrence and costs. In Figure 1, a
systems
38 System of System Failures
are as follows [1]:
ageing processes.
theoretical considerations.
incurring these losses.
where R(t)—probability of failure to the catastrophic damage occurrence.
E[O]—expected value of the cost of developing the construction and reliability test (in order to eliminate catastrophic damage).
E[S]—expected value of the risk resulting from the frequency of the failure and losses as a result of the occurrence of catastrophic damage in the operation process.
E[K]—expected value of the total cost E[K] = E[O] + E[S].
Figure 1 shows the possibility of moving from the zero-risk policy of a threat to the 'acceptable' risk policy based on the principle 'as low the risk, as it is reasonably achievable'.
The basis for estimating the risk of threat is a forecast of the occurrence of negative phenomena during operation resulting in catastrophic damage.
In this chapter, the selected models, which can be used for estimating the reliability and risk of the occurrence of catastrophic damage in the accepted time period of the device operation, will be presented.
| doab | 2025-04-07T04:13:03.845013 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 23
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.24 | 3. Types of damage causing threats and models for assessing the probability of their occurrence
#### 3.1. Classification of construction systems and wear processes
By assumption, the aviation technology has high reliability requirements, which, in practice, are implemented through special inspection procedures and appropriate design solutions involving the introduction of excesses of structure, strength, power, information, etc. The structural excess is characterised by elements or functional systems, basic and reserveprotective ones. After damage to the basic system, the protection systems start functioning. It ensures a high safety level of aircraft flights, which is one of the most important issues in the air transport. Despite these protections and great efforts of technical services, the failures that cause accidents occur.
increase in the number of experiments. Based on operational tests, it can be concluded that the normal distribution provides an approximate (asymptotic) description of the random variable of T time of proper operation of the device's element to damage, and it can be used when the element's wear and ageing parameters create a continuous random process to achieve the limit state.
The random variable of T life time of tested objects has normal distribution, if its probability
The shape of f(t) density function curve of normal distribution characterises the population of objects in terms of homogeneity. The homogeneity of the population of the same elements of devices in terms of their durability in operation is represented by the coefficient of variation
For v small values, it is possible to accurately predict the moment of time for achieving the
þ ð∞
t
In order to simplify the calculations in practice, the so-called standardised variable is adopted:
<sup>u</sup> <sup>¼</sup> <sup>t</sup> � <sup>m</sup>
it indicates a number of average (standard) deviation in terms of which the random variable Tt being the implementation of life time of the particular i element differs from its expected value
f tð Þdt <sup>¼</sup> <sup>1</sup>
σ ffiffiffiffiffiffi <sup>2</sup><sup>π</sup> <sup>p</sup> þ ð∞
t
<sup>2</sup><sup>π</sup> <sup>p</sup> � exp � ð Þ <sup>t</sup> � <sup>m</sup>
2
Probabilistic Methods for Damage Assessment in Aviation Technology
http://dx.doi.org/10.5772/intechopen.72317
ð Þ T —variance of the random variable are the
exp � ð Þ <sup>t</sup> � <sup>m</sup>
<sup>σ</sup> (3)
� � (4)
� �du (5)
2
dt (2)
2σ<sup>2</sup> " # (1)
41
2σ<sup>2</sup> " #
f tðÞ¼ <sup>1</sup>
σ ffiffiffiffiffiffi
density is given by the following formula:
where <sup>m</sup> <sup>=</sup> <sup>E</sup>(T)—expected value and <sup>σ</sup><sup>2</sup> <sup>¼</sup> <sup>D</sup><sup>2</sup>
limit state in the operating time interval (0, t),
R tðÞ¼ 1 � F tðÞ¼ 1 �
The reliability function value is calculated as follows:
ðt
f tð Þdt ¼
With t=m+ σu, taking into account that dt = σdu, the above formula is as follows:
f tðÞ¼ <sup>1</sup>
R tðÞ¼ <sup>1</sup>
ffiffiffiffiffiffi <sup>2</sup><sup>π</sup> <sup>p</sup>
However, the formula for the reliability function is as follows:
ffiffiffiffiffiffi
þ ð∞
mþσ�u
<sup>2</sup><sup>π</sup> <sup>p</sup> � exp � <sup>u</sup><sup>2</sup>
2
exp � <sup>u</sup><sup>2</sup> 2
�∞
distribution parameters.
v = σ/m (Figure 2).
m = E(T).
Due to the fact that the integral
The protecting systems constituting the reserve of basic systems significantly increase the production costs and reduce the overall performance, such as capacity, range, fuel consumption, etc. They also require special treatment in the operation of aircraft, so that they have very high probability of correct functioning at the very low probability of use.
The accuracy of continuous or periodic identification of the state of usability is an important issue. The person stating the state of usability of basic and reserve technical systems can make two types of errors:
The result of the erroneous statement on the system activating the emergency release of the landing gear was the emergency landing of PLL LOT plane, Boeing 767-300ER, on November 1, 2011 at Warsaw Chopin Airport, which will be discussed in the further part of the chapter.
The wear and ageing processes of various elements are correlated with time or the functioning duration, or with calendar time in a varying degree. Generally, the construction elements and functional systems may be classified into three types:
#### 3.2. Elements strongly correlated with functioning time
In case of elements of the first group, it is possible to create the technical condition trajectory and to expect a moment of time, in which the limit state will occur. It is also possible to predict a moment of the element or unit secure taking out of service. In this case, a process of damage can be described with a suitably selected model for normal distribution, even with a small variance [2]. The suitable quantile of the random variable of functioning duration between damage can be a basis for developing a programme of diagnosing, maintenance and repairs. This group of elements can include slide bearings, gears, tyre treads of gear wheels, etc. A good model describing the time of the correct operation is normal distribution.
#### 3.2.1. Normal distribution
The normal distribution sometimes constitutes limit distribution, to which many other types of distribution asymptotically approach in the operational processes of devices, together with an
increase in the number of experiments. Based on operational tests, it can be concluded that the normal distribution provides an approximate (asymptotic) description of the random variable of T time of proper operation of the device's element to damage, and it can be used when the element's wear and ageing parameters create a continuous random process to achieve the limit state.
The random variable of T life time of tested objects has normal distribution, if its probability density is given by the following formula:
$$f(t) = \frac{1}{\sigma\sqrt{2\pi}} \cdot \exp\left[-\frac{(t-m)^2}{2\sigma^2}\right] \tag{1}$$
where <sup>m</sup> <sup>=</sup> <sup>E</sup>(T)—expected value and <sup>σ</sup><sup>2</sup> <sup>¼</sup> <sup>D</sup><sup>2</sup> ð Þ T —variance of the random variable are the distribution parameters.
The shape of f(t) density function curve of normal distribution characterises the population of objects in terms of homogeneity. The homogeneity of the population of the same elements of devices in terms of their durability in operation is represented by the coefficient of variation v = σ/m (Figure 2).
For v small values, it is possible to accurately predict the moment of time for achieving the limit state in the operating time interval (0, t),
The reliability function value is calculated as follows:
air transport. Despite these protections and great efforts of technical services, the failures that
The protecting systems constituting the reserve of basic systems significantly increase the production costs and reduce the overall performance, such as capacity, range, fuel consumption, etc. They also require special treatment in the operation of aircraft, so that they have very
The accuracy of continuous or periodic identification of the state of usability is an important issue. The person stating the state of usability of basic and reserve technical systems can make
The result of the erroneous statement on the system activating the emergency release of the landing gear was the emergency landing of PLL LOT plane, Boeing 767-300ER, on November 1, 2011 at Warsaw Chopin Airport, which will be discussed in the further part of the chapter. The wear and ageing processes of various elements are correlated with time or the functioning duration, or with calendar time in a varying degree. Generally, the construction elements and
• Elements having strongly correlated parameters determining the state of usability with the functioning duration or time, which can be identified with the existence of the mem-
• Elements having poorly correlated parameters of the state of usability with the functioning duration or time, which imply weak relationships of operating time with the technical
• Elements without correlation with the operating time, number of activation, or other
In case of elements of the first group, it is possible to create the technical condition trajectory and to expect a moment of time, in which the limit state will occur. It is also possible to predict a moment of the element or unit secure taking out of service. In this case, a process of damage can be described with a suitably selected model for normal distribution, even with a small variance [2]. The suitable quantile of the random variable of functioning duration between damage can be a basis for developing a programme of diagnosing, maintenance and repairs. This group of elements can include slide bearings, gears, tyre treads of gear wheels, etc. A
The normal distribution sometimes constitutes limit distribution, to which many other types of distribution asymptotically approach in the operational processes of devices, together with an
measure of the functioning duration, with randomly occurring damage.
good model describing the time of the correct operation is normal distribution.
high probability of correct functioning at the very low probability of use.
• an error of the first type consists of qualifying the usable device as unfit;
functional systems may be classified into three types:
condition change, wear and damage.
3.2. Elements strongly correlated with functioning time
ory related to the past.
3.2.1. Normal distribution
• an error of the second type consists of qualifying the unfit device as usable;
cause accidents occur.
40 System of System Failures
two types of errors:
$$R(t) = 1 - F(t) = 1 - \int\_{-\infty}^{t} f(t)dt = \int\_{t}^{+\infty} f(t)dt = \frac{1}{\sigma\sqrt{2\pi}} \int\_{t}^{+\infty} \exp\left[-\frac{(t-m)^2}{2\sigma^2}\right]dt\tag{2}$$
In order to simplify the calculations in practice, the so-called standardised variable is adopted:
$$
\mu = \frac{t - m}{\sigma} \tag{3}
$$
it indicates a number of average (standard) deviation in terms of which the random variable Tt being the implementation of life time of the particular i element differs from its expected value m = E(T).
With t=m+ σu, taking into account that dt = σdu, the above formula is as follows:
$$f(t) = \frac{1}{\sqrt{2\pi}} \cdot \exp\left(-\frac{\mu^2}{2}\right) \tag{4}$$
However, the formula for the reliability function is as follows:
$$R(t) = \frac{1}{\sqrt{2\pi}} \int\_{u + \sigma \cdot u}^{+\infty} \exp\left(-\frac{u^2}{2}\right) du\tag{5}$$
Due to the fact that the integral
Figure 2. The probability density of normal distribution for different values of v coefficient of variation.
$$\frac{1}{\sqrt{2\pi}} \cdot \int\_{-\infty}^{0} \exp\left(-\frac{u^2}{2}\right) du = 0.5\tag{6}$$
The use of the truncated normal distribution in the reliability tests of technical objects has the following practical sense. The equation for the density function of the normal distribution applies for all t values, from �∞ to þ∞. In the operational reliability tests of cars and their elements, there is always the relationship that t > 0, for which the density function is given by
<sup>2</sup><sup>π</sup> <sup>p</sup> � exp � ð Þ <sup>t</sup> � <sup>m</sup>
2
2
Probabilistic Methods for Damage Assessment in Aviation Technology
http://dx.doi.org/10.5772/intechopen.72317
<sup>σ</sup> <sup>¼</sup> <sup>u</sup> (13)
dt ¼ σdu (14)
� �du (15)
� � (16)
2σ<sup>2</sup> " # (11)
43
dt (12)
2σ<sup>2</sup> " #
exp � ð Þ <sup>t</sup> � <sup>m</sup>
f tðÞ¼ <sup>1</sup>
however, R(t) reliability function is provided by the following formula:
R tðÞ¼ <sup>1</sup>
The solution of the above integral includes the expression:
3.3. Elements poorly correlated with functioning time
damage threatening the safety or the most common ones.
aσ ffiffiffiffiffiffi <sup>2</sup><sup>π</sup> <sup>p</sup>
R tðÞ¼ <sup>ð</sup> ∞
substituting these figures, it is possible to obtain:
t
aσ ffiffiffiffiffiffi
f tð Þdt <sup>¼</sup> <sup>1</sup>
aσ ffiffiffiffiffiffi <sup>2</sup><sup>π</sup> <sup>p</sup> ð ∞
t � m
ð ∞
exp � <sup>1</sup> 2 u2
mþσu
R tðÞ¼ <sup>F</sup>0ð Þ <sup>u</sup> F<sup>0</sup> <sup>m</sup> σ
The second group includes elements and structures operating in the variable conditions that are subject to the material fatigue, vibration, corrosion, etc. The process of damage to the other group's elements can be described by the models with variable parameters and high dispersion, such as: gamma, log-normal, Weibull and others [2]. The selection of operating programmes is very difficult, especially in cases of aviation technology, where the failure of a function of the object's construction system threatens the safety of people, the environment or causes significant material losses. In this case, it is important to apply the density of services, matching them to the
t
the above formula
Because
where
and the function
$$\frac{1}{\sqrt{2\pi}}\int\_0^{u+\alpha u} \exp\left(-\frac{u^2}{2}\right) du = \Phi(u) \tag{7}$$
are called the Laplace function (integral), the final form of the equation of the reliability function will be as follows:
$$R(\mu) = 0.5 - \Phi(\mu) \quad \text{and} \quad F(\mu) = 0.5 + \Phi(\mu) \tag{8}$$
The above presented formulas for normal distribution of life time of the aircraft elements provide the right accuracy of calculations at a high degree of homogeneity of a feature and tested objects, which is characterised by small deviation values and the standard one, that is, when the expected value E(T) = m, where m >> σ, m > ð Þ 2÷3 σ is practically accepted.
In these cases, for which <sup>m</sup> <sup>σ</sup> < 2, it is recommended to use truncated normal distribution with the parameters of m, σ, for which the probability density function is as follows:
$$f(t) = \frac{1}{a\sigma\sqrt{2\pi}} \cdot \exp\left[-\frac{\left(t - m\right)^2}{2\sigma^2}\right] \tag{9}$$
where m means the average life time of the object to damage, t means a current variable, <sup>σ</sup><sup>2</sup> <sup>¼</sup> <sup>D</sup><sup>2</sup> ð Þ T means a variance, while σ > 0 and t > 0, and a constant is determined on the basis of the following formula:
$$a = \frac{1}{F\_0 \frac{m}{\sigma}} \tag{10}$$
The use of the truncated normal distribution in the reliability tests of technical objects has the following practical sense. The equation for the density function of the normal distribution applies for all t values, from �∞ to þ∞. In the operational reliability tests of cars and their elements, there is always the relationship that t > 0, for which the density function is given by the above formula
$$f(t) = \frac{1}{a\sigma\sqrt{2\pi}} \cdot \exp\left[-\frac{(t-m)^2}{2\sigma^2}\right] \tag{11}$$
however, R(t) reliability function is provided by the following formula:
$$R(t) = \bigcap\_{t}^{\alpha} f(t)dt = \frac{1}{a\sigma\sqrt{2\pi}} \left[ \exp\left[ -\frac{(t-m)^2}{2\sigma^2} \right] dt \tag{12}$$
Because
1 ffiffiffiffiffiffi <sup>2</sup><sup>π</sup> <sup>p</sup> �
1 ffiffiffiffiffiffi <sup>2</sup><sup>π</sup> <sup>p</sup> mþ ðσu
0
and the function
42 System of System Failures
function will be as follows:
In these cases, for which <sup>m</sup>
of the following formula:
<sup>σ</sup><sup>2</sup> <sup>¼</sup> <sup>D</sup><sup>2</sup>
ð 0
Figure 2. The probability density of normal distribution for different values of v coefficient of variation.
exp � <sup>u</sup><sup>2</sup> 2 � �
exp � <sup>u</sup><sup>2</sup> 2 � �
are called the Laplace function (integral), the final form of the equation of the reliability
The above presented formulas for normal distribution of life time of the aircraft elements provide the right accuracy of calculations at a high degree of homogeneity of a feature and tested objects, which is characterised by small deviation values and the standard one, that is,
<sup>2</sup><sup>π</sup> <sup>p</sup> � exp � ð Þ <sup>t</sup> � <sup>m</sup>
ð Þ T means a variance, while σ > 0 and t > 0, and a constant is determined on the basis
where m means the average life time of the object to damage, t means a current variable,
<sup>a</sup> <sup>¼</sup> <sup>1</sup> F<sup>0</sup> <sup>m</sup> σ
when the expected value E(T) = m, where m >> σ, m > ð Þ 2÷3 σ is practically accepted.
the parameters of m, σ, for which the probability density function is as follows:
aσ ffiffiffiffiffiffi
f tðÞ¼ <sup>1</sup>
R uð Þ¼ 0:5 � Φð Þ u and F uð Þ¼ 0:5 þ Φð Þ u (8)
<sup>σ</sup> < 2, it is recommended to use truncated normal distribution with
2
2σ<sup>2</sup> " #
du ¼ 0:5 (6)
du ¼ Φð Þ u (7)
(9)
(10)
�∞
$$\frac{t-m}{\sigma} = \mu \tag{13}$$
where
$$dt = \sigma du\tag{14}$$
substituting these figures, it is possible to obtain:
$$R(t) = \frac{1}{a\sigma\sqrt{2\pi}} \int\_{m+\sigma u}^{\infty} \exp\left(-\frac{1}{2}u^2\right) du\tag{15}$$
The solution of the above integral includes the expression:
$$R(t) = \frac{F\_0(u)}{F\_0\left(\frac{m}{\sigma}\right)}\tag{16}$$
#### 3.3. Elements poorly correlated with functioning time
The second group includes elements and structures operating in the variable conditions that are subject to the material fatigue, vibration, corrosion, etc. The process of damage to the other group's elements can be described by the models with variable parameters and high dispersion, such as: gamma, log-normal, Weibull and others [2]. The selection of operating programmes is very difficult, especially in cases of aviation technology, where the failure of a function of the object's construction system threatens the safety of people, the environment or causes significant material losses. In this case, it is important to apply the density of services, matching them to the damage threatening the safety or the most common ones.
With the development of the construction, it is important to mount the diagnosing systems for tracing (monitoring) the technical condition and signalling the pre-failure states in the units and functional systems. A certain way out of the situation involves monitoring of the course of induced forces with the use of a system of recorders adapted to record all relevant operational events, especially those threatening the safety of use. With the diagnosing and IT system for monitoring the state and the process of damage, it is possible to determine the area, in which the technical condition trajectory is placed, or to identify the durability resource.
#### 3.3.1. Gamma distribution
In this distribution, it is assumed that for randomly selected moments of t time in the object, the energy with the same value of individually operating induced forces (external loads) is cumulated, and that after putting k number of induced forces, the object is damaged.
The density function of this probability is as follows:
$$f(t) = \begin{cases} \frac{1}{\Gamma(k)}^{\lambda^k t^{k-1} e^{-\lambda t}} for & t \ge 0\\ 0 & for \quad t < 0 \end{cases} \tag{17}$$
p tðÞ¼ p ¼ λΔt þ 0Δt (21)
Probabilistic Methods for Damage Assessment in Aviation Technology
http://dx.doi.org/10.5772/intechopen.72317
45
N tð Þ ≥ SG (22)
<sup>y</sup> (23)
dt (24)
does not depend on the number of such increases, which occurred in time preceding t moment. In other words, the condition 'without consequences'that is significant for the simple Poisson's stream of damage is assumed. The above assumptions remain valid also for the normal distribution.
In case of the assumption that SG is a maximum permissible level of cumulation of n(t) stimuli, which result in ageing of a tested element of the aircraft and that for the number of stimuli
this object becomes unfit for further correct operation in the system, k number of induced forces, the cumulated energy of which is necessary for causing its damage, is calculated from
<sup>k</sup> <sup>¼</sup> SG
where y means the value, by which the ageing takes place (e.g., wear) of the element in a stepped manner under the impact of a single stimulus. However, λ magnitude is characterised
<sup>y</sup> <sup>¼</sup> dEf g <sup>η</sup>ð Þ<sup>t</sup>
By using the formula for the function of F(t) cumulated density of damage and R(t) = 1 � F(t) relationship, the element's reliability function for the Erlang distribution will be expressed by
> ð Þ <sup>λ</sup> � <sup>t</sup> <sup>n</sup> <sup>n</sup>! <sup>e</sup>
�λ<sup>t</sup> <sup>¼</sup> <sup>e</sup> �λt X k�1
n¼0
ð Þ <sup>λ</sup> � <sup>t</sup> <sup>n</sup>
<sup>n</sup>! (25)
<sup>λ</sup> <sup>¼</sup> <sup>1</sup>
k�1
n¼0
R tðÞ¼ <sup>1</sup> � F tðÞ¼ <sup>X</sup>
the following relationship:
the following formula:
by the average intensity of the aircraft ageing:
Figure 3. Gamma distribution density with different values k and λ.
where
k—number of events enforcing the ageing process, the cumulated effects of which cause the occurrence of damage in the object,
Γð Þk —gamma function is determined by the following formula:
$$
\Gamma(k) = \int\_0^\infty x^{k-1} e^{-x} dx \tag{18}
$$
For total k, there is the relationship:
$$
\Gamma(k) = (k-1)! \tag{19}
$$
and the gamma distribution is called the Erlang distribution.
In this case, F(t) distribution function has the following form:
$$F(t) = 1 - \sum\_{n=0}^{k-1} \frac{(\lambda \cdot t)^n}{n!} e^{-\lambda t} \tag{20}$$
In Figure 3, the gamma distribution density for various values k and λ was presented. At the same time, it is characteristic that the individual induced force (load) action results in the aircraft ageing (or increase of the energy cumulated in it in a stepped manner). The individual increase of effects of such an induced force has a constant value. Furthermore, the probability of the occurrence of the aircraft ageing increases in the time interval (t, t + Δt):
Figure 3. Gamma distribution density with different values k and λ.
With the development of the construction, it is important to mount the diagnosing systems for tracing (monitoring) the technical condition and signalling the pre-failure states in the units and functional systems. A certain way out of the situation involves monitoring of the course of induced forces with the use of a system of recorders adapted to record all relevant operational events, especially those threatening the safety of use. With the diagnosing and IT system for monitoring the state and the process of damage, it is possible to determine the area, in which
In this distribution, it is assumed that for randomly selected moments of t time in the object, the energy with the same value of individually operating induced forces (external loads) is
cumulated, and that after putting k number of induced forces, the object is damaged.
1 Γð Þk 0
Γð Þ¼ k
F tðÞ¼ <sup>1</sup> �<sup>X</sup>
of the occurrence of the aircraft ageing increases in the time interval (t, t + Δt):
k�1
ð Þ <sup>λ</sup> � <sup>t</sup> <sup>n</sup> <sup>n</sup>! <sup>e</sup>
n¼0
In Figure 3, the gamma distribution density for various values k and λ was presented. At the same time, it is characteristic that the individual induced force (load) action results in the aircraft ageing (or increase of the energy cumulated in it in a stepped manner). The individual increase of effects of such an induced force has a constant value. Furthermore, the probability
8 < :
λk t <sup>k</sup>�1e�λ<sup>t</sup> for for
k—number of events enforcing the ageing process, the cumulated effects of which cause the
ð ∞
xk�<sup>1</sup> e �x
0
t ≥ 0 t < 0
dx (18)
�λ<sup>t</sup> (20)
Γð Þ¼ k ð Þ k � 1 ! (19)
(17)
The density function of this probability is as follows:
f tðÞ¼
Γð Þk —gamma function is determined by the following formula:
and the gamma distribution is called the Erlang distribution. In this case, F(t) distribution function has the following form:
the technical condition trajectory is placed, or to identify the durability resource.
3.3.1. Gamma distribution
44 System of System Failures
occurrence of damage in the object,
For total k, there is the relationship:
where
$$p(t) = p = \lambda \Delta t + 0 \Delta t \tag{21}$$
does not depend on the number of such increases, which occurred in time preceding t moment. In other words, the condition 'without consequences'that is significant for the simple Poisson's stream of damage is assumed. The above assumptions remain valid also for the normal distribution.
In case of the assumption that SG is a maximum permissible level of cumulation of n(t) stimuli, which result in ageing of a tested element of the aircraft and that for the number of stimuli
$$N(t) \ge S\_G \tag{22}$$
this object becomes unfit for further correct operation in the system, k number of induced forces, the cumulated energy of which is necessary for causing its damage, is calculated from the following relationship:
$$k = \frac{\mathbf{S}\_G}{y} \tag{23}$$
where y means the value, by which the ageing takes place (e.g., wear) of the element in a stepped manner under the impact of a single stimulus. However, λ magnitude is characterised by the average intensity of the aircraft ageing:
$$
\lambda = \frac{1}{y} = \frac{d\mathbb{E}\{\eta(t)\}}{dt} \tag{24}
$$
By using the formula for the function of F(t) cumulated density of damage and R(t) = 1 � F(t) relationship, the element's reliability function for the Erlang distribution will be expressed by the following formula:
$$R(t) = 1 - F(t) = \sum\_{n=0}^{k-1} \frac{(\lambda \cdot t)^n}{n!} e^{-\lambda t} = e^{-\lambda t} \sum\_{n=0}^{k-1} \frac{(\lambda \cdot t)^n}{n!} \tag{25}$$
The expected value E(T), D<sup>2</sup> (T) variance and v(T) coefficient of variation for this distribution is as follows:
$$E(T) = \frac{k}{\lambda};\ D^2(T) = \frac{k}{\lambda^2};\ \mathbf{v}(T) = \frac{D^2(T)}{E(T)} = \sqrt{\frac{1}{k}}\tag{26}$$
does not decrease with time of operation. The value of λ parameter affects the shape of the
When (Tk) time of proper operation to damage is treated discretely [3] (e.g., by K�1 number of activating the object without damage to the moment of K activation, during which the failure will occur). Then, the geometric distribution is used. The F(t) distribution function of (Tk) time
p—means the probability of damage to the unit at K activation, it can be also calculated from
K
F tðÞ¼ P Tf g¼ <sup>k</sup> <sup>≤</sup><sup>K</sup> <sup>1</sup> � ð Þ <sup>1</sup> � <sup>p</sup> <sup>k</sup>þ<sup>1</sup> (31)
Probabilistic Methods for Damage Assessment in Aviation Technology
http://dx.doi.org/10.5772/intechopen.72317
47
E Tð Þ (32)
E Tð Þ (33)
R tð Þþ F tðÞ¼ 1 (34)
E Tð Þ <sup>¼</sup> <sup>1</sup> � exp � <sup>K</sup>
exponential distribution density curve presented in Figure 4.
for proper operation is calculated in the following way:
the relationship providing the approximate values:
F tðÞ¼ P Tf g ≤K ¼ 1 � e
Therefore, for the purposes of operation, it is possible to use the following formula:
Figure 4. Example courses of the exponential distribution density for different values of the parameter λ.
R tðÞ¼ exp ½ �¼ �λ<sup>t</sup> exp � <sup>t</sup>
In relation to the fact that F(t) distribution function is the complement to the reliability function
E Tð Þ—expected value of time of proper operation to damage.
When t < 0, the function is f(t) = 0,
where
where
unity:
#### 3.4. Elements without correlation with operating time
The elements of the third group are subject to the exponential reliability law, in which the constant intensity of damage is assumed. The damage have a random nature and most often come from:
The elements of the third group include bodies, glass housings and covers made of plastic, electronics components, etc.
#### 3.4.1. Exponential distribution
If (T) time of correct operation to damage is recorded in a continuous manner and the intensity of damage λð Þt is constant and does not depend on time in the entire interval (0, t), that is,
$$
\lambda(t) = \lambda = \text{const} \tag{27}
$$
the exponential distribution is used.
The F(t) distribution function of this distribution of (T) time of the correct operation in the interval (0,t) is calculated on the basis of the following relationship:
$$F(t) = P\{T \le t\} = 1 - e^{-\lambda t} = 1 - \exp\left[-\lambda t\right] \tag{28}$$
and f(t) function of distribution density for t > 0 is calculated on the basis of the relationship:
$$f(t) = \frac{dF(t)}{dt} = \lambda e^{-\lambda t} = \lambda \exp\left[-\lambda t\right] \tag{29}$$
where
λ > 0—means the distribution parameter (intensity of damage).
Moments of the exponential distribution are given by the following formula:
$$E(T) = \frac{1}{\lambda} \quad \text{and} \quad D^2(T) = \frac{1}{\lambda^2} \tag{30}$$
The equality E Tð Þ¼ <sup>1</sup> <sup>λ</sup> is true only for those elements of the device, for which the intensity of damage in the entire range of operation (0, t) is constant, and therefore, it does not increase or does not decrease with time of operation. The value of λ parameter affects the shape of the exponential distribution density curve presented in Figure 4.
When t < 0, the function is f(t) = 0,
When (Tk) time of proper operation to damage is treated discretely [3] (e.g., by K�1 number of activating the object without damage to the moment of K activation, during which the failure will occur). Then, the geometric distribution is used. The F(t) distribution function of (Tk) time for proper operation is calculated in the following way:
$$F(t) = P\{T\_k \le K\} = 1 - (1 - p)^{k+1} \tag{31}$$
where
The expected value E(T), D<sup>2</sup>
• overloads of a different nature;
electronics components, etc.
3.4.1. Exponential distribution
the exponential distribution is used.
E Tð Þ¼ <sup>k</sup>
3.4. Elements without correlation with operating time
• manufacture errors (material and technological errors);
<sup>λ</sup> ; D<sup>2</sup>
• non-compliance with the instructions for use or operation technology.
interval (0,t) is calculated on the basis of the following relationship:
F tðÞ¼ P Tf g ≤ t ¼ 1 � e
f tðÞ¼ dF tð Þ
Moments of the exponential distribution are given by the following formula:
E Tð Þ¼ <sup>1</sup>
λ > 0—means the distribution parameter (intensity of damage).
ð Þ¼ <sup>T</sup> <sup>k</sup>
The elements of the third group are subject to the exponential reliability law, in which the constant intensity of damage is assumed. The damage have a random nature and most often
The elements of the third group include bodies, glass housings and covers made of plastic,
If (T) time of correct operation to damage is recorded in a continuous manner and the intensity of damage λð Þt is constant and does not depend on time in the entire interval (0, t), that is,
The F(t) distribution function of this distribution of (T) time of the correct operation in the
and f(t) function of distribution density for t > 0 is calculated on the basis of the relationship:
<sup>λ</sup> and <sup>D</sup><sup>2</sup>
damage in the entire range of operation (0, t) is constant, and therefore, it does not increase or
ð Þ¼ <sup>T</sup> <sup>1</sup>
<sup>λ</sup> is true only for those elements of the device, for which the intensity of
dt <sup>¼</sup> <sup>λ</sup><sup>e</sup>
<sup>λ</sup><sup>2</sup> ; vTð Þ¼ <sup>D</sup><sup>2</sup>
as follows:
46 System of System Failures
come from:
where
The equality E Tð Þ¼ <sup>1</sup>
(T) variance and v(T) coefficient of variation for this distribution is
ð Þ T E Tð Þ <sup>¼</sup>
ffiffiffi 1 k r
λðÞ¼ t λ ¼ const (27)
�λ<sup>t</sup> <sup>¼</sup> <sup>1</sup> � exp ½ � �λ<sup>t</sup> (28)
�λ<sup>t</sup> <sup>¼</sup> <sup>λ</sup>exp ½ � �λ<sup>t</sup> (29)
<sup>λ</sup><sup>2</sup> (30)
(26)
p—means the probability of damage to the unit at K activation, it can be also calculated from the relationship providing the approximate values:
$$F(t) = P\{T \le K\} = 1 - e^{\frac{K}{E(T)}} = 1 - \exp\left[-\frac{K}{E(T)}\right] \tag{32}$$
where
E Tð Þ—expected value of time of proper operation to damage.
Therefore, for the purposes of operation, it is possible to use the following formula:
$$R(t) = \exp\left[-\lambda t\right] = \exp\left[-\frac{t}{E(T)}\right] \tag{33}$$
In relation to the fact that F(t) distribution function is the complement to the reliability function unity:
$$R(t) + F(t) = 1\tag{34}$$
Figure 4. Example courses of the exponential distribution density for different values of the parameter λ.
Figure 5. Reliability function courses; distribution: a—geometric, b—exponential.
then, R(t) for the geometric distribution is:
$$R(t) = 1 - \left[1 - \exp\left(-\frac{K}{E(T)}\right)\right] = \exp\left[-\frac{K}{E(T)}\right] \tag{35}$$
The following system of equations is right:
where
time interval of Δt;
following system of equations:
interval of Δt;
for n > 0
λ—proportionality factor that represents the damage risk;
P 0
P 0
P 0
For the system of equations (38), the initial conditions are as follows:
P0ð Þ¼ 0 1
By having the knowledge of P0ð Þt , then, P1ð Þt is determined, and so on.
P0ðÞ¼ t e
PnðÞ¼ t
�λ ðt
1 <sup>n</sup>! <sup>λ</sup> ðt
0
⋮
2 4
0
N tð Þdt
3 5
n
e �λ ðt
0
N Lð Þdt
The solution of the system of equations (39) takes the form of:
8
>>>>>>>>>>>>><
>>>>>>>>>>>>>:
P0ð Þ¼ t þ Δt P0ð Þt ð Þ 1 � λN tð ÞΔt
⋮
N tð Þ—number of operated aircraft, in which the considered damage may occur;
Δt—adopted time interval of aircraft operation (or the aircraft's flying time length).
<sup>0</sup> ¼ �λN tð ÞP0ð Þt
⋮
P1ð Þ¼ t þ Δt P1ð Þt ð1 � λN tð ÞΔtÞ þ P0ð Þt λN tð ÞΔt
Probabilistic Methods for Damage Assessment in Aviation Technology
http://dx.doi.org/10.5772/intechopen.72317
Pnð Þ¼ t þ Δt Pnð Þt ð1 � λN tð ÞΔtÞ þ Pn�<sup>1</sup>ð Þt λN tð ÞΔt
P0ð Þ t; t þ Δt —probability of non-occurrence of damage to basic and protection systems in the
Pið Þ t; t þ Δt , (i = 1,…n)—probability of the occurrence of 'i' number of damage in the time
By dividing Eq. (37) by Δt and going to the border with Δt ! 0, it is possible to obtain the
<sup>1</sup> ¼ �λN tð ÞP1ð Þþ t λN tð ÞP0ð Þt
<sup>n</sup> ¼ �λN tð ÞPnð Þþ t λN tð ÞPn�<sup>1</sup>ð Þt
Pnð Þ¼ 0 0 for n > 0
Equation (38) is a linear differential equation and it is solved recursively. First, P0ð Þt is found.
N tð Þdt
(
(37)
49
(38)
(39)
(40)
and for the exponential distribution, R(t) is:
$$R(t) = 1 - \left[1 - \exp\left(-\lambda t\right)\right] = \exp\left(-\lambda t\right) \tag{36}$$
In Figure 5, the reliability functions R tðÞ¼ P Tf g > t of the geometric distribution and the exponential distribution were presented. In the first case, the graph constitutes a step curve. However, the second graph constitutes a continuous curve. It was assumed that the parameters p and λ of both functions are the same, and their value is p = λ = 0.1.
Another way, which makes it possible to estimate the probability values of the occurrence of catastrophic damage in the aircraft devices, can include the use of models, including the limit state.
| doab | 2025-04-07T04:13:03.845605 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 24
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.25 | 4. Estimation of the average number of the aircraft failure within a given period
A quantitative description and probability evaluation of damage to the basic and protection systems of the aircraft can be carried out in accordance with the postulates of the Poisson process [4].
Assuming that:
The following system of equations is right:
$$\begin{aligned} P\_0(t + \Delta t) &= P\_0(t) \ (1 - \lambda N(t)\Delta t) \\ P\_1(t + \Delta t) &= P\_1(t) \ (1 - \lambda N(t)\Delta t) + P\_0(t)\lambda N(t)\Delta t \\ &\vdots \\ P\_n(t + \Delta t) &= P\_n(t) \ (1 - \lambda N(t)\Delta t) + P\_{n-1}(t)\lambda N(t)\Delta t \\ \text{for } n &> 0 \end{aligned} \tag{37}$$
where
then, R(t) for the geometric distribution is:
48 System of System Failures
and for the exponential distribution, R(t) is:
and the number of operated aircraft;
period
process [4].
Assuming that:
R tðÞ¼ <sup>1</sup> � <sup>1</sup> � exp � <sup>K</sup>
Figure 5. Reliability function courses; distribution: a—geometric, b—exponential.
ters p and λ of both functions are the same, and their value is p = λ = 0.1.
• proportionality factor identifying the risk of damage is constant;
E Tð Þ
In Figure 5, the reliability functions R tðÞ¼ P Tf g > t of the geometric distribution and the exponential distribution were presented. In the first case, the graph constitutes a step curve. However, the second graph constitutes a continuous curve. It was assumed that the parame-
Another way, which makes it possible to estimate the probability values of the occurrence of catastrophic damage in the aircraft devices, can include the use of models, including the limit state.
4. Estimation of the average number of the aircraft failure within a given
A quantitative description and probability evaluation of damage to the basic and protection systems of the aircraft can be carried out in accordance with the postulates of the Poisson
• probability of damage is directly proportional to the length of the concerned time period
<sup>¼</sup> exp � <sup>K</sup>
R tðÞ¼ <sup>1</sup> � <sup>1</sup> � exp ð Þ �λ<sup>t</sup> <sup>¼</sup> exp ð Þ �λ<sup>t</sup> (36)
E Tð Þ
(35)
P0ð Þ t; t þ Δt —probability of non-occurrence of damage to basic and protection systems in the time interval of Δt;
Pið Þ t; t þ Δt , (i = 1,…n)—probability of the occurrence of 'i' number of damage in the time interval of Δt;
N tð Þ—number of operated aircraft, in which the considered damage may occur;
λ—proportionality factor that represents the damage risk;
Δt—adopted time interval of aircraft operation (or the aircraft's flying time length).
By dividing Eq. (37) by Δt and going to the border with Δt ! 0, it is possible to obtain the following system of equations:
$$\begin{aligned} P\_0 &= -\lambda \mathcal{N}(t) P\_0(t) \\ P\_1' &= -\lambda \mathcal{N}(t) P\_1(t) + \lambda \mathcal{N}(t) P\_0(t) \\ &\vdots \\ P\_n' &= -\lambda \mathcal{N}(t) P\_n(t) + \lambda \mathcal{N}(t) P\_{n-1}(t) \end{aligned} \tag{38}$$
For the system of equations (38), the initial conditions are as follows:
$$\begin{cases} P\_0(0) = 1 \\ P\_n(0) = 0 \end{cases} \begin{cases} \\ \\ \qquad \text{for } n > 0 \end{cases} \tag{39}$$
Equation (38) is a linear differential equation and it is solved recursively. First, P0ð Þt is found. By having the knowledge of P0ð Þt , then, P1ð Þt is determined, and so on.
The solution of the system of equations (39) takes the form of:
$$\begin{cases} \begin{array}{c} \begin{array}{c} \begin{array}{c} t \\ \end{array} N(t)dt \end{array} \\ P\_0(t) = e \end{array} \\ \begin{array}{c} \begin{array}{c} \vdots \\ \end{array} \\ P\_n(t) = \frac{1}{n!} \left[ \begin{array}{c} t \\ \end{array} N(t)dt \right] \end{array} \end{cases} \tag{40}$$
The probability that n damage requiring the launch of protection systems will occur in the time interval (0,t) is described with the Poisson distribution, whereas the role of the expression 'λt' is replaced with the following magnitude λ Ðt 0 N tð Þdt due to the low frequency of the occurrence of this type of damage in the process of the aircraft operation.
The integral <sup>Ð</sup><sup>t</sup> 0 N tð Þdt can be replaced with the following total:
$$\int\_{0}^{t} N(t)dt \leftrightarrow \sum\_{i=1}^{N} t\_{i} \tag{41}$$
E n½ �¼ <sup>X</sup><sup>∞</sup>
ti—flying time within a given period of i aircraft
the observation, the following was obtained:
• in the interval (0, t1), n<sup>1</sup> of damage occurred; • in the interval (t1, t2), n<sup>2</sup> of damage occurred;
• in the interval (ti�1, ti), n<sup>i</sup> of damage occurred;
<sup>L</sup> <sup>¼</sup> ð Þ <sup>λ</sup>T<sup>1</sup> <sup>n</sup><sup>1</sup>
<sup>¼</sup> <sup>λ</sup><sup>n</sup>1þn2þ…þni
ni! <sup>e</sup>
equation in this manner, it is possible to find the relationship for λ.
maximum likelihood method is determined.
N—number of operated aircraft.
⋮
where
where
Hence
n¼1
We are often interested not only in the probability that n damage will occur for given flying time, but in the magnitude of λ coefficient characterising the intensity (risk) of the damage occurrence. In order to determine the estimator of λ parameter, a maximum likelihood method will be used. It should be supposed that we observed and recorded the occurrence of damage in several separate time intervals, when the aircraft's flying time was: t1, t2,…, ti. As a result of
The probability of the occurrence of the said number of damage, that is, n1+ n<sup>2</sup> +…+ n<sup>i</sup> during
ð Þ <sup>λ</sup>T<sup>2</sup> <sup>n</sup><sup>2</sup> <sup>n</sup>2! <sup>e</sup>
The above recorded probability, considered as a function of λ variable, at defined n1, n2, …, ni, T1,T2, …Ti is called the likelihood. Currently, such a value of λ, for which L likelihood adopts the greatest value, is found. For this purpose, relationship (47) is subjected to logarithms and a derivative in relation to λ, which is equated to zero, is calculated. By solving the obtained
> <sup>λ</sup><sup>b</sup> <sup>¼</sup> <sup>n</sup><sup>1</sup> <sup>þ</sup> <sup>n</sup><sup>2</sup> <sup>þ</sup> … <sup>þ</sup> ni T<sup>1</sup> þ T<sup>2</sup> þ … þ Ti
With the help of the relationship (49), the estimator of λ coefficient with the use of the
<sup>2</sup> … <sup>T</sup>ni i
�λT2<sup>⋯</sup> ð Þ <sup>λ</sup>Ti ni
�λð Þ T1þT2þ…þTi
ni! <sup>e</sup>
Ti ¼ ti � ti�<sup>1</sup> (48)
�λT<sup>1</sup>
(47)
(49)
operation with the intensity of their occurrence of λ is expressed by the relationship:
<sup>n</sup>1!n2! …ni! <sup>e</sup>
�λT<sup>1</sup> �
T<sup>n</sup><sup>1</sup> <sup>1</sup> <sup>T</sup><sup>n</sup><sup>2</sup> nPnðÞ¼ <sup>t</sup> <sup>λ</sup><sup>b</sup> <sup>X</sup>
N
Probabilistic Methods for Damage Assessment in Aviation Technology
ti (46)
http://dx.doi.org/10.5772/intechopen.72317
51
i¼1
where
N—number of aircraft operated within the considered time;
ti—flying time of the aircraft within the considered time.
For a single aircraft, the probability of the damage occurrence during the considered t flying time will be:
$$q\_1 = 1 - e^{-\lambda t} \tag{42}$$
where
q <sup>1</sup>—probability of damage in one aircraft;
t—aircraft's flying time.
Since λ risk of damage to both systems (basic and protection) that causes the failure is low, the expression e�λ<sup>t</sup> can be expanded into a power series.
Hence
$$e^{-\lambda t} \cong \ 1 - \lambda t \tag{43}$$
By substituting Eq. (43) to (42), the following is obtained:
$$\mathcal{V}\_2 \cong \mathcal{U} \tag{44}$$
With the relationship (44), it is possible to estimate the probability of the failure occurrence in a single aircraft.
The probability of correct aircraft functioning is expressed by the following relationship:
$$P\_1 = 1 - \lambda t\tag{45}$$
In order to estimate the average number of failures during a given period for the operated aircraft park, the following relationship can be used:
Probabilistic Methods for Damage Assessment in Aviation Technology http://dx.doi.org/10.5772/intechopen.72317 51
$$E[n] = \sum\_{n=1}^{\infty} nP\_n(t) = \hat{\lambda} \sum\_{i=1}^{N} t\_i \tag{46}$$
where
The probability that n damage requiring the launch of protection systems will occur in the time interval (0,t) is described with the Poisson distribution, whereas the role of the expression 'λt'
N tð Þdt \$ <sup>X</sup>
For a single aircraft, the probability of the damage occurrence during the considered t flying
q<sup>1</sup> ¼ 1 � e
Since λ risk of damage to both systems (basic and protection) that causes the failure is low, the
e
q
The probability of correct aircraft functioning is expressed by the following relationship:
With the relationship (44), it is possible to estimate the probability of the failure occurrence in a
In order to estimate the average number of failures during a given period for the operated
N
i¼1
N tð Þdt due to the low frequency of the occurrence
ti (41)
�λ<sup>t</sup> (42)
�λ<sup>t</sup> ffi <sup>1</sup> � <sup>λ</sup><sup>t</sup> (43)
<sup>2</sup> ffi lt (44)
P<sup>1</sup> ¼ 1 � λbt (45)
Ðt 0
is replaced with the following magnitude λ
The integral <sup>Ð</sup><sup>t</sup>
50 System of System Failures
where
time will be:
t—aircraft's flying time.
where q
Hence
single aircraft.
0
of this type of damage in the process of the aircraft operation.
N—number of aircraft operated within the considered time;
ti—flying time of the aircraft within the considered time.
<sup>1</sup>—probability of damage in one aircraft;
expression e�λ<sup>t</sup> can be expanded into a power series.
By substituting Eq. (43) to (42), the following is obtained:
aircraft park, the following relationship can be used:
N tð Þdt can be replaced with the following total:
ðt
0
ti—flying time within a given period of i aircraft
N—number of operated aircraft.
⋮
We are often interested not only in the probability that n damage will occur for given flying time, but in the magnitude of λ coefficient characterising the intensity (risk) of the damage occurrence. In order to determine the estimator of λ parameter, a maximum likelihood method will be used. It should be supposed that we observed and recorded the occurrence of damage in several separate time intervals, when the aircraft's flying time was: t1, t2,…, ti. As a result of the observation, the following was obtained:
The probability of the occurrence of the said number of damage, that is, n1+ n<sup>2</sup> +…+ n<sup>i</sup> during operation with the intensity of their occurrence of λ is expressed by the relationship:
$$\begin{array}{rcl} L & = & \frac{\left(\lambda T\_1\right)^{n\_1}}{n\_i!} \quad e^{-\lambda T\_1} \cdot \frac{\left(\lambda T\_2\right)^{n\_2}}{n\_2!} \quad e^{-\lambda T\_2} \cdots & \frac{\left(\lambda T\_i\right)^{n\_i}}{n\_i!} \quad e^{-\lambda T\_1} \\ & = & \frac{\lambda^{n\_1 + n\_2 + \ldots + n\_i} T\_1^{n\_1} T\_2^{n\_2} \cdots}{n\_1! n\_2! \ldots n\_i!} \quad e^{-\lambda \left(T\_1 + T\_2 + \ldots + T\_i\right)} \end{array} \tag{47}$$
where
$$T\_i = t\_i - t\_{i-1} \tag{48}$$
The above recorded probability, considered as a function of λ variable, at defined n1, n2, …, ni, T1,T2, …Ti is called the likelihood. Currently, such a value of λ, for which L likelihood adopts the greatest value, is found. For this purpose, relationship (47) is subjected to logarithms and a derivative in relation to λ, which is equated to zero, is calculated. By solving the obtained equation in this manner, it is possible to find the relationship for λ.
Hence
$$\widehat{\lambda} = \frac{n\_1 + n\_2 + \dots + n\_i}{T\_1 + T\_2 + \dots + T\_i} \tag{49}$$
With the help of the relationship (49), the estimator of λ coefficient with the use of the maximum likelihood method is determined.
Hence, relationship (47) takes the following form:
$$
\widehat{q} = \widehat{\lambda}t\tag{50}
$$
Eq. (51) in the function notation adopts the following form:
u zð Þ ; t —density function of the diagnostic parameter z at the time of t.
u zð Þ¼ ; t
solving this equation, the following density function is obtained:
b—average increase in the diagnostic parameter per time unit;
zd—diagnostic parameter value specifying the limit state.
determined after transformation of relationship (54) as follows [5]:
be presented in the following way:
a—average increase square of the diagnostic parameter per time unit;
Q tð Þ¼ ; zd
After taking into account the physics of the diagnostic parameter increase and appropriate transformation, the Fokker-Planck differential equation is obtained from Eq. (52). As a result of
> 1 ffiffiffiffiffiffiffiffiffi <sup>2</sup>πat <sup>p</sup> <sup>e</sup>
The probability of the catastrophic damage occurrence with the use of the relationship (53) can
1 ffiffiffiffiffiffiffiffiffi <sup>2</sup>πat <sup>p</sup> <sup>e</sup>
ð ∞
zd
The risk level of the catastrophic damage occurrence in the operating time function can be
ðt
0 f tð Þzd
1 ffiffiffiffiffiffiffiffiffi <sup>2</sup>πat <sup>p</sup> <sup>e</sup>
� <sup>z</sup>ð Þ <sup>d</sup>�bt <sup>2</sup>
Q tð Þzd ¼
f tð Þzd <sup>¼</sup> zd <sup>þ</sup> bt 2t
6. Assessment of a chance of the catastrophic damage occurrence with the constant level along the increasing diagnostic parameter value
In point 5, a case of the device operation, when the catastrophic damage occurred only after exceeding the limit state by the diagnostic parameter value, was considered. Currently, the next case is considered, when the opportunity of additional one (the second type of catastrophic
�ð Þ <sup>z</sup>�bt <sup>2</sup>
�ð Þ <sup>z</sup>�bt <sup>2</sup>
<sup>2</sup>at (53)
http://dx.doi.org/10.5772/intechopen.72317
53
<sup>2</sup>at dz (54)
dt (55)
<sup>2</sup>at (56)
where
where
where
where
u zð Þ¼ ; t þ Δt ð Þ 1 � λΔt u zð Þþ ; t λΔtu zð Þ � Δz; t (52)
Probabilistic Methods for Damage Assessment in Aviation Technology
where
t—aircraft's flying time within the year.
Relationship (50) makes it possible to estimate the probability of the damage occurrence in a single aircraft within a given time interval.
| doab | 2025-04-07T04:13:03.848031 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 25
} |
ffc5706f-49f6-414c-bb4e-8db2086cf9ea.26 | 5. Catastrophic damage model of the device including the limit state
These models can be used for determination of the probability of the occurrence of various negative events in the devices for the following cases:
It is assumed that:
May Uz,t mean the probability that in the moment of t, the diagnostic parameter value will be equal to z. For example, it can be assumed that z may mean, for example, the crack length or the surface wear value.
In order to describe an increase in the parameter value in the random basis, the following differential equation was adopted:
$$
\mathcal{U}\_{z,t+\Delta t} = (1 - \lambda \Delta t)\mathcal{U}\_{z,t} + \lambda \Delta t \mathcal{U}\_{z-\Delta z,t} \tag{51}
$$
where
Δz—increase in the diagnostic parameter value during one flight of the aircraft;
λΔt—probability of the aircraft flight in the time interval of Δt, whereas λΔt ≤ 1;
λ—intensity of the aircraft flights.
Eq. (51) in the function notation adopts the following form:
$$u(z, t + \Delta t) = (1 - \lambda \Delta t)u(z, t) + \lambda \Delta t u(z - \Delta z, t) \tag{52}$$
where
Hence, relationship (47) takes the following form:
t—aircraft's flying time within the year.
single aircraft within a given time interval.
negative events in the devices for the following cases:
parameter, which evaluates its state;
current value is determined by z.
• The parameter z is non-decreasing.
differential equation was adopted:
λ—intensity of the aircraft flights.
variables. It is assumed that:
the surface wear value.
where
in the parameter, which evaluates its state;
where
52 System of System Failures
<sup>b</sup><sup>q</sup> <sup>¼</sup> <sup>λ</sup>b<sup>t</sup> (50)
Relationship (50) makes it possible to estimate the probability of the damage occurrence in a
These models can be used for determination of the probability of the occurrence of various
• when a chance of the catastrophic damage occurrence is constant along the increasing
• when a chance of the catastrophic damage occurrence increases together with an increase
• when the parameters determining a chance of the damage occurrence constitute random
• The device's technical condition is determined by one dominant diagnostic parameter. Its
May Uz,t mean the probability that in the moment of t, the diagnostic parameter value will be equal to z. For example, it can be assumed that z may mean, for example, the crack length or
In order to describe an increase in the parameter value in the random basis, the following
Uz,tþΔ<sup>t</sup> ¼ ð Þ 1 � λΔt Uz,t þ λΔtUz�Δz,t (51)
• A change in the diagnostic parameter value occurs only during the aircraft flight:
Δz—increase in the diagnostic parameter value during one flight of the aircraft; λΔt—probability of the aircraft flight in the time interval of Δt, whereas λΔt ≤ 1;
5. Catastrophic damage model of the device including the limit state
• when the parameter, specifying their state, will exceed the limit state;
u zð Þ ; t —density function of the diagnostic parameter z at the time of t.
After taking into account the physics of the diagnostic parameter increase and appropriate transformation, the Fokker-Planck differential equation is obtained from Eq. (52). As a result of solving this equation, the following density function is obtained:
$$
\mu(z,t) = \frac{1}{\sqrt{2\pi at}} e^{-\frac{(z-bt)^2}{2at}} \tag{53}
$$
where
b—average increase in the diagnostic parameter per time unit;
a—average increase square of the diagnostic parameter per time unit;
The probability of the catastrophic damage occurrence with the use of the relationship (53) can be presented in the following way:
$$Q(t, z\_d) = \bigcap\_{z\_d}^{\infty} \frac{1}{\sqrt{2\pi at}} e^{-\frac{(z-bt)^2}{2at}} dz \tag{54}$$
where
zd—diagnostic parameter value specifying the limit state.
The risk level of the catastrophic damage occurrence in the operating time function can be determined after transformation of relationship (54) as follows [5]:
$$\left(Q(t)\_{z\_d} = \int\_0^t f(t)\_{z\_d} dt\right) \tag{55}$$
where
$$(f(t)\_{z\_d} = \frac{z\_d + bt}{2t} \frac{1}{\sqrt{2\pi at}} e^{-\frac{(z\_d - bt)^2}{2at}} \tag{56}$$
| doab | 2025-04-07T04:13:03.848905 | 1-12-2023 18:25 | {
"license": "Creative Commons - Attribution - https://creativecommons.org/licenses/by/3.0/",
"book_id": "ffc5706f-49f6-414c-bb4e-8db2086cf9ea",
"url": "https://mts.intechopen.com/storage/books/6197/authors_book/authors_book.pdf",
"author": "",
"title": "System of System Failures",
"publisher": "IntechOpen",
"isbn": "9781789230475",
"section_idx": 26
} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.